Biology of Subterranean Fishes

Biology of
Subterranean Fishes
© 2010 by Science Publishers

Biology of
Subterranean Fishes
Editors
Eleonora Trajano
Departamento de Zoologia
Instituto de Biociências da USP
São Paulo – SP
Brazil
Maria Elina Bichuette

Departamento de Ecologia e Biologia Evolutiva
Universidade Federal de São Carlos
São Carlos – SP
Brazil
B.G. Kapoor
Formerly Professor of Zoology
Jodhpur University
India
6000 Broken Sound Parkway, NW

CRC Press Suite 300, Boca Raton, FL 33487
Taylor & Francis Group 270 Madison Avenue
an informa business New York, NY 10016
Science Publishers
2 Park Square, Milton Park
www.crcpress.com Abingdon, Oxon OX 14 4RN, UK
Enfield, New Hampshire

Published by Science Publishers, P.O. Box 699, Enfield, NH 03748, USA
An imprint of Edenbridge Ltd., British Channel Islands
E-mail: info@scipub.net
Website: www.scipub.net
Marketed and distributed by:

6000 Broken Sound Parkway, NW
CRC Press Suite 300, Boca Raton, FL 33487
Taylor & Francis Group 270 Madison Avenue
an informa business New York, NY 10016
2 Park Square, Milton Park
www.crcpress.com Abingdon, Oxon OX 14 4RN, UK
Copyright reserved © 2010
ISBN: 978-1-57808-670-2
Cover Illustrations
A small troglobitic catfish, Rhamdiopsis sp., from caves in Chapada Diamantina,
Northeastern Brazil’’ and “Poço Encantado Cave, habitat of Rhamdiopsis sp., in Bahia
State, Northeastern Brazil” Reproduced by permission of Adriano Gambarini.
Library of Congress Cataloging-in-Publication Data

Biology of subterranean fishes / editors, Eleonora Trajano, Maria Elina Bichuette,
B.G. Kapoor. -- 1st ed.
p. cm.
ISBN 978-1-57808-670-2 (hardcover)
1. Amblyopsidae. 2. Fishes--Ecophysiology. 3. Fishes--Adaptation. 4. Fishes--Habitat.
5. Deep sea fishes. I. Trajano, Eleonora. II. Bichuette, Maria Elina. III. Kapoor, B.G.
QL638.A366B56 2010
597.17584--dc22
2009051455
All rights reserved. No part of this publication may be reproduced,

stored in a retrieval system, or transmitted in any form or by any
means, electronic, mechanical, photocopying or otherwise, without
the prior permission of the publisher, in writing. The exception to
this is when a reasonable part of the text is quoted for purpose of
book review, abstracting etc.
This book is sold subject to the condition that it shall not, by way
of trade or otherwise be lent, re-sold, hired out, or otherwise circulated
without the publisher's prior consent in any form of binding or cover
other than that in which it is published and without a similar condition
including this condition being imposed on the subsequent purchaser.
Printed in the United States of America
This book is dedicated to Dr. Thomas L. Poulson, in appreciation for his pioneering
and inspiring work on cave fishes and other subterranean creatures
Preface
Subterranean fishes are among the most instigating, scientifically relevant

and yet fragile organisms. Their specialization to the selective regime
typical of subterranean habitats, highly contrasting to epigean (surface)
ones, results in unique combinations of features, of which the most
conspicuous are the reduction of eyes and dark pigmentation. Envisaged
as bizarre creatures, cave fishes have been known and are the object
of interest for centuries, but studies intensified after the 1920’s, with
the advent of modern genetics and, more recently, molecular biology,
phylogeny and conservation.
Due to that selective regime, with permanent darkness as a major
directive force throughout the hypogean realm, diverging rates of
populations genetically isolated from their epigean colonizers are
generally high. At the same time, homoplastic evolution of cave-related
characters is quite common, producing sets of convergent character states
in more or less distantly related taxa. High divergence rates in relation
to close epigean relatives, homoplastic evolution of several characters,
and the simplified nature of subterranean ecosystems, with fewer trophic
levels, fewer species interacting and a tendency toward environmental
stability (i.e. fewer varying abiotic factors), render subterranean organisms
excellent subjects for evolutionary and ecological studies. Models of
regressive evolution of characters, a rather frequent phenomenon among
epigean animals, as Darwin himself was aware of, may be tested using
troglobitic (exclusively subterranean) species. Likewise, physiological
and behavioral models, in areas such as chronobiology and stoichiometry,
benefit from the study of hypogean animals.
However, these characteristics, which we find so interesting, may
impose serious problems for survival under conditions of unnatural
environmental stress. Small population sizes due to geographic restriction
and adaptations to cope with food scarcity, including K-selected life cycle
traits and low population densities, result in low population turnovers
and, consequently, a reduced capacity to recover from population losses.

vi Preface
Moreover, many troglobites, having evolved under relatively constant

abiotic conditions, lost their ability to cope with environmental fluctuations,
such as those provoked by human disturbance. A robust knowledge of the
ecology and biology of these animals is a necessary and most urgent step
for efficient actions that target their protection.
Knowledge about subterranean fishes progresses at a steady pace, with
novelties arising every day. This book begins with an overview of the main
evolutionary and ecological concepts applied to the study of subterranean
fishes, in the context of a retrospective and prospective of the field (Chapter
1). In the following chapters, we present updated approaches, integrating
new and previously known data, on general topics such as evolutionary
genetics (Chapter 5) and development (Chapter 6), behavior (Chapter
7), biodiversity (Chapter 2), and conservation (Chapter 3). A worldwide
checklist with 164 troglomorphic species represents a significant increase
in relation to the most recent one, also by Proudlove (104 species described
up to 2003) -- an increase that is mainly due to additions to the subterranean
ichthyofauna of China and Brazil. Regional faunas, in totum or focusing on
selected taxa, of countries such as China (Chapter 11), India (Chapter 12),
Brazil (Chapter 9), Mexico (Poecilia mexicana; Chapter 8), and the African
continent (Chapter 10), are discussed in terms of biodiversity and its causes,
natural history and distribution, and ecology. A comprehensive chapter
on the Amblyopsidae, a small North American family with one epigean,
one troglophilic and four troglobitic species, summarizes information on
almost every aspect of the systematics and biology of this family, which
are currently the best known cave fishes (Chapter 7).
The contributors to this book are from leading speleobiologists
dedicated to the study of subterranean fishes around the world. The
editors of this project acknowledge the prompt response and enthusiasm
of these scientists, who were essential to its success as a new and vibrant
publication showing the utmost progress in the field. We also thank the
specialists who reviewed the texts, in particular Drs. David C. Culver and
John Holsinger.
This book is dedicated to Dr. Thomas L. Poulson, who earned our
deepest admiration and respect for his pioneering and inspiring work on
amblyopsid fishes and other subterranean animals, always keeping a fresh
enthusiasm for these wonderful creatures. Thank you, Tom, for being such
an example of a scientist and a person.
Eleonora Trajano and Maria Elina Bichuette

Foreword
Tom Poulson and Speleobiology—A Reflection
David C. Culver
Department of Environmental Science
American University
4400 Massachusetts Avenue NW
Washington, DC 20016, USA
E-mail: dculver@american.edu
I have been asked by the editors to summarize Tom Poulson’s contributions

to speleobiology and I am honored to do so. I mentioned this honor to
a student who replied, “Oh yes, he is the one who works with Kathy
Lavoie on the beetle-cricket interaction.”1 I was reminded of the often
quoted comment about the 1960’s iconic rock and roll band, The Beatles.
Oh yes, someone from the next generation said, “Isn’t that the band Paul
McCartney was in before Wings?” I realize that the anecdote dates me as
well, but the point is that I want to place Poulson’s contributions, which
have spanned generations, in a larger context. Tom himself, both in this
volume and elsewhere (Poulson 2001a) has already commented on his
contributions and their relevance to the research questions facing hypogean
fish researchers. I certainly don’t intend to comment on his comments, but
1
For the record, Kathy Lavoie is an accomplished cave microbiologist and a former graduate
student of Poulson’s.
Photograph of Tom Poulson reproduced by Courtesy of Dante Fenolio

viii Foreword
rather to focus on what I think is the real reason we celebrate Poulson’s

contributions. Simply put, he put not only the study of subterranean fishes,
but speleobiology in general, in a neo-Darwinian context by his seminal
work on adaptation in the amblyopsid cavefishes.
In most habitats, adaptations is the single most obvious aspects of an
organism’s phenotype. Cheetahs are adapted for running and capturing
prey; monarch butterflies are brightly coloured to warn potential predators
of their toxicity; etc. The theme of most nature films is the adaptation
of organisms to their environment, and by implication the triumph of
evolution by natural selection. Subterranean animals are different in
this respect. The most obvious feature of many subterranean animals
are losses, not adaptations. Even Darwin saw subterranean animals as
degenerates: examples of eyelessness and loss of structure in general. For
him, the explanation was a straightforward Lamarckian one, and one that
did not involve adaptation and the struggle of existence.
“It is well known that several animals which inhabit the caves of
Carniola [Slovenia] and Kentucky, are blind….As it is difficult to
imagine that eyes, though useless, could be in any way injurious to
animals living in darkness, their loss may be attributed to disuse”
(Darwin 1859)2.
Small wonder then that for decades following Darwin, adaptation was
not associated with subterranean organisms. Much confusion followed,
which I will only briefly consider.
At the end of the 19th century, one of the leaders of the neo-Lamarckian
school of evolution was Packard, who was also the leading American
speleobiologist of his time. He was convinced that use and disuse governed
the evolution of subterranean animals and gave virtually no role to natural
selection and adaptation. He also held that evolution of what we would
now call troglomorphy was rapid. Packard knew cave animals well. He
had visited several dozen caves in North America and described many
species (Packard 1888).
In 1907 Banta supported what seems to a modern biologist a bizarre
theory—that of orthogenesis:
“Animals do not possess degenerate eyes and lack pigment because
they are cave animals….They are cave animals because their eyes are
degenerate and because they lack pigment….They are isolated in caves
and other subterranean abodes because they are unfit for terranean
life…” (Banta 1907).
In other words, animals are not blind because they are in caves, they are
in caves because they are blind! The major proponent of orthogenesis in
2
This quotation remained the same through all six editions of the book On the Origin of
Species.

Foreword ix
subterranean species was Vandel (1964). Vandel also minimized the role of
natural selection and adaptation with the following analogy, which links
the idea of aging individuals to senescent phyletic lines:
“The idea of adaptation has grown to the point where it has been
written that depigmentation and anophthalmy represent ‘adaptations
to subterranean life’. This is like saying that cararrh [common colds],
rheumatism, and presbyopia [far-sightedness] are adaptations to old
age”.
The final approach to the evolution of the morphology of cave animals
that is not selectionist came from the Kosswigs. Working in the 1930’s
they were very interested in genetic polymorphism and believed it held
the key to understanding regressive evolution. Based on their studies of
the highly polymorphic isopod Asellus aquaticus in the Postojna Planina
Cave System in Slovenia (Kosswig and Kosswig 1940), they believed that
mutation was the key to understanding this variability, and they held that
the presence of highly polymorphic populations of stygobionts was the
result of accumulating mutations that were not subject to selection.
In the late 1950’s, when Poulson began his graduate studies on
amblyopsid fish, it was hard to find anyone who studied cave animals
that looked to natural selection as an explanation of morphology, life
history, or behavior. In fact, it was hard to find anyone who looked at
any aspect of cavefish biology other than eye and pigment degeneration.
The integration of Darwinian ideas of evolution by natural selection with
population genetics (neo-Darwinism) was very much at the forefront in
biology at that time (see Gould 2002 for an extended discussion of the neo-
Darwinian synthesis and what he calls its hardening). Cave life was being
left behind.
What Poulson did, and it truly transformed the study of subterranean
animals, was to focus on adaptations, not on losses of eyes and pigment.
Using the comparative method, he analyzed all the known species of
Amblyopsidae (Poulson, 1963). The five3 species ranged in habitat from
freshwater swamps to streams deep in caves. This work resulted in
what is probably the most quoted single research paper on cave biology
ever4. In the paper, Poulson did a thorough comparative study of the
morphology, physiology, and behavior of the five species of fish and
demonstrated that the differences could only be explained by differences
in the selective environment. In particular, Poulson showed that the
food-poor, aphotic environment of caves imposed a selective regime that
3
Since his initial work, a sixth species has been discovered—Spleoplatyrhinus poulsoni.
4
According the Web of Science©, his paper has been cited 106 times since 1983 (as far back as
digital records go). I can find only one research paper (some reviews are cited more) that is
cited even 50 times—Wilkens (1971), which was cited 71 times.

x Foreword
resulted in an increase in extra-optic sensory structures, reduced metabolic

rate, increased longevity, larger eggs, and a host of other features. The
key result of this work (also summarized in Poulson and White 1969) is
that cavefish are adapted to their environment, just as the cheetah is to its
environment. I believe this is why his work is so often cited. Now, it is a
matter of course to read in taxonomic descriptions of cave animals from
all over the world and in a variety of languages that the species under
consideration is adapted to the subterranean environment because of a set
of elaborated or elongated morphological features. Poulson made the idea
of adaptation to the cave environment respectable.
Of course, he said more than this in his 1963 paper, and has substantially
expanded this work since that time (Poulson 1985, 2001b). I will consider
some of these expansions and elaborations below, but first I want to
complete the context of his work in the 1960’s. Two other North American
speleobiologists working in the 1960’s added to and complemented the
neo-Darwinian view of cave animals. Kenneth Christiansen (1961, 1965)
also used a comparative approach to study the morphology and behavior
of cave Collembola, wingless hexapods. He also gave a convincing
demonstration of adaptation in caves. The primary difference (except, of
course, for the choice of organisms) was that Christiansen emphasized the
problem Collembola faced in moving across wet mud in water in caves
rather than on food scarcity or darkness. Thomas Barr (1967, 1968), coming
from a more taxonomic tradition, developed the dominant paradigm for
speciation in caves.
There was more to Poulson’s seminal paper. In particular, he argued
that the degree of eye and pigment loss was a measure of the length of
time the different amblyopsid species have been isolated in caves. It may
seem quaint to try to measure time of isolation using morphological
rather than molecular measures, but of course no molecular techniques
for sequencing or measuring genetic distance were available at the time.
It is an indication of how much he changed the agenda of studies of cave
organisms that his use of regressive features was only in this context.
He then argued, with considerable data to back it up, that the degree of
morphological, physiological, and behavioral change exhibited by the
cave amblyopsid fish was correlated with the amount of eye and pigment
reduction, and that the greater the change, the greater the adaptation to
cave life. The interesting question of how long cave animals have been
isolated in caves continues to resonate and continues to be controversial
(e.g. Trontelj 2007). The argument that more troglomorphic species are
more highly adapted is likewise controversial because different species
may be adapted to different subterranean environments (e.g., Noltie and
Wicks 2001). Whatever the final answers to these questions, it was Poulson
who set the agenda and first asked them.

Foreword xi
I would be remiss if I left the impression that all or even most of

Poulson’s contributions to speleobiology have involved fish. For more
than 30 years, he has championed the idea that the amount, nature, and
distribution of food in caves were key elements in organizing terrestrial
cave communities and in determining the selective environment. This is
not the appropriate place to review this work, but I want to note that he
trained and was assisted by a remarkably talented group of Ph.D. students,
including the late Thomas C. Kane (Kane and Poulson 1976), Kathleen
Lavoie (Poulson and Lavoie 2000), David Griffith (Griffith and Poulson
1993), and Kurt Helf (Poulson, Lavoie, and Helf 1995). I was also honored
to be his graduate student, and we did joint work on the determinants of
species of richness in caves (Poulson and Culver 1969) and metabolic rates
in cave amphipods (Culver and Poulson 1971).
My own experiences with him in the field remain vivid and important,
now nearly forty years since most of them transpired. I have never
encountered a more careful observer of cave life—he was and is a great
naturalist. I have also never encountered a better critic of bad ideas and a
more enthusiastic supporter of new ideas and hypotheses. But more than
anything it is his enthusiasm for studying cave life that dominates time in
the field with Tom.
His influence on graduate students extended far beyond his own and far
beyond his published papers. Several generations of American graduate
students had the experience of being in the field with Poulson, or hearing
one of his inimitable seminars. And their work was better as a result. Mine
certainly was.
Literature Cited
Banta, A.M. 1907. The fauna of Mayfield’s Cave. Carnegie Institution of Washington
Publications 67: 1-114.
Barr, T.C. 1967. Ecological studies in the Mammoth Cave system of Kentucky. I.
The biota. International Journal of Speleology 3: 147-204.
Barr, T.C. 1968. Cave ecology and the evolution of troglobites. Evolutionary Biology
2: 35-102.
Christiansen, K.A. 1961. Convergence and parallelism in cave Entomobryinae.
Evolution 15: 288-301.
Christiansen, K.A. 1965. Behavior and form in the evolution of cave Collembola.
Culver, D.C. and T.L. Poulson. 1971. Oxygen consumption and activity in closely
related amphipod populations from cave and surface habitats. American Midland
Naturalist 85: 74-84.
Darwin, C. 1859. On the Origin of Species by Means of Natural Selection, or the
Preservation of Favoured Races in the Struggle for Life, John Murray, London, first
edition.

xii Foreword
Gould, S.J. 2002. The Structure of Evolutionary Theory. Harvard University Press,
Cambridge.
Griffith, D.M. and T.L. Poulson. 1993. Mechanisms and consequences of intraspecific
competition in a carabid cave beetle. Ecology 74: 1373-1383.
Kane, T.C. and T.L. Poulson. 1976. Foraging by cave beetles: spatial and temporal
heterogeneity of prey. Ecology 57: 793-800.
Kosswig, C. and L. Kosswig. 1940. Die Variabilität bei Asellus aquaticus unter
besonderer Berucksichtigung der Variabilität in isolierten unter- und
oberirdischen Populationen. Revue de Facultie des Sciences (Istanbul), ser. B, 5:
1-55.
Noltie, D.B. and C.M. Wicks. 2001. How hydrogeology has shaped the ecology of
Missouri’s Ozark cavefish, Amblyopsis rosae, and southern cavefish, Typhlichthys
subterraneus: insights of the sightless from understanding the underground.
Environmental Biology of Fishes 62: 171-194.
Packard, A.S. 1888. The cave fauna of North America, with remarks on the anatomy
of brain and the origin of the blind species. Memoirs of the National Academy of
Sciences (USA) 4: 1-156.
Poulson, T.L. 1963. Cave adaptation in amblyopsid fishes. American Midland
Poulson, T.L. 1985. Evolutionary reduction by neutral mutations: plausibility
arguments and data from amblyopsid fishes and linyphiid spiders. Bulletin of
the National Speleological Society 47: 109-117.
Poulson, T.L. 2001a. Adaptations of cave fishes with some comparisons to deep-
sea fishes. Environmental Biology of Fishes 62: 345-364.
Poulson, T.L. 2001b. Morphological and physiological correlates of evolutionary
reduction of metabolic rate among amblyopsid fishes. Environmental Biology of
Fishes 62: 239-249.
Poulson, T.L. and D.C. Culver. 1969. Diversity in terrestrial cave communities.
Ecology 50: 153-158.
Poulson, T.L. and W.B. White. 1969. The cave environment. Science 165: 971-981.
Poulson, T.L. and K.H. Lavoie. 2000. The trophic basis of subsurface ecosystems.
In: Subterranean Ecosystems. H. Wilkens, D.C. Culver and W.F. Humphreys (eds).
Elsevier Press, Amsterdam, The Netherlands. pp. 231-250.
Poulson, T.L., K.H. Lavoie and K. Helf. 1995. Long term effects of weather on the
cricket guano community in Mammoth Cave National Park. American Midland
Trontelj, P. 2007. The age of subterranean crayfish species. A comment on Buhay
and Crandall (2005): subterranean phylogeography of freshwater crayfishes
shows extensive gene flow and surprisingly large population sizes. Molecular
Ecology 16: 2841-2843.
Vandel, A. 1964. Biospéologie: la Biologie des Animaux Cavernicoles. Gauthier-Villars,
Paris.
Wilkens, H. 1971. Genetic interpretation of regressive evolutionary processes:
studies on hybrid eyes of two Astyanax cave populations (Characidae, Pisces).

Contents
Preface v
Foreword – Tom Poulson and Speleobiology—A Reflection
David C. Culver vii
1. Cavefish: Retrospective and Prospective 1

Thomas L. Poulson
2. Biodiversity and Distribution of the Subterranean
Fishes of the World 41
Graham S. Proudlove
3. Conservation of Subterranean Fishes 65
Maria Elina Bichuette and Eleonora Trajano
4. Behavioral Patterns in Subterranean Fishes 81
Jakob Parzefall and Eleonora Trajano
5. The Evolutionary Genetics of Cave Fishes:
Convergence, Adaptation and Pleiotropy 115
Richard Borowsky
6. Development as an Evolutionary Process in
Astyanax Cavefish 141
William R. Jeffery and Allen G. Strickler
7. Subterranean Fishes of North America: Amblyopsidae 169
Matthew L. Niemiller and Thomas L. Poulson
8. Subterranean Fishes of Mexico (Poecilia mexicana,
Poeciliidae) 281
Martin Plath and Michael Tobler
9. Subterranean Fishes of Brazil 331
Eleonora Trajano and Maria Elina Bichuette
10. Subterranean Fishes of Africa 357
Roberto Berti and Giuseppe Messana

xiv Contents
11. Subterranean Fishes of China 397

Chen Zi-Ming, Luo Jing, Xiao Heng and Yang Jun-Xing
12. Subterranean Fishes of India 415
Atanu Kumar Pati and Arti Parganiha
Subject Index 441
Index of Scientific Names 451
Author Index 455
About the Editors 457
Color Plate Section 459

CHAPTER 1
Cavefish: Retrospective and
Prospective
Thomas L. Poulson
Professor Emeritus, Department of Biological Sciences, University of
Illinois-Chicago. E-mail: tomandliz@bellsouth.net
I. INTRODUCTION
A. My Aims
One of my aims is to take a retrospective and prospective view that will
help colleagues identify areas of cavefish research that need increased
attention. In this context I urge everyone to think about how to solve
‘mysteries’ in their research area and to explain any new insights (eurekas)
they have had. My second aim is to urge everyone to place their studies
in the context of ecological and evolutionary concepts so that others will
use cavefish as examples. In this chapter I identify what I consider to be
exemplary examples from both my own work and that of others. I pay
attention to how new techniques are giving better answers to old questions.
And I consider what questions can and cannot be answered with different
study systems, especially Astyanax species vs the Amblyopsidae. I urge
you to do the same in your chapters and future work. In all cases try
to update what you have published in readily accessible books. And of
course build on past work of others especially in The Biology of Hypogean
Fishes (Romero 2001) and Subterranean Fishes of the World (Proudlove
2006) but also build on a number of fine chapters in Encyclopedia of Caves
(Culver and White 2005), Subterranean Ecosystems (Wilkens, Culver, and
Humphreys 2000), and the Special Issue of The National Speleological
Society Bulletin on Regressive Evolution (Culver 1985).

2 Biology of Subterranean Fishes
B. Ecological Concepts
I like the following definition of ecology from Charles Krebs ecology
text, with key words in capitals. ECOLOGY is the SCIENTIFIC STUDY of
INTERACTIONS that determine the DISTRIBUTION and ABUNDANCE
of organisms. By the process of ‘SCIENTIFIC STUDY’ I mean starting
with a question or patterns, like distribution and abundance, and then
using prior knowledge and observations to formulate a hypothesis that
has simple enough predictions to be potentially falsifiable. The next step
is using natural or manipulative experiments to test the predictions from
each hypothesis. I strongly believe in formulating multiple hypotheses to
avoid the trap of a favorite hypothesis that can at best lead to inadvertent
bias and controversy and at worst lead to advertent fraud. For natural
experiments I follow Jared Diamond in using both ‘snapshot’ and
‘trajectory’ experiments. ‘INTERACTIONS’ are potential agents of natural
selection. They include all combinations of abiotic-abiotic interactions
(e.g. temperature and oxygen concentration in water), abiotic-biotic
interactions (e.g. flooding and injury to fish), and biotic-biotic interactions
(e.g. competition, predation, parasitism, mutualism, and commensalism).
‘DISTRIBUTION’ includes ‘address’, i.e. habitat, and geographic location
at all spatial scales. This includes the spatial scale of biogeographic patterns
that are generated at evolutionary time scales. And ‘ABUNDANCE’
includes frequency across potentially suitable habitats and both density
and population size in each habitat occupied. It also includes niche.
Niche is a way of life or ‘profession’ (e.g. top predator in a food web vs
omnivore).
C. Tradeoffs
At the interface of ecology and evolution I have always found it useful
to consider tradeoffs and how they relate to rigor, variability, and
predictability of agents of natural selection. A common misconception
is that plus, i.e. advantageous, tradeoffs occur in different species or
circumstances than minus, i.e. disadvantageous, tradeoffs. In fact each
trait or circumstance has both plus ands minus tradeoffs. For example
floods often are associated with transport of allochthonous organic matter
into caves (+) but the rigor of turbulence and velocity of flood-water may
injure cavefish (–). And a generalist, in habitat or feeding niche, is a ‘jack-
of-all-trades’ (+) but a ‘master’ of none (–). The opposite tradeoffs occur
for a specialist. Low variability in abiotic conditions or highly predictable
timing of cycles of variability may favor specialization. But high variability
with low predictability may favor generalization.

Thomas L. Poulson 3
D. Temporal and Spatial Scale

Whether experiments, differences, or events are biologically significant
depends on relevant spatial and temporal scales. Unusual ecological events
may be unimportant on an evolutionary time scale. And experiments that
have statistically significant results may be unrealistic at the large spatial
scales of cave streams or the long generation times of troglobitic cavefish.
On the other hand predictability of rare bad or good events on a generation
time scale may not be statistically significant but could have huge and
important effects as agents of selection. Think of catastrophic capture of
a surface stream by a subterranean cave or a 50-year flood that renews
allochthonous organic matter to an otherwise low nutrient cave stream.
An important issue for temporal variability and predictability is the
scale of cavefish longevity, or better generation time, in relation to return
time for floods. With evolution of long cavefish lives, associated with
slow growth and great metabolic economies, the unpredictability of food
renewal with 10-year or 25-year floods can change from unpredictable
to predictable. Thus the chance of successful reproduction is increased
especially since long life is usually associated with spread of risk of
reproductive failure by repeated reproduction – iteroparity – with few
large eggs hatching to large young (quality rather than quantity). In short
lived species with high fecundity and small eggs there may be only one
reproduction – semelparity – and here spread of risk is among many small
young (quantity rather than quality).
An important spatial scale issue is the relation of manipulative
experiments in aquaria or artificial streams to natural experiments in
orders of magnitude larger cave streams and the suite of caves a species
occupies in a watershed. Interpretations of the importance of chemical
cues and territoriality are especially subject to using unrealistic spatial
scales.
E. Evolutionary Concepts
Evolution is the central paradigm of biology. As Theodosius Dobzhansky
wrote, “nothing makes sense except in the light of evolution.” Evolution
is change in inherited traits in a population over generations. The
appropriate time unit is the generation (e.g. bacteria evolve faster than
cavefish in the same environment). The unit of evolutionary change is
the population and the unit of natural selection is the individual. To my
knowledge nobody has considered whether or how there might be group
selection in any cavefish.

Implicitly Darwin used the scientific method to develop his hypothesis

of evolution by natural selection. He observed that organisms have many
more offspring than needed to replace themselves and so hypothesized
that there should be unlimited population growth. But tests of the
prediction falsified the hypothesis because he did not observe continued
exponential population growth of any species in nature. Instead he noted
that population numbers of a species vary over time but tend to remain
stable. So Darwin further hypothesized that there must be “a struggle
for existence”. And he suggested that those individuals that survive
competition, predation, and other exigencies are naturally selected. These
individuals will leave the most offspring. Based on the ‘strong principle
of inheritance’, that Darwin observed in domestic animals – offspring
resemble their parents – he further suggested that there will be evolution.
And so species may evolve new or improved adaptations as environments
change.
Over time Darwin’s hypothesis of evolution by natural selection has
changed from a hypothesis to a continually supported theory. First his
hypothesis was tested in myriad species in many ways and could not be
falsified. So his hypothesis became a theory (note the misunderstanding
by creationists that ‘evolution is only a theory’).
It does not matter, to the validity of the theory, that Darwin did not
know about genes or that mutation and recombination are sources of
genetic variation. He in fact thought that degeneration or rudimentation of
cavefish eyes was due to disuse. We now know that the real explanation of
cavefish eye and pigments rudimentation involves both neutral mutations
of loss and concurrent positive and negative pleiotropic effects of a few
genes with large developmental effects.
II. GENETIC AND DEVELOPMENTAL BASES FOR

RUDIMENTATION
A. Background
Our earliest model for a gene was simplistic and survives only as the reality
that there are some ‘Mendelian’ genes with recessive and/or dominant
alleles. We now know that most traits are coded by additive effects of
quantitative trait loci (QTLs) in a polygenic manner. We also know that
some QTL genes have very large effects by controlling development
often pleiotropically by affecting several different traits. And, as Stephen
J. Gould argued for years, changes in developmental control genes may
have and have had huge evolutionary effects. In fact, for a complex eye in
an amphipod a catalog of mutants has only one with slight increases, for

Thomas L. Poulson 5
amount of inter-ommatidial pigment, but dozens that are moderately to

strongly deleterious (Sexton and Clark 1936). To me that means not only
that accumulation of point mutations can help explain rudimentation
but also that mutations that can act to increase elaboration of a structure
are rare. Hence developmental control mutations that are polygenic and
recombination of such genes would seem to be the main way to elaborate
structures in general and in cavefish in particular. We will see that ongoing
studies in the Astyanax fasciatus system support this scenario.
B. Alternative Hypotheses for Rudimentation

What is/are the mechanism(s) for the evolutionary mechanism of eye
and pigment rudimentation seen convergently in all cavefish species? Is
rudimentation due to accumulation of neutral loss mutations in the absence
of stabilizing selection? And/or is it due to indirect selection against eyes
and pigment by antagonistic pleiotropy, a kind of genetic hitchhiking?
Rudimentation of eye and pigment systems is the pattern, evolution is the
process, and either drift or selection is the mechanism.
Why do many invertebrate troglobites completely lack eyes or melanin
pigment whereas this is never the case for vertebrates? I agree with many
workers who suggest that any mutations that would eliminate eyes and
pigment are selected against in vertebrates because of the close relation
among developmental pathways. Put another way stabilizing selection
maintains genes important for eye, pigment, and head morphogenesis in
vertebrates. I suggest that this is not the case for invertebrates because
eyes and pigment developmental pathways occur relatively late and so
mutations of loss are not selected against.
I prefer the words elaboration for evolution of non-visual sense organs
in cavefish and the terms reduction in size and rudimentation for changes
in eyes and pigment. These terms have less implied etymological baggage
than the terms constructive and degenerative or regressive. The terms –
elaboration and rudimentation – do not imply mechanism and so they
avoid the issues of whether there has been selection against or for traits
or accumulation of loss mutations with loss of stabilizing selection. In fact
evidence is strong that there are pleiotropic tradeoffs for elaboration and
reduction in an amphipod, Gammarus minus, and in a cavefish, Astyanax
fasciatus. However there is some disagreement about how pleiotropy is
working and its relative importance for pigment vs eye rudimentation at
early vs late stages of the evolution of troglomorphy.
Two recent reviews critically evaluate all mechanisms of rudimentation.
Both reinforce my plausibility arguments, based on relative sizes and
savings of parts and metabolic rates (Poulson 1985), that metabolic

savings is not the basis for antagonistic pleiotropy (Culver and Wilkens
2000, Jeffery 2005 and this volume). I find Protas et al. (2008) suggestion
that high cost of retinal metabolism may help account for eye reduction
unconvincing and unlikely based on the kinds of analyses I have made
earlier (Poulson 1985).
C. Ongoing Studies of Rudimentation and Elaboration

Jeffery (2005), Wilkens (2007), and Protas et al. (2008) propose three variants
of the pleiotropy mechanism at the developmental control gene level for
eye rudimentation in Astyanax fasciatus. In all cases eyes are indirectly
selected against by selection for elaboration of non-visual sense organs, a
kind of genetic hitchhiking.
All genetic/developmental mechanisms for rudimentation in Astyanax
fasciatus have to account for several patterns that have been repeatedly
documented. One pattern is that, compared to the surface ancestor, cavefish
eye primordia are small, that eyes tissues proliferate and differentiate
normally to a point, and that eyes then stop differentiation and become
highly variable in size and structure with continued body growth (e.g.
Figures 1 and 2 in Wilkens 2007). As Culver and Wilkens (2000) point
out this high variability is common to all vestigial structures in many
organisms. Another common pattern is that hybridization studies show
that polygenic inheritance is predominant especially with eye size and
pigment cell numbers. And hybridization between different blind cave
populations shows that there can be partial recovery of eye size.
Even the best worker’s research groups, each doing exciting and
important work, often fail to deal with all of the general patterns. In
Jeffery’s (2005) review I do not believe he addresses the ontogenetic
increase in variability of eye size and structure so emphasized by Wilkens.
And in Wilkens’ (2007) review I do not believe he addresses Jeffery’s results
that show that even late acting genes are still ‘functioning’ for eye and
pigment systems. Nor does Jeffery’s review address Wilkens' points about
mutations that change functionality of pigment genes. And Protas et al.’s
new hypothesis does not address either Wilkens' or Jeffery’s suggestions
about pleiotropy. Instead they present a new twist of ‘multi-trait evolution’
where both ‘genes’ that are associated with elaboration and rudimentation
are clustered but not so closely linked to preclude recombination. The
elaborated traits include metabolic economy, taste buds, and other traits
important for surviving in a food-poor cave environment. I suggest that
these gene clusters may be the animal analog of regulatory plant genes
with wide-ranging effects on growth and metabolic economies that code

Thomas L. Poulson 7
for what has been called the ‘stress-resistance syndrome’ genes in plants
(Chapin et al. 1993, West-Eberhard 2003).
Hybridization studies show that eye genes associated with
rudimentation are different in different blind cave populations of Astyanax
fasciatus. Wilkens (2007) showed early on that laboratory hybrids of
different blind Astyanax populations have larger eyes than either parent.
And Borowsky’s group (Borowsky 2008, Protas et al. 2008, and Borowsky,
this volume) have just shown that different eye rudimentation genes,
in populations independently isolated in caves, can complement each
other to restore vision. The degree of complementation, from 1-40% in
different hybrids, increases with geographic distance between parental
cave populations and thus with the increasing likelihood that related but
different quantitative genes have been involved. These results are as we
would predict for parallel evolution leading to convergent phenotypes via
different genes affecting the same developmental processes.
In Astyanax some of these control genes, especially ‘sonic hedgehog’,
have pleiotropic effects by both inhibiting eye development and
promoting development of taste buds (Jeffery 2005). Borowsky’s group
(Protas et al. 2008) have shown that the traits that are reduced and enhanced
by a control gene are clustered on Asytanax chromosomes but are not so
closely linked that recombination is inhibited. Jeffery (2005) reviews how
these control genes operate by using elegant transplant experiments of eye
anlage to one side of an early embryo, using the other side of the head of
the donor recipient as a control. Remarkably, transplants of early eye lens
anlage, or even lens messenger RNAs, can both restore eyes and vision of
blind forms, if the donor is eyed, and result in eye degeneration in eyed
forms, if the donor is blind.
Microarray and differential display measures of gene expression show
that as many genes are up-regulated as down-regulated (Jeffery 2005).
Generally, elaborated traits are affected positively and rudimenting traits
are affected negatively. Jeffery concludes that this is the mechanism for
antagonistic pleiotropy. Surprisingly all genes continue to be expressed,
as measured by gene arrays and messenger RNAs, even though their
down-regulation results in abnormal and small eyes. As with hens teeth
(“as rare as hens teeth”) I expect someone to eventually find an eyed fish
among a wild blind cavefish population when its suppressing control genes
fail to function. In addition using such atavistic mutants, researchers are
currently using genetic and developmental manipulation to resurrect teeth
in birds and show that genetic programs for other reptilian traits are still
present. Thus the studies resurrecting eyes in blind cavefish by Jeffery’s
group are just an example of an emerging generality i.e. that evolutionary
losses are often a matter of modulating and suppressing genetic circuitry

which is not lost. But so far in these studies workers have not been able
to resurrect teeth, tails, or scales that are completely normal. To me this
suggests that there has also been some loss of function in structural genes.
So too, I suspect, in cavefish.
I suggest that, if we could do developmental studies, in more
troglomorphic cavefish than Astyanax we would find that many of the
latest acting eye and pigment genes would not even be expressed. I
suggest this is the case for species that may even completely lose eyes and
pigmented melanophores as adults (Phreatichthys andruzzii and probably
Speoplatyrhinus poulsoni).
The early developmental constraints seem to me to be least for pigment
system genes and so I expected and documented that they apparently
show faster evolutionary rudimentation than for visual systems in the
Amblyopsidae (Poulson 1985). Jeffery agrees that pigment genes at the
end of a developmental cascade should be least subject to stabilizing
selection and so was surprised to find that they too are expressed early
on in Astyanax. But Wilkens (2007) documents three different kinds of
Mendelian pigment genes that have apparently ceased to function in some
Astyanax cavefish populations. Wilkens suggests that Jeffery’s results
may have been complicated because the Pachon population of Astyanax
studied has had both ancient and recent introgression with surface fish.
Thus Pachon fish may be less troglomorphic than other Astyanax cave
populations and certainly much less troglomorphic than many cavefish
species such as the cave Amblyopsids which have much more reduced
pigment systems.
A detailed summary of data on Amblyopsids (Poulson 1985, Table 2,
Niemiller and Poulson, this volume) show that visual systems and pigment
systems continue to become more rudimented even though advantageous
troglomorphic traits show much less further elaboration in the grades of
increasing troglomorphy seen among species.
My hypothesized increasing grades of troglomorphy in Amblyopsid
fishes are from Typhlichthys subterraneus to Amblyopsis spelaea to Amblyopsis
rosae to Speoplatyrhinus poulsoni. For this series there is continued and clear
increased rudimentation with smaller size of eyes and optic lobes, number
of pigmented melanophores and their sizes and melanin densities with an
increasing degree of negative allometry for each. But Figure 2c in Poulson
(1985) shows that, aside from neoteny (see the section below Is Neoteny
a Route to Troglomorphy?), there are relatively small and inconsistent
increases of adaptive elaborating traits, both for size and positive allometry,
among the three well-studied Amblyopsid troglobites. These traits
include head size, pectoral fin length, neuromast numbers, cerebellum
size, and forebrain size. Another adaptive trait that does not change is

Thomas L. Poulson 9
that egg volume is close to the same in all troglobites. After 23 years my
interpretation is still that there has been continued accumulation of loss/
simplifying mutations for visual and pigment systems in Amblyopsid
troglobites (Poulson 1985). My logic is that the early developmental
constraints and stabilizing selection are not so important when eyes and
pigment systems have already become so reduced evolutionarily.
A corollary is that continued rudimentation of eye and pigment
systems support my hypothesis that the four troglobitic Amblyopsids
indeed represent increasing evolutionary time in caves. My interpretation
for elaborated traits is that with strong selection at the time of isolation in
caves at least the sensory systems quickly got about as good as possible.
At present, I see no independent way of testing this hypothesis because,
even if molecular clocks can be shown to be accurate, we have no way of
dating the time when current troglobite species were isolated in caves.
D. Sources of Genetic Variation

The relative contributions of mutation, recombination, and hybridization
to genetic variation, the fuel for evolutionary change, have not been
adequately considered in studies of cavefish. Recent work, especially on
Galapagos finches by Peter and Rosemary Grant, has shown the importance
of rare hybridization as a source of enhanced variation that increases rates
of adaptation to environmental change. Thomas Barr has suggested to me
that some species of his trechine cave beetles do occasionally hybridize.
And Wilkens’ colleagues have shown, using mitochondrial DNA, that at
least one Astyanax cave population, Pachon, probably had both ancient and
recent introgression of genes from surface fish. However neither Barr nor
Wilkens speculate as to whether rare hybridization might have increased
the rate or degree of adaptation to life in the cave or slowed it.
E. Population and Quantitative Genetic Approaches

Astyanax researchers and others, would benefit by taking a population
genetic approach. Culver effectively uses theory to explore the actual
and potential effects of neutral and pleiotropic genes in Gammarus minus
(Amphipoda). See Tables 20.2 – 20.4 in Culver and Wilkens (2000). It would
also be of interest to calculate the plausibility of and potential rates of
evolutionary elaboration with Mendelian genes vs structural polygenes,
vs developmental control genes with large pleiotropic effects.

III. A HIERARCHY OF ADJUSTMENT TO STRESS

I find it useful to consider a temporal hierarchy of possible adjustment
modes to physical-chemical stress; first by behavior, then physiology, then
development, and finally genetic. In the hierarchy for adjustment the four
modes take increasing time and are decreasingly reversible from behavior
to physiology to development to genetic. There is some overlap in time
taken but in principle they are ‘employed’ in sequence if stress is not
alleviated. If behavioral change does not alleviate the stress over a period
of minutes to hours then physiological adjustments are made over periods
of days to seasons. And if physiological adjustment does not alleviate
stress then developmental change over a lifetime may work. If the stress
continues or is predictable and rigorous then genetic change may alleviate
the stress over generations i.e. evolutionary adaptation. Of course even
the three modes that operate within a lifetime have a genetic basis.
In most cases mobile animals like cavefish can alleviate stress using
just behavior and development. Sedentary plants and sedentary colonial
animals often have striking developmental adjustments that are adaptive.
A common example is sun leaves vs shade leaves in trees. These are of
course changes in morphology within the developmental phase of a
lifetime.
Looking at behavioral and physiological adjustments is especially
relevant to determining whether a species is preadapted to living in the
complete darkness of caves with attendant food restriction. It is useful to
study these adjustments even if we do not have the real epigean ancestor
of a cavefish species to study. For example Chologaster cornuta is the only
possible living model for a preadapted Amblyopsid ancestor. And the
right model for troglobitic Astyanax fasciatus is the surface stream Astyanax
fasciatus.
In the context of low food supply as a stressor, assessment of behavioral
and physiological adjustment are the first lines of defense. So I suggest
the following research agenda: (1) Start by putting the eyed form in the
dark and quantifying its food searching and acquisition behavior. (2) At
the physiological time scale, assess the resistance to starvation, by rate of
weight loss and ability to recover from different degrees of weight loss
(Huppop 2000). This may relate to whether the intestinal epithelium and
intestine length remain intact during starvation. And determine whether
resistance to starvation can be prolonged by providing detritus or even
cave mud. Finally, quantify the rate of and amount of fat deposition when
preferred food is available ad libitum. (3) At the developmental time scale,
determine whether allometric increase in sense organs with low food
supply and low growth rate lead to enhanced food finding abilities.

Thomas L. Poulson 11
IV. EVOLUTION OF TROGLOMORPHY IN CAVEFISH
A. Preadaptation: The First Step in Cavefish Evolution

Virtually all cavefish researchers agree that species that successfully made
the transition from the surface (epigean) to caves (hypogean) were in some
way preadapted. Preadaptations are not necessarily exaptations with new
functions in a new environment. A problem, as Culver and Poulson (1971)
first pointed out, is that we often do not know the relevant surface species
to use as a comparison.
If we do know a close surface relative or, in a few cases know the actual
surface ancestor, then too often we have not done the experiments to see
how much of the cave species’ behavior and food finding ability is just
due to the surface species compensating, within a few hours. weeks, or
months to being in darkness. We all know that, as humans that rely heavily
on vision and are not preadapted to life in darkness, we can get better
at navigating in the dark over time. See the cartoon of a cavefish using
a cane (Romero 2001 page 6). And we never have trouble making love
in complete darkness. Some humans that become blind at an early age
develop extraordinary abilities to navigate without sight by kinesthetic
sense, touch, and a crude echolocation. And some rare individuals can
even do extraordinary feats of balance on a tightrope. The Great Wallenda,
a world renown trapeze performer, could even walk blindfolded on a
swaying rope. So it should not be surprising that epigean fish species that
do not rely on sight and are nocturnal and have well developed non-visual
sensory systems, might do well in caves if there is sufficient food. This is
the case for most ancestors of troglobitic cavefish but not all.
B. Astyanax fasciatus vs the Amblyopsidae as Models

for Cavefish Evolution
I suggest that each model system has advantages and disadvantages. I
focus on these two systems because they are by far the most thoroughly
studied of all cavefish.
Astyanax fasciatus troglobites would seem to be of relatively recent
origin because there are no reproductive barriers (but see Porter
et al. (2007) suggesting as long as 1-2 million years of isolation for some
cave populations). And genetics and developmental analyses can be done
because they are easy to breed in the laboratory. But Astyanax fasciatus is
unusual for cavefish in several ways: the mode of cave isolation, heavy
reliance on vision in surface streams, and relatively modest troglomorphy.
Either they have not been isolated in caves long enough or have not had

enough food limitation to evolve the kinds of sensory elaboration and

metabolic efficiencies possible for cavefish.
As treated in detail by Niemiller and Poulson (this volume), the
Amblyopsid species represent the entire spectrum of epigean to hypogean
existence and extremes of troglomorphy. And the one troglophile species,
Forbesichthys agassizi, is a possible model for the what most researchers
agree is the more usual transition from epigean to hypogean life. Even the
strictly surface species, Chologaster cornuta, has obvious preadaptations for
life without light and Forbesichthys is somewhat more preadapted (Poulson
1961, 1963). All the Amblyopsid troglobites are very troglomorphic and
their surface ancestors are no longer extant and they are apparently ancient
based on molecular data (Niemiller, pers. comm.). So we do not know the
extent to which any of their surface ancestors were preadapted. The lack
of a surface ancestor also means that genetic studies of the evolution of
troglomorphy are impossible. Even genetic studies of any Amblyopsid
troglobite species are currently impossible because we do know how to
get any of the species to reproduce in the laboratory.
Astyanax fasciatus seems to be an exception to my hypothesis that
the habitat shift model does not apply to troglobitic fish. There are two
reasons. First, its isolation in caves is likely associated with catastrophic
loss of surface streams as they were ‘captured’ by subterranean drainage.
This is apparently an unusual mode of isolation and, because it is sudden,
it certainly resulted in extreme selection for traits that would allow
survival and reproduction in caves. Wilkens and Huppop (personal
communication) seem to agree by using the phrase “disruptive selection
at an abrupt epigean to hypogean border”. The second reason Astyanax is
unusual is that initial selection in caves was especially strong because it is
not ‘pre-adapted’ by having small eyes and being nocturnal.
If surface Astyanax fasciatus did not also have preadaptations of a lateral
line system, that acted in concert with vision to allow striking schooling
behavior and agonistic interactions, and taste buds, for discriminating
edible and inedible food, then I suggest that it would not have survived in
caves. When placed in the dark an eyed fish changes its body orientation
and searching behavior. This immediate behavioral change allows it to find
food but not quickly or efficiently compared to the cave form (Huppop
1985). Parzefall (2000) describes how the cave form, compared to the
surface form, searches with a more oblique body angle that brings its chin,
that has the densest taste buds, in closer contact with the substrate. In
the dark the eyed form can use its lateral line system to avoid obstacles
and perhaps sense living prey (Montgomery et al. 2001 and Coombs, pers.
comm.).
Physiological preadaptations of Astyanax are that it can put on some
fat and resist starvation for a while, though not as well as the blind cave

form (Huppop 2000). I suggest that the selection for increased short-
term resistance to starvation and fast fat deposition during good times
were selected by periodic but not completely predictable flood inputs of
allochthonous organic matter. The surface streams that were captured by
underground drainage still accumulate organic litter in the dry stream
bed and this litter will flush into the cave during especially heavy and
prolonged rain. In between times food supply would decline quickly. If
there had been slow and steady decline in food input over time after stream
capture then I expect that the fish would have evolved lower metabolic
rates and much more efficient food finding capacities. Even the egg size is
only marginally greater than in the epigean ancestor suggesting that even
post-yolk absorption hatchlings are not severely food limited.
An alternative hypothesis for the lack of lowered metabolic rate and for
small eggs is that cave Astyanax fasciatus populations have been isolated
too few generations to evolve the extreme food-finding and metabolic
economies seen in Amblyopsid troglobite species (but see Porter et al.
2007 suggesting > 1 million year lineage age for some cave populations).
Still another alternative is that the Astyanax caves, even those without bat
colonies, are not very food-limited when averaged over several years. And
still another hypothesis is that rare introgression with the surface form
slows the evolution of troglomorphy. But introgression is only apparent
for some cave populations: Pachon, Chica, and Micos.
An updated overview of the Amblyopsidae as a model system (from
Poulson 1963, 1985) including preadaptations in behavior, physiology, and
development is the chapter by Niemiller and Poulson, in this volume. This
chapter covers the efficiencies of search, detection, capture, and utilization
using some heretofore unpublished data. In summary, the troglophile
cannot survive long if in the food-poor cave environment. The springs
and spring runs where it goes from springs or caves to feed at night have
especially abundant food supplies. When there is enough food they have
no trouble making chance contact with live prey and grabbing them no
matter what part of their body is contacted.
C. Phenotypic vs Genotypic Troglomorphy

One persisting challenge is to separate phenotypic change during
development, for example due to allometric growth with different food
supplies, from genotypic differences selected over generations. And we
need to see if allometric increases in sense organs and brain parts during
the development of the troglophile, in a low food cave, translate into better
food-finding abilities.

As in Astyanax fasciatus cave forms, each Amblyopsid cavefish species

has opposite allometry of eye and pigment systems (–) vs sensory systems
(+). Eye and optic lobe size and pigment cell number show negative
allometric growth and head size and non-visual sensory systems, especially
neuromast number, show positive allometric growth. I spent considerable
effort measuring allometric growth in my dissertation because I had been
impressed by Heuts’ suggestion of its importance in understanding cave
adaptation in Caecobarbus geertsi (Heuts 1951). Subsequently I decided that
my effort had been largely a waste of time. But I am now newly excited
about my allometry results since I see how to relate them to the work
of Jeffery’s group on developmental control genes in Astyanax fasciatus.
Remember that Jeffery’s work was partly influenced by Stephen J. Gould’s
emphasis on the importance of developmental genes in major evolutionary
change. And the interface of developmental plasticity and evolution is of
great interest currently (West-Eberhard 2003).
To me, my dissertation data suggest that changes in positive and
negative allometry in Amblyopsids represents the result of antagonistic
pleiotropic effects. It could mean that the same processes that are being
discovered by genetic and developmental studies in Astyanax fasciatus
are reflected in allometric differences within and among species of
Amblyopsids. Data on allometry are important because in Amblyopsids,
and many other troglobitic fishes, we cannot do breeding experiments.
This means we cannot thoroughly determine the relative contributions of
developmental phenotypic flexibility and genetic determinism when we
see differences between populations of a species. I now see a possible way
around this conundrum.
I suggest that we could do ‘common garden’ experiments and measure
the degree of positive and negative allometry for traits. We would start
with the smallest individuals available and give them low and high food
rations; unfortunately this may be practicable only with the short-lived
troglophiles, like Forbesichthys agassizi among the Amblyopsids.
However I already have several natural experiments that show that
especially low or high food supply and growth rate have opposite effects
on rudimenting structures and elaborating structures. Rudimenting
structures are eyes, brain optic lobes, and pigment cells. And elaborating
structures are lateral line neuromasts and cerebellum. In my dissertation
(Poulson 1961, Figure 6) I compared epigean and cave populations of the
troglophile Forbesichthys agassizi in the Mammoth Cave area. A briefly
surviving population in the lowest food supply area of Mammoth Cave
had by far the lowest growth rate and smallest known eyes and number of
pigmented melanophores each with the lowest known negative allometric
coefficients for the species. And it had the most neuromasts and largest

cerbellum with the highest known positive allometric coefficients of all

populations.
Also in my dissertation I compared nine populations of Typhlichthys
subterraneus (Poulson 1961, Figure 7) including three with opposite extremes
of food supply. Typhlichthys in Blowing Cave (west central TN), with an
unusual amount of allochthonous organic matter, had the highest growth
rate with the largest optic lobes and most pigmented melanophores, with
correspondingly highest positive allometry. They also had small heads and
low numbers of neuromasts, with correspondingly low positive allometry.
In contrast Missouri populations of Typhlichthys (River and Welch’s Caves)
had among the lowest growth rates, had the smallest optic lobes with the
lowest positive allometry, and had either few pigmented melanophores
(Welch’s Cave) or the only population with no pigmented melanophores
even at the smallest body size (River Cave). These two cave populations
had the largest heads, longest pectoral fins, and most neuromasts of any
Typhlichthys populations.
How much of the Missouri Typhlichthys morphology is due to phenotypic
flexibility and how much of it is due to especially great evolutionary
troglomorphy in populations that have been independently isolated in
caves? Based on the analyses of cave and karst of the Missouri Salem
Plateau (Noltie and Wicks 2001), I predict that Missouri Typhlichthys caves
have a low food supply compared to Kentucky or Tennessee caves. This
could have selected for faster evolutionary increases in troglomorphy and
it could also have phenotypic effects on allometry during each generation
via low growth rates. As a test there is a natural experiment at The Gulf,
a deep karst window in Missouri with a very high food supply and a
large Typhlichthys population (Poulson 2001). It could show us whether all
Missouri Typhlichthys are especially troglomorphic and/r whether the low
food supply presumed from the geology is partly responsible. I predict
that the fast growth rates and small body size of The Gulf Typhlichthys
(Poulson, unpublished) will be associated with relatively smaller optic
lobes and eyes, and more neuromasts than for the Tennessee and Kentucky
Typhlichthys. But the Gulf population will also have fewer neuromasts and
larger eyes and optic lobe and more pigmented melanophores than any
other Missouri Typhlichthys. New data on melanophores (Niemiller and
Poulson, this volume) are consistent with this hypothesis.
D. Is Neoteny a Route to Troglomorphy?

Neoteny (aka paedomorphosis or paedogenesis) is the retention in adults
of traits seen in juveniles. Based on the grades in evolution of troglomorphy
I identified in my PhD dissertation (Poulson 1961) I predicted the next
grade in troglomorphy for the family if evolution were to continue. In

1967 I received a photograph in a plain brown envelope of a white cavefish

with an extremely large head relative to its body size (Figure 1A in Cooper
and Kuehne (1974), Color plate in Niemiller and Poulson, this volume )
with the cryptic question ‘Is this what you predicted’? Thinking that my
colleagues were putting me on, I went so far as to say it was a small ~ 10
mm Typhlichthys because of its relatively large head and un-branched fin
rays. If I had noted the oxygen bubbles as a clue to scale I would have
realized that it was a much larger fish. In fact it was the 58.3 mm holotype
of Speoplatyrhinus poulsoni named in my honor! In discussions with John
Cooper while examining the type series I suggested that it represented
an even more extreme neoteny than in any other the other Amblyopsid
troglobites and he and his coauthor agreed in their description of this new
species (Cooper and Kuehne 1974). In fact Table 2 in my 1985 paper, where
I discussed the basis for increasing troglomorphy, does not adequately
show how large the head is relative to the body or the strikingly low
weight to length ratio both of which, along with absence of branched fin
rays in the adult, are the strongest neotenic trends of any of the troglobitic
Amblyopsids (see drawings by J. E. Cooper in Niemiller and Poulson,
this volume). In addition there is preliminary evidence that both visual
and pigment systems are the most rudimentary of any of the troglobites.
Pigmented melanophores are very small, very sparse, and disappear by
adult size. And the sunken, spathulate snout and narrow head possibly
reflect absence of circumorbital bones and change in other related parts
of the skull. On this basis I predict that histological studies will show that
there is not even an eye vestige in the adult.
The drawings in Weber (2000) of convergently large spathlulate heads
of Speoplatyrhinus poulsoni, other highly troglomorphic cavefish, and even
the most troglomorphic salamander known (Eurycea rathbuni) all suggest
extreme neoteny to me. In addition to reflecting extreme eye rudimentation,
I have suggested (as early as in Mohr and Poulson 1967) that the large head
of E. rathbuni allows hypertrophied sensory systems, especially lateral line,
that are associated with enhanced detection of moving prey. When and if
we are ever able to determine the developmental – genetic basis for neoteny
in extreme troglobites we can test my prediction (Poulson 2001) that the
basis will be developmental genes with pleiotropic effects that work in an
analogous way to the developmental genetic changes responsible for the
‘stress-resistance syndrome’ in higher plants (Chapin 1991). And I now
suggest that the basis will be similar to the processes that Jeffery’s group
is so elegantly showing at the early stages of troglomorphic evolution in
Astyanax fasciatus (see earlier section Ongoing Studies of Rudimentation
and Elaboration).

V. ECOLOGY AND LIFE HISTORY

Generally slow growth, long life, late age at first reproduction, and
iteroparity with few large young at each reproduction are life history traits
associated with neoteny and troglomorphy, at least in the Amblyopsidae.
Ecology of other cavefishes is especially well-treated by Trajano (2001).
A. Life History Patterns

The + and – tradeoffs of both low r and high r life history extremes have
been reviewed earlier (Spatial and Temporal Scale) in the context of the
Tradeoffs section. I have recently (Poulson 2001) made a detailed critique
of the nature and quality of evidence, especially from analysis of scale
marks and recapturing marked fish, for differences in life history patterns
of deep sea fishes and epigean, troglophilic, and troglobitic Amblyopsids.
In brief the extreme patterns are as follows.
Fish with a high r life history (r in the logistic equation for population
growth) are characterized by fast growth, early reproduction, and short
life. The main mode of risk spreading of risk is among many ‘low quality’
young at a single reproduction (semelparity). This pattern occurs when
and where there are definite and predictable seasons. In Amblyopsids
this is the case for both the epigean species, Cholgaster cornuta, and the
troglophile Forbesichthys agassizi (see Niemiller and Poulson, this volume).
The + tradeoff of this life history is that a species can quickly increase
in numbers during good times. The – tradeoff is that populations will
decline rapidly in poor times. In words this has been called boom (+) and
bust (–).
Fish with a low r life history are characterized by slow growth, late age
at first reproduction, long life, and repeated reproduction (iteroparity).
The main mode of risk spreading, to ensure some successful reproduction,
is by repeated reproduction attempts each with few high quality young.
This pattern occurs when times of favorable food supply are rare or
unpredictable. The – tradeoff is that populations cannot respond quickly
to good times. The + tradeoff is that populations decline slowly during
bad times. For students, I have called this pattern ‘fizz (–) and fizzle (+)’.
It is seen in all troglobitic Amblyopsids but it is accentuated in Amblyopsis
spelaea (see Pearson, this volume) and probably in Speoplatyrhinus
poulsoni.

B. Abiotic Agents of Selection, Including Food Supply

I have rejected the so-called ‘constant cave’ paradigm that I supported
early in my studies of cave organisms (e.g. Poulson 1964, Poulson and
White 1969). Caves are clearly less variable physically and chemically than
overlying surface environments but I suggest that rare or unusual events can
have great effects, good and bad. Because water flows and mixes, aquatic
habitats are much more constant spatially than fully terrestrial habitats
in the same cave. But they are more variable temporally, also because of
flow. Thus allochthonous input of organic matter to aquatic habitats is
much more variable temporally than input to terrestrial habitats. During
spring floods, water temperature and chemistry can change rapidly. For
both terrestrial and aquatic cave habitats the rigor of low food supply
with darkness and lack of plant primary productivity seems to be almost
universal even though there is certainly temporal variability at a scale of
cavefish life-spans. Thus I was not surprised, in retrospect, to find that all
Amblyopsid species tested had the same very low change of metabolic
rates and spontaneous activity when acclimated to temperatures from 5 to
25°C (Poulson 1985 and Niemiller and Poulson, this volume).
It will be useful to infer how much the rigor of temperature change,
flooding, and food supply has varied over glaciation cycles, that encompass
evolutionary time for cavefish. And whether it has changed in the past
several centuries with forest clearing and agriculture and associated
increased of fine sediments in cave streams. In addition reservoirs have
flooded or partly flooded caves along some river valleys. Have these
anthropogenic impacts had negative effects on cavefish populations?
Though it is clear that caves are food-limited compared to surface
habitats, it is not clear to me that the degree of troglomorphy seen among
cavefishes is related to differences in food supply. For example, Trajano
(2001) suggests that troglobites from phreatic habitats are especially
troglomorphic and I agree that we would expect deep phreatic habitats to
have less food input than shallow caves with open channel streams. But
the deepest phreatic habitat known is the Edwards Aquifer in Texas and it
has perhaps the highest troglobite diversity including apparently among
the highest densities of both a predatory troglobitic salamander and fish
(reviewed by Trajano 2001). The Edwards Aquifer ecosystem may be
unusual in being based on bacteria and fungi that somehow derive their
energy in situ, maybe from hydrocarbons (Poulson and Lavoie 2000). And
among and within the putative Typhlichthys species and Astyanax cave
species, there is no clear relation of the degree of troglomorphy and the
presence or absence of allochthonous particulate plant or bat fecal organic
matter. Of course we do not know what the food supplies were at the

time of isolation in caves when selection for troglomorphy would have

been the most intense (Poulson 1985). However some have argued, myself
included (e.g. Poulson and Lavoie 2000), that caves in the wet tropics are
and were always less food limited than temperate caves.
Are caves in the wet tropics really less food limited than temperate caves
and if so are troglobitic tropical fish in these areas less troglomorphic than
temperate species? Almost all wet tropic caves have many bats of several
species year around. And with high rainfall and warm temperatures cave
formation and fragmentation are fast; this results in many entrances and
so many chances for abiotic input of allochthonous organic matter both
by gravity and flooding. At least in Brazil caves Trajano’s (2001) review of
cavefish ecology documents much organic matter, some bat populations,
regular sinking streams, and Siluriform (‘catfish’) troglobites of several
lineages that are not very troglomorphic. In addition surface sister species
exist for many of the troglobites. This last observation could be interpreted
to mean, as an alternative to low selection intensity with high food supplies,
that the troglobites are recent and have not had long enough to evolve
extreme troglomorphy. In the case of Amblyopsids in North America the
lineages of the troglobitic species are ancient and the surface ancestors are
no longer extant. And, in addition to apparently long evolutionary time in
caves the caves occupied by Amblyopsids overall have much lower food
supplies than in the Brazilian tropics.
C. Biotic Interactions as Agents of Selection
Background
Though I regard decreased interspecific competition in caves as an
implausible explanation for fish actively invading caves (section below
Mechanisms of Isolation in Caves), I do believe that competition,
predation, and parasitism are likely much less intense for cavefish than for
their surface ancestors. If true then there are predictions to be made about
the kinds and intensities of –/– interactions (competition), +/– interactions
(predation and parasitism), +/0 interactions (commensalism), and +/+
interactions (mutualisms). Culver’s elegant experimental studies and
theoretical analyses of aquatic stream communities in the Appalachians of
USA are testimony to the surprising diversity of both direct and indirect
biotic interactions including mutualism (summarized in Culver 1982).

Predation
Cavefish are either the top predator in caves or they have no predators
even if they are mainly detritivores. This has three consequences. First
they may evolve rudimentation of ‘fright’ or startle reactions as in the
Amblyopsidae (Poulson 1963). For the more troglobitic Amblyopsids
I can often pick them up by hand by moving slowly and carefully! It
would be interesting to see if they have lost chemical skin exudates with
injury i.e. ‘shreckstoffs” known from Ostariophycine schooling epigean
fish (Astyanax is an Ostriophyscine but its cave populations do not have
a functioning shreckstoff). But even if cavefish have water borne chemical
ways of distinguishing themselves from other species it does not mean
they are biologically useful at the scale of cave streams with huge dilution
and mixing. Chemical contact recognition is a more realistic possibility
(e.g. Parzefall 2000 for Poecilia).
A second consequence of being the top predator is that, because of
trophic transfer inefficiencies, food becomes increasingly limiting as one
moves up a food chain. It is known that cannibalism can sometimes be
important in Amblyopsis spelaea in population regulation (Poulson 1969)
and in Forbesichthys agassizi with virtual absence of alternative food in
caves (Hill 1966). Though many cavefish are predators it does not mean
that they might not get some nutrition by eating detritus or carrion when
starving. Whoever wrote ‘predator by choice and saprovore by necessity’
had a point.
A final consequence of being a top predator, especially with a long life, is
damage by toxins including heavy metals and pesticides. If even parts per
billion or trillion of mercury or DDT are present, they will biomagnify up a
food chain over a few days and very slowly bioaccumulate in individuals
over many years of life. It seems possible that ‘broken back syndrome’
discussed by Keith and Poulson (1981) for Ambylopsis spelaea was due to
pesticide runoff from agriculture in the cave watershed.
Parasitism and Disease

Possible reduction in parasitism and disease has not been adequately
studied and it has not been suggested, heretofore, as a possible advantage
to life in caves. What little is known for Amblyopsids is reviewed in
Niemiller and Poulson (this volume). For the troglophile Forbesichthys
agassizi, Whittaker and Hill (1968) have documented the high frequency
and high density of a cestode but there are not as extensive data on
parasites for troglobitic Amblyopsids (but see Whittaker and Zobel 1978).
I predict that parasitism will be found to be less in troglobites than in

troglophiles and less in troglophiles than in epigean species. Another side

of this coin is that immunological resistance to disease will have become
lower in troglobites.
I urge all cavefish researchers to start taking data on diversity, frequency
(parasitologists use the term prevalence) and abundance (parasitologists
use the term intensity) of parasitism and disease. Pearson (this volume)
has data showing a high prevalence of fin erosion in one Amblyopsis
spelaea population. The fin erosion got better in marked fish over a year
or two. It would be useful to know if captive cavefish are susceptible to
and able to immunologically respond to some of the common aquarium
fish fungal and bacterial diseases. Relating intensity of parasitism to an
immunological marker, like globulins, would be a useful non-destructive
way to indirectly assay prevalence and intensity of parasitism.
Competition
The presence or absence of competition has been too loosely considered
for cavefish because too few recognize the distinction between direct
and indirect competition and evidence needed to show the importance
of each (Griffith and Poulson 1993). Indirect competition is also known
as resource competition or exploitation competition and in fish operates
mainly through differential efficiency of locating food; it need not involve
even proximity of individuals to operate. Direct competition is also known
as interference competition and in fish operates mainly by defense of a
territory and maintaining exclusive access to local food resources; it must
involve contact between individuals. There are three kinds of evidence
needed to support a hypothesis that exploitation competition is ongoing:
1. Fish must reduce the abundance of a food resource; 2. Reduced food
abundance must reduce food-finding; and 3. Reduced food-finding must
result in reduced fecundity and/or survival. To demonstrate ongoing
interference competition one criterion must be met: either 1. Fish must
directly reduce each other’s survival e.g. by fighting, or 2. Fish must
indirectly reduce each other’s survival or fecundity by reducing each
other’s foraging effectiveness. If this reduction is dependent on food
density then it overlaps with criterion 3 for exploitation competition. Of
course absence of evidence for ongoing competition does not exclude the
importance of a ‘ghost of competition past’ as an agent of selection.
Astyanax fasciatus (Parzefall 2000) and Rhamdia transiitoria (Trajano 2001)
are the only species where agonistic behavior (a kind of direct competition)
has been compared in a troglobite and what is known to be its surface
ancestor. The surface catfish shows clear preadaptations and its agonistic
behaviors did not differ from its troglobiotic relative. However the surface

characin is only marginally preadapted and researchers argue that its

agonistic behavior in caves is reduced and different than for the surface
fish where vision is so important. Parzefall writes that surface Astyanax
defend small territories when they are not schooling or shoaling but it is
hard to imagine how this would help an individual. It seems to me that
food in a surface stream would not be extremely patchy or defendable.
Eyed surface fish in the dark show no aggression but blind cavefish do
show some aggression. However it is not clear to me that increasing
‘defense’ of a few cm larger areas as cave individuals decrease activity
with food deprivation (Parzefall 2000) will allow them more access to food
in a laboratory tank much less in a cave.
Bechler (pers. comm.) found clear intraspecific territorial behavior in
4 Forbesichthys or 4 Typhlichthys in the realistic scale of a 6.3 m long 4 m2
artificial stream only where there were clearly defendable rock piles. But
both species of Amblyopsis moved through all parts of the stream, were
rarely in a rock pile, and only showed subtle and occasional agonistic
behavior.
The general reduction in agonistic behavior in troglobitic fish (Parzefall
2000 and Bechler 1983) is consistent with my hypotheses that there are
no defendable resources in caves that would favor interference/direct
competition. But the evolution of increased foraging and metabolic
efficiencies does not mean that resources cannot still be limiting for
troglobites. In the case of one Amblyopsis spelaea population (Poulson
1969) over a number of years numbers fluctuated 2.3 fold, total mass 1.3
fold, and total metabolic demand only 1.15 fold. This natural experiment
suggests to me that exploitation competition is regulating populations in
a density dependent manner. And the increase in reproduction in another
Amblyopsis spelaea population after collecting many large fish is consistent
with this hypothesis (Poulson 1969). This was an inadvertent manipulative
experiment.
Commensalism and Mutualism

Despite the fact that there can always be cascading indirect effects
along food chains (Culver 1982), nobody has explored possibilities of
commensalisms and mutualism affecting cavefish. These +/0 and +/+
biotic interactions may not be important for cavefish that are always at the
top of the food chain. Top predators are the trophic level that is likely to
affect other species. For example if a cavefish controlled the population of
an isopod species that was a competitive dominant over another species
then that would be an indirect mutualism; the enemy of my enemy is my
friend. Of course there are rarely two species of cavefish in one cave and
when there are they are in different habitats. This means there is little
chance for mutualism or commensalisms between cavefish species.

VI. ALLOPATRY, SYMPATRY, AND SYNTOPY IN

CAVEFISH
A. Terminology
I start by reviewing terms that describe the geographic relation among
incipient species, newly evolved species, or distantly related species.
In the classic model of speciation the process starts when a barrier to
dispersal arises due to a vicariant event like a prolonged drought or
glacial cycle. This splits a species range so that the separated populations
(now allopatric) begin to evolve differences because their environments
are different. If these incipient species, designated as subspecies if they are
morphologically recognizable, are separated long enough to accumulate
sufficient differences they become reproductively isolated and are now
full species. Species that are separated geographically are allopatric. If the
geographic barrier to dispersal disappears then the species can become
sympatric. And if they occur in the same same cave they have become
syntopic. If these syntopic species are not different enough they may
compete or interbreed. If hybrids are viable then introgression may occur.
This is the case for Astyanax fasciatus cave species today in eastern Mexico. If
newly syntopic species are reproductively isolated they may still compete.
And if competition is rigorous they may evolve greater microhabitat
preferences. This is called character displacement or reinforcement and is
one hypothesis to explain the syntopy of two troglobitic Amblyopsids in
Mammoth Cave, Kentucky. This brings us to consideration of alternative
hypotheses to explain the few cases of syntopy in cavefish.
B. Local Organic Enrichment in a Large Cave

If organic enrichment is natural and localized in a large cave system, as
with a maternity colony of bats, then food may not be limiting and several
related species may coexist in syntopy. Alternatively one of two troglobites
could have a refuge from competition in an area of the same cave with
much less food. Shelta Cave is a large system in northern Alabama that
may exemplify both of these possibilities. Before the gray bat colony was
lost there were five species of macroscopic troglobites: an Amblyopsid
fish, an Atyid shrimp, and three crayfish (Cooper 1975). This cave has
unusually stable abiotic conditions with very infrequent and localized
abiotic input of allochthonous organic matter. Even with a small gray bat
colony one of the cave crayfish species had an age at first reproduction
estimated as 45 years with only 1-3 successful reproductions for adults
that may live to 100 years!

C. Gradients of Food Supply in a Large Cave

In the huge Mammoth Cave system three Amblyopsid species occur
and two troglobites seem to replace one another along an upstream to
downstream gradient of increasing depth and decreasing food supply
(Poulson 1992). Possible hypotheses to explain this syntopy are discussed
in the chapter on The Amblyopsidae (Niemiller and Poulson, this volume).
One hypothesis involves competition and the other is that the species
had different habitat and food needs before they became sympatric and
syntopic.
VII. PROBLEMS, MYSTERIES, AND CONTROVERSIES
A. Problems of Measuring Food Supply for Cavefish

First I echo Trajano’s (2001) lament that it is rare for cavefish researchers
to describe much less quantify habitat characteristics. I add that it is even
rarer for them to quantify food supply. I suggest that even the best studies
could be better by considering rigor, variability and predictability of
allochthonous food input.
Since most of us argue that food limitation is a central agent of selection
in caves, it is incumbent on all of us to better characterize food supply.
We should at least try to compare amounts of food in caves with large
vs small populations or densities of each cavefish species we study. One
way to measure potential food input from outsIde is the ‘openness’ and
depth of the karst – cave system (e.g. Noltie and Wicks 2001, Niemiller
and Poulson, this volume). More directly there are a number of ways to
quantify the amount and patchiness, of fine and coarse particulate organic
matter in cave streams in space and over time (e.g. Poulson 1992). If there
are rocks it is easy to measure the frequencies and densities macroscopic
potential prey, like isopods and amphipods.
Of course the fish are surely better than we are at finding food and
we can measure this success, if large populations justify sacrifice, by the
frequency and volume of different sized food items in the gut. Using fish
success as a measure of food supply and food limitation has the danger of
some circularity if comparing two cavefish species since one expectation
of evolved troglomorphy is more efficient location and capture of food.
Put another way, the same cave could support different densities of a more
and less troglomorphic species.
In many caves we need to consider the + tradeoff of food input and
the – tradeoff of injury or washout from flooding. Pearson (this volume)

has followed up on my observations of Amblyopsis spelaea from upstream

to downstream of a system that floods regularly and has input of epigean
prey from the surface (a + tradeoff). Pertinent to risk from flooding (a –
tradeoff) is the increased volume and current speed in the downstream
cave. These caves (Spring Mill State Park Indiana where Eigenmann
studied in the early 1900s) are separated by short karst windows. My
starting observation on one date was the change in abundance and size-
frequency distribution going downstream (Poulson 1961, Figure 4). The
numbers, over comparable stream lengths, were 86 to 36. The sizes ranged
from 20-83 mm SL upstream to 45-87 downstream. Interpreted today in
the context of meta-populations this suggests that the upper cave is a
source and the lower cave section is a sink. One explanation is that those
small fish that are washed out of the upper cave perish; in the upper cave
small fish are in streamlets and backwaters that are not as frequent in the
lower cave. There is certainly plenty of food in the lower cave as there
are high densities of isopods and amphipods in the shallow rocky riffles
upstream of pools where the fish live. Consistent with this high food
supply in the lower cave, in one year Pearson found about 20 percent of
the adult females carrying eggs in their branchial cavities on one date. But
the apparent risks of fast currents and injury are such that there was no
recruitment of immature fish the next year.
B. What is the Primary Habitat for Cavefish?
Background
Even for species that have been found in many caves, like three of the four
troglobites in the Amblyopsidae, there are few caves with large populations
of fish. Put another way, in most localities we see only a few and often
only one fish at a time. Before discussing alternative hypotheses to explain
these patterns we need to review some observations about cave passages.
Here I greatly simplify the details provided in Palmer’s complete treatise
on Cave Geology (2007).
Caves Enterable by Humans

Humans can enter few caves because most have no entrances and even in
caves with entrances we often cannot access much of the upstream areas
of the subterranean drainage basins. On the other hand there is not likely
to be a ‘woodwork’ around enterable passages that we cannot census for
fish. Above a critical dimension of a few cm cave passage enlargement
occurs along a few routes that quickly ‘capture’ >95% of water volume.
Even in cave streams with lots of loose rocks many capture-mark-recapture

studies reviewed by Trajano (2001) show that visual census at one time
finds a third to half of the real population. And in many of the best-studied
Amblyopsid cavefish caves there are no obvious hiding places and so visual
censuses miss a maximum of 10-20 percent of the population estimated by
mark-recapture statistics (Pearson, this volume).
Hypotheses
So this brings us to two main alternative hypotheses about where most
individuals of a troglobitic cavefish species occur: 1. In areas accessible
to census; or 2. In areas not accessible to census. Either hypothesis has
to explain the general patterns that only a small percent of all known
localities for a troglobitic fish species have more than a few fish seen at
one time.
Let me start with the best mark-recapture study of a cavefish population
I have read. This is a multi-year near monthly study of an Amblyopsis rosae
cave with at least 35 fish seen and as many as 29 fish caught on each trip
(Brown and Johnson 2001). Logan Cave has a gray bat maternity colony
(the only other Amblyopsis rosae cave with large populations is also a gray
bat cave; see Figure 4 in Poulson 1961, Cave Springs Cave). Logan has
about 1.6 km of accessible stream but only about 10 percent of fish are
seen in the upper third of the stream which is not optimum habitat and
another 10 percent are found in the lower third of the stream near the
entrance. From their two-year study with 23 trips Brown and Johnson infer
a regular movement of unmarked fish from the aquifer, but not into the
aquifer. And they conclude that most fish are in the inaccessible aquifer.
They suggest that this conclusion helps explain why a few fish at a time
are found in springs or wells and even in caves. In each case it is rare to
see as many as five fish at a time even over multiple visits to the same site.
They infer that these few fish are at the edge of aquifers that we cannot
access. They rationalize the obvious issue of food scarcity in the aquifer
by observations by video camera in newly-dug wells of extensive mats of
organic, bacteria like material.
I find Brown and Johnson’s (2001) observations and deductions
intriguing but not convincing evidence for most fish being in the
inaccessible aquifer. The organic matter in the well could be due to local
organic pollution enrichment in the shallow aquifer. In the accessible cave
they do not preclude the hypothesis that new unmarked fish were in the
rubble bottom of the stream. This is consistent with studies in the other
population in a grey bat cave where 27-72 fish, all those catchable, were
collected by a Tulane University researcher at three times 1-6 months apart
(Poulson 1961, Figure 4, Cave Springs Cave). Also they infer, from size-

frequency data, that the fish are reproducing in the accessible parts of
Logan Cave. I agree with this last interpretation.
So why do I come to an opposite conclusion that caves, but not the
water table aquifers, are the primary habitat for Amblyopsis rosae (and for
all cavefish)? First, in terms of energetics the two bat caves should be the
primary habitats because they have the highest food supplies. Furthermore
~ 80% of censused fish are seen in the stream area below the bat roosts. At
a larger scale each bat cave is a source population, rather than a sink. It
is recognized as a source by having all fish sizes represented with at least
some very small fish every few years. This means that the population is
reproducing and at least maintaining itself. It is likely to be a larger source
of emigrants than of immigrants.
On the other hand, the many sink populations each have only one or
a few fish seen at any one time. For any one locality and in aggregate
their size distribution is skewed toward larger individuals and if there are
any small individuals they are rare over time and among sites. Thus these
populations are probably not reproducing and so must be maintained by
immigration. Thus they are sinks.
Sources and Sinks

Population ecologists argue that both sources and sinks in a meta-
population are important. Sink populations can become sources if some
disaster wipes out large populations in a source cave. Thus the many sink
populations spread the risk of population loss so that the meta-population
always survives. For cavefish the individuals in the sink populations are
not necessarily doing poorly since they have few competitors for a low
food supply and have metabolic economies and excellent food-searching
behaviors (Niemiller and Poulson, this volume). And even if they have
no food they lose weight slowly and recover quickly when they find food
with their efficient search patterns (Niemiller and Poulson, this volume).
In fact they may rarely starve even in the aquifer since the condition of
Amblyopsis rosae in the bat caves, in terms of fat deposits and gut contents,
is only subtly greater than for fish collected from springs, wells, or small
caves with no obvious allochthonous organic matter (Poulson 1961).
Another perspective on sink populations is that they include individuals
that are potential colonizers of new habitats. It is probable that sink
populations are much larger than apparent from the “windows” at which
we see them, like springs or wells, but they are surely spread out over
large areas. When they encounter an open cave stream with a good food
supply their search patterns will keep them there (Niemiller and Poulson,
this volume). They are the successful individuals in what biogeographers
call ‘sweepstakes dispersal’.

Pearson (this volume) addresses the issue of primary habitat for at least
six populations of Amblyopsis spelaea where he or I have visually censused
60-100 fish on every visit over 40-50 years. His data and mine are consistent
with the inference that individuals seen at the edge of the aquifer are
outliers of sink populations with high dispersion of individuals. And we
know for this species that it carries eggs and newly hatched young in its
branchial chamber for almost six months so a ‘pregnant’ female has a long
time to disperse to a new cave through the water table and found a new
population.
A corollary of all of the above discussion is that calculations of cavefish
population sizes and densities (Trajano 2001) mean little or nothing without
a lot of context about stream characteristics and food supply. And home
range and movements (Trajano 2001) are only meaningful within a species
where one can relate the data to habitat structure and food supply.
C. Why do Some Highly Troglomorphic Fish Species

Show Relict Distributions?
Given the troglomorphic adaptations of cavefish and the existence of
meta-populations, it is a mystery as to why some species come to have
relict distributions.
The most likely hypothesis to me is that they always had restricted
distributions. This is only an extension of what we now know to be
a common pattern of different populations of one epigean ancestor
becoming independently isolated in different cave systems in a karst
region. If the cave region/limestone formation is limited in extent then
there never could have been a widespread meta-population. We have seen
that cavefish can disperse far and well but only within a limestone area
without subterranean barriers.
We have several models of widespread surface species that have incipient
troglomorphy in restricted parts of a karst region without obvious barriers
to dispersal. Two of these are not cavefish (Gyrinophilus salamanders,
Niemiller and Fitzpatrick, pers. comm., and Gammarus minus amphipods,
summary in Culver and Wilkens 2001) but they have been exhaustively
studied and analyzed so the interpretations could be generally applicable.
The amphipod has been studied in most detail. The small karst area where
they have evolved troglomorphy has extensive caves with a lot of detrital
food and barriers to dispersal between the spring and cave populations. So
if some climatic even wipes out the surface spring populations the species
complex will have a relict distribution. Cottus sculpins may be a fish
example parallel to the amphipod. Incipient troglomorphic populations
occur only in a small karst area of Missouri with extensive caves and lots

of detrital wash-in (Burr et al. 2001). I suspect there may be more such
examples from areas other than the United States when sufficient detailed
surveys the distribution of surface–cavefish complexes are done.
The hypothesis just discussed to explain relict distributions, of a small
area where isolation in caves can occur, does not seen to fit the most
troglomorphic Amblyopsid, my namesake Speoplatyrhinus poulsoni. It is
restricted to Key Cave in an intensively studied karst area of northern
Alabama with many caves and no apparent barriers to subterranean
dispersal (Kuhajda and Mayden 2001).
One hypothesis relates to the character of the cave. Key Cave has a
maze structure typical of caves formed at the water table in flat bedded
limestones. As such the deeper parts of the cave may have been flooded by
impoundments in the adjacent Tennessee River so that suitable habitat has
been restricted. But this does not explain why many nearby caves have
Typhlichthys which has occurred as a single individual for at least seven
years in Key Cave.
Another hypothesis is that Speoplatyrhinus was out-competed by what
is now a widespread species Typhlichthys subterraneus (e.g. Woods and
Inger 1957). I think this is unlikely based on my observation of density
dependent population regulation in Amblyopsis spelaea (Poulson 1969)
which indicates that effective exploitation competition is operating in a
highly troglomorphic species. If this is general then less troglomorphic
fish, like Typhlichthys subterraneus, should never be able to competitively
exclude more evolved species, like Speoplatyrhinus poulsoni. The reason is
that the more efficient species will always have a refuge in finding food.
As the less efficient, but perhaps faster feeding, species reduces food to
lower and lower abundances it may suffer exploitation competition but
the more efficient troglobite will not yet be food limited.
Another hypothesis to explain relict distributions comes from a
possible answer to another mystery i.e. how do troglophiles apparently
outcompete troglobites when there is pollution by organic enrichment
(not just in cavefish)? The parallel on a longer time scale is that a new
species, originally troglophilic, might be able to ‘invade’ a cave region with
a pre-existing troglobite if the food supply in caves increased at a time that
surface climate resulted in isolation of the new species. We cannot test this
paleogeographic hypothesis but we can do experiments relevant to the
observation that organic enrichment can result in a troglophile replacing a
troglobite.
I suggest a testable model of observed replacement of troglobites by
troglophiles with organic enrichment due not to competition but rather
to demographic swamping. This is analogous to a human minority
group replacing the original ethnic or racial group simply due to higher

reproductive rate. Such replacement would proceed slowly and eventual

loss of the original fish species would be due to random processes. This
demographic swamping could be speeded by harassment if the less
efficient troglophile is more aggressive than the troglobite even if resources
are not defendable. Though I believe that this demographic swamping
mechanism operates generally with organic enrichment, I have no
evidence that it explains the restriction of Speoplatyrhinus to a single cave.
If the swamping hypothesis were correct I would expect caves around
Key Cave to have much higher food supplies than Key Cave and large
populations of Typhlichthys.
D. Dating Isolation in Caves
Molecular Clocks
The accuracy of molecular clocks is controversial and in cavefish may be
complicated by increasing generation times with evolution of troglomorphy.
Another problem is finding cave independent traits for which mutations
are neutral even in early stages of isolation in caves. We may have to wait
for complete genome sequencing to find single nucleotide changes that
are not part of genes.
Neutral Mutations
Despite problems using molecular clocks, our best candidates for cave
independent traits associated with rudimentation in caves may give the
relative timing of isolation in caves. To check such inferences we need
to also use at least geo-morphological evidence (e.g uppermost – oldest
passage sediments in Mammoth Cave dated back to 3.5 my by cosmogenic
Al and Be) and palynological evidence (pond bottom cores dating back
several hundred thousand years) to match timing of divergence inferred
from genetic evidence with times of change in landscapes and climates. A
major problem is that we do not have geomorphological or isotope data
about climates and land forms that go far enough back in time, to the early
Pliocene or even Miocene times, that molecular data suggest are times of
Amblyopsid lineage divergence (Niemiller, pers. comm.)
E. Phylogeography and Parallel Evolution of Cavefish
Species Complexes
It is becoming increasingly clear from molecular data that many or
most species of trogolobite are actually species complexes that represent
independent isolations in caves (e.g. Trajano 2001). Until the relevant

molecular studies are done (e.g. chapters in this volume by Borowsky

and by Niemiller and Poulson) Trajano has suggested (2001) that we use
geologically and hydrologically defined subterranean drainage basins as
criteria for occurrence of sister species.
A related area of evolutionary research on cavefish is inferring
phylogenies and current biogeographic history within and among species
or species complexes using continually improving computer algorithms
to analyze molecular data. This approach is emphasized in this volume by
Borowsky and by Niemiller and Poulson. I hope experts will enlighten us
about the relative efficacy of different genetic markers, e.g. mitochondrial
vs nuclear genes and single nucleotide polymorphisms vs RAPDs vs
microsatellite variation for answering questions at different temporal and
spatial scales. It is not clear to me that any of these markers can tell us even
the relative times that different troglobites have been isolated in caves.
However molecular markers may help us avoid the problem of
distinguishing cave dependent characters subject to convergent and
parallel evolution, from cave independent characters (see the benchmark
paper by Christiansen 1961). Unfortunately no one has yet studied any
demonstrably cave independent character(s) in enough depth to infer
absolute evolutionary time in caves for any cavefish. In this context no
morphological characters should be neglected in constructing phylogenies
even if they reflect the action of many quantitative genes that we cannot
identify with present techniques (but see Protas et al. 2008).
Phylogeography of the Amblyopsidae

One of the ‘Mysteries of the cavefish’ we will explore for the Amblyopsidae
(Niemiller and Poulson, this volume) is how or whether successive glacial
cycles resulted in isolation of different sympatric or different allopatric
surface ancestors in different places resulting in very different current
geographic distributions. Culver (pers. comm.) seems to favor the view
that the differences in adaptation among Amblyopsid troglobites (Poulson
1963, 1985) may be due to differences in cave type and food supply in the
areas where each of allopatric ancestors were isolated. I agree that some
of the differences among troglobites can be explained by differences in
geology and cave character among cave regions (see also the section above
on Phenotypic vs Genotypic Troglomorphy).
I have also discussed how the relative time of isolation, based on
overall eye and pigment rudimentation, can be used to infer different
evolutionary time in caves (Poulson 1985). This supposition is supported
by the great differences in millions of years of separation of Amblyopsid
lineages as deduced from molecular data in which Amblyopsis rosae is the

oldest lineage (Niemiller and Poulson, this volume). Niemiller’s ongoing

molecular genetic studies will give us some answers as to the relative
ages of species divergence but cannot tell us even the relative durations of
time since isolation in caves. What we can do is look at different species
of Amblyopsids and other cavefish groups and ask whether the present
abiotic and biotic agents of selection are different enough to expect more
troglomorphy in some species or in some karst regions.
F. Mechanisms for Evolutionary Isolation of Cavefish
Hypotheses
Though we may not be able to date the time or even the relative times of
isolation we can use natural experiments to evaluate alternative hypotheses
for the mechanism of isolation. Many authors have discussed hypotheses
about how epigean species came to be cave species but Holsinger (2000)
provides the best review I have read of the two main competing processes
(models for isolation).
Climate Effects
For cavefish the climate relict model aka Pleistocene effect model seems
to me more applicable than the adaptive shift model aka habitat shift
model.
Drying and warming during Pleistocene cycles has apparently been
just as important to evolutionary isolation in caves in parts of the tropics,
South America at least (Trajano 2001), as in the temperate zones of the
world.
In North America the troglophilic spring cavefish Forbesichthys agassizi
may or may not be a good model for the transition as you will read in our
chapter on the Amblyopsidae (Niemiller and Poulson, this volume). In
Brazil, despite likely loss of surface streams with drying, during Pleistocene
times (Trajano 2001) there is and probably were much higher food supplies
in caves than in periglacial temperate areas. With warm temperatures and
high rainfalls overall I was the first to argue in detail that caves form faster
and have more entrances, including karst windows, than in temperate
zones. In addition many if not most tropical caves have bat populations
(Poulson and Lavoie 2000, Poulson 2005). So perhaps it is not surprising
that tropical troglobitic fish, many of which have extant ancestors that
are demonstrably preadapted (e.g. Trajano 2001 and Wilkens 2005) have
not been shown to have the extreme metabolic economies or sensory
hypertrophies documented for temperate troglobitic fish, especially the
Amblyopsidae.

Habitat Shift
I suggest that this is a rare mode of isolation for cavefish. Examples may be
cases where the actual or putative surface ancestor is a diurnal, schooling
fish with large eyes. Possible examples are Caecobarbus geertsi and Astyanax
fasciatus. And even in these cases there were some preadaptations with
some sensory systems useful in darkness. In an earlier section (Astyanax
fasciatus vs Amblyopsidae as Models) I have discussed the catastrophic
stream capture that all agree was the mode of isolation for Astyanax
fasciatus. This is in no way analogous to the active invasion of lava tubes
by ancestors of cave Arthropods shown by Howarth’s (1981) studies.
VIII. PROTOCOL FOR STUDY OF A NEW OR

ENDANGERED SPECIES
A. Rationale
I provide the following outline both as a summary of some parts of this
chapter and as a suggestion for priority of study for a newly discovered
species, an endangered species, or fish in caves where study is difficult (e.g.
remote and need for SCUBA). In sequence I suggest field observations,
non-destructive studies of captured individuals in the cave, behavioral
observations in an aquarium in situ, physiological and behavioral studies
of live individuals held temporarily in the laboratory, and detailed studies
of preserved specimens.
B. Field Observations
Both descriptions of the cave and its fauna and estimates of size-frequency
distribution of the fish are critical. The likelihood of flooding, amounts
of allochthonous organic matter, and substrate types in riffles, pools, or
ponds are combined with measures of live food (see Poulson 1992 for
detailed methods). The nature of the karst may also be important (e.g.
Noltie and Wicks 2001). And the estimated size-frequency distribution is
a powerful clue as to whether the cave is a source or sink and whether the
species is long- or short-lived.
C. Field Data on Captured Individuals

I start with observations of behavior as you try to capture a fish; remember
that troglomorphy is often associated with loss of ‘fright’ reactions. Once
a fish is caught measure standard length and combine this with volume

displacement to estimate condition factor. Assess the amount and location

of visible fat deposits and external parasites or lesions. Then take a scale
sample for possible determination of minimum age from scale marks that
show slowing of growth. Also take a fin clip for genetic studies. With care
these studies can be done without anesthesia (e.g. Pearson, this volume).
D. In situ Aquarium Observations

Start with a hand-carried mini aquarium in which you watch a fish after
taking data on the fish in hand. Note how calmly or frenetically the
fish swims and how fast it calms down. Once calm, time the frequency
and estimate the depth of respiratory movements; this may give a clue
to metabolic rate. In a larger aquarium on the stream bank, with a rock
for structure and possible hiding place, take notes on swimming and
exploratory behavior. Does the fish swim in the open water as wll as along
the sides or bottom? Is it highly thigmotactic? Does it investigate or pause
under the rock? If there are live prey available in the cave put some in
the aquarium and watch possible searching behavior and prey capture
details.
E. Laboratory Aquarium and Respirometer Studies
Short-term Studies of Days to a Week

Use video (e.g. Bechler 1983 and Sheryl Coombs’ website) to better
quantify swimming, exploring, searching, and prey capture behaviors and
efficiencies with different prey types and prey densities (e.g. Niemiller and
Poulson, this volume). Then use a torus respirometer with IR light activity
sensor to relate metabolic rate to swimming activity. This gives a measure
of adaptation to chronic food scarcity.
Longer-term Studies of Weeks to Months

If you have several individuals then keep them in separate aquaria with a
rock and then study agonistic behavior of home and intruder fish (Bechler
1983). Repeat the experiments with well-fed and fish deprived of food for
a week. In fish derived of food, follow weight loss and condition factor
for a month, or longer if weight loss is slow; this gives a measure of any
adaptations to periodic food scarcity. At the end of the starvation period
provide food ad libitum and measure the recovery of weight and condition
factor along with any obvious fat storage.

F. Study of Preserved Specimens

Even endangered species will have specimens in collections upon
which the species description was based. Urge the museum curators to
allow detailed studies and even partial dissections to get the maximum
data possible. There are some data that can only come from preserved
specimens.
Data without Dissection

Describe external sense organs and quantify their number and degree of
development (e.g. Poulson 1961, 1963 and 1985). If a size series is available
also determine their positve or negative allometric coefficients. Do the
same for fin and head lengths. Determine the weight length relation,
relative head size, and branching of fin rays as clues to possible neoteny
(e.g. Cooper and Kuehne 1974). Quantify the density, sizes, and character
of pigmented melanophores and their allometric coefficient.
Data from Dissections

At a minimum open the body cavity and measure fat deposits, gono-
somatic index, and mature ova number and maximum size. If allowed,
determine frequency and volume of detritus and prey items in the stomach
and intestine and relate this to data on food supply in the cave where
the specimens were collected. And, if allowed, carefully open the cranial
cavity to measure and draw brain parts, olfactory rosette, any visible eye
remnant, and semi-circular canals and otoliths (e.g. Poulson 1963 and
Cooper and Kuehne 1974).
Acknowledgements
I start by thanking the editors of this volume, especially ‘Leo’ or Eleonora
Trajano, both for inviting me to write this introductory chapter and for
organizing this book in my honor. I was surprised and am still touched.
But the acme of my biospeleological career was the naming of the fish
I ‘predicted’ as Speoplatyrhinus poulsoni by two of my closest colleagues
(cOOp and Bob Kuehne 1974) aided and abetted by the good humor and
photos of Tom Barr. My namesake has even made the Burger King kids’
collectible cards on endangered species. Next I thank all of my students
and colleagues over the past ~50 years who have been sounding boards
for my ideas about cave biology and cavefishes as they have changed
and matured. Especially influential have been Tom Barr, Dave Bechler,
Richard Borowsky, Ken Christiansen, cOOp Cooper, Dave Culver, Dave

Griffith, John Holsinger, Bill Jeffery, Tom Kane (deceased), Kathy Lavoie,
Bill Pearson, Leo Trajano, and Horst Wilkens. Next, I thank in advance
any and all who build on my contributions by considering my suggestions
and testing my hypotheses. I especially throw down the gauntlet to my
new young colleague Matt Niemiller. And finally I admonish all readers
to practice my mantra of using multiple hypotheses and doing long-term
studies as exemplified in one of my limericks.
Tom always uses multiple hypotheses
To avoid those that smack of sophistries
To Tom long studies aren’t dippity and he eschews serendipity
The right times and right places aren’t wizardries
References
Bechler, D.L. 1983. The evolution of agonistic behavior in Amblyopsid fishes.
Behavioral Ecology and Sociobiology 12: 35-42.
Borowsky, R. 2008. Restoring sight in blind cavefish. Current Biology 18: R23-R24.
Brown, J.Z. and J.E. Johnson. 2001. Population biology and growth of Ozark
cavefishes in Logan Cave National Wildlife Refuge, Arkansas. Environmental
Biology of Fishes 62: 161-169.
Burr, B.M., G.L. Adams, J.K. Krejca, R.J. Paul and M.L. Warren, Jr. 2001.
Troglomorphic sculpins of the Cottus carolinae species group in Perry County,
Missouri: Distribution, external morphology, and conservation status.
Chapin, F.S. III, A. Kellar and F. Pugnaire. 1993. Evolution of suites of traits in
response to environmental stress. American Naturalist 142 (Supplement): 61-84.
Christiansen, K. 1961. Convergence and parallelism in cave Entomobryinae.
Cooper, J.E. 1975. Ecological and behavioral studies in Shelta Cave, Alabama with
emphasis on decapod crustaceans. PhD Dissertation, Department of Zoology,
University of Kentucky, Lexington. 364 pp.
Cooper, J.E. and R.A. Kuehne. 1974. Speoplatyrhinus poulsoni, a new genus and
species of subterranean fish from Alabama. Copeia 1974: 486-493.
Culver, D.C. 1982. Cave Life: Evolution and Ecology. Harvard University Press,
Cambridge.
Culver, D.C. 1985 (ed.). Regressive Evolution, Bulletin of the National Speleological
Society 47: 70-162.
Culver, D.C. and T.L. Poulson. 1971. Oxygen consumption and activity in closely
related Amphipod populations from cave and surface habitats. American
Midland Naturalist 85: 74-84.
Culver, D.C. and H. Wilkens. 2000. Critical review of the relevant theories of the
evolution of subterranean animals. In: Ecosystems of the World, Volume 30,

Subterranean Ecosystems, H. Wilkens, D.C. Culver and W.F. Humphreys (eds.),

Elsevier, Amsterdam, pp. 381-398.
Culver, D.C. and W.B. White (eds.). 2005. Encyclopedia of Caves. Elsevier,
Amsterdam.
Culver, D.C., T.C. Kane and D.F. Fong. 1995. Adaptation and Natural Selection in Caves:
The Evolution of Gammarus minus. Harvard University Press, Cambridge.
Griffith, D.M. and T.L. Poulson. 1993. Mechanisms and consequences of intraspecific
competition in a Carabid cave beetle. Ecology 74: 1373-1383.
Heuts, M.J. 1951. Ecology, variation, and adaptation of the blind African cave fish
Caecobarbus geertsi Blgr. Annales Musee Royale Zoologie Belgique 82: 155-230.
Hill, L.G. 1966. Studies of the biology of the spring cavefish, Chologaster agassizi
Putnam. PhD Dissertation, Department of Biology, University of Louisville. 104
pp.
Holsinger, J.R. 2000. Ecological derivation, colonization, and speciation. In:
Ecosystems of the World, Volume 30, Subterranean Ecosystems, H. Wilkens, D.C.
Culver and W.F. Humphreys (eds.), Elsevier, Oxford, pp. 399-415.
Howarth, F.G. 1980. The zoogeography of specialized cave animals: A bioclimatic
model. Evolution 34: 394-406.
Huppop, K. 1985. The role of metabolism in the evolution of cave animals. The
National Speleological Society Bulletin 47: 136-146.
Huppop, K. 2000. How do cave animals cope with the food scarcity in caves? In:
Jeffery, W.R. 2005. Adaptive evolution of eye degeneration in the Mexican blind
cavefish. Journal of Heredity 96: 185-196.
Keith, J.H. and T.L. Poulson. Broken-back syndrome in Amblyopsis spelaea, Cave
Research Foundation 1979 Annual Report: 45-48.
Kuhajda, B.R. and R.L. Mayden. 2001. Status of the federally endangered Alabama
cavefish Speoplatyrhinus poulsoni (Amblyopsidae), in Key Cave and surrounding
caves, Alabama. Environmental Biology of Fishes 62: 215-222.
Mohr, C.E. and T.L. Poulson. 1967. The Life of the Cave. McGraw-Hill, New York,
232 pp.
Montgomery, J.C., S. Coombs and C.F. Baker. 2001. The mechanosensory lateral
line system of the hypogean form of Astyanax fasciatus. Environmental Biology of
Fishes 62: 87-96.
Niemiller, M.L. and B.M. Fitzpatrick. Phyologenetics of the southern cavefish,
Typhlichthys subterraneus: Implications for conservation and management.
Proceedings of the 2007 National Cave Management Symposium. (In Press)
Noltie, D.B. and C.B. Wicks. 2001. How hydrogeology has shaped the ecology of
subterraneus: Insights on the sightless from understanding the underground.
Palmer, A.N. 2007. Cave Geology. Cave Books, Dayton, Ohio, 454 pp.

Parzefall, J. 2000. Ecological role of aggressiveness in the dark. In: Ecosystems of the
World, Volume 30, Subterranean Ecosystems, H. Wilkens, D.C. Culver and W.F.
Humphreys (eds.), Elsevier, Amsterdam, pp. 221-228.
Porter, M.L., K. Dittmar and M. Perez-Losada. 2007. How long does evolution of
troglomorphic form take? Estimating divergence times in Astyanax mexicanus.
Acta Carsologica, Time in Karst, Postojna, pp. 173-182.
Poulson, T.L. 1960. Cave adaptation in Amblyopsid fishes. PhD Dissertation,
Department of Zoology, University of Michigan, Ann Arbor. University
Microfilms 61-2787.
Poulson, T.L. 1963. Cave adaptation in Amblyopsid fishes. American Midland
Poulson, T.L. 1964. Animals in aquatic environments: Animals in caves. In:
Handbook of Physiology: Environment. D.B. Dill (ed.), American Physiological
Society, Washington, pp. 749-771.
Poulson, T.L. 1969. Population size, density, and regulation in cave fishes. Actes of
the 4th International Congress of Speleology, Ljubljana, Yugoslavia 4-5: 189-192.
Poulson, T.L. 1985. Evolutionary reduction by neutral mutations: Plausibility
arguments and data from Amblyopsid fishes and Linyphiid spiders. Bulletin of
Poulson, T.L. 1992. The Mammoth Cave ecosystem. In: The Natural History of
Biospeleology, A. Camacho (ed.), Museo Nacional de Ciencias Naturales, Madrid,
pp. 569-611.
Poulson, T.L. 2001. Adaptations of cave fishes with some comparisons to deep-sea
fishes. Environmental Biology of Fishes 62: 345-364.
Poulson, T.L. 2005. Food Sources. In: Encyclopedia of Caves, D.C. Culver and W.B.
White (eds.), Elsevier, Amsterdam, pp. 255-263.
Poulson, T.L. and K.H. Lavoie. 2000. The trophic basis of subsurface ecosystems.
In: Ecosystems of the World, Volume 30, Subterranean Ecosystems, H. Wilkens, D.C.
Culver and W.F. Humphreys (eds.), Elsevier, Amsterdam, pp. 118-136.
Protas, M., I. Tabansky, M. Conrad, J.B. Gross, O. Vidal, C.J. Tabin and R. Borowsky.
2008. Multi-trait evolution in a cave fish, Astyanax mexicanus. Evolution and
Development 10: 196-205.
Proudlove, G.S. 2006. Subterranean Fishes of the World. International Society for
Subterranean Biology, Moulis.
Romero, A. 2001 (ed.). The Biology of Hypogean Fishes. Kluwer Academic Publishers,
Dordrecht, 376 pp.
Sexton, E.W. and A.R. Clark. 1936. A summary of the work on the amphipod
Gammarus chevreuxi, Sexton, carried out at the Plymouth Laboratory (1912-
1936). Journal of the Marine Biological Association, United Kingdom 21: 357-414.
Trajano, E. 2001. Ecology of subterranean fishes: An overview. Environmental
Weber, A. 2001. Fish and Amphibia. In: Ecosystems of the World, Volume 30,

West-Eberhard, M.J. 2003. Developmental Plasticity and Evolution. Oxford University

Press, Oxford, 816 pp.
Whittaker, F.H. and L.G. Hill. 1968. Proteocephalus chologaster sp. N. (Cestoda:
Proteocephalidae) from the spring cavefish Chologaster agassizi Putnam (Pisces:
Amblyopsidae) of Kentucky. Proceedings of the Helminthological Society of
Washington 35: 15-18.
Whittaker, F.H. and S.J. Zober. 1978. Proteocephalus poulsoni sp. N (Cestoda:
Proteocephalidae) from the northern cavefish Amblyopsis spelaea Dekay, 1842
(Pisces: Amblyopsidae) of Kentucky. Folia Parasitologica 25: 277-280.
Wilkens, H. 2007. Regresssive evolution: Ontogeny and genetics of cavefish eye
rudimentation. Biological Journal of the Linnean Society 92: 287-296.
Wilkens, H., D.C. Culver and W.F. Humphreys (eds.). 2005. Ecosystems of the World,
Volume 30, Subterranean Ecosystems, Elsevier, Amsterdam, 791 pp.
Woods, L.P. and R.F. Inger. 1957. The cave, spring, and swamp fishes of the family
Amblyopsidae of central and eastern United States. American Midland Naturalist
58: 232-256.

CHAPTER 2
Biodiversity and Distribution of the
Subterranean Fishes of the World
Graham S. Proudlove
Department of Zoology, The Manchester Museum, The University of
Manchester, Manchester M13 9PL, UK
E-mail: g.proudlove@manchester.ac.uk
INTRODUCTION
The fishes are the largest vertebrate group with something over 28,000
described species. The majority (around 16,000 species) are marine
but around 12,000 species are found only in the freshwaters found on
continental and island landmasses. These freshwaters are found as
steams, rivers, lakes and ponds and eventually all freshwater finds its way
to the oceans in the global hydrological cycle. During their course from
deposition as rainwater, to eventual arrival at the oceans, some freshwater
penetrates the surface to flow through various types, and various depths,
of subterranean habitats. These vary from the tiny interstices between
sand grains, through caves enterable by man, to vast groundwater bodies
at great depth below the surface. Evolution has been at work within
these subterranean places and animals which have found themselves
in the darkness have adapted and evolved. They have evolved a set of
very distinctive characteristics typical of, and found only in, subterranean
environments. These are found in all taxa, vertebrate and invertebrate,
which are now obligate inhabitants of the subterranean (or hypogean)
realm. Most obviously these animals loose their eyes and any melanin
and carotenoid pigmentation. They very often appear white in colour, or
even transparent. These obvious losses have lead to the oft-used phrase
“blind white cave animals”. In addition to these losses are a set a gains
which are adaptations to life without light and most often very food-poor
conditions. For example, the limbs of Crustacea are elongated, the lateral

line of fishes is expanded, and the sensory setae of insects are lengthened.
There are parallel behavioural, physiological and neurological alterations
all aimed at survival in the hypogean. All of these changes are referred to
as troglomorphic. The relative degree of troglomorphy is related directly
to the time of evolution – the phylogenetic age – of the particular taxon
in question (see Porter et al. 2007 and Trajano 2007 for discussion of this
point).
Out of all 12,000 species of freshwater fish only a very few, currently just
over 150, have successfully colonised the hypogean. The number known is
climbing fast but it is probably not likely to pass 1000 species. The purpose
of this opening chapter is to set out, in broad terms, the overall patterns
of biodiversity in subterranean fishes and to examine their distribution
at global (biogeographic) scale. It sets the scene for the more detailed
chapters making the main body of the book.
There is a large literature (of more than 2500 publications) on all
aspects of the biology of subterranean fishes. The most recent synthesis
of this literature is Proudlove (2006) which provides detailed accounts
for 104 species and includes the complete bibliography to 2004. Wilkens
(2005) is a highly readable account covering all aspects of their biology.
Similar, though briefer, accounts are provided by Proudlove (2004) and
Romero (2004). Romero (2001) is a volume similar to this one with a set of
contributed papers covering many aspects. The chapters by Weber (2000)
and Weber et al. (1998) are both excellent syntheses. Though now very out
of date, and out of print, the book by Thines (1969) is well worth seeing.
THE CHECKLIST
The following checklist contains taxonomic details of all known
subterranean species. Most are formally described but a few are known but
have yet to receive a formal description. The most recent species description
(of Trichomycterus uisae) was published after the Tables in this chapter
were compiled and is not included in these Tables. All authorship details
were checked with the Catalogue of Fishes (Eschmeyer, W.N. (ed.) Catalog
of Fishes, electronic version (updated 29 August 2008) www.calacademy.
org/research/ichthyology/catalog/fishcatsearch.html).
The Development of Knowledge: How Many

Subterranean Fishes are There?
The first subterranean fish species (Amblyopsis spelaea in the USA) was
formally described in 1842. Since then a further 149 species (to December
2008) have been discovered. We can plot the rate at which they have been

Graham S. Proudlove 43
discovered (as cumulative totals) and then use the curve to extrapolate in
order to see how many species may exist. The curves are shown in Fig. 1
(an arithmetic plot) and Fig. 2 (a logarithmic plot). Several features are
evident in the Figures. First the gradient changes dramatically around 1981
(year 140 in the figures). This is almost certainly the result of more species
being found because more caving expeditions were being mounted from
this time onwards. China especially was almost unknown speleologically
before the 1980s but since then very large mileages of passage have been
explored and many cave fish species discovered. (Subterranean fishes are
unusual in that most are found by non-biologists whereas this is not the
case for nearly all other newly described fish species.)
Fig. 1 An arithmetic plot of the cumulative number of subterranean fish species

versus the date they were described. Year 0 = 1842 when the first subterranean
fish was described.
Color image of this figure appears in the color plate section at the end of the
book.
The fitted curve can be used to extrapolate in order to see how many
species may exist in total. In order to see if the curve is a good fit to the
data we can compare the expected total species for a certain year with the
actual observed number. It is easy to se that the curve is a very good fit.
We can then predict that there will be 166 species by 2010 (16 more than
2008) and 209 species by 2015 (59 more than 2008). Unfortunately there is

Fig. 2 A logarithmic plot of the cumulative number of subterranean fish species

fish was described.
book.
no evidence that the curve has reached an asymptote so it is not possible

to predict the total global number of species.
1981 1991 2006 2007 2010 2015
Expected 44 69 138 145 166 209
Observed 43 68 136 150 – –
SYSTEMATIC DISTRIBUTION
Summaries of the systematic composition of subterranean species are
provided in Table 1 (Orders) and Table 2 (Families). There is an obvious
split into two groups: fishes within the clade Otophysi and those not in this
clade. Seventy-nine per cent of all species are otophysans, with only 21%
not being, despite there being many more non-Otophsyi. The explanation
for this is that otophysans make up the majority of the primary freshwater
fishes (Helfman et al. 2009) and it is mainly freshwater fishes that can get
into the subterranean domain. Only the families Bythitidae, Gobiidae and
Eleotridae have invaded caves from the sea. Table 3 lists all genera which
have only one subterranean representative (38 genera and species, 60% of
genera, 25% of species). There are 20 monotypic genera (32%) which are

restricted to subterranean habitats (marked HR = hypogean restricted).

There are eight monotypic genera which are currently known only from
the type locality (marked TLO). Table 4 lists genera with two or more
hypogean species and those where all species are hypogean restricted
(HR), and known only from the type locality. The genera (25, 40%) and
species (35, 23%) which are known only from subterranean sites (HR on
the previous tables) are summarised in Table 5.
Table 1 The numbers of genera and species in each Order which contains
hypogean fishes
Order Genera Species Monotypic Multitypic genera
genera (species)
Characiformes 2 3 1 0
Cypriniformes 23 66 11 0 Otophysi
Siluriformes 19 49 3 3(7) 45 genera (71%)
Gymnotiformes 1 1 0 0 119 species (79%)
Persopsiformes 3 4 2 1(2)
Ophidiiformes 3 8 1 1(6)
Cyprinodontiformes 1 1 0 0 18 genera (29%)
Synbranchiformes 2 5 0 0 31 species (21%)
Scorpaeniformes 1 2 0 0
Perciformes 8 11 2 1(3)
10 orders 63 150 20 6(18)
Table 2 The numbers of genera and species in each Family that contains
hypogean fishes
Family Genera Species Monotypic Multitypic genera
genera (species)
Characidae 2 3 1 0
Cyprinidae 12 37 8 0
Nemacheilidae 8 26 1 0
Balitoridae 1 1 1 0
Cobitidae 2 2 1 0 Otophysi
Ictaluridae 3 4 2 1(2) 45 genera (71%)
Siluridae 1 3 0 0 119 species (79%)
Clariidae 3 4 1 1(2)
Heptapteridae 3 13 0 0
Callichthyidae 1 1 0 0
Trichomycteridae 5 16 0 0
Loricariidae 1 3 0 0
Astroblepidae 1 2 0 0
Table 2 Contd..

Table 2 Contd..
Siluriformes in. sed. 1 3 0 1(3)
Sternopygidae 1 1 0 0
Amblyopsidae 3 4 2 1(2)
Bythitidae 3 8 1 1(6)
Poeciliidae 1 1 0 0 18 genera (29%)
Synbranchidae 2 5 0 0 31 species (21%)
Cottidae 1 2 0 0
Gobiidae 3 4 1 0
Eleotridae 5 7 1 1(3)
21 Families 63 150 20 6(18)
plus 1 in. sed.
Table 3 Genera with only a single known hypogean species. The figure in
brackets is the total number of species in the family (from Nelson 2006). HR =
hypogean restricted, TLO = known from the type locality only
Genus Family Notes
Aspidoras Callichthyidae (177)

Barbopsis Cyprinidae (2420) HR, monotypic
Bostrychus Eleotridae (155)
Caecobarbus Cyprinidae (2420) HR, monotypic
Caecocypris Cyprinidae (2420) HR, TLO, monotypic
Caecogobius Gobiidae (1950) HR, TLO, monotypic
Clarias Clariidae (90)
Copionodon Trichomycteridae (201)
Cryptotora Balitoridae (170) HR, monotypic
Eigenmannia Sternopygidae (28)
Eleotris Eleotridae (155)
Glossogobius Gobiidae (1950)
Glaphyropoma Trichomycteridae (201)
Indoreonectes Nemacheilidae (420)
Iranocypris Cyprinidae (2420) HR, TLO, monotypic
Milyeringa Eleotridae (155) HR, monotypic
Nemacheilus Nemacheilidae (420)
Neolissocheilus Cyprinidae (2420)
Ogilbia Bythitidae (107)
Oxeleotris Eleotridae (155)
Paralepidocephalus Cobitidae (177)
Phreatichthys Cyprinidae (2420) HR, monotypic
Poecilia Poeciliidae (304)
Poropuntius Cyprinidae (2420)
Table 3 Contd..

Table 3 Contd..
Protocobitis Cobitidae (177) HR, TLO, monotypic
Satan Ictaluridae (46) HR, monotypic
Silvinichthys Trichomycteridae (201)
Speoplatyrhinus Amblyopsidae (6) HR, TLO, monotypic
Stygichthys Characidae (962) HR, monotypic
Sundoreonectes Nemacheilidae (420)
Troglocobitis Cobitidae (177) HR, TLO, monotypic
Troglocyclocheilus Cyprinidae (2420) HR, TLO, monotypic
Trogloglanis Ictaluridae (46) HR, monotypic
Typhliasina Bythitidae (107) HR, monotypic
Typhlichthys Amblyopsidae (6) HR, monotypic
Tyhlobarbus Cyprinidae (2420) HR, TLO, monotypic
Tyhlogarra Cyprinidae (2420) HR, monotypic
Uegitglanis Clariidae (90) HR, monotypic
38 genera and species 20 genera and species HR
32% of genera, 13% of species
Table 4 Genera with two or more hypogean species. The figure in brackets is
the total number of species in the family (from Nelson 2006). HR = hypogean
restricted, TLO = known from the type locality only
Genus Family No. of hypogean species Notes
Amblyopsis Amblyopsidae (6) 2 All HR
Ancistrus Loricariidae (684) 3
Astroblepus Astroblepidae (54) 2
Astyanax Characidae (962) 2
Cottus Cottidae (275) 2
Garra Cyprinidae (2420) 2
Horaglanis Clariidae (90) 2 All HR
Ituglanis Trichomycteridae (201) 5
Luciogobius Gobiidae (1950) 2
Lucifuga Bythitidae (107) 6 All HR
Monopterus Synbranchidae (17) 3
Ophisternon Synbranchidae (17) 2
Oreonectes Nemacheilidae (420) 2
Paracobitis Nemacheilidae (420) 4 All TLO
Phreatobius Siluriformes in. sed. 3
Pimelodella Heptapteridae (175) 2
Prietella Ictaluridae (46) 2 All HR
Pterocryptis Siluridae (97) 3 All TLO
Rhamdia Heptapteridae (175) 8
Rhamdiopsis Heptapteridae (175) 3
Table 4 Contd..

Table 4 Contd..
Schistura Namacheilidae (420) 8

Sinocyclocheilus Cyprinidae (2420) 26
Trichomycterus Trichomycteridae (201) 8
Triplophysa Nemacheilidae (420) 8
Typhleotris Eleotridae (155) 3 All HR
25 genera 113 species
Table 5 The species and genera which are hypogean restricted

Family Genus and species
Cyprinidae Stygichthys typhlops

Caecobarbus geertsii
Barbopsis devecchii
Phreatichthys andruzzii
Iranocycpris typhlops
Typhlogarra widdowsoni
Caecocypris basimi
Typhlobarbus nudiventris
Troglocyclocheilus khammouanensis
Nemacheilidae Troglocobitis starostini
Balitoridae Cryptotora thamicola
Cobitidae Protocobitis typhlops
Ictaluridae Trogloglanis pattersoni
Satan eurystomus
Prietella phreatophila
Prietella lundbergi
Clariidae Uegitglanis zammaranoi
Horaglanis krishnai
Horaglanis alikunhii
Amblyopsidae Amblyopsis spelaea
Amblyopsis rosae
Typhlichthys subterraneus
Speoplatyrhinus poulsoni
Bythitidae Lucifuga subterraneus
Lucifuga dentatus
Lucifuga spelaeotes
Lucifuga simile
Lucifuga teresinarum
Lucifuga lucayana
Typhliasina pearsei
Gobiidae Caecogobius cryptophthalmus
Table 5 Contd..

Table 5 Contd..
Eleotridae Typhleotris madagascariensis

Typhleotris pauliani
Typhleotris undescribed
Milyeringa veritas
10 families 25 genera and 35 species
Table 6 The number of hypogean species of fishes in countries where they are
known. It is highly significant that half of the total known species are from only
three countries, and that over a quarter are from China. It is also significant that all
Chinese species have been discovered since 1979
Country Number of species % Proportion
China 42 28
Brazil 23 15 50%
Mexico 11 7
Thailand 9 6
India 8 5
USA 8 5
Venezuela 4 3
Cuba 4 3
Madagascar 4 3
Somalia 3 2
Iran 2 1
Iraq 2 1
Oman 2 1
Laos 2 1
Vietnam 2 1
Bolivia 2 1
Colombia 2 1
Bahamas 2 1 50%
Australia 2 1
Japan 2 1
Democratis Rep. Congo 1 0.5
Malaysia 1 0.5
Turkmenistan 1 0.5
Namibia 1 0.5
Trinidad 1 0.5
Bolivia 1 0.5
Argentina 1 0.5
Ecuador 1 0.5
Peru 1 0.5
Table 6 Contd..

Table 6 Contd..
Galapagos 1 0.5
Philippines 1 0.5
Papua New Guinea 1 0.5
Guam 1 0.5
Indonesia 1 0.5
34 countries 150
Table 7 The genera, and number of species per genus, for the three countries
with the largest number of known species. HR = hypogean restricted.
Country Genera No. of species Notes
China Sinocyclocheilus 26
Paracobitis 3
Oreonectes 2
Triplophysa 8
Protocobitis 1 HR
Paralepidochephalus 1
Pterocryptis 1
7 genera 42 species
Brazil Stygichthys 1 HR
Pimelodella 2
Rhamdia 1
Rhamdiopsis 3
Aspidoras 1
Trichomycterus 3
Ituglanis 5
Glaphyropoma 1
Copionodon 1
Ancistrus 2
Phreatobius 2 All HR
Eigenmannia 1
12 genera 23 species
Mexico Astyanax 2
Prietella 2 All HR
Rhamdia 4
Typhliasina 1 HR
Poecilia 1
Ophisternon 1
6 genera 11 species

GEOGRAPHIC DISTRIBUTION
Subterranean fishes are known from 34 countries. Three of these countries
(China, Brazil and Mexico) have very large areas of karstic rocks, mainly in
limestones but also in other lithologies such as quartzite, and many caves.
They also have very active cave exploration and study. A combination of
these factors means that these three countries account for half of all known
subterranean fishes with China having 28%, Brazil 15% and Mexico 7% of
all known species (Table 6). The remaining 50% are found in 31 countries,
with 14 countries with one species, 10 with two species, one with three
species, three with four species, two with eight species and one with nine
species. The countries with three or more species (10 in all) account for
76% of all known species. In most of these countries cave exploration is
ongoing, and advancing, and more species are expected in all of them. Table
7 lists the numbers of species in each of genera known from China, Brazil
and Mexico. Most genera have only a few known subterranean species.
Exceptions are Sinocyclocheilus with twenty-six and Ituglanis with five.
Acknowledgements
Figures 1 and 2 were developed for me by Professor Malcolm Elliott of
the Freshwater Biological Association and I am most grateful for this very
valuable assistance.
References
Helfman, G., B.B. Collette, D.E. Facey and B.W. Bowen. 2009. The diversity of Fishes.
Biology, Evolution and Ecology. Wiley-Blackwell, Oxford, 2nd Edition.
Nelson, J.S. 2006. Fishes of the World. John Wiley & Sons, Inc., Hoboken. 4th
Edition.
Porter, M.L., K. Dittmar and M. Perez-Losada. 2007. How long does evolution of the
troglomorphic form take? Estimating divergence times in Astyanax mexicanus.
In: Time in Karst, A. Kranjc, F. Gabrovsek, D.C. Culver and I.D. Sasowsky (eds.).
Karst Water Institute Special Publication 12. Karst Waters Institute, Leesburg,
pp. 173-182.
Proudlove, G.S. 2004. Pisces (Fishes). In: Encyclopaedia of Caves and Karst Science. J.
Gunn (ed.). Fitzroy Dearborn, New York, pp. 593-595.
Proudlove, G.S. 2006. Subterranean Fishes of the World. An Account of the Subterranean
(Hypogean) Fishes Described up to 2003 with a Bibliography 1541–2004. Includes Note
added in press with details 2003-2006. International Society for Subterranean
Biology, Moulis.
Romero, A. (ed.). 2001. The biology of hypogean fishes. Environmental Biology of
Fishes 62: 1-364.
Romero, A. 2004. Pisces (Fishes): Amblyopsidae. In: Encyclopaedia of Caves and Karst
Science. J. Gunn (ed.). Fitzroy Dearborn, New York, pp. 595-597.

Slechtova, V., J. Bohlen and H.H. Tan. 2007. Families of Cobitoidea (Teleostei:
Cypriniformes) as revealed from nuclear genetic data and the position of
the mysterious genera Barbucca, Psilorhynchus, Serpenticobitis and Vaillantella.
Molecular Phylogenetics and Evolution 44: 1358-1365.
Tang, Q., H. Liu, , R.L. Mayden and B. Xiong. 2006. Comparison of evolutionary
rates in the mitochondrial DNA cytochrome b gene and control region and
their implications for phylogeny of the Cobitoidea (Teleostei: Cypriniformes).
Thines, G. 1969. L’evolution regressive des poissons cavernicoles et abyssaux. Masson
et Cie, Paris.
Trajano, E. 2007. The challenge of estimating the age of subterranean lineages:
Examples from Brazil. In: Time in Karst, A. Kranjc, F. Gabrovsek, D.C. Culver
and I.D. Sasowsky (eds.). Karst Water Institute Special Publication 12. Karst
Waters Institute, Leesburg, pp. 191-198.
Weber, A. 2000. Fish and amphibia. In: Ecosystems of the World 30. Subterranean
Ecosystems, H. Wilkens, D.C. Culver and W.F. Humphreys (eds.). Elsevier,
Amsterdam, pp. 109-132.
Weber, A., G.S. Proudlove and T.T. Nalbant. 1998. Morphology, systematic diversity,
distribution and ecology of stygobitic fishes. In: Encyclopaedia Biospeologica, C.
Juberthie and V. Decu (eds.). Societe de Biospeologie, Moulis and Bucarest, pp.
1179-1190.
Wilkens, H. 2005. Fish. In: Encyclopedia of Caves, D.C. Culver and W.B. White (eds.).
Elsevier, Burlington, pp. 241-251.

CHECKLIST OF SUBTERRANEAN FISHES OF THE

WORLD TO DECEMBER 2009
Checklist of subterranean fishes as discovered to December 2009. Generic
placement and suprageneric classification follow Nelson (2006), with the
following exceptions: Nemacheilidae, Balitoridae and Cobitidae follow
Tang, Liu, Mayden and Xiong (2006) and Slechtova, Bohlen and Tan (2007);
Amblyopsiformes follows Murray and Wilson (1999). Currently accepted name
combinations are numbered and are followed by their original author and
date and by revisers (if any). Other names which have been used (synonyms,
previous combinations and misidentifications) are indented and follow
immediately after the current name for the species. Author names and dates
in square brackets are for species which have not been formally described, and
identify the source with most information about that particular species.
Note: The analyses in this chapter were made on the 150 species known to
me at the time the chapter was prepared. Since then further species have been
described, or discovered in the literature. This checklist contains all troglobiotic
fish species known at December 2009. The additional 14 species over those
analysed in the text do not in any way alter the conclusions of the chapter.
Additional notation is as follows:
MG
Monotypic subterranean-restricted genus.
MuG
Subterranean-restricted genus with two or more species.
TLO
Species known only from its type locality.
TMP
Troglomorphic (and probably troglobiotic) hypogean population of a
previously known, and named, epigean fish. These species exhibit some, or
many, of the characteristics termed troglomorphy, i.e. they are depigmented
and may have a reduction in eye size. Further work is required accurately to
asses their status.
DRC = Democratic Republic of Congo (formerly Zaire), Madagasc. =
Madagascar, PNG = Papua New Guinea, UAE = United Arab Emirates
Order Characiformes
Family Characidae
1. Astyanax jordani (Hubbs and Innes, 1936) Buckup, 2003 México
Anoptichthys jordani Hubbs and Innes, 1936
Anoptichthys antrobius Alvarez, 1946
Anoptichthys hubbsi Alvarez, 1947

Astyanax fasciatus (Cuvier, 1819) cave form

Astyanax mexicanus (De Filippi, 1853) cave form
2. Astyanax undescribed species [Espinasa, Rivas and
Espinosa Perez, 2001] TLO México
3. Stygichthys typhlops Brittan and Bohlke, 1965 MG Brazil
Order Cypriniformes
Family Cyprinidae
4. Caecobarbus geertsii Boulenger, 1921 MG DRC
MG
5. Phreatichthys andruzzii Vinciguerra, 1924 Somalia
6. Barbopsis devecchii Di Caporiacco, 1926 MG Somalia
Eilichthys microphthalmus Pellegrin, 1929
Barbopsis stefaninii Gianferrari, 1930
Zaccarinia stefaninii Gianferrari, 1934
7. Iranocypris typhlops Bruun and Kaiser, 1944 MG TLO Iran
MG
8. Typhlogarra widdowsoni Trewavas, 1955 Iraq
9. Caecocypris basimi Banister and Bunni, 1980 MG TLO Iraq
10. Garra barreimiae Fowler and Steinitz, 1956 TMP Oman
11. Garra barreimiae wurayahi Khalaf, 2009 TMP UAE
TLO
12. Garra dunsirei Banister, 1987 Oman
13. Typhlobarbus nudiventris Chu and Chen, 1982 MG TLO China
14. Sinocyclocheilus guilinensis Ji, 1985 China
15. Sinocyclocheilus anatirostris Lin and Luo, 1986 China
16. Sinocyclocheilus bicornutus Wang and Liao, 1997 China
17. Sinocyclocheilus anophthalmus Chen, Chu, Luo and
Wu 1988 TLO China
18. Sinocyclocheilus cyphotergous (Dai, 1988) Kottelat and
Brehier, 1999 TLO China
Gibbibarbus cyphotergous Dai, 1988
19. Sinocyclocheilus microphthalmus Li, 1989 TLO China
20. Sinocyclocheilus angularis Zheng and Wang, 1990 TLO China
21. Sinocyclocheilus altishoulderus (Li and Lan, 1992)
Wang and Chen, 1998 TLO? China
Anchicyclocheilus altishoulderus Li and Lan, 1992
22. Sinocyclocheilus hyalinus Chen and Yang, 1993 TLO China

23. Sinocyclocheilus rhinocerous Li and Tao, 1994 TLO? China

24. Sinocyclocheilus furcodorsalis Chen, Yang and Lan,
1997 TLO? China
25. Sinocyclocheilus guangxiensis Zhou and Li, 1998 China
26. Sinocyclocheilus albeoguttatus Zhou and Li, 1998 China
27. Sinocyclocheilus lingyunensis Li, Xiao, Zan, Luo and
Li, 2000 TLO China
28. Sinocyclocheilus jiuxuensis Li and Lan, 2003 China
29. Sinocyclocheilus tianeensis Li, Xiao and Luo, 2003 China
30. Sinocyclocheilus tianlinensis Zhou, Zhang and He, 2004 China
31. Sinocyclocheilus tileihornes Mao, Lu and Li, 2003 China
32. Sinocyclocheilus hugeibarbus Li and Ran, 2003 China
33. Sinocyclocheilus xunlensis Lan, Zhao and Zhang, 2004 China
34. Sinocyclocheilus liboensis Li, Chen and Ran, 2004 China
35. Sinocyclocheilus donglanensis Zhao, Watanabe and
Zhang, 2006 China
36. Sinocyclocheilus halfibindus (Li and Lan, 1992)
Wang and Chen, 1998 TLO? China
Anchicyclocheilus halfibindus Li and Lan, 1992
37. Sinocyclocheilus aquihornes Li and Yang, 2007 China
38. Sinocyclocheilus brevibarbatus Zhao, Lan and
Zhang, 2008 TLO China
39. Sinocyclocheilus broadihornes Li and Mao, 2007 China
40. Sinocyclocheilus huaningensis Li, 1998 China
41. Sinocyclocheilus longifinus Li, 1996 China
42. Poropuntius speleops (Roberts, 1991) Kottelat and
Brehier, 1999 Thailand
Barbus speleops Roberts, 1991
43. Troglocyclocheilus khammouanensis Kottelat and
Brehier, 1999 MG TLO Laos
44. Neolissochilus subterraneus Vidthayanon and
Kottelat, 2003 TLO Thailand
45. Longanalus macrochirous Li, Ran and Chen, 2006 China
Family Nemacheilidae
46. Paracobitis smithi (Greenwood, 1976) Nalbant and
Bianco, 1998 TLO Iran

Nemacheilus smithi Greenwood, 1976

Noemacheilus smithi Greenwood, 1976
47. Paracobitis posterodarsalus Li, Ran and Chen, 2006 TLO China
TLO
48. Paracobitis maolanensis Li, Ran and Chen, 2006 China
49. Nemacheilus troglocataractus Kottelat and Gery, 1989 TLO Thailand
50. Oreonectes anophthalmus Zheng, 1981 TLO China
51. Oreonectes furcocaudalis Zhu and Cao, 1987 China
52. Oreonectes translucens Zhang, Zhao and Zhang, 2006 China
53. Oreonectes microphthalmus Du, Chen and Yang, 2008 China
54. Oreonectes retrodorsalis Lan, Yang and Chen, 1995 China
55. Triplophysa gejiuensis (Chu and Chen, 1979) Chen,
Yang and Xu, 1992 TLO China
Noemacheilus gejiuensis Chu and Chen, 1979
56. Triplophysa xiangxiensis (Yang, Yuan and Liao,1986)
Chen, Yang and Xu, 1992 TLO Noemacheilus xiangxiensis
Yang, Yuan and Liao, 1986 China
TLO
57. Triplophysa yunnanensis Yang, 1990 China
58. Triplophysa shilinensis Chen and Yang, 1992 TLO China
59. Triplophysa nandanensis Lan, Yang and Chen, 1995 China
60. Triplophysa aluensis Li and Zhu, 2000 China
61. Triplophysa longibarbata (Chen, Yang, Sket and
Aljancic, 1998) Du, Chen and Yang, 2008 TLO
Paracobitis longibarbatus Chen, Yang, Sket and
Aljancic, 1998 China
Nemacheilus liboensis Ran, 2000 and Chen,
2000 nomen nudum
Oreonectes liboensis Wang, 1991 nomen nudum
62. Triplophysa tianeensis Chen, Cui and Yang, 2004 China
63. Triplophysa rosa Chen and Yang, 2005 China
Linemacheilus wulong Wu, He and Li, 2005
nomen nudum
Heminoemacheilus wulongensis Wu, He and
Li, 2007 nomen nudum
64. Triplophysa qiubeiensis Li, Yang, Chen, Tao, Qi and
Han, 2008 China
65. Triplophysa longipectoralis Zheng, Du, Chen and
Yang, 2009 TLO China

66. Triplophysa sinensis [Chen, Li and Yang, 2008] China

TLO
67. Schistura sijuensis (Menon, 1987) Kottelat, 1990 India
Noemacheilus sijuensis Menon, 1987
68. Schistura oedipus (Kottelat, 1988) Kottelat, 1990 Thailand
Nemacheilus oedipus Kottelat, 1988
69. Schistura jarutanini Kottelat, 1990 TLO Thailand
70. Schistura kaysonei Vidthayanon and Jaruthanin, 2002 TLO Laos
71. Schistura spiesi Vidthayanon and Kottelat, 2003 TLO Thailand
72. Schistura deansmarti Vidthayanon and Kottelat, 2003 TLO Thailand
73. Schistura spekuli Kottelat, 2004 Vietnam
74. Schistura papulifera Kottelat, Harries and Proudlove,
2007 India
TLO
75. Sundoreonectes tiomanensis Kottelat, 1990 Malaysia
76. Indoreonectes evezardi (Day, 1872) Kottelat, 1990 TLO TMP India
Noemacheilus evezardi Day, 1872
77. Troglocobitis starostini (Parin, 1983) MG TLO Turkmenia
Noemacheilus starostini Parin, 1983
78. Heminoemacheilus hyalinus Lan, Yang and Chen, 1996 China
Family Balitoridae
79. Cryptotora thamicola (Kottelat, 1988) Kottelat, 1998 MG Thailand
Homaloptera thamicola Kottelat, 1988
Family Cobitidae
80. Protocobitis typhlops Yang and Chen, 1993 MG TLO China
81. Protocobitis polylepis Zhu, Lu, Yang and Zhang, 2008 China
Order Siluriformes
Family Ictaluridae
82. Trogloglanis pattersoni Eigenmann, 1919 MG USA
83. Satan eurystomus Hubbs and Bailey, 1947 MG USA
84. Prietella phreatophila Carranza, 1954 MuG México
85. Prietella lundbergi Walsh and Gilbert, 1995 MuG México

Family Siluridae
86. Pterocryptis cucphuongensis (Mai, 1978) Ng and
Kottelat, 1998 TLO Vietnam
Silurus cucphuongensis Mai, 1978
87. Pterocryptis buccata Ng and Kottelat, 1998 TLO Thailand
88. Pterocryptis undescribed species [Clarke pers.
comm. 2001] TLO China
Family Clariidae
89. Uegitglanis zammaranoi Gianferrari, 1923 MG Somalia
90. Clarias cavernicola Trewavas, 1936 TLO Namibia
MuG
91. Horaglanis krishnai Menon, 1950 India
92. Horaglanis alikunhii Babu and Nayar, 2004 MuG India
Family Heptapteridae
93. Pimelodella kronei (Miranda Ribeiro, 1907) Pavan, 1946 Brazil
Typhlobagrus kronei Miranda Ribeiro, 1907
Pimelodella lateristriga var. kronei Haseman, 1911
Caecorhamdella brasiliensis Borodin, 1927
94. Pimelodella spelaea Trajano, Reis and Bichuette, 2004 Brazil
95. Rhamdia quelen urichi (Norman, 1926) Mees, 1974 TLO Trinidad
Caecorhamdia urichi Norman, 1926
Caecorhamdella urichi (Norman, 1926) Hubbs,
1938 lapsus calami
96. Rhamdia laticauda typhla Greenfield, Greenfield, and
Woods 1982 TLO Belize
97. Rhamdia reddelli Miller, 1984 México
98. Rhamdia zongolicensis Wilkens, 1993 TLO México
99. Rhamdia macuspanensis Weber and Wilkens, 1998 TLO México
100. Rhamdia laluchensis Weber, Allegrucci and Sbordoni,
2003 TLO México
101. Rhamdia guasarensis DoNascimiento, Provenzano and
Lundberg, 2004 Venezuela
102. Rhamdia enfurnada Bichuette and Trajano, 2005 Brazil
103. Rhamdia undescribed species [Cordeiro pers. comm.] Brazil

104. Rhamdiopsis undescribed species 1

[Trajano and Bichuette, 2007] Brazil
Family Callichthyidae
107. Aspidoras sp.
Family Trichomycteridae
108. Trichomycterus chaberti Durand, 1968 TLO Bolivia
109. Trichomycterus itacarambiensis Trajano and
De Pinna, 1996 TLO Brazil
110. Trichomycterus spelaeus DoNascimiento, Villarreal
and Provenzano, 2001 TLO Venezuela
111. Trichomycterus sandovali Ardila Rodriguez, 2006 Colombia
112. Trichomycterus santanderensis Castellanos-Morales,
2007 TLO Colombia
TLO
113. Trichomycterus uisae Castellanos-Morales, 2008 Colombia
114. Trichomycterus undescribed species 1
115. Trichomycterus undescribed species 2
116. Trichomycterus undescribed species 3 [Schultz, 1949] TLO Venezuela
TLO
117. Ituglanis passensis Fernandez and Bichuette, 2002 Brazil
118. Ituglanis bambui Bichuette and Trajano, 2004 Brazil
119. Ituglanis epikarsticus Bichuette and Trajano, 2004 Brazil
120. Ituglanis ramiroi Bichuette and Trajano, 2004 Brazil
121. Ituglanis mambai Bichuette and Trajano, 2008 Brazil
122. Ituglanis undescribed species
[Trajano pers. comm. 11.2009] Brazil
123. Glaphyropoma spinosum Bichuette, de Pinna
and Trajano, 2008 TLO Brazil

124. Copionodon undescribed species

125. Silvinichthys bortayro Fernandez and De Pinna, 2005 Argentina
Family Loricariidae
126. Ancistrus cryptophthalmus Reis, 1987 Brazil
127. Ancistrus galani Perez and Viloria, 1994 TLO Venezuela
128. Ancistrus formoso Sabino and Trajano, 1997 Brazil
Family Astroblepidae
129. Astroblepus pholeter Collette, 1962 TLO Ecuador
TLO
130. Astroblepus riberae Cardona and Guerao, 1994 Peru
Siluriformes incertae sedis

131. Phreatobius cisternarum Goeldi, 1905 MuG Brazil
132. Phreatobius dracunculus Shibatta, Muriel-Cunha
and De Pinna, 2007 MuG Brazil
133. Phreatobius sanguijuela Fernandez, Saucedo,
Carvajal and Schaefer, 2007 MuG Bolivia
Order Gymnotiformes
Family Sternopygidae
134. Eigenmannia vicentespelaea Triques, 1996 TLO Brazil
Order Amblyopsiformes
Family Amblyopsidae
135. Amblyopsis spelaea DeKay, 1842 MuG USA
136. Amblyopsis rosae (Eigenmann, 1898) Woods and
Inger, 1957 MuG USA
Typhlichthys rosae Eigenmann, 1898
Troglichthys rosae (Eigenmann, 1898)
Eigenmann, 1899
Amblyopsis rosae whitae Romero, 1998
nomen nudum

137. Typhlichthys subterraneus Girard, 1859 MG USA

Typhlichthys osborni Eigenmann, 1905
Typhlichthys wyandotte Eigenmann, 1905
Typhlichthys eigenmanni Charlton, 1933
Typhlichthys eigenmanni Hubbs, 1938 nomen nudum
138. Speoplatyrhinus poulsoni Cooper and Kuehne,
1974 MG TLO USA
Order Ophidiiformes
Family Bythitidae
139. Lucifuga subterranea Poey, 1858 MuG Cuba
Lucifuga subterraneus Poey, 1858
Lucifuga teresinarum Diaz Perez, 1988
140. Lucifuga dentata Poey, 1858 MuG Cuba
Lucifuga dentatus Poey, 1858
Stygicola dentatus Gill, 1863
141. Lucifuga simile Nalbant, 1981 MuG Cuba
MuG
142. Lucifuga spelaeotes Cohen and Robins, 1970 Bahamas
143. Lucifuga lucayana Moller, Schwarzhans, Iliffe and
Nielsen, 2006 MuG Bahamas
144. Typhliasina pearsei (Hubbs, 1938) Whitley, 1951 MG México
Typhlias pearsei Hubbs, 1938
Ogilbia pearsei (Hubbs, 1938) Cohen and
Nielson, 1978
145. Ogilbia galapagosensis (Poll and Leleup, 1965)
Cohen and Nielson, 1978 Galapagos
Caecogilbia galapagosensis Poll and Leleup, 1965
Order Cyprinodontiformes
Family Poeciliidae
146. Poecilia mexicana Steindachner, 1863 TMP México
Poecilia sphenops Valenciennes in Cuvier and
Valenciennes, 1846
Mollienisia sphenops (Valenciennes in Cuvier and
Valenciennes, 1846) Regan, 1913

Order Synbranchiformes
Family Synbranchidae
147. Ophisternon infernale (Hubbs, 1938) Rosen and
Greenwood, 1976 México
Pluto infernalis Hubbs, 1938
Furmastix infernalis (Hubbs, 1938) Whitley, 1957
Synbranchus infernalis (Hubbs, 1938) Mees, 1962
148. Ophisternon candidum (Mees, 1962) Rosen and
Greenwood, 1976 Australia
Anomatophasma candidum Mees, 1962
149. Monopterus eapeni Talwar in Talwar and Jhingran, 1992 TLO India
Monopterus indicus Eapen, 1963
Monopterus “indicus” (Eapen, 1963) Rosen
and Greenwood, 1976
150. Monopterus roseni Bailey and Gans, 1998 TLO India
TLO
151. Monopterus digressus Gopi, 2002 India
Order Scorpaeniformes
Family Cottidae
152. Cottus carolinae species group
[Burr, Adams, Krejca and Warren, 2001] TMP USA
153. Cottus “bairdii-cognatus” complex
[Espinasa and Jeffery, 2003] TLO TMP USA
Order Perciformes
Family Gobiidae
154. Luciogobius pallidus Regan, 1940 Japan
155. Luciogobius albus Regan, 1940 Japan
156. Caecogobius cryptophthalmus Berti and Ercolini,
1991 MG TLO Philippines
TLO
157. Glossogobius ankaranensis Banister, 1994 Madagasc
158. Bostrychus microphthalmus Hoese and Kottelat, 2005 Indonesia

Family Eleotridae
159. Typhleotris madagascariensis Petit, 1933 MuG Madagasc.
160. Typhleotris pauliani Arnoult, 1959 MuG Madagasc.
161. Typhleotris undescribed species
[Sparks and Stiassny, 2004] MuG Madagasc.
162. Milyeringa veritas Whitley, 1945 MG Australia
163. Oxyeleotris caeca Allen, 1996 TLO PNG
164. Eleotris undescribed species
[Tibbatts pers. Comm. 2006] Guam

CHAPTER 3
Conservation of Subterranean Fishes
Maria Elina Bichuette1 and Eleonora Trajano2
1
Departamento de Ecologia e Biologia Evolutiva, Universidade Federal
de São Carlos, São Carlos, SP, Brazil. E-mail: bichuette@uol.com.br
2
Departamento de Zoologia, Instituto de Biociências, Universidade de
São Paulo, São Paulo, SP, Brazil. E-mail: etrajano@usp.br
INTRODUCTION
Subterranean ecosystems pose special problems for conservation due
to their intrinsic fragility and the distinctive features of subterranean
communities, including a high degree of endemism and morphological,
ecological and behavioral specialization. Such fragility is a consequence of
the relatively low biological diversity of subterranean ecosystems and their
general dependence on nutrients imported from the surface, making them
strongly influenced by disturbances in epigean habitats. Furthermore,
many troglobites have small geographic ranges and low population
densities, resulting in small population sizes, K-selected life histories, thus
population turnover is slow and their ability to recover from population
losses is decreased compared to epigean related taxa, and they may be
highly susceptible to environmental fluctuations (Trajano 2000).
Around 98% of unfrozen freshwater in the world is subterranean
and Juberthie (2004) claims at least 50% is inhabited by troglobitic
and troglophilic species. These aquatic habitats include karst aquifers
(including the epikarst zone), phreatic aquifers in valleys in many kinds
of sediments and rocks, interstitial and hyporheic habitats, and base-level
streams. Troglobitic fishes have been found in all these kinds of habitats
(see, for instance, Trajano and Bichuette, this volume).

1. THREATS AND VULNERABILITY OF AQUATIC

SUBTERRANEAN HABITATS AND SPECIES
Aquatic subterranean habitats are often disturbed by human actions,
such as habitat degradation, hydrological manipulations, environmental
pollution, overexploitation and impacts of introduced fauna (Proudlove,
2001). Main threats to subterranean fishes throughout the world are
discussed and exemplified below (see also Sullivan 1956, Poulson 1968,
Culver 1986, Proudlove 2001, Juberthie 2004, Elliott 2005).
1.1 Habitat Destruction: Quarrying and Open Pit Mining

The ultimate threat for subterranean faunas as a whole, as well as for
epigean communities, is the removal of rock layers for limestone and ore
(iron, gold etc.) exploitation. Removal of gravel from stream-beds may
destroy the interstitial and stream habitats as well. As far as we know,
there is no reported instance of extinction of troglobitic fishes caused by
this extreme threat, but it is possible that populations have already been
destroyed before being noticed. This is particularly true for species living
in less obvious, non-cave habitats, whose existence has been recently
acknowledged, such as troglomorphic catfishes, genus Phreatobius, from
underground waters in the Amazon basin.
1.2 Habitat Changes: Use of Water
1.2.1 Dams and Reservoirs

Building of dams and reservoirs for hydro-electrical plants and water
storage for human use has profound effects over the structure and
physical-chemical characteristic of aquatic habitats, changing from lotic
to lentic conditions: velocity of water current decreases, water depth
increases and soft bottom predominates, dissolved oxygen may drop to
quite low levels. Due to the elevation of the regional water table, caves
in the surrounding areas may also be affected. Increased sedimentation
may be harmful for aquatic organisms with gills, like fishes, making it
difficult for rock-dwellers to find proper substrates, low oxygen levels will
exclude many organisms, and so on. As a consequence, stream-dwellers
may disappear.
The construction of dams in the Green River in Kentucky, U.S.A.,
raised the water level in the Mammoth Cave, and this may have affected
the amblyopsid Amblyopsis spelaea, either because of silt deposition on
previously sand bottomed areas and/or due to the reduction of particulate
organic matter, a major food source for aquatic communities (Proudlove
2001).
Maria Elina Bichuette and Eleonora Trajano 67
1.2.2 Over-pumping of Subterranean Aquifers

Withdrawal of large amounts of water for human use, for domestic,
agricultural or industrial purposes, is one of the most serious threats for
aquatic troglobites throughout the world. When recharge rates become
insufficient for replenishing the aquifer, the water table gradually lowers,
water quantity and quality declines and habitat for subterranean organisms
is lost. In the extreme, it becomes an instance of habitat destruction
(see Section 1.1).
Proudlove (2001) mentions several cases of actual or potential
habitat loss for troglobitic fishes due to over-pumping, including the
North American ictalurid catfishes Satan eurystomus and Trogloglanis
pattersoni, endemic to the San Antonio pools of the Edwards aquifer,
one of the main water sources in Texas, U.S.A., and Prietella phreatophila,
from Mexico; the African clariids Clarias cavernicola, from Namibia, and
Uegitglanis zammaranoi, from Somalia, as well as the cyprinids Barbopsis
devecchii and Phreatichthys andruzzi; the Australian Ophisternon candidum
(Synbranchidae) and Milyeringa veritas (Eleotridae).
Over-pumping also threatens non-troglomorphic fish in caves. In Devil’s
Sinkhole, Nevada, U.S.A., the Devil´s Hole pupfish, Cyprinodon diabolis, is
endangered by over-pumping of the regional aquifer, which has reduced
this cave fish to a tiny population at the bottom of a sinkhole (Elliott 1981).
Other North American pupfishes, totally dependent on karst aquifers,
are also endangered, as is the case of Cyprinodon elegans, from Comanche
Springs, and C. bovines, from Leon Springs. (Elliott, op. cit.).
Moreira and colaborators (In Press) called the attention for the highly
specialized Brazilian phreatic fish, Stygichthys typhlops, from eastern Brazil
(Jaíba karst area). Information provided by town inhabitants, well-drilling
operators, local farm owners and workers indicates that S. typhlops was
quite common and widespread in the Jaíba area, at least up to the early
1990’s, occurring in springs, shallow hand dug wells, and also coming
out during drilling of artesian wells. Almost all of these reports referred
to sightings at least ten years ago. Most of the wells referred to are now
dry, and in one case where the well was not dry it was clogged with clay
preventing (or severely decreasing) the influx of water. In the past twenty
years, the Jaíba region has undergone extensive irrigation projects for
agriculture, mostly drawing from subterranean water. In the first half of
the 1980’s, the intensity and frequency of utilization of this resource was
low, probably one-tenth the recharge capacity of the aquifer (Silva 1984).
However, there has been a marked decrease in water table levels, with
several wells and springs reportedly dried out in the past few years. This
has promoted a shift from traditional wells or open springs to artesian
wells as source of water in recent years.

This situation poses a major concern for the future of S. typhlops. The
lowering of the water level may result in significant reduction of available
habitat and reduce the input of nutrients from the surface, what could
lead to a rapid population decline. In 2004, S. typhlops was included in the
Brazilian official list of threatened species (MMA 2004), which hopefully
will help preserving this species.
Further lowering of two meters in the water level will result in drying
of the two wells where we collected specimens. Unless deeper access to
the fish habitat is found, collecting of additional material for future studies
may become impossible in a few years.
1.3 Pollution
Pollution includes eutrophication and contamination from agricultural and
industrial runoff (e.g. fertilizers, pesticides and heavy metals – Proudlove,
2001). Underground waters generally are very sensitive to pollution
because pollutants may rapidly spread into subterranean aquifers due to
their limited filter capacity. Hence, the detection and removal of pollutants
from groundwater is especially difficult.
1.3.1 Organic Pollution: Eutrophication due to Garbage

Dumping, Inadequate Sewage Treatment and
Introduction of Fauna
Organic pollution of subterranean habitats may be caused by legal and
illegal thrash and garbage dumping and infiltration of wastewater from
faulty sewage (Culver 1986, Bolner and Tardy 1989). Making waste
“disappear” by dumping it into caves and sinkholes has been a widespread
practice, and it is still practiced in many places throughout the world.
This illusion is especially dangerous due to the facility with which the
pollutants disperse and their general invisible nature after dispersed.
In many studies (e.g. Cruickshank et al. 1980, Varela et al. 1980), an
evaluation of water quality from samples collected in caves showed
chemical and bacteriological contamination affecting cave fish populations.
Proudlove (2001) mentions cases of contamination by high levels of
nitrogen compounds affecting subterranean fishes. The proliferation of
bacteria may also affect the density of some aquatic organisms, causing
an imbalance on their population parameters. The skin of fishes is
protected by mucus and it is relatively thin, representing a main entrance
for pathogens, and any injury on fish bodies allows the proliferation of
diseases, for example, the Lucifuga species from Cuba.
Specialized troglobitic fishes exhibit a better feeding performance when
density of prey is low than troglophilic and epigean close relatives, the

opposite being true when prey density is high. This has been shown for
the amblyopsids Typhlichthys subterraneus, Amblyopsis rosae and A. spelaea,
when compared to Forbesichthys agassizi and C. cornuta (Poulson 1963). As a
consequence, eutrophication favors the invasion of naturally food-limited
subterranean habitats by epigean species, which may displace troglobites
by successfully competing with them.
The impacts caused by the introduction, either accidental or deliberate,
of alien animals may be considered a special case of organic pollution.
Proudlove (2001) cites some examples of epigean fishes introduced in
cave areas, where they may threat troglobitic species due to competition
(Clarias gariepinus versus C. cavernicola, in Namibia) and/or predation
(some marine species versus Lucifuga spp., in Cuba), or hybridization
(Astyanax mexicanus versus troglobitic Astyanax, Mexico).
In Brazil, carps (Cyprinus carpio) introduced in the epigean drainage of
Intervales State Park, São Paulo State, invaded caves and are presently
threatening, probably due to competition and/or habitat change (siltation),
cave decapods, genus Aegla. In the neighbor Alto Ribeira Tourist State Park,
these decapods are apparently endangered by the expansion of freshwater
shrimps, genus Macrobrachium (E. Trajano et al., pers. obs.).
1.3.2 Chemical Pollution

Chemical pollution by heavy metals, fertilizers and pesticides is a major
environmental problem worldwide. Vesper (2005) summarizes the problem
of contamination of cave waters by heavy metals. There are probably
many aquatic troglobites threatened by chemical pollution throughout the
world, but few cases have been well-documented.
The blind catfish, Pimelodella kronei, occurs in several caves at both sides
of the Betari stream, in the Upper Ribeira karst area, SE Brazil: the Areias
cave system (Areias de Cima – type-locality, by far with the largest known
population – Areias de Baixo and Ressurgências das Areias da Água
Quente), Ressurgência das Bombas and Córrego Seco Cave, in at least
partially isolated aquifers at the right margin of the Betari stream; and the
Alambari cave system (Alambari de Cima and Gurutuva caves), at the left
margin. The populations in the two sides of the Betari are separated by an
insoluble, diabase dyke and may represent distinct, cryptic species. During
the 1980’s, the Alambari system population practically disappeared,
probably due to pollution caused by uncontrolled mining activities
uspstream from the cave system (Trajano 1997, 2000). The Alambari de
Cima cave population did not recover till now, more than two decades after
its sharp decline. Apparently, other aquatic troglobites also disappeared
or almost, including undescribed decapods, genus Aegla, and planarians.
This well illustrates how a single pollution event may have catastrophic
consequences for the fragile subterranean fauna.

1.4 Human Visitation

Excessive human visitation is one of the major causes of impact for
subterranean faunas. It is the result of the considerable development
of speleology as sports and adventure, overcrowding many caves. This
activity can cause a significant impact on river margins, trampled by
visitors and changing the bottom of streams and rivers through siltation.
Considering that troglobitic fishes use chemical and mechanical reception
to detect food, reproductive partners and to avoid any threat, alteration
of water physical conditions can cause a serious damage to these animals.
For example, the Brazilian cave catfish Ituglanis bambui has a relatively
large population but it occurs in a vadose tributary at Angélica Cave,
central Brazil (Fig. 1). This kind of habitat is extremely fragile; the bottom
is formed by silt and clay and the main visitation in this cave cross the
small tributary. For another species of Ituglanis from central Brazil, the
concern is high – I. epikarsticus is extremely rare in the cave habitat, and the
number of observed specimens conspicuously decreased since the 1970’s.
So far, it has only been found in a set of rimstone pools situated right in the
pathway for visitors reaching the far end of the cave (Fig. 2). Thus, there is
reason for major concern about the future of this highly endemic species.
A strict control of visitation is urgently needed to assure its protection,
until evidence appears that the species is safe in non-cave habitats.
Fig. 1 Locality of the Brazilian cave catfish Ituglanis bambui – a vadose tributary
at Angélica Cave, central Brazil (Photo: Maria Elina Bichuette).

Other negative consequences of visitation are pollution of water as

result of toxic products left in caves (such as carbide and batteries) and
accumulation of non-biodegradable products, topoclimatic changes
caused, for instance, by hot light sources and the opening of artificial
passages during cave exploration and management; introduction of alien
species and materials, such as metal, cement and treated wood; trampling
of cavernicoles and soil compacting of sediment banks, the easiest pathway
inside caves; destruction of specific habits, like rimstone pools formed
Fig. 2 Locality of the Brazilian cave catfish I. epikarsticus – a set of rimstone

pools situated at São Mateus Cave, central Brazil (Photo: Flavio Dias Passos).

by infiltration of epikarstic aquifers; and direct disturbance (lights, loud

noises, handling of specimens, introduction of pathogens) (Trajano 2000).
1.5 Deforestation
In many regions, karst depressions have been used for cultivation since
its soil is commonly deep and fertile, differently from karst slopes where
the soil is generally shallow (Ford and Williams 1989). The removal of
the natural vegetation cover causes changes in karst hydrological regimes,
particularly by disruption storage and release mechanisms in the regolith.
Fire associated with hazard reduction in agricultural areas can cause sealing
of the epikarst, reduction infiltration capacity, so that surface waters may
carry high sediment loads, impacting on groundwater quality (Holland
1994). The main damage is the siltation of the water bodies, in some cases
filling and obstructing the caves, and destroying an entire habitat.
Tree roots hold the soil in the fissures of the rock and prevent denudation
and subsidence of the terrain. Besides buffering erosion, natural vegetation
around karst areas is source of input of allochthonous food items for the
subterranean biota.
In Madagascar, the gobiid Glossogobius ankaranensis may be threatened
by increasing the level of catastrophic floods in some areas and siltation in
others (Proudlove 2001). The same may be true for several other troglobitic
fishes.
1.6 Overcollecting
The harmful effects of overcollecting and handling cave animals were
an early concern of biospeleologists. Sullivan (1956) cautioned against
excessive collecting of specimens, and the National Speleological Society
– NSS (U.S.A.) adopted a conservation policy that discouraged collecting
by amateurs or for any commercial use.
Overcollections may threaten small populations with low reproductive
rates, such as the crayfish, Orconectes sheltae, in Shelta Cave, Alabama,
U.S.A., which may not become sexually mature until reaching the age of
40+ years (Cooper 1975). In general, larger troglobites, such as cavefish,
salamanders and crayfish, are long-lived, have small population sizes,
and reproduce slowly; therefore collecting should be highly restrained.
Smaller, more abundant troglobites, such as some amphipods, isopods,
millipedes and beetles, may withstand intensive collections. However,
some small species are exceedingly rare. Culver (1982) emphasized the
need of restraining collections by admitting that he had caused a severe
decline in populations of several cave isopods.
Overcollecting was suspected to have caused the decline of the cavefish
Amblyopsis spelaea in Mammoth Cave, which were sold as curios in the

19th Century (Poulson 1968). The Mexican characins Astyanax fasciatus and
the African cyprinid Caecobarbus geertsii have been subject to collection for
aquarism. The former bred very easy in captivity. Apparently, the later
cannot bred in captivity and probably all the individuals sold were from
wild stocks. In 1993, 1,500 individuals were reported as imports by the
U.S.A. (Proudlove 2001). Culver (1986) considered Amblyopsis rosae and
Prietella phreatophila threatened by overcollecting.
In Brazil, the Pimelodella kronei catfish population from the type-locality,
Areais de Cima Cave, was subject to over-collections by an amateur during
the mid-1970’s, whose effects were visible ten years later (Trajano and Britski
1992, Trajano 2000). Tourist visitation was forbidden and, although this
prohibition was not vigorously enforced, the population recovered after
15-20 years (E. Trajano, pers. obs.). It is interesting to note that this is the
maximum lifespan estimated for P. kronei (Trajano 1991). As a preliminary
approach, we suggest that the minimum time for a troglobitic population
recovering should be two times the average estimated lifespan.
1.7 Climate Change

In general, stream-dwelling troglobitic fish depend, at least in part, on
floods carrying organic matter into the subterranean biotope. In regions
subject to highly seasonal climates, flash floods typically occur at the peak
of the rainy season (January and February in Brazilian karst areas, in the
Cerrado and Caatiga domains), and there is no significant precipitation
during all the dry season (May to September). Population studies have
shown that the dry season may represent a highly stressful period for
cavefish due to the progressive depletion of food resources, with many
individuals loosing weight and some even shrinking, as observed for
the catfishes Trichomycterus itacarambiensis (Trajano 1997) and Pimelodella
spelaea (Trajano et al. 2004). As indicated by the relatively low degree of
troglomorphism (reduction of eyes and pigmentation), these species may
have been isolated for a relatively short time in caves, lacking specializations
to cope with a pronounced, long-term shortage of nutrients, and living in
a delicate energetic equilibrium.
A decrease in precipitation throughout the last two decades was reported
by local inhabitants of Itacarambi region (locality for T. itacarambiensis),
with flash floods becoming progressively rarer. This is coincident with a
significant decrease in the catfish density observed during visual censuses
in 2006 and 2007 (Trajano et al. 2009) at the resurgence end of the 5,000
m long cave where the only known population is found. On the other
hand, the population density at the 3,000 m upstream was similar to that
recorded in 1994 (Trajano 1997). Such decrease may be due to a deficiency
in restoration of nutrient sources far from the sinkhole. Because the Olhos

d’Água stream is normally a small water course, with a short carrying

capacity, such restoration should be done mainly by flash floods, which
occur every year or nearly so.
A similar situation was recently observed for P. spelaea: after a particular
long dry season, with a two-month delay in the beginning of the rainy
period, the population studied in April 2008 has apparently decreased in
relation to previous years and the few individuals captured were very thin
(Trajano et al., unpubl. data).
It is still unclear whether these phenomena are part of a natural cycle
of climatic fluctuations, illustrating an evolutionary event (bottleneck
effect), or a consequence of progressive climatic changes due to human
activities, perhaps within the global warming context. This being the case,
measures for protection of these interesting cave species would be the
same necessary for the whole planet.
2. ENDANGERED TROGLOBITIC FISHES AROUND THE

WORLD
There are 164 species of troglobitic fishes in the world (Proudlove, this
volume), several still undescribed. The formal description of new species
may be a slow work, especially in the case of those belonging to taxa ill-
defined or in need of revision, or for which close relatives are not clearly
recognized (the so-called “phylogenetic relicts”). For this reason many
species are not included in the Red Books published by the International
Union for the Conservation of Nature and Natural Resources (IUCN),
which categorizes the threat levels and produces the lists of nominal
endangered species. These lists are very helpful in the protection and
proposal of conservationist actions. The first list of cave fishes under the
categories proposed by the IUCN in 1994 was published in 1996 (IUCN
1996). After this, an updated list was published ten years later in a
specialized publication (Proudlove 2006).
There is another difficulty when revisions of the list are elaborated.
Some data about the life history, ecology and behaviour are essential to
propose species like in an endangered category, and for many species, very
little has been studied and are few references in the literature. However,
a number of species has been extensively studied and the literature is
large.
The rarity of the blindfish Amblyopsis spelaea in Mammoth Cave and
its absence from adjacent areas to the north led to speculation that it was
either introduced or decimated during the long period when blindfish were
sold as curios. Poulson (1968) examined historical and scientific records
and found that most early records from Mammoth were for A. spelaea,

not Typhlichthys subterraneus, which also inhabits the cave. A. spelaea was
the dominant species in the Echo and Roaring River areas around 1890,
and it is still common in Roaring River. The present rarity of A. spelaea
is probably related to silting and flooding associated with deforestation,
forest fires, and Lock and Dam (Poulson 1968).
3. ENDANGERED AQUATIC TROGLOBITES: HOW TO

PROTECT THEM?
Caves are windows to the subterranean realm, but the actual subterranean
habitats occupied by troglobitic species are usually much larger,
encompassing inaccessible spaces. In order to guarantee the protection
of these species, it would be necessary to estimate the habitat extension
needed for a minimum sized effective population. Especially for species
living in phreatic waters, it is expected that such minimum habitat size
will be larger than the water bodies appearing in large spaces accessible
to humans, i.e., caves. In order to be really effective, protection of aquatic
troglobites should focus on cave systems and/or aquifers. Unfortunately,
conservation policies throughout the world generally focus on caves, and
not on the entire underground environment.
A major problem for the protection of many troglobitic fishes with
multiples populations is the lack of detailed genetic and morphological
allowing for the recognition of the taxonomic status of each of these
populations. It has been observed that the degree of isolation, indicated by
differentiation, may vary between populations of a same nominal species
(see, for instance, Reis et al. 2006), and the possibility of replacement of loss
in a population by immigrants from cospecific ones will vary accordingly.
In cases with good evidence of isolation, even partial, the isolated sets of
populations should be the taxonomic units for conservation purposes. As
well, undescribed species should be included in red lists when the available
evidence point to important menaces threatening their survival. In view
of their generally high degree of endemism, an acceptable descriptive
reference for troglobites would be its locality.
The Pimelodella kronei population occurring in the right margin
of the Betari stream (see Section 1.3.2) was included in the Red List of
Threatened Fauna in the State of São Paulo (Secretaria do Meio Ambiente
1998) separately from the populations found in the left margin (P. kronei
sensu stricto). It received the status of critically endangered, whereas
the latter were considered as vulnerable, assuring the proper protection
for each population. Unfortunately, the Federal environmental agency
decided not to include, in the Brazilian Red List, undescribed species. As
a consequence, many Brazilian troglobitic species, some in urgent need of

protection actions such as the above mentioned blind catfishes and several
invertebrates, were left aside.
As a general plan for an effective protection of subterranean species, we
suggest the following steps (not necessarily chronological):
• field, taxonomic and genetic studies focusing on distribution,
taxonomic status and connectiveness of populations, population
dynamics (total and effective population sizes and densities, home
ranges) and life history; complementary laboratory research on
behavior and physiology are also important to establish the species
vulnerability to disturbance;
• when actual or potential threats are detected, conservation actions
directed to the species/population ecology and biology, including
but not restricted to habitat protection;
• long-term monitoring of disturbed populations, lasting for at least
two generations-time after conservation actions are taken;
• education aiming to developing public awareness of the beauty and
importance of subterranean organisms and ecosystems, and self-
consciousness of individual, group and institutional responsibility
for environmental protection.
4. RESTORATION PROJECTS
Cave restoration projects are beneficial in removing harmful materials,
trash, and graffiti. Some precautions are needed to avoid stressing cave
communities that may have colonized organic materials, particularly
old woodpiles that may be a heaven for invertebrates. Elliott (1982) and
Hubbard (1995) emphasized that wood should be examined by a biologist
during cleanup projects or left alone. Such woodpiles may have attracted
a large population of invertebrates over decades of time, and to remove
it suddenly may, in effect, trap out a significant portion of the population
in that area of the cave. Oftentimes such wood can be removed gradually
but a small residue should be kept in order to provide habitat for the
remaining fauna (Elliott 1997a).
An important concept is to give some caves or microhabitats time
to “rest”. In fact, it has been observed that a moderate, but continuous
visitation has more harmful effects to cave ecosystems than peaks of
intensive visitation intercalated with relatively long (weeks to months)
periods without or with few visitations. For example, there is a proposal
of seasonal visitation in some rooms of the Kartchner Caverns, in the State
of Arizona (U.S.A.). There is a Big Room area in this cave which is the
summer home of the cave bat, Myotis velifer; the colonies reach to 1,000-
2,000 bats and, in order to protect these bats, cave development work

ceases in this room between May and mid-September (Toomey 2006).

As mentioned above, it is herein suggested that the minimum time for
a troglobitic fish population recovering should be two times the average
lifespan estimated for this population.
5. FUTURE PERSPECTIVES
The ultimate survival of a cave or karst community depends on the proper
protection and management of the cave and the surrounding terrain. In
the case of some endemic, terrestrial cave invertebrates, a one-hectare
preserved area unit may be sufficient as long as nutrient and water inputs
are not altered and disturbances are minimal. However, often we must
reach beyond the immediate karst area to contributing watersheds or,
in the case of bats, to alternate roost habitats. With migratory bats, these
considerations must be on a continental scale to insure success. Many
conservation units established for cave protection have been set aside
only to protect the entrance area, and not the entire watershed for the
cave system. Many recognize the need for ecosystem management, which
considers the whole karst area. The trend by governments has been toward
increasing aspirations for karst ecosystem management, but in practice
it is very difficult financially and politically. In the Austin, area (Texas),
even though seven terrestrial troglobites have been listed as endangered,
the federal government has not acquired any area to preserve them, even
though it has spent millions of dollars to purchase a national wildlife refuge
nearby for the benefit of two endangered bird species. Sixty-five areas that
were planned to preserve caves will have to await funding through a local
development fee collected by the county and city (Elliott 1997c).
Some believe that caves cannot be managed, and that they should be
left alone (except for occasional caving trips). Such a passive conservation
ideal cannot work in developed areas where drainage patterns are
changed, native vegetation and faunas are gradually perturbed, nutrient
and moisture inputs are altered, and pests invade. In managing cave
preserved areas, it is important to be able to measure the results from
time to time. Therefore, baseline surveys, census surveys, and written
management plans are essential.
Karst groundwater issues are so economically and politically
important that they move beyond the influence of the technical experts
who best understand them (Elliott 1994, 1996b). In Texas, the regulation
of overpumping of the Edwards Aquifer, one of the most important
karst aquifers in North America, has been gradually improved through
a succession of state rulings and legislations that were contradicted by
federal court decisions. It is quite possible that the San Marcos and Comal

Springs will go dry within the next decades, endangering many species
and the human economy (Elliott 1993, 1996a).
It has become axiomatic that baseline faunal surveys with simultaneous
monitoring of ecological factors such as temperature, humidity, moisture,
air movements, and nutrient inputs, are essential for understanding
changes in cave faunas (Poulson and Kane 1977, Elliott and Reddell
1989, Perkins 1990, Elliott 1994, Northup and Welbourn 1995, Lewis
1996). Subterranean faunas change seasonally and also in response to
climate changes at larger temporal scales. Initial faunal surveys can be
accomplished in several ways, including hand collecting, baiting, Berlese
extraction, and pitfall trapping. Once a faunal list is compiled, monitoring
censuses in different seasons and years increases the range and value of
the data set. Large fluctuations in some trogloxenic populations can occur
on an annual time scale; for instance, cave cricket populations in some
Central Texas caves dropped by half after a drought year in 1996 (Elliott,
unpublished data). Historical data from the Mammoth Cave ecosystem
over 100 years have been useful in understanding the current ecology of
the cave. Comparative studies of different types of cave communities give
us a better idea of how to manage a particular cave, such as a show cave
that has been perturbed for many years.
References
Bolner, K. and J. Tardy. 1989. Bacteriological and chemical investigations of dripping
waters in the caves of Budapest. In: International Symposium on Physical, Chemical
and Hydrological Research of Karst, Slovenská Speleologická Spločnos, Liptovský
Mikulaš. pp. 102-111.
Cooper, J.E. 1975. Ecological and behavioral studies in Shelta Cave, Alabama, with
emphasis on decapods crustaceans. Ph.D. dissertation, University of Kentucky,
USA.
Cruickshank, G., J. Aguirre, D. Kraemer and E. Craviota. 1980. Bacterial
contamination of the limestone aquifer beneath Merida, Mexico. In: Aquifer
Contamination and Protection, R.E. Jackson (ed.). Project 8.3 of the International
Hydrological Programme, UNESCO, Paris. pp. 341-345.
Culver, D.C. 1982. Cave life: evolution and ecology. Harvard University Press,
Cambridge, 189p.
Culver, D.C. 1986. Cave faunas. In: Conservation Biology: The Science of Scarcity and
Diversity, M. Soule (ed.). Sinauer Associates, Sunderland. pp. 427-442.
Dickson, G.W., L.A. Briese and J.P. Giesy, Jr. 1979. Tissue metal concentrations in
two crayfish species cohabiting a Tennessee cave stream. Oecologia 44(1): 8-12.
Elliott, W.R. 1981. Damming up the caves. Caving International 10: 38-41.
Elliott, W.R. 1982. An introduction to biospeleology. In J. Hassemer (ed.), Caving
Basics. 1987. National Speleological Society, Huntsville, pp.100-108.

Elliott, W.R. 1993. Cave fauna conservation in Texas. In D.L. Foster (ed.), 1991
Proceedings of the National Cave Management Symposium, Bowling Green, Kentucky,
pp. 323-337.
Elliott, W.R. 1994. Biodiversity and conservation of North American cave faunas:
an overview. In B. Mixon (ed.), 1994 NSS Convention Program, Abstracts, National
Speleological Society, Huntsville, p. 48.
Elliott, W.R. 1996a. The Barton Spring salamander. American Caves, 9(1): 14-15.
Elliott, W.R. 1996b. The evolution of cave gating – how the philosophy and
technology have changed. American Caves, 9(2): 9-15.
Elliott, W.R. 1997a. A survey of ecologically disturbed areas in Carlsbad Cavern,
New Mexico. Report to Carlsbad Caverns National Park, 10p.
Elliott, W.R. 1997b. Biological, geological and engineering aspects of Mayor
Elliott Cave, Georgetown, Texas. Report to Paul Price Associates and the City of
Georgetown, Texas, 8p.
Elliott, W.R. 1997c. The caves of the Balcones Canyonlands Conservation Plan,
Travis County, Texas. Report to Travis County, Texas, Transportation and Natural
Resources Department, Balcones Canyolands Conservation Plan, 156p.
Elliott, W.R. 2005. Protecting caves and cave life. In: Encyclopedia of Caves, D.C.
Culver and W.B. White (eds.). Academic Press Elsevier. pp. 458-467.
Elliott, W.R. and J.R. Reddell. 1989. The status and range of five endangered
arthropods from caves in the Austin, Texas, Region. Austin Regional Habitat
Conservation Plan, 100p.
Ford, D.C. and P.W. Williams. 1989. Karst Geomorphology and Hydrology. Unwin
Hyman, London.
IUCN, 1996. 1996 IUCN Red list of Threatened Animals. IUCN, Gland. 368 pp.
Holland, E. 1994. The effects of fire on soluble rock landscapes. Helictite 32(1): 3-9.
Hubbard, D.A. 1995. Cave conservation and cave clean-ups: not always one in the
same. National Speleological Society News, 53:31.
Juberthie, C. 2004. Conservation of subterranean habitats and species. In:
Ecosystems of the World 30, Subterranean Ecosystems. H. Wilkens, D.C. Culver and
W.F. Humphreys (eds.). Academic Press, Elsevier. pp. 691-700.
Lewis, J.J. 1996. Cave bioinventory as a management tool. In G.T. Rea (ed.), 1995
Proceedings of the National Cave Management Symposium. Spring Mill State Park,
Mitchell, Indiana, pp. 228-236.
MMA–—Ministério do Meio Ambiente. (2004). Lista Nacional das Espécies de
Invertebrados Aquáticos e Peixes Ameaçados de Extinção. Instrução Normativa
no 5, 21st May of 2004. Brasília, Brazil.
Moreira, C.R., M.E. Bichuette, O.T. Oyakawa, M.C.C. de Pinna and E. Trajano.
Rediscovery of Stygichthys typhlops Brittan and Böhlke, 1965, a subterranean
characiform fish from Brazil. Journal of Fish Biology (In Press).
Northup, D.E. and W.C. Welbourn. 1995. Conservation of Invertebrates and
Microorganisms in the Cave Environment. In Dale L. Pate (ed) 1993 Proceedings
of the National Cave Management Symposium, Carlsbad, pp. 292-301.


Naturalist, 70(2): 257-290.
Poulson, T.L. 1968. Aquatic cave communities. Cave Research Foundation Annual
Report: 16-18.
Poulson, T.L. and T.C. Kane. 1977. Ecological diversity and stability. In T. Aley
and D. Rhodes (eds.), 1976 Proceedings of the national Cave management
Symposium, Mountain View, Arkansas, pp.18-21.
Proudlove, G.S. 2001. The conservation status of hypogean fishes. In Romero, A.
(ed.). The biology of hypogean fishes. Kluwer Academic Publ., Dordrecht, pp. 201-
213.
(Hypogean) Fishes Described up to 2003 with a Bibliography 1541-2004. International
Society for Subterranean Biology, Moulis, 300 pp.
Silva, A.B. 1984. Análise morfoestrutural, hidrogeológica e hidroquímica no estudo
do aqüífero cárstico do Jaíba, Norte de Minas Gerais. Unpubl. Ph.D. Thesis,
Universidade de São Paulo, São Paulo.
Sullivan, G.N. 1956. Biological aspects of cave conservation. The News, National
Speleological Society 14(3): 30-31.
Reis, R.E., E. Trajano and E. Hingst-Zaher. 2006. Shape variation in surface and
cave populations of the armoured catfish Ancistrus (Siluriformes: Loricariidae)
from the São Domingos karst area, Upper Tocantins River, Brazil. Journal of Fish
Biology 68: 414-429.
Toomey, R.S. 2006. Defining limits and protocol at Karchner Caverns. In: Cave
Conservation and Restoration, V. Hildreth-Werker and J.C. Werker (eds.). National
Speleological Society, Huntsville, pp. 141-145.
Trajano, E. 1997. Synopsis of Brazilian troglomorphic fishes. Mémoires de Biospéologie,
24: 119-126.
Trajano, E. 2000. Cave faunas in the Atlantic tropical rain forest: composition,
ecology and conservation. Biotropica, 32, (4b): 882-893.
Trajano, E. and H.A. Britski. 1992. Pimelodella kronei (Ribeiro, 1907) e seu sinônimo
Caecorhamdella brasiliensis Borodin, 1927: Morfologia externa, taxonomia
e evolução (Teleostomi, Siluriformes). Boletim de Zoologia, São Paulo 12: 53-89.
Trajano, E., R.E. Reis and M.E. Bichuette. 2004. Pimelodella spelaea, a new cave catfish
from Central Brazil, with data on ecology and evolutionary considerations
(Siluriformes: Heptapteridae). Copeia 2004(2): 315-325.
Trajano, E., S. Secutti and M.E. Bichuette. 2009. Population decline in a Brazilian
cave catfish, Trichomycterus itacarambiensis Trajano & Pinna, 1986 (Siluriformes):
reduced flashflood as a probable cause. Speleobiology Notes 1: 24-27.
Varela, E., O. Barros and S. Jimenez. 1980. The effects of sugar-waste disposal on a
karstic limestone aquifer. In: Aquifer Contamination and Protection, R.E. Jackson
(ed.). Project 8.3 of the International Hydrological Programme, UNESCO, Paris.
pp. 329-339.
Vesper, D.J. 2005. Contamination of cave waters by heavy metals. In: Encyclopedia
of Caves. D.C. Culver and W.B. White (eds.). Academic Press, Elsevier.
pp. 127-131.

CHAPTER 4
Behavioral Patterns in Subterranean
Fishes
Jakob Parzefall1 and Eleonora Trajano2
1
University of Hamburg, Biozentrum Grindel, Germany.
E-mail: parzefall@zoologie.uni-hamburg.de
2
Departamento de Zoologia, Instituto de Biociências da Universidade de
São Paulo, Brazil. E-mail: etrajano@usp.br
I. INTRODUCTION
The ecological conditions in the subterranean environment are characteriz-
ed by two main factors: permanent darkness, thus absence of green plants
and of photoperiods, and a more or less constant temperature. Subterranean
habitats in different parts of the world have been colonized successfully
by teleostean fishes from 20 families (Proudlove, this volume). In the
majority of these families, epigean (surface) members are known to be
active at dusk and/or night, so it is not surprising that those species with a
preference for activity in darkness colonize subterranean habitats such as
caves. But among the ancestors of troglobitic fish we also note members of
the families Characidae (tetras), Cobitidae (loaches), Balitoridae (loaches),
Cyprinidae (carp-fish) and Poeciliidae (toothcarps), which include many
species with clear diurnal activity.
Mainly for these species the question arises what enables certain
families to survive and to reproduce in caves. One possibility would be
that members of these families show behavioral preadaptations to life in
darkness. Preadaptations (or exaptations) may be defined as character
states conferring performance advantage in a given selective regime,
but which have been selected in another, independent previous regime
(Trajano 2005). In addition, or associated with the predominantly nocturnal
activity, preadaptations in epigean fishes include enhanced mechano

and/or chemosensorial systems, as in the amblyopsids and siluriforms

respectively, and an opportunistic, generalist feeding.
Nevertheless, in order to distinguish preadaptations, which are
plesiomorphic states for troglobites, from autopormorphies resulting from
the differentiation in the subterranean habitats, it is necessary to establish
the direction of transformation series into cladograms, by applying the
comparative method. For behavior, both the outgroup comparison and
the ontogenetic methods may be used, but they require phylogenetic
trees and ontogenetic series respectively. Unfortunately, there are few
ontogenetic series available for cavefish, because relatively few species
reproduce in laboratory and collection of juveniles directly in the natural
habitat is rare, and even fewer behavioral data for the younger stages. As
well, phylogenies including troglobitic fish species are scanty.
The evolutionary interpretation of character states observed in
subterranean organisms is particularly difficult for the so-called
phylogenetic relicts, i.e., species without known epigean close relatives,
as is the case with the Brazilian phreatic characiform, Stygichthys typhlops.
At first considered a characid, recent taxonomic studies pointed to a more
basal position within the characiforms, but its exact familial position
was not established yet. As well, the phylogenetic positions of the four
recognized Brazilian species of troglobitic Rhamdiopsis (see Trajano and
Bichuette, this volume) within the genus and that of the phreatic African
catfish, Uegitglanis zamaranoi, within the Clariidae, were not established.
Consequently, hypotheses on direction of evolutionary change for character
states observed in this highly modified species are questionable.
At the moment, the most parsimonious solution is to compare the
studied troglobitic species with the geographically closest epigean
congeneric species, in the assumption that this would be a less modified
descendant of a common ancestor with the differentiated cave species
(the hypothesis of extinction of the epigean sister-species followed by
dispersion of less related species to a closer geographic area requires more
evolutionary steps). For the outgroup comparison method, a third species
is required. In the absence of phylogenies, available data on any other
epigean congeners has been used (e.g. Trajano and Bockmann 1999).
In this chapter, we compare behavior patterns of troglobitic fishes with
their closest epigean relatives, aiming to detect apomorphic character states
resulting from specialization to the subterranean way of life, and discuss
common patterns among troglobitic species, in search of homoplastic
traits indicating convergent adaptation to this peculiar life due to similar
selective regimes. We examine whether differences observed in relation
to epigean relatives are phenotypically plastic responses to cave life, or
whether the observed behavioral changes are genetically fixed.

Jakob Parzefall and Eleonora Trajano 83
II. BEHAVIORAL PATTERNS IN SUBTERRANEAN

FISHES
The ethology of subterranean fishes is a comparatively well developed
discipline of Subterranean Biology in great part thanks to the pioneering
work of the Belgium researcher G. Thinés and collaborators, who
published a most impressive experimental work focusing on different
species from the 1950’s and 1980’s, and the German geneticist C. Kosswig,
who carried out the first studies on the genetics of behavioral traits, from
the late 1940’s to the early 1960’s. In the United States, on the early 1960’s
T. Poulson published the first comprehensive comparative study on cave
fishes, which included behavioral aspects. Kosswig’s students, conducted
by J. Parzefall and H. Wilkens, continued his work with special focus
on Mexican cavefish. This legacy was augmented by Italian researchers
leaded by R. Berti from the 1970’s on and by Brazilian (E. Trajano and
collaborators) and Indian researchers (J. Biswas, A. Pati and collaborators)
from the late 1980’s on.
As a consequence, several species, mainly from Mexico, Brazil, India and
northern Africa, have been studied in greater or less detail. Nevertheless,
behavioral studies on subterranean fishes still present a great taxonomic
heterogeneity. A few species have been intensively studied, especially
the Mexican tetra characins, genus Astyanax, whereas important groups
with many troglobitic derivatives, such as the cyprinids, are scarcely
known with regard to their behavior in particular and biology in general.
Siluriformes (catfishes), another important fish group in subterranean
habitats, are in the mid term: several Brazilian species belonging to
three families (Heptapteridae, Trichomycteridae and Loricariidae) have
been studied in detail, but most species from other countries call for
investigation. Therefore, general patterns are not quite clear.
The reasons why Mexican blind tetras are so intensively studied are
mainly practical, because these fish are easily maintained and bred in
laboratory, being an excellent material for genetic studies. Nevertheless,
they do not provide a good paradigm because, being one among only
three known troglobitic characiforms (in contrast with 78 cypriniforms
and 52 siluriforms – Proudlove, this volume), the Mexican blind Astyanax
represent a taxonomic exception. In fact, their putative epigean sister-
species, A. mexicanus, active at dusk and with a chemically oriented
reproductive behavior, is an exception among the generally diurnal,
visually oriented characids. As well, several behavioral traits of the blind
tetras do not have clear equivalents in many other subterranean fishes.
Another difficulty arises from the fact that congeneric troglobitic
organisms may present a mosaic distribution of character states related to

an exclusively subterranean life, indicating not only a rather independent

differentiation but also different evolutionary rates affecting these
characters. A mosaic distribution of character states (although not presented
as such), including behavioral traits, was first noticed for amblyopsids
(Poulson 1985). This small North American family, which encompasses an
exclusively epigean, a troglophilic and four troglobitic species, provides
an excellent material for comparative research (see Niemiller and Poulson,
this volume).
Well documented cases of a mosaic distribution of character states
in fishes include Brazilian armored catfish, genus Ancistrus (Bessa and
Trajano 2002), and Thai cave balitorids, genus Schistura (Trajano and
Borowsky 2003). The same is true for several confamilial taxa such as the
neotropical heptapterid catfishes (Trajano and Bockmann 1999). Besides
morphological traits, such mosaics affect behavioral characters as diverse
as vertical distribution, cryptobiotic (hiding) habits, reaction to light and
agonistic interactions. The factors underlying this phenomenon are still
poorly understood because, at a coarse-grained scale, the selective regimes
under which these species are seem not to differ significantly.
Among the troglobitic fishes treated below, the Mexican cave Astyanax
were compared to the epigean A. mexicanus, the cave form of Poecilia
mexicana, characterized by variable eyes and pigmentation, to the
normally eyed and pigmented epigean form, Thai cave balitorids with
epigean species belonging to the same and different genera, troglobitic
amblyopsids with epigean and troglophilic. Among the Brazilian
catfishes, the heptapterids Pimelodella kronei was compared to its putative
epigean sister-species, P. transitoria, and to other congeneric catfish living
in other basins (including the troglobitic P. spelaea), Rhamdia enfurnada to
the epigean widely distributed nominal species, Rhamdia quelen, and the
troglobitic trichomycterids and loricariids with any epigean congeners for
which behavioral data were available.
These studies included mainly laboratory research, complemented by
naturalistic observations in the habitat in several cases, including most of
the Brazilian species so far investigated. They showed that, although it is
not clear how general are the detected patterns for troglobitic fishes, some
phenomena are quite common and would represent patterns at least for
some taxa. In fact, among the aspects discussed below, there are many cases
of loss or reduction of behavioral traits, similar to the reduction of eyes
and pigmentation, well know and characteristic of troglobites in general.
Such reductions would represent instances of regressive characters, i.e.,
characters where the derived states correspond to a decrease in the size,
intensity and/or complexity of structures, behaviors or physiological
processes (Trajano 2005), a process frequently attributed to the loss of

function or, in other words, to the relaxation of stabilizing selective

pressures (Wilkens 1992).
II.1 Spontaneous Behavior

Most epigean siluriforms and balitorids, two taxa with many subterranean
representatives, are bottom-dwellers. Except for highly specialized
benthonic taxa such as the armored catfishes (Loricariidae), a tendency to
increase the midwater activity, in parallel with a decrease in cryptobiotic
habits, has been observed in the studied troglobitic catfishes and also in
balitorids from Thailand.
Both in caves and in laboratory, the catfish Pimelodella kronei
(Heptapteridae), from southern Brazil, is frequently observed swimming
in the midwater and near the surface. In contrast, Pimelodella transitoria
is a typical bottom-dweller, very shy, elusive and photophobic catfish.
In a quantitative laboratory study comparing these species, with a few
exceptions the blind catfish spent at least half the observation time calmly
swimming, and most specimens stayed hidden under dens less than one
third that time; in contrast, the epigean catfish remained almost all time
inside their hiding places (Trajano 1989).
Not quantified observations point to a similar adaptation in the
highly troglomorphic heptapterids of the Nemuroglanis subclade (sensu
Trajano and Bockmann 1999), Rhamdiopsis sp. from Chapada Diamantina
(cited as “new heptapterine” in Trajano 2003) and Rhamdiopsis sp. from
Campo Formoso (formerly cited as Taunayia sp.), and the trichomycterid
Trichomycterus sp. from Serra da Bodoquena (see Trajano and Bichuette
this volume), and, to a lesser degree, in the Rhamdiini heptapterids, Rhamdia
enfurnada (Bichuette and Trajano 2005) and Rhamdia macuspanensis (Weber
and Wilkens 1998), the trichomycterid Trichomycterus itacarambiensis and a
new population of Aspidoras, the first known troglomorphic callichthyid
catfish (Trajano and Secutti, pers. obs.). As expected, the morphologically
less specialized Pimelodella spelaea and Rhamdiopsis sp. from Cordisburgo
(Trajano and Bichuette, op. cit.) retained the basically bethonic habits.
The increased midwater activity probably enhances the chances
of finding food under conditions of low density of prey, especially if
associated to low predation rates and low competition at midwater. As
a matter of fact, the proportion of allochthonous (terrestrial) food items
in the diet of Pimelodella kronei is higher than the observed for its eyed
relative, P. transitoria (Trajano 1989). In the case with P. kronei, midwater +
bottom activity may be a plesiomorphic character state, since it was also
observed for other Pimelodella species, including P. transitoria (Trajano and
Bockmann 1999). As well, constant swimming in the water column in

addition to the bottom may be a plesiomorphic state for Trichomycterinae

catfishes. Thus, high activity levels with frequent midwater swimming
may be a plesiomorphic trait for cave trichomycterids. On the other hand,
midwater activity in the two above mentioned Rhamdiopsis species is
better interpreted as an autapomorphic adaptation of these species to the
cave habitat (Trajano and Bockmann 1999). Apparently, this is also the case
with the cave Aspidoras catfish.
Unlike their epigean relatives, several Brazilian troglobitic catfishes
rarely hide. These are possible examples of behavioral regression related
to cave life, which affects cryptobiotic habits. Such evolutionary reduction
is consistent with the hypothesis that cryptobiotic habits are part of a
strategy against visually-oriented predators in the surface habitat. Hiding
is not effective against non-visual, chemo-oriented predators, and other
strategies are required against predators able to feed in caves, such as
nocturnal predaceous fishes. Therefore, selection for maintenance of
hiding habits may be relaxed and the character regress, as much as eyes
and pigmentation would regress in the absence of light (Trajano 2003).
A mosaic distribution of behavioral traits was reported for Brazilian
loricariids, genus Ancistrus. The morphologically specialized Ancistrus
formoso is conspicuously cryptobiotic (although less than epigean
Ancistrus spp.), spending almost 70% of the time hidden, whereas an inter-
population variation is observed in the less specialized Ancistrus
cryptophthalmus, in parallel with the degree of eye reduction (but not with
the pigmentation reduction – Reis et al. 2006), from 60% of the time hidden
in catfish from Angélica Cave to less than 15% in specimens from Passa
Três Cave (Bessa and Trajano 2002).
For some Brazilian troglobitic fishes, potentially important predators
were found in the cave habitat, as is the case with Rhamdiopsis sp.
from Campo Formoso (Trajano and Bockmann 2000) and Ancistrus
cryptophthalmus (Trajano and Bichuette, unpubl. data). This may explain
the strong avoidance of mechanical stimuli in the former, which presents
enlarged cephalic lateral line canals. Probably due to the absence of
important predation rates, other cave fishes, such as P. kronei, R. enfurnada,
Rhamdiopsis sp. from Chapada Diamantina and T. itacarambiensis, do not
show strong phobic reactions to stimuli as those from water movements
and sudden illumination.
A mosaic of behavioral character states was also observed for four
troglobitic balitorids from Thailand, Schistura oedipus, S. jaruthanini, S.
spiesi and Nemacheilus troglocataractus (which would also be a Schistura –
R. Borowsky, pers. comm.), studied in laboratory (Trajano and Borowsky
2003). The epigean balitorids, S. mahnert, S. similis and Acanthocobitis
zonalternans, were studied under the same standardized conditions,

allowing for the application of the out-group comparison method. The

epigean balitorids were typical bottom-dwellers, with occasional incursions
onto the midwater. S. jaruthanini, the least troglomorphic cave balitorid
we studied, presented a significant activity at midwater in addition to the
bottom, a specialized behavior convergent with those observed in several
Brazilian troglobitic catfishes. In contrast, the highly troglomorphic species
N. troglocataractus and S. oedipus retained the predominant bottom activity.
Schistura spiesi is intermediate in this respect. Cryptobiotic habits varied
among cave balitorids, with no pattern in relation to those of the epigean
ones: S. oedipus was the most cryptobiotic, and S. jaruthanini, the least.
II.2 Feeding Behavior

Suitable food sources and food quantity vary among different subterranean
habitats, and also from cave to cave. In general, subterranean animals
depend upon food brought in from outside the cave and tend to be
opportunistic carnivores or omnivores. Even though there are exceptions
represented by caves with abundant food (Langecker et al. 1996), most
subterranean habitats, especially the phreatic ones, have less food
available than surface habitats. Food sources can be widely distributed
or concentrated in patches and are often unpredictable in space and time.
Therefore, food finding abilities are often improved and additional storage
tissues help cave fishes to survive prolonged starvation periods.
The blind cave form of Astyanax mexicanus from the Pachon cave was
studied in competition experiments in darkness. In the case of the cave form,
the focal fish detected 80% of small pieces of meat that were distributed
on the bottom of an aquarium, whereas the epigean fish managed to find
only 20% of food items (Hüppop 1987). Schemmel (1980) found that the
improved food finding ability in the cave form of A. mexicanus is due to
a behavioral adaptation: the cave fish swim and feed in a typical angle of
about 45° to the substrate (Fig. 1), whereas the epigean fish swim vertically
while feeding. Studies using population hybrids confirmed that this
difference is based on genetic factors. In addition, the taste buds, which
are restricted to the mouth region in the epigean fish, are distributed over
the lower jaw and cover also ventral areas of the head in the cave form. A
similar feeding mode is exhibited by the Texan blind catfish Trogloglanis
pattersoni (Ictaluridae), which have a very specialized mouth used for
the chemical localization of food on the bottom (Langecker and Longley
1993).
When investigating the feeding behavior of amblyopsid fishes, Poulson
(1963) obtained results similar to those reported for cave Astyanax: the
blind Amblyopsis spelaea detected a small prey item hours before the

Fig. 1 The position in cave and epigean form of Astyanax mexicanus to pick food
from the bottom (after Schemmel 1980).
troglophilic Forbesichthys agassizi. A. spelaea ate all ten Daphnia presented

in the test aquarium before C. agassizi had eaten half of them (Niemiller
and Poulson, this volume; Poulson and White 1969).
On the other hand, no important difference in the food finding ability
per se was observed for P. kronei when compared to P. transitoria – the
higher searching times recorded in laboratory for the latter are due to its
hesitation to leave the den, once outside both species are equally efficient
in food taking, even under dim light. This supports the idea that the
epigean catfish are highly preadapted to colonize subterranean habitats
and few adaptations were necessary for a successful cave life in P. kronei.
With regard to the feeding tactics, in addition to the crepuscular to
nocturnal preying of bottom animals and grubbing (excavating while
moving) usual for epigean bottom-dwelling catfish, the troglobitic catfishes
with increased midwater activity are also surface pickers. For instance, P.
kronei catfish are frequently observed swimming with the open maxillary
barbels sweeping the water surface and promptly catching prey near or
at the water surface. In the natural habitat, allochthonous items are more
important for this species than for P. transitoria (Trajano 1989).
In subterranean amblyopsids, density dependent cannibalism seems to
account for most of the mortality of the young (Poulson 1963). This would
be an important mechanism of population control, in view of the food
scarcity and lack of predators observed in the cave habitat for these fish
(Poulson, op. cit.). Cannibalism was also observed for the Brazilian fishes,
Rhamdiopsis sp. from Toca do Gonçalo and Stygichthys typhlops. Both are
highly specialized troglobites living in phreatic habitats apparently subject
to severe food shortage, and conspecifics could represent an important
source of food. Rhamdiopsis sp. catfish are able to ingest very large prey,
even larger than themselves, an adaptation enabling a quite wide feeding
niche. A specialized behavior was observed for the ingestion of large
organisms, including conspecifics, which are swallowed posterior part
first and digestion starts with the prey only partially swallowed (Trajano
and Bockmann 2002).

II.3 Biological Rhythms

The cyclic activity of organisms is internally controlled by oscillators,
which are organized in internal timing systems (“biological clocks”).
These oscillators produce endogenous rhythms which are synchronized
by external cyclic stimuli, the zeitgebers (time givers) or forcing signals,
resulting in the observed rhythmic behaviors. Cycles with a period close
to 24 h are circadian rhythms, those with period close to four weeks are
circalunar, those close to 12 months are circannual and so on. The most
powerful zeitgebers for activity, including the locomotor and feeding
behaviors, are the daily cycles of light and temperature. The seasonal
variation in the length of photoperiods and monthly temperatures are
important zeitgebers of reproductive cycles, migrations etc.
Troglobitic species, especially those evolving in very constant
environments, provide good opportunities to test the two main hypotheses
on the evolution of circadian rhythms, the internal versus external,
ecological selection (Lamprecht and Weber 1992). Having evolved in the
absence of 24-h zeitgebers, it is predicted that, if external factors provide
the main selectives forces for circadian rhythms, these could regress at
some degree in many troglobites like the observed for eyes, melanic
pigmentation and other characters related to light. On the other hand, if
internal order were of prime importance for the maintenance of circadian
rhythms, these would not be lost in troglobites (Volpato and Trajano
2006).
Several studies pointed to the loss or weakening of circadian components
of activity in troglobites as diverse as beetles, crustaceans and fishes,
suggesting that such rhythmicity is not necessary for the maintenance
of internal temporal order, thus corroborating the hypothesis of external
selection as a major selective force for the appearance and maintenance of
endogenous circadian rhythms, at least for locomotor activity. Such pattern
is quite clear for the several studied subterranean fishes. In general, these
species were studied by monitoring their locomotor activity in laboratory,
both under constant (free-run) conditions, usually darkness, and LD
cycles, using an infrared photocells setup connected to a computer system
which automatically records the number of beam crossings along the time.
All these troglobitic species belong to typically nocturnal taxa and, when
known and available, the epigean sister-species were also studied for
comparison.
In Astyanax mexicanus, free-running activity rhythms were detected,
as expected for an epigean species. Light/dark (LD) cycles with different
periods (12:12h, 6:6h, 4:4h, 16:8h etc.) acted as zeitgebers, entraining the
locomotor activity, and residual oscillations (post-oscillations) were

observed after the transition from LD to DD, during one or a few cycles.
This suggests the existence of an endogenous circadian oscillator, which
becomes obvious in free-running conditions. On the other hand, in the
highly specialized troglobite, “A. antrobius” (from El Pachón Cave),
although activity was entrained by all applied LDs, the signal energies were
lower than in the tests with epigean fishes, the rhythms of total activity
disappeared immediately after the transition from LD to DD (no residual
oscillations), and in no LD with a period frequency differing from 24-h a
circadian rhythm could be observed in addition to the entrained frequency.
It is concluded that the internal clock of “A. antrobius” was simplified in
relation to its epigean ancestor: the passive system has developed into
an extremely passive one, incapable of synchronizing, thus the circadian
oscillator was subject to regression, but it was not completely lost (Erckens
and Martin 1982). The hybrid population, “Astyanax jordani”, from La
Chica Cave, seems to be intermediate in this aspect, because one or two
residuals oscillations were observed after a transition from LD (12:12) to
DD (Erckens and Weber 1976). The persistence of a circadian clock in this
cave fish was pointed out by Cordiner and Morgan (1987), who recorded
free-running circadian activity rhythms (not shown by Erckens and Weber
1976), often masked by apparently random, infradian oscillations.
Several Brazilian catfishes, with different degress of troglomorphy
(see Trajano and Bichuette, this volume) had their locomotor rhythmicity
investigated (Trajano and Menna-Barreto 1995, 1996, 2000, Trajano et al.
2005, unpubl. data). In most cases, individual variation concerning the
presence of free-running circadian rhythms and the frequency and
periodicity of other rhythms was reported. In the less specialized, stream-
dwelling troglobitic Pimelodella kronei, P. spelaea, Rhamdia enfurnada,
Rhamdiopsis sp. from Cordisburgo and Trichomycterus itacarambiensis, at
least 2/3 of the monitored specimens did present significant circadian
rhythms under constant darkness. It is notheworthy that all the studied
epigean catfishes, P. transitoria and Taunayia bifasciata, exhibited strong,
significant free-running circadian components of locomotor activity.
On the other hand, in highly troglomorphic Brazilian subterranean
fishes, the proportion of individuals with free-running circadian rhythms
was much lower, as in Rhamdiopsis sp. from Chapada Diamantina, or none,
as is the case with Stygichthys typhlops and Rhamdiopsis sp. from Toca do
Gonçalo. The two Rhamdiopsis species were also studied under LD cycles,
12:12 h. In both species, activity was entrained by these cycles, but no
residual oscillations were observed, indicating a possible masking effect.
Activity rhythms in six species of balitorids from Thailand, two
epigean Schistura species and four troglobitic species (Schistura oedipus, S.
jaruthanini, S. spies and Nemacheilus troglocataractus) were investigated in

laboratory. All species exhibited an approximate 24 hour rhythms, except

S. oedipus, which had a longer cycle, 38-40 hours (Borowsky and Trajano
2003).
The Indian balitorid, Nemacheilus evezardi, was subject to several studies
on circadian and circannual rhythmicity, with focus on locomotor activity,
air gulping behavior and phototactic behavior (Pradhan et al. 1989, Biswas
et al. 1990, Biswas 1991). The epigean “form” N. evezardi is basically a dusk
active bottom, with maximum locomotor activity at the early part of the
dark phase, concealing themselves under stones during most of the day.
This is another example of epigean fish preadapted to the subterranean
life.
Biswas et al. (1991) studied the locomotor activity and surfacing
frequency of N. evezardi during the pre-spawning, spawning and post-
spawning phases, comparing epigean (kept under natural LD cycles)
and hypogean individuals (kept under DD). Significant rhythms were
detected for total activity in both epigean and hypogean loaches during
the pre-spawning phase, but only for the epigean fishes during the
spawning and post-spawning phases. Significant circadian rhythms of
surfacing behavior related to air-gulping were observed during the pre-
spawning and spawning phases only. In hypogean loaches, the levels of
total and surfacing activity were lower in the pre-spawning and spawning
phases, suddenly increasing during the post-spawning phase, whereas
no change in total activity was noticed for the epigean loaches. These
results demonstrate the influence of the reproductive condition over the
expression of circadian rhythmicity, which differs among cave and epigean
populations.
In conclusion, the comparison between troglobitic fishes and their
epigean close relatives (troglobitic Astyanax versus A. mexicanus, P. kronei
versus P. transitoria; troglobitic Rhamdiopsis versus Taunayia bifasciata, S.
oedipus versus epigean and other cave balitorids, N. evezardi cave form
versus epigean form) analyzed under the same conditions and using
the same protocols, provides good evidence for the hypothesis of an
evolutionary regression of time-control mechanisms in troglobitic species,
either affecting the oscillator(s) itself (themselves) or due to an uncoupling
between the oscillators and at least one of their related functions; in the
case, the locomotor activity. Regression of retina and, at least for some of
the studied species, possibly also of the photoreceptor of the pineal organ,
where circadian oscillators of fishes would be located, may be involved in
the disorganization of the circadian system verified in several troglobitic
species (Volpato and Trajano 2006).
Besides differences in general locomotor activity, cavefishes do not
reproduce as regularly as their epigean relatives: e.g., amblyopsid cave-

fishes reproduce only when food is abundant in the cave (Poulson 1963).
In various cave populations of Astanax mexicanus we could observe
offspring or very young fish only occasionally during 25 years of field
work (Parzefall, pers. obs.).
II.4 Reaction to Light

Reaction to light is understandably one of the earliest and most intensively
studied behaviors in subterranean fishes. So far, the phototactic behavior
has been studied in troglobitic fishes from ten families: Characidae
(genus Astyanax), Cyprinidae, Balitoridae, Heptapteridae (formerly in
Pimelodidae), Trichomycteridae, Loricariidae, Clariidae, Amblyopsidae,
Bythitidae (formerly in Brotulidae) and Poeciliidae (Parzefall 1998,
Gerhard and Trajano 1997, Bessa and Trajano 2002, Trajano and Borowsky
2003).
Many cavefishes have epigean relatives that are photophobic, as is the
case with the catfishes. A decreased photonegative response in relation to
the highly photophobic epigean relative, P. transitoria, was described for
P. kronei, with a considerable individual variation. A tendency towards a
weak photonegative or even a photopositive response was also recorded
for other catfishes such as the Somalian clariid, Uegitglanis zammaranoi,
and the Brazilian trichomycterid, Trichomycterus itacarambiensis. For a
discussion on the reaction to light in siluriforms, see Trajano (2003).
It is noteworthy that the mosaic distribution in morphological
and behavioral characteristics observed in the Brazilian Ancistrus
catfish includes the reaction to light (Trajano and Bessa 2002): adult A.
cryptophthalmus catfish from Passa Três Cave are indifferent, the reaction
to light in Ancistrus formoso varies individually, and A. cryptophthalmus
catfish from Angélica and Bezerra caves are basically photophobic,
although less than the epigean Ancistrus from the same and other
regions too. Juvenile from Passa Três, with eyes still not regressed, are
photonegative, indicating that the absence of reaction to light in choice-
chamber experiments exhibited by the adults, with regressed eyes sunken
into the head (Secutti and Trajano 2009), may be a consequence of poor
light sensitivity; nevertheless, the possibility of an ontogenetic behavioral
change may not be ruled out.
In members of the families Amblyopsidae, Bythitidae, Characidae,
Claridae, Cyprinidae, photonegative behavior with respect to light
intensities above 30 lux has been described. Only the presumed
phylogenetically young cave form of the poeciliid Poecilia mexicana showed
a photopositive reaction at three different light intensities (Parzefall et al.
2007, Plath and Tobler, this volume).

The most detailed study has been conducted in different populations

of the characid Astyanax mexicanus by Langecker (1992). Langecker could
demonstrate that this behavior in the epigean fish depends on light
intensity. Juveniles showed a slightly positive response under all tested
conditions (Fig. 2), while adult fish were photonegative. Similar results
were found in juvenile and adult cave Garra barreimiae (Timmermann
and Plath, unpublished data). The photonegative behavior in surface
dwelling Astyanax was distinctly stronger in blinded fish. Blinded fish
that had been pineal-ectomized showed a weaker response, suggesting
an important role of the pineal organ for light perception and phototactic
behavior. In contrast to the epigean fish, the removal of the pineal organ
had no significant effect on the phototactic behavior of the cavefish. Thus,
it could be argued that differences in the phototactic behavior between
surface and cavefish would be caused by a reduction process in parts of
the pineal organ. Ultrastructurally, however, surprisingly little change
in this organ was detected (Langecker 1990). In addition, the degree of
differentiation of the photoreceptor cells indicates that they are probably
still light sensitive (Langecker 1990). Therefore, it seems that the different
phototactic responses of surface and cave Astyanax represent a reduction
process on the behavioral level. The same mechanisms may also be
operating in the case of several other cavefishes, but data are still lacking
for most systems.
Thai balitorids were also investigated with focus on their reaction to
light in a series of tests under different light intensities and in different
phases of 12:12 h light-dark cycles (Trajano and Borowsky, unpubl. data).
Fig. 2 Preference for the lighted area in a light-dark choice experiment of Astyanax
mexicanus. Mean and variability of eyed, blinded and pineal ectomized epigean
fish and blind Pachon cave fish under different light conditions (after Langecker
1989, 1992).

In general Schistura loaches, both epigean (S. moiensis, S. similis and S.

mahnert) and troglobitic (S. oedipus, S. jaruthanini and S. spiesi), exhibited
photonegative responses in choice-chamber experiments, although non-
significant responses were observed in some tests. As well, individual
variation was observed in the three studied N. troglocataractus fish. Except
for S. jaruthanini, the troglobitic species showed a weaker photophobia
in the night phase of the 12:12 LD cycle; whether these differences
between the phototactic behavior in the light and in the night phases are
endogenous or due to masking is an open question. On the other hand,
Acanthocobitis specimens, studied as another balitorid outgroup, did not
show a photonegative behavior, indicating a diurnal activity (Trajano and
Borowsky, unpubl. data). This is consistent with field observations on
epigean balitorids in Thailand (Trajano et al. 2002), and points to a possible
nocturnal activity as a specialized behavior in Schistura. Such preadaptation
to the subterranean life could explain the diversity of troglobitic Schistura
species.
II.5 Dorsal Light Reaction

With some exceptions, nearly all fish swim with their back oriented
towards the light. Overall, the vertical orientation of the dorso-ventral
axis is triggered by the direction of light and gravity. Illumination from
different angles relative to an upright position of the test fish result in a
deviation of this normal position and can be measured experimentally as
the angle of inclination from 0° (= vertical orientation).
To date this reaction has been studied in a comparative approach,
including surface and cave forms of the same species, in Astyanax mexicanus
only. Using hybrids between cave and epigean populations as well as
the phylogentically young Micos cave population with functional eyes,
Langecker (1990) found that the angle of inclination has changed in the
cave fish (Fig. 3). The dorsal light reaction is almost completely reduced
in the cave fish, and absence of a response did not depend on the degree
of eye development (Langecker 1990). Using crossbred fish, the regression
was found to be genetically based and to underlie polygenic inheritance
involving at least three genetic factors (Langecker 1993)
II.6 Chemical Communication and Alarm Reaction

Chemical signals are good candidates for social communication in the
absence of light. In fact, most catfishes are especially dependent on
chemical orientation and several studies pointed to their high sensitivity
to chemical cues, showing its importance in feeding behavior and social

60
Teapao
50 F1 (TEA×PA) n = 18
40 n = 18
30
20
10
60
Frequency (%)
50
40 F2 (TEA×PA)
30 n = 36
20
10
60
50
Micos
40
N = 24
30
20
10
s4 9 14 19 24 29 34 39 44 49°
Angle of inclination (degrees)
Fig. 3 Dorsal light reaction of Astyanax mexicanus. Frequency distribution in the

epigean population, F1 and F2 hybrids with the blind Pachon Cave fish, and in the
Micos Cave fish. All specimens tested had functional eyes (data from Langecker
1990).
communication. Several authors studied the role of chemoreception in

the detection at distance of food and conspecific substances in troglobitic
vertebrates. Mexican blind tetras, Astyanax sp., and cyprinids, Caecobarbus
geertsi and Phreatichthys andruzzii, tested in choice-chamber setup, showed
generalized attraction to water conditioned by conspecifics (Trajano 2003).
In a food-poor environment with a low density of conspecifics, it would be
advantageous for non-aggressive troglobitic fish to approach any stimuli
that could signalize the presence of food or potential mates.
The situation is different for aggressive troglobites such as Pimelodella
kronei, which showed a more complex behavior when compared to P.
transitoria (Hoenen and Trajano 1995). While all tested eyed catfishes

showed consistent repulsion to chemical cues from conspecifics and from

blind catfishes as well, individual responses of P. kronei greatly varied
in accordance with sex and size of the catfish: large males tended to be
attracted and small fish (less than 12.0 mm SL) tended to be repelled by
conspecifics. This indicates that social interactions are more complex
in P. kronei than in P. transitoria – the latter simply avoid other fish in
most situations, whereas blind catfish reactions vary according to the
circumstances (Hoenen and Trajano, op. cit.).
The generalized repulsion of Pimelodella transitoria to other catfishes
is possibly a correlate of its cryptobiotic habits and part of a generalized
phobic behavior, including photophobia. It may be regarded as a
mechanism allowing individual spacing in order to minimize direct
competition for hiding places. If this is characteristic of the genus (or
higher groups), then P. kronei behavior would represent a specialization
to the subterranean life, through an increase in the frequency of positive
responses to the presence of conspecifics. As in Astyanax sp. and other
troglobitic fishes, positive responses to conspecifics cues would enhance
the chance of finding food and reproductive partners. However, tendency
to aggregate increase the intraspecific competition at food sites and, thus,
the probability of agonistic interactions. Therefore, small individuals,
whose chance of meeting a large one and being chased away is higher,
would tend to avoid conspecifics in order to prevent disadvantageous
interactions.
The alarm or fright reaction, i.e., the response to specific alarm
substances, appears to be confined to the ostariophysan fishes (Pfeiffer
1963). Alarm pheromones are released from club cells in the epidermis
when damaged (e.g., by a predator), and lead to a distinct set of behavioral
responses: the alarmed conspecifics react by seeking cover, closer shoaling,
rapid swimming or ‘freezing’ behavior (i.e. they stop moving).
In a detailed comparative study with different populations of epigean
and hypogean Astyanax mexicanus and their hybrids (Fricke 1988, Fricke
and Parzefall 1989), it could be demonstrated that fish from all populations
studied stop feeding when schreckstoff is introduced into the tank. In the cave
fish, some of the behavior patterns found in surface-dwelling populations
seem to be reduced: cover seeking on the bottom, rapid swimming
(zigzags) and freezing are still found in the cave fish. All individual cave
fish tested avoided the site where the alarm substance was released (Fig.
4). In addition, a higher swimming activity was registered. It seems that
the alarm behavior in Astyanax mexicanus was modified as an adaptation
to life in darkness, i.e., in the absence of visual communication: the use of
visual signals to inform conspecifics about the predation event is reduced.
However, the avoidance reaction per se is still existent, even though avian

Fig. 4 Feeding rate at the water surface in two epigean and three cave populations
of Astyanax mexicanus. The dot marks introduction of food, the circle with dot
introduction of alarm substance, the square with dot introduction of skin extract of
Pterophyllum scalare a species without alarm substance (after Fricke 1988).
and piscine predation on cave fishes seems to be an exception (see Plath

and Tobler, this volume). The higher shoaling activity can be interpreted
as a general response to predation threat (Plath and Schlupp, 2008), and
may also lead the alarmed animals away from the predation area.
Up to now it was not possible to release this behavior in the field
(Langecker and Fricke, unpubl. data). Thinés and Legrain (1973) described
for the blind cave cyprinid Caecobarbus geertsi that alarm substance causes
foraging behavior to be orientated towards the bottom of the aquarium. It
seems tempting to speculate that these cavefish could experience predation

from above, namely by fish-eating bats, but this result could also indicate
the evolutionary persistence of a now functionless behavior.
II.7 Social Groups

For three species, Astyanax mexicanus (Characidae), Garra barreimiae
(Cyprinidae) and Poecilia mexicana (Poeciliidae), social groups have been
reported as the common form of social organization in the epigean forms.
Such groups of fish that remain together for social reasons are termed by
Pitcher (1983) as shoals. Following this definition, the epigean fish of these
species are mostly shoaling, and sometimes, when disturbed by predators,
they switch to schooling, which includes synchronized moving and escape
maneuvers (Parzefall 1979, 1983). Shoaling is strongly reduced in all cave
fish examined so far.
A detailed study has been made with the cave population of P. mexicana
to answer the question whether this lack of shoaling behavior is just due to
the absence of visual orientation, or whether a genetically based reduction
on the behavior level can be found. Two epigean populations of P. mexicana
have been compared with the extreme cave form from chamber XIII of
the Cueva del Azufre (see Plath and Tobler, this volume) and with their
hybrids (Parzefall 1993a). In light, the epigean fish from the Río Teapao
(PST) were strongly attracted by a stimulus shoal, but the tendency
to follow the shoal of conspecifics is already reduced in the epigean
population living near the cave entrance (El Azufre, PSO). This creek has
a milky appearance due to suspended sulfur, which may indeed provide
some protection from predation. Furthermore, increased food competition
in sulfidic habitats may explain this reduction (Plath and Schlupp 2008).
The cave fish showed similarly low shoaling (Fig. 5). In both populations
there was strikingly more variability than in the epigean fish. However, in
darkness no preference for the stimulus shoal has been registered in any
population examined. The F1 hybrids exhibited the same level of shoaling
as the epigean parental generation and significantly differed from the cave
fish. The F2 generation differed significantly from the putative epigean
ancestor in the mean shoaling tendency, as well as in the high variability.
There is only a significant difference in the cavefish backcross from the
epigean and the F1 generation. These data demonstrate a genetically
fixed reduction which takes already place in the milky Arroyo del Azufre
outside of the cave.
Also in a comparable study in different populations of Astyanax
mexicanus (Characidae), a genetically based reduction of the schooling/
shoaling in the cave forms of that species was found (Parzefall 1993a).

min)
Fig. 5 Tendency to follow a school of subadult Poecilia mexicana in min/10 min.

Test of single specimen in light (L) and darkness (D) of two epigean, one cave
population and hybrids (after Parzefall 1993a).
Timmermann et al. (2004) provide another example for the reduction of

shoaling behavior during the evolution of a cave fish from a previously
surface dwelling species. The surface form of Garra barreimiae (Cyprinidae)
showed a strong shoaling tendency even in darkness, but the cave form
showed only a weak preference to associate with a stimulus shoal.
In the subterranean species studied by Berti and Thinés (1980) and
Jankowska and Thinés (1982), shoaling/schooling was also found to be
absent: Caecobarbus geertsi (Cyprinidae), Barbobsis devecchi (Cyprinidae) and
Uegitglanis zammeranoi (Claridae). All these cave-dwellers show a random
spatial distribution in their habitats as well as in aquaria. However, in
the latter cases comparative data for the epigean relatives are needed to
evaluate if shoaling was indeed reduced. For example, in several catfishes,
also the epigean forms do not shoal.
After mapping behavioral character states into a phylogeny proposed
for the Heptapteridae, Trajano and Bockmann (1999) hypothesized
that the solitary habit observed for the Rhamdiini Pimelodella kronei is
an autapomorphy related to the cave life, probably as a response to an
increased intraspecific competition for food, since the epigean P. transitoria
catfishes aggregate while feeding. Nevertheless, the same solitary habits

exhibited by non-Rhamdiini heptapterids such as the Rhamdiopsis

spp. from Chapada Diamantina and from Toca do Gonçalo, would be
plesiomorphic traits inherited from a small sized, probably pedomorphic,
stream-dwelling ancestor. Therefore, for the latter the solitary habits are
not an adaptation for the food poor subterranean environment. This well
illustrates the need for phylogenetic schemes in order to interpret character
states observed in troglobites into a evolutionary context.
II.8 Reproductive Behavior

Cavefishes require behavioral adaptations that allow for effective
mate finding and subsequent fertilization in the absence of any visual
orientation.
In species living at high population density, males usually have an easy
game to find a female. The livebearing poeciliid fish Poecilia mexicana lives
in high population densities in its cave habitat. Like in other species of
“shortfin mollies”, already in the epigean form of P. mexicana any visual
sexual display is lacking (Parzefall 1969, Plath et al. 2007). The mating
behavior of P. mexicana males thus appears preadapted to cave life. The
males check conspecific females by nipping at the enlarged genital region,
the so-called genital pad (Parzefall 1969). Mature females that are in the
receptive stage of their monthly sexual cycle advertise their receptivity
chemically by substances emanating from their genital pad (Parzefall 1973).
A female accepting a male stops fleeing and allows the male to copulate.
Normally bigger males are preferred based on visual cues. Only in the cave
fish, females are able to perform mate choice in darkness (Fig. 6), which is
probably accomplished by switching from using the visual system to using
the lateral line and possibly also chemical cues (Parzefall 2001, Plath et al.
2004). For the female choice behavior we found some other differences
between cave and epigean populations in P. mexicana: only the females
from the cave population exhibited a preference for well nourished males
even under completely dark conditions (Plath et al. 2005). Overall, the
sexual activity of males is also reduced in cave living P. mexicana (Plath et
al. 2003). These behavioral changes are discussed in an ecological context
by Plath and Tobler (this volume).
Species with a lower population density in the cave habitats like the
characid fish Astyanax mexicanus (Fraipont 1992, Wilkens 1972) attract
conspecifics by chemical signals emitted into the water. In addition, sexually
active females mark the substrate by a chemical substance produced in the
genital region and restrict their movement to a small area, which may help
males localize the females. The swimming activity of the male increases
after contact with such a marked substrate or with a female with ripe

100
90
**
80 n.s.
*
70
**
% time near male 60

50
40
30
20
10
LM SM
0
(a) POx PS0 PSXIIIL PSXIIID
100
90
80
*
70 n.s.
n.s.
% time near male
*
60
50
40
30
20
10
LM SM
0
(b) POx PS0 PSXIIIL PSXIIID
100
90 ***
80 *** ***
70 ***
% time near male
60
50
40
30
20
10
LM SM
0
(c) POx PS0 PSXIIIL PSXIIID
Fig. 6 Choice tests with females of P. mexicana. Epigean and cavefishes had
the possibility to associate with a large (LM, black bars) or a small male (SM,
white bars). Surface population POx, cave entrance population PS0 and cave
population, reared in light PSXIIIL, or in darkness PSXIIID, * p < 0,05, ** p < 0,01,
n.s. p > 0,05. Paired t-test. a) stimulus males in wire-mesh cylinders in light, b) in
darkness, c) in clear Plexiglas cylinders in light (after Plath et al. 2004).

oocytes. The male is searching for the sexually active female. Upon finding
a female, it swims parallel to the female and tries to achieve contact using
fine hooks on his anal fin in order to spawn in the free water (Wilkens
1972). In a comparative study of six cave and one epigean population,
the female preference for large male body size in A. mexicanus has been
tested under different light conditions. Females from the surface form as
well as females from the phylogentically young and eyed Micos cave form
showed a strong preference for large males. When only non-visual cues
were presented in darkness, a preference for large males was found in
the case of two of the cave populations (Micos and Yerbaniz populations;
Plath et al., 2006 b). This shows that not all cavefishes necessarily switch
to using non-visual cues for mate choice, and that the evolution of mating
preferences may be different among different cave systems.
As already pointed out, most troglobitic fish are descendents of nocturnal
species, thus few modifications are expected to affect their reproductive
behavior. Unfortunately, few troglobitic fishes reproduce in laboratory and,
with the exception of the above discussed studies, detailed descriptions
of reproductive are mostly lacking. Among Brazilian subterranean fish,
spontaneous reproduction of three species occurred in laboratory, the
trichomycterids Ituglanis passensis and Trichomycterus itacarambiensis, and
the loricariid, Ancistrus cryptophthalmus from Passa Três Cave, which bred
three times in consecutive years. For the first two, the reproduction was
noticed only after the juveniles were independent (but still with non-
regressed eyes, supporting the notion that the ontogenetic eye regression
is a rule for troglobitic fish), but the whole reproductive events could
be recorded for the latter (Secutti and Trajano 2009). No differences in
reproductive behavior were observed between A. cryptophthalmus and
epigean Ancistrus catfish, which are typically rocky guarders presenting
parental care usually displayed by the male, including territorialism and
tending of eggs and newly hatched larvae. These would be preadaptive
traits retained by the cave species.
II.9 Agonistic Behavior

Aggressive behavior consists of different patterns: threatening postures
and attacks, which may be followed by fights if the attacked individual
does not take flight. The threatening animal in general enlarges his body
shape (e.g., by spreading its fins), shows striking, dark colors (in surface
animals) or sends impressive calls. Then the attacker typically attempts
to bite, strike or beat the opponent. To prevent an injury, fights can be
ritualized to techniques which allow measuring how strong an opponent
is. The weaker partner then has the chance to exhibit submissive behaviors

to bring the agonistic interactions to an end without being injured. Both

types of behavior patterns are combined under the term agonistic behavior
here. Agonistic behavior is used in a variety of contexts. One of them
is competition for food. Epigean Astyanax mexicanus (both males and
females) defend small feeding territories of ten to twenty centimeters in
diameter, depending on body size, by fin-spreading, snake-swimming and
ramming attempts. In escalated fights, biting and circling can be observed.
Submission is signaled by a head-up-position. In darkness the epigean fish
do not perform these behavior patterns (Burchards et al. 1985, Hausberg
1995). In the cave form, territories are smaller, and only signals effective
in close body contact are reported: biting, circling and tail-beating have
been observed, but no fin-spreading, snake-swimming and ramming.
Hausberg (1995) argues that the aggressive behavior of the cave fish is of
a new type that has evolved as an adaptation to life in a lightless habitat
after the reduction of the visually mediated aggressive behavior shown by
the epigean ancestor.
The maintenance, or even enhancement of agonistic interactions,
seems to be common for troglobitic fishes. Ercolini et al. (1981) studied
the aggressive behavior in the clariid catfish, Uegitglanis zammaranoi,
based on pairing individuals in a resident × intruder scheme. Using
the same laboratory method, complemented with observations in the
natural habitat, Trajano (1991) compared the agonistic behavior of P.
kronei and its epigean sister-species, P. transitoria. Both U. zammaranoi and
P. kronei present a dominant-subordinate relationship established through
aggressive interactions usually starting immediately after unknown catfish
are paired and expressed as the possession of the bottom by the dominant
fish, which banishes the subordinate to the upper levels of the aquarium.
At least in P. kronei, the dominance is primarily based on size, previous
residence being a secondary determinant. Agonistic interactions were also
observed in caves, with smaller catfish being chased away by the larger
ones. Likewise, some of the aggressive acts described for U. zammaranoi in
laboratory were also observed in the wells where the specimens have been
captured.
In laboratory, Pimelodella transitoria shows more or less the same type of
aggressive behavior as P. kronei (Fig. 7), but centered on dens. In epigean
habitats, however, the eyed catfish are seen sharing hiding places and
frequently aggregate during foraging, which is concentrated at dusk and
dawn (Gerhard 1999). It is possible that the exclusively solitary habits of
P. kronei were accompanied by an enhancement of aggressiveness towards
conspecifics.
A rich behavioral repertoire was observed for these species, and
behavioral components common for U. zammaranoi and P. kronei, also

Fig. 7 Aggressive behavior in Pimelodella kronei: zigzagging in head-to-tail (a)

and head-to-head position (b), ventral shove (c) and fin biting (d) (after Trajano
1991).
reported for epigean catfish such as Ictalurus spp., included “zig-

zagging”, “snout/head, caudal and lateral shoves”, “bites”, “patrolling”,
“chasing”, “mutual chasing”, “flight”, “avoidance”, “ascent/subordinate
swimming”. Categories exclusive to the former include “mouth-locking”
and “gapping”, and to the latter, “arched posture”. This would correspond
to a general pattern for aggressive siluriforms. A simplified version of this
repertoire have been recorded for Ancistrus cryptophthalmus catfish, which
has been seen defending food sources (edible leaves and commercial food

in laboratory, grazing stones in the cave habitat) from conspecifics using

lateral strokes, aided by extended opercular spines (odontodes), which
are particularly developed in Ancistrus catfish; some chasing of smaller
individuals by larger ones is also observed, mutual chasing and attempts
to thrust from beneath the opponent, corresponding to “ventral shoves”
are rare.
Less frequent but more intense agonistic interactions are exhibited by R.
enfurnada catfish: in laboratory, when two catfish meet, usually the larger
one chases away the smaller in a quick pursuit, or the smaller simply
avoids the other one. However, in some occasions, intense fights occur,
leading to the death of one fish in less than 24 h. The juveniles are more
tolerant to each other proximity (Bichuette and Trajano 2005).
Agonistic interactions were observed for all studied balitorids (Trajano
and Borowsky 2003), but they were more frequent and intense among the
cave species, especially in N. troglocataractus, which presented an apparent
territorial behavior. The latter are quite intolerant to the proximity of
conspecifics, displaying a rich repertoire of agonistic interactions which
include lateral and ventral shoves, lateral zig-zagging, bites, pursuits
(including circular pursuits) and body elevation (Trajano and Borowsky,
unpubl. data).
Competition for mates can result in defending breeding territories that
only females are allowed to enter. Another possibility is that the males
compete within a social group of conspecifics and eventually establish
a rank order, which is often correlated with the physical strength of the
group members, such or just their body size. In such a case only the high-
ranking male has unlimited access to the females.
In epigean Poecilia mexicana, the dominant male responds to an attractive
female by following behavior and dark body coloration. In the field, the
pair separates from the shoal and the male defends a territory around the
female, where the male is repeatedly nipping and attempts to copulate.
During aggressive encounters with more or less equal sized males, small
males adopting a female-like body coloration try to sneak copulations
(Parzefall 1973, 1979). Males in the cave population of P. mexicana do not
defend territories. In laboratory studies with epigean fish and cavefish
(which have functional eyes), a quantitative genetically based reduction
of the aggressive behavior was demonstrated (Parzefall 1974, 1979, 1993a).
The reduction of the aggressive pattern was found to be highly variable
within the population. Some of the cavefish tested in light did not perform
any aggressive pattern and it seemed that they did not “understand”
attacks by other (surface) males. Instead, they were answering aggression
by nipping and copulation attempts against the aggressor. Attractive

females can be seen followed by several males without any aggression

(Parzefall, unpubl. data). There is no other example in cave-dwellers up to
now with a quantitative reduction of the whole aggressive repertoire and
a reduction of the rank order.
No aggressive interactions clearly recognizable as such have been
observed for Brazilian troglobitic trichomycterids but, considering that
trichomycterid catfish in general are not aggressive, this is probably a
plesiomorphic trait of the troglobitic species. Occasional chasing of smaller
individuals by the larger ones was observed in Rhamdiopsis sp. from
Chapada Diamantina. On the other hand, it is not clear whether chasing
in Rhamdiopsis sp. from Toca do Gonçalo kept together are agonistic
interactions or cannibalistic attempts (Trajano and Bockmann 2000). In any
case, the absence of robust comparable data for epigean relatives hampers
further evolutionary interpretations of these behaviors.
A case for reduction of aggressive behavior in troglobitic fishes was
made by Bechler (1983) for amblyopsids, based on parings in laboratory.
The troglophilic Forbesichthys agassizi exhibit the whole set of observed
agonistic acts, which encompassed both aggressive (tail-beat, head-butt,
attack, bite, chase and jaw lock) and submissive (freeze, escape), while the
exclusively epigean C. cornuta exhibited none. F. agassizi repertoire was
also present in Amblyopsis spelaea, but some components were missing in
Typhlichthys subterraneus (jaw-lock was not observed) and A. rosae (only tail-
beat observed among the aggressive acts). It is noteworthy that in species
having similar overall repertoires, frequency, duration and intensity of
different acts vary between species. Hence, among cave species the higher
specialized amblyopsids, Typhlichthys subterraneus and Amblyopsis rosae,
are engaged in simpler, less intense bouts, which were considerably
shorter in length. A first step in the process of reduction of agonistic acts
could be a decreasing in intensity and duration of the interactions.
Considering F. agassizi as a representative of the behavior present in the
ancestor(s) of these troglobitic species, Bechler (op. cit.) concluded for a
reduction of the aggressiveness in T. subterraneus and, in a more advanced
form, in A. rosae, hypothesizing that such regression was correlated with a
reduction in the metabolic rates of these species – as the metabolic needs
decreased, these fish would be less food-limited, therefore intraspecific
competition would decrease. This is a very interesting and compelling
hypothesis, although the lack of generally accepted phylogenies for the
family (see Niemiller and Poulson, this volume) difficultates any proposal
of direction of evolutionary character changes in the troglobitic species.

III. HOW TO EXPLAIN BEHAVIORAL CHANGES IN

CAVEFISHES? A GENETIC APPROACH
The studies of changes in the behavior of cave-dwellers reported here
reveal that a complex trigger system releases the right behavioral response
on the basis of chemical and tactile stimuli. Many animals can therefore
survive in complete darkness by using non-visual signals. In some cases
the existing behavior like aggression in Astyanax mexicanus has changed
to a more efficient behavioral system. In all cases these behavioral
innovations seem to be derived from behavior patterns already existent
in the surface forms, and to date no completely new character has been
found in cavefishes.
For the traits presented in this chapter, the question arises whether the
behavioral changes can be explained by selection or on the basis of the
neutral mutation theory as developed by Kosswig (1948, 1963). Kosswig
hypothesized that the absence of stabilizing selection in darkness allows
for an accumulation of selectively neutral mutations, leading to increased
variance in non-selected traits, such as eye structure. Over time the
variability then decreases within a population because of genetic drift. In
phylogenetically old cave animals this may lead to homozygosity of alleles
that do not allow for the expression of a given trait, and thus, diminished
variability for the reduced patterns (Wilkens 1988). Other authors
favor selective explanations involving selection for increased metabolic
economy (Poulson 1963) or indirect effects of pleiotropy (Barr 1968, Jeffery
2005). A more detailed review of the relevant theories of the evolution
of subterranean animals has been given by Culver and Wilkens (2000).
There is only a general consensus that differentiation of cave populations,
especially with regard to regressive characters, cannot occur when there is
substantial gene flow between the cave and surface populations.
It seems that for the reduction of eyes, pigmentation, the circadian clock,
the dorsal light reaction, phototactic behavior and – in part – the lack of
schooling/shoaling, the neutral mutation theory is a good explanation
(but see Plath and Tobler, this volume, for an effect of food shortage on
the reduction of shoaling). For other characters presented here, directional
selection seems a plausible explanation.
IV. GENERAL CONCLUSIONS

The majority of troglobitic fishes belong to epigean taxa which are active at
dusk and/or night, photophobic, cryptobiotic (with strong hiding habits),
most of them bottom-dwellers. The analysis of behavioral character states

in these troglobitic species into an evolutionary context reveals several

apomorphies which can be related to their exclusively subterranean life:
• Weaker cryptobiotic habits, probably as a consequence of the
absence of visually oriented predation;
• Increased midwater activity observed in catfishes and balitorids,
probably as an adaptation to enhance the chances of finding food,
in the absence of ecologically important predators and midwater-
active competitors;
• Weaker internal time-control mechanisms at the circadian scale,
possibly due to an evolutionary regression affecting the oscillator(s)
itself (themselves) or to an uncoupling between the oscillators
and at least one of their related functions, such as the locomotor
activity;
• Decreased photonegative response to light;
• with some exceptions, tendency towards solitary habits and
moderate to high aggressiveness, probably related to a high
intraspecific competition for food and potential mates, which may
represent the maintenance of the character state already present
in the epigean ancestor or a behavioral specialization to the
subterranean life.
Part of the above-mentioned behavioral patterns would represent
adaptive (constructive) character states, as seems to be the case with
the increased midwater activity, the loss of generalized phobic reactions
and the enhanced aggressiveness, and part would be regressive traits
resulting from evolutionary processes similar to those affecting eyes and
melanic pigmentation, as in the case with the weaker cryptobiotic habits,
photophobia and internal time-control mechanism, when compared to
epigean relatives.
Only a minority has epigean ancestors with diurnal activity. After
colonizing subterranean habitats, they developed striking behavioral
changes like the reduction of schooling, dorsal light reaction and alarm
behavior. For the reproductive behavior, chemical and tactile releasers,
already working together with optical ones in the epigean ancestors,
assume a important position. It is striking that within these families up to
now no species has been found having a optical display which was able to
colonize the dark habitat successfully.
Acknowledgements
The first author (JP) would like to thank the DFG (German research board)
for financial support since many years, the Mexican government for

providing research and collecting permits over many years and M. Hänel
(Zoological Institute, Hamburg) for drawing the figures. Many thanks also
to Martin Plath for comments on the manuscript. The junior author (ET)
is greatly indebted to the continuous financial support from the Fundação
de Amparo à Pesquisa do Estado de São – FAPESP and Conselho Nacional
de Desenvolvimento Científico e Tecnológico – CNPQ.
References
Barr, T.C. 1968. The ecology and the evolution of troglobites. Evoutionary Biology
2: 35-102.
Bechler, D.L. 1983. The evolution of agonistic behavior in amblyopsid fishes.
Berti, R. and G. Thinés. 1980. Influences of chemical signals on the topographic
orientation of the cavefish Caecobarbus geertsi (Pisces, Cyprinidae). Experientia
36: 1384-1385.
Bessa, E. and E. Trajano. 2002. Light reaction and cryptobiotic habits in armoured
catfishes, genus Ancistrus, from caves in Central and Northwest Brazil
(Siluriformes: Loricariidae). Mémoires de Biospéologie 28: 29-37.
Bichuette, M.E. and E. Trajano. 2005. A new cave species of Rhamdia (Siluriformes:
Heptaperidae) from Serra do Ramalho, northeastern Brazil, with notes on
ecology and behavior. Neotropical Ichthyology 3(4): 587-595.
Biswas, J. 1991. Annual modulation of diel activity rhythm of the dusk active loach
Nemacheilus evezardi (Day): A correlation between day length and circadian
parameters. Proceedings of the Indian National Science Academy B, 5: 339-346.
Biswas, J., A.K. Pati, K. Pradhan and R.S. Kanoje. 1990. Comparative aspects of
reproductive phase dependent adjustments in behavioural circadian rhythms
of epigean and hypogean fish. Comparative Physiology and Ecology 15(4): 134-
139.
Borowsky, R. and E. Trajano. 2003. Biological rhythms in cavefishes from Thailand.
In: 2003 Joint Meeting of Ichthyologists and Herpetologists, ASIH.
Burchards, H., A. Dölle and J. Parzefall. 1985. Aggressive behaviour of an epigean
population of Astyanax mexicanus (Characidae, Pisces) and some observations
of three subterranean populations. Behavioral Processes 11: 225-235.
Cordiner, S. and E. Morgan. 1987. An endogenous circadian rhythm in the swimming
activity of the blind Mexican cavefish. In: Chronobiology and Chronomedicine. G.
Hildebrandt, R. Moog and F. Raschke (eds.). Verlag Peter Lang, Frankfurt am
Main. pp. 177-181.
Culver, D.C. 1982. Cave Life. Evolution and Ecology. Harvard University Press,
Cambridge, MA, USA and London.
Culver, D.C. and H. Wilkens. 2000. Critical review of the relevant theories of the
evolution of subterranean animals. In: Ecosystems of the World, Volume 30,
Subterranean Ecosystems, H. Wilkens, D.C. Culver and W.F. Humphreys (eds.).
Elsevier, Amsterdam pp. 381-398.

Erckens, W. and F. Weber. 1976. Rudiments of an ability for time measurement in the
cavernicolous fish Anoptichthys jordani Hubbs and Innes (Pisces, Characidae).
Experientia 32: 1297-1299.
Erckens, W. and W. Martin. 1982. Exogenous and endogenous control of swimming
activity in Astyanax mexicanus (Characidae, Pisces) by direct light response and
by a circadian oscillator. II. Features and time controlled behaviour of a cave
population and their comparison to an epigean ancestral form. Zeitschrift für.
Naturforschung C 37: 1266-1273.
Ercolini, A., R. Berti and A. Cianfanelli. 1981. Aggressive behaviour in Uegitglanis
zammaranoi Gianferrari (Clariid, Siluriformes), anophthalmic phreatic fish from
Somalia. Monitore Zoologico Italiano n.s. 14: 39-56.
Espinasa, L, P. Rivas-Manzano and H. Espinosa-Perez. 2001. A new blind cave
fish population of genus Astyanax: Geography, morphology and behavior.
de Fraipont, M. 1992. Réponse d’Astyanax mexicanus aux stimulation chimiques
provenant de groupes de congénères à différents stades du développement.
Mémoires de Biospéologie 19: 209-214.
Fricke, D. 1988. Reaction to alarm substance in cave populations of Astyanax
fasciatus (Characidae, Pisces). Ethology 76: 305-308.
Fricke, D. and J. Parzefall. 1989. Alarm reaction, aggression and schooling in
cave and river populations of Astyanax fasciatus and their hybrids. Mémoires de
Biospéologie 26: 177-182.
Gerhard, P. 1999. Ecologia de populações e comportamento de quatro espécies
de bagres Heptapterinae (Teleostei: Siluriformes) em riachos do Alto Vale do
rio Ribeira (Iporanga, São Paulo). M.Sc. Thesis, Instituto de Biociências, São
Paulo.
Hausberg, C. 1995. Das Aggressionsverhalten von Astyanax fasciatus (Characidae,
Teleostei): Zur Ontogenie, Genetik und Evolution der epigäischen und
hypogäischen Form. Ph.D. thesis, University Hamburg.
Heuts, M.J. 1951. Ecology, variation and adaptation of the blind fish Caecobarbus
geertsi Blgr. Annales Société Royale Zoologique Belgique: pp. 155-230.
Heuts, M.J. 1953. Regressive evolution in cave animals. Symposium Society of
Experimental Biology 7: 290-309.
Hoenen, S.M.M. and E. Trajano. 1995. Locomotor responses of Pimelodella spp.
from Southeastern Brazil to chemical signals of other individuals (Teleostei:
Siluriformes: Pimelodidae). Mémoires de Biospéologie 22: 57-63.
Hüppop, K. 1987. Food finding ability in cave fish (Astyanax fasciatus). International
Journal of Speleology 16: 59-66.
Hüppop, K. 1988. Phänomene und Bedeutung der Energieersparnis bei dem Höhlenfisch
Astyanax fasciatus (Characidae). Ph.D. thesis, University Hamburg.
Hüppop, K. and H. Wirkens. 1991. Bigger eggs in subterranean Astyanax fasciatus
(Characidae, Pisces). Zeitschrift für Zoologische Systematik und Evolutionsforschung
29: 280-288.

Jankowska, M. and G. Thinés. 1982. Etude comparative de la densité de groupes de

poissons cavernicoles et épigés (Characidae, Cyprinidae, Claridae). Behavioral
Processes 7: 281-294.
Jeffery, W.R. 2005. Evolution of eye degeneration in cavefish: the return of
pleiotropy. Subterranean Biology 3: 1-11.
Junge, P. 1993. Egg size and larval development of epigean and cave forms
of Rhamdia guatemalensis (Pimelodidae, Pisces). In: Trends in Ichthyology. H.
Schröder, J. Bauer and M. Schartl (eds.). GSF-Bericht 7/92, pp. 165-173.
Kosswig, C. 1948. Genetische Beiträge zur Präadaptationstheorie. Review of Faculty
of Science (Istanbul) Series B 5: 176-209.
Kosswig, C. 1963. Genetische Analyse konstruktiver und degenerative
Evolutionsprozesse. Zeitschrift für Zoologische Systematik und Evolutionsforschung
1: 205-239.
Lamprecht, G. and F. Weber. 1992. Spontaneous locomotion behavior in
cavernicolous animals: The regression of the endogenous circadian system. In:
The Natural History of Biospeleology, A.I. Camacho (ed.). Monografias Museo.,
Nacional de Ciencias Naturales, pp. 295-326.
Langecker, T.G. 1989. Studies on the light reaction of epigean and cave populations
of Astyanax fasciatus (Characidae, Pisces). Mémoires de Biospéologie 16: 169-176.
Langecker, T.G. 1990. Der Einfluss des Lichts bei der Evolution von Höhlenfischen.
Ph.D. thesis, University Hamburg.
Langecker, T.G. 1992. Light sensitivity of cave vertebrates. Behavioral and
morphological aspects. In: The Natural History of Biospeleology, A.I. Camacho
(ed.). Monografias Museo Nacional de Ciencias Naturales, pp. 295-326.
Langecker, T.G. 1993. Genetic analysis of the dorsal light reaction in epigean and
cave dwelling Astyanax fasciatus (Teleostei, Characidae). Ethology, Ecology and
Langecker, T.G. and G. Longley. 1993. Morphological adaptations of the Texas blind
catfishes Trogloglanis pattersoni and Satan eurystomus (Siluriformes, Ictaluridae)
to their underground environment. Copeia 1993: 976-986.
Langecker, T.G., H. Wilkens and J. Parzefall. 1996. Studies on the trophic structure
of an energy-rich Mexican cave (Cueva del Azufre) containing sulphurous
water. Mémoires de Biospéologie 23: 121-127.
Niemiller, M.L. and T.L. Poulson. Subterranean fishes of North America:
Amblyopsidae. This volume.
Parzefall, J. 1969. Zur vergleichenden Ethologie verschiedener Mollienesia- Arten
einschließlich einer Höhlenform von Mollienesia sphenops. Behaviour 33: 1-37.
Parzefall, J. 1973. Attraction and sexual cycle of Poeciliidae. In: Genetics and
Mutagenesis of Fish, J.H. Schröder (ed.). Springer, Berlin, pp. 177-183.
Parzefall, J. 1974. Rückbildung aggressiver Verhaltensweisen bei einer Höhlenform
von Poecilia sphenops (Pisces, Poeciliidae). Zeitschrift für Tierpsychologie 35: 66-
84.
Parzefall, J. 1979. Zur Genetik und biologischen Bedeutung des Aggressions-
verhaltens von Poecilia sphenops (Pisces, Poecilidae). Untersuchungen an

Bastarden ober- und unterirdisch lebenden Populationen. Zeitschrift für

Tierpsychologie 50: 399-422.
Parzefall, J. 1983. Field observation in epigean and cave population of the Mexican
characid Astyanax mexicanus (Pisces, Characidae). Mémoires de Biospéologie 10:
171-176.
Parzefall, J. 1993a. Schooling behaviour in population—hybrids of Astyanax
fasciatus and Poecilia mexicana (Pisces, Characidae, Poeciliidae). In: Trends
in Ichthyology. H. Schröder, J. Bauer and M. Schartl (eds.). GSF-Bericht 7/92,
pp. 297-303.
Parzefall, J. 1993b. Behavioural ecology of cave dwelling fishes. In: The Behaviour of
Teleost Fishes. T.J. Pitcher (ed.). Chapman and Hall, London, pp. 573-606.
Parzefall, J. 1998. Behavioural adaptation of cavefishes. In: Encyclopaedia
Biospeologica, v. II, C. Juberthie and V. Decu (eds.). Société de Biospéologie,
Moulis, pp. 1191-1200.
Parzefall, J. 2001. A review of morphological and behavioural changes in the cave
molly, Poecilia mexicana, from Tabasco, Mexico. Environmental Biology of Fishes
62: 263-275.
Parzefall, J., C. Kraus, M. Tobler and M. Plath. 2007. Photophilic behaviour in
surface- and cave-dwelling Atlantic mollies, Poecilia mexicana (Poeciliidae).
Journal of Fish Biology 71: 1225-1231.
Pati, A. 2001. Temporal organization in locomotor activity of hypogean loach,
Nemacheilus evezardi, and its epigean ancestor. Environmental Biology of Fishes
62: 119-129.
Peiffer, W. 1963. Vergleichende Untersuchung über die Schreckreaktion und
den Schreckstoff der Ostariophysen. Zeitschrift für Vergleichende Physiologie
47: 111-147.
Pitcher, T.J. 1983. Heuristic definitions of shoaling behavior. Animal Behaviour
31: 611-613.
Plath, M. and I. Schlupp. 2008. Parallel evolution leads to reduced shoaling
behavior in two cave dwelling populations of Atlantic mollies (Poecilia mexicana,
Poeciliidae, Teleostei). Environmental Biology of Fishes (online first).
Plath, M., J. Parzefall and I. Schlupp. 2003. The role of sexual harassment in cave-
and-surface dwelling populations of the Atlantic molly, Poecilia mexicana
(Poeciliidae, Teleostei). Behavior Ecology and Sociobiology 54: 303-309.
Plath, M., J. Parzefall, K.E. Körner and I. Schlupp. 2004. Sexual selection in
darkness? Female mating preferences in surface- and cave-dwelling Atlantic
mollies, Poecilia mexicana (Poeciliidae, Teleostei). Behavior Ecology and Sociobiology
55: 596-601.
Plath, M., K.U. Heubel, F.J. Garcia de Leon and I. Schlupp. 2005. Cave molly
females (Poecilia mexicana, Poeciliidae, Teleostei) like well-fed males. Behavior
Ecology and Sociobiology 58: 144-151.
Plath, M., U. Seggel, H. Burmeister, K.U. Heubel and I. Schlupp. 2006a.
Choosy males from underground: male mating preferences in surface-
and cave-dwelling Atlantic mollies (Poecilia mexicana). Naturwissenschaften
93: 103-109.

Plath, M., M. Rhode, T. Schröder, A. Taebel-Hellwig and I. Schlupp I. 2006b.

Female mating preferences in blind cave tetras Astyanax fasciatus (Characidae,
Teleostei). Behaviour 143: 15-32.
Plath, M., I. Schlupp, J. Parzefall and R. Riesch. 2007. Female choice for large body
size in the cave molly, Poecilia mexicana (Poeciliidae, Teleostei): influence of
species- and sex-specific cues. Behaviour 144: 1147-1160.
Plath, M. and M. Tobler. Subterranean fishes of Mexico (Poecilia mexicana,
Poeciliidae), this volume.
Poulson, T.L. 1963. Cave adaptation in amplyopsid fishes. American Midland
Poulson, T.L. 1985. Evolutionary reduction by neutral mutations: plausibility
arguments and data from amblyopsid fishes and lyniphiid spiders. NSS Bulletin
47(2): 109-117.
Proudlove G. Biodiversity and distribution of the world.
Pradhan, R.K., A.K. Pati and S.M. Agrawal. 1989. Meal scheduling modulation of
circadian rhythm of phototactic behaviour in cave dwelling fish. Chronobiology
International 6: 245-249.
Biology 68: 414-429.
Secutti, S. and E. Trajano. 2009. Reproductive behavior, development and
eye regression in the cave armored catfish, Ancistrus cryptophthalmus Reis,
1987 (Siluriformes: Loricariidae), breed in laboratory Neotropical Ichthyology
7(3): 479-490.
Schemmel, C.H. 1980. Studies on the genetics of feeding behaviour in the cave fish
Astyanax fasciatus f. anoptichthys. Zeitschrift für Tierpsychologie 53: 9-22.
Thinés, G. and J.M. Legrain. 1973. Effects de la substance d`alarme sur le
comportement de poissons cavernicoles Anoptichthys jordani (Characidae) et
Caecobarbus geertsi (Cyprinidae). Annales de Spéléologie 28: 271-282.
Timmermann, M., I. Schlupp and M. Plath. 2004. Shoaling behaviour in a surface-
and a cave-dwelling population of a barb, Garra barreimiae (Cyprinidae,
Teleostei). Acta Ethologica 7: 59-64.
Trajano, E. 1989. Estudo do comportamento espontâneo e alimentar e da dieta do
bagre cavernícola, Pimelodella kronei, e seu provável ancestral epígeo, Pimelodella
transitoria (Siluriformes, Pimelodidae). Revista Brasileira de Biologia 49: 757-769.
Trajano, E. 1991. The agonistic behaviour of Pimelodella kronei, a troglobitic catfish
from Southeastern Brazil (Siluriformes, Pimelodidae). Behavioural Processes
23: 113-124.
Trajano, E. 2003. Ecology and ethology of subterranean fishes. In: Catfishes, V. 2.
G. Arratia, B.G. Kapoor, M. Chardon and R. Diogo (eds.). Science Publs., Enfield,
812 p.
Trajano, E. 2005. Evolution of lineages. In: The Encyclopedia of Caves. D.C. Culver
and W.B. White (eds.). Elsevier Academic Press, San Diego, pp. 230-234.

Trajano, E. and L. Menna-Barreto. 1995. Locomotor activity pattern of Brazilian

cave catfishes under constant darkness (Siluriformes, Pimelodidae). Biological
Rhythm Research 26(3): 341-354.
Trajano, E. and L. Menna-Barreto. 1996. Free-running locomotor activity rhythms in
cave-dwelling catfishes, Trichomycetus sp., from Brazil (Teleostei, Siluriformes).
Biological Rhythm Research 27(3): 329-335.
Trajano, E. and F.A. Bockmann. 1999. Evolution of ecology and behaviour in
Brazilian cave Heptapterinae catfishes, based on cladistic analysis (Teleostei:
Siluriformes). Mémoires de Biospéologie 26: 123-129.
Trajano, E. and F.A. Bockmann. 2000a. Ecology and behaviour of a new cave
catfish, genus Taunayia, from northeastern Brazil (Siluriformes, Heptapterinae).
Ichthyological Explorations of Freshwaters 11: 207-216.
Trajano, E. and L. Menna-Barreto. 2000b. Locomotor activity rhythms in cave
catfishes, genus Taunayia, from eastern Brazil (Teleostei: Siluriformes:
Heptapterinae). Biological Rhythm Research 31(4): 469-480.
Trajano, E. and R. Borowsky. 2003. Behavior of cave fishes from Thailand
(Cypriniformes: Balitoridae, Cyprinidae). In: 2003 Joint Meeting of Ichthyologists
and Herpetologists, ASIH.
Trajano, E., L. Duarte and L. Menna-Barreto. 2005. Locomotor activity rhythms
in cave fishes from Chapada Diamantina, northeastern Brazil (Teleostei:
Siluriformes). Biological Rhythm Research 36(3): 229-236.
Trajano, E., N. Mugue, J. Krejca, C. Vidthayanon, D. Smart and R. Borowsky. 2002.
Habitat, distribution, ecology and behavior of cave balitorids from Thailand
(Teleostei: Cypriniformes). Ichthyological Explorations of Freshwaters 13(2): 169-
184.
Volpato, G.L. and E. Trajano. 2006. Biological rhythms. In: Fish Physiology, Volume
21. The Physiology of Tropical Fishes, A.L. Val, V.M. Val and D.J. Randall (eds.).
Elsevier, San Diego, pp. 101-153.
Weber, A., G.S. Proudlove and T.T. Nalbant. 1998. Morphology, systematic diversity,
distribution, and ecology of stygobitic fishes. In: Encyclopaedia Biospeologica, v. II,
C. Juberthie and V. Decu (eds.). Société de Biospéologie, Moulis, pp. 1179-1190.
Wilkens, H. 1972. Über Präadaptationen für das Höhlenleben, untersucht am
Laichverhalten ober- und unterirdischer Populationen des Astyanax mexicanus
(Pisces). Zoologischer Anzeiger 188: 1-11.
Wilkens, H. 1988. Evolution and genetics of epigean and cave Astyanax fasciatus
(Characidae, Pisces). Support for the neutral theory. Evolutionary Biology
23: 271-367.

CHAPTER 5
The Evolutionary Genetics of Cave
Fishes: Convergence, Adaptation
and Pleiotropy
Richard Borowsky
Department of Biology, New York University, NY 10003, USA
E-mail: rborowsky@nyu.edu
ABSTRACT
Cave animals have converged repeatedly on a troglomorphic suite of
characters including regression of eyes and pigmentation, augmentation
of other senses, and changes in morphology, metabolism and behavior.
Convergences are evolution’s “experimental replicates” and their study
permits the testing of hypotheses about the evolutionary process. The
genetic bases for these changes are the focus of this review. We cover
the Mendelian, quantitative and molecular genetics of troglomorphic
evolution, and the population genetics of cave forms, with an emphasis
on the Mexican tetra, Astyanax mexicanus. Molecular data is useful for
reconstructing the relationships of cave populations, but gene flow into and
out of cave populations can obscure relationships. We examine the roles
of selection and drift in troglomorphic evolution and conclude that both
processes are important. Direct selection against eyes in the cave probably
reflects energy conservation and the high metabolic cost of maintaining
the retina. Indirect selection through pleiotropy is probably a major factor
in all evolutionary adaptation to new environments, including the cave.
INTRODUCTION
The greatest tool of modern science is the replicated experiment, which
permits the testing and informed revision of hypotheses. Evolutionary

biology, however, is generally an observational science and most often lacks

this tool to test hypotheses about the evolutionary process. As detailed
by other authors in this volume and elsewhere, the typical evolutionary
responses to cave life, regardless of taxonomic group, are the reductions of
eyes and pigmentation, the augmentation of other compensatory senses,
and behavioral and metabolic changes relating to reduced food supply.
Thus, cave animals converge on a troglomorphic suite of characters, and it
is this convergence that makes them uncommonly valuable for the study
of evolutionary processes.
Figure 1 illustrates an example of this convergence, showing three
individual Astyanax mexicanus (Mexican tetra; Characidae), one from a
surface population and two from independently derived cave populations.
The two examples of cave fish are clearly convergent in reduction of eyes
and melanin pigmentation. Mutations in how many genes are involved in
these changes? Are they the same genes in the two different populations?
Are the mutations in coding or regulatory sequences? Can these examples
of evolution in response to a defined ecological change tell us something
general about the evolutionary process?
The cave fish system is a powerful one to approach such questions
because of convergence, which brings to evolutionary biology much of the
power of the replicated experiment. In effect, each cave species is a replicate
of an experiment that asks what happens when you turn out the lights.
Given replication, hypotheses formed by the study of one cave species can
be tested through their predictions by using other cave species. Unlike
parallel evolution, where several populations evolve towards the same
phenotype having started with similar or identical genetics (Schluter et al.
2004), the ancestral population genetic states of convergent populations
may be completely uncorrelated. Thus, converging replicates can be
completely independent and permit us to learn about the different genetic
ways that evolution can bring about phenotypic change. The power of
this approach in revealing the genetic changes that underlie multitrait
evolution is one of the main points of this chapter.
This chapter will touch on each of the three distinct but related areas
of evolutionary genetics as they relate to cave animals: the genetic
changes underlying phenotypic macroevolution, how ecology influences
population genetics, and the use of genetic information to infer the
phylogenetic relationships among populations or species. Most examples
will come from the well studied Astyanax mexicanus system, but information
from other cave animals will be added, where appropriate. We begin
with a discussion of what is known of the genetic bases of troglomorphic
traits, move on to discuss the population genetic and phylogeographic
implications, and finally what these findings reveal about troglomorphic
evolution, and evolution in general.

Richard Borowsky 117
Fig. 1 Individuals of Astyanax mexicanus from a surface population and from

the Molino and Pachón cave populations. The cave populations independently
converged on eye reduction, albinism, and decreased numbers of melanophores.
book.
THE GENETIC BASIS OF CAVE FISH EVOLUTION AND
ADAPTATION
Sadoglu was the first to study the genetics of troglomorphy in A.
mexicanus by crossing cave (in this paper, “cave” forms are troglobites)
and surface forms to obtain F1 and F2 generations (Sadoglu 1956, 1957b).
For most characters that differ between cave and surface A. mexicanus,
the F1 is intermediate between the two parents and the F2 exhibits a range
of variation. Some traits, however, exhibit unifactorial inheritance. The F2
progeny (n = 1065) derived from a cross between an albino Pachón cave
fish and a surface fish exhibited a 3:1 ratio of melanic (ab+/_) to albino
individuals (ab/ab), demonstrating that albinism in this population was
inherited as a simple recessive (Sadoglu 1957a). All melanic individuals

at seven days post-fertilization had black eyes while all albinos had pink
eyes. Thus, the same gene affecting melanin formation in the melanophores
also controlled melanin production in the retinal pigmented layer of eye
(RPE). The melanic fish exhibited a full range of pigmentation from almost
no melanophores in the skin to as many as in surface fish. Both classes had
individuals exhibiting a broad range of eye sizes from greatly reduced to
ones as large as in surface fish.
A second locus (“brown”) that affected the amount of melanin
produced was also documented by classic genetics (Sadoglu and Mckee
1969). Progeny of cave/surface crosses that were able to make melanin
(ab+/_) either had black or brown eye color as seven day larvae, according
to their genotype at this second locus (bw+/_ or bw/bw, respectively). Thus,
the presence or absence of melanin as well as its quantity or state were
both shown to be under control of single unlinked loci, while control of
melanophore numbers and eye size were demonstrably multifactorial.
The locus responsible for albinism in the Pachón and Molino
populations was later shown to be oca2 (Protas et al. 2006). The loss-of-
function mutations in both cave populations are deletions of whole or
major portions of exons in oca2, but interestingly, the mutational sites in
the two populations are different. Thus, albinism in the two populations
was achieved independently, although the same gene is involved.
Albinism is also known from the Japonés, Yerbaniz, Sabinos, and Curva
populations (Sadoglu 1957a, Wilkens and Strecker 2003, Protas et al.
2006, unpublished observations, R. Borowsky). The Japonés albino gene
does not complement the Pachón albino mutation, suggesting its basis
is also loss of oca2 function, but the oca2 coding sequence in the Japonés
population is intact and the basis for albinism in the population remains
unclear (Protas et al. 2006). The molecular bases of albinism in Sabinos,
Yerbaniz and Curva populations are unknown. Interestingly, oca2 is the
locus responsible for ocular and cutaneous albinism, the most common
form of albinism in humans (King et al. 2001).
The frequency of the albino allele may have changed in the Pachón
population in recent times. Sadoglu’s wild stock from Pachón was
obtained prior to 1956. She reports the genotypes of six fish used in three
crosses as five heterozygotes (ab/ab+) and one homozygote (ab/ab) and in
the discussion it is implied that the six were a random sample from her
stock (Sadoglu and Mckee 1969). The frequency of the wild type allele
in this sample was 0.417. In contrast, in a recent collection (2008) of cave
fish from Pachón only four out of 69 individuals were melanic, giving an
estimate of the frequency of the wild type allele of 0.029 (unpublished
observations, R. Borowsky).

Breeding and complementation studies have shown that the brown

mutation is present in the Pachón, Yerbaniz, Piedras, Sabinos, and Curva
caves, but not the Molino cave population (Sadoglu and Mckee 1969,
Wilkens and Strecker 2003). In Pachón, it is at high frequency (Sadoglu
and Mckee 1969). The gene responsible for the phenotype is as yet
unidentified.
Studies of the polarities of quantitative trait loci (QTL) suggest that
eyes are selected against in caves and that this selection is direct (Protas
et al. 2007). Selection may also operate against the production of melanin
in cave animals, but it might be indirect, as a consequence of the loss or
alteration of melanin in the eye. Melanin is an important constituent of the
retinal pigmented epithelium (RPE), which is in intimate contact with the
photoreceptor layer of the retina. Changing or eliminating melanin may
alter the ability of the RPE to interact properly with the photoreceptors and
alter eye structure. Thus, oca2 or bw alleles affecting melanin production
or structure may be selected for in the cave environment through negative
pleiotropic affects on eye development. This potential pleiotropic pathway
would be independent of genes affecting the number or distribution of
melanophores.
In contrast to the genes that affect melanin production, other
troglomorphic traits studied in this species are under multigenic control;
the F2 of cave/surface crosses exhibit broad variation in trait values
because of independent assortment of cave and surface alleles during
meiosis in the F1. Wilkens estimated the minimum numbers of genes
affecting four troglomorphic traits in several cave populations using
Lande’s modification of the Castle/Wright quantitative genetic method
(Lande 1981, Wilkens 1988). He determined that six to eight genes
controlled eye size differences between cave and surface populations, two
to three controlled differences in melanophore numbers, two controlled
differences in feeding behavior, and 11 to 12 controlled differences in
numbers of taste buds. These estimates are based on the assumptions of
full additivity and equal effects of the genes and, thus, are underestimates
of the true number of genes involved. Wilkens’ work clearly established
that many troglomorphic traits in A. mexicanus have a multigenic basis
and evolved because of changes at numerous loci. Presumably, this is
generally true of troglomorphic vertebrates.
Quantitative trait locus analyses of both the Molino and Pachón
populations confirmed that multiple unlinked genetic loci controlled
phenotypic differences between cave and surface fish for many traits
(Borowsky and Wilkens 2002, Protas et al. 2006, 2007, 2008). These include
the classic regressive traits of eye size and number of melanophores,
other traits that are presumably adaptive, like condition factor, olfactory

sensitivity to dissolved amino acids, and numbers of maxillary teeth, and

those of unknown significance such as depth of the caudal peduncle,
skull bone structure and number of anal fin rays. Differences between
cave and surface populations for these traits are controlled by between
two and thirteen identifiable QTL, and they are more numerous than the
earlier estimates from quantitative genetics. Thus, different components
of the suite of troglomorphic traits range from unifactorial to greatly
multifactorial.
Hybrids between cave populations of A. mexicanus, especially
geographically distant ones, exhibit some recovery of wild type phenotype
compared to their parents. Eye rudiments, for example, are larger in F1
hybrids between Sabinos and Pachón caves than they are in the parental
stocks (Wilkens 1971). This recovery of phenotype is exhibited in numbers
of melanophores as well (Wilkens and Strecker 2003), and derives from
complementation, reflecting the fact that the genes involved in character
change differ among populations. Hybridization between individuals
from Molino and Tinaja cave produced F1 with recovery of the eye so
complete that up to 40% of them had visual function as fry (Borowsky
2008). Complementation analyses revealed that the loci involved in eye
loss differ nearly completely among Molino, Pachón and Tinaja caves
(Borowsky 2008). Thus, evolution’s palette is enormously varied and
numerous genes are available to effect convergent trait changes.
GENETIC DIVERSITY IN CAVE FISH POPULATIONS

Populations of troglobites generally have lower genetic variability than
those of related surface species, but exceptions to the generalization exist,
especially among cave invertebrates (Culver 1982). For fishes, however,
the pattern of lower diversity in cave populations appears consistent.
Three studies of allozyme variation in amblyopsids, characids, and
trichomycterids directly compared related troglobitic, troglophilic
and epigean forms and showed that the troglobitic populations were
consistently less variable, averaging heterozygosity levels of about 0.01,
one sixth the average in surface populations (Avise and Selander 1972,
Swofford et al. 1980, Perez and Moodie 1993). Nucleotide diversity (p)
was estimated using RAPD technology (Borowsky 2001) in four cave and
eight surface species of balitorid fishes from Thailand (Borowsky and
Vidthayanon 2001). The cave species had significantly lower p values,
averaging about 20% of those found in surface species. Thus, in four
different families, estimates of nuclear DNA diversity using a variety
of methods revealed a consistent pattern of reduced variability in cave
fishes.

Astyanax mexicanus is particularly interesting in this regard because it

has multiple cave populations, some of independent origin. Thus, it allows
for replicate measures of the effects of cave occupation on population
genetic diversity. Comparisons of mitochondrial gene diversity between
multiple surface and eight cave populations of A. mexicanus showed
that the majority of the cave populations had fewer haplotypes and
lower nucleotide diversities than surface populations (Dowling et al.
2002, Strecker et al. 2003, 2004). There were a few exceptions: some cave
populations had diversities as high as in surface populations and the two
studied genes, cytochrome b and ND2, did not reveal consistent patterns
for the same cave populations. In general, however, the cave populations
with higher diversity were those known to receive eyed surface
conspecifics, such as Rio Subterráneo and Cueva Chica, or receive waters
from arroyos supporting surface populations, such as Tinaja and Sabinos.
Isolated, perched populations, such as Molino, Pachón and Curva had no
detectable mtDNA diversity. A study of nuclear DNA diversity in seven
cave populations of A. mexicanus and eight surface populations using both
RAPDs and microsatellite typing again showed lower variability in cave
than in surface populations (Panaram and Borowsky 2005).
There are three main determinants of genetic variability in isolated
populations: effective population size, fluctuations in size (bottleneck
effects), and mutation rate (Kimura and Ohta 1971). In general, effective
population size in cave species is probably significantly smaller than in
surface species because of limitations in extent of available habitat and
food supply, and the species’ restricted geographic range. Population
sizes of balitorid cave species in Thailand were estimated by visual census
and estimates of suitable habitat, and were generally on the order of
102 to 103 (Trajano et al. 2002). Mark-release-recapture of A. mexicanus in
Pachón cave led to estimates of a total population size on the order of
103 to 104 (Mitchell et al. 1977); although this estimate appears large, it is
dwarfed by the numbers of the same species in the surface rivers (personal
observation). Bottlenecks may also play a significant role in reducing
variability in cave populations, but more likely these would be episodic
in nature, as opposed to an initial founder effect supposed by some to be
significant in the origin of cave forms (Barr 1968). Differences in generation
time between surface and cave populations or other factors might lead to
differences in mutation rates, but it remains to be tested whether such
factors are significant in reducing cave fish genetic diversity.
In A. mexicanus, the issue of genetic diversity is complicated by gene
flow between cave and surface populations. Populations of A. mexicanus
in caves known to receive migrants from the surface populations were
more genetically variable than those in caves isolated from the surface,

suggesting that they had received measurable gene flow from epigean
populations (Panaram and Borowsky 2005). The same study showed
through analyses of microsatellite variation that gene flow between cave
and surface populations was bidirectional, with a greater flow of cave
alleles to surface populations than of surface to cave. These results are
significant for two reasons: First, gene flow from surface to subterranean
populations greatly increases their evolutionary potential by seeding
them with the genetic variation needed to evolve. Even a modest flow
of genes into a cave population from a much larger pool of surface fish
would introduce variants at a rate far exceeding that from mutation alone.
Second, the opportunity for genes to leave cave populations and return to
the surface means that evolutionary novelties evolved in cave populations
might have futures independent of their source populations.
As a single, although perhaps atypical, example of influx of genetic
material into cave populations from the surface, the Pachón cave
mitochondrion was derived from recent surface populations and is distinct
from those of more closely related cave populations (Dowling et al. 2002,
Strecker et al. 2003, 2004). No example of the ancestral cave haplotypes has
been detected in the Pachón population, although they may exist at low
frequency. Thus, it appears that the surface haplotype was fixed, or nearly
so, by a selective sweep. In contrast, the nuclear markers from Pachón
show clear affinities to other cave populations. One possible explanation
for this state is that early evolutionary changes in Pachón ancestors in
nuclear genes related to metabolic efficiency created a genetic background
within which the surface mitochondrial haplotype was more efficient
than the cave haplotype. Surface migrants into the cave population could
have then introduced nuclear and mitochondrial genes. Selection against
surface nuclear genes and for the surface mitochondrion would have led
to the present day situation. The phylogeography of cave populations and
the placement of Pachón are discussed below.
How can cave genes enter into surface populations? Cave fish have
little chance of survival in surface waters and less chance of successful
reproduction. The example of Cueva Chica suggests the answer. Chica’s
entrance is an insurgence, but one that receives runoff from an intermittent
arroyo, which contains no fish. The first pools encountered upon entering
the cave typically have only troglomorphic fish and it is only deeper in
the cave that hybrids between troglomorphic and surface fish are found.
Eyed fish are typically found in the last pool, which sumps (Breder 1943).
There is general agreement that eyed fish enter the cave at its deepest pool,
confluent with the aquifer and at the same level as the nearby Rio Tampaon
(Mitchell et al. 1977). Thus, the aquifer is at least sparsely populated
with eyed fish, which could serve as a conduit for genes moving from

one location to the next, within the aquifer. Surface fish occupying and
alternating between the two habitats at the edges of rivers or the shallow
pools (= tinajas) that are windows into the aquifers, could bring genes out
of cave populations.
In summary, cave fish populations generally have lower genetic
diversity than surface populations. It is probable that this reflects their
relatively small effective population sizes, resulting either from reduced
actual population sizes or from periodic fluctuations in numbers. Effective
population sizes of cave populations may be somewhat increased by gene
flow among cave populations. The balitorid, Schistura oedipus, is known
from five caves in NW Thailand. Mitochondrial DNA sequence data
reveals isolation by distance among the five populations and suggests
interpopulation gene flow (Borowsky and Mertz 2001).
Genetic diversity of cave populations may be significantly increased
by migration from surface populations of conspecifics, if they exist. If
so, this would enhance the evolvability of cave populations. Genes may
leak out of cave populations into populations of surface conspecifics, if
they exist, potentially allowing cave derived novelties to outlive the cave
populations.
MOLECULAR PHYLOGENETICS OF ASTYANAX

MEXICANUS POPULATIONS
In order to interpret the evolutionary history of cave adaptation in this
species and the genetic bases of the striking morphological convergences
exhibited by the present day cave populations, we must understand their
origins and relationships. How closely are they related to the surface
populations and when did they diverge?
The 29 known cave populations of A. mexicanus have a geographic range
of over 130 km in NE Mexico (Fig. 2). Although most of the populations
exhibit similar degrees of troglomorphy, including extreme rudimentation
of eyes and pigmentation, there are subtle morphological differences among
them that probably make each population unique and discriminable. For
example, multivariate statistical analysis of morphometric data from eight
of these populations, some geographically proximate, established that
every individual studied grouped unambiguously with other individuals
from its own population (Mitchell et al. 1977). The observation of reliable
differentiation among distinct populations that currently experience
barriers to gene flow raises legitimate questions of their evolutionary
relationships. Was each cave population derived from a separate invasion
of subterranean waters or were there fewer invasions, perhaps even one,
with subsequent expansion of the range though underlying aquifers? If

Fig. 2 The Sierra de El Abra region of NE Mexico illustrating the locations of

caves and other geographic features mentioned in the text.
there were range expansions, did they occur before the troglomorphic
phenotype was attained or was adaptation to subterranean life a necessary
step before the species could spread?
Two factors complicate unraveling these phylogenetic relationships:
the first is that karst terrain is continuously remodeling, and the present
geography and hydrology may differ greatly from that of the past when
the cave populations were evolving. As a single example of this, the
caves of the Sierra de Guatemala currently have no obvious underground
connection with those of the Sierra de El Abra; the ridges are separated

by Servilleta Canyon, cut by the Rio Boquilla (Fig. 2). A walk through the
canyon, however, affords views of numerous fossil channels of phreatic
origin on opposing walls which clearly provided past subterranean
connections between what are now the ridges (personal observation). The
second complicating factor is that at least eleven of the populations are
known to receive immigration from surface populations (Mitchell et al.
1977, Langecker et al. 1991, Panaram and Borowsky 2005), and there is
evidence of bidirectional gene flow between cave and surface populations
(Panaram and Borowsky 2005). Thus, information from geography is
certainly incomplete and even genetics may be a source of misinformation
about phylogeny because of possible lateral transmission of alleles.
Four studies have addressed the phylogenetic relationships of A.
mexicanus populations using a variety of molecular data including RAPDs,
mtDNA sequences and SSCPs, and microsatellite variants. Although the
results differed in detail from one to the next and sometimes different
cave populations were employed, an overall picture of relationships has
emerged. The studies show that there were at least four separate invasions
of the underground and that at least two different ancestral stocks of
surface A. mexicanus gave rise to the present day populations. Subsequent
discussion is best followed with reference to Fig. 2, which shows the
geographic locations of the Mexican populations.
An early study used RAPDs to examine the relationships of various A.
mexicanus cave and nearby surface populations (Borowsky and Espinasa
1997). It showed that cave populations of the Sierra de Guatemala
(Molino, Vasquez and Caballo Moro) were distinct from those of the El
Abra (Pachón and Tinaja), that populations of both of these cave groups
were distinct from surface fish, and that the cave fish of Rio Subterráneo
clustered with the surface populations and not with the other cave groups.
The data did not resolve whether the observed relationships were due
to shared ancestry or ongoing gene flow. Jeffery and his coworkers used
SSCP and sequence data from the mitochondrial ND2 gene to look at six
cave and eleven nearby surface populations (Dowling et al. 2002). They
found that three of the cave populations, Rio Subterráneo, Cueva Chica
and Pachón, clustered strongly with the surface populations while three
others, Sabinos, Tinaja and Curva made a separate cluster. The first surprise
in their results is that Subterráneo, Chica and Pachón are widely separated
geographically and surround the other cave cluster, including Sabinos,
Tinaja and Curva (Fig. 2). A second surprising result was the placement of
Pachón with surface fish, rather than Tinaja, as in the earlier study.
Strecker et al. studied some of the same populations using sequence
data from the mitochondrial gene cytochrome b and microsatellite markers
and were able to shed light on the apparently discrepant placement of

the Pachón population. Cytochrome b data clustered Pachón with Rio

Subterráneo, Chica and surface populations and put Sabinos and Tinaja in
a separate cluster, as in Dowling et al. (2002). In contrast, the microsatellite
data clearly placed Pachón with Sabinos and Tinaja, as in the earlier RAPD
study. Thus, the difference seems to reflect different histories of mtDNA
and nuclear DNA. The authors suggest that Pachón’s true affinities are
to the other populations of the central El Abra (including Tinaja, Sabinos
and Curva) and that Pachón’s surface mitochondrial genotype reflects
subsequent hybridization and fixation in the population via a selective
sweep. They also conclude that the central populations of the El Abra are
distinct from the flanking populations (Subterráneo, Chica and Molino)
and were derived from an ancestral surface stock inhabiting the region
earlier than the present-day stocks, which more recently gave rise to
the peripheral populations. Their second, and more extensive study of
cytochrome b sequence variation in numerous cave and surface populations
from the area and from southern Mexico and Central America (Strecker
et al. 2004) bears out these conclusions and shows that the present day
surface populations of Guatemala and Belize are more closely related to
the ancestral stock of the central El Abra cave populations than the surface
populations presently surrounding the El Abra.
In sum, these studies suggest that the present day cave populations of
A. mexicanus come from at least four separate subterranean invasions. The
present day populations of both central and northern El Abra originated
from the more ancient surface stock. It is not known whether this stock
made a single subterranean invasion and then subsequently expanded
underground to its present range, from Pachón in the north to Curva in
the south, or whether there were multiple invasions. The original surface
stock is believed to have gone extinct locally through climate change
(Strecker et al. 2004). Subsequently, a later invasion of a distinct stock
provided ancestors to the present day surface fish and the flanking cave
populations to the north (Molino), the south (Chica), and the west (Rio
Subterráneo). The geographic separation of the flanking cave populations
indicates that they were derived from at least three separate invasions of
the underground.
REGRESSIVE EVOLUTION: SELECTION OR DRIFT?

Darwin famously doubted that natural selection played a role in eye loss
in cave fishes; “As it is difficult to imagine that eyes, although useless,
could be in any way injurious to animals living in darkness, I attribute
their loss wholly to disuse” (Darwin 1859). Since then, there has been
much discussion and speculation about the roles of selection and drift in

the losses of eyes and pigmentation in cave animals [see papers in (Culver
1985) and review in (Culver et al. 1995)]. Recent work shows that both drift
and selection play important roles in trait regression. Selection may be
direct, or indirect, through pleiotropy.
If we focus on any cave associated trait such as eye reduction, we can ask
if the derived state came about through selection or genetic drift. We can
distinguish between the two possibilities, at least in theory, by determining
the polarities of the effects of allelic substitutions at the gene loci involved
in determining the trait state. In practice, this is done through QTL analysis,
which reveals the approximate genomic positions of the responsible loci.
We can define as a positive QTL one in which the homozygote for the
cave allele is more troglomorphic than the homozygote for the surface
allele, while a negative QTL is the opposite. Orr (1998) observed that QTL
analysis of a troglomorphic trait should reveal a consistent positive pattern
if the trait evolution was driven by selection but more mixed polarities if
the trait evolved through drift.
We compared polarities for eye size (E) and melanophore number
(M) QTL in a hybrid cross between Pachón cave and surface fish (Protas
et al. 2007). QTL affecting eye size were completely consistent in their
effects: in every case detected in the Pachón cross, and in others detected
subsequently in different crosses (unpublished observations, R. Borowsky),
the set of individuals homozygous for the cave allele had smaller eyes on
average then the set homozygous for the surface allele (Fig. 3). In contrast,
the polarities for M QTL were mixed. Most of the time the cave alleles
were associated with decreased numbers of melanophores, but sometimes
they were associated with increased numbers of melanophores. The
difference between the patterns for E and M was statistically significant
and we concluded that eye regression was driven by selection, but that
melanophore numbers evolved through drift (Protas et al. 2007). By which
we mean that melanophore number itself was not subject to selection. It
could be that some (or even all) of the allelic substitutions affecting M
were driven by selection on pleiotropic characters. While the hypothesis
of selection can be tested by experiment or observation, neutrality (drift)
is the null hypothesis and cannot be directly tested.
PLEIOTROPY IN MULTI-TRAIT EVOLUTIONARY CHANGE

The term “pleiotropy” has two related but distinct meanings: in the context
of functional biology it describes a locus with a gene product participating
in two or more developmental pathways (Carroll et al. 2005). In the context
of quantitative genetics, it refers to the phenomenon in which a single
allelic substitution affecting two or more distinct traits (Lynch and Walsh
1998). The second meaning is used here.

Scaled effects
Fig. 3 Scaled phenotypic values for genotypes at QTL affecting eye or lens
size (A) or melanophore numbers (B). SS = Surface homozygote; CC = cave
homozygote; SC = heterozygote. (From Protas et al. 2007.)
Pleiotropy has long been considered a potential mechanism to drive

regressive evolution in cave animals through indirect selection (Allee et al.
1949, Krekeler 1958, Wright 1964, Barr 1968, Culver 1982, Poulson 1985, Sket
1985, Borowsky and Wilkens 2002, Yamamoto et al. 2004) and evidence of
its more general importance in multitrait evolution is accumulating (Protas
et al. 2008). Jeffery (this volume) offers a clear example of a potential source
of pleiotropy in A. mexicanus regressive evolution: Hedgehog (Hh) gene
expression is upregulated during Pachón cave fish development resulting
in an expanded midline zone of expression. Increased Hh expression
enhances the development of midline structures involved with feeding
and gustation, increasing fitness, but at the expense of eye development.
Without stabilizing selection to maintain eye structure, the inevitable
result of this antagonistic pleiotropy is evolution towards smaller eyes.
While the elements of this model have only been tested for the Pachón
population and the genetic lesions responsible for eye regression are
different in the Tinaja and Molino populations (Borowsky 2008), the
proposed mechanism may prove to be an important contributing factor
in eye regression in other populations also because numerous genes can
affect the Hedgehog pathway.
Other ways in which pleiotropy might come about include signaling
systems with multiple signaling molecules and receptors having similar
structures and promiscuous binding (Ducrest et al. 2008), or through
changes in transcription factors with multiple roles (Carroll et al. 2005).
These are discussed in greater detail, below.
Pleiotropic effects are commonly observed in new mutations (Fitzpatrick
2004) and we should look beyond the narrow question of the degree to
which pleiotropy drives regressive evolution and inquire into its more

general roles in multi-trait evolution. The Astyanax model is a powerful

tool to investigate this question because we can identify numerous traits
that co-evolved during a defined period of time and also study their
genetics.
A key observation is that single trait QTL mapped for numerous traits
are significantly clustered in the genome (Protas et al. 2008). In addition,
more complex analyses cited in the same study detect numerous multitrait
QTL. A multitrait QTL is essentially a multivariate phenotype, based on
two or more traits, derived from transformation of the initial trait space
into a lower dimension in a manner analogous to principal component
analysis (Jiang and Zeng 1995, Korol et al. 1995, 1998, 2001, Ronin et al.
1998, 1999). For example, a model incorporating information on variation
in eye size (E), melanophore number (M) and maxillary tooth count (T),
in this case named EMT, would find genomic regions that had significant
effects on each of the component traits. Five significant EMT QTL were
detected in the Pachón cross (Protas et al. 2008). Five others were found
that control condition factor, number of anal fin rays, and numbers of
melanophores. Numerous other multitrait QTL were detected and in some
cases different multitrait QTL cluster revealing that as many as eight traits
can be influenced by genetic variation in the same region in the genome
(Protas et al. 2008).
The existence of these clusters and their varied phenotypic associations
were unanticipated results of our QTL mapping studies. While multitrait
QTL represent areas in the genome where allelic variation influences more
than one phenotypic character, they do not necessarily represent clusters of
different genes. Any cluster might have a multigenic basis or, alternatively,
result from a single genetic locus with pleiotropic effects. In the absence
of fine scale molecular characterization of the QTL we do not know how
much each mechanism contributes to the overall picture. Nevertheless, we
argued that the large numbers of such trait clusters, their rich phenotypic
composition, and the relatively short time to create gene clusters through
genomic rearrangement, suggest that a good proportion of them must be
due to single loci with pleiotropic effects (Protas et al. 2008).
An example of how such broad pleiotropic substitution effects could
come about is provided by the melanocortin system, in which the
product of one gene, pro-opiomelanocortin (POMC), is processed post-
translationally to yield a variety of biologically active peptides collectively
called melanocortins. These include the melanocyte-stimulating-hormones,
a-ACTH, and b-endorphin, and collectively have wide ranging effects on
a variety of systems, including skin pigmentation, stress response, energy
budgeting, feeding behavior, and the immune response (Takahashi and
Kawauchi 2006). The melanocortins stimulate cellular activity through

interaction with membrane bound melanocortin receptor molecules (MC1R

through MC5R), which have tissue specific expression patterns. Because
the MCRs exhibit minor expression in the “wrong” tissues and because of
a general promiscuity in binding of melanocortins to receptors, alterations
of POMC function may have widespread pleiotropic phenotypic effects
(Ducrest et al. 2008). As a second example, transcription factors, typically
play roles in multiple pathways at different stages of development or in
different tissues (Carroll et al. 2005). Allelic substitutions in such genes are
likely to have pleiotropic effects in diverse pathways.
Wright (1964), building on the observations of Emerson (Allee et al.
1949), predicted the existence of multitrait clusters and their importance
for regressive evolution, but his argument was more general. Wright
suggested a “practically universal pleiotropy,” using “pleiotropy” to refer
to the multiple affects of a single allelic substitution. He argued that the
selective coefficient of any new mutation would depend, therefore, upon
the aggregate fitness effects of its component traits.
Figure 4 illustrates the effects of a single mutation at a locus as
vectors showing the polarity of multiple phenotypic changes resulting
from substitution of the new allele for the old. Increased or decreased
phenotypic expression can be either beneficial or deleterious. In the
hypothetical example illustrated, beneficial changes include increased
sensitivity to dissolved amino acids in the water (A) and decreased eye
size (E), while deleterious changes include increased rate of weight loss on
low rations (W) and fewer maxillary teeth (T). The selective effect of the
deceased numbers of melanophores (M) is unknown. Depending upon
the relative importance of the beneficial and deleterious effects, this new
mutation could be selected for or against.
Figure 5 illustrates similar clusters of trait effects detected in the F2 of
a cross between a Pachón cave individual and a surface fish (Protas et
al. 2008). In each of seven different linkage groups we identified regions
with QTL affecting at least four of the following traits, E, A, T and M
(defined above). In six of the seven clusters there were effects on eye
size. Consistent with our previous observations (Protas et al. 2007), in
all six cases the substitution affect of the cave allele was to decrease eye
size, strengthening our earlier conclusion that eyes are directly selected
against in the cave environment. In three of the four clusters that had
affects on olfactory sensitivity, the cave allele conveyed greater sensitivity
(presumably advantageous for finding food in the dark). In one case
however, LgP14, sensitivity decreased, but this deleterious affect could be
offset by the other affects of decreased eye size and increased numbers of
maxillary teeth. In each of the seven clusters the cave genotype always was
associated with phenotypic changes conferring fitness benefits. Reading

Mutation
Open here
Fig. 4 A hypothetical example of the pleiotropic effects of mutation. An allele

at ground state (unmodified circle) is mutated to an allele with pleiotropic effects
(negative or positive) on five phenotypic traits: Amino acid sensitivity (A), number of
maxillary teeth (T), relative eye size (E), rate of weight loss on reduced rations (W),
and melanophore count (M). At the lower right, the circle is opened and stretched
into the form presented in Figure 5.
Fig. 5 Scaled values of cave allele substitution effects for four traits in QTL
clusters on seven linkage groups in Astyanax mexicanus: Amino acid sensitivity
(A), number of maxillary teeth (T), relative eye size (E), and melanophore count
(M). Substitution effects were scaled against phenotypic range in the mapping
F2 (Pachon × Surface cross). Positive values are phenotypic increases, negative
values are decreases. (From Protas et al. 2008.)
book.

Fig. 6 Contd..
the linkage groups from left to right in Fig. 5, there were: 2+ 1–; 1+ 1–; 3+;
2+ 1–; 2+; 2+; 2+; 2+.
While these clusters could represent regions with tightly linked genes
affecting each trait separately, the weight of the evidence is that pleiotropy
is making a strong contribution to the phenomenon (Protas et al. 2008).
Of course, one of the strongest arguments for the role of pleiotropy in
these trait clusters is the presence of deleterious effects (LgP7, LgP9, and
LgP14). It is hard to explain how alleles causing deleterious effects could
come to predominate in the cave environment without indirect selection
through pleiotropic beneficial phenotypic correlates.
These observations on multitrait QTL clusters in the Pachón cave
fish genome are consistent with Wright’s predictions. This hypothesis,
which has broad implications for the study of morphological evolution,
multitrait adaptation and adaptive radiation, makes testable predictions
about the presence and distribution of similar multi-trait clusters in other
populations of cave fishes. The predictions are that multitrait QTL of

Fig. 6 Contd..
Fig. 6 Genetic map of Astyanax mexicanus showing alignment of multi-trait QTL

detected in the mapping progenies of crosses between surface fish and cave fish
from Pachón, Tinaja, or Molino. QTL were detected as previously described (Protas
et al. 2008) and hand placed on an integrated map, based on positions determined
on cross specific maps. The bars cover + one LOD intervals. Multitrait QTL were
based on combinations of two or three traits: E = eye size; M = melanophore count;
T = number of maxillary teeth; D = depth of the caudal peduncle; P = placement of
the dorsal fin. QTL are coded, PACH, MOLI and TINA to identify the cross in which
they were detected. Traits are fully defined and methods are detailed in (Protas
et al. 2008).
book.
similar trait composition will be observed in crosses between surface fish

and other cave populations, but that they will map to different areas of
the genome. The similarity in trait composition is expected because the
same environmental factors will be driving evolution of the separate cave
populations. The expectation is that they will map to different regions
reflects the independent origins of the relevant mutations.
These predictions were tested by mapping multiple-trait QTL in
separate crosses between surface fish and cave fish from three different
populations. Two of the mapping progenies were previously reported on:

Pachón X Surface F2 and Molino X Surface BC (Protas et al. 2006, Protas et

al. 2008). The third cross was a Tinaja X Surface F2, with 285 young with
207 markers, not previously reported on. In all of these crosses, multi-trait
QTL were detected and mapped using the methods described in Protas
et al. (2007, 2008). The QTL mapped were EM (eye size and melanophore
number), ET (T = number of maxillary teeth), EC (C = condition factor),
ED (D = depth of the caudal peduncle), EP (P = placement of the dorsal
fin), and EMT. The measurement of these traits and the methods of QTL
detection have been described previously (Protas et al. 2008). The present
analysis is based on an increased number of markers compared to the
earlier analyses. Because the mappable markers differed somewhat among
the crosses, the QTL were hand placed on a composite map to make
their distributions comparable among crosses. Thus, the positions are
approximate, but accurate enough to test for positional overlaps between
populations.
A total of 44 QTL were detected in the three crosses (Fig. 6). Most (30)
were in Pachón, which was the largest cross, and fewest were in Molino
(2), which was the smallest cross. Considering all three crosses, eleven
sites were found in the genome with EMT QTL. One pair of EMT QTL for
Pachón and Tinaja co-mapped to linkage group B (LgB) and another pair
for Molino and Tinaja co-mapped to LgX. The remaining three Tinaja and
seven Pachón EMT QTL did not co-map. One pair of EP QTL from Pachón
and Tinaja overlap in LgD. The remaining three from each of the two
populations map to different regions. In addition, one Pachón EMT QTL
barely overlaps with Tinaja ED and EM QTL in LgA. Aside from these,
there is little evidence that multi-trait QTL from the different crosses that
contain the same constituent traits co-map. This supports the pleiotropy
hypothesis that predicts that multi-trait QTL can arise in many different
places throughout the genome.
SUMMARY AND CONCLUSIONS

The shift from surface life to cave life poses two challenges for animals: they
lose the ability to see, and they must survive on relatively restricted rations.
Adaptations to cave life include regression of the eyes, augmentation of
other senses, and changes in metabolism. Correlated changes include
reductions in pigmentation and changes in body proportions and meristic
characters. These may or may not be adaptations. In Astyanax mexicanus,
some of these evolutionary changes, such as albinism or the production
of phaeomelanin are due to mutations at single loci. Most of the other
changes, such as eye and melanophore regression, changes in body
proportions and number of teeth, increased olfactory acuity, and changes
in both anabolic and catabolic metabolism, have a multigenic basis.

Populations of cave fishes are generally smaller than populations of

related surface fishes, presumably because of reduced habitat and food
availability. In consequence, they have lower genetic variability. Bottlenecks
caused by periodic reductions in population size may contribute to
reduced effective population sizes and lower genetic variability further. In
A. mexicanus, there is good evidence of bi-directional gene flow between
surface and cave populations. Gene flow into cave populations helps
maintain genetic diversity and their potential for further evolutionary
change. Gene flow from cave to surface populations potentially allows
cave derived alleles an evolutionary longevity greater than that of the
cave population.
In the case of A. mexicanus, phylogeographic analyses show that current
cave populations were derived from at least two separate stocks of surface
fish and at least four separate subterranean invasions. Thus, fish from the
29 existing surface populations resemble one another not only because of
common ancestry, but because of strong convergences.
Evidence from QTL analysis shows that eye regression occurred because
of direct selection against eyes in the cave environment. In contrast, the
pattern of reduction in numbers of melanophores was more haphazard
and does not suggest direct selection against the cell type. Reduction in
numbers of melanophores might have come about through drift or indirect
selection through pleiotropy.
Pleiotropy plays a significant role in both regressive and constructive
evolution, as evidenced by the existence in the cave fish genome of
numerous areas having concentrations of QTL for both types of cave
related traits. While some of these trait clusters undoubtedly reflect the
effects of tightly linked genes, there are too many to explain through
linkage. The pleiotropy hypothesis predicts the existence of multitrait QTL
with similar trait contents in independently derived populations and that
these independently derived multitrait QTL will map to different genomic
regions. Data comparing the locations of multitrait QTL in three different
cave populations are consistent with this prediction.
While pleiotropy is often cited in discussions of the evolution of
troglomorphy as a mechanism to drive trait regression through indirect
selection, this is a narrow view of the importance of the phenomenon.
Pleiotropy is probably a key feature of multitrait evolution in all species.
One of the principal features of cave species is their evolutionary
convergence on troglomorphy. This gives their study some of the power of
experimental replication to test hypotheses about evolutionary processes.
In Astyanax mexicanus there is the extra advantage that the multiple
convergent cave populations are interfertile with one another and with
eyed surface fish. These features make the A. mexicanus system an especially

powerful tool for illuminating the genetic bases of cave adaptation and for
testing evolutionary hypotheses and, in general.
Acknowledgements
This work was supported by grants from the NIH (R03EY016783-01) and
the NSF (IOS-0821939).
References
Allee, W.C., A.E. Emerson, O. Park, T. Park and K.P. Schmidt. 1949. Principles of
Animal Ecology. W.B. Saunders Co., Philadelphia.
Avise, J.C. and R.K. Selander. 1972. Evolutionary genetics of cave-dwelling fishes
of genus Astyanax. Evolution 26: 1-19.
Barr, T.C. JR. 1968. Cave ecology and the evolution of troglobites. In: Dobzhansky,
T., Hecht, M.K. and Steere, W.C. (eds). Evolutionary Biology, Plenum Press, New
York, pp. 35-102.
Borowsky, R. 2008. Restoring sight in blind cavefish. Current Biology 18: R23-24.
Borowsky, R. and L. Espinasa. 1997. Antiquity and Origins of Troglobitic Mexican
Tetras, Astyanax fasciatus. In: 12th International Congress of Speleology.
International Union of Speleology/Swiss Speleological Society, La Chaux-de-
Fonds, Switzerland. pp. 359-361.
Borowsky, R. and H. Wilkens. 2002. Mapping a cave fish genome: Polygenic
systems and regressive evolution. The Journal of Heredity 93: 19-21.
Borowsky, R.B. and L. Mertz. 2001. Genetic differentiation among populations
of the cave fish Schistura oedipus (Cypriniformes: Balitoridae). Environmental
Borowsky, R.L. 2001. Estimating nucleotide diversity from random amplified
polymorphic DNA and amplified fragment length polymorphism data.
Borowsky, R.L. and C. Vidthayanon. 2001. Nucleotide diversity in populations of
balitorid cave fishes from Thailand. Molecular Ecology 10: 2799-2805.
Breder, C.M. 1943. Apparent changes in phenotypic ratios of the characins at the
type locality of Anoptichthys jordani Hubbs and Innes. Copeia 1943: 26-30.
Carroll, S.B., J.K. Grenier and S.D. Weatherbee. 2005. From DNA to Diversity:
Molecular Genetics and the Evolution of Animal Design. Blackwell Science, Oxford.
2nd edition.
Cambridge, Mass.
Culver, D.C. (ed.). 1985. Regressive Evolution. National Speleological Society
Bulletin 47, No. 2.
Culver, D.C., T.C. Kane and D.W. Fong. 1995. Adaptation and Natural Selection in
Caves. Harvard University Press, Cambridge, Mass.

Darwin, C. 1859. On the Origin of Species by Means of Natural Selection, or the

Preservation of Favoured Races in the Struggle for Life. John Murray, London.
Dowling, T.E., D.P. Martasian and W.R. Jeffery. 2002. Evidence for multiple genetic
forms with similar eyeless phenotypes in the blind cavefish, Astyanax mexicanus.
Molecular Biology and Evolution 19: 446-455.
Ducrest, A.L., L. Keller and A. Roulin. 2008. Pleiotropy in the melanocortin system,
coloration and behavioural syndromes. Trends in Ecology & Evolution (Personal
edition) 23: 502-510.
Fitzpatrick, M.J. 2004. Pleiotropy and the genomic location of sexually selected
genes. American Naturalist 163: 800-808.
Jiang, C.J. and Z.B. Zeng. 1995. Multiple-trait analysis of genetic-mapping for
quantitative trait loci. Genetics 140: 1111-1127.
Kimura, M. and T. Ohta. 1971. Theoretical Aspects of Population Genetics. Princeton
University Press, Princeton.
King, R.A., V.J. Hearing, D.J. Creel and W.S. Oetting. 2001. Albinism. In: The
Metabolic and Molecular Bases of Inherited Disease. C.R. Scriver, A.L. Beaudey, W.S.
Sly and D. Valle (eds.). McGraw Hill, New York. pp. 5587-5627.
Korol, A.B., Y.I. Ronin and V.M. Kirzhner. 1995. Interval mapping of quantitative
trait loci employing correlated trait complexes. Genetics 140: 1137-1147.
Korol, A.B., Y.I. Ronin, E. Nevo and P.M. Hayes. 1998. Multi-interval mapping of
correlated trait complexes. Heredity 80: 273-284.
Korol, A.B., Y.I. Ronin, A.M. Itskovich, J. Peng and E. Nevo. 2001. Enhanced
efficiency of quantitative trait loci mapping analysis based on multivariate
complexes of quantitative traits. Genetics 157: 1789-1803.
Krekeler, C.H. 1958. Speciation in cave beetles of the genus Pseudanophthalmus
(Coleoptera: Carabidae). American Midland Naturalist 59: 167-189.
Lande, R. 1981. The minimum number of genes contributing to quantitative
variation between and within populations. Genetics 99: 541-553.
Langecker, T.G., H. Wilkens and P. Junge. 1991. Introgressive hybridization in
the Pachon cave population of Astyanax fasciatus (Teleostei: Characidae).
Ichthyological Explorations of Freshwaters 2: 209-212.
Lynch, M. and B. Walsh. 1998. Genetics and Analysis of Quantitative Traits. Sinauer,
Sunderland, M.A.
Mitchell, R.W., W.H. Russell and W.R. Elliott. 1977. Mexican Eyeless Characin Fishes,
genus Astyanax: Environment, Distribution, and Evolution. Texas Tech Press,
Lubbock.
Orr, H.A. 1998. Testing natural selection vs. genetic drift in phenotypic evolution
using quantitative trait locus data. Genetics 149: 2099-2104.
Panaram, K. and R. Borowsky. 2005. Gene flow and genetic variability in cave
and surface populations of the Mexican Tetra, Astyanax mexicanus (Teleostei:
Characidae). Copeia 2005: 409-416.
Perez, J.E. and G.E. Moodie. 1993. Genetic variation in a cave-dwelling Venezuela
catfish. Zoologia 44: 28-31.


arguments and data from amblyopsid fishes and linyphiid spiders. The NSS
Bulletin 47: 109-117.
Protas, M., M. Conrad, J.B. Gross, C. Tabin and R. Borowsky. 2007. Regressive
evolution in the Mexican cave tetra, Astyanax mexicanus. Current Biology 17:
452-454.
Protas, M., I. Tabansky, M. Conrad, J.B. Gross, O. Vidal, C.J. Tabin and R. Borowsky.
2008. Multi-trait evolution in a cave fish, Astyanax mexicanus. Evolution &
Protas, M.E., C. Hersey, D. Kochanek, Y. Zhou, H. Wilkens, W.R. Jeffery, L.I. Zon, R.
Borowsky and C.J. Tabin. 2006. Genetic analysis of cavefish reveals molecular
convergence in the evolution of albinism. Nature Genetics 38: 107-111.
Ronin, Y.I., A.B. Korol and E. Nevo. 1999. Single- and multiple-trait mapping analysis
of linked quantitative trait loci: Some asymptotic analytical approximations.
Genetics 151: 387-396.
Ronin, Y.I., A.B. Korol and J.I. Weller. 1998. Selective genotyping to detect
quantitative trait loci affecting multiple traits: Interval mapping analysis.
Theoretical and Applied Genetics 97: 1169-1178.
Sadoglu, P. 1956. A preliminary report on the genetics of the Mexican cave
characins. Copeia 1956: 113-114.
Sadoglu, P. 1957a. A Mendelian gene for albinism in natural cave fish. Experientia
13: 394.
Sadoglu, P. 1957b. Mendelian inheritance in the hybrids between the Mexican blind
cave fishes and their overground ancestors. Verhandlungen Deutsche Zoologische
Gesellschaft, Graz, 1957. Zool. Anz., Supplement: 21: 432-439.
Sadoglu, P. and A. Mckee. 1969. A second gene that affects eye and body color in
Mexican blind cave fish. Journal of Heredity 60: 10-14.
Schluter, D., E.A. Clifford, M. Nemethy and J.S. McKinnon. 2004. Parallel evolution
and inheritance of quantitative traits. American Naturalist 163: 809-822.
Sket, B. 1985. Why all cave animals do not look alike a discussion on adaptive
value of reduction processes. The NSS Bulletin 47: 78-85.
Strecker, U., L. Bernatchez and H. Wilkens. 2003. Genetic divergence between
cave and surface populations of Astyanax in Mexico (Characidae: Teleostei).
Molecular Ecology 12: 699-710.
Strecker, U., V.H. Faundez and H. Wilkens. 2004. Phylogeography of surface and
cave Astyanax (Teleostei) from Central and North America based on cytochrome
b sequence data. Molecular Phylogenetics and Evolution 33: 469-481.
Swofford, D.L., B.A. Branson and G.A. Sievert. 1980. Genetic differentiation of
cavefish populations. Isozyme Bulletin 13: 109-110.
Takahashi, A. and H. Kawauchi. 2006. Evolution of melanocortin systems in fish.
General and Comparative Endocrinology 148: 85-94.
Trajano, E., N. Mugue, J.K. Krejca, C. Vidthayanon, D. Smart and R. Borowsky. 2002.
Habitat, distribution, ecology and behavior of cave balitorids from Thailand
(Teleostei: Cypriniformes). Ichthyological Explorations of Freshwaters 13: 169-184.

Wilkens, H. 1971. Genetic interpretation of regressive evolutionary processes –

Studies on hybrid eyes of 2 Astyanax cave populations (Characidae: Pisces).
(Characidae, Pisces) – support for the neutral mutation theory. Evolutionary
Biology 23: 271-367.
Wilkens, H. and U. Strecker. 2003. Convergent evolution of the cavefish
Astyanax (Characidae: Teleostei): Genetic evidence from reduced eye size and
pigmentation. Biological Journal of the Linnean Society 80: 545-554.
Wright, S. 1964. Pleiotropy in the evolution of structural reduction and of
dominance. American Naturalist 98: 65-69.
Yamamoto, Y., D.W. Stock and W.R. Jeffery. 2004. Hedgehog signalling controls eye
degeneration in blind cavefish. Nature 431: 844-847.

CHAPTER 6
Development as an Evolutionary
Process in Astyanax Cavefish
William R. Jeffery* and Allen G. Strickler
Department of Biology, University of Maryland, College Park,
MD 20742 USA
E-mail: Jeffery@umd.edu
ABSTRACT
Astyanax mexicanus cavefish and their surface dwelling counterparts are
a model system for studying the role of development in the evolution
of cave related phenotypes. In contrast to many other cave animals,
Astyanax is an excellent laboratory organism with frequent and copious
spawning and large robust embryos. Here we review molecular, cellular
and developmental mechanisms underlying the loss of eyes and
pigmentation in Astyanax cavefish. Eye degeneration involves enhanced
activity of the Hedgehog signaling system along the embryonic midline,
which induces lens apoptosis and ultimately causes the arrest of optic
growth and changes in craniofacial morphology. The loss of melanophore
pigmentation is caused by a specific block in the ability of undifferentiated
melanoblasts to convert L-tyrosine to L-DOPA, a precursor of melanin. We
discuss the importance of a developmental perspective in understanding
the evolutionary mechanisms underlying eye and pigment regression.
INTRODUCTION
Evolution creates novel phenotypes by changing development. Therefore,
knowledge of development is a prerequisite for understanding the
mechanisms of evolutionary change. Most studies of cave animals
have focused on problems in evolution, behavior, ecology, taxonomy,

and conservation. There are several reasons for the virtual absence of
developmental perspectives in cave biology. First, embryonic and juvenile
stages are required for the study of development, and they are difficult
to collect in caves. Second, many cave animals are known to produce a
reduced complement of embryos that develop very slowly relative to
their surface dwelling counterparts. Developmental biologists prefer to
focus on species that have many progeny and develop rapidly. Third,
some cave animals tend to brood embryos rather than freely spawn their
eggs, compounding the difficulty in studying early developmental events.
Finally, unlike many other areas of cave biology, developmental biology is
a laboratory science that requires experimental analysis and an organism
that thrives and reproduces in the laboratory. There are few such model
organisms available in cave biology. One of these is the Mexican cavefish
Astyanax mexicanus (sometimes called A. fasciatus and referred to here as
Astyanax), which has many favorable attributes for developmental analysis
(Jeffery 2001, 2008).
Astyanax has a surface dwelling form (surface fish) and 30 known
populations of cave dwelling forms (cavefish). Surface fish adults have
large eyes and three different types of pigment cells, whereas cavefish have
reduced or lost these traits, a phenotype shared with many diverse cave
animals (Fig. 1). Cavefish have also gained some less obvious constructive
features, such as large jaws, more taste buds, and possibly a more highly
developed olfactory system. Background information on the morphology,
Fig. 1 Astyanax surface fish (top) and Pachón cavefish (bottom) adults. Note the
absence of eyes and black melanin pigment in cavefish.

William R. Jeffery and Allen G. Strickler 143
biogeography, evolution, and classical genetics of Astyanax cavefish can be

found in Wilkens (1988).
Astyanax cavefish are named after their cave of origin (Mitchell et al.
1977). For example, Pachón, Chica, Los Sabinos, and Subterráneo, and
Molino cavefish are found in La Cueva de El Pachón, La Cueva Chica,
La Cueva de los Sabinos, La Cueva de la Subterráneo, and Sótano de El
Molino, respectively. In this article, unless another cavefish population is
named specifically, reference to cavefish will imply the Pachón population,
which has been the subject of considerable developmental analysis.
Evidence from a variety of sources suggests that some Astyanax cavefish
populations have evolved cave related phenotypes independently
(Wilkens 1971, Dowling et al. 2002, Strecker et al. 2003, 2004, Borowsky
2008). Thus, Astyanax cavefish are also an excellent model system to study
convergent evolution of developmental mechanisms.
Astyanax can be raised on a simple diet, spawns frequently and
abundantly, and has a relatively short generation time (about 4-6 months).
It has large, clear robust embryos that develop similarly to those of other
teleosts and are suitable for a variety of experimental manipulations,
such as microsurgery (Yamamoto and Jeffery 2000, Yoshizawa and Jeffery,
2008) and in vitro culture of developing tissues and organs (McCauley et al.
2004). Cavefish embryos can be contrasted to surface fish embryos in the
same way that mutant and wild-type phenotypes are compared in genetic
analysis. The fairly close phylogenetic relationship between Astyanax and
zebrafish permits rapid cloning of genes identified in the latter, providing
a host of potential molecular markers for developmental studies. Finally,
the procedures for functional gene analysis pioneered in zebrafish can
also be used in Astyanax.
Cavefish and other cave animals present two major challenges
to evolutionary developmental biologists: (1) to understand the
developmental mechanisms underlying the phenotypic changes related
to cave life and (2) to determine the evolutionary forces responsible for
driving these developmental and phenotypic changes. This chapter
reviews Astyanax embryonic and larval development in the context of
these challenges. We focus on the molecular, cellular, and developmental
mechanisms responsible for loss of eyes and pigmentation. This information
is then used to address the mechanisms underlying the evolution of these
regressive traits.
ASTYANAX DEVELOPMENT
Astyanax development involves (1) fertilization and cleavage, (2) epiboly
and gastrulation, (3) morphogenesis, segmentation, and formation of

the body axis, and (4) organogenesis. During these stages, the fertilized
egg becomes an embryo, the embryo hatches and develops into a larva
(or fry), and the fry becomes an adult. Cahn (1958) carried out the first
detailed analysis of Astyanax development comparing surface fish to
Chica and Los Sabinos cavefish. She showed that embryonic and larval
development resembles that of other teleosts, such as zebrafish. Early
development of surface fish and cavefish is superficially similar, except for
slight differences in timing events and the reduced size of eye primordia.
As the embryo hatches and the fry grow, differences between the cave and
surface forms become more pronounced. Below we describe embryonic
and larval development in Pachón cavefish.
Figure 2 shows Pachón cavefish development from fertilization to
hatching. Eggs are deposited in a clutch of several hundred and fertilized
externally. It has been reported that Subterráneo cavefish eggs are slightly
larger than those of surface fish (Hüppop and Wilkens 1991). However, egg
size does not seem to differ markedly between Pachón cavefish and surface
fish. The egg cytoplasm and yolk platelets are distributed homogenously in
the unfertilized egg. Fertilization begins a series of contractile movements
in which the yolk is concentrated into a large vegetal region, the yolk mass,
and the yolk-free cytoplasm is squeezed into a smaller animal region, the
blastodisc (Fig. 2, 1-2). Subsequently, incomplete cleavages occur only in
the blastodisc, which becomes the embryo proper. The uncleaved yolk
mass becomes the yolk sac.
Teleosts exhibit discoidal cleavage, which is characterized by
incomplete cleavage furrows that do not extend through the yolk mass.
Discoidal cleavage is also meroblastic, meaning that the early blastomeres
retain cytoplasmic continuity with each other through the yolk mass. This
continuity, which persists until about the 16-cell stage, is important for
functional experiments because molecules microinjected into a restricted
region of the embryo can disperse into every blastomere. The early
cleavages are synchronous, and the cleaving embryo is marked by a rapid
increase in cell number with no increase in the size of the blastodisc (Fig. 2,
3-14). The first cleavage furrow divides the blastodisc into two equal sized
blastomeres (Fig. 2, 3). The second, third, and fourth cleavage furrows are
positioned perpendicular to the first cleavage plane, producing a 4 × 4
array of 16 cells in the blastodisc (Fig. 2, 4-6). After the 16-cell stage, the
four inner cells (non-marginal cells) are separated completely from the
underlying yolk mass, whereas the 12 outer cells (marginal cells) retain
cytoplasmic links to the yolk mass. Every cleavage up to the sixth is in a
vertical plane through the blastodisc (Fig. 2, 3-8). The sixth cleavage is the
first horizontal cell division (Fig. 2, 8). Beyond the sixth cleavage, there are
no stereotypical arrangements of blastomeres (Fig. 2, 9-14). Blastomeres

Fig. 2 Early development of Astyanax cavefish (Pachón) eggs and embryos.

Cavefish embryos are shown from the unfertilized egg to the recently hatched
larva. Stages are ordered from left to right in each row. Numbers at the left
indicate successive stages to the right in each row. Stages are named below the
corresponding photograph.
(marginal and non-marginal) facing the outside of the embryo become

the enveloping layer (EVL), whereas cells that have no external surface
are known as the deep cells. As cleavage continues, the deep cells greatly
outnumber those of the EVL.
The blastula is distinguished from previous stages by the loss of division
synchrony, the beginning of cell motility, and other changes associated
with the so-called mid-blastula transition (Fig. 2, 12). During the blastula
stages, the blastodisc forms a crescent shaped structure on top of the yolk

mass (Fig. 2, 13). Blastulae pass through the oblong, sphere, and dome
stages. During these stages, the blastodisc elongates and spreads out over
the yolk mass (Fig. 2, 14-16). Blastodisc expansion involves increased
adherence of the marginal cells to the underlying yolk mass, forming a
border zone known as the yolk syncytial layer (YSL). In zebrafish, other
teleosts, and presumably Astyanax, blastomeres are still uncommitted to
particular developmental fates at the blastula stage (Long 1983, Langeland
and Kimmel 1997).
During epiboly the EVL moves over the yolk mass to create an embryo
whose outer surface is almost completely covered with a thin epithelium.
The extent of epiboly is measured as the percentage of yolk mass covered
by extending epithelial cells. Epiboly is driven by the YSL, which slowly
contracts and pulls the attached EVL over the yolk mass (Fig. 2, 17-22).
Gastrulation is complex in teleosts, and will be only briefly considered
here. Gastrulation begins at about 50% epiboly, when cells of the surface
layer move under the rim of the extending blastodisc, then converge
and extend toward the animal pole, thus forming the embryonic axis.
The movements of gastrulation are most extensive in a region called
the embryonic shield, which is the teleost counterpart of the amphibian
organizer. At the end of gastrulation (and neurulation), the embryonic
body plan is clearly visible, with future head and tail on opposite sides
of the anterior-posterior axis. This is called the (tail)bud stage (Fig. 2, 22).
At the tailbud stage, the embryonic midline divides the embryo into left
and right halves. The midline (prechordal plate and the notochord) is an
important signaling center throughout embryogenesis. The commitment
of cells to distinct fates occurs during gastrulation, and the fate map is
now restricted.
During the next stage of embryogenesis, the embryonic body plan is
elaborated. This period includes the enlargement and elongation of the
embryo, the gradual disappearance of the yolk sac, the formation of
organ primordia, and the appearance of somites, the progenitors of body
wall muscle and the vertebral column (Fig. 2, 23-37). Somites appear in
sequence from the base of the head toward the tail and can be used to
classify embryos into different stages (e.g. 20 somite stage, Fig. 2, 32). Many
morphogenetic events occur during this period of development, including
brain formation from the anterior neural tube and the migration of neural
crest cells. Organogenesis begins after the embryo hatches by breaching
the wall of the chorion, and thus becomes a free-swimming larva (Fig. 2,
38). During the larval stage, fry develop the phenotypic characteristics of
Astyanax adult surface fish and cavefish (Fig. 3).
As in other teleosts (Fernald, 1991), the eyes of surface fish grow
continuously to match the growth of the body. In contrast, cavefish eyes

Fig. 3 Post-hatching development of Astyanax surface fish and cavefish (Pachón)

larvae. Surface fish (above) and cavefish (below) are shown in each frame. A. One
day post-fertilization (dpf). B. Three dpf. C. One week post-fertilization (wpf). D.
Two wpf. E. Three wpf. F. One month post-fertilization (mpf). Note developmental
arrest and progressive loss of eyes in cavefish, development and rapid growth of
eyes in surface fish, body pigment cell development in surface fish, and absence
of eye and body pigmentation in cavefish. Scale bars are 62.5 mm (A), 125 mm
(B), 250 mm (C-D), and 500 mm (F).
reach a certain size and then growth is arrested (Figs. 3, 4A). Because
the body continues to grow, the degenerating eyes gradually disappear
into the orbits (Fig. 3). Surface fish eyes become pigmented, due to
melanin deposition in the retinal pigment epithelium (RPE), the cell layer
surrounding the retina, whereas the cavefish RPE either lacks pigmentation
or the number of pigment cells is greatly reduced (Fig. 3C-F). In surface
fish, melanin-containing body pigment cells also differentiate and form
distinctive patterns in the dermal layers of the skin, whereas these cells
are absent or decreased in cavefish (see Figs. 1A and 3C-F). Many of the
constructive features of cavefish, including changes in jaw structure and
taste buds, also appear during the larval stages.

EYE REGRESSION: FROM OPTIC PRIMORDIUM TO

CRANIOFACIAL SKELETON
Eye Development and Degeneration

Teleost eyes develop similarly to the eyes of other vertebrates. Three
different parts of the ectoderm contribute to the eye. The ocular lens is
formed from the lens placode, a thickening in the surface ectoderm. The
retina and RPE are formed from the optic vesicles, which are bilateral
protrusions of the forebrain neuroectoderm. The optic vesicles arise from
bilaterally symmetric optic fields in the anterior neural plate. Each optic
vesicle rotates about 90° and then buckles inward to form the optic cup,
with the future retina on the concave side and the RPE on the convex side.
The isthmus connecting the optic cup to the forebrain becomes the optic
stalk, which will form the casing of the optic nerve fibers. Simultaneously,
the lens placode reorganizes into a vesicle, which detaches from the surface
epithelium and enters the concave opening of the optic cup. The third
part of the ectoderm responsible for eye development is the neural crest.
Cranial neural crest cells migrate from the developing neural tube region
into spaces between the lens and surface epithelium to form the inner
parts of the cornea, between the lens and the distal edges of the retina
to contribute to the iris and ciliary body, and into the areas surrounding
the RPE to form the choroid and sclera. Neural crest cells probably also
contribute to the ocular dermal bones that develop much later around the
orbit, becoming part of the adult craniofacial skeleton.
The eye primordium consists of three parts, lens, retina, and RPE,
which differentiate in concert. The lens vesicle produces fiber cells, which
synthesize crystallin proteins, and becomes transparent, leaving behind a
layer of undifferentiated stem cells. The retina differentiates into several
layers. From distal to proximal, they consist of (1) the ganglion cell layer,
which transmits neural signals to the brain via axons extending through the
optic stalk into the optic tectum, (2) the intermediate layers, which consist
of inter-neurons and glial cells, and (3) the photoreceptor layer, where rod
and cone cells translate photons into neural signals. The RPE forms tight
connections with the photoreceptor layer and produces melanin pigment.
Body pigment cells containing melanin have a different origin from those
of the RPE, and will be discussed later in this chapter.
The sequence of events during surface fish and cavefish eye development
are compared in Fig. 4A. The cavefish eye primordium is slightly smaller
than its surface fish counterpart (Fig. 4B, C; Fig. 6C, D). This difference in
size is due to a smaller lens and optic cup, which appears to be missing
its ventral sector (Fig. 6G, H). In contrast to the surface fish eye, cavefish

24 h 36 h 5 days 3 months
12 h 16 h 20 h
24 h 48 h 72 h 10 days 1 month 3 months
Fig. 4 Eye development and degeneration in Astyanax. A. Diagram showing

similar development of the eye primordium from up to about 20 hrs post fertilization
(hpf) in cavefish and surface fish (left), followed by their subsequent divergence
(right). The surface fish eye differentiates and rapidly increases in size (top) from
1 dpf to 1 mpf, whereas the cavefish eye arrests in growth, degenerates, and
gradually sinks into the orbit (bottom). B, C. Size differences in the 24 hpf surface
fish (B) and cavefish (C) eye primordia. D-G. Sections of 2 (D, E) and 3 (F, G)
dpf surface fish (D. F) and cave fish (E, G) eye primordia showing apoptosis
(dark stained spots) in various eye tissues. In cavefish, apoptosis begins in the
lens (arrowheads) and spreads to the retina (arrows). There is no apoptosis in
these tissues in surface fish. H. Diagram showing the roles of cell proliferation
and apoptosis during retina/RPE growth in surface fish and cavefish. Clear
retinal areas: embryonic retina and central part of growing retina derived from
embryonic retina. Shaded area: part of retina derived from cell proliferation at the
CMZ after the embryonic stages. X: Apoptotic areas. The surface fish retina grows
continuously due to cell proliferation at the CMZ, whereas the cavefish retina is
arrested in growth because the products of cell proliferation at the CMZ die before
they contribute to the differentiated retina.

optic tissues either fail to be induced (cornea, iris, ciliary body) or begin to
differentiate and then partially or completely degenerate (lens, retina, and
probably RPE). However, perhaps the most important flaw in the cavefish
eye primordium is the absence of optic growth from the larval to the
adult stages (Fig. 4A). Eventually, the arrested cavefish eye primordium is
overgrown by head epidermis and connective tissue, and disappears into
the orbit, making adult cavefish appear superficially eyeless.
Because cavefish eye development involves growth arrest, it is
important to consider the origin of new cells. Stem cells in the epithelial
layer are the source of new lens fiber cells. The source of most new retinal
and all new RPE cells is stem cell niche at the edge of the optic cup, a
region known as the ciliary marginal zone (CMZ). As the eye enlarges
during larval development, it is surrounded by orbital bones, which form
a part of the craniofacial skeleton. The number, size, and organization of
the orbital bones is distinct between surface fish and cavefish, and even
among different cavefish populations (Alvarez 1947, Yamamoto et al.
2003). As described below, the presence or absence of a functional eye is
critical in the morphogenesis of the orbital bones and organization of the
craniofacial skeleton.
Cellular Mechanisms of Eye Degeneration

A block in cell proliferation, an increase in programmed cell death, or
a combination of these processes could cause the arrest of cavefish eye
development. In many developing systems, programmed cell death
or apoptosis is an alternative to cell differentiation. Investigations of
apoptosis during Astyanax eye development (Jeffery and Martasian 1998,
Yamamoto and Jeffery 2000) have led to a breakthrough in understanding
the regulation of eye degeneration.
If cell death is restricted to a single eye tissue, or begins in one tissue and
later spreads to others, then the tissue that dies first is a candidate to initiate
the degeneration process. The early cavefish eye primordium is largely
free of cell death, except for one tissue: the lens (Fig. 4E). No apoptosis is
detected in the surface fish lens (Fig. 4D). Although cell proliferation does
not cease in the cavefish lens, the rate of apoptosis is very high, eventually
obliterating the lens, or reducing it to a tiny vestige in the adult (Soares et al.
2004). A few days after the initiation of lens apoptosis, the cavefish retina
also begins to undergo apoptosis (Fig. 4G) (Alunni et al. 2007, Strickler et
al. 2007a). Cell death is restricted to the intermediate retinal layers and
regions adjacent to the CMZ. Later in development, dying cells are also
seen in the cavefish RPE (Strickler et al. 2007a). Like in the lens, cell death
is not observed in the surface fish retina (Fig. 4F) or RPE. Thus, the lens

is the first tissue to undergo cell death in the cavefish eye primordium,
suggesting that its absence may be responsible for eye degeneration.
The cessation of retinal growth in cavefish could be caused by failure
of the dying lens to produce a growth-promoting factor. Surface fish have
an active CMZ, in which proliferating cells can be detected by labeling
replicating DNA or the presence of DNA replication enzymes, such
as proliferating cell nuclear antigen (PCNA). As newly born cells are
displaced from the CMZ into the retinal layers and RPE, they differentiate
and increase the general mass of the retina/RPE. Surprisingly, cell
proliferation markers are expressed normally in the cavefish CMZ during
the period in which the retina does not markedly increase in size (Strickler
et al. 2002, 2007a, Alunni et al. 2007). However, the new cells are quickly
removed from the retina by apoptosis, which persists during cavefish
larval development and into adult life (Strickler et al. 2007a). Thus, the
cavefish eye is arrested in growth because newly born cells die before they
are able to differentiate and join the retinal layers (Fig. 4H). The possibility
that lens dysfunction is responsible for retinal cell death and has a general
role in eye degeneration was tested by transplantation of embryonic lenses
between surface fish and cavefish.
The lens transplantation method is illustrated in Fig. 5A. The embryonic
lens is removed from a donor embryo shortly after its formation (about
a day before the first detection of apoptosis in the cavefish lens) and
transplanted into the optic cup of a host embryo. Lens transplantation is
done unilaterally, with the unoperated eye of the host serving as a control,
and reciprocally: a surface fish lens is transplanted into a cavefish optic
cup and vice versa (Yamamoto and Jeffery 2000). When a cavefish lens
was transplanted into a surface fish optic cup it died on schedule, just as
if it had not been removed from the donor embryo. In contrast, when a
surface fish lens was transplanted into a cavefish optic cup it continued
to grow and differentiate, just as it would have done in the surface fish
host. These results indicate that the cavefish lens is autonomously fated
for apoptosis.
After obtaining a surface fish lens, the eyes of Pachón or Los Sabinos
cavefish began to grow (Fig. 5C, H) (Yamamoto and Jeffery 2000, Jeffery et
al. 2003). Eventually, the cornea and iris appeared, and the enlarged retina
became more highly organized. Further growth resulted in the presence
of a highly developed eye containing a cornea, iris, and photoreceptor
cells. In contrast to the eye with a transplanted lens, the unoperated eye of
the cavefish host degenerated and disappeared into the orbit (Fig. 5B, G).
Likewise, after obtaining a cavefish lens, development of the surface fish
eye was retarded, the cornea and iris did not differentiate, and the size and
organization of the retina was reduced. The degenerate surface fish eye

Fig. 5 Lens transplantation experiments. A. Diagram showing the transplantation

method in which a donor lens is removed from the optic cup of one form of Astyanax
embryo and transplanted unilaterally into the optic cup of another form after the
host lens is removed. This operation is carried out at about 1 dpf. B-J. Changes
in eye development after lens transplantation during embryogenesis. B, C, F, G.
Surface fish lens was transplanted into a Pachón (B, C) or Los Sabinos (F, G),
cavefish host. D, E, I, H. Changes in eye development after a Pachón (D, E) or
Los Sabinos (H, I) cavefish lens was transplanted into a surface fish host. J, K.
Changes in orbital bone structure after a surface fish lens was transplanted into a
cavefish optic cup unilaterally. B, D, F, H, J: Control (unoperated) side. C, E, G, I,
K. Transplant side.

eventually disappeared into the orbit, mimicking the cavefish eye (Fig. 5E,
I), whereas the unoperated eye developed normally, producing a one eyed
surface fish (Fig. 5D, H).
The following conclusions can be made from the lens transplantation
experiments. First, the lens is required for normal development of the
retina, cornea, and iris. Second, as a result of apoptosis the cavefish lens has
lost the ability to organize eye development. Third, the cavefish optic cup
(RPE/retina) has retained the ability to respond to signals generated by
a normal surface fish lens. Lens transplantation also allowed two further
insights. First, the lens has a role in promoting the survival of retinal cells: a
transplanted surface fish lens can protect the cavefish retina from apoptosis
(Strickler et al. 2007a). Second, the lens has an indirect role in determining
craniofacial morphology. When a surface fish lens is transplanted into
a cavefish optic cup, an orbital bone phenotype is obtained resembling
surface fish rather than cavefish (Fig. 5J, K) (Yamamoto et al. 2003). Thus,
the cavefish host develops with hybrid craniofacial morphology, one side
(the lens transplant side) resembling surface fish and the other (the control
side) cavefish.
Clearly, the lens has a major role in regulating cavefish eye degeneration.
Whether the death of the lens is the only cause of eye degeneration, or
other optic alterations, such as independent changes in the retina or RPE,
are also involved (Strickler et al. 2007a), remains to be determined.
Molecular Mechanisms of Cavefish Eye Degeneration

Understanding the molecular mechanisms of eye degeneration requires
identification of the genes involved in this process and how they
function during development. Many eye development genes are known
in vertebrates, allowing a candidate gene approach to be used for gene
identification. Genes encoding transcription factors that function near the
top of eye gene hierarchies, as well as structural genes encoding proteins
that function at the bottom of these cascades have been isolated in Astyanax.
In situ hybridization was used to determine their expression patterns.
Most of the surveyed genes and/or their encoded proteins do not show
expression changes in surface fish and cavefish embryos. For example, the
Prox1 transcription factor is expressed normally in the developing lens
and retina of cavefish until after the eye begins to degenerate (Jeffery et
al. 2000). Likewise, prior to lens degeneration, genes encoding ßB- and
gM-crystallins and the membrane proteins MIP and MP19 are expressed
normally (Jeffery et al. 2000, Strickler et al. 2007b). Many genes also show
the same or similar expression patterns in the developing surface and cave
fish retinas (Jeffery et al. 2000, Strickler et al. 2002, Menuet et al. 2007). For

example, Langecker et al. (1993) noted that an opsin gene is expressed

in the degenerating photoreceptor layer of the cavefish retina. The most
important changes in optic gene expression patterns during cavefish
development are described below.
The pax6 gene encodes a transcription factor that is expressed in the lens,
retina, RPE, and their precursors early in teleost eye development (Krauss
et al. 1991, Püschel et al. 1998). Later, pax6 expression becomes restricted
to the lens epithelial cells, some of the retinal layers, and the corneal
epithelium. In surface fish embryos, the pax6 expression domains in the
bilateral optic fields connect across the midline at their anterior margins
(Fig. 6 A, B). In cavefish embryos, however, the corresponding pax6 domains
are diminished in size and show a large gap across the midline (Strickler
et al. 2001) (Fig. 6C, D). The division of the optic vesicle into the optic cup
and stalk is controlled by reciprocal antagonistic intereactions between
the Pax6, Pax2, and Vax1 transcription factors (Schwarz et al. 2000). Pax6
directs optic cup development, whereas Pax2 and Vax1 control optic stalk
development. Accordingly, a reduction of pax6 levels (or an increase in
pax2 and vax1 levels) increases the optic stalk at the expense of the optic
cup. The reduction of pax6 expression coupled with the overexpression
of Pax2 and Vax1 accounts for the ventrally reduced optic cup in cavefish
embryos (Fig. 6).
The wider gap between optic fields in the cavefish neural plate provides
further insight into how eye degeneration is controlled. During vertebrate
development, the presumptive optic cup is initially determined as a
single medial optic field, which is subsequently split into two bilateral
eye domains by Hedgehog (Hh) signals emanating from the underlying
prechordal plate (Macdonald et al. 1995, Ekker et al. 1995). Hh signaling
inhibits pax6 expression along the midline to divide the original eye
domain into bilateral eyes. Teleosts have at least two-hh midline signaling
genes, sonic hedgehog (shh) and tiggy winkle hedgehog (twhh), which show
overlapping expression patterns (Ekker et al. 1995). Yamamoto et al. (2004)
compared shh and twhh expression patterns during surface fish and
cavefish development, showing that the midline expression domains of
both genes are expanded in cavefish relative to surface fish (Fig. 7A, B).
Later in cavefish development, shh expression is also expanded anteriorly,
curling around the rostrum in the presumptive oral area (Fig. 7C, D). The
expression patterns of genes acting downstream of shh and twhh in the Hh
midline signaling pathway, such as patched, encoding a Shh receptor, and
nkx2.1, encoding an Shh dependent transcription factor, are also expanded
(Yamamoto et al. 2004), suggesting that a general increase in midline Hh
signaling has evolved in cavefish.

Fig. 6 Optic vesicle (A-F) and optic cup (G-L) development in surface fish and
cavefish. A, B. Neural plate stage embryo showing differences in pax6 expression
in the surface fish and cavefish optic fields (OF). Arrowhead shows the midline
pax6 expression gap, which is wider in cavefish. C, D. Optic vesicles (OV) showing
size and pax2a expression (arrowheads) differences in surface fish and cavefish.
In A-D. embryos are viewed dorsally with anterior at the top. E, F. Diagram showing
size differences in the surface fish and cavefish optic vesicles. Territories fated to
form optic stalk are lightly shaded and those fated to form retina/RPE are darkly
shaded. G, H. Surface fish and cavefish optic cups (OC) showing ventral size
reduction in the latter. L: lens. I, J. The vax1 gene is overexpressed ventrally in
the cavefish optic cup relative to surface fish. In G-J, embryos are viewed laterally
with dorsal at the top. K, L. Diagram showing size and relative optic cup territorial
differences between cavefish and surface fish. The optic stalk (OS) is lightly shaded
and the optic cup is darkly shaded.
Two additional genes appear to be associated with cavefish lens

apoptosis. First, the heat shock protein90a (hsp90a) gene is specifically
activated in the cavefish lens vesicle just prior to apoptosis (Hooven et
al. 2004). Inhibition of hsp90a expression suppresses lens apoptosis and
rescues lens differentiation. Second, the aA-crystallin gene, which encodes
a potent anti-apoptotic factor, is strongly downregulated in the lens
vesicles of Piedras (Behrens et al. 1998) and Pachón cavefish (Strickler et
al. 2007b). aA-crystallin may normally protect the lens from apoptosis.
Further studies will be necessary to determine if aA-crystallin and Hsp90
a interact in a cascade leading to lens apoptosis.

Fig. 7 Role of Hh midline signaling in cavefish eye degeneration. A-D. The

cavefish embryonic midline shows a wider shh expression domain than its surface
fish counterpart. The expression of dlx3 and pax2a marker genes does not change.
A, B. Tailbud stage. C, D. Ten somite stage. E-J. Effects of shh overexpression in
surface fish. E, F. Increased shh expression (compare F with C) and reduced pax6
expression (E) on one side of the midline of an embryo injected with shh mRNA.
G, H. As a result of shh overexpression, the optic cup (retina/RPE) is missing its
ventral sector (G) and the adult eye has degenerated (H). Arrowhead in G: missing
ventral sector of the retina. Arrowhead in H: missing eye. I, J. Lens apoptosis
(J) after injection of an embryo with shh mRNA. Arrowheads: lens. H. Diagram
showing antagonistic relationship between Pax6, Pax2, and Vax1 transcription
factors, ventralization of the optic cup, and lens apoptosis in cavefish. Arrows:
activations. Blocked lines: inhibitions.

The Key Role of Hedgehog Signaling

The role of enhanced Hh signaling in cavefish eye degeneration was further
investigated by Yamamoto et al. (2004). Shh signaling was increased by
injection of shh mRNA into surface fish embryos. When shh mRNA was
injected into one side of a cleaving embryo, shh expression was expanded
along that side of the prechordal plate (anterior embryonic midline), and
pax6 expression was down regulated unilaterally in the corresponding
optic field (Fig. 7E, F). When surface fish embryos overexpressing Shh
developed into larvae and adults, the eye was missing on one side of the
head (Fig. 6G, H). Thus, blind cavefish were phenocopied by increasing
the levels of shh gene expression in surface fish, demonstrating a key role
for Shh signals in eye degeneration. Importantly, lens apoptosis is also
induced by shh overexpression in surface fish embryos (Yamamoto et al.
2004) (Fig. 7I, J). A diagram of the gene network leading to eye degeneration
via hyperactive Shh signaling, reduction of the optic vesicle, and lens cell
death is shown in Fig. 7K.
It can be concluded that a sequence of regulatory events beginning
with expanded midline signaling, proceeding through reduction in size
of the eye primordia, lens apoptosis, and retinal apoptosis, and resulting
in arrested eye growth and alteration of craniofacial morphology, is
responsible for cavefish optic degeneration (Fig. 8). Alterations in the
activity of many different genes and their upstream regulators control these
changes. Recent genetic analysis has shown that at least 12 quantitative
trait loci (QTL) are involved in the loss of eyes in Pachón cavefish (Protas
et al. 2007).
(Alpha A-crystallin
and Hsp90alpha)
Fig. 8 Summary of early and late events in cavefish eye degeneration and
consequences on craniofacial development.

PIGMENT CELL REGRESSION: FROM NEURAL CREST

TO MELANOGENESIS
Astyanax surface fish has three types of body pigment cells: light reflecting
iridophores, yellow-orange xanthophores, and black melanophores.
Pigmentation normally functions in protection from the damaging effects
of sunlight, in camouflage, and in species and sex recognition. Selective
pressure for retaining these functions is relaxed in the absence of light.
What are the consequences in Astyanax cavefish?
The early studies of Rasquin (1947) showed that melanophores are
decreased but xanthophores seem to be present at the same levels in
cavefish. Very little is known about changes in iridophores. Of the three
types of pigment cells, most is known about melanophores (Wilkens 1988,
McCauley et al. 2004). Subterráneo cavefish show a modest reduction in
melanophore pigmentation, Curva, Los Sabinos, and Tinaja cavefish show
substantial decreases in melanophore pigmentation, and Molino and
Pachón cavefish show little if any melanophores. In addition to changes in
the number of melanophores, cavefish also show defects in the ability to
produce melanin, the pigment found in melanophores.
Migratory Neural Crest Cells in Cavefish

All types of body pigment cells are derived from the neural crest, migratory
cells that arise from the border of the neural tube and surface ectoderm
(Erickson 1993, LeDouarin and Kalcheim 1999). Vertebrate neural crest
cells have a myriad of different derivatives, including sensory ganglia,
the peripheral nervous system, cranial cartilage and bone, as well as body
pigment cells. Cell tracing, immunological, and tissue culture methods
were used to follow neural crest development in cavefish (McCauley et al.
2004, Jeffery 2006). In both cavefish and surface fish, similar numbers of
labeled neural crest cells migrate into the epidermis (Fig. 9A-D), suggesting
that there is no defect in neural crest cells during cavefish development.
Neural crest cells that do not migrate properly or receive normal
differentiation signals often die by apoptosis (Morales et al. 2005).
Therefore, apoptosis could remove neural crest derived precursors in
cavefish embryos before they differentiate into pigment cells. When this
possibility was tested only a few dying neural crest cells were observed
in cavefish embryos, and their number was about the same as in surface
fish embryos (Jeffery 2006). Therefore, melanophores or their progenitor
cells do not undergo massive apoptosis during cavefish embryogenesis.
Cavefish pigmentation defects must arise downstream of the generation,
migration, and divergence of pigment cell types.

Fig. 9 Neural crest and melanophore development in cavefish. A-C. Detection of

migratory neural crest cells in cavefish embryos by DiI injection and subsequent
tracing of labeled cells. A. A 1.5 dpf cavefish embryo showing the site of Dil injection
(arrowhead). B. Fluorescence image of the embryo in A showing migration of
Dil injected cells. DiI: original injection site. C, D. Higher magnification images
of insets in B showing morphology of injected cells (arrows). E, F. A cavefish
embryo after L-DOPA assay showing tyrosinase positive melanoblasts (arrows).
E. Whole mount viewed laterally. F. Section through the trunk. G-J. Whole mounts
of tail fins of adult surface fish (G) and cavefish (H-J). G. The surface fish fin has
melanophores (arrows). H. The cavefish fin lacks melanophores. I. The cavefish
fin has melanoblasts (arrows) that can convert exogenously supplied L-DOPA to
melanin. J. Cavefish melanoblasts lack the ability to convert L-tyrosine to L-DOPA
and melanin. K. Diagram of pigment cell development from the neural crest derived
precursor cells showing the location of the pigmentation block in cavefish. L. The
melanin biosynthesis pathway showing the relationships between L-tyrosine,
tyrosinase, OCA2, L-DOPA, downstream substrates and enzymes, and melanin.

Tyrosinase-Positive Melanoblasts in Cavefish

The early events of pigment cell formation and diversification are not
completely understood in vertebrates. However, it seems likely that
iridophores, xanthophores, and melanophores are derived from the same
neural crest cell lineage (Fig. 9K). Since cavefish lacking melanophores
contain appreciable numbers of other pigment cell types (Rasquin 1947,
McCauley et al. 2004), the lesion in melanophore development probably
lies downstream of the split between the pigment cell progenitors (Fig.
9K).
Melanophore differentiation involves the initial formation of colorless
melanoblasts, which subsequently synthesize black melanin pigment
and become functional melanophores. The biochemical steps involved in
melanin synthesis during the transition from melanoblast to melanophore
are well known (Fig. 9L). First, the essential amino acid L-tyrosine
is transported from the cytoplasm into the melanosome, where it is
converted to L-DOPA by the multifunctional enzyme tyrosinase. Next,
L-DOPA is converted into melanin within the melanosome by a series
of enzymatic reactions, the first of which is also catalyzed by tyrosinase.
Most subsequent reactions in this biosynthetic pathway are spontaneous.
Tyrosinase is the key enzyme in melanogenesis. Do cavefish pigment
progenitor cells have functional tyrosinase? Tyrosinase activity was
determined by the L-DOPA assay, in which melanin production is
determined after exogenous L-DOPA is provided to fixed specimens. The
L-DOPA assay showed that Pachón, Chica, Los Sabinos, Tinaja, and Curva
cavefish all exhibit active tyrosinase in cells resembling melanoblasts in
their morphology and location within the embryo (Fig. 9E, F). Tyrosinase
positive melanoblasts were also observed in the scales and fins (Fig. 9G-I)
of adult cavefish. The inability of these melanoblasts to synthesize melanin
is due to a block in the melanogenic pathway immediately upstream of
the tyrosinase dependent steps.
Cavefish Melanoblasts are Unable to Convert L-tyrosine

into L-DOPA
The first step in melanin synthesis is the conversion of L-tyrosine to
L-DOPA, which is also catalyzed by tyrosinase (Fig. 9L). Cavefish must
have L-tyrosine itself because it is required for protein synthesis. Because
cavefish seem to lack endogenous melanosomal L-DOPA there may
be a deficiency in the ability of L-tyrosine to be converted to L-DOPA.
This possibility was investigated by a modified L-DOPA assay in which
exogenous L-tyrosine was provided to fixed specimens instead of L-DOPA

(McCauley et al. 2004). If cavefish are able to convert L-tyrosine to L-DOPA

then black pigment would be deposited in the same cells that have active
tyrosinase. However, melanin deposition could not be detected in cavefish
exposed to excess L-tyrosine (Fig. 9H-J). The results show that cavefish
melanoblasts are unable to convert L-tyrosine to melanin, implying that
melanogenesis is blocked because cytoplasmic L-tyrosine cannot be
transported into cavefish melanosomes (Fig. 9L).
Developmental and Genetic Studies Converge to Define

the Defect in Cavefish Melanogenesis
In lacking the ability to convert L-tyrosine to L-DOPA, cavefish resemble
the most common type of human tyrosinase-positive albinism: type II
oculocutaneous albinism (OCA2) (Oeting and King 1999). OCA2 albinism
is caused by mutations in the oca2 gene, the human homologue of the
mouse pink-eyed dilution (p) gene, which encodes a 100-kDa integral
membrane protein (Rosenblatt et al. 1994). Although its precise function is
unclear, the OCA2 protein has been proposed either to facilitate L-tyrosine
transport into the melanosome or to generate a proton flux regulating
melanosome pH, which is important in melanin synthesis (Brilliant 2001,
Toyofuku et al. 2002). A single recessive gene controls cavefish tyrosinase-
positive albinism (Sadoglu 1957). QTL analysis and a candidate gene
survey show that different mutations in the oca2 gene are responsible
for albinism in Molino and Pachón cavefish (Protas et al. 2006). Thus,
genetic and developmental studies have pinpointed the defect in cavefish
melanophore pigmentation: the conversion of L-tyrosine to L-DOPA
within the melanosome mediated by OCA2.
DEVELOPMENTAL INFERENCES INTO EVOLUTION

Comparative studies of Astyanax surface fish and cavefish development
provide important insights into the processes and mechanisms of regressive
evolution. In the final section of this chapter, we discuss these insights.
Developmental Constraints
Developmental constraints restrict the amplitude of evolutionary changes,
or make them unlikely or impossible, by limiting developmental flexibility.
This lack of flexibility appears to have a very important role in cavefish
regressive evolution. If the eye is ultimately lost, why is it necessary to
construct an eye in the first place? The answer may be that early steps in
eye development are required to induce other parts of the embryo, and thus

the modification of these steps would be fatal. To our knowledge, eyes are
initially formed and then degraded during larval or adult development
in all cave dwelling vertebrates (Eigenmann 1909, Durand 1976, Berti
et al. 2001). Indeed, we feel that cave vertebrates lacking embryonic eye
primordia will not be discovered because of this strong developmental
constraint.
Exactly how the eyes are lost, which we have seen to be under the
control of midline signaling in Astyanax, reflects another developmental
constraint. Because all vertebrates have bilateral eyes arising from a single
medial optic field, the subsequent separation of optic fields is likely to be
an ancient vertebrate trait that evolved in concert with other head features.
Thus, if the Shh midline-signaling pathway is altered, as we have seen in
cavefish, there may be automatic consequences on eye development, in
this case leading to degeneration (Jeffery 2005).
The way in which retinal development is arrested in cavefish may be
a further example of developmental constraint. We have shown that the
arrest of retinal development is not caused by inhibition of cell division
at the CMZ, which would seem to be the simplest way to stop growth.
Instead, retinal growth is curtailed by apoptosis of newly born cells
(Strickler et al. 2007a). This must be a very costly process in terms of energy
expenditure, so why has inhibition of cell proliferation, the most simple
and least expensive route to preventing retina development, not been
taken? The probable answer lies in the fact that the retina is actually a part
of the brain. In both retina and brain, stem cells replenish the laminated
areas through the same course of action, which may not be easy to modify.
Accordingly, killing new cells after they proliferate in the retina may be
more allowable than blocking stem cell division in the CMZ because of
an ancient constraint on how different parts of the brain grow in concert
during development.
A final example of a developmental constraint occurs in the melanophore
development pathway. In cavefish, a block in the pigment cell-generating
pathway occurs relatively late in the pigment cell development pathway,
during the conversion of melanoblasts into melanophores. Earlier steps
in this pathway, such as the determination and migration of neural
crest cells, the restriction to pigment cell fate, and the diversification of
different pigment cell lineages is apparently not changed, even though
the usefulness of any pigment cell type is questionable in cavefish. The
reason neural crest cells must be formed is clear: they have many critical
derivatives, the loss of which would be lethal. Why the other pigment
cell types are formed in cavefish, albeit possibly at reduced levels, is a
mystery. They may have some unknown and indispensable function even
in the dark cave environment.

It is clear that cavefish regressive evolution is channeled to a large extent

by developmental constraints. Only those pathways that are not linked to
other essential pathways can be degraded or truncated.
Developmental Amplification
A major question in contemporary evolutionary developmental biology
is how large-scale changes in the phenotype can occur rapidly during
evolution. Cavefish show that slight changes in early development can
have major impacts at later stages and in the adult. An example is the
craniofacial skeleton.
The differences in craniofacial skeletons between cavefish and surface
fish, in particular the ocular bones surrounding the eye, are so extreme
that they were formerly used to support their designation as separate
genera (Alvarez 1947). However, the majority of these changes are
related to whether or not a large eye punctuates the craniofacial skeleton.
When the eye is absent from the surface of the head, as in cavefish, the
craniofacial skeleton is patterned differently from when an eye is present.
Major changes in the craniofacial skeleton can be elicited by transplanting
a surface fish lens into a cavefish optic cup during early development
(Yamamoto et al. 2003). The sequence of events is as follows: a normal lens
induces anterior eye parts and promotes the growth of a normal retina,
producing a large growing eye, which in turn dictates the morphology
of surrounding bones in the adult (Fig. 8). This scenario illustrates that
amplification of small changes in early development can produce large
differences in adult morphology.
Pleiotropy
Pleiotropy, the condition in which a single gene may control multiple,
seemingly unrelated phenotypes, has long been considered as a possible
mechanism for the evolution of regressive traits in cave animals (Barr 1968).
Accordingly, if downregulation of genes controlling eye development
simultaneously increases the development of a beneficial trait, such as
olfaction or another sensory system, the latter might be adaptive and
subject to natural selection. The discovery of enhanced midline signaling
mediated by highly pleiotropic hh genes (Yamamoto et al. 2004) opens
many possibilities that may be able to explain eye degeneration in
cavefish. As we have seen, Hh overexpression has a negative effect on eye
development, and it is known from studies on other vertebrates that Hh
signaling has positive effects on many other developmental traits. Thus,
selection for the positive traits would automatically affect the negative

ones. In the future, it will be important to determine the identity of positive

traits influenced by Hh signaling. Pleiotropic events could also explain the
loss of melanophores, if the precursors of these cells are used for another
purpose that benefits survival in the cave environment.
Evolutionary Forces
Two hypotheses have been championed to explain the regressive features
of cave animals: (1) the accumulation of selectively neutral (loss of function)
mutations and (2) indirect selection based on energy conservation and/
or antagonistic pleiotropy (Culver 1982, Jeffery 2005). Although neither
hypothesis has been proved in the case of Astyanax cavefish, developmental
studies generally support one or the other for loss of pigmentation and
eyes respectively.
In the case of eye loss, the developmental information seems to
support selection over neutral mutation. First, the genes involved in
eye development that have been studied thus far do not appear to have
mutated to a degree in which they have lost function. In addition, the
restoration of eyes by lens transplantation suggests that all genes that act
downstream of lens function are present and potentially active in cavefish.
Also supportive of selection is that most genes with modified expression
patterns, such as those in the Shh signaling pathway and hsp90a, increase
rather than decrease their activity in cavefish.
In contrast to eye regression, developmental studies on loss of
pigmentation could support either selection or neutral mutation.
On one hand, the accumulation of neutral mutations resulting in
loss of melanophores might be possible, especially if the oca2 gene
is not pleiotropic and its disruption does not affect other important
developmental pathways. Recent genetic analysis, in which individual
QTL governing the extent of melanophore development has been shown
to either increase or decrease melanophore abundance, support the role of
neutral mutation and genetic drift (Protas et al. 2007). On the other hand,
as we have discussed above, melanogenesis may be disrupted because it is
adaptive, allowing pigment cell precursors to be shunted into other, more
beneficial differentiation pathways. Some of these possibilities are testable
and predict a bright future for developmental biology in addressing both
how and why regressive features have evolved in Astynanx cavefish.
Acknowledgements
The research from the Jeffery laboratory described in this article was
supported by grants from NIH (R01-EY014619) and NSF (IBN-0542384).

References
Alunni, A., A. Menuet, E. Candal, J-B. Pénigault,, W.R. Jeffery and S. Rétaux. 2007.
Developmental mechanisms for retinal degeneration in the blind cavefish
Astyanax mexicanus. Journal of Comparative Neurology 505: 221-233.
Alvarez, J. 1947. Descripción de Anoptichthys hubbsi caracinindo ceigo de La Cueva
de Los Sabinos. San Luis Potosi Sociedad Mexicana de Historia Natural 8: 215-219.
Barr, T. 1968. Cave ecology and the evolution of troglobites. Evolutionary Biology
2: 35-102.
Behrens, M., H. Wilkens and H. Schmale. 1998. Cloning of the aA-crystallin genes
of the blind cave form and the epigean form of Astyanax fasciatus: A comparative
analysis of structure, expression and evolutionary conservation. Gene 216: 319-
326.
Berti, R., J.P. Durand, S. Becchi, R. Brizzi, N. Keller and G. Ruffat. 2001. Eye
degeneration in the blind cave-dwelling fish Phreatichys andruzzi. Canadian
Journal of Zoology. 79: 1278-1285.
Borowsky, R. 2008. Restoring sight in blind cavefish. Current Biology 18: R23-R24.
Brilliant, M.H. 2001. The mouse p (pink-eyed dilution) and human P genes,
oculocutaneous albinism type 2 (OCA2), and melanosomal pH. Pigment Cell
Research 14: 86-93.
Cahn, P.H. 1958. Comparative optic development in Astyanax mexicanus and in
two of its blind cave derivatives. Bulletin of the American Museum of Natural
History 115: 73-112.
Culver, D. 1982. Cave Life. Evolution and Ecology. Harvard, Cambridge, MA.
Dowling, T.E., D.P. Martasian, and W.R. Jeffery 2002. Evidence for multiple
genetic lineages with similar eyeless phenotypes in the blind cavefish, Astyanax
mexicanus. Molecular Biology and Evolution 19: 446-455.
Durand, J.P. 1976. Ocular development and involution in the European cave
salamander, Proteus anguinus Laurenti. Biological Bulletin of the Marine Biological
Laboratory, Woods Hole 151: 450-466.
Eigenmann, C.H. 1908. The eyes of the blind vertebrates of North America. V. The
history of the eye of blind Amblyopsis from its appearance to its disintegration
in old age. Contributions of the Zoological Laboratory, Indiana University Mark
Anniversary Volume: 167-204.
Ekker, S.C., A.R. Ungar, P. von Greenstein, J.A. Porter, R.T. Moon and P. Beachy.
1995. Patterning activities of vertebrate hedgehog proteins in the developing
eye and brain. Current Biology 5: 944-955.
Erickson, C.A. 1993. From the crest to the periphery: Control of pigment cell
migration and lineage segregation. Pigment Cell Research 6: 336-347.
Fernald, R.D. 1991. Teleost vision: Seeing while growing. The Journal of
Experimental Zoology 5: 167-180.
Hooven, T.A., Y. Yamamoto and W.R. Jeffery. 2004. Blind cavefish and heat shock .
protein chaperones: A novel role for hsp90a in lens apoptosis. International Journal
of Developmental Biology 48: 731-738.

Hüppop, K. and H. Wilkens. 1991. Bigger eggs in subterranean Astyanax fasciatus

(Charcidae: Pisces). Their significance and genetics. Zeitschrift für zoologische
Systematik und Evolutionsforschung 29: 280-288.
Jeffery, W.R. 2001. Cavefish as a model system in evolutionary developmental
biology. Developmental Biology 231: 1-12.
Jeffery, W.R. 2005. Adaptive evolution of eye degeneration in the Mexican blind
cavefish. Journal of Heredity 96: 185-196.
Jeffery, W.R. 2006. Regressive evolution of pigmentation in the cavefish Astyanax.
Israel Journal of Ecology and Evolution 52: 405-422.
Jeffery, W.R. 2008. Emerging systems in Evo/Devo: cavefish and mechanisms of
microevolution. Evolution and Development 10: 265-272.
Jeffery, W.R. and D.P. Martasian. 1998. Evolution of eye regression in the cavefish
Astyanax: Apoptosis and the Pax-6 gene. The American Zoologist 38: 685-696.
Jeffery, W.R., A.G. Strickler and Y. Yamamoto. 2003. To see or not to see: Evolution of
eye degeneration in Mexican blind cavefish. Comparative and Integrative Biology
43: 531-541.
Jeffery, W.R., A.G. Strickler, S. Guiney, D. Heyser and S.I. Tomarev. 2000. Prox1 in
eye degeneration and sensory organ compensation during development and
evolution of the cavefish Astyanax. Genes Development and Evolution 210: 223-
230.
Krauss, S., T. Johannsen, V. Korzh and A. Fijose. 1991. Zebrafish pax[zf-a]: A paired
box gene expressed in the neural tube. European Molecular Biology Organization
Journal 10: 3609-3619.
Langecker, T.G. H. Schmale and H. Wilkens. 1993. Transcription of the opsin gene
in degenerate eyes of cave-dwelling Astyanax fasciatus (Teleostei: Characidae)
and its conspecific ancestor during early ontogeny. Cell and Tissue Research 273:
183-192.
Langeland, J.A. and C.B. Kimmel. 1997. Fishes. In: Embryology: Constructing the
Embryo. S.F. Gilbert and A.M. Raunio (eds). Sinauer, Sunderland, MA.
Le Douarin, N.M. and C. Kalcheim. 1999. The Neural Crest. Cambridge University
Press, New York. 2nd edition.
Long, W.L. 1983. The role of the yolk syncytial layer in determination of the plane
of bilateral symmetry in the rainbow trout, Salmo gairdneri, Richardson. The
Journal of Experimental Zoology 228: 91-97.
McCauley, D.W., E. Hixon and W.R. Jeffery. 2004. Evolution of pigment cell
regression in the cavefish Astyanax: A late step in melanogenesis. Evolution and
Macdonald, R., K. Anukampa Barth, Q. Xu, N. Holder, I. Mikkola and S. Wilson.
1995. Midline signalling is required for Pax6 gene regulation and patterning of
the eyes. Development 121: 3267-3278.
Menuet, A., A. Alunni, J-S. Joly, W.R. Jeffery and S. Rétaux. 2007. Shh overexpression
in Astyanax cavefish: Multiple consequences on forebrain development and
evolution. Development 134: 845-855.

Mitchell, R.W., W.H. Russell and W.R. Elliot. 1977. Mexican eyeless characin fishes,
genus Astyanax: Environment, distribution, and evolution. Special Publications
of the Museum, Texas Tech University 12: 1-89.
Morales, A.V., J.A. Barbas and M.A Nieto. 2005. How to become neural crest: From
segregation to delamination. Seminars in Cell and Developmental Biology 16: 655-
662.
Oeting, W.S. and R.A. King 1999. Molecular basis of albinism: Mutations and
polymorphisims of pigmentation genes associated with albinism. Human
Mutation 13: 99-113.
Protas, M., M. Conrad, J.B. Gross, C. Tabin and R. Borowsky. 2007. Regressive
evolution in the Mexican cave tetra, Astyanax mexicanus. Current Biology 17:
452-454.
Protas, M.E., C. Hersey, D. Kochanek, Y. Zhou, H. Wilkens, W.R. Jeffery, L.T. Zon,
R. Borowsky and C.J. Tabin. 2006. Genetic analysis of cavefish reveals molecular
Püschel, A.W., P. Gruss and M. Westerfield. 1992. Sequence and expression pattern
of pax-6 are highly conserved between zebrafish and mice. Development 114:
643-651.
Rasquin, P. 1947. Progressive pigmentary regression in fishes associated with cave
environments. Zoologia 32: 35-44.
Rosenblatt, S., D. Durham-Pierce, J.M. Garner, Y. Nakatsu, M.H. Brilliant and S.J.
Orlow. 1994. Identification of a melanosomal membrane protein encoded by
the pink-eyed dilution (type II oculocutaneous albinism) gene. Proceedings of the
National Academy of Sciences of the United States of America USA 91: 12071-12075.
Sadoglu, P. 1957. A Mendelian gene for albinism in natural cave fish. Experientia
13: 394.
Schwarz, M., F. Cecconi, G. Berneir, N. Andrejewski, B. Kammandel, M. Wagner and
P. Gruss. 2000. Spatial specification of mammalian eye territories by reciprocal
transcriptional repression of Pax2 and Pax6. Development 127: 4325-4334.
Soares, D., Y. Yamamoto, A.G. Strickler and W.R. Jeffery. 2004. The lens has a
specific influence on optic nerve and tectum development in the blind cavefish
Astyanax. Developmental Neuroscience 26: 308-317.
Strecker, U., L. Bernachez and H. Wilkens. 2003. Genetic divergence between
cave and surface populations of Astyanax in Mexico (Characidae, Teleostei).
Molecular Ecology 12: 699-710.
Strecker, U., V.H. Faúndez and H. Wilkens. 2004. Phylogeography of surface and
cave Astyanax (Teleostei) from Central and North America based on cytochrome
b sequence data. Molecular Phylogenetics and Evolution 33: 469-481.
Strickler, A.G., Y. Yamamoto and W.R. Jeffery. 2001. Early and late changes in
Pax6 expression accompany eye degeneration during cavefish development.
Development Genes and Evolution 211: 138-144.
Strickler, A.G., K. Famuditimi and W.R. Jeffery. 2002. Retinal homeobox genes and
the role of cell proliferation in cavefish eye degeneration. International Journal of
Developmental Biology 46: 285-294.

Strickler, A.G., Y. Yamamoto and W.R. Jeffery. 2007a. The lens controls cell survival
in the retina: Evidence from the blind cavefish Astyanax. Developmental Biology
311: 512-523.
Strickler, A.G., M.S. Byerly and W.R. Jeffery. 2007b. Lens gene expression analysis
reveals downregulation of the anti-apoptotic chaperone a A crystallin during
cavefish eye degeneration. Development Genes and Evolution 217: 771-782.
Take-uchi, M., J.D. Clarke and S.W. Wilson. 2003. Hedgehog signalling maintains
the optic stalk-retinal interface through the regulation of Vax gene activity.
Toyofuku, K., J.C. Valencia, T. Kushimoto, G-E. Costin, V.M. Virador, W.D. Viera,
V.J. Ferrans and V.J. Hearing. 2002. The etiology of oculocutaneous albinism
(OCA) type II: The pink protein modulates the processing and transport of
tyrosinase. Pigment Cell Research 15: 217-224.
(Characidae, Pisces). Evolutionary Biology 23: 271-367.
Yamamoto Y. and W.R. Jeffery 2000. Central role for the lens in cavefish eye
degeneration. Science 289: 631-633.
Yamamoto, Y., D.W. Stock and W.R. Jeffery. 2004. Hedgehog signalling controls eye
degeneration in blind cavefish. Nature 431: 844-847.
Yamamoto, Y., L. Espinasa, D.W. Stock and W.R. Jeffery. 2003. Development
and evolution of craniofacial patterning is mediated by eye-dependent and-
independent processes in the cavefish Astyanax. Evolution and Development
5: 435-446.
Yoshizawa, M. and W.R. Jeffery. 2008. Shadow response in the blind cavefish
Astynanx reveals conservation of a functional pineal eye. The Journal of

CHAPTER 7
Subterranean Fishes of North
America: Amblyopsidae
Matthew L. Niemiller1 and Thomas L. Poulson2
1
Department of Ecology and Evolutionary Biology, University of
Tennessee, Knoxville, Tennessee, 37996, USA
E-mail: mniemill@utk.edu
2
Emeritus Professor, University of Illinois-Chicago
E-mail: tomandliz@bellsouth.net
INTRODUCTION
The Amblyopsid cavefishes, family Amblyopsidae, have been viewed as
a model system for studying the ecological and evolutionary processes
of cave adaptation because the four cave-restricted species in the family
represent a range of troglomorphy that reflect variable durations of
isolation in caves (Poulson 1963, Poulson and White 1969). This group has
both intrigued and excited biologists since the discovery and description
of Amblyopsis spelaea, the first troglobitic fish ever described, in the
early 1840s. Other than the Mexican cavefish (Astyanax fasciatus), cave
Amblyopsids are the most comprehensively studied troglobitic fishes
(Poulson, this volume).
The Amblyopsidae (Fig. 1) includes species with some unique features
for all cavefish. Typhlichthys subterraneus is the most widely distributed of
any cavefish species. Its distribution spans more than 5° of latitude and
1 million km2 (Proudlove 2006). Amblyopsis spelaea is the only cavefish
known to incubate eggs in its gill chamber. In fact, this species is the only
one of the approximately 1100 species in North America with this behavior.
The Amblyopsidae is the most specious family of subterranean fishes in the
United States containing four of the eight species recognized. Two other

Fig. 1 Members of the Amblyopsidae. The family includes (A) the surface-
dwelling swampfish (Chologaster cornuta), (B) the troglophile spring cavefish
(Forbesichthys agassizii), and four troglobites: (C) the southern cavefish
(Typhlichthys subterraneus), (D) the northern cavefish (Amblyopsis spelaea), (E)
the Ozark cavefish (A. rosae), and (F) the Alabama cavefish (Speoplatyrhinus
poulsoni). Photos courtesy of Uland Thomas (A), Dante Fenolio (E), and Richard
Mayden (F).
book.

Matthew L. Niemiller and Thomas L. Poulson 171
families have species adapted to subterranean habitats, the Ictaluridae

and Cottidae. Two species of ictalurids are endemic to the Edwards
Aquifer of Texas, the widemouth blindcat (Satan eurystomus) and the
toothless blindcat (Trogloglanis pattersoni); and troglomorphic populations
of sculpins (Cottus sp.) are known from Missouri and Pennsylvania.
Our primary aims of this chapter are to provide an up-to-date,
comprehensive review of the pertinent literature and unpublished research
(Niemiller on phylogenetics and phylogeography and Poulson on pigment,
foraging patterns, metabolic efficiencies, and metabolic acclimation to
temperature) regarding Amblyopsid cavefishes, including their surface
and troglophilic relatives in the family. We begin with a synopsis of the
history of Amblyopsid research from the early 1840s to the present day.
Here we discuss the significant studies and researchers that have shaped
the knowledge base. Next, we review the systematic relationships and
taxonomy of the family. The Amblyopsidae have only recently been the
subject of phylogenetic examination using molecular markers and we
include preliminary data here. We then discuss the biogeography, habitat,
and distribution of the family. In the next section, we review morphology
of the family in general and morphology of cave adaptation in particular.
We include some new data on pigment systems. Subsequently, we
summarize aspects of ecology, life history, and behavior including topics
such as reproduction, diet, parasites and disease, longevity, metabolism,
and demography. Here we also include some new data. Finally, we end
with a discussion of the conservation status and threats to members of
the family. In each section, we suggest avenues of future research that are
needed.
Our review is intended to be as detailed and comprehensive as possible
and to include the majority of relevant references. With this in mind, we
hope that this chapter will serve as both a stimulus for future research and
an exhaustive bibliographic reference regarding Amblyopsid biology.
I. HISTORY OF AMBLYOPSID RESEARCH

In this section, we provide a brief history of Amblyopsid research beginning
with the discovery of the first member of the family in the early 1840s. By
no means is this review meant to encompass all papers on Amblyopsids.
Instead, we attempt to outline the major studies and prominent players
in Amblyopsid research, particularly during the early years from 1840 to
1910. For more complete reviews on the history of hypogean fish research
including Amblyopsids, readers should peruse the works of Romero
(2001) and Proudlove (2006).

The description of the northern cavefish, Amblyopsis spelaea (= A.

spelaeus), from the River Styx in Mammoth Cave, Kentucky, by DeKay (1842)
represents the first scientifically acknowledged report of a troglomorphic
fish species (Romero 2001, Proudlove 2006). However, probable reports of
A. spelaea occurred earlier. James Flint in 1820 recounted a conversation
with a local in Indiana who stated that a neighbor found ‘blind fishes’
when digging a well (Flint 1820, Romero 2001, Proudlove 2006). The first
published account of A. spelaea from the Mammoth Cave area appeared
in the book An Excursion to the Mammoth Cave, and the Barrens of Kentucky
by Robert Davidson (Davidson 1840). His observations represent the first
biological observations of the species, and of a troglobitic fish (Proudlove
2006).
The original description of A. spelaea by DeKay was not very detailed
but incited the interest of other researchers. Wyman (1843ab, 1851, 1854ab,
1872) conducted first detailed examination of the internal anatomy and
brain. Tellkampf (1844) provided detailed descriptions of A. spelaea from
Mammoth Cave. Amblyopsids were of particular interest to Jean Louis
Agassiz who published the first insight on the potential importance of
troglomorphic fishes to biological research in 1847 (Romero 2001). Agassiz
(1847) described a research plan involving A. spelaea to investigate the
embryology, anatomy, and effects of light on the species. Although he
never followed through with his plans, several of his students investigated
aspects of Amblyopsid biology (see below). Agassiz (1853) did, however,
describe the epigean swampfish, Chologaster cornuta (= C. cornutus).
Between 1850 and 1900, troglomorphic fishes, and Amblyopsids in
particular, were prominent in the debate on evolution. Several prominent
biologists including Frederick Ward Putnam (a former student of Agassiz),
Alpheus Packard, Edward Lankester, Alpheus Hyatt, Jeffries Wyman,
and Carl H. Eigenmann discussed the evolution of troglomorphy in cave
fauna, and of cavefish especially. Their influence on and contributions to
cave biology are eloquently detailed by Romero (2001). Also during this
time period, four Amblyopsid species were described (three are currently
recognized). The southern cavefish, Typhlichthys subterraneus, was
described by Girard (1859) from a well near Bowling Green, Kentucky.
Girard thought this new species was a transitional form between A. spelaea
and the entirely epigean species, C. cornuta. In 1872, Frederick Ward
Putnam described the spring cavefish from a well near Lebanon, Tennessee
(Putnam 1872). Ten years later, Forbes (1882) described C. papilliferus.
At the end of nineteenth and into the twentieth century, Carl H.
Eigenmann was the prominent figure in cave research. Between 1887
and 1909, much of his work focused on understanding the loss of visual
structures in cave vertebrates. Amblyopsids were included in many of

his studies. In 1898, Eigenmann described the Ozark cavefish, Amblyopsis

rosae (= Typhlichthys rosae). In 1905, Eigenmann (1905) described both
T. osborni and T. wyandotte. Both species would later be synonymized
under T. subterraneus. Between 1898 and 1905 alone, he published
at least 39 papers and abstracts on cave vertebrates focusing on the
morphology and development of loss of vision (Romero 2001). His work
culminated in his 1909 book titled Cave Vertebrates of North America: A
Study in Degenerative Evolution (Eigenmann 1909) in which he detailed
the structure, development, and degeneration of the eye in several cave-
dwelling vertebrates known to science at the time. Amblyopsids were a
major focus of his investigations in this volume. The volume also includes
an important chapter by Fernandus Payne on laboratory observations and
experiments on the feeding and sensory capabilities of A. spelaea. Between
the writings of Eigenmann and the early 1960s, few studies centered on
Amblyopsids.
However, the family received renewed interest from biologists
beginning with the systematic investigation by Woods and Inger
(1957). In 1955, Thomas L. Poulson began a comparative study of the
Amblyopsidae. His Ph.D. dissertation included aspects of morphology,
physiology, life history, ecology, and behavior. Using Amblyopsids
as a ‘natural evolutionary experiment’, Poulson (1960, 1963) inferred
evolutionary and ecological patterns of cave adaptation and attempted to
outline a step-by-step process of increasing subterranean (troglomorphic)
specialization from preadapted surface to obligate cave-dwelling species.
Poulson has continued to study evolutionary and ecological aspects of
cave adaptation to the present day. His multiple hypotheses and tests of
these hypotheses have influenced many other researchers. In fact, the most
recent Amblyopsid to be described, the Alabama cavefish (Speoplatyrhinus
poulsoni), was named in his honor (Cooper and Kuehne 1974), partly
because of his predictions of what the next stage of troglomorphy would
look like. Much of our knowledge about the biology of Amblyopsid
cavefishes has been obtained by the numerous studies he has conducted,
many of which are referenced in more detail in this chapter. He continues
to study these fishes and several new workups of old unpublished data
are included in this chapter.
Amblyopsid cavefishes continue to intrigue and excite biologists
attempting to discern the ecological and evolutionary facets of cave
adaptation. Although many studies in the last twenty years have focused
on the demography, distribution, conservation status, and threats of the
cave-dwelling species, other aspects of biology including the phylogenetic
relationships, biogeography, and life history continue to evoke interest.
We have attempted to summarize what we feel are the most relevant

references for each of these areas in subsequent sections of this chapter.

We apologize in advance to authors of any extensive or insightful studies
we may have missed and hope the authors will work with us to redress
any such omissions.
II. TAXONOMY AND SYSTEMATICS

The Amblyopsidae have been known to science since the 1840s. The family-
group name (Amblyopsidae-Amblyopsini) was first used by Bonaparte
(1846). Previously, the family names Heteropygii and Hypsaeidae were
used by Tellkampf (1844, 1845) and Storer (1846) respectively, but these
names are considered unavailable because they were not formed from
the stem of an available genus-group name (Poly and Proudlove 2004).
Aspects of taxonomy and systematics of the family have been examined
by Cox (1905), Eigenmann (1909), Woods and Inger (1957), Swofford et
al. (1980), Swofford (1982), Bergstrom et al. (1995), Bergstrom (1997), and
Niemiller and Fitzpatrick (2008), Near et al. (unpublished data). The
latter four studies were molecular in nature. More recently, Poly and
Proudlove (2004) reviewed taxonomy and systematic relationships of the
Amblyopsidae.
Presently five genera and six species are recognized within the
Amblyopsidae (Fig. 1). These include the epigean swampfish (Chologaster
cornuta Agassiz), a troglophile, the spring cavefish (Forbesichthys agassizii
Putnam), and four troglobitic species in order of presumed increasing
time of isolation in caves: southern cavefish (Typhlichthys subterraneus
Girard), northern cavefish (Amblyopsis spelaea DeKay), Ozark cavefish
(A. rosae Eigenmann), and the Alabama cavefish (Speoplatyrhinus poulsoni
Cooper and Kuehne). We will briefly discuss the higher-level relationships
of the Amblyopsidae before discussing nomenclature and interspecific
relationships within the family.
A. Higher-level Relationships
The Amblyopsidae have been considered the sister group to the pirate
perches, family Aphredoderidae (Rosen 1962, Patterson 1981, Patterson
and Rosen 1989) but also share a close affinity with the trout perches,
family Percopsidae (Rosen 1962). Together, these three families have been
included in the order Percopsiformes (Greenwood et al. 1966, Nelson
1984, 2006). However, some have questioned the monophyly of this order.
Murray and Wilson (1999) suggested Amblyopsids might be more closely
related to the Anacanthines and proposed recognition of Amblyopsids
as a distinct order, the Amblyopsiformes. Poly in Poly and Proudlove

(2004) suggested that Amblyopsids might be related to Gobioids because

of the distribution of Amblyopsids in relation to the former Mississippi
Embayment (see below), similarities in patterns of neuromasts on the
body, and similarities in morphology, particularly when compared
with troglobitic Gobioids. McAllister (1968) also suggested a possible
relationship between Gobioids and Amblyopsids. Wiley et al. (2000)
resolved a sister relationship between trout perches and pirate perches
based on a combined analysis of morphological and molecular characters
but did not include any Amblyopsid species. More recently, a molecular
study using both mitochondrial and nuclear markers and including
members of all three families supports the inclusion of the Amblyopsidae
in the order Percopsiformes with a sister relationship between cavefishes
and pirate perches (Smith and Wheeler 2006). Immunological evidence also
supports the inclusion of the Amblyopsidae in the order Percopsiformes
(Kalayil and Clay 1976).
B. Taxonomy
Agassiz (1853) described the swampfish (Fig. 1A) from ditches in rice
fields in South Carolina. Since the description, taxonomy has remained
stable. Only two synonyms exist: Chologaster cornutus (Agassiz 1853) and
C. avitus (Jordan and Jenkins in Jordan 1889). The swampfish generally is
considered the most basal Amblyopsid. Like the other Amblyopsids, little
molecular work has been conducted on the species. The most substantial
is the allozyme study by Swofford (1982) who found considerable
differentiation among C. cornuta populations both among and within
drainages. Preliminary evidence using mitochondrial and nuclear DNA
markers also indicate significant genetic differentiation across the species
range (Niemiller and Fitzpatrick 2008, Near et al., unpublished data).
Poulson (1960, 1963) also found considerable variation in adult body size
and meristic characters among drainages.
The spring cavefish (Fig. 1B) was originally described as C. agassizi by
Putnam (1872) from a well near Lebanon, Wilson Co., Tennessee. Later,
Forbes (1882) described C. papilliferus from a spring in western Union
Co., Illinois, on the basis of coloration differences between the Tennessee
and Illinois populations. Jordan and Evermann (1927) erected a new
genus, Forbesella, citing that the subterranean nature of spring cavefish
warrants separate recognition. Jordan (1929) later replaced Forbesella with
Forbesichthys, as the former was preoccupied in tunicates. This genus is
still considered a junior synonym of Chologaster by some authors, however.
Woods and Inger (1957) noted that populations of spring cavefish from
southern Illinois, central Kentucky, and central Tennessee all differed
slightly but did not warrant specific or subspecific designation. Therefore,

C. papilliferus was synonymized under C. agassizi and their revision has

been followed by most subsequent authors with the exception of Clay
(1975) who maintained that C. agassizi and C. papilliferus are specifically
distinct. Allozyme analyses by Swofford (1982) revealed considerable
differentiation between populations that justified resurrection of the genus
Forbesichthys, which was elevated later by, Page and Burr (1991). Spring
cavefish have not been the subject of phylogenetic studies since Swofford
(1982); however, recent evidence suggests that populations from Illinois
(formerly papilliferus) and Tennessee may be phylogenetically distinct at
both mitochondrial and nuclear markers (Near et al., unpublished data).
Kentucky populations have yet to be examined.
The southern cavefish (Fig. 1C) was described by Girard (1859) from a
well near Bowling Green, Warren Co., Kentucky. Later, Eigenmann (1905)
described both T. osborni and T. wyandotte based on differences in head
width and eye diameter. Typhlichthys osborni was described from Horse
Cave, Kentucky. Typhlichthys wyandotte was described from a well near
Corydon, Indiana, that was later destroyed. Recently, a well-like entrance
into a cave on the property of a car dealership in Corydon was discovered
and is believed to represent the type locality (Black in Lewis 2002).
Regardless, this species is generally considered invalid and was not listed
as a locality in Woods and Inger (1957). Recent surveys in the vicinity of
Corydon have failed to document T. subterraneus, finding only A. spelaea
(Lewis 1998, Lewis and Sollman 1999). Typhlichthys eigenmanni (nomen
nudeum) was described as a fourth species in the genus from Camden
Co., Missouri (likely River Cave). Recently, Parenti (2006) proposed that
T. eigenmanni Charlton (1933) is a subjective synonym of T. subterraneus.
Woods and Inger (1957) synonymized all species under T. subterraneus
on the basis of lack of any clear geographic pattern in morphological
variation. A population from Sloans Valley Cave, Pulaski Co., Kentucky,
differs in several ways from populations to the southwest in Tennessee
and may represent an undescribed species (Cooper and Beiter 1972, Burry
and Warren 1986). However, further details regarding this population
have not been published.
Because of the extensive distribution of T. subterraneus and the results of
molecular studies of other troglobites (especially Culver et al. 1995), some
authors have speculated that T. subterraneus actually represents several
independent invasions and, therefore, distinct lineages (Swofford 1982,
Barr and Holsinger 1985, Holsinger 2000, Poulson, this volume). Indeed,
electrophoretic allozyme analyses by Swofford (1982) showed considerable
differentiation among morphologically similar populations of Typhlichthys,
even those that are geographically close, suggestive of multiple,
independent invasions and limited gene flow. In fact, six of the thirteen

populations Swofford sampled were monomorphic and for different

allozyme alleles. However, owing to small sample size, Swofford’s study
was limited in its ability to distinguish modular or hierarchical subdivision
from a continuous relationship between genetic and geographic distance.
More recently, Bergstrom et al. (1995) and Bergstrom (1997) investigated
phylogenetic relationships of populations west of the Mississippi River
using mitochondrial DNA. Although limited, these studies revealed
considerable variation among populations. Likewise, Niemiller and
Fitzpatrick (2008) examined genetic variation among eastern populations
of Typhlichthys in Alabama, Georgia, and Tennessee. Again, significant
genetic divergence was observed and both mitochondrial and nuclear
DNA variation was structured among hydrological drainages. Molecular
and morphological evidence also indicate that Arkansas populations
warrant recognition of a distinct species (Graening et al., unpublished
data). These studies support the hypothesis that morphological similarity
is the result of parallel evolution rather than significant dispersal and gene
flow across major drainage and river divides (Cooper and Iles 1971, Barr
and Holsinger 1985).
The northern cavefish (Fig. 1D) was the first troglobitic fish ever
described in the scientific literature. The formal description is based on
fish from Mammoth Cave, Edmonson Co., Kentucky by DeKay (1842).
Originally spelled Amblyopsis spelaeus, Woods and Inger (1957) corrected
the name to A. spelaea and provided the most important definition of the
species (Proudlove 2006). Swofford (1982) found that populations north
and south of the Ohio River are monomorphic for the same allozyme
alleles with no heterozygosity. The data do not contradict Poulson (1960)
who observed a distinct split in morphological variation across the Ohio
River. Rather than indicating pre-Pleistocene dispersal as Poulson (1960)
proposed, Swofford concluded that populations north and south of the
river likely are of more recent origin during the Pleistocene and insufficient
time has elapsed for significant genetic differentiation. Clearly more work
is needed to elucidate the biogeographic history of the species. No studies
to date have examined mitochondrial or nuclear DNA differentiation
within the species.
The Ozark cavefish (Fig. 1E) was first reported from specimens collected
from caves near Sarcoxie, Jasper Co., Missouri, by Garman (1889) and
identified as T. subterraneus. The name Typhlichthys rosae was applied to
the species in several of Eigenmann’s papers in 1898 (for a review see Poly
and Proudlove 2004). The Ozark cavefish was reclassified as Troglichthys
rosae by Eigenmann (1898, 1899a) and this name stood until Woods and
Inger (1957) placed the species in the genus Amblyopsis, on the basis
of morphological similarity with A. spelaea. Phylogenetic studies that

included samples of A. rosae have revealed considerable divergence across

the range. Although Swofford (1982) sampled just two populations, one
from the Neosho River drainage in Jasper Co., Missouri, and the other
from the Illinois River drainage in Benton Co., Arkansas, these populations
were well differentiated and with very low heterozygosities for allozyme
loci. The phylogenetic studies of Bergstrom et al. (1995) and Bergstrom
(1997) also revealed distinct genetic structure associated with each
major drainage. Bergstrom (1997) proposed that mitochondrial genetic
divergences were so pronounced that at least subspecific designation
was warranted. At least four genetically differentiated groups associated
with distinct drainage basins were identified: 1) Illinois River drainage in
northwestern Arkansas, 2) White River drainage in southwestern Missouri,
3) Neosho River drainage in southwestern Missouri, and 4) Neosho River
drainage in northeastern Oklahoma. Genetic variation within drainages
also suggested that many localities are genetically isolated and, therefore,
constitute distinct endemic populations (Noltie and Wicks 2001).
Cooper and Kuehne (1974) described the Alabama cavefish (Fig. 1F),
Speoplatyrhinus poulsoni, from Key Cave, Lauderdale Co., Alabama, in honor
of Thomas L. Poulson. Little is known about how this species is related to
other Amblyopsids. Because S. poulsoni is the most troglomorphic of all the
subterranean Amblyopsids and thought to represent a relict population,
the single cave endemic probably is not closely related to any of the other
species (Proudlove 2006). It has at least three features that are completely
unlike other Amblyopsids: arrangement and relative sizes of otoliths, size
and extent of caudal papillae (Cooper and Kuehne 1974), and absence
of any tactile receptors on the head (Poulson, personal observation). In
addition, there are two extreme neotenic features: the absence of branched
fin rays in adults and the absence of a bilateral supraopercular papilla
opening to the head subdermal lateral line system (Cooper and Kuehne
1974).
C. Interspecific Relationships
Although Amblyopsids have received considerable attention from
evolutionary biologists and ichthyologists, comparatively little work
has been conducted on the systematics of the family, particularly at
the molecular level. Prior to the study of Woods and Inger (1957), nine
species in five genera were recognized. As mentioned above, the authors
synonymized all species of Typhlichthys under T. subterraneus and
moved Troglichthys rosae into the genus Amblyopsis and Forbesichthys into
Chologaster. Likewise, C. papilliferus was synonymized with C. agassizi.
Woods and Inger (1957) suggested two phylogenies representing the

interspecific relationships of the family (Fig. 2A–B) relying primarily on

the presence or absence of two characters: (i) scleral cartilages in the eye
and (ii) postcleithrum in the pectoral girdle. However, a reanalysis by
Swofford (1982) using the same characters as Woods and Inger (1957) does
not support either proposed phylogeny. Instead, Swofford (1982) proposed
the phylogenies presented in Fig. 2C–D. Accordingly, the resurrection
of the genus Forbesichthys was recommended because C. agassizi and C.
cornuta are considerably divergent morphologically according to meristic
characters (Woods and Inger 1957) and genetically (Swofford 1982). This
treatment was preferred over the more drastic alternative, which involved
synonymizing Typhlichthys under Chologaster.
Uyeno (pers. comm.) examined karyotypes of three Amblyopsid
species. Both F. agassizii from Rich Pond, Kentucky, and from Wolf Lake,
Illinois, and A. spelaea had 2n = 24 chromosomes noting that both of these
species had very similar karyotypes. Uyeno also examined the karyotype
Fig. 2 Phylogenetic hypotheses of the interspecific relationships of the

Amblyopidae and according to the (A and B) morphological examinations of
Woods and Inger (1957), (C and D) allozyme studies of Swofford (1982), and
(E) mitochondrial DNA (Niemiller and Fitzpatrick 2008, Near et al., unpublished
data). Faga = F. agassizii, Ccor = C. cornuta, Tsub = T. subterraneus, Aros = A.
rosae, and Aspe = A. spelaea. Forbesichthys agassizii and S. poulsoni are not
represented in the mtDNA phylogeny.

of two populations of T. subterraneus, Cave City, Kentucky, and Thomason

Cave, Alabama. In T. subterraneus, the karyotype was 2n = 46 including
one pair of rather large metacentric chromosomes. Karyotypes have not
been examined for C. cornuta, A. rosae, and S. poulsoni.
Bergstrom et al. (1995) and Bergstrom (1997) were the first to examine
interspecific relationships of Amblyopsids using DNA sequence data.
Using a portion of the mitochondrial NADH-dehydrogenase subunit-2
(NAD2) gene, their analyses resolved a sister relationship between A. rosae
and all other Amblyopsids (S. poulsoni was not included), but low levels
of differentiation were found between C. cornuta, F. agassizii, A. spelaea,
and some T. subterraneus populations. Their ND2 phylogeny clearly
differs from the phylogenetic hypotheses of Woods and Inger (1957) and
Swofford (1982). Unfortunately, DNA samples from C. cornuta, F. agassizii,
and A. spelaea were extracted from formalin-preserved tissues and were
contaminated. Therefore, the relationships resolved likely do not reflect
the true history of the ND2 gene. However, the intraspecific relationships
reported within A. rosae and western populations of T. subterraneus are
believed to be accurate given that extractions were made from freshly
collected tissue.
Recent phylogenetic work utilizing both mtDNA and nuclear DNA
markers has revealed a different relationship (Fig. 2E) among the
Amblyopsids (Niemiller and Fitzpatrick 2008, Near et al., unpublished
data). MtDNA supports the basal position of C. cornuta within the
Amblyopsidae. Monophyly of Amblyopsis is not support with the mtDNA
dataset. However, other nuclear markers need to be examined before
taxonomic revisions are made.
The Alabama cavefish has not been included in any systematic treatment
of the family. Boschung and Mayden (2004) state that Typhlichthys
and Speoplatyrhinus form a monophyletic group that is sister to a clade
comprising Forbesichthys and Chologaster. This statement likely is a typo
and Amblyopsis should be replaced for Speoplatyrhinus. Proudlove (2006)
speculated that S. poulsoni is probably not closely related to other species in
the family. On several counts it is the most neotenic, most troglomorphic,
and most distinct species in the family (Cooper and Kuehne 1974, Poulson
1985; and Taxonomy). The exact placement of Speoplatyrhinus within the
family remains unknown and will remain so until a fresh tissue sample
can be obtained.
III. DISTRIBUTION, BIOGEOGRAPHY AND HABITAT

The present-day distributions of the Amblyopsid species have been
influenced by climatic, geological, and ecological factors throughout their

evolution. All Amblyopsids occur in eastern North America (Fig. 3). All
but one species, the epigean C. cornuta, occur in the Interior Low Plateau
or Ozark Plateau where the ancient former position of the Cretaceous
Mississippi Embayment bounds their collective distributions to the south
and the Pleistocene glacial advance bands the distribution to the north.
In this section we discuss the distribution, phylogeographic history,
and habitat preferences of the each species within the Amblyopsidae
beginning with epigean swampfish and advancing our discussion toward
the troglobitic forms. In the process, we discuss the several hypotheses
that have been posited explaining the biogeographic patterns observed.
The swampfish is the only entirely epigean Amblyopsid and the only
species of the family found in the Coastal Plain of the southeastern United
States (Fig. 4). It occurs from southeastern Virginia to east-central Georgia
(Cooper and Rohde 1980, Rohde et al. 1994). It typically occurs in heavily
vegetated and shaded lowland swamps, swampy creeks, and backwater
habitats that are tannin-stained and acidic (Cooper and Rohde 1980, Rohde
et al. 1994, Ross and Rohde 2003). Water temperatures in these habitats
rarely exceed 25°C (Poulson 1963) and pH ranges 5.7–6.8 (M.D. Norman
in Jenkins and Burkhead 1994). Swampfish also have been collected
from sites with salinities up to 5% in North Carolina but are not found in
brackish water. Woods and Inger (1957) hypothesized that the ancestor to
C. cornuta migrated from the Interior Low Plateau southward and around
the southern Appalachians via continuous swamp habitat that fringed the
location of the border of the ancient Mississippi Embayment. However,
recent molecular evidence suggests an opposite route, as C. cornuta is
the most basal member of the Amblyopsidae (Niemiller and Fitzpatrick
2008, Near et al., unpublished data). A more plausible scenario would be
that a C. cornuta-like ancestor inhabited swamps and wetlands adjacent
to the karst regions of the Interior Low and Ozark Plateaus and likely
the Coastal Plain during the Miocene. As the climate became more arid
during this period, some populations became isolated in springs and caves
eventually giving rise to the troglobitic forms, whereas other populations
remained living in a swamp-like habitat.
However, the absence of C. cornuta from the Gulf Coastal Plain has
puzzled past researchers. Woods and Inger (1957) felt that C. cornuta formerly
inhabited this area in the past because several community associates, such
as Aphredoderus sayanus and Umbra pygmaea, have distributions that range
into the Gulf Coastal Plain. They speculated that C. cornuta might have
been extirpated from the Gulf Coastal Plain when their swamp-like habitat
dried up because of prolonged drought and was not able to recolonize
because of poor dispersal ability. According to the scenario outlined above,
we also believe C. cornuta or a cornuta-like ancestor inhabited portions of
the Gulf Coastal Plain and was extirpated. Meristic data (Poulson 1960)


Fig. 3 Distribution by county of the Amblyopsidae in the eastern United States. Only the swampfish, C. cornuta, is found outside
the Interior Low Plateau or Ozark Plateau.
Color image of this figure appears in the color plate section at the end of the book.
and both allozyme data (Swofford 1982) and DNA sequence data (Near
and Niemiller, unpublished data) indicate that dispersal in C. cornuta is
limited. Swofford (1982) observed significant differentiation even among
populations in the same hydrological drainage and concluded that C.
cornuta is highly specific in its habitat requirements, and, consequently, is
unable to disperse over long distances unless corridors of suitable habitat
are available. Because C. cornuta cannot tolerate brackish water (Woods
and Inger 1957), dispersal may only be achieved by headwater exchange
(Swofford 1982). However, fine-scale genetic analyses are needed to test
this hypothesis.
The spring cavefish occurs from south-central Tennessee northward
into central and western Kentucky, then westward following the Shawnee
Hills of southern Illinois and the Benton Hills west of the Mississippi
River in southeastern Missouri (Fig. 3). Spring cavefish occur at the
interface of epigean and subterranean habitats bridging the threshold
toward a troglobitic life. Although it occurs in caves throughout most of
its range, F. agassizii is most abundant in springs, spring runs, and spring
seeps (Smith and Welch 1978). Both habitat types are utilized, however, as
fish often emerge from subterranean haunts at dusk to feed and then later
retreat back underground before dawn. Fish also can be found underneath
rocks in springs and spring runs during the day. In Tennessee, F. agassizii
is common in dense vegetation associated with springs and spring-fed
streams (Etnier and Starnes 1993) and have been collected in very low
numbers in caves. Illinois populations in the LaRue-Pine Hills Ecological
Area have been taken at springs in the swamp during winter (Smith and
Welch 1978), but it is unknown whether fish actively use or successfully
disperse through swampy, lentic habitats.
In Kentucky and Tennessee, spring cavefish occur in springs and
caves, including the Mammoth Cave system, from the Highland Rim
physiographic province of the Tennessee River drainage, middle and
lower Cumberland River drainage, and the upper Barren-Green River
system of Kentucky (Etnier and Starnes 1993). Spring cavefish have not
been collected in Alabama, but the close proximity of a population in the
Elk River drainage in Franklin Co., Tennessee (Armstrong and Williams
1971), led Boschung and Mayden (2004) to speculate that the species likely
occurred in springs associated with the Elk River system in Alabama in
the past. Forbesichthys agassizii has been collected from springs issuing
from the base of limestone bluffs of the Shawnee Hills and Benton Hills
at the edge of the Mississippi River alluvial plain in southwestern Illinois
(Weise 1957) and southeastern Missouri (McDonald and Pflieger 1979)
respectively. Until recently, the Shawnee Hills and Benton Hills populations
likely were continuous and were isolated around 2,000 years ago when
the Mississippi River was diverted through the narrow gap between these
regions (McDonald and Pflieger 1979).

Forbesichthys shows a little less allozyme differentiation than C. cornuta

with an average heterozygosity of 0.028 compared to 0.040 (Swofford
1982). This suggests that F. agassizii populations in Kentucky and portions
of the Highland Rim in Tennessee utilize surface corridors for dispersal
(Swofford 1982), particularly when compared with more differentiation
in troglobitic T. subterraneus populations over a similar geographic area.
However, molecular evidence indicates that populations of Forbesichthys
from Illinois and Tennessee are distinct and likely isolated and suggestive
that long distance dispersal is rare. Although cave populations of F. agassizii
exist, subterranean dispersal seems unlikely, given low abundance in caves,
low tolerance to starvation, relatively poor food finding ability, and short
life span (see below). But surface dispersal along rivers between springs
may be possible since Woods and Inger (1957) report several specimens
collected from surface streams.
Of the four troglobitic Amblyopsids, three occur in the Interior Low
Plateau and the other occurs in the Ozark Plateau (Fig. 3). Collectively,
their distributions lie north of the boundary of the ancient Mississippi
Embayment and south of the Pleistocene glacial boundary. Woods and
Inger (1957) proposed that a surface ancestor entered caves sometime
during the Tertiary Period and gradually dispersed into the areas where
they occur today.
However, it is unlikely that subterranean dispersal alone can account
for the widespread collective distribution of the troglobitic Amblyopsids,
and of T. subterraneus in particular. Dispersal through surface watercourses
is unlikely (Poulson 1960, Woods and Inger 1957) given the separation
of some populations by over 1000 surface river miles (Woods and Inger
1957). Fish are very rarely observed in surface streams just downstream of
springs after flood events. Although most fish likely perish, some might
find their way back underground further downstream. It is much more
likely that fish could move through cavernous river bottoms or solution
channels (Poulson 1960), however. Although little direct evidence supports
this avenue as a significant form of dispersal, indirect evidence indicates
that surface rivers do not significantly impede dispersal of subterranean
species. First, the thick horizontal Ordovician and Mississippi limestone
formations of the Interior Low Plateau and Ozark Plateau and the
groundwater systems contained therein pass underneath major surface
rivers, including the Cumberland, Tennessee, Ohio, and Mississippi
rivers. Populations of the troglobitic species occur on either side of these
big rivers (e.g., A. spelaea north and south of the Ohio River; T. subterraneus
north and south of the Tennessee River in Alabama) and, therefore, rivers
do not appear to pose significant barriers to dispersal (Woods and Inger
1957). Rotenoned Southern cavefish apparently floated to the river surface

from conduits in a limestone riverbed during a fish survey at a prospective

dam site (Mohr and Poulson 1966). Second, populations on opposite
sides of major rivers often show little genetic variation. Populations of T.
subterraneus in the same hydrological drainage but on opposite sides of a
river are poorly genetically differentiated (Niemiller and Fitzpatrick 2008)
indicating that either these populations have been recently isolated or that
some migration is occurring between them.
Contemporary surface drainage divides also may not impede
subterranean dispersal. Woods and Inger (1957) reported the presence of T.
subterraneus from a spring well near the crest of the Arkansas/White River
watershed divide. The spring well lies at an elevation of 1360 ft at the top of
the Springfield Plateau and fish potentially can move from one watershed
to the other. This is supported by the observation of very genetically
different southern cavefish within the same population in the Salem Plateau
(Bergstrom 1997). Although this is consistent with movement between
watersheds, there are other explanations for considerable genetic diversity
within a population such as lineage sorting. The recharge basins in this area
are large and contiguous with few dry or faulted barriers between basins.
Therefore, some individuals can potentially move between recharge basins
(Noltie and Wicks 2001). However, other molecular evidence indicates
that many eastern populations of T. subterraneus are confined to distinct
hydrological drainage basins (Niemiller and Fitzpatrick 2008), as genetic
structure is highly correlated with hydrological structure. Poulson (1960)
argued that populations of T. subterraneus along the Eastern Highland Rim
and western escarpment of the Cumberland Plateau were more similar than
elsewhere in the range. He hypothesized that the observed homogeneity
was because of occasional dispersal by cave headwater stream capture.
However, both the electrophoretic data of Swofford (1982) and genetic
analyses of (Niemiller and Fitzpatrick 2008, Niemiller et al., unpublished
data) indicate that morphological similarity of these populations is not
due to homogenizing gene flow.
The southern cavefish is the most widely distributed troglobitic fish
in North America. Its range is discontinuous and divided into two
main components east and west of the Mississippi River: an eastern
component that extends along the Cumberland Plateau and through
Interior Low Plateau from central Kentucky (Mammoth Cave region)
southward into central Tennessee, northern Alabama, and northwestern
Georgia, and a western component that includes the Ozark Plateau of
central and southeastern Missouri and northeastern Arkansas (Fig. 3).
The distribution of the species was once thought to include southern
Indiana and northeastern Oklahoma. These records are now thought
to be erroneous. Mayden and Cross (1983) showed that all specimens
of southern cavefish from northeastern Oklahoma are in fact A. rosae.

Similarly, records of T. subterraneus from Green Co., Missouri, are incorrect

and actually represent A. rosae (Jones and Taber 1985). Therefore, A. rosae
and T. subterraneus should no longer be considered sympatric (Proudlove
2006). Eigenmann (1905) described T. wyandotte from a well near Corydon,
Indiana, in the heart of the range of A. spelaea and this was almost certainly
a misidentification.
There do not appear to be clear differences among the habitats of
Amblyopsid cavefish. Most cavefish habitats are at or near the water table
(Poulson 1963) and the types of organic and inorganic substrates vary
enormously within and among caves for all four troglobitic Amblyopsids
(e.g., Fig. 3 in Poulson 1963). Also, water velocity among seasons and among
years varies tremendously for most cavefish habitats. Nevertheless, some
intriguing regional differences in geology and related size and integration
of caves in the Ozarks exist for T. subterraneus and A. rosae. These differences
may relate to food supply and degree of isolation (see below and Poulson,
this volume). In addition, the geology of the karstic strata inhabited
by T. subterraneus, at least in the Salem Plateau, suggests that southern
cavefish reside at appreciable depth below the surface (Noltie and Wicks
2001). As mentioned above, evidence suggests that this species can move
between recharge basins; however, dispersal between basins appears to
be low (Swofford 1982, Niemiller and Fitzpatrick 2008, Niemiller et al.,
unpublished data). In contrast, A. rosae lives in cave systems that are
smaller, shallower, and much less likely to be interconnected (Noltie and
Wicks 2001).
The extensive but highly disjunct geographic range of T. subterraneus
may be the result of multiple, independent invasions and subsequent
dispersal (Swofford 1982, Barr and Holsinger 1985, Holsinger 2000,
Poulson, this volume). Woods and Inger (1957) hypothesized a single
invasion of subterranean waters by Typhlichthys and a separate invasion for
Amblyopsis. However, others have proposed a multiple-invasion scenario
for Typhlichthys as well as the other troglobitic species. Swofford (1982)
offered two possible scenarios to explain the distribution of T. subterraneus.
First, the ancestor of T. subterraneus was a widely distributed epigean species
preadapted to cave life and possibly similar to F. agassizii today. During the
Pleistocene, it originally occupied a more northern distribution but was
forced southward during glacial periods then underground during the
warmer and drier interglacial periods. Alternatively, the surface ancestor
may have resembled F. agassizii, having occupied a variety of habitats
from streams to springs to caves. As climatic conditions changed, surface
populations were extirpated leaving only cave populations scattered over
a broad area but genetically isolated from one another. In both of these
scenarios, some range expansions through subterranean dispersal may
occur accounting for some of the current distribution patterns that are
difficult to interpret otherwise (Swofford 1982).

Swofford’s alternative scenarios also can be applied to both Amblyopsis

species. Willis and Brown (1985) hypothesized a similar vicariant scenario
in which the surface ancestor of cave Amblyopsids inhabited swamps
during the wetter periods of the Pleistocene becoming restricted to
swamp-like springs and caves during drier periods. Rather than just one
or two invasions, possibly each current cavefish population represents a
separate invasion event. The actual number of independent subterranean
invasions likely falls somewhere in between the two extremes. Recent
molecular evidence also indicates that the major lineages within the
Amblyopsidae originated well before the Pleistocene (Bergstrom 1997,
Near et al., unpublished data), however the glacial cycles likely had a
profound effect on the current distributions and intraspecific relationships
within each species. These studies also support a multi-invasion scenario
and support the hypothesis that morphological similarity is the result of
parallel evolution rather than significant dispersal and gene flow across
major drainage and river divides.
If current taxonomy is indicative of past evolutionary history, then
a plausible biogeographic hypothesis explaining the distributions of
both Amblyopsis species is difficult to frame. Woods and Inger (1957)
proposed two hypotheses to explain the distribution of Amblyopsis. First,
the disjunct ranges of the two species could have resulted by dispersal
from a more northern former distribution above the limit of the glacial ice
sheets. The ancestor was subsequently split and diverged as a direct result
of Pleistocene glaciations. Since no evidence has ever indicated that any
Amblyopsid existed north of the karst regions of the Interior Low Plateau,
they favored a second hypothesis that postulated an ancestral Amblyopsis
that ranged widely through the limestone plateaus and were isolated into
two groups by range expansion of the more competitive Typhlichthys.
Bergstrom (1997) places the timing of the split between Typhlichthys and
Amblyopsis around 8 mya and provides a biogeographic scenario consistent
with this and subsequent splits within the respective lineages (Noltie and
Wicks 2001).
Recent molecular is inconsistent with the above scenarios since
Amblyopsis, as currently recognized, does not form a monophyletic group.
“with A. rosae diverging much earlier.’’ Obviously, more research is
needed to elucidate the large-scale biogeographic patterns of the family.
At a finer scale, considerable work has been conducted on the distribution
and habitat of A. spelaea and A. rosae.
The northern cavefish follows a narrow arc of karst, from near the
city of Bedford in Lawrence Co., south-central Indiana, southward to the
Mammoth Cave area in central Kentucky. It has a distribution through
portions of ten counties (Fig. 3). Karst landscapes start with the Mitchell

Plain in Indiana and continue as the Pennyroyal Plain in Kentucky. In the

most recent assessment of the species distribution, Pearson and Boston
(1995) reported 76 localities from five counties in Indiana and 38 localities
from five counties in Kentucky; however, some localities no longer exist
because of quarry operations and highway construction. Northern cavefish
have the largest populations in caves with deep pools and moderately
deep shoals. The best habitats have ledges, overhangs, and backwater
areas that serve as refugia during floods. Substrate type and particle size
are quite variable (Fig. 3 in Poulson 1960). The cave types flood regularly
and have high food supplies of plankton and isopods (see below, Food
Supplies in Caves).
The distribution of the northern cavefish is bounded by the East Fork
White River to the north. Harsh periglacial conditions and alteration
and impounding of major rivers north of this area may have limited the
northern range of A. spelaea (Keith 1988). The southern extent of the species’
range is the Mammoth Cave system. It is not clear why its distribution
does not extend further to the south or immediately east to the sinkhole
plain where the range of T. subterraneus starts. In Mammoth Cave itself, T.
subterraneus and A. spelaea occur in different but contiguous habitats.
Mammoth Cave is one of only a few sites with syntopic species of
cavefish. Typically only a single species of troglobitic fish is found within
a cave system (Weber 2000) and this is the case within the Amblyopsidae.
However, two instances of syntopy have been observed in the family. First,
A. spelaea, T. subterraneus, and the troglophile F. agassizii co-occur within
the Mammoth Cave system in Kentucky. In the Mammoth Cave system, T.
subterraneus and A. spelaea are syntopic only where habitat is sufficiently
diverse to allow habitat segregation (Poulson and White 1969, Poulson
1992, Poulson, this volume).
In the Mammoth Cave system, segregation of the three species of
Amblyopsids has been hypothesized as follows. Spring cavefish are
washed into the cave from sinking streams during flood events but only
survive as long as a year when the subterranean food supply is well
above average. Although the species can subsist on the increase in food
availability, reproduction is unlikely as not enough food can be consumed
to support its energetic demands and higher reproductive output compared
with the troglobitic species. With regards to the two troglobitic species
that inhabit the Mammoth Cave system, their distributions can possibly
be explained by competitive interactions (Woods and Inger 1957, Poulson
1992). Southern cavefish are found in the master drains of vertical shafts,
whereas A. spelaea inhabit deeper water with decreasing food availability
further downstream. It remains unclear whether A. spelaea is precluded
from upstream sections because of its larger size or by aggressive

dominance of T. subterraneus (Poulson 1992). Alternatively, T. subterraneus

may be excluded from downstream habitats by decreased food supply
or by competitive dominance by A. spelaea. The other observed syntopic
interaction between two Amblyopsids involves T. subterraneus and S.
poulsoni and is discussed below.
The distribution of A. rosae is limited to caves in the Springfield Plateau
of the Ozark Highlands Province of northwestern Arkansas, southwestern
Missouri, and northeastern Oklahoma. (Fig. 3) The Springfield Plateau
encompasses ca. 21,000 km2 that is drained by the White River to the south
and east, the Neosho River to the west, and the Osage River to the north.
Historically, A. rosae occurred in 24 caves in nine counties with unconfirmed
reports increasing occurrence to 52 caves in 14 counties (Brown and Todd
1987). Willis and Brown (1985) confirmed A. rosae in just four caves in
Arkansas, five in Missouri, and three in Oklahoma. Brown and Todd (1987)
reported three additional localities in Benton Co., Arkansas. Currently, A.
rosae is known from 47 localities (G.O. Graening, unpublished data): 11
caves in one county in Arkansas, 24 in seven counties in Missouri, and
12 in two counties Oklahoma. These localities occur in the main section
of the plateau; however, no localities are known from the section of the
plateau that extends southeast into Arkansas and from the central area in
and around McDonald Co., Missouri (Noltie and Wicks 2001).
Ozark cavefish are confined to the karstic layer at or immediately
adjacent to the surface based on the stratigraphy of the Springfield Plateau
limestones and underlying shale layer (Noltie and Wicks 2001). The
majority of cavefish localities are caves developed in the shallow and thin
Boone and Burlington limestones and intersect the water table close to the
surface (Noltie and Wicks 2001). Fish are generally observed in two types
of subterranean habitats: 1) laterally-oriented cave streams that occupy
conduits accessible by humans that intersect the surface near the base of
Springfield Plateau escarpment either along the bluff line or of river valley
walls, and 2) karst windows and wells that typically intersect conduits
that are too small for humans to penetrate (Poulson 1960, USFWS 1989,
Noltie and Wicks 2001). Cave streams inhabited by A. rosae typically are
small. Karst window and well localities often occur toward the interior
of the plateau and are not associated with the plateau escarpment. These
are windows to human-inaccessible habitat that may be phreatic. These
areas may serve as refugia, but also may provide avenues for dispersal
and gene flow. Poulson (this volume and below) believes that these areas
contain a very small portion of the overall cavefish population. Genetic
studies are consistent with both the hypothesis of small numbers and
limited dispersal since there is pronounced variation among populations
(Swofford 1982, Bergstrom et al. 1995, Bergstrom 1997).

Optimum Ozark cavefish habitat occurs in caves with large colonies

of gray bats, Myotis grisescens (Brown and Todd 1987) or comparatively
large sources of allochthonous matter (USFWS 1989). Five of the 14 caves
with A. rosae populations reported by Willis and Brown (1985) contained
bat colonies and another six caves contained guano piles indicating past
use. These caves had the highest food supplies based on isopod counts in
leaf pack traps and copepods and cladocerans in plankton samples. Ozark
cavefish occur in small cave streams where they reside in quiet pools from
a few centimeters to 4 m in depth with chert rubble or silt-sand substrates
(Poulson 1963).
The Alabama cavefish, Speoplatyrhinus poulsoni, is the most troglomorphic
Amblyopsid and its single occurrence (Fig. 3) in Key Cave, Lauderdale
Co., Alabama, is thought to be a relict population of a once more widely
distributed species. It is known from just five pools within Key Cave,
Lauderdale Co., Alabama (Kuhajda and Mayden 2001). The maze-like
cave system is developed in the Mississippian-aged Tuscumbia Limestone
that approximately lies along the ancient Cretaceous shoreline of the
Mississippi Embayment (Cooper and Kuehne 1974). The pools inhabited
by S. poulsoni lie in a zone of seasonal oscillation of the water table where
pools that form during high water become isolated during drier conditions
(Trajano 2001). Many of the pools are extremely deep with depths up to
5 m (Kuhajda and Mayden 2001) and some are situated adjacent to bat
roosts where guano occasionally slides into and likely enriches the aquatic
environment (Cooper and Kuehne 1974). Recently, Kuhajda and Mayden
(2001) reported the capture in 1995 of a single T. subterraneus from Key
Cave. Until this observation, the two species were thought to be allopatric.
Although the effect of T. subterraneus on S. poulsoni is unknown, the
absence of S. poulsoni from adjacent caves where T. subterraneus have been
observed suggests that competitive interactions have and might currently
be influencing the distribution of S. poulsoni. This hypothesis has not been
examined but Poulson (this volume) argues that it is unlikely. Rather he
proposes demographic swamping as an alternative hypothesis.
IV. MORPHOLOGY
A. Family Characteristics
Members of the Amblyopsidae (Fig. 4; Table 1) are characterized by
possessing (i) a large, flat head and a tubular, non-streamlined body, (ii) an
oblique mouth with the lower jaw protruding beyond the upper jaw, (iii)
a segmented premaxilla, (iv) jugular position of the anus and urogenital

Fig. 4 The Amblyopsid cavefishes include an epigean, troglophilic, and four

troglobitic species that exhibit a range of troglomorphy. The fish illustrated are
drawn to scale of average-sized adults. Modified from drawing by John Ellis in
Romero (2004).
pore, (v) reduced head lateral line canals and absence of the trunk lateral
line canal but the presence of superficial papillae (neuromasts) arranged
in distinct rows on the head and body, and papillae of unknown function
in 2–4 rows on the caudal fin, (vi) small, embedded cycloid scales except
on the head, (vii) six branchiostegal rays, (viii) presence of a swim bladder,
(ix) tubular anterior nostrils, (x) absence of pelvic fins (except rudimentary
in A. spelaea). Troglobitic species also can be characterized by (i) lack of
externally visible eyes, (ii) reduced pigmentation, (iii) hypertrophy of
the superficial lateral line system that includes an extensive system of
elevated neuromasts arranged in distinct ridges, (iv) hypertrophy of the
semicircular canals and otoliths, and (v) presence of highly-developed


Table 1 Meristic characters for species in the Amblyopsidae.
Species Dorsal fin Anal fin Pectoral fin Caudal fin Caudal fins Pelvic fin Vertebrae Post- Penultimate
rays rays rays rays rays cleithrum vertebrae bend
Chologaster 11 (9–12) 9 (9–10) 10 (10–11) 10 (9–11) branched absent 27–29 present straight
cornuta
Forbesichthys 10 (9–11) 10 (9–11) 9–10 (9–11) 14–15 (11–17) branched absent 33–35 present upwards
agassizii slightly
Typhlichthys 8–9 (7–10) 8–9 (7–10) 10–11 (9–12) 12–13 (10–15) branched absent 28–29 present upwards
subterraneus slightly
Amblyopsis 10 (9–11) 9–10 (8–11) 10 (9–11) 12–13 (11–13) branched 4 (0–6)1 29–30 absent upwards
spelaea strongly
Amblyopsis 7–8 (7–9) 8 (8–9) 10 (10–12) 10 (9–11) branched absent 28 absent straight
rosae
Speoplatyrhinus 9 (9–10) 8 (8–9) 9 (9–11) 22 (21–22) not absent 27–28 absent
poulsoni branched
1
Frequency distribution of pelvic fin rays in A. spelaea:
0 rays – 1, 1 ray – 1, 2 rays – 2, 3 rays – 18, 4 rays – 58, 5 rays – 7, 6 rays – 1
caudal sensory papillae. Amblyopsids are also characterized by having

dorsal and anal fins that are similar in shape with the dorsal fin origin
anterior to the anal fin origin. All fins lack spines (although the first ray
in the dorsal fin of T. subterraneus has been reported as spinuous). Dorsal
fin rays range 7–12, anal fins 7–11 rays, and pectoral fins 9–12 rays. Pelvic
fins are absent except in A. spelaea that have 0–6 rays. The caudal fin may
be elliptical, lanceolate, or rounded with 9–22 branched rays. No fin rays
are branched in S. poulsoni. The urogenital pore is positioned just anterior
to the anal fin at hatching and migrates anteriad until it occupies a jugular
position in adults (Woods and Inger 1957).
B. Sexual Dimorphism
Most Amblyopsids cannot be sexed using external morphology. Spring
cavefish cannot be sexed using external morphology, although Weise
(1957) noted that in the spring almost all adult individuals could be sexed
by observing the gonads through the translucent body wall. This is rarely
the case for cave Amblyopsids that seem to have low clutch sizes (see
Fecundity).
Male swampfish can be distinguished from females by the presence
of a Y-shaped appendage that develops on the snout. This appendage
is believed to be a contact-stimulatory organ implicated in courtship
behavior (Poulson 1960) and pheromone chemoreception (Jenkins and
Burkhead 1994), although its function is unknown.
C. Neoteny
Poulson believes that there is good evidence for neoteny in the
Amblyopsidae. He predicted the characteristics of his namesake based on
increasing neoteny from Typhlichthys to Amblyopsis. Absence of bifurcate
fin rays in adult S. poulsoni is the most extreme neotenic trait for troglobitic
Amblyopsids (Cooper and Kuehne 1974, Weber 2000) as this condition is
also found in 15-17 mm SL T. subterraneus. Moreover, adult S. poulsoni and
15-20 mm T. subterraneus and A. spelaea are similar in head shape and body
proportions (Fig. 5 and Poulson in Culver 1982). Some of the specialized
morphological and behavioral traits in troglobitic Amblyopsids, such as
longer fins, larger heads, and increased exposure of neuromast organs,
may be explained by neoteny which is one kind of heterochrony (Cooper
and Kuehne 1974, Trajano 2001).

Fig. 5 Scale drawings (Poulson and Cooper) of the body shapes and relative
head sizes of Amblyopsid fishes. With increasing cave adaptation heads become
relatively larger and bodies smaller. Dorsal views from left to right are F. agassizii,
T. subterraneus, A. spelaea, and S. poulsoni. That these are neotenic trends is
suggested by the side views of a 48 mm SL Speoplatyrhinus and a 13 mm SL
Amblyopsis. Can you tell which is which?
D. Eyes
Carl Eigenmann’s classic study (1897) of The Eyes of the Amblyopsidae
remains the best ever. It is a paradigm for the comparative approach to
understanding a process; degeneration of eyes both phyletically, among
the five species known at the time, and ontogenetically, across almost
the known body lengths of each species. Eigenmann’s discussions of
mechanisms also are a paradigm for scientific study. He used evidence
to falsify some but not reject others of six alternative hypotheses
(see Poulson’s this volume).
Eigenmann’s conclusions (1897, p. 587 and Summary Points 1–23), based
on the best-developed eyes for each species, are that phyletically there
is increasing degeneration and simplification of eyes from Chologaster to
Typhlichthys to Amblyopsis spelaea to Troglichthys (=Amblyopsis) rosae. Of the
parts of fish eyes missing or indistinguishable histologically (Table 2) the
numbers are C. cornuta 0, F. agassizii 1, Typhlichthys (Kentucky) 7, A. spelaea
(Indiana) 11, and A. rosae (Missouri) 12. The maximum eye sizes in micra
are 1100 for C. cornuta, 760–930 for F. agassizii, 180 for T. subterraneus, 200
for A. spelaea, and 85 for A. rosae. Figure 6 shows these differences in parts
and in eye size.

Table 2 Summary of parts of the eye of members of the Amblyopsidae. The eyes of cave Amblyopsids are reduced in different
ways but overall the number of parts absent, reduced/vestigial/undefined/merged has increased from T. subterraneus to A. spelaea
to A. rosae. This pattern is consistent with the hypothesis that rudimentation of eyes is an index of evolutionary time isolated in
caves. From Eigenmann 1897. The eye of S. poulsoni has not been histologically examined.
Species C. cornuta F. agassizii T. subterraneus A. spelaea A. rosae
Habitat Swamp Spring Cave Cave Cave

Retina
1. Pigment epithelium (PE) Defined Defined Defined Defined Imperfect
Pigment in PE Present Present Absent Present Reduced
2. Rods and cones Defined Defined Undefined/Absent Undefined Cones absent

3. Outer nuclear layer 1 layer of cells 3 layers of cells Undefined Merged Undefined/Absent
4. Outer plexiform layer Defined Defined Defined Undefined Absent
5. Horizontal cells Defined Defined Undefined Undefined Undefined
6. Inner nuclear layer 1 layer 3 layers Defined Merged Undefined
8. Inner plexiform layer Defined Defined Defined Defined Defined
9. Ganglion layer Defined Defined Defined Funnel-shaped Reduced
Optic nerve connection to brain Yes Yes Yes No No No
Focusing
Lens Present Present Vestigial/Absent Vestigial/Absent Vestigial/Absent
Ciliary muscles Present Absent Absent Absent Absent
Eye muscles Normal Normal Absent Reduced Reduced/Absent
Nutritive
Vitreous body Present Present Vestigial Vestigial Vestigial
Hyaloid membrane Present Present Present Absent Absent
Scleral cartilages Absent Absent Absent Present Present
Pupil Open Open Open Closed Closed
Maximum eye diameter (mm) 1.10 0.76 – 0.93 0.18 0.2 0.09
Eye parts absent 0 1 3 3 7
Eye parts vestigial 0 0 4 8 5
Variability within and between none none some great great
individuals
Fig. 6 Scale drawings (from Eigenmann 1897) of the vestigial eyes of cave
Amblyopsids with the retinas of Chologaster to the same scale (thus complete
eyes would be ~10x as large). The numbers of the retinal layers are identified
on Table 2 of eye parts. “x” is the degenerate lens or its parts. Not shown are the
vestigial eye muscles of Amblyopsis spelaea.

Eigenmann (1897) stressed the fact that eyes of each species “degenerate
along different lines”. For example, Typhlichthys eyes are the least degenerate
overall for a cave Amblyopsid but lack eye muscles, scleral cartilages,
and pigment epithelium. On the other hand, the overall more degenerate
eyes of A. spelaea have some eye muscles, clear scleral cartilages, and well
developed pigment epithelium. Nonetheless, in general, Eigenmann stated
that “the eye of Amblyopsis spelaea represents one of the stages through
which the eye of Troglichthys passed (phyletically)”.
Ontogenetically each species’ eyes develop to a maximum state of
tissue differentiation and then become more simplified and degenerate
with advancing age (Eigenmann 1897). In addition, older fish of the same
size have more variable eye development and there is more right to left
side variation of eyes in the same individual. In A. spelaea, for which the
largest range of sizes was studied (6 to 130 mm), Eigenmann found that
the degeneration started in the earliest stages of eye development and that
the eyes had degenerated completely in the largest fish with one eye even
being absent in a 130 mm individual. These patterns are consistent with
that for cave Astyanax fasciatus with increased variability in the largest
individuals among individuals and in right to left side asymmetry within
individuals.
Poulson (1963, Table 2 in 1985) reported on the sizes and allometric
growth constants (b) of eyes and brain optic lobes for five species and optic
lobe volume in all six species including S. poulsoni. The values reported
here are for the populations of each species with the greatest available size
range since accuracy of allometric constant depends on having a full size
range; the size values are for 45 mm SL fish on the fitted line of part size
vs. body length. The size ranges in mm SL and localities are as follows:
C. cornuta, Roquist Creek, North Carolina, 7–50; F. agassizii Rich Pond,
Kentucky, 15–55; T. subterraneus, Shelta Cave, Alabama, 17-65; A. spelaea,
Letch Cave, Kentucky, 8–90; and A. rosae, Cave Springs Cave, Arkansas,
18–55. There are differences among populations related to both locality
and growth rate that are discussed elsewhere (Poulson 1960 and this
volume).
As for eyes and optic lobe length, there is a decline in optic lobe volume
in cubic mm from 2.12 in C. cornuta to 1.53 in F. agassizii to similar values
for the first three troglobites (1.01, 1.37, and 0.96) to the minimum value of
0.38 for Speoplatyrhinus. The values for allometric growth constant of optic
lobe length are ~0.80 and ~0.60 for the nontroglobitic species to ~0.40 for
Typhlichthys to ~0.60 for A. spelaea to 0.20 for A. rosae. There was too small
a range in body size of S. poulsoni specimens to estimate an allometric
constant but it appears to be the lowest of all.

There is a trend of increasing and then decreasing variance of eye

and optic lobe size from epigean to troglophilic to the most troglobitic
Amblyopsids (Fig. 13 in Poulson 1960, Table XIV in Poulson 1963). Along
with this pattern there is a decrease in size and allometric constant for
eyes and optic lobe length. In 1960, Poulson interpreted these trends
in the context of a balance between natural selection and mutations of
rudimentation. Selection for maintenance of vision or against mutations
of rudimentation is seen in both non-troglobitic species since size and
allometric constants are high but with distinctly greater variability in F.
agassizii. Typhlichthys eyes are about 80 percent smaller with much less
positive allometry and still greater variance. Compared to F. agassizii, its
optic lobe is about 30 percent smaller, variance is the same and allometry
is a little less positive. In A. spelaea, eyes and optic lobes and their
allometric constants are slightly larger than for Typhlichthys despite a more
degenerate eye and an increased variability of eye size in adults. Poulson’s
interpretation for Typhlichthys and A. spelaea was, and still is, of mutations
of rudimentation predominately influencing the patterns since eyes had
already been reduced so much. In A. rosae, eyes and optic lobes and their
allometric constants are all lower than for A. spelaea but variance for both
is as low as for C. cornuta! This decreased variability is clearly related to
a longer evolutionary time in caves but its mechanistic explanation is
elusive. One hypothesis for decrease in visual system variability is that
the visual system is already about as reduced as possible (Fig. 2 in Poulson
1985). But Speoplatyrhinus has even more reduced optic lobes!
It may be relevant to the trends in visual system variability that the
average heterozygosities for 14 allozyme loci also decline with increasing
troglomorphy (Swofford 1982). Poulson (1985) hypothesized that “selection
for some alleles and chance fixation of the most common alleles gave rise
to increasing uniformity.” Swofford (pers. comm.) believes that we cannot
differentiate between stochastic effects and selection with his data.
E. Pigmentation
Other than the rudimentation of eyes, reduction of pigment cells and
pigment are the most commonly observed reductive troglomorphies in
vertebrate troglobites (Langecker 2000). Reduction of melanin can occur
at many steps with many controlled by single genes (Jeffery, this volume).
Size and number of pigmented melanophores is controlled by polygenic
systems in an additive manner (Wilkens 1988). In many cases, the genetic
bases for reduction are not the same in different subterranean fish species.
In the Mexican cavefish, albinism in different populations is caused by
independent mutations in a single gene, Oca2 (Protas et al. 2006). Mutations

that affect the polygenic system result in continuous variation towards

complete depigmentation at slower rates than mutations that result in
the loss of the ability to synthesize melanin (Trajano 2007). Accordingly, it
has been argued that the loss of pigmentation as a proxy for relative age
of subterranean inhabitation should be restricted to related species that
have retained melanin where the extent of depigmentation is caused by
mutations in the polygenic system (Trajano 2007).
Aspects of the developmental, physiological, and genetic mechanisms of
pigment rudimentation have not been investigated in the Amblyopsidae.
Only Poulson (Table 15 in 1960, 1963, Table 2 in 1985, and new data herein)
has quantified the melanophore sizes and numbers and dispersion of
melanin granules within melanophores in the Amblyopsidae. As with
eyes, there is a decline for all of these measures in numbers (densities),
allometry for number, and size through the epigean, troglophilic, and
troglobitic species. All of the data are consistent with increasing durations
of evolutionary time and increases in elaborated traits from Typhlichthys to
A. spelaea to A. rosae to S. poulsoni.
The original observations (Poulson 1960, 1985) were made with a
dissecting microscope and the new data are on excised skin wet mounts
examined with a compound microscope at 500× and 970× (oil immersion).
At 500×, 5–8 melanophores were scored for shape and melanin particle
dispersion in each individual. Dispersion scores are as used in zebra fish
from 1 punctate (maximum aggregation of melanosomes) to 2 amoeboid
to 3 stellate to 4 stellate-reticulate to 5 reticulate (maximum dispersion
of melanosomes). At 970×, for each individual 2–3 melanophores were
scored for intensity of melanin deposition in melanosomes, the integrity of
limiting membranes, presence of a nucleus, and details of the protoplasmic
projections (dendrites or filaments).
With the exception of several individuals in one population of A. rosae
(Roy Pierson Cave, Missouri) and one population of Typhlichthys (River
Cave, Missouri) with no pigmented melanophores visible with a dissecting
microscope (Fig. 7; Table 3), the melanophores in all species appear to have
fully developed melanin granules (melanosomes) with dense deposition
of melanin (electron microscope examination may reveal differences in the
ontogeny of melanophores and melanosomes among the cave species).
There is a trend, with increasing troglomorphy among the species, for
the troglobites to have an increasing proportion of melanophores in the
superficial muscles, body cavity mesenteries, and brain meninges but this
has not been quantified. The other trend common to all three troglobites
is for expanded melanophores (with dispersed melanin granules) to
be less symmetric/stellate and more irregularly elongate. They have
more projections having open ends, and they have less distinct limiting
membranes overall (Fig. 7).

500× 500× 500× 500× 500×
970× 970× 970× 970× 970×
Fig. 7 Drawings of pigmented melanophores. For each of 1–2 localities for each
species of Amblyopsid, extremes of melanophore size (microns) and shape are
shown for three fish examined at 500×. Also, the range of branching patterns
are shown from examination at 970× (oil immersion). See text for methods and
interpretations.
Chologaster cornuta possesses the most melanophores that are also the
largest. Chologaster cornuta is an attractive fish with a striking color pattern
due partly to three kinds of chromatophores (Figs. 1A and 7). Despite being
nocturnal, the pattern is a combination of disruptive and counter-shading
camouflage with the swampfish’s habit of resting during the day among
dense vegetation with dappled sunlight. In live fish (but not in preserved
fish) there appear to be three kinds of chromatophores: xanthophores and,

Table 3 Pigmented melanophores (45mm SL fish). See Fig. 7 for appearance of melanophores at 500× and 970×. See text for
methods and detailed results. The decline in density and size of melanophores is consistent with the hypothesis that the amount
of rudimentation is an index of evolutionary time isolated in caves.

Trait C. cornuta F. agassizii (spring) F. agassizii (cave) T. subterraneus A. spelaea A. rosae S. poulsoni
Number on cheek <=60 35 25 0 0 0 0

b 0.71 0.42 – 0.40 – – – –
Number on body >103 > 100 ~ 50 12 6 0.7 0.6
b +? +? +? 1.5 – 1.1 – 2.4 –?
Micra2 100 83 50 21 16 15 10?
Micron length
Mean 32 65 55 42 31 35 20?
Range 15–50 20–150 15–100 33–200 21–45 8–200 na
Dispersion
Score (1–5) 5 3–5 4–5 1–5 2–3 1–5 4–5
perhaps, two types of melanophores. Swampfish are chocolate brown

dorsally with a sharp border grading to orange-yellowish ventrally and
is reddish on the opercles (xanthophores). Two to three stripes are present
laterally with one at the brown to orange-yellowish border. Only the
melanophores are easily quantified in this species. With full dispersion of
melanin (score of 5), the cheek melanophores average 0.10 mm in diameter
with up to ≥ 60 in a reference area of 0.175 mm2. The allometric constant
of increase in numbers with standard length is 0.71 for a population that
grows up to 50 mm SL. Melanophores on the body are too numerous to
count and determine dispersion pattern accurately. Subjectively, however,
these melanophores all have fully dispersed melanin with a score of 5.
Swampfish body melanophores are dense, dark, and either packed like
paving stones or overlapping, especially along the three black stripes. They
range in estimated maximum dimension from 15 to 65 microns with most
40–50. It remains to be seen how dispersion would change under different
light and background conditions, but the disruptive color pattern and
counter-shading suggests that adapting to light levels and backgrounds
may not be necessary.
Forbesichthys agassizii is dull brown dorsally to slightly lighter ventrally
(Fig. 1B) with only one kind of chromatophore-melanophores. Fish have
poorly defined stripes, one mid-dorsal and two lateral, and slightly more
melanophores along the edges of the myomeres than elsewhere. The range
in size of melanophores in micra among individuals and populations
is 20–150. All melanophores have many projections (10–21 per cell with
most blunt-ended and few thinning at the ends) with melanin granule
dispersion of 3 < 4 < 5. The melanophores on the cheek of the Pine Bluffs
fish (spring-cave N = 9) increase from 35 to 45 per .175 mm2 with growth
and an allometric constant of 0.42 while for Mammoth Cave spring
cavefish (in cave N = 6), cell numbers decline from 25 to 15 per .175 mm2
with an allometric constant of –0.60.
These data suggest that light intensity will affect number, density and
perhaps range of dispersion within a fish ontogenetically. In the spring-
cave populations, melanophores have a score of 3–5 for range of dispersion
from reticulate to fully dispersed melanin granules. Poulson does not
have comparable data for the Mammoth Cave fish but the pale color
suggests that the range would include punctate (score of 1). To extend
this argument, Poulson expects to see that spring-cave populations will
increase melanophore melanin dispersion and density of melanophores if
kept in the light.
Live T. subterraneus from all localities appear white to pearly opalescent
(some guanine?) and in high-resolution photographs outlines of the tiny
embedded scales are visible as slightly grayer color (Fig. 1C). Kentucky

populations have visible blobs of fat around the vestigial eyes. Compared
to Chologaster, all Typhlichthys populations (three from east of the
Mississippi River and two from west of the Mississippi River sample sizes
4–13) have comparable to somewhat higher densities of melanophores but
much smaller surface areas of each pigmented melanophore. Unlike in
Chologaster, it is rare to see melanophores contiguous and there is only
one individual where some melanophores overlapped. In preserved
specimens, the visible melanophores are concentrated along the dorsal
myomere borders and scattered elsewhere on the body. Melanophores are
not visible on the cheek and the number in the reference 0.175 mm2 pectoral
area increase from 2 to 12 with growth with an allometric constant of 1.50.
Melanin dispersion in these melanophores ranges from mostly stellate
(score of 3) to reticulate (score of 4 < 5). Their longest axis (6–130 micra) is
much greater than the narrowest (3–20 micra). There are some pigmented
melanophores on the surface of muscle under the dermis and relatively
more in the internal connective tissues.
For the populations with the greatest size range and sample sizes,
Shelta Cave, Alabama (N = 13, 13–60 mm SL) and The Gulf, Missouri
(N = 13, 21–30 mm SL), there are interesting differences (Fig. 7), which
Poulson hypothesizes are related to the open, partially sunlit, and very
deep sinkhole of The Gulf (the same differences are seen with the Shelta
fish 21–32 mm SL as a control for size). Of the Gulf fish melanophores,
the dispersion scores are one (0), two (0), three (7), four (19) and five (16);
whereas for Shelta fish, the scores are one (20), two (9), three (9), four (4),
and five (5). The Shelta fish are not only more variable between individuals
(e.g., one with no pigmented melanophores, one with all score one and
one with all score 5) but also within individuals (e.g. one with all 1, one
1-3, and one 3-5). Poulson suggests that the greater uniformity and darker
melanophores in The Gulf are due to the low to medium light levels. This
is consistent with substantial increases in pigmentation seen by several
workers in Typhlichthys kept in the light in the laboratory (e.g., Woods and
Inger 1957).
Compared to Typhlichthys, A. spelaea melanophores are much reduced
in numbers, sizes, and ranges of melanin granule dispersion. The data
are mean melanophore area in mm2 × 10–3 from T. subterraneus to A.
spelaea (0.21 to 0.06), range of longest dimension in micra of melanophores
(8–135 with a median of 75 to 23–65 with a median of 36), number per
0.175 mm2 reference area (90 to 6), and range of dispersion (1–5 with a
mean of 4.4 to 2–3 with a mean of 2.4). In addition, for the same dispersion
scores of 3, A. spelaea melanophores have about half as many projections
(4.0) as T. subterraneus (7.0). And unlike for Typhlichthys, neither number
nor darkness of melanophores increases in the light (Eigenmann 1897,

Poulson, personal observations). In fact, the allometric constant for

number of melanophores is negative (–1.10) and some larger fish have no
melanophores (vs. 1.50 for Typhlichthys). All of these data are consistent
with increasing accumulation of rudimentation mutations at the A. spelaea
grade of evolutionary troglomorphy.
Compared to A. spelaea, A. rosae has a somewhat different pattern of
pigment loss but overall, in contrast to a much greater rudimentation
of eyes, shows only somewhat more rudimentation of pigmented
melanophores. Mean melanophore area in mm2 × 10–3 is not as reduced in
A. rosae compared to A. spelaea (0.15 to 0.06) the range of longest dimension
in micra of melanophores is greater (12–200 with a median of 42 to 23–65
with a median of 36), and the range of melanin granule dispersion is greater
(1–5 with a mean 4.1 to 2–3 with a mean of 3.4). However, melanophore
numbers per 0.175 mm2 reference area is less in A. rosae (0.7 < 6.0). There
was no difference in number of melanophore projections for the same
dispersion (5.3 ~= 4.0). As for A. spelaea, A. rosae does not become darker in
the light and the negative allometric constant is still more negative (–2.37)
with some large individuals with no melanophores.
F. Brain Pattern as a Clue to Sensory Systems

In teleost fish, many brain sensory input areas are visible as separate parts
so sensory abilities can be inferred from brain pattern (Evans 1940). All the
following comparisons are for 45 mm SL Amblyopsid fish from double log
plots of size of part versus standard length.
Vision and Reactions to Light

Even the surface-living, nocturnal Amblyopsid with the largest eyes,
C. cornuta, has very small eyes and optic lobes compared to the diurnal
Fundulus notatus topminnow (Figure 11 in Poulson 1960). Both estimated
eye and optic lobe volumes for F. notatus are ~12× that for C. cornuta.
Both C. cornuta (Poulson, personal observations) and F. agassizii (Weise
1957, Poulson, personal observations) hide in vegetation or under objects in
even low levels of ambient light. Even blinded F. agassizii show no change
in swimming patterns or food-finding ability (Payne in Eigenmann 1909).
Spring cavefish certainly uses their eyes to detect light intensity since the
species exits caves into springs after dark and retreats back underground as
dawn approaches (Weise 1957, Hill 1966; Poulson, personal observation).
Amblyopsis spelaea show very weakly developed negative photokinesis
even when vestigial eyes of 15–25 mm TL fish are removed (Payne 1907).

Comparative data on Amblyopsid visual systems can be found in

Table 2 and Fig. 6). Maximum eye diameters in mm are greatest in the
non-troglobitic species and decrease in the troglobitic species. No data are
available for S. poulsoni but much smaller eyes are expected based on very
small optic lobe length. Optic lobe volumes in mm3 also are greatest in the
nontroglobitic species and smallest in the most troglomorphic species, S.
poulsoni. If we could measure optic tectum volume, Poulson expects that
lamellar organization of visual fish will be absent in all the troglobites
even if some optic tracts have been taken over by other sensory systems.
Even further optic tectum rudimentation is predicted with increasing
evolutionary time in caves. See the details of eye structure rudimentation
in the earlier section on eyes.
Response to light has been investigated for several Amblyopsid
species. Poulson (1963) noted both C. cornuta and F. agassizii are negatively
phototactic, i.e., directed movement away from light. Among the other
Amblyopsid species, A. spelaea is reported to be negatively photokinetic,
i.e., lower activity in the dark resulting in more time spent in the dark
end of a gradient (Payne 1907, Eigenmann 1909, Green and Romero 1997),
whereas T. subterraneus is indifferent to light (Eigenmann 1909, Verrier
1929, Green and Romero 1997). Photokinetic response may be linked
to functioning of the pineal organ in Amblyopsids (Green and Romero
1997). The pineal organ is poorly developed and functionally ineffective
in T. subterraneus (McNulty 1978b), which is indifferent to light. Although
the pineal organ of A. spelaea has not been examined, Green and Romero
(1997) predict that the organ is photosensitively functional, although role
of extraocular, extrapineal photoreception cannot be readily dismissed.
Photokinetic responses have not been investigated for S. poulsoni. Likewise,
variation among populations within species has not been examined.
Poulson believes that the weak and varying photokinetic responses
to low light intensities among troglobitic Amblyopsids (e.g., Payne 1907,
Poulson, unpublished data) have no ecological or evolutionary significance.
Part of his reasoning is that A. rosae and T. subterraneus change from
negatively to positively or neutrally photokinetic with increase in size.
This could be because of decrease in transparency of the head tissues above
the brain. Nerves and the brain of many organisms respond electrically to
bright light. Payne removed eyes of small A. spelaea and found no change
in their weak negative photokinesis. Furthermore, it is extremely rare that
any cave Amblyopsid is exposed to a light gradient where water flows
into or out of a cave. In one such case (the caves of Spring Mill State Park
where Eigenmann and Payne worked) A. spelaea avoids the silty areas in
a 150 m stretch before the stream exits a cave but can be seen swimming
in the twilight zone of rocky areas at the entrance to the next cave 50 m

downstream across a karst window (Payne 1907, Poulson, unpublished

data). In another case (The Gulf in Missouri), T. subterraneus preferentially
forages in the light areas along the sides of a steep-sided and deep karst
window, probably because of higher food densities.
Taste
The sense of taste (contact chemoreception) does not seem to be important
for Amblyopsids. Anatomically the sense organs on the head do not look
like taste buds and neither the skin brain centers (a median ‘facial lobe’ of
the medulla) nor mouth brain center (paired ‘vagal lobes’ of the medulla)
can even be identified visually. Smith and Welch (1978) noted that spring
cavefish rely on gustation rather than visual or olfactory cues to determine
prey edibility. Forbesichthys that ingest and then spit out inedible items that
researchers wiggle near them may be using some mouth taste, but there
is sometimes inedible organic matter or bat guano as a small component
of gut contents in troglobitic Amblyopsids. This suggests that they are not
as good as human babies that can selectively sort and spit out peas from a
mush of potatoes or oatmeal (Poulson, personal observations).
Olfaction
The sense of smell (distance chemoreception) is inferred to be better
developed in all Amblyopsids than in Fundulus notatus based both on
external olfactory rosette area and internal brain olfactory lobe length
(Table 4, Fig. 8, Fig. 11 in Poulson 1960, Fig. 4 in Poulson 1963). All
Amblyopsids have a tubular intake and pore output for water that somehow
is drawn across the olfactory rosette. Since every rosette surface cell has
a cilium (Claude Baker, personal communication) Poulson presumes that
the surface area of olfactory rosettes is a good index of olfactory capacity.
Among Amblyopsid species, there is no discernable change with increased
troglomorphy in olfactory lobe size, no pattern for number of lobes of
the olfactory rosette, and only modest increase in estimated olfactory
rosette area (Table 4, Fig. 8, Fig. 11 in Poulson 1960, Fig. 4 in Poulson 1963,
Table 2 in Poulson 1985). This lack of a trend suggests to Poulson that cave
Amblyopsids do not exhibit much enhanced chemoreceptive abilities.
Fernandus Payne (in Eigenmann 1909), Poulson (1960), and Hill (1966)
have done simple behavioral feeding experiments and neither F. agassizii
nor A. spelaea react positively to smells of live or dead prey that are the
main food items in their guts and in their environments. However, both
the troglophile and troglobite react to and even seize moving cotton balls,
sticks or wires. This is not surprising with the overwhelming importance
of lateral line neuromast and tactile senses (see below).

250 ×
400 ×
Fig. 8 Drawings by Poulson of sensory system anatomy of Amblyopsid fishes.

Shown are “tactile” receptors of Amblyopsis spelaea, the olfactory apparatus
of Forbesichthys agassizii, and the difference in exposure of neuromasts in
Forbesichthys agassizii and Chologaster cornuta.
If not used in feeding, what is the function of olfaction in Amblyopsids?

At present this is an unsolved mystery. Experiments on neurophysiological
responses to serial dilutions of amino acids have not been conducted but
lack of behavioral response to live but immobile prey only millimeters
away suggest that olfactory detection of prey is not important. Aumiller
and Noltie (2003) provided evidence that T. subterraneus detects and
is attracted to exudates of both conspecifics and prey (amphipods) in a
laboratory setting but the study did not differentiate between olfaction
and gustation, nor is there any behavioral responses to an injured fish
that suggests the presence of alarm odors (i.e., shreckstoff). Likewise,
sex-related variation in responses has not been addressed. However,
Aumiller and Noltie suggest that cavefish may rely on chemoreception
when locating conspecifics during the reproductive season or detecting

Table 4 Sensory systems of Amblyopsid fishes listed in order of increased (+++) to no (0) evolutionary elaboration with life in

caves. Based on Poulson (1960, 1963, 1985, unpublished data) and Cooper (unpublished data). Standardized to 45 mm SL.
Where there is a trend, species are given rank from least elaborated (1) to most elaborated (6). Cave species are listed in order of
increasing time of isolation in caves (1–4) based on eye and pigment rudimentation.
Species C. cornuta F. agassizii T. subterraneus A. spelaea A. rosae S. poulsoni
Habitat Swamp Spring Cave1 Cave2 Cave3 Cave4
H2O Movement (Lateral Line)
Lateral line neuromast 21.6 (1) 28.5 (2) 45.6 (6) 37 (4) 32.5 (3) 41.9 (5)
num. in rows 2–4
Head surface area (mm2) 443 (2) 329 (1) 666 (3) 615 (3) 683 (3) 666 (3)
Height of cupula + neuromast (mm) 0.3 (1) 0.66 (2) 0.84 (3) 1.14 (6) 0.8 (3) 0.96 (5)
Forebrain ‘computer’ volume (mm3) 0.5 (1) 0.9 (2) 1.15 (3) 1.3 (5) 1.25 (5) 1.15 (3)
Equilibrium (Otoliths & Semicircular Canals)
Saggitus length (mm) 0.69 (1) 0.96 (2) 2.1 (4) 2.05 (4) 2.07 (4) 1.3 (3)
SS Canal complex (mm) 1.65 (1) 2.15 (2) 4.1 (5) 4 (5) 3.48 (4) 3.15 (3)
Cerebellum ‘computer’ volume (mm3) 0.67 (1) 0.7 (1) 1.04 (4) 1.48 (5) 1.49 (5) 0.9 (2)
Caudal Papillae (Unknown Function)
Total number 24? (1) 36? (2) 89 (3) 96 (5) 111 (6) 88 (3)
Relative size 1 1 1.5 1.5 1.7 3.2
Tactile Receptors
Number on head (0.175 mm2) 8 (1) 14 (2) 21 (4) 16 (2) 18 (4) 0 (0)
Relative size 1 1.5 1.5 2.5 2 0
Number of types 1 1 1 2–3 1–2 0
Smell (Distance Chemoreception)
Number of lobes in olfactory rosette 4 5 6 8 5 5
Olfactory rosette surface area (mm2) 0.045 0.05 0.062 0.07 0.071 0.066
Olfactory ‘computer’ lobe length (mm) 0.53 0.55 0.47 0.60 0.62 0.45
potential food sources via exudates dispersed in groundwater flows over

long distances. These hypotheses remain to be critically tested.
Poulson (this volume) argues that critical testing of the importance
of olfaction must involve realistic scales of volume, mixing of currents,
and extreme dilution effects expected in nature. Even in a moderate-sized
artificial stream, Bechler (personal communication) saw only occasional
agonistic responses of ripe males and females of four Amblyopsid species
tested intraspecifically. In addition, Poulson (in Mohr and Poulson 1966)
once spent 24 hours watching six presumed pairs of large A. spelaea in
Upper Twin Cave, Indiana. Even though they were close to each other,
he could detect no change in behavior when they closely approached
each other, seemingly by chance. Poulson has watched newly hatched A.
spelaea freeze when approached by an adult in the field. One hatchling that
moved was eaten immediately and cannibalism is known to occur when
population densities are high (Poulson 1969; Hill 1966). Bechler found
that a ‘freeze’ reaction is a submissive act seen in all Amblyopsid species
that show agonistic behavior (Bechler 1983). All of these observations are
consistent with the importance of lateral line sensing and not olfaction.
Mechanosensory Lateral Line System

Eigenmann (1909) thought that the large number of lines of pores or
ridges on the head and body in all Amblyopsids were tactile organs. In
fact, each ridge or line of pores is one part of a lateral line system that
is most sensitive to water movement. The lateral line system has sense
organs called neuromasts that are part of two systems. The head canal
system has only three pores opening to the surface visible with a dissecting
microscope. With histological sections there appear to be neuromasts
under each of a number of tiny openings all along the subdermal head
canals (Claude Baker, personal communication). The much more extensive
superficial system has large neuromasts that are arranged in lines on the
body surface. In many teleosts, superficial neuromasts are scattered, much
smaller, and much more numerous than in the Amblyopsids (Montgomery
et al. 2001 for Astyanax fasciatus; Claude Baker, personal communication,
for Aphredoderus). In all cases, each neuromast consists of several sensory
cells with cilia embedded in a gelatinous cupula (Fig. 8A and B; Fig. 4 in
Poulson 1963).
Montgomery et al. (2001) provide a short and lucid summary of the
mechanosensory lateral line system anatomy and how it works in teleost
fish. The authors go on to summarize a variety of neuroanatomical,
behavioral, selective pharmacological inhibition, and neurophysiological
evidence including some on Astyanax fasciatus cavefish that suggest
different functions for subdermal canal neuromasts, mostly on the

head, and superficial neuromasts, all over the body. The canal system
neuromasts filter out low frequency noise and are most sensitive to high
frequency water movements generated by localized moving prey. In
contrast, the superficial neuromasts are most sensitive to low frequency
water movement generated by changes in water flow velocity around
the fish (‘hydrodynamic imaging’) as it approaches an obstacle or swims
upstream in a current (rheotaxis). Nonetheless, superficial neuromasts in
some species react to water displacement rather than to velocity and this
seems to be the case in Amblyopsids.
Poulson’s (1960, 1963) data are consistent with the head neuromast
ridges detecting live prey. In fact, the cupulae of anesthetized Amblyopsids
can be seen visibly to move under dark field of a dissecting microscope to
tethered moving prey. For large water fleas (Daphnia magna), F. agassizii
cupulae move at 10 mm distance while for T. subterraneus and A. spelaea
the distance is 20–40 mm. For small amphipods (Hyalella azteca), F.
agassizii cupulae move at 20–30 mm, A. spelaea cupulae at 30-45 mm,
and T. subterraneus cupulae at 40–50 mm. These are a little less than the
distances at which live fish orient toward and then capture the same prey
items. The longer distance for reaction of the troglobites is related to more
elevated neuromast ridges and longer cupulae in T. subterraneus and A.
rosae compared to F. agassizii. In C. cornuta, the neuromasts are in lines but
they are not elevated and the cupulae are very short (Fig. 8 and Fig. 4 in
Poulson 1963).
The system of superficial ridges of neuromasts, each with 10–20
neuromasts, is especially hypertrophied in all the troglobites (Figs. 9 and
10) and so Poulson (1960, 1963, 1985) asserted that it was their primary
sensory system. This is consistent with relative enlargement of the brain
primary sensory nerve projections of lateral line to the cristae cerebelli
on the medulla and eminentia granularis on the sides of the cerebellum.
Higher sensory processing and integration of the lateral line inputs with
other sensory systems is in the forebrain (telencephalon), which is also
hypertrophied in all the troglobites (Fig. 9; Fig. 3 in Poulson 1963; Table 2
in Poulson 1985). The many neuromast ridges on the head, with vertical
and horizontal orientations of adjacent ridges (Fig. 10) must give rich
detail of water movement information for the forebrain to interpret.
In addition to the advantage of a relatively large head having more
elaborate superficial neuromasts, there is a side benefit of better
‘hydrodynamic imaging’. A larger head allows slower swimming to
detect and avoid obstacles. Slower swimming generates less noise for
the neuromasts to detect moving prey. Montgomery et al. (2001) write,
“Although hydrodynamic imaging may not be the sole prerogative of
blind cavefish, these fish appear to have evolved rather sophisticated
mechanisms for processing and using the images.”

Fig. 9 Scale drawings (from Poulson and White 1969) contrasting the troglobitic
Amblyopsis spelaea with its troglophilic relative Forbesichthtys agassizii. Compared
to Amblyopis, each of the lateral line rows in Forbesichthys has fewer neuromasts,
they are less elevated above the skin, and their brain ‘computer’ centers are less
hypertrophied, but its eyes and optic lobes are larger. Also, the presumed ‘tactile’
receptors (shown by the enlarged lateral line rows), are smaller and fewer in kind
than for Amblyopsis. Note also the difference in semicircular canals and otoliths.
However, there is no difference in olfactory lobes of the brain.
Rheotaxis Depends on the Lateral Line System

Except perhaps for C. cornuta, all Amblyopsids show positive rheotaxis
(e.g., Tables 13 and 14 in Poulson 1960). Chologaster cornuta rarely is
found in flowing streams. The difference among species in rheotaxis is of
degree. Allochthonous food renewal is almost always associated with late
winter/spring increases in water input to caves. Amblyopsis spelaea will
swim in the strong currents and lives in cave streams that almost always
have appreciable current. As floods begin and the current becomes too
strong (up to 7 m3 sec–1) they move to the bottom of deep pools and into
backwaters. Typhlichthys in L&N Cave, Kentucky, are exquisitely sensitive

Fig. 10 Scale drawings (courtesy of John E. Cooper) of the heads of Typhlichthys

subterraneus and Speoplatyrhinus poulsoni. Note the relatively longer head and
shovel-nosed shape of S. poulsoni head, which are neotenic trends. Both species
show the convergent similarity in lateral line ridges or “stitches”. Each stitch has
four to as many as 39 neuromasts cups.
to slight increases in current; they seek shelter hours before we can detect
an oncoming flood (Pearson, personal communication). Forbesichthys rely
on swimming upstream to return to caves every day (Weise 1957) or at the
end of the growing season (Hill 1966).
Tactile Receptors and Thigmotaxis

In fish, the brain primary input for tactile information, the somato-sensory
lobe, is hard to distinguish from that for lateral line input since they go
to the same medulla area and share cranial nerves V and VII. A further
complication is that the short neuromasts could function as tactile receptors
when Amblyopsids strike at moving prey or are touched by moving prey
(think about seeing light when you are hit on the eye in the dark). Also,
the small size of what are presumed to be tactile receptors on the head

and around the mouth of both Chologaster and Forbesichthys makes it very
difficult to quantify numbers using a dissecting microscope, especially
with the heavily pigmented epithelium.
Despite the above problems, Poulson has attempted to quantify putative
tactile receptors in all the Amblyopsids. All have what are presumed to be
tactile sense organs concentrated on the heads with relative densities on
the body << side of head < top of head above the mouth << under the
mouth and lips. The incomplete data for number of tactile receptors per
0.175 mm2 are as follows: for the side of the head, T. subterraneus 3–8 =
A. spelaea 4–16 = A. rosae 4–7; for the top of the snout, A. spelaea 1–4 << A.
rosae 11–15; and for under the mouth and lips, A. spelaea 16–20 < A. rosae
25–35. On the side of the head, both F. agassizii and all three troglobites
have similar densities of these presumed tactile receptors but they are
larger in the troglobites and especially larger and more diverse in shape
in A. spelaea (Fig. 8, Fig. 4 in Poulson 1963). S. poulsoni has none of these
receptors anywhere on its head or body.
Despite relatively low numbers of tactile receptors and superficial
neuromasts, both C. cornuta and F. agassizii behaviorally show strong
thigmotaxis and Typhlichthys shows stronger thigmotaxis than either
species of Amblyopsis. In spring cavefish, Weise (1957) found that strong
thigmotaxis overshadows negative phototaxis. Spring cavefish exhibit
strong preferences for cover objects, even if those objects are well
illuminated. When resting, F. agassizii and especially C. cornuta are often
in groups touching each other. In aquaria and in the field, C. cornuta rests
amongst dense vegetation during the day and F. agassizii rests under sticks
and rocks during the day. In an artificial stream F. agassizii spends > 90%
of its time under rocks and Typhlichthys spends much more of its time
under or next to objects resting than either species of Amblyopsis (Bechler,
personal communication).
During foraging and prey capture, lateral line and touch receptors
probably act in a complementary way. Poulson’s (1960) and others’
observations (Eigenmann 1909, Weise 1957) show that Amblyopsids
capture the same live prey by a combination of lunge and grab. They
do not appear to use gape and suck feeding. Troglobitic cavefish initiate
capture at the same distances that cupulae can be seen to move but F.
agassizii appears to contact the prey before attack. Chologaster cornuta will
not capture prey unless it contacts the head or lips; whereas F. agassizii
will turn and grab prey touching any parts of its body especially if it has
previously captured one or more prey. Weise (1957) notes for F. agassizii
that “an amphipod is taken by a vicious sidewise jerking of the head and
is immediately swallowed.” Spring cavefish will take three to four 10–15
mm amphipods in a few seconds. Amblyopsis spelaea also ‘jerks and grabs’
as it captures a large prey item.

Static and Dynamic Equilibrium Reception

As with all vertebrates, teleost fish and Amblyopsids have semicircular
canals and otoliths. The sensory organs associated with each are modified
neuromasts. The semicircular canals detect rates and directions of
movement while the otoliths detect body position even at rest.
The three semicircular canals are all filled with a viscous fluid and are
oriented at right angles to each other; the neuromasts in the ampullae at
the end of each canal detect inertial movement of fluid in the three canals
in each of three dimensions as a fish swims or stops. The three calcium
carbonate otoliths detect position in three planes even while a fish is
not moving. The otoliths rest on a group of neuromasts such that the
null or usual body orientation results in no nerve impulses. In the null
position the otolith presses evenly on all the sensory hairs but any change
in position results in an uneven loading of the otolith on the hairs and
increasing frequency of nerve impulses with increasing departure from
the null position. The saggitta is the largest otolith and it detects deviation
from a horizontal position. The much smaller astericus and lapillus give
information when the fish is oriented in other planes.
Fig. 11 Scale drawings (Poulson) of eyes and optic nerve (where present), brains,
and semicircular canals (SSC) with otoliths of Amblyopsid fishes. Chologaster
cornuta is the small-eyed swampfish with largest eyes and optic lobes (ol) and
smallest forebrain, cerebellum, SSCs and otoliths. Forbesichthys agassizii
has intermediate brain proportions. Amblyopsis rosae shows the convergent
proportions also seen in Typhlichthys subterraneus and Amblyopsis spelaea
but has the smallest eyes and optic lobe and the largest forebrain, cerebellum,
SSCs and otoliths. Speoplatyrhinus poulsoni shows still smaller optic lobes but its
cerebellum and SSCs are not as hypertrophied. Note that it has a different relative
size of the otoliths than for the other Amblyopsids. Also note that there is no trend
in relative development of olfactory lobes at the front of the brain. See Fig. 9 for
the relation of brain proportions to body shape and relative development of the
lateral line system.

Ordinarily, the static equilibrium system, the otoliths, and dynamic

equilibrium system, the semicircular canals, operate in concert with
vision. So it is perhaps not surprising among the Amblyopsids that the
reduced eye Forbesichthys, that cannot see images (Eigenmann 1909), and
blind cavefish have hypertrophied semicircular canals, otoliths, and the
cerebellum brain integrative area for these two systems (Figs. 8 and 11,
Fig. 3 in Poulson 1963). As with the lateral line system and its brain input
and integrative centers, there is little further increase in relative size of
these systems with increased evolutionary time in caves from F. agassizii
to T. subterraneus to A. spelaea to A. rosae to S. poulsoni (Table 4, Table 2 in
Poulson 1985).
It is of interest to discuss why the increase in semicircular canals
and otoliths seen in essentially blind and nocturnal Forbesichthys is not
sufficient to compensate for lack of vision. The answer appears to be that
the relatively larger head in all the troglobites has been selected in part to
increase the lateral line sensory system (above). The larger head in turn
makes it necessary to have further hypertrophy of the semicircular canals
and otoliths. Fine et al.’s (1987) analysis of the largest known semicircular
canals and otoliths in any teleost is instructive. The fish in question is a
deep-sea Ophidiid with an even larger head relative to body size than in
S. poulsoni. It carries to extremes the theoretical need for large canals in
fish generally, because fish have no neck and swim in a viscous medium
(Jones and Spells 1963). With increased relative head size the radius of
curvature and canal diameters must be larger to have the same neuro-
physiological gain in sensitivity. In addition, a larger head results in less
inertial change when turning and so requires larger canals to detect. The
deep sea Ophidiid has a very large head, approximately 10 × the volume
of the body. The relative head to body volumes in Amblyopsids range
from about equal in S. poulsoni to about 10% in C. cornuta. Even C. cornuta
has a relatively large head for teleosts.
A Mystery Sensory System: The Caudal Papillae

All Amblyopsids have a vertical row of papillae at the base of the caudal
plus 2–5 horizontal rows extending out along the fin. These are larger and
more rounded than the largest presumed tactile sense organs elsewhere
on the body. Based on drawings in Woods and Inger (1957), papillae are
slightly larger on T. subterraneus and both Amblyopsis species than on
either C. cornuta or F. agassizii. In addition, total numbers of these papillae
are T. subterraneus 27 <= A. spelaea 30 < A. rosae 36. Based on very accurate
drawings by John Cooper (Fig. 12), the densities, placement, and sizes
of the papillae are strikingly different in T. subterraneus (Alabama) and
S. poulsoni. Typhlichthys papillae are much smaller and more crowded than

Fig. 12 Scale drawing (courtesy of John E. Cooper) of the caudal fins of

Typhlichthys subterraneus and Speoplatyrhinus poulsoni. The unbranched fin rays
of S. poulsoni is a neotenic trait. The especially large caudal papillae (function
unknown) of S. poulsoni are much larger than those of T. subterraneus and all
other cave Amblyopsids.
in S. poulsoni; with shorter total row length (8.5 mm T. subterraneus << 24.8
mm S. poulsoni) and fewer rows (3 vs. 5) over less fin area. Typhlichthys has
70 papillae with 12–17 per mm and S. poulsoni has 88 papillae with 3–5 per
mm. Obviously, we need equally careful measurements and counts on all
the other species to see if there are differences among other species.
V. PREY DETECTION AND AVAILABILITY

All of the available field and gut content evidence is consistent with the
laboratory feeding observations (reviewed above) that show that the
superficial lateral line system is necessary and probably sufficient to
explain the diet of Amblyopsid fish. All items in the gut are certain to
have been alive and moving prey, and predominant prey in the gut match
the distribution and abundance of prey in the field. This is true in terms

of frequency of occurrence, abundance, and volume. If prey are similar in

these regards, then the species that is most active, and thus most easily
detected with the lateral line system, is more frequent and abundant in
fish guts. Thus, the fish are not showing a preference for particular food
items because they take whatever they can detect.
A. Data for Chologaster and Forbesichthys

Gut Contents
The data are excellent, in terms of sample sizes (hundreds to thousands of
fish) and time of year of sampling for C. cornuta (Ross and Rohde 2003) and
F. agassizii (Weise 1957; Hill 1966). The match of prey in the gut and prey
in the field was remarkably similar in C. cornuta from Black Swamp Creek,
North Carolina, and F. agassizii from Pine Hills, Illinois, cave-spring runs.
Both sites have dense vegetation and hiding places so visual predators
of the Amblyopsids are probably unimportant. In both sites Gammarid
amphipods were the most frequent, most numerous, and largest food items
in the Amblyopsid guts and in the field. Of the items eaten the amphipods
are also the most active and so most easily detectable using the superficial
lateral line system. The Rich Pond spring runs did not have amphipods
but did have abundant Chironomid midge larvae and small oligochaetes
with a high diversity of other potential prey. The Forbesichthys guts had
more midges than worms. Hill (1966) attributed this to the greater activity
levels of the midges, which made them more detectable by the fish.
At Rich Pond, Kentucky, there is a huge difference in food supply and
feeding outside the cave and inside (Hill 1966). All surface-collected fish
had very high condition factors and full guts. The cave source of the Rich
Pond spring run had almost no live prey compared to the super-abundance
of prey outside. Guts of fish 35–40 mm SL spending their first winter in
the cave were 99.9% empty and fish had low condition factors. Guts of fish
spending their second winter in the cave were only 18% empty with the
rest having fish (cannibalism) and only slightly poorer condition factors
than when they were outside the spring before.
B. Cave Amblyopsid Gut Contents as Samplers of

Prey Density
For Amblyopsid troglobites the fish are a better sampler than we are of
prey availability in the caves with their very low food supplies (Poulson
and Lavoie 2000, Poulson 2005). Poulson (1960, unpublished data) has
some gut content data for several localities with sample sizes in the 50–
100 range and where we can both easily census fish and crayfish and can

occasionally census amphipods and isopods by examining loose rocks.

Harpactacoid and Cyclopoid copepods are virtually impossible to census
accurately and they are the most frequent and numerous prey in guts of
cavefish.
Here we present some representative data for T. subterraneus from Shelta
Cave, Alabama, A. spelaea from Donaldson Cave, Indiana, and A. rosae from
Cave Springs Cave, Arkansas (a gray bat maternity cave). For four fish
each of 37–43 mm SL the gut contents were as follows with total numbers
of each prey and % frequency (25, 50, 75, or 100%). All the data were taken
at low water times well after the most recent flood. Four T. subterraneus
had 167 copepods (100% frequency), 13 tiny ostracods (25%), 1 amphipod
(25%), and 1 tiny shrimp (25%). Four A. spelaea had 65 copepods (100%
frequency), 7 isopods (75%), and 1 amphipod (25%). Larger A. spelaea had
more isopods, some amphipods, and one fish even had a tiny crayfish.
Four A. rosae had 65 copepods (100% frequency), 8 cladocerans (25%), 4
isopods (50%), and one 10 mm crayfish (25%).
Extensive checks of guts in museum collections for other caves
with lower censused numbers of each Amblyopsid species show that
troglobites NEVER had an empty rectum or intestine and rarely had an
empty stomach. In contrast, six of ten Forbesichthys from Mammoth Cave
had completely empty guts in the year when Poulson censused fifteen fish
that had survived after being washed in from far upstream sinking creeks.
For those four individuals with any food in their guts, the maximum was
6 copepods!
C. Is Forbesichthys Truly Troglophilic?

Hill’s (1966) data on Forbesichthys at Rich Pond reviewed above is consistent
with our current view that spring cavefish are not ordinarily able to
complete their life history in caves because of inadequate food supplies.
Rich Pond has many thousands of Forbesichthys that eat frequently and
grow quickly in the spring run. However, 6–30 mm SL fish in the cave had
empty guts and had poor condition factors. Even 40–60 mm individuals
had only cannibalized fish in their guts. In caves with the same very low
food densities, of copepods >> isopods > amphipods > tiny crayfish or
salamanders, all troglobitic Amblyopsids have food in their guts.
D. Gut Contents in Relation to Prey Densities

An early review of relative live food occurrence in relation to gut contents
for all the Amblyopsids (Poulson 1960) is consistent with our present
generalization that fish eat the most available and fastest moving prey.

The habitats of C. cornuta and F. agassizii have abundant prey with 2–4
orders of magnitude greater numbers and volume than any cave.
Here are some data for surface habitats for C. cornuta and F. agassizii.
A 0.25 ft2 Ekman dredge sample, in Rockfish Creek, North Carolina, in
April had three odonate nymphs, 15 small amphipods, 121 copepods, 52
ostracods, and many rotifiers. In the guts of the 16 C. cornuta collected at
the same time there were 17 taxa but they were predominately amphipods
(60% frequency, 41% numbers, and 80% volume). In a Pine Bluffs, Illinois,
spring run with a cave-spring population of around 150 Forbesichthys,
large (10–15 mm) Gammarus troglophilus have astounding densities year
around of 10–25 per 0.25 ft2. Isopods and flatworms make up the rest of
the macroscopic fauna but neither is found in F. agassizii stomachs. As
Weise (1957) states the fish are monophagous on amphipods.
The best comparative data on food scarcity in caves is from Hill (1966)
since he made weekly censuses of Forbesichthys abundance, sizes, and gut
contents along with Ekman dredge samples in both the cave and surface
spring run at Rich Pond, Kentucky. During the spring months as the water
table rises and the cave waters flow out into the huge spring run, surface
prey became abundant but inside the cave there was almost no food. The
extremes of prey in Ekman dredge samples outside in the spring run vs.
inside the cave were as follows: chironomid larvae 300–1200 versus 0–6,
oligochaetes 600–1200 versus 0–2, nematodes 0–500 versus 0–1, copepods
1–250 versus 0–20, and ostracods 1–120 vs. none. Outside fish guts were
completely full with 70–80% chronomids and 10–20% oligochaetes by
volume; inside the cave fish guts were empty.
Only some caves with Amblyopsid troglobites studied by Poulson
(1960) have more live prey than in the Rich Pond cave. Even the best A.
spelaea cave (Upper Twin Cave in Indiana) in terms of abundance with
86–130 cavefish (Fig. 4 in Poulson 1960) and 45–80 cave crayfish had 1–5
isopods and 0–1 amphipods in ten Ekman samples. However, plankton
samples of 800 liters ranged from 0–300 ml ETOH displacement (Scott
1907). During a spring flood, the 300 ml consisted of 70% silt and 15% leaf
fragments by volume. The remaining 45 ml had an incredible number of
plankters including 1998 Bosmina cladocerans, 410 adult copepods + 319
nauplii, and 717 rotifers. This cave is fed in part by a sinking stream and
both the plankton and the benthos are more than an order of magnitude
more than in the most food rich caves Poulson sampled with other
species of Amblyopsids. We do not have data on food supply for the two
Amblyopsis spelaea caves with the highest densities of cavefish known:
Under the Road Cave had an estimated 4199 fish per hectare in a 300 m
stream segment with virtually no visible organic matter (Table 5, Poulson
1969) and Penitentiary Cave had an estimated 2643 fish per hectare in a

1200 m series of rimstone pools with 10-30 cm of leaf, twig, and acorn
fragments (Pearson, personal communication). Seven caves with the
highest visually censused populations of Typhlichthys had peak numbers
of 4, 6, 11, 16, 20, 23, and 64 fish and for A. rosae, three caves had 7, 11, and
72 fish. The two most abundant Typhlichthys caves had only 2–4 copepods,
0–3 ostracods, and 0–2 rotifers in one 800 liter plankton sample and no
fauna in ten Ekman dredge samples. However, 40–50 cave crayfish were
sampled. The most abundant A. rosae cave had 40 copepods, 6 ostracods,
and 1 isopod in an 800-liter plankton sample and the benthos could not be
sampled among the large rock rubble; however fish gut samples show that
isopods, amphipods, and small crayfish are present.
VI. ECOLOGY, LIFE HISTORY AND BEHAVIOR

In this section, we discuss aspects of Amblyopsid ecology, life history,
and behavior, including such topics such as reproduction, metabolism,
demography, and agonistic behavior. Much of the information assembled
to date is from the works of Poulson, Bechler, and colleagues. However,
many questions remain unanswered.
A. Life History Patterns
Hypotheses and Theory

On the basis of scale marks, size-frequency distributions, egg size and
number, and metabolic rates Poulson (1960, 1963, 1985) inferred that there
is a trend of increasing longevity, decreasing growth rates, and decreasing
metabolic rates from the epigean C. cornuta, to the troglophile F. agassizii,
to the three troglobitic cavefish known at the time. For the troglobites, he
hypothesized that increasing life span, lowered metabolic rates, and large
eggs resulting in large young are adaptations to low food supply in caves.
On the basis of gaps in the size-frequency distribution with no small fish
and some data on gill surface area, he later inferred that his namesake
S. poulsoni has an even longer and more energy efficient life history and
metabolic pattern.
In the context of life history theory (Poulson 2001b and this volume)
the Amblyopsids represent the extremes of very high r to very low r
(instantaneous population growth rate in the logistic equation dN/dt
= rN (K–N/K). At the C. cornuta extreme of high r, there is a 1–2 year
lifespan and one-time reproduction (semelparity) at an age of <= 1 year.
The tradeoffs are of population booms when times are good and busts
when times are bad. Associated with this pattern are spreading of risk

of mortality among numerous small eggs that result in small young.

Amblyopsis spelaea represents the other extreme of low r with a 20–30 year
lifespan (see below) and an age at first reproduction no earlier than 6 years.
The tradeoffs are of population ‘fizz’ when times are good and population
‘fizzle’ when times are bad. Associated with this pattern are spreading of
risk of reproductive failure among many attempts (iteroparity) with few
large eggs and large resulting young at each reproductive attempt.
Iteroparity likely evolved in cave Amblyopsids, and in many other cave
species, because of extreme selective pressures on energy economy at all
life history intervals and because adults are top predators and essentially
immune from predation (Poulson 2001b). In the surface species, C. cornuta
and F. agassizii, death likely occurs after a single reproductive event shortly
after reaching sexual maturity. For both these species, almost all adult
females breed in a given year (Poulson 1963, Ross and Rohde 2003, Hill
1966). Older individuals in their second or third years probably represent
those that failed to reproduce at an earlier age.
For cave species food limitation is the likely selective force limiting
reproduction and reproductive output. In wet years when food availability
is higher, a larger proportion of females likely reproduce. Contrastingly, in
poor years when food availability is severely limited, perhaps all females
fail to reproduce. Iteroparity and increased longevity reduce the risk of
population extirpation when food supplies are too low for an extended
period of time to allow females to reproduce. However, mark-recapture
measures of growth rate (below) show that cave Amblyopsids are much
older than Poulson (1963, 1985) originally suggested based on scale marks
and so the reproductive potential of each species may be considerably less.
This means that conservation of various populations may be more critical
than we currently think (D. Bechler, personal communication).
Tests of Hypotheses: Life History Patterns

In cave Amblyopsids, only a fraction of all sexually mature females
reproduce in a given year. In the three species for which reproductive data
existed, Poulson (1963) estimated that 50% of T. subterraneus, 20% of A. rosae,
and only 10% of A. spelaea females reproduce in a given year. Based on
better data and inferred longer longevities, his revised estimates (Poulson
1985) were 30% T. subterraneus, 5% A. spelaea, and 25% A. rosae. Based on
a thorough review of even more data and the knowledge from growth
rates of even greater longevities, he now estimates the percent of mature
females reproducing in a year averaged over a decade for all Amblyopsids
as 100% C. cornuta, 70% F. agassizii, 20% A. rosae, 10% Typhlichthys, and 5%
A. spelaea. These numbers may be too high for the troglobites if preliminary
data hold up showing sex ratios of all Amblyopsids as almost 2 to 1 for

adults. The revised percents of females reproducing over a decade then

become 13 % A. rosae > 7% T. subterraneus > 3% A. spelaea).
The Ross and Rohde (2003) life history study of C. cornuta is a paradigm
for rigor and completeness. It redresses the odd situation that this most
common and accessible (epigean) species of Amblyopsid was the least
studied and least understood. The life history of this Amblyopsid was
previously only inferred from size-frequency distributions and ovarian
egg counts from small samples from many localities each on usually only
one date (Fig. 4 in Poulson 1960; Ross and Rohde 2003). From collections
at one site on 25 dates with samples of 24–41 fish over one year, Ross and
Rohde could follow the growth rates of young of the year over 65 days in
spring from 8–10 mm SL to 13–18 mm SL for males (~4 mm growth per
month) and to 23–30 mm for females (~8 mm growth per month). By 14
June, the few surviving adults were gone and presumably dead. The first
year fish growth rate slowed in summer and increased in fall. By early
January, males were 19–25 mm and females 29–37 mm SL. We now think
that troglobitic Amblyopsids of these sizes are as old as 3-6 years even
with energetic adaptations to orders of magnitude lower food supplies.
From all of their data, Ross and Rohde (2003) agree with Poulson’s
earlier inference that C. cornuta has a one year life cycle and dies after
reproduction (semelparity). Their data show a sharp 2-month decline of
mean gonosomatic index from 11% to 1% and from 1% to 0% in the next
month (GSI = weight of ovaries as % total body weight less ovary weight).
The initiation of this short spawning season coincided with a sharp
February to March rise in stream temperature from 5 to 15ºC by mid May
when 8–10 mm young of the year were sampled. There were only two of
this size so the smallest young were undoubtedly missed with relatively
coarse meshed seines.
Even Ross and Rohde’s extensive data on ovarian egg number and
diameter are difficult to interpret and compare to other Amblyopsids.
They wisely used only ovaries that were “tight and full of eggs” to estimate
fecundity in 158 females 20–47 mm SL. There was huge variation in egg
diameter from 1.14–1.78 mm (mean 1.48) and in number of eggs from 6–98
(mean 24.7). Together this indicates great variation in both clutch size
and size of smallest free-swimming hatchlings. These data are difficult to
compare to other Amblyopsids that grow much larger (females assayed
for fecundity are 45–65 mm SL). Females from the population Ross and
Rohde studied mostly grew to a maximum of 39 mm with 3 of 156 larger
(45 mm = 33 eggs, 41 mm = 71 eggs, and 47 mm = 98 eggs). From a sample
size of only three females from two other populations where average
maximum SL is in the mid 40s, the fecundities are much higher (42 mm
= 339 and 45 mm = 426 eggs in one population and 44 mm = 252 eggs in

another population) than the fecundities Poulson used for comparison to

other Amblyopsids (Table 6 in Poulson 1960). Also the egg diameters are
higher than the 0.9 –1.2 mm Poulson (1960) reported and used to estimate
reproductive risk (Poulson 1985).
Loren G. Hill’s (1966) Ph.D. dissertation on F. agassizii is in many ways
the most complete study ever done on one population of Amblyopsid.
In his Rich Pond, Kentucky, study site there is a population size of many
thousands of fish and he collected samples of up to 100 on each date
throughout the year. He followed growth, feeding, gut contents, condition
factor, fecundity, and ontogeny of vent migration and scale development
weekly for the growing season. He studied fish both in the cave and in
the spring-run into which the fish moved in late winter, as water started
to flow from the cave. Fish returned to the cave in summer, as the surface
waters dried up, and spawned in the cave the following winter.
Rich Pond has similarities and differences to the Pine Bluffs, Illinois,
spring-cave populations (Weise 1957). Pine Bluff populations are similar
in fecundity, growth rates, and 3–4 year classes (size cohorts and scale
annuli). However, it is different in the much smaller habitat area and
much smaller populations (Weise 1957, Smith and Welch 1978). In eight
Pine Bluffs springs studied by Petersen mark–recapture, the population
estimates were 1–302 with a conservative total of 862 (Smith and Welch
1978). The food available was abundant under rocks and in watercress in
epigean spring runs with 291–902 individuals per square foot of five taxa.
In every monthly sample over a year a large Gammarus species (10–15 mm)
comprised 50–80% of numbers and 90–95 % of volume in the field and >
98% by volume in adult fish guts (Weise 1957). Both Weise and Smith and
Welch agree that most fish exit the underground into the spring a half
hour after dark and return about one hour before dawn (a few remain
under rocks outside during the day). Thus, they are extremely negatively
phototactic and positively thigmotactic. They also agree that spawning
occurs underground since gravid fish disappear from the surface in
January and February and no gravid fish are found on the surface after
that.
Tests of Hypotheses: Growth Rates and Age

Since Poulson’s (1985) latest summary of data and hypotheses based on
scale marks and presumed cohorts, various workers have done mark-
recapture studies to obtain data on growth rates of F. agassizii (Smith and
Welch 1978), A. spelaea (N. Welch, personal communication), T. subterraneus
and A. spelaea (W. Pearson, personal communication), and A. rosae (Brown
and Johnson 2001). All of these studies show that the ages and growth
rates match those inferred from scale marks and distinct size classes in

Forbesichthys (2–3 years) but are underestimates of age by 2–4 × for all
the troglobites. Thus, potential lifespans are 16–24 years for Typhlichthys,
30–45 years for A. spelaea, 8–12 years for A. rosae, and 16–24 years for
Speoplatyrhinus. A large range in age for the same size fish within each
species is due in part, Poulson now believes, to the great variation in
growth rate in fish eating just copepods, fish also eating a few isopods and
amphipods, and fish that show huge growth rate spurts when they are
lucky enough to capture and eat a large prey item (Poulson 2001b).
Tests of Hypotheses: Fecundity, Egg Size and

Cost of Reproduction
Cave Amblyopsids tend to have not only fewer, but also larger and
potentially more nutrient-rich eggs than their surface counterparts. As
in other cave organisms, larger eggs also contain more yolk and produce
larger larvae. By investing more yolk volume per egg, larvae hatch at
larger size and are larger in size when all yolk is absorbed. Therefore,
individuals have greater starvation tolerance, greater swimming ability
when foraging but also when avoiding predation, and can accommodate a
wider range of prey sizes when all yolk reserves are depleted. All of these
are adaptations that increase survival in a food-poor environment.
How well does a more careful evaluation fit Poulson’s (1985) hypotheses
about the advantages of increasing egg size with decreasing clutch size
with cave adaptation? Data on size of fry when their yolk sac are absorbed
is a better measure of sizes of eggs laid than is the sizes of eggs in ovaries
of gravid females. The reason is that Forbesichthys ovarian eggs at Rich
Pond range from 1.5 to 2.0 mm but the sizes of hatchling fry after yolk
sac absorption are all 6.0 mm (Hill 1966) suggesting that laid eggs are
2.0 mm. The only other species where we are certain about size of post
yolk absorption fry is in A. spelaea where we can occasionally find females
brooding 2.3 mm eggs in their branchial cavities and find free-swimming
8.5 mm hatchlings (e.g., Eigenmann 1909, Poulson 1960, Pearson, this
volume). Thus, an increase in egg diameter from 2.0 to 2.3 translates to an
increase in post yolk absorption fry size from 6.0 to 8.5 mm SL. What does
this size difference mean to fry?
From data on fish (Blaxter 1983, Hempel 1965, Hunter 1972) an increase
in egg size of 2.5 mm translates to the following advantages for fry. There
is an estimated 2.5-fold increase in maximum size of prey and increase in
range of food item sizes that can be eaten (mouth size), a 3-fold increase
in volume of water searched per time (swimming speed), a better ability
to escape predators (swimming speed), and a 3-fold decrease in density
of prey needed to maintain weight (resistance to starvation). This is about
the advantage of increased egg size in troglobitic Amblyopsids (from

1.8 mm eggs in C. cornuta to 2.0 mm eggs in F. agassizii to 2.2 mm eggs

in A. rosae to 2.3 mm eggs in T. subterraneus and A. spelaea) that Poulson
(1985) calculated. It does not seem like a big difference but it still seems
evolutionarily significant with the extremely low food supplies in caves.
For an egg-laying female, the evolutionary reduction in estimated clutch
volume from C. cornuta to A. rosae is an important aspect of overall increase
in energy economy. Poulson has seen photos or read about live females
with egg-swollen bodies and easily visible eggs but only for C. cornuta and
F. agassizii (Ross and Rohde 2003, Weise 1957). Poulson now believes that
the estimates of clutch size based on ovarian eggs (i.e., Table 2 in Poulson
1985) are too high for cave species but we have no way of knowing by
how much.
Poulson has recalculated aspects of reproductive cost that he presented
in 1985. The new data (Table 5) are based on average clutch sizes,
maximum ova diameter, and average weight of reproducing females. For
a C. cornuta of 42 mm SL and a weight of 1.0 g with a clutch of 152 eggs of
1.8 mm diameter, the total egg volume is 466 mm3 with 466 mm3 per gram
of female. For a F. agassizii of 60 mm SL and a weight of 3.0 g with a clutch
of 154 eggs of 2.0 mm diameter, the total egg volume is 645 mm3 with 215
mm3 per gram of female. For a Typhlichthys of 50 mm SL and a weight of
1.6 g with a clutch of 58 eggs of 2.3 mm diameter, the total egg volume is
370 mm3 with 231 mm3 per gram of female. For an A. spelaea of 65 mm SL
and a weight of 5.2 g with a clutch of 65 eggs of 2.3 mm diameter, the total
egg volume is 414 mm3 with 80 mm3 per gram of female. Finally, for an
A. rosae of 45 mm SL and a weight of 2.7 g with a clutch of 34 eggs of 2.2
mm diameter, the total egg volume is 190 mm3 with 70 mm3 per gram of
female. In summary, the cost of eggs measured as mm3 per gram of female
declines with increasing cave adaptation with C. cornuta > F. agassizii <= T.
subterraneus >> A. spelaea >= A. rosae.
We get a different perspective if we compare these volumes converted
to calories per day over four months (the time observed for eggs to mature
in vivo in nature) with routine metabolic rate converted to calories per day
(Table 5). The data are calories per day routine metabolic rate + calories
per day to grow eggs (% due to egg growth): C. cornuta 18.4 + 2.9 (16%)
<< F. agassizii 45.0 + 4.0 (9%) >> A. spelaea 18.7 + 2.6 (14%) > T. subterraneus
13.4 + 2.3 (17%) > A. rosae 9.8 + 1.2 (12%). Thus, from this perspective
Poulson’s (1969) statement would seem to be wrong: “The high cost of
reproduction is a crucial problem to be solved in adapting to the cave
environment”. The low absolute values for A. rosae only helps explain
why it has the highest estimated percent of females reproducing each
year (20% compared to 10% in T. subterraneus and 5% in A. spelaea). But
comparing this to C. cornuta (100%) and F. agassizii (70%) it does appear


Table 5 Calculation of reproductive costs and comparison to percent of females reproducing and maximum population density.
See text for methods. The advantage of large egg size for fry size is not offset by a high cost of egg production and total mm3 of
eggs per g of female declines. In addition, the restrictions of low food supply in cave (F. agassizii spring vs. cave) are offset among
the cave species by increasing efficiencies that is reflected by increasing population densities and percent females breeding.
Species C. cornuta F. agassizii F. agassizii T. subterraneus A. spelaea A. rosae S. poulsoni
Habitat Surface Spring Cave Cave Cave Cave Cave

Free-swimming fry (mm SL) 4.5a 6.0 na 8.5a 8.5 8.0a
Maximum egg diameter (mm) 1.8 2.0 na 2.3 2.3 2.2
Egg volume (mm3) 3.05 4.19 4.19 6.38 6.38 5.58
Clutch size 152 154 0 < 58? 65 < 34?
Female size (mm SL) 42 60 40 50 65 45
Egg volume per g female mass 466 254 na 231 80 70
Cal eggs 2.9 4.0 na 2.3 2.6 1.2
RMR Cal 18.4 45.0 na 13.4 18.7 9.8
Estimated % females breeding/yr 100 70 0 10 5 20
Estimated reproductions/lifetime 1 1 0 2–3 3–5 2–3
Lifespan 1–2 2–4 2–4 16–24 30–45 16–24 16–24
Max density hectare–1 42,000 80,000 0–56 864 4199 2134
a
estimated
that the statement is correct. The reason is that caves have such low food
supplies compared to swamp and spring-run habitats.
With the low volume of eggs per gram of female in troglobites, we would
not expect to notice large swollen bodies in females ready to spawn. Also,
there would not seem to be a high energetic risk given the great energetic
economy in all the species. Poulson now predicts a flexibility of clutch size
in all the troglobites based on differences among years in food supply (e.g.,
Pearson this volume for a rare year with ~20% of adult female A. spelaea
with branchial eggs and hatchlings versus usual maximum of 5%) and his
lucky large meal hypothesis (Poulson 2001b).
B. Branchial Brooding and Reproductive Behavior

There is a morphological reason to expect all Amblyopsids to branchially
brood eggs but it has only been seen in A. spelaea. The jugular position
of the genital papilla in adults and the attachment of gill membranes
to the isthmus in all Amblyopsids suggest that all species brood eggs
in their branchial cavities (Woods and Inger 1957, Poulson 1963). Some
have thought A. rosae brood eggs in the branchial cavity (Breder and
Rosen 1966, Balon 1975). It has been suggested that non-incubated eggs
are more susceptible to predation in the cave environment (Noltie and
Wicks 2001); however, direct evidence is lacking. Conversely, evidence
against branchial incubation in the other Amblyopsids exists. At least two
researchers have suggested that total egg volume in ripe Amblyopsids
exceeds branchial volume in Amblyopsids other than A. spelaea. Here we
do some calculations to determine whether we agree.
We start by calculating branchial volume of a 65 mm SL incubating A.
spelaea with 60 eggs each 2.3 mm diameter: 6.38 mm3 × 60 = 383 mm3.
This is about 4 × the branchial volume Poulson (2001a) estimated from
linear estimates for a 45 mm fish. If we proportionally increase the 45 mm
estimate to 65 mm/45 mm the estimate is still only 121 mm3 and about
a third of the observed volume of eggs in the females branchial cavity.
We suggest that this continued discrepancy is due to the likelihood
that 121 mm3 is a large underestimate because brooding females have
obviously puffed up cheeks with a completely distended branchial cavity.
If we apply a proportional correction to Poulson’s estimate of A. rosae
branchial volume for a 45 mm SL female we get 260 mm3 for a branchial
cavity that is completely distended (383 mm3/121 mm3 × 82 mm3 linear
estimate for A. rosae). If this 45 mm fish had a clutch size of 34 eggs each
2.2 mm in diameter (Table 2 in Poulson 1985) the clutch volume would be
190 mm3 or about 73 percent of the estimated distended branchial volume.
On this basis, we disagree with Ginny Adams (personal communication)

that total ovarian egg volume of A. rosae exceeds its distended branchial
volume. In fact, we believe that Amblyopsis clutch volumes are even less
(see reasoning in the previous section). On the other hand, we agree with
Jenkins and Burkhead (1994) that 339 and 426 ova would not fit in the
branchial cavities of female C. cornuta of 42 and 46 mm SL. Our estimate
of distended branchial volume of 42–46 mm SL C. cornuta is about 110–130
mm3 and using a 1.8 mm egg of 3.05 mm3 volume (since we believe that
females only lay eggs at or near maximum size) the clutch volumes are
600 mm3 and 754 mm3. Unlike the case for either Amblyopsis species, these
clutch volumes hugely exceed our newly estimated distended branchial
volume (~1220 mm3 >>>> 130 mm3). Ross and Rohde’s (2003) observations
are consistent with our conclusion that C. cornuta cannot carry its large
clutches of eggs in its branchial cavity. Ross and Rohde were very careful
to hold many newly captured large gravid females in jars but never saw
eggs or fry in or released from a branchial cavity in 13 years of field study.
They speculate that swampfish likely deposit eggs somewhere in the
stream. Swampfish have been observed building a nest-like depression
under a rock in an aquarium (J.S. Smith, personal communication in Ross
and Rohde 2003).
Poulson thinks that the especially low reproductive effort of eggs per
gram for A. spelaea (~80) may be related to the constraint of branchial
brooding that has been verified many times for this species (Eigenmann
1898, Poulson 1960, Pearson, this volume). Clearly there could be no more
eggs than observed in fish with extremely distended branchial cavities.
In this context, it is of interest that Eigenmann reported a few cases of
decrease of eggs in the branchial chamber in three cases where he kept
careful track (57 34, 61 43, and 65 51). This could be due to failure
of development and/or swallowing some eggs. It seems clear that the +
tradeoff of branchial incubation is protection of eggs and new hatched fry
but the potential – tradeoff of not being able to feed while incubating may
not be important. The metabolic rate of A. spelaea is low and fish survive
more than a year without feeding. Whatever the negatives, it is clear that
females could not hold any more eggs in their fully distended branchial
cavity than the reported numbers of 61–70 for 60–75 mm SL females.
The mode of egg-laying and question of parental care both remain
mysteries to be solved in the future. Dave Bechler has, in our opinion,
the most ‘green thumb’ for maintaining troglobitic Amblyopsids for long
periods with high food supplies. But, despite many females developing
some visible eggs, Bechler (personal communication) has never seen
evidence of reproductive behavior much less egg-laying.

C. Population Size and Population Density

Adaptation to low food supply in caves can perhaps be best seen by
comparing the population density per hectare for the best habitats for all
the species with the maximum numbers estimated from 4–12 censuses
over 1–12 years. For estimates we used a multiplier factor of 1.5–2.5 ×
from mark-recapture population size calculations in habitats of different
complexities and numbers of hiding places. The C. cornuta population
studied by Ross and Rohde (2003) was in an area of only 50 m2 and had an
estimated 42,000 fish per hectare (2.5 × maximum number of 84 censused)!
The F. agassizii population studied by Weise (1967) was in only 25 m2 and
had an estimated 80,000 fish per hectare (2.5 × maximum number of 104
censused) while the Mammoth Cave population studied by Poulson was
in a 5400 m2 habitat that had 0 fish per hectare (11 of 12 years) and an
estimated 56 fish per hectare one year (2.0 × maximum number of 15
censused). Given the highest caloric cost for metabolic rate and eggs for
all species (45 and 4.0 cal per day) it is not surprising that Forbesichthys
rarely if ever reproduces in caves.
Compared to the spring-cave Forbesichthys, all the troglobites had higher
densities in caves due to their lower metabolic rate and greater foraging
efficiencies. Typhlichthys in Shelta Cave, Alabama, were in a 7,800 m2 of
prime habitat and had an estimated 229 fish per hectare (2.5 × maximum
of 64 censused) plus as many as 116 crayfish and 30 shrimp per hectare.
Typhlichthys in Blowing Springs Cave, Tennessee were in an 1100 m2 area
and had an estimated 524 fish per hectare (2.5 × maximum of 23 censused).
With much lower metabolic rates both species of Amblyopsis had much
higher estimated population sizes and much higher densities. Amblyopsis
spelaea in Upper Twin Cave, Indiana, were in a 3600 m2 habitat and had
an estimated 417 fish per hectare (2.5 × maximum of 130 censused) and
in Under the Road Cave, Kentucky in only 300 m2 of habitat there were
an incredible estimated 4199 per hectare (1.5 × maximum of 84 censused).
And Pearson and Boston (1995) estimated 2643 A. spelaea per hectare in
3702 m2 of Penitentiary Cave (1.9 × 519 censused). Amblyopsis rosae in
Logan Cave, Arkansas, in 1100 m2 of prime habitat had an estimated
1160 fish per hectare (2.5 × maximum 51 censused) and in Cave Springs
Cave, Arkansas, in 1300 m2 of prime habitat had an estimated 2134 fish
per hectare (2.5 × maximum of 111 censused). In summary, the estimated
maximum density of fish in the 1 - 2 best habitats for each species is as
follows. In food-rich surface habitats there were 42,000 C. cornuta per
hectare and 80,000 F. agassizii per hectare. Forbesichthys in a cave had only
0–56 fish per hectare <<< 524 and 864 for T. subterraneus << 2643 and 4199
for A. spelaea >= 932 and 2134 for A. rosae.

Despite striking energy economies that have allowed high population

densities, there appears to be density-dependent population regulation
for the populations that have been most accurately censused over a period
of years (Poulson 1969). The first evidence is the irregular appearance of
small size classes in all troglobites (Fig. 4 in Poulson 1960). Next, there
are slow growth rates when eating predominately copepods. Finally, the
density-dependent population regulation inferred for A. spelaea, including
cannibalism of 8–10 mm fry, suggests energetic limitation on reproduction
despite low metabolic rates, efficient foraging, and absence of empty guts
at any time.
The data that led to Poulson’s (1969) inference of population regulation
was based on a nine censuses from 1957–1964 of all sizes of A. spelaea in
Upper Twin Cave, Indiana. The numbers fluctuated from 84–130 (1.55 ×)
but the total mass fluctuated less from 363–465 grams (1.28 ×) and the total
metabolic demand (based on routine metabolic rates) fluctuated still less
from 153–180 ml 02 per day (1.17 ×). In addition, the number of incubating
females ranged from 0–3 per census and two instances of new hatched
young. At the time, Poulson suggested that in years with highest metabolic
demand there might be density-dependent cannibalism. An inadvertent
manipulative experiment confirmed this hypothesis. About 30 fish of 50–
70 mm SL were collected from Under the Road Cave (population 84 fish at
the time) for histological and behavioral studies and guts of five fish had
small fish ~ 1–2 years of age that had been cannibalized. Two years later,
Poulson censused 67 fish including nine fry 9–11 mm SL and there were no
cannibalized fish in the guts of the eight large fish collected for anatomical
studies.
D. Timing of Reproduction
Troglobitic species in many systems are subject to pronounced seasonality
in food availability and water levels (flooding). Accordingly, many species
have reproductive cycles synchronized with seasonality. In Amblyopsids,
peaks in reproduction occur just after spring floods (Poulson and Smith
1969). Synchronization during this time is adaptive because offspring
survival is maximized. Young are produced shortly after spring floods
when food availability is still high, yet mortality due to extreme flows
is reduced (Poulson and Smith 1969). However, timing of these cues is
unpredictable and may occur from late fall into spring. Rises in water
level and alkalinity coupled with subtle drops in water temperature may
be triggers to reproduction and synchronization of circannian rhythms
of reproductive readiness (Poulson 1963, Jegla and Poulson 1970).
Synchronization of circannian rhythms of reproductive cycles would allow

cave Amblyopsids to maximize reproductive readiness when chances for

reproductive success also are at a maximum (Poulson and White 1969).
The reproductive cycle in the epigean Amblyopsid is clear and is
probably cued by photoperiod and water temperature. In a detailed year
long study at one site Ross and Rohde (2003) showed that C. cornuta
exhibits a fairly short, spring spawning season that coincides with a fast
increase in water temperature from 5 to 15°C. Ovaries begin to enlarge in
late fall, show peak development in winter, and all females are spent in
April. Throughout the range, reproduction occurs from early March to
late-April (Poulson 1963, de Rageot 1992, Jenkins and Burkhead 1994; Ross
and Rohde 2003) although a few individuals may spawn in late February.
Gravid females with large ova have been collected from 29 January–20
February in the Tar River drainage (Jenkins and Burkhead 1994), and free-
living young as small as 9 mm have been collected in mid-May (Rohde
and Ross in Jenkins and Burkhead 1994).
The seasonality of reproduction in F. agassizii is also clear. In the Illinois
populations studied by Weise (1957) and Smith and Welch (1978), most
adults presumably spawn underground in late winter based on near
disappearance of adults from springs. Ova begin to enlarge in the fall
reaching mature size in January when adults move underground (Weise
1957, Poulson 1963). Subterranean spawning is believed to occur from
January through April and peaks in February when water levels typically
are at their maximum. Fry appear and adults return to the surface by early
May. The evidence for seasonality of reproduction in Rich Pond, Kentucky
populations is even more clear because Hill sampled biweekly year-round
in both the surface and cave habitats. In the cave in winter, both males and
females have developed full gonads. Spent males and immatures leave
the cave as water levels increase and start to spill out to form the seasonal
surface spring run. But spent females and newly hatched young do not
exit the cave until 1.5 months later. From this observation, Hill reasonably
inferred some parental care though fecundities and egg volumes are much
too high for branchial brooding to be possible.
Data on the yearly reproductive cycle for the cave Amblyopsids are
comparatively less detailed with A. spelaea being the best-documented
species. Poulson and Smith (1969) inferred seasonal reproductive cycles for
A. spelaea based on times of year when gravid females, incubating females
with different stages of egg development, and newly hatched and very
small young are observed (unpublished data; Eigenmann 1899a). Thus in A.
spelaea, breeding appears to occur during high water levels from February
through April. Females brood eggs in their gill cavities until hatching and
hold young until yolk reserves are used up. Eigenmann (1909) showed that
this takes 4–5 months. Fry begin to appear in late summer and early fall

(e.g., Fig. 4 in Poulson 1960 for Under the Road Cave). The reproductive
cycle of A. rosae is thought to closely parallel that of A. spelaea but the
evidence is scant. Gravid females of A. rosae have been found from late
August through December (Poulson 1963, Adams and Johnson 2001). On
one occasion three 10 mm SL fish were observed in mid-July (Adams and
Johnson 2001). Except for one other case of an 11 mm SL fish, the smallest
fish ever observed are in the mid 20 mm range. The reproductive cycle of
T. subterraneus appears to be similar but even less evidence is available.
Breeding likely occurs in the spring when water levels are at their highest.
However, no newly hatched fish have ever been seen or collected and the
smallest fish ever observed documented are 15-20 mm (Fig. 4 in Poulson
1960 for Shelta Cave, Alabama) that are possibly nearly a year old. No
data on the reproductive cycle of S. poulsoni exist.
E. Growth and Development

Generally, cave organisms exhibit reduced growth rates and delayed
development and maturity compared to related surface species. Reduced
growth rates represent an adaptive response to low food supplies in cave
environments because less energy over a given amount of time is needed
(Hüppop 2000). Within the Amblyopsidae, growth and developmental
rates decrease with increasing cave adaptation (Poulson 1963).
The fastest growth rates occur in the epigean species. First-year
swampfish from North Carolina grow 9.0 mm SL on average from June to
January and reach a mean size of 32.6 mm SL by 22 months of age (Ross
and Rohde 2003). Slow to no growth occurs during summer, presumably
because of lower feeding rates during this time period. An explanation
for slow or no growth in the summer might be elevated metabolic rates
because of elevated habitat temperature combined with reduced food
supplies.
First year spring cavefish grow 10–20 mm per year on average and also
exhibit variable growth rates from season to season (Smith and Welch 1978,
Hill 1966). Hill (1971) studied squamation and pigmentation development
in F. agassizii. Scale primordia first appear on the caudal peduncle at around
six weeks. By 12 weeks, both squamation and pigmentation pattern are
well developed.
Growth rates for troglobitic Amblyopsids are substantially slower
with estimates of 1.0–1.25 mm month–1 for T. subterraneus, 1.0 mm
month–1 for A. spelaea, and 0.9 mm month–1 for A. rosae based on putative
annual scale marks (Poulson 1963 estimated the number of annuli on
the scales photographs shown in Figure 8 of Woods and Inger (1957) as
4+ T. subterraneus, 3+ to 4 A. rosae, and 4+ to 6 A. spelaea). Using mark-

recapture, Brown (1996) observed growth rates in A. rosae at Logan Cave

lower than those inferred by Poulson from scale marks (1963). Some fish
species do not always show annulus formation because limited feeding
precludes annulus development. Thus, it may be reasonable to assume
that some cave Amblyopsid populations fail to form annuli in a given year
because of low food availability (D. Bechler, personal communication).
Therefore, age and growth estimates based on annuli counts are surely
underestimates.
Excluding 20 individuals that decreased in size, Brown and Johnson
(2001) observed growth rates of 48 A. rosae that were recaptured at least
once. Growth rates declined with increasing size. Individuals 30–39 mm
TL grew, on average, 0.7 mm month–1; 40–49 mm TL, 0.3 mm month–1,
and 50+ mm TL, only 0.06 mm month–1. However, some individuals
experienced much higher growth rates >1 mm month–1. Observed spurts
in growth were not always correlated with increasing food availability,
and are influenced by other factors including age, sex, reproduction, and
other exogenous factors (Brown 1996, Trajano 2001). One explanation for
observed growth spurts is the rare capture of large prey, such as salamander
larvae or crayfish (Poulson 2001b). With a highly efficient metabolism,
these rare feeding events could lead to significant spurts of growth much
higher than average.
Sexual maturity is also delayed in cave Amblyopsids. Both C. cornuta
and F. agassizii reach sexual maturity around 12 months of age. The
development of the Y-shaped snout appendage in male C. cornuta parallels
sexual maturity. Ross and Rohde (2003) documented the development of
this appendage noting that it is not discernable by 16 mm SL, but can
be readily observed by 23–25 mm SL with full development by 25 mm
SL at 13–14 months. On the other hand based on scale mark estimates of
growth rate, T. subterraneus, A. rosae, and A. spelaea may take a minimum
of 2, 3, and 4 years respectively, to reach sexual maturity. Estimates for the
cave species are based on Poulson’s (1963) original examination of growth
annuli, and, therefore, are very conservative. More reasonable estimates
are 4, 6, and 10 years for age at first reproduction of T. subterraneus, A.
rosae, and A. spelaea.
Vent migration is characteristic of the Amblyopsidae and
Aphredoderidae (Poulson 1963). Like other developmental stages, time
from hatching to completion of vent migration is longest in the troglobitic
species (Poulson 1963). Ross and Rohde (2003) have traced migration of
the vent in swampfish. Anterior migration of the vent with increasing
body size is most rapid from 10–17 mm SL and decreases significantly with
very little change ≥ 19 mm SL. Hill (1966) also found that vent migration
to the jugular position was complete in 16-18 mm SL Forbesichthys. And,

Eigenmann (1909) found that the process took longer for A. spelaea. As in
all species the vent is located posterior to the pelvic fins in newly hatched
young. But in A. spelaea 25 mm in length, the vent has only migrated
forward in front of the pelvic fins and by 35 mm, it is still positioned just
posterior to the pectoral fins.
F. Longevity
Increased longevity of cave organisms compared to their surface relatives
is one of several life history adaptations toward a low-r strategy by which
cave organisms cope with limited food resources. Prolonged life spans,
coupled with a trend from semelparity to iteroparity, increases the chance
of population persistence over time, as a population is less likely to be
extirpated during times of extremely low food supplies that result in little
to no recruitment (Hüppop 2000).
This pattern is evident in the Amblyopsidae as longevity inferred from
scale marks increases with increased adaptation to cave environments
(Poulson 1963). The shortest life spans are observed in the epigean
C. cornuta and troglophile F. agassizii. Previously, maximum longevity of
C. cornuta was estimated at 14–15 months (Poulson 1963) with a few two-
year-old fish reported (Rhode and Ross 1986). However, a recent study
indicates that several swampfish lived up to 22 months with one living
26 months (Ross and Rohde 2003). Spring cavefish are known to live up
to three years (Hill 1966, Smith and Welch 1978). It is likely that, as in
many short-lived species, death occurs after a single reproductive attempt
(semelparity). Therefore, older individuals are those that simply did not
acquire enough resources to reproduce at a younger age.
Troglobitic species live considerably longer. Conservatively based on
scale marks, southern cavefish were estimated to live 3–4 years (Poulson
1963); however, individuals have been maintained in captivity for over a
decade and likely live considerably longer than initial estimates in nature
(Noltie and Wicks 2001). Poulson (1963) originally estimated the average
longevity of the more troglomorphic species, A. spelaea and A. rosae, at about
7 and 5 years, respectively. However, Poulson (2001) later questioned his
original longevity estimates of the troglobitic species stating they may be
off by a factor of 3–4, partly because of the difficulty in determining scale
annuli in larger individuals but primarily because of observed growth rates
of marked individuals in nature (see above Growth Rates). Accordingly,
maximum life spans for A. spelaea and A. rosae may actually be as long as
24–28 and 15–19 years, respectively. The maximum life span of S. poulsoni
is unknown, although the largest specimen collected, a 58.3 mm SL female,
was estimated to be as old as 8 years (Cooper and Kuehne 1974, USFWS

1982). This estimate was based on the analysis by Poulson of scale marks in
a 42 mm SL individual based on the same the criteria used to estimate age
in other Amblyopsids. One clearly larger S. poulsoni have been observed
but not measured in Key Cave (J. Cooper, personal communication) and
unquestionably represents an older individual.
G. Metabolic Rates
A trend toward reduction in metabolic rate with increasing cave
adaptation exists within the Amblyopsidae (Poulson 1963, 1985). Poulson
(1963) found that standard, routine, and active metabolic rates decrease
with increasing cave adaptation (Fig. 13). The cause of decreased standard
metabolic rate is a combination of a decrease in gill surface area and
reduction in the volume and rate of ventilation (Poulson and White 1969,
Poulson 2001a). Poulson (2001a) found that the strongest correlates of
metabolic rate reduction in Amblyopsids were reduction in ventilation
frequency and volume > reduction in brain metabolic rate > reduction
Fig. 13 Standard and routine metabolic rates (MR) in Amblyopsid cavefishes in

ml O2 g–1 h–1. Routine metabolic rate for sizes of fish specified are in ml O2 hr–1.

in gill surface area. No single physiological or morphological trait was

in the same rank order as the overall decline in metabolic rate. However,
collectively the reductions were in the same order as metabolic rates. To
Poulson this is consistent with a general pattern of different pathways to
convergent evolution of complex traits. Among Amblyopsid species there
were no differences with cave adaptation in histological indices of thyroid
activity in adults, although subtle increases in thyroid follicle cell height
and colloid vacuolization from the non-growing season (fall/winter)
to the growing season (spring/summer) were detected. This was most
noticeable within A. rosae.
Adams and Johnson (2001) documented a statistically significant
effect of body mass on metabolic rate that differed among seasons for a
population of A. rosae at Logan Cave in Arkansas. A positive relationship
was observed during summer and autumn followed by a negative
relationship during winter and spring. The authors speculated that
seasonal relationships between body mass and metabolic rate may reflect
alterations in environmental conditions, such as food availability and
ambient dissolved oxygen, however other factors, such as biased sex
ratios or seasonal and size susceptibility to handling stress may influence
metabolism. But Poulson wonders whether the effects are biologically
significant given the small sample sizes and no plausible explanations.
In a study on acclimation of metabolic rate to temperature (next section)
the routine metabolic rate of many individuals of each species was
measured at 15°C. These data are more extensive than Poulson’s (1960)
data and include fish from other localities so it is of interest to compare the
results. The data are for C. cornuta (4 of 23–32 mm SL from Cashie Creek,
South Carolina), F. agassizii (10 of 27–32 mm from Pine Bluffs, Illinois
and 14 of 32–68 mm from Rich Pond, Kentucky), T. subterraneus (13 of
29–41 mm from Cave City, Kentucky), A. spelaea (13 of 35–70 mm from Sig
Chatlet Cave, Kentucky), and A. rosae (7 of 36–46 mm from Cave Springs
Cave, Arkansas).
Here are minimum and maximum routine metabolic rates ml O2 g–1 hr–1
at 15°C with the average and the range among individuals (standardized
to a 1 gram fish):
C. cornuta maximum 0.190 (0.142–0.298) and minimum 0.116
(0.095–0.132)
F. agassizii maximum 0.137 (0.039–0.396) and minimum 0.087
(0.029–0.092)
T. subterraneus maximum 0.082 (0.026–0.181) and minimum 0.061
(0.015–0.136)
A. spelaea maximum 0.037 (0.015–0.093) and minimum 0.023
(0.010–0.023)

A. rosae maximum 0.031 (0.012–0.070) and minimum 0.026

(0.008–0.070)
This decline in routine metabolic rate C. cornuta > F. agassizii >> T.
subterraneus > A. spelaea = A. rosae is really close to that reported by Poulson
in (1985).
H. Tolerance of Temperatures from 5 to 25°C

In these studies of acclimation, data were taken in a torus respirometer.
An oxygen electrode was used to measure metabolic rate and an infrared
detector was used to measure spontaneous swimming activity.
The results are that metabolic rate and spontaneous activity of all
Amblyopsid species acclimate to a wide range of temperatures. This is
yet another bit of evidence that belies the constant cave paradigm. In
retrospect, it is not surprising that even troglobitic Amblyopsids acclimate
since they must deal with floods that bring in prey or allochthonous
organic matter that their troglobitic prey need. At the end of winter, rain
and snow melt result in cold water entering the cave. In more rare summer
floods, the water entering the cave is warm. The Pine Bluffs populations
of F. agassizii are in springs that only vary from 11–15°C but they acclimate
to temperatures of 5, 10, 15, 20, and 25°C the same way as fish from Rich
Pond, Kentucky that experience water temperatures of 7–26°C.
The data (Poulson, previously unpublished) show that there is no
difference in degree of acclimation among the species and each species
shows acclimated metabolic rates and spontaneous activity that are
essentially the same from 5 to 25°C. Thus, each species shows almost
perfect compensation for temperature. Put another way, metabolic rates
and activity are the same in fish acclimated to 5, 10, 15, 20, and 25°C (no
significant differences between rates at any temperature for any species).
The process of acclimation takes 1–3 weeks depending on the temperature
difference. Initially, fish taken form 15 to 5°C show a reduced metabolic
rate and activity and fish taken from 15 to 25°C show an elevated metabolic
rate and activity.
I. Starvation Resistance
Poulson (1961, 1963) crudely compared starvation resistance in
Amblyopsids and found increasing ability to cope with food scarcity from
the epigean C. cornuta to the most troglobitic species, A. rosae. Chologaster
cornuta showed the greatest loss in body mass and lowest starvation
resistance, expiring after only 45 days (Table 6). On the other hand, A.
rosae lost only 9% body mass after 20 days and survived over 13 months

Table 6 Effects of starvation on Amblyopsid cavefishes under aphotic conditions

(from Poulson 1960). Decreasing standard metabolic rate (SMR) reflects decreased
rate of weight loss during starvation and decreased routine metabolic rate (RMR)
reflects increased time until spontaneous activity stops during starvation. Data
standardized to 45 mm SL fish. Metabolic rates in ml O2 g–1 hr–1.
Species Habitat Body mass1 SMR Activity2 Minimum – Maximum RMR
C. cornuta Swamp 34 0.065 46 0.085–0.298

F. agassizii Spring 24 0.028 116 0.029–0.396
T. subterraneus Cave 15 0.016 230 0.015–0.181
A. spelaea Cave 13 0.018 280 0.006–0.093
A. rosae Cave 9 0.011 400 0.006–0.070
1
% loss of body mass after 20 days
2
End of spontaneous activity in days
without food. Amblyopsis spelaea can survive up to 20 months without food

(Sloan in Eigenmann 1909).
Starvation resistance in cavefish can be achieved by four primary
means, acting alone or in concert (Hüppop 2000): 1) consumption of large
amounts or a single large prey item at a single feeding event, 2) reduction
of activity and metabolic rate, 3) storage of large amounts of fat, and 4)
reduction of the digestive absorptive epithelium or reduction of gut length.
The first three mechanisms are evident in Amblyopsids. Large prey items
occasionally have been observed in the stomach contents of collected
fish. Second, metabolic rate decreases within increasing cave adaptation
(see Metabolic Rates above and Poulson 1960, 1985, 2001). However, in
Poulson’s studies, activity of some of the troglobites actually increased over
the first six weeks before decreasing later. The observed initial increase in
activity has been attributed to searching for food. Lastly, although fat was
not directly measured, Poulson (1960) noted that the troglobitic species
possessed larger fat stores compared to C. cornuta and F. agassizii. Thus,
fat reserves (though much lower than in Prietella and Astyanax cavefish),
coupled with low metabolic rates explain the increased survival time of
troglobitic Amblyopsids under starvation (Hüppop 2000).
However, more detailed studies, like those on other cavefishes
(especially by Hüppop 2000), are needed to assess the ability to recover
weight after varying amounts of weight loss and to store large amounts of
fat quickly when food is available. In addition, it is important to compare
the mechanisms of resistance to starvation seen in epigean fish to determine
the degree of preadaptation. This has been done for Astyanax fasciatus
(Hüppop 2000) but not for Amblyopsids. For example, all vertebrates,
including humans, when deprived of food first use glycogen, then fat, and
as a last resort protein resulting in muscle wasting. Associated with these
changes is a reduction in activity and basal metabolic rate.

J. Circadian and Circannian Rhythms

Little information exists about endogenous cycles in cave organisms
(Langecker 2000) but some evidence is available for cave Amblyopsids.
Some troglobitic species have retained circadian oxygen consumption
but have lost circadian activity, as their rhythms cannot be entrained
to light-dark cycles (Poulson and Jegla 1969, Poulson and White 1969).
This has been viewed as evidence that entrainment of activity cycles to
the environment is not maintained by selection in subterranean habitats
(Poulson and White 1969) as has been observed in F. agassizii (Poulson and
Jegla 1969). With regard to endogenous annual (circannian) rhythms, both
A. spelaea and A. rosae exhibit a well-defined yearly reproductive cycle
(see above ‘Timing of Reproduction’). Because caves do not have reliable
seasonal cues (Poulson and Smith 1969), circannian rhythms are adaptive
to maximize reproductive readiness when chances for reproductive
success also are at a maximum at unpredictable times of the year (Poulson
and White 1969).
K. Agonistic Behavior
Agonistic behavior of Amblyopsids has been investigated in detail by
Bechler (1980, 1981, 1983). He examined intraspecific dyadic interactions
(1 resident and 1 intruder) in 80 liter or 160 liter aquaria with a rock hiding
place in each. Only C. cornuta showed no agonistic behavior and Poulson
suggests that this is because no resources are defendable in its epigean
plant and debiris-choked habitat with extremely high food densities.
Among the four other species Bechler documented two submissive acts,
“freeze” and “escape”, in all species. This is the only study to document
“freezing” in hypogean fishes (Parzefall 2000). Poulson has observed
freezing by recently hatched A. spelaea under threat of cannibalism. Large
Forbesichthys also cannibalize smaller fish (Hill 1966) and freezing should
be an excellent defense since Amblyopsids use only lateral line to detect
other fish and prey. Thus it is not surprising that fish that perceived that
they are losing in an agonistic encounter ‘froze’ more often than escaping
by fleeing. This allowed them to avoid the most intense kinds of acts.
In order of increasing intensity and decreasing frequency Bechler
documented tail beat > chase > attack = head butt > bite and jaw lock
(Table 7). The only metric that was in the order of increasing time isolated
in caves was total number of agonistic bouts with F. agassizii 158 >> T.
subterraneus 59 = A. spelaea 61 >> A. rosae 21. Mean duration of bouts had
T. subterraneus and A. spelaea reversed in rank order with F. agassizii 53
sec = A. spelaea 57 sec >> T. subterraneus 26 sec > A. rosae 18 sec. Of the six

Table 7 Agonistic behavior in Amblyopsid fishes. From Bechler (1983). In paired

interaction tests (1 resident and 1 intruder) C. cornuta showed no agonistic
behavior. Among the one troglophile and three troglobites, three species showed
five to six of six kinds of acts (F. agassizii = A. spelaea <= T. subterraneus) but
the most cave-adapted species, A. rosae, showed only one of six. See text for
interpretation.
Habitat Surface Spring/Cave Cave Cave Cave

Aggressive acts
Tail-beat – +++ +++ ++ ++
Head-butt – + + ++ –
Attack – +++ + + –
Bite – + + ++ –
Chase – + ++ ++ –
Jaw-lock – +(2) – +(1) –
Submissive acts
Freeze – +++ +++ +++ +++
Escape – ++ ++ ++ ++
Total no. acts 0 158 59 61 21
Mean duration of acts (sec) 0 53 26 57 18
– not observed; + rare; ++ regular; +++ frequent; ++++ always
possible aggressive acts there was less of a trend among species with F.
agassizii, T. subterraneus, and A. spelaea showing 5–6 acts and A. rosae with
showing only tail-beating (Table 7).
One clear difference among species is that F. agassizii and T. subterraneus
always initiated agonistic acts from under or next to rock shelter whereas
A. spelaea and A. rosae patrolled the whole tank and did not set up
“stations”. This is consonant with the high importance of thigmotaxis to
F. agassizii and T. subterraneus (see Thigmotaxis above). Bechler (personal
communication) noted the same prevalence of thigmotaxis with long
periods of rest in F. agassizii and T. subterraneus in a 6.3 m long 4.1 m2
artificial stream with many rocks. And he noted that both A. spelaea and A.
rosae were active almost all the time and did not set up “stations”. To make
these observations, Bechler recorded the behavior of four individuals of
each species periodically by videotape for 30 days. The four individuals
of each species “encountered each other” (within 15 cm) much more
infrequently and showed many fewer agonistic interactions per encounter
than did the two individuals in aquaria of orders of magnitude smaller
volumes. In fact, in aquaria pairs of individuals always were agonistic
when encountering each other whereas in the stream the percentages of
encounters with any agonistic behavior were 55% T. subterraneus > 27% A.
spelaea = 26% A. rosae >= 21% F. agassizii. In addition, the bout lengths were

shorter and the agonistic acts less intense than in aquaria. The density
of four fish in 4.1 m2 of artificial stream translates to 9,662 per hectare,
which is twice the highest we have ever estimated in nature for any cave
Amblyopsid. So, it seems to us that in caves encounters would be even less
frequent. In fact, neither we or Bechler (personal communication) have
ever seen agonistic interactions in the field with hundreds of observations
of fish within 15 cm of one another.
Bechler (1983) viewed degenerative evolution as the most likely
mechanism for reduction of agonistic behavior in Amblyopsids suggesting
conservation of energy due to other adaptations to subterranean life.
Thus reduction in metabolic rate and fecundity with increased swimming
efficiency reduced selective pressures caused by limited food availability
and allowed for reduction in agonistic behavior. He suggested that
retention of some levels of agonistic behavior in A. spelaea might reflect
lower levels of subterranean adaptation or the possibility of defense of
prime foraging areas in riffle and pool cave streams. Agonistic behavior
has not been observed in A. spelaea in nature and Poulson thinks that it is
unlikely.
Poulson (this volume) argues that there seems to be no defendable
resource in caves and so we should not expect territoriality and associated
agonistic behavior to be important. This suggestion is consistent with
the lack of any agonistic behavior in C. cornuta where there is clearly no
defendable resource or food in the dense vegetation in which it lives.
L. Territoriality and Social Groups

Although epigean species are often territorial or form social groups,
troglobitic species generally are found in low population densities and
usually solitary with a large home range (Langecker 2000). In general,
populations of A. spelaea and T. subterraneus are low in density (but see
Poulson 1969 and above TESTS OF HYPOTHESES: FECUNDITY AND
EGG SIZE) and individuals are irregularly distributed over suitable
habitat (Poulson 1963). Individual fish have large home ranges, cover long
distances in search of food, and never defend areas (Poulson 1963, Mohr
and Poulson 1966). However, during aggression trials by Bechler (1983),
F. agassizii and T. subterraneus established distinct territories in aquaria
with rocks (“stations”), whereas the less sedentary species, A. spelaea
and A. rosae, patrolled the entire tank. Cavefish do not display schooling
behavior, although individuals are sometimes observed in close proximity
typically around food sources (e.g., underneath a bat roost). Even in caves
with the largest populations, 70-80% of fish are found in perhaps only
10-20% of the accessible cave stream (Brown and Johnson 2001, Poulson,
unpublished data, Pearson, personal communication).

M. Food Habits
The reported diet of Amblyopsids has been well characterized and was
treated in the context of prey detection earlier in this chapter (V. Prey
Detection and Availability). All Amblyopsids eat live, moving prey with
invertebrates comprising most of the diet. However, cannibalism has been
documented in the family and small amounts of nonliving food such as
bat guano and detritus have been observed in the stomach contents of
some species. These are probably ingested along with live prey but would
provide very much lower nutritional benefit per volume than live prey.
Swampfish are likely crepuscular or night feeders (Poulson 1963)
principally feeding on amphipods, chironomids, and cladocerans with
amphipods comprising the majority of diet (G.B. Pardue and M.T. Huish
in Jenkins and Burkhead 1994, Ross and Rohde 2003). Other potential
prey items include nematodes, ostracods, ephemeropterans, odonates,
plecopterans, megalopterans, coleopterans, trichopterans, and other
dipterans. Fifty-three percent of 289 stomachs of day-collected fish
examined by Ross and Rohde (2003) were empty from one North Carolina
site; however, several fish had noticeable food in the intestines. Prey
diversity and percent occurrence and abundance of each organism in the
diet is greatest from January–April then substantially decreases in summer
and fall (Ross and Rohde 2003). Cannibalism has not been reported in C.
cornuta.
The diet of F. agassizii varies geographically and between cave and
surface populations. Illinois populations feed almost exclusively on
Gammarus amphipods (Forbes and Richardson 1908, Layne and Thompson
1952, Gunning and Lewis 1955, Weise 1957), although amphipods (Weise
1957), insect remnants (Gunning and Lewis 1955), and detritus (Gunning
and Lewis 1955) also have been found in stomachs. In Kentucky, surface
fish feed principally on chironomids but also copepods, oligochaetes,
nematodes, and ostracods (Hill 1969a). On the contrary, individuals of
the same population are strongly cannibalistic on younger individuals
when in the cave part of the habitat. In caves cannibalism may represent
an alternative feeding strategy in response to competition for more typical
but extremely rare invertebrate food sources. In rare years when F. agassizii
occurs in the Mammoth Cave food supply is much greater than usual
(Poulson, unpublished data).
The diets of the troglobitic Amblyopsids consists largely of copepods
and isopods but rare, larger meals, such as crayfish, salamander larvae,
or conspecifics, may result in high growth efficiency and a burst in
growth rate (Poulson 2001). An assortment of prey have been reported
in stomach contents of T. subterraneus including copepods, amphipods,

isopods, trichopteran and tendepedid larvae, cladocerans, isopods, and

crayfish (Poulson 1960, 1963, Cooper and Beiter 1972); however, copepods
are the primary food source accounting for 60–90% of the diet by volume
(Poulson 1963). The diet of A. spelaea includes copepods, amphipods,
isopods, and salamander larvae (Clay 1975). Amblyopsis spelaea smaller
than 45 mm total length chiefly eat copepods, whereas those in excess of
45 mm may consume amphipods, isopods, and crayfish (Poulson 1963).
Larger fish also occasionally cannibalize smaller fish. Few studies have
investigated the food habits of A. rosae, but gut contents show that the
diet of Ozark cavefish comprises the majority of available food sources
in the caves they inhabit. Copepods constitute up to 90% of the diet by
volume, but isopods, amphipods, cladocerans, ostracods, small crayfish,
coleopteran larvae, plecopterans, crickets, and small salamanders also are
taken (Poulson 1960, 1963). Bat guano has also been found in stomach
contents. It has been suggested that younger individuals are cannibalized.
However, Adams and Johnson (2001) observed a presumed parental-
young aggregation of A. rosae for more than six weeks with no evidence of
cannibalism. It remains unknown how long such parent-kin interactions
occur and whether kin recognition exists. Based on stomach contents,
extent of fat deposits, and annulus formation, A. rosae fed less in the fall
than other times of the year (Poulson 1960). The diet of S. poulsoni has
not been examined but likely consists of amphipods, isopods, and small
crayfish (Boschung and Mayden 2004).
N. Predators
Troglobitic Amblyopsids are at the top of the food chain in most cave
systems that they inhabit, and, therefore, have few natural predators.
Epigean fishes may potentially prey on cave Amblyopsids (Brown
1996, Poly and Boucher 1996), as well as crayfish (Romero 1998a), and
troglobitic salamanders. Young cavefish are susceptible to cannibalism by
larger adults (Poulson 1963), as has been documented in A. spelaea and
one cave population of F. agassizii. Cannibalism may serve as one means
to regulate population densities in a food-limited environment (Poulson
1969). Epigean populations of F. agassizii likely are occasional prey for other
fish, snakes, birds, and mammals (Smith and Welch 1978). At Rich Pond in
Kentucky, natural predation is seasonally heavy (J.E. Cooper in Smith and
Welch 1978). Epigean predators, such as raccoons, fish, and water snakes,
also may be potential threats to cave-inhabiting Amblyopsids; however,
no cases of predation on cave-dwelling Amblyopsids by epigean predators
have been documented even in cave entrance areas.

O. Parasites and Diseases

Like many cave vertebrates, few parasites have been observed afflicting
Amblyopsid hosts. Proteocephalan cestodes have been collected from the
pyloric caeca of A. spelaea and F. agassizii. Three Proteocephalus poulsoni
were collected from two northern cavefish from Under the Road Cave in
Breckinridge Co., Kentucky (Whittaker and Zober 1978). Whittaker and Hill
(1968) described P. chologasteri from spring cavefish. In southern Illinois, 71
percent of fish examined were parasitized by cestodes and other internal
parasites (G. Garoian in Smith and Welch 1978). The acanthocephalan
Neoechinorhynchus cylindratus has been reported from the intestines of A.
spelaea (Nickol and Whittaker 1978) but neither frequency nor incidence
was reported. In cursory examination of many Amblyopsid guts for food
item analyses, Poulson has never seen an obvious parasite.
Reports of external parasites are even fewer. A species of copepod,
Cauloxenus stygius, is an obligate ectoparasite found on the upper lip of
northern cavefish. Little is known about the species. Its distribution is
believed to be that of its host but the parasite has rarely been observed
and few records exist (Cope 1872, Blatchley 1897, Giovannoli 1933,
Lewis 2002ab, Poulson, unpublished data). In other Amblyopsids, small,
unidentified leeches also have been reported for F. agassizii in Illinois (Smith
and Welch 1978). Additional studies using a comparative framework are
needed to not only document additional parasitic taxa, but to also compare
frequency and density of occurrence between the epigean, troglophilic,
and troglobitic Amblyopsids.
Few diseased Amblyopsid cavefishes have been reported in nature.
Fournie and Overstreet (1985) reported on an adult F. agassizii from Union
Co., Illinois, with a retinoblastoma on the right side of the head. This
condition may be related to chromosomal abnormalities. At least one other
individual collected at the same spring had a similar tumor in appearance
and eventually died after the tumor involved the entire head. However,
this specimen was not available for histological examination (Bechler,
pers. comm. in Fournie and Overstreet 1985).
In troglobitic Amblyopsids, Pearson and Boston (1995) documented a
bacterial infection of the fins of A. spelaea at Donaldson Cave, Lawrence
Co., Indiana. All individuals observed had extremely shortened pectoral
and caudal fins with ragged margins. Many fish also had small, red spots
scattered over the body. During a recent trip to Donaldson Cave, this
condition was not observed (Niemiller, unpublished data). Broken-back
syndrome was noted for only three years in perhaps 10% of fish from
this site. The likely cause was exposure to pesticide and other chemical
contamination since the surface watershed was in row crops and pasture

(Keith and Gray 1979, Keith and Poulson 1981). Gas bubble disease has
been documented in recently collected T. subterraneus at a spring site in
Missouri (Schubert et al. 1993). Southern cavefish from Missouri may be
particularly susceptible to this disease because of the depths at which
individuals reside (Schubert et al. 1993, Noltie and Wicks 2001).
There is some logic to believe that caves may be a refuge from many
kinds of diseases or parasites so Poulson (this volume) has suggested that
we consider this hypothesis carefully and at least keep good records. A
corollary is that absence of parasitism and low incidence of disease may
have led to evolutionary rudimentation of immune defenses. However,
the recovery of many individual A. spelaea from fin rot in one cave
(W. Pearson, personal communication) may be evidence of perfectly
adequate immune responses. It will be worthwhile to compare the
Amblyopsids for incidence of and recovery from common aquarium fish
diseases.
P. Abundance and Population Sizes

Few studies have attempted to quantify population sizes and relative
abundance of Amblyopsids, including C. cornuta and F. agassizii. The few
studies that have attempted to quantify population sizes via techniques
such as mark-recapture or survey removal have focused on caves that
are known to contain relatively large populations. Other studies for
which the most reliable estimates of abundance have been obtained have
focused on the species of conservation concern. Additional demographic
studies, including long-term censuses, are needed for both epigean and
subterranean populations.
In general, the majority of cave Amblyopsids localities yield few fish
sightings during single surveys. Although this may be a real reflection
of actual abundance in some instances, it is important to realize that the
distribution and abundance of these troglobitic species, perhaps with the
exception of S. poulsoni, likely is greater than currently realized. Localities
for which cavefish have been reported represent but a fraction of total
available habitat accessible to fish. This fact was clearly illustrated during
the Maramec fertilizer pipeline break that resulted in the death of nearly
1,000 southern cavefish and likely many more (see discussion in Noltie
and Wicks 2001) from a drainage basin with few records documented
previously. The problem with inferring population densities from such
fish kill is that we do not know the volume or extent of habitat impacted.
Most observations of cavefish are restricted to caves near the surface
and there is some controversy as to whether even the best cavefish caves
are sources or sinks (see Poulson, this volume). Poulson believes that
populations that we can sample make up the majority of the total number

of fish in a cave watershed. Part of his logic is that cavefish will come to
be found in the highest available food areas and these are likely to be in
shallow caves with allochthonous inputs and a combination of deep pools
and shoals. Further we usually find most fish in the highest food supply
areas of these caves (see above ‘Population Size and Population Density’).
He also argues that only few fish will be in deep phreatic habitats with little
food input or the most upstream parts of caves inaccessible to humans. He
bases this second inference on both Cave Springs Cave, Arkansas, with
about 100 A. rosae seen on each visit, and Upper Twin Cave, Indiana, with
84–130 A. spelaea seen on each visit. As one goes far upstream in both these
caves the number of cavefish drops to none as the water gets shallower
and faster flowing with no refuges during floods.
Consequently, habitats where we see no fish or only a few at each visit
may be population sinks and not sources. Wells and short stream segments
encountered in an otherwise dry cave may not be representative of the
habitat that most troglobitic Amblyopsids inhabit. To be sure cavefish can
disperse through and occupy submerged passages inaccessible to humans
but these habitats are probably neither usual for the fish nor optimal. This
does not mean, however, that these fish are doomed. They may at least
be potential dispersers. They could move long distances given their long
lives, low metabolic rates, and foraging efficiencies. In those caves where
we always see 6–10 fish they may even be slowly reproducing. If so, these
sinks could become a source for re-colonization if some disaster befalls the
fish in the best caves or habitats.
Swampfish are reported as generally rare or uncommon throughout
their range (Poulson 1963, Cooper and Rohde 1980, Shute et al. 1981,
Jenkins and Burkhead 1994). However, its reported rarity may be more of
a sampling artifact than a reflection of true abundance (Ross and Rohde
2003). Abundance estimates may be biased as most sampling occurs
during daylight (Jenkins and Burkhead 1994), whereas C. cornuta are most
active at night and occur in habitats difficult to seine or electro-fish. In
Virginia, the largest series was taken in the Blackwater River drainage via
a nonselective ichthyocide (Jenkins and Burkhead 1994). Poulson found
that repeated kick-seining at the edges of dense weed beds routinely led
to capture of 1–3 swampfish at each attempt for up to six repeats at the
same spot. Ross and Rohde’s made many collections over a year in a short
section of stream and the maximum density estimated from their data is
42,000 per hectare (84 × 2.5 in 50 m2).
Historically, F. agassizii has been considered rare to uncommon
throughout much of its range. Smith and Welch (1978) estimated less
than a thousand individuals from eight springs in Union Co., Illinois, and
around 40 individuals at Cave Springs Cave in Union Co. However, many

hundreds of specimens have been accessioned from a large spring run at

Rich Pond in Warren Co., Kentucky. In Tennessee, F. agassizii can be locally
abundant in ideal habitats (Etnier and Starnes 1993), and, like C. cornuta, is
easily overlooked because of its nocturnal and reclusive habits (Smith and
Welch 1978). Apparent abundance of F. agassizii is dependent upon light
levels, groundwater velocity, and season (N.M. Welch in Smith 1979). Pine
Hills IL spring runs have as many as 80,000 fish per hectare whereas in
Mammoth Cave most often had 0 fish per hectare and 50 per hectare only
in a year with high food input (Poulson 1969).
We suggest that most localities for cave Amblyopids are population
sinks and a very small percentage are population sources. Figure 14
shows that there were <10 fish per survey in 68% of A. spelaea caves, 72%
of Typhlichthys caves, and 87% of A. rosae caves. Conversely, there were
30+ fish per survey in only 5% of A. rosae caves, 8% of Typhlichthys caves,
and 14% of A. spelaea caves. This pattern is even more accentuated if we
consider the 2–3 caves that have 46–87% of all individuals observed for
each species.
For Typhlichthys the top three caves account for 46% of the 354 fish
censused in 39 caves by Niemiller. One population in Putnam County,
Tennessee, had 121 fish in a 400 m stream section (Niemiller, personal
observation) with an estimated 864 per hectare. In Shelta Cave, Alabama,
on one trip there were 64 fish in an area under a gray bat colony (Poulson
1960) with an estimated 229 per hectare (64 × 2.5 in 7000 m2). In several
caves, we routinely see > 40 fish per visit and in two caves these moderate
numbers are consistent over a period of years. From most to least numbers
per trip, Herring Cave in Rutherford Co, Tennessee had 47, 39, 37, and 32
fish. Blowing Springs Cave in Coffee Co., Tennessee had 52, 37, 31, and 26
Fig. 14 Portion of cave Amblyopsid localities with the maximum number of

observed fish during a single survey correspond to the following cateogories: 1
fish, 2–5 fish, 6–10 fish, 11–20 fish, 21–30 fish, and 30+ fish.

fish including recent surveys by Niemiller and surveys from the 1960s by
Poulson.
For A. spelaea, the top three caves by number of fish observations account
for 87% of the 1591 fish censused in 47 caves, and these caves have had high
populations at every survey over the past 50 years (Poulson 1969, Pearson
and Boston 1995). During their surveys in the early 1990s, Pearson and
Boston (1995) conducted mark-recapture and census studies in several A.
spelaea caves in Indiana and Kentucky. During a single trip in March 1994,
they counted 515 fish at Penitentiary Cave in Breckinridge Co., Kentucky.
Over 100 fish were observed during single surveys of two other caves
in Kentucky and the Donaldson-Twin Cave complex in Lawrence Co.,
Indiana. Blue Springs Caverns is also reported to contain a large A. spelaea
population in excess of 150 individuals (Welch in Keith 1988). Poulson
(1969 and VI C this chapter) estimated maximum population densities of
A. spelaea from 417 per hectare in Upper Twin Cave, Indiana (130 ¥ 2.5 in
3600 m2), to 4199 per hectare in Under the Road Cave, Kentucy (84 ¥ 1.5 in
300 m2).
For A. rosae the top three caves by maximum number of fish per census
account for 55% of the 417 fish in 44 caves (G.O. Graening, unpublished
data). These three caves have been regularly censused over a period of 20–
50 years (Graening unpublished data, Poulson 1960, Brown and Johnson
2001). The top three censuses were 100, 122, and 139 for Cave Springs
Cave, Arkansas, 48, 50, and 51 for Logan Cave, Arkansas, and 27, 26, and
33 for Ben Lassiter Cave, Missouri. The densities for the two best caves for
the prime habitats with > 80% of all fish in each cave are 2134 per hectare
for Cave Springs Cave (0.8 ¥ 139 ¥ 2.5 in 1300 m2) and 932 per hectare for
Logan Cave (0.8 ¥ 51 ¥ 2.5 in 1100 m2).
Speoplatyrhinus poulsoni may be the rarest fish in North America,
restricted to Key Cave in Lauderdale Co., Alabama. Alabama cavefish
have only been observed in five pools within the cave and is extremely
difficult to capture because of difficulty in sampling in deep pools in
which it resides. Accordingly, its rarity and difficulty in capture make
mark-recapture studies to estimate population size infeasible. Therefore,
abundance has been inferred from sight observations during pool surveys.
Kuhajda and Mayden (2001) summarized their survey efforts and those of
past researchers since the 1970s. Although the numbers of fish observed
are comparable from the 1970s to the 1980s and the most recent surveys
by Kuhajda and Mayden in the 1990s, abundance is extremely low. Only
two of the five pools have been routinely surveyed since the 1970s. A total
of 25 survey trips of these two pools have been conducted in the last 30
years (reviewed by Kuhajda and Mayden 2001). Up to nine fish have been
observed between these two pools during a single visit with the majority

of observations occurring in the mid-1980s. In the most recent surveys that

included examination of five pools, Kuhajda and Mayden (2001) observed
from 5 to 10 fish with an average of 7.8 fish per survey. Although low in
numbers, three different broad size classes were observed indicating that
recruitment is occurring. In this cave we can clearly only access a very
small amount of habitat and the geology indicates that there is considerable
inaccessible habitat.
VII. FORAGING EFFICIENCY AND METABOLIC

EFFICIENCY
A. Past and New Studies on Metabolic Efficiency

Poulson (1960, 1963, 1985) has documented a trend of increasing metabolic
efficiency from the epigean C. cornuta, to the troglophile F. agassizii, to
Typhlichthys and A. spelaea, to A. rosae to S. poulsoni. Earlier in this chapter,
we have reviewed the old and new evidence, from eyes and pigment,
which corroborates Poulson’s early inference that the four troglobites show
an increasing period of evolutionary time in caves from T. subterraneus to
A. spelaea to A. rosae to S. poulsoni. To summarize the original data and new
data (see Metabolic Rates) there is a trend of about a five-fold decrease in
routine metabolic rates but with a five to ten-fold increase in daily body
lengths swimming (Table 2 in Poulson 1985).
Poulson (1985) hypothesized that part of the increased swimming
activity is due to longer and inferred bigger relative areas of pectoral and
caudal fins that results in greater distance moved per coordinated caudal-
pectoral fin stroke. However, only A. spelaea has average body lengths
moved per fins stroke (0.62–1.25) that is higher than the overlapping body
lengths per stroke of F. agassizii (0.32–0.72), T. subterraneus (0.35–0.67), and
A. rosae (0.42–0.81).
The spontaneous swimming levels, measured at the same time as
metabolic acclimation studies in a 0.5 liter torus respirometer, shows that
the lowered metabolic rates of at least T. subterraneus and A. spelaea is not
at the expense of lowered swimming activity. The ranges of swimming
activity indices are C. cornuta 1–14 <<< F. agassizii 19–195 < T. subterraneus
70–320 = A. spelaea 19–300 >> A. rosae 6–55. These studies were done only
over about a 20–30-hour period.
A more realistic measure of activity is to follow spontaneous activity in
a 0.5-liter torus container with continuously renewed water over a period
of 6–12 days. These data show that activity changed in different ways for
the Amblyopsid species over the study. In general, over time Forbesichthys

activity declined drastically (n=2), Typhlichthys activity decreased slightly

(n=4), A. spelaea activity actually increased on average (n=3), and A. rosae
activity remained the same (n=3). The patterns were consistent within
species.
An even more realistic measure of activity is in a 50-liter aquarium.
Poulson used an elongate aquarium with a bank of infrared sensors at mid-
point to bias for swimming over spontaneous activity. The percent of hours
with at least some swimming was measured for two fish of each species
for 7 days. The percent of hours swimming were as follows: C. cornuta 51
& 53% = F. agassizii 52 & 58% < T. subterraneus 76 & 86% < A. spelaea 92
& 99% = A. rosae 86 & 100%. Along with data on swimming speed, this
gave estimates of body lengths swum per day while foraging (Table 2 in
Poulson 1985) of: C. cornuta 1000 body lengths per day < F. agassizii 1854
<< T. subterraneus 5454 << A. spelaea 10980 > A. rosae 8082.
And a still more realistic measure is provided by Bechler
(personal communication) for fish in a 6.3 × 1.5 m artificial stream with
rock shelters. Using videotape he followed four individuals at a time
periodically for 30 days. The average cm moved per 15 minutes was F.
agassizii 141 << A. rosae 408 < T. subterraneus 647 << A. spelaea 1274.
B. Efficiency of Searching for and Capturing Prey

Poulson and Michael Barnett (unpublished data from 1967–1968) studied
the searching patterns of Amblyopsids and the efficiency of finding and
eating two types of prey at different densities. The prey were ten large
cladoceran (Daphnia magna) and or five small isopods (Lirceus) in a 10-liter
aquarium or one cladoceran or one isopod in a 100-liter aquarium. Fish
sample sizes were 3–5 individuals of each species.
Poulson made a priori predictions of capture efficiency using simulated
foraging of average-sized fish for each species (Table 8). He calculated the
volumes of water that foraging fish could search in the following way.
The width and depth (mm) across which prey could be detected was
predicted based on head dimensions, plus projection of neuromast ridges
and cupulae above head skin plus distance at which neuromast cupula
move to moving prey. This was multiplied by two for head width and for
head depth. These two sensitivity distances (horizontal and vertical) were
multiplied to get a sensitivity area in mm2 and this area was multiplied
by body lengths swum per minute while foraging length to get a volume
of sensitivity in mm3 per minute. The swimming speed used was that
observed when ten Daphnia or five Lirceus were first introduced to the
aquarium with one fish.

Table 8 Predictions of foraging effectiveness. We start by calculating the

surface area (in mm2) of sensitivity to detection of prey around the head. Then,
we convert this to volume (in mm3) searched per time by multiplying surface area
by body lengths swum per minute. Compare to Table 9 Time to capture prey and
Table 9 % of hours some swimming in 7 days.
1. Body length (mm) 42 60 50 65 45
2. Neuromast cup + 0.3 0.66 0.84 1.14 0.96
cupula length (mm)
3. Distance (mm) at 10 16 40 35 30
which cupula moves to
moving prey
4. Head width +2 + 3 × 283 646 2536 1944 2076
head depth + 2 + 3 (mm2)
5. Body length × min–1 × 105 222 425 423 320
Body length (mm)
6. Volume searched 40 143 1078 822 664
(4 × 5) × min–1 × 10–3
= ml min–1
7. Time (min) to search a 249 70 9.3 12.2 15.1
10-l aquarium
8. Time (min) to search a 2487 699 93 122 151
100-l aquarium
The volume of sensitivity per minute in mm3 was converted to cm3 to

compare to the average volume occupied by different densities of the two
prey types in the 10,000 cm3 oor 100,000 cm3 aquarium to get a measure
of how long it would take for the fish to encounter one prey if it did not
repeat searching any volume. This gave predictions of how the species
would differ. This prediction was tested for each species by comparing
the times to capture of each prey until all were gone. In tests with multiple
prey of each type, data were time at which successive prey were captured
along with swimming speed in body lengths per minute and turning rate
(right or left by 90 degrees) in turns per minute.
With increasing troglomorphy among the Amblyopsids, there is a
trend of increasing head size, increased projection of the neuromast cups
and cupulae above the head, and increased distance at which cupulae
move in response to moving prey. This trend is striking when comparing
Forbesichthys to any troglobite but the differences among the troglobites are
subtle (Table 8; Table 2 in Poulson 1985). There was no trend in swimming
speed except that Forbesichthys swam slower than any of the troglobites
and Typhlichthys was the fastest swimming species.

The same relative differences among species were predicted for volume
potentially searched in cm3 per minute (73 cm3 per minute for Forbesichthys
<<< 1300 cm3 for Typhlichthys > 864 cm3 for A. spelaea >= 839 cm3 for A.
rosae). Thus, all of the 10,000 cm3 test aquarium could be searched, if the
fish used no overlapping paths, in 137 minutes for Forbesichthys >>> 8
minutes for Typhlichthys < 12 minutes for A. spelaea = 12 minutes for A.
rosae. How does this compare to the time taken to actually find and eat
different densities and kinds of prey? The answer is not well. There were
unexpected differences in time to capture ten vs. one Daphnia magna and
the species best at finding and eating Lirceus isopods was not the same
species best at foraging for Daphnia cladocerans (Table 9).
A single Daphnia is the foraging challenge that most closely mimics the
extremely low prey densities of zooplankton in caves where the maximum
ever recorded is ~ 5 × 10–2 per 10,000 cm3! With this test, Forbesichthys
did worse than predicted as it took 34–2100 minutes to find and eat the
single Daphnia. Amblyopsis rosae was like the other troglobites and better
Table 9 Comparison of computed volume searched (Table 8) with time to capture

one isopod and one water flea in a 100-l aquarium and first isopod of five total and
first water flea of ten total in a 10-l aquarium. Ranges are reported in parentheses
where available.

Time to search 100-l 2487 699 93 122 151
aquarium (min)
Time to catch one 2160 300 48 25 18
water flea in 100-l
(min)
Time to catch 620 50 (45–74) 6.2 (1.5–12) 2.8 (1.5–4) 14 (4–36)
one isopod in 100-l
(min)
Time to catch first of na 6.8 ± 9.1 1.5 ± 0.5 5.6 ± 5.4 11.0 ± 2.0
10 water fleas
in 10-l (min)
Time to catch first na 4.4 (3.7–5.0) 3.6 (2.2–5.0) 1.8 (0.5–4.0) 9.0 (7.0–11.0)
of 5 isopods in
10-l (min)
Hrs (%) swimming 52 55 81 96 93
over one week
Body lengths × hr–1 1000 1854 5454 10980 8000
in 100-l with no rocks
Body lengths × hr–1 in na 94 557 784 362
6.3 m stream w/rocks
(4 fish over 30 days)

than predicted as it took 3–42 minutes to find and eat the single Daphnia.
What accounts for these departures from prediction? It is not differences
in searching speed or turning rate.
All the Amblyopsid species showed behavioral changes as soon as prey
were introduced though no controls were used adding inanimate objects.
That they were reacting to live prey is shown by sustained increases in
swimming speed. Forbesichthys increased its speed 20% from 3.8 to 6.1
body lengths per minute. Typhlichthys increased its speed 30% from 6.6
to 8.6. Amblyopsis spelaea increased its speed 49% from 4.1 to 6.1 and A.
rosae increased its speed 51% from 4.7 to 7.7 body lengths per minute.
All species increased their turning rates by three-fold which helped keep
them in the same area where they had caught a prey when multiple prey
were in the test aquarium. This behavior would be even more adaptive in
caves with much lower densities and much higher patchiness of prey.
When a fish detects a living prey there are also similarities among the
species. If a large food item like an isopod or white worm is dropped into
an aquarium. Most individuals of all species move toward the area of
disturbance (Eigenmann 1909, Hill 1966, Barnett and Poulson, unpublished
data). Also when a fish approaches a prey on the bottom, it will pause
and back up if the prey stops moving. Then it will “jerk and grab” as the
prey starts moving again. All species use a lunge and grab motion as they
capture a prey item; none use gape and suck to catch prey.
We have especially detailed observations of the changes in swimming
rate, turning rate, and general behavior of A. spelaea while foraging for
and eating ten Daphnia magna. In Michael Barnett’s words the fish initially
show “lots of turning and jerking’ and one fish was “furious at four near
misses”. After the fish had reduced the number of Daphnia to 4–6 their
swimming and turning rates increased from 0–3 to 4–18 body lengths per
minute and turning increased from 0–2 to 2–5 turns per minute. These
rates slowed some when they had reduced prey number to 2-4 and the
fish seemed “much less frustrated” if they missed a first strike. One
interpretation is that they were becoming “full” since the interval between
captures increased.
The relative success of Amblyopsid species (Table 9) was different
with ten Daphnia in a 10-liter aquarium than with only one in a 100-liter
aquarium and this more closely matched the predictions from simulated
volume searched per time. Though Forbesichthys did better and A. rosae
worse than predicted, all species were effective foragers. They showed
statistically but not ecologically significant differences. The times to catch
the first prey ranged in minutes from 1.5 ± 0.5 SD T. subterraneus to 11.0
± 2.0 SD A. rosae, with T. subterraneus < A. spelaea = F. agassizii < A. rosae.
The relative differences were the same for the 5th and 10th prey captured

though F. agassizii and T. subterraneus showed no decline in intervals

between prey capture and A. spelaea and A. rosae capture intervals increased
after the 5th prey was captured. Poulson believes that T. subterraneus was
best at capturing Daphnia once detected because its neuromast cupulae
are knobbed at the tip and seem to be especially sensitive to slight water
movements. We cannot explain why A. rosae was the slowest at eating
Daphnia since it was the fastest by far at finding and eating a single Daphnia
(above).
Interestingly, A. spelaea was fastest at finding and eating isopods whether
with five or one per test aquarium. We think this is because isopods are its
principal prey in nature (see Food Habits) and its search patterns in nature
suggest they are concentrating on finding isopods (see Mohr and Poulson
1966). For sample sizes of three for each species, the range in minutes for
capture of the first of five isopods was 0.5–4 for A. spelaea to 7–11 for A.
rosae with A. spelaea < T. subterraneus = F. agassizii < A. rosae. For capture of
the third of five isopods, A. spelaea was still fastest but now Forbesichthys
was slower than A. rosae: A. spelaea <= T. subterraneus < A. rosae < F. agassizii.
These differences were accentuated with capture times of a single isopod.
Times to capture were 2–4 minutes for A. spelaea and 7–74 minutes for
F. agassizii with A. spelaea < T. subterraneus << A. rosae << F. agassizii. Part
of the explanation is that both A. spelaea and T. subterraneus search almost
only at the bottom where the isopods are located, whereas F. agassizii and
A. rosae continue to search the entire tank. This is a more effective strategy
for finding Daphnia that swim throughout the water column.
Poulson’s perspective from these foraging experiments is that his
predictions based on morphological troglomorphy missed some interesting
differences in effectiveness when foraging for different prey types. These
differences seem to be due partly to the types of prey most common in
each species habitat. We have just mentioned this for A. spelaea, which
lives in caves that have many isopods. We predict that Forbesichthys from
Pine Hills will outperform even the troglobites with amphipods as prey
because it eats only amphipods in the field. Also, Forbesichthys from Rich
Pond eats mainly chironomid larvae and oligochaetes and so will also be
effective foraging for amphipods with their strong thigmotaxis and rapid
grabbing during prey capture. We have watched them eat white worms in
the lab and they are especially effective. With worms, midge larvae, and
amphipods the thigmotaxis used in prey capture complements lateral line
detection of prey. We also predict that A. rosae and Typhlichthys may be most
effective at foraging for copepods that are their main prey. This should also
be true for small A. spelaea < 40 mm SL that eat mainly copepods and few
isopods in the field. We think Typhlichthys will do the best with copepod
prey because our lab observations show that it is especially sensitive to

slight water movement, even ‘excitedly’ moving towards a previously

quiescent isopod when it started to wave its antennae!
C. Hypothesized Tradeoffs of Living in Caves

Despite the negative (–) tradeoff of low food supply in caves, it is clear
that Forbesichthys spawns in caves. Hill’s (1966) study of Forbesichthys was
much more complete than Weise’s (1957) though both provide excellent
laboratory observations of feeding and behavior. In both sites, the fish
reproduce in the cave since adults disappear from the springs for a few
months starting in late fall and then adults and small young appear
outside in the spring. Unlike Weise, Hill was able to study the fish in
both the spring run and in the cave and the contrasts provide important
insights into the tradeoffs of the cave and spring run for feeding, refuge,
and reproduction.
The Rich Pond Cave may be a refuge from parasitism and predation
for newly hatched young but it has 3–4 orders of magnitude less food that
Hill sampled outside and the new hatched fish of 6 mm SL stop growing
and die if they do not leave the cave. If they leave the cave, they grow
to about 35 mm SL their first year eating > 80% chironomid larvae by
volume. When fish then overwinter in the cave their guts are 99.9% empty.
When fish leave the cave the following spring, they grow to about 50mm
SL with a diet of approximately 20% Forbesichthys and 78% chironomid
larvae. When fish return to the cave again, about 82% of their guts have
Forbesichthys and the rest are empty. This indicates that the spring cavefish
is especially food-limited in caves and may not be a good model for the
way other Amblyopsids became isolated in caves.
Forbesichthys may only be a marginal troglophile because there is no
documentation of it reproducing in caves. The best-studied populations
emerge nightly or seasonally from food-poor caves to feed in spring
runs that have abundant live prey. If the surface spring runs dried up
with climatic warming and drying then these populations might well go
extinct. So a better habitat type to allow gradual evolution of food finding
and metabolic efficiencies would be a sinking stream. These would have
reliable allochthonous organic input and so much higher food supplies
than in caves that exit to springs. Forbesichthys is occasionally found in
such caves, in the Mammoth Cave Region, but these populations have not
been studied to see if they have incipient adaptations to low food supply.

D. Matching Amblyopsid Adaptations to Cave Type

Poulson has had a bit of an epiphany as he has spent three months reviewing
old data, working up old unpublished data, and carefully re-reading what
he thinks are the best studies available. He agrees with Noltie and Wicks
(2001) that A. rosae and T. subterraneus occupy very different kinds of caves
and karst in Missouri and here extend their suggestions about how cave
type relates to all troglobitic Amblyopsid biology.
Typhlichthys occurs over such a wide geographic area with such a great
difference in cave and karst type that it is difficult to make generalizations.
If the hypothesis of some match of adaptations to cave type is correct then
we predict differences between Missouri populations and Kentucky–
Tennessee populations. The Missouri Typhlichthys are in caves that are
far below the surface, have extensive underground watersheds, and have
relatively low input of allochthonous organic matter due to often great
distances from surface inputs and low currents. One prediction is that
they will have more efficient metabolic patterns and foraging with longer
lives and larger sizes than reported thus far for T. subterraneus. The studied
Typhlichthys east of the Mississippi River occur in caves that are close to the
surface, have small underground watersheds, and often have streams that
flood and bring in allochtonous organic matter. For these Typhlichthys, we
may expect to find local adaptation of populations that have been isolated
in caves independently. Local adaptation is especially likely since genetic
evidence indicates less gene flow than among cave populations of other
Amblyopsid troglobites.
Amblyopsis rosae occur in caves that are shallow with small underground
watersheds (Noltie and Wicks 2001) and small rocky streams that never
flood severely. The shallow caves often have extensive allochthonous
organic input and guano from gray bat maternity colonies in the habitats
with the largest populations of fish. Poulson (1985) has previously
suggested that this high food supply is related to the relatively short
life and relatively frequent reproduction of A. rosae compared to other
Amblyopsid troglobites. He argued that small maximum size (smallest of
all cave Amblyopsids) is associated with less food needed to get the extra
energy needed to allow reproduction. Here we extend the connection of
high food supply to relatively fast growth rates allowed by an especially
energy efficient metabolic pattern. This along with an inferred small
clutch size (see Fecundity) allow for more frequent reproduction that
is manifested in a fairly even size-frequency distributions in the best
habitats.

Amblyopsis spelaea is the largest Amblyopsid and occurs in large streams

that have the highest input of allochthonous organic matter and often
strong currents. The extreme of input of coarse particulate organic matter
is in Penitentiary Cave, Kentucky (Pearson and Boston 1995), with 519
fish censused in 3702 meters of rimstone pools with 10–30 cm of leaf, twig,
and acorn fragments. And the extreme for live prey input is in Twin Cave,
Indiana, with 130 fish censused in 3600 m2 of riffles and pools with 70 ml
of plankton including hundreds of cladocerans and copepods and 500 ml
of fine silt and organic matter in an 800 liter sample during one flood (Scott
1909); this cave also has very high densities of troglobitic isopod prey.
Poulson now hypothesizes that even those cave stream A. spelaea habitats
now predominately silt-bottomed and slow-flowing still have high food
supplies and once had higher stream velocities and greater flooding.
His reasoning is that the range of A. spelaea is just south of the farthest
glacial advances where there would have been much greater flooding
and faster cave enlargement than at present. He thinks that the dense
populations of isopods in many A. spelaea caves are explained by cave
stream character and that the efficient foraging for isopods by fish > 35
mm SL is associated with isopods as the most frequent and abundant prey
larger than copepods. Regular flooding with strong currents has selected
for large size and abundant isopods have allowed large size and a large
clutch size with branchial incubation. Consonant with this hypothesis is
Bechler’s observation (personal communication) that in an artificial stream
A. spelaea was the most active swimmer of the four Amblyopsids studied.
An interesting corollary of large size of A. spelaea may be that it allowed
syntopic coexistence with the much smaller Typhlichthys in the Mammoth
Cave System (Poulson 1992). Recall that T. subterraneus is in the more
food-rich small shaft drain streams and is replaced by A. spelaea as you
census downstream to the larger and more food-poor base level streams.
These A. spelaea in the Mammoth Cave System grow to much larger sizes
(115–130 mm SL) than farther north (80–90 mm SL) and this may reflect
character displacement. As a final note, Poulson thinks that he missed the
significance of large size in A. spelaea because he always reported traits
at the same 45 mm SL size. This was a kind of control when comparing
troglomorphic trends (Table 2 in Poulson 1985). Poulson (2001b) has
discussed the + and – tradeoffs of large body size for cavefish and for A.
spelaea the positives are much greater than the negatives.
What kinds of specializations to the character of Key Cave might we
look for in S. poulsoni? Key Cave is the only locality for S. poulsoni and
its character may have changed due to anthropogenic impacts (see below
Threats). Another complication is that we cannot access much of its habitat
even using SCUBA. Nonetheless, Key Cave is a maze system developed at

and well below the water table. It does not flood and the only allochtonous
organic input seems to be bat guano. The predominant prey is undoubtedly
copepods though a large fish might be able to eat a very small crayfish. The
maximum size estimated for visual sightings is around 60 mm SL. They
are apparently very slow growing based on a conservative scale mark age
estimate of 5 years for a 42 mm SL fish. If Alabama cavefish are like other
Amblyopsids, we can multiply this estimate by 2–3 to get a probable age of
10–15 years. Also, since adults of other Amblyopsids slow down greatly in
growth rate a 60 mm SL fish (estimated size of largest observed fish) may
be 20–30 years old. This suggests extremely low available food supply and
is consonant with estimates that its metabolic rate is even lower than for
A. rosae. It is also consonant with the extreme neoteny inferred from its
huge relative head size and unbranched fin rays in adults. These extremes
may represent both adaptation to its very low food supply and the
inference that it has been isolated in caves the longest of all the troglobitic
Amblyopsids based on pigment and optic system rudimentation (Table 4;
Table 2 in Poulson 1985). Whether we will be allowed to test the predicted
genetic, physiological, and behavioral correlates of these morphological
trends is problematic because the species is listed as ‘critically endangered’.
Poulson would at least love to see his namesake alive before he dies (but
see Poulson, 2009a)!
VIII. CONSERVATION
The conservation status of subterranean fishes has received increasing
attention in the past few years. According to Proudlove (2006), 63 of the
104 known species of subterranean fishes are listed by the International
Union for the Conservation of Nature and Natural Resources (IUCN 1996,
2000). All four troglobitic Amblyopsids are included on the list. Three are
considered Vulnerable and the other, S. poulsoni, is considered Critically
Endangered. Two species, A. rosae and S. poulsoni, are listed by the United
States Fish and Wildlife Service (USFWS) under the Endangered Species
Act.
In this section, we review the conservation status of Amblyopsid
fishes including the two nontroglobitic species, examine the major threats
facing each species, and conservation measures that have either been
implemented or proposed. This section largely follows that of Proudlove’s
(2006) chapter titled “The Conservation Status of Subterranean Fishes.” We
encourage readers seeking more information about the conservation of
other species of troglobitic fishes to examine the works of Proudlove (2006)
and the ‘Threatened Fishes of the World’ series in the journal Environmental
Biology of Fishes.

A. Conservation Status
All troglobitic Amblyopsids are considered vulnerable or endangered
across their respective distributions (Table 10). In general, there is trend
towards increasing threat status with increased cave adaptation. The non-
troglobitic species are apparently secure throughout their respective ranges,
although disjunct populations of F. agassizii in southeast Missouri are listed
as endangered (Missouri Natural Heritage Program 2008). Typhlichthys
subterraneus (as currently recognized) is the most widely distributed and
least cave-adapted (Poulson 1963) of the cave Amblyopsids. As such, it is
considered the most secure (although it is afforded protection in several
states) and is considered endangered only in Georgia where it ranges only
into the extreme northwest corner of the state. Amblyopsis spelaea is afforded
protection at the state level and is a species of concern by USFWS (USFWS
1996). Amblyopsis rosae is listed as endangered throughout its range and
was listed as threatened by USFWS in 1984 with a recovery plan written in
1989 (USFWS 1989). The most cave-adapted Amblyopsid, S. poulsoni, has
been described as “the rarest American cavefish and one of the rarest of all
freshwater fish” (USFWS 1996) and is found only in Key Cave, Lauderdale
Co., Alabama. Because of its extremely limited distribution and suspected
low population size, S. poulsoni is designated as Critically Endangered by
IUCN and was listed as threatened by USFWS in 1977 and later elevated
to endangered in 1988.
B. Threats
Proudlove (2006) listed five broad threats that subterranean fishes can face.
This list includes: (1) habitat degradation, (2) hydrological manipulations,
(3) environmental pollution, (4) overexploitation, and (5) impacts of
introduced aquatic animals. Many of the threats discussed below are
interrelated because of their wide range of potential effects. For example,
dam construction can result in direct destruction and degradation of
cavefish habitat, alter hydrological patterns, and allow surface species
to colonize and either compete or prey on existing cavefish populations.
Here we generally follow the broad classification of threats listed by
Proudlove (2006) and review the threats to Amblyopsid populations and
focus the majority of our discussion on three troglobitic species, A. rosae,
A. spelaea, and S. poulsoni, for which aspects of conservation have been
most thoroughly examined. We focus on the first four of Proudlove’s list
as little work has investigated the effects of introduced species on cavefish
populations.


Table 10 Conservation status of each Amblyopsid species throughout its respective distribution
Listing C. cornuta F. agassizii T. subterraneus A. spelaea A. rosae S. poulsoni
Federal Listing Not listed Not listed Not listed Not listed Threatened Endangered
Global Status G5 G4G5 G4 G4 G3 G1
IUCN Red List Not listed Not listed Vulnerable Vulnerable Vulnerable Critically
Category Endangered
Alabama Protected (S3) Endangered,
Protected (S1)
Arkansas Inventory Element (S1) Endangered (S1)
Georgia Not listed (SU) Endangered (S1)

Illinois Not listed (S1S2)
Indiana Endangered (S1)
Kentucky Not listed (S4S5) Special Concern (S2S3) Special Concern (S3)
Missouri Endangered (S1) Not listed (S2S3) Endangered (S2)
North Carolina Not listed (S4)
Oklahoma Endangered (S1)
South Carolina Not listed (SNR)
Tennessee Not listed (S4) Deemed in Need of
Management (S3)
Virginia Not listed (S3)
Sources: Alabama Natural Heritage Program 2007, Illinois Endangered Species Protection Board 2006, Indiana Department of Natural Resources 2007, Kentucky State
Nature Preserves Commission 2005, Missouri Natural Heritage Program 2008, Oklahoma Department of Wildlife Conservation 2008, Romero and Conner 2007, Withers
et al. 2004.
Habitat Degradation and Alteration

Habitat degradation and alteration can result from the direct destruction
or manipulation of habitat during quarrying and mining operations,
highway construction, and urban development. Keith (1988) reported
two A. spelaea localities were either partly or completely destroyed by
quarrying operations. However, the majority of habitat degradation
and alteration threats are indirect resulting in loss of habitat because of
siltation, sedimentation, and alteration of hydrological flow patterns and
levels. Many caves in the eastern Interior Low Plateau have massive silt
banks along streams that are likely associated with farming that began
in the 1800s (Poulson, personal observation). However, some caves
contain cavefish populations that are found entirely on silt substrates
and have high population sizes and frequent reproduction. Likewise, on
a much longer time scale, huge changes in habitat composition and food
availability must have occurred with glacial cycles during the Pleistocene.
Studies are needed to assess the actual rather than the potential impacts
on cavefish populations from increased siltation and sedimentation.
Land development within cave recharge zones can alter surface runoff
patterns or even block or destroy major recharge points. This can result
in dramatic habitat alteration because of increased or decreased water
volume, water velocity, sedimentation, or stream scouring depending
on local hydrological patterns. In forested areas, increased erosion and
production of sediment because of logging can result in increased siltation
and sedimentation or the complete blockage of a cave passage. For surface
Amblyopsids, loss of forested areas can cause the decline or loss of local
populations. The installation of an electric transmission line has been
cited as the cause of population decline at a C. cornuta locality in Virginia
(de Rageot 1992). Removal of the surrounding forest caused increased
insolation and drying of aquatic habitat.
Reduced input of surface runoff in recharge zones could have dramatic
impacts on reproduction of Amblyopsid cavefishes. Cavefish are believed
to rely on increased flow and small temperature changes associated with
cave flood events during winter and spring to coordinate reproduction
and spawning (Poulson 1963, 1969). Because most populations of cavefish
are reportedly small (S. poulsoni in particular), the importance of successful
timing of sexual maturation and spawning is magnified (Kuhajda 2004).
Reductions in surface runoff may disrupt the environmental cues necessary
for successful reproduction leading to greater susceptible to extirpation.
Impoundments have resulted in many problems and threats for cavefish
populations. A primary example is the construction of Lock and Dam #6
constructed on the Green River below Mammoth Cave in 1906. Although
the Green River naturally back-floods into the cave system, flood levels

have increased since dam construction (Lisowski and Poulson 1981). The
Styx and Echo River areas in Mammoth Cave experienced an apparent
decline in cave biota, including cavefish, from the late 1800s to the 1920s
(Elliott 2000). From the 1950s to the 1970s, cavefish were still observed
but were only large in size and low in abundance. The present rarity of A.
spelaea in the Echo and Styx River parts of the system likely is related to
flooding and silting associated with deforestation and construction of Lock
and Dam #6 and other impoundments along the Green River and the Nolin
River, a tributary to the Green River, in the 1970s (Poulson 1969, 1996, Elliott
2000). In addition to alteration and loss of habitat, river impoundments
may have resulted in decrease in within cave downstream transport of
particulate organic matter by hydraulic damming when releases from
the upstream Green River dam keep the river level at Mammoth Cave
from declining completely after flood peaks. This exacerbates the decline
of cave biota in the base-level stream of Mammoth Cave (Poulson 1996).
This is discussed in more detail by Poulson (1992).
Hydrological Manipulations
Hydrological manipulations can include underground water removal for
human consumption, irrigation, or industry. However, some hydrological
manipulations, such as impoundments or increased surface runoff, can
raise water tables and alter habitats (see above). Lowering of the water
table resulting from direct human consumption, irrigation, or industrial
use may threaten cavefish populations. Because S. poulsoni lives in a zone
of seasonal oscillation of the water table (Trajano 2001), lowering of the
water table (and drainage alterations) may isolate fish in these pools
exposing them to decreased oxygen levels if decomposable organic mater
is present and increased concentrations of contaminants and death if pools
dry out during the dry season (Kuhajda 2004). Accordingly, lowering of the
water table because of a proposed industrial park for the city of Florence,
Alabama, within the recharge area of Key Cave has been cited as a threat
to S. poulsoni (USFWS 1977, Kuhajda and Mayden 2001).
Groundwater Pollution
Groundwater pollution has been listed as factor negatively affecting
populations for all cave Amblyopsids. This threat includes eutrophication
and contamination from agricultural and industrial runoff containing
pesticides, fertilizers, and heavy metals, sewage effluent, spills and illegal
dumping of hazard materials, and thermally altered runoff. Although
few studies have examined the direct effects of groundwater pollution
on cavefish populations in detail, several studies implicate this threat in
population declines.

Groundwater pollution may be acute in nature, such as a toxic spill

resulting in a large impulse of contaminants, or chronic occurring over
several months to years (Proudlove 2001). Both forms have been attributed
to cavefish declines or extirpations from cave systems. Nearly 1,000 dead or
dying T. subterraneus were expelled from Meramec Spring in Missouri after
a fertilizer pipeline rupture in November 1981 caused acute, catastrophic
deoxygenation of groundwater (Vandike 1984, Crunkilton 1985). In
contrast, several decades worth of gross pollution by decomposable organic
matter (creamery waste) and heavy metal contamination (electroplating
waste) is the suspected cause of the apparent extirpation of T. subterraneus
along with other cave life at Hidden River Cave in Kentucky (Lewis
1996). However, T. subterraneus and other cave life have re-colonized areas
previously affected from far upstream refuges.
Several sources have been implicated in heavy metal and hazardous
chemical contamination of groundwater throughout the distributions
of cave Amblyopsids. These sources include drilling related to oil and
gas development, inactive wells, industrial effluents, accidental spills,
underground storage tank leaks, sinkhole dumping, and runoff from
roads (road salt) and agricultural fields (pesticides and herbicides) (Keith
1988, Pearson and Boston 1995, Kuhajda and Mayden 2001, Kuhajda 2004).
Heavy metal runoff from a local landfill may threaten populations of T.
subterraneus in Pulaski Co., Kentucky (Tercafs 1992). Increased contaminant
levels may cause increased susceptibility to disease. Pesticides were
attributed as the cause of “broken back syndrome” that affected perhaps
10 percent of a population of A. spelaea in Indiana (Keith and Poulson
1981).
At least four of these threats, industrial effluents, underground storage
tank leaks and sinkhole dumping, have been connected to the decline of
T. subterraneus and other cave life from Hidden River Cave (Pearson and
Boston 1995, Lewis 1996). Organic enrichment from sewage treatment plant
effluents and septic tank leaks also have been implicated at Hidden River
Cave and other caves with Amblyopsids. Organic enrichment can increase
nutrients in an otherwise low-nutrient environment and drastically alter
food web dynamics, increase risk of disease, and dramatically decrease
dissolved oxygen levels. Brown et al. (1998) attributed a 30% decline
in a population of A. rosae to increased levels of inorganic and organic
compounds.
Groundwater pollution is a major threat to S. poulsoni (Kuhajda 2004)
and has been considered as one of the factors likely to cause the decline
of the species. Herbicide and pesticide application to cotton fields in
the recharge zone of Key Cave have been shown to have direct access
to into the cave. Moreover, a sewage sludge disposal operation also lies
within the recharge zone of the cave (Aley 1986). Thermal pollution in

the form of thermally altered runoff may alter reproductive cycles of

Alabama cavefish (USFWS 1982, Kuhajda 2004) and potentially reduce or
eliminate recruitment. These and other threats associated with either the
current or potential alteration and degradation of the Key Cave recharge
area and aquatic habitat, along with the perceived small population size
and extremely small distribution have resulted in the listing of S. poulsoni
as ‘Critically Endangered’ because of its high risk of extinction (Hilton-
Taylor 2000). These threats have been addressed by Kuhajda (2004) and
outlined in detail in three different versions of the USFWS Recovery Plan
for S. poulsoni (USFWS 1982, 1985, 1990).
Collection and Cave Visitation

The collection of cavefish, illegal or otherwise, for the aquarium trade or
scientific purposes may pose a threat to all cave Amblyopsids. Because of
their uniqueness to hobbyists and the ease at which individuals can be
captured, cave Amblyopsids can be easily exploited. Over-collection of
fish can reduce or even eliminate local populations. The rarity of A. spelaea
in the Echo River and River Styx sections of Mammoth Cave system and
its presumed absence from adjacent caves to the north have led some to
speculate that the species was either introduced or decimated during the
1800s when it was sold as a novelty (Poulson 1968, Elliott 2000). Over-
collection by both the scientific community and amateurs is thought be a
concern for populations of A. rosae (Culver 1986, USFWS 1989).
Commercial exploitation of caves can either alter or even destroy
considerable amounts of cavefish habitat. Commercial caves increase
human traffic and disturbance in addition to increased light levels. At least
five populations of A. spelaea are indirectly affected by commercial cave
tours including Blue Springs Caverns and Upper Twin Cave in Indiana
and Hidden River Cave and Mammoth Cave in Kentucky (Pearson
and Boston 1995). However, the exact impacts and long-term effects of
commercial cave operations remains to be examined. At Upper Twin
Cave, no differences in apparent abundance of A. spelaea exist between
times when tours are conducted and times when tours are not in operation
(Poulson, personal observation). And the continued abundance of fish
in Upper Twin Cave since Eigenmann’s studies suggests that pole boat
tours in the downstream part of the cave, starting in the 1950s, have not
compromised the populations.
Human disturbance caused by increased traffic is more of a concern than
commercial exploitation. The activities of even the most cautious caver
may have serious impacts on cave organisms in shallow, silt-bottomed
streams. Disturbance caused by substantial cave visitation may alter
breeding of cavefish populations, disturb food sources, and unknowingly

stress individual fish by increasing fish activity. However, no evidence has

been obtained for any of the above.
Increased disturbance caused by human disturbance is thought to
negatively affect grey bat (Myotis grisescens) colonies by increasing bat
mortality or the eventual abandonment of a cave. If bat colonies are
extirpated, cavefish populations may lose an important source of food and
nutrients (USFWS 1989). This could be a very serious threat given that
the caves with the largest A. rosae populations are also grey bat maternity
colony caves. However, more studies are needed to assess the impacts and
lasting effects of intense cave visitation on cave fauna. The huge decline
in the rich aquatic fauna of Shelta Cave, Alabama, including Typhlichthys,
has been attributed to the loss of a major grey bat roost (Hobbs, personal
communication).
Mechanisms of Impact: Community Signatures

Poulson (1992) has hypothesized community signatures both as early
warning signs and to identify the four most common kinds of pollution
impact. These are meant to supplement Indices of Biological Integrity (IBIs)
developed by William Pearson and his students to detect any changes in
aquatic cave communities over time. As with IBIs, community signatures
require un-impacted baselines. And impacted caves serve as inadvertent
experiments to test the community signature hypotheses. Poulson (2009b)
proposes that aquatic cave communities will show different responses
to siltation, eutrophication, acute toxicity, and chronic toxicity. These
signatures are based both on first principles of ecology and toxicology and
the differences in life history and metabolic rate of different aquatic cave
fauna. The obvious caveat for using community signatures is that several
kinds of pollution may occur in the same cave. In these cases, there should
be extremely low IBIs. Historically, Hidden River Cave, Kentucky, lost all
its troglobites in the main stream. The normal species were replaced by
sheets of colonial sewage bacteria and by dense mats of sludge worms
that were not killed by the massive inputs of decomposable organic
matter with attendant low dissolved oxygen and/or high concentrations
of heavy metals.
Differences are expected with chronic toxicity at low levels and
acute toxicity at high levels. Heavy metals, like lead and mercury, and
organic compounds, like chlorinated hydrocarbons, are usually toxic
because they are not naturally occurring and so organisms have not
evolved physiological ways of sequestering or detoxifying them. As with
herbicides used on weeds as a model, the species with fastest metabolic
and growth rate will be most quickly and seriously affected by a pulse of
heavy metals or pesticides. Also, species with low metabolic and growth

rates may not be affected by pulsed toxicity but are at particular risk with
chronic low levels of toxins due to continued bioaccumulation over a long
lifetime. If they are also top predators they are in double jeopardy due
to biomagnification of toxins along food chains. With acute toxicity their
prey may be killed.
Aquatic cave communities impacted by acute toxicity may be missing
the organisms at the beginning of the food chain. Troglophiles at the base of
food chains will be more affected than troglobites because the troglophiles
have the highest growth and metabolic rates. If there is a pulse of toxin
input, as in a railroad car derailing or truck accident on a highway near a
sinkhole, then Forbesichthys is most at risk and A. rosae and Speoplatyrhinus
at least risk.
The community signature for low-level toxic pollution will be
different with opposite vulnerabilities than for high-level toxic pollution.
Among Amblyopsids, troglobites will be most affected due to both
biomagnification and bioaccumulation. Size frequency distributions will
show that the largest and oldest fish are missing or under-represented
compared to smaller fish. Even troglobitic crayfish may be at risk if the
food supply is low and since longevities are documented to be as long as a
century in food-poor caves! Short-lived species like copepods and isopods
may be unaffected even if they are troglobitic. They may even increase
in density if predatory fish and crayfish are reduced in numbers due to
chronic toxicity.
It is difficult to find an unimpacted control cave to provide a comparison
for the expected community signature for siltation. The problem is that a
great number of caves have had increases in silt levels associated with
land clearing for agriculture over the past 200 years. Nonetheless, the
expected impact of siltation is homogenization of the stream bottom
habitat. At the extremes, silt can cover rock and gravel refuges for isopods
and amphipods in riffles and so there will be fewer prey washing into
pools and shoals deep enough for fish and crayfish. Silt is also likely to
cover or mix with fine particulate organic matter that copepods, isopods,
and amphipods graze or ingest. Also, at low levels it may foul the biofilms
on rocks that are grazed by everything except predators. Poulson (1992)
has provided detailed examples for the lower levels of the Mammoth
Cave where siltation is due to the combined effects of downstream and
upstream dams on the hydropattern of Green River into which the cave
streams flow.
The community signature for organic enrichment (eutrophication) is
the most clear since it is sensitive to the level of pollution. The mechanism
is that decomposition of organic matter provides increased food at low
levels but uses up dissolved oxygen. At low levels of organic enrichment

species at the beginning of the food chain increase in numbers and their
size frequency distributions are skewed more to smaller size classes
with increased rates of reproduction. This is especially true if they are
troglophiles. Over time the faster reproducing troglophiles simply
outproduce the troglobites even though more energy efficient troglobites
cannot be outcompeted. See Poulson (this volume) for an explanation of
how this may lead to demographic swamping of troglobites by the faster
reproducing troglophiles.
At high levels of organic enrichment, especially if pulsed in time, all
the normal aquatic fauna is replaced by species tolerant of extremely low
dissolved oxygen. At the extreme there are only stringy mats of colonial
sewage bacteria, like Sphaerotilus, and tubificid worms. The red tubificids
have a hemoglobin that can bind oxygen at very low concentrations and
so these worms can be seen as waving mats at the stream edge where a
little bit of dissolved oxygen remains.
C. Conservation Measures
Several conservation measures have been proposed or implemented for
populations of cave Amblyopsids. Fencing or gating of cave entrances have
been proposed or implemented to reduce and control human visitation
to sensitive cave ecosystems including Amblyopsid caves. Special bat
gates are needed to allow entry and exit by bats but stop human entry.
Bat Conservation International and The National Speleological Society
have been leaders in the improvement and installation of such bats on an
increasing number of bat caves.
Protection of cave surface and subsurface watersheds is probably the
most important intervention for cavefish caves. Thomas Aley (Ozark
Underground Laboratory, Protem, Missouri) is one of the best practioners of
state-of-the-art water tracing that is critical to delineating cave watersheds.
Among others his studies have led to the protection of watersheds of at
least Key Cave, Alabama, the only locality for Speoplatyrhinus poulsoni,
and for the best Amblyopsis rosae cave, Cave Springs Cave in Arkansas.
Watershed protection has included establishing preserves as well as
institution of best land management practices around sinkholes and
sinking creeks that includes reforestation. In other cases water tracing
has identified the source of pollutants and so allowed legal action that
remedied the situation. Hidden River Cave in Kentucky is one example.
We suggest that what we have called source caves deserve complete
protection of their watersheds. Recall that a few caves for each species
have the vast majority of all individuals ever censused. To us, attention to
protecting these caves is a number one priority for the near future.

Despite the fact that the population of S. poulsoni appears to have

remained stable over the past 30 years, perturbations within the recharge
basin could alter the status of this species (Kuhajda and Mayden 2001).
Accordingly, USFWS purchased 1060 acres within the recharge basin in
January 1997 and established the Key Cave National Wildlife Refuge.
Likewise, the Logan Cave population of A. rosae is protected by the 123-
acre Logan Cave National Wildlife Refuge, and another population is
protected by the 40-acre Ozark Cavefish National Wildlife Refuge.
Introduction of cavefish to new localities or to caves that were historic
localities is worth considering. Until we learn to breed Amblyopsids the
only source for introductions is existing caves with thriving populations.
To protect genetic integrity these source caves should only be in watersheds
that include the recipient cave. In the case of Speoplatyrhinus, there is only
one cave so spreading of risk of extinction by introductions to adjacent
caves will require very careful consideration.
Acknowledgements
First we thank the editors of this book, especially Eleonora Trajano,
for inviting our participation. Second, we thank a number of friends
and colleagues who have allowed us to cite unpublished observations,
answered questions, and discussed ideas while we have been writing
this chapter. In alphabetical order these are Ginny Adams, Tom Aley,
Claude Baker, Richard Borowsky, Mike Barnett (deceased), Tom Barr,
David Bechler, Ken Christiansen, Cheryl Coombs, John (‘cOOp’) Cooper,
Dave Culver, Bill Elliott, Ben Fitzpatrick, Beep Hobbs, Bill Jeffery, Jim
Keith, Bob Kuehne (deceased), Bernie Kuhajda, Brian Miller, Tom Near,
Bill Pearson, Al Romero, “Leo” Trajano, Ted Uyeno, and Fred Whittaker.
Third, we thank John (‘cOOp’) Cooper, Dante Fenolio, Bernie Kuhajda,
Rick Mayden, Al Romero, and Uland Thomas for use of fish photographs
and drawings. Fifth, we thank authors of what we consider to have been
especially thorough, insightful, groundbreaking, and helpful research
papers, reviews or dissertations concerning one or more species in
the Amblyopsidae. These include Dave Bechler, Zach Brown and Jim
Johnson, Carl Eigenmann (deceased), Michael Fine, Loren Hill, Kathrin
Hüppop, Bill Pearson, Doug Noltie and Carol Wicks, Fernandus Payne
(deceased), Graham Proudlove, Steve Ross and Fred Rohde, Al Romero,
Dave Swofford, John Weise, Norbert Welch, and Loren Woods and Bob
Inger. Last we thank all of our many colleagues over the years who have
been helpful in the field and with discussions about cave biology. You
know who you are and we thank you one and all.

References
Adams, G.L. and J.E. Johnson. 2001. Metabolic rate and natural history of Ozark
cavefish, Amblyopsis rosae, in Logan Cave, Arkansas. Environmental Biology of
Fishes 62: 97-105.
Agassiz, J.L.R. 1847. Plan for an investigation of the embryology, anatomy and
effect of light on the blind-fish of the Mammoth Cave, Amblyopsis spelaeus.
Proceedings of the American Academy of Arts and Sciences 1: 1-180.
Agassiz, J.L.R. 1853. Recent researches of Prof. Agassiz. American Journal of Science
and Arts 16: 134.
Alabama Natural Heritage Program. 2007. Alabama inventory list: The rare,
threatened and endangered plants and animals of Alabama. Alabama Natural
Heritage Program, Montgomery, Alabama, 55 pp.
Aley, T. 1986. Hydrogeologic investigations for a proposed landfill near Florence,
Alabama. A contract study and report for Waste Contractors, Inc.
Aley, T and C. Aley. 1979. Prevention of adverse impacts on endangered, threatened,
and rare animal species in Benton and Washington counties, Arkansas. Phase 1.
Report, Ozark Underground Laboratory, Protem, Missouri. 35 pp.
Armstrong, J.G. 1969. A study of the spring dwelling fishes of the southern bend of
the Tennessee River. M.Sc. Thesis, University of Alabama, Tuscaloosa, Alabama.
60 pp.
Armstrong, J.G. and J.D. Williams. 1971. Cave and spring fishes of the southern
bend of the Tennessee River. Journal of the Tennessee Academy of Science 46: 107-
115.
Aumiller, S.R. and D.B. Noltie. 2003. Chemoreceptive responses of the southern
cavefish Typhlichthys subterraneus Girard, 1860 (Pisces, Amblyopsidae) to
conspecifics and prey. Subterranean Biology 1: 79-92.
Barr, T.C. and J. Holsinger. 1985. Speciation in cave faunas. Annual Review of Ecology
and Systematics 16: 313-337.
Bailey, V. 1933. Cave life of Kentucky, mainly in the Mammoth Cave region.
American Midland Naturalist 14: 385-635.
Balon, E.K. 1975. Reproductive guilds of fishes: A proposal and definition. Journal
of the Fisheries Research Board of Canada 32: 821-864.
Bechler, D.L. 1976. Typhlichthys subterraneus Girard (Pisces: Amblyopsidae) in the
Jackson Plain of Tennessee. National Speleological Society Bulletin 38: 39-40.
Bechler, D.L. 1980. The evolution of agonistic behavior in Amblyopsid fishes. Ph.D.
dissertation. Saint Louis University, Saint Louis, Missouri. 160 pp.
Bechler, D.L. 1981. Agonistic behavior in the Amblyopsidae, the spring, cave, and
swamp fishes. Proceedings of the International Congress of Speleology 8: 68-69.
Bechler, D.L. 1983. The evolution of agonistic behavior in Amblyopsid fishes.
Bergstrom, D.E. 1997. The phylogeny and historical biogeography of Missouri’s
Amblyopsis rosae (Ozark cavefish) and Typhlichthys subterraneus (southern
cavefish). M.S. Thesis. University of Missouri, Columbia, Missouri.

Bergstrom, D.E., D.B. Noltie and T.P. Holtsford. 1995. Ozark cavefish genetics:
The phylogeny of Missouri’s Ozark cavefish (Amblyopsis rosae) and southern
cavefish (Typhlichthys subterraneus). Final report to Missouri Department of
Conservation.
Blatchley, W.S. 1897. Indiana caves and their fauna. Report of the Indiana State
Geologist for 1896, pp. 120-212.
Blaxter, J.H.S. 1965. The feeding of herring larvae and their ecology in relation to
feeding. California Cooperative Oceanic Fisheries Investigations, Reports 10: 79-88.
Bonaparte, C.L. 1846. Catalogo metodico dei pesci europei. Atti Scienz. Ital. Settima
Adunanza (1845): 1-95.
Boschung, H.T. and R.L. Mayden. 2004. Fishes of Alabama. Smithsonian Institution
Press, Washington.
Branson, B.A. 1991. The Mammoth Cave blindfish. Tropical Fish Hobbyist 40: 35,
39-40.
Breder, C.M. and D.R. Rosen. 1966. Modes of Reproduction in Fishes. Natural History
Press, New York.
Brown, A.V. and L.D. Willis. 1984. Cavefish (Amblyopsis rosae) in Arkansas:
Populations, incidence, habitat requirements and mortality factors. Final
Report, Federal Aid Project E-1-6, Arkansas Game and Fish Commission, Little
Rock, Arkansas. 61 pp.
Brown, A.V. and C.S. Todd. 1987. Status review of the threatened Ozark cavefish
(Amblyopsis rosae). Proceedings of the Arkansas Academy of Science 41: 99-100.
Brown, A.V., G.O. Graening and P. Vendrell. 1998. Monitoring cavefish population
and environmental quality in Cave Springs Cave, Arkansas. Final Report to
Arkansas Game and Fish Commission, Little Rock, Arkansas.
Brown, J.Z. 1996. Population dynamics and growth of Ozark cavefish in Logan Cave
National Wildlife Refuge, Benton County, Arkansas. M.S. Thesis, University of
Arkansas, Fayetteville, Arkansas. 105 pp.
Brown, J.Z. and J.E. Johnson. 2001. Population biology and growth of Ozark
cavefishes in Logan Cave National Wildlife Refuge, Arkansas. Environmental
Burr, B.M. and M.L. Warren. 1986. A distributional atlas of Kentucky fishes.
Kentucky Nature Preserves Commission Scientific and Technical Series
4. 399 pp.
Charlton, H.H. 1933. The optic tectum and its related fiber tracts in blind fishes. A.
Troglichthys rosae and Typhlichthys eigenmanni. Journal of Comparative Neurology
57: 285-325.
Clay, W.M. 1975. The Fishes of Kentucky. Kentucky Department of Fish and Wildlife
Resources, Frankfurt, Kentucky.
Cooper, J.E. and D.P. Beiter. 1972. The southern cavefish, Typhlichthys subterraneus
(Pisces, Amblyopsidae), in the eastern Mississippian Plateau of Kentucky.
Copeia 1972: 879-881.
Cooper, J.E. and R.A. Kuehne. 1974. Speoplatyrhinus poulsoni, a new genus and
species of subterranean fish from Alabama. Copeia 1974: 486-493.

Cooper, J.E. and F.C. Rohde. 1980. Chologaster cornuta Agassiz. In: Atlas of North
American Freshwater Fishes, D.S. Lee, C.R. Gilbert, C.H. Hocutt, R.E. Jenkins,
D.E. McAllister and J.R. Stauffer Jr. (eds.), North Carolina State Museum of
Natural History, Raleigh, North Carolina, p. 481.
Cope, E.D. 1871. Life in the Wyandotte Cave. The Annals and Magazine of Natural
History 8: 368-370.
Cope, E.D. 1872. On the Wyandotte Cave and its fauna. American Naturalist 6: 109-
116.
Cox, U.O. 1905. A revision of the cave fishes of North America. Appendix to the
Report of the Commission for Fisheries for 1904: 377-393.
Crunkilton, R. 1985. Subterranean contamination of Maramec Spring by
ammonium nitrate and urea fertilizer and its implication on rare cave biota.
Missouri Speleologist 25: 151-158.
Cambridge.
Culver, D.C. 1986. Cave faunas. In: Conservation Biology: The Science of Scarcity and
Diversity, M.E. Soule (ed.). Sinauer Associates, Sunderland, pp. 427-443.
Culver, D.C., T.C. Kane and D.W. Fong. 1995. Adaptation and Natural Selection in
Caves: The Evolution of Gammarus minus. Harvard University Press, Cambridge
and London.
Davidson, R. 1840. An excursion to the Mammoth Cave and the Barrens of Kentucky,
with some notices of the early settlement of the state. Published in the same
year by both A.T. Skillman and Son, Lexington and Thomas Cowperthwait and
Co., Philadelphia.
Dearolf, K. 1956. Survey of North American cave vertebrates. Proceedings of the
Pennsylvania Academy of Science 30: 201-210.
de Rageot, R. 1992. Notes on the swampfish (Chologaster cornuta Agassiz) in the
Dismal Swamp of Virginia. Banisteria 1: 17-18.
DeKay, J.E. 1842. Zoology of New York or the New York fauna; comprising
detailed descriptions of all the animals hitherto observed within the state of
New York, with brief notices of those occasionally found near its borders, and
accompanied by appropriate illustrations. Part IV, Fishes. Printed by W. and A.
White and J. Visscher, Albany.
Eigenmann, C.H. 1897. The Amblyopsidae, the blind fish of America. Report of the
British Association for the Advancement of Science 1897: 685-686.
Eigenmann, C.H. 1898. A new blind fish. Proceedings of the Indiana Academy of
Sciences 1897: 231.
Eigenmann, C.H. 1899a. A case of convergence. Science 9: 280-282.
Eigenmann, C.H. 1899b. The blind fishes of North America. Popular Science Monthly
56: 473-486.
Eigenmann, C.H. 1899c. Blind Fish. Biological Lectures at The Marine Biological
Laboratory, Woods Hole. Athenaeum Press, Boston.
Eigenmann, C.H. 1899d. The eyes of the blind vertebrates of North America. I. The
eyes of the Amblyopsidae. Archiv. fiir. Entwickelungsmechanik 8: 545-617.

Eigenmann, C.H. 1900. The blind-fishes. Biological Lectures from the Marine Biological
Laboratory of Woods Hole 2: 113-126.
Eigenmann, C.H. 1905. Divergence and convergence in fishes. Biological Lectures
from the Marine Biological Laboratory of Woods Hole 8: 59-66.
Eigenmann, C.H. 1909. Cave Vertebrates of America. A Study in Degenerative Evolution.
Carnegie Institution of Washington, Washington, D.C.
Elliott, W.R. 2000. Conservation of North American cave and karst biota. In:
Etnier, D.A. and W.C. Starnes. 1993. The Fishes of Tennessee. The University of
Tennessee Press, Knoxville, Tennessee.
Evans, M.H. 1940. Brain and Body of Fish. Technical Press, London.
Fine, M.L., M.H. Horn and B. Cox. 1987. Acanthonus armatus, a deep-sea teleost
fish with a minute brain and large ears. Proceedings Royal Society of London B
230: 257-265.
Flint, J. 1820. Letters from America July 10, 1818 – November 9th 1820. W. and C. Tate,
Edinburgh (Reprinted in 1822).
Forbes, S.A. 1882. The blind cave fishes and their allies. American Naturalist 16:
1-5.
Forbes, S.A. and R.E. Richardson. 1908. The Fishes of Illinois. Illinois State Laboratory
of Natural History, Urbana, Illinois.
Fournie, J.W. and R.M. Overstreet. 1985. Retinoblastoma in the spring cavefish,
Chologaster agassizi Putnam. Journal of Fish Diseases 8: 377-381.
Fuller, T.C. 1958. Rare, blind, white cave fish are found in Kentucky cave by
“amateur spelunkers”. Kentucky Happy Hunting Ground 14: 16-17.
Garman, S. 1889. Cave animals from southwestern Missouri. Bulletin of the Harvard
Museum of Comparative Zoology 17: 225-240.
Georgia Natural Heritage Program. 2007. Tracking list of special concern animals
of Georgia. Georgia Natural Heritage Program, Georgia Department of Natural
Resources, Atlanta, 11 pp.
Giovannoli, L. 1933. Invertebrate life of Mammoth and other neighboring caves.
Girard, C.F. 1859. Ichthyological notices. Proceedings of the Academy of Natural
Sciences of Philadelphia 1859: 56-68.
Green, S.M. and A. Romero. 1997. Responses to light in two blind cave fishes
(Amblyopsis spelaea and Typhlichthys subterraneus) (Pisces: Amblyopsidae).
Greenwood, P.H., D.E. Rosen, S.H. Weitzman and G.S. Myers. 1966. Phyletic
studies of teleostean fishes, with a provisional classification of living forms.
Bulletin of the American Museum of Natural History 131: 339-456.
Gunning, G.E. and W.M. Lewis. 1955. The fish population of a spring-fed swamp
in the Mississippi bottoms of southern Illinois. Ecology 36: 552-558.
Hempel, G. 1965. On the importance of larval survival for the population dynamics
of marine food fish. California Cooperative Oceanic Fisheries Investigations Reports.
10: 13-23.

Hill, L.G. 1966. Studies on the biology of the spring cavefish, Chologaster agassizi
Putnam. PhD Dissertation, Department of Biology, University of Louisville.
104 pp.
Hill, L.G. 1968. Oxygen preference in the spring cavefish, Chologaster agassizi.
American Fisheries Society Transaction 97: 448-454.
Hill, L.G. 1969a. Feeding and food habits of the spring cavefish, Chologaster agassizi.
Hill, L.G. 1969b. Distributional and population isolation of the spring cavefish,
Chologaster agassizi (Pisces: Amblyopsidae). Proceedings of the Oklahoma Academy
of Science 48: 32-36.
Hill, L.G. 1971. Scale development and patterns of squamation on the spring
cavefish, Chologaster agassizi (Amblyopsidae). Proceedings of the Oklahoma
Academy of Science 51: 13-14.
Holsinger, J.R. 2000. Ecological derivation, colonization, and speciation. In:
Hubbs, C.L. 1938. Fishes from the caves of Yucatan. Carnegie Institution of Washington
Publication 491: 261-295.
Hüppop, K. 2000. How do cave animals cope with the food scarcity in caves? In:
Ecosystems of the World, Volume 30, Subterranean Ecosystems. H. Wilkens, D.C.
Hilton-Taylor, C. 2000. 2000 IUCN Red List of Threatened Species. IUCN, Gland,
Switzerland and Cambridge, United Kingdom.
Hunter, J.R. 1972. Swimming and feeding behavior of larval anchovy Engraulis
mordax. Fishery Bulletin 70: 821-838.
Hutchings, J.A. 1991. Fitness consequences of variation in egg size and food
abundance in brook trout Salvelinus fontinalis. Evolution 45: 1162-1168.
Illinois Endangered Species Protection Board. 2006. Endangered and threatened
species list. Illinois Endangered Species Protection Board, Springfield, Illinois.
Available online at http://dnr.state.il.us/espb/datelist.htm.
Indiana Department of Natural Resources. 2007. Indiana’s species of greatest
conservation need. Division of Fish and Wildlife, Indiana Department of
Natural Resources, Indianapolis, Indiana. 2 pp.
IUCN 1996. 1996 IUCN Red List of Threatened Animals. IUCN, Gland and
Cambridge.
IUCN 2000. 2000 IUCN Red List of Threatened Species. IUCN, Gland and
Cambridge.
Jegla, T.C. and T.L. Poulson. 1970. Circadian rhythms, I: Reproduction in cave
crayfish Orconectes pellucidus inermis. Comparative Biochemical Physiology 33: 347-
355.
Jenio, F. 1980. Predation of freshwater gammarids (Crustacea: Amphipoda).
Proceedings of the West Virginia Academy of Science 51: 67-73.
Jenkins, R.E. and N.M. Burkhead. 1994. Freshwater Fishes of Virginia. American
Fisheries Society, Bethesda, Maryland, 1079 pp.

Jones, G.M. and K.E. Spells. 1963. A theoretical and comparative study of the
functional dependence of the semicircular canal upon its physical dimensions.
Proceedings Royal Society of London B 157: 403-419.
Jones, S.R. and C.A. Taber. 1985. A range revision for western populations of the
southern cavefish Typhlichthys subterraneus (Amblyopsidae). American Midland
Jordan, D.S. 1889. Descriptions of fourteen species of fresh-water fishes collected
by the U.S. Fish Commission in the summer of 1888. Proceedings of the United
States National Museum 11: 351-362.
Jordan, D.S. 1929. Forbesichthys for Forbesella. Science 70: 68.
Jordan, D.S. and B.W. Evermann. 1927. New genera and species of North American
fishes. Proceedings of the California Academy of Science 16: 501-507.
Kalayil, P.K. and W.M. Clay. 1976. Immunological characteristics and relationships
of tissue antigens in Amblyopsid fishes. Federation Proceedings 35: 751.
Keith, J.H. 1988. Distribution of northern cavefish, Amblyopsis spelaea DeKay, in
Indiana and Kentucky and recommendations for its protection. Natural Areas
Journal 8: 69-79.
Keith, J.H. and L.M. Gray. 1979. A preliminary study of the occurrence of broken-
back syndrome in the northern cavefish (Amblyopsis spelaea) at Spring Mill State
Park, Mitchell, Indiana. Proceedings of the Indiana Academy of Sciences 88: 163.
Keith, J.H. and T.L. Poulson. 1981. Broken-back syndrome in Amblyopsis spelaea,
Donaldson-Twin Caves, Indiana. Cave Research Foundation 1979 Annual
Report: 45-48.
Kentucky State Nature Preserves Commission. 2005. Rare and extirpated biota of
Kentucky. Kentucky State Nature Preserves Commission, Frankfurt, Kentucky.
19 pp.
Kuhajda, B.R. 2004. The impact of the proposed Eddie Frost Commerce Park on
Speoplatyrhinus poulsoni, the Alabama cavefish, a federally endangered species
restricted to Key Cave, Lauderdale County, Alabama. Endangered Species Update
21: 57.
Kuhajda, B.R. and R.L. Mayden. 2001. Status of the federally endangered Alabama
cavefish, Speoplatyrhinus poulsoni (Amblyopsidae), in Key Cave and surrounding
caves, Alabama. Environmental Biology of Fishes 62: 215-222.
Langecker, T.G. 2000. The effects of continuous darkness on cave ecology and
cavernicolous evolution. In: Ecosystems of the World, Volume 30, Subterranean
Ecosystems, H. Wilkens, D.C. Culver and W.F. Humphreys (eds.), Elsevier,
Amsterdam, pp. 135-158.
Layne, J.N. and D.H. Thompson. 1952. Recent collections of the Amblyopsid fish
Chologaster papillifera in Illinois. Copeia 1952: 39-40.
Lewis, J.J. 1996. The devastation and recovery of caves affected by industrialization.
Proceedings of the 1995 National Cave Management Symposium: 214-227.
Lewis, J.J. 1998. The subterranean fauna of the Blue River area. Final Report, The
Natural Conservancy, 266 pp.

Lewis, J.J. 2002a. Conservation assessment for northern cavefish copepod

(Cauloxenus stygius). USDA Forest Service, Eastern Region, Report. 10 pp.
Lewis, J.J. 2002b. Conservation assessment for southern cavefish (Typhlichthys
subterraneus). USDA Forest Service, Eastern Region, Report. 9 pp.
Lewis, J.J. and T. Sollman. 1998. Groundwater monitoring in significant aquatic
caves that lie beneath impending residential developments in the Blue River
basin of southern Indiana. Final Report, U.S. Fish and Wildlife Service. 89 pp.
Lisowski, E.A. and T.L. Poulson. 1981. Impacts of Lock and Dam Six on base level
ecosystems in Mammoth Cave. Cave Research Foundation 1981 Annual Report:
48-54.
Mayden, R.L. and F.B. Cross. 1983. Reevaluation of Oklahoma records of
the southern cavefish, Typhlichthys subterraneus Girard (Amblyopsidae).
Southwestern Naturalist 28: 471-473.
McAllister, D.E. 1968. The evolution of branchiostegals and associated opercular,
gular, and hyoid bones and the classification of teleostome fishes, living and
fossil. Bulletin of the National Museum of Canada 221: 1-239.
McDonald, E.F. and W.L. Pflieger. 1979. The spring cavefish, Chologaster agassizi
(Pisces: Amblyopsidae), in southeastern Missouri. American Midland Naturalist
102: 194-196.
McNulty, J.A. 1978a. The pineal organ of the troglophilic fish, Chologaster agassizi:
An ultrastructural study. Journal of Neural Transmission 43: 47-71.
McNulty, J.A. 1978b. Fine structure of the pineal organ in the troglobytic fish,
Typhlichthys subterraneus (Pisces: Amblyopsidae). Cell and Tissue Research 195:
535-545.
Means, M.L. and J.E. Johnson. 1995. Movement of threatened Ozark cavefish in
Logan Cave National Wildlife Refuge, Arkansas. Southwestern Naturalist 40:
308–313.
Missouri Natural Heritage Program. 2008. Missouri species and communities of
conservation concern checklist. Missouri Department of Conservation, Jefferson
City, Missouri, 55 pp.
Mohr, C.E. and T.L. Poulson. 1966. The Life of the Cave. McGraw-Hill, New York.
Montgomery, J.C., S. Coombs and C.F. Baker. 2001. The mechanosensory lateral
line system of the hypogean form of Astyanax mexicanus. Environmental Biology
of Fishes 62: 87-96.
Murray, A.M. and M.V.H. Wilson. 1999. Contributions of fossils to the phylogenetic
relationships of the percopsiform fishes (Teleostei: Paracanthopterygii): Order
restored. In: Mesozoic Fishes 2—Systematics and Fossil Record. G. Arratia and H.-P.
Schultz (eds.), Verlag Dr. Friedrich Pfeil, Munich, Germany, pp. 397-411.
Nelson, J.S. 1994. Fishes of the World. Third Edition. John Wiley Sons, New York.
Nelson, J.S. 2006. Fishes of the World. Fourth Edition. John Wiley Sons, New York.
Nickol, B.B. and F.B. Whittaker. 1978. Neoechinorhynchus cylindratus
(Acanthocephala) from the troglodytic fish, Amblyopsis spelaea, in Kentucky.
Proceedings of the Helminthological Society of Washington 45: 136-137.

Niemiller, M.L. and B.M. Fitzpatrick. 2008. Phylogenetics of the southern cavefish,
Typhlichthys subterraneus: Implications for conservation and management.
Proceedings of the 2007 National Cave and Karst Management Symposium (In
Press).
Noltie, D.B. and C.B. Wicks. 2001. How hydrogeology has shaped the ecology of
subterraneus: Insights on the sightless from understanding the underground.
Oklahoma Department of Wildlife Conservation. 2008. Oklahoma’s endangered
and threatened species and species of special concern. Natural Resources
Section, Oklahoma Department of Wildlife Conservation, Oklahoma City.
Available online at http://www.wildlifedepartment.com/endanger.htm.
Page, L.M. and B.M. Burr. 1991. A Field Guide to Freshwater Fishes of North America
North of Mexico. Houghton Mifflin, Boston.
Parenti, L.R. 2006. Typhlichthys eigenmanni Charlton, 1933, an available name
for a blind cavefish (Teleostei: Amblyopsidae), differentiated on the basis of
characters of the central nervous system. Zootaxa 1374: 55-59.
Parzefall, J. 2000. Ecological role of aggressiveness in the dark. In: Ecosystems of the
World, Volume 30, Subterranean Ecosystems. H. Wilkens, D.C. Culver and W.F.
Humphreys (eds.), Elsevier, Amsterdam, pp. 221-228.
Patterson, C. 1981. The development of the North American fish fauna—a problem
of historical biogeography. In: The Evolving Biosphere. P.L. Forey (ed.), Cambridge
University Press, Cambridge, pp. 265-281.
Patterson, C. and D.E. Rosen. 1989. The Paracanthopterygii revisited: Order and
disorder. In: Papers on the Systematics of Gadiform Fishes. D.M. Cohen (ed.),
Science Seriees, Natural History Museum Los Angeles Co., No. 32, pp. 5-36.
Payne, R. 1907. The reactions of the blind fish, Amblyopsis spelaeus, to light. Biological
Bulletin 13: 317-323.
Pearson, W.D. and C.H. Boston. 1995. Distribution and status of the northern
cavefish, Amblyopsis spelaea. Final report, Nongame and Endangered Wildlife
Program, Indiana Department of Natural Resources, Indianapolis.
Poly, W.J. and C.E. Boucher. 1996. Nontroglobitic fishes in caves: Their
abnormalities, ecological classification and importance. American Midland
Poly, W.J. and G.S. Proudlove. 2004. Family Amblyopsidae Bonaparte
1846—cavefishes. California Academy of Sciences Annotated Checklists of Fishes 26.
7 pp.
Poulson, T.L. 1960. Cave adaptation in Amblyopsid fishes. PhD Dissertation,
Department of Zoology, University of Michigan, Ann Arbor. University
Microfilms 61-2787.
Poulson, T.L. 1967. Comparison of terrestrial cave communities. Cave Research
Foundation Annual Report 1967: 39.

Poulson, T.L. 1968. Aquatic cave communities. Cave Research Foundation Annual
Report 1968: 16-18.
Poulson, T.L. 1969. Population size, density, and regulation in cave fishes. Actes of
the 4th International Congress of Speleology, Ljubljana, Yugoslavia 4-5: 189-192.
arguments and data from Amblyopsid fishes and Linyphiid spiders. Bulletin of
Poulson, T.L. 1992. The Mammoth Cave ecosystem. In: The Natural History of
Biospeleology. A. Camacho (ed.), Museo Nacional de Ciencias Naturales, Madrid,
pp. 569-611.
Poulson, T.L. 1996. Research aimed at management problems should be hypothesis-
driven: Case studies in the Mammoth Cave Region. In: Proceedings of the 1995
National Cave Management Symposium, Spring Mill State Park, Mitchell, Indiana,
G.T. Rea (ed.). Indiana Karst Conservancy, Indianapolis, Indiana, pp. 267-273.
Poulson, T.L. 2001a. Morphological and physiological correlates of evolutionary
reduction of metabolic rate among Amblyopsid cave fishes. Environmental
Poulson, T.L. 2001b. Adaptations of cave fishes with some comparisons to deep-
Poulson, T.L. 2005. Food sources. In: Encyclopedia of Caves, D.C. Culver and W.B.
White (eds.), Elsevier, Amsterdam, pp. 255-263.
Poulson, T.L. 2009a. New studies of Speoplatyrhinus poulsoni (Pisces: Amblyopsidae).
Proceedings of the 15th International Congress of Speleology 3: 1337-1342.
Poulson, T.L. 2009b. Field metrics for cave stream bio-integrity. Proceedings of the
15th International Congress of Speleology 3: 1343-1348.
Poulson, T.L. and P.M. Smith. 1969. The basis for seasonal growth and reproduction
in aquatic cave organisms. Actes of the 4th International Congress of Speleology,
Ljubljana, Yugoslavia 4-5: 197-201.
Poulson, T.L. and T.C. Jegla. 1969. Circadian rhythms in cave animals. Actes of the
4th International Congress of Speleology, Ljubljana, Yugoslavia 4-5: 193-195.
Poulson, T.L. and K.H. Lavoie. 2000. The trophic basis of subterranean ecosystems.
In: Ecosystems of the World, Volume 30, Subterranean Ecosystems. H. Wilkens, D.C.
Protas, M.E., C. Hersey, D. Kochanek, Y. Zhou, H. Wilkens, W.R. Jeffery, L.I. Zon, R.
Borowsky and C.J. Tabin. 2006. Genetic analysis of cavefish reveals molecular
Proudlove, G.S. 2001. The conservation status of hypogean fishes. Environmental
Proudlove, G.S. 2006. Subterranean Fishes of the World. International Society for
Subterranean Biology, Moulis.
Putnam, F.W. 1872. The blind fishes of the Mammoth Cave and their allies. American
Naturalist 6: 6-30.

Rohde, F.C., R.G. Arndt, D.G. Lindquist and J.F. Parnell. 1994. Freshwater Fishes
of the Carolinas, Virginia, Maryland, and Delaware. University of North Carolina
Press, Chapel Hill, North Carolina.
Romero, A. 1998a. Threatened fishes of the world: Amblyopsis rosae (Eigenmann,
1898) (Amblyopsidae). Environmental Biology of Fishes 52: 434.
Romero, A. 1998b. Threatened fishes of the world: Typhlichthys subterraneus Girard,
1860 (Amblyopsidae). Environmental Biology of Fishes 53: 74.
Romero, A. 1998c. Threatened fishes of the world: Speoplatyrhinus poulsoni Cooper
and Kuehne, 1974 (Amblyopsidae). Environmental Biology of Fishes 53: 293-94.
Romero, A. 2001. Scientists prefer them blind: The history of hypogean fish
research. Environmental Biology of Fishes 62: 43-71.
Romero, A. 2002. Between the first blind cave fish and the last of the Mohicans: The
scientific romanticism of James E. DeKay. Journal of Spelean History 36: 19-29.
Romero, A. and L. Bennis. 1998. Threatened fishes of the world: Amblyopsis spelaea
DeKay, 1842 (Amblyopsidae). Environmental Biology of Fishes 51: 420.
Romero, A. and M. Conner. 2007. Status report for the southern cavefish,
Typhlichthys subterraneus, in Arkansas. Final report to the Arkansas Game and
Fish Commission, Little Rock, Arkansas.
Rosen, D.E. 1962. Comments on the relationships of the North American cave
fishes of the family Amblyopsidae. American Museum Novitates 2109: 1-35.
Ross, S.W. and F.C. Rohde. 2003. Life history of the swampfish from a North
Carolina stream. Southeastern Naturalist 2: 105-120.
Schubert, A.L.S., C.D. Nielsen and D.B. Noltie. 1993. Habitat use and gas bubble
disease in southern cavefish (Typhlichthys subterraneus). International Journal of
Speleology 22: 131-143.
Scott, W. 1909. An ecological study of the plankton of Shawnee Cave. Biological
Bulletin 17: 386-402.
Shute, J.R., P.W. Shute and D.G. Lindquist. 1981. Fishes of the Waccamaw River
Drainage. Brimleyana 6: 1-24.
Smith, P.W. 1979. The Fishes of Illinois. University of Illinois Press, Urbana, Illinios.
Smith, P.W. and N.M. Welch. 1978. A summary of the life history and distribution
of the spring cavefish, Chologaster agassizi, Putnam, with population estimates
for the species in southern Illinois. Illinois Natural History Survey Biological Notes
104: 1-8.
Smith, W.L. and W.C. Wheeler. 2006. Venom evolution widespread in fishes: A
road map for the bioprospecting of piscine venoms. Journal of Heredity 97: 206-
217.
Storer, D.H. 1846. Synopsis of the fishes of North America. Memoirs of the American
Academy 2: 436.
Swofford, D.L. 1982. Genetic variability, population differentiation, and biochemical
relationships in the family Amblyopsidae. M.S. Thesis. Eastern Kentucky
University, Richmond, Kentucky.
Swofford, D.L., B.A. Branson and G. Sievert. 1980. Genetic differentiation of
cavefish populations (Amblyopsidae). Isozyme Bulletin 13: 109-110.

Tellkampf, T.G. 1844. Ueber den blinden Fisch der Mammuth-Hohle in Kentucky,
mit Bemerkungen ueber einige undere in dieser Hohle lebenden Thiere. Archiv
fur Anatomie, Physiologie und wissenschaftliche Medicin 4: 381-394.
Tellkampf, T.G. 1845. Memoirs on the blind-fishes and some other animals living
in the Mammoth Cave in Kentucky. New York Journal of Medical and Collateral
Sciences 5: 84-93.
Tercafs, R. 1992. The protection of the subterranean environment. Conservation
principles and management tools. In: The Natural History of Biospeleology,
A.I. Camacho (ed.), Monografias del Museo Nacional de Ciencias Naturales,
Madrid, pp. 481-524.
Trajano, E. 2007. The challenge of estimating the age of subterranean lineages:
examples from Brazil. Acta Carsologica 36: 191-198.
United States Fish and Wildlife Service (USFWS). 1977. Final threatened status
and critical habitat for five species of southeastern fishes. Federal Register 42:
45526-45530.
United States Fish and Wildlife Service (USFWS). 1982. Recovery plan for the
Alabama cavefish, Speoplatyrhinus poulsoni Cooper and Kuehne, 1974. U.S.
Department of Interior, U.S. Fish and Wildlife Service, Atlanta, Georgia.
United States Fish and Wildlife Service (USFWS). 1985. Revised recovery plan for
the Alabama cavefish, Speoplatyrhinus poulsoni Cooper and Kuehne, 1974. U.S.
United States Fish and Wildlife Service (USFWS). 1989. Ozark cavefish recovery
plan. U.S. Department of Interior, U.S. Fish and Wildlife Service, Atlanta,
Georgia.
United States Fish and Wildlife Service (USFWS). 1990. Alabama cavefish,
Speoplatyrhinus poulsoni Cooper and Kuehne, 1974 (second revision) recovery
plan. U.S. Department of Interior, U.S. Fish and Wildlife Service, Atlanta,
Georgia.
United States Fish and Wildlife Service (USFWS). 1996a. Proposed establishment
of Key Cave National Wildlife Refuge, Lauderdale County, Alabama. U.S.
United States Fish and Wildlife Service (USFWS). 1996b. Region 3 decision on
the status recommendation for the northern cavefish (Amblyopsis spelaea). U.S.
Department of Interior, U.S. Fish and Wildlife Service, Bloomington, Indiana.
Vandike, J.E. 1981. The effects of the November 1981 liquid fertilizer pipeline break
on groundwater in Phelps County, Missouri. Report for Water Resources Data
and Research, Missouri Department of Natural Resources, Division of Geology
and Land Survey, Rolla. 24 pp.
Vandike, J.E. 1984. Hydrogeologic aspects of the November 1981 liquid fertilizer
pipeline break on groundwater in the Meramec Spring recharge area, Phelps
County, Missouri. Proceedings of the 1984 National Cave Management Symposium,
Missouri Speleology 25: 93-101.

Verrier, M.L. 1929. Observations sur le comportement d’un poisson cavernicole

Typhlichthys osborni Eigenmann. Bulletin of the National Museum of Natural
History Paris (Series 2) 1: 82-84.
Weber, A. 2000. Fish and amphibia. In: Ecosystems of the World, Volume 30,
Weise, J.G. 1957. The spring cave-fish, Chologaster papilliferus, in Illinois. Ecology
38: 195-204.
Whittaker, F.H. and L.G. Hill. 1968. Proteocephalus chologaster sp. n. (Cestoda:
Proteocephalidae) from the spring cavefish Chologaster agassizi Putnam, 1782
[error = Putnam 1872] (Pisces: Amblyopsidae) of Kentucky. Proceedings of the
Helminthological Society of Washington 35: 15-18.
Whittaker, F.H. and S.J. Zober. 1978. Proteocephalus poulsoni sp. n. (Cestoda:
Proteocephalidae) from the northern cavefish Amblyopsis spelaea DeKay, 1842
(Pisces: Amblyopsidae) of Kentucky. Folia Parasitologica 25: 277-280.
Wiley, E.O., G.D. Johnson and W.W. Dimmick. 2000. The interrelationships
of acanthomorph fishes: A total evidence approach using molecular and
morphological data. Biochemical Systematics and Ecology 28: 319-350.
Wilkens, H. 1988. Evolution and genetics of epigean and cave Astyanax mexicanus
(Characidae, Pisces): Support for the neutral mutation theory. Evolutionary
Biology 23: 271-367.
Willis, L.D. and A.V. Brown. 1985. Distribution and habitat requirements of the
Ozark cavefish, Amblyopsis rosae. American Midland Naturalist 114: 311-317.
Withers, D.I., K. Condict and R. McCoy. 2004. A guide to the rare animals of
Tennessee. Division of Natural Heritage, Tennessee Department of Environment
and Conservation, Nashville, Tennessee. 16 pp.
Woods, L.P. and R.F. Inger. 1957. The cave, spring, and swamp fishes of the family
Amblyopsidae of central and eastern United States. American Midland Naturalist
58: 232-256.
Wyman, J. 1843a. Description of a blind-fish from a cave in Kentucky. American
Journal of Science and Arts 45: 94-96.
Wyman, J. 1843b. Description of a blind-fish from a cave in Kentucky. Annals and
Magazine of Natural History 12: 298-299.
Wyman, J. 1851. Account of dissections of the blind fishes (Amblyopsis spelaeus)
from the Mammoth Cave, Kentucky. Proceedings of the Boston Society of Natural
History 3: 349, 375.
Wyman, J. 1854a. The eyes and organs of hearing in Amblyopsis spelaeus. Proceedings
of the Boston Society of Natural History 4: 149-151.
Wyman, J. 1854b. On the eye and the organ of hearing in the blind fishes (Amblyopsis
spelaeus DeKay) of the Mammoth Cave. Proceedings of the Boston Society of Natural
History 4: 395-396.
Wyman, J. 1872. Notes and drawings of the rudimentary eyes, brain and tactile
organs of Amblyopsis spelaeus, in Putnam (1872). American Naturalist 6: 6-30.

CHAPTER 8
Subterranean Fishes of Mexico
(Poecilia mexicana, Poeciliidae)
Martin Plath1 and Michael Tobler 2,3
1
Institute of Ecology, Evolution and Diversity, Department of Ecology and
Evolution, J.W. Goethe University Frankfurt, Siesmayerstrasse 70-72,
D-60054 Frankfurt am Main, Germany
E-mail: mplath@bio.uni-frankfurt.de
2
Texas A&M University, Department of Fish and Wildlife Sciences, 2258
TAMU, College Station, TX 77843-2258 USA
3
University of Zürich, Institute of Zoology, Winterthurerstrasse 190
CH-8057 Zürich, Switzerland
INTRODUCTION
The Atlantic molly (P. mexicana) is a widespread freshwater fish living along
the Atlantic coast of Mexico and Central America (Miller 2005). It inhabits
coastal lagoons, estuaries, lowland ponds and rivers up to highland streams
(Bussing 1998, Miller 2005). Mollies are small fish less than 10 centimeters
long, and belong to the family Poeciliidae (livebearers). Poeciliids give
birth to fully developed young, which are instantly independent. Males
use a modified anal fin, the so-called gonopodium, to transfer sperm
bundles (spermatozeugmata) to the female during copulation (Rosen and
Bailey 1963, Meffe and Snelson 1989).
In the southern Mexican Cueva del Azufre system, P. mexicana has
colonized at least two caves. Cavernicolous P. mexicana have been referred
to as the cave molly (Gordon and Rosen 1962, Parzefall 2001). Cave mollies
are the only known poeciliids that naturally inhabit subterranean habitats
(Proudlove 2006). The only other known poeciliid species occurring in a
cave is Gambusia affinis introduced into the ‘Spunnulate’ dolinas system of
Torre Castiglione in southeastern Italy (Camassa 2001). Various aspects of

the evolutionary ecology of P. mexicana from the Cueva del Azufre system
have been studied thoroughly during the past decades. Today, the cave
molly ranks among the best-studied cavefishes worldwide along with
the cave form(s) of the Mexican tetra Astyanax mexicanus (see chapters 5
and 6 in this volume). This chapter provides a synopsis particularly over
some recent advances in the understanding of the evolutionary ecology of
cavernicolous P. mexicana.
HISTORY AND OVERVIEW

The cave molly has always been tightly connected to the local native
culture. The Cueva del Azufre (also known as Cueva de las Sardinas
and Cueva Villa Luz) is sacred to the Rain God of local Zoque Indians.
Once a year around Easter (towards the end of the dry season), the Zoque
penetrate into the cave, and poison and capture cave mollies with the aid
of rotenone- and deguelin-containing barbasco root (Lonchocarpus sp.,
Fabaceae). The fish are then cooked and eaten. During this act, which is
accompanied by prayers, the dead fish act as messengers to the Rain God,
bringing on the first rains of the rainy season.
Scientific research in the cave and the recognition of the uniqueness
of its inhabitants, commenced in the late 19th century. According to an
unpublished manuscript of the recently deceased American ichthyologist
Robert Rush Miller, the first scientific collections of fish in the Cueva
del Azufre date back to 1896, when A. Dugas sent specimens to the
US National Museum. M.W. Stirling, Chief of the Bureau of American
Ethnology (Smithsonian Institution), and his collaborators researched the
cave and collected cave mollies several times between 1944 and 1948. The
first map of the cave and a scientific description of the cave molly – not as
a distinct species but a divergent population of Poecilia sphenops1 – were
published by Gordon and Rosen in 1962. A second, more detailed map
was later published by Hose and Pisarowicz (1999).
Ever since the 1960s, several research groups have worked in the Cueva
del Azufre, specifically on the fish inhabiting the cave. This research was
spearheaded by a group from the University of Hamburg led by Jakob
Parzefall. Parzefall pioneered maintaining and breeding the cave molly
in captivity and published numerous scientific studies on this fish. He
established the cave molly as a model system to study the evolution of
cave animals, with special emphasis on behavioral adaptations to darkness
(Parzefall 1993, 2001).
1
Later research indicated that, based on the shape of their teeth, the molly populations in and
around the cave are P. mexicana not P. sphenops (Miller, R.R. 1983. Checklist and key to the
mollies of Mexico (Pisces, Poeciliidae, Poecilia, subgenus Mollienesia). Copeia: 817-822.

Martin Plath and Michael Tobler 283
The pioneering work on the cave molly was primarily based on laboratory
studies. In the course of time, this led to a somewhat simplifying view in
two directions: (1) The predominant view of the Cueva del Azufre system
was that it consists of two distinct habitat types, the Cueva del Azufre and
normal surface habitats. (2) The absence of light was assumed to be the
most important selective factor. More recent work led to a redefinition of
the Cueva del Azufre system to amend these views by including additional
habitats in comparative analyses and by recognizing the importance of
hydrogen sulfide (H2S), which is present in some habitats of this system, as
an important selective factor. The discovery of a non-sulfidic cave (Cueva
Luna Azufre) and independent sulfidic springs in surface habitats (El
Azufre) in 2006 allowed extending the comparative approach of studying
the ecology and evolution of P. mexicana under divergent environmental
conditions. The inclusion of a wider array of populations – especially from
adjacent surface habitats – and the recognition of a multitude of both abiotic
and biotic environmental factors as potential sources of natural selection
allowed for a better understanding of the ecological and evolutionary
divergence in cave-inhabiting populations of P. mexicana.
In this chapter, we review the current knowledge on the evolutionary
ecology of P. mexicana in the Cueva del Azufre system. We first compare
various aspects of the abiotic environments in which populations of
P. mexicana occur and discuss divergent traits occurring in the different
populations as well as their potential adaptive significance. We then
examine differences in the biotic environment of the different habitat
types, discuss the consequences for the ecology of different populations,
and highlight again the evolutionary responses of P. mexicana. By the end of
this chapter, we synthesize the current knowledge and scrutinize support
for the idea that parapatric ecological speciation due to local adaptation
to divergent environmental conditions may be occurring in this system.
Throughout, we make an attempt to disentangle the effects of different
selective forces – primarily darkness (i.e., evolutionary trends that may be
shared also with other cave fishes), toxic hydrogen sulfide, and selective
pressures stemming from ecological differences among habitat types – on
the evolution of different populations of P. mexicana. Most information
currently available considers the behavior, ecology and evolution of the
cave molly population from the sulfidic Cueva del Azufre, but in some
cases, fish from the recently discovered Cueva Luna Azufre have already
been included in recent research projects, and – where available – first
results are presented.

AN INTRODUCTION TO THE CUEVA DEL AZUFRE

SYSTEM
The Cueva del Azufre system is located near the village of Tapijulapa in
the state of Tabasco, Mexico (Figure 1). All watercourses in the study area
eventually drain into the Río Oxolotan, which is part of the Río Grijalva
drainage system (Tobler et al. 2006b). The Cueva del Azufre system provides
an unparalleled ‘natural experiment’ with two abiotic environmental
factors – the presence or absence of light and H2S – occurring in a fully
2 × 2 factorial design. All habitat types are connected and located within
a perimeter of roughly 3 km. Most importantly, all of the habitat types are
inhabited by P. mexicana (Tobler et al. 2006b, Tobler et al. 2008a). The four
habitat types typically recognized for comparative analyses include:
• The Cueva del Azufre itself is a sulfidic cave structured into 13
different chambers, the nomenclature of which follows Gordon and
Rosen (1962; Figure 2). The front chambers obtain some dim light
through breaks in the ceiling, whereas the rearmost cave chambers
are completely dark. The cave is drained by a creek fed by a number
of springs throughout the cave, most of which contain high levels
of dissolved H2S (Tobler et al. 2006b, Tobler et al. 2008b).
• The recently discovered Cueva Luna Azufre is – despite its name –
a non-sulfidic cave (Figure 3). The creek in the Cueva Luna Azufre
is also fed by springs, but none of these contain H2S (Tobler et al.
2008a). The Cueva Luna Azufre is substantially smaller than the
Cueva del Azufre. Although the two caves are in close proximity,
they are located within different hills that are separated by a surface
valley.
• The El Azufre is a sulfidic surface habitat. The creek originates in
the hills southwest of the two caves and is fed by multiple sulfidic
as well as non-sulfidic springs. The El Azufre flows through the
valley that separates the two caves. Both caves drain into the El
Azufre, which eventually joins the Río Oxolotan (Figure 1).
• Fish from several non-sulfidic surface habitats in the vicinity of
the divergent habitats are used for comparative analyses (Figure 1).
In earlier studies, primarily fish from the Río Teapao (about 25 km
south of Tapijulapa) were used for comparison with cave mollies
(Parzefall, 2001). Our recent research has concentrated on more
proximate surface populations from the nearest large rivers, namely
the Río Oxolotan and the Río Amatan, as well as some of their
tributaries that are similar in size and structure to the El Azufre
(Tobler et al. 2008b). Poecilia mexicana from non-sulfidic surface
habitats represent the ancestral form that resembles the fish that
must originally have colonized the sulfidic and cave habitats.

Fig. 1 (A) Location of field sites near Tapijulapa, Tabasco, Mexico. Blue indicates
non-sulfidic surface sites (AT Arroyo Tres, AA Arroyo Tacubaya, RA Río Amatan,
AC Arroyo Cristal, RO Río Oxolotan, AB Arroyo Bonita), yellow indicates sulfidic
surface sites (EA I, II El Azufre, sites I and II, red: the sulfidic Cueva del Azufre,
brown: the non-sulfidic Cueva Luna Azufre. The inserted figure shows the location
of the Cueva del Azufre system in Mexico. (B-E) Habitat types inhabited by P.
mexicana in the Cueva del Azufre system. (B) non-sulfidic surface habitat (Arroyo
Bonita), (C) sulfur sources in the springhead area of El Azufre, (D) Cueva del
Azufre. (E) P. mexicana occur at high abundance inside the Cueva del Azufre.
book.

Fig. 2 Simplified map of the Cueva del Azufre. Simplified by Laura Rosales with
permission of Bob Richards from Hose and Pisarowicz (1999); chambers (I-XIII)
approximate location from Gordon and Rosen (1962).
Like in other caves around the world, the Cueva del Azufre and the Cueva
Luna Azufre are characterized by the absence of light and consequently
by the reduction of fluctuations in temperature. H2S, however, although
also present in other cave systems (Sarbu et al. 1996, Macalady et al. 2006),
is less known as environmental factor influencing organisms, and some
biologically relevant aspects of this chemical are thus introduced here
briefly.
H2S as an Environmental Factor

Already in the first description of the Cueva del Azufre, Gordon and Rosen
(1962) reported the presence of toxic hydrogen sulfide (H2S) in the Cueva
del Azufre as a faint odor noticeable throughout the cave, but the effects
of H2S only received marginal attention in early studies on cave mollies,
and previous studies have mainly focused on darkness as the driving
evolutionary force in this system.
Because of its lipid solubility, H2S freely penetrates biological
membranes and readily invades organisms (Reiffenstein et al. 1992). Like
cyanide, sulfide is an inhibitor of the cytochrome c oxidase blocking the
electron transport in aerobic respiration, thereby hampering the function

Fig. 3 Simplified map of the Cueva Luna Azufre Azufre (after Pisarowicz 2005).
of mitochondria and the production of ATP (Lovatt Evans 1967, Nicholls

1975, Petersen 1977, National Research Council 1979). H2S is also able
to modify oxygen transport proteins, such as globins in vertebrates,
with which it forms non-functional sulfhemoglobin and sulfmyoglobin
(Carrico et al. 1978, Park et al. 1986). It further can inhibit about 20 other
enzymes and – at least in mammals – has neurotoxic effects that arrest
the central respiratory drive (Bagarinao 1992, Reiffenstein et al. 1992).
Due to its biochemical effects, H2S is highly toxic for aerobic organisms
even in micromolar amounts (Torrans and Clemens 1982, Bagarinao 1992,
Grieshaber and Völkel 1998). It has been shown to play a role in natural
fish kills (Bagarinao and Lantin-Olaguer 1999, Luther et al. 2004) and to
aggravate mortality due to hypoxia and low pH (Bagarinao and Lantin-
Olaguer 1999). Lower concentrations of H2S may have a chronic toxic
effect (National Research Council 1979).
H2S is highly reactive at room temperature and is spontaneously
oxidized to polysulfides, elemental sulfur, thiosulfate, sulfite, and sulfate
in water (Cline and Richards 1969, Chen and Morris 1972). This reaction is
biologically relevant because the presence of H2S leads to and aggravates

hypoxia in aquatic systems, resulting in a distinct inverse correlation in

the concentrations of H2S and oxygen (Bagarinao 1992).
In a systematic survey of abiotic environmental conditions inside the
Cueva del Azufre and the El Azufre, high concentrations of H2S were found,
reaching more than 300 µM in spring areas (Tobler et al. 2006b, Tobler et al.
2008b). Further away from the spring areas, the concentrations are lower
(10-80 µM/l) because of the oxidation of sulfide and the emission from the
water body to the atmosphere; these concentrations, however, still have to
be considered as highly toxic. H2S concentrations in the Cueva del Azufre
system are high compared to other systems with naturally containing H2S,
such as deep sea hydrothermal vents (Van Dover 2000, Nybakken 2001).
Sulfide concentrations in the Cueva del Azufre system seem to vary to
some extent over time and are likely dependent on the discharge of the
springs relative to the precipitation in the area (Tobler et al. 2006b). Some
sulfidic surface springs, for example, only carry water during the rainy
season (authors, personal observation).
Although the water in sulfidic habitats of the Cueva del Azufre system
still contains low concentrations of oxygen, the conditions are considered
hypoxic. This is different from systems like swamps and tidal pools,
in which H2S originates biogenically from sulfate reduction of bacteria
(Jorgensen and Fenchel 1974, Jorgensen 1982, 1984). Here, anoxic conditions
are a prerequisite for the production of H2S.
DIVERGENCE OF P. MEXICANA ALONG ABIOTIC

GRADIENTS
Genetic Differentiation Among Habitats

All habitat types in the Cueva del Azufre system are physically connected.
In theory, P. mexicana should be able to move freely within and between
habitat types. Molecular genetic analyses, using microsatellites as well
as cytochrome b sequences, gave insights not only into the phylogenetic
relationship among mollies from the different habitat types, but also on
migration patterns and genetic differentiation due to differences in abiotic
conditions.
Cytochrome b sequence data in P. mexicana from the Cueva del Azufre
system suggests that fish from the different habitat types are closely
related to each other (Tobler et al. 2008b). There is extensive haplotype-
sharing among habitat types and private alleles have only diverged by a
few point mutations among habitat types. The current data suggest that
the divergent habitats (including the sulfidic surface habitats and the

two caves) were originally colonized by fish from adjacent non-sulfidic

surface habitats. Both caves were likely colonized independently from the
El Azufre.
Likewise, population genetic analyses based on nuclear markers (10
unlinked microsatellite loci) indicated that P. mexicana in different habitat
types are genetically distinct from one another and that gene-flow between
habitats is low despite their spatial proximity (Plath et al. 2007a, Tobler et al.
2008b). Thus, neither are P. mexicana populations in the divergent habitats
sink populations that persist only through continuous immigration of
individuals from normal epigean habitats, nor are P. mexicana in the Cueva
del Azufre system panmictic. The current data suggest that each habitat
type harbors a genetically distinct population (Figure 4). Thus, P. mexicana
did not only diverge into cave and surface populations, but the presence of
H2S also seems to restrict gene flow. Genetically detected migration events
were more common between sites of the same habitat type than between
sites differing in at least one of the two abiotic environmental parameters
(Figure 5).
0.14
Luna Azufre
coordinate 2
0.07 Non-sulfidic surface sites

Cueva del Azufre AB AC
X AT
0.00
XIII EA I RA RO
V
XI
El Azufre
–0.07 EA II
–0.10 –0.05 0.00 0.05 0.10

coordinate 1
Fig. 4 Genetic differentiation in the Cueva del Azufre system based on

microsatellite data (10 unlinked loci). The principal component analysis is based
on Fst values among populations (GenAlEx). Note the gap between the “sulfur
plateau” cluster (left) and all surrounding populations from non-sulfidic surface
sites (right). Data from Tobler et al. (2008b).
Cavernicolous P. mexicana consequently represent one of the few known

systems in which cave-dwellers coexist parapatrically with their recent
ancestors, to which they are still connected by low rates of gene-flow
(see also Schilthuizen et al. 2005). The system thus provides a unique and
rare opportunity not only to study ecological and evolutionary processes
associated with cave colonization, but it is an emergent model system
to investigate mechanisms leading to local adaptation and reproductive

AB 2
2 3 5 2 3 4 1
6 7 6
3
EA II 1 7
2 1
2
RA RO AC 5
2
V X XI XIII
14
4
LA
3 2 2
4
2
EA I 2
1 3 1
1
AT 2
4
1 3
1
Fig. 5 Population assignment using microsatellite data (10 unlinked loci). Given
are numbers of 1st generation migrants and the direction of migration events,
as calculated from GENECLASS. For abbreviations (population code) see
Figure 1. Note that genetically detected “migrants” do not necessarily represent
actual migration events.
isolation due to divergent natural selection, i.e., ecological speciation

(Schluter 2000, Rundle and Nosil 2005). Both of these aspects will be
discussed in more detail below.
Phenotypic Differentiation and Adaptation Along Abiotic

Gradients
Poecilia mexicana from different habitat types did not only diverge
genetically, but also phenotypically, for example in terms of their
morphology and their behavior. We first discuss divergent traits that are
likely driven by and adaptive to differences in the abiotic environmental
conditions among habitat types. Below we will also look at traits that likely
evolved in response to different ecological conditions among habitats.
Each habitat type in the Cueva del Azufre system harbors P. mexicana
with distinct morphological features (Tobler et al. 2008a, Tobler et al.
2008b). Most importantly, there is independent variation explained along
the axis surface-cave as well as along the non-sulfide-sulfide gradient
(Figure 6). The observed morphological differences among P. mexicana
from different habitat types are not caused by environmentally induced
phenotypic variation alone, since laboratory stocks maintained under
identical conditions retained the distinct morphology of wild-caught fish
from the respective habitat types for multiple generations, indicating that
morphological traits have a heritable basis (Peters and Peters 1968, Tobler
et al. 2008b). Likewise, divergent behavioral traits of P. mexicana from
different habitat types have been shown to be heritable (Plath et al. 2004c,
Plath et al. 2006c, Plath 2008, Plath and Schlupp 2008).

cave surface
non-sulfidic sulfidic
Fig. 6 Morphological divergence due to (A) darkness, and (B) hydrogen sulfide.
Cartesian transformation grids depicting the effects of both selection pressures
(deformations are exaggerated three times). Note the smaller eye size and more
slender body in cave dwelling P. mexicana, and the larger head (and larger gills,
respectively) in mollies from sulfidic habitats.
Differentiation in Caves
Like cave-dwelling species in other taxonomic groups, P. mexicana have
diverged in multiple traits that are commonly associated with cave
evolution, both in terms of regression of traits that are no longer under
selection in darkness as well as in the evolution of novel traits. Compared to
fish from surface habitats, P. mexicana from the two caves are characterized
by a reduction in eye size (Gordon and Rosen 1962, Walters and Walters
1965, Peters et al. 1973, Plath et al. 2007a, Tobler et al. 2008b). Relatively
large eyes were found in surface habitats, smaller eyes in the anterior cave
chambers (V, X), and the smallest eyes were detected in the innermost
cave chamber XIII (Plath et al. 2007a). This pattern shows some congruence
with a previously proposed morphocline in eye size (Parzefall 1970, Peters
et al. 1973, Parzefall 2001). However, the morphological and genetic data
did not support the proposed mechanism for this morphocline, namely
its maintenance by migration from both directions into the middle cave
chambers. This would have led to an increased variance in eye size in the
middle cave chambers, which was not found (Plath et al. 2007a).
Unlike the eyes of other cave dwellers (Porter and Crandall 2003),
those of cavernicolous P. mexicana are still functional, and cave mollies
readily show behavioral responses similar to conspecifics from surface
populations to visual stimuli (Parzefall 2001, Körner et al. 2006, Parzefall
et al. 2007). A study of visual pigments in cave mollies as well as various

Fig. 7 Morphological divergence in P. mexicana. (A) Female and (B) male

from a non-sulfidic surface site (Arroyo Cristal). Note the bright coloration of
Fig. 7 Contd.

related Poecilia species/populations using microspectrophotometry

indicated that all species/populations exhibit four cone types including
UV-sensitive cones as described for other poeciliid species (Körner et
al. 2006). Differences in the absorption characteristics between species/
populations were small and, most importantly, the maximum absorbance
wavelength (lmax) values of the visual pigments of the cave form of
P. mexicana are almost unchanged, although stabilizing selection on visual
pigments is not evident in this population.
Besides the reduction of eye size, cave populations of P. mexicana also
have reduced body pigmentation (Figures 7A-E, Gordon and Rosen 1962,
Peters et al. 1973, Tobler et al. 2008a) and have more slender bodies than
conspecifics from epigean habitats (Figure 6, Tobler et al. 2008b) a trait
commonly observed in cave organisms (Langecker 2000). The reduction
in body height is not associated to poor nutritional condition of cave
populations (Tobler 2008), but could be driven by divergent predatory
regimes or resource scarcity (see below).
Poecilia mexicana from the Cueva del Azufre have also evolved a series
of constructive traits most of which seem to improve communication
and orientation in darkness. So far, only individuals from the Cueva del
Azufre but not the Cueva Luna Azufre have been investigated, so that it is
unclear whether the newly discovered population exhibits similar traits.
Female cave mollies exhibit a distinct genital pad that is absent in epigean
fish (Figure 7D). Supposedly, this pad secrets chemical signals that play a
role during reproduction (Walters and Walters 1965, Zeiske 1968, Parzefall
1970, 1973) and are detected by males by the increased number of taste
buds on the head (Parzefall 1970). Also, cave mollies exhibit a hyper-
developed cephalic lateral line system (Walters and Walters 1965; Parzefall
2001). The mechano-sensory lateral line system has been documented to
mediate spatial orientation in other cave fishes (Abdel-Latif et al. 1990,
Burt de Perera 2004). Last but not least, a part of the otolith morphology
– the suculus – diverged in cavernicolous P. mexicana (Schulz-Mirbach
Fig. 7 Contd..
large,dominant males in (B). (C) Male from chamber XI of the Cueva del Azufre,
“golden morph”. Arrow shows the gonopodium, the male copulatory organ. (D)
female from chamber XI with a pronounced genital pad (arrow) around the genital
pore. (E) Cave molly female from chamber V, “sunken eye type”. (F) The role of
Ophisternon aenigmaticum as a molly predator in the Cueva del Azufre remains
to be examined. This dead individual was found in El Azufre and had probably
been flushed out of the cave. The ruler is 30 cm long. (G) Inside the Cueva del
Azufre, cave mollies face predation by giant water bugs (Belostoma sp., be). ch:
chironomid larvae, one of the food sources of cave mollies.
book.

et al. 2008). The sulcus is a furrow-like structure on the inner face of an

otolith and is attached to the sensory epithelium. Otoliths from surface-
and cave-dwelling populations show pronounced differences in the depth
of the sulcus. Fish from surface habitats have otoliths with a flat sulcus
(Figure 8A1-2), whereas cave mollies possess otoliths characterized by a
deeper sulcus (Figure 8B1-2). In addition, otolith weight in fish from cave
chamber X is significantly higher than in fish from El Azufre. Greater
otolith weight in cave mollies (relative to the body size of the fish) may
be an effect of the increased depth of the sulcus. This difference in the
depth of the sulcus may reflect a greater hair cell area for sensory cells.
As otoliths are static position receptors, these differences may allow for a
better position perception in absence of visual cues (see Poulson 1963 for
divergent otolith morphology in troglobitic amblyopsids).
The modification of sensory structures in the cave molly is also reflected
in its behavior. Cave mollies from the Cueva del Azufre have evolved the
ability for communication in darkness, a trait that is lacking in surface
populations. While surface-dwelling P. mexicana predominantly rely on
visual signals, e.g., during mate choice, cave mollies are able to assess
mate quality in darkness (Plath et al. 2004c, 2006c). Most importantly, the
ability for non-visual communication is heritable and largely independent
of ontogenetic effects, i.e. whether fish were reared in light or permanent
darkness. This indicates a sensory shift in the detection of traits relevant
for mate choice as an evolutionary response to life in darkness. Using
visual cues for mate choice is the ancestral state in P. mexicana, and the
Fig. 8 Otoliths of P. mexicana from the sulfidic El Azufre at the surface (A), and
from cave chamber X of the Cueva del Azufre (B) (from Schulz-Mirbach et al.
2008).

use of non-visual cues has evolved as a novel trait in the cave population.
Our data indicate that the cave molly population from the independently
colonized Cueva Luna Azufre has not evolved the ability for comparable
communication in darkness and, in that respect, are more similar to surface
fish than the population from the Cueva del Azufre (Tobler et al. 2008c).
In summary, cave mollies are characterized by a number of regressive
and constructive traits when compared to surface populations. Although
currently far less understood, mollies from sulfidic habitats also diverged
from mollies in non-sulfidic habitats, irrespective of whether they live in a
cave or a surface habitat.
Adaptations to Sulfidic Habitats

Individuals from sulfidic habitats are characterized by an increase in head
size (Figure 6), which is correlated with an increase in gill filament length
(Tobler et al. 2008b). An increase in gill size facilitates oxygen uptake in
hypoxic environments (Graham 2005) and has been documented in various
fish species living in other types of hypoxic environments (Chapman et al.
1999, Chapman et al. 2000, Chapman and Hulen 2001, Timmerman and
Chapman 2004).
This highlights the importance of respiratory adaptations facilitating
efficient oxygen acquisition for survival in sulfidic habitats. Sulfide
detoxification in organisms capable of tolerating high and sustained
concentrations of H2S is primarily achieved through its oxidation to less
toxic sulfur compounds and subsequent excretion through the kidneys
(Curtis et al. 1972, Bagarinao 1992, Ip et al. 2004). Due to the hypoxic
conditions in sulfidic habitats, oxygen available for respiration is generally
limited, but at the same time oxygen is required for coping with the toxic
effects of H2S. Organisms inhabiting sulfidic habitats should therefore be
under selection for maximizing oxygen uptake. While some fish species
rely on air-breathing using specialized breathing organs or derived
respiratory epithels in the stomach or guts to do so (Bagarinao and Vetter
1989, Brauner et al. 1995, Affonso and Rantin 2005), P. mexicana from the
Cueva del Azufre rely on aquatic surface respiration (ASR), where fish
exploit the more oxygen-rich air-water interface using their gills (Plath et
al. 2007c).
Indeed access to the water surface and the possibility to perform ASR
critically influence survival in sulfidic water even over short periods
of time. When exposed to sulfidic water, cave mollies invest more than
75% of their time budget performing ASR. An experiment in the natural
habitat revealed that access to the water surface is the major predictor of
the survival even over short periods of time (24 h) (Plath et al. 2007c).

Physiological adaptations of P. mexicana to detoxify H2S remain to

be studied in detail. Invertebrates living in sulfidic environments often
harbor symbiotic sulfide-oxidizing bacteria (which detoxify H2S) on their
respiratory organs or as endosymbionts, or show unusually structured
(tube-like) mitochondria in response to exposure to H2S (Bagarinao 1992,
Grieshaber and Völkel 1998, Van Dover 2000). A pilot study using electron-
microscopic techniques, however, did not reveal any symbiotic bacteria
on or in the gills of cave mollies, and mitochondria are normally shaped
(Figure 9; Giere et al., unpublished data).
Fig. 9 EM detail of a gill filament of P. mexicana from the sulfidic Cueva del
Azufre showing regular mitochondria and nuclei. The bar equals 1 micrometer
(Giere et al. unpublished data).
There is evidence though that P. mexicana from sulfidic habitats have

physiological adaptations absent in conspecifics from non-sulfidic
habitats. While fish from sulfidic habitats can sustain in sulfidic water,
P. mexicana from non-sulfidic habitats lost motion control within seconds
after exposure to H2S (Tobler et al. 2008d). Similar results were obtained
when re-examining data from Peters et al. (1973) on sulfide tolerances
of laboratory-reared animals (Figure 10), suggesting that differences in
sulfide tolerance in P. mexicana is at least in part genetic.

2500
Estimated marginal means [h] 2000
1500
1000
500
0
surface cave
Fig. 10 Molly populations differ in their tolerance to H2S. We reanalyzed data

published by Peters et al. (1973). Sulfide tolerance was measured in fish from
non-sulfidic (surface) habitats and the sulfidic Cueva del Azufre. All fish used were
maintained in the laboratory in non-sulfidic water prior to the study. We reanalyzed
the data using a General Linear Model (GLM) with time sustained in sulfidic
water as a dependent variable, population as an independent variable and H2S
concentration as a covariate. H2S concentration did not have an effect (F = 0.563,
P = 0.463) but there were significant differences among populations (F = 5.168,
P = 0.035). In the figure, the estimated marginal means (± SD) from the GLM
are shown. Overall, P. mexicana from non-sulfidic surface populations (left) had a
lower sulfide tolerance than conspecifics from the Cueva del Azufre (right).
ECOLOGICAL DIFFERENCES AMONG HABITATS

The different habitat types in the Cueva del Azufre system differ in more
than abiotic environmental conditions. The energy basis among habitats
likely differs qualitatively and quantitatively. Due to the lack of light,
photosynthetic primary production is absent in cave habitats (Poulson
and Lavoie 2000) and likely impaired in sulfidic surface habitats, since H2S
is also toxic for most phototrophic primary producers (Bagarinao 1992).
On the other hand, the presence of sulfide allows for chemoautotrophic
primary production through sulfide oxidizing bacteria (Sarbu et al. 1996,
Hose et al. 2000, Colaco et al. 2002). Allochthonous input probably occurs
in all habitat types, but is dominated by leaf litter and terrestrial insects in
surface habitats (Allan 1995) and bat guano in the case of the two caves (cf.
Poulson and Lavoie 2000).
The food web of the ecosystem of the Cueva del Azufre appears to rely
to some extent on in situ chemoautotrophic bacterial primary production
(Langecker et al. 1996, Hose et al. 2000), whereas most other caves studied
so far mainly rely on energy input from the neighboring surface habitats.
Ecosystems relying on chemoautotrophic primary production are known

from deep-sea hydrothermal vents and cold seeps (Peek et al. 1998, Van
Dover 2000). A sulfurous chemoautotrophic cave-ecosystem, the Movile
cave, has been described from Romania (Sarbu et al. 1996). There, only
invertebrates – many of them endemic to the cave – use this unusual
food base. The Cueva del Azufre is the only known ecosystem in which a
vertebrate, P. mexicana, inhabits a chemoautotrophic cave.
The different habitat types also differ dramatically in the composition
of ecological communities. The current knowledge on the biodiversity
in the different aquatic habitats is still fragmentary and data are mostly
restricted to fish. In non-sulfidic surface habitats, diverse fish communities
can be found, which – as is typically for the wider region – are dominated
by cichlid and poeciliid species (Table 1, Tobler et al. 2006b). By contrast,
the diversity of fishes is heavily reduced in the divergent habitats, and
P. mexicana occurs as the single dominant species. Heterandria bimaculata
and Xiphophorus hellerii (Poeciliidae), Astyanax aeneus (Characidae) as well
as “Cichlasoma” salvini and Thorichtys helleri (Cichlidae) occur in areas of
the El Azufre where H2S in not measurable (interestingly, P. mexicana is
basically absent in the El Azufre as soon as other fish species are present).
Furthermore, a facultatively air-breathing synbranchid eel (Ophisternon
aenigmaticum) occurs in sulfidic surface habitats and has occasionally been
reported in the Cueva del Azufre (Parzefall, personal communication;
Figure 7F). In the Cueva Luna Azufre, P. mexicana has so far been the only
recorded fish.
Other taxonomic groups, such as snails and other aquatic invertebrates,
were only surveyed in the Cueva del Azufre, but a comparative analysis of
different habitat types remains to be carried out. However, it is likely that
the invertebrate diversity follows a similar pattern with reduced species
diversity in the sulfidic and cave habitats. A reduced species diversity and
dominance of a few specialists have been documented from other caves
(Gibert and Deharveng 2002) and other sulfidic habitats (Tobler et al. 2008d
and unpublished data).
Interspecific Interactions
Predation
Poecilia mexicana in non-sulfidic surface habitats are primarily exposed
to piscivorous fishes and birds as predators. Piscivorous birds are also
present in sulfidic surface habitats, but neither bird or fish predators
occur in the two caves (Tobler et al. 2006b, Tobler et al. 2007a). However,
in the Cueva del Azufre (but not the Cueva Luna Azufre), an aquatic
heteropteran (Belostoma sp.) occurs at high densities (over 1 individual/

Table 1 Fish communities in the Cueva del Azufre system. Table after Tobler
et al. (2006b) updated with additional data from 2007 and 2008.
Cueva Cueva El Non-
del Luna Azufre sulfidic
Azufre Azufre surface
habitats
Characidae
Astyanax aeneus (Günther 1860) x1 x
Brycon guatemalensis Regan 1908 x
Ariidae
Potamarius nelsoni (Evermann & Goldsborough 1902) x
Pimelodidae
Rhamdia guatemalensis (Günther 1864) x
Rhamdia laticauda (Kner 1858) x
Loricariidae
Pterygoplichthys cf. disjunctivus (Weber, 1991) x2
Batrachoididae
Batrachoides goldmani Evermann & Goldsborough 1902 x
Atherinopsidae
Atherinella alvarezi (Díaz Pardo 1972) x
Belonidae
Strongylura hubbsi Collette 1974 x
Poeciliidae
Heterophallus milleri Radda 1987 x
Heterandria bimaculata (Heckel 1848) x1 x
Poecilia mexicana Steindachner 1863 x x x x
Priapella chamulae Schartl, Meyer & Wilde 2006 x
Xiphophorus hellerii Heckel 1848 x1 x
Synbranchidae
Ophisternon aenigmaticum Rosen & Greenwood 1976 x3 x
Centropomidae
Centropomus undecimalis (Bloch 1792) x
Cichlidae
‘Cichlasoma’ salvini (Günther 1862) x1 x
Oreochromis cf. aureus (Steindachner 1864) x2
Paraneetroplus gibbiceps (Steindachner 1864) x
Theraps lentiginosus (Steindachner 1864) x
Thorichthys helleri (Steindachner 1864) x1 x
Vieja bifasciata (Steindachner 1864) x
Vieja intermedia (Günther 1862) x
Eleotridae
Gobiomorus dormitor Lacépède 1800 x
1
Only in non-sulfidic microhabitats
2
Introduced species
3
Parzefall, personal communication

m2 in cave chamber V) and seems to be the predominant predator of

P. mexicana (Tobler et al. 2007b). Belostoma are large aquatic hemipterans
(Figure 7G). As generalist feeders, they usually prey on aquatic insects,
snails, amphibians, and fish (Menke 1979). Belostomatids are sit-and-wait
predators that catch bypassing prey items with their raptorial forelegs that
are strongly incrassate, with the femora often grooved to accept the tibiae.
Upon capture, Belostoma inject toxins causing prey paralysis and digestive
enzymes causing tissue necrosis (Swart and Felgenhauer 2003).
Several experiments assessing the prey choice of Belostoma gave insights
into the potential role of predation on the evolution of cave mollies.
Belostoma exhibit a prey preference for larger cave mollies over smaller
fish (Plath et al. 2003b, Tobler et al. 2007b) as well as for males over females
of the same size (Tobler et al. 2008e). These preferences parallel the prey
preference of other predators (such as birds and fish) in surface habitats that
also preferentially attack large and male prey items (Endler 1987, Trexler
et al. 1994). In other poeciliids, differences in susceptibility to predation
among sexes as well as among size classes have been documented to
have profound impacts on various traits, such as life history strategies,
sex ratios and size distributions. In cave mollies, the higher susceptibility
to predation of large individuals, for example, has been hypothesized to
maintain male size polymorphisms in the Cueva del Azufre population.
Here, large males are favored by female choice (see below) but more likely
to be preyed upon by Belostoma, which gives an advantage in survival
to small males (Figure 11). It remains to be studied how strong the
contribution of size- and sex-specific predation by Belostoma is. Given that
sex ratios are more equal in cave mollies than in surface habitats (Table
2), the impact of predation by Belostoma on sex ratios is probably much
weaker than predation by herons and predatory fishes on the same species
in surface habitats.
Parasites
Differences in the ecological communities as well as the abiotic
environmental conditions among habitats have been hypothesized to
cause differences in parasite infection in P. mexicana. Parasites may become
locally extinct either directly, through selection by adverse environmental
conditions (such as H2S) on free-living parasite stages, or indirectly, through
selection on other host species involved in a parasite’s life cycle (Figure 12,
Tobler et al. 2007a). This hypothesis was tested by comparing the infection
rates of the digenean trematode Uvulifer sp. among several populations of
P. mexicana (Tobler et al. 2007a). The cercariae of Uvulifer sp. infect fish by
penetrating their skin, causing the production of black cysts (“black spots”)
in which the host encapsulates the parasite (Spellman and Johnson 1987,

((A) SURFACE FORM (e.g., Río í Oxolotan)

Natural Selection
(size-selective predation by herons etc.)
Male-male
- competition +
Large Small
male
+
+ -+male
“Alternative“
Female mating
choice behaviour
Female
(B) CAVE MOLLY (Cueva del Azufre)

(B
Natural Selection
(size-
(size-selective predation by Belostoma))
Male-male
- competition
+
Large Small
male male
+
Alternative
Female mating
choice behaviour
Female
Fig. 11 Direction of sexual and natural selection on male body size (i.e., size
at maturation) in mollies from (A) non-sulfidic surface sites and (B) the Cueva
del Azufre. (+) indicates an advantage for this male class due to the respective
selection factor, while (–) indicates an evolutionary disadvantage. After Plath et
al. (2004a).
Bush et al. 2001). Consistent with the predictions, we found populations

from sulfidic habitats to be significantly less parasitized by Uvulifer sp.
than populations from non-sulfidic waters, and cave mollies showed no
signs of this parasite at all. A subsequent, yet unpublished study, however,
indicated that this pattern in not consistent across different parasite
species. For example, P. mexicana from the Cueva del Azufre show a high

Table 2 Sex ratios in several populations of P. mexicana in Mexico. The rationale

behind female-biased sex ratios in the recently discovered non-sulfidic Cueva
Luna Azufre despite the apparent absence of predators in that cave remains to
be studied.
Habitat Sampling site Year Sex ratio n
(males/females) (adult)
Surface, Río Purificacion, stream 1998 0.07 49

non-sulfidic Río Purificacion, pool 1998 0.08 68
Río Purificacion, stream 2002 0.50 81
Río Purificacion, stream 2002 0.08 13
(Barretal)
Río Mante tributary 2002 0.63 31
Río Mante, Ciudad Mante 2002 0.46 19
Buena Vista (Río Tuxpan) 2002 0.38 95
Arroyo Cristal 2004 0.28 100
Arroyo Cristal 2006 0.36 19
Arroyo Bonita 2006 0.27 19
Río Amatan 2007 0.38 66
Sulfur El Azufre 1998 2.00 313
creek El Azufre 2002 1.98 173
El Azufre, Clear Creek1 2002 0.94 35
El Azufre 2004 0.99 100
Cave Cueva del Azufre, chambers
III-V 2004 1.33 191
X 1998 1.09 148
XI 2004 0.56 74
XIII 1998 0.49 138
XIII 2002 1.65 45
XIII 2004 0.56 145
Cueva Luna Azufre 2007 0.19 31
1
A small freshwater tributary of the sulfidic El Azufre.
prevalence of Dactylogyrus sp. (a monogenean gill parasite) infections, a

parasite that is far less frequent in other populations. Overall, it cannot
be concluded that P. mexicana from divergent habitats are generally less
parasitized, but the parasite communities clearly differ among habitat
types both in species composition as well as the frequency of occurrence
of different parasite species (Tobler et al., unpublished data).

Intestinal worm, sexual

reproduction
Final host: piscivorous bird
Metacercaria, no Miracidium, no
reproduction reproduction
Second intermediate Free living stage
host: fish
Sporocyst, asexual
Cercaria, no reproduction
reproduction First intermediate
Free living stage host: watersnail
direct selection on free-living parasite stages

Reduction in the parasite load in the fish host can be
caused by indirect selection on the parasite’s other host
species
Fig. 12 Life cycle of Uvulifer sp., a typical trematode parasite of Poecilia mexicana.
Arrows indicate pathways by which extreme environmental conditions (H2S and/or
darkness) can affect the prevalence of the parasite directly or indirectly, i.e., due to
selection against further hosts in its life cycle. From Tobler et al. (2007a).
Trophic Ecology
Resource Use and Trophic Morphology

Poecilia mexicana in non-sulfidic surface habitats primarily fed on detritus
and algae, which is consistent with previous studies investigating the
food habits of closely related species and P. mexicana in other parts of its
range (Darnell 1962, Winemiller 1993, Kramer and Bryant 1995, Bussing
1998, Miller 2005). The most pronounced difference in diet was observed
between non-sulfidic surface habitats and the other habitat types, where
invertebrates were consumed by P. mexicana (Tobler 2008). Especially in
the cave habitats, invertebrates made up the majority of the gut content.

A substantial amount of this invertebrate diet (about 50%), however,

stemmed from bat guano, not living invertebrates. This finding is also
consistent with a previous study that investigated the gut content of
P. mexicana from the sulfidic cave only (Langecker et al. 1996). Consequently,
the colonisation of extreme habitats in P. mexicana was accompanied by a
shift in resource use.
Likely, both a change in resource availability and differences in the
competitive regime (i.e., competitive release) were driving the shift in
resource use of P. mexicana. Indeed, specimens collected in different habitat
types did not only differ in what they fed on, but also the diversity of
food items present in their intestines; fish from cave habitats exhibited a
significantly higher diversity (i.e., a higher individual trophic niche width).
It is unlikely that this increase in the trophic niche width was caused by
a higher diversity in resources available, since there is ample evidence
that caves (and sulfidic habitats for that matter) exhibit a reduced species
diversity compared to surface habitats (Poulson and White 1969, Gibert
and Deharveng 2002). Niche expansion as a response to competitive
release (MacArthur et al. 1972, Schluter and McPhail 1992, Begon et al.
1996) or as a response to resource scarcity (MacArthur and Pianka 1966,
Schoener 1971, Fenolio et al. 2006) may drive the increase in the diversity
of resources used. Both scenarios, however, are not necessarily consistent
with P. mexicana from sulfidic surface habitats having a low individual
dietary niche width. At least in terms of fishes, the communities in the
sulfidic surface habitats are comparatively poor compared to those in the
cave habitats (Tobler et al. 2006b, Tobler et al. 2008, Riesch et al. 2009a).
The major difference in trophic morphology among populations living
in different habitat types was found between P. mexicana from non-sulfidic
surface habitats and those from extreme habitats, which parallels the major
dietary shift with incorporation of invertebrates into the diet. Intestinal
tract lengths in fishes are typically correlated with the amount of plant
material ingested (Kramer and Bryant 1995), and P. mexicana from divergent
habitats (which consume less plant material) had shorter intestinal tracts
(Tobler 2008, Tobler et al. 2009). They were also characterized by wider and
thicker jaws, which may be advantageous in handling larger (Wainwright
1996) and/ or more evasive prey items (Hulsey and Garcia de Leon 2005,
Higham et al. 2007).
Differentiation in skull morphology among populations inhabiting
different habitat types is less pronounced than the differentiation in
general body shape, which seems to be driven predominantly by abiotic
environmental factors, i.e., the lack of light (eye size reduction) as well
as the presence of H2S and hypoxia (increase in head and gill size). Thus,
differences in jaw morphology (Tobler 2008) are not simply explained

by a correlated response to selection on other characteristics of the head

(Chapman et al. 2000), but may actually be adaptive to differential use
of resources among populations. The finding that skull morphology is at
least partly determined by genetics supports the idea that evolutionary
divergence in skull morphology by natural selection is possible.
Energy Limitation
Early studies examining body condition in this system (by comparing
length-weight regressions) found P. mexicana from normal surface habitats
to exhibit the highest body condition, while cave populations had the
lowest, and specimens from sulfidic surface habitats were intermediate
(Tobler et al. 2006b, Plath et al. 2007c, Tobler et al. 2008a). Morphological
differences among populations, however, affected these results since cave
population are generally more slender bodied. In fact, P. mexicana from
sulfidic surface habitats have equally low amounts of storage lipids as fish
from the cave populations (Tobler 2008).
Fish from cave and sulfidic habitats may have a low condition for
different reasons. Poecilia mexicana from the non-sulfidic cave exhibited
low amounts of storage fats likely because resources are scarce. Caves
relying on energy input from surface habitats are known to be energy
limited (Streever 1996, Hüppop 2000, Poulson and Lavoie 2000). Bat
guano is thought to be the trophic base of cave food webs whenever bats
are present and provide an energy-rich food base (Culver 1982, Willis and
Brown 1985), but recent work indicates that this is not necessarily the case
(Graening and Brown 2003).
Sulfidic habitats in turn have been suggested to be resource-rich
due to the presence of chemoautotrophic bacterial primary production
(Langecker et al. 1996). The paradox of fish with low body condition
living in an apparently resource-rich environment may be explained in
two (not mutually exclusive) ways. (1) Although resource-rich, sulfidic
habitats may lack particular nutrients for fish or provide an imbalanced
diet, which may negatively affect condition (Jeyasingh 2007). (2) Coping
with the toxic environment may be energetically costly. Although the
physiological mechanisms of sulphide-tolerance are not well understood
in P. mexicana, detoxifying H2S has been shown to be energetically costly
under hypoxic conditions in the mudskipper, Boleophthalmus boddaerti
(Ip et al. 2004). Short-term survival of P. mexicana in sulfidic water is
directly dependent on energy-availability and possibility to perform ASR
(see above) (Plath et al. 2007c). ASR itself is physiologically costly and
constrains an individual’s energy budget, leaving less time for foraging
(Kramer 1983, Weber and Kramer 1983, Chapman and Chapman 1993,
Tobler et al. 2009). Poecilia mexicana in sulfidic habitats thus seem to be

living in a resource-rich habitat but paying a cost for coping with the toxic
conditions. The high resource availability in these habitats may, in fact, be
one of the factors making life in such extreme environments possible at
all.
Low energy availability and/or high energetic costs of sustaining in a
stressful environment likely are strong selective factors acting on P. mexicana
from divergent habitat types. In fact, a multitude of traits, ranging from
shift in life history strategies to the reduction of costly behavioral traits,
may have evolved as a response to energy limitation in the divergent
habitats.
Life History Modifications

Life history traits, such as female fecundity, can evolve rapidly in response
to environmental changes (Reznick et al. 1990, Reznick et al. 2001). Some
recent studies highlight that life history modifications also occurred in
P. mexicana from the Cueva del Azufre system. In a pilot study, female
fecundity (number of ripe oocytes and developing embryos) was assessed
in wild-caught P. mexicana from the Cueva del Azufre and a non-sulfidic
surface creek (Arroyo Cristal) as well as in laboratory-reared individuals
from the Cueva del Azufre and the Río Oxolotan, which were reared in
either light or darkness (Riesch et al. 2009b). Compared to mollies from
surface habitats, cave molly females exhibited a dramatically reduced
fecundity, suggesting an evolutionary shift in energy allocation by cave
molly females (Figure 13). On the other hand, oocytes are larger. Differences
between populations were found to be stronger than those described for
guppies (Poecilia reticulata) from different habitat types in Trinidad, where
divergent predatory regimes appear to be the main driver in life history
divergence (Reznick and Endler 1982).
Even in the laboratory setting, the fecundity of cave mollies was
generally lower compared to that of surface-dwelling mollies. In both
populations fecundity was slightly higher under the respective “natural”
light conditions (light for the Río Oxolotan population and darkness in the
case of cave mollies) (Riesch et al. 2009b, in press-a). Our results suggest a
genetic component to the reduction in female fecundity in the cave molly,
because in the laboratory cave mollies were reared in the absence of toxic
H2S and the overall effect was largely independent of the light treatment.
The results from a recent study including Luna Azufre fish (Riesch et al. in
press-b) corroborate these findings: extremophile populations generally
have fewer, but larger eggs.
Differences in egg size among populations are also accompanied
by differences in maternal nutrient provisioning after conception

Fig. 13 Life history trait evolution in P. mexicana. Assuming that the volume of
the ovary constrains the number of oocytes a female can produce, P. mexicana fe-
males have two options: they can either produce numerous small oocytes (above,
female from Arroyo Cristal) or few, but large oocytes containing more yolk (below,
cave molly female from cave chamber X of the Cueva del Azufre (compare Trexler
and DeAngelis, 2003).
book.
(matrotrophy). The extent of maternal provisioning can be estimated by

comparing the weight of neonates to the weight of ova at conception. In
purely lecitotrophic species (i.e. species without matrotrophy), embryos
loose weight during gestation due to metabolic losses, but weight

loss should be minimal or even absent in species with matrotrophy. In

P. mexicana, the weight of neonates as a percentage of the weight of the egg
at fertilization was approximately 58% in fish from non-sulfidic surface
sites, 86% in the Cueva Luna Azufre population, and 57% in the fish from
the Cueva del Azufre. Our data indicate that this double-provisioning of
nutrients (via more yolk and more matrotrophy) leads to larger body size
of the neonates in the cave populations compared to fish from non-sulfidic
surface sites (Riesch et al. in press-a, in press-b).
Female reproductive investment is usually considered much more
energy demanding than male investment into spermatogenesis (Bateman
1948, Trivers 1972); however, also sperm production is not cost-
neutral (Nakatsuru and Kramer 1982). Thus, male sperm stores in five
populations of P. mexicana from naturally sulfidic and non-sulfidic habitats
were examined (Franssen et al. 2008). Pronounced differences among
populations in the number of sperm stripped per male were also reflected
by differences in gonad weights. The largest sperm stores were detected in
males from non-sulfidic surface creeks (Arroyo Cristal and Arroyo Tres),
while males from sulfidic surface waters (El Azufre) had smaller sperm
stores. Inside the sulfur cave (Cueva del Azufre), males were collected in
two cave chambers, and male sperm stores were similarly low. Hence, the
extreme environmental conditions in sulfidic habitats appear to constrain
male sperm production.
In the laboratory male sperm production of cave mollies (Cueva del
Azufre) and surface mollies (Río Oxolotan) was studied in response to
nutrient availability (Franssen et al. 2008). Males of both populations
were maintained both under a high and a low food treatment. Surface
dwelling and cave molly males produced comparable amounts of sperm
when fed ad libitum, suggesting that unlike female fecundity male sperm
production is to a large extent a plastic response of the males to adverse
environmental conditions. Interestingly, less sperm was stripped from
surface males after one week of food deprivation, while no such decrease
was found in the case of cave molly males. This may point towards an
increased starvation resistance in cave mollies.
BEHAVIORAL ECOLOGY
Reduction of Costly Behavioral Traits
Anti-predator Behaviors
Shoaling behavior protects fishes from avian and piscine predation, but
at the same time costs of group living arise, e.g. due to increased food

competition (Krause and Ruxton 2002). Cave mollies (both caves) live in
an environment in which avian and piscine predators are lacking, and they
are energy limited. Altogether, this should favor the reduction of shoaling.
When shoaling behavior (i.e. the tendency to associate with a stimulus
shoal) was compared among surface dwelling populations of (Poecilia
mexicana) and the two cave forms, shoaling was found to be reduced
in both cave forms compared with surface dwelling mollies (Plath and
Schlupp 2008). An analysis of shoaling in population hybrids revealed
that this reduction is heritable (Parzefall 1993).
Also a plastic response of shoaling behavior to starvation was found
in surface fish (Río Purificación population); individuals that had been
food deprived for one week prior to the tests spent less time shoaling
than well-fed ones, probably because food competition is more relevant to
starved fish (Plath and Schlupp 2008). Moreover, starved fish are probably
physically weaker, and hence, inferior in food competition (Krause et al.
1999). From the latter result it can be concluded that sustained malnutrition
over evolutionary times may indeed have selected for a genetic reduction
of shoaling behavior in cave mollies in an environment with absence of
avian and piscine predation.
Shoaling behavior often comes along with ‘conspecific cueing’ (Kiester
1979, Wolf 1985), where phenotypically similar individuals, such as
members of the same species, attempt to form homogenous groups,
because individuals in uniform shoals are worse targets to predators.
Heterospecific fishes, such as swordtails (Xiphophorus hellerii) are common
in all non-sulfidic waters around the Cueva del Azufre and the Cueva
Luna Azufre (Table 1) and occur in some non-sulfidic parts of the El
Azufre, but are absent in the caves. When surface-dwelling P. mexicana
females (Río Oxolotan) were given an opportunity to associate with a
conspecific or a swordtail female, they strongly preferred the conspecific
female when visual cues were available to the female (Riesch et al. 2006b).
No association preference was observed when only non-visual cues were
provided. Females from El Azufre showed only a slightly weaker response.
In contrast, cave-dwelling females showed no preference under all testing
conditions, suggesting that the preference to associate with conspecifics
was reduced. These results were corroborated by a subsequent study
using conspecific vs. guppy (P. reticulata) females as stimulus fish (Tobler
et al. 2006a).
Aggression
Parzefall (1974) demonstrated that aggressive male interactions occur
only in surface populations, but not in the cave molly. In surface-dwelling
Atlantic mollies the fish normally form shoals with a size dependent

rank-order in males. When females become receptive and ready to be

inseminated, large dark-colored males vigorously defend such females.
Aggression is most pronounced when a male of the same size tries to
approach such a female, but similar-sized males will generally show
intense fighting when they first meet, i.e., before a rank-order is established.
The aggressive repertoire consists of seven different behavior patterns
(aggressive fin erection, tail lifting, s-positioning, tail beating, ramming,
biting and circling; Parzefall 1974, 2001). Small, subdominant males with
more cryptic body coloration typically hide in a head down position and
try to sneak copulations (Parzefall 1969).
In the cave molly (Cueva del Azufre), all patterns of aggressive behavior
are reduced. This reduction was found to increase gradually from the
animals originating from the front chambers of the Cueva del Azufre to
those found in the rear cave chambers (Figure 8 in Parzefall 2001). When
cave specimens were tested with males of either their own or the epigean
population, the reduction of aggressive behavior remained unchanged
(Figure 9 in Parzefall 2001). In all tests, the mean number of aggressive
behaviors of the fish from the innermost cave pool (XIII) was about zero.
In mixed tests the majority of the cave molly males answered the initial
attacks of the epigean males with sexual patterns (following, nipping and
copulation), only a few reacted defensively for a short time. No cave molly
male ever attacked.
When Parzefall (1974, 1979) tested hybrids between cave and epigean
fish, the frequency distribution for aggressive fin erection and s-position
revealed a genetically based reduction of the aggressive behavior in the
cave molly. On the basis of quantitative genetic data, Parzefall (1979)
concluded that a closely linked polygenic system exhibiting additive
gene interaction exists and controls the aggressive behavior in P. mexicana.
Given that males do not fight for access to females, intrasexual selection
via aggressive male-male competition breaks down (Plath et al. 2004a;
Figure 11).
When comparing the aggressive behavior of cave mollies with that of
other cave-dwellers, reduced aggression seems to be the exception rather
than the rule. For example, in the cave form of the Mexican tetra Astyanax
mexicanus, males and females may defend small feeding territories and use
signals effective in close body contact to communicate with competitors.
Biting, circling and tail-beating have been observed, but no fin-spreading,
snake-swimming and ramming like in the epigean relatives (Parzefall and
Trajano, this volume). It seems that the aggressive behavior of these cave
fish evolved as an adaptation to the lightless habitat after the reduction
of the visually mediated aggression of the epigean ancestor (see chapter
4 by Parzefall and Trajano). Also other cave fishes, like the Somalian

cyprinid Phreatichthys andruzii, can be highly aggressive (authors, personal

observation).
The reduction of aggressive behavior in the cave molly has previously
been interpreted as an adaptation to darkness, but this argumentation failed
to explain why other cave fishes can be highly aggressive (Parzefall 1993).
Therefore, it seems reasonable to assume that the extreme environmental
conditions in the Cueva del Azufre act as a strong selection factor, favoring
the reduction of any energy-demanding behaviors.
Male Mating Behavior

Male poeciliids from normoxic surface habitats usually spend most of
their time either defending females or attempting to mate (Parzefall 1969,
Houde 1997). Poecilia mexicana males show more sexual behavior than
wanted by the females, leading to a sexual conflict (Plath et al. 2003b).
Indeed, P. mexicana females from all populations examined so far (Río
Oxolotan, El Azufre and Cueva del Azufre), tend to avoid the presence of
a male when given a choice to associate with a male or a female stimulus
fish in a simultaneous, dichotomous choice situation (Plath et al. 2001,
Plath and Tobler 2007). Only during few days of receptivity during their
monthly sexual cycle, cave molly females appear to be attracted by males
(Plath et al. 2001).
Nevertheless, poeciliid males also attempt to mate with non-receptive
females, i.e., they show coercive mating. Sexual harassment by males
has been reported from several livebearing fishes (Poeciliidae) and has
been shown to inflict costs to females (Plath et al. 2007b). For example,
poeciliid females have reduced feeding opportunities when accompanied
by a male, because females need to dedicate attention to avoiding male
copulation attempts (Plath et al. 2007b). When comparing male sexual
activity, and female feeding time reduction by male harassment in
Río Oxolotan, El Azufre and Cueva del Azufre fish (Plath et al. 2003b),
we could demonstrate that (a) unlike surface mollies, cave molly and
El Azufre males show low sexual activity even in the laboratory (i.e. in
light and in the absence of H2S), (b) females do not suffer feeding time
reductions from male harassment in those populations, and (c) small Río
Oxolotan males show especially high rates of sexual behavior (probably
to compensate for their disadvantage in female choice, Figure 11), but this
pattern breaks down in the cave molly and the El Azufre population (see
Plath et al. 2004b, Plath et al. 2005b, Plath et al. 2006a for further laboratory
and field studies).
After the discovery of the second cave form of P. mexicana from the non-
sulfidic Cueva Luna Azufre, male sexual behavior, and potential costs of
male harassment in terms of reduced feeding rates of the females in the

presence of a male, could also be examined in the latter cave form, and was
compared to previous data from two populations inhabiting clear-water
surface sites (Río Oxolotan and Tampico in central Mexico), the sulfidic
El Azufre, and the Cueva del Azufre (Plath 2008). In all populations with
at least one physicochemical stressor present (H2S or absence of light), a
reduction in male sexual activity was recorded. While females from both
habitats without physicochemical stressors spent less time feeding around
males, no such costs of male harassment were detected in populations
from extreme habitats, including the Cueva Luna Azufre cave fish. This
study corroborated the interpretation that energy-limitation is indeed the
driving force selecting for reduced male sexual activity, since fish from
all habitats with presence of physicochemical stressors were found to
show a low body condition factor under natural conditions (Tobler 2008).
As revealed by the analysis of male sperm production (see above), those
males need to allocate relatively more energy to somatic maintenance, and
thus allocate relatively less energy into reproduction.
Another study examined the question of whether low male sexual
activity, lack of sexual harassment and lack of size-dependent mating
behavior in the cave molly (Cueva del Azufre population) is phenotypically
plastic (Plath et al. 2004b). When cave molly males were sexually deprived
for one week, male sexually activity increased slightly, but still no costs
of harassment for females were detected. Also when female feeding
motivation was manipulated by food deprivation or when a combination
of both treatments was used, females did not suffer from male harassment.
The number of sexual behaviors was not correlated with male body size in
any experiment, indicating that even after sexual deprivation small cave
molly males do not switch to the pattern of size-dependent (“alternative”)
mating behavior known in surface-dwelling P. mexicana. Also when
another male, i.e., a rival, could interact with the same female did small
males not change their behavior, and still showed less sexual behavior
than large males (Riesch et al. 2006a).
When plasticity in male sexual behavior in response to sexual
deprivation was observed in the El Azufre population (Plath et al. 2006a),
costs of male harassment for females were also not detectible following
the described treatments. However, although El Azufre males also
maintained relatively low sexual activity after sexual deprivation, mating
behavior was found to be negatively correlated with male body size. In
conclusion, reduced sexual activity has a strong genetic component in
P. mexicana from sulfidic habitats, and the genetic basis for size-dependent
mating behavior is lacking at least in the cave molly (Figure 11), while it
still has a phenotypically plastic component in the El Azufre population.
Altogether, these results suggest a disruption in the antagonistic
coevolution between male persistence traits (harassment) and female

resistance traits (vigilance and flight behavior) in cave mollies. It seems

that female vigilance behavior is reduced compared to surface fish, because
males show extremely low sexual activity under natural conditions (Plath
et al. 2005b), and even experimentally increased male sexual activity no
longer elicits a response in cave molly females.
Sexual Selection
Size Preferences
The cave molly is one of very few cave fishes in which the evolution of
mating preferences has been investigated so far (but see Plath et al. 2006b
for mating preferences in cave tetras). One male trait females are commonly
responsive to throughout the Animal Kingdom is large male body size
(Ryan and Keddy-Hector 1992, Andersson 1994). Like most poeciliids,
epigean P. mexicana females exhibit a strong mating preference for large
bodied males (Plath et al. 2004b). Mating preferences for large males are
visually mediated in surface populations, but cave molly females from
the Cueva del Azufre have evolved the ability to discriminate between
males of different size using non-visual cues. The ancestral visually
mediated preference has not been lost and can be observed when females
are tested under light conditions (Plath et al. 2003a, Plath et al. 2004c, Plath
et al. 2004d). Interestingly, cave mollies from the independently colonized
Cueva Luna Azufre have not evolved the ability to perform mate choice in
darkness (Tobler et al. 2008c).
The differences in the ability to discriminate between different sized
males in darkness had consequences for male traits under selection (male
size at maturity). Pronounced male size polymorphisms are characteristic
for epigean populations as well as the cave molly population from the
Cueva del Azufre. Costs of growing large and being preferentially preyed
upon by predators are balanced in large males by a mating advantage
mediated by female choice (Figure 11). In the Cueva Luna Azufre, however,
large males appear to be absent. The lack of the ability in females to
discriminate between large and small males might have led to a reduction
of advantages of being large, such that growth in males has been counter-
selected (Tobler et al. 2008c).
Mating preferences for body size cannot only be found in females,
but also in males. Males likely prefer to mate with large females because
their fecundity – which is a direct function of body size – is higher. Like
females, male cave mollies from the Cueva del Azufre population have
evolved the ability to determine female size in darkness, while epigean
populations are unable to do so. Male mating preferences are clear-cut

both in simultaneous (Plath et al. 2006) and sequencial (full contact) mate
choice experiments (Plath 2008). Males from the Cueva Luna Azufre have
so far not been tested for their ability to discriminate between differently
sized females in darkness.
Indicator Value of Nutritional State

Theory predicts that female preferences may evolve relative to the ‘indicator
value’ of the male trait (Harvey and Bradbury 1991, Kirkpatrick and Ryan
1991, Andersson 1994, Grether 2000). Traits that reliably reflect male body
condition could function as indicators of the male’s (in)ability to cope with
its environment, such as environmental fluctuation, food shortage, or other
stressful conditions. One such male trait is its nutritional state, namely, the
extension of its abdomen. If males are unable to gather enough food, they
will inevitably look malnourished. Female P. mexicana from the different
populations clearly differed in mating preferences for well nourished vs.
starved males. River-dwelling females never showed a clear preference,
but sulfur creek females did so in the visual cues treatment. Cave molly
females strongly preferred to associate with the well-fed male both when
visual and non-visual communication was possible, i.e., they showed this
preference also in darkness (Plath et al. 2005a).
As discussed before, fish from both the sulfidic surface and cave
habitats are energy limited, whereas mollies from non-sulfidic rivers have
access to sufficient food (Plath et al. 2005a). The weak response in the river
population reflects that good male nutritional state has a low ‘indicator
value’ in such habitats (basically, every male is well fed), such that females
do not base they mate choice on this trait. By contrast, a male that shows
a good nutritional state despite the environmental harshness in its habitat
will possibly provide indirect benefits to the female’s offspring (in the
form of “good genes”). In all populations studied, there was variation in
male nutritional state (Plath et al. 2005a). Hence, under natural conditions,
there is a potential for benefits to females by female choice, because males
differ in “quality”.
SYNTHESIS
Insights into the Origin of Cave Organisms
Active Colonization
Originally, all cave organisms invariantly descended from epigean
ancestors that colonized subterranean habitats and subsequently diverged

into distinct evolutionary lineages. Cave colonization has long been viewed
as a passive process, where organisms were accidentally washed into the
underground and got trapped (Wilkens 1979, Langecker 1989). According
to Romero and Green (2005), this notion predominantly derived from the
– faulty – logic that caves provide a harsh environment that organisms
would try to avoid if possible. In fact, most cave habitats are continuous
with adjacent epigean habitats, and no permanent physical barriers
prevent organisms from returning to their original habitats (e.g., Romero
et al. 2002, Reis et al. 2006). Alternatively, cave colonization by epigean
organisms may be active and advantageous, and potential advantages
include environmental stability, exploitation of unoccupied niches, as well
as a reduction in competition and predation (Romero and Green 2005).
Empirical evidence for active colonization, however, has been scarce to
date.
Negative phototactic behavior, which has been reported as a potential
mechanism for active cave colonization, does not explain cave colonization
in P. mexicana as cave forms – just like their epigean ancestors – consistently
exhibit positive phototactic behavior under various light conditions
(Parzefall et al. 2007). Thus, phototactic behavior alone neither explains
the initial colonization of subterranean habitats by P. mexicana, nor the
restricted gene flow among populations.
The comparative analysis of the biotic environment of the different
habitat types as well as the ecology of P. mexicana in the Cueva del
Azufre system, however, has provided evidence that advantages of cave
colonization may exist. Firstly, the colonization of divergent habitat types
was accompanied by a shift in resource use (Tobler 2008), indicating the
exploitation of resources underused in ancestral habitats. Secondly, the
presence of predators differs vastly among habitat types. Avian as well
as piscine predators are absent in the cave habitats, but are common at
least in non-sulfidic surface habitats (Tobler et al. 2006b, Tobler et al. 2007a,
Riesch et al. 2009a), indicating that subterranean habitats may offer some
protection against predation. Lastly, it was also hypothesized that cave
habitats may offer an advantage in terms of reduced parasite exposure.
Some parasites (e.g., the trematode Uvulifer sp.) that are highly abundant
in non-sulfidic surface habitats have reduced prevalence in sulfidic surface
habitats and are even completely absent in the sulfidic cave (Tobler et al.
2007a). Active colonization conferring benefits for cave colonizers thus
seems a plausible scenario for the colonization of the caves in the Cueva
del Azufre system.
Parapatric Divergence and Ecological Speciation

A predominant view in the literature is that obligate cave-dwellers evolved
from facultative cave colonizers that became isolated through extinction

of the surface ancestor (Barr 1968). Fluctuating environmental conditions

during periods of climate change are thought to drive surface populations
to extinction while cave populations sustain in suitable habitats (e.g., in
terms of temperature or humidity); hence this hypothesis is known as the
climate-relict model (Barr 1968, Wilkens 1973, Humphreys 1993, Trajano
1995, Allegrucci et al. 2005). Essentially, this is a two-step model of cave
evolution, where caves were first colonized without restriction of gene flow
with surface populations followed by allopatric genetic and phenotypic
differentiation in geographically isolated caves after the local extinction of
surface-dwelling ancestors.
The climate-relict model is inherently difficult to test in specific cases,
since it fails to provide exclusive testable predictions. Even if phylogenetic
and biogeographic patterns are congruent with the reconstruction of
climatic changes in the past, it is impossible to exclude the possibility
that troglobites diverged parapatrically from epigean ancestors before
the respective surface populations went extinct. So, although many caves
with their geological age and their environmental stability may indeed
act as climatic refugia for phylogenetic lineages long extinct in adjacent
surface habitats, the climate-relict hypothesis only offers limited insight to
the ultimate mechanisms of the origin of troglobites.
An alternative hypothesis prevalent in the biospeleological literature is
the ‘adaptive shift’ model, where epigean species invade caves to exploit
novel resources and subsequently adapt to the divergent environmental
conditions present (Howarth 1987). Divergence between epigean ancestors
and cave inhabitants is usually assumed to arise parapatrically along the
steep environmental gradient. Additional, sometimes very specific, models
have been proposed to explain the origins of troglobites (see Danielopol
and Rouch 2005 for an overview).
Generally, the investigation of the origins of subterranean organisms
has been plagued by the lack of appropriate study systems. Many lineages
of troglobites are phylogenetically old, making the elucidation of the
mechanisms of initial colonization and divergence an improbable task.
Future work will have to put more emphasis on phylogenetically young
systems, where cave colonization occurred recently or is still in progress
(e.g., the Cueva del Azufre system, see also Schilthuizen et al. 2005).
Molecular genetic work highlights that different populations of
P. mexicana are still connected by low rates of gene flow, and, most
importantly, all evidence points towards parapatric divergence of cave
populations. However, the isolating mechanisms leading to genetic
differentiation among populations of P. mexicana from different habitat
types are only poorly understood so far. It is unlikely that populations in
the divergent habitats are genetically incompatible (i.e. not inter-fertile),

since the absence of intrinsic post-zygotic reproductive isolation has been

documented even in more distantly related poeciliid species (Hubbs 1959,
Schartl 1995, Ptacek 2002, Dries 2003, Rosenthal et al. 2003, Alexander and
Breden 2004, Kittell et al. 2005). Likewise, a genetically-based preference for
separate habitat types (Rice and Salt 1990, Johnson et al. 1996) is unlikely at
least for the separation between surface- and cave-dwelling populations,
as P. mexicana from the Cueva del Azufre – like surface-dwelling fish from
El Azufre – exhibit photophilic behavior (Parzefall et al. 2007).
We propose that divergent selection caused by the abiotic environmental
conditions gives rise to reproductive isolation. Several mechanisms,
which may act in synchrony, seem possible in this system. (1) Selection
could act directly on immigrants from divergent populations causing
prezygotic isolations (Nosil et al. 2005). For example, P. mexicana from non-
sulfidic habitats have been shown to have a high susceptibility to the toxic
effects of H2S (Peters et al. 1973, Tobler et al. 2008d). Likewise, fish from
surface habitats have an impaired ability for intraspecific communication
in darkness, which may lower their reproductive success in the cave
environment (Plath et al. 2004c, 2006c). Selection against immigrants,
however, should be unidirectional, unless adaptations to sulfidic and cave
habitats come at a cost in non-sulfidic and surface habitats, respectively.
For example, the reduced eye size could make cave P. mexicana more
susceptible to predation or less efficient foragers in light compared to
conspecifics from surface habitats. (2) Poecilia mexicana from different
habitat types may be less attracted to conspecifics from divergent habitat
types, which again may cause prezygotic isolation (Schluter 2000, Rundle
and Nosil 2005). For example, in poeciliid fishes of the genus Gambusia,
assortative mating for divergent body shapes caused by different
predatory regimes mediates reproductive isolation among allopatric
populations (Langerhans et al. 2007). Whether mating preferences for
divergent phenotypes are present in the P. mexicana system remains to be
studied. Besides intersexual selection, intrasexual selection by male-male
competition may be operating. For example, all populations from cave
and/or sulfidic habitats have reduced male sexual activity (Plath et al.
2003b, Plath 2008), and Cueva del Azufre fish are less aggressive (Parzefall
1974). Since reproductive success in poeciliids is determined by the number
of successful copulations (Magurran and Seghers 1994), such males may
be outcompeted by males from non-sulfidic surface fish due to higher
aggression and higher rates of sexual behavior. (3) Divergent selection
could act on hybrids of P. mexicana from different habitats (Hatfield and
Schluter 1999, Schluter 2000), but to date no empirical evidence for this
mechanism is available.
Future studies will need to pay careful attention to the mechanisms
of reproductive isolation causing the observed small-scale population

differentiation in the Cueva del Azufre system in order to test whether

parapatric ecological speciation is indeed occurring. The current data,
however, provide strong evidence for population differentiation and
local adaptation along abiotic environmental gradients. Most other model
systems studied to date found biotic ecological factors, i.e. interactions
between species, such as resource use (Funk 1998, Ryan et al. 2007),
predation (Langerhans et al. 2007), or parasitism (Blais et al. 2007), to be
driving adaptive divergence among lineages.
Conservation Issues
Cave mollies are only known from the Cueva del Azufre system. The
wider region is currently under rapid development, especially in terms of
tourism. Deforestation, change in land use, and general population growth,
which is accompanied by the spread of human settlements ever closer to
the caves, must currently be considered as the main risk for the long term
conservation of the caves and their inhabitants. During the last few years,
a substantial proportion of the forests around the Cueva del Azufre were
cut down, and signs of severe erosion can be seen. Furthermore, cattle
farming as well as agriculture have increased in frequency. Like in other
places in Mexico, the number of inhabitants is still increasing around the
Cueva del Azufre. Settlements are now found directly above the Cueva
del Azufre. Although no data are currently available, an increase of soil
particle-, nutrient-, and potentially contaminant-influx (like human wastes
and fertilizers) into the cave has to be expected, which may be a direct
threat to the fragile cave ecosystems.
Also, the number and frequency of people visiting the caves may affect
the organisms living in it. Nowadays, the Cueva del Azufre is a major
source of attraction for tourists, and an ever-larger number of visitors
come to the cave. Green tourisms clearly has developed as a major source
of income for the local community, however, it needs to be assured that
current and future development will be sustainable. Especially lamps of
visitors that enter the deeper parts of the cave disturb bats and the cave
fish. Also the waste left in the cave can attract alien species, which may
be competitors to the cave adapted organisms, thereby threatening their
existence.
FUTURE DIRECTIONS
Work on the cave molly, in comparison to other populations of Poecilia
mexicana, has provided insight into the behavioral and evolutionary
ecology of an extremophile cave fish. Moreover, molecular biological

analyses so far indicated that extremophile poeciliids provide a unique

opportunity to examine divergent local adaptation, a process that may
ultimately cause (parapatric) small-scale speciation in cave and sulfidic
habitats. By conducting similar analyses in other extremophile populations
of P. mexicana, it may become possible to test a central hypothesis proposed
here, namely that genetic differentiation and restricted gene-flow are more
pronounced along harsh environmental gradients.
Another focus of future work should lie on the evolution of potential
isolation mechanisms. For example, it seems promising to examine if
assortative mating occurs at the interface of sulfidic and non-sulfidic, or
surface and cave habitats, i.e. if females (and males) have evolved mating
preferences for mates of their own, locally adapted population. Assortative
mating might further promote speciation in sulfidic systems by preventing
gene flow. Generally, research in the Cueva del Azufre system is one of
the few examples where the ecology, behavior and evolution of several
surface and cave populations were simultaneously considered. Applying
multi-biological approaches to study other cave fish systems is clearly a
promising challenge for future research.
Acknowledgements
We are deeply indebted to the people in the community of Tapijulapa, the
Universidad Intercultural del Estado de Tabasco in Oxolotan as well as the
Municipal de Tacotalpa for their continuous support and their hospitality
during our visits. A number of people are acknowledged for contributing
to this project: C.M. Franssen, F.J. García de León, O. Giere, K.E. Körner, A.
and D. Möller, J. Parzefall, R. Riesch, L. Rosales Lagarde, and I. Schlupp. T.
Schulz-Mirbach kindly provided information about otoliths. We thank J.K.
Langhammer for providing access to an unpublished manuscript of R.R.
Miller. Financial support came from the Deutsche Forschungsgemeinschaft
(DFG) to M.P. and from the Swiss National Foundation (SNF PBZHA-
121016) to M.T. The Mexican government and the local authorities provided
permits to conduct our research (the most recent ones are: Permiso de
pesca de fomento numbers: 291002-613-1577, DGOPA/5864/260704/-
2408). The authors are indebted to C. Lamptey (M.P.) and C.M. Franssen
(M.T.) for their support during the writing process. Finally, we wish to
thank J. Parzefall for introducing us to the biology of the cave molly.
References
Abdel-Latif, H., E. Hassan and C. von Campenhausen. 1990. Sensory performance
of blind Mexican cave fish after destruction of the canal neuromasts.
Naturwissenschaften 77: 237-239.

Affonso, E. and F. Rantin. 2005. Respiratory responses of the air-breathing fish

Hoplosternum littorale to hypoxia and hydrogen sulfide. Comparative Biochemistry
and Physiology C - Toxicology and Pharmacology 141: 275-280.
Alexander, H.J. and F. Breden. 2004. Sexual isolation and extreme morphological
divergence in the Cumana guppy: A possible case of incipient speciation. Journal
of Evolutionary Biology 17: 1238-1254.
Allan, J. 1995. Stream Ecology: Structure and Function of Running Waters. Kluwer
Academic Publishers, London.
Allegrucci, G., V. Todisco and V. Sbordoni. 2005. Molecular phylogeography of
Dolichopoda cave crickets (Orthoptera, Rhaphidophoridae): A scenario suggested
by mitochondrial DNA. Molecular Phylogenetics and Evolution 37: 153-164.
Andersson, M. 1994. Sexual Selection. Princeton University Press, New Jersey.
Bagarinao, T. 1992. Sulfide as an environmental factor and toxicant: Tolerance and
adaptations in aquatic organisms. Aquatic Toxicology 24: 21-62.
Bagarinao, T. and R.D. Vetter. 1989. Sulfide tolerance and detoxification in shallow
water marine fishes. Marine Biology 103: 291-302.
Bagarinao, T. and I. Lantin-Olaguer. 1999. The sulfide tolerance of milkfish and
tilapia in relation to fish kills in farms and natural waters in the Philippines.
Hydrobiologia 382: 137-150.
Barr, T.C. 1968. Cave ecology and the evolution of troglobites. Evolutionary Biology
2: 35-102.
Bateman, A. 1948. Intra-sexual selection in Drosophila. Heredity 2: 349-368.
Begon, M.E., J.L. Harper and C.R. Townsend. 1996. Ecology. Blackwell Science,
Oxford, 3rd edition.
Blais, J., C. Rico, C. Van Oosterhout, J. Cable, G.F. Turner and L. Bernatchez. 2007.
MHC adaptive divergence between closely related and sympatric African
cichlids. PLoS ONE 2:e734.
Brauner, C.J., C.L. Ballantyne, D.J. Randall and A.L. Val. 1995. Air-breathing in the
armored catfish (Hoplosternum littorale) as an adaptation to hypoxic, acidic, and
hydrogen sulfide-rich waters. Canadian Journal of Zoology—Revue Canadienne de
Zoologie 73: 739-744.
Burt de Perera, T. 2004. Spatial parameters encoded in spatial map of the blind
Mexican cave fish, Astyanax fasciatus. Animal Behaviour 68: 291-295.
Bush, A.O., J.C. Fernández, G.W. Esch and J.R. Seed. 2001. Parasitism: The Diversity
and Ecology of Animal Parasites. Cambridge University Press, Cambridge.
Bussing, W.A. 1998. Peces de las aguas continentales de Costa Rica—Freshwater
fishes of Costa Rica. Editorial de la Universidad de Costa Rica, San José.
Camassa, M.M. 2001. Responses to light in epigean and hypogean populations
of Gambusia affinis (Cyprinodontiformes: Poeciliidae). Environmental Biology of
Fishes 62: 115-118.
Carrico, R., W. Blumberg and J. Peisach. 1978. The reversible binding of oxygen to
sulfhemoglobin. Journal of Biological Chemistry 253: 7212-7215.

Chapman, L.J. and C.A. Chapman. 1993. Desiccation, flooding, and the behavior
of Poecilia gillii (Pisces: Poeciliidae). Ichthyological Exploration of Freshwaters 4:
279-287.
Chapman, L.J. and K.G. Hulen. 2001. Implications of hypoxia for the brain size and
gill morphometry of mormyrid fishes. Journal of Zoology 254: 461-472.
Chapman, L., F. Galis and J. Shinn. 2000. Phenotypic plasticity and the possible
role of genetic assimilation: Hypoxia-induced trade-offs in the morphological
traits of an African cichlid. Ecology Letters 3: 387-393.
Chapman, L.J. C.A. Chapman, D.A. Brazeau, B. McLaughlin and M. Jordan.
1999. Papyrus swamps, hypoxia, and faunal diversification: variation among
populations of Barbus neumayeri. Journal of Fish Biology 54: 310-327.
Chen, K. and J. Morris. 1972. Kinetics of oxidation of aqueous sulfide by O2.
Environmental Science and Technology 6: 529-537.
Cline, J. and F. Richards. 1969. Oxygenation of hydrogen sulfide in seawater at
constant salinity, temerature, and pH. Enviromental Science and Technology 3:
838-843.
Colaco, A., F. Dehairs and Desbruyeres. 2002. Nutritional relations of deep-sea
hydrothermal fields at the Mid-Atlantic ridge: a stable isotope approach. Deep-
Sea Research Part I—Oceanographic Research Papers 49: 395-412.
Culver, D.C. 1982. Cave Life—Evolution and Ecology. Harvard Uiversity Press,
Cambridge.
Curtis, C., T. Bartholomew, F. Rose and K. Dodgson. 1972. Detoxication of sodium
35
S-sulphide in the rat. Biochemical Pharmacology 21: 2313-2321.
Danielopol, D.L. and R. Rouch. 2005. Invasion, active versus passive.
In: Encyclopedia of Caves, D.C. Culver and W. White (eds.) Elsevier Academic
Press, Amsterdam.
Darnell, R.M. 1962. Fishes of the Rio Tamesi and related coastal lagoons in east-
central Mexico. Publications of the Institute of Marine Sciences, University of
Texas 8: 299-365.
Dries, L.A. 2003. Peering through the looking glass at a sexual parasite: Are
Amazon mollies Red Queens? Evolution 57: 1387-1396.
Endler, J.A. 1987. Predation, light-intensity and courtship behavior in Poecilia
reticulata (Pisces, Poeciliidae). Animal Behaviour 35: 1376-1385.
Fenolio, D., G. Graening, B. Collier and J. Stout. 2006. Coprophagy in a cave-adapted
salamander: The importance of bat guano examined through nutritional and
stable isotope analyses. Proceedings of the Royal Society B-Biological Sciences 273:
439-443.
Franssen, C., M. Tobler, R. Riesch, F.J. García de León, R. Tiedemann, I. Schlupp
and M. Plath. 2008. Sperm production in an extremophile fish, the cave molly
(Poecilia mexicana Poeciliidae, Teleostei). Aquatic Ecology 42: 685-692.
Funk, D.J. 1998. Isolating a role for natural selection in speciation: Host adaptation
and sexual isolation in Neochlamisius bebbianae leaf beetles. Evolution 52: 1744-
1759.
Gibert, J. and L. Deharveng. 2002. Subterranean ecosystems: A truncated functional
biodiversity. Bioscience 52: 473-481.

Gordon, M.S. and D.E. Rosen. 1962. A cavernicolous form of the Poeciliid fish
Poecilia sphenops from Tabasco, México. Copeia 1962: 360-368.
Graening, G. and A. Brown. 2003. Ecosystem dynamics and pollution effects in
an Ozark cave stream. Journal of the American Water Resources Association 39:
1497-1507.
Graham, J. 2005. Aquatic and aerial respiration. In: The Physiology of Fishes, D.
Evans and J. Claiborne (eds.). CRC Press, London.
Grether, G.F. 2000. Carotenoid limitation and mate preference evolution: a test of
the indicator hypothesis in guppies (Poecilia reticulata). Evolution 54: 1712-1724.
Grieshaber, M.K. and S. Völkel. 1998. Animal adaptations for tolerance and
exploitation of poisonous sulfide. Annual Review of Physiology 60: 33-53.
Harvey, P. and J. Bradbury. 1991. Sexual selection. In: Behavioural Ecology: An
Evolutionary Approach, J. Krebs and N. Davies (eds.). Blackwell, Oxford, pp.
203-233.
Hatfield, T. and D. Schluter. 1999. Ecological speciation in sticklebacks: Environment
dependent hybrid fitness. Evolution 53: 866-873.
Higham, T.E., C.D. Hulsey, O. Rican and A.M. Carroll. 2007. Feeding with speed:
Prey capture evolution in cichlids. Journal of Evolutionary Biology 20: 70-78.
Hose, L. and J. Pisarowicz. 1999. Cueva de Villa Luz Tabasco, Mexico: reconnaissance
study of an active sulfur spring cave and ecosystem. Journal of Cave and Karst
Studies 61: 13-21.
Hose, L., A.N. Palmer, M. Palmer, D.E. Northup, P.J. Boston and H.R. DuChene.
2000. Microbiology and geochemistry in a hydrogen-sulphide-rich karst
environment. Chemical Geology 169: 399-423.
Houde, A. 1997. Sex, Color, and Mate Choice in Guppies. Princeton University Press,
Princeton.
Howarth, F.G. 1987. The evolution of non-relictual tropical troglobites. International
Hubbs, C. 1959. Population analysis of a hybrid swarm between Gambusia affinis
and G. heterochir. Evolution 13: 236-246.
Hulsey, C.D. and F. Garcia de Leon. 2005. Cichlid jaw mechanics: Linking
morphology to feeding specialization. Functional Ecology 19: 487-494.
Humphreys, W.F. 1993. Cave fauna in semi-arid tropical Western Australia: A
diverse relict wet-forest litter fauna. Mémoires de Biospéologie 20: 105-110.
Hüppop, K. 2000. How do cave animals cope with the food scarcity in caves?
In: Ecosystems of the World 30: Subterranean Ecosystems, H. Wilkens, D.C. Culver
and W.F. Humphries (eds.). Elsevier Science, Amsterdam.
Ip, Y.K., S.S.L. Kuah and S.F. Chew. 2004. Strategies adopted by the mudskipper
Boleophthalmus boddaerti to survive sulfide exposure in normoxia or hypoxia.
Physiological and Biochemical Zoology 77: 824-837.
Jeyasingh, P.D. 2007. Plasticity in metabolic allometry: The role of dietary
stoichiometry. Ecology Letters 10: 282-289.
Johnson, P., F. Hoppensteadt, J. Smith and G. Bush. 1996. Conditions for sympatric
speciation: A diploid model incorporating habitat fidelity and non-habitat
assortative mating. Evolutionary Ecology 10: 187-205.

Jorgensen, B.B. 1982. Ecology of the bacteria of the sulphur cycle with special
reference to anoxic/oxic interface environments. Philosophical Transactions of the
Royal Society of London Series B-Biological Sciences 298: 543-561.
Jorgensen, B.B. 1984. The microbial sulfur cycle. In: Microbial Geochemistry, W.
Krumbein (ed.). Blackwell Science Publishers, Oxford.
Jorgensen, B.B. and T. Fenchel. 1974. The sulphur cycle of a marine sediment model
system. Marine Biology 24: 189-201.
Kiester, R. 1979. Conspecifics as cues: a mechanism for habitat selection in the
Panamanian grass anole (Anolis auratus). Behavioral Ecology and Sociobiology 5:
323-330.
Kirkpatrick, M. and M.J. Ryan. 1991. The evolution of mating preferences and the
paradox of the lek. Nature 350: 33-38.
Kittell, M., M. Harvey, S. Contreras-Balderas and M.B. Ptacek. 2005. Wild-
caught hybrids between sailfin and shortfin mollies (Poeciliidae, Poecilia):
morphological and molecular verification. Hidrobiologica 15: 131-137.
Körner, K.E., I. Schlupp, M. Plath and E.R. Loew. 2006. Spectral sensitivity of
mollies: comparing surface- and cave-dwelling Atlantic mollies, Poecilia
mexicana. Journal of Fish Biology 69: 54-65.
Kramer, D. and M. Bryant. 1995. Intestine length in the fishes of a tropical stream: 2.
Relationships to diet—the long and short of a convoulted issue. Environmental
Kramer, D.L. 1983. The evolutionary ecology of respiratory modes in fishes: An
analysis based on the costs of breathing. Environmental Biology of Fishes 9: 145-
158.
Krause, J. and G. Ruxton. 2002. Living in Groups. Oxford University Press, Oxford.
Krause, J., N. Hartmann and V. Pritchard. 1999. The influence of nutritional state
on shoal choice in zebrafish, Danio rerio. Animal Behaviour 57: 771-775.
Langecker, T.G. 1989. Studies on the light reaction of epigean and cave populations
of Astyanax fasciatus (Characidae, Pisces). Mémoires de Biospéologie 16: 169-176.
Langecker, T.G. 2000. The effect of continuous darkness on cave ecology and
cavernicolous evolution. In. Ecosystems of the World 30: Subterranean Ecosystems,
H. Wilkens, D.C. Culver and W.F. Humphreys (eds.). Elsevier Science,
Amsterdam.
Langecker, T.G., H. Wilkens and J. Parzefall. 1996. Studies on the trophic structure
of an energy-rich Mexican cave (Cueva de las Sardinas) containing sulfurous
water. Mémoires de Biospéologie 23: 121-125.
Langerhans, R.B., M. Gifford and E. Joseph. 2007. Ecogical speciation in Gambusia
fishes. Evolution 61: 2056-2074.
Lovatt Evans, C. 1967. The toxicity of hydrogen sulphide and other sulphides.
Quarterly Journal of Experimental Physiology 52: 231-248.
Luther, G.W., S. Ma, R. Trouborst, B. Glazer, M. Blickley, R.W. Scarborough and
M.G. Mensinger. 2004. The roles of anoxia, H2S, and storm events in fish kills of
dead-end canals of Delaware inland bays. Estuaries 27: 551-560.

Macalady, J.L., E.H. Lyon, B. Koffman, L.K. Albertson, K. Meyer, S. Galdenzi and S.
Mariani. 2006. Dominant microbial populations in limestone-corroding stream
biofilms, Frasassi cave system, Italy. Applied and Environmental Microbiology 72:
5596-5609.
MacArthur, R. and E. Pianka. 1966. On optimal use of patchy environment.
American Naturalist 100: 603-609.
MacArthur, R., J. Diamond and J. Karr. 1972. Density compensation in island
faunas. Ecology 53: 330-342.
Magurran, A.E. and B.H. Seghers. 1994. Sexual conflict as a consequence of ecology:
Evidence from Guppy, Poecilia reticulata, populations in Trinidad. Proceedings of
the Royal Society of London Series B-Biological Sciences 255: 31-36.
Meffe, G.K. and F.F. Snelson (eds.). 1989. Ecology and Evolution of Lifebearing Fishes
(Poeciliidae). Prentice Hall, New Jersey.
Menke, A. 1979. Family Belostomatidae—Giant water bugs. In: The Semiaquatic
and Aquatic Hemiptera of California (Heteroptera: Hemiptera), A. Menke (ed.).
University of California Press, Berkeley.
Miller, R.R. 1983. Checklist and key to the mollies of Mexico (Pisces, Poeciliidae,
Poecilia, Subgenus Mollienesia). Copeia 1983: 817-822.
Miller, R.R. 2005. Freshwater Fishes of Mexico. University of Chicago Press,
Chicago.
Nakatsuru, K. and D.L. Kramer. 1982. Is sperm cheap? Limited male-fertility and
female choice in the lemon tetra (Pisces, Characidae). Science 216: 753-755.
National Research Council. 1979. Hydrogen Sulfide. University Park Press,
Baltimore.
Nicholls, P. 1975. The effect of sulphide on cytochrome aa3. Isosteric and allosteric
shifts of the reduced alpha-peak. Biochemistry Biophysics Acta 396: 24-35.
Nosil, P., T.H. Vines and D.J. Funk. 2005. Perspective: Reproductive isolation caused
by natural selection against immigrants from divergent habitats. Evolution 59:
705-719.
Nybakken, J. 2001. Marine Biology, an Ecological Approach. Benjamin Cummings,
San Francisco.
Park, C., R. Nagel, W. Blumberg, J. Peisach and R. Maliozzo. 1986. Sulfhemoglobin:
Properties of partially sulfurated tetramers. Journal of Biological Chemistry 261:
8805-8810.
Parzefall, J. 1969. Zur vergleichenden Ethologie verschiedener Mollienesia-Arten
einschliesslich einer Höhlenform von M. sphenops. Behaviour 33: 1-37.
Parzefall, J. 1970. Morphologische Untersuchungen an einer Höhlenform von
Mollienesia sphenops (Pisces, Poeciliidae). Zeitschrift für Morphologie der Tiere 68:
323-342.
Parzefall, J. 1973. Attraction and sexual cycle of poeciliids. In: Genetics and
Mutagenesis of Fish, J. Schroeder (ed.). Springer Verlag, Berlin.
Parzefall, J. 1974. Rückbildung aggressiver Verhaltensweisen bei einer Höhlenform
von Poecilia sphenops (Pisces, Poeciliidae). Zeitschrift für Tierpsychologie 35: 66-
84.

Parzefall, J. 1979a. Genetics and biological significance of the aggressive behavior

of Poecilia sphenops (Pisces, Poeciliidae): Studies on hybrids of epigeous and
hypogeous living populations. Zeitschrift für Tierpsychologie 50: 399-422.
Parzefall, J. 1979b. Zur Genetik und biologischen Bedeutung des
Aggressionsverhaltens von Poecilia sphenops (Pisces, Poeciliidae). Zeitschrift für
Tierpsychologie 50: 399-422.
Parzefall, J. 1993. Behavioural ecology of cave-dwelling fishes. In: Behaviour of
Teleost Fishes, T.J. Pitcher (ed.). Chapman and Hall, London, 2nd edition.
Parzefall, J. 2001. A review of morphological and behavioural changes in the cave
molly, Poecilia mexicana, from Tabasco, Mexico. Environmental Biology of Fishes
62: 263-275.
Parzefall, J., C. Kraus, M. Tobler and M. Plath. 2007. Photophilic behaviour in
surface- and cave-dwelling Atlantic mollies Poecilia mexicana (Poeciliidae).
Journal of Fish Biology 71: 1225-1231.
Peek, A.S., R. Feldmann, R. Lutz and R.C. Vrijenhoek. 1998. Conspeciation of
chemoautotrophic bacteria and deep-sea clams. Proceedings of the National
Academy of Sciences USA 95: 9962-9966.
Peters, N. and G. Peters. 1968. Zur genetischen Interpretation morphologischer
Gesetzmässigkeiten der degenerative Evolution. Zeitschrift für Morphologie der
Tiere 62: 211-244.
Peters, N., G. Peters, J. Parzefall and H. Wilkens. 1973. Über degenerative und
konstruktive Merkmale bei einer phylogenetisch jungen Höhlenform von
Poecilia sphenops (Pisces, Poeciliidae). Internationale Revue der Gesamten
Hydrobiologie 58: 417-436.
Petersen, L. 1977. The effect of inhibitors on the oxygen kinetics of cytochome c
oxidase. Biochimica et Biophysica Acta 460: 299-307.
Plath, M. 2008. Male mating behavior and costs of sexual harassment for females
in cavernicolous and extremophile populations of Atlantic mollies (Poecilia
mexicana). Behaviour 145: 73-98.
Plath, M. and M. Tobler. 2007. Sex recognition in surface- and cave-dwelling
Atlantic molly females (Poecilia mexicana, Poeciliidae, Teleostei): Influence of
visual and non-visual cues. Acta Ethologica 10: 81-88.
Plath, M. and I. Schlupp. 2008. Parallel evolution leads to reduced shoaling
behavior in two cave dwelling populations of Atlantic mollies (Poecilia mexicana,
Poeciliidae, Teleostei). Environmental Biology of Fishes 82: 289-297.
Plath, M., K.E. Korner, I. Schlupp and J. Parzefall. 2001. Sex recognition and female
preferences of Cave mollies Poecilia mexicana (Poeciliidae, Teleostei) in light in
darkness. Mémoires de Biospéologie 28: 163-167.
Plath, M., K. Körner, J. Parzefall and I. Schlupp. 2003a. Persistence of a visually
mediated mating preference in the cave molly, Poecilia mexicana (Poeciliidae,
Teleostei). Subterranean Biology 1: 93-97.
Plath, M., J. Parzefall and I. Schlupp. 2003b. The role of sexual harassment in
cave- and surface-dwelling populations of the Atlantic molly, Poecilia mexicana
(Poeciliidae, Teleostei). Behavioral Ecology and Sociobiology 54: 303-309.

Plath, M., M. Arndt, J. Parzefall and I. Schlupp. 2004a. Size-dependent male

mating behaviour in the Cave molly, Poecilia mexicana (Poeciliidae, Teleostei).
Subterranean Biology 2: 59-64.
Plath, M., A. Brümmer and I. Schlupp. 2004b. Sexual harassment in a livebearing
fish (Poecilia mexicana): Influence of population-specific male mating behaviour.
Acta Ethologica 7: 65-72.
Plath, M., J. Parzefall, K. Körner and I. Schlupp. 2004c. Sexual selection in darkness?
Female mating preferences in surface- and cave-dwelling Atlantic mollies,
Poecilia mexicana (Poeciliidae, Teleostei). Behavioral Ecology and Sociobiology 55:
596-601.
Plath, M., M. Tobler and I. Schlupp. 2004d. Cave fish looking for mates: A visual
mating preference in surface- and cave-dwelling Atlantic mollies (Poecilia
mexicana, Poeciliidae). Zeitschrift für Fischkunde 7: 61-69.
Plath, M., K.U. Heubel, F.G. de Leon and I. Schlupp. 2005a. Cave molly females
(Poecilia mexicana, Poeciliidae, Teleostei) like well-fed males. Behavioral Ecology
and Sociobiology 58: 144-151.
Plath, M., K.U. Heubel and I. Schlupp. 2005b. Field observations on male mating
behavior in surface- and cave-dwelling Atlantic mollies (Poecilia mexicana,
Poeciliidae). Zeitschrift für Fischkunde 7: 113-119.
Plath, M., A. Brümmer, J. Parzefall and I. Schlupp. 2006a. Size-dependent male
mating behaviour and sexual harassment in a population of Atlantic mollies
(Poecilia mexicana) from a sulphur creek. Acta Ethologica 9: 15-21.
Plath, M., M. Rohde, T. Schroder, A. Taebel-Hellwig and I. Schlupp. 2006b. Female
mating preferences in blind cave tetras Astyanax fasciatus (Characidae, Teleostei).
Behaviour 143: 15-32.
Plath, M., U. Seggel, H. Burmeister, K.U. Heubel and I. Schlupp. 2006c. Choosy
males from the underground: Male mating preferences in surface- and cave-
dwelling Atlantic mollies (Poecilia mexicana). Naturwissenschaften 93: 103-109.
Plath, M., S. Hauswaldt, K. Moll, M. Tobler, F. Garcia de Leon, I. Schlupp and R.
Tiedemann. 2007a. Local adaptation and pronounced genetic differentiation in
an extremophile fish, Poecilia mexicana, inhabiting a Mexican cave with toxic
hydrogen sulfide. Molecular Ecology 16: 967-976.
Plath, M., A. Makowicz, I. Schlupp and M. Tobler. 2007b. Sexual harassment in
live-bearing fishes (Poeciliidae): Comparing courting and noncourting species.
Behavioral Ecology 18: 680-688.
Plath, M., M. Tobler, F.J. Garcia de Leon, O. Giere and I. Schlupp. 2007c.
Survival in an extreme habitat: the role of behaviour and energy limitation.
Naturwissenschaften 94: 991-996.
Porter, M. and K. Crandall. 2003. Lost along the way: The significance of evolution
in reverse. Trends in Ecology & Evolution 18: 541-547.
Poulson, T.L. 1963. Cave adaptation in amblyopsid fishes. American Midland
Poulson, T.L. and W. B. White. 1969. The cave environment. Science 165: 971-981.

Poulson, T.L. and K.H. Lavoie. 2000. The trophic basis of subterranean ecosystems.
In: Ecosystems of the World 30: Subterranean Ecosystems, H. Wilkens. D.C. Culver
and W.F. Humphries (eds.). Elsevier Science, Amsterdam.
Proudlove, G. 2006. An account of the subterranean (hypogean) fishes described up
to 2003 with a bibliography 1541-2004. International Society for Subterranean
Biology, Moulis.
Ptacek, M.B. 2002. Patterns of inheritance of mating signals in interspecific hybrids
between sailfin and shortfin mollies (Poeciliidae: Poecilia: Mollienesia). Genetica
116: 329-342.
Reiffenstein, R., W. Hulbert and S. Roth. 1992. Toxicology of hydrogen sulfide.
Annual Reviews of Pharmacology and Toxicology 1992: 109-134.
Reis, R., E. Trajano and E. Hingst-Zaher. 2006. Shape variation in surface and cave
populations of the armoured catfishes Ancistrus (Siluriformes: Loricariidae)
from the Sao Domingos karst area, upper Tocantins River, Brazil. Journal of Fish
Biology 68: 414-429.
Reznick, D. and J. Endler. 1982. The impact of predation on life history evolution in
Trinidadian guppies (Poecilia reticulata). Evolution 36: 160-177.
Reznick, D., H. Bryga and J.A. Endler. 1990. Experimentally induced life-history
evolution in a natural population. Nature 346: 357-359.
Reznick, D., M.J. Butler and H. Rodd. 2001. Life-history evolution in guppies. VII.
The comparative ecology of high- and low-predation environments. American
Rice, W. and G. Salt. 1990. The evolution of reproductive isolation as a correlated
character under sympatric conditions: Experimental evidence. Evolution 44:
1140-1152.
Riesch, R., I. Schlupp and M. Plath. 2006a. Influence of male competition on male
mating behaviour in the cave molly (Poecilia mexicana, Poeciliidae, Teleostei).
Journal of Ethology 24: 27-31.
Riesch, R., M. Tobler, F.J. Garcia de Leon, I. Schlupp and M. Plath. 2006b. Reduction
of the association preference for conspecifics in surface- and cave-dwelling
Atlantic mollies, Poecilia mexicana. Behavioral Ecology and Sociobiology 60: 794-
802.
Riesch, R., V. Duwe, N. Herrmann, L. Padur, A. Ramm, K. Scharnweber, M. Schulte,
T. Schulz–Mirbach, M. Ziege and M. Plath. 2009a. Variation along the shy-bold
continuum in extremophile fishes (Poecilia mexicana, P. sulphuraria). Behavioral
Riesch, R., M. Tobler, I. Schlupp and M. Plath. 2009b. Reduced offspring number
in the cave molly (Poecilia mexicana, Poeciliidae). Environmental Biology of Fishes
84: 89-94.
Riesch, R., M. Plath, F.J. García de León and I. Schlupp. In press-a. Convergent
life–history shifts: toxic environments result in big babies in two clades of
poeciliids. Naturwissenschaften.
Riesch, R., M. Plath and I. Schlupp. In press-b. Toxic hydrogen sulfide and dark
caves: life history adaptations to extreme environments in a livebearing fish
(Poecilia mexicana, Poeciliidae). Ecology.

Romero, A. and S.M. Green. 2005. The end of regressive evolution: Examining and
interpreting the evidence from cave fishes. Journal of Fish Biology 67: 3-32.
Romero, A., A. Singh, A. McKie, M. Manna, R. Baker, K.M. Paulson and J.E.
Creswell. 2002. Replacement of the troglomorphic population of Rhamdia
quelen (Pisces: Pimelodidae) by an epigean population of the same species in
the Cumaca Cave, Trinidad, West Indies. Copeia 2002: 938-942.
Rosen, D. and R. Bailey. 1963. The poeciliid fishes (Cyprinodontiformes), their
structure, zoogeography and systematics. Bulletin of the American Museum of
Natural History 126: 1-176.
Rosenthal, G. G., X. de la Rosa, S. Kazianis, M.J. Stephens, D.C. Morizot, M.J. Ryan
and F.J. Garcia de Leon. 2003. Dissolution of sexual signal complexes in a hybrid
zone between the swordtails Xiphophorus birchmanni and Xiphophorus malinche
(Poeciliidae). Copeia 2003: 299-307.
Rundle, H.D. and P. Nosil. 2005. Ecological speciation. Ecology Letters 8: 336-352.
Ryan, M.J. and A. Keddy-Hector. 1992. Directional patterns of female mate choice
and the role of sensory biases. American Naturalist 139: S4-S35.
Ryan, P., P. Bloomer, C. Moloney, T. Grant and W. Delport. 2007. Ecological
speciation in south Atlantic island finches. Science 315: 1420-1422.
Sarbu, S.M., T.C. Kane and B.K. Kinkle. 1996. A chemoautotrophically based cave
ecosystem. Science 272: 1953-1955.
Schartl, M. 1995. Platyfish and swordtails: A genetic system for the analysis of
molecular mechanisms in tumor formation. Trends in Genetics 11: 185-189.
Schilthuizen, M., A. Cabanban and M. Haase. 2005. Possible speciation with gene
flow on tropical cave snails. Journal of Zoological Systematics and Evolutionary
Research 43: 133-138.
Schluter, D. 2000. The Ecology of Adaptive Radiation. Oxford University Press,
Oxford.
Schluter, D. and J. McPhail. 1992. Ecological character displacement and speciation
in sticklebacks. American Naturalist 140: 85-108.
Schulz-Mirbach, T., C. Stransky, J. Schlickeisen and B. Reichenbacher. 2008.
Differences in otolith morphologies between surface- and cave-dwelling
populations of Poecilia mexicana (Teleostei, Poeciliidae) reflect adaptations to
life in an extreme habitat. Evolutionary Ecology Research 10: 1-22.
Schoener, T. 1971. Theory of feeding strategies. Annual Review of Ecology and
Systematics 2: 369-404.
Spellman, S.J. and A.D. Johnson. 1987. In vitro encystment of the black spot
trematode Uvulifer ambloplitis (Trematoda, Diplostomatidae). International
Journal of Parasitology 17: 897-902.
Streever, W.J. 1996. Energy economy hypothesis and the troglobitic crayfish
Procambarus erythrops in Sim’s sink cave, Florida. American Midland Naturalist
135: 357-366.
Swart, C. and B. Felgenhauer. 2003. Structure and function of the mouthparts
and salivary gland complex of the Giant Waterbug, Belostoma lutarium (Stal)
(Hemiptera: Belostomatidae). Annals of the Entomological Society of America 96:
870-882.

Timmerman, C.M. and L.J. Chapman. 2004. Hypoxia and interdemic variation in
Poecilia latipinna. Journal of Fish Biology 65: 635-650.
Tobler, M., H. Burmeister, I. Schlupp and M. Plath. 2006a. Regressive evolution of
visually mediated preferences in the Cave molly (Poecilia mexicana, Poeciliidae,
Teleostei). Subterranean Biology 4: 59-65.
Tobler, M., I. Schlupp, K. Heubel, R. Riesch, F.J. Garcia de Leon, O. Giere and M.
Plath. 2006b. Life on the edge: Hydrogen sulfide and the fish communities of a
Mexican cave and surrounding waters. Extremophiles 10: 577-585.
Tobler, M., I. Schlupp, F.J. Garcia de Leon, M. Glaubrecht and M. Plath. 2007a.
Extreme habitats as refuge from parasite infections? Evidence from an
extremophile fish. Acta Oecologica 31: 270-275.
Tobler, M., I. Schlupp and M. Plath. 2007b. Predation of a cave fish (Poecilia
mexicana, Poeciliidae) by a giant water-bug (Belostoma, Belostomatidae) in a
Mexican sulfur cave. Ecological Entomology 32: 492-495.
Tobler, M. 2008. Divergence in trophic ecology characterises colonisation of extreme
habitats. Biological Journal of the Linnean Society 95: 517-528.
Tobler, M., I. Schlupp, R. Riesch, F.J. Garcia de Leon and M. Plath. 2008a. A new
and morphologically distinct cavernicolous population of Poecilia mexicana
(Poeciliidae, Teleostei). Environmental Biology of Fishes 82: 101-108.
Tobler, M., T.J. DeWitt, I. Schlupp, F.J. Garcia de Leon, R. Herrmann, P. Feulner,
R. Tiedemann and M. Plath. 2008b. Toxic hydrogen sulfide and dark caves:
phenotypic and genetic divergence across two abiotic environmental gradients
in Poecilia mexicana. Evolution 62: 2643-2659.
Tobler, M., I. Schlupp and M. Plath. 2008c. Does divergence in female mate choice
affect male size distributions in two cave fish populations? Proceedings of the
Royal Society London: Biology Letters 4: 452-454.
Tobler, M., R. Riesch, F.J. Garcia de Leon, I. Schlupp and M. Plath. 2008d. Two
endemic and endangered fishes, Poecilia sulphuraria (Alvarez, 1948) and
Gambusia eurystoma Miller, 1975 (Poeciliidae, Teleostei), as only survivors in a
small sulfidic habitat. Journal of Fish Biology 72: 523-533.
Tobler, M., C. Franssen and M. Plath. 2008e. Male-biased predation on a cave fish
by a giant water bug. Naturwissenschaften 95: 775-779.
Tobler, M., R. Riesch, C.M. Tobler and M. Plath. 2009. Compensatory behavior in
response to sulfide–induced hypoxia affects time budgets, feeding efficiency,
and predation risk. Evolutionary Ecology Research 11: 935-948.
Torrans, E. and H. Clemens. 1982. Physiological and biochemical effects of acute
exposure of fish to hydrogen sulfide. Comparative Biochemistry and Physiology
71C: 183-190.
Trajano, E. 1995. Evolution of tropical troglobites: Applicability of the model of
Quaternary climatic fluctuations. Mémoires de Biospéologie 22: 203-209.
Trexler, J., R. Tempe and J. Travis. 1994. Size-selective predation of Sailfin mollies
by two species of heron. Oikos 69: 250-259.
Trivers, R. 1972. Parental investment and sexual selection. In: Sexual Selection and
the Descent of Man. B. Campbell (ed.). Aldine Publishing Company, Chicago.

Van Dover, C.L. 2000. The Ecology of Deep-sea Hydrothermal Vents. Princeton
University Press, Princeton.
Wainwright, P.C. 1996. Ecological explanation through functional morphology:
The feeding biology of sunfishes. Ecology 77: 1336-1343.
Walters, L. and V. Walters. 1965. Laboratory observations on a cavernicolous
poeciliid from Tabasco, Mexico. Copeia 1965: 214-233.
Weber, J.M. and D.L. Kramer. 1983. Effects of hypoxia and surface access on growth,
mortality, and behavior of juvenile Guppies, Poecilia reticulata. Canadian Journal
of Fisheries and Aquatic Sciences 40: 1583-1588.
Wilkens, H. 1973. Über das phylogenetische Alter von Höhlentieren:
Untersuchungen über die cavernicole Süsswasserfauna Yucatans. Zeitschrift für
zoologische Systematik und Evolutionsforschung 11: 49-60.
Wilkens, H. 1979. Reduktionsgrad und phylogenetisches Alter: Ein Beitrag zur
Besiedlungsgeschichte der Limnofauna Yukatans. Zeitschrift fur zoologische
Systematik und Evolutionsforschung 17: 262-272.
Willis, L. and A. Brown. 1985. Distribution and habitat requirements of the Ozark
cavefish, Amblyopsis rosae. American Midland Naturalist 114: 311-317.
Winemiller, K.O. 1993. Seasonality of reproduction by live-bearing fishes in tropical
rainforest streams. Oecologia 95: 266-276.
Wolf, N. 1985. Odd fish abandon mixed-species groups when threatened. Behavioral
Zeiske, E. 1968. Prädispositionen bei Mollienesia sphenops (Pisces, Poeciliidae) für
einen Übergang zum Leben in subterranen Gewässer. Zeitschrift für Vergleichende
Physiologie 58: 190-222.

CHAPTER 9
Subterranean Fishes of Brazil
Eleonora Trajano1 and Maria Elina Bichuette2
1
Departamento de Zoologia, Instituto de Biociências,
Universidade de São Paulo, São Paulo, SP, Brazil
E-mail: etrajano@usp.br
2
Departamento de Ecologia e Biologia Evolutiva,
Universidade Federal de São Carlos, São Carlos, SP, Brazil
E-mail: bichuette@uol.com.br
INTRODUCTION
Brazilian subterranean ichthyofauna is remarkable worldwide not only due
to the species richness, including troglomorphic and non-troglomorphic
fishes, but also for the high ecological and evolutionary diversity, without
a taxonomic correlation. More than 25 troglomorphic species showing
different degrees of morphological specialization are currently known to
occupy a large diversity of habitats, from epikarst to fast flowing base-
level streams, in different karst regions. Such areas are submitted to
distinct, sometimes contrasting climates, thus subterranean fishes must
adapt to different ecological conditions, varying from accentuated food
shortage, frequently seasonal, to a relative climatic stability and, in some
cases, abundance of food resources.
The first known Brazilian troglobite was the heptapterid Pimelodella
kronei, from the Upper Ribeira karst area, southern São Paulo State (SE
Brazil), described as Typhlobagrus kronei Ribeira, 1907. The blind catfish
from Iporanga was also the first South American subterranean organism
studied in great detail, by Pavan (1945) and later on, by Trajano (1987)
for their PhD theses, which included morphology, behavior and ecology
analyzed within an evolutionary context. Therefore, this troglobitic

fish represents an important historical reference in the development of

Brazilian Subterranean Biology.
In the last two decades, studies on the taxonomy, natural history,
population ecology and behavior of Brazilian troglobitic fishes have
multiplied, encompassing several species living in different kinds of
subterranean habitats and showing different degrees of reduction of eyes
and melanic pigmentation (classical “troglomorphisms”, Trajano 2005).
However, this still represents a relatively small part of what there is to be
investigated.
Herein, we present an overview on the state of art of the knowledge
on Brazilian subterranean fishes, including both troglomorphic
(morphologically specialized), probably troglobitic (exclusively
subterranean) species, and non-troglomorphic, either troglophilic,
trogloxenic or accidental fishes in caves.
I. BIODIVERSITY
An updated list of Brazilian subterranean troglomorphic fishes, with data
on localities, predominant type of habitat (according to the classification in
Trajano 2001a) and degree of reduction of eyes and melanic pigmentation,
is presented in Table 1. As a consequence of the advancement of
speleological activities throughout the country, information on “pale fish”
is continuously arriving, and confirmed new records on subterranean
species are added in a rate of 1-2 each year. Therefore, the present list and
related inferences about patterns must be regarded as a current status of
the knowledge on Brazilian subterranean fishes.
I.1 Taxonomic Diversity
I.1.1 Troglomorphic and Found Exclusively in Subterranean

Habitats: Brazilian Troglobitic Fishes
The neotropical freshwater ichthyofauna is one of the most diversified in
the world. Among the two dominant taxa, the Siluriformes (catfish) are
mainly nocturnal and chemo-oriented, whereas most Characiformes are
diurnal, oriented predominantly by vision. Therefore, it is not surprising
the fact that the great majority of neotropical troglobitic fishes, including
those found in Brazil, are siluriforms.
So far, 25 species of troglomorphic fishes, i.e., species with eyes and
pigmentation reduced at least at some degree beyond that observed in
their epigean (surface) congeners, indicating a troglobitic (exclusively
subterranean) status, are known for Brazil (Table 1). All but two species are

Table 1 Brazilian troglobitic fishes, with localities (karst area/county, State), types of habitat (according to classification in Trajano
2001) and degree of troglomorphism. 1. slightly, but significantly reduced eyes/melanic pigmentation, with narrow intrapopulation
variation; 2. wide variation, from slightly reduced to not visible externally eyes/pigmentation; 3. most individuals with deeply
reduced eyes/pigmentation; 4. all individuals completely (externally) anophthalmic and/or pigmented; 4+. DOPA positive (see text);
4– DOPA negative; Polim. Polymorphic phenotypes
Taxon Locality Habitat Eyes Pigm.
CHARACIFORMES, incertae sedis
1 Stygichthys typhlops Jaíba, MG Upper phreatic zone (fissures) 4 4+
GYMNOTIFORMES
2 Eigenmannia vicentespelaea São Domingos, GO Base-level stream 1 1
Eleonora Trajano and Maria Elina Bichuette 333

SILURIFORMES
LORICARIIDAE
3 Ancistrus cryptophthalmus São Domingos, GO Base-level streams 2 2
4 Ancistrus formoso Bonito, MS Flooded cave 4 4
CALLICHTHYIDAE
5 Aspidoras sp. Posse, GO Base level streams 2 1
TRICHOMYCTERIDAE:
6 Trichomycterus itacarambiensis Itacarambi, MG Base-level stream 2 Polim.
7 Trichomycterus sp.1* Bonito, MS Flooded caves 3 3
8 Trichomycterus sp. 2 Serra do Ramalho, BA Vadose tributaries 3 3
9 Ituglanis passensis São Domingos, GO Base-level stream 3 3
10 Ituglanis bambui idem Vadose tributaries 3 3
11 Ituglanis epikarsticus idem Epikarst 3 3
12 Ituglanis ramiroi dem Vadose tributaries 3 3
Table 1 Contd..

Table 1 Contd..
13 Ituglanis mambai** Posse, GO Base-level stream 1 1
14 Ituglanis sp. Serra do Bodoquena, MS Base-level + Vadose streams 2 2
15 Glaphyropoma spinosum Igatu, BA Vadose tributary 2 3
16 Copionodon sp. Igatu, BA Vadose tributary 2 2
HEPTAPTERIDAE
17 Pimelodella kronei Alto Ribeira, SP Base-level streams 3 2
18 Pimelodella spelaea São Domingos, GO Vadose tributary 1 1
19 Rhamdia enfurnada Serra do Ramalho, BA Base-level streams 2 2
20 Rhamdia sp. Serra da Bodoquena, MS Base-level streams 3 2
21 Rhamdiopsis sp. 1*** Chapada Diamantina, BA Upper phreatic zone (caves) 4 4+
22 Rhamdiopsis sp. 2 (= Taunayia sp.) Campo Formoso, BA Upper phreatic zone (cave) 4 4–
23 Rhamdiopsis sp. 3 Cordisburgo, MG Base-level stream 1 1
24 Phreatobius cisternarum Amazonas delta (PA/AM) Phreatic zone (alluvial fan) 3 3
25 Phreatobius dracunculus Rio Pardo (RO) Phreatic zone (alluvial fan) 4 4
* cited as Trichomycterus undesc. sp. 2 in Trajano (1997)
** cited as Trichomycterus sp. 3 in Trajano and Bichuette (2005)
*** cited as “new genus” in Trajano (1997, 2003)
siluriforms belonging to four families: the Trichomycteridae, Heptapteridae

and Loricariidae (in this order) also include the great majority of troglobitic
fishes found elsewhere in South and Central America, whereas the new
Callichthyidae, genus Aspidoras, represents the first record of a troglobitic
representative in the family. The only catfish family with troglobitic
derivatives in South America (Ecuador and Peru – see Proudlove, this
volume), which does not occur in Brazil, is Astroblepidae. Therefore, at
least for the troglobitic species, the Brazilian subterranean ichthyofauna
is representative of the general tendencies for the continental Neotropical
region.
The Trichomycteridae, with 11 troglobitic species in Brazil (four
genera – Trichomycterus, Ituglanis, Glaphyropoma and Copionodon), are well
represented in other South American countries (Trichomycterus spp. in
Venezuela, Colômbia, Bolívia and Argentina). Among the heptapterids (9
species in Brazil), the genus Rhamdia, with two troglomorphic species in
Brazil and one in Venezuela, encompasses the highest number of troglobitic
catfishes in Mexico. It is noteworthy that the Mexican species are related to
the epigean R. laticauda, whereas the Brazilian species is closely related to
R. quelen. The heptapterid genera Rhamdiopsis (3 species) and Pimelodella (2
species) do not have troglobitic derivatives in other neotropical countries;
Pimelodella catfishes are widely distributed in South America, but the
poorly known genus Rhamdiopsis would be restricted to Brazil.
The Amazonian genus Phreatobius encompasses two Brazilian
subterranean species sampled in artificial wells, Phreatobius cisternarum,
the first described troglomorphic species in the Neotropical region, and
P. dracunculus; in addition, at least three undescribed species were found
living in submerged litter banks at the margins of the Rio Negro. P.
cisternarum was object of a detailed anatomical study by Reichel (1927),
without any similar. This species is widely distributed around the Rio
Amazonas delta, from northern State of Amapá to Belém, Pará State, also
occurring in the island of Marajó. The recently described P. dracunculus
was also collected in wells, in Rondônia, at the right margin of the Rio
Amazonas. A third subterranean species accessible through wells, P.
sanguijuela, was described from Bolívia, illustrating the wide distribution of
this genus. The systematic position of Phreatobius within the Siluriformes is
under study using molecular techniques; apparently it forms a trichotomy
within the Pimelodoidea clade (Conorhynchos + Heptapteridae (Phreatobius
+ Pimelodidae + Pseudopimelodidae)) (Muriel-Cunha 2008).
The Loricariidae, by far the most speciose neotropical siluriform family,
are relatively poorly represented among the troglobitic fishes, with three
species (two in Brazil, one in Venezuela) belonging to the genus Ancistrus.
The armored catfishes are grazers, generally feeding on aquatic plants

and periphyton, therefore not expected to be frequent in subterranean

habitats. Nevertheless, some species would be able to survive and form
populations in caves rich in plant detritus, feeding on the thin layer of
finely particulate organic covering rocky substrates.
The Neotropical electric-fishes, Order Gymnotiformes (knifefishes), are
typically preadapted to the subterranean life, as they live in turbid waters
and/or have nocturnal activity, orienting by electroreception. Hence, it is
not surprising to find a troglomorphic species, Eigenmannia vicentespelaea,
in a cave system in Central Brazil, and several non-troglomorphic in other
caves (see below).
In addition to the troglomorphic catfishes and knifefish, there is one
Brazilian characiform, Stygichthys typhlops, whose familial position is still
uncertain. When described, with basis on a single specimen accidentally
collected during a well drilling operation in the early 1960’s and sent to
the USA, this species has been hypothesized to be a tetra characin (Brittan
and Bohlke 1965). No additional specimen was collected until 2003, when
a research team visited the type-locality, Jaíba, and obtained fish from two
artificial wells (there were reports on the past occurrence of “white fish”
in other springs and wells, which are currently dry – see Bichuette and
Trajano, this volume). Taxonomic studies on this material revealed that S.
typhlops is not a characin, but it familial position within the characiforms
is still unclear (Moreira et al., in press).
Brazilian subterranean ichthyofauna is distinguished worldwide by its
diversity at the familial (six families with troglobitic representatives) and
generic (12 genera) levels. This is probably due, at least in part, to the high
epigean fish diversity providing a great array of potential colonizers.
Nearly 2,600 freshwater fish species are estimated to occur in Brazil
(M.T. Pizza, pers. comm.), corresponding to 20 percent of the freshwater
species in the world (ca. 13,000 – Lévêque et al. 2008). Therefore, the
number of known Brazilian troglobitic fishes is relatively modest. This is
partly because the largest and richest region for the fish fauna, which is
the Amazon basin, is relatively poor in subterranean habitats suitable for
colonization and differentiation of hypogean populations.
The high diversity of Brazilian subterranean fishes when compared to
other countries may be due to one or all the following factors: 1) general
high ichthyofaunistic diversity in Brazil, providing a wide variety of
potential colonizers of hypogean habitats; 2) high research efforts than
in most other regions; and/or 3) the essentially comparative approach
by Brazilian researchers, using phylogenetic criteria for the recognition
of the troglobitic status, which is primarily a distributional attribute (the
restriction to the subterranean environment). Because it is not scientifically
possible to prove their absence from surface habitats, the second best

criterion for recognition of troglobites is the occurrence of differentiation,

expected to occur sometime after genetic isolation.
Therefore, in a preliminary classification, we consider troglobitic any
subterranean fish showing eyes significantly smaller than in the closest
known epigean relative (phenotypic plasticity renders pigmentation less
suitable for this purpose). Likewise, wider intrapopulation variability than
in epigean relatives indicates taxonomic identity. This approach, which
may inflate the list of Brazilian troglobitic fishes compared to those from
other countries, avoids the traps of subjectively establishing a minimum
degree of eye and pigmentation reduction acceptable for a “true” troglobite
– this is basically a gradistic approach, whose inadequacy for evolutionary
studies has long been demonstrated. Or worst, of confusing a small eye
with a reduced eye and, conversely, a relatively large eye with a non-
reduced one.
I.1.2 Troglomorphic, but not Found in Typical Subterranean

Habitats
As observed in other large rivers, such as those in the Congo basin
(Thinès and Proudlove 1986), troglomorphic fishes have also been
found in epigean habitats in the Amazonas basin. Such fish may not be
classified as troglobites, which are, by definition, inhabitants of hypogean
habitats. Amazonian troglomorphic fishes include psammophilous (sand-
dwelling) and fossorial species, such as Pygidianops spp. and Typhlobelus
spp. (Trichomycteridae), and deep/turbid water species, as the siluriforms
Bathycetopsis oliverai (Cetopsidae), Micromyzon akamai (Aspredinidae)
and an undescribed pimelodid, and the gymnotiform Orthosternarchus
tamandua (Trajano 1997a).
I.1.3 Non-troglomorphic, but Found in Subterranean

Habitats (Caves)
Very little attention has been devoted to non-troglomorphic fishes found
in subterranean habitats and few publications (e.g., Poly 2001) mention
their occurrence. Therefore, it is unclear whether the relatively rich non-
troglomorphic ichthyofauna recorded in several Brazilian caves is a
particularity of such karst areas or simply an artifact due to the prevalent
lack of information for other countries.
It may be difficult to distinguish between troglophilic, trogloxenic and
accidental fish in caves, because this requires case-to-case population
studies focusing on the distribution, permanence and reproduction in the
subterranean environment. The long-term permanence of a considerable
number of individuals at relatively large distances from the cave

entrances indicates the presence of self-sustained source populations in

the subterranean biotope, defining the troglophilic status (sensu Holsinger
and Culver 1988). Based on these criteria, Table 2 contains the species
hypothesized as forming troglophilic populations in at least one Brazilian
cave. As expected, since such fish are also subterranean, they mostly belong
to taxa considered as preadapted to the cave life, i.e., the siluriforms and
gymnotiforms. An exception is the Erythrinidae characiforms, sit-and-
wait predators rather frequent in some caves harboring large populations
of potential prey, represented, for instance by the troglobitic R. enfurnada
and A. cryptophthalmus catfishes. No troglomorphic erythrinid has been
Table 2 Updated list of Brazilian troglophilic fishes, with localities (karst area/
county, State) and types of habitat (Trajano 2001). Based on Mattox (2008) and
unpublished data (marked with asterisk, locality including cave).
Taxon Locality Habitat
CHARACIFORMES: ERYTHRINIDAE
1 Erythrinus sp. Altamira-Itaituba, PA Base-level stream
2 Hoplerythrinus unitaeniatus São Domingos, GO Base-level stream
3 Hoplias cf. malabaricus* Serra do Ramalho, BA1 Base-level stream
GYMNOTIFORMES
4 Gymnotus cf. carapo* Cordisburgo, MG 2 Base-level stream
SILURIFORMES:
LORICARIIDAE
5 Isbrueckerichthys alipionis * Alto Ribeira, SP3 Vadose tributary
6 Hypostomus sp.* São Domingos, GO4 Vadose tributary
7 Parotocinclus sp. Presidente Olegário, MG Vadose tributary
TRICHOMYCTERIDAE
8 Trichomycterus sp. A* Mambaí, GO5 Base-level stream
9 Trichomycterus sp. B* Cordisburgo, MG2 Vadose tributary
10 Trichomycterus aff. T. mimonha* Montes Claros, MG6 Base-level stream
HEPTAPTERIDAE
11 Pimelodella transitoria Alto Ribeira, SP Base-level stream
12 Pimelodella sp.* Cordisburgo, MG2
13 Rhamdia sp.* Varzelândia, MG7 Base-level stream
14 Imparfinis hollandi São Domingos, GO Base-level stream
1
Enfurnado Cave; 2Morena Cave; 3Santana Cave; 4São Bernardo Cave; 5Penhasco and Nova Esperança
caves; 6Lapa do Zu cave; 7Zé Avelino Cave.
*
unpubl. data
reported, but a few individuals without eye balls were found in a sandstone
cave in Amazonia (Trajano and Moreira 1991).

Table 2 shows only species presenting good evidence to form troglophilic

populations. Many other species with an unclear status were recorded in
Brazilian caves (see, for instance, Trajano 1991, Bichuette and Trajano 2003).
Once more, most are siluriforms belonging to the same families as the
troglobitic species (Loricariidae, Callichthyidae, Trichomycteridae, and
Heptapteridae), but the diversity of characiforms, especially characids,
is greater. Tetra characins are relatively frequent, but usually as isolated,
emaciated individuals, probably trapped by accident in caves, after floods
for instance. Their frequency in Brazilian caves would just reflect their
abundance in epigean water bodies.
I.2 Habitat Diversity

Troglomorphic fish have been found in practically all kinds of Brazilian
subterranean habitats, from non-karst alluvial sediments to typical karst
habitats, from dozens of meters deep into the water table to shallow
streams. Apparently, type of habitat is not correlated with taxonomy. In
fact, congeneric fish such as Ituglanis and Trichomycterus catfish may occupy
habitats as diverse as streams, phreatic waters and the epikarst (Table 1).
I. epikarsticus is the first fish species reported to live in an epikarst aquifer,
and distribution evidence points to dispersion through the epikarst in I.
bambui and I. ramiroi. For another genus of Trichomycteridae, Glaphyropoma
sp., the type locality is a quartzitic cave. The local terrain is composed of
arenites, siltites, argilites and conglomerates and represents the unique
record of a Brazilian troglobitic fishes in this kind of rock. Caves in this
area are formed by the erosion of soft rock components by rainwater
penetrating through surface cracks and leaving spaces delimited by the
harder components.
Streams are likely the ancestral habitat for the Brazilian troglobitic
siluriforms, except perhaps for the large Rhamdia and Pimelodella species,
which may also be frequent in large rivers (Trajano and Bockmann 1999).
Therefore, the colonization of phreatic spaces is probably an ecological
specialization (autapomorphy) of species such as Rhamdiopsis sp. 1 and
sp. 2, Ituglanis epikarsticus, I. bambui and I. ramiroi. Wide spaces such as
large subterranean lakes and cave pools would represent a secondary
habitat for the above mentioned Rhamdiopsis species.
Another sequence of evolutionary events, without a necessary phreatic
step, seems logical for Ancistrus formoso and Trichomycterus sp. 1 (Table 1),
that live in the lower part of the Serra da Bodoquena karst area and have
bulky bodies, adapted to fast-flowing waters in wide spaces. Probably,
their ancestors lived in streams crossing the respective localities at Bonito
County and progressively adapted to slow-moving water as the caves

were flooded due to the subsidence of the Pantanal region (Bonito Co. is
located at the border of this region).
All known Brazilian troglobitic fishes live in subterranean habitats
accessible through caves, except for S. typhlops and Phreatobius spp.,
collected in artificial wells. For both, the colonization of ground waters
was probably vertical, associated with droughts of epigean water bodies,
streams in a karst area in the case of S. typhlops and submerged litter banks
in the case of Phreatobius species.
I.3 Morphological Diversity

The habitat diversity promotes an enhancement of variation in body
size and shape. Fish adapted to life in narrow subterranean spaces tend
to be smaller and more elongated than their stream-dwelling relatives,
both epigean and hypogean. Rhamdiopsis spinosum 1, Ituglanis epikarsticus
and, to a lesser extent, also I. bambui and I. ramiroi would be instances of
evolutionary reduction in body size allowing for life in phreatic waters,
before (preadaptation) or after (apomorphy) the colonization of this
habitat (Bichuette and Trajano 2004). An opposite tendency was observed
for Trichomycterus sp. 1 found in flooded caves and possessing a bulky
body with a well developed pre-dorsal fold. Life in small spaces is also
probably the cause for loss of the lateral line in the body observed in
Stygichthys typhlops, Phreatobius spp. and Rhamdiopsis sp. 1 (from Chapada
Diamantina).
However, the most striking morphological characteristic of the Brazilian
troglobitic ichthyofauna is the inter- and intrapopulation diversity in
degree of reduction of eyes and melanic pigmentation (Table 1). As in
the case with the habitat diversity, this variation has no clear a taxonomic
correlation, but an apparent geographic correlation has been recognized
(Trajano and Bichuette 2005).
Fish such as Pimelodella spelaea (Fig. 1), Rhamdiopsis sp. 3 (from
Cordisburgo) and Glaphyropoma spinosum (Fig. 1) have slightly reduced
eyes and pigmentation, and only the comparative analysis allowed for
detection of significant differences in relation to epigean congeners. Eyes
and pigmentation are also slightly reduced in Eigenmannia vicentespelaea
(Fig. 1), but a few anophthalmic, very pale individuals have been found
(Bichuette and Trajano 2006). On the other hand, S. typhlops, A. formoso,
Rhamdiopsis sp. 1 (Fig. 1) and Rhamdiopsis sp. 2 (Fig. 1) populations are
homogeneously unpigmented and eyeless (at least in the external view).
All degrees of variation are observed between these extremes, frequently
with a mosaic distribution of character states revealing an independent
differentiation of eyes and pigmentation. In Rhamdia enfurnada, both

Fig. 1 Contd..

Fig. 1 Contd..
Fig. 1 Brazilian troglobitic fishes, showing inter-specific differences in degree

of reduction of eyes and melanic pigmentation: a. Pimelodella spelaea (Photo:
Maria Elina Bichuette); b. Eigenmannia vicentespelaea (Photo: José Sabino); c.
Rhamdiopsis sp. 1 (Photo: Adriano Gambarini); d. Rhamdiopsis sp. 2 (Photo: José
Sabino); e. Ancistrus formoso (Photo: José Sabino); f. Stygichthys typhlops (Photo:
Cristiano Moreira); g. Glaphyropoma spinosum (Photo: Adriano Gambarini).
book.

characters show a normal distribution within the only known population,

but with independent states in each individual: paler specimens may have
relatively large eyes, dark fish may exhibit deeply reduced eyes, and so
on. In Pimelodella kronei, pigmentation has a normal distribution in the
best known population, found in Areias de Cima cave, but 90 percent of
the fish do not have externally visible eyes and the remaining 10 percent
have variable vestigial eyes. In R. enfurnada, Rhamdia sp. and P. kronei,
vestigial eyes are frequently assymetrical in size, which is in accordance
with neutralistic theories of regressive evolution of eyes in cave fish.
All collected specimens of P. cisternarum have tiny vestigial eyes and
traces of pigmentation, whereas P. dracunculus catfish are more advanced
in degree of troglomorphism, having no pigment cells or eyes. Eyes are
also absent in P. sanguijuela.
The Ituganis catfish species, represented by five formal species
(I. passensis I. bambui, I. epikarsticus, I. ramiroi and I. mambai) (Fernandez
and Bichuette 2002, Bichuette and Trajano 2004, 2008), present a mosaic
of characters considering the eyes and body pigmentation. I. bambui and
I. mambai, which present a lower degree of regression of body pigmentation
and eyes, possibly have been isolated in the subterranean environment
for a shorter time than I. passensis, I. ramiroi and I. epikarsticus. I. bambui
and I. mambai have similar body pigmentation patterns and eye sizes with
irregular light brown spots along the dorsum and flanks and the head
with a dorsal portion darker than the rest of the body; the eyes are visible
externally as round black spots, which are covered by skin and reduced in
relation to epigean Ituglanis species. The more troglomorphic species also
present a mosaic, but, apparently the variability is lower. I. ramiroi species
has a pale yellowish color pattern with scattered chromatophores spots
along the body; the eyes appear as very small round black spots covered
by thin skin, in some cases no eyes are externally visible. I. passensis can
be considered an intermediary case, showing coloration from yellowish
to light gray, with scattered chromatophores. The eyes are reduced and,
sometimes, not visible externally. I. epikarsticus is the most troglomorphic
species of Ituglanis, with individuals totally depigmented, showing a
translucid aspect. The eyes are small round black spots covered by thin
skin, with a lower diameter compared to the other Ituglanis cave species
(0.1 mm).
A most remarkable instance of mosaic distribution of characters states,
including eyes, pigmentation and body shape, was reported for four
populations of the armored cave catfish, Ancistrus cryptophthalmus, from
São Domingos karst area, Central Brazil (Reis et al. 2006). The populations
found in the caves Angélica, Bezerra, São Vicente I and Passa Três (part of
São Vicente cave system) were compared to Ancistrus catfish found in the
epigean stream reaches and ranked according to their degree of reduction

of eyes and pigmentation. Different sequences were observed for eyes

(from less to more reduced): epigean (eyes as in epigean congeners)
Bezerra Angélica São Vicente I Passa Três (eyes not externally
visible in adults), and for pigmentation: epigean (light black to brownish)
Bezerra (brownish to light brownish) Passa Três (light brownish
to dark yellow-greenish) São Vicente I and Angélica (light brownish
to light yellow). General body shape was compared through geometric
analysis, and a third sequence was found when associating head shape
and position of eyes and nostrils, and length of caudal region (Reis et al.
2006: Fig. 5): epigean + Angélica + Bezerra + Passa Três São Vicente I.
This mosaic distribution of character states among the four studied cave
populations indicates a certain degree of isolation, with some independent
differentiation in these populations.
An interesting case of polymorphism is represented by the only known
population of Trichomycterus itacarambiensis (Fig. 2). Two distinct color
morphotypes are present, one represented by completely depigmented
individuals (one third of the population as verified in 1994) and another
by variably pale, but clearly pigmented catfish (the remaining two thirds);
15 years later, this proportion was half to half, due to a population decrease
affecting mostly the pigmented fish (Trajano et al. 2009). Variably reduced
eyes are present in both phenotypes, showing a normal distribution in
relation to eye size; externally visible eyes in depigmented fish are red
colored (Trajano and Pinna 1996).
Such diversity in troglomorphic degree among related taxa provides
excellent opportunities for the investigation of mechanisms of character
regression. In fact, morphological and physiological studies indicate that
the diversity in modes of pigmentation variation (continuous versus
discontinuous, polymorphism etc.) results from a combination of different
mechanisms of pigmentation reduction. The continuous variation in color,
from darker to paler individuals, is associated with a decrease in density
(and, in some cases, also size) of melanin cells in the skin. In populations
with a discontinuous variation, as T. itacarambiensis, the depigmented
phenotypes correspond to individuals that lose the ability to synthesize
melanin. Physiological studies showed that highly specialized troglobitic
species, with homogeneously depigmented populations, respond
differently to the administration of L-DOPA in vitro: in most, classified as
DOPA(+), melanin is then synthesized and melanin cells become visible
(e.g., Stygichthys typhlops, Ancistrus formoso, Rhamdiopsis sp. 1), but in some,
as Rhamdiopsis sp. 2 and the depigmented individuals of T. itacarambiensis,
there is no response, and they are classified as DOPA (–). Apparently, a
polygenic system is involved in the differentiation of melanophores,
determining their density in the skin. On the other hand, monogenic
systems would be involved in the ability to synthesize melanin, which

Fig. 2 Contd..

Fig. 2 Contd..
Fig. 2 Intra-specific variability in the degree of reduction of eyes and pigmentation

in Trichomycterus itacarambiensis; a-b: pigmented fish with variably sized eyes;
c-d: light pigmented versus albino specimens. (Photos: José Sabino).
book.

could be lost due to the blocking of different steps in this synthesis (Felice
et al. 2008).
I.4 Regional Diversity and Speciation in Brazilian

Troglobitic Fishes
Neither the troglobitic nor the troglophilic subterranean ichthyofaunas
are evenly distributed throughout the Brazilian karst areas, with some
regions conspicuously richer than others. So far, São Domingos karst
area, in the Rio Tocantins drainage, Amazonas basin, is the richest in both
troglomorphic (seven species) and non-troglomorphic species (Bichuette
and Trajano 2003). However, the most specialized troglobites occur in
semiarid karst areas of central and northern Bahia State (northeastern
Brazil), where epigean drainages are mainly temporary and the blind fish
are generally confined to phreatic water bodies.
The lack of a strict geographic correlation between total richness
(troglobites + non-troglobites) and richness of troglobites alone indicate
the existence of different factors: whereas general subterranean diversity
seems to be mainly related to ecological factors, such as habitat diversity
and availability of food resources, in addition to richness of potential
colonizers (all very high in São Domingos caves), richness of troglobitic
species and their degree of differentiation is better explained by historical
factors, mainly vicariance leading to speciation (Trajano 2001b).
Ideally, different kinds of evidence, distributional, geological,
geomorphological, paleoclimatic, morphological etc., should be put
together in order to uncover the most likely processes leading to the
diversification of the subterranean fauna and the origin of troglobites. It
is generally accepted that allopatric speciation is the main cause of animal
diversity, and there is no reason to suppose that subterranean organisms
are an exception.
Disruption of connection between epigean (source of colonizers) and
hypogean drainages is the most obvious, and probably the commonest
cause of genetic isolation of aquatic organisms in subterranean habitats. In
the case of cyclical events, such as the Quaternary glacial cycles, geographic
isolation in dry phases would intercalate with the re-establishment of
drainage continuity in wet phases, when surviving epigean populations
may come back in contact with the subterranean ones. In this case,
either hybridizing between these populations or secondary contact with
maintenance of their taxonomic identities (if reproductive incompatibility
was achieved by differentiation) may occur.

Isolation during past drier climates seems to be the case of the blind
catfish, Pimelodella kronei, currently in secondary contact with its epigean
relative, P. transitoria. The Alto Ribeira karst area is presently situated in a
wet zone, but there is good evidence of dry, isolation phases in subtropical
Brazil during the last 116,200 years (Cruz-Jr. et al. 2005). On the other hand,
karst areas in Bahia State where troglobitic fish are found (Table 1) are located
in currently semiarid regions, thus in a phase favoring differentiation.
Among these species, Rhamdiopsis sp. 1 (Chapada Diamantina) and sp.
2 (Campo Formoso), R. enfurnada and Trichomycterus sp. 2 live in water
bodies without direct epigean connections. For Campo Formoso, there
is also evidence pointing to accentuated paleoclimatic fluctuations in the
past 210,000 years (Wang et al. 2004), but dry phases were longer than in
the Alto Ribeira. Longer isolation periods may account for the high degree
of troglomorphism observed in Rhamdiopsis spp. from Bahia.
The localities with Stygichthys typhlops as well have no permanent,
direct connections with epigean drainages. In this case, lowering of the
water table due to extensive water pumping for irrigation projects in the
last five decades is speeding up the disconnection process and threatening
the species by habitat loss (Moreira et al., in press).
Since no epigean congener was found in São Domingos karst area, in
spite of intensive collecting efforts (Bichuette and Trajano 2003), extinction
of epigean relatives may explain the origin of Pimelodella spelaea and the
four troglobitic Ituglanis species found in the area. Topographical isolation
may contribute for differentiation in these species, since they live basically
in the upper karst level (vadose tributaries and epikearst – Bichuette and
Trajano 2004). Topographic isolation may also be a cause for differentiation
in Rhamdiopsis sp. 3 (from Cordisburgo), which live in a stream several
meters above the base level in the area (unpubl. data).
More complicated is the case with the armored catfish, Ancistrus
cryptophthalmus, from São Domingos. Topographic isolation may explain
the morphological differentiation between the four cave populations
(see above), because they inhabit stream reaches separated by waterfalls
(Reis et al. 2006). However, there is a parapatric contact between epigean
Ancistrus and A. cryptophthalmus at the large cave entrances of Angélica
and Bezerra caves (sinkholes and resurgences). In the absence of strong
evidence for extinction of epigean populations sometime in the past, with
a secondary contact nowadays, parapatric speciation cannot be ruled out.
II. NATURAL HISTORY AND BIOLOGY

From about 1,700 references dealing strictly with natural history and
ecology of subterranean fishes (G. Proudlove, pers. comm.), ca 35 are related

with Brazilian cavefish, including habitat descriptions and comments on

habitat use, population studies, reproduction and behavior. The first work,
more extensive, was published by Pavan (1946) focusing the heptapterid
catfish Pimelodella kronei. From 1990’s on, a series of ecological studies were
published focusing the Brazilian cavefishes (e.g., Trajano 1991, 1997a, b,
Trajano and Bockmann 1999, Trajano 2001a,b, Trajano and Bichuette 2007).
These works include information on population sizes and densities, fish
movements and life stories aspects, such as feeding, reproduction and
growth rates. Some of these aspects will be considered herein.
II.1 Diet and Feeding Behavior

Brazilian subterranean fishes are frequently generalist carnivores feeding
on an opportunistic basis (trichomycterids, heptapterids and the electric
fish species). Potentially these fishes can explore many kinds of resources,
like crustaceans (amphipods and isopods), aquatics insects (juveniles to
adults), mollusks, worms (annelids), besides several allochthonous items
(terrestrial invertebrates); in the few cases where direct comparisons were
made, it was shown that subterranean fishes belong to similar trophic
guilds of their epigean relatives (Trajano 1989, Bichuette, unpubl. data).
This suggests that epigean ancestors were already preadapted to feed in
the hypogean environment.
Special cases must be considered herein, for example, cannibalism
was recorded for Stygichthys typhlops (Moreira et al., in press). There are
records of cave catfish foraging on submerged bat guano (Rhamdiopsis sp.
1 – Mendes 1995, M.E. Bichuette, pers. obs.), but is not clear whether these
fishes are feeding directly on the guano or on the invertebrates living on it,
since most of the stomach items are invertebrates. The armoured catfishes
Ancistrus cryptophthalmus are typically detritivorous (Trajano 2001a),
probably feeding on the film of detritus covering rocks.
The high proportion of specimens captured with empty stomachs or
few items inside is directly related with the effect of food scarcity of caves.
For instance, Trichomycterus itacarambiensis and Ituglanis populations
showed this pattern during the dry season (Trajano 1997b, Bichuette,
unpubl. data). Furthermore, it was observed a decrease in the frequency of
insects and an increase of oligochaete items throughout the dry period for
T. itacarambiensis, probably related with the decrease in insect populations
along the dry period.
In relation to feeding behavior, different foraging styles were observed for
Brazilian cavefishes, related to the peculiar conditions of the subterranean
realm. The heptapterids also feed at midwater and surface (Trajano 1989,

Fig. 3 Contd..

Fig. 3 Contd..
Fig. 3 Habitat and population study methods for Brazilian cavefishes (Photos:
Adriano Gambarini): a-b. habitat of Glaphyropoma spinosum from Chapada
Diamantina, northeastern Brazil; c. collecting procedure for mark-recapture studies
and; d. measuring standard length for mark-recapture studies.
book.

M.E. Bichuette, pers. obs.), contrasting to the epigean relatives which are
typical bottom-dwellers.
II.2 Life History

Troglobitic organisms usually show a precocial lifestyle, including life
history traits such as large and yolky eggs, low fecundity, delayed and
infrequent reproduction, slow growth rates and increase longevity
(Culver 1982). For Brazilian cavefishes, some of these traits were verified,
suggesting a precocial lifestyle: T. itacarambiensis and A. cryptophthalmus
showed low proportions of reproductive females, producing relatively
large eggs, slow individual growth rates and high longevity (Trajano
1997b, Trajano and Bichuette 2007). However, a precocial lifestyle may
be a characteristic already present in epigean ancestors, favoring the
colonization of the subterranean environment, as observed for Pimelodella
kronei and P. transitoria (Trajano 1989).
In general, reproductive data for cavefishes are limited since several
subterranean habitats are not reachable during the rainy periods and
the population densities are frequently low. Seasonal reproduction was
observed for some Brazilian cave catfishes: Rhamdiopsis sp. 1 and T.
itacarambiensis showed females with developed gonads in the end of the
rainy season, suggesting a reproductive peak in this period (Mendes 1995,
Trajano 1997b). Trajano (1991) verified a very low number of Pimelodella
kronei catfish showing mature gonads and no reproductive patterns were
detected.
In relation to individual growth, very low rates (0.06 mm × month–1
in average) or negative growth, were recorded for Trichomycterus
itacarambiensis (Trajano 1997b). This study was carried out along the
dry season when the condition factor decreased, demonstrating the
influence of food limitation in the life cycle of this species. Low rates of
individual growth and negative growth was also observed for Ancistrus
cryptophthalmus (Trajano and Bichuette 2007). For Pimelodella kronei,
average growth rates were 1.0 mm × month–1, and longevity around 10-
15 years (Trajano 1991). These data are similar with that recorded for the
putative sister-species P. transitoria (Gerhard 1999), showing a precocial
lifestyle and possibly configurating a preadaptation to the cave life.
III. POPULATION ECOLOGY

The population ecology of several Brazilian cavefishes was studied using
either mark-recepture methods (MR) or visual census (VC). MR produces

more reliable data on population sizes, and population densities may be

calculated based on the one occupied by these populations. In the case of
VC, densities are estimated directly from counts.
Trajano (2001a) proposed a classification based on mean population
densities of cavefishes: (1) species with low population densities (< 0.1
ind × m–2), for example, Rhamdiopsis sp. 1 from Chapada Diamantina, (six
populations from 10 studied) Rhamdiopsis sp. 2 from Campo Formoso,
and Pimelodella kronei; (2) species with median population densities
(0.1-1.0 ind × m–2), for example Trichomycterus itacarambiensis, Ancistrus
cryptophthalmus, Ituglanis bambui, I. passensis, I. ramiroi and Rhamdiopsis
sp. 1; there are no records for the third category proposed by Trajano
(op. cit.), i.e., high population densities (> 1.0 ind × m–2, on average).
Population sizes, based on MR studies, also varied from very small
to comparable to relatively small populations of epigean species. The
population size of the stream-dweller Trichomycterus itacarambiensis from
Olhos D’água cave (5,000 m of habitat extension) was estimated to be
1,500 to 2,000 adults; Pimelodella kronei from Areias cave (5,000 m of habitat
extension), 900-1,200 adults; Ancistrus cryptophthalmus from Angélica
cave (8,000 m of habitat extension) reached 20,000 individuals but the
population from Passa Três cave (2,000 m of habitat extension), had only
1,000 individuals; the trichomycterids from Goiás, Ituglanis spp., showed
relatively small populations, with hundreds of individuals.
References
Bichuette, M.E. and E. Trajano. 2003. Epigean and subterranean ichthyofauna from
São Domingos karst area, Upper Tocantins river basin, Central Brazil. Journal of
Fish Biology 63: 1100-1121.
Bichuette, M.E. and E. Trajano. 2004. Three new subterranean species of Ituglanis
from Central Brazil (Siluriformes: Trichomycteridae). Ichthyological Explorations
of Freshwaters 15(3): 243-256.
Bichuette, M.E. and E. Trajano. 2006. Morphology and distribution of the
cave knifefish Eigenmannia vicentespelaea Triques, 1996 (Gymnotiformes:
Sternopygidae) from Central Brazil, with an expanded diagnosis and comments
on subterranean evolution. Neotropical Ichthyology 4(1): 99-105.
Brittan, M.R. and J.E. Böhlke. 1965. A new blind characid fish from southeastern
Brazil. Notulae Naturae 380: 1-4.
Cruz Jr., F.W., S.J. Burns, I. Karmann, W.D. Sharp, M. Vuille, A.O. Cardoso,
J.A. Ferrari, P.L.S. Dias and O. Viana Jr. 2005: Insolation-driven changes in
atmospheric circulation over the past 116,000 years in subtropical Brazil. Nature
434: 63-66
Culver, D.C. 1982. Cave Life. Evolution and Ecology. Harvard Press, Cambridge.

Felice, V., M.A. Visconti and E. Trajano. 2008. Mechanisms of pigmentation loss in
subterranean fishes. Neotropical Ichthyology 6(4): 657-662.
Gerhard, P. 1999. Ecologia de populações e comportamento de quatro espécies
de bagres Heptapterinae (Teleostei: Siluriformes) em riachos do Alto vale
do rio Ribeira (Iporanga, São Paulo). MSc. Thesis – Instituto de Biociências,
Universidade de São Paulo, São Paulo. 129 pp.
Holsinger, J.R. and D.C. Culver. 1988. The invertebrate cave fauna of Virginia and a
part of Eastern Tennessee: Zoogeography and ecology. Brimleyana 14: 1-162.
Lèvêque, C., T. Oberdorff, D. Paugy, M.L.J. Stiassny and P.A. Tedesco. 2008. Global
diversity of fish (Pisces) in freshwater. Hydrobiologia 595: 545-567.
Mattox, G.M.T., M.E. Bichuette, S. Secutti and E. Trajano. 2008. Surface and
subterranean ichthyofauna in the Serra do Ramalho karst area, northeastern
Brazil, with updated lists of Brazilian troglobitic and troglophilic fishes. Biota
Neotropica 8(4): 145-152.
Mendes, L.F. 1995. Ecologia populacional e comportamento de uma nova espécie de
bagres cavernícolas da Chapada Diamantina, BA (Siluriformes, Pimelodidae).
MSc. Thesis-Instituto de Biociëncias, Universidade de São Paulo, São Paulo. 86
pp.
Moreira, C.R. M.E. Bichuette, O. Oyakawa, M.C.C. Pinna and E. Trajano.
Rediscovery of Stygichthys typhlops Britton and Bohlke, 1965, an enigmatic
subterranean characiform fish from Jaiba karst area, eastern Brazil. Journal of
Fish Biology, in press.
Muriel-Cunha, J. 2008. Biodiversidade e sistemática molecular de Phreatobiidae
(Ostariophysi, Siluriformes) - com uma proposta sobre sua posição filogenética
em Siluriformes e uma discussão sobre a evolução do hábito subterrâneo,
Unpubl. Ph.D. thesis, Universidade de São Paulo, Sao Paulo, 144 p.
Pavan, C. 1945. Os peixes cegos das cavernas de Iporanga e a evolução. Boletim da
Faculdade de Filosofia Ciências e Letras, Biologia Geral 6: 79: 1-104.
Poly, W.J. 2001. Nontroglobitic fishes in Bruffey-Hills Creek Cave, West Virginia,
and other caves worldwide. Environmental Biology of Fishes 62: 73-83.
Reichel, M. 1927. Étude anatomique du Phreatobius cisternarum Goeldi, silure
aveugle du Brésil. Revue Suisse de Zoologie 34(16): 285-403 + 6 pl.
Biology 68: 414-429.
Thinès, G. and G. Proudlove. 1986. Pisces. In: Stygofauna Mundi, L. Botosaneanu
(ed.). E.J. Brill, Leiden, pp. 709-733.
Trajano, E. 1987. Biologia do bagre cavernícola, Pimelodella kronei, e de seu
provável ancestral, Pimelodella transitoria (Siluriformes, Pimelodidae). Ph.D.
thesis, Instituto de Biociências, Universidade de São Paulo, São Paulo, 211 pp.
Trajano, E. 1989. Estudo do comportamento espontâneo e alimentar e da dieta do
bagre cavernícola, Pimelodella kronei, e seu ancestral epigeo, Pimelodella transitoria
(Siluriformes, Pimelodidae). Revista Brasileira de Biologia 49(3): 757-769.

Trajano, E. 1991. Populational ecology of Pimelodella kronei, troglobitic catfish from

southeastern Brazil (Siluriformes, Pimelodidae). Environmental Biology of Fishes
30: 407-421.
Trajano, E. 1997a. Synopsis of Brazilian troglomorphic fishes. Mémoires de
Biospéologie 24: 119-126.
Trajano, E. 1997b. Population ecology of Trichomycterus itacarambiensis, a cave
catfish from Eastern Brazil (Siluriformes, Trichomycteridae). Environmental
Trajano, E. 2001a. Ecology of subterranean fishes: An overview. Environmental
Biology of Fishes 62(1-3): 133-160.
Trajano, E. 2001b. Mapping subterranean biodiversity in Brazilian karst areas.
Karst Waters Institute Special Publication 6: 67-70 (Proceedings of the Workshop
“Mapping Subterranean Biodiversity”, Moulis).
Trajano, E. 2005. Evolution of lineages. In: The Encyclopedia of Caves, D.C. Culver
and W.B. White (eds.). Academic Press, San Diego, pp. 230-234.
Trajano, E. and J.R.A. Moreira. 1991. Estudo da fauna de cavernas da Província
Espeleológica Arenítica Altamira-Itaituba, PA. Revista Brasileira de Biologia 51(1):
13-29.
Trajano, E. and F.A. Bockmann. 1999. Evolution of ecology and behaviour in
Brazilian cave Heptapterinae catfishes, based on cladistic analysis (Teleostei:
Siluriformes). Mémoires de Biospéologie 26: 123-129.
Trajano, E. and M.E. Bichuette. 2005. Diversity of subterranean fishes in Brazil. In:
World Subterranean Biodiversity, J. Gibert (ed.). Proceedings of an International
Symposium held on 8-10 December 2004 in Villeurbaine, pp. 161-163.
Trajano, E. and M.E. Bichuette. 2007. Population ecology of cave armoured
catfish, Ancistrus cryptophthalmus Reis, 1987, from central Brazil (Siluriformes:
Loricariidae). Ecology of Freshwater Fish 16: 105-115.
Trajano, E., S. Secutti and M.E. Bichuette. 2009. Population decline in a Brazilian cave
catfish. Trichomycterus itacarambiensis Trajano and Pinna, 1996 (Siluriformes):
reduced flashflood as a probable cause. Speleobiology Notes 1: 24-27.
Wang, X., A.S. Auler, R.L. Edwards, H. Cheng, P.S. Cristalli, P.L. Smart, D.A.
Richards and C.-C. Shen. 2004. Wet periods in northeastern Brazil over the past
210 kyr linked to distant climate anomalies. Nature 432: 740-743.

CHAPTER 10
Subterranean Fishes of Africa
Roberto Berti1 and Giuseppe Messana2
1
Department of Evolutionary Biology,
University of Florence, Italy
2
Institute for the study of Ecosystems, National
Research Council (CNR), Florence, Italy
INTRODUCTION
The extension of Africa comprising temperate to tropical latitudes, with a
great quantity of different ecosystems, makes it an extremely variegated
continent with high levels of biodiversity. A few zones have been explored
from a biospeleological point of view, nonetheless the results obtained so far
show a great subterranean diversity. Most of the biospeleological research
begun in the first decades of the twentieth century and was achieved on
Eastern (Ethiopia, Kenya and Somalia) Northern (especially Morocco)
Central (Democratic Republic of the Congo) and Southern (Namibia and
S. Africa) Africa and on Madagascar. Several researchers explored these
territories and collected a large amount of zoological material.
It is in these years that the geographic and naturalistic exploration led
to the discovery of the majority of the species of African stygobitic fishes,
distributed only in the sub-Saharan hydrographical systems. The first cave
dwelling fish was a little cyprinid discovered in the Democratic Republic
of the Congo (Belgian Congo, at that time) and it is the first subterranean
fish in the world that was officially protected.
Among the eight valid species of fishes known for the Afrotropical
Region, six were described between 1921 and 1936. The beginning of the
World War II, followed by the instability due to decolonization, was the
cause of a long interruption of naturalistic research in Africa. This is the

reason why for the description of the two new African subterranean fishes,
we have to wait for the second half of the century.
It is also in this period that an extensive biological, physiological and
behavioural research was conducted by Belgian and Italian researcher on
the stygobitic fishes respectively from Central and Eastern Africa.
The present chapter enlightens how much is known about the
distribution, biology and eco-physiology of each of the fish species
inhabiting the subterranean waters of the African continent.
Fig. 1 Distribution of African subterranean fishes. The type localities of the

eight species are indicated. 1: Caecobarbus geertsii; 2: Phreatichthys andruzzii;
3: Barbopsis devecchii; 4: Uegitglanis zammaranoi; 5: Clarias cavernicola;
6: Typhleotris madagascariensis; 7: Typhleotris pauliani; 8: Glossogobius
ankaranensis.
book.

Roberto Berti and Giuseppe Messana 359
FAMILY CYPRINIDAE
Caecobarbus geertsii Boulenger, 1921
Fig. 2 Caecobarbus geertsii (after M. Poll; from Leleup 1956a, modified).
Historical Notes
The fish was discovered by some amateur cave explorers in 1917 in a
cave near the town of Thysville, now Mbanga-Ngungu. One of them was
Mr G. Geerts, at that time Managing Director of the Railway Company
that exploited the railway from Matadi to Leopoldville via Thysville.
Three years later Dr Henri Schouteden, a Belgian zoologist responsible
for the zoological collections of the Congo Museum in Tervuren, went to
Congo for a collecting trip. He met Mr Geerts, who told him about a blind
depigmented fish living in some caves near Thysville and arranged for
him the collection of three specimens (Thys van den Audenaerde 1999).
Shipped to Belgium, the fish was described by Georges Albert Boulenger,
the world’s expert on African fishes, who established the new genus and
species.
Caecobarbus is the first subterranean fish in the world to be officially
protected, since the Belgian colonial authorities, in 1937, placed it on the list
of fully protected animals, in the same category as gorilla and chimpanzee
(Thys van den Audenaerde 1999). Moreover, this is the only hypogean fish
whose trade is regulated by CITES, having been listed in CITES Appendix
II in 1981 (Proudlove and Romero 2001, Proudlove 2006).

Derivatio Nominis
Caecobarbus: from caecus (blind, in Latin) + barbus (because of its close
relationships with the genus Barbus);
geertsii: as a tribute to Mr G. Geerts.
Common Name
Congo blind barb (English), barbu aveugle (French).
Size
Up to 110 mm (Thinès 1955a, 1969)
Distribution Area and Ecological Notes

The species inhabits (Heuts 1952, Heuts and Leleup 1954, Leleup 1956a,
1956b) seven between caves and cave complexes (Table 1) located
near and south of the town of Mbanga-Ngungu (formerly Thysville;
05o15’08”S, 14o 52’09”E), Western Democratic Republic of the Congo.
Table 1 Caves inhabited by Caecobarbus geertsii (from Heuts and Leleup 1954,
modified).
Cave Water Water pH Altitude Tributary
temperature hardness (m a.s.l.) to
(°C)
B 5 (Grotte des Gaz) 23 11.5 7.4 595 Fuma
B 6c (Grotte des Cascades) - - - 528 Fuma
B 7 (Grande Grotte de 22.5-23.4 °C 8-13.5 7.4-7.6 682.5 Fuma
Thysville)
B 11 (Grotte de Nenga) 22 15.5 7.3 540 Fuma
B 15a 22 10.75-12.75 7.5-7.6 515 Fuma
B 15b 21-22 13-13.75 - - Fuma
B 15e - - - - Fuma
B 15f - - - 500 Fuma
B 16a 21.5 15.5 7.8 640 Kokosi
B 16b - - - - Kokosi
B 20 21.5 12 - - Kokosi
The caves colonized by Caecobarbus belong hydrographically to two

streams (Fuma and Kokosi, respectively), both tributary to the Kwilu
River (Lower Congo basin). The distribution area is quite small, since all
collection sites are placed within a radius of 7/8 km at the most. The area
is included in limestone formations, laid down during the Pre-Cambrian

or at least during the Cambrian; the genesis of the caves is thought to take
place between late Tertiary and mid-Holocene (Heuts 1952).
The Caecobarbus habitats show pronounced seasonal cycles of activity.
During the rainy season they receive heavy inflows of water carrying
nutrients from the surface and eventually leading to complete submergence;
during the dry season the subterranean water flow either is not perceivable
or it is weak and of endogenous origin. Therefore, periodic changes occur
in food resources availability, fish density and water chemical features,
the latter ones showing cyclic variations (1 to 3 times wide) in Ca(HCO3)2
concentration and in free CO2 contents. No subterranean continuity is
thought to exist between cave complexes, thus each complex seems to be
an isolated subterranean unit; this is confirmed by both chemical features
and population analyses (Heuts 1952, Heuts and Leleup 1954).
The aquatic associated fauna is quite rich, and almost exclusively
composed of invertebrates. These include both potential preys as culicids
larvae, planarians, oligochaetes, crustaceans (ostracods, copepods,
isopods), the dryopid beetle Troglelmis leleupi Jeannel, and potential
predators as the crab Potamonautes biballensis or dytiscids larvae (Leleup
1956a, 1957). The only available data about vertebrates concern (Heuts
and Leleup 1954, Leleup 1956a) the presence of few epigean fishes, as
some siluriforms – most likely Clarias sp. (B15 and B6 cave complexes)
and Labeo sp. specimens (B6 cave complex).
Systematics and Phylogenetic Relationships

The species is generally thought to be phylogenetically related to the
genus Barbus Cuvier (Boulenger 1921, Pellegrin 1928, 1929a, 1929b,
Gianferrari 1932, 1934, Petit 1938a, Bruun and Kaiser 1943, Heuts 1953,
Matthes 1963, Vandel 1965, Thinès 1969). In Caecobarbus the pharyngeal
teeth are arranged over two instead of three rows as it occurs in Barbus; on
the ground of this feature some authors (Pellegrin 1932, Gianferrari 1937a)
suggested its close relationships with a primitive Barbus type.
More recently Howes (1991) proposed the species to belong to a barbin
lineage of the subfamily Cyprininae.
Morpho-physiological Notes
When adult, Caecobarbus is a blind, apparently anophthalmic, cavefish.
In small individuals (under 10 mm in length) the eye is still visible
by transparency through the skin of the orbital region as a dark spot
(Heuts 1952); almost all histological components of the eye of epigean
fish larvae are recognizable, moreover both optic nerve and chiasma are
still present (Quaghebeur 1955). Then a degenerative process occurs, and

the adult eye is vestigial, deeply receded inside the skull and consisting
only of histological remnants, disorderly arranged and covered by
connective tissue and skin (Gérard 1936, Quaghebeur 1955). Moreover,
eye degeneration is accompanied by the reduction of optic lobes
(Quaghebeur 1957) as well as by a rearrangement of the circumorbital
dermal bones (Petit 1941).
The metabolic rate, as measured by oxygen consumption per unit
weight, is three times lower than in the epigean Barbus conchonius, and
the existence of a low growth rate was established by scale readings
(Heuts 1952, 1953).
The scales are still present, but very thin and soft (Boulenger 1921,
Petit 1938a). The skin colour is reddish, due to the loss of all external
melanine pigmentation, while black pigment cells are present on the
peritoneum. A spot of guanine pigment on the operculum is shown by
the populations of two cave systems (B16 and B20, respectively); it is
absent in the remnant ones (Heuts 1952). Other differences exist between
different populations; this is the case of a) the highest ray of the dorsal fin,
which is not serrated in six out of the seven studied populations, and it is
in the seventh (from cave B5) population, b) the age distribution, deduced
from scale readings, which appears to be typical for each population, and
c) the growth rate curve (Heuts 1952). The variability existing between
populations for these morphological and physiological features strongly
supports the hypothesis that the species could be structured in distinct
local populations (Heuts 1952, Heuts et Leleup 1954).
Histological surveys showed that a) the thyroid is little developed, with
a lowered functional activity (Olivereau and Francotte-Henry 1955), b) the
hypophysis lacks thyrotropic cells, and this could explain the reduction
of the thyroid gland (Olivereau and Herlant 1954). The aforementioned
findings appear to be in complete accordance with both physiological
features and morphological traits, as lowered growth rate, reduced
metabolic rate and loss of external melanine pigmentation (Olivereau and
Francotte-Henry 1955).
Besides these degenerative features, one compensative trait is shown.
Thus, whereas the trunk lateral line system is normally developed
(Boulenger 1921), an increased amount of sensory buds – neuromasts or
taste buds – is present on the head (Heuts 1952, 1953).
Ethological Notes
The first behavioural observations on Caecobarbus were few remarks
on its spontaneous behaviour in captivity describing the swimming
characteristics and the feeding behaviour (Petit and Besnard 1937).

The first quantitative studies are those by Thinès (1953, 1955b) on the
light sensitivity: tested in an alternative chamber apparatus Caecobarbus
shows an evident photophobic behaviour, which doesn’t vary in intensity
as a function of the light wavelength. In the field, fishes observed in a light
gradient at the cave entrance exhibited different behaviours as a function
of their age, the young fishes being apparently indifferent to light and
the old ones becoming photophobic for light intensities higher than 8 lux
(Thinès 1958a).
Feeding behaviour studies highlighted how Caecobarbus react to food
administration with an immediate polarization of its swimming activity
towards the bottom where the fish starts an active food search (Thinès
and Wissoq 1972). The stereotype of polarisation on the substrate has
an important role in the behavioural repertoire of Caecobarbus, since it
appears both after mechanical stimulation (Thinès and Wissoq 1972) and
as response to alarm substance, which is thought to elicit a real feeding
behaviour (Thinès and Legrain 1973).
Chemical stimulations not only trigger feeding behaviour of Caecobarbus
(Thinès and Wissoq 1972), they are also effective in influencing its
topographic orientation. In fact, when placed in a choice-apparatus, adult
fishes show a significant tendency to orient towards the section in which
water from a tank occupied by familiar conspecifics is introduced (Berti
and Thinès 1980). In despite of this apparent tendency to aggregation,
neither schooling nor shoaling behaviour is shown, as evidenced by
Jankowska and Thinès (1982).
FAMILY CYPRINIDAE
Phreatichthys andruzzii Vinciguerra, 1924
Fig. 3 Phreatichthys andruzzii (photo by R. Innocenti).

book.
Historical Notes
The first six specimens were collected at an unspecified date by lieutenant
Giovanni Zaccarini, a geographer who played an active role in the
ichthyologic exploration of the inland waters of Italian Somaliland. Dr
Alcibiade Andruzzi, Navy medical major and at that time director of
the Colonial Health Service of Somalia, took care of some naturalistic
investigations on behalf of the Museo Civico di Storia Naturale di Genova.
Thus, when he received the material collected by Zaccarini, shipped the
fishes to Prof. Decio Vinciguerra, ichthyologist of that museum, who
examined and described them as belonging to a new genus and species.
Derivatio Nominis
Phreatichthys: from jre’ ar phrear (well, in Greek) + cq V ichthys (fish, in
Greek);
andruzzii: in homage to A. Andruzzi.
Common Name
None
Size
Up to 106 mm in total length and to 87 mm in standard length (personal
observations).

The species inhabits the subterranean waters of the Central Somalia. From
a geological point of view, the hydrographic system where Phreatichthys
occurs is included in evaporative formations of the middle-lower Eocene
Age, connected in the West with biogenic limestone massive formations
of the lower Eocene.
The distribution area of Phreatichthys is rather small, the distance
between the two farthest collection sites being of about 30 km. The habitat
is probably a continuous hydrographic system, but still the species seems
to be structured in distinct local populations, each probably occupying
its own area (Berti and Zorn 2001), and among which gene flow seems
to be either restricted or absent (Cobolli Sbordoni et al. 1996). The water
table is very superficial, lying at a depth varying from some meters to few
tens of centimeters under the ground surface. Besides the type locality, six
collection sites are known (Table 2).

Table 2 Known collection sites of Phreatichthys andruzzii; type locality: Bud Bud
(from Ercolini et al. 1982, modified).
Locality Latitude Longitude Altitude (m) T(°C)
Bud Bud (3 sites) 04°11’19”N 46°28’27”E 137 29
Gal Ef (2 sites) 04°11’53”N 46°28’16”E 139 -
El Dirri 04°20’12”N 46°36’24”E 136 29
Diptias 04°07’02”N 46°27’16”E 137 31
Damairor 04°08’15”N 46°27’52”E 137 30
Gheriale 04°08’22”N 46°29’05”E 133 29
El Dauao (2 sites) 04°10’58”N 46°29’18”E 136 31
The collection stations are either hand dug wells (Bud Bud 1, Gal Ef
2, Damairor) or springs (El Dirri, Diptias, El Dauao 1) or sites where the
water table become superficial after the caliche had collapsed (Gal Ef 1,
Gheriale, El Dauao 2, Bud Bud 2 and 3).
In the wells of Bud Bud, together with Phreatichthys, were collected the
isopod Acanthastenasellus forficuloides new species and genus (Chelazzi and
Messana 1985) and a new amphipod of the genus Afridiella, A. pectinicauda
(Ruffo 1982).

The species was suggested to be phylogenetically related to the genus
Barbus on the ground of a purely morphomeristic examination (Vinciguerra
1924). The descriptor’s statement was in principle shared by other
authors (Pellegrin 1932, Gianferrari 1932, 1934, 1937a, Bruun and Kaiser
1943, Matthes 1963, Vandel 1965, Thinès 1969, Ercolini et al. 1982, Howes
1991) and was never rejected until Banister (1984) pointed out evident
osteological similarities to the genus Garra.
The hypothesis of a close phylogenetic relationship between Phreatichthys
and the genus Garra is supported by a recent molecular phylogenetic
study (Colli et al. 2008). The study, conducted on mitochondrial sequences
corresponding to the whole cytochrome b gene and to a fragment of the
12s rRNA gene, shows that Phreatichthys is firmly placed within a well-
supported clade composed exclusively by Asiatic and Middle Eastern
species of the genus Garra.
Finally, a rather close relationship with the other Somalian cave
cyprinid, Barbopsis devecchii, is shown by the results of an electrophoretic
study (Cobolli-Sbordoni et al. 1996). Allozymes analyses suggest that both
species could originate from a common epigean ancestor, widely spread
in the area in the past.

Phreatichthys andruzzii is one of the best adapted forms to the subterranean
habitat, as shown by the extent of its degenerative features. Scales and
pigmentation are fully absent (Vinciguerra 1924). The maximal eye
differentiation is reached 36 hours after egg laying, followed by a
degeneration so rapid and hard that one month later a rudimentary cyst
is all that remains (Berti et al. 2001). In the adult the anophthalmia is
complete, accompanied by the loss of the optic nerves; a strong reduction
of the entire encephalon, and in particular of the optic lobes, is also present
(Ercolini and Berti 1975).
No compensative improvement is revealed by any other sensory system.
A comparative survey of structural and ultrastructural features of the
olfactory epithelium revealed no significant qualitative nor quantitative
compensations in relation to the degeneration of the visual organ (Delfino
et al. 1981). While the cephalic lateral line system is perfectly developed,
according to the typical teleostean pattern, the trunk system appears
heavily degenerated. The canal is not continuous, wide interruptions
are present and only few and short tubes persist along its course (Berti,
unpublished data).
Finally, an histological and histochemical survey carried out on the
epidermis cells (Bianchi 1978a) showed, besides the characteristics of the
muciparous cells, the presence of club cells, the alarm substance cells.
The high degree of Phreatichthys adaptation to the subterranean life is
also confirmed by the presence of an extremely reduced metabolic rate.
During the tests carried out with a respirometer Phreatichthys shows an
oxygen consumption 2 to 2.5 times lower than Puntius lateristriga, while
under anoxia the time survival results 3 times higher (Ercolini et al. 1987).
Such a high capability to survive under anoxia is probably to be related to a
very peculiar mechanism of respiratory compensation. In all Osteichthyes
so far studied, the spleen stores and releases erythrocytes according to
the animal’s respiratory needs. Phreatichthys, on the contrary, uses its liver
rather than spleen as site of accumulation in the respiratory compensation
processes, like it occurs in the amphibian Rana esculenta (Frangioni et al.
1997).
Phreatichthys seems to tolerate large variations of water salinity, as
indicated by the 1:3 ratio of the conductivity values recorded at Bud Bud
and, respectively, at El Dirri (Ercolini et al. 1982)
The species is very easy to breed; both spontaneous and hormone
induced reproductions were repeatedly obtained in captivity. Phreatichthys
is a long-lived fish; at present wild individuals collected in 1977 and 1980
are still alive (personal observations).

Ethological Notes
In spite of the complete anophthalmia, the adult is clearly photosensitive.
When the fish had to choose between a dark and a light sector
(monochromatic equi-energetic lights of different wavelength), a clear
photophobic behaviour is shown, particularly under blue light (Ercolini
and Berti 1975).
Compared with Barbus filamentosus tested in absence of any visual
sensory stimulation, Phreatichthys shows a more efficient feeding
behaviour. The difference seems to be related to both the tendency to swim
closer to the bottom and the behavioural stereotypes of food localization
and intake. The food search is clearly guided by chemical stimulation, but
some data indicate that also mechanical stimulation could play some role
(Berti and Masciarelli 1993).
The chemical information is effective in orienting the fish locomotory
activity also. Phreatichthys andruzzii can discriminate both between
the odour of familiar and unfamiliar conspecifics (Berti et al. 1982) and
between the odour of unfamiliar conspecifics and heterospecifics (Berti et
al. 1989), showing different behavioural responses to the three chemical
cues: indifference to the odour of familiar conspecifics, low avoidance for
unfamiliar conspecifics, higher avoidance for heterospecifics (Berti and
Zorn 2001). It was hypothesised that avoidance of areas in which odours
of unfamiliar fishes are detected indirectly assures that each specimen
remains within its habitual living area. The response by P. andruzzii to
extraneous fish odours could be part of a behavioural mechanism of
spatial orientation based on chemical information.
Recently, it has been shown that fish swimming can be topographically
polarised by self-odour perception. When an unfamiliar area is
experimentally scented with fish self-odour, Phreatichthys behaves as if the
area was previously explored. The fish prefers an odour-free area to a self-
odour-scented one, and when offered the choice between a familiar and an
unfamiliar area, it prefers the unexplored environment. Avoidance of self-
odour-scented areas would allow effective exploration of the subterranean
environment, thus minimizing the risks of repeatedly exploring the same
water volumes (Paglianti et al. 2006).

FAMILY CYPRINIDAE
Barbopsis devecchii Di Caporiacco, 1926
Fig. 4 Barbopsis devecchii (photo by R. Innocenti; from Ercolini & Berti 1978).
Historical Notes
During the spring of 1924 (Di Caporiacco 1927) the Florentine professors
G. Stefanini and N. Puccioni made several naturalistic investigations in
the northern part of the Italian Somalia. Both were whole naturalists, a
paleontologist the former and an anthropologist the latter. When they
visited the village of Talèh they didn’t fail to collect the cyprinids living
in those wells. Back home, they delivered the fishes to Di Caporiacco who
described them as a new genus and species.
Derivatio Nominis
Barbopsis: from Barbus (because of its likeness to the genus Barbus) + yiV
opsis (semblance, likeness, in Greek)
devecchii: as a tribute to Cesare Maria De Vecchi, governor of Italian Somalia
in those years.

Synonyms
Barbopsis devecchi (Di Caporiacco 1926)
Eilichthys microphthalmus (Pellegrin 1929a, 1929b)
Barbopsis stefaninii (Gianferrari 1930)
Zaccarinia stefaninii (Gianferrari 1934)
Common name
None
Size
Up to 103.4 mm in total length and to 83.4 mm in standard length (Poll
1961).

The distribution area of Barbopsis is quite large, stretching over a wide
part of the Wadi Nogal Valley, in the North-Eastern Somalia. According
to the collection sites so far known (Table 3), it extends along a NW-SE
axis from Talèh to Eil, for at least 200 km in length. Eil is the site where
the two specimens of Barbopsis described by Pellegrin (1929a, 1929b) as
Eilichthys microphthalmus were collected. A survey of the region, made in
1980, greatly improved the knowledge of the species distribution, but was
unable to confirm the presence of the fish in that site (Ercolini et al. 1982).
Table 3 Known collection sites of Barbopsis devecchii; type locality: Talèh (from
Ercolini et al. 1982, modified).
Talèh 09°09’21”N 48°24’22”E 645 28

Gibaganle 08°30’51”N 48°39’14”E 395 29
El Goddomei 08°31’52”N 48°44’57”E 380 26
Bug Der 08°35’51”N 48°47’34”E 376 30
Callis 08°23’15”N 49°05’09”E 312 30.5
Eil 07°58’44”N 49°49’00”E 25 -
The geological nature of the territory, which includes the collection

wells, is rather uniform: it is formed by evaporitic formations of the lower
and middle Eocene (Talèh evaporites: anhydrites, gypsum, dolomites and
clays), except in the zone of Eil where an oligocenic formation prevails,
with marls and biological limestones.
The habitat of Barbopsis is probably constituted by a continuous water
layer, with fissures and small intercommunicating cavities. In Talèh, the

type locality, and surroundings, are present caves sometimes conspicuous

(Messana et al. 1985). Barbopsis devecchii is able to tolerate wide variations
of some environmental characteristics showing, as total salinity and
hardness, 1:2 ratios among different sites (Ercolini et al. 1982).
Some interesting stygobitic Crustacea have been found in the waters
where Barbopsis lives. The wells of Bug Der and Gibaganle are inhabited
also by the microphthalmic decapod Caridina lanzana (Ercolini et al. 1982),
while in Gibaganle, Callis and Eil the cirolanid isopod Haptolana somala
was found (Messana and Chelazzi 1984). Finally the springs of Eil host,
together with H. Somala, also the isopod Stenasellus migiurtinicus (Messana
and Chelazzi 1984). Probably, these crustaceans are the main trophic
resource of Barbopsis, which has a carnivorous attitude in captivity (personal
observations). In our opinion should be rejected the hypothesis of a mainly
phytophagous habit, based on the study of the feeding mechanism and of
the gut content of a single specimen (Matthes 1963). Moreover such a kind
of diet doesn’t fit (Proudlove 2006) to a species so much adapted to cave
life.

The phylogenetic affinity between Barbopsis and the genus Barbus was
stated since the original description of B. devecchii (Di Caporiacco 1926)
and was also suggested for both Eilichthys microphthalmus (Pellegrin 1929a,
1929b, 1932) and Barbopsis stefaninii (Gianferrari 1930, 1932, 1934, 1937a).
Both latter species were considered invalid by Poll (1961) and synonimised
with B. devecchii. However the close relationship of Barbopsis with the
genus Barbus was widely validated (Bruun and Kaiser 1943, Vandel 1965,
Thinès 1969, Ercolini et al. 1982, Howes 1991).
In the years some doubts arose about the value of such phylogenetic
affinities. Matthes (1963) stated that Barbopsis derived from a “primitive
Barbus type”, thus sharing the opinion Gianferrari (1937a) had formed on
the ground of the arrangement of the pharyngeal teeth, while Banister
(1984), in his paper on subterranean populations of Garra barreimiae, stated
that the relationships of Barbopsis are unclear, and that it does not have
any close relatives among the epigean fishes of north-eastern Africa. Some
indication on the topic can be given by a genetic study carried out with
allozyme molecular markers (Cobolli-Sbordoni et al. 1996) on Barbopsis and
Phreatichthys, the other cave cyprinid from Somalia. The paper revealed a
rather close relationship between the two species suggesting a common
origin from an extinct surface-dwelling ancestor.
We consider it worth, in this chapter, to take the opportunity to clarify
a small taxonomic mystery. Gianferrari in a paper of 1934, where she lists
all the African cavefishes, cites the cyprinids from Callis – previously

described as Barbopsis stefaninii (Gianferrari 1930) – as Zaccarinia stefaninii,

thus creating a new taxon in a completely anomalous manner as no
description has ever appeared in any journal so far. The fact should have
been brought to the author's attention as there is no trace of Zaccarinia
stefaninii in her successive articles.
Barbopsis devecchii shows less degenerative features than other African
cave cyprinids. The depigmentation is incomplete with different degrees
of reduction, the scales are thin but never lacking, the eyes are variously
reduced in size, sometimes sunken under the skin but still discernible.
Optic nerves and chiasma are present, but the optic lobes are widely
reduced and the retina is more or less degenerated according to the
degree of eye reduction. In cases of marked microphthalmia the retina
is strongly disrupted and neither retinal layers nor cones and rods are
always recognizable; on the contrary, slightly microphthalmic fish show
the typical succession of cellular layers and both types of visual cells are
still noticeable (Ercolini and Berti 1978).
In teleosts the pseudobranch is an organ directly involved in the eye’s
functioning. In Barbopsis the pseudobranch tends to degenerate and
disappears sooner than the eye; among the 64 specimens examined by
Bianchi and Ercolini (1984), 59 had both eyes, 4 had one eye and 1 had
no externally visible eyes, while 14 had both pseudobranchs, 9 had one
pseudobranch and 41 had no pseudobranchs. Furthermore, no correlation
existed between number of acidophil cells, pseudobranch volume and
external diameter of the ipsilateral eye; thus pseudobranch degeneration
and eye reduction seem to occur independently. Similarly, eye’s reduction
in size doesn’t seem to be related to the degree of depigmentation (Della
Croce 1963).
Finally, a comparative survey of structural and ultrastructural features
of the olfactory epithelium (Delfino et al. 1981) revealed no significant
degenerative nor compensative traits.
As well as the other Somalian cave cyprinid Phreatichthys andruzzii does,
Barbopsis exhibits a reduced metabolic rate. When tested in a respirometer
chamber specimens of the latter species showed an oxygen consumption
about 2 times lower than Puntius lateristriga, while under anoxia their time
survival resulted 1.5 to 2 times higher (Ercolini et al. 1987).
Ethological Notes
At present, the behaviour of Barbopsis is almost unexplored. As far as we
know only two studies, on the response to light and, respectively, on the

tendency to shoaling, were performed as yet. The former (Ercolini and

Berti 1978) dealt with fish responses to equi-energetic, achromatic and
monochromatic light stimuli; the results suggest that Barbopsis, although
it exhibits an evident photophobia, can still adapt itself to light even in
cases of marked eye reduction; this capacity determines a substantially
photo-indifferent response in animals previously kept under a normal
light-darkness cycle. The latter one (Jankowska and Thinès 1982) showed
the absence of any tendency to aggregate, since the mean density of fish
group fell within the theoretical limits of random distribution values.
FAMILY CLARIIDAE
Uegitglanis zammaranoi Gianferrari, 1923
Fig. 5 Uegitglanis zammaranoi (photo by R. Berti; from Ercolini et al. 1982,

modified).
book.
Historical Notes
The species was discovered by an Italian army officer, the major Vittorio
Tedesco Zammarano, who collected the first specimens from a well of
the village of Uegit, Southern Somalia, at an unknown date. Sent to Italy,
the fish was described by Dr Luisa Gianferrari, curator of zoology at the
Museo di Storia Naturale di Milano.
Derivatio Nominis
Uegitglanis: from Uegit (the type locality) + gla’ niV glanis (catfish, in
Greek);
zammaranoi: in honour of the collector.

Synonyms
Uegitglanis zammaranoi subsp. baidoaensis (Parenzan 1938)
Common Name
None
Size
Up to 250 mm in total length (Thinès 1958b).

The species inhabits the subterranean waters of a vast plateau of Southern
Somalia, between the course of Juba and Webi Shebeli, the two main rivers
of the region. At present four collection sites (Table 4) are known; the
presence of Uegitglanis in stations quite distant from one another (about
130 km between Manas and Siggia) suggests that the fish habitat is a wide
and continuous underground hydrographic system, with tunnels, fissures
and cavities, even of large dimensions such as in Manas (Thinès 1958b,
Messana et al. 1985).
Table 4 Known collection sites of Uegitglanis zammaranoi; type locality: Uegit

(from Ercolini et al. 1982, modified).
Uegit (2 sites) 03°48’26” N 43°14’49”E 371 30.5

Baidoa (2 sites) 03°06’58” N 43°38’48”E 433 28.5
Manas 02°53’28” N 43°27’42”E 428 29.5
Siggia 03°59’18” N 43°24’48”E 416 27
The distribution area is in Upper Jurassic formations, with organogenetic

and oolitic limestones, marls and gypsum (Merla et al. 1979). In spite of the
uniform geological structure, however, the physico-chemical features of
the waters inhabited by Uegitglanis show a wide variability. The species
clearly tolerates remarkable variations in water composition, at least as
regards salinity and hardness, and in some cases even highly foul waters,
with high contents of H2S (Thinès 1958b, Ercolini et al. 1982). Such wide
tolerance towards some chemical environment factors is confirmed by
Authors’ experience in breeding the species, kept in laboratory for about
10 years in very good conditions (unpublished data).
Reproduction in captivity has never been observed, in the wild
reproduction is presumed to occur between January and March, a period

characterized by a dry, hot climate; females with developed ovaries were

collected only in March, from the well of Manas.
Together with Uegitglanis cohabits Stenasellus pardii, a stygobitic isopod
found both at Baidoa and Uegit (Ercolini et al. 1982), which is probably the
principal if not the exclusive trophic resource of the fish.

Systematic position and phylogenetic affinities of Uegitglanis are quite
uncertain. The original description (Gianferrari 1923) stated a direct
derivation from the genus Clarias, on the basis, among other things, of
the presence of a suprabranchial organ and a bony capsule enveloping the
swim bladder. The presence of both structures was yet denied by other
studies, and thus the close relationship with Clarias was questioned. Besides
this, the characteristics also of both the skull and the Weberian apparatus
induced David (1936) to suggest that Uegitglanis occupies an intermediate
position between clariids and bagriid, and Chardon (1968) to create the
new family Uegitglanidae. Contrarywise, a quite different relationship
was proposed by Trewavas, when she stated that the closest genus was to
be identify with Gymnallabes (Trewavas 1936).
Molecular studies could clarify the phylogenetic relationships of this
species, at the moment considered a primitive Clariidae (Teugels and
Adriaens 2003)
Finally, the validity of the subspecies baidoaensis, erected by Parenzan
(1938) on the ground of both differences in size and a supposed habitat
fragmentation, has to be excluded (Ercolini et al. 1982).
The adaptation degree to the hypogean environment shown by Uegitglanis
is one of the highest among all cavefish. The eye disappeared and the
depigmentation is almost complete, light traces of pigment being shown
by both the meninx and the tegument between the nostrils. Besides the
absence of any ocular residue, anatomical and histological studies showed
the total disappearance of optic nerves and chiasma. The whole encephalon
is reduced in size, and the reduction is particularly enhanced at the level of
the optic lobes (Ercolini and Berti 1977).
The thyroid follicles (Olivereau 1960a, 1960b) are reduced in number
and not fully differentiated; moreover, they are characterized by some
lowering of activity levels.
On the contrary, at macroscopic, structural and ultrastructural level the
olfactory organ (Bianchi et al. 1978, Bianchi 1978b) does not exhibit any

degenerative nor compensative traits as regards the olfactory rosette and

the cytological components of the olfactory neuroepithelium. Likewise,
histological and histochemical studies performed about skin cells (Bianchi
1975) as well as about encephalic neurosecretory cells (Bianchi 1978c) do
not show any significant divergences from epigean species.
Many other aspects of the fish morphology and anatomy have been
taken into consideration, such as dentition (Gianferrari 1925), swim
bladder (Gianferrari 1937b), cranial osteology (David 1936) and Weberian
apparatus (Chardon 1968). The suprabranchial organ is fully absent (David
1936, Chardon 1968); the tendency to lose this air-breathing structure
is common to few other clariid genera and, with regard to Uegitglanis,
is probably to be related to its cave-dwelling lifestyle. Tortonese (1952)
suggested that the resultant inability to perform aerial respiration should
be compensated by the improved vascularisation of both anal and dorsal
fin.
Ethological Notes
Behavioural aspects so far investigated concern photic sensitivity and
aggressive behaviour.
Notwithstanding its complete anophthalmia Uegitglanis, as well as
other cave fish do, revealed to be photosensitive. Kept for long time under
12/12 light-dark cycle, Uegitglanis showed slightly photophobic responses
to achromatic light during the lighted phase, whereas during the dark
phase it didn’t exhibit any definite reaction (Ercolini and Berti 1977). Some
field observations, made in the wells of Uegit, Iscia Baidoa, and Manas
in the course of both day- and night-time, confirmed the laboratory data
(Berti, personal observations).
Agonistic interactions have been noticed in the field also (Berti and
Ercolini 1979), and laboratory observations allowed to define an exhaustive
description of the aggressive behaviour. Specimens raised in isolation and
tested in pairs immediately exhibit distinct agonistic behaviour, setting up
clear dominant-subordinate relationships. The immediate effect of dominance
seems to be the possession of the bottom by the dominant fish which
banishes the subordinate to midwater (Ercolini et al. 1981). Morphological
and quantitative analysis of the various aggressive patterns show that the
behavioural repertoire is very rich and highly articulate, similar to that of
epigean siluriforms (Berti et al. 1983).
As expected for a fish characterized by a such territorial behaviour no
tendency to school or shoal is exhibited, as evidenced by Jankowska and
Thinès (1982).

FAMILY CLARIIDAE
Clarias cavernicola Trewavas, 1936
Fig. 6 Clarias cavernicola (from Skelton & Teugels 1991; reproduced by kind
permission of the South African Institute for Aquatic Biodiversity).
Historical Notes
The first report about this species is probably that of the geologists Jaeger
and Waibel (1921). It is however only in 1933 that the first specimens were
collected, during the Dr Karl Jordan’s expedition. Later on the species was
described by Dr Ethelwyn Trewavas, ichthyologist at the British Museum
of Natural History of London.
Derivatio Nominis
cavernicola: from caverna (cave, in Latin) + colere (to dwell, in Latin).
Common Name
Cave catfish (English), spelonkbaber (Afrikaans), blinde Höhlenfisch
(German)
Size
Up to 161 mm in standard length (Bruton 1995).

Clarias cavernicola inhabits the crystal-clear waters of a subterranean small
lake, lying at the bottom of the Aigamas Cave (19°26’30”S, 17°17’26”E;
Skelton and Teugels 1991), the type locality, about 25 km north of Otavi,
Namibia. The distribution area is very restricted, although there are reports
of similar fishes from other caves of the same dolomitic limestone region,
such as Ajgab Cave, about 50 km west of Aigamas Cave, or Gamarab Cave,
near Outjo (Jubb 1969), the type locality is at present the one site where the
presence of the species was ascertained.

The cave has two entrances, about 100 m distant; the first one is a 4 m
chimney opening into a large chamber, the second one is a narrow opening
giving access to the cave whose floor descends steeply before dropping
almost vertically for about 20 m to the water surface (Skelton 1987, 1990a,
1990b). When the geologists Jaeger and Waibel (1921) visited the cave, the
water surface was about 18 m in length and 2.5 m wide, the water depth
was about 70 m, the temperature 24.5°C. Since then, some important
alterations probably occurred in the subterranean habitat, as suggested by
the fact that the water level has fallen by about 20 m (Skelton 1987, 1989),
and the water temperature was considerably increased, reaching in 1987
the value of 27°C (Bruton 1995).
The fishes seem to spent most part of their activity in the upper water
layers, seldom descending deeper than 15 m. They congregate and feed
exclusively over a shelf that slopes from just below the water surface
to several meters depth, and where trophic materials falling from the
cave tend to collect (Skelton 1990a, 1990b). The food supply seems to be
principally of exogenous nature, and consists of the carcasses of dead bats
and wild animals trapped in the cave, their parasites and these insects
feeding on such remains, of insects contained in bat and baboon feces, of
coprophagous invertebrates (Trewavas 1936, Skelton 1987, 1990a, 1990b,
Bruton 1995). Moreover, it seems very likely that fish feed also on the
small shrimps present in the lake as well as that the larger fish cannibalize
smaller individuals (Skelton 1989), an event not unusual in the cave
environment (Berti and Zorn 2001).
Both the aforementioned presence of small shrimps (Skelton 1989) and
the finding inside the fish stomach of some invertebrates such as a curved
white grub and a flatworm (Trewavas 1936), get us to hypothesize that
an associated fauna could inhabit the waters of the cave. In our opinion it
should be of great interest to investigate these aspects of the Agamas Cave
ecology.
So far no studies appeared on the ecology of Aigamas Cave, and the
population size is not exactly known. According to different assessments
the presence of 150-200 (Sefton et al. 1986) or 200-400 individual (Bruton
1995) was hypothesized.

Although some species had been proposed as closely related to C.
cavernicola the identification of its epigean ancestor, as well as it occurs
for many other cave fish, is still largely doubtful. The original description
(Trewavas 1936) hypothesized a relationship of C. cavernicola with either
C. allaudi or C. submarginatus or C. dumerilii. Jubb (1958, 1969) substantially
agreed on the hypothesis of a phylogenetic relationships with C. dumerilii,

because of both the morphological resemblance between the two fishes

and the distribution area of the epigean species, inhabiting the Cunene
and Ocavango rivers, to the north of the Otavi district. On the contrary,
in his systematic outline of the African species of the genus Clarias,
Teugels (1982a, 1982b, 1986) rejected the relationships with C. dumerilii,
confirming that one with C. allaudi and C. submarginatus, i.e. the two other
fishes mentioned by Trewavas (1936). According to Teugels (1982, 1986),
very closely related species could be C. nigromarmoratus and yet more C.
theodorae. The latter one was proposed as the closest species because of its
extraordinary similarity in shape and size of both head and body, and in
the position of the anal fin origin (Skelton and Teugels 1991). Moreover the
distribution area of C. theodorae is partially situated in two basins (Cunene
and Ocavango) adjacent to the Aigamas Cave.
Although C. cavernicola shows some degenerative traits clearly related to
cave life, the extent of its adaptation to the underground habitat is not so
high as in Uegitglanis zammaranoi, the other African cave clariid.
In some specimens the eyes are absent (Teugels 1986) but in most
individuals they are variously developed, often sunk in the socket, with
the surface skin more or less opaque (Trewavas 1936). The great variability
that seems characterize the eye degeneration is confirmed by the results of
a morphological study about brain and sense organs in C. cavernicola and
in the epigean species C. gariepinus and C. ngamensis (Bok 1968). While
the brain doesn’t show the striking reductions in size characterizing other
cave fishes, the cerebellum being its only reduced region, and while the
olfactory apparatus shows a slight decrease in development, not so marked
as to suggest that the cavefish has poorly developed olfactory powers,
the eyes and optic nerves of C. cavernicola have degenerated to a large
degree. Furthermore, there was found to be considerable variation in the
degree of degeneration of these two structures both among different fish
and also in the same individual, and the large variation in diameter of the
optic nerve was found to correlate with that one of the eye. Histological
sections of the most degenerated eyes show that choroid and retinal layers
are completely disorganized, and a few large vacuolated cells are all that
remains of the lens tissue; no rods or cones were identified (Bok 1968).
C. cavernicola as well as most cave-dwelling fishes, is devoid of pigment;
in life the skin colour is creamish flushed with pink or orange (Skelton and
Teugels 1991).
Unlike Uegitglanis the suprabranchial organ is still present but, in
common with many clariid genera (Teugels and Adriaens 2003), is greatly
reduced; therefore the suprabranchial chamber is only partly filled (Teugels

1986, Skelton and Teugels 1991). The reduction of the suprabranchial

organ has probably to be related to a loss of its functional significance, due
to either a lowered oxygen consumption or an high oxygen content of the
water or both, since C. cavernicola very seldom even comes to the surface
to gulp air (Hennig 1977).
In spite of the great attention this species deserves, no real scientific
studies were made on its biology and life history (Skelton 1987, 1990a,
1990b). No data are available on breeding habits, and several attempts
to breed captive individuals artificially have failed (Skelton 1990a, 1990b,
Bruton 1995).
Ethological Notes
At present, the behaviour of C. cavernicola is completely unexplored. The
only available records concern a few extemporary observations about its
spontaneous behaviour in both the field and aquarium (Scheide 1977,
Hennig 1977, Skelton 1987, 1990).
FAMILY ELEOTRIDAE
Typhleotris madagascariensis Petit, 1933
Fig. 7 Typhleotris madagascariensis (from Arnoult 1959a).
Historical Notes
The species was discovered about 1930 (the collection date is unknown)
by Henry Perrier de la Bathie and Mr Ursch, controller of the Eaux et
Forêts. Georges Petit described the fish as new genus and species three
years later.
Derivatio Nominis
Typhleotris: from tujlóV typhlos (blind, in Greek) + Eleotris (because of its
close relationship to that genus);
madagascariensis: from Madagascar.

Common Name
None
Size
Up to 67 mm in standard length (Arnoult 1959a) and to 100 mm in total
length (Maugé 1986).

The species inhabits the subterranean waters of a karst, included in the
Eocene or upper Cretaceous limestone formations (Arnoult 1959b, Kiener
1964, Decary and Kiener 1970) of the Mahafaly Plateau, southwestern
Madagascar. The type locality is the Mitoho Cave (24°03’S, 43°45’E, 50 m
a.s.l.; Goodman et al. 2005), Toliara (formerly Tuléar) Province. Besides
the type locality, six other sites are known as inhabited by the fish. From
north to south, they are the well of Ambilanilalika (about 40 km south
of the Onilahy River mouth), and the caves of Malazamanga, Lavaboro,
Lalia, Andramanoetse and Nikotsy, just north of Itampolo. The size of
the distribution area is rather uncertain; the eastern boundaries are still
unknown but, however, the fish habitat probably lengthen over a large
area south of the Onihaly River, since the distance between the two farthest
sites is about 90 km (Kiener 1964, Dekary and Kiener 1971).
The water table lies at a depth varying from 15 (Mitoho Cave) to 27
meters (well of Ambilanilalika) under the ground surface (Petit et al. 1935,
Poisson and Besairie 1947, Angel 1949).
Mitoho Cave, the type locality, was an underground chamber whose
ceiling weakened and partially collapsed, exposing a small lake that
represents the only accessible part of the subterranean hydrographic
system. Observed in the field, the fishes didn’t show any photophobic
behaviour, searching for food even in the lighted portion of the cave (Petit
1933, Petit et al. 1935). The diet is suggested to consist of either an aquatic
micro-fauna of epigean origin (Petit et al. 1935) or plankton (Kiener 1964)
or, more likely, cave crustacean and insects (Arnoult 1959a, 1959b). In
Mitoho Cave, the associated aquatic fauna comprehends both crustaceans
and insects: among the former, the cave decapod Typhlopatsa pauliani,
the isopod Anopsilana poissoni, some copepods (Tropocyclops confinis
and Diaptomus sp.) and some eyed tanaidaceans, among the latter the
hemipteran Microvelia mitohoi (Decary and Kiener 1970, Remillet 1973).


The validity of the genus and its close phylogenetic relationships with the
epigean genus Eleotris, stated by the original description (Petit 1933), has
always been shared by all authors dealing with this matter. On the contrary,
its inclusion in the family Eleotridae (Petit 1933, Petit et al. 1935, Petit 1941,
Thinès 1955a, Arnoult 1959a, Maugé 1986, Bauchot et al. 1991, Stiassny and
Raminosoa 1994, Lévêque 1997, Eschmeyer 1998, Romero and Paulson
2001) has been repeatedly called into question by some author assigning
the genus to the family Gobiidae (Thinès 1969, Thinès and Proudlove
1986). Finally, Hoese and Gill (1993), on the strength of a study of the
phylogenetic relationships of eleotridid fishes, suggested that Typhleotris
together with other eleotrid genera was to be considered belonging to
the family Gobiidae, subfamily Butinae. The matter is quite controversial
(Proudlove 2006), therefore we prefer to follow the traditional systematic
status and to consider Typhleotris as belonging to the family Eleotridae.
Typhleotris madagascariensis appears to be highly adapted to cave life. The
adult results fully blind: eyes are neither externally visible (Petit 1933,
Petit et al. 1935, Arnoult 1959a) nor found, sunken in the head soft tissues,
under dissection (Petit 1941). Unfortunately no histological studies were
performed to verify the degeneration degree, and the presence of eye
remnants can’t be excluded at all. As it occurs in the cavefishes of the
genus Astyanax (Breder 1944, Yamamoto et al. 2003) and in Caecobarbus
geertsii (Petit 1941) the eye degeneration induces a deep rearrangement
of the craniofacial skeleton, that in T. madagascariensis causes a reduced
mandibular mobility (Petit 1941). Degenerative features are shown by the
tegumentary system also. The scales are still present but very thin and soft;
they are both ctenoid and, predominantly, cycloid (Petit 1933, Petit et al.
1935, Angel 1949), with the latter ones being degenerated ctenoid scales
(Petit 1938b). The fishes are variously depigmented, a high intraspecific
variability occurring; the skin colour (light yellow, in alcohol) varies in
life from white, with a blood-red spot on the operculum (Petit 1933, Petit
et al. 1935), to brown-pinkish to brown (Angel 1949). On the contrary, no
variability is shown by the amount of vertebrae, without exception in
number of 25 (Decary and Kiener 1970).
Up to now, no physiological studies were conducted on T.
madagascariensis.

Ethological Notes
In the occasion of the fish discovery, Perrier de la Bathie observed for few
days its spontaneous behaviour in the field (Petit 1933, Petit et al. 1935).
Those observations, together with the few ones made in the same cave
some forty years later (Kiener 1964), constitute the only behavioural
records till now known.
FAMILY ELEOTRIDAE
Typhleotris pauliani Arnoult, 1959
Fig. 8 Typhleotris pauliani (from Arnoult 1959a).
Historical Notes
The first specimens were collected in July 1956 by Renaud Paulian,
Director of the I.R.S.M. (Institut de Recherche Scientifique de Madagascar), in
a cave near the Baie des Assassins (Murderers’ Bay), between the towns of
Morombe and Toliara (formerly Tuléar), south-western Madagascar. The
new species was described three years later by Jacques Arnoult, of the
zoological laboratory of the Muséum National d’Histoire Naturelle de Paris.
Derivatio Nominis
pauliani: in homage to R. Paulian, collector of the first specimens.
Common Name
None
Size
Up to 56 mm in standard length (Arnoult 1959a, 1959b) and to 65 mm in
total length (Maugé 1986).


The type locality is the Andranomaly Cave, near the Baie des Assassins (about
22°12’30’’S–43°17’30”E) and south-east of the village of Andolombezo.
The fish habitat is constituted by some pools, 150 m in diameter, 1,5 m in
depth; the water temperature is about 22°C. The fish diet probably consists
of small aquatic animals and insects, and bat guano (Arnoult 1959a, 1959b,
Remillet 1973).
On the outskirts of Andolombezo two other caves named Safora and
Ankilivona (Decary and Kiener 1971) are known as inhabited by the fish.
At present no other sites are known as housing the species, it is therefore
impossible to define or even hypothesize the extent of the fish distribution
area. Although Typhleotris pauliani and T. madagascariensis both inhabit
the subterranean waters of the same karst, in the Mahafaly Plateau, their
habitats seem to be fully distinct. Not only the nearest collection sites of
the two species are more than 300 km apart but also, and above all, any
hypothetical connection should be broken off by the valleys of Onilahy
and Fiherenana rivers (Arnoult 1959b, Paulian 1961, Kiener 1964).

On the ground of the high similarity and the evident close relationships
with Typhleotris madagascariensis, Arnoult didn’t erect any new genus for
the cavefish he described and only established a new species.
The validity of the genus and its close phylogenetic relationships with
the epigean genus Eleotris stated by Petit (1933), has always been shared
by all authors dealing with this matter. On the contrary, its inclusion in
the family Eleotridae (Petit 1933, Petit et al. 1935, Petit 1941, Thinès 1955a,
Arnoult 1959a, Maugé 1986, Bauchot et al. 1991, Stiassny and Raminosoa
1994, Lévêque 1997, Eschmeyer 1998, Romero and Paulson 2001) has been
repeatedly called into question by some author assigning the genus to the
family Gobiidae (Thinès 1969, Thinès and Proudlove 1986). Finally, Hoese
and Gill (1993), on the strength of a study of the phylogenetic relationships
of eleotridid fishes, suggested that Typhleotris together with other eleotrid
genera was to be considered belonging to the family Gobiidae, subfamily
Butinae. The matter is quite controversial (Proudlove 2006), therefore we
prefer to follow the traditional systematic status and to consider Typhleotris
as belonging to the family Eleotridae.
Typhleotris pauliani appears to be well adapted to the hypogean environment.
No traces of eyes are externally visible and the pigmentation is fully

absent, thus the skin colour is white with a light pink spot under the
operculum in life whereas the fish colour turns to “old ivory” in alcohol
(Arnoult 1959a, 1959b). The scales are almost exclusively cycloid and just
a few ctenoid scales are present (Arnoult 1959b).
Neither morpho-anatomical nor histological nor physiological studies
were done.
Ethological Notes
With the only exception of a few field observations on the swimming
modality (Arnoult 1959b), the behaviour of T. madagascariensis is still fully
unknown.
FAMILY GOBIIDAE
Glossogobius ankaranensis Banister, 1994
Fig. 9 Glossogobius ankaranensis (photo by Jane Wilson).

book.
Historical Notes
The fish was firstly seen and collected by members of the 1986 Crocodile
Caves Expedition to Madagascar (Wilson 1995, 1996). Deposited in the
collections of the Natural History Museum of London, was described
by Keith E. Banister who established the new species Glossogobius
ankaranensis.

Derivatio Nominis
ankaranensis: from the Ankarana karst, where the species lives.
Common Name
None
Size
Up to about 70 mm in total length (Wilson et al. 1988, Wilson 1996).

The species inhabits the subterranean waters of the Ankarana massif, an
outcrop of Middle Jurassic limestones in the northern Madagascar. At
present the fish is known only from the type locality, the Second River
Cave (about 12°50’S, 49°10’E). This is a cavity about 150 m long, and
the river flows slowly through about 100 m of low passage. The river is
about 2 m deep and 5 m wide, in September the water temperature is
21°C. The Second River Cave is the only accessible part of the fish habitat
so far known, but the species is thought to people also other sites of the
hydrographic subterranean system (Wilson 1996). Despite the small
dimensions of the massif (about 28 km in length and 8 km in width), the
cave complex is very long, with 98 km of cave passages already surveyed
(Wilson et al. 1988) and more still to be found (Wilson 1985, 1987a): it
seems very probable that the fish habitat extends over other parts, either
still unknown or inaccessible, of the karst.
G. ankaranensis occurs in great numbers in the watercourse, calmly
swimming even in a lighted portion of the cave, and lives in sympatry
with G. callidus, a related species normally eyed and pigmented (Wilson
1996). The epigean gobiid, smaller in size than G. ankaranensis, does not
seem to compete with this one; plausibly arrived in the cave with the rainy
season flood waters (Wilson 1987b, Wilson et al. 1988) it appears to be well
adapted to cave life, since it is present with very large populations (Wilson
1996). Finally an undetermined eel, probably belonging to a new species,
lives in most caves; up to now no specimens have been caught (Wilson
1987b).
The Ankarana underground waters house also a rich invertebrate fauna.
The copepod Mesocyclops leuckarti, the epigean decapod Macrobrachium
moorei, an undetermined depigmented crab and eleven decapods (eight
belonging to the genus Caridina and three to the genus Parisia) are at
present known (Remillet 1973, Wilson 1987b). Some decapods with
various degrees of cave life adaptation are endemic of the Ankarana

streams (Gurney 1984), while some new species belonging to the genera
Caridina and Parisia are waiting for their description (Wilson 1987b).
Although no fish had been observed feeding on these crustaceans, such
a rich invertebrate fauna probably constitutes the main part of the diet of
G. ankaranensis, together with organic matter washed in by floods, dead
insects, bat guano, swift droppings and feces of forest mammals coming
in to drink (Banister 1994, Wilson 1996).

In the original description (Banister 1994) the author pointed out that
the new species fitted within the definition of the genus Glossogobius
as outlined by Koumans (1953), hence there were no good reasons for
erecting a new genus. With regard to its phylogenetic relationships the
descriptor suggested that the new species is closely related to G. giuris, a
gobiid very similar in many characters to G. ankaranensis and widespread
over the Indo-Pacific area.
Glossogobius ankaranensis shows a high degree of adaptation to cave life.
The eye degeneration is almost complete, the only traces of eyes are two
minute black spots deeply embedded in the tissues. The fish are devoid
of pigment and are pink in colour when alive, ivory when preserved
in alcohol. The scales, both ctenoid and cycloid, are thin and weak; no
perforated lateral line scales are present (Banister 1994, Wilson 1996). At
present no histological data are available.
Ethological Notes
The ethology of Glossogobius ankaranensis is up to now completely
unexplored. The only known behavioural features are some field
observations on the swimming modalities. They are slow-moving in a
browsing swimming pattern, gently oscillating in the horizontal plane
(Wilson 1996).
Acknowledgements
We would like to thank the following colleagues and friends who kindly
provided help in the preparation of this chapter: Jane Wilson, who provided
the Glossogobius ankaranenensis picture; Gie Robeyns, who supplied
precious information about Caecobarbus geertsii; Alessandro Cianfanelli,
for his valuable help in both text revision and figures arrangement.

References
Angel, F. 1949. Contribution à l’étude du Typhleotris madagascariensis, poisson
aveugle, cavernicole, du Sud-Ouest de Madagascar. Bulletin du Muséum national
d’histoire naturelle, 2e série 21: 56-59.
Arnoult, J. 1959a. Poissons des eaux douces. Faune de Madagascar: Publications de
l’Institut de Recherche Scientifique, Tanarive-Tsimbazaza. Volume 10.
Arnoult, J. 1959b. Une nouvelle espèce de poisson aveugle de Madagascar:
Typhleotris pauliani n. sp. Memoires de l’Institut Scientifique de Madagascar. Série
A 13: 133-138.
Banister, K.E. 1984. A subterranean population of Garra barreimiae (Teleostei:
Cypinidae) from Oman, with comments on the concept of regressive evolution.
Journal of Natural History 18: 927-938.
Banister, K.E. 1994. Glossogobius ankaranensis, a new species of blind cave goby
from Madagascar (Pisces: Gobioidei: Gobiidae). Aqua 1: 25-28.
Bauchot, M.L., M. Desoutter, D.F. Hoese and H.K. Larson. 1991. Catalogue critique
des types de Poissons du Muséum national d’Histoire naturelle - (Suite) - Sous
ordre des Gobioidei. Bulletin du Muséum national d’histoire naturelle, 4e série,
Section A 13 (1-2, Supplément): 1-82.
Berti, R. and A. Ercolini. 1979. Aggressive behaviour in the anophthalmic phreatic
fish Uegitglanis zammaranoi Gianferrari (Claridae: Siluriformes). Monitore
zoologico italiano (N.S.) Supplemento 13: 197.
Berti, R. and G. Thinès. 1980. Influence of chemical signals on the topographic
orientation of the cave fish Caecobarbus geertsi Boulenger (Pisces, Cyprinidae).
Experientia 36: 1384-1385.
Berti, R. and L. Masciarelli. 1993. Comparative performances of non-visual food
search in the hypogean cyprinid Phreatichthys andruzzii and in the epigean
relative Barbus filamentosus. International Journal of Speleology 22: 121-130.
Berti, R. and L. Zorn. 2001. Locomotory responses of the cave cyprinid Phreatichthys
andruzzii to chemical signals from conspecifics and related species: New
findings. Environmental Biology of Fishes 62: 107-114.
Berti, R., G. Thinès and B. Lefèvre. 1982. Effets des informations chimiques provenant
d’ un milieu habité par des congénères sur l’ orientation topographique du
poisson cavernicole Phreatichthys andruzzii Vinciguerra (Pisces, Cyprinidae).
International Journal of Speleology 12: 103-117.
Berti, R., A. Ercolini and A. Cianfanelli. 1983. The aggressive behavior repertoire of
an anophthalmic phreatic fish from Somalia. Experientia 39: 217-219.
Berti, R., R. Vezzosi and A. Ercolini. 1989. Locomotory responses of Phreatichthys
andruzzii Vinciguerra (Pisces: Cyprinidae) to chemical signals of conspecifics
and of closely related species. Experientia 45: 205-207.
Berti, R., J.P. Durand, S. Becchi, R. Brizzi, N. Keller and G. Ruffat. 2001. Eye
degeneration in the blind cave-dwelling fish Phreatichthys andruzzii. Canadian
Journal of Zoology 79: 1278-1285.
Bianchi, S. 1975. Preliminary observations on the epidermis of Uegitglanis zammaranoi
Gianferrari (Claridae: Siluriformes.) Monitore zoologico italiano (N. S.) Supplemento
6: 91-101.

Bianchi, S. 1978a. Osservazioni preliminari sulle cellule mucose di Phreatichthys

andruzzii Vinciguerra, pesce freatico anoftalmo della Somalia. Rivista di Istochimica
normale e patologica 22: 15-18.
Bianchi, S. 1978b. Le cellule mucose e l’istochimica dei mucopolisaccaridi nell’epitelio
olfattivo di Uegitglanis zammaranoi Gianferrari (Clariidae, Siluriformes), pesce
freatico anoftalmo della Somalia. Rivista di Istochimica normale e patologica 22: 101-
107.
Bianchi, S. 1978c. Indagine preliminare sulla topografia delle cellule neurosecernenti
a secreto fucsinoparaldeide positivo nel S.N.C. di Uegitglanis zammaranoi
Gianferrari, pesce freatico anoftalmo della Somalia. Boll. Zool. 45 (Supplement):
8.
Bianchi, S. and A. Ercolini. 1984. Observations on the morphology and regression
of the pseudobranchs in Barbopsis devecchii Di Caporiacco microphthalmic
phreatobic cyprinid from Somalia. Monitore zoologico italiano (N.S.) Supplemento
18: 223-230.
Bianchi, S., G. Delfino and A. Ercolini. 1978. Morphology and fine structure of the
olfactory organ in Uegitglanis zammaranoi Gianferrari (Claridae Siluriformes),
anophthalmic phreatic fish from Somalia. Monitore zoologico italiano (N.S.)
Supplemento 10: 157-171.
Bok, A. 1968. A Morphological Study of the Brain and Sense Organs of the Cave Fish,
Clarias cavernicola Trewavas. Zoology Honours Project, Rhodes University,
Grahamstown.
Boulenger, G.-A. 1921. Description d’un poisson aveugle découvert par M. G.
Geerts dans la grotte de Thysville (Bas-Congo). Revue Zoologique Africaine 9:
252-253.
Breder, C.M. Jr. 1944. Ocular anatomy and light sensitivity studies on the blind fish
from Cueva de los Sabinos, Mexico. Zoologica 29: 131-144.
Bruton, M.N. 1995. Threatened fishes of the world: Clarias cavernicola Trewavas,1936
(Clariidae). Environmental Biology of Fishes 43: 162.
Bruun, A.F. and E.W. Kaiser. 1943. Iranocypris typhlops n. gen., n. sp., the first true
cave-fish from Asia. Danish Scientific Investigations in Iran 4: 1-8.
Chardon, M. 1968. Anatomie comparée de 1’appareil de Weber et des structures
connexes chez les Siluriformes. Annales du Musée royal de l’Afrique centrale. Série
in 8vo (Sciences zoologiques) 169: 1-277.
Chelazzi, L. and G. Messana. 1985. Acanthastenasellus forficuloides n. gen. n. sp., a
stenasellid isopod (Asellota) from Somalian phreatic layer. Monitore zoologico
italiano (N.S.) Supplemento 20: 43-54.
Cobolli Sbordoni, M., E. De Matthaeis, M. Mattoccia, R. Berti and V. Sbordoni.
1996. Genetic variability and differentiation of hypogean cyprinid fishes from
Somalia. J. Zool. Syst. Evol. Research 34: 75-84.
Colli, L., A. Paglianti, R. Berti, G. Gandolfi and J. Tagliavini. 2008. Molecular
phylogeny of the blind cavefish Phreatichthys andruzzii and Garra barreimiae
within the family Cyprinidae. Environmental Biology of Fishes doi: 10.1007/
s10641-008-9393-z.
David, L. 1936. Uegitglanis, Silure aveugle de la Somalie Italienne - Chainon entre
Bagrides et Clariides. Revue de Zoologie et de Botanique Africaines 28: 369-388.

Decary, R. and A. Kiener. 1970. Les cavités souterraines de Madagascar. Annales de

Spéléologie 25: 409-440.
Decary, R. and A. Kiener. 1971. Inventaire schématique des cavités de Madagascar.
Annales de Spéléologie 26: 31-46.
Delfino, G., S. Bianchi and A. Ercolini. 1981. On the olfactory epithelium in
cyprinids: A comparison between hypogean and epigean species. Monitore
zoologico italiano (N.S.) Supplemento 14: 153-180.
Della Croce, N. 1963. III Spedizione del Prof. G. Scortecci in Somalia. Osservazioni su
Barbopsis devecchii Di Caporiacco (Cyprinidae). Bollettino dei musei e degli istituti
biologici dell’Università di Genova 32 (1962-1963): 15-28.
Di Caporiacco, L. 1926. Un nuovo genere di Ciprinide somalo delle acque di pozzo.
Monitore zoologico italiano 37: 23-25.
Di Caporiacco, L. 1927. Pesci raccolti in Somalia nel 1924, dalla spedizione Stefanini e
Puccioni. Monitore zoologico italiano 38: 84-89.
Ercolini, A. and R. Berti. 1975. Light sensitivity experiments and morphology
studies of the blind phreatic fish Phreatichthys andruzzii Vinciguerra from
Somalia. Monitore zoologico italiano (N.S.) Supplemento 6: 29-43.
Ercolini, A. and R. Berti. 1977. Morphology and response to light of Uegitglanis
zammaranoi Gianferrari, anophthalmic phreatic fish from Somalia. Monitore
zoologico italiano (N.S.) Supplemento 9: 183-199.
Ercolini, A. and R. Berti. 1978. Morphology and response to light of Barbopsis
devecchii Di Caporiacco (Cyprinidae) microphthalmic phreatic fish from
Somalia. Monitore zoologico italiano (N.S.) Supplemento 10: 299-314.
Ercolini, A., R. Berti and A. Cianfanelli. 1981. Aggressive behaviour in Uegitglanis
zammaranoi Gianferrari (Clariidae, Siluriformes), anophthalmic phreatic fish
from Somalia. Monitore zoologico italiano (N.S.) Supplemento 14: 39-56.
Ercolini, A., R. Berti, L. Chelazzi and G. Messana. 1982. Researches on the phreatobic
fishes of Somalia: Achievements and prospects. Monitore zoologico italiano (N.S.)
Ercolini, A., R. Berti, L. Chelazzi and G. Messana. 1987. Oxygen consumption
in hypogean and epigean cyprinids (Pisces). Monitore zoologico italiano (N.S.)
Eschmeyer, W.N. (ed.). 1998. Catalog of Fishes. California Academy of Sciences, San
Francisco, 3 vols.
Frangioni, G., R. Berti and G. Borgioli. 1997. Hepatic respiratory compensation and
haematological changes in the cave cyprinid, Phreatichthys andruzzii. Journal of
Comparative Physiology B 167: 461-467.
Gérard, P. 1936. Sur l’existence de vestiges oculaires chez Coecobarbus geertsi.
Mémoires du Musée royal d’histoire naturelle de Belgique, 2e série 3: 549-552.
Gianferrari, L. 1923. Uegitglanis zammaranoi un nuovo Siluride cieco africano. Atti
della Società italiana di Scienze naturali 62; 1-3.
Gianferrari, L. 1925. Su la dentatura di Uegitglanis zammaranoi Gianf. Atti della Società
italiana di Scienze naturali 63: 326-334.
Gianferrari, L. 1930. Un nuovo ciprinide somalo (Barbopsis stefaninii Gianf.). Atti della
Società italiana di scienze naturali e del Museo civico di storia naturale 69: 106-111.

Gianferrari, L. 1932. Elenco dei pesci cavernicoli e loro distribuzione geografica.

Atti della Società italiana di scienze naturali e del Museo civico di storia naturale
71: 217-223.
Gianferrari, L. 1934. I pesci cavernicoli ciechi africani. Rassegna faunistica 1: 33-36.
Gianferrari, L. 1937a. Nuove osservazioni sui Ciprinidi di Callis (Somalia Italiana).
Atti della Società italiana di scienze naturali e del Museo civico di storia naturale
76: 198-204.
Gianferrari, L. 1937b. Su la vescica natatoria di Uegitglanis zammaranoi Gian. Atti
della Società italiana di Scienze naturali 76: 195-197.
Goodman, S.M., D. Andriafidison, R. Andrianaivoarivelo, S.G. Cardiff, E. Ifticene,
R.K.B. Jenkins, A. Kofoky, T. Mbohoahy, D. Rakotondravony, J. Ranivo,
F. Ratrimomanarivo, J. Razafimanahaka and P.A. Racey. 2005. The distribution
and conservation of bats in the dry regions of Madagascar. Animal Conservation
8: 153-165.
Gurney, A.R. 1984. Freshwater shrimp genera Caridina and Parisia (Decapoda:
Caridea: Atyidae) of Madagascar, with descriptions of four new species. Journal
of Natural History 18: 567-590.
Hennig, H. 1977. Note on the blind catfish Clarias cavernicola. S.W.A. Wissenshaftliche
Gesellschaft. Verein für Höhlenforschung, Arbeitsberichte 10: 11-12.
Heuts, M.J. 1952. Ecology, variation and adaptation of the blind African cave fish
Caecobarbus geertsii Blgr. Annales de la Société royale zoologique de Belgique 82
(1951): 155-230.
Heuts, M.J. 1953. Regressive evolution in cave animals. Symposia of the Society for
Heuts, M.J. and N. Leleup. 1954. La Géographie et l’Ecologie des Grottes du Bas-
Congo. Les Habitats de Caecobarbus geertsi Blgr. Annales du Musée royal du Congo
Belge. Série in 8vo (Sciences Zoologiques) 35: 1-71.
Hoese, D.F. and A.C. Gill. 1993. Phylogenetic relationships of eleotridid fishes
(Perciformes: Gobioidei). Bulletin of Marine Science 52: 415-440.
Howes, G.J. 1991. Systematics and biogeography: An overview. In: Cyprinid
Fishes — Systematics, Biology and Exploitation, I.J. Winfield and J.S. Nelson (eds).
Chapman & Hall, London, pp. 1-33.
Jaeger, F. and L. Waibel. 1921. Beiträge zur Landeskunde von Südwestafrika — Band II.
Landschaften des nördlichen Südwestafrika. E.S. Mittler und Sohn, Berlin (quoted
by Skelton 1990a).
Jankowska, M. and G. Thinès. 1982. Etude comparative de la densité de groupes de
poissons cavernicoles et épigés (Characidae, Cyprinidae, Clariidae). Behavioural
Processes 7: 281-294.
Jubb, R.A. 1958. A cave-dwelling cat-fish Clarias cavernicola Trewavas 1936, in
South West Africa. Piscator 43: 56-57.
Jubb, R.A. 1969. Fishes of the dolomitic limestone caves and sink holes of southern
Africa. Piscator 75: 15-20.

Kiener, A. 1964. De la présence de certaines populations ichthyologiques dans les

eaux souterraines des formations karstiques de la côte Ouest du Madagascar.
Bulletin du Madagascar 219: 573-592.
Koumans, F.P. 1953. Gobioidea. In: The Fishes of the Indo-Australian Archipelago, M.
Weber and L.F. De Beaufort (eds). E.J. Brill Ltd, Leiden. Volume 10.
Leleup, N. 1956a. La Faune cavernicole du Congo Belge et Considérations sur les
Coléoptères reliques d’Afrique intertropicale. Annales du Musée royal du Congo
Belge. Série in 8vo (Sciences Zoologiques) 46: 1-171.
Leleup, N. 1956b. Description de deux grottes du Bas-Congo. Notes Biospéologiques
11: 73-80.
Leleup, N. 1957. La faune cavernicole du Congo Belge. Folia scientifica Africae
centralis 3: 31-33.
Lévêque C. 1997. Biodiversity Dynamics and Conservation: The Freshwater Fish of
Tropical Africa. Cambridge University Press, Cambridge.
Matthes, H. 1963. A comparative study of the feeding mechanisms of some African
Cyprinidae (Pisces, Cypriniformes). Bijdragen tot de dierkunde 33: 3-35.
Maugé, A.L. 1986. Eleotridae. In: Checklist of the Freshwater Fishes of Africa (CLOFFA),
J. Daget, J.-P. Gosse and D.F.E. Thys van den Audenaerde (eds). ISNB, Bruxelles,
MRAC, Tervuren, ORSTOM, Paris, Volume 2, pp. 389-398.
Merla, G., E. Abbate, A. Azzaroli, P. Bruni, P. Canuti, M. Fazzuoli, M. Sagri and P. Tacconi.
1979. A Geological Map of Ethiopia and Somalia (1973) 1: 2.000.000 and Comment, with
a Map of Major Landforms. Consiglio Nazionale delle Ricerche, Firenze.
Messana, G. and L. Chelazzi. 1984. Haptolana somala n. sp., a phreatobic cirolanid
isopod (Crustacea) from the Nogal Valley (Northern Somalia). Monitore zoologico
italiano (N.S.) Supplemento 19: 291- 298.
Messana, G., L. Chelazzi and N. Baccetti. 1985. Biospeleology of Somalia. Mugdile
and Showli Berdi caves. Monitore zoologico italiano (N.S.) Supplemento 20: 325-
340.
Olivereau, M. 1960a. Quelques aspects anatomiques et physiologiques de la glande
thyroïde des Poissons. Annales de la Société royale zoologique de Belgique 90(2):
83-98.
Olivereau, M. 1960b. Etude anatomique et histologique de la glande thyroïde
d’Uegitglanis zammaranoi Gianferrari, poisson aveugle et cavernicole,
comparaison avec un Clariidae voisin, Clarias buthupogon A. Dum. Annales de la
Société royale zoologique de Belgique 90(2): 99-116.
Olivereau, M. and M. Herlant. 1954. Étude histologique de l’hypophyse de
Caecobarbus geertsii Blgr. Bulletin de l’Académie royale de Belgique. Classe des
sciences 40: 50-57.
Olivereau, M. and M. Francotte-Henry. 1955. Étude histologique et biométrique
de la glande thyroïde de Caecobarbus geertsi Blgr. Annales de la Société royale
zoologique de Belgique 86(1): 129-150.
Paglianti, A., G. Messana, A. Cianfanelli and R. Berti. 2006. Is the perception of
their own odour effective in orienting the exploratory activity of cave fishes?
Canadian Journal of Zoology 84: 871-876.

Parenzan, P. 1938. Pesci delle acque sotterranee della Somalia. Bollettino di Pesca, di
Piscicoltura e di Idrobiologia 14: 755-768.
Paulian, R. 1961. La zoogéographie de Madagascar et des îles voisines. Faune de
Madagascar: Publications de l’Institut de Recherche Scientifique, Tanarive-
Tsimbazaza. Volume 13.
Pellegrin, J. 1928. Les Poissons cavernicoles aveugles d’Afrique. Compte rendu de
l’Association française pour l’avancement des sciences: 409-412.
Pellegrin, J. 1929a. L’Eilichthys microphthalmus Pellegrin, Poisson cavernicole de la
Somalie italienne. Bulletin du Muséum national d’histoire naturelle, 2e série 1: 363-
364.
Pellegrin, J. 1929b. Sur un poisson cavernicole africain microphthalme. Compte rendu
hebdomadaire des séances de l’Académie des sciences 189: 204-205.
Pellegrin, J. 1932. Les Cyprinidés cavernicoles d’Afrique. Archivio zoologico italiano
16: 622-629.
Penney, A.J. 1988. Notes on the collection of blind cave catfish from Aigamas Cave
during Swex ’87. The Bulletin of the South African Spelaeological Association 29:
14-15.
Petit, G. 1933. Un poisson cavernicole aveugle des eaux douces de Madagascar:
Typhleotris madagascariensis gen. et. sp. nov. Compte rendu hebdomadaire des séances
de l’Académie des sciences 197: 347-348.
Petit, G. 1938a. Au sujet de Caecorbarbus geertsi Blgr. Bulletin de la Société zoologique
de France 63: 135-141.
Petit, G. 1938b. Sur Typhleotris madagascariensis G. Petit. Bulletin du Muséum national
d’histoire naturelle, 2e série 10: 491-495.
Petit, G. 1941. Sur la morphologie cranienne de deux poissons cavernicoles
aveugles: Typhleotris madagascariensis G. Petit et Coecobarbus geertsi Blgr. Compte
rendu des travaux de la Faculté des sciences de Marseille 1: 36-40.
Petit, G. and W. Besnard. 1937. Sur le comportement en aquarium du Caecobarbus
geertsii Boulenger. Bulletin du Muséum national d’histoire naturelle, 2e série 9: 50-
53.
Petit, G., L. Germain, Th. Monod and F. Angel. 1935. Contribution à l’étude
faunistique de la réserve naturelle du Manampetsa (Madagascar). Annales des
Sciences Naturelles, Zoologie, 10e série 18: 421-481.
Poisson, H. and H. Besairie. 1947. Note sur une nouvelle station d’un poisson
cavernicole aveugle du plateau calcaire Mahafaly. Bulletin de l’Academie Malgache
N.S. 26: 167-168.
Poll, M. 1961. Contribution a l’étude des poissons d’eau douce de Somalie
appartenant aux genres Gobius et Barbopsis. Bollettino dei musei e degli istituti
biologici dell’Università di Genova 31: 15-35.
Proudlove, G.S. 2006. Subterranean Fishes of the World — An Account of the
Subterranean (Hypogean) Fishes Described up to 2003 with a Bibliography 1541-2004.
International Society for Subterranean Biology, Moulis.
Proudlove, G.S. and A. Romero. 2001. Threatened fishes of the world: Caecobarbus
geertsii Boulenger, 1921 (Cyprinidae). Environmental Biology of Fishes 62: 238.

Quaghebeur, M. 1955. Onderzoekingen over blinde grotvissen in verband met de reductie

in de ontwikkeling van hun ogen. Thesis, University of Louvain (quoted by Thinès,
1969).
Quaghebeur, M. 1957. Onderzoek over de reductie van de ogen, van de oogkas en van
het tectum opticum bij blinde grotvissen. Thesis, University of Louvain (quoted by
Thinès, 1960).
Remillet, M. 1973. Aperçu de la faune souterraine à Madagascar. In: Livre du
cinquantenaire de l’Institut de Spéologie “Émile Racovitza”, T. Ordigan (ed.). Editura
Academiei Republicii Socialiste România, Bucureşti, pp. 135-160.
Romero, A. and K.M. Paulson. 2001. It’s a wonderful hypogean life: A guide to the
troglomorphic fishes of the world. Environmental Biology of Fishes 62: 13-41.
Ruffo, S. 1982. Studi sui Crostacei Anfipodi. 92. Nuovi Anfipodi di acque sotterranee
della Somalia. Monitore zoologico italiano (N.S.) Supplemento 17: 97-113.
Sefton, M., J. Martini and R. Ellis. 1986. Cave description—Aigamas Cave. The
Bulletin of the South African Spelaeological Association 27: 91-92 (quoted by Penney,
1988).
Skelton, P.H. 1987. South African Red Data Book—Fishes. South African National
Scientific Programmes Report No. 137
Skelton, P.H. 1989. Gems of a different kind. Rössing (Winhoek, Namibia) October
1989: 6-11.
Skelton, P.H. 1990a. The status of fishes from sinkholes and caves in Namibia.
Namibia Scientific Society, Journal 42: 75-83.
Skelton, P.H. 1990b. The status of fishes from sinkholes and caves in Namibia.
Bulletin of the South African Speleological Association 31: 77-81.
Skelton, P.H. and G.G. Teugels. 1991. A review of the clariid catfishes (Siluroidei,
Clariidae) occurring in southern Africa. Revue d’Hydrobiologie Tropicale 24: 241-
260.
Stiassny, M.L.J. and N. Raminosoa. 1994. The fishes of the inland waters of
Madagascar. In: Biological Diversity of African Fresh- and Brackish Water Fishes.
G.G. Teugels, J.-F. Guégan and J.-J. Albaret (eds). Annales du Musée royal de
l’Afrique centrale (Sciences zoologiques) 275: 133-148.
Teugels, G.G. 1982a. A systematic outline of the African species of the the genus
Clarias (Pisces: Clariidae), with an annotated bibliography. Annales du Musée
royal de l’Afrique centrale. Série in 8vo (Sciences zoologiques) 236: 1-249.
Teugels, G.G. 1982b. Preliminary data of a systematic outline of the African species
of the genus Clarias (Pisces: Clariidae). Revue de Zoologie africaine 96: 731-748.
Teugels, G.G. 1986. A systematic revision of the African species of the genus Clarias
(Pisces: Clariidae). Annales du Musée royal de l’Afrique centrale (Sciences zoologiques)
247: 1-199.
Teugels, G.G. and D. Adriaens. 2003. Taxonomy and phylogeny of Clariidae — An
overview. In: Catfishes, G. Arratia, B.G. Kapoor, M. Chardon and R. Diogo (eds),
Science Publishers, Inc., Enfield (NH), USA, pp. 465-487.

Thinès, G. 1953. Recherches expérimentales sur la photosensibilité du poisson

aveugle Caecobarbus geertsii Blgr. Annales de la Société royale zoologique de Belgique
84: 231-265.
Thinès, G. 1955a. Les poissons aveugles (I)—Origine, taxonomie, répartition
géographique, comportement. Annales de la Société royale zoologique de Belgique
86(1): 5-128.
Thines, G. 1955b. Étude comparative de la photosensibilité des poissons aveugles
Caecobarbus geertsii Blgr et Anoptichthys jordani Hubbs et Innes. Annales de la
Société royale zoologique de Belgique 85 (1954): 35-58.
Thinès, G. 1958a. Beobachtungen über die Phototaxis und die Thermotaxis des
blinden Höhlenfisches, Caecobarbus geertsi Blgr (Cyprinidae). Experientia
14: 381-382.
Thinès, G. 1958b. Observations sur les habitats de l’Uegitglanis zammaranoi Gianferrari
1923, Clariidae aveugle de la Somalie Italienne. Revue de Zoologie et de Botanique
africaines 52: 117-124.
Thinès, G. 1960. Sensory degeneration and survival in cave fishes. Symposia of the
Zoological Society of London 3: 39-52.
Thinès, G. 1969. L’évolution régressive des Poissons cavernicoles et abyssaux. Masson
et Cie., Paris.
Thinès, G. and N. Wissocq. 1972. Etude comparée du comportement alimentaire de
deux poissons cavernicoles (Anoptichthys jordani Hubbs et Innes et Caecobarbus
geertsi Blgr). International Journal of Speleology 4: 139-169.
Thinès, G. and J.-M. Legrain. 1973. Effets de la substance d’alarme sur le
comportement des poissons cavernicoles Anoptichthys jordani (Characidae) et
Caecobarbus geertsi (Cyprinidae). Annales de Spéléologie 28: 291-297.
Thinès, G. and G. Proudlove. 1986. Pisces. In: Stygofauna mundi—A Faunistic,
Distributional, and Ecological Synthesis of the World Fauna Inhabiting Subterranean
Waters (including the Marine Interstitial). L. Botosaneanu (ed.), E.J. Brill/Dr. W.
Backhuys, Leiden, pp. 709-733.
Thys van den Audenaerde, D.F.E. 1999. A peacock, a blind barbel, a railway wagon
and a disappointed ichthyologist. Ichthos 62: 8-11.
Tortonese, E. 1952. Intorno al Siluroide cieco Uegitglanis zammaranoi (Gianf.) con
particolare riguardo ai dispositivi respiratorii. Bollettino dell’Instituto e Museo di
Zoologia della Università di Torino 3 (1951-1952): 163-170.
Trewavas, E. 1936. Dr. Karl Jordan’s expedition to South-West Africa and Angola:
The fresh-water fishes. Novitates Zoologicae 40: 63-74.
Vandel, A. 1965. Biospeleology—The Biology of Cavernicolous Animals. Pergamon Press
Ltd., Oxford.
Vinciguerra, D. 1924. Descrizione di un ciprinide cieco proveniente dalla Somalia
Italiana. Annali del Museo civico di storia naturale Giacomo Doria 51 (1923): 239-243.
Wilson, J.M. 1985. Ecology of the crocodile caves of Ankarana. Cave Science 12:
135-138.
Wilson, J.M. 1987a. The crocodile caves of Ankarana, Madagascar. Oryx 21: 43-47.

Wilson, J. 1987b. The crocodile caves of Ankarana: expedition to northern

Madagascar, 1986. Other cave fauna. Cave Science 14: 115-116.
Wilson, J. 1995. Lemurs of the Lost World — Exploring the Forests and Crocodile Caves
of Madagascar. Impact Books, London. 2nd edition.
Wilson, J.M. 1996. Conservation and ecology of a new blind fish Glossogobius
ankaranensis from the Ankárana Caves, Madagascar. Oryx 30: 218-221.
Wilson, J.M., P.D. Stewart and S.V. Fowler. 1988. Ankarana—a rediscovered nature
reserve in northern Madagascar. Oryx 22: 163-171.
Yamamoto, Y., L. Espinasa, D.W. Stock and W.R. Jeffery. 2003. Development
and evolution of craniofacial patterning is mediated by eye-dependent
and-independent processes in the cavefish Astyanax. Evolution & Development
5: 435-446.
Further Reading
Botosaneanu, L. (ed.). 1986. Stygofauna Mundi—A Faunistic, Distributional, and
Ecological Synthesis of the World Fauna Inhabiting Subterranean Waters (Including
the Marine Interstitial). E.J. Brill, Leyden.
Culver, D.C. and W.B. White (eds.). 2005. Encyclopedia of Caves. Elsevier/Academic
Press, Amsterdam/Boston.
Gunn, J. (ed.). 2005. Encyclopedia of Caves and Karst Science. Fitzroy Dearbone, New
York.
Jeannel, R. and E.-G. Racovitza. 1914. Énumération des Grottes visitées, 1911-1913
(5e série) Biospeologica XXXIII. Archives de Zoologie Expérimentale et Générale 53:
325-558.
Juberthie, C. and V. Decou (eds.). 1994-2001. Encyclopaedia Biospeologica. Société
Internationale de Biospéologie, Moulis. Volume 1 (1994), Volume 2 (1998),
Volume 3 (2001).
Wilkens, H., D.C. Culver and W.F. Humphreys (eds.). 2000. Subterranean Ecosystems,
Ecosystems of the World. Elsevier, Amsterdam. Volume 30.

CHAPTER 11
Subterranean Fishes of China
Chen Zi-Ming1, Luo Jing1, Xiao Heng1 and Yang Jun-Xing2
1
School of Life Science, Yunnan University, Kunming,
Yunnan, 650091, P. R. China. E-mail: cziming@hotmail.com
2
Kunming Institute of Zoology, Chinese Academy of Sciences, Kunming,
Yunnan, 650223, P. R. China
INTRODUCTION
Subterranean or hypogean fishes are herein defined as those living in [an
occluded, semi-occluded] a cave or subterranean river system during
their whole life. Among these, species restricted to subterranean habitats
(troglobites) usually show a series of autapomorphies (troglomorphisms),
including morphological, physiological, behavioral and ecological features,
which can be directly or indirectly related to the hypogean life, being
either adaptive or not (Trajano 2001). The most common morphological
traits characterizing cavefishes and other troglobites are the reduction of
eyes and pigmentation when compared to their surface relatives. Eyes are
vestigial or absent and dark pigmentation may reduce to the point that
the body appears to be transparent. In addition, the numbers and sizes of
mechano- and chemo-reception organs are often increased (Greenwood
1967, Chen et al. 1997, Chen et al. 1998, Chen et al. 2001).
To a certain degree, it can be said that cavefishes are living fossils
in the process of evolution (Chen et al. 2001). Because of their unusual
appearance, local people also call them “transparent fishes”, “glass
fishes”, “blind fishes” or “eyeless fishes”. Hypogean fishes have always
been a source of amazement and amusement throughout the history of
ichthyology (Romero and Paulson 2001).
The authors contributed equally to this work

There are 3.44 million km² of karst landform areas in China, the largest
part, about 0.55 million km², concentrated in southwest China in [namely]
Yunnan, Guizhou, Guangxi, Chongqing, Sichuan, Hubei and Hunan
Provinces (Li et al. 2002). In these regions, well developed karst landforms,
abundant precipitation and multitudinous river systems provide good
conditions for the formation of caves and evolution of troglobitic fishes
(Chen et al. 2001).
I. A BRIEF HISTORY OF DISCOVERY AND

DESCRIPTION
The earliest record of the hypogean fishes in the world is Sinocyclocheilus
hyalinus Chen and Yang, 1993 (Chen et al. 2001). This fish species occurs in
Alugu Cave, Luxi County, Yunnan Province and was called ‘transparent
fish’ on a stone epigraph located at the entrance of Alugu Cave and dated
to China’s Ming Dynasty in 1,541 (Xie 1541). It was not formally described
until 1993. After 1993, many Chinese cavefishes have been discovered and
described.
On 23 April 1976, a little fish (Typhlobarbus nudiventris Chu and Chen
1982) was found 100 meters deep in Jianshui County, Yunnan Province,
where the local people were digging to find a new water source. Following
that, nine blind loaches, Triplophysa gejiuensis (Chu and Chen 1979), were
discovered in an underground river about 400 m deep in Gejiu County,
Yunnan Province, in 1978. After these two discoveries, the interest in
Chinese hypogean fishes has been aroused among ichthyologists in China
and abroad resulting in an upsurge of exploration and study. The number
of troglomorphic fishes described for China increased rapidly, from eight
species in 1990 (Chen 1990), to 11 in 1993 (Chen and Yang 1993), 21 in
2001 (Chen et al. 2001), 31 in 2006 (Zhao and Zhang 2006 ) and 51 in 2008
(Fig. 1).
II. DIVERSITY OF HYPOGEAN FISHES OCCURRING IN

CHINA
II.1 Features of Hypogean Fishes

So far, 51 species of troglomorphic fishes have been discovered and
formally described (Table 1). The composition of the Chinese hypogean
fish fauna has the following features:
(1) The richness of troglobitic fish species is great compared to the total
number (151) recorded worldwide (Proudlove, this volume);

Chen Zi-Ming et al. 399
60
51
50
Number of species
40
31
30
21
20
11
10 8
0
1990 1993 2001 2006 2008
Fig. 1 The increased number of described hypogean fish species occurring in

China through time.
(2) All the troglobitic fishes occurring in China belong to the Order
Cypriniformes;
(3) Most of these troglobitic fishes are in two genera in two families.
The genus Sinocyclocheilus (Cyprinidae) has the largest number of
troglobitic species (28), accounting for 54.9% of the total. The genus
Triplophysa (Balitoridae) has 10 troglobitic species, accounting for
19.6% of the total;
(4) Many species, especially in Cyprinidae, have more developed
barbels than their epigean congeners. As one of the important
sense organs, barbels may function importantly in the evolutionary
transition from the earth’s epigean water to the dark hypogean
water environments (Chen et al. 2001);
(5) All the known troglobitic species seem to be endemic to very
small areas and to have small populations. Some species, such as
T. nudiventris Chu and Chen, 1982, have been collected only from
wells. Several species, such as T. nudiventris Chu and Chen, 1982,
Longanalus macrochirous Li, Ran and Chen, 2006, T. rosa Chen and
Yang, 2005, are only known from a single specimen. Only a few
species are represented by several small populations, such as S.
rhinocerous Li and Tao, 1994 (Li et al. 2000);
(6) The number of specimens by which new species have been described
is usually low: 27 species (52.9% of the total) were described from
five specimens or less, and eight (15.7%) were described from a
single specimen (Table 1); in other cases, the type material is in
different collections (at least 13 up to now), hampering comparisons
and further research (Zhao and Zhang 2006a).

Table 1. The taxa, distribution localities, habitats, degree of troglomorphism and specimens of the hypogean fishes occurring in

China (the species were grouped by family following Nelson’s (1994) systematic order), types of habitat (according to classification
in Trajano 2001) and degree of troglomorphism. Abs.= not visible externally; Var. = variable in the population; Red. = slightly
reduced; Polim. = population with distinct morphs. 1. slightly, but significantly reduced eyes/melanic pigmentation, with narrow
intrapopulation variation; 2. wide variation, from slightly reduced to not visible externally eyes/pigmentation; 3. most individuals with
deeply reduced eyes/pigmentation; 4. all individuals completely (externally) anophthalmic and/or depigmented; 4+. DOPA positive
(melanin is then synthesized and melanin cells become visible when in light); 4– DOPA negative (there is no response to light);
Polim. Polymorphic phenotypes (according to classification in Trajano and Bichuette, this volume)
Taxon Locality Habitat Eyes Pigm. Specimens
Order Cypriniformes
Family Cyprinidae
Subfamily Barbinae
1 Sinocyclocheilus albeoguttatus Leye, Guangxi Upper phreatic 4 4 1
(Zhou and Li 1998) zone (cave)
2 Sinocyclocheilus altishoulderus Taiping township, Base-level 3 3 13
(Li and Lan 1992) Donglan, Guangxi stream
#
3 Sinocyclocheilus anatirostris Lingyun and Leye Base-level 4 4 6
(Lin and Luo 1986) County, Guangxi stream
#
4 Sinocyclocheilus angularis Panxian, Guizhou Flooded caves 3 3 3
(Zheng and Wang 1990)
#
5 Sinocyclocheilus anophthalmus Jiuxiang township, Base-level 4 4 8
(Chen and Chu 1988) Yiliang, Yunnan stream
6 Sinocyclocheilus aquihornes Qiubei, Yunnan Base-level stream 4 4 1
(Li and Yang 2007)
7 Sinocyclocheilus bicornutus Xingren and Zhenfeng Base-level stream 3 4+ 9
(Wang and Liao 1997) County, Guizhou
8 Sinocyclocheilus brevibarbatus Gaoling Town, Du’an Base-level stream 1 1 7
(Zhao et al. 2008) County, Guangxi
Table 1 Contd..
Table 1 Contd..
9 Sinocyclocheilus broadihornes Shilin, Yunnan Base-level stream 3 3 9

(Li and Mao 2007)
#
10 Sinocyclocheilus cyphotergous (Dai 1988) Luodian, Guizhou Base-level stream 3 4 4
11 Sinocyclocheilus donglanensis Donglan, Guangxi Flooded cave Abs Abs 5
(Zhao et al. 2006)
12 Sinocyclocheilus furcodorsalis Tian’e, Guangxi Upper phreatic zone 4 4 4
(Chen et al. 1997) (cave)
13 Sinocyclocheilus guangxiensis (Zhou and Li Longhuang cave, Tongle Base-level stream 4 4 3
1998) township, Leye, Guangxi
14 Sinocyclocheilus guilinensis (Ji 1985) Guilin, Guangxi Upper phreatic zone 1 1 7
(cave)
15 Sinocyclocheilus halfibindus (Li and Lan Fengchen township, Base-level stream 4 4 25
1992) Fengshan, Guangxi
16 Sinocyclocheilus huaningensis (Li 1998) Dalongtan,Huaning, Yunnan Upper phreatic zone 1 1 26
17 Sinocyclocheilus hugeibarbus (Li and Ran 2003) Libo, Guizhou Base-level stream 3 3 5
#
18 Sinocyclocheilus hyalinus (Chen and Yang 1993) Alugu cave, Luxi, Yunnan Base-level stream 4 4+ 2
19 Sinocyclocheilus jiuxuensis (Li and Lan 2003) Jiuxu township, Hechi, Guangxi Base-level stream 3 4 6
20 Sinocyclocheilus liboensis Libo, Guizhou Upper phreatic zone 1 1 2
(Li et al. 2004) (cave)

21 Sinocyclocheilus lingyunensis Lingyun, Guangxi Base-level stream 3 4 13
(Li et al. 2000)
22 Sinocyclocheilus longifinus (Li 1998) Dalongtan, Huaning, Yunnan Upper phreatic zone 1 1 2
#
23 Sinocyclocheilus microphthalmus (Li 1989) Luolou township, Base-level stream 3 4 15
Lingyun, Guangxi
#
24 Sinocyclocheilus rhinocerous (Li and Tao 1994) Xinzhai, Luoping, Yunnan Flooded caves 2 3 2
25 Sinocyclocheilus tianeensis Tian’e, Guangxi Upper phreatic 4 4 11
(Li et al. 2003) zone (cave)
26 Sinocyclocheilus tianlinensis Pingshan township, Upper phreatic 4 4 1
(Zhou et al. 2004) Tianlin, Guangxi zone (cave) Table 1 Contd..
Table 1 Contd..

27 Sinocyclocheilus tileihornes
(Mao et al. 2003) A’gan, Luoping, Yunnan Base-level stream 3 4 7
28 Sinocyclocheilus xunleensis Xunle township, Upper phreatic 4 4 5
(Lan et al. 2004) Huangjiang, Guangxi zone (cave)
#
29 Typhlobarbus nudiventris Yangjieba, Jianshui, Yunnan Upper phreatic 4 4 3
(Chu and Chen 1982) zone (cave)
Subfamily Labeoninae
30 Longanalus macrochirous Libo, Guizhou Upper phreatic Abs Abs 1
Family Balitoridae
Subfamily Nemacheilinae
31 Heminoemacheilus hylinus Bao’an Township, Base-level stream 4 4 6
(Lan et al. 1996) Du’an County, Guangxi
#
32 Oreonectes anophthalmus (Zheng 1981) Qifeng mountaion, Upper phreatic 4 4 7
Wuming, Guangxi zone (cave)
33 Oreonectes furcocaudalis (Zhu and Cao 1987) Rongshui County, Guangxi Base-level stream 4 4 3
34 Oreonectes microphthalmus Du’an County, Guangxi Upper phreatic zone 4 4 6
(Du et al. 2008)
#
35 Oreonectes retrodorsalis Liuzhai township, Base-level stream 4 4 7
(Lan et al. 1995) Nandan, Guangxi
36 Oreonectes translucens Xia’ao Village, Du’an Upper phreatic 4 4 3
(Zhang et al. 2006) County, Guangxi zone (cave)
#
37 Paracobitis longibarbatus Libo, Guizhou Flooded caves 4 4 3
(Chen et al. 1998)
38 Paracobitis maolanensis Maolan, Guizhou Upper phreatic 4 4 1
39 Paracobitis posterodarsalus Nandan, Guangxi Upper phreatic 4 4 1
40 Triplophysa aluensis (Li and Zhu 2000) A’lugu cave, Luxi, Yunnan Base-level stream 3 3 1
Table 1 Contd..
Table 1 Contd..
#
41 Triplophysa gejiuensis Kafan, Gejiu, Yunnan Base-level stream 4 4 9
(Chu and Chen 1979)
42 Triplophysa nandanensis Nandan, Guangxi Upper phreatic 4 4 11
(Lan et al. 1995) zone (cave)
43 Triplophysa qiubeiensis Nijiao, Qiubei, Yunnan Upper phreatic 4 4 11
44 Triplophysa rosa (Chen and Yang 2005) Wulong, Chongqing Upper phreatic 4 4 1
zone (cave)
*
45 Triplophysa sp. Leye, Guangxi Upper phreatic 3 3 10
(Chen et al. 2010) zone (cave)
#
46 Triplophysa shilinensis Weiboyi, Shilin, Yunnan Upper phreatic 4 4 2
(Chen and Yang 1992) zone (cave)
47 Triplophysa tianeensis Tian’e, Guangxi Base-level stream 4 4 6
(Chen et al. 2004)
48 Triplophysa xiangxiensis Longshan, Xiangxi, Base-level stream 4 4 3
(Yang et al. 1986) Hunan
#
49 Triplophysa yunnanensis (Yang 1990) Jiuxiang township, Base-level stream 1 1 4
Yiliang, Yunnan
Family Cobitidae

Subfamily Cobitinae
50 Protocobitis polylepis Wuming, Guangxi Upper phreatic zone 4 4 2
(Zhu et al. 2008)
51 Protocobitis typhlops (Yang and Chen 1994) Xia’ao, Du’an, Guangxi Upper phreatic 4 4 7
zone (cave)
Total: 1 order 3 families 4 subfamilies 8 genera 51 species.

*unpublished
#
Being listed in ‘China Species Red List’ (Wang and Xie 2004).
II.2 Geographic Distribution Patterns

All the 45 described Chinese troglobitic fishes occur in south and southwest
China with 27 species in Guangxi, 14 in Yunnan, 8 in Guizhou, one in
Hunan and one in Chongqing (Fig. 2). There are 0.55 million km² of karst in
southwest China, namely Yunnan, Guizhou, Guangxi, Chongqing, Sichuan,
Hubei and Hunan Provinces, where factors such as the well developed
karst landforms, abundant precipitation and multitudinous river systems
provide good conditions for the evolution of troglobitic populations, and
abundant potential for the discovery of more troglobitic species. By 1987,
there were more than 70 cave species being described in China. Besides
these, many other new species have been discovered. These include snails,
shrimps, spiders, crabs, and diplopods. There were two new families of
diplopod (Paracortinidae, Sinocallipodidae), five new genera of diplopod
and one new genus of crabs (Ran and Chen 1998). Just in Guizhou,
there were 85 cave animal species reported by 2007 (Internet resource,
http://gzsb.gog.com.cn/system/2007/10/27/010151816. shtml).
Chongqing
1 species, 2%
Guizhou
8 species,
15.7%
Hunan
1 species, 2%
Yunnan Guangxi
14 species, 27.5% 27 species, 52.9%
Fig. 2 The distribution pattern of the number of described hypogean fish species
in different provinces (or regions) in China.
Although so many cave species have been discovered, still there are
likely that many species remain to be discovered because: (1) the area of
karst is so vast that not all caves potentially harbouring troglobites have
been explored [or been explored] to the extent needed to discover them;
and (2) in other cases, cavefishes and other animals have been reported

but no specimens were sent to ichthyologists or other experts. There is

no doubt that more and more troglobitic species (including fishes) will be
discovered through time.
II.3 The Questionable Status of Some Cavefish Species

The status and validity of some cavefish species is questionable:
(1) Sinocyclocheilus guilinensis Ji, 1985. This species was first mentioned
in Zhou’s (1985) paper about cavefishes occurring in Guangxi
and has never been mentioned in later literature. From its simple
description it is similar to S. jii Zhang and Dai, 1992. However Zhao
et al. (2008a) compared the specimens to those of S. jii and found
that Sinocyclocheilus guilinensis is a valid species.
(2) Sinocyclocheilus yimenensis Li and Xiao, 2005. Although all the
Sinocyclocheilus species are associated with cave or other underground
habitats, this does not mean that they are all troglobites. Only
those with typical troglobite characteristics, such as degeneration
of eyes and greatly reduced pigmentation etc., are treated as such.
Although S. yimenensis was collected from an underground river
(Li et al. 2005), it has no characteristics of troglobites. Hence, we
treat it as an epigean species found accidentally in caves.
(3) Paralepidocephalus yui Tchang, 1935. This species was first reported
from Yilong Lake in Yunnan (Tchang 1935). It could be also collected
from cave environments, such as Jiuxiang Limestone Cave (Yang
and Chen 1993b), but it has no typical troglobite characteristics,
and is treated as an epigean species herein.
III. CURRENT RESEARCH STATUS ON CAVEFISHES IN

CHINA
Currently, most studies in China are focused on discovering and
describing new species. Detailed studies in depth on other subjects, such
as physiology, behavior, ecology, etc., are scarce or lacking. The limited
information on some of these subjects is herein documented:
III.1 Biology and Ecology

When kept in light in the laboratory, individuals of species such as
S. anophthalmus, may become dark due to the synthesis of melanin
or development of more pigmented melanophores (Chen et al. 2001).
Li et al. (2000) described the conditions of three collecting localities of
Sinocyclocheilus rhinocerous. One locality is a seasonal lake, about 7000 m2 in

summer and dries in winter. The second locality is a cave about 80 m2 and
15 m below ground; the water level is consistent all [around the] year. The
third locality is an outlet of a spring in summer, flowing only in the rainy
season, from May to October. S. malacopterus, Discogobio brachyphysallidos
and Rhodeus sinensis are also found in this spring outlet. Li and Tao (2005)
have bred S. rhinocerous in captivity. In the lab, these fish are omnivores,
eating steamed rice, raw noodles, paste and earthworms. The optimum
water temperature is 12-20°C; the high lethal temperature is 28°C and
when the temperature is lower than 4°C and higher than 0°C the fish stop
feeding and moving.
III.2 Morphology, Anatomy and Troglomorphic

[Characteristic] Evolution
There are 59 species in genus Sinocyclocheilus, but only 28 species are
hypogean. This genus provides an ideal basis for research on the evolution
of troglomorphic characters. By comparing epigean and hypogean
species in Sinocyclocheilus, an evolutionary trend from surface dwellers to
underground dwellers has been observed for characteristics such as size of
the eyes, pigmentation, scales, barbels, lateral line and frontal organ (Chen
et al. 1994). Eyes tend to become vestigial in subterranean species. The
barbels, which the fish use to probe for food, are longer and thicker than in
surface water dwellers and so on (Chen et al. 1994, Wang and Chen 2000,
Chen et al. 2001).
Li et al. (1997) observed the tissues of the horn-like frontal organ of
S. rhinocerous and found no nerves or glands. They concluded that this
strange derived organ had no relationship with feeling; it may act as an
organ to protect the soft head if the fish collides with the stony wall in the
dark.
Chen et al. (2001) pointed out that cavefish, as well as some troglobitic
invertebrates, are often distantly related phylogenetically, but have
evolved similar characteristics. This is a clear example of the convergent
evolution of adaptation to the dark cave environment.
III.3 Morphological and Genetic Evolution

The genus Sinocyclocheilus, which is endemic to China, occurs in karst
cave waters and surface rivers or lakes in Yunnan, Guizhou and Guangxi.
It includes at least 60 species of which 28 are hypogean. It provides a
good model system, in which closely related species specialized to the
life in subterranean habitats possess similar phenotypes. This allows
us to study evolutionary developmental mechanisms and the role of

natural selection and adaptation in cave animals. Based on osteological

and external characteristics, Shan and Yue (1994) revealed that all
Sinocyclocheilus species in their study were grouped into two monophyletic
clusters, one including all surface species and the other all cave species
except S. anophthalmus that groups with the first cluster. The distinctive
morphological divergences between these two groups were variations in
the frontal organ, number of lateral line scales and presence of radii in
the scales. Species in the group containing all but one of the cave species
shared the following character states: lower lateral line scale numbers (less
than 45 or body naked resulting from scale regression), radii present only
in the scales apical region, and an acutely upheaved frontal–parietal area.
They regarded these characters as having a phylogenetic significance.
Thus they divided the genus Sinocyclocheilus [was subdivided] into two
subgenera, Sinocyclocheilus (for the surface species plus S. anophthalmus)
and Gibbibarbus (for the remaining troglobitic ones).
Wang and Chen (1999) re-examined the distinguishing characters
between the two subgenera and found out that scale numbers within
Sinocyclocheilus varied in a continuous way and the divergence in scale
radius between the two groups was probably due to convergence.
Consequently, they did not support the view that this genus could be
subdivided into those two subgenera, Sinocyclocheilus and Gibbibarbus.
However, they also treated the states of frontal–parietal area as one of
phylogenetically informative characters. Thus, all cave species in their
study clustered as a subclade and nested within the surface species.
Xiao et al. (2005) determined nucleotide sequences of complete
mitochondrial cytochrome b gene (1140 bp) and partial ND4 gene (1032
bp) of 31 recognized in group Sinocyclocheilus species and one outgroup
species, Barbodes laticeps. Phylogenetic trees were reconstructed using
maximum parsimony, Bayesian, and maximum likelihood analyses.
Their results showed that cave species were polyphyletic and occurred
in all the five major clades, indicating that adaptation to subterranean
environments had occurred multiple times during the evolutionary history
of Sinocyclocheilus. In addition, Li et al. (2008) added five more species,
but their results did not change the number of clades within the genus.
They also estimated the divergence time of the genus Sinocyclocheilus as
the late Miocene, about 11 million years ago (Ma). In addition, except for
S. rhinocerous and S. hyalinus, all species distributed in Yunnan formed
a strongly supported monophyletic group. Their results supported that
the diversification of Sinocyclocheilus in Yunnan may correlate with the
uplifting of the Yunnan Plateau, and Yunnan might be the diversification
centre of Sinocyclocheilus. Obviously, based on the genetic results, it is not
acceptable to subdivide the genus Sinocyclocheilus into Sinocyclocheilus
and Gibbibarbus based on convergent troglobitic traits, such as reduction
of eyes and pigmentation.

IV. CONSERVATION
Apparently, all the hypogean fish species occurring in China are endemic
to small areas and have small populations. For some species, such as
Typhlobarbus nudiventris, the only access to them has been limited to
occasional collections from wells. As already mentioned, several species,
such as Typhlobarbus nudiventris, Longanalus macrochirous, Triplophysa rosa,
are known from a single specimen each; only a few species are represented
by several small populations, such as Sinocyclocheilus rhinocerous (Li et al.
2000). Proudlove (2001) concluded that the hypogean fishes are susceptible
to five main threats: habitat degradation, hydrological manipulations,
environmental pollution, overexploitation, and introduced exotic species.
Any threat could have serious consequences for cave species survival.
In 1998, four hypogean fish species (Sinocyclocheilus anophthalmus,
Typhlobarbus nudiventris, Triplophysa gejiuensis, Oreonectes anophthalmus)
were included in the “China Red Data Book of Endangered Animals,
Pisces” (Yue and Chen 1998). In 2004, this number had increased to 14
(Wang and Xie 2004) (Table 1). Hypogean fish species are receiving more
and more attention for their conservation. But, in view of the large number
of troglobitic species and their vulnerability, conservation actions are not
sufficient.
Cavefish have evolved in relatively stable cave environments for
long periods of time so when caves are altered by human activities the
hypogean fishes living there may not survive. For example, after Alugu
Cave was developed as a scenic spot, S. hyalinus almost became extinct
because of environmental deterioration (Yang and Chen 1994). With the
fast development of China’s economy and fast increase in population,
such problems may increase in the future. So, for the hypogean fish
survival, a wider range of consequences resulting from human activities—
such as irrigation, hydropower construction, tourism, and underground
water use—should be circumspectly considered. And much more positive
conservation measures—including habitat protection, population recovery
and scientific research—should be supported and implemented.
The distributional locality of hypogean fishes is usually a point-like
area and most of them are located in remote rural mountains. So far, there
is no report that introduced exotic species pose threat for their survival.
V. DISCUSSION
The unique morphology and rarity of hypogean fishes make them very
precious. Their discovery in China not only adds a new feature to the
Chinese fresh-water fish fauna, but also enriches the content of Chinese
speleology (Chen et al. 2001). Considering that 150 hypogean fish species

have been listed for the world (Proudlove, this volume), the diversity of
45 hypogean fish species in China is high compared with the total number.
Additionally, there are 3.44 million km² of karst landform area in China,
and most of these karst landform areas are poorly explored by experts on
hypogean creatures. It is undoubtedly true that the number of hypogean
fish species will increase with continued exploration and study. The
increasing number of new species discovered and described enrich Chinese
fish fauna and fish fauna of the world, and provide many opportunities for
research. On the other hand, many hypogean fish species may disappear
due to habitat disturbance (irrigation, hydropower construction, tourism
exploitation, and underground water use) if conservation measures are
not taken.
Although many hypogean fish species have been described, there
are still numerous problems with the published papers. In the case of
Sinocyclocheilus, for instance, some species descriptions were too simple
without proper comparisons with other related species. In addition, papers
were not reviewed by peers. Also, type specimens are scattered in different
research institutions or private collections and this poses obstacles for the
examination of specimens for new species description. Finally, experts in
the same field do not openly communicate.
In spite of the rich Chinese hypogean fish fauna, studies on evolution,
ecology, physiology, conservation and behaviour are just beginning or
have not started. We need not only more study but also more finanacial
support.
Acknowledgments
We thank Dr. Eleonora Trajano and Dr. Thomas Poulson for improving the
manuscript and helping with the language. This study was supported by
the National Science Foundation of China (30970326, U0936602, 30870291),
the Cooperation Project between the Chinese Academy of Sciences and
Guizhou Province (2005-2008), The National Basic Research Program
(also called 973 Program) (2003CB415105) and the Scientific Research
Foundation of Yunnan University (2008YB004).
References
Chen, X.Y. and J.X. Yang. 2005. Triplophysa rosa sp. nov.: A new blind loach from
China. Journal of Fish Biology 66: 599-608.
Chen, X.Y., G.H. Cui and J.X. Yang. 2004. A new cave-dwelling fish species of genus
Triplophysa (Balitoridae) from Guangxi, China. Zoological Research 25: 227-231.
(In Chinese with English abstract.)

Chen, Y.R. 1990. Research status on Chinese cavefishes. Biology Information 2(3):
117-119. (In Chinese)
Chen, Y.R. and J.X. Yang. 1993. A synopsis of cavefishes from China. Proceedings of
XI International Congress of Speleology: 121-122.
Chen, Y.R. and J.X. Yang. 1994. Exploitation and conservation of cave fish in tourist
resources. In: Proceedings of International Symposium on the Exploitation and
Protection of Karst and Cave Scenic Tourist Resources, L.H. Song (Ed.). Earthquake
Press, Beijing, pp. 149-153.
Chen, Y.R., J.X. Yang and G.C. Xu. 1992. A new blind loach of Triplophysa from
Yunnan stone forest with comments on its phylogenetic relationship. Zoological
Research 13(1): 17-23. (In Chinese with English abstract.)
Chen, Y.R., J.X. Yang and Z.G. Zhu. 1994. A new fish of the genus Sinocyclocheilus
from Yunnan with comments of its characters adaptation. Acta Zootaxonomica
Sinica 19(2): 246-253. (In Chinese with English abstract.)
Chen, Y.R., J.X. Yang and J.H. Lan. 1997. One new species of blind cavefish from
Guangxi with comments on its phylogenetic status (Cypriniformes: Cyprinidae:
Barbinae). Acta Zootaxonomica Sinica 22(2): 219-223. (In Chinese with English
abstract.)
Chen, Y.R., J.X. Yang and Z.M. Chen. 2001a. Species and convergent evolution of
cave-dwelling Sinocyclocheilus fishes. Memoires de Biospeologie (International
Journal of Subterranean Biology), Tome XXVIII : 185-189.
Chen, Y.R., Z.M. Chen and J.X. Yang. 2001b. Cave fishes of china. Memoires de
Biospeologie (International Journal of Subterranean Biology), Tome XXVIII :
181-184.
Chen, Y.R., X.L. Chu, Z.Y. Luo and J.Y. Wu. 1988. A new blind cyprinid fish from
Yunnan, China with a reference to the evolution of its characters. Acta Zoologica
Sinica 34(1): 64-70. (In Chinese with English abstract.)
Chen, Y.R., J.X. Yang, B. Sket and G. Aljancic. 1998. A new cave loach of Paracobitis
with comments of its characters evolution. Zoological Research 19(1): 59-63. (In
Chinese with English abstract.)
Chen, Z.M., L.X. Li and J.X. Yang. 2010. Description of a new cave-dwelling species
of the genus Triplophysa (Balitoridae: Nemacheilinae) from Guangxi, China.
(Unpublished.)
Chu, X.L. and Y.R. Chen. 1978. The discovery of the first blind fish from China.
Journal of Chinese Nature 1(6): 343. (In Chinese)
Chu, X.L. and Y.R. Chen. 1979. A new blind cobitid fish (Pisces, Cypriniformes)
from subterranean waters in Yunnan, China. Acta Zoologica Sinica 25(3): 285-
287. (In Chinese with English abstract.)
Chu, X.L. and Y.R. Chen. 1982. A new genus and species of blind cyprinid fish from
China with special reference to its relationships. Acta Zoologica Sinica 28(4): 383-
388. (In Chinese with English abstract.)
Du, L.N., X.Y. Chen and J.X. Yang. 2008. A review of the Nemacheilinae genus
Oreonectes Günther with descriptions of two new species (Teleostei: Balitoridae).
Zootaxa 1729: 23-36.
Greenwood, P.H. 1967. Blind cave fishes. Studies in Speleology 1(5): 262-275.

Lan, J.H., Y.H. Zhao and C.G. Zhang. 2004. A new species of the genus Sinocyclocheilus
from China (Cypriniformes, Cyprinidae, Barbinae). Acta Zootaxonomica Sinica
29: 377-380. (In Chinese with English abstract.)
Li, W.X. and J.H. Lan. 1992. A new genus and three new species of Cyprinidae
from Guangxi, China. Journal of Zhanjiang Fisheries College 12(2): 46-51. (In
Li, W.X. and J.N. Tao. 1994. A new species of Cyprinidae from Yunnan—
Sinocyclocheilus rhinocerous. Journal of Zhanjiang Fisheries College 14(1): 1-3. (In
Li, W.X. and Z.G. Zhu. 2000. A new species of Triplophysa from cave Yunnan. Journal
of Yunnan University 22: 396-398. (In Chinese with English abstract.)
Li, W.X. and J.N. Tao. 2002. Local dissection of body of the fish Sinocyclocheilus
rhinocerous. Journal of Yunnan Agricultural University 17(3): 207-209, 219. (In
Li, W.X. and J.N. Tao. 2005. Preliminary research on the cultivation of Sinocyclocheilus
rhinocerous. Yunnan Agricultural Science and Technology 1: 46. (In Chinese)
Li, W.X., D.F. Wu and A.L. Chen. 1998. Two new species of Sinocyclocheilus from
Yunnan (Cypriniformes, Cyprinidae). Journal of Zhanjiang Ocean University
18(4): 1-5. (In Chinese with English abstract.)
Li, Y.B., J.J. Hou and D.T. Xie. 2002. The recent development of research on karst
ecology in Southwest China. Scientia Geographica Sinica 22(3): 365-370. (In
Li, W.X., J.H. Lan and S.Y. Chen. 2003a. A new species of cave Sinocyclocheilus
from Guangxi—Sinocyclocheilus jiuxuensis. Journal of Guangxi Normal University
Li, W.X., J.C. Ran and H.M. Chen. 2003b. A new species of cave Sinocyclocheilus
in Guizhou and its adaptation comment. Journal of Jishou University (Natural
Science Edition) 24(4): 61-63. (In Chinese with English abstract.)
Li, W.X., H. Xiao, R.G. Zan, Z.Y. Luo, C.H. Ban and J.B. Ben. 2003c. A new species
of Sinocyclocheilus from caves in Guangxi. Journal of Guangxi Normal University
Li, W.X., H.M. Chen and J.C. Ran. 2004. A new species of Sinocyclocheilus from
the cave in Guizhou, China. Journal of Zhanjiang Ocean University 24(3): 1-3. (In
Li, W.X., D.F. Wu, A.L. Chen and J.N. Tao. 2000. The preliminary investigation of
geographical distribution and ecological adaptation for cave environments of
Sinocyclocheilus rhinocerous. Journal of Yunnan Agricultural University 15(1): 1-4.
(In Chinese with English abstract.)
Li, W.X., H.F. Yang, H. Chen, C.P. Tao, S.Q. Qi and F. Han. 2008. A new blind
underground species of the genus Triplophysa (Balitoridae) from Yunnan, China.
Zoological Research 29(6): 674-678.
Li, Z.Q., B.C. Guo , J.B. Li, S.P. He and Y.Y. Chen. 2008. Bayesian mixed models and
divergence time estimation of Chinese cavefishes (Cyprinidae: Sinocyclocheilus).
Chinese Science Bulletin 53(13): 1560-1569.

Mao, W.N., Z.M. Lu, W.X. Li, H.B. Ma and G. Huang. 2003. A new species of
Sinocyclocheilus (Cyprinidae) from cave of Yunnan, China. Journal of Zhanjiang
Ocean University 23(3): 1-3. (In Chinese with English abstract.)
Nelson, J.S. 1994. Fishes of the World. John Wiley & Sons, New York.
Ran, J.C. and H.M. Chen. 1998. A survey of speleobiological studies in China.
Carsologica Sinica 17(2): 151-159.
Shan, X.H. and P.Q. Yue. 1994. The study on phylogeny of the Sinocyclocheilus fishes.
Zoological Research 15(Supplement): 36-44. (In Chinese with English abstract.)
Tchang, T.L. 1935. A new genus of loach from Yunnan. Bulletin of the Fan Memorial
Institute of Biology 7: 17-19.
Wang, D.Z. and J.W. Liao. 1997. A new species of Sinocyclocheilus from Guizhou,
China (Cypriniformes: Cyprinidae: Barbinae). Acta Academiae Medicinae Zunyi
20(2, 3): 1-3. (In Chinese with English abstract.)
Wang, D.Z. and Y.Y. Chen. 2000. The origin and adaptive evolution of the genus
Sinocyclocheilus. Acta Hydrobiologica Sinica 24: 630-634. (In Chinese with English
abstract.)
Wang, D.Z. and D.J. Li. 2001. Two new species of the genus Triplophysa from
Guizhou, China (Cypriniformes: Cobitidae). Acta Zootaxonomica Sinica 26:
98-101. (In Chinese with English abstract.)
Wang, S. and Y. Xie. 2004. China Species Red List, Vol. I. Higher Education Press,
Beijing. (In Chinese)
Wang, D.Z., Y.Y. Chen and X.Y. Li. 1999. An analysis on the phylogeny of the genus
Sinocyclocheilus. Acta Academiae Medicinae Zunyi 22(1): 1-6. (In Chinese with
English abstract.)
Xiao, H, S.Y. Chen, Z.M. Liu, R.D. Zhang, W.X. Li, R.G. Zan, Y.P. Zhang. 2005.
Molecular phylogeny of Sinocyclocheilus (Cypriniformes: Cyprinidae) inferred
from mitochondrial DNA sequences. Molecular Phylogenetics and Evolution 36:
67-77.
Xie, Y.J. 1541. Travel Notes on Alu Cave (on a stone epigraph located at the entrance
of Alugu Dong and dated in Ming Dynasty). (In Chinese)
Yang, J.X. and Y.R. Chen. 1993a. The cavefishes from Doan, Guangxi, China with
comments on their adaptation to cave habitats. Proceedings of XI International
Congress of Speleology: 124-126.
Yang, J.X. and Y.R. Chen. 1993b. The cavefishes from Jiuxiang limestone cave of
Yunnan, China with reference to the character evolutions. Proceedings of XI
International Congress of Speleology: 126-128.
Yang, J.X. and Y.R. Chen. 1994. Characters, exploitation and conservation of the
fish resources in karst areas of Yunnan. In: Proceedings of International Symposium

on the Exploitation and Protection of Karst and Cave Scenic Tourist Resources, L.H.
Song (Ed.). Earthquake Press, Beijing, pp. 145-148.
Yue, P.Q. and Y.Y. Chen. 1998. China Red Data Book of Endangered Animals, Pisces.
Science Press, Beijing.
Zhang, C.G. and D.Y. Dai. 1992. A new species of Sinocyclocheilus from Guangxi,
China. Acta Zootaxonomica Sinica 17(3): 377-379. (In Chinese with English
abstract.)
Zhang, C.G. and Y.H. Zhao. 2001. A new fish of Sinocyclocheilus from Guangxi,
China with a comment on its some biological adaptation (Cypriniformes:
Cyprinidae). Acta Zootaxonomica Sinica 26: 102-108. (In Chinese with English
abstract.)
Zhang, Z.L., Y.H. Zhao and C.G. Zhang. 2006. A new blind loach, Oreonectes
translucens (Teleostei: Cypriniformes: Nemacheilinae), from Guangxi, China.
Zoological Studies 45(4): 611-615.
Zhao, Y.H. and C.G. Zhang. 2006a. Cavefishes, concept, diversity and research
progress. Biodiversity Sciences 14(5): 451-460. (In Chinese with English
abstract.)
Zhao, Y.H., K. Watanabe and C.G. Zhang. 2006b. Sinocyclocheilus donglanensis, a
new cavefish (Teleostei: Cypriniformes) from Guangxi, China. Ichthyological
Research 53: 121-128.
Zhao, Y.H., C.G. Zhang and J. Zhou. 2008a. Sinocyclocheilus guilinensis, a new
species from an endemic cavefish group (Cypriniformes: Cyprinidae) in China.
Environmental Biology of Fishes DOI 10.1007/s10641-008-9344-8.
Zhao, Y.H., J.H. Lan and C.G. Zhang. 2008b. A new cavefish species, Sinocyclocheilus
brevibarbatus (Teleostei: Cypriniformes: Cyprinidae), from Guangxi, China.
Environmental Biology of Fishes DOI 10.1007/s10641-008-9338-6.
Zhou, J. 1985. Fishes of Karst cave in Guangxi. Carsologica Sinica 4: 377-385. (In
Chinese)
Zhou, J., C.G. Zhang and A.Y. He. 2004. A new species of the genus Sinocyclocheilus
from Guangxi, China (Cypriniformes, Cyprinidae). Acta Zootaxonomica Sinica
29: 591-594. (In Chinese with English abstract.)
Zhou, S.B. and G.L. Li. 1998a. A new species of the genus Sinocyclocheilus from cave
of Guangxi, China. In: Contributions from Tianjin Natural History Museum.
Ocean Press, Beijing, pp. 9-12.
Zhou, S.B. and G.L. Li. 1998b. A new species of Sinocyclocheilus from Guangxi
(Cypriniformes: Cyprinidae: Barbinae). Guangxi Sciences 5: 139-141, 149. (In
Zhu, S.Q. and W.X. Cao. 1987. The noemacheiline fishes from Guangdong and
Guangxi with descriptions of a new genus and three new species (Cypriniformes:
Cobitidae). Acta Zootaxonomica Sinica 12: 323-331.
Zhu, Y., Lü, Y.J., Yang, J.X. and S. Zhang. 2008. A new blind underground species
of the genus Protocobitis (Cobitidae) from Guangxi, China. Zoological Research
29(4): 452-454.

CHAPTER 12
Subterranean Fishes of India
Atanu Kumar Pati and Arti Parganiha
School of Life Sciences, Pt. Ravishankar Shukla University
Raipur 492 010, India
E-mail: akpati19@gmail.com
INTRODUCTION
The term “subterranean” is an adjective meaning something/somebody
existing or occurring under the earth’s surface. It is a larger domain that
includes caves, caverns, lava tubes and phreatic spaces. Habitats, such as
deep burrows, ant hill and other similar underground biogenic structures,
of several burrowing and cryptobiotic animals, could be included in the
subterranean category. Manmade underground tunnels, namely concrete
pipelines carrying domestic sewage, could also be termed as a type of
subterranean niche. In general, subterranean spaces are poorly explored
as regards mapping of the biodiversity. The main limitations are: (a) lack
of easy access to these underground spaces, (b) prevalence of darkness
with high humidity, and (c) the fear of unknown and uncertain things
and/or events. In India, so far no concerted efforts have been made to
survey the already known subterranean habitats. Further a dedicated
research institute/center in the realm of biospeleology (= science of cave/
hypogean organisms) is conspicuously absent.
SEARCH FOR A CAVE (KARST) MAP OF INDIA

We have a hunch that in India more than thousands of caves may be
present. Unfortunately for this subcontinent neither a karst nor a cave
map is available. Important agencies, such as Geological Survey of India
(GSI), National Thematic and Mapping Organization of India (NATMO),

Survey of India and Remote Sensing Organization of India, were contacted

through email. Also the websites of these organizations were visited. The
result of these efforts was in the negative. The websites of different State
Governments of India were also browsed with an emphasis to look into the
forest and tourism related data bases, but nothing was found. However,
wild life researchers could not be contacted through questionnaire. The later
is a daunting task. Nonetheless many caves in India have been explored
by amateur cave enthusiasts, biologists, geologists and personnel from the
forest departments and NGOs. The only available source of information
about the Indian caves is the South Asia Cave Registry compiled by H.D.
Gebauer, Germany.
WHY FISHES BELONGING TO CERTAIN FAMILY ARE

FOUND IN CAVES?
The organisms belonging to the phylum Arthropod constitute the bulk
of the subterranean biodiversity. Among the sub-phylum vertebrata the
members of the class Pisces are the most prominent invaders of subterranean
niche. The Order Cypriniformes stand out distinctly having the maximum
number of reported subterranean species, belonging to the family, namely
Cyprinidae, Balitoridae, and Cobitidae (Proudlove 2006). The species
belonging to the former two families have been reported abundantly
from the subterranean habitats worldwide (Figure 1). The question: What
make them to survive in nutrition-deficient and perpetually dark hypogean
habitats has not yet been answered adequately.
SUBTERRANEAN FISH FAUNA OF INDIA

Although India is the seventh largest country in the world by geographical
area, only seven distinctly subterranean fish species have been described
so far (Figure 2; Table 1). Of those, three are cave dwellers and the
remaining four inhabit shallow phreatic spaces, notably in the network of
ground water channels connected to wells in Kerala. Most of the reports
on these species deal with their distribution, systematics and phylogenetic
relationship. To some extent biology of only two species has been studied.
Pati and his colleagues (see Pati 2008) have studied ecophysiological and
ethological aspects in Nemacheilus evezardi in the last 20 years and have
published more than 20 original research papers/reviews. Mercy and
her colleagues have studied on various morphological and physiological
aspects in Horaglanis krishnai and came out with about 9 original research
papers (see Proudlove 2006). Biological studies on other five subterranean
species have never been carried out. Therefore, this chapter includes

Atanu Kumar Pati and Arti Parganiha 417
60 30
20
Frequency
40 10
0
Cyprinidae Balitoridae Cobitidae
20
es
es
es
es
es
es
es
es
es
es
rm
m
rm
rm
m
ym orm
rm
rm
m
or
or
or
or
ifo
fo
ifo
ifo
f
iif
cif
rif
tif
sif
i
ni
nt
en
rc
d
no
ra
lu
nc
ri
op
do
i
Pe
ph
pa
yp
Si
ha
ra
rc
no
or
O
C
nb
C
Pe
G
ri
Sc
yp
Sy
C
Fig. 1 Subterranean biodiversity among different orders of the class Pisces. Inset
represents the maximum number of reported subterranean species, belonging to the
family, namely Cyprinidae, Balitoridae, and Cobitidae of the Order Cypriniformes.
Based on data given in Proudlove (2006).
mainly a brief review of the work on biology of N. evezardi and H. krishnai.

It also includes brief notes on other five species without going into details
regarding their systematics, taxonomy, and conservation status as these
aspects have been taken care of firstly by Romero and Paulson (2001) and
later by Proudlove (2006) in the recent past.
NEMACHEILUS EVEZARDI (DAY, 1872)
Historical Account
There is complete absence of any scientific report on the exact year of
discovery of the loach, Nemacheilus evezardi inside the Kotumsar Cave
located on the jurisdiction of the Kanger Valley National Park, Bastar,
India. There was also no mention about the Kotumsar Cave itself in the
Chhattisgarh Feudatory States Gazetteer, 1909 (Tiwari 2000). Possibly
it is the geographer Prof. Shankar Tiwari, who spotted the species in
1959 while mapping the cave for the first time between 1958 and 1959
(Tiwari 2000). Subsequently in 1959, Prof. S.M. Agarwal (misspelt as A.M.
Agarwal in Proudlove 2006) wrote to G. Thinès about the presence of this
species in the Kotumsar Cave (Agarwal – personal communication) and
thereafter Thinès (1969) reported about its existence in this cave (Thinès

Fig. 2 Distribution of Indian subterranean fishes. Numbers indicate the location

of the species found in India. 1: Nemacheilus evezardi; 2: Schistura sijuensis; 3:
Schistura papulifera; 4: Horaglanis krishnai; 5: Horaglanis alikunhii; 6: Monopterus
eapeni; 7: Monopterus roseni.
book.
1969). Agarwal also sent specimens of this species to the then Director of
the Bombay Natural History Society, E.G. Silas, who identified the species
as Nemacheilus evezardi Day, 1872 (Agarwal Telephonic Communication in
April 2009).

Table 1 List of subterranean fish fauna of India

Name of Synonyms Distribution Latitude; Vulnerability Red list
species Longitude status
Nemacheilus Indoreonectes Kotumsar 18°52’09”N; Low Data
evezardi evezardi, Cave, Bastar 81°56’05”E vulnerability deficient
Oreonectes district,
evezardi Chhattisgarh
Schistura Noemacheilus Siju Cave, 25°21’02”N; Low Vulnerable
sijuensis sijuensis, Garo Hills, 90°41’04”E vulnerability
Mesonoemachilus Meghalaya
sijuensis,
Nemacheilus
multifasciatus
Schistura None Krem 25°12’48”N; Low Not
papulifera Synrang 92°21’48”E vulnerability evaluated
Pamiang
Cave, Jaintia
Hills,
Meghalaya
Horaglanis None Kottayam Between Low Vulnerable
krishnai district, 8°04’N, vulnerability
Kerala 76°13’E and
10°21’N,
77°38’E
Horaglanis None Parappukara, 10°23’N; Low Not
alikunhii Trichur 76°15’E vulnerability evaluated
district,
Kerala
Monopterus Monopterus Kottayam 9°30’N; Low Data
eapeni indicus, district, 76°33’E vulnerability deficient
Amphipnous Kerala
indicus
Monopterus None Periyam 10°38’N; Low to Not
roseni village, 76°22’E moderate evaluated
Kerala vulnerability
Controversy on the Nomenclature and Systematics/

Taxonomic Position of Nemacheilus evezardi Day
N. evezardi is a hill stream loach and found in both hypogean (cave) and
epigean (river) forms. There have been discrepancies about its nomenclature
and systematics/taxonomic position. Kottelat (1990), without giving
any supportive reasons, placed it in the genera Indoreonectes (Romero
and Paulson 2001) and it is still described as Indoreonectes evezardi by
Proudlove (2001, 2006). Singh and Yazdani (1993) named it as Oreonectes
evezardi; however, it was valid for a short time only. Pati and colleagues

also described it as Oreonectes in some of their earlier publications (Biswas

et al. 1990b, Biswas and Pati 1991) obviously following Singh and Yazdani
(1993). Likewise, contradictions were also made for the systematic
position of Nemacheilus evezardi. Day (1958) included it within the family
Cyprinidae. Biswas (1990) placed it in the family Cobitidae, whereas,
Eschmeyer (1998) positioned it within the family Balitoridae. Singh
(personal communication) kept it within the family Homalopteridae.
According to Pati (2008) it belongs to the family Balitoridae. The generic
name of the species N. evezardi was never spelt as Noemacheilus by Day
(1958) as has been mentioned by Proudlove (2006). The genus has been
subdivided into two groups strictly based on number of barbels and
Nemacheilus evezardi is the solitary species with eight (8) barbels. Further,
surprisingly in Day (1958) there is no mention of the genus Indoreonectes as
has been reported by Proudlove (2006).
Hypogean versus Epigean Controversy

The both hypogean and epigean forms found abundantly in caves of
the Kanger Valley National Park and adjacent river (Kanger Dhara),
respectively, continued to be Nemacheilus evezardi. Although the
exact age of the evolution of the cave population is not known, many
distinct differences relating to physiology and behavior do exit when
both populations are compared. Prominently, the reduction of eyes,
pigmentation and acquisition of air gulping behavior in the hypogean
population offer adequate justification for its status as a separate species.
Habitat Ecology
The population of hill stream loach, N. evezardi has successfully colonized
the Kotumsar Cave on the Kanger Valley National Park. It is found in small
water pools and perennial streams of number of other limestone caves
located on this Park. It has established itself as a distinct cave population
and differs from its epigean population in many ways (Biswas 1990, Pati
2008). The epigean population inhabits rapidly flowing water in hill
streams, rivulets and river, “Kanger Dhara,” flowing inside the National
Park (Biswas 1990,1992). The hypogean N. evezardi is a bottom dweller, but
prefers to stay in shallow waters.
The Kotumsar Cave – the natural habitat of the cave population, is
situated in the left flank of the Kanger River in the Kanger Valley National
Park (18° 52’ 09” N; 81° 56’ 05” E), at an altitude of 560 m, Jagdalpur,
Bastar, India. It is one of the best-studied caves in India and was
discovered in 1958. Inside the cave one can notice numerous stalactite and

stalagmite formations. The air and water temperatures of the cave remain
relatively constant at an annual average of 28.25±1.23°C and 26.33±0.96°C,
respectively (Biswas 1990).
Distribution
N. evezardi has an ample distribution in the Godavari and Kolab watersheds,
of which the Kanger River is a part. It has also been reported from Pune,
India (Day 1958) and may inhabit rivers in the Deccan Plateau.
Morphological Features
N. evezardi has a cylindrical body with dorsally located small-sized eyes. It
has eight barbels: one pair is nasal, two rostral and one maxillary. This is
the distinguishing character of the species. The lateral line is indistinct. The
hypogean form is short (3-4 cm) in size as compared with its epigean one
(5-7 cm) (Figure 3). There is also a remarkable difference in the pigment
distribution pattern. The cave form is either totally (albinic) or sparsely
pigmented. Thus, it lacks the vertically directed small dark blotches of
the epigean form. A dark spot at the base of the caudal fin is present in
both forms. The hypogean form has reduced eyes as compared with its
Fig. 3 Albinic subterranean loach, Nemacheilus evezardi.

book.

epigean counterpart (Biswas 1990). The spines are absent on the head in
both forms.
Population Density and Reproduction

N. evezardi does not breed in captive conditions of the laboratory.
We failed number of times in our attempt to have effective induced
breeding. Therefore, we left with the single alternative option to study
its reproductive rhythm in its natural habitat. However, it was extremely
difficult to remain inside the cave continuously for longer period of time
during and after the monsoon. Further, lack of sexual dimorphism in the
species makes the matter more complicated. Biswas et al. (1990c) reported
occurrence of fingerlings of hypogean fishes in July-August-September. He
studied seasonal frequency of adults (pigmented or albino) and juveniles
(pigmented or albino) in the water pools located in the first floor of the
Kotumsar Cave during four different seasons. He performed the study
over two consecutive years to validate the reproducibility of spawning
rhythm. He revealed that the total number of hypogean adults increased
(approximately doubled) every year soon after the end of monsoon as
compared to other seasons of the study. However, juveniles were noticed
only during and after the monsoon. Further, it was noticed that at any
given time of the year the population of adult albinos does not exceed 10%
of the entire hypogean population restricted to the water pools in the first
floor of the cave.
The period of July-August-September receives maximum rainfall in
the Kanger Valley National Park. Therefore, the spawning in cavefish
could be associated with the flooding of the water courses of the caves.
The flooding signal might be working as a trigger for the initiation of
breeding. Until we observe a free-running rhythm in its reproductive
behavior it would be difficult to talk about an underlying circannual clock
mechanism. However, in contrast, Biswas (1990) observed an increase in
the population of epigean fingerlings in the river in April. It seems that
the trigger for the initiation breeding may be different for the epigean
population.
Adaptation to Food-deficient Conditions

It is well known that organisms inhabiting caves and caverns develop
abilities to withstand prolonged periods of starvation. They exhibit
reduced respiratory intensity and decreased activity pattern under food-
limited environment. Does hypogean Nemacheilus evezardi also exhibit the
similar phenomenon? Biswas (1991) performed studies under starvation or

restricted feeding schedules in both hypogean and epigean fish to answer

this question. Fishes were exposed to these conditions for 30 days and
body mass loss in both forms, irrespective of feeding status, was recorded.
He demonstrated that reduction in body weight was always greater in
epigean fish following starvation or restricted feeding as compared to
their hypogean counterparts. On the other hand, hypogean fish gained
more weight than epigean fish when food was provided ad libitum. He
further observed that when food was supplied at the rate of 1 mg feed g–1
body weight, the hypogean N. evezardi lost less weight as compared to its
epigean counterpart. However, the epigean fish perceived the same food
regime (1 mg feed g–1 body weight) as good as starvation (Figure 4). This
clearly suggests that in a food-limited environment the hypogean loach
has abilities to optimize energy expenditure.
Energy Metabolism
The hypogean species of several systematic groups exhibit lower metabolic
rates than their epigean relatives (Dickson and Franz 1980, Culver 1982,
Hüppop 1985, 1986, 1989, Culver et al. 1995, Poulson 2001a,b). Perhaps
Starvation 1 mg/g body wt. 3 mg/g body wt. ad libitum

8 a
4
b
% change in body mass
0
–4
–8 a
a
– 12 b
– 16
– 20
– 24
Fig. 4 Percent change in body mass of hypogean (cave) and epigean (river) fish
following starvation or restricted feeding. Restricted feeding schedules consisted
of provision of 1 mg/g and 3 mg/g body weight food daily during the course of the
experiment. The percent change in body mass was computed relative to the initial
body weight of an individual fish. a, differs from the mean value of the starved
hypogean fish; b, differs from the mean value of the starved epigean fish (based
on Biswas, 1991).

lower metabolic rate is one of the important adaptive features among

the cavernicoles. This aspect was examined in the hypogean loach N.
evezardi (Biswas 1991, Biswas and Pati 1991). The metabolic efficiency
of both hypogean and epigean population of N. evezardi was studied
keeping individuals of both populations under either restricted feeding
regimens or starved conditions. Effects of T4 (thyroxine) and T3 (3,5,3’
triiodothyronine) on the rate of muscle tissue (that constitutes 70% of
total body mass) respiration were also investigated in both populations.
Results indicate that the rate of oxygen consumption by muscle tissue
of the hypogean N. evezardi is remarkably lower (about 2.7 times) than
the epigean fish (Biswas 1991, Biswas and Pati 1991). Further, period
of one week starvation was found to have significant effect on muscle
tissue respiration in epigean population as compared to the respiratory
rate in fully-fed condition. Moreover, T4 or T3 in vivo did not induce any
significant change in the rate of oxygen consumption by muscle tissue of
hypogean population. In contrary, in epigean population, both T4 and T3
significantly elevated the rate of oxygen consumption by the muscle tissue
(Biswas and Pati 1991). Could it be that the muscle tissue of the hypogean
fish lost its sensitivity to thyroid hormones during adaptation to the
subterranean environment? The second possible explanation is that there
may be some other hormone(s) and/or substance(s) that may oppose the
action of thyroid hormones. This speculation looks more appropriate as it
is supported by the results obtained from in vitro study in that the rate of
oxygen consumption by the muscle tissue of both hypogean and epigean
fish increased significantly, irrespective of dose level employed. The
hypogean population of N. evezardi might have evolved some specialized
endocrine mechanism(s) to ensure that the rate of oxidative metabolism
does not elevate remarkably even if levels of thyroid hormones shoot up
in the circulation abruptly.
Cryptobiotic Behavior
Burying behavior in epigean forms is a common phenomenon that protects
them from predators, strong light, and extreme temperatures (Rita Kumari
et al. 1979). Biswas et al. (1990b) examined this behavior in both epigean
(under natural day length, NDL) and hypogean (under constant darkness,
DD) populations of N. evezardi as function of time of the day, type of
substratum, water current and body size. In general they concluded that:
(i) frequency of burying varies as function of time of the day in both forms,
(ii) the frequency of burying is considerably less in the hypogean fish as
compared to its epigean counterpart, (iii) fish of both populations prefer
to bury themselves under the stones; however, hypogean forms also bury
themselves in the mud (Figure 5), (iv) the pattern of burying behavior

b
100
Percent buried
80
a
60
40
20 c
0
Cavefish River fish
Stone Mud
Fig. 5 Burying behavior in relation to substrata choice. Histograms represent

mean percent buried in each substrata ± 1 SE. Histograms bearing dissimilar
superscripts differ statistically significantly at p < 0.05 level (based on Biswas et al.
1990b).
with reference to water current is different in hypogean loach than that of

epigean ones, and (v) burying behavior has direct relationship with body
size in the epigean loach. Exhibition of burying behavior in hypogean
N. evezardi may also reflect its resting phase. Furthermore, behavior
of burying in mud for the hypogean loach may be a reflection of food-
searching behavior. It is known to feed upon soil nematodes present in the
water pools inside the cave.
In addition, a statistically significant circadian variation in the rate of
burying behavior in both populations of N. evezardi has been demonstrated
(Biswas et al. 1990b). The peak burying time was found to be early hours of
the day in both (Figure 6). Could it be that the predator pressure is more
intense at this time of the day?
Surfacing Activity/Behavior
The phenomenon of surfacing activity has been reported in hypogean fish
N. evezardi (Biswas et al. 1990a). This activity in fish is one of the examples
of adaptation to its environment for survival. The term ‘surfacing
activity’ is synonymous with other terms, namely air-breathing activity,
air-breathing behavior, air-gulping behavior, surfacing behavior that are
frequently used in the literature (Maheshwari 1998). It is characterized
sequentially by a fast upward movement, a quick air gulping at the water-
air interface and an equally swift descent.

Clock hour
00:00 06:00 12:00 18:00 00:00 06:00
a
Burying activity
Surfacing activitya
Phototactic activityb
Locomotor activityc
Muscle glycogend
Muscle lactated
Liver lactated
Alkaline phosphatase (gill)d
Acetylcholinesterase (gill)d
Acetylcholinesterase (brain)d
Punctate staged
Punctate stagea
Punctostellate staged
Punctostellate stagea
Stellate staged
Stellate stagea
Reticulostellate staged
Reticulostellate stagea
Reticulate staged
Reticulate stagea
Dark Light Dark
Dark Light Dark
Fig. 6 Peak map for circadian rhythms in several behavioral and physiological
variables in the cave loach, N. evezardi acclimated to LD 12:12 or DD. Solid triangle
represents estimated acrophase (peak) of a given variable. The horizontal line on
either side of the solid triangle defines the 95% confidence limits of acrophase.
Absence of horizontal line indicates non-detection of statistically validated
circadian rhythm by Cosinor. aDD; bchoice chamber consisting of both photic (LL)
and aphotic (DD) zones; cLD 12:12 (Lights-on at 06:00); dLD 12:12 (Lights-on at
08:00) (based on Pradhan 1984, Pradhan et al. 1989, Biswas 1990, Biswas et al.
1990a, b, Pradhan and Biswas 1994b, Pati, 2001b).
Biswas (1990) reported that, the epigean N. evezardi, unlike its hypogean
counterpart, never exhibits air-gulping behavior either in its natural habitat
or under laboratory conditions. It seems that the hypogean loaches have
undergone behavioral divergence. Probably this behavior is of immense
adaptive significance for the species. This unique behavior probably
enables the hypogean loach to survive in stagnant water pools of the cave

that becomes sporadically hypoxic over an annual time scale. The habit
of gulping atmospheric air has also been reported in many other epigean
tropical fish species those inhabit shallow ponds and streams that dry
up periodically or in stagnant water pools that become oxygen deficient
by decaying vegetation during summer. Thus, air-gulping behavior is
indispensable for survival of the species and helps it to overcome the
problem of poor availability of oxygen in the stagnant water pools.
For the first time Biswas et al. (1990a) documented the multifrequency
rhythms in air-gulping behavior in hypogean N. evezardi both under
laboratory and natural cave conditions (Figure 6). They recorded circadian
variation in air-gulping activity once in every month for 14 consecutive
months under carefully simulated cave conditions in the laboratory in
respect of darkness, temperature and humidity. A statistically significant
circadian rhythm in air-gulping activity was validated during nine out
of 14 months of observations. The peak activity was recorded between
14:00 and 18:00 (Biswas et al. 1988). Furthermore, a 6-monthly rhythm
in circadian Mesor was detected, whereas circannual rhythms were
witnessed for circadian amplitude and acrophase. Two prominent peaks
for circadian Mesor were observed, one in September and the other in
March. They further documented that probably air-gulping behavior
exhibited free-running rhythm. However, the method employed to record
air-gulping activity was subjective and the duration of study was short,
i.e., two consecutive cycles (48-h span) every month. Therefore, at this
juncture it is difficult to conclude that air-gulping behavior in hypogean
N. evezardi is endogenous in nature.
Biswas (1990) made an attempt to study air-gulping behavior in natural
conditions of the cave only over a single 24-h transverse scale in the month
of April. In this condition fish exhibited a statistically significant circadian
rhythm in air-gulping activity. The circadian Mesor was remarkably more
than that obtained under the laboratory conditions in April. In another
study effect of photoperiods (DD, LL, LD 12:12 and DL 12:12) on the air-
gulping behavior was also investigated (Biswas et al. 1990a, Pradhan and
Biswas 1994a). They demonstrated that this behavior was significantly
influenced by different photoperiods.
Phototactic Behavior
Phototaxis refers to the movement of an organism toward (positive
phototaxis) or away (negative phototaxis) from the source of light. The study
of phototactic behavior in cavernicolous fish is of great interest, especially
because they undergo depigmentation and have either regressed or
vestigial eyes. This behavior has been studied extensively in the hypogean

N. evezardi (Pradhan et al. 1989, Biswas 1993, Pradhan and Biswas 1994a).
Biswas (1993) documented that the propensity of photophilic nature is
strong in troglophilic (sparsely pigmented) N. evezardi than its troglobitic
(depigmented) counterparts.
A statistically significant 24-h rhythm in phototactic behavior has been
documented in N. evezardi (Figure 6) (Pradhan et al. 1989, Biswas 1993,
Parganiha et al. 2006). Pradhan et al. (1989) documented a high amplitude
circadian rhythm in phototactic behavior in hypogean population of N.
evezardi. It is suggested that locomotor activity may reflect phototactic
behavior of organisms as phototaxis is associated with some degree of
locomotion. However, a difference between peaks of phototactic behavior
(Pradhan et al. 1989) and locomotor activity rhythm (Biswas 1990)
was documented. Therefore, phototactic behavior is established as an
independent behavior. Furthermore, the population of hypogean loach
fed strictly between 07:00 and 11:00 exhibited a peak in phototactic rhythm
at 17:00, whereas peak occurred at 02:42 in those fed between 19:00 and
23:00. This clearly suggests that periodic restricted feeding schedule may
modulate the rhythm in phototactic behavior. Pati and Agrawal (2002) also
witnessed a significant circadian rhythm in phototactic behavior. They
examined phototactic rhythm in the population of a cave loach for eight
consecutive days and documented that significant circadian rhythm in
phototactic behavior persisted for about five days only. They demonstrated
a free-running pattern in the occurrence of peaks. The disappearance of
rhythm in phototaxis after 5th day could be explained as a consequence of
starvation and it could be a strategy to conserve energy when food is not
available. In another experiment they reported photophobia in the same
population of cave loach under starved condition (Agrawal and Pati 2002).
The availability of food seems to be an important modifier of phototactic
behavior in hypogean N. evezardi (Pati and Agrawal 2002).
Parganiha et al. (2006) reproduced comparable results on phototactic
rhythm in an experiment conducted under different light intensities (1- or
10- or 100-lux). The animals were not provided with the food during the
entire period of the study. A statistically significant circadian rhythm in
the phototacic activity of the population under 1- and 10-lux conditions,
but not under 100-lux condition, was demonstrated. The mean level of
activity gradually declined from 1-lux condition to 100-lux condition. In
general, amplitude declined and peak advanced as function of increase
in light intensity. In another study choice made by the population was
recorded every five minutes over a 90-minute session either at midday
or at midnight under all three 1-, 10- and 100-lux conditions from a
remotely placed monitor connected to a video camera. The Phototactic
Index (PI) was computed (Camassa 2001). Results obtained for phototactic

index supported the findings on the characters of rhythmic activity of the

population of loach in the light zone. The loaches were more negatively
phototactic in the midnight (versus midday) and at 100-lux condition
(versus 1- or 10-lux conditions). These results corroborate earlier findings
reported from our laboratory in that Nemacheilus evezardi exhibits a distinct
circadian rhythm in phototactic rhythm. Parganiha et al. (2006) concluded
that the background light intensity in the photic zone could alter circadian
rhythm parameters of phototactic rhythm considerably.
Rhythmic Behavior in Locomotor Activity

Number of studies has been performed by Pati and colleagues on
locomotor activity in N. evezardi. Pati (2001) reported a bimodal pattern in
locomotor activity rhythm in the hypogean population of cave loach kept
under LD 12:12 cycles. The onset and cessation of activity coincided with
the lights-off (18:00) and lights-on (06:00) timings, respectively. However,
it was more precise at the onset of darkness. The phase angle (y) was
computed with reference to both time points; it was found to be almost
zero to lights-off time. This indicates harmonization between beginning
of activity and onset of darkness, whereas such phenomenon could not
be validated with reference to lights-on timing. Most of the test fishes
exhibited a large amount of activity before and after lights-on time (Pati
2001). The bimodal pattern was further confirmed from the results of
Cosinor rhythmometry. Although a statistically significant 24-h rhythm in
locomotor activity was discerned in this hypogean form (Figure 6), a 12-h
rhythm was more pronounced. It can be explained in terms of peak spread
and amplitude. Peak spread was large (about 11.7 h) when computed at
24-h time window, whereas it reduced to 1.1 h when 12-h time window
was considered for the analysis. Likewise, difference was also observed
for amplitude. Amplitude was higher in 12-h rhythm than that of 24-h
rhythm. Further, results of spectral analysis also complemented the above
findings. In some specimens the locomotor activity rhythm exhibited
12-h period as the prominent component of its waveform. Interestingly,
this bimodality in locomotor activity pattern was absent in epigean
loach though they exhibited comparable activity rhythm. The onset and
cessation of activity were equally precise in epigean loach. In contrast to
hypogean loach, the population of epigean loach exhibited shorter peak
spread in 24-h fitted period as compared to that with 12-h period. Further,
amplitude of 24-h rhythm was also observed to be higher than that of
12-h rhythm. Circadian period (t = 24 h) was the prominent feature in
epigean loach (Pati 2001). This noticeable difference between hypogean
and epigean forms concerning bimodality and characteristics of rhythm
parameters of locomotor activity is difficult to explain at this moment.

Pati and colleagues conducted a number of experiments exposing the

hypogean and epigean populations of N. evezardi to LD 12:12 for about six
to eight days and thereafter to constant darkness (DD) or continuous light
(LL) for varying length of time. In one of the experiments the majority of
the hypogean loach exhibited free-running rhythms in locomotor activity
following their transfer from LD 12:12 to DD (Pati 2001). The population of
epigean loach also exhibited free-running rhythms in locomotor activity.
However, in both cases, it was difficult to locate the accurate timings of
activity onset, activity cessation or activity midpoints in the actogram.
A diminished 24-h mean activity level was observed in hypogean
loach following their transfer from LD 12:12 to LL (Pati 2008). A weak
and indistinct free-running rhythm with 24-h or nearly 24-h period in
locomotor activity was noticed in this loach under LL. These findings
clearly suggest that the hypogean loach, N. evezardi has an endogenous
circadian rhythm that is characterized by persistence of locomotor activity
rhythm when exposed to either DD or LL. Further, it strongly suggests
that the hypogean fish still possesses a functional circadian oscillator
underlying its overt circadian rhythm in locomotor activity under LD
12:12. It could be concluded that the circadian machinery is still conserved
in the population of the hypogean loach.
In another experiment Pati (2008) documented higher level of activity in
epigean population following their exposure to DD for 18 days. The 24-h
mean locomotor activity, expressed as % change from 24-h mean activity
in LD, increased many folds on individual basis as well as at group level
after onset of darkness. However, the significant changes in mean activity
level were observed only after 10th day. Surprisingly, hypogean loaches
did not exhibit any noticeable changes in the average level of locomotor
activity under DD. Another interesting significant difference was noticed
between the two populations regarding occurrence of peaks of locomotor
activity rhythm under DD. The hypogean fish always exhibited phase
advance, whereas the epigean fish exhibited phase delay under constant
darkness (Figure 7). Such differences between the hypogean and the
epigean loaches are good examples of divergence in the behavioral trait
following colonization of subterranean habitat.
Color Change Behavior

Chromatophore dependent color change behavior is a physiological
function that is present in many organisms, such as fish, frog, lizard, etc.
It has been reported that the epigean population of N. evezardi is able to
darken its body color very quickly with little excitement. Pradhan and
Biswas (1994b) demonstrated that this chromatophore dependent color

(computed with reference to average

12
Phase advance or delay in DD
4
peak in LD)
–4
–8
– 12
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Days after DD onset
Fig. 7 Day-to-day change in peak of the locomotor activity of the hypogean

loach (solid triangle) and the epigean loach (solid circle) in DD, computed from
their respective overall 24-h mean peak in LD. An asterisk indicates statistically
significant difference in the response to DD between the populations of the
hypogean and the epigean loaches, based on t-tests (p<0.05). Vertical line on
either side of the mean value (either solid triangle or solid circle) represents 1
standard error. The hypogean fish always exhibited phase advance, whereas the
epigean fish exhibited phase delay in locomotor activity rhythm under constant
darkness (DD).
change phenomenon is also functional in the hypogean population kept

under LD 12:12 photoperiod. Further, they observed that different stages
of chromatophore, such as punctate (stage-I), punctostellate (stage-II),
reticulostellate (stage-IV) and reticulate (stage-V) vary as a function of time,
except stellate (stage-III) (Figure 6). Pigments dispersed maximally during
later half of the light phase, whereas concentration was the maximum
during dark span of the LD 12:12 cycles (Pradhan 1984, Pradhan and
Biswas 1994b). Conversely, a statistically significant rhythm was validated
only in the occurrence of stage-I chromatophore (punctate type) with the
peak located at 06:48 under DD. Non-detection of significant rhythm in
other types of chromatophore in DD suggests absence of any particular
advantage of color change in subterranean mode of life. Notwithstanding
subterranean life inside the cave, chromatophore of the hypogean loach
still retained their ability to respond to light and darkness.
Bhargava et al. (1984) reported that the cavefish N. evezardi is extremely
sensitive to its background color/illumination. Under black background
it becomes dark, whereas on a white background it becomes pale. In
addition, they also demonstrated that the rate of color change is rapid
initially; later on it becomes slow and gradual, irrespective of its responses
to background (black or white).

Biochemical Constituent
A statistically significant circadian variation in a number of biochemical
constituents of different tissues has been documented in the hypogean
N. evezardi (Figure 6) (Pradhan 1984). Particularly content of muscle
glycogen, muscle lactate and acetylcholinesterase enzyme activities of gill
and brain tissues of the hypogean form exhibited a significant 24-h rhythm
when exposed to LD 12:12 cycles. However, this phenomenon could not
be validated for content of liver lactate or alkaline phosphatase activity
of gill tissue. Nonetheless, it suggests that biochemical clocks of this fish
could be synchronized by the artificially maintained photoperiods despite
the fact that it has been adapted to lifestyle inside the perpetually dark
cave. However, these biochemical variables in the hypogean population of
loach have not yet been studied under constant conditions (DD or LL).
SCHISTURA SIJUENSIS (MENON, 1987)

This fish belonging to the family Balitoridae is known only from its type
locality, the Siju Cave, Garo Hills, Meghalaya, India (25°21′N; 90°41′E).
There are number of papers that describe its systematic and phylogenetic
relationships (Kemp and Chopra 1924, Hora 1924, 1935, Pillai and Yazdani
1977, Menon 1987, Kottelat, 1990, Talwar and Jhingran 1992, Brooks and
Smart 1995). Interestingly, these authors describe this species under
various synonyms (refer Proudlove 2006). Further, none of these papers
deals with the study on its system biology or physiology or ethology.
SCHISTURA PAPULIFERA (MENON, 1987)

This species has been described (new species) recently from the Krem
Synrang Pamiang system in the Jaintia Hills, eastern Meghalaya (between
25°11′14′′N, 92°21′03′′E and 25°12′48′′N, 92°21′48′′E) (Kottelat et al. 2007).
The authors have claimed that this is the first strictly hypogean loach from
India and that it is not known if it has an epigean ancestor. It displays
few troglomorphic characters, such as regression in the pigment system
and vestigial lateral eyes (Kottelat et al. 2007). It differs from other known
species of Schistura characterized by the lower half of the head covered
by small skin projections and presence of five pores in the supratemporal
canal of the cephalic lateral-line system (Kottelat et al. 2007). This species
has been reported to be light insensitive (Kottelat et al. 2007), although so
far no empirical studies have been conducted to support its photoneutral
behavior.

HORAGLANIS KRISHNAI (MENON 1950)

Horaglanis krishnai is a unique freshwater fish belonging to the Family
Clariidae. It inhabits subterranean channels that anastomose with an
underground network of wells in the Kottayam district of Kerala, India
(between 8°04′N, 76°13′E and 10°21′N, 77°38′E). It has been reported to
migrate from one well to another through this network of subterranean
channels. It has also been collected from a locality at Ettumanur, which
is about 12 km from Kottayam town (Subhash Babu and Nayar 2004). It
could be considered endemic to Kottayam district. The systematics and
phylogenetic relationships of this species has been outlined elsewhere
(Romero and Paulson 2001, Proudlove 2006). There are two genera of
the family Clariidae in which the eyes are absent; they are Uegitglanis
Gianferrari and Horaglanis Menon. The former has been reported from
the wells in the village Uegit, Southern Somalia. Horaglanis is different
from Uegitglanis in having relatively shorter dorsal and anal fins, vestigial
pectoral fins without spines and gill membranes united with isthmus
(Subhash Babu and Nayar 2004).
H. krishnai (Maximum size: 4.2 cm) is blind and it is presumed that the
absence of the eyes could be attributed to the degeneration of the optic
lobes. There are no scales on the skin (Figure 8). It possesses bulbous
stomach that facilitates storage of food. It has specialized ileo-sphincter that
retains the digested food to maximize the rate of absorption (Mercy and
Pillai 1984). It has been speculated that it feeds on insects and crustaceans
available in the phreatic spaces. The alimentary canal has been suitably
modified to suit to its carnivorous habit. The skull of this fish is strong
and bones are firmly articulated. Perhaps this is essential for its migratory
Fig. 8 Subterranean Clariidae, Horaglanis krishnai (Source: www.planetcatfish.

com).
book.

habit inside narrow subterranean passages. Mercy et al. (1983) studied the
oxidative metabolism in this species and reported a linear relationship
between respiratory metabolism and body weight.
HORAGLANIS ALIKUNHII (SUBHASH BABU AND

NAYAR 2004)
H. alikunhii has an elongated eel-like body (3.2 cm) belonging to the
family Clariidae. It was discovered in a well in the void laterite soil of
Parappukara (10°23′N, 76°15′E), Trichur district, Kerala. It has four pairs of
barbels. Eyes are completely absent. Gill membranes united at the anterior
end near the lower jaw. The H. alikunhii differs from H. krishnai remarkably
with reference to shape of the head (elongated vs. globular), pectoral fins
(highly vestigial), and number of caudal fin rays (30 vs. 24) (Subhash
Babu and Nayar 2004). The red color of the live fish has been attributed
to the abundance of erythrocytes in the superficial blood capillaries and
is explained as an adaptation to the oxygen-deficient waters in its habitat
(Subhash Babu and Nayar 2004). However, so far no physiological or
behavioral studies have been conducted on this species.
MONOPTERUS EAPENI (TALWAR IN TALWAR AND

JHINGRAN 1992)
It is commonly known as Malabar swampeel and belongs to the family
Synbranchidae. The major synonym of this species is Monopterus indicus
(Eapen 1963). It is distributed in wells of Kottayam district (9°30′N,
76°33′E). The troglomorphic fish, H. krishnai and M. eapeni coexists in
the same wells. The fish is scaleless and has sunken eyes (Romero and
Paulson 2001). The fish lives in the water table up to 10 m below the
surface (Proudlove 2006).
MONOPTERUS ROSENI (BAILEY AND GANS 1998)

It belongs to the family Synbranchidae. It was found in a well at Periyam
village (10°38′N, 76°22′E), Kerala, India, about 100 km away from the type
locality of the M. eapeni. No details of its habitat and biology are available
except that it differs significantly from the other subterranean Monopterus
species, M. eapeni (Bailey and Gans 1998).

CONCLUSIONS AND FUTURE DIRECTIONS

Many other fish species have been reported from different caves of
India. However, they are presumably nonstygobitic fish fauna (Table 2;
Figure 9) without any distinct troglomorphic features. These fish species
might have been accidentally discovered in the caves. There are many
rivers that flow through the caves. The fish fauna found in the part of
the river inside the cave are not necessarily truly subterranean fauna.
In general, biological studies on subterranean fish fauna of India are
restricted to Nemacheilus evezardi and Horaglanis krishnai only. Further, the
subterranean Monopterus species, M. eapeni and M. roseni are known only
from single individuals. The biological data available on subterranean
species from India are meager. Although there may be thousands of caves
located on the Indian subcontinent, adequate efforts have never been
made to carry out biospeleological surveys. There is an urgent need to
lay emphasis on research on caves and cave fauna of India. The funding
agencies of the country should keep biospeleology in the list of the thrust
areas of research. Fully dedicated research centers should be created in the
already identified karstic areas of the country to pursue biospeleological
research.
Fig. 9 Nonstygobitic fish species collected from the Nelabhilum Cave, Andhra
Pradesh (courtesy: Ranga Reddy).
book.
Table 2 Suspected cave fauna from Siju cave and other caves of India*
Scientific name Species authority Cave State
Barilius barna Hamilton, 1822 Siju cave Meghalaya1

Barilius bendelisis Hamilton, 1807 Siju cave Meghalaya1
Danio aequipinnatus McClelland, 1839 Siju cave Meghalaya1
Acrossocheilus McClelland, 1839 Siju cave Meghalaya1
hexagonolepis
Tor tor Hamilton, 1839 Siju cave Meghalaya1
Noemacheilus Day, 1878 Siju cave Meghalaya1
multifasciatus
Channa orientalis Bloch & Siju cave Meghalaya1
Schneider, 1801
Ambassis nama Hamilton, 1822 Siju cave Meghalaya1
Schizothoracine fish Not known A cave near Rajasthan2
Udaipur
Tiny pigmented fish Not known Gupta Madhya Pradesh3
Godavari
Gufa (lower
cave)
Small pigmented fish Not known Gupta Madhya Pradesh3
Godavari
Gufa (upper
cave)
Puntius sarana sarana Hamilton, 1822 Nelabhilum Andhra Pradesh3
Puntius sarana subnasutus Valenciennes, 1842 Nelabhilum Andhra Pradesh3
Channa orientalis Bloch & Schneider, 1801 Nelabhilum Andhra Pradesh3
*
Based on: 1Pillai and Yazdani (1977), 2Tehsin et al. (1988), 3Reddy (2009 – Personal communication)
Acknowledgements
We thank Dr. Rohit Kumar Pradhan and Dr. Jayant Biswas, for permitting
us to use data from their Ph.D. theses submitted to Pt. Ravishankar Shukla
University, Raipur, India. We thank Dr. Ranga Reddy for providing us
with the pictures of nonstygobitic fish fauna of Nelabhilum Cave, Andhra
Pradesh. Some of the findings of the researches reported in this chapter
have been supported by the Department of Science & Technology, New
Delhi, and the University Grants Commission, New Delhi, through
its DRS-SAP sanctioned to the School of Life Sciences, Pt. Ravishankar
Shukla University, Raipur, India, in the thrust of area of research in
Chronobiology.

References
Agrawal, A. and A.K. Pati. 2002. Feeding schedule modulates phototactic responses
in cave populations of Nemacheilus evezardi. XVI International Symposium of
Biospeleology, Museo Civico di Storia Naturale di Verona, September 8-15: 27.
Bailey, R.M. and C. Gans. 1998. Two new synbranchid fishes, Monopterus roseni
from penisular India and M. desilvai from Sri Lanka. Occasional Papers of the
Museum of Zoology of the University of Michigan 726: 1-18.
Bhargava, H.N., A.K. Jain and D. Singh. 1984. On background related chromatic
response in the cave fish, Nemacheilus evezardi (Day). Journal of Animal Morphology
and Physiology 31: 203-209.
Biswas, J. 1990. Biospeleology: The behavioral and physiologic adaptations in a
cavernicole. Ph.D. Thesis, Pandit Ravishankar Shukla University, Raipur, India,
136 pp.
Biswas, J. 1991. Metabolic efficiency and regulation of body weight: A comparison
between life in hypogean and epigean ecosystems. International Journal of
Biswas, J. 1992. Influence of epigean environmental stress on a subterranean cave
ecosystem: Kotumsar. Biome 5: 39-43.
Biswas, J. 1993. Constructive evolution: Phylogenetic age related visual sensibility
in the hypogean fish of Kotumsar Cave. Proceedings of the National Academy of
Sciences India B63: 181-187.
Biswas, J. and A.K. Pati. 1991. Influence of thyroid hormones on muscle tissue
respiration in hypogean and epigean populations of loach Oreonectes evezardi.
Indian Journal of Experimental Biology 29: 933-936.
Biswas, J., A.K. Pati and R.K. Pradhan. 1990a. Circadian and circannual rhythms
in air-gulping behavior of cave fish. Journal of Interdisciplinary Cycle Research 21:
257-268.
Biswas, J., R.K. Pradhan and A.K. Pati. 1990b. Studies on burying behavior in
epigean and hypogean fish, Oreonectes evezardi: An example of behavioral
divergence. Mémoires de Biospéleologie 17: 33-41.
Biswas, J., A.K. Pati, R.K. Pradhan and R.S. Kanoje. 1990c. Comparative aspects of
reproductive phase dependent adjustments in behavioral circadian rhythms of
epigean and hypogean fish. Comparative Physiology and Ecology 15: 134-139.
Biswas, J., A.K. Pati, R.K. Pradhan and R.S. Kanoje. 1988. Reproductive phase
dependent adjustments in behavioral circadian rhythm characteristics: An
example of constructive evolution in the hypogean fish. Symposium on Recent
Trends in Comparative Endocrinology, Centre of Advanced Study in Zoology,
Banaras Hindu University, Varanasi, March 11-13: 18.
Brooks, S.J. and C.M. Smart. 1995. Caving in the Abode of the Clouds. The Caves and
Karst of Meghalaya, North East India. Bristol Exploration Club and Orpheus
Caving Club, Somerset and Derbyshire.
Camassa, M.M. 2001. Responses to light in epigean and hypogean populations
of Gambusia affinis (Cyprinodontiformes: Poeciliidae). Environmental Biology of
Fishes 62: 115-118.

Cambridge.
Culver, D.C., T.C. Kane and D.W. Fong. 1995. Adaptation and Natural Selection
in Caves: The Evolution of Gammarus minus. Harvard University Press,
Cambridge.
Day, F. 1872. Monograph of Indian Cyprinidae, Part IV. Journal of the Asiatic Society
41: 171-198.
Day, F. 1958. The Fishes of India; Being a Natural History of the Fishes Known to Inhabit
the Seas and Fresh Waters of India, Burma and Ceylon. William Dawson & Sons
Ltd., London, Vol. 1, 778 pp.
Dickson, G.W. and R. Franz. 1980. Respiration rates, ATP turnover and adenylate
energy charge in excised gills of surface and cave crayfishes. Comparative
Biochemistry and Physiology A65: 375-379.
Eapen, K.C. 1963. A new species of Monopterus from South India. Bulletin of the
Department of Marine Biology and Oceanography, University of Kerala 1: 129-132.
Eschmeyer, W.N. 1998. Catalog of Fishes. California Academy of Sciences, San
Francisco.
Hora, S.L. 1924. Fish of the Siju cave, Garo Hills, Assam. Records of the Indian
Museum 26: 27-31.
Hora, S.L. 1935. Notes on fishes in the Indian Museum. XXIV. Loaches of the genus
Nemacheilus from eastern Himalayas, with the description of a new species from
Burma and Siam. Records of the Indian Museum 37: 49-68.
Hüppop, K. 1985. The role of metabolism in the evolution of cave animals. Bulletin
of the National Speleological Society 47: 136-146.
Hüppop, K. 1986. Oxygen consumption of Astynax fasciatus (Characidae, Pisces):
A comparison of epigean and hypogean populations. Environmental Biology of
Fishes 17: 299-308.
Hüppop, K. 1989. Genetic analysis of oxygen consumption rate in cave and surface
fish of Astyanax fasciatus (Characidae, Pisces): Further support for the neutral
mutation theory. Mémoires de Biospéleologie 16: 163-168.
Kemp, S. and B. Chopra. 1924. The Siju Cave, Garo Hills, Assam Part 1. Introduction.
Records of the Indian Museum 26: 3-22.
Kottelat, M. 1990. New species and populations of cave nemacheilines in south
and south-east Asia (Osteichthyes, Balitoridae). Mémoires de Biospéleologie 17:
49-55.
Kottelat, M., D.R. Harries and G.S. Proudlove. 2007. Schistura papulifera, a new
species of cave loach from Meghalaya, India (Teleostei: Balitoridae). Zootaxa
1393: 35-44.
Maheshwari, R. 1998. An analysis of the air-gulping behavior of the catfish,
Heteropneustes fossilis, with reference to hormonal regulation. Ph.D. Thesis,
Pandit Ravishankar Shukla University, Raipur, India, 112 pp.
Menon, A.G.K. 1950. On a remarkable blind siluroid fish of the family Clariidae
from Kerala (India). Records of the Indian Museum 48: 59-66.

Menon, A.G.K. 1987. Noemacheilus sijuensis sp. nov. The fauna of India and adjacent
countries. Pisces, Zoological Survey of India, Calcutta. IV, Teleostei–Cobitidae,
Part 1, Homalopteridea, 175.
Mercy, T.V.A. 1981. The unique disposition of the valvula cerebelli in Horaglanis
krishnai Menon (Teleostei). Geobios 8: 95-96.
Mercy, T.V.A. 1982. On the presence of a forked barbel in the blind catfish Horaglanis
krishnai Menon. Geobios New Reports 1: 60-61.
Mercy, T.V.A. 1983. Studies on the air bladder of the blind clariid Horaglanis krishnai
Menon. Geobios New Reports 2: 111-115.
Mercy, T.V.A. 2000. Biology of an endemic blind catfish Horaglanis krishnai Menon.
In: Endemic Fish Diversity of Western Ghats, A.G. Ponnaih and A. Gopalakrishnan
(eds.). NBFGR/NATP Publishers, Lucknow, India, pp. 278-279.
Mercy, T.V.A. and N.K. Pillai. 1984. The anatomy and histology of the alimentary
tract of the blind catfish Horaglanis krishnai. International Journal of Speleology 14:
69-85.
Mercy, T.V.A. and N.K. Pillai. 2001. Studies on the cranial osteology of the blind
catfish Horaglanis krishnai Menon (Pisces, Clariidae). International Journal of
Mercy, T.V.A., K.G. Padmanabhan and N.K. Pillai. 1982. Morphological studies on
the oocytes of the blind catfish Horaglanis krishnai Menon. Zoologischer Anzeiger
209: 211-223.
Mercy, T.V.A., N.K. Pillai and N.K. Balasubramonian. 1983. Studies on the oxygen
consumption of the blind catfish Horaglanis krishnai Menon in light and dark
conditions. Matsya 9 and 10.
Mercy, T.V.A., N.K. Pillai and N.K. Balasubramonian. 2001. Studies on certain
aspects of behavior in the blind catfish Horaglanis krishnai Menon. International
Parganiha, A., A. Ramteke, A. Kar and A.K. Pati. 2006. Background light intensity
modulates circadian rhythm in phototactic behavior in the cave loach,
Nemacheilus evezardi. 18th International Symposium of Biospeleology, Cluj-
Napoca, Romania, July 10-15: 51-52.
Pati, A.K. 2001. Temporal organization in locomotor activity of the hypogean
loach, Nemacheilus evezardi, and its epigean ancestor. Environmental Biology of
Fishes 62: 119-129.
Pati, A.K. 2008. Circadian rhythms in hypogean fish: With special reference to the
cave loach, Nemacheilus evezardi. In: Fish Life in Special Environments P. Sébert,
D.W. Onyango and B.G. Kapoor (eds.). Science Publishers, New Hampshire,
pp. 83-130.
Pati, A.K. and A. Agrawal. 2002. Circadian rhythm in phototactic behavior in cave
loach Nemacheilus evezardi. XVI International Symposium of Biospeleology,
Museo Civico di Storia Naturale di Verona, September 8-15: 58.
Pillai, R.S. and G.M. Yazdani. 1977. Ichthyo-fauna of Garo Hills, Meghalaya (India).
Records of the Zoological Survey of India 72: 1-22.
Poulson, T.L. 2001a. Morphological and physiological correlates of evolutionary
reduction of metabolic rate among amblyopsid cave fishes. Environmental

Poulson, T.L. 2001b. Adaptations of cave fishes with some comparisons to deep-
Pradhan, R.K. 1984. Biochemical studies of some tissues of Nemacheilus evezardi
Day from Kotumsar Cave. Ph.D. Thesis, Pandit Ravishankar Shukla University,
Raipur, India, 195 pp.
Pradhan, R.K. and J. Biswas. 1994a. The influence of photoperiods on the air-
gulping behavior of the cave fish Nemacheilus evezardi (Day). Proceedings of the
National Academy of Sciences India B 64: 373-380.
Pradhan, R.K. and J. Biswas. 1994b. Towards regressive evolution: The periodic
color change behavior of a troglophilic fish Nemacheilus evezardi (Day).
International Journal of Speleology 23: 191-201.
Pradhan, R.K., A.K. Pati and S.M. Agarwal. 1989. Meal scheduling modulation of
circadian rhythm of phototactic behavior in cave dwelling fish. Chronobiology
International 6: 245-249.
(Hypogean) Fishes Described up to 2003 with a Bibliography 1541 – 2004. International
Society for Subterranean Biology, Moulis.
Rita Kumari, S.D., N.B. Nair and N.K. Balasubramanium. 1979. On the burying
behavior of the tropical loach, Lepidocephalus thermalis (Cuv. & Val.). Proceedings
of the Indian National Science Academy B 45: 129-141.
Singh, D.F. and G.M. Yazdani. 1993. Studies on the ichthyofauna of Nasik District,
Maharashtra, India. Records of the Zoological Survey of India 90: 195-201.
Subhash Babu, K.K. and C.K.G. Nayar. 2004. A new species of the blind fish
Horaglanis Menon (Siluroidea: Clariidae) from Parappukara (Trichur District)
and a new report of Horaglanis krishnai Menon from Ettumanur (Kottayam
district) Kerala. Journal of the Bombay Natural History Society 101: 296-298.
Talwar, P.K. and A.G. Jhingran. 1992. Monopterus (Monopterus) eapeni Talwar, nova
nom. In: Inland Fishes of India and Adjacent Countries, P.K. Talwar and A.G.
Jhingran (eds.). A.A. Balkema, Rotterdam.
Tehsin, R., V.S. Durve and M. Kulshreshtha. 1988. Occurrence of a Schizothoracine
fish (snow trout) in a subterranean cave near Udaipur, Rajasthan. Journal of the
Bombay Natural History Society 85: 211-212.
Thinès, G. 1969. L’Evolution regressive des poissons cavernicoles et abyssaux. Mason et
Cie, Paris.
Tiwari, P. 2000. Bastar ki Kotumsar Guphayein. in Hindi. English translation, The
Caves of Bastar. Sharda Publishing House, New Delhi, 22 pp.

About the Editors
Eleonora Trajano, zoologist and speleobiologist, is Professor at the Instituto

de Biociências da Universidade de São Paulo, Brazil, dedicating to the study of
subterranean fishes since her Doctorate in the 1980´s. She is a leading researcher
on Subterranean Biology in Brazil, coauthoring more than 120 papers published
in specialized refereed journals, books, annals of scientific meetings and in the
popular media, focusing mainly on the biology, ecology and evolution of bats,
fishes and subterranean invertebrates, also including faunistic surveys on caves
from several karst areas in Brazil and abroad. E. Trajano is an active member
of the International Society for Subterranean Biology and other international
scientific associations, and a member of several editorial boards. The conservation
of subterranean habitats is currently one of her major areas.
Maria Elina Bichuette, zoologist and speleobiologist, is Professor at the
Departamento de Ecologia e Biologia Evolutiva da Universidade Federal de São
Carlos, Brazil, since 2006. Her research area is the study of subterranean fishes,
mainly population ecology and behaviour. She is author of many Brazilian cave
fish descriptions, papers published in refereed journals, book chapters focusing
on threatened cave fauna, annals of scientific meetings focusing mainly on the
systematics, population ecology and evolution of fishes, besides subterranean
invertebrates. M. E. Bichuette is starting work in some Brazilian regions which
have never been studied until now; she is also council member of the International
Society for Subterranean Biology and other national scientific and speleological
associations.
B.G. Kapoor was formerly Professor and Head of Zoology in Jodhpur University
(India). Dr. Kapoor has co-edited 20 books published by Science Publishers,
Enfield, NH, USA. The most recent ones are Fish Defenses (Volume 1), 2009
(with Giacomo Zaccone, José Meseguer Peñalver and Alfonsa García-Ayala); The
Biology of Blennies, 2009 (with Robert A. Patzner, Emanuel J. Gonçalves and
Philip A. Hastings); Development of Non-Teleost Fishes, 2009 (with Yvette W.
Kunz and Carl A. Luer); Fish Defenses (Volume 2), 2009 (with Giacomo Zaccone,
C. Perrière and A. Mathis); Fish Locomotion: An Eco-ethological Perspective,
2010 (with Paolo Domenici). Dr. Kapoor has also co-edited The Senses of Fish:
Adaptations for the Reception of Natural Stimuli, 2004 (with G. von der Emde
and J. Mogdans); and co-authored Ichthyology Handbook, 2004 (with B. Khanna),
both from Springer, Heidelberg. He has also been a contributor in books from
Academic Press, London (1969, 1975 and 2001). His E-mail is: bhagatgopal.
kapoor@rediffmail.com

Color Plate Section
Foreword
Photograph of Tom Poulson reproduced by Courtesy of Dante Fenolio

460 Color Plate Section
Chapter 2
(1981-2007)
160
140 y = 0.7e0.046x
R2 = 0.99
120 (All data)
Cum. species
100
80
(1842-1981) y = 1.18e0.027x
R2 = 0.96
60
40
y = 1.22e0.026x
R2 = 0.94
20
0
0 50 100 150 200
Year
Fig. 1 An arithmetic plot of the cumulative number of subterranean fish species

fish was described.
1000
(1981-2007)
y = 0.07e0.046x
2
100 R = 0.99
(All data)
Cum. species
(1842-1981)
y = 1.22e0.026x y = 1.18e0.027x
2 2
R = 0.94 R = 0.96
10
1
0 50 100 150 200
Year
Fig. 2 A logarithmic plot of the cumulative number of subterranean fish species

fish was described.

Chapter 5
Fig. 1 Individuals of Astyanax mexicanus from a surface population and from

the Molino and Pachón cave populations. The cave populations independently
converged on eye reduction, albinism, and decreased numbers of melanophores.

Fig. 5 Scaled values of cave allele substitution effects for four traits in QTL
clusters on seven linkage groups in Astyanax mexicanus: Amino acid sensitivity
(A), number of maxillary teeth (T), relative eye size (E), and melanophore count
(M). Substitution effects were scaled against phenotypic range in the mapping
F2 (Pachon × Surface cross). Positive values are phenotypic increases, negative
values are decreases. (From Protas et al. 2008.)
A
Fig. 6 Contd..

Fig. 6 Contd..
Fig. 6 Genetic map of Astyanax mexicanus showing alignment of multi-trait QTL

detected in the mapping progenies of crosses between surface fish and cave fish
from Pachón, Tinaja, or Molino. QTL were detected as previously described (Protas
et al. 2008) and hand placed on an integrated map, based on positions determined
on cross specific maps. The bars cover + one LOD intervals. Multitrait QTL were
based on combinations of two or three traits: E = eye size; M = melanophore count;
T = number of maxillary teeth; D = depth of the caudal peduncle; P = placement of
the dorsal fin. QTL are coded, PACH, MOLI and TINA to identify the cross in which
they were detected. Traits are fully defined and methods are detailed in (Protas
et al. 2008).

Chapter 7
Fig. 1 Members of the Amblyopsidae. The family includes (A) the surface-
dwelling swampfish (Chologaster cornuta), (B) the troglophile spring cavefish
(Forbesichthys agassizii), and four troglobites: (C) the southern cavefish
(Typhlichthys subterraneus), (D) the northern cavefish (Amblyopsis spelaea), (E)
the Ozark cavefish (A. rosae), and (F) the Alabama cavefish (Speoplatyrhinus
poulsoni). Photos courtesy of Uland Thomas (A), Dante Fenolio (E), and Richard
Mayden (F).
C. cornuta
F. agassizii

T. subterraneus
A. spelaea + T. subterraneus
A. spelaea
A. rosae
S. poulsoni
Fig. 3 Distribution by county of the Amblyopsidae in the eastern United States. Only the swampfish, C. cornuta, is found outside
the Interior Low Plateau or Ozark Plateau.
Chapter 8
B C
D E
Fig. 1 (A) Location of field sites near Tapijulapa, Tabasco, Mexico. Blue indicates
non-sulfidic surface sites (AT Arroyo Tres, AA Arroyo Tacubaya, RA Río Amatan,
AC Arroyo Cristal, RO Río Oxolotan, AB Arroyo Bonita), yellow indicates sulfidic
surface sites (EA I, II El Azufre, sites I and II, red: the sulfidic Cueva del Azufre,
brown: the non-sulfidic Cueva Luna Azufre. The inserted figure shows the location
of the Cueva del Azufre system in Mexico. (B-E) Habitat types inhabited by P.
mexicana in the Cueva del Azufre system. (B) non-sulfidic surface habitat (Arroyo
Bonita), (C) sulfur sources in the springhead area of El Azufre, (D) Cueva del
Azufre. (E) P. mexicana occur at high abundance inside the Cueva del Azufre.

A B
C D
G
ch be
Fig. 7 Morphological divergence in P. mexicana. (A) Female and (B) male

from a non-sulfidic surface site (Arroyo Cristal). Note the bright coloration of
large,dominant males in (B). (C) Male from chamber XI of the Cueva del Azufre,
“golden morph”. Arrow shows the gonopodium, the male copulatory organ. (D)
female from chamber XI with a pronounced genital pad (arrow) around the genital
pore. (E) Cave molly female from chamber V, “sunken eye type”. (F) The role of
Ophisternon aenigmaticum as a molly predator in the Cueva del Azufre remains
to be examined. This dead individual was found in El Azufre and had probably
been flushed out of the cave. The ruler is 30 cm long. (G) Inside the Cueva del
Azufre, cave mollies face predation by giant water bugs (Belostoma sp., be). ch:
chironomid larvae, one of the food sources of cave mollies.
Non-sulfidic surface sites
Cueva del Azufre

Fig. 13 Life history trait evolution in P. mexicana. Assuming that the volume of
the ovary constrains the number of oocytes a female can produce, P. mexicana
females have two options: they can either produce numerous small oocytes
(above, female from Arroyo Cristal) or few, but large oocytes containing more yolk
(below, cave molly female from cave chamber X of the Cueva del Azufre (compare
Trexler and DeAngelis, 2003).

Chapter 9
(a)
(b)
(c)
(d)
Fig. 1 Contd..

Fig. 1 Contd..
(e)
(f)
(g)
Fig. 1 Brazilian troglobitic fishes, showing inter-specific differences in degree

of reduction of eyes and melanic pigmentation: a. Pimelodella spelaea (Photo:
Maria Elina Bichuette); b. Eigenmannia vicentespelaea (Photo: José Sabino); c.
Rhamdiopsis sp. 1 (Photo: Adriano Gambarini); d. Rhamdiopsis sp. 2 (Photo: José
Sabino); e. Ancistrus formoso (Photo: José Sabino); f. Stygichthys typhlops (Photo:
Cristiano Moreira); g. Glaphyropoma spinosum (Photo: Adriano Gambarini).

(a)
(b)
Fig. 2 Contd..

Fig. 2 Contd..
(c)
(d)
Fig. 2 Intra-specific variability in the degree of reduction of eyes and pigmentation

in Trichomycterus itacarambiensis; a-b: pigmented fish with variably sized eyes;
c-d: light pigmented versus albino specimens. (Photos: José Sabino).

(a)
(b)
Fig. 3 Contd..

Fig. 3 Contd..
(c)
(d)
Fig. 3 Habitat and population study methods for Brazilian cavefishes (Photos:
Adriano Gambarini): a-b. habitat of Glaphyropoma spinosum from Chapada
Diamantina, northeastern Brazil; c. collecting procedure for mark-recapture studies
and; d. measuring standard length for mark-recapture studies.

Chapter 10
Fig. 1 Distribution of African subterranean fishes. The type localities of the

eight species are indicated. 1: Caecobarbus geertsii; 2: Phreatichthys andruzzii;
3: Barbopsis devecchii; 4: Uegitglanis zammaranoi; 5: Clarias cavernicola;
6: Typhleotris madagascariensis; 7: Typhleotris pauliani; 8: Glossogobius
ankaranensis.

Fig. 3 Phreatichthys andruzzii (photo by R. Innocenti).
Fig. 5 Uegitglanis zammaranoi (photo by R. Berti; from Ercolini et al. 1982,

modified).

Fig. 9 Glossogobius ankaranensis (photo by Jane Wilson).

Chapter 12
2
3
4
5
6
7
Fig. 2 Distribution of Indian subterranean fishes. Numbers indicate the location

of the species found in India. 1: Nemacheilus evezardi; 2: Schistura sijuensis; 3:
Schistura papulifera; 4: Horaglanis krishnai; 5: Horaglanis alikunhii; 6: Monopterus
eapeni; 7: Monopterus roseni.

Fig. 3 Albinic subterranean loach, Nemacheilus evezardi.
Fig. 8 Subterranean Clariidae, Horaglanis krishnai (Source: www.planetcatfish.

com).

Fig. 9 Nonstygobitic fish species collected from the Nelabhilum Cave, Andhra
Pradesh (courtesy: Ranga Reddy).

Biology of Subterranean Fishes

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Biology of Subterranean Fishes

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Biology of Subterranean Fishes

Uploaded by

Copyright:

Available Formats

Biology of

© 2010 by Science Publishers

Maria Elina Bichuette

6000 Broken Sound Parkway, NW

© 2010 by Science Publishers

Marketed and distributed by:

Copyright reserved © 2010

Library of Congress Cataloging-in-Publication Data

All rights reserved. No part of this publication may be reproduced,

Subterranean fishes are among the most instigating, scientifically relevant

© 2010 by Science Publishers

Moreover, many troglobites, having evolved under relatively constant

Eleonora Trajano and Maria Elina Bichuette

© 2010 by Science Publishers

I have been asked by the editors to summarize Tom Poulson’s contributions

© 2010 by Science Publishers

rather to focus on what I think is the real reason we celebrate Poulson’s

© 2010 by Science Publishers

© 2010 by Science Publishers

resulted in an increase in extra-optic sensory structures, reduced metabolic

© 2010 by Science Publishers

I would be remiss if I left the impression that all or even most of

© 2010 by Science Publishers

© 2010 by Science Publishers

1. Cavefish: Retrospective and Prospective 1

© 2010 by Science Publishers

11. Subterranean Fishes of China 397

© 2010 by Science Publishers

© 2010 by Science Publishers

© 2010 by Science Publishers

D. Temporal and Spatial Scale

© 2010 by Science Publishers

Implicitly Darwin used the scientific method to develop his hypothesis

II. GENETIC AND DEVELOPMENTAL BASES FOR

© 2010 by Science Publishers

amount of inter-ommatidial pigment, but dozens that are moderately to

B. Alternative Hypotheses for Rudimentation

© 2010 by Science Publishers

C. Ongoing Studies of Rudimentation and Elaboration

© 2010 by Science Publishers

© 2010 by Science Publishers

© 2010 by Science Publishers

D. Sources of Genetic Variation

E. Population and Quantitative Genetic Approaches

© 2010 by Science Publishers

III. A HIERARCHY OF ADJUSTMENT TO STRESS

© 2010 by Science Publishers

IV. EVOLUTION OF TROGLOMORPHY IN CAVEFISH

A. Preadaptation: The First Step in Cavefish Evolution

B. Astyanax fasciatus vs the Amblyopsidae as Models

© 2010 by Science Publishers

enough food limitation to evolve the kinds of sensory elaboration and

© 2010 by Science Publishers

C. Phenotypic vs Genotypic Troglomorphy

© 2010 by Science Publishers

As in Astyanax fasciatus cave forms, each Amblyopsid cavefish species

© 2010 by Science Publishers

cerbellum with the highest known positive allometric coefficients of all