Discovering

Evolutionary
Ecology
Bringing together ecology and evolution

Peter J. Mayhew
University of York, UK

1
3
Great Clarendon Street, Oxford OX2 6DP
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide in
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trade mark of Oxford University Press
in the UK and in certain other countries
Published in the United States
by Oxford University Press Inc., New York
© Oxford University Press 2006
The moral rights of the author have been asserted
Database right Oxford University Press (maker)
First published 2006
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,
without the prior permission in writing of Oxford University Press,
or as expressly permitted by law, or under terms agreed with the appropriate
reprographics rights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above
You must not circulate this book in any other binding or cover
and you must impose the same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Data available
Typeset by Newgen Imaging Systems (P) Ltd., Chennai, India
Printed in Great Britain
on acid-free paper by
Biddles Ltd., King’s Lynn

ISBN 0–19–857060–0 978–0–19–857060–8

ISBN 0–19–852528–1 (Pbk.) 978–0–19–852528–8 (Pbk.)

10 9 8 7 6 5 4 3 2 1
Preface

There’s more to this life than just living.

Frank Borman, Apollo 8 astronaut

The natural world is a place I escape to: a place that goes about its business
regardless of everyday individual human concerns. It is a place of beauty,
change, diversity, and endless fa scination. Like many who share these senti-
ments, I was never content to just be in nature: I had to watch, name, learn,
and understand. This book is about understanding how and why the natural
world works, thereby to appreciate it more for what it really is. For me, that is
one of the things that make life ‘more than just living’.
For naturalists, two fields of science feel especially comfortable: ecology
and evolution. Ecology is traditionally a science of the great outdoors, dealing
with the interactions between organisms and their environment (including
other organisms). Evolution is traditionally a science of museum specimens,
dealing with how lineages of organisms arise, change, and eventually go
extinct. Both ecologists and evolutionary biologists share a common goal:
they want to understand the diversity of life; how it arises, how it is main-
tained, and why sometimes it is not. They should have a lot to say to each
other. The field where ecologists and evolutionary biologists meet is called
evolutionary ecology and, despite having 150-year-old roots, it has only
recently matured into something that can fill books.
This book has one overriding aim: to synthesize the field of evolutionary
ecology; that is, to explain what the field as a whole has discovered, rather
than just all the little bits. Along the way there is some detail; the work of
scientists. While the detail can exist without the synthesis, the synthesis gives
the detail added value. While some of the detail may change, be lost, or added
to, the synthesis I hope will remain.
I have written primarily for the students of biology whom I meet at
undergraduate level. In 1998, as a new lecturer at the University of York, my
colleague Richard Law invited me to take over his lectures on evolutionary
ecology. However, I found no books that dealt with the field in the way I
needed and decided to write my own. I have written the book that I would
have wanted as a student: using a short, informal style, so some people might
viii PREFACE

get to the end. As a result this is not a compendium of evolutionary ecology

knowledge. There is always more detail in the world, or indeed in any
scientific field, than any one person can assimilate. From what little detail we
do have, however, we mortals must formulate pictures of the world that we
can apply to novel situations, of which the world is full. I hope this book has
just enough to do that. The book may also be more widely accessible than I
originally meant it to be. I hope that postgraduates and other researchers in
the field, who tend to stay within the bounds of a single chapter, will find it
useful to have an overall view that places their work in a broader context. The
public at large should also have a fighting chance, and I have tried to make
that more likely by including a glossary of the more technical terms. Terms
included in the glossary appear in bold on first mention.
The precise content of the book was shaped by three secondary desires.
First, I did not want to write yet another behavioural ecology book. But,
because most evolutionary ecologists study behaviour, if I had devoted
space in proportion to the amount of work carried out in the various
subdisciplines of the field, that is pretty much what would have happened.
However, a behavioural ecology book would not have achieved my broader
aims. Instead, I have tried to cover a wide range of topics to do justice to the
breadth of the field in ways that previous books have not. Each chapter serves
merely as an introduction to each topic, about which others have written
entire books. For those who feel like learning a bit more, I make a few
recommendations for further reading at the end of each chapter. Some of
the topics in the book are not normally considered to lie in evolutionary
ecology, but more solidly in mainstream evolution or ecology. I have
included them because I feel they should be here.
Second, I am aware that most biologists express a greater enthusiasm for
some organisms than others. They spend a lot of time trying to persuade
each other that their study organisms are the most interesting. I believe that
to appreciate evolutionary ecology to the full, you must be prepared to dis-
card taxonomic and functional prejudice. This does not mean that you
should not feel a special affection for some taxa; rather you should not feel
disaffection for other taxa. The reader should be prepared for a good mix of
the botanical, microbial and zoological, aquatic and terrestrial. To empha-
size this even more I have occasionally employed positive discrimination in
my choice of material.
Third, I have not made a special effort to emphasize applied questions.
Evolutionary ecology can help solve many problems that beset our planet
and our species, but my desire here is to help people to love the subject, and
not to plague them with worry or guilt. I have included applied questions
simply where they provide a fascinating perspective that improves under-
standing. As it turns out, there should be enough applied biology to keep
enthusiasts happy.
 PREFACE ix

The chapters should preferably be read in sequence from start to finish

since they build upon each other to provide the overall picture at the end.
Because I still wanted this book to be scientific, factual statements are
supported by citations from the primary scientific literature, though space
and flow limited the extent to which I could do this. Space limitations also
meant that I often had to reduce long complicated stories to a few salient
points, leaving out alternative viewpoints. This makes it virtually certain that
researchers in the field, and possibly other readers, will disagree with me at
least once somewhere in the book. I hope that you all find such moments
stimulating.
Many people helped in the creation of this book. Biology students at York
made comments on my teaching that shaped the way the book was written.
Several people, mostly anonymously, reviewed the initial proposal, and I am
grateful to all of them. I particularly thank Brian Husband, who convinced
me that speciation mechanisms had to be included. I am grateful to the
following persons for commenting on draft chapters: Peter Bennett, Calvin
Dytham, Ian Hardy, Richard Law, Geoff Oxford, Ole Seehausen, Jeremy
Searle, and Mark Williamson.
Permission to reproduce photographs was generously provided by John
Altringham, Craig Benkman, May Berenbaum, Didier Bouchon, Sarah
Bush, David Conover, James Cook,Angela Douglas,Andrew Forbes, Richard
Fortey, Niclas Fritzén, Leslie Gottlieb, Peter Grant, Angela Hodge, Greg
Hurst, Mike Hutchings, Ian Hutton, Eric Imbert, Colleen Kelly, E. King, Hans
Peter Koelewijn, Thomas Ledig, Mark Macnair, James Marden, Stephane
Moniotte, Camille Parmesan, Olle Pelmyr, Thomas Ranius, Loren Rieseberg,
Dolph Schluter, Ole Seehausen, Kim Steiner, Robert Vrijenhoek, Truman
Young, Arthur Zangerl, and Gerd-Peter Zauke.
I am grateful to the following for permission to reproduce various figures:
The American Association for the Advancement of Science, The Royal
Society of London, The Society for the Study of Evolution, and Springer
Science and Business Media. Ian Sherman at Oxford University Press opened
the door to what you are reading, gave valuable advice, displayed admirable
patience, and was above all a friendly face. I am grateful to Alastair Fitter,
for granting me the sabbatical term in which I made the majority of progress.
I was also supported by my colleagues at York who bore the brunt of my
‘normal’ work while I was on sabbatical, particularly Calvin Dytham and
Dale Taneyhill. Finally, thanks to my wife Emese and daughters Alice and
Lara, the former for understanding my need to write the book and support-
ing me in the struggle, and the latter for illustrating to me at first hand many
of the interesting concepts mentioned in the book.
 Contents

1 Where two fields meet 1

2 Evolutionary cover-stories 13
3 Brave new worlds 25
4 Traits, invariants, and theories of everything 37
5 Sons, daughters, and distorters 51
6 Voyagers, residents, and sleepers 64
7 Doing adaptive things 75
8 Evolution and numbers 86
9 A world of specialists 97
10 The good, the bad, and the commensal 108
11 Evolving together 120
12 Birth of species 132
13 Death of species 145
14 Big evolution 158
15 Big ecology 170
16 Combining in diversity 180

REFERENCES 186
GLOSSARY 204
INDEX 209
1 Where two fields meet

A teacher of mine once simplified his complex family history by saying that
he, like all of us, originated from Olduvai Gorge in Tanzania (the ‘cradle of
mankind’). Tropical Africa has been a cauldron of diversity not only for our
own species. It is, to take one example, surprisingly fishy. The Great Lakes of
East Africa (Figure 1.1), and surrounding rivers, contain a whopping 1500
species in just one fish family, the cichlids, familiar to freshwater aquarium
enthusiasts. This makes cichlids the most species-rich family of vertebrates,
beating such diverse and familiar groups as songbirds and mice. They are so
diverse that many still await proper scientific description, and many more
are doubtless completely undiscovered. Lakes Victoria and Malawi each con-
tain about 500 species, and about 250 species are found in Lake Tanganyika.
Diversity of this sort is what makes our planet such an interesting place, and
of course, we have to find out what caused it.
The cichlid species of the East African lakes have not each immigrated
there from the surrounding habitat; they were born there, and in most cases
they are endemics, being found in just one of the lakes (Fryer and Iles 1972).
They are a ‘radiation’ of species. This radiation is all the more remarkable
when the ages of the lakes are considered. Lake Tanganyika is the oldest (but
has fewest species) at about 10 million years. Lake Malawi, the second oldest
is a mere 1–2 million years old. Lake Victoria, amazingly, may have been
completely dry around 14,500 years ago, the end of the last ice age. Since
then, 500 cichlid species have been born. If species arose in a clockwork
linear fashion, that would mean one new species of fish every 29 years!
The varied lifestyles of the fish are equally impressive. In Lake Victoria, for
example, have been found cichlids with the following diets: adult fish,
fish larvae, fish scales, fish parasites, freshwater snails, insect and other
invertebrate larvae, plant and animal plankton, algae growing on rocks, and
vascular plants, all with specialized jaws to match (Figure 1.2). The most
impressive radiations have occurred among the ‘haplochromine’ cichlids
living on rocky shores in Lakes Victoria and Malawi (Kocher 2004). Clearly,
we need to know how so many species could have formed in such a short time
span, why it happened here, why cichlids, and why haplochromines most of
all? At stake is our understanding of species richness itself.
2 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 1.1 Seen from space, the Great Lakes of the East African Rift Valley are major landscape features.
The two largest ones here are Lake Victoria (right) and Lake Tanganyika (bottom)—Lake
Malawi is off the bottom of the picture. Lake Victoria is about 300 km across and its northern
tip is on the equator. Photo from the NASA Visible Earth image archive. Black lines indicate
national boundaries.

1.1 Alternative mechanisms

First, let us think briefly about how species are supposed to form. The
standard dogma is that this happens through geographic separation and
subsequent differentiation. One lineage splits into two distinct ones because
a spatial separation occurs, either through a dispersal event to an isolated
 WHERE TWO FIELDS MEET 3

Fig. 1.2 The diversity of jaw morphology of Lake Victoria cichlids. Clockwise from top left they eat,
snails, fish, fish larvae, algae on rocks, invertebrates on rocks, insect larvae.

new region, or through fragmentation of an existing one (vicariance). The

lineages evolve in isolation, through natural selection or other processes, and
eventually become distinct enough to be called new species. The differences
between related, but geographically isolated species are what gave Darwin
and Wallace many clues to their theory of evolution.
Could such processes be at work in the fastest vertebrate radiation?
Geographic separation and natural selection have undoubtedly contributed,
and a number of observations on geographic distribution and morpholo-
gical divergence among species are consistent with the process. For example,
closely related sister species in Lake Victoria sometimes have widely sep-
arated geographic ranges (Seehausen and van Alphen 1999); and different
populations of the same species have distinct jaw morphologies that match
local diets, suggesting local adaptation (Bouton et al. 1999). But there
remains a dearth of special explanation: why here and why haplochromines?
A growing weight of evidence suggests a role for additional mechanisms
and in particular in haplochromines.
4 DISCOVERING EVOLUTIONARY ECOLOGY

What additional mechanisms might be important? Can speciation, for

example, occur without geographic isolation? There are two problems that
need to be overcome. First, there has to be ecological divergence: the two
incipient species have to occupy different niches to prevent them from
competing and allow stable coexistence. Second, there has to be reproductive
divergence, so that interbreeding does not occur. Getting these events to
occur without geographic isolation is a conceptual challenge that has long
occupied evolutionary biologists. In the 1990s, this question was bothering
cichlid enthusiast, Ole Seehausen. Ole’s hunch was that species could diverge
in situ into reproductively isolated populations by assortative mating
based on male coloration. Over time, mate selection by different females for
different coloured males would produce two reproductively isolated species
living in the same ecological niche but differing in male coloration. Once sep-
arated like this, the way would be open for natural selection to allow niche dif-
ferentiation. The process could then repeat itself. The power of this
mechanism is its potential speed. Initial ecological differentiation need only
be small, and the constant disruptive power of female choice would drive
populations rapidly apart. It was a process that seemed capable of giving rise
to a multitude of species in a very short time.
What evidence supported this hypothesis? One source is patterns of
geographic overlap between species. If speciation has occurred in the
absence of geographic separation, there should also be groups of closely
related species that overlap in range a lot. In fact, there are many such cases
in Lake Victoria (Seehausen and van Alphen 1999). What about sexual
selection? In the field, sympatric sister species tended to be opposite colours
more commonly than allopatric pairs of species. This is consistent with the
origination of new species via selection on coloration in situ. These patterns
have also recently been demonstrated in Lake Malawi cichlids (Allender et al.
2003). In the laboratory, females from red species behaved preferentially
towards red males, as did females of blue species towards blue males. When
exposed to monochromatic light that hid the males’ bright nuptial hues,
females would no longer show a mate preference (Seehausen and van
Alphen 1998). This was indeed assortative mating based on colour. But why
should female mate choice be disruptive? One possible answer is percep-
tual bias: the colour-sensitive cone cells of haplochromines are particularly
sensitive to red and blue parts of the spectrum, and these different sensi-
tivities could lead females to perceive red or blue males preferentially
(Seehausen et al. 1997). However, other possible mechanisms could be at
work. What ever the mechanism, female haplochromines agree with
Winston Churchill when he said: ‘I cannot pretend to feel impartial about
colours. I rejoice with the brilliant ones and am genuinely sorry for the
poor browns’.
 WHERE TWO FIELDS MEET 5

25

Number of coexisting species

20

15

10

0
0 100 200 300 400 500
Width of the transmission spectrum (nm)

Fig. 1.3 Number of coexisting cichlid species against the clarity of the water at different sites in Lake
Victoria (a wide transmission spectrum represents clear water). After Seehausen et al. (1997),
with permission from AAAS.

Another feature of haplochromine cichlids is that if mating does take

place between individuals of different species, the offspring are normally
perfectly viable. The only reason they can be called separate species at
all is because of their fussy mate preferences. Evolutionary biologists
call this ‘pre-zygotic’ isolation. In the field, Ole started to find rather
disconcerting observations that mimicked what he was seeing in the
laboratory (Seehausen et al. 1997). Where the water was murky, and that
was often quite a recent phenomenon, he found few species of fish
(Figure 1.3) and of dull brown coloration. In clear waters, many species
coexisted together, and they were beautifully coloured. It looked as if
previous mating barriers were breaking down. Turn it on its head, and mate
choice in clear water seemed to have allowed divergence and maintenance
of species in the first place.
Could disruptive mate choice be the reason why it is the cichlids, and not
some other fish group, that have diverged in this way, and especially the
haplochriomine fish that radiated in lakes Victoria and Malawi? That too
appears to be the case. Comparing the incidence of mating system and male
nuptial coloration in different cichlid groups, Ole showed that there was a
significant association between the incidence of polygyny (where males
mate with more than one female, long associated with highly selective female
mate choice) and male nuptial coloration. Furthermore, the base of the
radiation that gave rise to the fish ‘superflocks’ of Lake Victoria and Malawi,
the haplochromines, was characterized by the origin of male nuptial
coloration (Seehausen et al. 1999).
Could not some other fish group possessing strong sexual selection also
have radiated? Put another way; is there anything else about the cichlids,
6 DISCOVERING EVOLUTIONARY ECOLOGY

which would lead to this mating system being particularly diversifying for
them? Part of the answer may have to do with that second essential process
of speciation without geographic isolation, ecological divergence. Some
kind of novel ecological flexibility might open up new niches, making each
new speciation experiment more likely to succeed. In fact cichlids have
long been known to possess a novel character that would lead to such
flexibility: the ‘decoupled pharyngeal jaw’ apparatus (Liem 1973). The bones
of the mouth have been freed to evolve into specialized food-gathering
implements, while the bones at the back have become very efficient grinding
elements. This novelty has given the cichlids jaw-evolvability as well as
behavioural plasticity. That it has played an important role in the present
diversity of cichlids is a very good bet.
Therefore, much evidence points towards a role for disruptive sexual selec-
tion acting on male coloration, followed by ecological differentiation as the
reason why cichlids, and particularly those in Lakes Victoria and Malawi, have
diversified so rapidly and why those species are still maintained.It is a nice idea.
But does a world with those simple conditions produce the desired result? Will
it also work in theory? This problem was tackled by a talented undergraduate
at the University of Utrecht, Sander van Doorn, who built a simulation model
of the process (van Doorn et al. 1998). This step is an important one, because
ultimately biologists want to put aside a small set of essential processes into a
body of theory that captures the essence of reality. We have to know what
processes are sufficient and important, and which are just noise.

1.2 Simulated lakes and simulated radiations

Theoretical models consist of assumptions, a best guess about how things work
in nature,and predictions,which are the model results.A good model will make
a few biologically reasonable assumptions and result in predictions that bear a
strong resemblance to reality, hence isolating the important mechanisms.
Van Doorn and colleagues started by assuming that individual fish can be
characterized by a colour preference (of females), a pigmentation (of males),
and by their niche use (represented for simplicity by a single number: think
of it as prey size, or water depth). Individuals compete, and are more likely to
die if their niche use is similar to that of other individuals. This keeps the
population size limited. Fish are born by sexual reproduction, which is
dependent on female mate preference, male pigmentation, and the degree
of niche overlap (similar niches increases the probability of mating). Mate
preference, male pigmentation, and niche use are also heritable, so that
offspring resemble their parents, but imperfectly, so small random changes
(mutations) are created in each generation. Finally, the more brightly
 WHERE TWO FIELDS MEET 7

coloured males are, the lower their survival as a result of natural selection
(such as predation). So far, so good.
One other important assumption is that females have peaks in perceptual
ability at both ends of the colour spectrum. Perceptual ability relates the
pigmentation of males to the colour perceived by females. In perfectly
clear water, there is a near-perfect match between the two, although females
perceive very bright pigments (at either end of the colour spectrum) slightly
better than others. In very murky water, all pigments appear brown to
females. In slightly murky water, only pigments that are close to female’s
perceptual peaks are perceived to be coloured.
The model was run by starting off a small population of a single species in
clear water and letting mating, reproduction, and death take its course.
Species were defined as groups of individuals that, because of their
niche, colour, or preference, were very unlikely to mate, and could hence
evolve independently of the others. After 2000 generations, five species
were coexisting from the original species in this simple and tiny virtual
lake. The process could clearly work. How exactly does it happen?
The key is female mate choice. As a result of the biased perception of red
and blue, on average females prefer males that have more-extreme-than-
average pigmentation. As a result, both pigments and preference become
more extreme over time (Figure 1.4). The process is a familiar concept in

Red
Female colour preference

Blue
Blue Red
Male colour pigments

Fig. 1.4 Speciation via sexual selection in the van Doorn et al. (1998) model. Individuals of different
species are represented by different symbols. The curved line represents female preference
for male colour and is biased towards red and blue (females on average prefer males that
are bluer or redder than the population average). Because of this bias, brown species
gradually split into two, one redder and one bluer, as can be seen with the species repre-
sented by the open squares. After van Doorn et al. (1998), with permission from the Royal
Society of London.
8 DISCOVERING EVOLUTIONARY ECOLOGY

sexual selection theory and is known as a ‘runaway’ process. Species that

have neutral colours and preferences, neither red nor blue, will split into
two species with slightly brighter (redder or bluer) colours. Once incipient
species no longer interbreed, their niches diverge as a result of competition
(this can not happen in a single species because interbreeding stops the niche
changing). The amount of niche space present limits the number of species
that can coexist, and it is for this reason that the model only produces a few
species. If species have different niches, they can have more similar colours
without losing their integrity as species. That of course is exactly what we see
in Lake Victoria: lots of species with the same nuptial colour.
The final triumph of the model is what happens when the water is made
turbid. Species cannot diverge, or any longer remain sexually isolated
because all males appear the same to females. Species number crashes, just
as in nature. The model is successful because by using the small pieces of
biology gathered so far, it successfully predicts many of the important
patterns in nature: it is a good conceptual cartoon for what goes on in nature.
However, the model appears not to be the last word in cichlid speciation.
Species in the model form from brown fish gradually splitting into slightly
less brown ones. In fact, individual species in nature often display a male
red/blue colour polymorphism, suggesting that speciation and colour
change are much more instantaneous. Thus, the model is in some respects
only a rough cartoon of some of the actual processes. In addition, there is a
second type of colour polymorphism within some species in which
females vary in colour and are associated with a rather interesting genetical
system (Seehausen and van Alphen 1999; Seehausen et al. 1999). Something
different must be going on in those.
Teaming up with theoretician Russ Lande, Seehausen devised a model that
incorporates these ‘instant’ novel female colour morphs with the strange
genetics in a sympatric speciation scenario (Lande et al. 2001). They showed
that given the way novel colour morphs and other traits are inherited
together, rapid speciation is likely to result even without ecological differen-
tiation. The female colour polymorphism is due to a gene that causes sex
reversal from male to female and is associated with a distinct colour pattern
(Seehausen et al. 1999) (Figure 1.5).
Imagine then that novel colours are only seen in females. Unusual males
that prefer, or do not discriminate against, this colour now have high mating
success for two reasons; they are rare male phenotypes, so get all the mating
with unusual coloured-females that normal males pass by. In addition, if the
sex-reversal gene is widespread, they will also be the rarer sex, so get more
mates anyway. This process, which favours the novel males through rarity of
the male sex, is called sex ratio selection. We will encounter this process again
in Chapter 5. An association between the new colour morph and preference
 WHERE TWO FIELDS MEET 9

Fig. 1.5 A cichlid, Paralabidochromis chilotes, from Lake Victoria (length 15 cm). Blotchy morphs like
these are, in most populations, female, and include sex reversed males that may play a role in
speciation by sex ratio selection. Photo courtesy of Ole Seehausen.

for that colour morph builds up. Over only a few dozen generations, a new
reproductively isolated species has arisen in situ.
It appears likely that at least two in situ processes can account for colour-
diverse haplochromine species richness: sexual selection and sex ratio
selection. Both these processes can cause speciation with geographic separa-
tion, but they can also do it in the absence of geographical separation. The
processes appear bizarre and extraordinary at first sight. However, both
processes are not unexpected in a wider context; we will come across them
again later in the book. What then has the cichlid story taught us?

1.3 Cichlids and evolutionary ecology

The cichlid story illustrates many of the broader features of evolutionary
ecology, the science that involves both ecological and evolutionary know-
ledge. Evolutionary biology is the field concerned with understanding how
biological lineages change through time (anagenesis), split (cladogenesis),
and ultimately go extinct. Ecology is concerned with the interaction of
organisms with their environment. The organisms can be considered at
various levels of a hierarchy, comprising the individual, the population
(groups of individuals of the same species), and the community (groups
of interacting populations from different species). Communities in turn
comprise the biotic component of ecosystems, which also include their
interactions with the abiotic world. Ecology asks how individuals behave in
different environments, what determines population size, and the properties
of communities and ecosystems, such as their diversity. Knowing all this,
why do ecology and evolution interact and how do they do so?
A basic answer, and one that does not require much in-depth study, is that
both fields are concerned with understanding similar characteristics. For
10 DISCOVERING EVOLUTIONARY ECOLOGY

example, both evolutionary biologists and ecologists would consider species

richness as one of the key variables they want to understand. Both too
would want to understand why species richness varies across environments,
such as different lakes in the case of cichlids, and across clades, such as
haplochromines versus other cichlids or cichlids versus other fish.
Another answer, that requires some knowledge of the subject, is that
evolutionary and ecological processes are affected by each other
(Figure 1.6). They do this in many ways: one way is through adaptation.
Darwin’s and Wallace’s greatest discovery was an understanding of the way
in which this occurs: evolution through natural selection. Organisms vary
in form (phenotype). These forms are heritable because of variation in
their underlying genetics (genotype). The phenotypes interact with their
environment, and some are more successful than others for a variety of
reasons: they may survive or reproduce better. This differential success is
called natural selection. Thus, the individuals that contribute to the gene
pool of the next generation are a subset of those that were born and will
pass on that subset of characteristics to the next generation through
their genotype. In this way the population changes through time. A second
type of selection process is normally distinguished from natural selection:
sexual selection. Sexual selection causes evolution of traits affecting
mating success in males and females. Both natural selection and sexual
selection come about from phenotypes interacting with their environ-
ment, and for this reason selection is generally viewed as an ecological
process. Natural selection is responsible for the evolution of traits, such as
cichlid jaw shape, which governs their ecological niche. Sexual selection is
responsible for traits, such as the bright male coloration of cichlids, that
influence their mating success.
So ecology, through the medium of selection, causes anagenesis, evolution
within lineages. Ecology can also influence cladogenesis, the other big evolu-
tionary process. We saw this in the role that water clarity plays in speeding up
or slowing down rates of cichlid speciation and extinction.
Evolution can also affect ecology, and this interaction occurs at several
levels of the ecological hierarchy (Figure 1.6). If you know about the envir-
onment, you can sometimes accurately predict, or at least in retrospect
understand, the phenotypes that are favoured. Evolutionary biologists
need to do this routinely, and it will be a repeated theme throughout this
book. Ecology at the level of the individual is largely concerned with trying
to predict how individual traits should be related to the environment
through selection pressure. Behavioural ecology is the field that asks what
behaviours would suit particular environments, such as the mate preferences
seen in haplochromine cichlids. It is one of the richest parts, but only one of
 WHERE TWO FIELDS MEET 11

Ecology Evolution

Individuals
Anagenesis
Populations

Communities
Cladogenesis
Ecosystems

Fig. 1.6 The interaction of ecology and evolution.

the parts of evolutionary ecology. Hence evolution by natural selection

affects ecology at the level of the individual.
The traits that evolve within species are often relevant to population
and community processes. For example, each species has a characteristic
reproductive rate, size, and length of life. These are important character-
istics in determining how many individuals of a species can exist in any
one place, and how variable the populations are. Species also vary in their
ecological specialization; for example, how many other species they eat or
which eat them. Haplochromine cichlids, for example, have very specialized
jaws. These interactions evolve through natural selection, but they also
structure communities. Knowing about one should help us to understand
the other.
The second major evolutionary process, cladogenesis is also important for
an understanding of ecology. To produce species-rich communities, such
as in East African lakes, species have to be formed and not go extinct. Both
evolution within lineages and the origin and death of lineages are processes
that might have contributed. Thus evolution influences every level of the
field of ecology and maybe key to understanding some of the basic ecological
properties of our planet.
In the following chapters, we will explore the ways in which the two fields
of ecology and evolution interact, see what we have learnt about the world as
a result, and along the way build up a picture of how exactly these interac-
tions occur. However, the book will describe something else about evolu-
tionary ecology that cannot be fully appreciated without an overall view
of the field. It is that the topics which the field addresses are mutually sup-
portive, such that understanding of one aids understanding of others. For
example, we can understand the rates of speciation in cichlids from a knowl-
edge of speciation and extinction mechanisms, and we can understand those
from a knowledge of sexual selection and sex determination. Ultimately
12 DISCOVERING EVOLUTIONARY ECOLOGY

then, workers in one area will benefit from an awareness of other areas. This
is what makes a synthesis worthwhile. Knowing how these interactions
between topics occur reveals interesting features about how our living uni-
verse is shaped, and provides another aspect to the bigger picture that the
field depicts. The next chapter looks at how organisms became complex
from very simple beginnings.

1.4 Further reading

The arguments here about cichlid speciation are well described in Seehausen
(2000), and much of the story is told in Seehausen et al. (1997) and the Galis
and Metz (1998) commentry on this. Meyer (1993) and Turner (1999) are
also useful. More general reviews about cichlids, including speciation and
sexual selection, are in Kornfield and Smith (2000) and Kocher (2004). The
general issue of sympatric speciation is reviewed by Via (2001).
2 Evolutionary cover-stories

Great things are not done by impulse, but by a series of small things
brought together.
Vincent van Gogh

People have long suspected that the first organisms must have been rela-
tively simple. Since the origin of life, some organisms must therefore have
undergone important evolutionary transitions that resulted in the kinds of
species with which we are now familiar. In recent years, biologists have come
to view these transitions not only as revolutions in the way living organ-
isms looked and behaved, but also as solutions to similar problems.
Understanding how they occurred brings us great insight into how natural
selection works, and why modern, complex, organisms live and behave as
they do.
Two people, John Maynard Smith and Eörs Szathmáry, did much to
promote this conceptual unification in the 1990s (Szathmáry and Maynard
Smith 1995; Maynard Smith and Szathmáry 1995, 1999). Together they
defined eight major transitions (Figure 2.1), united by changes in the way
that genetic information is transmitted between generations. In the origins
of life they postulated individual replicating molecules forming populations
of such molecules in compartments, such as cells (1). Later on these replicators
bound physically together into chromosomes (2). Eventually, RNA, acting
as both a replicator and metabolic catalyst, largely gave up these functions
to more specialist molecules: DNA and proteins (3). Some prokaryotes
(bacteria) eventually transformed into eukaryotes (4). Asexual clones
among the eukaryotes transformed into sexual populations (5). Some single-
celled protists transformed into multicellular organisms (plants, animals,
and fungi) (6). In a few groups, solitary individuals began to live in social
colonies (7), and in one of these, our own species, language emerged (8). We
therefore bear the distinction of being the only lineage that has undergone all
eight transitions. This would make us, in some quantifiable sense, the most
complex biological entities not only in what we have evolved but in how we
evolve.
14 DISCOVERING EVOLUTIONARY ECOLOGY

Cells Chromosomes DNA + Proteins

Eukaryotes
Multicellularity
Sex

Social colonies

Language

Fig. 2.1 The major transitions in evolution.

Explaining these individual transitions is challenging for three reasons,

summed up by three different senses in which they are ‘major’. The first, and
the one that Maynard Smith and Szathmáry stress, is in an intellectual sense;
the phenotypic changes we have to postulate are in themselves changes to the
genetic system. This requires us to think especially hard about how evolution
works because evolutionary biologists normally have the luxury of assuming
that the genetic system is a constant. The second use of the word ‘major’ is in
a structural sense: that the phenotypic changes were large. However, to be
consistent with Darwinian evolution, changes must proceed by a series
of small steps that retain functional integrity and which will be favoured
by selection in each generation. We must first therefore imagine possible
intermediate phenotypes, not all of which might be illustrated in the world
about us. Then we must imagine environments or circumstances in which all
the postulated intermediates would be favoured. In meeting these first two
challenges we are postulating solely the origins of the characters involved.
The third meaning of the word ‘major’, however, is that the changes were
in some sense ‘successful’ from a macro-evolutionary perspective. In this
sense we are implying that the transition was retained to the present day, and
usually retained in abundance. This creates special challenges because, as will
be shown later, the transitions can be seen to set up potential conflicts that
 EVOLUTIONARY COVER-STORIES 15

would disrupt the integrity of the new system. Many of the transitions
require formerly independent, or even totally new, genetic systems to come
together and cooperate as part of a larger system.Yet, biologists are now used
to the idea of genetic entities displaying selfish behaviour to ensure their own
persistence. Thus it is sometimes problematic to imagine persistence of
the novel unit. To cap off our problems, hypotheses must be consistent
with existing evidence. Thin though that often is, even a little evidence can
establish useful boundaries to possibilities, as fictional detectives are apt to
explain.
In the third sense, the transitions were not initially major, but with the
benefit of hindsight, having stood the test of time, many can now be seen to
be so. To be retained in abundance, there are four possible contributing
processes. First the transition might have happened on numerous occasions.
In fact this is normally not the case. All of the transitions have happened to
our knowledge only once, with the exception of multicellularity and social
colonies, which have both evolved a limited number of times. This relative
uniqueness is unsurprising given the drastic nature of the changes. The other
three processes are; (1) reversal to the ancestral state, which might have been
limited; (2) extinction of clades possessing the trait, which might have been
reduced; and (3) speciation of clades possessing the trait, which might have
been increased. The problem with explaining persistence is to find evidence
for or against these processes.

2.1 Sex as a major transition

Let us see how one of the transitions stands up to these challenges. Sex
technically refers to a special type of cell cycle (Figure 2.2), not, as is more
normally used, copulation. Understanding the evolution of sex therefore
means thinking hard about how cells replicate and divide and why this might
change. Since the way cells do that is normally taken for granted, it is useful
to be prepared for the unexpected in the paragraphs that follow.
Sex undoubtedly evolved in eukaryotes from a clonal ancestral state. A
normal (mitotic) cell cycle is comparatively simple: some time into its life,
each chromosome copies itself, and then the cell divides into two. In a sexual
(meiotic) life cycle, new diploid offspring are born by fusion of two haploid
gametes (syngamy). The gametes that fuse are normally very different in
form and behaviour (anisogamy), one small and motile (sperm), the other
large and immobile (egg). A number of mitotic cell cycles may then follow
(development in multicellular organisms). Then some homologous
chromosomes swap bits of DNA (recombination), in a process known as
‘crossing over’because of the appearance of the process under the microscope.
16 DISCOVERING EVOLUTIONARY ECOLOGY

2nd (mitotic-like)
Crossing over
division

Reduction
Pre-meiotic division Haploid mitosis
doubling

Division

Anisogamy
Pre-mitotic Syngamy
doubling

Fig. 2.2 A sexual life cycle. The dark lines are chromatids of a single homologous pair of chromo-
somes, drawn to indicate whether the cell is haploid or diploid, and whether the chromatids
have replicated or not. The circles are cells.

They then copy themselves and undergo two cell divisions to give rise to four
haploid cells. In some organisms these also undergo mitotic divisions before
syngamy (Figure 2.2).
Which of these steps came first,according to Maynard Smith and Szathmáry
(1995)? One of the surviving ancient protist lineages, Barbulanympha, which
lives inside the guts of insects, has a cycle that involves endomitosis (gain of
diploid state by copying of the haploid chromosomes) instead of syngamy.
This led Cleveland (1947) to suggest that the first stage might have been the
acquisition of a life cycle that alternated between a diploid stage acquired via
endomitosis, and a haploid stage via a single one-step reduction division.
Next, according to Maynard Smith and Szathmáry (1995), endomitosis
would be replaced by syngamy. This would leave an otherwise normal one-
step meiosis as seen in many sporozoans (the group to which the malaria
parasite belongs). Crossing over and chromosome doubling followed, giving
a two-step meiosis, and finally anisogamy (Figure 2.3). Let us see how we can
account for one of those steps.
The vast majority of work on the evolution of sex has addressed the advant-
age of crossing over, or recombination. There are two processes that might
have selected for its evolution. The first is that recombination can lower the
genetic load if mutations act synergistically (having two is more than twice as
bad as having one) (Kondrashov 1988). Imagine a distribution of deleterious
mutations at equilibrium in a clonal population. Most organisms have a few,
and a few have many (Figure 2.4). Now imagine that recombination occurs.
The mutations are redistributed among the population, and once, after selec-
tion has acted and equilibrium is achieved, there are fewer mutations
 EVOLUTIONARY COVER-STORIES 17

Ancestral eukaryote
Endomitosis plus one–step meiosis
mitosis?

Barbulanympha Syngamy

Sporozoa Premeiotic doubling

Crossing over?
Chlamydomonas Anisogamy

Multicellular metakaryotes

Fig. 2.3 Possible sequence of steps in the origin of sex, with intermediate states represented by some
extant organisms, after Maynard Smith and Szathmáry (1995).
Number of individuals

(1)

(2)

Number of deleterious mutations

Fig. 2.4 The genetic load in sexual and asexual populations. Recombination can reduce the load of
synergistic mutations (1). In small asexual populations, the genetic load of slightly deleterious
mutations can increase, ratchet-like (2).

(Figure 2.4).This is because recombination in each generation throws together

unlucky individuals with many such mutations, which suffer more severely
than their fellows with only a few because of their synergistic effects. The death
of these individuals purges the population somewhat of the mutations. This
process can work even in an infinite population,and only requires the presence
of synergistic mutations. There is presently little factual evidence for the syn-
ergistic effects of mutations, but theoretically it is a reasonable expectation
(Szathmáry 1993).According to metabolic theory,mutations affecting a meta-
bolic cycle should affect mainly the concentration of chemical intermediates,
18 DISCOVERING EVOLUTIONARY ECOLOGY

not that of end-products. If it is maximal end-product production that is

important, as is likely in small organisms whose fitness depends on fast
growth, then mutations are not synergistic. If it is some optimal balance of
intermediates that is important, as in large long-lived organisms, then muta-
tions will be synergistic and recombination will be favoured. Rather nicely, the
frequency of recombination varies markedly across species, and is most fre-
quent in large, long-lived organisms (Bell and Burt 1987), where we would
most expect the effects of mutations to be synergistic.
The second possible process that might have favoured recombination is
selection for change (directional selection) on polygenic traits (traits
controlled at several loci) (Maynard Smith 1979). Hamilton (1980) most
famously adhered to this hypothesis to explain not only the origin of sex but
also more specifically its maintenance. He regarded co-evolutionary arms
races (see Chapter 11) between hosts and parasites as a likely and widespread
source of such directional selection. This idea has become colloquially
known as the Red Queen theory (Van Valen 1973) after the Lewis Carol
character in ‘Through the Looking Glass’ who had to run as fast as she could
to stay in the same place.
The early protists were certainly not immune from such co-evolutionary
forces, though the selective pressure is much greater on long-lived macro-
scopic organisms with long generation times relative to their parasites, where
the traits under selection for change are those involved with defence and
resistance. Nicely but at the same time frustratingly, this also fits well with the
observation that long-lived organisms have higher rates of recombination.
The frustration is that both Hamilton’s and Kondroshov’s hypotheses make
the same prediction about the frequency of recombination relative to size
and lifespan, so the observation fits but does nothing to narrow the range of
plausible hypotheses. Rather more fortunately, there is independent evi-
dence for the Red Queen hypothesis, which we will examine later in relation
to the maintenance of sex.
Another step in the origin of sex is worth mentioning here. In many single
celled organisms, such as the single celled alga, Chlamydomonas reinhardii,
familiar in many school biology classrooms, the gametes that fuse to form
a diploid alga are of identical size (isogamy). In most sexual species, however,
one gamete of the pair (the egg) is larger and specialized to carry the
organelles. Cosmides and Tooby (1981) and later Hurst and Hamilton
(1992) argued that such specialization has evolved to prevent conflict
between organelles from different parents. Many organelles, such as mito-
chondria and chloroplasts, contain their own DNA, (in the latter cases they
were originally independent prokaryotic organisms). Such replicating
entities should presumably be selected in the short term to produce copies
of themselves at the expense of competing entities, and this is likely to be
 EVOLUTIONARY COVER-STORIES 19

detrimental to the eukaryote cell as a whole. The potential problem is

apparently real, for even in isogamous protists, uniparental inheritance of
the organelles occurs, and in Chlamydomonas is apparently controlled by
nuclear genes (central control, analogous to a police force in human society).
Hurst and Hamilton argue that uniparental inheritance is possible only if
there are two mating types and no more, for otherwise there is the danger of
offspring lacking organelles entirely. Thus, conflict between organelles has,
they claim, led to the origin of two (not three nor some other number) sexes.
In fact, some protists exchange genetic information without cytoplasmic
exchange (a process known as conjugation). In these cases there is no
possibility of organelle competition, and multiple ‘sex’ or ‘incompatability’
types are known.

2.2 The maintenance of sex

Having drawn a scenario for the origins of sex, and thought of ways in which
the various steps might be selected for, we are left with explaining its persist-
ence and prevalence. It is clear that those lineages that evolved sex have gone
on to diversify into millions of species that have largely retained sex. In some,
however, clonal reproduction has secondarily arisen (these are normally
called parthenogens). Is the commonness of sex due to the rarity of reversal
to the clonal state? It is undoubtedly part of the answer. Some animals and
plants, for example, have no known parthenogens, despite being species rich
and well known. They include birds, mammals, and gymnosperms (conifers
and their kin). In each of these three groups, mechanisms are known that are
likely to have prevented reversal to the clonal state.
In birds, parthenogenetic individuals sometimes arise but these fail to per-
sist as unisexual lines. The reason is that in birds the female is the hetero-
gametic sex (with a ‘Z’ and a ‘W’ chromosome). During parthenogenesis,
chromosome doubling occurs as normal, but there is only one subsequent
division, leaving diploid eggs. Many of these will contain two Z chromosomes,
leading to male production, and hence maintaining both sexes (Crews 1994).
This is a big pity for short-term poultry production! If in birds, females were
the homogametic sex (two X chromosomes), all parthenogenetic offspring
would also have to be female. Given that in mammals females are the
homogametic sex, one would imagine that cattle, sheep, and pig producers
would have had more luck, but here prevention of parthenogenesis comes
from another source.
In mammals the phenomenon that prevents parthenogenesis is known as
genomic imprinting. The phenomenon was first noticed when researchers
tried but failed to get embryos to develop by fusion of two egg nuclei. The
20 DISCOVERING EVOLUTIONARY ECOLOGY

reason is that early zygote development requires genes from both parents
that have different levels of activation: if the genes come from the same par-
ent the activation levels are all wrong. We will encounter this phenomenon
again in Chapter 7, but, briefly, the reason is likely to be that in mammals par-
ents conflict over what level of gene activation in the zygote is preferred, set-
ting up an offensive/defensive gene activation war (e.g. Burt and Trivers
1998).
In gymnosperms the mystery is more clear-cut: although the egg provides
most of the organelles for the zygote, it is the pollen that provides the chloro-
plasts. Unisexual gymnosperms would lack the ability to photosynthesize.
Provision of an essential organelle is also the likely reason for general rarity
of animal parthenogenesis, and also for the strange forms it takes when pre-
sent. In many animals, the sperm provides a centriole to the zygote. In many
parthenogenetic animals, including most clonal vertebrates, parthenogene-
sis is ‘sperm-dependent’: the eggs need to be ‘fertilized’ by sperm of another
species for successful development, though the sperm genome never makes
it into the next generation (Beukeboom and Vrijenhoek 1998).
Though obstacles to reversal are one important reason why sex is still
prevalent, it is not the whole story. Many plants, for example, could easily
persist from generation to generation in a clonal state by vegetative repro-
duction. Yet, where they exist, wholly clonal plants, and parthenogenetic
organisms in general, appear to be relatively recent phenomena. For example,
most are isolated species in genera that are predominantly sexual. In a few
cases, genetic and other techniques have been used to estimate the actual ages
of clones and their sexual parents. In the case of the genus Poeciliopsis,
a guppy fish that inhabits streams in southern United States and Mexico,
sexual species are up to 3 My old, but their sperm-dependent parthenogens
are mostly less than a few thousand (Figure 2.5). Interestingly, the oldest
surviving clone is also the only one where the sperm genome finds

Fig. 2.5 The unisexual fish Poeciliopsis 2monacha-lucida. This species is a sperm-dependent
parthenogen, meaning it relies on the sperm of another species for reproduction. However,
the genome of the sperm donor is not expressed in the offspring which are genetically identical
to their mother. This species suffers more from parasites than its sexual relatives, hence
supports the ‘Red Queen’ hypothesis for the maintenance of sex. Photo courtesy of Bob
Vrijenhoek.
 EVOLUTIONARY COVER-STORIES 21

expression in the fish phenotype,the so called ‘hybridogen’(it is later excluded

from gamete production though). Only recent originations persist probably
because earlier originations have largely gone extinct while their sexual rela-
tives have not (Vrijenhoek 1994). There are a relatively few higher taxa that
have persisted for millions of years, popularly known as the ‘ancient asexu-
als’. Our theories for why parthenogens might have high extinction rates
should ideally account for these exceptions.
The problem of the relative persistence of sexual species versus asexual
clones has traditionally been thought of in terms of the so-called ‘two-fold
cost’ of sex. Female parthenogens can increase in number at twice the rate
of their sexual counterparts because they do not give birth to males, which
cannot themselves give birth (Figure 2.6). Sexual organisms may even suffer
a number of additional costs, such as finding a mate. All of this suggests
that sexual organisms should be the ones with the higher extinction rates. It
was this problem that eventually became the focus of Hamilton’s research:
why was it that clones, once they arose, did not quickly send their sexual
parents extinct? Hamilton became convinced that the answer lay with one
of the short-term advantages of recombination, in particular the Red
Queen hypothesis. There is evidence to support his contention. First, some

Generation 1 Generation 2 Generation 3

Unisexual

Sexual

Fig. 2.6 The two-fold cost of sex. Here a species is shown which always has two offspring per
generation, except that in the sexual form half of these are male, which mate with the females
(dotted arrows), while the unisexual form can have offspring without mating. By the third
generation, there are four females in the unisexual line but only one female in the sexual line.
22 DISCOVERING EVOLUTIONARY ECOLOGY

of the so-called ancient asexuals are obligate mutualists. These include some
mycorrhizal fungi, which inhabit the roots of plants, and the fungi that
are the food for leaf-cutter ants. In contrast to parasites, which experience
directional selection from their hosts, mutualists would be expected to
experience stabilizing selection to aid the efficiency of the interaction. This,
as Maynard Smith showed, can select for an absence of recombination. In
addition, we have direct measures from some asexual clones that they carry
higher parasite loads than their sexual counterparts. This is true, for example,
of the Poeciliopsis clones (Vrijenhoek 1994, Figure 2.5).
Clones, of course, may also suffer a higher load of deleterious mutations,
as Kondroshov showed, contributing to their extinction rate. In Poeciliopsis,
there is also evidence for this. By a clever series of crosses, it has been possible
to express the hybridogen genes that are normally dominated by those of the
sperm donor. These show several developmental defects compared with the
parental or hybrid genotypes.
In addition to Kondrashov’s mechanism, an alternative long-term mech-
anism can account for this, known as Müller’s ratchet. Müller’s ratchet is in
some ways more general than Kondrashov’s theory, for it does not rely on
synergistic mutations; mutations merely have to be of small effect. It also
only works in small populations, but that is probably a fairly general phe-
nomenon. When mutations are of small effect, the distribution of those
mutations at equilibrium among individuals will be approximately bell-
shaped: few will be completely free of them, most will have a few, and only a
few will have a lot (Figure 2.4). In a small population, however, the categories
of individual with no mutations is easily lost by chance events, even if they
are the most fit. In a clonal population, these can never be recovered. Of
course, the category of individuals with most mutations can also be lost, but
those are replaced by subsequent mutation.
Overall then, in a clonal population, the load of slightly deleterious muta-
tions continually cranks up, ratchet-like. In a sexual population, however,
recombination recreates individuals free of mutation, and the genetic load
remains stable despite stochastic loss of the fittest individuals (Figure 2.4).
Note that because of the long-term nature of the mechanism, it cannot be
invoked to explain the origination of recombination, merely its persistence
relative to clonal reproduction.
In general, should we be searching for long- or short-term mechanisms to
explain the persistence of sex? Hamilton was convinced that the latter was
necessary largely because of competition between clone and sexual parent.
Is there in fact evidence for this? Do clones actually displace their sexual
parents, and is there a risk of them being sent entirely extinct? In general, we
might expect, in the absence of short-term advantage, that the clone would
successfully displace the parent from part of its former range, but that
 EVOLUTIONARY COVER-STORIES 23

existing genetic diversity among the parent would allow the parent to persist in
parts of its range to which the clone is less adapted. In Poeciliopsis, the diversity
of species is inversely related to the diversity of clones, suggesting some com-
petitive exclusion. In many plants, such as the dandelions, parthenogenetic
forms also appear prevalent in some environments, especially at high altitudes
and latitudes. This in turn suggests that the costs and benefits of each form of
reproduction vary spatially, and also that short-term advantages of sex are not
always initially enough to compensate for any costs.
To summarize, the maintenance of sex has been influenced by the follow-
ing processes: first, constraints to reversal among a number of lineages,
particularly animals. Second, clones can successfully displace their sexual
competitors from some environments, suggesting a severe cost to sex.
However, directional selective forces can compensate for these costs in some
environments, and, combined with increased genetic loads, send most
clones extinct relatively rapidly.
How does the evolution of sex compare with the other major transitions?
Maynard Smith and Szathmáry identify several common features of the
transitions (Table 2.1). Of these, sex is very illustrative. Entities have com-
bined together, through the evolution of syngamy, to form a sexual popula-
tion from previously independent clones. Further, reversal of sex to the
clonal state is sometimes difficult because sex has developed a complex
machinery for reproduction. Sex has probably led to conflict between
entities, such as between organelles in the parent gametes for representation

Table 2.1 The common features of the major transitions in evolution

Feature Molecules in Chromosomes DNA Eukaryotes Multicellular Sex Social Language

compartments ⫹ life colonies
protein

Entities Yes Yes No Yes Yes Yes Yes No

combine?
New ways No No Yes No No Yes No Yes
of
information
transmission?
Reversal Yes Yes Yes Yes Yes Yes Yes ?
difficult?
Conflict Yes Yes No Yes Yes Yes Yes No
between
entities?
Mechanisms Yes Yes No Yes Yes Yes Yes No
to prevent
conflict?
Division of Yes Yes Yes Yes Yes Yes Yes No
labour?
24 DISCOVERING EVOLUTIONARY ECOLOGY

in the zygote. To solve this, mechanisms, such as uniparental inheritance

have evolved. Sex has led to division of labour among the combining entities,
such as male and female gametes. In the evolution of sex, there has been no
new method of transmitting information developed. That has only occurred
in three of the transitions: in the origin of DNA and protein from RNA, in the
origin of language, and in the origin of epigenesis (gene activation) in the
origin of multicellular life.
In this chapter we have been postulating processes that have caused
change within lineages through natural selection, a theme that will continue in
the next several chapters. I cannot help but end here with a well-known quote
from Aldous Huxley who once said that ‘an intellectual is a person who has
discovered something more interesting than sex’. There is a certain irony in
this quote for evolutionary ecologists. Many of them would argue, on purely
intellectual grounds, that there is in fact nothing more interesting than sex,
full stop.

2.3 Further reading

Beginners should try Maynard Smith and Szathmáry (1999) first, followed
by Szathmáry and Maynard Smith (1995). Maynard Smith and Szathmáry
(1995) is quite heavy going, but also more complete. Useful works on the
evolution of sex include Maynard Smith (1984), Bell (1982), Stearns (1987).
Two recent special issues cover the subject: in Science (25 September 1998,
vol. 281: 1979–2008) and Trends in Ecology and Evolution 1996 vol. 11.
Poeciliopsis is reviewed by Vrijenhoek (1994).
3 Brave new worlds

The eight novel ways of transmitting information, outlined in the previous

chapter as ‘major transitions’, by no means exhaust evolution’s extraordinary
feats. Over the history of life, evolutionary events have also radically changed
the characteristics of the biosphere. As in the previous chapter then, which
considered how evolutionary changes increased the complexity of organ-
isms, this chapter will consider how evolution has increased the complexity
of planetary ecology. Four things generally indicate that major ecological
changes have occurred in the past, identifiable, for example, from the fossil
record (Vermeij 1995; Kanygin 2001): First, changes in species richness.
Second, organisms living in new places. Third, ecosystems with new
functional groups. Fourth, new geochemical cycles. For present purposes
we are only interested in such changes that are linked with evolutionary
events, as most (but not all) of them are by definition. The changes need col-
lectively to create the essential and complex features of modern planetary
ecology, which are worth briefly describing.
Species richness is currently (recent extinctions excepted) higher than
ever before (Signor 1990; Sepkoski 1999). About half of described macro-
scopic species are insects, a quarter green plants, and most of the rest sundry
invertebrates (Southwood 1978). Life exists and flourishes nearly every-
where on the face of the globe; in the marine, freshwater and terrestrial
realms, anoxic mud, animal guts, hypersaline lakes, volcanic springs, desert
dunes, within frozen antarctic rock and in rocks hundreds of metres below
ground. Some organisms even live an essentially atmospheric existence,
feeding on other flying organisms.
The most productive and diverse ecosystems on the planet are terrestrial.
They comprise: green plant producers; a diverse assemblage of vertebrate
and insect herbivores; predators, parasites and parasitoids, and a soil fauna
of scavengers, detritivores, decomposers, and nutrient cycling bacteria. In
the marine realm the most productive and species-rich ecosystems are those
with sessile producers, such as coral reefs, kelp forests, and seagrass beds. The
open ocean is a comparative desert, but consists of pelagic phytoplankton,
zooplankton, and macroscopic predators. A benthic fauna consists of filter
feeders, scavengers, and predators. The biosphere’s energy comes, almost
exclusively, from light captured by plants, and its carbon source is atmospheric
26 DISCOVERING EVOLUTIONARY ECOLOGY

Ma
0
CENOZOIC
65
CRETACEOUS
MESOZOIC

142
JURASSIC Possible date Novelty Why a ‘major transition’?
(Ma)
205
140 Flowers New trophic relationships, increase
TRIASSIC
250 in species richness
PERMIAN 350–50 Flight Spatial expansion of species,
increase in species richness
292
350 Phytophagy New trophic niches, increase in
CARBONIFEROUS species richness, changes to
354 biogeochemical cycles
PALAEOZOIC

DEVONIAN 350 Trees Increase in species richness,

417 changes to biogeochemical cycles
SILURIAN 430–390 Terrestrialization Spatial expansion of species.
440 535 Major animal Increase in species richness, new
lineages trophic niches, change to
ORDOVICIAN
biogeochemical cycles,
495
3500–2200 Phagotrophy New trophic niche
CAMBRIAN
3500–2500 Aerobic respiration Change to biogeochemical cycles
535
3500 Oxygenic Change to biogeochemical cycles
PROTEROZOIC
photosynthesis
Oxic world
3500 Photosynthesis Change to biogeochemical cycles
2500
3500 Heterotrophy New trophic niche
ARCHEAN
Life begins
3500
HADEAN
Earth forms
4500

Fig. 3.1 Some major transitions in ecology resulting from evolutionary novelty, and when they occurred.

carbon dioxide, which is eventually returned by (high energy-yielding) aer-

obic respiration. Where, then, did all this come from and what evolutionary
novelties helped it get there? Below I tentatively identify eleven major
changes that take us from the origin of life to an ecologically modern planet
(Figure 3.1).

3.1 Evolution of the biosphere: a brief history

The origin of the biosphere and of earth’s ecology occurred between 3.8 and
3.5 billion years ago. Both autotrophic and heterotrophic origins have been
proposed. Previously the heterotrophic use of organic molecules synthesized
in the pre-biotic broth was a popular idea (see Lazcano and Miller 1999). More
recently, autotrophic theories have re-emerged, for two reasons: first, early cell
membranes would probably have lacked sufficient permeability to transport
large molecules. Second, it is now realized that metabolic cycles that grow and
reinforce themselves can emerge spontaneously. If such metabolic cycles
 BRAVE NEW WORLDS 27

occurred in the pre-biotic world, an autotrophic ancestral metabolism would

by definition result (Wächtershauser 1988, 1990). Because photosynthesis
today requires a more complex biochemistry, an ecosystem consisting initially
of only chemo-autotrophs is likely. It is also likely that a second trophic level
would quickly be added, consuming waste products (and dead remains) of
these producers. There is evidence also that photosynthesis may nonetheless
have been a very early acquisition. Early life probably lacked protein enzymes
to catalyse reactions, using instead RNA. It turns out that chlorophyll synthe-
sis involves (non-intuitively) molecules bound to RNA, a likely relic of the
ancient involvement of RNA in catalysis (Benner et al. 1989). The origin of
photosynthesis provided a new energy source, light, for the world’s ecosystems
that would have massively increased its potential productivity.
The early photosynthesizers probably used hydrogen sulphide or
molecular hydrogen as their source of electrons, rather than water (Xiong
et al. 2000). Using water as an electron source releases molecular oxygen. We
know that cyanobacteria, which today are the predominant oxygenic
prokaryotes, were among the earliest prokaryotes, at least 3.5 billion years
old. However, it was not until about 2.5–2 billion years ago that a great
increase in atmospheric oxygen occurred, marking the end of the Archean
era (Figure 3.1). Until then, some process must have removed the oxygen
produced by cyanobacteria. One possible process is aerobic respiration
(Towe 1990). Aerobic respiration is many times more energy-yielding
than any of its anaerobic alternatives, and although at this stage there were
certainly no food chains as we recognize them today, aerobic respiration
made possible the future advent of the higher trophic levels with substantial
biomass (Fenchel and Finlay 1995). By this very early time, 3.5 billion years
ago, all basic bioenergetic processes had probably evolved, many of them
several times, and the biogeochemical cycling of carbon, nitrogen, and
sulphur was established as we know it today.
Although the earliest recognizable eukaryote fossils date from the time of
transition to oxic atmosphere, 2100 Ma, the lineage from which modern
eukaryotes are derived is a very deep evolutionary branch, and probably
dates right back to the time of the earliest evidence for life, 3.5 Ga.
Somewhere in this interval, one of our ancestral bacterial lineages (biomole-
cular evidence suggests it resembled an archaebacterium) developed a
cytoskeleton and lost its cell wall. With this came the ability to engulf large
organic particles. The first eukaryote lineages were doubtless anaerobic, a
fact indicated by the many extant anaerobic eukayotes belonging to basal
branches of the eukaryotic tree. These early predators represented a trophic
interaction quite different to anything in the prokaryotic world, in which
organic material had to pass through cell walls and membranes. However,
only with the advent of mitochondria, probably at the Proterozoic boundary
28 DISCOVERING EVOLUTIONARY ECOLOGY

(Figure 3.1), could these predators fully reap photosynthetic productivity by

integrating with the aerobic world.
There existed now a period of about 1 Ga that is relatively featureless in
terms of historical evidence, and indeed may have been relatively stable in
ecological terms, until 535 Ma. The next 100 My period, known as the
Cambrian explosion and subsequent Ordovican radiation, is one of endur-
ing interest for biologists (Figure 3.1). Over the next 100 My appeared the
major animal lineages, including the first benthic and pelagic macropreda-
tors and the first animals capable of burrowing more than a few millimetres
into sediments.
The consequences were various and significant. The Earth became much
more species rich. Use of skeletal materials based on calcium, phosphorus,
and silica led to greater control of these minerals by organisms, as opposed to
by inorganic processes. Disturbance of sediments by burrowers recovered
carbon and other nutrients from sediments for recycling rather than burial.
All these new animals produced masses of faeces. Faeces dropped to the
ocean floor rather than remaining in the water column, and in doing so con-
sumed less dissolved oxygen. Flow of oxygen from the surface waters to the
ocean floor would have been facilitated, as suggested by geological evidence
(Logan et al. 1995). This may itself have contributed to the Cambrian radia-
tion by facilitating skeletal formation, large bodies, and active metabolisms.
Macropredators might have stimulated novel defences in prey, which them-
selves would be a cause of selection on predators. Such ‘co-evolution’ may
have been a stimulus for diversification in lifestyle and structure.
By the end Ordovician, marine ecosystems would have looked pretty
modern. Soon after, however, multicellular organisms then began to form
complex terrestrial ecosystems. At the time, colonization of the land was
notable primarily for an expansion of earth ecospace. However, eventually,
the progressive evolution of terrestrial communities led to major alterations
of biogeochemical cycles, and terrestrial domination of global biodiversity
and production. The earliest land plants were relatively small in stature.
About 380 Ma the earliest trees appeared and by 350 Ma, forests composed
of horsetails, clubmosses, ferns, progymnosperms, and seed plants had a
widespread global distribution and covered a number of clearly distinguish-
able biomes. The consequences for the biosphere appear to have been
immense. Global productivity probably soured to unprecedented levels.
Coal was deposited in massive amounts never again attained. Global carbon
dioxide levels dropped to 10% of their previous levels in about 50 My,
eventually resting about their present level. This set the scene for subsequent
periods of significant global cooling.
It is possible that the earliest terrestrial plants were relatively free from
natural enemies, such as herbivores. By mid-Carboniferous there was
 BRAVE NEW WORLDS 29

abundant evidence that the onslaught had begun. Insects with characteristic
mouthparts, and fossil leaves with evidence of bite marks, along with
vertebrates with teeth designed for chewing all suggest that plants had begun
their war against animal attack that continues today. This was a significant
new trophic level, for now some of the plant productivity was available to
other organisms. Terrestrial ecosystems had come of age.
Flight has probably evolved four times in the history of life: in insects,
pterosaurs, birds, and bats. Most biologists would agree that flight has had
major ecological repercussions. First, the atmosphere could at last be
properly utilized. While flying species are variably adapted to an aerial
existence, a few birds, such as the swifts, live the vast majority of their (often
considerable) lives on the wing. Flight may also have contributed signific-
antly to global diversity. Bats, birds, and winged insects are all species rich
(see de Queiroz 1998). These organisms are likely to have contributed to
diversification of other species, such as the plants they pollinate and disperse.
The evolution of flowering plants is our final major transition. Angiosperms
are the most species-rich division of plants today, they dominate global
productivity, and their origins coincided with a rise in plant diversification
that shows no signs of abating. Effects on insect diversification are also
detectable and non-trivial (Farrell 1998). We have now arrived at an
essentially modern ecology. How and why did these changes occur and why
have they been retained? Let’s look at an example, the evolution of flight in
birds, insects, bats, and pterosaurs.

3.2 The evolution of animal flight: understanding a major

transition in ecology
There is now little doubt among most biologists that birds derive from a
group of theropod dinosaurs. The theropods were a bipedal carnivorous
group that share many anatomical features with birds. A series of recent
fossils, most notably from Liaoning province of China and described by
Xu Xing and colleagues, include therapods with epidermal feather-like
structures that we might collectively refer to as ‘fuzz’. They were pre-adapted
for flight through a fast cursorial predatory lifestyle. This resulted in a
shortening and stiffening of the tail, reduction in the size of the midbody,
lengthening of the raptorial arms, swivel-wrist joint, light hollow bones,
and a reduction in body size (Sereno 1999).
What subsequently happened? There are several ecological scenarios. The
arboreal hypothesis states that birds evolved from ancestors that lived
in trees and gained the ability to glide from tree to tree. Arboreal gliding
30 DISCOVERING EVOLUTIONARY ECOLOGY

organisms are common today, and in general provide a plausible intermedi-

ate stage to flight because the energy for lift is supplied by gravity. The
discovery of Microraptor ghui, with its apparently four gliding limbs (Xu
et al. 2003), has recently renewed interest in this scenario. However,
theropods were primarily bipedal ground dwelling runners and this has also
focussed attention on a possible cursorial origin. By flapping their forearms
as they ran, rather like a swan taking-off from water, therapods could have
increased their running speed by taking weight off the hind legs, allowing
the hind legs to provide more forward thrust. In this way the wings would
gradually take over from the hind legs until both lift and forward thrust
could be provided by the wings alone (Burgers and Chiappe 1999).
Other more complex scenarios have been also proposed. Garner and
co-workers (1999) have suggested the ‘Pouncing Proavis’ hypothesis. They
envisage therapods dropping onto prey from a perch, in the manner of
modern owls and buzzards. The forearms could have assisted balance and
the feather surface would develop initially to control the drop, rather than
provide lift. Selection for greater horizontal range of drop (a swoop, involv-
ing lift generation) could then have transformed the role of the wing. Like the
cursorial hypothesis this retains the functional distinctiveness of the fore-
and hindlegs since both have separate roles, which is less likely in an arboreal
origin. Another possible scenario (Dial 2003) is that wing flapping devel-
oped to assist therapods to climb to elevated refuges, such as trees or bolders,
as it is still used today in birds, such as quails and chickens, even in their
flightless chicks.
If birds developed flight from bipedal and basically ground-dwelling
ancestors, what of insects? There is intriguing evidence that insect wings may
have developed from leg segments supporting gills that originally developed
along the length of the body. Some fossil mayfly nymphs possessed these,
and the thoracic ones would then be homologous with modern wings
(Kukalova-Peck 1978). In modern insects, genes are present that switch off
the development of these structures in all but the wing segments, but the
potential to form wings is present in every segment (Carroll et al. 1995). But
how could an aquatic gill come to function as a wing? Some mayflies and
stoneflies use their wings in a unique way (Figure 3.2): to sail or skim across
the water surface to reach land after emerging as adults onto the water surface
(Marden and Kramer 1995). Close to the water’s surface, wing beating provides
more power because the air is compressed between the wing and the water.
The aquatic skimming origin therefore postulates ancestral apterygotes and
pterygotes having aquatic larvae with moveable gills, and air-breathing
adults. Retaining gills through to the adult stage aided sailing to land, and
selected for larger but also fewer structures to aid directional stability.
Skimming developed further speed and control via flapping, and eventually
 BRAVE NEW WORLDS 31

Fig. 3.2 This male stonefly, Allocapnia vivipara, is flightless but has raised its short wings to sail across
the water surface to dry land. Flight in insects may have originally evolved through such a
stage. Photo courtesy of Jim Marden.

adults were created that were fully capable of flight.The hypothesis provides an
explanation for the somewhat mysterious observation that while the primitive
wingless insects and most derived insects are terrestrial, the extant primitive
winged insects (mayflies and dragonflies) all retain aquatic larvae.
The origins of bats and pterosaurs are much more obscure. Neither have
clearly identifiable fossil ancestors, nor do transitional fossil forms exist. In
both taxa the flight membrane and lack of cursorial hind legs are much more
suggestive of an arboreal gliding ancestor than for the birds and insects
(Figure 3.3). Some candidate pterosaur ancestors may have been bipedal
however, and bipedality was certainly common among the stem archosaur
groups. In the bats, most scenarios envisage a nocturnal, arboreal, and
insectivorous ancestor for the following reasons: the hind limbs help support
the flight membrane (Figure 3.3), making a cursorial ancestor very unlikely;
all bats are nocturnal hence that is a likely ancestral state; and the ancestral
eutherians were doubtless insectivorous. Recently, Speakman (2001) has
proposed an alternative hypothesis: that of an arboreal, diurnal, frugivorous
ancestor. The advantage of this hypothesis is that one can imagine the ances-
tral bat leaping from branch to branch using vision effectively for foraging.
Some then developed insectivory, and all were forced into nocturnality to
escape raptorial birds, after which fruit bats specialized the visual system,
and the microbats the echolocation system. There is agreement that the set-
ting was arboreal and gliding, but beyond this many scenarios are possible.
The origin arguments presented above are summarized in Table 3.1. It is
exciting that of the three extant flying taxa, three completely different evolu-
tionary scenarios may have played out.
32 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 3.3 A lesser mouse-eared bat, Myotis blythii, taking to the air. Note the short hind limbs and
flight membrane stretching from the forelimbs to the hind limbs. This is very un-bird-like and
reminiscent of a quadrupedal glider. No bats have become flightless. Photo courtesy of John
Altringham.

So we can imagine plausible scenarios for how these origins may have
happened. Why to these organisms though, and why at those moments in
time? There are dozens of gliding animals in today’s forests: why have they
not all developed powered flight? Have they simply not hit on the necessary
mutations to transform a gliding animal into one with powered flight? It is
unlikely. Our evolutionary understanding of the transitions involved is that
they have been of a continuous nature, with small change building upon
small change. This is well illustrated by the fossil record for bird evolution.
None of the changes involved can be seen as particularly remarkable.
Furthermore, birds, insects, and bats all have a different flight apparatus sug-
gesting that selection can work through multiple routes. More likely then,
specific external influences may be necessary. Recently, Dudley (2000) has
championed the view that increases in atmospheric oxygen concentration
may have facilitated the origins of powered flight. The Late Carboniferous
period, characterized by the first pterygote insect fossils, represented the
historical peak in Earth’s oxygen concentration, about double that of today.
The Mesozoic era, characterizing the origins of pterosaurs and birds, was a
period of increasing oxygen concentration, reaching a secondary peak in the
Late Cretaceous, reducing somewhat since. This secondary peak coincides
 BRAVE NEW WORLDS 33

Table 3.1 Hypotheses on the origins of flight

Group Who What Where did it Why did it Controversies

changed? changed? change? change?

Pterygote Apterygotes Larval gills Newly Improved No consensus

insects with aquatic derived from emerged speed over on the origins
larvae pleural adults sailing the water of wing
structures or skimming surface structures, or
became wing over the on aquatic
water surface apterygote
ancestors
Pterosaurs Unknown Gliding Arboreal Improved Absence of
basal membrane setting distance and any fossil
archosaur supported by control of ancestors
a finger glide makes this an
became wing open question
Birds Feathered Feathered Cursorial Flapping Primarily
dromeosaurs forearm setting aids running about the
became a speed by ecological
wing providing setting and
life reasons for
change
Bats Unknown Gliding Arboreal Improved Absence of
basal membrane setting distance and any fossil
eutherian supported by control of ancestors
fingers glide makes this an
became wing open question

with the origins of flight in vertebrates. High oxygen concentrations would

have two important effects: first, they would have increased the density of the
air, and the flight surface would thus provide more lift. Second, they would
increase metabolic capacity and hence provide more power per effort. Since
flight is energetically expensive, slight increases in power and lift might have
been sufficient to turn net cost into net benefit. Rather interestingly, changes
in oxygen concentration may have been stimulated by other evolutionary
transitions, such as the increase in productivity due to terrestrialization.
Recently, Lenton et al. (2004) have suggested that a chain of such events
might have occurred in the history of life, with evolutionary changes giving
rise to environmental changes, which in turn give rise to evolutionary
changes.

3.3 The maintenance and ecological effects of flight

Once flight originated, for its ecological effects to be expressed it had to be
maintained. Loss of flight is an interesting phenomenon, for it has occurred
very frequently in insects and birds, and not at all in bats (nor probably in
34 DISCOVERING EVOLUTIONARY ECOLOGY

pterosaurs). The obvious difference between these two groups of organisms

is that the former (insects and birds) retained a functional distinction
between the flight apparatus and the legs: they can both walk or run without
using their wings. In pterosaurs and bats this is not the case and both groups
would be relatively ineffective on the ground, hindering the transition to a
terrestrial lifestyle again. In birds and insects, loss of flight could mean a
reallocation of energy away from the flight apparatus and increases in
reproductive expenditure (Roff 1990, 1994). They lost much of course,
and in birds loss of flight is only viable under special circumstances. It
has happened mostly in a few taxonomic groups (rails notably) and under
special ecological circumstances (notably on islands, see Figure 13.3) (Roff
1994). There are three likely reasons for the latter: first, an absence of land
predators that makes escape (and especially nesting off the ground) less
important. Second, on islands, high dispersal tendencies might increase
the risk of mortality through loss of individuals at sea. Third, a less active
metabolism might be very advantageous in surviving long periods of food
shortage on islands, where birds cannot simply move elsewhere.
Insects have lost their powers of flight many more times than birds, and
they have done so in a variety of ecological circumstances (at least once in
nearly all major habitats). There are many flightless island insects, no
doubt many for the same reasons as birds, but there are also many flightless
insects in other habitats. Loss of flight is very rare in freshwater insects. This
is unsurprising given the ephemeral nature of many freshwater habitats,
most of which stand a good chance of temporary or permanent drought.
Conversely, flightlessness is very common among parasitic insects, particu-
larly of vertebrates (Figure 3.4). In fact only two large radiations of second-
arily flightless insects have occurred: among the fleas and the lice, which
primarily use mammals and birds as hosts. These do not require flight for
dispersal to new hosts. Wing loss may also bring a particular advantage in
facilitating movement within the fur and feathers of their hosts. Given that
loss of flight is apparently so easy in insects, one may wonder why only two
large radiations of flightless insects have occurred. One possibility is that
speciation is frequently associated with niche shifts that would favour the
reversal of winglessness again. Another part of the answer may be that, in
general, flightlessness leads to poorer net rates of cladogenesis. Both extinc-
tion risk and speciation probability may be affected and both these subjects
will be the focus of later chapters.
The maintenance of flight enabled it to become a major transition. Why
did this transition have ecological repercussions? I have argued above that
these were manifested in three main ways: first, flight in itself represents
an occupation of ecospace (the atmosphere) previously unoccupied, just as
terrestrialization does. This requires no special explanation: the innovation
 BRAVE NEW WORLDS 35

Fig. 3.4 A body louse, Campanulotes compar, 1 mm long, which infests feral pigeons, Columba livia.
Lice and fleas represent some of the few major radiations of insects that have lost their wings.
Photo courtesy of Sarah Bush and E. King.

and the ecological change are synonymous. Second, flight probably also
increased the species richness of clades that evolved it (de Queiroz 1998). It
is interesting to speculate why this might have been. Dispersal is often linked
with the speciation process because it can transport organisms to isolated
areas where they can differentiate from their ancestors (Chapter 1). Dispersal
may also inhibit extinction by allowing areas in which local extinction has
occurred to be recolonized. Flight may also have had interesting repercus-
sions on life history (see Chapter 4). Birds, for example, have considerably
extended lifespans compared with mammals of the same mass. Bats also have
much longer lifespans than other mammals of the same mass. It does not
necessarily follow that an increase in lifespan will reduce the extinction risk of
species, but it is possible. This theme is investigated in more detail in Chapter
14. Of course, reduction in mortality, such as from predators, may have
helped select for powered flight in the first place. Third, flying organisms per-
form important ecological roles that may have stimulated the evolution of
other groups, the most obvious case being plant–pollinator interactions.
Families of angiosperm with animal pollination are significantly more
36 DISCOVERING EVOLUTIONARY ECOLOGY

species rich than those without animal pollination (Ricklefs and Renner
2000). Insects, bats, and birds are the major pollinating animals and they can
also fly. Their dispersal abilities give them much greater potential to visit
other flowers and promote outcrossing.
The evolution of flight then, illustrates well the challenges, but also the
rewards, of understanding a major transition in ecology. We have to consider
origins, and in particular, scenarios of progressive selective advantage in the
face of small continuous changes. Those are available for birds and insects,
and to a lesser extent for bats and pterosaurs, and the scenarios are diverse.
We must also explain the timing of the events and what stimulated them.
Invoking high metabolic requirements for flight and periods of high oxygen
concentration may explain that, implying that changes in the environment
were critical, perhaps stimulated by previous evolutionary novelties. We also
have to consider the frequency of loss of the character and its relative effects
on speciation and extinction rates. Losses of the character are few in some
groups, where the flight mechanism is at odds with terrestrial locomotion,
and common in others where it is not. Where it is common, losses do not
seem to confer any consistent benefit in terms of increased speciation or
reduced extinction rates. Finally, we need to explain why the ecological
effects of the transition have been great. Some of these themes may also be
found in some of the other transitions, and it will be exciting to discover if a
consensus on this emerges in the coming years of the sort that Maynard
Smith and Szathmáry provided in the previous chapter.

3.4 Further reading

Southwood (2003) provides an introduction to the history of evolutionary
and ecological change. Fenchel and Finlay (1995) and Willis and McElwain
(2002) are useful more specialist sources. Padian and Chiappe (1998) give a
good introduction to bird origins, Thomas and Norberg (1996) summarize
the insect story, and Speakman (2001) bats. Dudley (2000) deals with the
when. Roff (1990) covers many issues of flightlessness.
 4 Traits, invariants, and theories
of everything

The last two chapters were concerned with how natural selection can change
the way a lineage looks (its phenotype). These changes have had profound
effects, first on the complexity of organisms (Chapter 2), and second on the
complexity of the ecology of the planet (Chapter 3). This chapter continues
the theme, but deals with traits that are individually far more ordinary: life
history traits. Nonetheless, they account for much of the diversity in form
across species, thus collectively are of immense interest.
A life history is a description of the major characteristics of an organism
from its birth to its death. The traits are variables that can be measured or
categorized across individuals and that collectively make up the description
(Figure 4.1). In the past, most work on life history evolution proceeded on
a trait-by-trait basis. Each of the traits is associated with a large body of
literature asking the following question: what selective forces have driven the

Indeterminate growth
Determinate growth
Size at maturity
RIP (adult body size)
Body size

Semelparity

Iteroparity

Reproduction

Offspring size

Birth Age at maturity Age at death

Time
Fig. 4.1 A cartoon of a life history, showing size of an individual against time, and various life
history traits highlighted. Semelparity means reproducing only once, iteroparity many times.
Determinate growth is when growth stops at reproductive maturity, indeterminate growth
when it continues.
38 DISCOVERING EVOLUTIONARY ECOLOGY

trait into its present states? Typically, theoretical models are constructed for
each trait, describing the kinds of environment that favour different values of
the trait. The models are then tested by seeing whether their assumptions
and predictions are met in nature. If the answer is yes in both cases, our
model’s assumptions provide us with the understanding we are looking for.
If the answer is no, either the model or the data are inadequate and must be
re-examined.
In general we might expect certain values of a trait to be favoured rather
universally. For example, it is clearly best, in a Darwinian sense, to produce
many offspring, all other things being equal. Happily for those who like
diversity, all other things are not equal: high fecundity must come at a cost to
some other trait of importance to the organism’s fitness, such as offspring
size. This phenomenon is known as a trade-off. The organism is thus faced
with the more complex choice of finding the value of the trait that maximizes
fitness in the face of trade-offs with other traits of importance. In the above
example, evolution would essentially select between organisms that have
large offspring but few of them, and organisms that have many but small off-
spring. Following this, our evolutionary model will generally make some
assumptions about the nature of any trade-offs involved (a constraint), and
about the consequences of each value of the trait for fitness (a currency). In
general, it is assumed that the value of the trait (the strategy) that leads to
greatest fitness in a given environment will be that which is selected for,
and should, all being well, represent that observed in nature (Figure 4.2).
Such models are commonplace in evolutionary ecology and are known as

Optimum

Constraints
Fitness currency

Strategy set

Fig. 4.2 The characteristics of an optimization model. The arrow shows peak predicted fitness and
therefore what would be expected in nature.
 TRAITS, INVARIANTS, AND THEORIES OF EVERYTHING 39

optimization or optimality models (Parker and Maynard Smith 1990). Let us

start by seeing how this approach works.

4.1 Single trait optimization: reproductive lifespan

Some organisms display ‘big bang’ reproduction, after which they die
(semelparous), while others reproduce more consistently over a long
period of time (iteroparous) (Figure 4.1). The question of whether to be
semelparous or iteroparous is a rather simple one because there are only
two strategies to consider. We simply have to calculate, for any given
organism, which should lead to the highest fitness. Charnov and Schaffer
(1973) produced a simple calculation to help, which is derived by asking
which strategy produces the highest rate of population growth (their
fitness currency). They assumed a plant could display either an annual or
a perennial strategy. The fitness of the annual is determined by its own
survival to maturity and fecundity once mature. However, that of a
perennial is determined by both its juvenile survival, annual fecundity,
and adult survival from year-to-year. These are the constraints. If annuals
are to be more fit than perennials, annuals must have higher annual
fecundities than perennials, because they do not survive for more than
one flowering season. In contrast, high adult survival should select for
perennials because this allows them to increase population growth rate,
the fitness currency.
The data provide confirmation of these predictions. Annual plants gener-
ally have higher annual fecundities than their perennial close relatives (gen-
erally between 1.5 and 5 times) (Young 1990). Such fecundity differences are
most apparent between close relatives inhabiting adjacent environments. On
Mount Kenya in Africa, for example, are two species of Lobelia, one semel-
parous and the other iteroparous (Figure 4.3). The semelparous forms have
high fecundity and grow on dry rocky slopes, where adult mortality is very
high. Iteroparous forms have lower annual fecundity and grow on moister
valley bottoms where adult survival is higher (Young 1990).
The development of simple (classical) models like the above has been
repeated for all of the common traits under investigation. But a ‘single trait’
view of the world is also somewhat limiting, for each organism represents a
specific and sometimes characteristic combination of traits. To understand
the organism as a whole we would have to apply and test theory for each of
the life history traits separately. This is clearly an impractical task on a large
scale. Instead, we need an approach that considers, in a single framework, the
whole life history of an organism. That has been the recent challenge in life
history evolution, and it has led to great things.
40 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 4.3 Lobelias growing on the slopes of Mount Kenya, Africa. The two tree-like plants are giant
groundsels, Senecio keniodendron. The tall feathery spikes are Lobelia telekii, a semelparous
species, about 2 m tall. The prickly plant in the right foreground with the shorter flower spike
is another species, Lobelia deckenii ssp.keniensis. This is iteroparous and tends to dominate
in moist valley bottoms (see distance) where adult mortality is low. Photo courtesy of Truman
Young.

4.2 Invariants, combinations, and comparative studies

Researchers studying particular groups of organisms have long known that
certain traits are correlated in very specific ways with other traits across
species. The most famous of these correlations are those in the mammals,
which display the so-called ‘fast–slow continuum’ (Promislow and Harvey
1990). Large-bodied mammals, such as elephants and buffalo, have long
adult lifespans, take a long time to mature, have small litters of large
offspring. In contrast, small-bodied mammals, such as mice and shrews,
have short adult lifespans, mature very rapidly, and have large litters of small
offspring. They therefore ‘live fast and die young’. All mammals fit relatively
neatly somewhere onto this continuum. We will examine how widespread
this pattern is in other groups later on.
The differences just described are known because of comparative studies
(Harvey and Pagel 1991), and much of the rest of this chapter relies on
comparative studies. Comparative studies take variation among different
species as their source of data. This variation has evolutionary origins, and
much of the time the forces causing the variation we see among species have
 TRAITS, INVARIANTS, AND THEORIES OF EVERYTHING 41

Species1 Species2 Species3 Species4

Fig. 4.4 Four species plotted onto an evolutionary tree, or phylogeny, showing their evolutionary
relationships. The comparison between species 1 and species 2 is evolutionarily independent
of the comparison between species 3 and 4. Both may be used as data in a comparative study.

been ecological ones. We can use the variation we see therefore as the results
of some grand experiment in evolution played out over time, and ask what
factors correlate with the variation. We have been using comparative data in
a rather loose sense throughout the previous chapters: for example, we com-
pared the species richness of different cichlid groups with different traits, the
rates of recombination among species of different size, and the fecundities of
annual verses perennial plants. More formally, comparative biologists like to
use their data in some kind of statistical analysis, and the power of such
analyses increases with the quantity of data. Normally, this means finding
many different taxonomic groups that have changed independently of other
groups. To ensure this is the case, comparative biologists often include in
their study a ‘phylogeny’ that describes the relationships between the species
in the dataset, and helps to identify independent evolutionary events
(Figure 4.4).
Comparative studies on mammals have revealed the fast–slow continuum
just described. The more specific details of this continuum are also interest-
ing. First, many of the life history traits are related to body size across species
by a characteristic power function, known as an allometric exponent
(Figure 4.5). The birth rate of a mammal species, for example, is propor-
tional to its body size to the power of ⫺0.25 (thus, birth rate declines with
body size across species). Amazingly, all of the allometric exponents for
mammalian life history traits are close to some multiple of a quarter. This
seems to require explanation. Furthermore, but rather less surprisingly given
what we have just seen, certain traits, when multiplied together have values
that are unrelated to body size. For example, if one trait with the exponent
0.25 is multiplied by another trait with the exponent ⫺0.25, the result is
a number that is not related to the body size of the species, and may be
relatively constant. Eric Charnov, who has done most to bring these facts
42 DISCOVERING EVOLUTIONARY ECOLOGY

Linear plots Log–log plots

0.25
0.25

Log other trait

Other trait 0
0

–0.25 –0.25

Body size Log body size

Fig. 4.5 Allometry. Traits are allometric if they are related to body mass by a power function (linear plot,
numbers indicate the exponent). If both the trait and body mass are logged, the relationship
is a straight line where the slope of the line is equal to the exponent. Both 0.25 and ⫺0.25
exponents are common in mammals. If a trait with an exponent of 0.25 is multiplied by a trait
with an exponent of ⫺0.25, the resulting product is an ‘invariant’ because it is unrelated to
body mass (exponent of zero).

to the attention of other researchers, called these interesting numbers

‘invariants’. In mammals, for example, one invariant is age at maturity times
the adult annual mortality rate (this is in fact, relatively invariant within
many groups of organisms but varies across groups). These strange facts
seem to be telling us something pretty fundamental about life history
evolution; the trouble is in knowing exactly what.
Charnov (1991) was the first to build a theoretical evolutionary model
that would be able to predict these relationships. The basis of the model was
natural selection on age at maturity governed primarily by adult mortality.
One of the studies that most influenced this assumption is a study by Harvey
and Zammuto (1985) using data on mortality rates of mammal species in
their natural habitats. Getting data on that rather tricky variable for enough
species has entailed centuries of man-hours in the field by dedicated
researchers. Rather interestingly, the data showed that adult mortality rates
in the wild strongly predicted where a mammal stood on the fast–slow
continuum. Small-bodied mammals suffered high adult mortality and
large-bodied mammals low mortality. Even more interesting, when body
size was controlled for statistically, adult mortality rate was still related to
most other life history traits: for example, mammals with low adult mortal-
ity for their size also had low fecundity for their size. This seemed to suggest
that adult mortality might be a controlling selective force in shaping the
fast–slow continuum.
In mammals offspring reach independence from their mothers at a
certain size but then grow further until, when mature, they stop growing
(Figure 4.6). Charnov then reasoned that the mortality rates they experienced
after independence would govern when would be the best time to mature: if
 TRAITS, INVARIANTS, AND THEORIES OF EVERYTHING 43

Maturity
Body size

Independence Death
Time

Fig. 4.6 Charnov’s (1991) model. The solid line represents a mammal suffering heavy mortality after
independence. It is best for it to mature early even though this means it can devote fewer
resources to reproduction (vertical arrows). A mammal suffering lower mortality (dotted line)
can afford to wait longer before maturing. As a result, it lives longer, is bigger when mature,
and can devote more resources to reproduction. This model oversimplifies mammalian growth
and metabolism (see Figure 4.8), yet makes accurate predictions across species.

they matured late, the benefit would be achieving a bigger body size, and
consequent larger reproductive resources per unit time. However, this had to
be traded-off against the risk of death before maturity. In general then, a late
maturation time would only be selected if adult mortality was low.
Where do the allometric slopes in the model all come from? Charnov
assumed a within-species allometry of metabolic rate on body size of 0.75.
Before maturity, this metabolism causes growth, after maturity the metabo-
lism is channelled into reproduction. Charnov had to make two more
assumptions: first, about offspring size. He assumed that size at weaning is a
constant proportion of maturation size across species. Finally, he assumed
that the populations of mammals were of constant size, so that juvenile mor-
tality balances fecundity. Density-dependent mortality acts only after the
strategies have been decided on, so this does not contradict the idea of opti-
mization. These assumptions are the origins of all the quarter-power scalings
in the model.
The model can successfully predict the fast–slow continuum. For example,
imagine an organism with low adult mortality. As a result of this, it lives a
long adult life. It is selected to mature late and large because it can afford to
delay reproduction to gain reproductive power due to the low risk of mortal-
ity before maturation (Figure 4.6). Since offspring size at weaning scales with
body size to the power 1 in the model, but reproductive power only 0.75, it
leaves fewer but larger offspring, which have a low risk of mortality prior to
44 DISCOVERING EVOLUTIONARY ECOLOGY

Table 4.1 Scaling exponents, relationships between variables and invariants predicted or
assumed by Charnov (1991) and observed by Purvis and Harvey (1995)

Variables Theory Observed

Assumptions Allometric exponent

Growth rate 0.75 0.82✓
Biomass offspring per year 0.75 0.66✓
Size at weaning to size at maturity 1 0.89✘
Predictions Allometric exponent
Age at maturity 0.25 0.24✓
Annual fecundity ⫺0.25 ⫺0.24✓
Adult mortality rate ⫺0.25 ⫺0.24✓
Juvenile mortality rate ⫺0.25 ⫺0.32✓
Pairs of traits Relationship
Adult mortality, age at maturity Negative Negative✓
Juvenile mortality, age at maturity Negative Negative✓
Juvenile mortality, adult mortality Positive Positive✓
Birth rate, adult mortality Positive Positive✓
Birth rate, juvenile mortality Positive Positive✓
Age at maturity times adult mortality Size invariant Size invariant✓
Age at maturity times juvenile mortality Size invariant Size invariant✓
Age at maturity times fecundity Size invariant Size invariant✓

Note: Ticks indicate that the data match the prediction, crosses indicate a mismatch.

independence. The model manages, via a bit of algebra, to predict nearly all
of the allometric exponents and invariants seen in the data (see Table 4.1).
If we were to accept the model as an approximate description of the
evolutionary forces at work, it would be a major achievement. It might, for
example, explain the human life history (large body size, long adult lifespan,
late maturation,few large offspring) in terms of low adult mortality in our evo-
lutionary past. Other curious facts would be explicable. The longest lifespans
of any mammal are achieved by the bowhead whale which typically lives to
over one hundred years, and several may well have lived over 200 years (George
et al. 1999). Such findings have raised some public interest: why do they live so
long? If Charnov is right, then the answer is ridiculously simple: they live long
because they experience low adult mortality. The model, of course, makes a
number of simplifying assumptions, and several developments have since
been made (Kozlowski and Weiner 1997; Charnov 2001, 2004).
How far do other organisms fit the Charnov model? One would think not
many. For example, in many organisms, body size is not determined by a
decision on when to mature, but by the size of a food patch allocated to them
by their parents and the number of offspring with which they share it. Many
insects, such as parasitic wasps and seed eating beetles, possess such a life
history (Mayhew and Glaizot 2001). In others, such as altricial birds (which
feed their offspring at the nest until fledging), offspring do not grow after
independence at all but are raised to maturation size by their parents. In
 TRAITS, INVARIANTS, AND THEORIES OF EVERYTHING 45

still others, such as most perennial plants, growth is not determinate, so

organisms not only grow after maturation but also do not divert all their
metabolic effort away from growth and into reproduction.
It would seem therefore that so many organisms break the assumptions of
the model that it could not possibly be general. However, some of the model’s
predictions are turning out to be sufficiently general to suppose that some of
the principles are common to many superficially different life histories.
Flowering plants, for example, show many features of the fast–slow contin-
uum predicted by Charnov’s model, such as a negative relationship between
adult mortality and age at maturity, despite breaking the assumption of
deterministic growth (Franco and Silvertown 1997). Enquist et al. (1999)
have even used Charnov’s results for mammals to derive relationships
between adult lifespan, age and size at maturity, and the density of wood
across tropical forest tree species. These relationships are positive as pre-
dicted by the theory. Many indeterminate growers show some of Charnov’s
predicted invariants; parasitic nematodes for example, display an invariant
maturation time multiplied by adult mortality rate, indicating that these are
negatively related to each other with exponents of equal magnitude but
opposite sign (Gemmil et al. 1999).
In contrast however, it is equally clear that it would not be valid to view
all organisms as possessing mammal-like life histories. In birds, there is a
general dichotomy between species that nest in safe places, such as
albatrosses on offshore islands, and those that nest in dangerous places,
such as grouse on the ground. The former suffer only low juvenile mortal-
ity, display low annual fecundity, grow slowly, breed late, and survive well
as adults, while the latter display opposite traits. However, adult size is not
consistently correlated with adult survival or with fecundity, but is
robustly related to age at first breeding (Bennett and Owens 2002). These
relationships are readily understood from the theory of single trait opti-
mization. For example, high pre-breeding mortality, when maturation and
growth are flexible, has long been known to select for high growth rates
and early maturation, and early maturation selects for increased invest-
ment in reproduction and decreased investment in survival. In parasitoid
wasps, there is no association between adult body size and either develop-
ment time, adult lifespan, or fecundity (Blackburn 1991). Instead, these
variables are correlated with traits specific to parasitoids: whether they
develop as endo- or ectoparasitoids, and whether or not they perma-
nently paralyze their host (Mayhew and Blackburn 1999). These traits
probably exert their influence through constraints, such as on host range
and egg size (Godfray 1994). Body size is related instead to host size and
clutch size (see Mayhew and Hardy 1998). Thus, one of the exciting
prospects of the next few years in life history evolution will be how many
46 DISCOVERING EVOLUTIONARY ECOLOGY

fundamentally different life histories there really are, and what the major
selective influences are. Trying to understand invariant relationships,
where they exist, is likely to be valuable.

4.3 The adaptive nature of metabolic scaling and

its consequences
‘I make no attempt to explain the 0.75 production scaling itself. I believe that
it represents a fundamental coupling between an organism and its environ-
ment, but my attempts to derive it from even more basic considerations have
thus far failed’. (Charnov 1993, p. 19)

Fig. 4.7 Optimal scaling of metabolic rate, following West et al. (1999). The problem is how to pack as
many metabolic ‘exchange units’ of constant size (circles) into a three-dimensional volume
(for simplicity the figure only shows two dimensions) while minimizing transport distance along
the network. Two networks are shown: for a small organism (bottom) and a large organism (top).
Note that the network for the small organism resembles a small piece of the network of the large
organism. This ‘self-similarity’ or ‘fractal’ assumption helps generate the three-quarter power
metabolic scaling. Metabolic rate is simply determined by the number of exchange units the
organism contains, and is optimized with a three-quarter power relationship to body mass.
 TRAITS, INVARIANTS, AND THEORIES OF EVERYTHING 47

An unexplained assumption in Charnov’s model was the three-quarter

scaling exponent of metabolic rate with body mass. Recently, Jim Brown
and co-workers have suggested an adaptive explanation for this (but see
Kozlowski et al. 2003; Kozlowski and Konarzewski 2004), and have gone on
to use it to not only explore many features of life history evolution but also
other biological phenomena, including both ecological and evolutionary
patterns, at large and small scales. The number of uses of the explanation is
so large that Jim Brown and co-workers have to deny that it is a theory of
‘everything’, presumably settling for a ‘large bundle of important things’.
Work derived from this original theory even has its own name, dubbed
‘metabolic ecology’.
The three-quarter scaling is most easily understood as arising from a
packaging problem (West et al. 1999). Imagine that the total area provided by
exchange units of constant size limits metabolic rate. These units might
be alveoli in the lungs, capillaries, leaves, or root hairs. We want to know how
the density of these units should scale with body mass so as to maximize
metabolic rate while minimizing the delivery distance from the surface to the
point of demand and while being packed into a three-dimensional volume
(Figure 4.7). According to West et al., this occurs when the number of
exchange units is proportional to mass to the three quarters. The maths, for
those so inclined, is explained in Box 4.1.

Box 4.1 The surface area of an object is proportional to its mass to the power 2/3,
because area is two-dimensional and mass is three-dimensional. The area
of exchange surfaces within a volume scales with slightly greater freedom
provided by two parameters, one relating to the number of exchange units
(␧a) and one to the distance between them (␧l):

a ⬀ M (2⫹␧a)/(3⫹␧a⫹␧1).

In a three-dimensional shape like a body, these parameters can vary

between 0 and 1. It is easily seen that a is maximized when ␧l ⫽ 0 and
␧a ⫽ 1. This makes sense because when ␧a ⫽ 1 the area is maximized and
when ␧l ⫽ 0 the distance between exchange units is minimized.When these
values are taken, the above becomes

a ⬀ M 3/4.

If metabolism depends directly on area of exchange units, it takes on this

three quarters exponent.
48 DISCOVERING EVOLUTIONARY ECOLOGY

Table 4.2 A few theoretical predictions that have arisen out of the three-quarter scaling of
metabolic rate and body mass

Phenomenon Assumptions Predictions Reference

Scaling of plant Plants compete for Scaling of plant density Enquist

population resources and grow until with size is –4/3; energy et al. (1998)
density limited by them; use per area is therefore
resource supply per unit size invariant
area is constant
Scaling of For a given mass, stem For trees of the same Enquist
production and diameter depends on diameter, production is et al. (1999)
life history traits wood density; independent of wood
in plants proportional investment density, relative growth
in reproduction is rate scales as ⫺0.25, trees
constant; maturity and with dense wood live
lifespan determined by longer and mature later
assuming determinate
growth
Above ground Plant design is limited Scaling of leaf number West
structure of by branching supply (0.75), height of tree et al. (1999)
plants networks; trunks and (0.25), flow rate per tube
branches resist buckling (0). Maximum tree height
while minimizing is about 100 m, limited by
energy dissipation the required width of the
supply tubes
Ontogenetic Growth depends on Growth is sigmoidal, and West
growth rates metabolic supply and at a given proportion of et al. (2001)
also on the demands of asymptotic mass, all
maintenance organisms spend about the
same proportion of
metabolism on growth
Latitudinal Total energy flux per Number of species Allen
gradient in area is body size increases with temperature et al. (2002)
species richness invariant, number of in ectotherms
of guilds across individuals in guild is
communities constant
Metabolic rates Cells within a body, Knowing metabolic rate at West
in mammals, mitochondria within a one level of organization et al. (2002)
cells, cell, and respiratory is sufficient to predict the
mitochondria, complexes within others
and molecules mitochondria are all
supplied by fractal-like
space-filling networks

Thus natural selection may provide an explanation for the scaling, and
hence, yet another contribution to the explanation of mammalian life
histories, as well as other quarter-power scaling exponents that derive from
it. However, combined with other assumptions, it can be used as the starting
point for the explanation of many other phenomena, most of which are sup-
ported by comparative data. Table 4.2 lists some of these. I cannot help but
 TRAITS, INVARIANTS, AND THEORIES OF EVERYTHING 49

quote Churchill again: ‘All the great things are simple, and many can be
expressed in a single word’. In this case, the word is ‘quarter’.
Let us see how one of these arguments works for a life history trait: growth
rate (West et al. 2001). The aim is to predict how animals grow from basic
assumptions about how metabolic rate scales with body mass. First, assume
that growth is fuelled by metabolism, and occurs by cell division. Then
assume that during growth a proportion of metabolism is devoted to making
new tissue and a certain amount to maintenance. The addition of new mass
depends on supply, proportional to mass to the three quarters (how the units
of exchange scale), and on demand, which is directly proportional to mass
(the number of cells). Thus, as growth occurs, and mass increases, demand
begins to outstrip supply, ultimately limiting growth (Figure 4.8).
This is beguilingly simple. If this simple model of growth is correct, it can
be shown that all animals should fit a single sigmoidal curve equation, and
this equation contains no less than three quarter-power exponents resulting
from the original metabolic assumptions. The values of some of the
equations’ parameters vary from species-to-species and this is what gives
different organisms apparently different growth profiles. Data for several
vertebrates do indeed fit the predicted curves, although in practice they cover
different parts of it: at birth a cow is already expending half its energy in
maintenance, and quickly approaches asymptotic size. A cod, however, only
spends half its energy on maintenance after 6 years of age. The cod can

Cow
Mass

Cod

Time
Fig. 4.8 The sigmoid curve of animal growth and development, following West et al. (2001). The curve
is asymptotic simply because demand increasingly outstrips supply as size increases. The
underlying equation of the line is the same for all species. However, because some of the
parameters take species-specific values, the curves would appear to differ for different species
when plotted on the same axes shown. Cod live most of their lives on the left part of their curve
maturing well before the asymptote (indeterminate growth), and cows mostly on the right,
maturing closer to it (determinate growth). Compare this with Figure 4.6.
50 DISCOVERING EVOLUTIONARY ECOLOGY

never really expect to reach asymptotic size, and in turn devotes a substantial
proportion of total lifetime energy to growth (about 40%). A cow, however,
expends only about 1% of its lifetime energy use to growth, and only 10% of
its pre-maturation energy use. Thus, according to the model determinate
and indeterminate growth simply reflect whether an organism reaches
asymptotic size before death.
The model, even more so than Charnov’s model, emphasizes the similari-
ties among organisms rather than their differences, something that applies to
most of the models that rely on common scaling assumptions. The model is
mechanistic rather than evolutionary, and merely invokes natural selection
for the original metabolic scaling assumption. Rather nicely, Charnov
(2001), following on from the above work, has incorporated these new
growth assumptions into a modified theory of mammalian life history
evolution. The new model is able to retain the original quarter-power scaling
predictions of his original model (Charnov 1991) but now makes more
realistic assumptions about growth and metabolism that themselves derive
from first principles. Life history theory, therefore, traditionally concentrat-
ing on individual traits, moving into invariant relationships and a more
whole organism view, has come full circle again. Throughout however,
the big message the field brings is that natural selection has played a
fundamental role in generating both the basic similarities that organisms
share, and the differences that make each unique. The field has made exciting
claims in recent years about how evolution shapes us all, and there is
more excitement to come. The following chapters consider traits that do
not find traditional placement in life history evolution, but might easily have
done so.

4.4 Further reading

Optimization theory and comparative studies are introduced by Krebs and
Davies (1993). Stearns (1992) and Roff (1992) summarize traditional life
history theory, and Lessells (1991) provides a useful quicker overview. The
mammal story is introduced by Harvey and Purvis (1999), after which try
Charnov (1993) for a broader overview to invariants. Brown and West
(2000) summarize some of the scaling work, and special issues of Functional
Ecology 18(2) and Ecology 85(7) debate some matters arising.
 5 Sons, daughters, and distorters

Most of us take for granted three facts about our gender: (1) that we each
produce only male or female gametes, not both; (2) that about half of us are
male and half are female; and (3) that this is a result of the mechanism of
sex determination known as male heterogamety. In male heterogamety,
possession of two similar (X) chromosomes determines a female individ-
ual, whereas possession of two non-identical chromosomes (X and Y) deter-
mines a male. Because haploid sperm are produced by fair (Mendelian)
segregation of the male sex chromosomes, half produce daughters and half
sons. This keeps the sex ratio at approximate equality.
In the natural world as a whole however, none of these facts may be taken
for granted; many organisms are cosexual, or display biased sex ratios, or
have different sex determining mechanisms. Instead of assuming the above
three facts, we should instead ask the following three questions:
1. What makes an organism dioecious as opposed to cosexual?
2. What favours equal or biased reproductive investment in the sexes?
3. How do sex determining systems evolve and why?
This chapter is primarily about answering these three questions. Sex alloca-
tion (the first two questions) is often considered the ‘jewel in the crown of
evolutionary ecology’ (West and Herre 2002). It was one of the first fields
to clearly match a knowledge of ecological circumstances to evolutionary
outcomes. For this reason alone it must be discussed. However, recent
advances are suggesting a far more interesting role for sex allocation: as a
driver of other evolutionary and ecological phenomena. The chapter will
explore this possibility as well.

5.1 Male and female in one individual or separate individuals?

Let us start with the first question. We will take a broad fitness-based
approach first outlined by Charnov et al. (1976). They viewed the question
of whether to express both sexes in the same individual or in separate indi-
viduals as analogous to asking if an organism should specialize in one role or
52 DISCOVERING EVOLUTIONARY ECOLOGY

a Convex

Female fitness
Fitness

Concave

Investment in gender x Male fitness

Fig. 5.1 The evolution of hermaphrodites, following Charnov et al. (1976). On the left are fitness
curves detailing how much fitness derives from a given investment in resources. Curve ‘a’ is
asymptotic, meaning that each extra investment gives less extra fitness return, whereas curve
‘b’ accelerates, meaning that only a high level of investment in that sex will return much
fitness. On the right are fitness trade-offs between the sexes. Curve ‘a’ translates into a convex
trade-off, meaning that switching investment to the alternative sex does not greatly reduce
the fitness of the other sex. For such curves, hermaphroditism is optimal (black circle), because
it gives the greatest total fitness derived from both sexes. If the fitness curve accelerates (b),
trade-offs become concave, and the optimum is to be either male or female but not both
(open circles).

be a generalist, assuming more than one role.We will encounter this problem
again in a more ecological context in Chapter 8, so the discussion here serves
as a useful introduction. First, imagine a trade-off between investment in
male versus female fitness. This can be a linear trade-off, convex or concave
(Figure 5.1). It is simple to show that when the curve is convex, cosexuality is
stable: a pure female, reallocating resource to ‘male-ness’, gains more fitness
than it loses, because the cosexual is almost as good a female as the pure
female itself while there is some male-ness added that also gives a fitness
return. Similarly, a pure male allocating resources to female-ness would be
almost as good as being male, and yet, get a bit of fitness through female
function too. What might cause the trade-off to be convex? We can under-
stand this from an examination of the relationship between investment in
one of the sexes and fitness. If the added fitness from extra investment in one
sex tends to decline, selection will increasingly favour investing in the other
sex (Figure 5.1). This means that just a bit of investment in either sex will
make the plant both a perfectly good male and a perfectly good female, and
the fitness trade-off will be convex.
The distribution of cosexuals in plants makes sense in this framework.
Cosexuality, for example, is associated with animal as opposed to wind
pollination, and with wind as opposed to animal-dispersed seeds (e.g. Bawa
1980; Charnov 1982; Vamosi et al. 2003). All these associations can be
related to the shape of the fitness gain curves. For example, wind-dispersed
seeds need to be light and therefore fitness asymptotes strongly with extra
 SONS, DAUGHTERS, AND DISTORTERS 53

investment. Animal-dispersed seeds, however, need to be resource-rich to

attract animals, and may even display an accelerating gain curve
(Figure 5.1). Thus, selection favours dioecy in animal-dispersed plants, and
cosexuality in wind-dispersed plants. With pollination, however, it is the
reverse: successful wind pollination demands pollen en mass, thus, the male
gain curve is unlikely to saturate at all strongly, thus selecting for dioecy.
Although this approach can claim some intuitive success, theory of the
evolution of cosexuality verses dioecy remains poorly tested in general: it is
very challenging to measure the gain curves necessary to confirm or refute
hypotheses, and many hypotheses are consistent with a single set of gain
curves (see Charnov 1982). We will encounter similar problems in later
chapters.

5.2 Sex allocation bias

A second sex allocation problem is understanding bias in allocation towards
one of the sexes overall, such that in dieocious populations there is a biased
sex ratio, or in cosexuals greater investment in one sex function. Our under-
standing of this problem originates largely with Fisher (1930) (though
see Edwards (1998); Seger and Stubblefield (2002)). Fisher pondered the
reasons why organisms tend to display 50 : 50 sex ratios. His argument is
one of frequency-dependent mating success. Consider a gene that biases
offspring production towards one sex. In the next generation that sex will
find itself more common, and will on average have lower mating success than
the opposite sex, hence, production of grand-offspring is reduced. Selection
will favour genes that bias the sex ratio in the opposite direction, until the sex
ratio reaches 50 : 50 again. An equal sex ratio is therefore an Evolutionarily
Stable Strategy (ESS) because selection will immediately tend to counter any
bias. This verbal argument of evolutionary stability is slightly different to
the fitness optimization problems mentioned in the last chapter in that the
fitness of any strategy depends on the strategies adopted by the rest of the
population, so the optimum shifts as evolution proceeds. This presents novel
challenges when attempting to predict evolutionary outcomes, and these
have led to a range of theoretical innovations.
Observing Fisher’s equalizing selective process in action has not been
widely possible because of a general absence of genetic variation in the sex
ratio in organisms displaying a 50 : 50 sex ratio. However, a small marine
fish, the Atlantic Silverside (Figure 5.2), showed its practical validity for the
first time (Conover and Vanvorhees 1990). In this fish, the temperature
at which offspring develop determines sex. Since the fish has a large geo-
graphic distribution, the temperature that results in an equal sex ratio is
54 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 5.2 The Atlantic silverside, Menidia menidia, which displays environmental sex determination
under frequency dependent selection, thus maintaining a equal sex ratio. Photo by R. George
Rowland and provided courtesy of David Conover.

lower in higher latitudes where the water is colder. In an elegant experiment,

populations of fish from different regions were reared at abnormal water
temperatures, and the population sex ratio was naturally biased, but over
time the sex ratio equalized because of selection on the temperature
threshold at which sex switches. This is exactly the process Fisher envisaged,
showing it works in nature as well as on paper.
In the general absence of studies like this, the validity of Fisher’s argument
has been largely tested indirectly from the association between an observed sex
ratio bias and the breaking of one of Fisher’s implicit theoretical assumptions.
Charnov (1993) has extended his ideas of ‘invariance’relationships (Chapter 4)
by reference to Fisher’s assumptions: the 50 : 50 sex ratio can be understood as
a life history invariant resulting from two underlying ‘symmetries’. The first is
that of inheritance: an offspring receives half of its genes from its father and
half from its mother. This tends to make both male and female offspring
equally valuable because half of any offspring’s genes will come from males and
the other half from females. If one sex ever became a more profitable way of
transmitting genes than the other, this symmetry would be broken and there
would be the potential for sex ratio bias. The second symmetry is that of pro-
portional gains through sons and daughters: when mothers switch investment
from one sex to another, their reproductive gains and losses are constant for all
mothers and in all circumstances. The alternative is that certain mothers may
gain more fitness than others from investing in a particular sex.
 SONS, DAUGHTERS, AND DISTORTERS 55

5.3 Inheritance asymmetry

The inheritance symmetry is known to be broken in two contexts. First, some
nuclear genes, such as those on the sex chromosomes, are transmitted more
effectively through one sex (Hamilton 1967).Y chromosomes that are able to
bias the sex ratio towards the heterogametic sex (XY males in mammals)
should be favoured over those that do not, because the homogametic sex
(XX females in mammals) never transmits them. Similarly, the X chromo-
some will be favoured by a bias towards the homogametic sex, but not as
strongly since it is also transmitted through the heterogametic sex (only half
as efficiently).
These so-called driving sex chromosomes, or meiotic drivers, are now
known from a number of organisms. Most, like ‘segregation distorter’ in
Drosophila fruit flies, are X chromosomes that cause female bias. One possible
reason is that Y chromosome drive is likely to be very strong and may cause
extinction of a population (through elimination of one sex) before counter-
measures can evolve. Driving Y chromosomes will therefore usually not persist
long enough to be observed. Countermeasures may come from modifier genes
on other chromosomes that suppress the action of the driver gene. In the fly
Drosophila simulans, X-linked drive is suppressed by modifier genes on the
Y chromosome and all the major autosomes (see Capillon and Atlan 1999).
Other, non-nuclear genetic elements can also select for sex ratio bias.
Mitochondria are cytoplasmic organelles originating from symbiotic bac-
teria (see Chapter 2). They are inherited only through eggs which carry the
cytoplasm for the zygote. In plants, mitochondrial mutants are responsible
for the phenomenon known as cytoplasmic male sterility (CMS), whereby
otherwise cosexual plants become entirely female (see Saumitou-Laprade
et al. 1994). CMS is common in agricultural plants, such as maize, rice,
sunflowers, and beans. Many insects similarly possess a symbiotic bac-
terium called Wolbachia, which is cytoplasmically inherited (Figure 5.3).
Wolbachia can have several effects on insects, including the production of
all-female lines (parthenogenesis). Incredible evidence for this first came
when males were discovered for the first time in some parasitic wasp
species after adding antibiotics to their diet, which killed the Wolbachia
(Stouthamer et al. 1990). Another class of maternally inherited microbes
are known as male-killers because of the way they achieve female biased sex
ratios. They have been found in species in a number of insect orders, includ-
ing wasps, beetles, butterflies, and flies. Wolbachia can act as a male killer in
some insects and is responsible for producing the most extreme sex ratio
biases known: in the butterfly Hypolimnas bolina, on Upolu island in Samoa
(Figure 5.3). Ninety-nine per cent of individuals are female in this popula-
tion (Dyson and Hurst 2004).
56 DISCOVERING EVOLUTIONARY ECOLOGY

(a) (b)

Fig. 5.3 (a) The butterfly H. bolina (wingspan 7 cm), whose sex ratios are extremely female biased due
to male-killing. (b) Wolbachia (length 2 ␮m), shown here in an H. bolina egg. Photos courtesy
of Greg Hurst.

Therefore, breaking the one-father one-mother symmetry frequently

leads to sex ratio bias, as well as conflict between different genetic elements
over the sex ratio. This is important as will be seen later in this chapter and
later in the book, because conflict of interest may be a powerful driver of
evolutionary change.

5.4 Proportional gains asymmetry

The second Fisherian symmetry is that of proportional gains from sons and
daughters. There are two famous ways in which asymmetric gains can result
in biased sex allocation. The first was realized by Hamilton (1967). Many
organisms display a ‘subdivided’ mating structure, meaning that females
obtain their mates from a relatively local pool of males. In such groups there
is competition among males for mates, and mothers contributing offspring
to the mating pool will achieve greater production of grand-offspring if
they bias their production of offspring towards females. The reasons are
two-fold: first, they do not waste effort on sons with low mating success;
second, they instead produce females that will be mated (Figures 5.4, 5.5).
In populations with less spatially structured mating (i.e. more females
contribute offspring to each mating pool), sex allocation should approach
equality. This is observed in numerous wasp species that lay their eggs on
resource patches where the offspring also mate (hosts insects in the case of
parasitoids, figs in the case of fig wasps) (see Werren 1980; Herre 1985).
In plants that habitually self-pollinate, there should be similar selection
pressure for a reduction in male allocation. There is comparative evidence
from numerous plants that allocation to male function decreases with the
selfing rate (Charnov 1982).
 SONS, DAUGHTERS, AND DISTORTERS 57

1:1 1:3

Fig. 5.4 The evolution of female biased sex ratios under local mating. Two situations are shown,
one with a 1 : 1 male to female ratio (left), and one with a 1 : 3 ratio of males to females. Large
circles represent offspring groups which mate among themselves before dispersal to
other patches. Closed small circles are females and open small circles are males. Small arrows
represent mating. Under a 1 : 1 sex ratio, two mated females disperse to form two patches
of offspring in the next generation. Under a 1 : 3 sex ratio, however, three females disperse,
leaving more descendents. Thus, selection favours a female bias.

The second significant way in which asymmetric gains can emerge is

when one sex benefits more from investment than the other sex in
some environments (Trivers and Willard 1973; Charnov et al. 1981)
(Figure 5.6). Many polygynous cosexual fish, for example, are female early
in life, and switch to become males later in life when they are also larger.
Presumably, taking over and defending a harem of females is only possible
for males if they are large, but it is then highly profitable. Females, however,
can produce offspring even when small. Many cosexual plants also bias
their allocation towards pollen and away from ovules under conditions of
stress, presumably because pollen is small and much less resource-dependent
58 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 5.5 Subdivided mating structures in nature. Pollinating fig wasps, Pleistodontes froggatti, inside a
Moreton Bay fig, Ficus macrophylla, that has been cut open. The wasp in the right foreground is
laying an egg into a fig flower. The female wasps have no wings as these are lost while pushing
their way into the fig through the narrow entrance. The offspring of these females will mate with
each other inside the fig; hence there is local mate competition and selection for a female biased
sex ratio. The full width of the photo is approximately 2 cm. Photo courtesy of James Cook.

than fruit. Of course, underlying all these sex allocation biases is a sex
determining mechanism. Next we will investigate how these mechanisms
might evolve.

5.5 The evolution of sex determining systems

Sex determination systems are highly variable across taxa. Heterogamety has
already been mentioned. Actually, this is a general term that conceals much
underlying genetic diversity. Males are the heterogametic sex in mammals,
females in birds, but reptiles, amphibians, and fish display both variants,
often within the same family. Different genetic systems are even known to
underlie male heterogamety: in mammals dominant male determiners exist
on the Y chromosome. In common sorrel plants, the Y chromosome is inert,
and gender is determined by the ratio of X chromosomes to autosomes. In
the Hymenoptera and some other arthropods, haplodiploidy occurs,
whereby males are haploid, developing either from unfertilized eggs or from
 SONS, DAUGHTERS, AND DISTORTERS 59

Males

Fitness
Females

Proportion males

Condition

Fig. 5.6 The Trivers-Willard hypothesis of conditional sex expression. In organisms with polygynous
mating systems (mammals, some fish), males benefit more from good condition than females
(top). Therefore selection favours male production under good conditions and female produc-
tion under poor conditions (bottom). In other organisms (most plants, many insects), females
benefit more from good conditions and sex allocation is reversed.

loss of the paternal genome after fertilization. Females are ‘normal’ diploids.
In addition to such ‘genetic’ systems are the so-called ‘environmental’ sys-
tems whereby key variables experienced during development determine
gender. For example, in many reptiles it is temperature that determines
whether an egg becomes male or female. However, in some lizards and
alligators, males develop at high temperatures, while in some turtles it is the
other way around. In other turtles and crocodiles, males develop only at
intermediate temperatures. In the European pond turtle, both genetic and
environmental factors determine sex. How did this tremendous variety
evolve? This question is one in which researchers are still just beginning to
make headway, but the results so far are fascinating.
In principle, since equal sex ratios are often expected to be evolutionarily
stable, so will mechanisms that tend to result in equal sex ratios. This phe-
nomenon, whereby selection favours genes that lead to particular sex ratios,
is known as sex ratio selection (Chapter 1), and is also the mechanism by
which sex allocation evolves. Heterogamety, because of Mendelian inheri-
tance, will tend to result in equal sex ratios, hence, should often be selected
for. Heterogamety appears to be relatively conserved (it is the exclusive
mechanism in both birds and mammals, both groups with a long ancestry).
60 DISCOVERING EVOLUTIONARY ECOLOGY

Because sex determination and sex allocation both evolve through the
same broad mechanism of sex ratio selection, they might be expected to
evolve in response to the same environmental factors. For example, it is
widely supposed that haplodiploidy has been selected for in insects and
mites by the presence of a subdivided mating structure (i.e. a proportional
gains asymmetry), the same force that selects for variable sex ratios in subdi-
vided populations, by allowing mothers ease of behavioural control over sex
allocation. That certainly seems to be the case in many species; for example,
Environmental Sex Determination should also be favoured when offspring
quality is differentially affected in the two sexes by factors that act during
development, the same force that selects for conditional sex expression
(Charnov and Bull 1977).
Inheritance asymmetry may also select for changes in the sex-determining
mechanism through conflict between genetic elements over the sex ratio. We
have already seen that different genetic elements may differ over the pre-
ferred sex allocation strategy. For example, sex chromosomes should favour
biased sex ratios while autosomes should generally favour equal sex ratios.
Similarly, cytoplasmic elements, such as mitochondria, should favour female
biased sex ratios, conflicting again with autosomes. What effects do these
conflicts have? One immediate effect is that if a force other than the parental
autosomes takes control, this can result in a novel mechanism of sex deter-
mination. Under cytoplasmic male sterility in plants, sex is determined not
by the ancestral mechanism, but by the presence of mitochondrial mutants
and restorer genes which determine whether an individual is cosexual or
female. In populations of Plantago lanceolata (Figure 5.7) there are three
known CMS mutants and three restorer genes known (de Haan et al. 1997).
Some populations have become fixed for one particular mutant and restorer,
to the extent that hybrids between populations are sterile. This is rather
interesting as it suggests that conflict over sex allocation can lead to repro-
ductive isolation, and potentially even to speciation. The reason conflict can
have these effects is that there is selection on both parties in the conflict
to rapidly evolve countermeasures such that they have the upper hand in
the conflict (an arms race). This can lead to rapid genetic differentiation
between populations evolving in isolation.
The other interesting effect of conflict between autosomes and mitochron-
dria in CMS systems is that it may facilitate the transition from cosexuality to
dioecy. Taxonomically, there is an association between the presence of dioecy
and gynodioecy (individuals either cosexual or female). Furthermore, some
species appear to have evolved towards dioecy as a result of CMS.When some
individuals in a population are fully female as a result of CMS, there can be
selection of cosexuals to reduce their allocation to female function in order
to restore the population equilibrium of equal allocation. In the common
 SONS, DAUGHTERS, AND DISTORTERS 61

(a) (b)

Fig. 5.7 Cytoplasmic male sterility in P. lanceolata. (a) A normal hermaphrodite flower with both stigmas
and anthers, and (b) a male-sterile lacking anthers. Photos courtesy of Hans Peter Koelewijn.

thyme, Thymus vulgaris, cosexual individuals in populations with CMS,

indeed, bias their allocation towards males in this way (Atlan 1992). Thus,
conflicts as a result of inheritance asymmetry can not only cause biased sex
allocation, but perhaps also shifts from cosexuality to separate sexes.
Another interesting example of how conflict over sex allocation can cause
the evolution of sex-determining systems comes from the common wood-
louse, Armadillidium vulgare (Figure 5.8). This has an ancestral system
of female heterogamety, termed ZW females, ZZ males. Some females,
however, are infected with Wolbachia and produce an excess of females.
The Wolbachia actually works by converting ZZ males into females
(ZZ ⫹ Wo) (Rigaud 1997). Several field populations now lack the normal W
chromosome entirely and consist of ZZ males and ZZ ⫹ Wo females. At
this stage, these populations have evolved from a normal chromosomal
sex-determining system to a cytoplasmic one. A further feminizing factor
was then discovered, labelled f, in a population derived from a single ZW
female formerly inoculated with Wolbachia. However, it emerged after
failure of Wolbachia transmission, suggesting that it might be an autosomal
countermeasure to Wolbachia. f can also sometimes become fixed on a
male chromosome (Z). This effectively then becomes a new female (W*)
62 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 5.8 Male (right) and female (left) A. vulgare. Photo courtesy of Didier Bouchon.

chromosome. Thus, conflict can lead to restoration of the chromosomal

system again (Figure 5.9). The woodlouse system shows how conflict can
lead to potentially rapid turnover in sex determination.

5.6 The jewel in the crown?

We began this chapter with three questions. In brief, how have we answered
them? First, separate sexes may evolve when an individual functions better as
a single gender than as two. Second, biased sex allocation may evolve when
one of two symmetries are broken. Finally, sex-determining systems proba-
bly evolve under the same processes and circumstances as sex allocation. We
also began this chapter with the statement that sex allocation is the jewel in
the crown of evolutionary ecology. We have seen that in one respect it is not,
for our understanding of the evolution of cosexuality and dioecy, while
certainly promising, is not as advanced empirically as our understanding of
sex allocation bias. Both sex allocation and sex determination probably
evolve by a common mechanism, and under common ecological circum-
stances. Sex allocation takes place in the context of sex determination, and
sex determination can evolve in response to sex allocation decisions. Our
 SONS, DAUGHTERS, AND DISTORTERS 63

ZZ
ZW

Wolbachia

ZZ
ZZ + Wolbachia

Wolbachia Femininzing factor, f

ZZ
ZZ + f

ZZ
ZW

Fig. 5.9 Evolution of sex determination in A. vulgare (after Rigaud 1997).

understanding of the evolution of sex determination is not far advanced, but

an integrated study of sex allocation and sex determination may yet prove
to justify a central place in evolutionary ecology, not simply because it is a
shining example of what is possible, but because of numerous potential
impacts on ecology and evolution, such as population extinction and in
reproductive isolation. Sex allocation, from Hamilton’s extraordinary
example, illustrates the power of genetic conflict to change living systems.

5.7 Further reading

Excellent short summaries of sex allocation are by Frank (2002), Policansky
(1987), and other articles in the same issue, and Charnov (1993, ch. 2). Hardy
(2002) provides a useful series of reviews on many sex allocation and determi-
nation issues. For an approachable text about sex ratio distorters try Majerus
(2003).Werren and Beukeboom (1998) review sexual conflict in sex allocation
and sex determination. Charnov (1982) and Bull (1983) are classic works, full
of good stuff, but require confidence with the broad issues.
6 Voyagers, residents, and sleepers

There are two lasting bequests we can give our children. One is roots.
The other is wings.
Hodding Carter Jr

One of the most important developments in evolutionary biology came in

1962 with the publication of ‘Animal dispersion in relation to social behavi-
our’ by Vero Wynne-Edwards. This was probably the most controversial
biology book of its generation. In it, Wynne-Edwards proposed that animal
populations were regulated by dispersal of surplus, less fit, individuals,
which prevented overpopulation. Wynne-Edwards believed this behaviour
to have evolved through differential survival and extinction of populations;
those groups of individuals without this benevolent behaviour went extinct,
leaving the remaining habitat to become populated by those that had
(Figure 6.1). This process, group selection, ran counter to the Darwinian
notion of selection between individuals, and was robustly rejected by
the academic community. The principal legacy of the book was to focus
attention on the mechanisms of natural selection and the evolution of
social behaviour. The ideas that followed laid the foundations of modern
evolutionary ecology.
Though Wynne-Edwards’s ideas on group selection rightly never gained
acceptance, the questions on which his work was based remain as some of the
most challenging in evolutionary biology. Many animal and plant species
display risky and costly dispersal behaviour. Having been born in a habitat
that was obviously suitable for their parents, they undertake movements to
new breeding habitats, exposing themselves to predators and unsuitable
environments, often at great energetic cost, and with no guarantee of success.
How could such behaviour evolve? In particular, can we explain it in the stand-
ard Darwinian context of selection on individuals, rather than invoking
group selection? The issue is, as Wynne-Edwards appreciated, a very import-
ant one. Dispersal is the process that binds the populations of a species
together. It can not only regulate the dynamics of populations, but also their
genetic differentiation, and coexistence and evolution with other species
 VOYAGERS, RESIDENTS, AND SLEEPERS 65

No dispersal High dispersal High dispersal,

one non-dispersing mutant
(a)

(b)

(c)

Fig. 6.1 Wynne-Edwards’s view of the evolution of dispersal by group selection, and why it does
not work. In the left column is a population with no dispersers (a), which becomes
overpopulated (b) and goes extinct (c). A population with high dispersal propensity
(middle column) would thus survive, even though the dispersers have very low survival
chances. This was Wynne-Edwards’s argument. However (right column), such populations
are susceptible to cheats (open circle), with a low dispersal propensity. Cheats do not pay
the cost of dispersal but gain the benefits from the rest of the population, and hence
increase in frequency.

(see Chapter 11). Understanding why it evolves might unlock our under-
standing of a whole gamut of evolutionary and ecological phenomena.
In addition to dispersal via their seeds, plants display an equally puzzling
but related phenomenon: seed dormancy (delayed germination). There are
two obvious costs to dormancy. First, there is some year-to-year mortality of
dormant seeds, so the total number of germinating seeds is reduced. Second,
in increasing populations there is selection for early reproduction, giving
more descendents by virtue of more generation cycles. Dormancy delays
reproduction. Hence, it is another costly and curious phenomenon. And
seed dormancy, like dispersal, is a population parameter that can affect
dynamics and coexistence. Rather nicely, but with hindsight not surpris-
ingly, the evolutionary solutions to dormancy and dispersal are intricately
related at a theoretical level, and, in plants at least, can no longer be consid-
ered in isolation.
66 DISCOVERING EVOLUTIONARY ECOLOGY

6.1 Evolutionary stable dispersal strategies

One of the great developments in evolutionary biology that resulted from
Wynne-Edwards’s challenge was the notion of the evolutionary stable
strategy or ESS (Chapter 5). These are strategies that, when adopted by a
population, cannot be invaded by any alternative strategy. ESSs result from
the relative fitness of individuals displaying alternative strategies. They are
used to predict evolutionary outcomes when the fitness of a given strategy
depends on the strategies adopted by the remainder of the population.
Hamilton (1967), in his paper on extraordinary sex ratios, is normally
attributed with the first explicit ESS model in evolutionary biology. It was
Hamilton again who, together with Bob May, first applied the approach to
the evolution of dispersal (Hamilton and May 1977). Hamilton and May
imagined a very simple scenario in which the environment consisted of a
number of identical habitat patches that were each occupied by a single
individual. These adults died every year (an annual species), and the
offspring then competed to exploit the vacant patches. Adults would give
rise parthenogenetically to dispersive offspring, with a given frequency.
Dispersing offspring were then allocated evenly across the patches, while
non-dispersing offspring remained in the natal patch to compete. The
outcome of competition within a patch was random, and dispersive off-
spring suffered a mortality cost before arriving at their new patch.
Given these conditions, what is the ESS rate of dispersive offspring? The
solution turns out to be remarkably simple: 1/(2 ⫺ s) where s is the survival
of dispersers from their natal patch to the new patch. Thus, if there is no
mortality cost to dispersal, all individuals disperse. If the mortality cost is
extremely high (s →0) over 50% of individuals should still disperse! Why
should this happen? Imagine a scenario in which the resident strategy is for
zero dispersal (Figure 6.2). A mutant individual with some level of dispersal
will displace this strategy: its offspring will compete for, and some of the
time gain, new patches which the resident non-dispersers can never regain.
The specific genetic advantage from dispersive offspring is that offspring
compete less among themselves and more among alternative strategies
which they can displace; there by reducing competition between kin that
share the same genes.
When they added in the possibility that patches could become totally
vacant (by distributing offspring randomly, not evenly among patches),
Hamilton and May noticed an interesting point: the ESS dispersal strategy
was always greater than that which would maximize site occupancy. The
implications are, first that the predominance of individual over group selec-
tion increases dispersal, and second that what is best for the individual is not
always best for the population. This second theme is actually one of the most
 VOYAGERS, RESIDENTS, AND SLEEPERS 67

No dispersal Mutant disperser

(a)

(b)

(c)

Fig. 6.2 Hamilton and May’s model of the evolution of dispersal. In the left hand column, individuals
(small circles) are distributed among four patches (large circles). After reproduction (b),
offspring compete and mortality acts (c). In the right hand column, a mutant disperser (small
open circle) arises in one patch (a). Some of its offspring disperse to neighbouring patches (b),
and some of them are successful. After competition, its patch occupancy has increased, hence,
dispersal tendency spreads through the population.

important messages of the evolution of dispersal, one that has been rein-
forced by subsequent work. In fact dispersal evolution can theoretically both
increase the risk of population extinction and reduce it (see Chapter 13).
Since then, a number of authors have extended Hamilton and May’s
assumptions, normally recovering their basic result as a special case (see
Johnson and Gaines 1990). It was clear, however, that the mathematical com-
plexities of the ESS approach would sometimes become insurmountable, and
many subsequent authors have taken alternative modelling approaches, such
as evolutionary simulations (see Chapter 1). Comfortingly these models have
tended to reinforce the findings of the ESS models, such that dispersal the-
orists can present to the world an almost united front. What have they found?
First, variation in site suitability over time tends to favour dispersal.
This suggestion preceded the advent of formal modelling (e.g. Southwood
1962) but has been confirmed by it. Temporal variation rewards indi-
viduals that hedge their bets by placing offspring in a variety of patches
in case the current patch deteriorates. A special case of this is the so-called
‘Janzen–Connell’ hypothesis (Janzen 1970; Connell 1971), usually used to
68 DISCOVERING EVOLUTIONARY ECOLOGY

explain species coexistence. In some organisms local habitat patches may

become unsuitable simply because it is already occupied by the species. One
reason may be because parents harbour natural enemies that can affect
offspring, and the effect of those enemies decreases with distance from the
parent. This selects strongly for dispersal between individuals of the same
species, and hence, also generates a species-rich local mix of individuals.
Chaotic local population dynamics, which also make patch quality unpre-
dictable in time, can select for dispersal in the same way.
Spatial, rather than temporal, variation in the environment, has the
opposite effect on dispersal evolution: it will tend to reduce it because most
individuals will be in the best sites, so fewer individuals benefit from moving
from poor to good sites than lose from moving from good to poor sites. This
spatial variation includes habitat fragmentation, which can reduce dispersal
rates severely (Travis and Dytham 1999). Clearly, in a situation where popu-
lations persist through migration between subpopulations (a metapopula-
tion), habitat fragmentation will select for reduced dispersal. This will
enhance the probability of extinction, a nasty side effect of selection acting
on individuals rather than groups. This phenomenon, termed evolutionary
suicide, has been explicitly modelled by Gyllenberg et al. (2002). They show
that under a range of conditions in which habitat patches can become
unsuitable, evolutionary suicide can occur.
The type of competition per patch is also predicted to influence dispersal.
One of the benefits of dispersal is reduced competition among residents that
are kin (Hamilton and May 1977). Inbreeding and population structures
that promote kin-groupings should increase those benefits. Given these
expectations, interesting predictions can be made about the sex that is
expected to disperse under different mating systems (Perrin and Mazalov
2000). The sex that is expected to disperse the most, depends on which sex
experiences the most severe form of competition and for what resource.
When competition affects both sexes equally, no sex bias in dispersal is
expected. In contrast, under polygyny or promiscuity, competition for mates
among males exceeds competition for resources among females, thus males
should disperse more than females. Male dispersal reduces the chances of
females dispersing, because related females are unlikely to become inbred. In
many monogamous breeding systems, however, males must defend
resources to attract females, so females might be expected to disperse more.
A final set of predictions relates to populations not at equilibrium. Many
populations show transient fluxes in geographic range, either because of
changing climate or invasions into new habitats, or because they are declin-
ing (see Chapter 13). Populations whose ranges are expanding are selected
for an increase in dispersal at the expanding range front due to the local
appearance of newly suitable patches (Travis and Dytham 2002).
 VOYAGERS, RESIDENTS, AND SLEEPERS 69

6.2 Evidence for dispersal evolution

What evidence supports these predictions? It must be admitted that the
theoretical richness of dispersal evolution, while matching easily that on
sex ratios, is not yet supported by the same wealth of evidence. The reason is
primarily the difficulty of measuring the variables: sex ratios simply require
counts of males and females, whereas direct measurement of dispersal rates
or distances is extremely problematic. Most studies have used some sort of
morphological marker for dispersal ability, such as presence or absence of
wings in animals, and presence of dispersive seed structures, such as pappi in
plants. Nonetheless, many of the theoretical predictions have received
empirical support. The prediction with greatest support is undoubtably the
effect of temporal heterogeneity.
Southwood (1962) was one of the first researchers to gather evidence in
support of the temporal variability hypothesis. He assembled supporting
evidence from the existing literature on insect dispersal, including incidence
of migratory habits, and catches in airborne insect traps. His paper was not
supported by explicit statistics but are consistent with the hypothesis. For
example, most species of British dragonfly that migrate also use temporary
water bodies like lakes or ponds for breeding habitat. Of the species from
rivers, streams, or bogs, none migrate. Roff (1990) did a similarly compre-
hensive review, this time supported by statistics, using presence or absence of
wings as the marker for dispersal. The proportion of species without wings
differed significantly among habitats, with woodlands, deserts, and the
ocean surface having particularly high levels of winglessness, which he inter-
preted in favour of the hypothesis. In a survey of North American grasshop-
pers and crickets (Orthoptera), he categorized species as either being fully
winged, wing dimorphic, or fully wingless. The flightless forms were pre-
dominantly found in caves, ant nests, alpine areas, and tundra; again habitats
that can be interpreted to have low temporal variability.
In a similar study, Denno et al. (1991) surveyed the incidence of the
macropterous (fully winged) state in 35 species of planthopping bugs
(Delphacidae), and then quantified the persistence of their habitats in rela-
tion to the bug lifespan. Again, species have a higher incidence of flight capa-
bility in habitats with low persistence times, which happened to be mainly
agricultural crops.Within species, it would also make sense for individuals to
be able to adjust their dispersal rates according to their individual assessment
of local conditions, if such a mechanism exists. Some species indeed display
such a plastic dispersive response. For example, the ruderal weed Crepis
sancta, which grows in the Mediterranean area of France, produces a greater
proportion of seeds with dispersive structures under nutrient depletion
(Imbert and Ronce 2001) (Figure 6.3).
70 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 6.3 Seeds of C. sancta. The seed at the top (a ‘peripheral achene’) lacks a parachute and is non-
dispersive, while the seed at the bottom (a ‘central achene’) has a parachute and is dispersive.
Under nutrient stress, a greater proportion of dispersive seeds is produced. Both seeds are
about 4 mm long. Photo courtesy of Eric Imbert.

So some evidence supports the notion that variation over time in habitat
quality has selected for dispersal ability. What about spatial variability? Here
the evidence is more indirect because spatial variability is hard to quantify.
However, once again we may be able to use rough markers. Oceanic islands
are areas whose suitability unambiguously decreases rapidly in space, as
the ocean is reached! As mentioned in Chapter 1, in many bird species,
incidence on islands is associated with the development of flightlessness
(see Figure 13.3). The phenomenon, however, is not restricted to birds;
there are many known instances of plant species loosing their seed dispersal
structures once island-locked. Cody and Overton (1996) showed rapid
 VOYAGERS, RESIDENTS, AND SLEEPERS 71

reduction of seed dispersal structures in species from the daisy family

(Asteraceae) on islands in Vancouver Sound. The evidence from insects (Roff
1990) is more equivocal,since island communities seem to have approximately
the same levels of flightlessness as equivalent mainlands (of similar altitude
and latitude). However, given that wingless forms are likely to have been
under-represented in the initial colonization of the islands, this may still repre-
sent a substantial overall reduction in dispersal ability over time (Grant 1998).
Analogous cases might also come from other island-like habitats, whose
persistence is assured in the short term but which become rapidly unsuitable
over space. One such habitat may be hollow trees, which can persist for many
years but which are rather rare in a woodland landscape overall. The beetle
Osmoderma eremita (Figure 6.4) lives in such hollow trees, and only about
15% of the beetles disperse from the tree they were born in each year (Ranius
and Hedin 2001). This makes sense given the kind of landscape they live in:
the tree they were born in will likely be suitable for a number of generations,
and dispersing individuals may have difficulty finding another suitable tree.
Therefore it makes sense for most individuals to stay put, just like in many
oceanic island organisms. However, hollow trees are not fully permanent
habitats; new ones arise and old ones disappear. Thus individuals will also
gain some fitness if a small fraction of their offspring is dispersive.
The contributions of inbreeding and competition have been rather scantily
tested, but there is a strong relationship between the sex that disperses and
the mating system in birds and mammals, as predicted by theory. In a review

(a) (b)

Fig. 6.4 The beetle O. eremita (a), and its habitat (b), a hollow tree. The tree hollow contains a circular
pitfall trap, into which a beetle is about to fall. Trapping studies have shown that only about
15% of these beetles disperse from their native tree, a finding that is consistent with theory of
the evolution of dispersal. Photos courtesy of Thomas Ranius.
72 DISCOVERING EVOLUTIONARY ECOLOGY

of studies of dispersal in birds and mammals, Johnson and Gaines (1990)

found that polygynous and promiscuous species show male biased dispersal,
while in monogamous species either both species disperse or females only.
Finally, evidence is beginning to suggest that expanding populations do
indeed have higher dispersal ability towards the edge of their range. During
the expansion of the Lodgepole Pine in North America, the seed morphology
shifted towards those with higher dispersal propensity (Cwynar and
MacDonald 1987). In several British insects that have recently expanded
their range due to climate warming, populations at the edge of the range
may have evolved increased dispersal relative to older more established
populations. The speckled wood butterfly Pararge aegeria is one such species.
Populations at the range margin have more massive thoraxes, which house
the flight musculature and is known to be correlated with flight ability. They
also have reduced fecundity, suggesting a fecundity trade-off associated with
increased dispersal ability (Hughes et al. 2003). There have been no studies
on declining populations for comparison, but the studies on expanding
populations, while giving comfort for those particular species, give us cause
for considerable concern about other species that, despite climate warming,
have seen their ranges contract through habitat destruction (Warren et al.
2001). It may be that with increasing isolation they begin to behave like
island populations and disperse even less than before.

6.3 Dormancy and other seed strategies

So dispersal is predicted to increase or diminish in response to a number of
selective pressures and there is growing evidence in favour of those predictions.
Seed dormancy has a similarly rich theoretical history. Cohen (1966) imagined
an annual plant living in a temporally varying environment. Seeds can either
germinate or remain dormant to the next year, in which case they suffer a
mortality cost.He then asked what proportion of germination maximizes long-
term population growth rate (a non-ESS approach). Unsurprisingly, as the
probability of total reproductive failure in any year increases, so does the opti-
mal dormant fraction of seeds. Dormancy, like dispersal, is therefore favoured
by environments that vary in time. The parallels do not end there: dormancy is
also favoured by competition between occupants of a patch,and especially if the
occupants are sibs, just like dispersal (e.g. Ellner 1987).
There is evidence to connect temporal heterogeneity with dormancy
and competition. Pake and Venable (1996) quantified dormancy in winter
annuals of the Sonoran desert, and found that species with higher temporal
variation in reproductive success had lower germination fractions. Hyatt
and Evans (1998) found a weak but significant negative association between
 VOYAGERS, RESIDENTS, AND SLEEPERS 73

family size and germination fraction in the desert mustard Lesquerella

fendleri, consistent with the idea that increased levels of sibling competition
select for increased levels of dormancy.
There is, however, yet another seed characteristic that can be selected for
under the same pressures: seed size. Larger seeds are generally favoured under
adverse circumstances for germination, such as competition and unsuitable
weather. Therefore plants are actually faced with (at least) three alternative
seed strategies to cope with kin competition, crowding, and hetereogeneity:
dispersal, dormancy, or larger seeds. Venable and Brown (1988), in a wonder-
ful paper that unifies the theory on these three different life history character-
istics, have explored whether these alternative strategies actually have different
fitness consequences in the face of environmental variability. All three
can work as bet hedging strategies to spread risk. All reduce the year-to-year
variation in fitness at the expense of reducing average (arithmetic mean)
fitness: dormancy increases fitness in unfavourable years,hence reduces fitness
variance across years, but at the cost of reduced population growth rate in
favourable years and reduced total germination. Dispersal reduces the spatial
variance in fitness but at the cost of mortality or fecundity. Seed size lowers the
temporal variance in fitness because it improves fitness in unfavourable years,
but decreases seed yield in favourable years because large seeds cost more than
small seeds. Suppose now that in a variable environment, there is a single opti-
mum balance of the mean and variance in fitness. Increases in the level of one
seed characteristic, say dispersal, will demand reductions in one or both of the
others.There should therefore be a trade-off between these three variables,one
not governed by a constraint but one due to selection.
In their model, Venable and Brown consider what levels of germination
rate, seed yield per germinating seed, and the probability of dispersal maxi-
mize long-term fitness gain. The model considers five variables and how they
affect selection for dispersal, dormancy, or seed size. The results are under-
standable if one views dispersal as a way of escaping in space, dormancy as
escaping in time, and seed size as local endurance (roughing it).
First, consider variation in the number of habitat patches. Increasing this
creates selection for increased dispersal, decreased dormancy, and reduction
in seed size. The reason is simple: the opportunity to escape in space is
increased. As a correlated response, the other two traits decrease. Second,
consider decreases in the likelihood of favourable conditions. This always
increases seed size,initially decreases dormancy but later increases it again,and
initially increases dispersal but later decreases it. Increasingly unfavourable
conditions naturally favour endurance strategies. Dispersal is initially
favoured if there are only a few unfavourable patches, but less so when nearly
the entire habitat is unfavourable. Then conditions favour dormancy
instead. Third, localized dispersal creates selection for reduced dispersal and
74 DISCOVERING EVOLUTIONARY ECOLOGY

as a result more dormancy and larger seeds, a very intuitive result. Fourth, if
patch suitability is similar in space, this selects for decreased dispersal,
increased dormancy, and increased seed size. The effectiveness of dispersal as
a means for escaping unfavourable conditions is reduced. Finally, if patch
suitabilities are similar in time, this decreases dispersal and dormancy but
increases seed size. It is obviously not possible to escape in space, so dispersal
should not be favoured here.
Hence not only do seed size, seed dormancy, and seed dispersal trade-off
against each other, but the optimal balance between them can evolve in
response to subtle changes in the nature of the environment. Comparisons
between different species, particularly among the well-described British
flora, provide excellent evidence that these seed properties trade-off against
each other. Rees (1993, 1997) has compared dispersal measures, seed size,
and dormancy propensity in several datasets. All three traits are negatively
related to each other. Although there have been few explicit tests of the
environmental correlates, the theoretical predictions match anecdotal
observations on these traits. For example, in many island plants, not only
has dispersal decreased but seed size has increased (see Carlquist 1965,
1974). Indeed the largest seed of any plant, the Coco-de-mer, is an endemic
of the Seychelles islands (see Edwards et al. 2002). These observations fit the
predictions about the number of habitat patches. Seed dormancy is reputed
to be common in desert plants, where the probability of favourable years
(rain) is especially low. In general, however, there is a great opportunity for
these predictions to be tested by long-term experiments that either manipu-
late or at least measure these variables in a quantifiable fashion.
Dispersal and dormancy are both costly traits and this realization set in
motion two initially independent research programmes. Both traits are
exciting because both are favoured by the same broad sets of environmental
variables, and, at least in plants, need to be considered in concert. Dispersal
in both plants and animals impinges on numerous other life history traits
(seed size, fecundity), and perhaps it is time that both should rightly take
their place among the core of life history theory. In addition, dispersal and
dormancy evolution provide a clear link between life histories and other
aspects of ecology and evolution.

6.4 Further reading

Johnson and Gaines (1990) provide a useful overview of classical dispersal
evolution theory, and Rees (1997) provides a good introduction to dorm-
ancy. Hamilton and May (1977) and Hamilton’s own (1996) commentary
on this are worthwhile dips.
7 Doing adaptive things

Make the most of the best, and the least of the worst.
Robert Louis Stevenson

Most of the traits we have considered so far involve characteristics that differ
between populations or species, such as their average sex ratio, their average
lifespan, and whether they are sexual or asexual. Individual organisms also
possess the ability to modify what they do in response to changes they
sense in their environment during their lifetime. In animals, where much of
this change involves differences in what they do, we call this behaviour.
In plants the differences result from changes in growth, development, and
morphology.
Where animal behaviour is considered in the light of adaptation, this
is called behavioural ecology. Behavioural ecology actually accounts for
the majority of work in evolutionary ecology, largely because many of the
conceptual advances that combined evolutionary and ecological thinking
have come from considering animal behaviour. There have been numerous
reviews, both academic and popular, of the core behavioural ecology material,
and it is not my intention to repeat it all here. Instead, I want to show how
many of the concepts have a much wider application than just to animal
behaviour. Hence, to provide some balance, this chapter will focus on plants.
I will cover three key concepts that crop up in other areas of the book:
foraging, social evolution, and sexual selection.
Plants display numerous plastic responses to environmental stimuli during
their lives that are analogous to animal behaviour (Silvertown and Gordon
1989). These include widespread responses, such as etiolation (lengthening
of shoot under shade) and tropisms (growth towards or away from a gradient
stimulus), as well as more unique responses, such as sex determination in
ferns (females produce a chemical that induces male formation in nearby
plants).In addition,behavioural ecology is concerned with very broad features
of biological life: reproduction, interactions with relatives, resource acquisi-
tion, competition, to name but a few. These are, of course, features of plant
as well as animal biology, hence should enlighten both.
76 DISCOVERING EVOLUTIONARY ECOLOGY

7.1 Resource acquisition

Let us see how this works. One of the most obvious things that animals and
plants have in common is the need for resource acquisition in an environment
that is not uniform. In behavioural ecology perhaps the most influential
model surrounding this problem has been the marginal value theorem
(Charnov 1976). In outline, the marginal value theorem is an optimization
model based on economic principles. Imagine an animal foraging in an
environment consisting of patches of food of varying quality. After arriving
at a patch, how long should the animal continue to forage there, before
leaving for the next patch? It should not stay indefinitely in a single patch
because there will be diminishing returns. The model calculates the optimum
time to stay as that which maximizes the rate of return of energy across all
patches. The benefit of staying in a patch is that an animal encounters more
food without having to spend time travelling between patches, and the cost
is that it has to spend an increasing time within the patch finding the next
food item. It is intuitive, and simple to show, that the animal should stay
in the patch until its rate of gain of food drops below its average gain in the
environment as a whole (i.e. until it becomes more profitable to move on).
This will naturally result in spending more time in richer patches. If the
travel time between patches is longer, the average gains in the environment
as a whole will be reduced, and the animal should stay longer in each patch
before moving on (Figure 7.1). These results are intuitive to us, because if we
have to travel a long way to the nearest shop, we tend to do all our shopping

Loading curve

a b d c

Travel time Time spent in foraging patch

Fig. 7.1 The marginal value theorem, following Charnov (1976). The x-axis represents time spent
travelling (to the left of the y-axis) or time at the foraging patch (to the right of the y-axis).
The y-axis shows food gained, described by the loading curve (solid curve), which asymptotes
due to patch depletion. The optimal leaving time (vertical arrows) is found by drawing the
tangent from the central place to the loading curve, here shown for a long (a) and a short
(b) travel time. The optimal leaving time is longer for the former (c) than the latter case (d).
 DOING ADAPTIVE THINGS 77

in one go, whereas if the nearest shop is just around the corner, we are likely
to do more frequent but shorter visits. A range of animals display these
qualitative predictions, making the marginal value theorem a useful heu-
ristic tool.
What about plants? Plants of course are sessile, hence do not forage in
quite the same way. Resources tend not to be other organisms (though this
does happen in a number of species, see below). Instead, plants use their
above-ground modules (leaves, stems, and sexual organs) to capture light,
attract pollinators, and to disperse seeds. Below-ground parts capture water
and ions. Rather than moving bodily into new foraging patches, plants must
direct the growth of new modules into resource rich patches. An explicit use
of the marginal value theorem to help understand this problem was by Kelly
(1990), who studied host choice in Dodder. Dodder is a rootless, leafless, and
non-photosynthetic parasitic plant that coils around the stems of other
plants and taps into their vascular system (Figure 7.2). It is essentially just
a stem that grows from one host to the next. Dodder individuals can attack
a number of plant species over the course of a season, and cover many
square metres of ground. Plant species presumably vary in their suitability.
Kelly therefore drew parallels between the food patches of the marginal value
theorem, and host individuals, and between time spent in a food patch and
investment in coiling around a host plant. Dodder plants that maximize
their long-term rate of gain in resources will invest more coils around better

Fig. 7.2 A shoot of dodder, Cuscuta europaea, coiling around the stem of a nettle host, Urtica dioica.
Photo courtesy of Colleen Kelly.
78 DISCOVERING EVOLUTIONARY ECOLOGY

quality hosts. In a series of transfer experiments, Kelly showed that plant

species in which the Dodder invested longer coils gave greater growth per
unit coil length, as predicted. In addition, greater growth translated into
greater survivorship and fecundity of the Dodder plant.
Although Dodder illustrates that some models of adaptive behaviour can
add to understanding of plastic plant responses, parasitic plants are not
archetypal of the botanical world. Gleeson and Fry (1997) have made a
more representative comparison. They used the marginal value theorem to
understand root proliferation in soil patches varying in nutrient concentra-
tion. They assumed that each plant had a limited amount of root to invest,
soil patches deplete, and plants want to maximize total uptake rate across
patches. Here, the optimal investment is where the rate of return per unit root
invested is equal across patches. As long as richer patches give consistently
higher gains per root than poorer patches, there will be a positive relationship
between root proliferation and soil patch quality. Testing their prediction on
the grass Sorghum vulgare, they found that plants grown in soil patches that
differed more in nutrient concentration also showed greater differences
in root proliferation in the expected direction. Similar qualitative trends in
root growth have been observed in a number of species, and the above-
ground parts of plants similarly display growth strategies that enhance their
rate of light capture in areas with higher light intensities (de Kroon and
Hutchings 1995).
Clonal plants face similar problems at larger scales. These plants consist
of plant units, called ramets, which produce lateral extensions (stolons or
rhizomes) from which new ramets develop (Figure 7.3): strawberry plants
are a familiar example. The environment can vary at the scale of the whole
ramet, such that some ramets may develop in uniformly poor patches

(a) (b)

Fig. 7.3 Ground ivy, G. hederacea, grown in a pot—(a) stolons bearing leaves, with petioles, and (b) a plant
‘foraging’ with stolons and petioles, across the greenhouse bench. Photos courtesy of Mike
Hutchings.
 DOING ADAPTIVE THINGS 79

(nutrient poor soil or low light intensity) while others develop in uniformly
good patches (nutrient rich soil or high light intensity). Sutherland and
Stillman (1988) found that, as predicted by foraging theory, clonal plant
stolons or rhizomes are more likely to branch in rich patches, but do not
alter their angle of branch. However,Wijesinghe and Hutchings (1996) found
that the woodland herb ground ivy (Glechoma hederacea) increases alloca-
tion to leaves and stolon branching frequency in patches of high light
intensity, and lengthens its petioles in patches of low light intensity. Such
responses are mainly properties of ramets. They suggest that stolons may
not be adapted for optimal placement of ramets, but optimal sampling of
the overall habitat. They suggest therefore that clonal plants should be con-
sidered as integrated units which can use both ramet and stolon to optimize
their use of the habitat. Theory combining these concepts is therefore needed
if we are to understand clonal plant proliferation. Whatever the long-term
outcome of this suggestion, it is clear that foraging concepts, and particularly
those provided by the marginal value theorem have value when applied to
plant resource acquisition.

7.2 Social evolution

One of the most intriguing features of animals is their behaviour towards
conspecifics (social behaviour). This behaviour can vary from cooperation
and altruism to competition and lethal combat. Cooperation and altruism,
in particular, have been the focus of attention because selfish entities are,
by-and-large, expected through natural selection (Chapter 2). However,
several mechanisms have been identified by which cooperation and altruism
can evolve (see also Chapter 11). The most famous, kin selection, occurs
when interactions occur between relatives, and is largely due to work by
Hamilton (1963, 1964a, b).
Hamilton derived from population genetics a general rule that would
predict when a gene for altruism would be favoured: when br ⬎ c. The three
terms of this inequality are respectively, the fitness benefits accrued to the
recipient of altruism (b), the fitness cost accruing to the altruist (c), and
Wright’s coefficient of relatedness (r) which measures the probability that a
gene in the altruist is also in the recipient. We have already encountered the
idea of relatedness in passing in Chapter 5 and live with these values from
day-to-day: for sexual diplo-diploids it is 0.5 between parents and offspring,
0.5 between siblings sharing both parents, 0.25 between grandparent and
grandchild (Figure 7.4). The effect of r in Hamilton’s rule is that higher val-
ues make it more likely that the inequality will be satisfied, hence more likely
80 DISCOVERING EVOLUTIONARY ECOLOGY

Generation 1
0.5
0.5
0.5 0.5

Generation 2 0.5

r = (0.5 × 0.5) + (0.5 × 0.5) = 0.5

0.5

Generation 3

r = (0.5 × 0.5) = 0.25

Fig. 7.4 Working out coefficients of relatedness. The arrows trace the possible lines of descent of a gene,
while numbers indicate the probability of that line of descent under Mendelian inheritance.
The relatedness of a grandchild to its grandparent is 0.25 (left), while that of sibs is 0.5 (right).

that altruism will be favoured. This of course is a very intuitive result: we are
nearly always nice to ourselves (r ⫽ 1), we look after our offspring very well
(r ⫽ 0.5), while the concept of the wicked stepmother (r ⫽ 0) is found in
many cultures. Hamilton’s rule has found wide support in a number of ani-
mal systems, but it is worth remembering that it is only exactly correct under
certain conditions, which are not always fulfilled (see Grafen 1984) whereby
one may have to resort to basic population genetics.
While Hamilton’s rule is widely acknowledged to define limits to altruism,
the flip side is that it also defines limits to selfish behaviour. Imagine a
situation where the decision is to keep a resource or to donate it to a relative.
If the resource has the same value for donor and recipient (b ⫽ c), then
an individual should prefer the resource to belong to itself than another
individual, even a sibling or offspring. The lower the relatedness between
individuals, the stronger the selfish tendency. Thus, the interests of relatives
frequently may not coincide, leading to conflict between family members.
Clarence Darrow’s famous saying, that ‘the first half of our lives is ruined
by our parents and the last half by our children’ holds some truth for many
organisms.
Plants have, in this sense, plenty of opportunity for social interactions
between relatives. The application of Hamilton’s rule, and especially the
concept of relatedness, helps our understanding of many reproductive
phenomena. The reproductive biology of angiosperms outlines many oppor-
tunities for conflict (see Mock and Parker 1997). Pollen consists of two
haploid cells (microspores) produced by a meiotic division encased in a
 DOING ADAPTIVE THINGS 81

(a)

(b)

(c)

(d)

(e)

Fig. 7.5 The formation of an ovule in angiosperms (thick lines are chromosomes, dotted circles are
nuclei, solid circles are cells). A diploid mother cell (a) undergoes meiosis to form four haploid
megaspores (b). Three disintegrate, while the surviving functional megaspore undergoes three
mitotic divisions producing a single cell with eight haploid nuclei (c). An uneven cytoplasmic
division then produces six haploid cells and one diploid cell (d). Once fertilized by the two
pollen sperm, the diploid cell becomes the triploid endosperm, and one of the haploid cells
the egg nucleus (e).

tough capsule. On arrival at the female stigma one cell grows as a pollen tube,
while the other divides by mitosis into two sperm. Eventually the pollen
tube reaches the ovary which consists of one or more ovules, which when
fertilized will develop into seeds. Like pollen, ovules are rather complicated
entities (Figure 7.5). The ovule starts out as a single diploid mother cell which
divides by meiosis into four haploid megaspores. Three of the four mega-
spores disintegrate, leaving a single functional megaspore. This undergoes
mitosis three times without cytoplasmic division to give a single cell with
eight haploid nuclei. Then a cytoplasmic division occurs leaving six small
haploid cells and one central cell with two haploid nuclei. When fertilized by
one of the sperm, this cell will become the triploid endosperm that provides
nutrition for the developing seed. Meanwhile, one of the haploid cells (the
egg nucleus) is fertilized by the other sperm to become an embryo.
The triploid endosperm is a curious phenomenon. The endosperm pro-
visions the seed by drawing on maternal resources. It contains two identical
copies of the maternal genome and one paternal. There are at least two
reasons why this might have happened. First, it could have arisen to try to
counter sexual conflict between mother and father. We will encounter sexual
conflict again in a later chapter in a different context, but we have already
encountered one of its consequences: genomic imprinting (Chapter 2).
82 DISCOVERING EVOLUTIONARY ECOLOGY

If embryos all have different fathers, then the paternal genome is selected to
gather resources for the embryo it has fertilized at the expense of other
embryos, to which it is totally unrelated. A mother, however, is related to all
embryos equally since she is mother to them all. In short, fathers and mother
conflict about apportionment of resources between embryos. Genomic
imprinting results from this conflict: it is where the expression of a gene is
conditional on the sex that transmits it (Chapter 2). Fathers, for example,
should be selected to activate genes that result in resource transmission to the
embryo; mothers should be selected to counter this. This is found in maize
where the number of paternal and maternal genomes in the endosperm
can be manipulated to some extent (Haig and Westoby 1991). Genes on
chromosome 10 are only active when paternally derived and result in normal
sized kernels if the endosperm genome ratio is normal, small kernels if there
is maternal overdose, and when the endosperm is tetraploid with equal
maternal and paternal representation, the embryos abort. The latter is pos-
sibly a result of active maternal countermeasures. From this example, the
doubling of the maternal genome could therefore be a ‘trumping’ mechanism
for the maternal sporophyte to reassert control over the paternal genome.
The alternative view stems from conflict between kin. From an inclusive
fitness perspective, the mother is indifferent to which embryo receives
resources (but see below), since she is equally related to all of them (r ⫽ 0.5).
The embryo, however, is related to other embryos by between 0.5 (if they all
have the same father) and 0.25 (if they all have different fathers) so prefers a
skew of resources towards itself. Thus, embryos should be selected to try to
gain resources from the mother at the expense of other embryos. There is
thus a potential for kin conflict, and hence parent–offspring conflict. At the
centre of all this is the endosperm. The relatedness of the endosperm to these
players is key. If the ancestral endosperm were a diploid maternal genome,
then that would have put the mother in control of resource provisioning,
since the endosperm and the maternal interests would be exactly congruent.
If the ancestral endosperm was a diploid identical twin of the embryo, that
would put the embryo in control. What we have instead is an intermediate
situation: a gene in the endosperm is guaranteed to be in the embryo (r ⫽ 1),
but a gene in the endosperm has a two-third chance of being in the sporo-
phyte (r ⫽ 0.67), and is more related to other embryos than its own embryo
is. Hence, things obviously do not go all the mother’s way, but she is much
better off than if she only had a half share in a diploid endosperm.
There is evidence as well for kin conflict in plants. In Dalbergia sissoo, a
tropical tree from the pea family, multi-seed pods become single seeded by
progressive seed abortion, caused by water-soluble chemicals that diffuse from
the focal sibling. This increases the weight of the focal seedling, and the
removal of its sibs enables the pod to disperse farther by the wind (Ganeshaiah
 DOING ADAPTIVE THINGS 83

(a) (b)

Fig. 7.6 A flower (a) of M. guttatus, and (b) plants growing by a copper-polluted stream at
Copperopolis, California. Photos Courtesy of Mark Macnair.

and Uma Shaanker 1988). Sibling conflict of this kind is also probably in
maternal interests in many plants. It is characteristic of most flowering plants
that there is extravagant overproduction of early reproductive structures in
most years. There is strong evidence that mothers are able to select which
should stay, and which should go. These include selfed embryos, which are in
many plants more likely to be dropped than outcrossed embryos. However,
other characteristics may be selected.Amazingly, Mimulus guttatus growing in
copper-stressed soil can selectively abort offspring which are copper sensitive
(Searcy and Macnair 1993) (Figure 7.6). In general there is good evidence that
selective abortion improves progeny fitness, that is, that it is adaptive. For
example, plants of Lotus corniculatus allowed to abort seeds naturally have
more fecund offspring than ones which have equivalent numbers hand
thinned at random (Stephenson and Winsor 1986).
Even when the fruit has left the parent plant, interactions between the
parent and offspring may continue.As Ellner has pointed out (1986), parents
may be selected towards making a certain proportion of seeds dormant
(Chapter 6). If there is competition between germinating sibs, parents may
be selected to produce dormant seeds to reduce sib competition. She values
all her offspring equally. Each seed, however, is expected to give priority to
itself. This can select for reduced dormancy, giving rise to conflict between
parent and offspring over whether seeds should be dormant or not. Rather
interestingly, dormancy is commonly caused by a tough seed coat, which is
broken down by mechanical or chemical means. The seed coat is a maternally
derived tissue, giving her control over seed dormancy.
84 DISCOVERING EVOLUTIONARY ECOLOGY

So the strange reproductive life cycles of plants leave much room for
family interactions, through (1) maternal over-production, (2) multiple
siring of offspring on a plant, and (3) provisioning of seeds from maternal
resources. The concept of kin selection enhances our ability to understand
these peculiarities of life.

7.3 Sexual selection in plants

This discussion of plant reproduction is convenient for the next subject:
sexual selection. Sexual selection was another brainchild of Darwin (1871)
who realized that characters that made organisms successful at acquiring
a mate might be selected for, and that this can account for some of the extra-
ordinary exaggerated traits of organisms, such as antlers of deer and long
tails and colourful plumage of pheasants. The first person to explicitly model
female choice as the exaggerating process was Fisher, who imagined runaway
selection (Chapter 1). Females would initially prefer some males more
than others and these males would then gain greater reproductive success.
Over time the preference and the male trait would become coupled and
increase in magnitude until countered by some cost of the male trait.A second
more recent set of models, collectively known as good gene models, assume
that female choice is costly and that therefore it should not be arbitrary.
Instead, females should be selected to choose males that provide them
or their offspring with some kind of advantage in natural selection. The par-
ticular traits females choose should in fact be correlated with some fitness
advantage in natural selection.
Are there any plant traits that might have evolved through sexual selection?
The obviously exaggerated plant reproductive traits are large showy flowers
and long pistils, which may have evolved by male–male competition. In the
former case, one can imagine male flowers competing with each other for
pollinators to remove pollen. The phenomenon of fleur-du-mâle, whereby
male flowers are overproduced and then aborted, may well be a sexually
selected phenomenon: a plant makes itself as showy as possible to attract
as many pollinators as possible into the vicinity. For example, many Agave
mckelveyama flowers are aborted before fruit initiation, and these are always
functionally male, and give high nectar rewards (Sutherland 1987). Therefore
some flowers, indeed, seem to serve merely to attract pollinators and have no
fruiting function. Queller (1983) studied Asclepias exaltata and showed that
inflorescence size is positively related to pollen removal, while seed set is
unaffected. This therefore suggests that the showy trait benefits male function
much more than female function.
 DOING ADAPTIVE THINGS 85

Despite the possible role of sexual selection in flower evolution, plants

in general do not work through the showy media of vision and sound that
animals do, so we should expect the effects of sexual selection to be more
cryptic. In particular, we might expect evolution of competition between
pollen. Pollen grains are known to inhibit each other’s growth in some species,
and in other species pollen grain size is related to fertilization success. Female
choice and male–male competition are likely to interact through pollen
competition, as the female tissues provide the arena for pollen competition.
In peach trees, the base of the style reduces after pollination and this
increases competition between pollen grains by reduction in both resource
flow and space (Herrero and Hormaza 1996). Pollen competitiveness also
sometimes reflects male quality: pollen tube growth rate in violets increases
with the seed production of the donor in violets, and this translates into
offspring vigour (Skogsmyr and Lankinen 2000).
The interactions between pollen and pollen, and pollen and sporophyte
therefore provide intriguing examples of sexual selection at the boundary of
male–male competition and female choice. Much more clearly defined
examples of female choice occur in seed abortion. The last section has already
shown how this can be selective according to male genotype, and also increase
offspring vigour. Once again though, the effects of this choice are relatively
cryptic to human eyes.
It is clear that the biological differences between plants and animals
sometimes provide us with evolutionary outcomes that are relatively unique
to one kingdom or the other. It is equally clear, however, that theory developed
for applications in one kingdom can be usefully translated across to the
other and enhance understanding. While nobody can deny the success of
behavioural ecology, the term itself has limited the application of its findings
to different taxa. Evolutionary ecology of course has no taxonomic bound-
aries and it is on this wider stage that the full implications of theory will
be realized.

7.4 Further reading

Krebs and Davies (1993) introduce behavioural ecology. Silvertown and
Gordon (1989) discuss plant ‘behaviour’, and Mock and Parker (1997)
provide an interesting introduction to family issues in plants. Skogsmyr
and Lankinen (2002) review sexual selection in plants.
8 Evolution and numbers

So far we have covered how ecology causes anagenetic change, the place
of ecology being largely subsumed under the term ‘selection’. The present
chapter, like Chapter 3, explores more the effect of evolution on ecology,
thus closing a causation loop. Specifically, we will investigate the effects of
anagenetic change on population dynamics, an ecological characteristic,
and ask whether we can better understand population level phenomena
by incorporating evolutionary assumptions. Further, we will explore an
additional theoretical step in the evolutionary process by using assump-
tions about population dynamics to help estimate the path of evolutionary
change. Thus, evolution influences ecology, which influences evolution
again.
There are essentially two situations in which knowledge of adaptation
can help predict the dynamics of populations. The first is when organisms
display plastic phenotypic strategies (such as behaviours) that are adapted
to suit particular environments, and these behaviours affect population
ecology in some way. In this scenario no evolution occurs per se during
the timescale of the study, but the plastic responses are assumed to be
the outcome of past selection. The second way is, if evolution itself occurs
on ecological timescales, by which I mean roughly the length of a human
lifespan. This is more common than it might at first appear. It is not
a great leap of understanding to realize that changes to a trait can affect
a population’s dynamics. That then requires ecologists to think about
evolution. However, a further logical step is possible. The fitness of
phenotypes may be affected by the ecological interactions between
individuals, for example if fitness is density- or frequency-dependent.
This may then decide which genotypes spread or disappear. Thus, there
may be a case for incorporating assumptions about the effect of a trait on
population dynamics merely to work out what the final evolutionary state
itself is. The theoretical approach that incorporates these assumptions is
known as adaptive dynamics, it is relatively new, and we will examine how
it works later in the chapter.
 EVOLUTION AND NUMBERS 87

8.1 Adaptive decision-making and population dynamics

In the last chapter we discussed a foraging model known as the marginal
value theorem (Charnov 1976). The model considers a patchy environment
with patches differing in productivity and is asked how a single organism
should behave in response to them. We assumed that the final phenotypic
state of the organism would be that which maximized fitness, according to
the optimization principle.We will now consider a whole population of indi-
viduals inhabiting the environment, and assume that the food supply in
a patch replenishes so that organisms can inhabit them full-time. How will
the individuals distribute themselves between patches that vary in productivity?
This problem was first considered by Fretwell and Lucas (1970). Imagine
the first organism to arrive in the habitat. It should, following the optimiza-
tion principle, settle in the most productive patch, where its food intake will
be highest. By foraging, however, the first organism reduces the productivity
of the patch for subsequent foragers through competition (Figure 8.1). The
second individual to arrive in the habitat therefore faces a slightly different
decision because the productivity of the best patch will now appear lower
and hence closer to that of the other patches. If it is still higher than all the
others, however, the second organism should still settle there. Individuals
continue to occupy the most productive patch until its apparent productivity
to the subsequent individual equals that of the second most productive
patch.At that stage subsequent organisms to arrive should alternately choose
the first and then the second patch as one is depleted below the level of
the other in turn. If all individuals are equal competitors and they know

Good habitat
Individual food intake rate

Poor habitat

Number of individuals in habitat

Fig. 8.1 The ideal free distribution. The two curves represent food rewards per individual in a rich and poor
habitat as the habitat fills up with competitors. The first set of individuals enters the rich habitat,
continuing until such time as the individual rewards are the same as in the poor habitat (horizon-
tal line). Thereafter, individuals fill up both habitats such that rewards are the same in both. After
Fretwell and Lucas (1970), with permission from Springer Science and Business Media.
88 DISCOVERING EVOLUTIONARY ECOLOGY

both the productivities of the different patches and the distribution of

competitors among them, the end distribution, one in which no individual
can do better by moving to another patch, is called the ideal free distribution.
This is an evolutionary stable strategy, the concept that applies to max-
imization situations when the fitness of a given strategy depends on the
strategies adopted by other individuals in the population (Chapter 6).
At the ESS (the ideal free distribution here), the number of individuals
per patch is proportional to their productivity, with more individuals in
the best patches.
The ideal free distribution model allows us to predict the distribution
of organisms among patches and also their fitness in those habitats. The
implicit assumptions, that individuals have perfect knowledge of the
habitat, of the distribution of conspecifics, and that they can move freely
among habitats, are obviously inappropriate for less mobile organisms,
but for some, such as migratory birds, they might be fair approximations.
The ESS models of habitat selection like the ideal free distribution can
then allow us to predict the productivity of individuals in a population
in response to changes in density. These ‘density-fitness functions’ can
then be used to predict how the population will respond to changes in the
environment, such as removal of some of the available habitat or a decrease
in its quality.
Imagine then a migratory bird population that breeds at one location
in summer and then spends the winter in another site where it only
feeds. Suppose one summer some of the winter-feeding habitat for the
species is removed by a new housing development. What will happen to
the population? The modelling logic is roughly this (Goss-Custard and
Sutherland 1997): the population migrating to the feeding ground for
the winter must be packed into a smaller area of habitat. An ESS model
such as the ideal free distribution is used to predict the relationship
between bird density and per capita food intake rate in the new smaller
habitat. By making some appropriate assumptions relating food intake
to winter mortality, we can describe the relationship between density
and mortality for the new smaller habitat. An ESS model can also be used
to describe the way that bird breeding-habitat fills up and responds to
density, and using some appropriate assumptions, how birth rate relates
to density. The equilibrium density of the population will be when the
birth rate exactly balances the death rate (Figure 8.2). It is now simple to
predict the new equilibrium population size if we know how much habitat
has been lost, hence, how much the mortality–density function has
been shifted.
Sutherland (1996) has done this for the European oystercatcher (Figure 8.3),
a wading bird that winters on western European coasts and breeds throughout
 EVOLUTION AND NUMBERS 89

Mortality

Mortality or fecundity

Fecundity

Equilibrium
Density
Fig. 8.2 How to find the equilibrium population density from the per capita mortality and fecundity
curves.

Fig. 8.3 An oystercatcher, Haematopus ostralagus, foraging. Photo courtesy of Stephane Moniotte.

the north of the continent. He used ESS models to predict the winter
mortality–density function, using data from birds wintering on the Exe
estuary in Britain, which harbours around 2,000 birds. From long-term
studies of these birds, we know that at high density there is more interference
between foraging birds as well as greater food depletion, all of which increases
winter mortality. For the slope of the birth rate–density function he used data
on a breeding population from the Dutch island of Schriermonnikoog.When
density is low on the island, birds can all occupy productive territories
90 DISCOVERING EVOLUTIONARY ECOLOGY

by the sea, and the birth rate is high. When density is high, birds are either
forced to occupy lower quality territories away from the sea, from which they
have to commute to the coast to find food, or they do not breed at all but
wait at the coast in the hope of occupying a coastal territory if one becomes
vacant. As a result, the per capita birth rate is reduced. Consequently,
Sutherland was able to predict that if wintering habitat on the Exe estuary
were reduced by 1%, the population of oystercatchers would be reduced
by 0.69%. The reduction in population size is therefore subproportional to
the loss of wintering habitat.
How accurate might the predictions of such ‘behavioural-based models’
be? In general, we expect them to be better than the major alternative, so-
called ‘demographic models’ which make predictions through direct mea-
surements of the density–mortality and density–birth rate relationships.
There are two problems with the demographic approach; one practical and
one theoretical. The practical problem is that obtaining accurate data on
these relationships can be very difficult. In contrast, studying the behaviour
of individuals is much more practical. The theoretical problem is that the
new environment, whose effects we want to predict, is very likely to change
the density–fitness relationships. For example, reducing the amount of winter
feeding-habitat is likely to increase the mortality rate for a given density.
The overall behavioural strategy that underlies these changes, however, is
likely to remain fixed, therefore by studying behaviour we are studying
the mechanistic basis for any change and can detect them.
There is also now some empirical evidence emerging for such advantages
of the ‘behaviour-based’ approach. Stillman et al. (2000) have constructed
a very detailed behavioural-based model to predict the population res-
ponses of oystercatchers on the Exe estuary. The model was constructed
using behavioural observations in the late 1970s, when the population
was relatively low. Since then there has been an increase in the wintering
oystercatcher population. As a result mortality on the estuary has also
risen, and the behavioural-based model was able to predict the level
of increased mortality relatively well. In contrast, the density–mortality
relationship in the late 1970s was a poor predictor of the relationship
later on.
We therefore have grounds for optimism in the value of population mod-
els based on adaptive behaviour. They have now been applied to a range of
bird and some mammal populations, aimed at predicting effects of envi-
ronment change, such as human disturbance, habitat loss, habitat exploita-
tion, sea level rise, and changing agricultural practice (see Sutherland and
Norris 2002). Thus, using adaptive models to predict how organisms will
change their behaviour in response to changes in their environment can help
us to predict the consequences of those changes for the population biology.
This assumes that all the evolution has happened in the past leaving the
 EVOLUTION AND NUMBERS 91

organism with a plastic adaptive response to the environment. Another

possibility exists however, that organisms will evolve to changing circum-
stances on timescales that are relevant to population ecologists.

8.2 Evolution in ecological time

Scientists now agree that evolution can frequently occur over timescales that
matter to ecologists, such as a few decades or less. There are few more graphic
examples of rapid evolution than of the emergence of new human diseases.
In the case of human immunodeficiency virus (HIV), the disease it causes
(acquired immune deficiency syndrome or AIDS) was first recognized in
1981. The virus responsible (Figure 8.4) was characterized in 1983 as a
retrovirus with similarities to primate lentiviruses, also known as simian
immunodeficiency viruses (SIVs). SIVs are widespread among African apes
and monkeys, and phylogenetic comparisons of HIV and SIV sequences
suggest, with a large margin for error, a likely origin of the current epidemic
in the middle decades of the twentieth century (Hahn et al. 2000). Most
interestingly, SIVs are not known to cause disease in their hosts, whereas
HIVs are, as far as is currently known, 100% fatal. It seems therefore that
HIVs have evolved to become more virulent over the timescale of only decades.
That, however, is nothing compared to what goes on within a single host.

Fig. 8.4 Particles of HIV infecting a human tissue-culture cell. The HIV particles are about 100 nm across.
Photo courtesy of the Ripple Electron Microscope Facility, Dartmouth College, New Hampshire.
92 DISCOVERING EVOLUTIONARY ECOLOGY

When transmitted to an uninfected human, the virus undergoes rapid

replication in the so-called ‘primary infection’stage. Replication occurs largely
in white blood cells, though other tissues can be used. The immune system
of the host then rapidly responds and reduces virus to very low levels. At this
stage the host can be recognized as sero-positive. During the next, latent,
stage of infection, diversity of the virus slowly increases. New strains of the
virus arise because of the low fidelity of the enzyme reverse-transcriptase,
which converts the RNA virus genome into DNA that can integrate with
the host genome. The rate of substitution in the virus genome is roughly
one million times that of human genes. Typically 1010 virus particles can
be produced per victim per day. After a variable number of years of viral
replication and gradually increasing viral diversity,depletion of T-lymphocytes
is such that the virus replicates rapidly, and the virus load becomes
dominated by fast-replicating strains (Nowak et al. 1991). AIDS symptoms
occur, and the victim dies of some secondary infection that would not
normally be fatal.
We have observations suggesting evolution on two scales: an increase in
the virulence of HIVs relative to SIVs and a within-host evolution towards
increasing virulence towards the latter stages of infection. Has natural
selection been responsible for these trends? It is likely in both cases. For
example, resistance to individual antiviral drugs, such as AZT, occurs
typically in six months, and similar mutations characterize the resistant
strains. This strongly suggests selective evolution of the virus. What about
the increase in HIV virulence relative to SIV virulence? Both viruses have
high replication rates, mutation rates, and viral loads. These then do not,
on their own, seem to engender high virulence. However, the virus
envelope protein of SIVs is very conserved, but rapidly evolves in HIVs.
In addition SIVs only stimulate a weak immune response from their hosts,
while HIVs stimulate a strong immune response.
This suggests that SIVs experience stabilizing selection, while HIVs are
positively selected by the immune response. When a strong immune
response occurs, genetic variation is created by strains attempting to evade
it, that eventually leads to virulence and AIDS. For SIVs a weaker immune
response creates reduced selection on the virus, creating less viral compe-
tition and less virulence. In summary, the difference in virulence between
HIV and SIV seems to have been driven by the extent of arms race between
virus and immune response (Holmes 2001). Under this scenario pressures
that select for virulence of HIV relative to SIV are the same as those that
select for virulence during HIV infections.
There seems little doubt therefore, that HIV can and has rapidly evolved
virulence. It seems obvious, especially in the context of a disease like AIDS,
that there will be population consequences of short-term evolution. How in
 EVOLUTION AND NUMBERS 93

general can we predict such consequences? A taste for this comes from recent
work on the evolution of exploited fish stocks. Unlike in HIV, a convincing
case for short-term evolution in fish stocks has taken time to accumulate. In
contrast to HIV, however, we are much better equipped to predict the popu-
lation consequences because of the presence of high quality data on relevant
parameters.

8.3 Consequences of short-term evolution

Through harvesting, humans can impose impressive selective forces on
populations. The North East Arctic cod Gadus morhua (Figure 8.5) has
experienced a large reduction in age at maturity from 9 to 6 years in the last
50 years, a change that is typical of other exploited stocks of marine fish
in the North Atlantic (Law 2000). One likely reason for this has been the
change from fishing only at the spawning grounds off the coast of Norway to
fishing largely in the Barents Sea, which is the feeding grounds of the fish.
This change in fisheries practice was brought about by the introduction of
motorized deep-sea trawlers in the 1920s. Currently almost 40% of the stock

Fig. 8.5 A cod, G. morhua, on the slab. Fishing activity has led to a decrease in the age of maturity and
consequent reduction in yields. Photo courtesy of Gerd-Peter Zauke.
94 DISCOVERING EVOLUTIONARY ECOLOGY

is removed annually from the Barents Sea. Over the course of several years,
the cod have very low chance of maturing successfully. Genetic variants that
escape the feeding ground to spawn earlier have a much better chance of
reproducing. In contrast, when the fish were caught only at the spawning
grounds, mortality would have been lower at the feeding grounds, creating
selection for increased age at maturity.
Can these changes in fishing mortality exert effects on the fish population?
There is some evidence that it can: for example, Conover and Munch (2002)
studied the Atlantic silverside (see also Chapter 5). Removing smaller
individuals gave an increased yield over selecting larger individuals. There
were two reasons: first, when small fish were harvested the adults grew larger,
giving greater reproductive potential. Second, removing smaller individuals
selected for faster growth as the fish then spend less time in the more vulner-
able stages. Such studies show the potential for harvesting to cause evolu-
tionary changes that affect yield. They do not, however, easily suggest what
the consequences have been in cod, where the regime is not purely a size
selective one and the evolutionary changes have been in age at maturation.
For this a specific cod model is needed.
The intuitive effects of reduction in age at maturity are that the yield
of fish will decline: the fish mature earlier, hence are putting energy
into reproduction instead of growth at an earlier age. Law and Grey (1989)
developed the notion of the evolutionarily stable optimal harvesting strategy
(ESOHS) to help quantify the effects on yield. At the ESOHS, fishers adopt
a strategy that maximizes yield after evolution has reached its final state.
They showed that the ESOHS for the North East Arctic Cod was the tradi-
tional practice of fishing only at the spawning grounds, and that fishing at
both feeding and spawning grounds led to a reduction in long-term yield of
up to two fold. One problem, for fishermen, of following the ESOHS is that
they may only gain the benefits of maximizing yield after waiting for evolu-
tion to come to rest. In the meantime, it is possible that they may have to
pursue a suboptimal strategy from the perspective of maximizing yield.
Instead it would be sensible for fishers, at least in the short term, to pick the
strategy that maximizes sustainable yield at each point in time. In response
to this, the fish would experience a slightly different selective pressure, and
evolve to a new state, after which the fishermen would perhaps need to
change their optimal harvest strategy. This situation is thus, as in HIV infec-
tions, a co-evolutionary circuit, this time consisting of evolutionary changes
in fish, followed by changes in harvesting strategy by the fishermen.
Heino (1998) has modelled these co-evolutionary dynamics for cod.
He was interested to know what particular type of harvesting strategy
maximized sustainable yield and whether the co-evolutionary dynamics
also lead to an asymptotic fishing strategy that is the same as the ESOHS.
 EVOLUTION AND NUMBERS 95

The technique he used to determine the evolution of the fish is known as

adaptive dynamics (Dieckmann 1997; Waxman and Gavrilets 2005, see
below). Like Law and Grey, he found that fishing in both the spawning and
feeding grounds led to a reduction in yield relative to fishing in only the
spawning grounds. However, the co-evolutionary dynamics frequently
led to the ESOHS even if it was not pursued in the first place.

8.4 Adaptive dynamics

The techniques used in Heino’s model are interesting. In a normal ESS
approach there are no explicit assumptions about population dynamics
made in determining the final evolutionary state. If, however, we explicitly
want to predict the effect of evolution on population dynamics, we are more
likely to have to face up to this challenge. Adaptive dynamics adds in this
interaction. In brief, the fitness of new mutants is assessed in terms of their
population dynamic interaction with residents. If at the end of that interac-
tion, the mutant has a higher fitness than the resident, the mutant becomes
the resident strategy and the resident is displaced. This continues until a
so-called ‘singularity’ is reached (Figure 8.6). Much of the time a singularity
may be a fitness maximum (and ESS), thus representing a single resting state.
However, in some circumstances the singularity may, counter-intuitively, be
a fitness minimum. This means that evolution converges to a point that is less
fit than all surrounding strategies. This is surprising to evolutionary biolo-
gists who are used to the analogy of evolution as climbing an adaptive land-
scape, a view that is consistent with the notion of an ESS. However, this

(a) (b)
Fitness

Fitness

Phenotype Phenotype

Fig. 8.6 Adaptive dynamics leading to evolutionary branching. A lineage normally moves (arrow) over
an adaptive evolutionary landscape by climbing to areas of higher fitness (a), eventually
reaching an adaptive peak which is fitter than all other local phenotypes. However, when the
adaptive peak is reached, the dynamical interactions between the resident genotypes and
mutants may allow any alternative new mutant to invade, which means that the resident
genotype is now at a fitness minimum (b). Under certain conditions, two divergent phenotypes
can then coexist, the start of a polymorphism or even speciation.
96 DISCOVERING EVOLUTIONARY ECOLOGY

intuitive notion assumes that the landscape is fixed, at least during the
timescales involved in any one study. Adaptive dynamics changes all that.
Invasion of new mutants is assessed in relation to the ecological dynamics
of mutants and residents, at each step of the adaptive walk. The dynamics
of evolution change the ecological dynamics at each step, and the changing
ecological dynamics change the local fitness landscape because fitness can be
density- or frequency-dependent. Thus, it is perfectly possible for evolution
to proceed towards a fitness minimum because the ecological interaction that
determines evolution towards that point is different to the interactions that
occur once the point is attained. After climbing to what seems like a peak,
sometimes the peak sinks and becomes a trough.
What happens in the neighbourhood of a fitness minimum is parti-
cularly interesting, for then any alternative mutant strategy is fitter than
the resident: in other words, selection is disruptive. This can lead either to a
coexisting polymorphism, or even speciation in sympatry. This so-called
‘evolutionary branching’ can occur repeatedly. Thus, taking into account the
population dynamic interactions involved during evolution can lead to very
different predictions about the asymptotic state of evolution, which are
not only intuitively appealing but seem to offer explanations for observable
phenomena. For example, polymorphic strategies are commonly seen in
traits that relate to population level phenomena, such as body size, dispersal
rate, dormancy rate, and so on. In addition, the degree of evolutionary
branching has obvious implications for speciation, and has been implicated
as a major source of sympatric speciation (Dieckmann and Doebeli 1999).
The justification for adaptive dynamics at present is purely logical; it is not
yet known to increase predictability. For an evolutionary ecologist, however,
there is excitement in exploring the interaction between evolution and
population ecology. The perceived need in medical fields, caused by the
evolution of disease, has led to the notion of Darwinian medicine (medicine
with evolution in mind), and in fishing circles to Darwinian fisheries (fishing
with evolution in mind). The next few years should unravel how many areas
of population ecology Darwin’s name will attach to.

8.5 Further reading

The ideal free distribution is introduced by Krebs and Davies (1993). Goss-
Custard and Sutherland (1997) and Sutherland and Norris (2002) introduce
behaviour-based models. Holmes (2001) discusses HIV evolution. Law (2000)
introduces evolution of fish stocks, and Waxman and Gavrilets (2005) and
the many commentries that follow it discuss adaptive dynamics.
9 A world of specialists

Specialists are people who always repeat the same mistakes.

Walter Gropius

Imagine that most typical of ecological entities, a community of coexisting

species. The community may be characterized by variables, such as the
number of species and the number or type of interactions. The next several
chapters will examine how evolution affects these fundamental properties of
communities. This chapter specifically examines the evolution of the breadth
of use of resources that govern the size of an organism’s ecological niche.
Some species are generalist, having in some sense a broad niche (large
range of environmental tolerance, broad diet, etc.). Most species, how-
ever, are probably specialists in at least one sense. This is curious from an
evolutionary perspective; we can imagine that a wide ecological niche
would be extremely beneficial from a fitness perspective, allowing an
organism to survive unexpected changes in its environment. The problem
then is why so many species apparently forsake such benefits, leading to
a great subdivision of resources and environments among species. The
widespread evolution of ecological specialization is at least a partial
answer to the question ‘why are there so many different kinds of animals
[and plants]’ (Hutchinson 1959).
It is conceptually useful to distinguish two types of niche; the ‘fundamental’
niche was defined by Hutchinson as the set of environments in which
a species could maintain a positive growth rate (i.e. where it could in theory
persist). The smaller, ‘realized’ niche is the subset of the fundamental niche
actually occupied in nature. The fundamental niche is governed by an organ-
isms’ adaptation to the environment in terms of morphology and physi-
ology. Some of the fundamental niche may then not be used, either due to
constraints, such as dispersal ability, or because the organism displays plastic
behavioural rejection of some habitats.
Understanding the evolution of the realized niche therefore demands
that we (1) understand the extent of the fundamental niche through genetic
variation for the preference and exploitation of different numbers of environ-
ments, and (2) understand adaptive decision-making (see Chapter 7) that
98 DISCOVERING EVOLUTIONARY ECOLOGY

narrows the range of environments actually utilized. We will first explore the
very rich body of theory on these subjects before proceeding to the evidence.

9.1 Evolution of performance under trade-offs

The first attempt to model the evolution of the fundamental niche was
that of Levins (1968). He devised a model that should seem familiar, for in
structure it is very similar to the model of the evolution of co-sexuality by
Charnov et al. (1976) (Chapter 5). Imagine an organism faced with two
habitats, A and B, of differing frequency. Its fitness in the two habitats is
defined by a trade-off, such that when its fitness in A is high, its fitness in B
will be lower, and vice versa, a situation known as antagonistic pleiotropy.
A severe trade-off is of a concave shape, but a convex trade-off is less severe,
such that at best an organism which is very fit in B will also be pretty fit in A
(Figure 9.1). Now suppose that the organism wants to maximize its fitness in
the environment as a whole: what point on the fitness curve is best for it?
If an organism chooses to maximize fitness in a single habitat, at the expense
of fitness in other habitats, it will by definition be a specialist. If fitness
is approximately the same in all habitats, it will be a generalist. The basic

Convex
Fitness in habitat B

Concave

Fitness in habitat A

Fig. 9.1 Levin’s (1968) model of the evolution of specialization. Fitness in two habitats may be described
by a trade-off, which may be convex or concave. A convex trade-off leads to the evolution of
a generalist, which can use both habitats quite well (black circle), because its total fitness from
both habitats is greatest then. A concave trade-off leads to specialization on the most common
habitat (white circles), because any alternative generalist would achieve reduced total fitness.
 A WORLD OF SPECIALISTS 99

solutions are pretty intuitive: if the trade-off is concave then organisms evolve
to maximize their fitness in one habitat, whichever is the most common. If the
trade-off is convex, such that a generalist is almost as fit in both habitats as
a specialist, then as long as both habitats are reasonably common, generalists
will evolve.
According to Levin’s model, the evolution of specialism is thus predicted
to rely on the presence of a fitness trade-off, the more severe the better, and
will also depend on the frequency of alternative habitats, with increasing bias
in frequency of one habitat tending to lead to specialism in that habitat.
Of course, this model just considers the fundamental niche, and is not
explicitly genetic.

9.2 Other single species models

Most other models considering the evolution of the fundamental niche have
been explicitly genetic (reviewed in Hedrick 1986; Futuyma and Moreno
1988). In general, they consider quite simple genetics, such as one locus
and two alleles, each giving higher fitness in one habitat than the other.
They aim in general to predict under what circumstances the population
will become fixed for one allele (a specialist population) or retain both
(a generalist population), the former case being a ‘monomorphism’, and the
latter a ‘polymorphism’. Of course for those interested in specialization,
the circumstances favouring monomorphism are the subject of interest. The
same models, however, are interesting to another group of scientists: those
interested in speciation. They are interested in the conditions that favour
polymorphism (Chapter 1). The reason is that once a polymorphism has
arisen, with some individuals in a population favouring one habitat and
others favouring another, that might eventually lead to the generalist species
splitting into two specialist species (Maynard Smith 1966). We will consider
a case where speciation is likely to occur in this chapter, and another where
speciation already has occurred in Chapter 12.
The predictions of the one locus genetic models are also quite intuitive: if
there is temporal turnover in the habitats, a stable polymorphism can arise.
This is most people’s intuitive reasoning for the advantages of generalism:
that it allows an organism to survive changes to its habitat. If, however, there
is spatial variation in the distribution of habitats, which is normal, then
whether a polymorphism evolves depends on how fitness is affected by
population density. If the density of the entire population affects fitness,
regardless of which habitat they are in (known as ‘hard’selection), specialism
is the normal outcome. If, however, it is the density of individuals in each habi-
tat patch that affects fitness (known as ‘soft’ selection), then a polymorphism
100 DISCOVERING EVOLUTIONARY ECOLOGY

can be stable. The reason is again intuitive, since the fitness of each genotype
will be frequency dependent, favouring individuals that exploit unoccupied
habitats, just as in the ideal-free-distribution model (Chapter 8).
Trade-offs in performance (or preference) in the different habitats are
central to all the above models: they assume that a ‘jack-of-all-trades is
master of none’. Recently, however, attention has turned to models that do
not rely on this assumption. The stimulus, as we shall see later, comes from
the rather weak empirical support for the existence of fitness trade-offs.
For example, Whitlock (1996) imagined competition between a specialist
and a generalist, whose fitness was initially the same in the specialist habitat.
However, because it experiences only one habitat, the specialist would more
rapidly accumulate alleles beneficial in that habitat due to increased strength
of selection. Conversely there would be lower accumulation of deleterious
alleles. Thus, the specialist’s fitness rises faster than the generalist’s in
that habitat, such that it becomes the superior competitor over time.
Specialization thus enhances the speed of adaptation and reduces genetic
load: the same long-term advantages invoked for the maintenance of sex
(Chapter 2). A number of other models have since explored specialization
without trade-offs (reviewed in Futuyma 2001). All differ from the trade-off
models in concentrating more on the constraints imposed by the evolutionary
process itself.

9.3 Niche evolution under competition

Whitlock’s model invokes competition between species. A suite of other
models have investigated how competition affects niche evolution (e.g. Slatkin
1980; Taper and Case 1985). In many of these models, a species is assumed to
display a character that affects its niche use, and that this is under genetic
control. The extent of competition between individuals depends on the
overlap in niche use. The models run in two phases: first competition is
applied and the fitness of the different phenotypes is calculated. Then
quantitative genetics is applied to calculate the new phenotypes after
reproduction has taken place. A carrying capacity is also specified for each
environment that determines how the level of competition relates to the
number of individuals using that part of the niche.
When only one species is present, it evolves so that its niche gives the
highest carrying capacity under intraspecific competition (Figure 9.2).
When two species are present, displacement of the trait occurs through
intraspecific competition. Two general outcomes are possible: character
divergence and character convergence. Character divergence occurs when
the mean phenotypes of the species move away from each other, because the
 A WORLD OF SPECIALISTS 101

(a) (b)

Resource
distribution
Abundance

Abundance
Species 1 Species 1 Species 2

Niche width Niche width

Type of resource Type of resource

Fig. 9.2 Reduction of niche width by interspecific competition. (a) When just a single species uses
a resource, the average niche position is determined by the most abundant resource type.
b) A competitor species will tend to displace the resident, reducing the resident’s niche width.

fitness of the phenotypes that overlap is reduced. However, more rarely char-
acter convergence can occur. Two circumstances that favour this are (1)
when one resource is essential for both species, such that it pays to dominate
that resource, and (2) if one resource is particularly abundant while others
are limiting.
Thus, competition between species can lead to character divergence
among sympatric species, and subsequent specialization, or convergence
depending on the circumstances. Of course, there is no reason why competi-
tion need be the only interspecific interaction that can affect niche evolu-
tion; organisms might, for example, become specialized because their
fitness is reduced in some habitats via predation or parasitism. Historically
though, such interactions have yet to receive their due share of theoretical
attention.

9.4 Reduction of the realized niche by decision-making.

Once the genetics of the fundamental niche have been brought to equilib-
rium, another set of processes can reduce the actual set of resources used.
These include individual decision-making, which, though it acts within
the constraints of genetics, and hence cannot extend the fundamental niche,
can reduce it. In general, the solutions to the optimal decisions are to be
found within the realms of behavioural ecology theory (Chapter 7). Foraging
models that consider the optimal use of resources are particularly appropriate,
and we will consider one more such model here, the optimal diet model,
which evolved from work by MacArthur and Pianka (1966). This, like the
marginal value theorem, is a rate maximization model, and assumes simply
that two resources are available at different frequency in the environment,
102 DISCOVERING EVOLUTIONARY ECOLOGY

(a)

(b)

Fig. 9.3 Optimal diet selection. An individual arrives at a low value resource (grey circle). Should it use
it or move on in the hope of finding a better resource (large circles)? When high value resources
are abundant, it is best to move on, because of the opportunities that are lost by staying (a),
but when good resources are scarce, it loses nothing by staying (b).

which give different fitness returns (as a result of the outcome of the evolution
of the fundamental niche). The resources have different handling times
(times taken to exploit them). On encountering the lower quality resource,
should an organism accept it or wait until it finds the next higher quality
resource? The solution is simple: it should accept it as long as if when doing
so it does not gain energy at a lower rate than if it searched for and used the
higher quality resource. Therefore, if the better resource is rare or has a larger
handling time, or is not much better than the worse resource, all these favour
acceptance of the worse resource (Figure 9.3). However, if the best quality
resource is far superior, common, or takes little time to handle, then selection
favours rejection of the lower quality resource and hence reduction of the
realized niche.
Models based on such ‘time limited dispersers’ have also been written for
egg-laying taxa and those selecting habitats to develop in, which for some
taxa, such as most herbivorous insects, is the same thing (see Jaenike 1990;
Mayhew 1997). The general predictions of these and similar models are that
specialization (reduction of the realized niche) is favoured by (1) abundance
of the high quality resource; (2) low density of competitors; (3) low variability
in resource abundance; (4) longevity; and (5) high egg loads. Unsurprisingly
these predictions are very similar to those that predict a specialized funda-
mental niche, although the variables explored do not always overlap. How do
all these predictions stand up to data?
 A WORLD OF SPECIALISTS 103

9.5 Evidence for assumptions and predictions

Most of the evolutionary models for fundamental niche evolution assume
antagonistic pleiotropy, in other words negative genetic correlations between
fitness in different habitats.Some negative fitness trade-offs have been shown,
and these undoubtedly enhance the process of specialization. For example,
pea aphids with a genetic propensity for high fecundity on clover have a
genetic propensity for low fecundity on alfalfa, and vice versa (Hawthorne
and Via 2001). Rather excitingly, performance on those plants is also cor-
related with preference for those plants, and could represent an incipient
stage in speciation. In general, however, most genetic correlations are not
significantly negative and most are close to zero, indicating that high fitness
in one environment is relatively independent of fitness in other environments.
There are numerous potential reasons why this result may be rather arte-
factual, to do with the difficulties of conducting flawless and powerful
experiments to measure genetic correlations. On the other hand, the fact
that most genetic correlations are near zero, rather than positive, implies that
different genotypes favour different environments, rather than a single geno-
type being best in all environments. The former is needed for the alternative
‘neutral’ models. In addition, deleterious mutations with habitat-specific
effects are known in Drosophila, which is an assumption of many of the
neutral models (Kawecki 1994).
There is plenty of evidence that variation in resource abundance can favour
specialization. Several examples of rapid niche evolution have occurred as
introduced species have invaded new habitats and become abundant
(Thompson 1998). In most cases this involves evolution of preference for the
new habitat. One of these species is the Edith’s Checkerspot butterfly
Euphydryas editha (Figure 9.4), which has been the subject of a long-term
study in and around the state of California (Singer et al. 1993). Two popula-
tions, ‘Rabbit’ and ‘Schneider’ have shown long-term changes in preference
to lay eggs on novel host plants that have become more abundant in these
habitats due to human activity. In both cases the preference differences
were heritable, and at Schneider this preference was also associated with
elevated growth rates on the preferred host (Singer et al. 1988). In addition
to such genetic changes, which represent changes in the fundamental niche
of a population, there is also plenty of evidence from this and other
phytophagous insects that selection of host plants on which performance is
poor depends on the abundance of hosts on which performance is high
(Mayhew 1997).
Recently, Prinzing (2003) has shown that arthropod species living on tree
trunks in Germany are more likely to occupy a large number of microhabitats
if they have long generation times, move faster, and spent more of their life
104 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 9.4 Edith’s Checkerspot butterflies, Euphydryas editha, mating. The butterfly has rapidly evolved to
use new host plants as the abundance of hosts in the environment has shifted. Photo courtesy
of Camille Parmesan.

on the trunks, giving them better search capabilities and opportunities.

The trends might have come about via evolution of the fundamental niche or
from plastic behaviour: both sets of theory make the same prediction and only
genetic or behavioural studies could separate the two possibilities. Similar
correlations have been found between the variability of a habitat and host
plant range across herbivorous insect species (Brown and Southwood 1983).
The evidence that interspecific competition causes niche evolution comes
from studies of well-defined morphological characters that are functionally
linked to resource use—such as the beaks of Darwin’s finches (see Figure 12.4).
The majority of cases (there are at least 75) cite as evidence that the characters
are more divergent when species are in sympatry than when they are allopatric.
However, in most cases the case for character displacement by competition is
far from complete. Perhaps the very best evidence comes from experimental
work by Schluter (2001) on body size and shape of three-spined sticklebacks.
In lakes where one species exists, the species is generalized and feeds both in
the water column and on the lake bed. When two species coexist, however,
they specialize with a thin species in the water column, and a larger rounded
species on the lake bed (Figure 9.5).In an elegant series of experiments in artificial
ponds, Schluter was able to show selection operating in the predicted direction
when an intermediate form was made to coexist with specialists of either type:
if it faced benthic competition, the more benthic-like of the intermediate
 A WORLD OF SPECIALISTS 105

Fig. 9.5 Specialization under competition. The picture shows sticklebacks, Gasterosteus spp., from Enos
lake on Vancouver Island, British Columbia, Canada. From top to bottom are shown: limnetic
male, limnetic female, benthic female, benthic male. The benthic forms are larger and have
deeper bodies than the limnetic forms. Photo courtesy of Dolph Schluter.

phenotypes had reduced success, thus selection favoured evolution towards a

more pelagic existence, and the opposite occurred when a pelagic competitor
was introduced. This disruptive selection is a likely mechanism of speciation
(see Chapter 12).
In general, the list of character displacement examples, although admit-
tedly preliminary, is remarkable for the absence of character convergence, for
the fact that the species involved are closely related, and for the fact that
carnivores are very well represented. These may, of course, be artefacts of
observer and publication bias, but they may also have a biological basis, a
possibility that is intriguing. Can other interspecific interactions cause spe-
cialization? There are some intriguing pieces of evidence. There is a cross-
species association between host plant range and sequestration of nasty
chemicals which serve as defences against predators. This is consistent with
predation-driven selection of specialization (Dyer 1995).
There is little evidence that intraspecific competition within habitats can
widen fundamental niche use, but certainly the realized niche tends to be
106 DISCOVERING EVOLUTIONARY ECOLOGY

greater when densities are high in a large number of species. For example,
Pemphigus aphids preferentially occupy leaves of their host plants with a good
vascular supply which they defend despotically. Subsequent aphids are faced
with having to occupy less productive locations on the plant, and at higher
densities the range of microhabitats used increases (Whitham 1980). Similar
observations could be quoted for almost any territorial species.
In general then, a number of extrinsic ecological and intrinsic biological
forces are both predicted to favour restriction of the fundamental and realized
niches, and theory and evidence are coming together. It cannot be claimed
that we have a good understanding of the relative importance of these
forces yet, but a series of hypotheses with some support bodes well for the
future.

9.6 Evolutionary trends in specialization

Given that we can hypothesize circumstances that might favour specialization
over generalization and vice versa, how might the frequency of specialization
and generalism vary over evolutionary time? There are three commonly cited
hypotheses. First, specialists may be more prone to extinction than general-
ists. Second, specialists may be more prone to speciation than generalists.
Third, lineages may change anagenetically from generalist to specialist more
frequently than the reverse.The first hypothesis is intuitive.A generalist seems
more likely to maintain a positive rate of increase in the face of environmental
change, and a species with a large geographic range or population size would
seem less likely to be prone to extinction causing events. The prediction is
supported by metapopulation models (see Chapter 13), where the greater the
density of habitable patches, the less likely the population will go extinct. The
second hypothesis is motivated by the observation that a large number of
species, especially those in adaptive radiations are ecologically specialized.
The third hypothesis, an anagenetic trend to specialization, was hypothe-
sized by Simpson (1953) to be the dominant trend in adaptive radiation. It
could come about for a number of reasons: resource partitioning through
competition, adaptation might be quicker through specialists than generalists
(see above); specialization might allow speciation through host switching (see
above); specialization might canalize evolution through suites of co-adapted
traits acting as constraints (Schluter 2000).
Does specialization tend to increase the rate of extinction? Here the evidence
strongly supports the hypothesis. Species persistence times in the fossil record
are higher for generalist clades in crinoid worms, marine gastropods, marine
bivalves, shrews, and antelopes. In recent (anthropogenic) extinctions, there
 A WORLD OF SPECIALISTS 107

is a tendency for increased extinction risk with decreasing geographic range

in primates, carnivores, and birds, and increased extinction risk with increas-
ing dietary specialization in hoverflies, reptiles, birds, and primates (Purvis
et al. 2000), and with habitat specialization in a number of taxa (Fisher and
Owens 2004). Does specialization tend to increase the rate of cladogenesis?
Although there is some evidence consistent with this observation, there is
also potential counter-evidence. For example, Owens et al. (1999) found
that bird species richness is positively correlated with dietary and habitat
generalism. This, however, might be due to a lower extinction rate or a higher
speciation rate, or both.
As regards transitions, trends are seen across phylogenies for increases in
specialization over time, increases in generalization and for variable transition
rates between the two. Schluter (2000) in a survey of 20 plant, invertebrate,
and vertebrate phylogenies found no overall tendency for the ancestor of the
radiation to be a specialist rather than a generalist or vice versa. Some taxa,
such as the Galapagos finches, had a generalist ancestor, while others, such as
the Hawaiian silverswords, had a specialized ancestor. Only in one group, the
genus Aquilegia, was there a tendency from a generalist to specialist state.
In the other groups there was no detectable trend.
The evidence to date suggests that generalization reduces extinction rates,
that degree of specialization variably affects speciation rates, and that trans-
itions between specialists and generalists are also variable. The next chapter
asks what all these species are specialized at doing, and how that evolves.

9.7 Further reading

Futuyma (2001) and Schluter (2001) give introductions to single species
and multi-species questions. Futuyma and Moreno (1988) is older but still
useful. Mayhew (1997) reviews realized niche models in the context of
phytophagous insects. Schluter (2000) reviews macroevolutionary trends.
10 The good, the bad, and the commensal

The web of our life is of a mingled yarn, good and ill together.
William Shakespeare

The present chapter deals with another feature of interactions between

species: the degree of antagonism. Interactions between two species can be
classified according to the fitness effects of the interaction on each species.
Most people’s understanding of ecological communities is that they are full
of exploiter–victim relationships, where one species benefits at the expense
of another (⫹,⫺). Predation, herbivory, and parasitism are examples, and
are of course ubiquitous in communities: they are the building blocks of
food webs. Competition involves negative effects on both parties (⫺,⫺) and
is also common between members of the same trophic level. Again, such
relationships are to be expected in a Darwinian world where individual’s
primary concern is its own reproductive success, rather than that of
another individual of a different species. There are also more nearly neutral
relationships, such as commensalisms (⫹,0) and amensalism (⫺,0). In these
one species is affected by the interaction, but the other is not affected.
Commensalism is common where one species incidentally benefits from
another species’ activity: humans have their own community of such species
that includes house sparrows, house mice, house spiders, and most garden
weeds, which have little effect on human fitness most of the time.Amensalism
is found when one species consistently excludes another (asymmetric
competition), as found in many plant and insect communities.
Finally, there are mutualistic relationships (⫹,⫹) where both parties
benefit from each other. Mutualistic relationships of this kind are ubiquitous
and of great ecological and evolutionary significance (Chapters 2 and 3).
They include plants and their pollinators; plants and their fruit and seed
dispersers; plants and their mycorrhizal fungi; corals and their symbiotic
algae; lichens (symbioses between fungi and algae); leguminous plants and
rhizobia bacteria; and many animal species with nutritional or digestive
symbioses with microbes; including cows, termites, aphids, and humans. Put
simply, these mutualistic relationships are essential components of most
modern communities.
 THE GOOD, THE BAD, AND THE COMMENSAL 109

One striking fact that has emerged from the study of species interactions
is that they must sometimes evolve pretty rapidly: different populations of
the same species can be either mutualistic or parasitic, and closely related
species likewise. In contrast, other relationships are apparently ancient and
stable. What conditions favour such changes or else help maintain the status
quo? Our understanding of these questions has particularly benefited from
two types of study; studies of parasite–host relationships and studies of
the evolution of cooperation, which, as we will see, share considerable
conceptual overlap.

10.1 The evolution of parasite virulence

Pathogens and parasites by definition exert negative effects on the fitness of
their hosts. However, they vary in the severity of those effects, such that some
diseases allow their hosts to live a long time and suffer few adverse effects,
while other hosts are almost instantly overcome and die. Why this variety?
Earlier, conventional wisdom led many biologists to believe that parasites
would eventually evolve so as not to harm their hosts, on which they depend
for their persistence. But in fact the truth, as in the case of HIV (Chapter 8) is
often far from this: some parasites evolve to become more virulent over time.
So the story must be more complex than the conventional wisdom would
suggest.
The major breakthrough in understanding was made by Anderson and
May (1982). They derived a general relationship for the fitness of a horizon-
tally transmitted pathogen whose population is at equilibrium with that of
its host. In such cases, lifetime reproductive success, R0, is an appropriate
measure of fitness:

␤(N)
R0⫽
␮⫹␣⫹␯

where ␤ is the transmission rate of the disease, dependent on host density N,

␣ is the parasite induced mortality rate of the host (virulence), ␯ is the host
recovery rate, and ␮ is the death rate of uninfected hosts.
If all these parameters are independent of each other, it is clear that viru-
lence should be zero, so as to make R0 as large as possible. This is the sense
behind the intuitive, but discredited, conventional wisdom. However, this
assumes an absence of genetic correlations between traits. In particular, viru-
lence and the transmission rate of the disease might be positively related.
One likely source of such a correlation might be that transmission relies on
110 DISCOVERING EVOLUTIONARY ECOLOGY

(a)

(b)

Fig. 10.1 Virulence and host density. High virulence (black circles) kills off hosts (large circles) rapidly, but
that does not matter when hosts are common and the pathogen is easily transmitted (a).
However, when hosts are rare, pathogens must be less virulent (fewer black circles) to keep
their host alive long enough to encounter susceptible hosts (b).

rapid multiplication of the disease organism within the host to create many
disease propagules, or inducing adverse symptoms in the host, such as
coughing, that aid transmission. If this is the case, then the optimal virulence
for the parasite will be such that the marginal gains of virulence, in terms of
added propagule production and transmission, balance the marginal costs,
in terms of reduced host survival (Figure 10.1).
The best example of evolution of virulence so as to balance these costs and
benefits is that of the rabbit myxoma virus. This virus is endemic to South
America but was released into Australia in 1950 to control rabbits. The
rabbits had been introduced to Australia by Europeans but had multiplied to
plague proportions in the absence of natural predators and disease. On
release the myxoma virus did all that had been hoped; it was 99% fatal and
reduced rabbit numbers to one-sixth their previous levels in just two years.
After just a few years, the effectiveness of the disease had fallen, causing
between 70 and 95% fatalities in infected hosts, with a reduction in the
effectiveness of control (Fenner and Ratcliffe 1965). It is known that this
involved evolution of the virus and not just host evolution or an acquired
immunity on the part of the rabbits, because virulence was assessed by
exposing the disease to control rabbits from populations that had never
encountered the virus.
It is likely that the reduction in virulence seen in the myxoma virus in
Australia actually had beneficial effects on transmission rates of the disease.
Infected rabbits become weak and immobile and develop skin lesions that
aid transmission of the virus to other individuals through mosquito vectors
 THE GOOD, THE BAD, AND THE COMMENSAL 111

(Fenner and Ratcliffe 1965). Furthermore, the mosquito vectors are rather
seasonal and thus infected hosts must remain alive for extensive periods
for the disease to persist. Living infected individuals are thus efficient trans-
mitters of the virus. Although the virus may gain in the short term through
more rapid overwhelming of the host, a dead host is no use as a source of
transmission.
Hence, selection between hosts infected with strains of the diseases that
vary in virulence can cause virulence to evolve to varying levels depending
on the strength of the correlations between traits, especially between
virulence and transmission rate. There is, however, another potential source
of selection on virulence; selection between strains of the disease within a
single host. There are two ways in which an individual can acquire more than
one strain of a disease. First, an individual can acquire the disease more than
once from different sources. An alternative is if the disease mutates and
diversifies within the host following a single infection event. Multiple infec-
tions can then select for increased virulence. The reason is that now there is
competition between unrelated individuals for a single resource. A disease
strain that is prudent and is relatively benign to its host will be out-competed
by a disease strain that is less prudent but exploits the resource before others
can profit from it.
This tendency for a resource shared by unrelated individuals to be overex-
ploited is known as the ‘tragedy of the commons’ (Hardin 1968) and is best
known in the context of human overexploitation of marine fisheries. When
several individuals exploit a shared host and there is a trade-off between vir-
ulence and competitiveness it can be shown that the evolutionary stable vir-
ulence strategy is 1 ⫺ r (Frank 1996) where r is the average coefficient of
relatedness between individuals (Chapter 7). Multiple infections will tend to
decrease r and thus increase virulence. HIV infections consist of multiple
strains, and should select for increased virulence during the course of an
infection and may be the ultimate cause of AIDS onset (Bull 1994).
Therefore, trade-offs between virulence and transmission rates, combined
with selection between hosts, and selection within hosts between disease
strains can both contribute towards the evolution of virulence.A third major
variable is known to affect virulence: vertical as opposed to horizontal trans-
mission. If pathogens are vertically transmitted, they are passed from parents
to their offspring down the generations. If pathogens are horizontally
transmitted they are able to pass between unrelated individuals within a
single generation (Figure 10.2). If the proportion of new infections through
vertical as opposed to horizontal transmission is high, it pays the pathogen to
allow the host to reproduce as much as possible, so that the disease has the
opportunity to infect a larger number of host offspring. In this sense the
interests of the pathogen and those of the host become more closely aligned
112 DISCOVERING EVOLUTIONARY ECOLOGY

Parent 1 Parent 2

Horizontal
Generation 1 (Parents)

Horizontal
Vertical

Generation 2 (Offspring)

Offspring of parent 1 Offspring of parent 2

Fig. 10.2 Horizontal and vertical transmission of parasites (black circles) between hosts (large circles).

through horizontal transmission. Frank (1996) has also showed that a

further selective pressure may help reduce the virulence of vertically
transmitted pathogen: as transmission becomes more and more exclusively
vertical, the chances of multiple infections become very low. Thus, selection
for virulence within hosts becomes weaker and weaker.
Is there evidence that vertical transmission selects for benevolence towards
hosts? The answer is yes. In a study on fig wasps and their nematode parasites,
Herre (1993) found that variation in the opportunity for horizontal as
opposed to vertical transmission across species was correlated with the
virulence of the nematodes. The wasp species develop in figs that the females
must enter in order to lay their eggs. Wasp species differ in the typical
number of females that will enter a given fig: sometimes only a single female
will enter a fig (the single foundress situation), but at other times many
females will enter a single fig (multiple foundresses, see also Chapter 5).After
laying her eggs, the female wasp dies inside the fig and her nematodes are
released to infect other fig wasps. In the single foundress situation, the
nematode parasites of the fig wasp are exclusively transmitted from mother
to offspring since there is only one adult wasp per fig. However, if there are
multiple foundresses, the nematode progeny can infect the offspring not just
of their former host, but those of other wasps, a situation more akin to
horizontal transmission. Herre measured the virulence of the nematodes on
the wasps by comparing the fecundity of wasps in fig fruits infected with
nematodes or free of them. Across 11 species of wasp, he found a positive
correlation between the proportion of fruits that were multiple foundress
and the virulence of the nematodes. Frank (1996) has since argued that the
variation in virulence is most likely caused by changes in the frequency of
 THE GOOD, THE BAD, AND THE COMMENSAL 113

multiple infection resulting from differences in the mode of transmission,

rather than the mode of transmission per se.
So far our discussion of the evolution of antagonism has been restricted to
the relative virulence of pathogens and parasites. None of the organisms thus
far mentioned are known to have positive fitness effects on their hosts, but
vary in the extent of the negative effects they exert. In addition, because we
have been focussing on the evolution of parasitic organisms, we have also
been concerned exclusively with so-called symbiotic interactions: those that
concern long-term associations between the two species relative to the life of
one of them. We shall now widen the discussion to include mutualistic
relationships including those that do not involve any permanent lifetime
association between the two partners.

10.2 The origins and maintenance of mutualism

Mutualisms (e.g. Figures 10.3, 10.4) are those interactions in which both
partners accrue increased fitness relative to individuals that do not engage in
it. How might selection favour the expression of traits that help other indi-
viduals of a different species, given a prior relationship that was more nearly
neutral or even antagonistic? Frank (1997) has suggested that two processes
act to help the spread of such traits: first, an initially high level of expression
on first meeting in both partners. Second, spatial association between pairs
of individuals with positive synergism so that benefits can be returned to the
donor (see also Yamamura et al. 2004).
Imagine that when two species first meet, both species possess a trait that
provides a small benefit to another species at some cost to itself. If these bene-
fits are below a threshold value, selection will favour lowering expression of
the trait, such that a mutualism does not arise (Frank 1997). But if the
expression of the trait is above a certain threshold, the synergistic effects
between species can push expression of the trait to an equilibrium at higher
levels of expression. If the benefits of the traits are high relative to the costs,
the threshold value is lower and the equilibrium value higher.
Association in space is the second factor that can enhance the develop-
ment of mutualism. Two types of association are in fact important: associa-
tions within populations of each species, and associations between species.
Associations within species are important to raise the relatedness between
individuals and prevent antagonism between them, as seen earlier in the
chapter. Associations between species make it easier for the mutualism
threshold to be passed—once passed, continued spatial association may not
be required as long as associations within species remain fairly high. The
mechanisms of spatial binding between different species are various. For
114 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 10.3 A plant–mycorrhizal symbiosis. The picture shows a root of Plantago lanceolata, with the
fungus Glomus hoi (stained dark) growing through it. The dark blob (50 ␮m across) is a vesicle,
thought to be a storage organ for the fungus. The fainter rectangular structure to the right of
the vesicle is an arbuscule, believed to be the site of nutrient exchange between the fungus
and the plant. Photo courtesy of Angela Hodge.

example, aphids pass their mutualistic bacteria, Buchnera (Figure 10.4),

from mother to embryo (vertical transmission). Mitochondria and
chloroplasts are also transmitted vertically through gamete cytoplasm
(Figure 10.2). Spatial binding of a different nature occurs between genes in a
genome via linkage, and that contributes towards cooperation between
genes in a cell. These latter examples illustrate that mutualism has occurred
not just between what we now think of as different species, but that it was
involved in the major transitions leading to modern species (Chapter 2). In
many cases, the theory relevant to interspecies mutualisms has developed
from thinking about these major transitions (see Maynard Smith and
Szathmáry 1995; Frank 1997).
The models of Frank predict a starting threshold degree of mutual benefit
below which mutualism will not evolve, and many other models make the
starting assumption that mutual synergism already exists. This assumption
is rather restrictive, since in many and perhaps most cases mutualism has
evolved from previously antagonistic relationships. Recently, Law and
Dieckmann (1998) have developed a model in which a permanent symbiotic
unit with purely vertical transmission can develop from an antagonistic
 THE GOOD, THE BAD, AND THE COMMENSAL 115

Fig. 10.4 Cells of the symbiotic bacterium Buchnera, within the cytoplasm of a type of specialized
cell, bacteriocytes, within the body cavity of the pea aphid, Acyrthosiphon pisum. The
Buchnera are about 4 ␮m across. Buchnera are vertically transmitted and provide essential
amino acids to the aphid which are absent from their diet of plant sap. Photo courtesy of
Angela Douglas.

exploiter–victim relationship (Figure 10.5). The model specifies a mortality

cost to the victim in the free-living state of reducing the amount of resource
donated to the exploiter in the symbiotic state. With no such cost the
symbiosis breaks down: victims reduce their flow of resources to their
exploiters until the species no longer interact. When the cost is sufficiently
high, however, there is selection on both parties to increase the degree of
coupled replication until this is the sole form of reproduction, despite the
fact that one species continues to exploit the other. Such a system is one of
mutual dependence, since both species now do better inside than outside the
relationship, despite the fact that only one species transfers resources to the
other. An experiment by Jeon (1972), involving amoebae and a parasitic bac-
terium, has shown the evolution of such mutual dependence in the presence
of a viability cost to the amoeba in the free-living state, and despite the fact
that that viability of the amoeba was no higher in symbiosis than before.
Hence, mutualism can arise in several ways from other kinds of interaction,
and certain conditions make the origin of mutualism more likely. Once
mutualism has arisen, however, a subsequent problem is how such mutu-
alisms can maintain their stability over time. Where a relationship involves
benefit as well as cost, the stability of the relationship is susceptible to
invasion by mutant individuals that receive benefit from the other species
116 DISCOVERING EVOLUTIONARY ECOLOGY

(a)

(b)
D
D
D

(c)

D D

Fig. 10.5 The merger of lineages through exploitation. In stage (a), two species coexist, and one
(large circle) occasionally exploits the other (black circle, arrows show flow of resources) in a
symbiosis. However, the symbiont can also occur in the free-living state. In stage (b), the
symbiont develops defenses (D) that stem the flow of resource to the host somewhat
(shorter arrows). In stage (c), the free-living form of the symbiont becomes extinct because
of the cost of defences in the free-living state, giving rise to a permanent exploitative
symbiosis.

Partner
Cooperate Defect

Reward Sucker
Cooperate 3 0

You

Temptation Punishment

Defect 5 1

Fig. 10.6 The Prisoner’s dilemma. Four possible interactions are shown, and the pay-off to you in
arbitrary fitness units. If the partner cooperates, it is best to defect (the temptation); if the
partner defects, it is best to defect (the punishment) rather than cooperate (the sucker). Result:
everyone always defects, cooperation is absent.

without donating anything themselves. Theoretically, such a situation is a

Prisoner’s dilemma game (Figure 10.6), in which for each party it is best to
defect whether the partner cooperates or defects. Thus, the Prisoner’s
 THE GOOD, THE BAD, AND THE COMMENSAL 117

dilemma becomes a trap in which cooperation cannot be a stable outcome.

Indeed, there is considerable evidence that cheating is a common, though
not exclusive outcome of mutualistic relationships. For example, many
insects rob flowers of nectar without affecting pollination. The surprise is
not that such cheaters exist, but that some mutualisms have survived for a
long time without being destroyed by such processes—the relationship
between figs and their pollinating wasps, between corals and algae, between
mycorhizal fungi and terrestrial plants are all ancient. What maintains such
mutualisms?
One possible answer is once again vertical transmission. Clearly, however,
there are still many mutualisms that are not transmitted vertically. These
include mycorrhizal associations (Figure 10.3), the rhizobia bacteria of
legumes, pollination interactions (Figure 10.7), seed dispersal interactions,
most gut floral associations, and cleaner associations. Many of these
interactions also involve multi-species associations, which are again
predicted to favour exploitation of the mutualistic partner. Clearly, there
must be additional mechanisms at work that can maintain the stability of
cooperation across species (Wilkinson and Sherratt 2001). There are several
theoretical possibilities, but with somewhat limited supporting evidence at
present.
One possible mechanism that might maintain the stability of a mutualism
is if the beneficial donations are not costly to the donor. In fact, many
mutualists donate products, or by-products, that are available in excess or are
not limiting to them: for example, plants donate carbon, which is normally
not a limiting nutrient, both to their mycorrhizal fungi and to their
pollinators in the form of nectar. Lack of costs means that mutualists are
playing a game more closely akin to ‘nice guys win’ in which it pays both
partners to cooperate even in the short term because defection does not
increase pay-offs.
Another potential stabilizing factor is if the partners can retaliate to defec-
tion (sanctions). Under such conditions, the decision to cooperate or defect
can be made conditional on the previous actions of the other partner. It is
then relatively easy to generate conditions under which persistent coopera-
tion is the stable outcome because the lifetime pay-off, as opposed to the
short-term pay-off, is now greater if organisms do not purely defect
(Maynard Smith and Szathmáry 1995; Wilkinson and Sheratt 2001). One
example comes from the Yucca—Yucca moth pollination mutualism
(Figure 10.7). Here the pollinating moth also consumes some of the Yucca
seeds. In cases where a larger number of the seeds are being consumed, some
Yucca species are able to selectively abort those developing fruits (Pellmyr
and Huth 1994). In this case therefore, the cost of overexploitation is paid by
the moth’s offspring, but that is just as effective a mechanism of selection.
118 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 10.7 Two moths, Tegeticula yuccasella, in a yucca flower. The moth on the left is pollinating the
flower, while the one on the right is laying its eggs at the base of the flower. Fruits infested with
many moth larvae are more likely to be aborted by the plant, a mechanism that may help to
maintain the mutualism by prevention of cheating. Photo courtesy of Olle Pellmyr.

Sanctions have recently also been demonstrated in the Rhizobium/legume

mutualism and may help maintain it there in the absence of some other
forces, such as vertical transmission (Kiers et al. 2003). In this case, the
legume appears to withhold oxygen from root nodules containing ‘cheating’
bacteria that are not providing sufficient nitrogen to the plant.
Control over mixing also occurs via uniparental inheritance of organelles
(Chapter 2), and via the unicellular zygote stage through which all
multicellular organisms pass during their life cycle, which contributes to the
stability of the body. Another form of control that selects against symbiont
exploitation is reproductive fairness (Frank 1997). In meiosis and mitosis
the replication of the chromosomes is rigidly controlled such that each has
parity with others and an equal chance of being present in the next genera-
tion. Under such rigid control the only way that a chromosome can increase
its own success is to work for the benefit of the group (the genome). In some
insects that play host to symbionts, rigid control of which symbionts
are transmitted to the next generation also occurs. In the sucking louse
Haematopus, symbionts are separated early in development into a ‘somatic’
group which never enter the next generation and are involved in the gut
symbiosis, and a ‘germ’ line which are destined to enter the next generation
(Buchner 1965). Again, the somatic group can only increase its own success
 THE GOOD, THE BAD, AND THE COMMENSAL 119

by working for the benefit of the group. A final way to enforce control is via a
policing strategy that directly combats cheats. This occurs in the case of mito-
chondrial cheats that cause cytoplasmic male sterility in plants (Chapter 5)
and whose effects are countered in some strains by nuclear genes.
Finally, while the evolution of cheating may sometimes lead to dissolution
of an interaction, or extinction of species, another possible outcome is stable
coexistence of cheater and mutualist. In fact stable coexistence is known
from many systems, such as the yucca/yucca-moth system, and figs and fig
wasps, both of which are host to non-pollinating insects. Ferriere et al.
(2002) have used an adaptive dynamics model (Chapter 8) to investigate this
possibility, assuming that symbionts compete for the resources received in
the relationship in an asymmetric way such that mutant individuals are
either more or less successful than residents under competition. They find
that under certain conditions, evolutionary branching can occur, leading to
the stable coexistence of a cheating and more mutualistic species.
Thus, a number of conditions may favour the origin and stability of
mutualistic relationships, some of which are consistent with empirical data,
albeit mostly anecdotal. The challenge is now to begin to test these ideas
more rigidly. In addition, that further theoretical and empirical research may
uncover additional alternative mechanisms for maintaining the stability of
cooperation between species. The next few years should be exciting in terms
of furthering our understanding.
Our discussion of the evolution of species interactions in the last two
chapters has largely considered only one partner in the relationship evolving
at a time. Obviously though, and especially in intimate or obligate
associations, each partner can serve as a selection pressure for the other. Such
‘co-evolution’ can then potentially lead to outcomes that differ markedly
from the consideration of only a single evolving species. Co-evolution will be
the subject of the next chapter.

10.3 Further reading

Ebert and Herre (1996) and Bronstein (2001) introduce virulence and
mutualism, respectively. Frank (1996, 1997) reviews the theory in depth
along with evidence.
11 Evolving together

In much of the last two chapters we attempted to explain the outcome of

ecological interactions by assuming one species evolving in relation to con-
straints provided by another, evolutionarily static, species. Of course there is
no reason why this process need be so one-way, and if it is not, then it is essential
that the reciprocal evolution of both species in an interaction be considered.
Such reciprocal evolution of interacting species is called co-evolution.
The aim of co-evolutionary studies is to predict or understand the
evolutionary dynamics and evolutionary outcomes that result from species
interactions. The subject actually has one of the longest histories of any in
evolutionary ecology, with basic concepts and theory dating to the 1950s
(see Thompson 1999). However, the intensity of research has only picked up
in the last two decades, aided, particularly, by the molecular and cladistic
revolutions. The evolutionary dynamics of interacting species do not read-
ily lend themselves to study. Although use of the fossil record has been
attempted, such studies on their own are often deficient in key data relating
to the interaction. By far, the most promising progress has been made with
living species, sometimes in combination with the fossil record.
To observe the dynamics and outcomes from living species, evolutionary
ecologists take two approaches: a longitudinal (historical) one that reconstructs
the past, normally using cross-species phylogenies to estimate the pattern of
changes over time. Sometimes, however, the evolutionary dynamics are suf-
ficiently rapid that long term field studies can detect them.A complementary
(transverse) approach is to observe the interaction at different points in
space, which highlights possible alternative long-term temporal outcomes.
Recently, it has been suggested (Thompson 1994, 2001) that the spatial
dynamics are also the key to understanding the temporal dynamics. Before
we proceed to examining that hypothesis, let us have a look at some of the
evidence that describes how co-evolution can proceed.
From several case studies of interacting species in the field, several alter-
native dynamical outcomes are now known or may be implied (Table 11.1).
They are doubtless just a small sample of the co-evolutionary universe. These
outcomes have previously been referred to as the ‘modes of co-evolution’
(Thompson 1989, 1994). I restrict myself here to modes that relate specifically
 EVOLVING TOGETHER 121

Table 11.1 Some alternative modes of co-evolution, following Thompson (1989, 1994)

Empirical model Example Properties

Diversifying co-evolution Maternally inherited One species stimulates

symbionts (Thompson speciation in another
1987)
Escape-and-radiation Umbelliferae and their Radiation in one lineage
co-evolution phytophages (Berenbaum occurs while it
1983) temporarily escapes the
interaction with another
Arms race Flower morphology and Directional change in
morphology of quantitative traits, perhaps
pollinators (Steiner and reaching stable or local
Whitehead 1990) equilibria
Co-evolutionary alternation Cuckoos and their hosts The identity of interacting
(Davies and Brooke 1989) species changes as a result
of an adaptive response in
one species which initially
decreases but later
increases the interaction
Mutual dependence Eukaryotes and their A mutualistic relationship
mitochondria/chloroplasts shows temporal stability
(Margulis and Bermudes with increasing
1985) specialization and
obligation
Co-evolutionary successional Trifolium repens and Local succession creates
cycles grasses (Turkington changing patterns of
1989) association between species
creating local or temporal
interactions in which
co-evolution occurs
Co-evolutionary turnover Body size of Anolis Asymmetric competition
lizards in the Caribbean creates cycles of invasion,
(Roughgarden and Pacala co-evolution, extinction,
1989) evolution, and reinvasion

to outcomes (rather than assumptions). They are distinguished from each

other by four characteristics (Table 11.2).

11.1 The number of interacting species

The first characteristic is the number of interacting species at any point in
time. In some interactions this is static. For example, mutual dependence
interactions (Figure 11.1) characterize mutualisms that become increasingly
intimate and obligate over time. This has occurred most notably with
the mitochondria and chloroplasts of eukaryotes, all three of which were
122 DISCOVERING EVOLUTIONARY ECOLOGY

Table 11.2 The characteristics that differentiate between the modes of co-evolution.

Mode of Number of Dynamics Dynamics Dynamics of

co-evolution interacting of species of traits antagonism
species richness

Diversifying 2 Radiation NS NS
during the
interaction
Escape-and- (1), 2, ⬎ 2 Radiation in Punctuated Increasing
radiation absence of equilibria antagonism
interaction
Arms race NS NS Dynamic, Increasing
directional, antagonism
local
equilibria
Co-evolutionary (1), 2, ⬎ 2 Stable Dynamic Increasing
alternation antagonism
Mutual 2 Stable Stable Reduced
dependence antagonism
Co-evolutionary NS Stable Stable Reduced
successional antagonism
cycles
Co-evolutionary (1), 2 Cycles of Dynamic Stable
turnover invasion and antagonism
extinction

Notes: NS indicates where the characteristic is not integral to the model definition. (1) refers to the
case where the interaction ceases for one or more species.

previously independent-living bacteria. Now, however, eukaryotes cannot

survive without their mitochondria or chloroplasts and neither can mito-
chondria or chloroplasts survive independently of eukaryotic cells.
Other modes of co-evolution, however, involve a change in the number
of interacting species over time. Escape-and-radiation co-evolution
(Figure 11.1) refers to the process that may have occurred in many
plant–herbivore interactions, whereby a plant develops a novel defence
against the herbivores which ceases the interaction temporarily. In the
absence of herbivory, the plants diversify into a greater number of species,
perhaps because incipient species can more readily establish. Eventually,
a herbivore species evolves a counter-defence, and this can then speciate into
the new niches provided by the previously resistant plants. Thus, the process
involves first a decrease in the number of species, and later a compensating
increase. The best evidence for this process comes from the Umbelliferae
family and their lepidopteran herbivores (Berenbaum 1983) (Figure 11.2).
A similar dynamic occurs in co-evolutionary alternation (Figure 11.1).
This occurs when a host evolves defences against a parasite, and in response
the parasite is selected to look for alternative unresistant hosts. Having
 EVOLVING TOGETHER 123

(a) (b) (c)

Fig. 11.1 Co-evolution and the number of interacting species (represented by the different shapes
connected by double headed arrows). In mutual dependence, this remains steady (a), In
escape-and-radiation co-evolution (b), one of the interacting species goes, extinct, the other
radiates and finally becomes host to another species. In co-evolutionary alternation (c), one
species is added to the interaction as another is removed.

(a) (b)
Freq

Niche

Fig. 11.2 Speciation and extinction resulting from co-evolution. In diversifying speciation (a), such
as between a host and a symbiont, interactions with alternative symbionts can lead to
differentiation of the host. In co-evolutionary turnover (b) one species is driven extinct during
the interaction (in this case via competition), but the resulting changes make the survivor
vulnerable to extinction itself (in this case by subsequent competitors).
124 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 11.3 A parsnip webworm caterpillar, Depressaria pastinacella, feeding in the flower head of a wild
parsnip plant, Pastinaca sativa. Parsnips and webworms represent an advanced stage in
escape-and-radiation co-evolution, with very nasty plant and very specialized herbivore. The
parsnip contains both linear and angular furanocoumarins, which the webworm is able to
detoxify. Photo courtesy of May Berenbaum and Arthur Zangerl.

found one, it eventually switches hosts entirely. Over time the old host may
eventually lose its previous resistance. This process has been suggested to
occur in the European cuckoo, a brood parasite that lays its eggs into the
nests of passerine bird species (Davies and Brooke 1989). Again, the number
of interacting species changes for both host and parasite.
Co-evolutionary turnover (Figure 11.3) is another mode of co-evolution
that involves a change in the number of interacting species. This occurs when
a species invades a new habitat and finds itself superior to the native
competition. The invader evolves convergently to resemble the native
species, forcing divergent changes on the traits of the competitor, eventually
sending it extinct. Eventually, this species may fall victim to a new invader.
 EVOLVING TOGETHER 125

The process has been suggested from observations on the Anolis lizards of the
Caribbean islands, where body size is the trait under question (Roughgarden
and Pacala 1989). Invading lizards arriving at an occupied island tend to have
large bodies and outcompete the smaller competitors, evolving convergently
towards a smaller body size that supports the highest carrying capacity.
Eventually, the native lizard may be forced into small bodied niches that
cannot support large populations and make it vulnerable to extinction.
Again, the number of interacting species varies over time.

11.2 The dynamics of species richness

The second characteristic that distinguishes the modes of co-evolution is
whether speciation and extinction result from the interaction, leading to
temporal dynamics of taxa. In some interactions neither speciation nor
extinction is implied as a result of the co-evolutionary process. This is the
case for mutual dependence. It is also the case for arms races. Arms races
occur when quantitative traits in both species show reciprocal directional
evolution. A well-known example is that of floral traits and those of their
pollinators. For example, Rediviva bees in South Africa visit Diascia flowers
to gather oils, which they collect from the base of flower spurs by inserting
their forelegs (Figure 11.4). Some populations of bees have evolved extra-
ordinary long forelegs, which match the long spurs of the flowers they
visit (Steiner and Whitehead 1990). It is likely that both characters have
co-evolved directionally: the bee to collect oils, and the flower to force more
effective pollination service from the bee. In other cases, as we have already
seen, co-evolution alters species richness. Escape-and-radiation co-evolution
increases it. Another interaction that involves speciation is the aptly named
‘diversifying co-evolution’ (Figure 11.3). This occurs in some maternally
inherited symbionts, such as Wolbachia and their hosts: the symbiont pro-
motes reproductive isolation, through parthenogenesis or an incompatibil-
ity mechanism. As a result new taxa of hosts are produced, and, of course,
new symbiont taxa too. As well as speciation, extinction can occur, as in
co-evolutionary turnover (Figure 11.3).

11.3 The temporal dynamics of traits

The third characteristic of the co-evolutionary process is the temporal
dynamics of traits involved in the interaction. In mutual dependence one
prediction is that these dynamics are reduced since change hinders efficient
126 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 11.4 The bee Rediviva neliana collecting oil from a flower of Diascia fetcaniensis. The forelegs of this
bee vary in length from 9 to 15 mm in different populations correlating with the length of the
floral spurs of local Diascia flowers. Photo courtesy of Kim Steiner.

interaction between the species, leading to stabilizing selection. A possible

illustration is in the mutualisms that involve ancient asexuals; they include
some mycorrhizal fungi, which inhabit the roots of plants, and the fungi that
are the food for leaf-cutter ants (Chapter 2). Another type of interaction that
can stabilize traits involves competitive interactions between plant species
that occur in co-evolutionary successional cycles. In these, succession creates
patterns of association between species in which co-evolution occurs. This is
best documented from white clover in grassland ecosystems. The clover
becomes adapted to flourish best next to whatever species of grass it has
become associated with. This is a particular kind of local adaptation. As
a result, competition between the clover and the grass is ameliorated
(Turkington 1989). Other interactions involve different trait dynamics.
Escape-and-radiation co-evolution involves periods of trait stability and
instability. Co-evolutionary arms races imply trait dynamics, although local
equilibria may be achieved, and dynamics are also found in co-evolutionary
 EVOLVING TOGETHER 127

turnover and co-evolutionary alternation. Diversifying co-evolution implies

nothing particular about these dynamics.

11.4 The dynamics of antagonism

The fourth and final characteristic is the dynamics of antagonism. Many cases
of co-evolution involve escalating degrees of antagonism. Included here are
escape-and-radiation co-evolution, arms races, and co-evolutionary alter-
nation. Arms races are interesting because they can occur, as in pollination
interactions, within a mutualism, which serves to underline how fragile such
mutualisms are to exploitation. Other modes highlight decreasing degrees of
antagonism: mutual dependence and co-evolutionary successional cycles.
Co-evolutionary turnover, while antagonistic, does not imply any change in
the degree of antagonism during the interaction. Diversifying co-evolution
does not specify any particular degree of antagonism.
Hence, co-evolution may or may not result in changes in the specificity of
the interaction, in the degree of antagonism, in species richness through
speciation or extinction, and variable dynamics of the traits involved. These
are the things we wish to understand or predict. To do so we must make some
starting assumptions and formulate models with them. The models that
have attempted to predict co-evolutionary outcomes mainly address three types
of interaction: competition (Chapter 9), predator–prey, and host–symbiont
(Chapter 10). Some have been successful in identifying conditions under
which the observed modes of co-evolution occur. As we saw in the last chap-
ter, there is some theory that can now predict the conditions for mutual
dependence in a symbiosis. With regard to competition, Roughgarden
(1995, p. 110) and co-workers have extensively modelled scenarios of
invasion and competition with regard to the Anolis lizard patterns. They find
that a taxon-loop involving invasion of larger bodied species, followed by
displacement and extinction of the smaller resident can be predicted if
competition is asymmetric and the width of the carrying capacity small. This
may then explain the patterns of body size and species richness seen on
certain islands. In this model the trait, body size, is assumed to be controlled
by numerous genes of small effect (quantitative genetics), which is a
reasonable assumption.
Other types of interaction demand alternative genetic assumptions. Models
investigating plants and their pathogens, such as rust fungi, frequently assume
a single major gene locus controlling the interaction in each species, so-
called ‘gene-for-gene’ co-evolution. For example, a pathogen may be virulent
or avirulent and a host resistant or not resistant. This appears to be the case
in many plant/plant–pathogen interactions. Gene-for-gene co-evolution
128 DISCOVERING EVOLUTIONARY ECOLOGY

has elsewhere been described as an alternative type of co-evolution, but it refers

really to an assumption rather than an outcome. However, the assumption can
generate rather characteristic outcomes: stable or fluctuating genotypic
polymorphisms, or fixation of one allele in both populations. Similarly,
variable dynamics are implied from field studies of plant resistance across
populations (Thompson and Burdon 1992). One interesting outcome
occurs when the fitness of host or parasite is frequency-dependent; that is,
that it is highest when rare. Under such conditions there can be cycling of
host and parasite allele frequency, so-called ‘Red Queen’ evolution, the type
of conditions that can favour variability among offspring, and hence sexual
reproduction (Chapter 2) as well as polyandry (females mating with several
males). There is now increasing evidence for the frequency-dependent
fitness of hosts under attack from parasites, as well as for links between
polyandry or recombination, and susceptibility to parasites (Lively 2001).
The theory of predator–prey co-evolution has, in contrast, evolved largely
in the absence of any background empirical data (Abrams 2001). The long
timescales involved in detecting both ecological and evolutionary dynamics
have been the most prohibitive hurdle. Recently, however, some studies on
planktonic algae and their predators have been successful in generating
short-term evolution in the algae that affect the ecological dynamics of both
species (Johnson and Agrawal 2003). While this is not co-evolution, the use
of microcosm systems like this holds the potential for generating useful data
that can test and further develop predator–prey co-evolutionary theory.
Thus there has been some useful but limited interplay between theory and
data in studies of co-evolution, though one could argue that greater potential
still exists. In some cases, empirical studies have provided knowledge of out-
comes or assumptions that have led to the development of theory capable of
explaining them. In other cases theoretical developments have predicted
outcomes that demand further empirical study.

11.5 The geographic mosaic

One particular empirical observation has dominated much recent thinking
on co-evolutionary dynamics: the observation that species interactions are
spatially differentiated between subpopulations. For example, in the South
African Rediviva bees and Diascia flowers (Figure 11.4), across populations
of bees there is variation in the length of the bee legs that matches the length
of the local floral spurs. Thus, traits have become geographically differenti-
ated across space. This is what John Thompson (1994, 2001) has called
a selection mosaic. There may also exist across space, areas where co-evolution
is more intense than in other areas, perhaps simply because of variation in
 EVOLVING TOGETHER 129

Fig. 11.5 A crossbill, Loxia curvirostra, from South Hills, Idaho, south of the Rockies, feeding on the cone
of a lodgepole pine, Pinus contorta. In this area, devoid of red squirrels, lodgepole pine cones
differ from those found in the Rockies, where red squirrels are major seed predators. The cones
from South Hills are more cylindrical, and have thicker scales as a defence against crossbills.
The South Hills crossbills have large, stout bills as a counter-adaptation. Photo courtesy of
Craig Benkman.

the degree of overlap of geographic range. An example of such hot spots

comes from a study by Benkman (1999) on lodgepole pine and crossbills in
the Rocky Mountains (Figure 11.5). In areas where red squirrels occur,
the pines have developed rounded cones that deter predation of seeds by
squirrels. In areas lacking red squirrels, the pines have developed longer
cones that are an effective defence against predation by crossbills. In these
areas, the crossbills have developed larger stouter bills. Thus, these areas are
co-evolutionary hot spots between pines and crossbills.
Subpopulations with divergent traits may also differ in the degree of popu-
lation mixing. The crossbill populations have different songs which are likely
to increase reproductive isolation. Intermediate bill morphologies caused by
hybridization should also be selected against, because each population lies at
its own adaptive peak (Benkman 2003). In the field, the birds from different
populations mate assortatively and are beginning to differentiate genetically
(Benkman 2003). Over time this may lead to speciation. These results are
intuitively pleasing because crossbills seem to have differentiated across
much of the northern hemisphere into a number of incipient species, and
130 DISCOVERING EVOLUTIONARY ECOLOGY

locally adaptive diversifying co-evolution provides a potentially widespread

mechanism. It is additionally exciting that without considering spatial
structure it would be more difficult to imagine that particular mode of
co-evolution occurring.
Observations such as these led Thompson to suggest in his ‘geographic
mosaic theory of coevolution’ that geographic structure is influential in
the evolution of species interactions: ‘Any theory of coevolutionary
dynamics must therefore take into account this geographic structuring of
most taxa and interactions’ (Thompson 2001, p. 332). This stimulated
many workers to examine whether linking subpopulations together in
a spatial structure can lead to different predictions about co-evolution in
those subpopulations than would be predicted if one just considered
a single subpopulation in isolation.
So far, several studies suggest that geographic structuring can be influ-
ential at that local population scale. For example, Hochberg et al. (2000)
developed a model of an obligate symbiont and its host, and investigated
how productivity differences across landscapes influence the evolution of
virulence in the symbiont. When productivity is high, the host population
is high in the absence of the symbiont (a source population) and when
productivity is low the host population is low (a sink population). When
virulent strains of the symbiont have a competitive advantage over aviru-
lent strains (perhaps via increased transmission), the virulent strain is
most likely to be found in the source population, and the avirulent strain
persists in marginal habitats that do not favour cheaters or exploiters so
much. Rather interestingly, if parasites are a strong selective pressure on
the evolution of sex, as suggested by Hamilton, then asexual forms should
be relatively favoured in marginal habitats where virulent symbionts can-
not persist. This matches empirical observations on many taxa that asexual
forms are commonest in marginal habitats (Chapter 2).
It is pleasing when the addition of a single assumption, such as spatial
structure, can serve to alter local predictions in the direction indicated by
data. However, there is a potentially much greater prize at stake; if the
addition of spatial structure can also change the global direction or mode of
co-evolution over all subpopulations combined. So far, work on the
geographic mosaic has not addressed this in earnest, but observations, such
as seen in Rocky Mountain crossbills, suggests that it might do so, at least
sometimes. The case of local co-evolution of CMS mutants and nuclear
genes in Plantago lanceolata, leading to reproductive isolation between
subpopulations (Chapter 5), provides another potential example. It is also
interesting that the general theory of the evolution of mutualisms increas-
ingly involves consideration of spatial structure (Chapter 10). Perhaps this
 EVOLVING TOGETHER 131

represents some convergence of the theory of species interactions from two

different perspectives.
Throughout the last three chapters it has been clear that the evolution of
species interactions can result in speciation. This is a logically consistent
concept, for as the next chapter will show, speciation requires rather specific
ecological and environmental conditions, and other species comprise an
important part of the ecological environment. The next chapter looks at the
mechanisms of speciation more generally.

11.6 Further reading

Thompson (1989) reviews the modes of co-evolution, and Thompson
(1994) updates this. Thompson (1999) reviews sources of data. Thompson
(2001) reviews some recent work on the geographic mosaic. Abrams (2001)
and Lively (2001) deal with predator–prey theory and parasite–host theory.
A special issue of the American Naturalist was devoted to co-evolution in
May 1999 (vol. 153 Supplement). It contains a number of reviews and case
studies of interest.
 12 Birth of species

Sympatric speciation is like the measles; everyone gets it, and we all get
over it.
attributed to Theodosius Dobzhansky

Until now we have been concerned almost exclusively with evolutionary

change within lineages—anagenesis. Anagenesis creates differences between
lineages and hence the diversity of form and function that we observe as a
major feature of our planet. It is, however, just one of the two major
evolutionary processes, the other being cladogenesis. Cladogenesis is the
creation of new lineages, speciation. Cladogenesis and anagenesis are jointly
responsible for life’s diversity: without cladogenesis, anagenesis could not
realize life’s potential for phenotypic diversity. Without anagenesis, as we
shall see, speciation could not occur.
Biologists have identified four distinct phylogenetic processes that have
created new lineages in the history of life (Figure 12.1), of which we have
already discussed two. The first is the assembly of biological entities from
non-biological chemical processes (Chapters 2 and 3). We have no evidence
that this has occurred more than once. The second process is the fusion of
two lineages into a single biological entity via symbiosis (Chapters 2, 10,
and 11). This has occurred a limited number of times but has often had
important consequences. The third process is hybridization: two individuals

Origins of life Symbiosis Hybridization Lineage splitting

Fig. 12.1 Four phylogenetic processes giving rise to new lineages.

 BIRTH OF SPECIES 133

from different species produce viable and fertile offspring that become
reproductively isolated from their progenitors. The fourth process is when
one species divides into two or more. It is these last two processes that will
form the subject of this chapter.
In general biologists studying speciation are interested in three types of
change: reproductive isolation, genetic differentiation, and ecological
differentiation. Sometimes these act in concert, sometimes one results auto-
matically in another, and sometimes one or more are considerably delayed.
For example, in Chapter 1 we considered the case of Lake Victoria hap-
lochromine cichlids, where reproductive isolation occurred first via changes
in mate choice and coloration, followed by ecological change, while genetic
differentiation leading to irreversible separation of lineages has yet to occur
in many lineages. In the case of hybridization, all three processes (reproduct-
ive isolation, genetic differentiation, and ecological differentiation) may be
a simultaneous consequence of the formation of hybrids.
The architects of the great Neo-Darwinian Synthesis (such as Dobzhansky,
Fisher, Mayr, Simpson, Wright) laid down the groundplan for thinking about
speciation. Mayr, in particular, was of the view that speciation of the fourth
kind (a lineage splitting into two or more lineages) occurred allopatrically in
so-called peripheral isolates, via the founder-effect, a largely non-adaptive
process in which small initial isolates created a novel starting gene pool for an
incipient species. Genetic differentiation could then be enhanced by divergent
natural selection. He sternly denounced the possibility of sympatric specia-
tion.Wright also advocated a large role for non-adaptive processes in his ‘shift-
ing balance’ theory, whereby a small population could traverse a non-adaptive
valley by genetic drift, and then diverge through natural selection up a differ-
ent adaptive peak. Adaptive processes were simply not seen as necessary by
most for speciation in allopatry, sympatric speciation was deemed unlikely at
best, and natural selection could finish what other processes had started.
Since then we have gradually entered a phase in which the conceptual bar-
riers to sympatric speciation have been eroded, the role of adaptive evolu-
tionary change in the speciation process has become more widely
appreciated, and sexual selection increasingly plays a role in our views of spe-
ciation. The relevance of speciation by hybridization has had a similarly che-
quered history, being more popular today than formerly. We should not fall
into the trap of believing that the current vogue of research into these areas
represents in some sense a greater truth. When a field swings from one
extreme to another, and has yet to settle down, predicting where the equilib-
rium will be found is difficult and risky. There is, however, a sensible prevail-
ing tendency to consider alternative processes more widely, and to ask not if,
but in what circumstances, a process may be relevant. Data are still relatively
thin on this latter point, but not totally absent, and there are good prospects
for quantitative answers in the near future.
134 DISCOVERING EVOLUTIONARY ECOLOGY

12.1 Speciation by hybridization

There is undeniable evidence that hybrid species are sometimes established,
and indeed it has been estimated that 11% of plant species have evolved in
this way. Some of the most elegant recent studies have been conducted by
Rieseberg and co-workers on North American sunflowers. Two species of
sunflower, Helianthus annuus and Helianthus petiolaris, have overlapping
geographic ranges in western USA and have given rise to three hybrid
species, Helianthus anomalus, Helianthus paradoxus, and Helianthus
deserticola (Figure 12.2). Evidence supports a ‘recombinatorial’ model for
the speciation of these hybrids (Figure 12.3). Under this model crossing
events between the two parent species lead to novel chromosomal arrange-
ments, as a result of chromosomal breakages (Rieseberg et al. 1995). The
novel rearrangements cause reproductive isolation between the hybrids
(Rieseberg et al. 1999). Some novel rearrangements carry a selective advant-
age, which allows invasion of new habitats (i.e. heterozygote advantage,
Rieseberg et al. 2003). Thus, while the parent species prefer heavy clay and
dry sandy soils, respectively, H. anomalus and H. deserticola are found in
more xeric habitats than their parents, while H. paradoxus is found only in
saline habitats. In the recombinatorial model then, hybridization both
creates novel adaptive variation and simultaneously a mechanism for
reproductive isolation.
Another route for hybridization to lead to speciation is through allopoly-
ploidy; the creation of hybrids with double (or more) of the normal chro-
mosome complement (Figure 12.3). Doubling of chromosome number
allows a hybrid to escape infertility problems at meiosis by giving each
parental chromosome its own complementary partner. The origin of several
allopolyploid species has been documented in the last several decades, often
as a result of species being introduced outside of their native range, leading
to novel hybridizations. For example, the diploid species Tragopogon pratensis,
Tragopogon dubius, and Tragopogon porrifolius were introduced into North
America at the beginning of the twentieth century, and in about 1940 gave
rise to two tetraploid hybrids in eastern Washington: Tragopogon miscellus
(from Tragopogon pratensis ⫻ Tragopogon dubius) and Tragopogon mirus
(from T. dubius ⫻ T. porrifolius). Both these polyploids have since expanded
their range considerably (Levin 2000).
In recent years a major concern about the likelihood of hybrid speciation—
the low fitness of hybrids and their consequent inability to persist—has been
somewhat quelled by new data (Arnold and Emms 1998). One particularly
graphic study by Peter and Rosemary Grant on Darwin’s finches on the
Galapagos island of Daphne Major highlights how hybrids might not always
be less fit than their parent species. They studied the fitness of hybrids
 BIRTH OF SPECIES 135

Fig. 12.2 The sunflowers H. annuus and H. petiolaris (top), which have given rise to three hybrid species
(bottom). Of these, H. anomalus and H. deserticola are found in xeric habitats, while
H. paradoxus is found in saline habitats. Photos courtesy of Loren Rieseberg.

between three finch species (Geospiza fuliginosa, Geospiza fortis, and

Geospiza scandens) over a number of years. In most years the hybrids were
less fit than their parent species because they had intermediate beak sizes that
were not best adapted to the types of seeds available (Figure 12.4). However,
136 DISCOVERING EVOLUTIONARY ECOLOGY

(a)

(b)

Fig. 12.3 Two ways in which hybridization can give rise to new species. (a) The recombinatorial model,
in which chromosomal breakages occur giving rise to hybrids that are reproductively isolated
from their parents. (b) Allopolyploidy, in which diploid gametes give rise to fertile hybrid
offspring because each chromosome retains its complementary pair.

(a) (b)

(c)

Fig. 12.4 Hybridization between Darwin’s finches on the Galapagos islands. The Cactus finch,
G. scandens (a), the medium ground finch, G. fortis (c), and a hybrid with an intermediate beak
shape (b). In dry years, the hybrids do not survive well because their beaks are not well adapted
to utilize existing seeds. In rarer wet years, when seeds are more abundant, competition is
reduced and hybrids survive better. Photos courtesy of Peter Grant.
 BIRTH OF SPECIES 137

Species 1 Species 2

Selection

Hybrids

Beak size

Fig. 12.5 Hybrids are often selected against through disruptive selection (top), and this prevents
introgression of the two parent species. Occasionally, however, as in Darwin’s finches, hybrids
may survive due to environmental change relaxing selection against them (bottom).

in the El Niño year of 1982–83, the rainfall on the islands dramatically

increased, and a dramatic change occurred in the types of seed available to
the finches. In that year, the hybrids had equal or higher survivorship,
recruitment, and breeding success than their parents (Grant and Grant
1993). Thus, changes in the external environment can sometimes affect dras-
tically the fitness of parents and hybrids. It is not inconceivable that such
variations in fitness might sometimes allow a hybrid species to persist and
spread into new habitats (Figure 12.5).
It is also becoming clear that hybrids frequently do persist for long periods
and spread despite their normally low fitness. Some populations of the closely
related Drosophila simulans and Drosophila mauritiana share large segments
of DNA indicating past gene exchange between them, despite experimental
evidence showing strong barriers to gene exchange (Solignac and Monnerot
1986). Such observations led Arnold (1997) to suggest that the traditional
assumption of low hybrid fitness, while often true, does not apply to all
hybrids and that some hybrids can be fitter than their parents in a range of
niches. Even if these are established rarely they can give rise to persistent
lineages. Thus, the concept of hybrid speciation, not just in plants but also in
animals, is experiencing something of a revival.

12.2 Speciation by lineage splitting

The majority of research effort on speciation has been on how lineages split
into two. Speciation of this form has an important geographic context,
138 DISCOVERING EVOLUTIONARY ECOLOGY

Table 12.1 The geographic context of some modes of speciation (from Levin 2000)

Size of Sympatric Spatial Proximal Distal

speciating relationship
entity Contiguous

Local Sympatric Peripatric Disjunct

speciation speciation Speciation
Big Parapatric Vicariance
speciation speciation

which has been the subject of much controversy (Table 12.1). The last two
columns of Table 12.1 represent so-called allopatric speciation, speciation
through geographic isolation. Allopatric speciation can be further
subdivided according to the size of isolated populations and how far they are
separated from each other. Under the peripatric model of peripheral isolates
favoured by Mayr, the speciating entity is small, and close to the main parent
population. Disjunct speciation differs in that the speciating entity is dispersed
far from the original. In both these modes of speciation stochastic processes,
such as founder effects and drift, may be important as the speciating
population is small, though selection may also be operating to change gene
frequencies. In vicariance speciation, isolating barriers emerge in an existing
part of the geographic range, leading to isolation between the fragmented
populations. Selection is likely to predominate here as the force causing
genetic differentiation between species, and ultimately isolation, because the
isolated populations are usually larger.
Some authors regard these modes of speciation as theoretically trivial,
because we expect geographically isolated populations to differentiate over
time by one process or another. In addition, there is good evidence that all
the allopatric modes do exist in nature. For example, clear-cut cases of
vicariance come from marine organisms separated by the Isthmus of
Panama (approximately 3 Ma). A large number of closely related species are
separated by this barrier from fishes and crabs to shrimps and sea urchins
(Lessios 1998). Disjunct speciation has occurred in the 14 species of
Darwin’s finches on the Galapogos islands, all descended from vagrant
individuals arriving from the American mainland about 3 Ma (Grant and
Grant 1996). Several likely cases of peripatric speciation, occurring in birds
on offshore archipelagos which differ from nearby mainland relatives are
documented by Mayr (1940, 1963) in the Carribean and East Indies. Thus,
allopatric speciation in general is highly plausible.
 BIRTH OF SPECIES 139

12.3 Splitting in sympatry?

If speciation by geographic isolation was never in doubt then the same
cannot be said of sympatric speciation (Chapter 1), speciation in the absence
of geographic isolation. Until very recently, there were no clear-cut empirical
cases: Ernst Mayr in his 1963 book systematically attacked the best-known
potential examples of the time, including the East African cichlids mentioned
in Chapter 1, and another example discussed below, the apple maggot fly, as
being just as (if not more) consistent with speciation in allopatry. The early
theoretical work also suggested rather restrictive conditions for sympatric
speciation (e.g. Maynard Smith 1966). In general, models found two
problematic areas. First, it was difficult to explain assortative mating. Many
models assumed strong linkage, or pleiotropy of the mating system to an
ecological trait under selection, but in general pleiotropy is expected to be
weak. Second, ecological differentiation is normally required to prevent
competitive exclusion of one of the incipient species. Even in the presence of
disruptive selection (intermediate phenotypes are selected against while
extreme phenotypes are favoured), there is a tendency for just one extreme
ecotype to evolve rather than two because a population will tend to move to
one side or the other of the new fitness minimum. If linkage between the
mating system and the ecological trait is weak, disruptive selection needs to
be very strong to overcome recombination of the ecological traits.
Since then, further evidence about particular case studies has strengthened
the case for sympatric speciation. In addition theoretical developments have
shown that sympatric speciation is more feasible than previously thought.
The theoretical advances have been (1) sexual selection leading to assortative
mating (Chapters 1 and 7), and (2) adaptive dynamics leading to evolution-
ary branching and hence coexisting ecological polymorphism (Chapters 8
and 10) (van Doorn and Weissing 2001). Models have tended to emphasize
solutions to one of these problems over the other, and hence can be termed
sexual selection or ecological models.
Sexual selection models explicitly take into account the interaction of
males and females and assume that males compete for mates while females
exert mate choice. Normally, Fisherian runaway selection (Chapter 7) is
assumed, such that male traits and female preference for those traits become
genetically correlated through non-random mate choice. More extreme traits
and preferences evolve until counteracted by natural selection. Mating
strategies need to become polymorphic to generate reproductive isolation,
and hence we need a source of disruptive selection. This might come from
innate female tendencies to prefer particular divergent phenotypes, as may
be the case in haplochromine cichlids. However, branching of male traits,
140 DISCOVERING EVOLUTIONARY ECOLOGY

and subsequently female preference, can also occur if rare phenotype males
experience less competition for mates from other males (van Doorn and
Weissing 2001, Chapter 1).
Ecological models emphasize disruptive selection on the ecological
phenotype. This can readily occur in adaptive dynamic models through
evolutionary branching, because the selection is not imposed externally but
dependent on other resident phenotypes (frequency dependent). Under
some circumstances then, evolution will drive the ecology of the species to a
point where it lies at a fitness minimum and branches into a stable polymor-
phism (Chapter 8). Resource competition can readily do this, as long as the
fitness advantage of utilizing rare resources through reduced competition
outweighs the lower abundance of those resources.
So incorporation of sexual selection and ecological branching into
models have shown that sympatric speciation can readily occur. In addition
the evidence in favour of some of the potential examples of sexual selection
has become much firmer (Chapter 1).A particularly influential case is that of
the apple maggot fly Rhagoletis pomonella (Figure 12.6). The fly normally
lays its eggs on developing apple fruit, in which the larvae develop, pupating
once the fruit falls to the ground. In the Hudson valley of New York state,
sometime during the mid-1800s, a host shift occurred and a new race
evolved, apparently in sympatry, that developed on hawthorn fruit.
In the 1960s,Guy Bush,developed a verbal model of sympatric speciation via
host race formation in the fly. Bush’s model bears many resemblances to some
of the classical sympatric speciation models. First, it assumes host-specific
mating (a pleiotropic effect), such that each race mates and oviposits on the
same fruit on which they fed as larvae. This would give assortative mating.
Second, he assumed host-associated fitness trade-offs, such that a gene

(a) (b)

Fig. 12.6 Apple maggot flies, R. pomonella, on an apple. Photos courtesy of Andrew Forbes.
 BIRTH OF SPECIES 141

giving a fitness advantage to development on one host was ill suited towards
development on another host. Hybrids would be less fit than both parents on
both hosts—a source of disruptive selection.
This model has subsequently been confirmed by data (Feder 1998). First,
comparison of allozymes between the apple and hawthorn races has shown
consistent differences. Races show genetic differences in host preference;
females of both races prefer hawthorn but hawthorn females are more averse
to apple. Once they have met, however, there is no pre-mating isolation, and
no evidence of sterility or inviability barriers to reproduction. The source of
the fitness trade-offs, and disruptive selection, is phenological: development
is faster in hawthorn flies. Hawthorns fruit later in the year than apples and
so hawthorn flies must develop quicker to complete development by
autumn. Apple flies, however, are exposed to warmer weather before the
winter arrives and must enter a deep diapause to prevent premature emer-
gence before the winter sets in. Interestingly, when pre-winter temperatures
are manipulated to expose hawthorn flies to ‘apple’-like conditions, a selec-
tion response was detected in the alleles controlling diapause such that the
flies became more like the apple race; in fact they did so almost completely in
a single generation of selection. Thus, there is antagonistic pleiotropy
between fitness determining loci, and this prevents gene flow between the
races. It appears that Bush was right, and the apple maggot fly remains the
single best example of sympatric speciation in action.
It remains plausible that sympatric speciation is common, but it will be
frustrating to wait until a large number of organisms have been studied in the
same degree of detail. Luckily, a broad-brush approach is available to deter-
mine the frequency of the different geographic modes of speciation: compar-
ison of geographic range overlap across phylogenies (Barraclough and Vogler
2000). Under sympatric speciation, recently formed species should share
large parts of their geographic range, whereas under allopatric speciation
there should be almost no sharing of geographic range. Deeper down in the
phylogeny, however, the degree of range overlap between sister taxa should
depend on the degree of subsequent change in range change after the specia-
tion event. By plotting the degree of overlap in the range against node height,
we obtain a signal about the geographic mode of speciation (Figure 12.7).
In a comparison of several insect and vertebrate phylogenies, Barraclough
and Vogler found that allopatric speciation was in fact the major signal, even
in the Rhagoletis phylogeny where there was evidence for just a single sym-
patric speciation event.A further finding was that the ranges of allopatric sis-
ter taxa were frequently very different in size, suggesting that peripatric
speciation is common. Broad-brush approaches like this lead us to hope that
we may in a short time be able to assess the frequency of different speciation
modes in a large number of taxa.
142 DISCOVERING EVOLUTIONARY ECOLOGY

Degree of sympatry
Sympatric speciation

Allopatric speciation

Node height (Ma )

Fig. 12.7 The degree of range overlap in species on a phylogeny, when they have been formed by either
sympatric and allopatric speciation. Young species formed sympatrically have a high degree of
range overlap but the overlap decreases with age due to range changes. Young allopatrically
formed species show zero range overlap which increases with time due to range changes.
After Barraclough et al. (1999).

12.4 Adaptive or non-adaptive differentiation?

Despite favouring different geographic modes of speciation, biologists have
differed considerably over the likely processes causing the initial genetic
differentiation between neospecies. Recently, there has been a dogmatic shift
away from non-adaptive processes towards adaptive selection as the major
process. The shift is a result of both evidence against non-adaptive processes,
such as founder effects, and evidence in favour of selection.
Evidence that counters the likelihood of a role for non-adaptive processes
comes from both theory and empirical evidence. For example, Barton
(1998) has argued from a population genetic viewpoint that population
bottlenecks, such that make founder effects and drift likely, are also likely to
decrease the chances of incipient species persistence due to lack of genetic
variation. Theoretical examination of Wright’s shifting balance has shown
that it is possible to cross a non-adaptive valley only if the valley is shallow
and population size is small. However, very shallow non-adaptive valleys are
unlikely to persist in a variable environment, so there are just as likely to be
episodes when selection can drive such a peak shift. Recent experiments on
Drosophila have also failed to produce significant population differentiation
from founder events (Rundle et al. 1998).
The pro-selection evidence comes from several sources (Schluter 1998). First,
in some adaptive radiations, ecological divergence apparently occurs as rapidly
as reproductive isolation. In many freshwater fish colonizing post-glacial lakes
 BIRTH OF SPECIES 143

and rivers in the Northern Hemisphere, new sympatric sister species are
strongly divergent ecologically, often with one benthic and one planktivo-
rous species coexisting (Figure 9.5). There is strong evidence that these dif-
ferences are adaptive, for they correlate with increased foraging efficiency
which gives an advantage in competition with other fish. In addition, some-
times mate discrimination is the result of natural selection: in sticklebacks,
body size is used as a trait in mate choice, but is also one of the major traits
responsible for ecological divergence. In some cases, divergent selection can
lead directly to post-zygotic isolation. Some monkeyflower (Mimulus gutta-
tus) populations have adapted to soils contaminated with copper (Figure 7.5),
but these alleles are lethal when combined with other alleles in hybrids with
other populations that are not tolerant to copper (Chapter 7).
A second area of evidence comes from examining hybridization events. If
differences between species are adaptive, hybrids should be less fit than their
parents in the native habitat, but this difference should disappear in the lab.
Genetic mechanisms reducing hybrid fitness, however, should be apparent
even in the lab. In fact, many groups produce highly viable and fertile labor-
atory hybrids, including Drosophila, Darwin’s finches, East African cichlids,
Hawaiian silverswords, and fishes from post-glacial lakes mentioned above.
A final source of evidence comes from measuring the rate of evolution across
populations or species as compared to the null hypothesis of neutral evolu-
tion through mutation and drift. Studies by Orr (1998) showed that of eight
loci affecting male genital structure in the sister species D. simulans and
D. mauritiana, all eight indicated evolution by selection. Studies on the
genetic variation in quantitative traits among populations also indicate
divergent selection in many species (Schluter 2000).
Thus, there is now growing evidence that the origin of species is indeed by
means of natural selection, as Darwin suggested in the title of his famous
book (Darwin 1859). Furthermore, lineage splitting in sympatry is looking
more likely in comparison to the dogma of previous decades, and sexual
selection is increasingly invoked to play a role in species divergence.
What then, is the role of ecology in the formation of species? First, ecology
sets the geographic stage for speciation of whatever kind. In sympatry,
speciation requires disruptive selection, which must have an ecological
dimension. For hybrid speciation geographic proximity is required, and
hybrid fitness and spread is determined by the ecological situation. Even in
allopatric cases, selection frequently creates the initial divergences between
incipient species that can lead to reproductive isolation. If stochastic
processes play a role, they too are dependent on specific ecological circum-
stances. Finally, natural selection is a major process driving further differen-
tiation between species such that it becomes irreversible. Thus, the birth of
species has a very fundamental ecological context. In the next chapter, we will
144 DISCOVERING EVOLUTIONARY ECOLOGY

see that the death of species, extinction, primarily a population ecology

phenomenon, has an evolutionary context.

12.5 Further reading

Schuilthuisen (2001) provides a readable introduction to speciation mechan-
isms. Howard and Berlocher (1998) has many good chapters, such as those
by Schluter (ecology), Feder (apple maggot fly), Arnold and Ems (hybrids),
and Rice (genomic conflict). A special issue of Trends in Ecology and
Evolution: July 2001. (vol. 16, no. 7), pp. 325–413 has many useful articles
covering chromosomes and speciation (Riesenberg), ecology (Schluter),
sexual selection (Panhuis et al.), and sympatry (Via). Levin (2000) covers
plants well. Gavrilets (2003) reviews theory for the confident.
13 Death of species

It is hard to have patience with people who say ‘There is no death’ or

‘Death doesn’t matter.’ There is death. And whatever is matters. And
whatever happens has consequences, and it and they are irrevocable and
irreversible.
C. S. Lewis

In the last chapter we saw how new evolutionary lineages may arise,
increasing the number of branches on the tree of life. We will now consider
how the number of branches might be depleted, thus ‘pruning’ the tree of
life; the process known as extinction. At its limit, extinction is a population
level phenomenon: it is the specific case of population dynamics when the
number of individuals becomes zero. This sounds like a purely ecological
process with little evolutionary perspective. Yet, as we shall see in both
this and the next chapter, extinction has evolutionary causes and con-
sequences. First, as populations decline, changes in genetic composition
often occur, some with impacts on the phenotype of the organism. These
changes may affect the probability of extinction, and are the subject of this
chapter. Second, certain characteristics may incidentally make some
species more likely to decline than others, or more vulnerable to extinction
once in decline. These differences have come about through evolution.
Finally, once a lineage has gone extinct, that may alter the likelihood of
extinction or speciation in other species. The latter topics are dealt with in
the next chapter.
There are several ways in which the processes of extinction might
be classified. From a phylogenetic perspective, extinction can occur by a
lineage simply ceasing to exist altogether, or alternatively by merger with
another lineage (e.g. hybridization). Another framework is to consider
what extrinsic processes in the environment might be causal agents (e.g.
Purvis et al. 2000). We will explore all these aspects of extinction. Another
useful and frequently-made distinction is between factors that make a
species rare and factors that cause extinction once a species is rare (Soulé
1987; Caughley 1994; Lawton 1995). This will form the framework for the
chapter.
146 DISCOVERING EVOLUTIONARY ECOLOGY

13.1 Reasons for rarity

Rarity is most commonly defined as some combination of geographic range
(small) and abundance (low) that might make a species vulnerable to extinc-
tion (see Gaston and Kunin 1997). Just because a species is rare, however,
does not mean it is declining. Species may be rare at three stages in their evolu-
tion: first, they may just have speciated and be in the process of expansion.
Second, they may have expanded their geographic range as far as possible,
but still remain rare. Third, they may be rare because their range has declined
from its maximum extent (Figure 13.1).
The danger of extinction while a species is still young is illustrated by the
fates of hybrid plant species that have formed in historical times (Chapter 12).
The Malheur Wire Lettuce, Stephanomeria malheurensis, is a recent derivat-
ive of Stephanomeria exigua known only from a single hilltop in Oregon
(Figure 13.2). The species is probably now extinct, largely due to changes in
the habitat and invasion of an introduced grass, Bromus tectorum.
A contrasting story is that of the cordgrass Spartina anglica, a tetraploid
formed through hybridization of Spartina maritima and Spartina alternifo-
lia, and first recorded in Lymington, UK, in 1892. Both the hybrid and its
parents grow in coastal mudflats. S. anglica spread rapidly following its
discovery, and in a few decades had spread along much of the southern
English coast and onto the French coast. It has since been introduced to other
European countries as well as China and Australia to stabilize low lying
coastal zones. The species cannot be considered at risk of extinction, and one
factor favouring its long-term survival has been its rapid exit from the
vulnerable early stage of its formation.

(3) (4)
(2)
(1) (5)

Fig. 13.1 Changes in geographic range during a species’ lifespan, giving rise to rarity. Species are
generally rare at their origination (1) and at the end of their lifespan (5), but in between may
be either widespread (3) or local (2), (4).
 DEATH OF SPECIES 147

Fig. 13.2 Malheur Wire Lettuce, S. malheurensis, at its only known site in Oregon. Photo courtesy of
Leslie Gottlieb.

Therefore, species that survive their initial rarity, like S. anglica, may enter
a second stage in their evolution, during which they expand their range to
some maximum because adaptations allow them to sustain positive rates of
increase beyond the area of their birth. Dispersal is obviously required
(Chapter 6) as are broad ecological tolerances (Chapter 8). Some species,
so-called endemics, will, because of their site of origin or specialization, never
attain a large geographic range. At least 44% of vascular plant species and
35% of vertebrates are endemic to 25 biodiversity ‘hot spots’ which represent
only 1.4% of the land area of the globe (Myers et al. 2000). Many of these hot
spots include island archipelagos, such as the Pacific islands, the Carribean,
New Zealand, the East Indies, and the Indian Ocean islands. Evolutionary
148 DISCOVERING EVOLUTIONARY ECOLOGY

changes may occur among rare species that make range expansion less likely.
For example, island species may evolve reductions in dispersal ability (Roff
1994; Grant 1998), such as flightlessness in birds (see Figure 13.3) and
insects, and reduction in seed dispersal in plants (Chapter 6), that reduce
their chances of increasing their geographic range. Rare species may also
develop adaptations that favour persistence. Kunin and Schmida (1997)
have identified changes to plant breeding systems and flower architecture as
possible adaptations to rarity. They tested the breeding systems of 52 Israeli
crucifers against their abundance and found that species with sparse popula-
tions tended to be more self-fertile. In addition, rare self-incompatible plants
had unusually large flowers, which should enhance the chances of cross-
pollination.
After expanding their range to some maximum, species might also
become rare later in their evolutionary history through declining abundance
or geographic range. The agents of decline can be either biotic or abiotic.
Biotic agents of decline are extremely well illustrated by the recent effects of
introduced species. The Lord Howe woodhen, Tricholimnas sylvestris
(Figure 13.3), for example, is a flightless rail that lives on Lord Howe Island,
about 600 km East of the Australian mainland. The island was first
discovered in 1788 and settled in 1834, with the consequent introduction of

Fig. 13.3 Two Lord Howe Woodhens, T. sylvestris, a species of flightless rail restricted to Lord Howe
Island in the South Pacific. Photo courtesy of Ian Hutton.
 DEATH OF SPECIES 149

pigs, dogs, cats, goats, and black rats. By 1853, the woodhen was restricted to
mountainous parts of the islands, and by 1920 the population was almost
entirely restricted to one single mountain top of 25 ha, containing no more
than 10 breeding territories. Miller found that the range of feral pigs did not
overlap with the range of the woodhen and could have been the cause of
decline through predation on nesting adults and eggs (Miller and Mullette
1985). After removal of the pigs, and a subsequent release of captive bred
birds into other areas of the island soon filled the entire available habitat on
the island. Declines due to introduced species illustrate the importance of
co-evolutionary forces (Chapter 11) on species abundance. Many island
birds, for example, are vulnerable to introduced ground predators because
they have evolved adaptations in the absence of predators that become
detrimental in presence of predators. These adaptations include flightless-
ness, ground-nesting, and lack of escape responses (Grant 1998).
Abiotic changes will typically involve changes in climate. Many species
have experienced well-documented contractions or expansions in their
range during the last several thousand years as a result of changes in global
temperature associated with the glacial and interglacial periods. The
redwood trees Sequioadendron giganteum and Sequoia sempervirens were
widespread in North America before the Pleistocene glaciations, but the
former is now restricted to a few valleys in the Sierra Nevada mountains in
California, and the latter to the fog belt of the coastal ranges. Aphodius
holderi was the most abundant dung beetle in Britain and other parts of
Eurasia during the colder parts of the last glaciation, but today is restricted to
a very small area in the high plateau of Tibet (Coope 1973).
Thus, at any stage of their evolutionary lifespan species may become rare.
Most will be rare at their origins, but if they can expand rapidly their long-
term prospects are good. Many species, however, may not ever achieve large
geographic ranges or population sizes, a contributory factor to which will be
their evolved characteristics. Finally, some species will experience declines
after reaching their range or abundance maximum, and one contributory
factor may be the evolutionary interactions with other species. When
populations are rare, what might deliver the final coup de grace?

13.2 Extinction when rare

A variety of processes may eventually cause population extinction once the
population has reached a small size or range. The importance of these
processes can vary with the size of the population and the characteristics of
the species. Many of these processes are stochastic, meaning they are not
exactly predictable but occur with a probability that can be known. Others
150 DISCOVERING EVOLUTIONARY ECOLOGY

are more predictable or ‘deterministic’. The processes can be environmental,

demographic, or genetic in nature (Lande 1988). We will deal with these
in turn.
Environmental stochasticity is the name given to unpredictable events
that affect all individuals in the population in similar ways. In any given year,
for example, weather conditions might promote or reduce reproduction;
and biotic agents might favour or hinder survival. In one well-documented
case, increases in environmental stochasticity have been strongly implicated
in population extinction. The Bay subspecies of Edith’s Checkerspot butter-
fly (Euphydryas editha bayensis) (see Figure 9.4) has recently experienced a
number of local population extinctions. This has been linked with increasing
variability in rainfall during recent decades. In dry years, the larvae cannot
develop sufficiently fast to enter diapause in the summer before their food
plants die, and there is massive larval mortality (McLaughlin et al. 2002).
Other stochastic events affecting whole populations have been implicated
in recent extinctions: the last population of the Heath Hen (a type of grouse
once endemic to North America) having been restricted by hunting and
habitat destruction to the island of Martha’s Vineyard at the turn of the
century, experienced a succession of unlucky events, including a drought
and fire, and unusually high predation by Goshawks. The population never
recovered. The risk of extinction from purely environmental stochasticity
can be modelled very simply by describing time to extinction versus
population size (carrying capacity) when the average vital rates are drawn at
random from a distribution (Lande 1993). Time to extinction depends on
the intrinsic rate of increase of the population, being larger when rate of
increase is large, but tends to asymptote with population size (Figure 13.4).
Thus, it remains an effective cause of extinction even in moderately sized
populations.
Demographic stochasticity in contrast is the result of changes in the aver-
age vital rates of a population due to differential success of individuals. Some
years by chance most of the individuals will be lucky, and at other times
unlucky, and these effects can be modelled as sampling variances drawn
from a distribution. The variance in average rates is inversely proportional to
the population size, and so time to extinction rises quickly (exponentially)
with population size (Figure 13.4). As a consequence, unlike environmental
stochasticity, demographic stochasticity is most important in very small
populations.
Other demographic causes of extinction may be more deterministic in
nature. The Allee effect refers to a deterministic decline in fitness with density
due to non-genetic reasons. In a population already struggling to maintain
positive rates of increase, the Allee effect can lead to an irreversible decline:
below a critical density, fitness is reduced, reducing density still further,
 DEATH OF SPECIES 151

Time to extinction
(a)

(b)

Population size

Fig. 13.4 Time to extinction against population size. In (a), the line is asymptotic, so even quite large
populations may be at risk. In (b), small populations are at risk, but even medium sized
populations are relatively safe.

reduces fitness still further.Two typical biological causes are density-dependent

mating success, and synergistic social interactions, such as group foraging or
defence from predators. The edge effect is a related process that occurs when
individuals at the edge of a population experience declining fitness. As the
geographic range of a population diminishes, a greater proportion of indi-
viduals is near the edge of the range and experience this decline in fitness. In
plants a typical cause of this non-genetic decline in fitness is a decline in
cross-pollination. In the rare Australian plant Banksia goodii, for example,
which grows about 20 cm high, the number of seeds set per plant is positively
correlated with population size, with populations covering less than 25 m2
setting no seed at all (Levin 2000). Only about 1000 plants of this species
remain and most populations are very small.
A final demographic cause of extinction is a change in the balance between
colonization and extinction of habitat patches or subpopulations. Networks
of ephemeral populations linked by emigration and immigration, and
persisting through a balance between the colonization and extinction of
subpopulations, are called ‘metapopulations’. Critical factors allowing the
persistence of such a metapopulation include the number, size and distribu-
tion of patches, and the dispersal rates between them. As a population
declines, either in the size or number of subpopulations, this can affect the
balance between local extinction and local colonization. As a result, below a
critical number of patches or size of subpopulations the whole system
collapses inevitably towards extinction. A system where these effects have
been modelled is in the Northern Spotted Owl Strix cavrinea occidentalis
(Lande 1988). The patches here are owl territories of 2.5–8 km2, which the
152 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 13.5 The Glanville Fritillary butterfly, Melitaea cinxia. Populations on the Åland Islands in Finland
are more likely to be inbred if small, and more likely to go extinct if inbred. Photo courtesy of
Niclas Fritzén.

owls occupy as monogamous pairs in old growth conifer forest in the Pacific
North-west. Plans for the conservation of the owl had originally envisaged
conserving 500 pairs to maintain sufficient genetic variation (see below).
A model based on habitat occupancy, however, showed that 500 pairs were
insufficient to halt extinction from demographic causes because coloniza-
tion of new territories would not balance loss of occupied territories. A
similar and more recent study on separate metapopulations of Glanville
Fritillary butterflies (Figure 13.5) in Finland showed that metapopulations
of a few small well separated habitat patches had gone extinct during recent
years while larger metapopulations with greater numbers of large proximate
habitat patches persisted (Hanski and Ovaskainen 2000) as predicted by a
simple model.

13.3 Genetic causes of extinction

Genetic factors leading to extinction include evolutionary changes in small
populations. If a species with a small surviving population is in close
proximity to more abundant congeners with which it can interbreed, a real
 DEATH OF SPECIES 153

Fig. 13.6 Extinction via hybridization. Here a rare species (white circles) has its range surrounded by
another species (black circles) with which it can hybridize, reducing its ability to replace itself.
Interbreeding occurs with close-neighbours, and eventually, all the reproduction of the rare
species is via hybridization.

risk of extinction comes from interspecific hybridization. There are several

reasons why this may be a threat to the less abundant species (Levin 2000).
First, crossing with other species reduces its potential to replenish its num-
bers because some reproductive potential is diverted to hybrid offspring
rather than pure-bred offspring (Figure 13.6). Because the small population
is less effective as a reproductive donor than the large population, the effect
on its reproductive potential is much greater. Second, the hybrids might
compete successfully against the parents reducing their fitness. Again, the
proportional effect is greater in smaller populations. Hybridization can also
increase pressure from natural enemies. Many plant hybrids are more
susceptible to pest exploitation than their parents and hence can support
large pest populations in proximity to their parent species (Levin 2000).
Finally, genetic introgression may be so one-way that the smaller popula-
tion simply becomes absorbed into the larger one. One likely future extinc-
tion from hybridization comes from the Catalina Mahogany tree,
Cercocarpus traskiae (Rosaceae) which occurs on Santa Catalina Island off
the Californian coast. It is restricted to a single population and is hybridizing
extensively with its close relative Cercocarpus betuloides. Eventually, this may
lead to total introgression of the mahogany such that no pure-bred plants
remain. There are several other botanical examples, but hybrid extinction is
also a possibility in animal populations. We encountered one example from
the haplochromine cichlids of Lake Victoria (Chapter 1) where mating bar-
riers in sympatry are breaking down due to water turbidity. Another animal
example comes from the endangered European White-headed duck, which
has been hybridizing extensively with Ruddy ducks, introduced from North
America and now considerably more abundant. Just as hybridization can
create species, so can it destroy them!
Inbreeding depression is another genetic process that may cause extinction.
This occurs when declining populations experience a decline in fitness
through increasing homozygosity of deleterious recessive alleles. It is normally
expressed through reduced fecundity and offspring viability. Inbreeding
154 DISCOVERING EVOLUTIONARY ECOLOGY

depression is readily created experimentally in captive populations. In

Drosophila melanogaster, about half the loss in fitness is due to recessive
lethal/semilethal mutations at about 5000 loci. The remainder is due to
many slightly deleterious alleles that are mildly recessive (Lande 1988).
When mating occurs among relatives, the chances of an offspring being
homozygous for these recessive alleles are increased, and their deleterious
effects are expressed. Inbreeding depression is not an automatic con-
sequence of small population size: gradual inbreeding tends to create little
depression of fitness because there is more opportunity for selection to
purge the population of homozygotes, thus removing the deleterious alleles.
Many insect and plant species consequently inbreed with little sign of fitness
depression. In contrast, when a population suddenly declines the effects of
inbreeding depression are more marked. As with Allee effects, the loss in
fitness can create a positive feedback ‘extinction vortex’, whereby in a small
population inbreeding occurs, fitness is reduced, which reduces population
size, which in turn increases inbreeding which reduces fitness and so on.
Ralls et al. (1979) looked for signs of inbreeding depression by measuring
juvenile survival in ungulate populations in captivity kept in either inbred or
outbred conditions. In 41 of 44 species examined, juvenile survival was lower
in the inbred populations. Inbreeding depression has also been inferred
from genetic and fitness studies of wild populations. The Chihuahua spruce
(Figure 13.7) was once widespread in Mexico but its range has contracted
into a few isolated populations of between 15 and 2400 trees. There is very
low gene flow between populations, and parents are more closely related
than half-sibs would be in a typical outbred population. Forty-five per cent
of seeds do not contain embryos, which is likely due to inbreeding depression
(Ledig et al. 1997). In some cases population extinction can be attributed to
inbreeding. In 42 subpopulations of the Glanville Fritillary (Figure 13.5) in
Finland, the risk of extinction was negatively correlated with the proportion
of heterozygosity in the populations (Saccheri et al. 1998), and one
generation of full sib-mating in the laboratory led to substantial inbreeding
depression.
In several wild populations the effects of inbreeding depression have been
reversed by transfer of individuals between populations (Hedrick and
Kalinowski 2000; Hedrick 2001). The Florida panther (Felis concolor coryi), a
subspecies of mountain lion, has been isolated in southern Florida since the
1900s and the current effective population size in recent years has been 25
or less. The males showed a high frequency of undescended testes and
deformed sperm. After the introduction of females from Texas in 1995, 14
offspring were produced from crosses with the Florida population, with a
reduced frequency of sperm defects.
 DEATH OF SPECIES 155

Fig. 13.7 A stand of Chihuahua Spruce, Picea chihuahuana, at El Realito, Chihuahua, Mexico. The trees,
once widespread, are now found in small isolated populations which show signs of inbreeding
depression. Photo courtesy of Thomas Ledig.

How big must a population be to be safe from inbreeding depression?

Franklin (1980) suggested the simple rule that an effective population size of
50 or less is at risk. This was based on very simple logic: in the absence of
selection the inbreeding coefficient increases by 1/2Ne per generation due
to random genetic drift, where Ne is the effective population size. Therefore,
to limit inbreeding to 1% per generation, a level tolerated in domesticated
animals, the effective population size should be 50. It is, however, only a
rough guide, and as Hedrick (2001) has pointed out, there are several cases of
endangered species, such as the Californian Condor, that have had founder
numbers of much less than 50 and that have rebounded from the brink of
extinction without apparent inbreeding depression.
Two other genetic changes can lower the survival prospects of small
populations: loss of genetic variation (and hence evolvability) through drift,
156 DISCOVERING EVOLUTIONARY ECOLOGY

and accumulation of deleterious mutations of small effect, also by drift. In

small populations, random fluctuations in gene frequency (drift) tend to
reduce genetic variation, leading to increased homozygosity and consequent
loss of adaptability. The process countering this loss of variation is mutation,
such that at any population size an equilibrium level of genetic variation is
achieved, which is lower for small populations. Franklin (1980) suggested
that an effective population size of 500 was sufficient to maintain genetic
variation that balanced the variation in the environment that necessitates
future evolutionary potential. This was based on data from Drosophila bristle
numbers suggesting that the rate of production of genetic variation in this
trait is about one thousandth of the environmental variance. However,
Lande (1995) has suggested that this rule of thumb may considerably over-
estimate the amount of genetic variation maintained. Recent data on
Drosophila suggest that only about 10% of the mutational variance is in fact
neutral or quasineutral (thus, replacing that lost through drift), 10 times less
than assumed above. The population size required to maintain sufficient
variation would then be 10 times as great (5,000).
In small populations, slightly deleterious mutations can also become fixed
by chance, a source of genetic stochasticity. Individually, these mutations do
not affect the risk of extinction, but collectively they can. Lande (1995) found
that if no selection is assumed (just mutation and drift), mean time to extinc-
tion from such new mutations is an exponential function of population size
(Figure 13.4). Thus, as for demographic stochasticity, at reasonably large
population sizes (Ne ⬎ 100), there is little risk of extinction from genetic
stochasticity. However, also assuming selection decreases the mean time
to extinction because it becomes a power of (asymptotically related to)
population size (Figure 13.4). As a result, genetic stochasticity might
be much more important at large population sizes, comparable to environ-
mental stochasticity, making very large populations necessary for long-term
viability.
Thus, a range of processes, both evolutionary and ecological, may
contribute towards extinction. These work both by making species rare and
by causing extinction once rare. Evolution can determine whether a species
becomes rare indirectly via selection on the ecological niche, and due to
co-evolutionary interactions with other species. Among small populations,
a range of evolutionary processes, such as drift, inbreeding depression, and
introgression, may finally cause extinction. Many of these may only be
important in very small populations, but others, such as loss of fitness due to
drift, may be active in comparatively large populations. In the next chapter,
we examine in more detail the characteristics of taxa that might be associated
with extinction, speciation, and species richness.
 DEATH OF SPECIES 157

13.4 Further reading

Caughley (1994) introduces extinction theory and explores many case
studies. Lande (1988) and Lawton (1995) also cover useful concepts. Levin
(2000) covers plants. Hedrick and Kalinowski (2000) and Hedrick (2001)
cover inbreeding.
 14 Big evolution

As long as you’re going to be thinking anyway, think big.

Donald Trump

Up until now the book has been concerned with the properties of individual
lineages and how they change through time: so-called ‘microevolution’. Our
focus in microevolution has been mostly on how the average properties of a
species may change and why they were of a certain observed value as opposed
to some other hypothetical values. This is a natural stance to take, because
intraspecific variation tends to be small in comparison with interspecific
variation, hence, we can justify to some extent treating species as single
points in the moving cloud of phenotype space (though see Chapter 11).
Evolutionary biologists can also consider the properties of clades of lineages:
so-called ‘macroevolution’ (Figure 14.1). Macroevolutionists tend to
Time

Morphospace

Fig. 14.1 Macroevolution versus microevolution. The main picture depicts a clade evolving through time
and morphospace. Species form and go extinct, and anagenesis occurs within species, and this
clade varies through time in species richness and disparity, the two major macroevolutionary
variables. Microevolution, in contrast, is concerned with individual species, and whether they
speciate, go extinct, or change in form (magnified section).
 BIG EVOLUTION 159

concentrate not on averages but on variation and diversity. Two types of

diversity, in particular, are of interest: the species richness of the clade
(number of lineages) and their diversity of form (disparity). While a single
species can in theory have a disparity, diversity of form among a number of
species can be much greater because the absence of interbreeding among
them means that evolutionary divergence can much more readily occur.
Species richness is a property that is obviously only interesting in clades as
opposed to individual species!
Underlying a clade’s species richness are the two cladogenetic processes of
speciation and extinction, the net effect of which (speciation minus extinc-
tion rate) gives rise to the net rate of diversification. Disparity is influenced
not only by these cladogenetic processes, which sprouts and prunes out the
different phenotypes, but also by the other evolutionary process of anagene-
sis, which moves the phenotypes of each lineage around. Macroevolution is
of interest to ecologists simply because the currencies of interest to
macroevolutionists, species richness, and diversity of form, are also curren-
cies that ecologists measure. Thus, ecologists attempting to explain numbers
of species or diversity of function frequently experience the need for
macroevolutionary explanation.
We can ask a number of questions about these properties and processes
affecting clades. How do they vary through time? How do they vary
geographically through space? How do they vary across clades? What prop-
erties of clades and the environment affect them? To answer these questions
we have the tried and tested combination of theory and data. We will discuss
the contribution of theory later in the chapter. The data come from two main
sources: the fossil record, and studies of extant clades. The studies of extant
clades come in three main kinds. First, studies of phylogenetic tree shape
enable us to derive information about cladogenetic processes. Second, stud-
ies of endangered or recently extinct species reveal underlying patterns in
extinction rates. Third, studies of recent adaptive radiations allow us to infer
the relative progress of diversity and disparity during a clade’s evolution. It
will be obvious from this discussion that not all sources of data impinge on
every question. In addition, some questions have received a relative dearth of
attention. In particular, most work on macroevolution has concentrated on
diversity rather than disparity, as will be obvious below (Table 14.1).

14.1 Macroevolutionary theory

Imagine a monophyletic clade evolving. What parameters are necessary to
describe its evolution? At minimum, we must describe the number of species,
but we may also wish to include a description of its disparity. Since describing
160 DISCOVERING EVOLUTIONARY ECOLOGY

Table 14.1 Sources of macroevolutionary data and questions they can address.

Clade Question Fossil Tree shape Historical Extant

property record extinctions radiations

Disparity When? ✓ ✓
Where? (✓) ✓
Who? (✓) (✓)
Speciation When? ✓ (✓)
Where? ✓ (✓)
Who? ✓ ✓
Extinction When? ✓ (✓)
Where? ✓ (✓) ✓
Who? ✓ ✓ ✓
Net When? ✓ ✓ ✓
diversification Where? ✓ ✓
Who? ✓ ✓

Notes: Brackets indicate current absence of studies.

disparity requires at minimum a description of cladogenesis, we will start by

simply describing the number of species and then add in disparity.
In essence the description of the species richness of a clade is very similar
to describing the number of individuals in a population. A clade starts off
with a single species, which speciates into two or more, which can also
speciate. If this is the only process acting on the clade, the clade multiplies
exponentially at a rate determined by the speciation rate. It is then simple to
describe the relationship between three parameters: the species richness of
the clade, its speciation rate, and the age of the clade. Knowledge of any two
of these parameters allows us to calculate the third. This simple exponential
growth model assumes a constant-rate deterministic speciation process with
no extinction. Of course, we can assume extinction without explicitly
modelling it, by just renaming the speciation parameter the net rate of diver-
sification, equal to the speciation rate minus the extinction rate.
This model has rather uninteresting dynamics but has two useful applica-
tions. First, if a clade’s growth conforms to it, it is an indication that no fur-
ther complex processes are at work, and therefore that the rate of
diversification is constant through time. By adding in the assumption of sto-
chasticity in rate, we can ask if the diversities of two clades imply simply sto-
chasticity in the same underlying rate of diversification, or if it implies
different underlying rates. Thus, this simple exponential model allows us to
answer two questions about the net rate of cladogenesis: is it constant (when)
and does it vary between clades (who)? The model was extensively used by
Stanley (1979) to compare radiation rates in different taxa, but originates
much earlier in work by the mathematician Yule (1924).
 BIG EVOLUTION 161

Log number of lineages

b
b–d

Origin of clade Time Present

Fig. 14.2 How to estimate speciation and extinction rates from a phylogeny of extant species. In a
semi-log plot of the number of lineages over time, the log number of lineages rises at a char-
acteristic rate, equal to the speciation minus the extinction rate (b ⫺ d). Species that are newly
formed, however, have not had a chance to go extinct, so the rate of formation of species near
the present is simply the speciation rate (b). The difference between the two slopes is therefore
the extinction rate (d). After Harvey et al. (1994), with permission from the Society for the
Study of Evolution.

A further simple step is to add in two discrete parameters describing the

rate of speciation and extinction separately. A clade whose speciation rate is
greater than its extinction rate grows in a roughly, but not exactly, exponential
fashion. Over much of its history growth rate is constant. However, early on
in its history its growth rate is faster, because extinction can only act on
lineages that have come into existence through speciation. This early rate of
increase represents the speciation rate unfettered by extinction. For clades
whose past history can only be inferred from the phylogeny of extant species,
this speciation rate can also be estimated as the rate of increase in branching
in the phylogeny near to the present (Figure 14.2), once again these species
are newly formed and have not had a chance to go extinct (Harvey et al.
1994). Once the speciation rate is known, the ‘normal’ rate of increase allows
the net rate of diversification to be calculated, and the difference between the
two is the extinction rate. Hence, knowledge of the dynamics of clade growth
allows the rate of speciation and extinction to be calculated from the
phylogenies of extant species. These rates can be calculated more simply
from the fossil record because origination and extinction are more directly
observable.
While a clade is radiating at constant (exponential) rate, what happens to
its morphological diversity? Perhaps the simplest model assumes that anage-
nesis can occur in a random walk fashion and that extinction and speciation
is random with respect to morphology. If morphological traits are assumed
162 DISCOVERING EVOLUTIONARY ECOLOGY

to be binary, as is practically the case with most morphological traits, then as

the clade radiates exponentially, its morphological diversity increases rapidly
at first but then asymptotes (Gavrilets 1999). The asymptote is caused simply
by the fact that the binary nature of the traits places strong (geometric)
constraints on continued morphological diversification. For continuous
traits, however, which are less constrained geometrically, disparity continues
to increase in a linear fashion (Foote 1996).
It seems unlikely in a finite environment that clades will grow unhindered
for ever at a constant rate. The simplest model that puts some constraint on
exponential growth is the so-called logistic model of Verhulst (1845),
whereby as the clade approaches a theoretical carrying capacity its net rate of
diversification is reduced by a feedback parameter. A wide variety of dynam-
ics are possible in the logistic model depending on the value of the feedback
parameter and the net rate of diversification, from stable equilibrium
through to chaotic dynamics. In addition, interactions between taxa are
possible within a logistic framework if the feedback is the result not just of
the clade’s own species richness but that of other clades too. This now allows
displacement dynamics whereby taxa can replace each other. More complex
models are possible but their necessity depends on the data to which we shall
now turn.

14.2 Trends in time

The most extensive data on temporal trends in macroevolution are from the
fossil record, in particular, that of shelly marine invertebrates, which has
been extensively documented. The data tend not to be analysed at the species
level because of various inaccuracies and biases that are accentuated at low
taxonomic levels.
The gross history of this group (Figure 14.3) shows a generally increasing
trend in family richness punctuated by sudden decreases, in particular, the
five big mass extinctions occurring at the end Ordovician, late Devonian, end
Permian, end Triassic, and end Cretaceous. These episodes were the end of
many groups. Over time, however, virtually every possible history of family
richness is demonstrated, from sudden increase and decline, to slow increase
and slow decline. Decreases in diversity following the mass extinctions are
then followed by increases in diversity, and this increase often takes the form
of a replacement of one taxon by another, a trend that suggests that interac-
tions between taxa are important in macroevolutionary dynamics, and
further that rates of diversification are limited in a logistic sense.
Extinction rates and speciation rates both exhibit a general decline over
time. This decline coincides with the replacement of taxa that exhibit
 BIG EVOLUTION 163

Terrestrial

1000

Number of families

0
e3
Marine
K
1000

O D
Pe Tr

e2

e1
P Mo
Ca
0
P M C

Fig. 14.3 How diversity has changed over time (P—Palaeozoic, M—Mesozoic, C—Cenozoic), according
to the fossil record. The well-known marine family record shows three possible equilibria
(e1, e2, e3) corresponding to the dominance of three faunas (Ca—Cambrian, P—Palaeozoic,
and M—Modern). The equilibria are punctuated by the five big mass extinction events
(O—Ordovician, D—Devonian, Pe—Permian, Tr—Triassic, and K—Cretaceous). The terrestrial
record in contrast looks much more exponential. After Benton (1995) with permission from
AAAS.

high rates of diversification and extinction and that are present at the start of
the Phanaerozoic (the Cambrian fauna of trilobites (Figure 14.4),
Monoplacophora and graptolites, the emergence of a second fauna that
dominates until the end of the Permian (the Palaeozoic fauna of crinoids,
cephalopods, and ostracods). This fauna displays intermediate speciation
and extinction rates. Finally, a modern fauna of gastropods, bivalve mol-
luscs, and echinoids dominates to the present with low rates of speciation
and extinction (Sepkoski 1999). A logistic model of clade growth of these
three faunas, with interaction purely through the logistic feedback term, with
these estimated diversification rates can predict the pattern of replacement
very accurately. For best accuracy the mass extinctions need to be imposed
on the system, but the pattern of faunal replacement and general trends in
164 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 14.4 The trilobite Dicranurus monstrosus, from the Devonian of Morocco, length about 4.5 cm
minus spines. Trilobites were a component of the ‘Cambrian Fauna’ one of the three great
marine faunas in the fossil record, characterized by a high origination rate and high extinction
rate. Photo courtesy of Richard Fortey and Claire Mellish, © The Natural History Museum,
London.

species richness occurs even without these. In fact the net effect of the mass
extinctions is to delay the replacements rather than to speed it up as is gener-
ally assumed (Kitchell and Carr 1985). Thus, the general large-scale trends
for the marine realm suggest logistic growth of clades with interaction and
replacement (Figure 14.3). Consistent with this, the patterns within clades
show a slow decline in rates of origination over time. There is also some
suggestion from the shape of phylogenies of extant taxa that rates of diversi-
fication tend to decline over time. The pattern of origination of bird families
is not consistent with a constant rate model of diversification, and suggests,
instead, that there has been a decline over time (Harvey et al. 1991). However,
 BIG EVOLUTION 165

the terrestrial fossil family-curve looks much more exponential (Figure 14.3),
and it also has been suggested that an exponential species-curve may under-
lie the logistic marine family-curve (Benton 1995). It does appear, therefore,
that not all groups have diversified in a logistic fashion, and that different
biotic realms may impose different constraints to diversification.
What of disparity? Again, variable patterns are seen in the fossil record. In
some groups, such as the Palaeozoic trilobites, crinoids, and insects, diversity
rose initially followed at length by a rise in disparity. In many groups how-
ever, including the Cambrian arthropods as a whole, many plant taxa,
Carboniferous ammonoids, and Paleozoic gastropods, disparity peaks early
and taxonomic richness rises afterwards (Foote 1997). This latter pattern is
perhaps the most common (cf. Schluter 2000, pp. 59–60), and is prima facie
consistent with the null pattern predicted by Gavrilets (1999). Foote (1997),
however, argues that for several reasons geometric constraints are unlikely to
be the only reason for the frequency of this pattern.
The relationship between disparity and taxonomic diversity can also be
addressed by comparisons of recent radiations. Schluter (2000) used com-
parisons of replicate radiations in finches, mammals, cichlid fishes, warblers,
and Anolis lizards to show that morphological diversity was greater in the
older clades, showing a continuing rise in disparity with time. However,
many of these clades do not show increased species richness with age, sug-
gesting that taxonomic richness reaches limits first, as paralleled in some fos-
sil groups. Studies of extant radiations also suggest that the evolutionary
rates both of morphological and taxonomic richness are higher when eco-
logical opportunity is greatest (Schluter 2000). For example, island radia-
tions often display greater morphological diversity and species richness than
their mainland counterparts.

14.3 Trends in space

Are there geographic patterns in macroevolutionary processes? The evidence
suggests so. Jablonski (1993) showed that the first occurrence of marine inver-
tebrate fossil groups was more often in the tropics than would be expected by
chance. This points to a higher origination rates of higher taxa there, perhaps
due to fostering of important morphological innovations that could give rise
to radiations. These innovations also occur in shallow water environments
more often than deep water environments. Stehli et al. (1972) also showed that
extant species of Foraminifera do not vary in age according to their present
day latitudes. This suggests that extinction rates do not vary with latitude, and
consequently invoking higher speciation rates in the tropics to explain their
higher diversity there. In contrast, the ages of higher taxa of corals, bivalves,
166 DISCOVERING EVOLUTIONARY ECOLOGY

More species Less species

Tropical Temperate

Fig. 14.5 Correlating diversification rates with latitude. Sister clades, sharing a common ancestor and
being the same age, are compared. One of the clades is largely tropical, the other temperate.
The difference in their species richness is also compared. If over a number of such pairs of sister
clades, the more tropical clade has the most species, then latitude affects the net rate of
diversification.

and barnacles decline towards the tropics. This suggests that speciation rates
are higher there (Sepkoski 1999). Studies of extant taxa also suggest similar
trends. Comparisons of the species richness of tropical and temperate sister
taxa in swallowtail butterflies and passerine birds suggests that the net rate of
diversification has been higher in the tropics (Cardillo 1999) (Figure 14.5).The
average age of avian tribes per latitudinal band increases towards the tropics
(Gaston and Blackburn 1996), suggesting a loss of older tribes towards higher
latitudes.In contrast,northern hemisphere freshwater fish species show a steep
decline in diversity north of 50 degrees, in areas recently glaciated.
Interestingly, however, the mitochondrial DNA distance between sister taxa is
lower at these high latitudes, suggesting a recent rise in speciation rate with lat-
itude following colonization of these lakes, a trend consistent with the notion
of ecological opportunity (Bernatchez and Wilson 1998).
Finally, reconstructing the history of geographic ranges in groups from
phylogenies of living species has suggested that some radiations have origi-
nated in the tropics (Bleiweiss 1998, Böhm and Mayhew 2005 ).What has hap-
pened to morphology in these communities? A few studies have compared the
morphospace occupied by temperate and tropical communities. Ricklefs and
O’Rourke (1975) found that disparity increased in tropical moth communi-
ties, with new morphologies not found in temperate communities. Similar
trends have been found in tropical marine communities (Roy and Foote 1997).

14.4 Trends across taxa

What of differences between taxa? There are consistent messages from both
fossil studies and studies of extant taxa. Across taxa, speciation rates and
extinction rates tend to be positively correlated (Stanley 1979; Sepkoski
 BIG EVOLUTION 167

Fig. 14.6 A weevil, family Curculionidae, walking on a pencil. Weevils are the most species-rich family,
in the most species-rich order, class, phylum, and kingdom. Photo © Peter Mayhew.

1999). One possible reason is that extinction and speciation are both
promoted by the same characteristics. Data from Jablonski (1986) on North
American Late Cretaceous gastropods illustrate a potential mechanism:
those taxa which have planktonic larvae, and hence are good dispersers, have
wide geographic ranges, and experience low extinction rates but also low
speciation rates. In contrast, taxa with direct development (no planktonic
stage) have small geographic ranges and tend to have high extinction rates,
but also high speciation rates. In many cases, however, we can identify taxa as
differing in rates of some cladogenetic process without knowing the reason
because the taxa are unreplicated; any feature that differs between the two
taxa could be responsible. For example, of the four major primate groups
(lemurs and their kin, new world monkeys, old world monkeys, and homi-
noids), only the old world monkeys differs from the others in net rate of
diversification, and furthermore, this is due to a higher rate of speciation
rather than lower rate of extinction (Purvis et al. 1995). This can be inferred
from the relative shapes of the phylogenies of the groups (Figure 14.2), which
for the old world monkeys shows a recent burst of branching indicating a
high rate of speciation over extinction. Similarly, based on their ages and
extant species richness, the beetles (Figure 14.6) show a higher net rate of
diversification than their sister group (Mayhew 2002), though there are
many possible reasons. Other such ‘significant radiations’ include the canids
(Bininda-Emonds et al. 1999), and the passerine and wading birds
(Figure 14.7) (Harvey et al. 1991).
168 DISCOVERING EVOLUTIONARY ECOLOGY

Fig. 14.7 A snipe, Gallinago gallinago, a member of the order Charadriformes; one of two bird orders
that are significant radiations. Photo courtesy of Stephane Moniotte.

One way of pinning down the reasons for differences in rates of cladogen-
esis is if replicate radiations have occurred in different groups with similar
characteristics (Figure 14.5). One can then compare the species richness of
these groups with those of their sister taxa which lack the characteristic or
‘key innovation’ in question. A large number of such studies have now been
done. A number of these are consistent with theory of speciation and
adaptive radition. For example, promiscuous insect groups have diversified
more rapidly than their sister groups with other mating systems (Arnqvist
et al. 2000). Promiscuous insects are those in which a single female is mated
by many males, and the genetic interests of male and female are predicted to
be divergent (see Chapter 7), leading to rapid evolution of sexual traits in an
intraspecific arms race. This rapid evolution could lead to rapid evolution of
reproductive isolation. Another trait that is linked with elevated species
richness is sexual selection. In passerine birds, a higher incidence of sexually
dimorphic coloration, indicative of sexual selection, is associated with
elevated species richness (Barraclough et al. 1995). Nectar spurs in plants
may have a similar effect by association with specific pollinators which
promote reproductive isolation (Hodges and Arnold 1995). Other traits
seem more likely to be linked with ecological opportunity. The link between
species richness and phytophagy, for example, seems likely to be related to
this (Mitter et al. 1988; Farrell 1998), as does latex and resin canals in plants,
which provide defence against herbivores (Farrell et al. 1991) allowing
‘escape and radiation’ (Chapter 11).
 BIG EVOLUTION 169

To get at extinction correlates we can also draw on the sobering recent

dataset on anthropogenic extinction. How do species that have recently gone
extinct differ from their close relatives that have not? In a number of groups,
correlates of extinction risk also match the theory of extinction (Fisher and
Owens 2004) (Chapter 13). In primates, carnivores and birds, small geo-
graphic range is correlated with extinction risk, as is dietary or habitat spe-
cialization in hoverflies, reptiles, birds, and primates, and low population
density in reptiles, carnivores, and primates. Large body size gives a higher
risk of extinction in birds, primates, and hoverflies.
There are few surprises here. What is particularly interesting, however, is
the interaction between the source of extinction threat and traits in birds.
Birds that are small are likely to be at risk from exploitation or introductions,
but not from habitat destruction. However, birds with specialized habitats
are more at risk of extinction from habitat destruction, but not from
exploitation or introductions (Owens and Bennett 2000). These threat inter-
actions may account for some of the contradictory results obtained from
some studies of extinction. For example, in mammals species in taxonomi-
cally isolated groups are most at risk of extinction whereas in angiosperms it
is species in species-rich groups. Body size has been reported to have positive
effects on extinction risk in primates and birds, but negative effects on all
mammals, and no effect in a large number of other studies. Generation time
has positive effects in carnivores, negative effects on birds, and no effects on
primates, reptiles, and hoverflies (Purvis et al. 2000).
The combination of fossil evidence and data from extant taxa then is
starting to make headway into the dynamics of clade characteristics and
differences between them. Taxonomic diversity and morphological diversity
are both promoted by ecological opportunity but they may not follow
directly parallel paths. In addition a number of characteristics affect
speciation and extinction rates that match the predictions of speciation and
extinction theory. In the next chapter we will examine large-scale patterns
from a more ecological perspective.

14.5 Further reading

Signor (1990) reviews the fossil evidence for trends in time, and Sepkoski
(1999) does the same across taxa. Purvis (1996), Barraclough et al. (1999),
Harvey et al. (1991), and Nee et al. (1996) review the use of phylogenies in
macroevolution. Purvis et al. (2000) and Fisher and Owens (2004) review
trends in extinction across taxa. Foote (1997) and Roy and Foote (1997)
review disparity.
15 Big ecology

Nature uses only the longest threads to weave her patterns, so that
each small piece of her fabric reveals the organization of the entire
tapestry.
Richard P. Feynman

Macroecology is the field that describes and attempts to explain statistical

patterns of abundance, distribution, and diversity (Brown 1995; Gaston and
Blackburn 2000). Macroecologists have identified a number of distinctive
patterns in these variables that seem to be telling us something important
about the rules governing ecology and evolution. In the development of the
field, identifying patterns constitutes what Gaston and Blackburn (1999)
describe as the ‘what’ stage of development, and is reasonably well advanced.
Ultimately, we would like to know ‘how’ and ‘why’ the patterns are generated.
This is not only because the patterns themselves seem to be important
features of our biotic universe, but also because underlying them are funda-
mental principles of the workings of ecology and evolution. We are groping
for the rules behind a great ecological and evolutionary game, perhaps the
greatest of them all. Since by far the most effort so far has gone into
documenting the patterns, the field is very much dominated by empirical
data and not by theory. Nonetheless, many hypotheses have now been
formulated to explain many of the patterns, and some data allow comment
on the merits of these hypotheses. One particular obstruction is that the
nature of the patterns restricts tests of the hypotheses. One particular
empirical option, experiment, is largely ruled out by the practical concerns
of scale. As we shall see, however, this does not exclude useful tests that can
distinguish between alternatives.
Several macroecological patterns rightly find prominence in ecology texts
but rather fewer find their way into evolutionary texts. This is probably a
mistake because, as I hope to show below, many of the patterns require us to
consider evolutionary processes and mechanisms as well as ecological ones.
Testing hypotheses about these mechanisms will therefore be aided by
evolutionary theory and data. Before discussing some of the theory and data,
however, let us outline some of the patterns we are trying to explain. The pat-
terns fall into two categories: associations between variables and frequency
 BIG ECOLOGY 171

Table 15.1 Some of the best known macroecological patterns

Type of pattern Variables Nature of pattern Ubiquity of pattern

concerned

Frequency Latitude of More species found Virtually ubiquitous

distribution occurrence at low latitudes than except at very small
high latitudes scales and taxa
Geographic range Bimodal or right Bimodal at small
size skewed and unimodal scales, unimodal and
right skewed at large
scales
Body size A right skewed Fairly ubiquitous
distribution with especially at large
many more small spatial and
than large species taxonomic scales
Abundance Most species rare Virtually ubiquitous
across scales and
taxa
Associations Body size and Larger bodied species Not ubiquitous but
between latitude at higher latitudes known from several
variables (Bergmann’s rule) bird assemblages
Body size and Negative relationship Very common, but
abundance found less frequently
at smaller scales
Geographic range Larger geographic Not ubiquitous but
and latitude ranges at high known from several
latitudes (Rapoport’s bird assemblages in
rule) the Holarctic
Abundance and More widespread Virtually ubiquitous
geographic range species achieve across scales and
higher local taxa
abundance
Species richness Positive relationship Virtually ubiquitous
and geographic area across scales and
taxa

distributions of single variables (Table 15.1). Each relationship can be

examined across a range of taxa, of different rank, and at a variety of spatial
scales. Spatial scale can mean the total area over which the relationship is
examined, and/or the area that comprises an individual data point (quadrat
size).When relationships are examined over large geographic areas, quadrats
naturally tend also to be large. For some relationships, data are relatively
taxonomically restricted; for example, geographic range is only well docu-
mented in a large number of species for a few higher taxa, mostly birds and
mammals. In contrast, data on local abundance or species richness come
from numerous studies on many taxa. Patterns are mostly examined over
quite large taxonomic ranks, such as classes. Changing the taxonomic rank
under consideration, may, as we shall see below, affect the pattern. Putting
172 DISCOVERING EVOLUTIONARY ECOLOGY

together frequency distributions is relatively straightforward and requires

few special considerations. However, relationships between two variables
often means comparative analyses of species characteristics, and then phylo-
genetically based comparative techniques should be used where possible
(Chapter 4). In the remainder of the chapter, we will consider two of the most
robust patterns and how evolutionary processes may contribute to them.

15.1 The evolution of frequency distributions in species

body sizes
Most species contributing to a large taxonomic rank are small. This appears
to be the overwhelming message from an examination of species body size
distributions (Figure 15.1). The distributions tend to be right skewed, even
on a logarithmic scale, and yet, the mode is seldom in the smallest size class
but slightly larger. These patterns probably hold regardless of potential
biases in the data (Blackburn and Gaston 1994). The skewness tends to be
less common at smaller spatial scales and when small taxonomic ranks
(i.e. families instead of classes) are examined (Maurer 1998a).
What processes might account for this pattern? Dial and Marzluff (1988)
and later McKinnley (1990) both developed verbal/graphical evolutionary
models based on evolutionary mechanisms. Dial and Marzluff assumed that
the difference between speciation and extinction (net diversification) is
greater at small than large body sizes (Figure 15.2). Then one can imagine
smaller-bodied lineages on the left-hand size of the frequency distribution
out-proliferating those on the right. McKinnley developed a passive
Frequency (species)

Body size

Fig. 15.1 A typical frequency distribution of species body sizes. The distribution is right skewed, with the
mode well to the left, but not at the smallest size class.
 BIG ECOLOGY 173

(a) (b) (c)

Fig. 15.2 Possible evolutionary pathways to the right-skewed frequency distribution of species body
sizes. (a) The ancestor was medium sized, and speciation occurs with equal frequency at all
body sizes, but larger species go extinct more rapidly. (b), Speciation occurs more frequently at
small body sizes. (c) The ancestral body size was small, and only rarely do some species become
large, hence anagenesis is biased.

diffusion scenario of body size evolution. He assumed first that clades

originate at small body sizes, and that there is a lower limit to body size.
Anagenetic processes must also play a role if a clade is to produce variable
body sizes; this might occur directionally (towards larger body sizes) or in a
random direction in a process akin to drift. As a result, the clade multiplies
fastest at smaller body sizes and develops a right-hand tail (Figure 15.2).
These ideas were first formalized in simulation studies of clade growth
and evolution by Maurer et al. (1992), whose approach was developed still
further by Kozlowski and Gawelczyk (2002). The simulation models contain
the following variables, which were examined to observe if they can produce
the shape and range of observed patterns. First, the body size of the clade
founder; second, the size-dependence of speciation rate; third, the size-
dependence of extinction rate, and fourth, the presence of a ‘reflecting
barrier’ at small body sizes. The latter represents a size limit below which
species cannot, or find it difficult to persist due to physiological constraints.
Kozlowski and Gawelczyk (2002) made this a gradual constraint such that
anagenesis to smaller body sizes becomes increasingly less likely below a
certain size. Fifth, body size change can occur either at speciation events, or
between them (anagenesis), although the direction of change was assumed
to conform to a random drift process (i.e. no trend).
The results of the models are mostly intuitive. The presence of the graded
reflecting barrier produces the short left-hand tail rather than a truncated dis-
tribution. The simulations produce a range of skewness as seen in nature due to
the stochastic nature of all the processes assumed. When the net diversification
174 DISCOVERING EVOLUTIONARY ECOLOGY

rate favours small species, more right skewed distributions are produced and
they tend to become more heavily right skewed. When the reflecting barrier
is present, small founder body size tends to increase skewness and propor-
tion of skew, although without it the opposite is the case. The timing of body
size change affects the relative importance of size-biased extinction and
speciation.
In general then, these evolutionary processes are capable of producing the
distributions seen in nature and certain combinations are more likely to do
so than others. For expiricists the challenge is now to quantify them in nature
and compare to theoretical predictions. Most work on this has been directed
at the macroevolutionary processes of speciation, extinction, and net
cladogenesis. Orme et al. (2002) have studied size-related trends in net
cladogenesis by observing the correlation between body size and species
richness across 38 species-level phylogenies across range of vertebrates and
invertebrates. Unfortunately, the results pose more questions than they
answer. There is no overall trend towards an association between body size
and net rate of cladogenesis, and in fact only one study does show this—a
genus of flies (Bitheca), where small body size increases the net rate of
cladogenesis. Furthermore, there was no relationship between body size and
distance from the root of the tree, so it appears that there are no strong and
consistent anagenetic trends. These do sometimes appear, however, in
particular clades; significant relationships are found in eight clades; in
primates, body size increases with distance from the root, while in carnivores
it decreases; overall, three of these cases show a positive relationship and five
are negative.
These results do seem at odds with the evolutionary models, and there are
several possible explanations; one is that there is actually no consistent skew
in the body size distributions of those taxa examined. If this is the case, the
studies can hardly be considered as explanations for positive skew. However,
the studies are also all much smaller than those generally used to describe the
patterns of interest, and, as described earlier, taxa of small taxonomic rank
tend to display less skewed body size distributions. It may be therefore that
we need to wait for larger phylogenies to become available. An alternative
possibility, not considered in the modelling studies, however, is that small
size is not a consistent correlate of species diversity but that large radiations
might tend to be small bodied. In fact this pattern has been documented in
mammalian clades by Gardezi and da Silva (1999). Such radiations would
have large effects on species richness but represent only single data points so
are unlikely to affect a general bias in cladogenesis. Another suite of studies
have examined the relationship between present day extinction risk and
body size in a number of groups. They generally conform to the studies of
cladogenesis in that some positive relationships are found, some negative
 BIG ECOLOGY 175

Proximate Ultimate
Patterns evolutionary ecological
processes forces

e.g. e.g. e.g.

frequency cladogenesis, climate,
distributions anagenesis niche space

Fig. 15.3 Processes contributing towards macroecological patterns at very large (e.g. global) scales. The
patterns are the direct result of evolutionary processes, which are thus proximate forces. These
forces occur in the way they do because of ecological forces, which are thus ultimate.

ones, and in some there is no relationship (Purvis et al. 2000). In general

then, there is presently scant evidence for any of the postulated evolutionary
processes that might have given rise to body size distribution patterns; the
reason for the mismatch must be found.
Even if we had a good knowledge of the evolutionary processes mentioned
above, that would remain an incomplete explanation, for underlying these
must themselves be causal processes (Figure 15.3). Three major ultimate
reasons have been postulated. First, there might be more resources or niches
available for smaller bodied species that might increase rates of cladogenesis
or affect anagenetic trends. In particular one can describe the relationship
between body size and number of individuals from measurements of
the habitat at different scales. Some studies of arthropod communities
(Morse et al. 1985; Shorrocks et al. 1991) show reasonably good fit between
the observed and predicted number of individuals. Of course the link
between amount of resource/niche space and evolutionary process remains
rather implicit. A second postulated ultimate process is that organisms
might be evolving towards some theoretical optimum size for the clade based
on reproductive power (Brown et al. 1993). The idea here is that within a
clade there will be some characteristic body size at which most energy
becomes available for reproduction, due to the differential between the mass
specific gains and losses of energy. Brown et al. (1993) predict that the
relationship between production rate (giving reproductive power) and body
size is shaped rather like species body size distributions. Other factors not in
the model can combine to drive a species away from the optimum and this is
more likely when the reproductive power is close to the maximum
achievable. Thus, the species distribution comes to resemble the body
size–production relationship.
The Brown et al. (1993) model apparently predicts the modes of some body
size distributions quite well (Maurer 1998; Roy et al. 2000). However, these
tests have themselves been criticized on the grounds of loose application of
the necessary parameters (Kozlowski 2002). In addition the model itself has
been criticized on a number of mathematical issues, points of internal
176 DISCOVERING EVOLUTIONARY ECOLOGY

consistency, and on biological and evolutionary grounds (for a review of

these points see Kozlowski 2002). In addition two further tests of the model,
both of which look for the proposed switch point in reproductive power at
the modal body mass, fail to confirm the model prediction (Jones and Purvis
1997; Symonds 1999). Once again, it is rather implicit which evolutionary
processes actually lead to the species body size distribution. Presumably ana-
genetic change is involved, as the model is really about what an individual
species’ body size should be. Presumably though, how many species get close
to that could be affected by speciation or extinction rates, which might
themselves be affected by the relatively reduced reproductive power away
from the optimum.
An alternative kind of optimality approach has been developed by
Kozlowski and Weiner (1997), in a model that developed Charnov’s (1991)
model of mammalian life history evolution. In Chapter 4, we saw how the
trade-off between production and mortality could combine to influence
when a species should mature, hence, its optimum adult body size. Kozlowski
and Weiner (1997) investigated a slightly more complex model than Charnov’s
in that they derive production from separate allometric relationships of
assimilation and respiration. Varying the parameters of this model by draw-
ing them at random from normal distributions happens to produce distribu-
tions of optimal species body sizes that are right skewed and shaped very like
those in nature (see also Kindlmann et al. 1999; Kozlowski and Gawelczyk
2002). Exactly how this optimization process affects the required evolution-
ary processes is again implicit; presumably ecological conditions make some
parameters more common than others, our third ultimate process, and this
affects the frequency of optima and hence of anagenetic change towards those
optima. Not all organisms optimize body size in this way (parasitoids and
birds are obvious exceptions—see Chapter 4), but the model illustrates the
general principle of how evolution on a per species basis towards individually
determined optima can lead to species body size distributions.
We will now turn to another well-known pattern and see how the same
sets of evolutionary principles can help explain that.

15.2 Evolution of the latitudinal gradient in species richness

Following on from the last chapter,we can refer to the same suite of evolutionary
processes to explain why, at larger taxonomic and spatial scales, species rich-
ness increases as latitude declines (Figure 15.4). Remarkably, the latitudinal
gradients are seldom considered in these terms, yet logically, they are likely to
be formed by the same suite of processes. Clades originate somewhere and,
all other processes being equal, their centre of species richness will remain
 BIG ECOLOGY 177

(a) (b)

Fig. 15.4 The latitudinal gradient in species richness seen from space. (a) Amazonia, with many species.
Deforestation along a road is visible in the bottom half of the picture. (b) Antarctica (the
McMurdo dry valleys) with very few species. Photos from the NASA Visible Earth image archive.

there.We must also consider how speciation and extinction vary with latitude
and the nature and frequency of latitudinal shifts in geographic range (ana-
genesis). Just as when body size distributions could be considered to be
bounded by a lower reflecting barrier, so geographical limits are placed on
latitudinal range shifts, provided by the positions of land masses and oceans,
for example, and on the fact that the latitude cannot vary below 0⬚ or above
90⬚ from the equator. Some of these geographical limits have been assessed in
a number of non-evolutionary models, and have been criticized exactly
because of their non-evolutionary nature. In fact the suite of evolutionary
processes that I have just described are the well-known processes of evolu-
tionary biogeography, and in a proximate sense explaining the latitudinal
gradient is an exercise in evolutionary biogeography. We would expect that
biased originations close to the equator as well as increased rates of cladoge-
nesis, reduced rates of extinction, and biased shifts in range towards the poles
would all contribute towards the observed latitudinal gradient (Figure 15.5).
Once again some progress has been made in documenting these processes.
In Chapter 14, we saw that the fossil record suggests higher origination rates
in the tropics, and sister-taxon comparisons of species richness suggest net
rates of cladogenesis that are sometimes higher in the tropics. Taxon age also
seems to vary with latitude in many instances, suggesting variation in under-
lying processes. In baboons, macaques, and their relatives (the Papionini),
the history of range changes has been reconstructed by mapping extant
distributions onto a phylogeny of species. These reconstructions show,
robustly, that ancestral distributions were equatorial, that tropical regions
have experienced more net cladogenesis, and that tropical regions have given
178 DISCOVERING EVOLUTIONARY ECOLOGY

(a) (b)

(c) (d)

Fig. 15.5 Four ways in which evolutionary processes may have given rise to the latitudinal gradient in
species richness (the circle with horizontal lines indicates the Earth with lines of latitude).
(a) Speciation is more frequent in the tropics. (b) Extinction is more frequent near the poles.
(c) Species move towards the equator over time. (d) The ancestor was tropical and changes in
range towards higher latitudes are rare.

rise to more dispersal, suggesting that they actually raise the species richness
of other regions (Böhm and Mayhew 2005).
In addition to documenting these processes, we will also need to find the
fundamental reasons behind them; ultimate processes, and once again we will
therein need to invoke ecological characteristics that have influenced them.
Here there have been so many different hypotheses that I cannot do justice to
them all, so I will mention one recent theory that makes quantitative predic-
tions, and has some empirical support (Allen et al. 2002). First, assume that
the total energy flux per species per unit area is relatively invariant (the ener-
getic equivalence rule). This is an empirically derived rule that appears to hold
and is derivable for some plants based on scaling relationships (Chapter 4).
Now assume that in a community of individuals made up of different
species of similar type and ecology (e.g. forest trees, reptiles, beetles), the
total number of individuals is determined by the number of species and the
number of individuals per species. If the energetic equivalence rule is to hold
then, at higher ambient temperatures, to maintain the same temperature
flux per species there must be fewer individuals per species. If there are fewer
individuals per species and the total number of individuals in the commu-
nity is approximately invariant with respect to temperature, then the num-
ber of species must increase. This logic actually allows quantitative
prediction of the relationship between species number and temperature for
ectothermic organisms once the exact metabolic relationships are taken into
account. The exact prediction is that a plot of ln species richness against
 BIG ECOLOGY 179

1000/K should give a slope of ⫺9. A number of tree, amphibian, fish, gastro-
pod, and parasite communities conform to this relationship across both lat-
itudinal and elevational gradients.
The model of Allen et al. (2002) is remarkable in several ways; first, it
makes quantitative predictions which can be more easily falsified than many
other models. Second, it derives species richness directly from a primary
environmental variable that is directly related to latitude, but actually varies
in a much broader sense. It is not specific about the evolutionary processes
but presumably the energetic equivalence rules could affect the likelihood of
range shifts, speciation, and extinction rates primarily by allowing species to
fit into a community in energetic terms or not. In many ways, the evolution-
ary processes can be considered to be irrelevant because as long as they pro-
duce enough species to fill up the energetic opportunities, the model will
hold. It is therefore a theory based largely on ecological equilibrium, with
some implicit evolutionary assumptions. Because of its novelty it has yet to
be fully assessed by the scientific community, so I will tempt fate by pointing
out two likely areas of criticism. The first is that it cannot predict the species
richness of endotherms, because, using the same logic, their metabolic rates
are invariant with ambient temperature, the number of individuals per
species does not change with temperature, and the number of species should
therefore remain the same. The second likely area of criticism is whether
most ecological communities are at energetic equilibrium. In general, this
goes counter to most ecologists’ notions of community assembly.
Thus, many macroecological patterns can be derived from consideration
of proximate evolutionary processes many of which have begun to be
quantified. Ultimately, these processes rely on the interplay between primary
ecological phenomena and the various proximate evolutionary processes.
Some possible primary phenomena have been identified and some data
supports their action, but the link between the ecological and evolutionary
phenomena remains poorly described theoretically and totally undescribed
empirically. There is much to be done to develop our understanding of how
and why these ecological patterns emerge and it is essential that evolutionary
biologists play a leading role.

15.3 Further reading

Brown (1995) and Gaston and Blackburn (2000) review macroecology, and
some recent reviews are in Blackburn and Gaston (2003). Kozlowski and
Gawelczyk (2002) review body size evolution, and Orme et al. (2002) cover the
phylogenetic evidence.Willig et al.(2004) provide an overview of the latitudinal
gradient in species richness from a statistical perspective.
 16 Combining in diversity

Coming together is a beginning; keeping together is progress; working

together is success.
Henry Ford

Over the previous 14 chapters, I have recounted the major questions that
evolutionary ecologists have addressed and some of the discoveries that they
have brought to light. In this chapter, I will draw on this experience to explain
what evolutionary ecology as a whole is trying to do. First, evolutionary ecol-
ogy is primarily about understanding biological diversity. Both ecology and
evolution are about understanding variation in a common set of variables, so
it is natural that they should sometimes need to work together to do that.
Second, evolution and ecology affect each other. Evolutionary ecology
explores the different ways in which this happens. Third, the different topics
in evolutionary ecology help explain each other. This occurs because many
evolutionary or ecological traits are dependent on each other, and because
many of the research tools are broadly applicable. Below I expand briefly on
these themes in the light of the preceding chapters.

16.1 Understanding biological diversity

I have stated that evolutionary ecology is collectively about understanding
diversity. Biological diversity is expressed through variation in the character-
istics of individuals, populations, communities, and clades. The characteris-
tics that differ include phenotypic traits, population size, and species richness
(Table 16.1). How has evolutionary ecology addressed these characteristics?

Table 16.1 The entities and traits that ecologists and evolutionary biologists study

Entity Individual Population Community Clade

Trait Phenotypes Number of Number of Number of

individuals species phenotypes or
species
 COMBINING IN DIVERSITY 181

Chapters 2 and 3 assessed how major innovations in phenotypic character-

istics occurred. These changes increased the total diversity of traits on the
planet, and many were successful in a macroevolutionary sense, increasing
species richness or other community properties. They explain why the world
is a rich and complex place. Chapters 4 to 7 considered how variation in phe-
notype arises even without major innovations discussed in earlier chapters.
They explain the little traits that make even closely related species or individ-
uals differ in form and function. In Chapter 8, we considered how population
size can be controlled through evolutionary processes. Rather interestingly,
we can not only explain variation in population size and dynamics through
trait evolution, but also trait evolution through population dynamics. The
result is not just variation in numbers but also variation in form and function.
In Chapters 9 to 11, we considered properties of species that are also prop-
erties of communities: their interactions with other species. Ecological inter-
actions evolve between generalism and specialism, between mutualism and
antagonism, or can follow a large number of co-evolutionary pathways, and
because of that the world is full of complex ecological communities and of
species that differ. In Chapters 12 to 15, we considered properties of com-
munities and clades, such as their species richness and constituent body
sizes. The rates of change in cladogenesis and morphological variation differ
over time and space and among lineages, and in doing so create the patterns
in traits and species richness that dominate our world. In these ways evolu-
tionary ecology addresses some of the major questions posed by ecology and
evolution, and also some of the major questions of any kind about the bios-
phere and its constituents.

16.2 How ecology and evolution interact

The previous chapters also illustrated how ecology and evolution affect each
other reciprocally and in diverse ways (Table 16.2). Throughout the book a
prominent issue has been how ecology shapes anagenetic evolution through
natural selection. In Chapters 4 to 7 we considered how the day-to-day
phenotypic variation among species evolves through selection pressures
exerted by their environment. These include changes in life history variables,
such as reproductive lifespan and age at maturity, changes in allocation of
reproductive effort to male verses female function and in the way sex is
determined, changes in dispersal ability and dormancy, and plastic changes
in phenotype exerted during the lifetime of individuals in response to
changes in their environment. In each of these cases, a consideration of the
fitness consequences of alternative phenotypes has led us to understand the
circumstances that would favour the genotypes that code for them.
182 DISCOVERING EVOLUTIONARY ECOLOGY

Table 16.2 How ecology and evolution affect each other

Influencing factor Influenced factor Reason

Ecology Phenotypic change Natural selection

Ecology Speciation Reproductive isolation,
ecological divergence
Ecology Extinction Reduction in range or
population size
Evolution Ecology of individuals Natural selection
Evolution Population ecology Rapid evolution or
adaptive behaviour
Evolution Community ecology Selection on species
interactions, changes in
species richness
Evolution Ecosystem ecology Evolutionary novelties
affect geochemical
cycling, ecospace
occupation

Ecology also affects cladogenesis through its impact on speciation.

Speciation involves reproductive isolation and ecological divergence, and
ecology is involved in both. Reproductive isolation results from the action of
selection or other ecological forces, such as drift, that require specific
ecological conditions, and are often a consequence of geographic isolation.
Speciation can also occur through hybridization, requiring proximity of two
close relatives. Ecological divergence occurs as a result of selection, or neutral
processes that require specific ecological conditions.
Ecology also affects evolution through its impact on extinction. Extinction
occurs when ecological forces combine to make the range or abundance of a
species small, after which a range of stochastic and deterministic events can
lead to irretrievable loss of fitness. Ecology also drives the diversity of form
and species richness among clades through competition and ecological
release.
Evolution also drives ecology. The ecology of individuals is affected by
plastic responses to changes in their environment. Evolutionary forces have
shaped these such that they are adaptive. Evolutionary forces, acting through
these plastic changes, can also affect how the size of a population will change
in response to changes in the environment. Rapid evolution can also affect
such changes. In addition, evolutionary forces can make a species rare or can
make it go extinct once rare. Evolution can also affect the properties of com-
munities: it can affect the number and type of species interactions, making
some species generalist and others specialist, some relationships antagonistic
and others mutualistic. In general, there has been an increase in the
 COMBINING IN DIVERSITY 183

complexity of earth’s communities and ecosystems over time, affected by

what might be termed major transitions in ecology. Evolutionary forces have
shaped those changes through stepwise addition of small evolutionary
changes, each of which was advantageous in the ecological circumstances of
the organisms concerned. Some of those changes involved an increase in
species richness, another property of communites. In fact anagenesis and
cladogenesis have affected many of the key macroecological patterns that
today dominate the living world.
Thus, evolutionary ecology has so far uncovered a rich array of interac-
tions between ecology and evolution, and knowledge of those interactions
has been the key to answering some important questions about our planet.

16.3 The interaction of questions in evolutionary ecology

Throughout the book several issues have been raised in more than one chapter,
implying that understanding one topic in evolutionary ecology can enhance
understanding of another. This occurs for two reasons. First, the topics are
biologically interdependent. Second, the techniques we can use to aid under-
standing are often general. These connections have also led many notable
evolutionary ecologists to move from one subject to the next. Let us now
recap the connections illustrated by the book (Figure 16.1).

Macroecology Major evolutionary

transitions
Macroevolution Major ecological
transitions

Extinction
mechanisms Life history
evolution

Speciation
Sex allocation
mechanisms

Co-evolution Dispersal
evolution
Evolution of
antagonism
Evolution of adaptive
plastic strategies
Evolution of
Evolution of
specialization
population dynamics

Fig. 16.1 The interactions between different subject areas in evolutionary ecology mentioned in this
book.
184 DISCOVERING EVOLUTIONARY ECOLOGY

In the last chapter, we saw that understanding macroecological patterns is

sometimes helped by an understanding of life history evolution. For exam-
ple, one macroecological pattern, the body size frequency distribution,
involves variation in a life history trait. Macroevolutionary forces, such as
rates of speciation and extinction, also affect many of these patterns, such as
why there are more species in the tropics. Macroevolutionary patterns are
also sometimes the result of major ecological or evolutionary transitions,
such as the evolution of sex or flight. They are often explicable by reference
to speciation and extinction theory, such as in the haplochromine cichlids
of Lake Victoria. Extinction is often the result of variation in life histories,
such as fecundity, ecological specialization, with specialists being more
extinction prone, or co-evolutionary forces, such that invasive species can
cause extinction of species that they have not co-evolved with. The theory of
extinction suggests that species that have small ranges are at risk of extinc-
tion, and this explains some macroevolutionary observations on extinction
rates across taxa.
Speciation can be enhanced by rapid evolutionary forces, such as sexual
selection, as in haplochromine cichlids, or conflict between evolutionary
entities, as in cytoplasmic male sterility, or changes in ecological specializa-
tion, as in apple maggot fly, or co-evolutionary forces, as in diversifying co-
evolution. Co-evolution characteristically results in changes in antagonism,
changes in specialization, changes in population dynamics. Changes in
antagonism have resulted in many of the major evolutionary transitions, can
affect specialization through expansion or contraction of the ecological
niche, can affect life history evolution through trade-offs with virulence,
population dynamics through its influence on vital rates, and can result in
extinction as a co-evolutionary process. Evolutionary changes in population
dynamics can be caused by adaptive changes in behaviour, as in many bird
and mammal species, can cause speciation through evolutionary branching
and extinction through adaptive suicide.
Plastic phenotypic responses can cause speciation, such as through mate
choice or conflict between entities. Dispersal and dormancy affect the degree
of antagonism between species and among species, and sex allocation
through determining population structure, and life history evolution
through trade-offs, and speciation, and extinction rates. Sex allocation is
affected by co-evolution and conflict of interest among entities and can
cause speciation. The major transitions in ecology are recognized by their
macroevolutionary effects, and some, such as the evolution of flowers and of
animal phyla, may have been the result of co-evolutionary forces. Others,
such as flight, have affected dispersal ability in many taxa.
Are any of these topics particularly pervasive? Cooperation and conflict,
dispersal, and life history evolution permeate many areas of evolutionary
 COMBINING IN DIVERSITY 185

ecology and might be considered central. Macroevolution is a force con-

tributing to phenotypic distribution of many of the traits of interest.
In addition to this web of interactions between the topics is a web of
mutually useful methods and tools. Optimization theory, evolutionary
stable strategies, adaptive dynamics, and population genetics are theoretical
tools that are broadly applicable in evolutionary ecology. Similarly, empirical
tools such as phylogenies and the comparative method, unmanipulated obser-
vations, and controlled experiments are useful in nearly all areas. Thus, the
tools of evolutionary ecology as well as the concepts are mutually supporting.
The picture provided by Figure 16.1 is a picture of evolutionary ecology, it
also describes an evolving biosphere and the forces that shape it. It is the
message of this book.

16.4 Prospects
While evolutionary ecologists have come a long way, there is obviously much
still to do. We still have a very sketchy empirical knowledge of many topics,
such as co-evolution, the evolution of dispersal, most of the early major tran-
sitions in ecology and evolution, macroevolution and even macroecological
trends for most taxa. Some areas are also theoretically poor, in some cases
due to a lack of data; macroecology and macroevolution notably so. Some
topics are recent and will inevitably see big advances in knowledge and direc-
tion; the evolution of population dynamics is one such area. Other areas are
mature in data and theory, and yet are still experiencing interesting new
changes, such as in sex allocation and life histories. Substantial progress in
understanding is therefore expected in all areas. Because of the interconnec-
tions between the different topics, we would expect progress in one area to
continue to advance others. Much of the progress will come through innov-
ative new techniques, both theoretical, like adaptive dynamics, and empiri-
cal, like molecular phylogenetics. However, evolutionary ecology has always
gained much, and will continue to gain much, by simple work that asks the
right questions.
Evolutionary ecology is wonderful for what it gives us and what it
demands of us. It gives us understanding of the natural world that we have
always striven for. It demands of us our best characteristics: ability, imagina-
tion, energy, and love. As contributors towards society and its progression,
we all deserve to appreciate and celebrate it.
References

Abrams, PA (2001). Predator–prey interactions. In CW Fox, DA Roff, and

DJ Fairburn, eds. Evolutionary ecology: concepts and case studies, pp. 277–289.
Oxford University Press, New York.
Allen, AP, Brown, JH, and Gillooly, JF (2002). Global biodiversity, biochemical
kinetics, and the energy equivalence rule. Science, 297, 1545–1548.
Allender, CJ, Seehausen, O, Knight, ME, Turner, GF, and Maclean, N (2003).
Divergent selection during speciation of Lake Malawi cichlid fish inferred from
parallel radiations in nuptial coloration. Proceedings of the National Academy of
Sciences, USA, 100, 14074–14079.
Anderson, RM and May, RM (1982). Coevolution of hosts and parasites.
Parasitology, 85, 411–426.
Arnold, ML (1997). Natural hybridization and evolution. Oxford University Press,
Oxford.
—— and Emms, SK (1998). Paradigm lost: natural hybridization and evolutionary
innovations. In DJ Howard and SH Berlocher, eds. Endless forms: species and speci-
ation, pp. 379–389. Oxford University Press, New York.
Arnqvist, G, Edvardsson, M, Friberg, U, and Nilsson, T (2000). Sexual conflict
promotes speciation in insects. Proceedings of the National Academy of Sciences,
USA, 97, 10460–10464.
Atlan, A (1992). Sex allocation in an hermaphroditic plant: the case of gynodioecy in
Thymus vulgaris. Journal of Evolutionary Biology, 5, 189–203.
Barton, NH (1998). Natural selection and random genetic drift as causes of
evolution on islands. In PR Grant, ed. Evolution on islands, pp. 102–123. Oxford
University Press, Oxford.
Barraclough, TG and Vogler, AP (2000). Detecting the geographical pattern of
speciation from species-level phylogenies. American Naturalist, 155, 419–434.
—— Harvey, PH, and Nee, S (1995). Sexual selection and taxonomic diversity
in passerine birds. Proceedings of the Royal Society of London, Series B, 259,
211–215.
—— Vogler, AP, and Harvey, PH (1999). Revealing the factors that promote
speciation. In AE Magurran and RM May, eds. Evolution of biological diversity,
pp. 202–219. Oxford University Press, Oxford.
Bawa, KW (1980). Evolution of dioecy in flowering plants. Annual Review of Ecology
and Systematics, 11, 15–39.
Bell, G (1982). The masterpiece of nature: the evolution and genetics of sexuality.
Croom Helm, London.
 REFERENCES 187

Benkman, CW (1999). The selection mosaic and diversifying coevolution between

crossbills and lodgepole pine. American Naturalist, 153 (Suppl.), S75–S91.
—— (2003). Divergent selection drives the adaptive radiation of crossbills.
Evolution, 57, 1176–1181.
Benner, SA, Ellington, AD and Tauer, A (1989). Modern metabolism as a palimpsest
of the RNA world. Proceedings of the National Academy of Sciences, USA, 86,
7054–7058.
Bennett, PM and Owens, IPF (2002). Evolutionary ecology of birds: life histories,
mating systems and extinction. Oxford University Press, Oxford.
Benton, MJ (1995). Diversification and extinction in the history of life. Science, 268,
52–58.
Berenbaum, MR (1983). Coumarins and caterpillars: a case for coevolution.
Evolution, 37, 163–179.
Bernatchez, L and Wilson, CC (1998). Comparative phylogeography of Nearctic and
Palearctic fishes. Molecular Ecology, 7, 431–452.
Beukeboom, LW and Vrijenhoek, RC (1998). Evolutionary genetics and ecology of
sperm-dependent parthenogenesis. Journal of Evolutionary Biology, 11, 755–782.
Bininda-Emonds, ORP, Gittleman, JL, and Purvis, A (1999). Building large trees by
combining phylogenetic information: a complete phylogeny of the extant
Carnivora (Mammalia). Biological Reviews, 74, 143–175.
Blackburn, TM (1991). A comparative examination of lifespan and fecundity in par-
asitoid Hymenoptera. Journal of Animal Ecology, 60, 151–164.
—— and Gaston, KJ (1994). The distribution of body sizes of the world’s bird
species. Oikos, 70, 127–130.
—— and —— eds. (2003). Macroecology: concepts and consequences. Blackwell,
Oxford.
Bleiweiss, R (1998). Origin of hummingbird faunas. Biological Journal of the Linnean
Society, 65, 77–97.
Böhm, M and Mayhew, PJ (2005). Historical biogeography and the latitudinal gradi-
ent of species richness in the Papionini (Primata: Cercopithecidae). Biological
Journal of the Linnean Society, 85, 235–246.
Bouton, N, Witte, F, van Alphen, JJM, Schenk, A, and Seehausen, O (1999). Local
adaptations in populations of rock-dwelling haplochromines (Pisces: Cichlidae)
from southern Lake Victoria. Proceedings of the Royal Society of London, Series B,
266, 355–360.
Bronstein, J (2001). Mutualisms. In CW Fox, DA Roff, and DJ Fairburn, eds.
Evolutionary ecology: Concepts and case studies, pp. 315–330. Oxford University
Press, New York.
Brown, JH (1995). Macroecology. Chicago University Press, Chicago, IL.
—— and West, GB, eds. (2000). Scaling in Biology. Oxford University Press,
Oxford.
—— Marquet, PA, and Taper, ML (1993). Evolution of body size: consequences of an
energetic definition of fitness. American Naturalist, 142, 573–584.
Brown, VK and Southwood, TRE (1983). Trophic diversity, niche breadth, and the
generation times of exopterygote insects in a secondary succession. Oecologia, 56,
220–225.
188 REFERENCES

Buchner, P (1965). Endosymbiosis of animals with plant microorganisms. Wiley

Interscience, New York.
Bull, JJ (1983). Evolution of sex determining mechanisms. Benjamin/Cummings,
Menlo Park.
—— (1994). Virulence. Evolution, 48, 1423–1437.
Burgers, P and Chiappe, LM (1999). The wing of Archaeopteryx as a primary thrust
generator. Nature, 399, 60–62.
Burt,A and Trivers, R (1998). Genetic conflicts in genomic imprinting. Proceedings of
the Royal Society of London, Series B, 265, 2393–2397.
Capillon, C and Atlan, A (1999). Evolution of driving X chromosomes and resistance
factors in experimental populations of Drosophila simulans. Evolution, 53,
506–517.
Cardillo, M (1999). Latitude and rates of diversification in birds and butterflies.
Proceedings of the Royal Society of London, Series B, 266, 1221–1225.
Carlquist, S (1965). Island life. The Natural History Press, New York.
—— (1974). Island biology. Columbia University Press, New York.
Carroll, SB, Weatherbee, SD, and Langeland, JA (1995). Homeotic genes and the
regulation and evolution of insect wing number. Nature, 375, 58–61.
Caughley, G (1994). Directions in conservation biology. Journal of Animal Ecology,
63, 215–244.
Charnov, EL (1976). Optimal foraging: the marginal value theorem. Theoretical
Population Biology, 9, 129–136.
—— (1982). The theory of sex allocation. Princeton University Press, Princeton.
—— (1991). Evolution of life history variation in female mammals. Proceedings of
the National Academy of Sciences, USA, 88, 1134–1137.
—— (1993). Life history invariants: some explorations of symmetry in evolutionary
ecology. Oxford University Press, Oxford.
—— (2001). Evolution of mammal life histories. Evolutionary Ecology Research, 3,
521–535.
—— (2004). The optimal balance between growth rate and survival in mammals.
Evolutionary Ecology Research, 6, 307–313.
—— and Bull, JJ (1977). When is sex environmentally determined? Nature, 266,
828–830.
—— and Schaffer, WM (1973). Life history consequences of natural selection: Cole’s
result revisited. American Naturalist, 107, 791–793.
—— Maynard Smith, J and Bull, JJ (1976). Why be an hermaphrodite? Nature, 263,
125–126.
—— Los-den Hartogh, RL, Jones, WT and van den Assem, J (1981). Sex ratio evolu-
tion in a variable environment. Nature, 289, 27–33.
Cleveland, LR (1947). The origin and evolution of meiosis. Science, 105,
287–289.
Cody, ML and Overton, JMC (1996). Short-term evolution of reduced dispersal in
island plant populations. Journal of Ecology, 84, 53–61.
Cohen, D (1966). Optimizing reproduction in a randomly varying environment.
Journal of Theoretical Biology, 12, 119–129.
 REFERENCES 189

Connell, JH (1971). On the role of natural enemies in preventing competitive

exclusion in some marine animals and in rain forest trees. In PJ Den Boer and
G Gradwell, eds. Dynamics of populations, pp. 298–312. Pudoc, Wageningen.
Conover, DO and Munch, SB (2002). Sustaining fisheries yields over evolutionary
time scales. Science, 297, 94–96.
—— and Vanvorhees, DA (1990). Evolution of a balanced sex ratio by frequency-
dependent selection in a fish. Science, 250, 1556–1558.
Coope, GR (1973). Tibetan species of dung beetle from late-Pleistocene deposits in
England. Nature, 245, 335–336.
Cosmides, LM and Tooby, J (1981). Cytoplasmic inheritance and intragenomic
conflict. Journal of Theoretical Biology, 89, 83–129.
Crews, D (1994). Constraints to parthenogenesis. In RV Short and E Balaban, eds.
The differences between the sexes, pp. 23–49. Cambridge University Press,
Cambridge.
Cwynar, LC and MacDonald, GM (1987). Geographic variation in lodge-pole pine in
relation to population history. American Naturalist, 129, 463–469.
Davies, NB and Brooke, M de L (1989). An experimental study of coevolution
between the cuckoo, Cuculus canoris, and its hosts, I and II. Journal of Animal
Ecology, 58, 207–236.
Darwin, C (1859). On the origin of species by means of natural selection. John Murray,
London.
—— (1871). The descent of man and selection in relation to sex. John Murray,
London.
de Haan, AA, Koelewijn, HP, Hundscheid, MPJ, and van Damme, JJM. (1997). The
dynamics of gynodioecy in Plantago lanceolata L. II. Mode of action and frequen-
cies of restorer genes. Genetics, 147, 1317–1328.
de Kroon, H and Hutchings, MJ (1995). Morphological plasticity in clonal plants: the
foraging concept revisited. Journal of Ecology, 83, 143–152.
de Queiroz, A (1998). Interpreting sister-group tests of key innovation hypotheses.
Systematic Biology, 47, 710–718.
Denno, RF, Roderick, GK, Olmstead, KL, and Döbel, HG (1991). Density-related
migration in planthoppers (Homoptera: Delphacidae): the role of habitat
persistence. American Naturalist, 138, 1513–1541.
Dial, KD (2003). Wing-assisted incline running and the evolution of flight. Science,
299, 402–404.
Dial, KP and Marzluff, JM (1988). Are the smallest organisms the most diverse?
Ecology, 69, 1620–1624.
Dieckmann, U (1997). Can adaptive dynamics invade? Trends in Ecology and
Evolution, 12, 128–131.
—— and Doebeli, M. (1999). On the origin of species by sympatric speciation.
Nature, 400, 354–357.
Dudley, R (2000). The evolutionary physiology of animal flight. Annual Review of
Ecology and Systematics, 62, 135–155.
Dyer, LA (1995). Tasty generalists and nasty specialists? Antipredator mechanisms in
tropical Lepidoptera larvae. Ecology, 76, 1483–1496.
190 REFERENCES

Dyson, EA and Hurst, GDD (2004). Persistence of an extreme sex-ratio bias in a

natural population. Proceedings of the National Academy of Sciences, USA, 101,
6520–6523.
Ebert, D and Herre, EA (1996). The evolution of parasitic diseases. Parasitology
Today, 12, 96–101.
Edwards, AVF (1998). Natural selection and the sex ratio: Fisher’s sources. American
Naturalist, 151, 564–569.
Edwards, PJ, Kollmann, J, and Fleischmann, K (2002). Life history evolution in
Lodoicea maldivica (Arecaceae). Nordic Journal of Botany, 22, 227–237.
Ellner, S (1986). Germination dimorphisms and parent-offspring conflict in seed
germination. Journal of Theoretical Biology, 123, 173–185.
—— (1987). Competition and dormancy—a reanalysis and review. American
Naturalist, 130, 798–803.
Enquist, BJ, Brown, JH, and West, GB (1998). Scaling of plant energetics and
population density. Nature, 395, 163–165.
—— West, GB, Charnov, EL, and Brown, JH (1999). Allometric scaling of
production and life history variation in vascular plants. Nature, 401, 907–911.
Farrell, BD (1998). ‘Inordinate fondness’ explained: why are there so many beetles?
Science, 281, 555–559.
—— Dussourd, DE, and Mitter, C (1991). Escalation of plant defenses: do latex and
resin canals spur plant diversification? American Naturalist, 138, 881–900.
Feder, JL (1998). The Apple Maggot Fly, Rhagoletis pomonella: flies in the face of con-
ventional wisdom about speciation? In DJ Howard and SH Berlocher, eds. Endless
forms: species and speciation, pp. 130–144. Oxford University Press, New York.
Fenchel, T and Finlay, BL (1995). Ecology and evolution in anoxic worlds. Oxford
University Press, Oxford.
Fenner, F and Ratcliffe, FN (1965). Myxomatosis. Cambridge University Press,
Cambridge.
Ferriere, R, Bronstein, JL, Rinaldi, S, Law, R, and Gauduchon, M (2002). Cheating
and the evolutionary stability of mutualisms. Proceedings of the Royal Society of
London, Series B, 269, 773–788.
Fisher, DO and Owens, IPF (2004). The comparative method in conservation biol-
ogy. Trends in Ecology and Evolution, 19, 391–398.
Fisher, RA (1930). The genetical theory of natural selection. Oxford University Press,
Oxford.
Foote, M (1996). Models of morphological diversification. In D Jablonski, DH Erwin,
and JH Lipps, eds. Evolutionary paleobiology, pp. 62–86. University of Chicago
Press, Chicago.
—— (1997). The evolution of morphological diversity. Annual Review of Ecology
and Systematics, 28, 129–152.
Franco, M and Silvertown, J (1997). Life history variation in plants: an exploration of
the fast-slow continuum hypothesis. In J Silvertown, M Franco, and JL Harper,
eds. Plant life histories: ecology, phylogeny and evolution, pp. 210–227. Cambridge
University Press, Cambridge.
Frank, SA (1996). Models of parasite virulence. Quarterly Review of Biology, 71, 37–78.
 REFERENCES 191

—— (1997). Models of symbiosis. American Naturalist, 150 (Suppl.), S80–S99.

—— (2002). A touchstone in the study of adaptation. Evolution, 56, 2561–2564.
Franklin, IR (1980). Evolutionary change in small populations. In ME Soulé and BA
Wilcox, eds. Conservation biology: an evolutionary-ecological perspective,
pp. 135–149. Sinauer, Sunderland.
Fretwell, SD and Lucas, HL (1970). On territorial behaviour and other factors
influencing habitat distribution in birds. Acta Biotheoretica, 19, 16–36.
Fryer, G and Iles, TD (1972). The cichlid fishes of the Great Lakes of Africa: their biology
and evolution. Oliver and Boyd, Edinburgh.
Futuyma, DJ (2001). Ecological specialization and generalization. In CW Fox,
DA Roff, and DJ Fairburn, eds. Evolutionary ecology: concepts and case studies,
pp. 177–189. Oxford University Press, New York.
—— and Moreno, G (1988). The evolution of ecological specialization. Annual
Review of Ecology and Systematics, 19, 207–233.
Galis, F and Metz, JAJ (1998). Why are there so many cichlids? Trends in Ecology and
Evolution, 13, 1–2.
Ganeshaiah, KN and Uma Shaanker, R (1988). Seed abortion in wind-dispersed
pods of Dalbergia sissoo: maternal regulation or sibling rivalry? Oecologia, 75,
135–139.
Gardezi, T and da Silva, J (1999). Diversity in relation to body size in mammals: a
comparative study. American Naturalist, 153, 110–123.
Garner, JP, Taylor, GK, and Thomas, ALR (1999). On the origins of birds: the
sequence of character acquisition in the evolution of avian flight. Proceedings of the
Royal Society of London, Series B, 266, 1259–1266.
Gaston, KJ and Blackburn, TM (1996). The tropics as a museum of biological
diversity: an analysis of the New World avifauna. Proceedings of the Royal Society of
London, Series B, 263, 63–68.
—— and —— (1999). A critique for macroecology. Oikos, 84, 353–368.
—— and —— (2000). Pattern and process in macroecology. Blackwell, Oxford.
Gavrilets, S (1999). Dynamics of clade diversification on the morphological
hypercube. Proceedings of the Royal Society of London, Series B, 266, 817–824.
—— (2003). Models of speciation: what have we learned in 40 years? Evolution, 57,
2197–2215.
Gemmill, AW, Skorping, A, and Read, AF (1999). Optimal timing of first reproduc-
tion in parasitic nematodes. Journal of Evolutionary Biology, 12, 1148–1156.
George, JC, Bada, J, Zeh, J et al. (1999). Age and growth estimates of bowhead whales
(Balaena mysticetus) via aspartic acid racemization. Canadian Journal of Zoology,
77, 571–580.
Gleeson, SK and Fry, JE (1997). Root proliferation and marginal patch value. Oikos,
79, 387–393.
Godfray, HCJ (1994). Parasitoids: behavioral and evolutionary ecology. Princeton
University Press, Princeton.
Goss-Custard, JD and Sutherland, WJ (1997). Individual behaviour, populations
and conservation. In JR Krebs and NB Davies, eds. Behavioural ecology: an
evolutionary approach, 4th edn, pp. 373–395. Blackwell, Oxford.
192 REFERENCES

Grafen, A (1984). Natural selection, kin selection and group selection. In JR Krebs
and NB Davies, eds. Behavioural ecology, an evolutionary approach, 2nd edn,
pp. 62–84. Blackwell, Oxford.
Grant, BR and Grant, PR (1993). Evolution of Darwin’s finches caused by a rare
climatic event. Proceedings of the Royal Society of London, Series B, 251, 111–117.
Grant, PR, ed. (1998). Evolution on islands. Oxford University Press, Oxford.
—— and Grant, BR (1996). Speciation and hybridization in island birds.
Philosophical Transactions of the Royal Society of London, Series B, 351, 765–772.
Gyllenberg, M, Parvinen, K, and Dieckmann, U (2002). Evolutionary suicide and
evolution of dispersal in structured metapopulations. Journal of Mathematical
Biology, 45, 79–105.
Hahn, BH, Shaw, GM, de Cock, KM, and Sharp, PM (2000). Aids as a zoonosis:
scientific and public health considerations. Science, 287, 607–614.
Haig, D and Westoby, M (1991). Genomic imprinting in endosperm: its effect on
seed development in crosses between species, and between different ploidies of the
same species, and its implications for the evolution of apomixis. Proceedings of the
Royal Society of London, Series B, 333, 1–13.
Hamilton, WD (1963). The evolution of altruistic behaviour. American Naturalist,
97, 354–356.
—— (1964a). The genetical evolution of social behaviour I. Journal of Theoretical
Biology, 7, 1–16.
—— (1964b). The genetical evolution of social behaviour II. Journal of Theoretical
Biology, 7, 17–52.
—— (1967). Extraordinary sex ratios. Science, 156, 477–488.
—— (1980). Sex versus non-sex versus parasite. Oikos, 35, 282–290.
—— (1996). Narrow roads of gene land. Vol. 1 Evolution of social behaviour.
W.H. Freeman Spektrum, Oxford.
—— and May, RM (1977). Dispersal in stable habitats. Nature, 269, 578–581.
Hanski, I and Ovaskainen, O (2000). The metapopulation capacity of a fragmented
landscape. Nature, 404, 755–758.
Hardin, G (1968). The tragedy of the commons. Science, 162, 1243–1248.
Hardy, ICW, ed. (2002). Sex ratios: concepts and research methods. Cambridge
University Press, Cambridge.
Harvey, PH and Pagel, MD (1991). The comparative method in evolutionary biology.
Oxford University Press, Oxford.
—— and Purvis, A (1999). Understanding the ecological and evolutionary reasons
for life history variation: mammals as a case study. In JM McGlade, ed. Advanced
Ecological Theory: Principles and Applications, pp. 232–248. Blackwell, Oxford.
—— and Zammuto, RM (1985). Patterns of mortality and age at first reproduction
in natural populations of mammals. Nature, 315, 319–320.
—— Nee, S, Mooers, AO, and Partridge, L (1991). These hierarchical views of life:
phylogenies and metapopulations. In RJ Berry, TJ Crawford, and GM Hewitt, eds.
Genes in Ecology, pp. 123–137. Blackwell, Oxford.
—— May, RM, and Nee, S (1994). Phylogenies without fossils. Evolution, 48, 523–529.
Hawthorne, DJ and Via, S (2001). Genetic linkage of ecological specialization and
reproductive isolation in pea aphids. Nature, 412, 904–907.
 REFERENCES 193

Hedrick, PW (1986). Genetic polymorphism in heterogeneous environments: a

decade later. Annual Review of Ecology and Systematics, 17, 535–566.
—— (2001). Evolutionary conservation biology. In CW Fox, DA Roff, and DJ
Fairburn, eds. Evolutionary ecology: concepts and case studies, pp. 371–383. Oxford
University Press, New York.
—— and Kalinowski, ST (2000). Inbreeding depression in conservation biology.
Annual Review of Ecology and Systematics, 31, 139–161.
Heino, M (1998). Management of evolving fish stocks. Canadian Journal of Fisheries
and Aquatic Sciences, 55, 1971–1982
Herre, EA (1985). Sex ratio adjustment in fig wasps. Science, 228, 896–898.
—— (1993). Population structure and the evolution of virulence in nematode
parasites of fig wasps. Science, 259, 1442–1445.
Herrero, M and Hormaza, JI (1996). Pistil strategies controlling pollen tube growth.
Sexual Plant Reproduction, 9, 343–347.
Hochberg, ME, Gomulkiewicz, R, Holt, RD, and Thompson JN (2000). Weak sinks
could cradle mutualisms—strong sources should harbor pathogens. Journal of
Evolutionary Biology, 13, 213–222.
Hodges, SA and Arnold, ML (1995). Spurring plant diversification: are floral nectar
spurs a key evolutionary innovation? Proceedings of the Royal Society of London,
Series B, 262, 343–348.
Holmes, EC (2001). On the origin and evolution of the human immunodeficiency
virus (HIV). Biological Reviews, 76, 239–254.
Howard, DJ and Berlocher, SH, eds. (1998). Endless forms: species and speciation.
Oxford University Press, New York.
Hughes, CL, Hill, JK, and Dytham, C (2003). Evolutionary trade-offs between
reproduction and dispersal in populations at expanding range boundaries.
Proceedings of the Royal Society of London, Series B, 270, S147–S150.
Hurst, LD and Hamilton, WD (1992). Cytoplasmic fusion and the nature of the
sexes. Proceedings of the Royal Society of London, Series B, 247, 189–194.
Hutchinson, GE (1959). Homage to Santa Rosalia, or why are there so many kinds of
animals? American Naturalist, 93, 245–249.
Hyatt, LA and Evans, AS (1998). Is decreased germination fraction associated with
risk of sibling competition? Oikos, 83,29–35.
Imbert, E and Ronce, O (2001). Phenotypic plasticity for dispersal ability in the seed
heteromorphic Crepis sancta (Asteraceae). Oikos, 93, 126–134.
Jablonski, D (1986). Larval ecology and macroevolution of marine invertebrates.
Bulletin of Marine Science, 39, 565–587.
—— (1993). The tropics as a source of evolutionary novelty through geological time.
Nature, 364, 142–144.
Jaenike, J (1990). Host specialization in phytophagous insects. Annual Review of
Ecology and Systematics, 21, 243–273.
Janzen, DH (1970). Herbivores and the number of tree species in tropical forests.
American Naturalist, 104, 501–528.
Jeon, KW (1972). Development of cellular dependence on infective organisms:
micrurgical studies in amoebas. Science, 176, 1122–1123.
194 REFERENCES

Johnson, ML and Gaines, MS (1990). Evolution of dispersal: theoretical models and

empirical tests using birds and mammals. Annual Review of Ecology and Systematics,
21, 449–480.
Johnson, MTJ and Agrawal, AA (2003). The ecological play of predator-prey dynam-
ics in an evolutionary theatre. Trends in Ecology and Evolution, 18, 549–551.
Jones, KE and Purvis, A (1997). An optimum body size for mammals? Comparative
evidence from bats. Functional Ecology, 11, 751–756.
Kanygin, AV (2001). The Ordovician phenomenon of explosive divergence of the
Earth’s organic realm: causes and effects on the biosphere evolution. Geologiya I
Geofizika, 42, 631–667.
Kawecki, TJ (1994). Accumulation of deleterious mutations and the evolutionary
cost of being a generalist. American Naturalist, 144, 833–838.
Kelly, CK (1990). Plant foraging: a marginal value model and coiling response in
Cuscuta subinclusa. Ecology, 71, 1916–1925.
Kiers, ET, Rousseau, RA, West, SA, and Denison, RF (2003). Host sanctions and the
legume-rhizobia mutualism. Nature, 425, 78–81.
Kindlmann, P, Dixon, AFG, and Dostálkova, I (1999). Does body size optimization
result in skewed body size distribution on a logarithmic scale? American
Naturalist, 153, 445–447.
Kitchell, JA and Carr, TR (1985). Non equilibrium model of diversification: faunal
turnover dynamics. In JW Valentine, ed. Phanerozoic diversity patterns: profiles in
macroevolution, pp. 277–309. Princeton University Press and Pacific Division,
AAAS, Princeton and San Fransisco.
Kocher, TD (2004). Adaptive evolution and explosive speciation: the cichlid fish
model. Nature Reviews Genetics, 5, 288–298.
Kondrashov, AS (1988) Deleterious mutations and the evolution of sexual
reproduction. Nature, 336, 435–440.
Kornfield, IL and Smith, PF (2000). African cichlid fishes: model systems for
evolutionary biology. Annual Review of Ecology and Systematics, 31, 163–196.
Kozlowski, J (2002). Theoretical and empirical status of Brown, Marquet and Taper’s
model of species-size distribution. Functional Ecology, 16, 540–542.
—— and Gawelczyk, AT (2002). Why are species’ body size distributions usually
skewed to the right? Functional Ecology, 16, 419–432.
—— and Konarzewski, M (2004). Is West, Brown and Enquist’s model of allometric
scaling mathematically correct and biologically relevant? Functional Ecology, 18,
283–289.
—— and Weiner, J (1997). Interspecific allometries are by-products of body size
optimization. American Naturalist, 149, 352–379.
—— Konarzewski, M, and Gawelczyk, AT (2003). Cell size as a link between non-
coding DNA and metabolic rate scaling. Proceedings of the National Academy of
Sciences, USA, 100, 14080–14085.
Krebs, JR and Davies, NB (1993). An introduction to behavioural ecology 3rd edn.
Blackwell, Oxford.
Kukalová-Peck, J (1978). Origin and evolution of insect wings and their relationship
to metamorphosis, as documented by the fossil record. Journal of Morphology,
156, 53–126.
 REFERENCES 195

Kunin, WE and Gaston, KJ (1997). The biology of rarity. Chapman & Hall, London.
—— and Schmida,A (1997). Plant reproductive traits as a function of local, regional,
and global abundance. Conservation Biology, 11, 183–192.
Lande, R (1988). Genetics and demography in biological conservation. Science, 241,
1455–1460.
—— (1993). Risk of population extinction from demographic and environmental
stochasticity and random catastrophes. American Naturalist, 142, 911–927.
—— (1995). Mutation and conservation. Conservation Biology, 9, 782–791.
—— Seehausen, O, and van Alphen, JJM (2001). Mechanisms of rapid sympatric
speciation by sex reversal and sexual selection in cichlid fish. Genetica, 112–113,
435–443.
Law, R (2000). Fishing, selection and phenotypic evolution. ICES Journal of Marine
Science, 57, 659–668.
—— and Dieckmann, U (1998). Symbiosis through exploitation and the merger of
lineages in evolution. Proceedings of the Royal Society of London, Series B, 265,
1245–1253.
—— and Grey, DR (1989). Evolution of yields from populations with age-specific
cropping. Evolutionary Ecology, 3, 343–359.
Lawton, JH (1995). Population dynamic principles. In JH Lawton and RM May, eds.
Extinction rates, pp. 147–163. Oxford University Press, Oxford.
Lazcano, A and Miller, SL (1999). On the origin of metabolic pathways. Journal of
Molecular Evolution, 49, 424–431.
Ledig, FT, Jacob-Cervantes,V, Hodgskiss, PD, and Eguiluz-Piedra T. (1997). Recent evo-
lution and divergence among populations of a rare Mexican endemic, Chihuahua
spruce, following Holocene climatic warming. Evolution, 51, 1815–1827.
Lenton, TM, Schellnhuber, HJ, and Szathmáry, E (2004). Climbing the co-evolution
ladder. Nature, 431, 913.
Lessells, CM (1991). The evolution of life histories. In JR Krebs and NB Davies, eds.
Behavioural ecology: an evolutionary approach, 3rd edn., pp. 32–68. Blackwell,
Oxford.
Lessios, HA (1998). The first stage of speciation as seen in organisms separated by the
Isthmus of Panama. In DJ Howard and SH Berlocher, eds. Endless forms: species
and speciation, pp. 186–201. Oxford University Press, New York.
Levin, DA (2000). The origin, expansion and demise of plant species. Oxford
University Press, Oxford.
Levins, R (1968). Evolution in changing environments. Princeton University Press,
Princeton.
Liem, KF (1973). Evolutionary strategies and morphological innovations: cichlid
pharyngeal jaws. Systematic Zoology, 22, 425–441.
Lively, CM (2001). Parasite–host interactions. In CW Fox, DA Roff, and DJ Fairburn,
eds. Evolutionary ecology: concepts and case studies, pp. 290–302. Oxford University
Press, New York.
Logan, GA, Hayes, JM, Hieshima, GB, and Summons, RE (1995). Terminal
Proterozoic reorganization of biogeochemical cycles. Nature, 376, 53–56.
MacArthur, RH and Pianka, ER (1966). On optimal use of a patchy environment.
American Naturalist, 100, 603–609.
196 REFERENCES

Majerus, MEN (2003). Sex wars: genes, bacteria, and biased sex ratios. Princeton
University Press, Princeton.
Marden, JH and Kramer, MG (1995). Locomotor performance of insects with
rudimentary wings. Nature, 377, 332–334.
Margulis, L and Bermudes, D (1985). Symbiosis as a mechanism of evolution: status
of cell symbiosis theory. Symbiosis, 1, 101–124.
Maurer, BA (1998a). The evolution of body size in birds. I. Evidence for non-random
diversification. Evolutionary Ecology, 12, 925–934.
—— (1998b). The evolution of body size in birds. II. The role of reproductive power.
Evolutionary Ecology, 12, 935–944.
—— Brown, JH, and Rusler, RD (1992). The micro and macro in body size evolu-
tion. Evolution, 46, 939–953.
Mayhew, PJ (1997). Adaptive patterns of host-plant selection by phytophagous
insects. Oikos, 79, 417–428.
—— (2002). Shifts in hexapod diversification and what Haldane could have said.
Proceedings of the Royal Society of London, Series B, 269, 969–974.
—— and Blackburn, TM (1999). Does development mode organize life history
evolution in the parasitoid Hymenoptera? Journal of Animal Ecology, 68,
906–916.
—— and Glaizot, O (2001). Integrating theory of clutch size and body size evolution
for parasitoids. Oikos, 92, 372–376.
—— and Hardy, ICW (1998). Nonsiblicidal behavior and the evolution of clutch size
in bethylid wasps. American Naturalist, 151, 409–424.
Maynard Smith, J (1966). Sympatric speciation. American Naturalist, 100, 637–650.
—— (1979). The effect of normalizing and disruptive selection on genes for recom-
bination. Genetical Research, 33, 121–128.
—— (1984). The ecology of sex. In JR Krebs and NB Davies, eds. Behavioural ecology: an
evolutionary approach, 2nd edn., pp. 201–210. Blackwell, Oxford.
—— and Szathmáry, E (1995). The major transitions in evolution. Freeman,
Oxford.
—— and —— (1999). The origins of life: from the birth of life to the origin of language.
Oxford University Press, Oxford.
Mayr, E (1940). Speciation phenomena in birds. American Naturalist, 74, 249–278.
—— (1963). Animal species and evolution. Harvard University Press, Cambridge.
McKinnley, ML (1990). Trends in body size evolution. In KJ McNamara, ed.
Evolutionary trends, pp. 75–118. University of Arizona Press, Tucson.
McLaughlin, JF, Hellmann, JJ, Boggs, CL, and Ehrlich, PR (2002). Climate change
hastens population extinctions. Proceedings of the National Academy of Sciences,
USA, 99, 6070–6074.
Meyer, A (1993). Phylogenetic relationships and evolutionary processes in African
cichlids. Trends in Ecology and Evolution, 8, 279–284.
Miller, B and Mullette, KJ (1985). Rehabilitation of an endangered Australian bird:
Lord Howe Woodhen Tricholimnas sylvestris. Biological Conservation, 34, 55–95.
Mitter, C, Farrell, B, and Wiegmann, B (1988). The phylogenetic study of adaptive
zones: has phytophagy promoted insect diversification? American Naturalist, 132,
107–128.
 REFERENCES 197

Mock, DW and Parker, GA (1997). The evolution of sibling rivalry. Oxford University
Press, Oxford.
Morse,DR,Lawton,JH,Dodson,MM,and Williamson,MH (1985).Fractal dimensions
of vegetation and the distribution of arthropod body lengths. Nature, 314, 731–733.
Myers, N, Mittermeier, RA, Mittermeier, CG, da Fonseca, GAB, and Kent J (2000).
Biodiversity hotspots for conservation priorities. Nature, 403, 853–858.
Nee, S, Barraclough, TG, and Harvey, PH (1996). Temporal changes in biodiversity:
detecting patterns and identifying causes. In KJ Gaston, ed. Biodiversity: A biology
of numbers and difference, pp. 230–252. Oxford University Press, Oxford.
Nowak, MA, Anderson, RM, McLean, AR, Wolfs, T, Goudsmit, J, and May, RM
(1991). Antigenic diversity thresholds and the development of AIDS. Science, 254,
963–969.
Orme, CDL, Isaac, NJB, and Purvis, A. (2002). Are most species small? Not within
species-level phylogenies. Proceedings of the Royal Society of London, Series B, 269,
1279–1287.
Orr, HA (1998). Testing natural selection versus genetic drift in phenotypic evolu-
tion using Quantitative Trait Locus data. Genetics, 149, 2099–2104.
Owens, IPF and Bennett, PM (2000). Ecological basis of extinction risk in birds:
habitat loss versus human persecution and introduced predators. Proceedings of
the National Academy of Sciences, USA, 97, 12144–12148.
—— —— and Harvey, PH (1999). Species richness among birds: body size, life
history, sexual selection or ecology? Proceedings of the Royal Society of London,
Series B, 266, 933–939.
Padian, K and Chiappe, LM (1998). The origin and early evolution of birds.
Biological Reviews, 73, 1–42.
Pake, CE and Venable, DL (1996). Seed banks in desert annuals: implications for
persistence and coexistence in variable environments. Ecology, 77, 1427–1435.
Parker, GA and Maynard Smith, J (1990). Optimality theory in evolutionary biology.
Nature, 348, 27–33.
Pellmyr, O and Huth, CJ (1994). Evolutionary stability of mutualism between yuccas
and yucca moths. Nature, 372, 257–260.
Perrin, N and Mazalov,V (2000). Local competition, inbreeding and the evolution of
sex-biased dispersal. American Naturalist, 155, 116–127.
Policansky, D (1987). Evolution, sex and sex allocation. Bioscience, 37, 466–468.
Prinzing, A (2003). Are generalists pressed for time? An interspecific test of the time-
limited disperser model. Ecology, 84, 1744–1755.
Promislow, DEL and Harvey, PH (1990). Living fast and dying young: a comparative
analysis of life history variation among mammals. Journal of Zoology, 220, 417–437.
Purvis,A (1996).Using interspecies phylogenies to test macroevolutionary hypotheses.
In PH Harvey, AJ Leigh Brown, J Maynard Smith, and S Nee, eds. New uses for new
phylogenies, pp. 153–168. Oxford University Press, Oxford.
—— and Harvey, PH (1995). Mammal life history evolution: a comparative test of
Charnov’s model. Journal of Zoology, 237, 259–283.
—— Nee, S, and Harvey, PH (1995). Macroevolutionary influences from primate
phylogeny. Proceedings of the Royal Society of London, Series B, 260, 329–333.
—— Jones, KE, and Mace, GM (2000). Extinction. Bioessays, 22, 1123–1133.
198 REFERENCES

Queller, DC (1983). Sexual selection in a hermaphroditic plant. Nature, 305,

706–708.
Ralls, K, Brugger, K, and Ballou, J (1979). Inbreeding and juvenile mortality in small
populations of ungulates. Science, 206, 1101–1103.
Ranius, T and Hedin, J (2001). The dispersal rate of a beetle, Osmoderma eremita,
living in tree hollows. Oecologia, 126, 363–370.
Rees, M (1993). Trade-offs among dispersal strategies in the British Flora. Nature,
366, 150–152.
—— (1997). Evolutionary ecology of seed dormancy and seed size. In J Silvertown,
M Franco, and JL Harper, eds. Plant life histories: ecology, phylogeny and evolution,
pp. 121–142. Cambridge University Press, Cambridge.
Ricklefs, RE and O’Rourke, K (1975). Aspect diversity in moths: a temperate-tropical
comparison. Evolution, 29, 313–324.
—— and Renner, SS (2000). Evolutionary flexibility and flowering plant diversity: a
comment on Dodd, Silvertown, and Chase. Evolution, 54, 1061–1065.
Rieseberg, LH, VanFossen, C, and Desrochers, AM (1995). Hybrid speciation
accompanied by genomic reorganization in wild sunflowers. Nature, 375,
313–316.
—— Whitton, J, and Gardner, K (1999). Hybrid zones and the genetic architecture
of a barrier to gene flow between two wild sunflower species. Genetics, 152,
713–727.
—— Raymond, O, Rosenthal, DM et al. (2003). Major ecological transitions in
annual sunflowers facilitated by hybridization. Science, 301, 1211–1216.
Rigaud, T (1997). Inherited microorganisms and sex determination of arthropod
hosts. In SL O’Neill,AA Hoffman, and JH Werren, eds. Influential passengers: inher-
ited microorganisms and arthropod reproduction, pp. 81–102. Oxford University
Press, Oxford.
Roff, DA (1990). The evolution of flightlessness in insects. Ecological Monographs, 60,
389–421.
—— (1992). The evolution of life histories: theory and analysis. Chapman & Hall,
New York.
—— (1994). The evolution of flightlessness: is history important? Evolutionary
Ecology, 8, 629–657.
Roughgarden, J (1995). Anolis lizards of the Caribbean: ecology, evolution and plate
tectonics. Oxford University Press, Oxford.
—— and Pacala, SW (1989). Taxon cycle among Anolis lizard populations: review of
the evidence. In D Otte and J Endler, eds. Speciation and its consequences,
pp. 403–432. Sinauer, Sunderland.
Roy, K and Foote, M (1997). Morphological approaches to measuring biodiversity.
Trends in Ecology and Evolution, 12, 277–281.
—— Jablonski, D, and Martien, KK (2000). Invariant size-frequency distributions
along a latitudinal gradient in marine bivalves. Proceedings of the National
Academy of Sciences, USA, 97, 13150–13155.
Rundle, HD, Mooers, AØ, and Whitlock, MC (1998). Single founder-flush events
and the evolution of reproductive isolation. Evolution, 52, 1850–1855.
 REFERENCES 199

Saccheri, I, Kuussaari, M, Kankare, M, Vikman, P, Fortelius, W, and Hanski, I

(1998). Inbreeding and extinction in a butterfly metapopulation. Nature, 392,
491–494.
Saumitou-Laprade, P, Cuguen, J, and Vernet, P (1994). Cytoplasmic male sterility in
plants: molecular evidence and the nucleocytoplasmic conflict. Trends in Ecology
and Evolution, 9, 431–435.
Schluter,D (1998).Ecological causes of speciation.In DJ Howard and SH Berlocher,eds.
Endless forms: species and speciation, pp. 114–129. Oxford University Press, New York.
—— (2000). The ecology of adaptive radiation. Oxford University Press, Oxford.
—— (2001). Ecological character displacement. In CW Fox, DA Roff, and DJ
Fairburn, eds. Evolutionary ecology: concepts and case studies, pp. 265–276. Oxford
University Press, New York.
Schuilthuisen, M (2001). Frogs, flies and dandelions: the making of species. Oxford
University Press, Oxford.
Searcy, KB and Macnair, MR (1993). Developmental selection in response to envi-
ronmental conditions of the maternal parent in Mimulus guttanus. Evolution, 47,
13–24.
Seehausen, O (2000). Explosive speciation rates and unusual species richness in
haplochromine cichlid fishes: effects of sexual selection. Advances in Ecological
Research, 31, 237–274.
—— and van Alphen, JJM (1998). The effect of male coloration on female mate
choice in closely related Lake Victoria cichlids (Haplochromis nyererei complex).
Behavioural Ecology and Sociobiology, 42, 1–8.
—— and —— (1999). Can sympatric speciation by disruptive sexual selection
explain rapid evolution of cichlid diversity in Lake Victoria? Ecology Letters, 2,
262–271.
—— —— and Witte, F (1997). Cichlid fish diversity threatened by eutrophication
that curbs sexual selection. Science, 277, 1808–1811.
—— Mayhew, PJ, and van Alphen, JJM (1999a). Evolution of colour patterns in East
African cichlid fish. Journal of Evolutionary Biology, 12, 514–534.
—— van Alphen, JJM, and Lande, R (1999b). Color polymorphism and sex ratio dis-
tortion in a cichlid fish as an incipient stage in sympatric speciation by sexual
selection. Ecology Letters, 2, 367–378.
Seger, J and Stubblefield, JW (2002). Models of sex ratio evolution. In Hardy, ICW,
ed. Sex ratios: concepts and research methods, pp. 2–25. Cambridge University
Press, Cambridge.
Sepkoski, JJ, Jr (1999). Rates of speciation in the fossil record. In AE Magurran and
RM May, eds. Evolution of biological diversity, pp. 260–282. Oxford University
Press, Oxford.
Sereno, PC (1999). The evolution of dinosaurs. Science, 284, 2137–2147.
Shorrocks, B, Marsters, J, Ward, I, and Evennett, PJ (1991). The fractal dimension of
lichens and the distribution of arthropod body lengths. Functional Ecology, 5,
457–460.
Signor, PW (1990). The geological history of diversity. Annual Review of Ecology and
Systematics, 21, 509–539.
200 REFERENCES

Silvertown, J and Gordon, DM (1989). A framework for plant behaviour. Annual

Review of Ecology and Systematics, 20, 349–366.
Simpson, GG (1953). The major features of evolution. Columbia University Press,
New York.
Singer, MC, Ng, D, and Thomas, CD (1988). Heritability of oviposition preference
and its relationship to offspring performance within a single insect population.
Evolution, 42, 977–985.
—— Thomas, CD, and Parmesan, C (1993). Rapid human-induced evolution of
insect-host associations. Nature, 366, 681–683.
Skogsmyr, I and Lankinen, Å (2000). Female assessment of good genes in stylar tis-
sue. Evolutionary Ecology Research, 2, 965–979.
—— and —— (2002). Sexual selection: an evolutionary force in plants? Biological
Reviews, 77, 537–562.
Slatkin, M (1980). Ecological character displacement. Ecology, 61, 163–177.
Solignac, M and Monnerot, M (1986). Race formation, speciation, and introgression
within Drosophila simulans, D. mauritiana, and D. sechellia inferred from
mitochondrial DNA analysis. Evolution, 40, 531–539.
Soulé, ME (1987). Viable populations for conservation. Cambridge University Press,
Cambridge.
Southwood, TRE (1962). Migration of terrestrial arthropods in relation to habitat.
Biological Reviews, 37, 171–214.
—— (1978). The components of diversity. In Mound LANW, ed. Diversity of insect
faunas, pp. 19–40. Blackwell, Oxford.
Southwood, TRE (2003). The story of life. Oxford University Press, Oxford.
Speakman, JR (2001). The evolution of flight and echolocation in bats: another leap
in the dark. Mammal Review, 31, 111–130.
Stanley, SM (1979). Macroevolution: pattern and process. Freeman, San Fransisco.
Stearns, SC, ed. (1987). The evolution of sex and its consequences. Birkauser, Basel.
—— (1992). The evolution of life histories. Oxford University Press, Oxford.
Stehli, FG, Douglas, RG, and Kafescioglu, IA (1972). Models for the evolution of
planktonic foraminifera. In Schopf, TJM, ed. Models in paleobiology, pp. 116–128.
Freeman, Cooper and Co, San Fransisco.
Steiner, KE and Whitehead, VB (1990). Pollinator adapation to oil-secreting
flowers—Rediviva and Diascia. Evolution, 44, 1701–1707.
Stephenson, AG and Winsor, JA (1986). Lotus corniculatus regulates offspring quality
through selective fruit abortion. Evolution 40: 453–458.
Stillman, RA, Goss-Custard, JD, West, AD et al. (2000). Predicting mortality in novel
environments: tests and sensitivity of a behaviour-based model. Journal of Applied
Ecology, 37, 564–588.
Stouthamer, R, Luck, RF, and Hamilton,WD (1990).Antibiotics cause parthenogenetic
Trichogramma (Hymenoptera/Trichogrammatidae) to revert to sex. Proceedings
of the National Academy of Sciences, USA, 87, 2424–2427.
Sutherland, S (1987). Why hermaphroditic plants produce many more flowers than
fruits: experimental tests with Agave mckelveyana. Evolution, 41, 750–759.
Sutherland, WJ (1996). Predicting the consequences of habitat loss for migrating
populations. Proceedings of the Royal Society of London, Series B, 263, 1325–1327.
 REFERENCES 201

—— and Norris, K (2002). Behavioural models of population growth rates: implica-

tions for conservation and prediction. Philosophical Transactions of the Royal
Society of London, Series B, 357, 1273–1284.
—— and Stillman, RA (1988). The foraging tactics of plants. Oikos, 52, 239–244.
Symonds, MRE (1999). Insectivore life histories: further evidence against an
optimum body size for mammals. Functional Ecology, 13, 508–513.
Szathmáry, E (1993). Do deleterious mutations act synergistically? Metabolic control
theory provides a partial answer. Genetics, 133, 127–132.
—— and Maynard Smith, J (1995). The major evolutionary transitions. Nature, 374,
227–232.
Taper, ML and Case, TJ (1985). Quantitative genetic model for the coevolution of
character displacement. Ecology, 66, 355–371.
Thomas, ALR and Norberg, RA (1996). Skimming the surface—the origin of flight
in insects? Trends in Ecology and Evolution, 11, 187–188.
Thompson, JN (1987). Symbiont-induced speciation. Biological Journal of the
Linnean Society, 32, 385–393.
—— (1989). Concepts of coevolution. Trends in Ecology and Evolution, 4,
179–183.
—— (1994). The coevolutionary process. University of Chicago Press, Chicago.
—— (1998). Rapid evolution as an ecological process. Trends in Ecology and
Evolution, 13, 329–332.
—— (1999). The raw material for coevolution. Oikos, 84, 5–16.
—— (2001). The geographic dynamics of coevolution. In CW Fox, DA Roff, and
DJ Fairburn, DJ, eds. Evolutionary ecology: concepts and case studies, pp. 331–343.
Oxford University Press, New York.
—— and Burdon, JJ (1992). Gene-for-gene coevolution between plants and
parasites. Nature, 360, 121–125.
Towe, KM (1990). Aerobic respiration in the Archaen? Nature, 348, 54–56.
Travis, JJM and Dytham, C (1999). Habitat persistence, habitat availability, and the
evolution of dispersal. Proceedings of the Royal Society of London, Series B, 266,
723–728.
—— and —— (2002). Dispersal evolution during invasions. Evolutionary Ecology
Research, 4, 1119–1129.
Trivers, RL and Willard, DE (1973). Natural selection of parental ability to vary the
sex ratio of offspring. Science, 179, 90–92.
Turkington, R (1989). The growth, distribution and neighbour relationships of
Trifolium repens in a permanent pasture. V. The coevolution of competitors.
Journal of Ecology, 77, 717–733.
Turner, GF (1999). Explosive speciation of African cichlid fishes. In AE Magurran
and RM May, eds. Evolution of biological diversity, pp. 113–129. Oxford University
Press, Oxford.
Vamosi, JC, Otto, SP, and Barrett, SCH (2003). Phylogenetic analysis of the ecologi-
cal correlates of dioecy in angiosperms. Journal of Evolutionary Biology, 16,
1006–1018.
van Doorn, GS and Weissing, FJ (2001). Ecological versus sexual selection models of
sympatric speciation: a synthesis. Selection, 2, 17–40.
202 REFERENCES

van Doorn, GS, Noest, AJ, and Hogeweg, P (1998). Sympatric speciation and
extinction driven by environment dependent sexual selection. Proceedings of the
Royal Society of London, Series B, 265, 1915–1919.
Van Valen, L (1973). A new evolutionary law. Evolutionary Theory, 1, 1–30.
Venable, DL and Brown, JS (1988). The selective interactions of dispersal, dormancy
and seed size as adaptations for reducing risk in variable environments. American
Naturalist, 131, 360–384.
Verhulst, P-F (1845). Recherches mathematiques sur la loi d’accroisse- ment de la
population. Nouveaux Memoires de l’Academie Royale des Sciences, des Lettres et des
Beaux-Arts de Belgique, 18, 1–32.
Vermeij, GJ (1995). Economics, volcanoes, and phanerozoic revolutions.
Paleobiology, 21, 125–152.
Via, S (2001). Sympatric speciation in animals: the ugly duckling grows up. Trends in
Ecology and Evolution, 16, 381–390.
Vrijenhoek, RC (1994). Unisexual fish. Annual Review of Ecology and Systematics, 25,
71–96.
Wächtershauser, G (1988). Before enzymes and templates: theory of surface
metabolism. Micobiological Reviews, 52, 452–484.
—— (1990). Evolution of the first metabolic cycles. Proceedings of the National
Academy of Sciences, USA, 87, 200–204.
Warren, MS, Hill, JK, Thomas, JA et al. (2001). Rapid responses of British butterflies
to opposing forces of climate and habitat change. Nature, 414, 65–69.
Waxman, D and Gavrilets, S (2005). Twenty questions on adaptive dynamics. Journal
of Evolutionary Biology, 18, 1139–1154.
Werren, JH (1980). Sex ratio adaptations to local mate comptetition in a parasitic
wasp. Science, 208, 1157–1159.
—— and Beukeboom, LW (1998). Sex determination, sex ratios and genetic conflict.
Annual Review of Ecology and Systematics, 29, 233–261.
West, GB, Brown, JH, and Enquist, BJ (1999). The fourth dimension of life; fractal
geometry and allometric scaling of organisms. Science, 284, 1677–1679.
—— —— and —— (2001). A general model for ontogenetic growth. Nature, 413,
628–631.
—— Woodruff, WH, and Brown, JH (2002). Allometric scaling of metabolic rate
from molecules and mitochondria to cells and mammals. Proceedings of the
National Academy of Sciences, USA, 99, 2473–2478.
West, SA and Herre, EA (2002). Using sex ratios: why bother? In ICW Hardy, ed. Sex
ratios: concepts and research methods, pp. 399–413. Cambridge University Press,
Cambridge.
Whitham, TG (1980). The theory of habitat selection examined and extended using
Pemphigus aphids. American Naturalist, 115, 449–466.
Whitlock, MC (1996). The red queen beats jack-of-all-trades: the limitations on the
evolution of phenotypic plasticity and niche breadth. American Naturalist, 148
(Suppl), S65–S77.
Wijesinghe, DK and Hutchings, MJ (1996). Consequences of patchy distribution of
light for the growth of the clonal herb Glechoma hederacea. Oikos, 77, 137–145.
 REFERENCES 203

Wilkinson, DM and Sherratt, TN (2001). Horizontally acquired mutualisms, an

unsolved problem in ecology? Oikos, 92, 377–384.
Willig, MR, Kaufman, DM, and Stevens, RD (2004). Latitudinal gradients of
biodiversity: pattern, process, scale, and synthesis. Annual Review of Ecology and
Systematics, 34, 272–309.
Willis, KJ and McElwain, JC (2002). The evolution of plants. Oxford University Press,
Oxford.
Wynne-Edwards, VC (1962). Animal dispersion in relation to social behaviour.
Hafner, New York.
Xiong, J, Fischer, WM, Inoue, K, Nakahara, M, and Bauer, CE. (2000). Molecular
evidence for the early evolution of photosynthesis. Science, 289, 1724–1730.
Xu, X, Zhou, ZH, Wang, XL, Kuang, XW, Zhang, FC, and Du, XK (2003). Four-
winged dinosaurs from China. Nature, 421, 335–340.
Yamamura, N, Higashi, M, Behera, N, and Wakano, JY (2004). Evolution of
mutualism through spatial effects. Journal of Theoretical Biology, 226, 421–428.
Young, TP (1990). Evolution of semelparity in Mount Kenya Lobelias. Evolutionary
Ecology, 4, 157–172.
Yule, GU (1924). A mathematical theory of evolution based on the conclusions of
Dr, JC Willis FRS. Philosophical Transactions of the Royal Society of London, Series
A, 213, 21–87.
 Glossary

Abiotic—not living. Cell cycle—the processes between the birth of a

Alleles—variant forms of a gene at a particular cell by division, and its growth and
locus. subsequent division into daughter cells.
Allozymes—identifiable forms of an enzyme Cell membrane—the layer surrounding and
coded for by alleles at a single locus. containing the fluid part of a cell.
Ammonoids—an extinct group of squid-like Cell wall—a non-living structure outside the
marine molluscs with coiled shells. cell membrane of plants, fungi, and bacteria
Anthers—the male parts of a flower, releasing that provides support and protection for
pollen. cells.
Apterygotes—primitively wing-less insects, Centriole—the organelle that forms the
including silverfish and bristletails. spindle fibres that separate the
Archaebacterium—a member of a diverse chromosomes during cell division.
group of bacteria differing biochemically Cephalopods—a class of marine molluscs,
from other ‘Eubacteria’, and today living in including squid, octopi, cuttlefish, nautili,
extreme environments, such as anaerobic as well as the extinct ammonoids and
mud, animal guts, salt lakes, and hot belemnoids.
springs. Chemo-autotrophs—Autotrophic bacteria
Archosaur—a member of the Archosauria, a that can synthesize organic compounds from
group of terrestrial vertebrates including inorganic raw materials in the absence of
crocodiles, birds, pterosaurs, and dinosaurs. sunlight. The energy is derived from the
Autosomes—chromosomes that are not sex oxidation of inorganic materials, such as
chromosomes. hydrogen sulphide, ammonia, and iron-bearing
Autotrophic—capable of utilizing inorganic compounds.
carbon as the main source of carbon and Chlorophyll—the green pigment in plants
of obtaining energy for life processes from that absorbs sunlight and uses its energy to
the oxidation of inorganic elements synthesize carbohydrates from CO2 and
(chemotrophic) or from radiant energy water, the process of photosynthesis.
(phototrophic). Chloroplasts—eukaryotic organelles,
Benthic—living on the bottom of a water body. originally derived from cyanobacteria, and
Biosphere—the components of planet Earth, that carry out photosynthesis.
including water, soil, and atmosphere, in Chromatids—The daughter strands of a
which organisms may be found. duplicated chromosome joined together at
Biotic—living. a structure called the centromere.
Bivalves—a class of mollusc, including clams, Chromosomes—a structure consisting of a very
oysters, cockles and their kin, surrounded by long piece of DNA carrying many genes.
a pair of hinged shells. Clade—a group of organisms that share a single
Bottlenecks—temporary reductions in common ancestor, representing a single
population size that reduce the genetic branch of the evolutionary tree.
diversity of a population. Cladistic revolution—a series of rapid scientific
Carrying capacity—the density above which a advances for estimating the evolutionary tree
population cannot be expected to increase. of life.
 GLOSSARY 205

Co-evolutionary—involving co-evolution, the Foraminifera—single-celled aquatic protists

reciprocal evolution of interacting species. with shells that are a prominent component
Cosexual—an individual that expresses both of marine ecosystems and of the fossil record.
gender functions, male and female, within its Functional groups—groups of species that
lifetime. perform a similar ecological role.
Cursorial—running on the ground. Gametes—the haploid cells, sperm, and egg
Crinoids—a group of marine animals, related that fuse to form a diploid cell in a sexual
to sea urchins and starfish, containing sea life cycle.
lilies and feather stars. Gastropods—snails and their relatives.
Cyanobacteria—a group of photosynthetic Genetic drift—an evolutionary process, more
bacteria, some of which became the important in small populations, in which the
chloroplasts of plants. genetic composition of a population changes
Cytoplasmic—in or of the cytoplasm, the fluid through random events.
interior of a cell. Genetic load—the extent to which the average
Cytoskeleton—a three-dimensional structure individual in a population is less fit than
within the fluid interior of eukaryotic cells the fittest individual. This equals the relative
that aids movement and stability. chance that an average individual will die
Detritivores—organisms that feed on before reproducing because of the deleterious
decomposing organic material. alleles that it possesses.
Dioecious—when only one sex is expressed in Genome—the total genetic content of a
a single individual. cell.
Diploid—cells containing chromosomes in Geochemical cycles—the movement and
homologous pairs, one derived from each transformation of materials around the
parent. Earth.
Directional selection—selection favouring Graptolites—extinct, stick-like, colonial marine
individuals with extreme traits (e.g. larger organisms, found in the early Palaeozoic.
than average) within a population, such that Possibly related to group of worms known as
the average trait shifts over time. hemichordates.
DNA—the acronym for deoxyribonucleic acid, Haploid—cells containing just one member of
the molecule that contains the genetic each homologous chromosome pair.
information within cells. Heterotrophic—organisms obtaining their
Echinoids—starfish, sea urchins, and their kin. carbon and energy from organic sources.
Effective population size—the size of an ideal Homonoids—the group containing gibbons,
population which acts the same, with respect great apes, and humans.
to the degree of genetic drift or inbreeding, Homologous—alike due to shared ancestry.
as the real population in question. Effective Homologous chromosomes in a cell derived
population sizes are usually smaller than from different parents but contain equivalent
real populations because of variation in sets of genes.
reproductive success among individuals or Horizontally transmitted—transmitted
deviation from a 50 : 50 sex ratio. between unrelated individuals within a single
Enzymes—proteins that speed up biochemical generation as well as across generations, as
reactions within cells, acting as metabolic opposed to vertical transmission (parent to
catalysts. offspring).
Eukaryotes—organisms distinct from bacteria, Inbreeding coefficient—The probability that a
which contain a nucleus containing non- zygote obtains copies of the same ancestral
circular chromosomes, a cytoskeleton, and allele from both its parents because the
other organelles, such as mitochondria and parents are related to each other.
chloroplasts. Leguminous plants—plants from the family
Eutherians—placental mammals, distinct from Leguminosae, containing peas, beans, and
monotremes and marsupials which both lack their relatives, which form a symbiotic
a placenta. association with rhizobia bacteria in their
Extant—still living, the opposite of extinct. root nodules.
206 GLOSSARY

Limnetic—the main water column of a lake. Ostracods—a class of aquatic crustacean, also
Linkage—the tendency for certain genes to be called seed shrimps.
inherited together because they are located Parasitoids—insects, mainly wasps and flies,
on the same chromosome. which develop as juveniles by feeding on the
Locus—the position on a chromosome where a body of another host organism, usually
particular gene is located. another insect.
Macroscopic—visible with the naked eye, as Phanaerozoic—the time period beginning
opposed to microscopic. about 535 Ma, consisting of the Palaeozoic,
Meiotic—to do with meiosis, the form of cell Mesozoic, and Cenozoic, in which fossils of
division in which a diploid cell gives rise to animals are abundant.
four haploid cells. Phenological—to do with phenology, the
Mendelian inheritance—the form of timing of biological events.
inheritance, discovered by Gregor Mendel, Phenotype—the physical parts of an organism,
which occurs in most sexual organisms, as opposed to the genotype, which is the
whereby each individual inherits two copies heritable blueprint for creating the
of a gene, one from each parent. phenotype, encoded in a cell’s DNA.
Metabolic catalyst—a molecule that facilitates Photosynthesize—carry out photosynthesis,
biochemical reactions within cells. the process by which some organisms
Metabolic cycle—the flow of material through synthesize organic molecules from inorganic
a series of biochemical reactions within an carbon using sunlight.
organism. Phytoplankton—the photosynthetic
Metapopulation—a population consisting of component of plankton.
many ephemeral subpopulations, which are Pistils—the central part of a flower, containing
linked by dispersal between them. The the female reproductive parts.
metapopulation persists by replacing extinct Pleiotropy—when one gene influences two or
subpopulations through recolonization. more phenotypic traits.
Mitochondria—eukaryotic organelles, Polymorphism—when individuals in a
originally derived from a group of bacteria, single species can be of two or more distinct
which perform aerobic respiration within phenotypes.
cells. Population genetics—the field that deals with
Mitotic—involving mitosis, a type of cell how genes influence the characteristics of a
division in which a (normally diploid) cell population, particularly the processes that
gives rise to two (normally diploid) daughter determine the frequency and distribution
cells. of alleles.
Molecular revolution—a series of rapid Power function—when a number is multiplied
scientific advances in understanding the by itself a specified number of times. The
biochemical basis of biology, and particularly number that specifies how many times is
in reading the genetic code. known as the exponent.
Monoplacophora—a primitive class of mollusc, Prokaryotes—bacteria, organisms without a
today living in deep-ocean trenches. nucleus, cytoskeleton, mitochondria,
Mycorrhizal fungi—fungi that form a symbiosis or chloroplasts, and with a single circular
by living within the roots of plants. chromosome. As opposed to eukaryotes.
Neutral evolution—evolution that has no Protein—a type of organic molecule, many of
effect on individual fitness, mainly consisting which function as enzymes, created from
of changes to non-coding parts of the an RNA template in a process known as
genome or synonymous changes to coding translation.
elements. Protist—eukaryotes, often single-celled, which
Nuclear—in or of the nucleus, the membrane- are not plants, animals, or fungi.
surrounded structure in eukaryotic cells that Pterygotes—the winged insects.
contains the chromosomes. Quantitative genetics—the field concerned
Organelles—the distinctive structural elements with measurable, continuous, or ‘quantitative’
within a cell. traits and their evolution, which attempts, in
 GLOSSARY 207

particular, to predict the response to selection Sympatric speciation—speciation in which

of a population. the two incipient daughter species have
Raptorial—predatory. When applied to birds, identical or substantially overlapping
implies a member of the Falconiformes geographic ranges.
(vultures, falcons, hawks, eagles). T-lymphocytes—vertebrate white blood cells
Rhizobia—a group of nitrogen-fixing bacteria involved in the immune response and
that form a symbiotic association in the roots which target particular foreign or cancer
of leguminous plants. cells.
RNA—the acronym for ribonucleic acid, a Trilobites—A group of extinct marine
single-stranded molecule formed from a arthropods, characterized by a three-lobed
DNA template via a process called body, which were a dominant component
transcription, and involved in protein of the marine fauna in the Early
synthesis. Palaeozoic.
Stabilizing selection—selection against the Unisexual—a species which has lost one of its
extremes of a population, which results in the sexes and become parthenogenetic through
population average staying the same. the surviving sex.
Substitution—changes to the DNA sequence of Zooplankton—the animal component of
an organism (mutations) in which one plankton.
nucleotide base (letter) is replaced by another Zygote—a fertilized egg.
different base (letter).
 Index

adaptive dynamics, see dynamics macroevolution 107, 164, 166–9

Agave mckelveyama 84 peripatric speciation 138
AIDS 91–2, 111 sex determination 58–9
Allee effect 150, 154 Blackburn 45, 166, 170, 172, 179
allele body size
beneficial 100 coevolution 121, 125, 127
deleterious 100, 154 life histories and scaling 37, 41–5, 48
lethal 143 macroecology 171–7, 179
recessive 153–4 macroevolution 169
Allen 48, 178–9 specialization in sticklebacks 104
allometry/allometric 41–4, 176 speciation 143
allozyme 141 Brown 47, 50, 73, 104, 170, 175, 179
altruism 79–80 Buchnera 114–15
anagenesis/anagenetic 9–11, 86, 106, 132, 158–9, 161, Bull 60, 63, 111
173–7, 181, 183 Bush 140–1
Anderson 109 butterfly 55–6, 72, 103–4, 150, 152, 166
angiosperms 29, 35, 80–1, 169
anisogamy 15–7 Cambrian 26, 28, 163–5
antagonism 108, 113, 122, 127, 181, 183–4 Carboniferous 26, 28, 32, 165
aphid 103, 106, 108, 114–15 Catalina Mahogany 153
apterygotes 30, 33 cell cycle 15–16
archaebacterium 27 Cercocarpus 153
Armadillidium vulgare 61–3 Charnov 39, 41–7, 50–4, 56–7, 60, 63, 76, 87, 98, 176
Arnold 134, 137, 144, 168 Chihuahua spruce 154–5
Asclepias exaltata 84 Chlamydomonas 17–19
assortative mating 4, 139–40 chlorophyll 27
atlantic silverside 53–4, 94 chloroplast 18, 20, 114, 121–2
autotrophic, see theory chromosome 13–16, 61–2, 81–2, 118, 134, 136
sex 51, 55, 60
Banksia goodii 151 X 19, 51, 55, 58
Barbulanympha 16–17 Y 19, 55, 58
Barraclough 141–2, 168–9 cichlid 1, 3–6, 9–12, 133, 139, 143, 153, 184
bat 29, 31–6 clade 158–63, 166, 169, 173, 175, 180
behavioural ecology, see ecology cladogenesis/cladogenetic 9–11, 34, 107, 132, 159–160,
Benkman 129 167–8, 174–5, 177, 181–3
biodiversity hot spots 147 cod 49, 93–4
birds co-evolution/co-evolutionary 128, 130–1, 183–5
behavioural-based models 88–90 alternation 121–3, 127
constraints to parthenogenesis 19 arms race 18, 126
dispersal 70–2, 148 diversifying 121, 125, 127, 130
evolution of flight 29–36, 148 escape and radiation 122–7
island 148–9 gene-for-gene 127
life-histories 44–5 geographic mosaic theory of, see theory
macroecology 171 hot spots 129
210 INDEX

co-evolution/co-evolutionary (cont.) dioecy/dioecious 51, 53, 60, 63

mutual dependence 115, 121, 123, 125, 127 disparity 158–60, 162, 165–6, 169
successional cycles 126–7 dispersal
turnover 121, 123–5, 127 evolution of, see evolution
Cohen 72 evolution of flight 34–6
commensalism 108 extinction 147–8, 151
comparative study 40–1, 50 macroecology 178
competition 8, 66–8, 71–3, 75, 79, 87, 100, 106, 108, 124, sex ratios 57
126, 182 distribution
asymmetric 119, 121, 123, 127 body size 175–6
between clone and sexual parents 22 frequency 172–3, 184
coevolution and 124, 126–7 geographic 3, 53
interspecific/between species 100–1, 104, 136, 143 ideal free, see model
intraspecific 100, 105 of competitors 88
kin 66, 73 of deleterious mutations 16
local mate 58 diversification 28–9, 159–67, 172–3
mate/male-male 56, 58, 68, 84–5, 140 DNA 13–15, 18, 23, 92, 137, 166
organelle 19 Dodder 77–8
pollen 85 dormancy 65, 72–4, 83, 96, 181, 184
resource 68, 140 Drosophila 55, 103, 137, 142–3, 154, 156
sibling 73, 83 dynamics
viral 92, 111 adaptive 86, 95–6, 119, 139–40, 182, 185
conflict chaotic 162
between organelles 18–19 displacement 162
between parents 20, 81 macroevolutionary 162
in the major transitions in evolution 23 of clade characteristics 169
kin 82 of clade growth 161
over sex allocation 56, 60–3 population 68, 86–7, 95–6, 145, 181, 183–5
parent-offspring 82
sexual 81 ecology 9–11, 25–6, 29, 36–7, 86, 143, 170, 180–5
sibling 83 behavioural 10, 75–6, 85, 101
Connell 67 community 182
Conover 53–4, 94 ecosystem 182
constraints 23, 38–9, 45, 73, 97, 100–1, 106, 120, 162, major transitions in 26, 29, 36, 183–5
165, 173 metabolic 47
cosexual 51–3, 55, 57, 60–2 of individuals 182
Crepis sancta 69 population 86, 96, 144, 182
Cretaceous 26, 33, 162–3, 167 ectotherm 48, 178
crinoid 163, 165 effective population size 154–6
crossbill 129–30 Ellner 72, 83
currency 38–9, 159 endomitosis 16–17
cyanobacteria 27 endosperm 81–2
Cytoplasmic Male Sterility/CMS 55, 60–1, 119, 130, 184 Enquist 45, 48
epigenesis 24
Dalbergia sissoo 83 ESS, see Model
Darwin 3, 84, 143 eukaryotes/eukaryotic 13–15, 17, 19, 23, 27, 121–2
Darwin’s finches 104, 107, 134–8, 143 Euphydryas editha 103–4, 150
decision making 87, 97, 101 Evolution
Dial 30, 172 and numbers 86
Diascia 125–6, 128 darwinian 14
Dieckmann 95–6, 114 flower 85
diet in ecological time 91
in cichlids 1, 3 life history 37, 39, 42, 45, 47, 50, 176, 183–84
optimal 101–2 major transitions in 14, 23
 INDEX 211

neutral 143 Felis concolor coryi 154

niche 100–1, 103–4 fig wasps 56, 58, 112, 119
of antagonism 113, 183 Fisher 53–4, 84, 107, 133, 169
of cheating 119 fleur-du-mâle 84
of cooperation 109 flight 26, 29–36, 69, 72, 184
of cosexuality 53, 62 Florida panther 154
of dispersal 65–7, 71, 185 Foote 162, 165–6, 169
of female biased sex ratios 57 fossil record 25, 32, 106, 120, 159, 161–5, 177
of flight 29, 36 foundress 112
of flowering plants 29 Frank 111–14, 118–19
of hermaphrodites 52 Fretwell 87
of mutualism 130 fruit
of population dynamics 183, 185 bats 31
of sex 15–16, 23–4, 130, 184 dispersers 108
of sex determining/determination 58, 61, 63 flies, see Drosophila
of social behaviour 64 Hawthorn 140–1
of specialization/specialism 98–9, 183 initiation 84
of species interactions 119, 130–1 Futuyma 99–100, 107
of syngamy 23
of terrestrial communities 28 Galapagos finches, see Darwin’s finches
of the biosphere 26 gamete 15, 18–19, 21, 23–4, 51, 114, 136
of the latitudinal gradient in species richness 176 Gasterosteus 105, see also stickleback
of virulence 110–11, 130 Gaston 146, 166, 170, 172, 179
rapid 91, 168, 182 gastropod 106, 163, 165, 167, 179
reciprocal 120 Gavrilets 95–6, 114, 162, 165
red queen 128 generalist 52, 97–100, 106–7, 182
short-term 92–3, 128 genetic drift 133, 155
social 75, 79 genetic load 16–17, 22, 100
through/by natural selection 10–11, 181 genomic imprinting 19, 81–2
within-host 92 geographic range 141, 146–8, 151, 169, 171, 177
evolutionary branching 95–6, 119, 139–40, 184 Geospiza, see Darwin’s finches
Evolutionary Stable Optimal Harvesting Glanville fritillary 152, 154
Strategy/ESOHS 94–5 Glechoma hederacea 79
Evolutionary Stable Strategy/ESS, see Model Gleeson 78
experiment 6, 41, 54, 74, 78, 103–4, 115, 142, 170, 185 Goss-Custard 88, 96
Extinction 9–11, 145–57, 159–169, 182–84 Grant 71, 134, 136–8, 148–9
body size and 172–6 Ground Ivy 78–9
co-evolution and 121–5, 127 growth
correlates of 169 clade 161, 163, 173
dispersal and 64–5, 67–8 determinate 37, 48–9
flight and 34–6 exponential 160, 162
latitude and 177–9 indeterminate 37, 49–50
mass 162–4 logistic 164
mechanisms of 145–57 mammalian 43
of clones 15, 21–3 ontogenetic 48
of specialists 106–7 population 39, 72–3
of symbiosis 116, 119 rate 39, 44–5, 48–9, 72–3, 85, 97, 103, 161
population 63, 67, 149–50, 154 sigmoidal 48–9
rate 159–69 gymnosperms 19–20
recent 25
risk 34–5, 107, 169, 174 Haematopus 118
Hamilton 18–19, 21–2, 55–6, 66–8, 74, 79, 130,
Farrell 29, 168 see also rule
fast-slow continuum 40–3, 45 haplochromines 1, 3–5, 10
212 INDEX

Harvey 40, 42, 44, 50, 161, 164, 167, 169 oceanic 70–1, 147
Hawthorne 103 radiations 165
Heath Hen 150 Santa Catalina 153
Hedrick 99, 154–5, 157 Schriermonnikoog 89
Heino 94 species 148
Helianthus 134 Upolu 55
Herre 51, 56, 112, 119 Vancouver 71, 105
heterogamety 51, 58–9, 61 iteroparity/iteroparous 37, 39–40
heterotrophic 26
HIV 91–4, 96, 109, 111 Jablonski 165, 167
horizontal transmission 109, 111–12, 178 Jaenike 102
Hurst 18–19, 55 Janzen 67
Hutchings 78–9 Johnson 67, 72, 74, 128
Hutchinson 97
hybrid 22, 134–7, 143, 146, 153 Kawecki 103
hybridization 134, 136, 143, 145–6, 153, 182 Kelly 77–8
Hypolimnas bolina 55–6 kin selection, see selection
hypothesis Kondrashov 16, 22
arboreal 29 Kozlowski 44, 47, 173, 175–6, 179
bat flight 31 Kunin 146, 148
cursorial 30
Janzen-Connell 67 Lake Malawi 1–2, 4–6
null 143 Lake Tanganyika 1–2
pouncing pro-avis 30 Lake Victoria 1–5, 8, 133, 153, 184
red queen 18, 20, 21, see also theory Lande 8, 150–1, 154, 156–7
surface skimming 30–1 latitudinal gradient 176–9
temporal variability 69 Law 93–6, 114
Trivers-Willard 59 Levin 134, 138, 144, 151, 153, 157
Levins 98
inbreeding life history 35, 37, 39, 41–2, 44–50, 54, 73–4, 176, 181,
coefficient 155 183–5
depression 153–6 linkage 114, 139
inheritance Lobelia 39–40
asymmetry 55, 60–1 local mate competition, see competition
cytoplasmic 55 Lodgepole pine 72, 129
maternal 55, 121, 125 Lord Howe woodhen 148–9
mendelian 59, 80 Lotus corniculatus 83
symmetry 54–5, 59–61 Loxia curvirostra 129
uniparental 19, 24, 118
insect MacArthur 101
dispersal 69, 71–2 McKinnley 172
flight 29–36 macroecology 170, 179, 183, 185
inbreeding 154 macroevolution 158–9, 162, 169, 183–5
interspecific interactions 108, 117–19 Malheur wire lettuce 146–7
life histories 44 mammals
macroevolution 165, 168 constraints to parthenogenesis 19–20
sex allocation 55–6, 59–60 dispersal 71–2
specialization 102–4, 107 life histories 40–5, 48, 50, 55
speciation 141 macroevolution 165, 169
invariant 42, 44–6, 48, 50, 54, 178–9 sex determination 55, 58–9
island sex ratios 59
archipelagos 147 marginal value theorem 76–9, 87, 101
Daphne Major 134 maturity 37, 39, 42–5, 48, 93–4, 181
Lord Howe 148 Maurer 172–3, 175–6
Martha’s Vineyard 150 May 66, 68, 74, 109
 INDEX 213

Mayhew 44–5, 102–3, 107, 166–7, 178 evolution of flight 34–5

Maynard Smith 13–14, 16–18, 22–4, 36, 39, 99, 114, evolution of symbiosis 115
117, 139 evolution of virulence 109
Mayr 133, 138–9 extinction 150
Melitaea cinxia 152 life histories 39–40, 42–5
Menidia menidia 54, see also atlantic silverside macroecology 176
metabolic cycle 17, 26 Mount Kenya 39–40
metabolic rate 43, 46–9 Müller’s ratchet 22
metapopulation 68, 106, 151–2 multiple infection 111–3
microevolution 158 mutations 6, 17, 22, 32, 92, 143, 156
Mimulus guttatus 83, 143 deleterious 16–17, 22, 103, 156
mitochondria 18, 27, 48, 55, 60, 114, 121–2 lethal/semilethal 154
Mock 80, 85 recessive 154
model synergistic 17, 22
Anderson and May’s 109 mutualism 113–15, 117–19, 121, 126–7, 130, 181, 185
adaptive dynamics 86, 95–6, 119, 139–40, 185 mycorrhizal fungi 22, 108, 114, 117, 126
behavioural-based 90 myxoma 110
Brown et al.’s 175
Bush’s 140–1 Ne 155–6
Charnov’s 43–4, 50 nematode 45, 112
Charnov and Schaffer’s 39 niche
Charnov et al.’s 51–2, 57, 98 ecological 4, 10, 97, 156, 184
constant rate 160, 162, 164 evolution, see evolution
Dial and Marzluff ’s 172 fundamental 97–9, 101–5
empirical 121 overlap 6
ESS 53, 66–7, 88–9, 95 realized 97, 101–2, 105, 107
exponential growth 160, 162 trophic 26
foraging 87, 101 Northen Spotted Owl 151–2
Frank’s 111–14, 118–19
Hamilton and May’s 66, 68, 74 optimal diet model, see model
Heino’s 94 optimization/optimality, see model
Hochberg et al.’s 130 Ordovican 28
Levin’s 98 Osmoderma eremita 71
logistic 162–5 ostracod 163
ideal free distribution 87–8, 96, 100 ovule 81
Kondrashov’s 16, 22 Owens 45, 107, 169
Kozlowski and Gawelczyk’s 173, 176, 179 oystercatcher 88–90
Kozlowski and Weiner’s 44, 176
Law and Dieckmann’s 114 Palaeozoic 163, 165
Maurer et al.’s 172, 173, 175–6 parasitoid 25, 45, 56, 176
McKinnley’s 172 Parker 39, 80, 85
optimal diet 101 passerine 124, 166–8
optimization/optimality 38–9, 76 perceptual bias 4
rate maximization 101 Permian 26, 162–3
recombinatorial 134, 136 phylogeny 41, 107, 120, 141–2, 161, 164, 166–7, 169, 174,
simulation 6, 173 177, 185
van Doorn et al.’s 6–7 Pianka 101
Venable and Brown’s 73 Picea chihuahuana, see Chihuahua spruce
West et al.’s 46–7, 49 Plantago lanceolata 60, 114, 130
Whitlock’s 100 pleiotropy 98, 103, 139, 141
monkey 91, 167 Poeciliopsis 20, 22–4
mortality pollen 20, 53, 57, 80–1, 84–5
dispersal 65–7 pollination
dormancy 72–3 co-evolution 108, 117–19, 121, 125, 127
evolution and population dynamics 88–90, 94 extinction 151
214 INDEX

pollination (cont.) group 64–6

macroevolution 168 habitat 88
major transitions in ecology 29, 35–6 hard 99
plant behaviour 74, 77, 84–5 kin 79, 84
sex allocation 52–3, 56, 58 natural 3–4, 7, 10–11, 181–2
polygyny 5, 68 mate 4
polymorphism 8, 95–6, 99, 139–40 mosaic 128
population genetics 79–80, 142, 185 runaway 84, 139
power function 41–2 sex ratio 8–9, 59–60
prisoner’s dilemma 116 sexual 4–12, 84–5, 139–40, 143–4, 168, 184
prokaryotes/prokaryotic 13, 18, 27 soft 99
protein 23–4, 27, 92 stabilizing 22, 92, 126
Proterozoic 26–7 within hosts 111
pterosaurs 29, 31–4, 36 semelparity/semelparous 37, 39–40
pterygotes 30, 32–3 Sepkoski 75, 163, 166, 169
Purvis 44, 50, 107, 145, 167, 169, 175–6 sex
allocation 51, 53, 56, 58–63, 183–5
r 79–80, 82, 111 determination 11, 51, 54, 58, 60–3, 73
R0 109 evolution of, see evolution
rabbit 110 heterogametic 19, 55, 58
radiation 1, 3, 5–6, 28, 34–5, 106–7, 121–7, 142, homogametic 19, 55
159–62, 165–8, 174 maintenance of 15, 18–20, 23, 34, 100, 113
rarity 8, 19–20, 146–8 origins of 17–19
recombination 15–18, 21–2, 41, 128, 139 ratios 51, 53–60, 63, 66, 69, 75
red queen, see theory ratio selection, see selection
Rediviva 125–6, 128 reversal 8
redwood 149 two-fold cost of 21
Rees 74 sexual selection, see selection
reversal 8, 15, 19–20, 23, 34 silversword 107, 143
Rhagoletis 140–1 Singer 103
Rhizobium 108, 117–18 singularity 95
Ricklefs 36, 166 SIV 91–2
Rieseberg 134–5 social evolution, see evolution
Rigaud 61, 63 Sorghum vulgare 78
RNA 13, 24, 27, 92 Southwood 25, 36, 67, 69, 104
Roff 34, 36, 50, 69, 71, 148 Spartina 146
Roughgarden 121, 125, 127 specialist 13, 97–100, 104, 106–7, 182, 184
Roy 166, 169, 175 speciation 4, 6–12, 15, 34–6, 132–44, 159–69, 172–9,
rule 182–4
Bergmann’s 171 allopatric 138, 141–2
energetic equivalence 178–9 allopolyploidy and 134, 136
Hamilton’s 79–80 adaptive dynamics and 95–6
Rapoport’s 171 CMS and 60
runaway process 8, see also selection coevolution and 121, 123, 125, 127, 129, 131
disjunct 138
scaling 44, 46–8, 50, 178 geographic distribution and 3, 138–42
Schluter 104–7, 142–4, 165 macroecology and 172–9
Seehausen 3–5, 8–9, 12 macroevolution and 159–69
selection parapatric 138
between hosts 111 peripatric 138, 141
directional 18, 22 recombinatorial model of 134, 136
disruptive 105, 137, 139–41, 143 sex ratio selection and 8–9
divergent 143 sexual selection and 4–12, 133, 139–44
frequency-dependent 54 shifting balance theory of 133, 142
 INDEX 215

specialization and 99, 103, 105–7 Thompson 103, 120–1, 128, 130–1
sympatric 8, 12, 96, 132–3, 139–42 Thymus vulgaris 60
vicariance 3, 138 trade-off 38, 52, 72–4, 98–100, 103, 111, 140–1,
without geographic isolation 4, 6 176, 184
Stephanomeria 146 tragedy of the commons 111
stickleback 104–5, 143 Tragopogon 134
stochasticity 150, 156 Triassic 26, 162–3
Strix cavrinea occidentalis 151 Tricholimnas sylvestris 148
Sutherland 79, 84, 88, 90, 96 trilobite 163–5
symbiosis 114–16, 118, 127, 132 Trivers 20, 57, 59
symmetry 54–6 Trivers-Willard 59
syngamy 15–7, 23
Szathmáry 13–14, 16–17, 23–4, 36, 117 unisexual 19–21

theory van Alphen 3–4, 8

autotrophic 26 van Doorn 6–7, 139–40
behavioural ecology 101, see also ecology variation
co-evolutionary 128, 130 genetic 10, 53, 92, 97, 142–3, 152, 155–6
dispersal 64–8 intraspecific 158
extinction 157, 169, 184 interspecific/across species 40–1, 158
foraging 76–9, 87–8 spatial 68, 99
geographic mosaic 128–31 temporal/over time 67–8, 70
in evolutionary ecology 6, 185 Venable 72–3
Kondrashov’s 16, 22 vertical transmission 111–12, 114–15, 117–18
life history 50, 74 Via 103
local mate competition 56–8 vicariance, see speciation
macroevolutionary 159 virulence 91–2, 109–13, 119, 127, 130, 184
metabolic 17
of everything 37, 47 Werren 56, 63
optimization 50, 185, see also model Wolbachia 55–6, 61, 63, 125
red queen 18, see also hypothesis Wynne-Edwards 64
sex allocation 51–8
sexual selection 8, 84 Xu 29–30
shifting balance 133, 142
speciation 168 Young 39–40
theropod 29–30 Yucca 117–19