(Ebooks PDF) Download Theoretical Ecology Principles and Applications 3rd Edition Robert May Full Chapters

Full download ebook at ebookgate.
com
Theoretical Ecology Principles and

Applications 3rd Edition Robert May
https://ebookgate.com/product/theoretical-ecology-
principles-and-applications-3rd-edition-robert-
may/
Download more ebook from https://ebookgate.com

More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Ecology Principles and Applications 2nd Edition Edition

J. L. Chapman
https://ebookgate.com/product/ecology-principles-and-
applications-2nd-edition-edition-j-l-chapman/
Macroeconomics Principles and Applications 6th Edition

Robert E. Hall
https://ebookgate.com/product/macroeconomics-principles-and-
applications-6th-edition-robert-e-hall/
Terrestrial Ecosystem Ecology Principles and

Applications 1st Edition Göran I. Ågren
https://ebookgate.com/product/terrestrial-ecosystem-ecology-
principles-and-applications-1st-edition-goran-i-agren/
Behavioural Ecology of Insect Parasitoids From

theoretical approaches to field applications 1st
Edition Eric Wajnberg
https://ebookgate.com/product/behavioural-ecology-of-insect-
parasitoids-from-theoretical-approaches-to-field-
applications-1st-edition-eric-wajnberg/
Principles of Contract Law 3rd Edition Robert Hillman
https://ebookgate.com/product/principles-of-contract-law-3rd-
edition-robert-hillman/
Principles of Tissue Engineering 3rd Edition Robert

Lanza
https://ebookgate.com/product/principles-of-tissue-
engineering-3rd-edition-robert-lanza/
Psychological Testing Principles Applications and

Issues 7th Edition Robert M. Kaplan
https://ebookgate.com/product/psychological-testing-principles-
applications-and-issues-7th-edition-robert-m-kaplan/
Psychological Testing Principles Applications and

Issues 8th Edition Robert M. Kaplan
https://ebookgate.com/product/psychological-testing-principles-
applications-and-issues-8th-edition-robert-m-kaplan/
Psychological Testing History Principles and

Applications 7th Edition Robert J. Gregory
https://ebookgate.com/product/psychological-testing-history-
principles-and-applications-7th-edition-robert-j-gregory/
Theoretical Ecology
This page intentionally left blank
Theoretical Ecology
Principles and Applications
EDITED BY
Robert M. May,
Department of Zoology, University of Oxford,
Oxford, UK
AND
Angela R. McLean,
Department of Zoology, University of Oxford,
Oxford, UK
1
1
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 Dare 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
# R. M. May and A. R. McLean 2007
The moral rights of the authors have been asserted
Database right Oxford University Press (maker)
First published 2007
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
Antony Rowe Ltd., Chippenham, Wiltshire
ISBN 978–0–19–920998–9 (Hbk) 978–0–19–920999–6 (Pbk)
10 9 8 7 6 5 4 3 2 1
Contents
Acknowledgements vii
Contributors ix
1 Introduction 1
Angela R. McLean and Robert M. May
2 How populations cohere: five rules for cooperation 7

Martin A. Nowak and Karl Sigmund
3 Single-species dynamics 17
Tim Coulson and H. Charles J. Godfray
4 Metapopulations and their spatial dynamics 35

Sean Nee
5 Predator–prey interactions 46
Michael B. Bonsall and Michael P. Hassell
6 Plant population dynamics 62

Michael J. Crawley
7 Interspecific competition and multispecies coexistence 84

David Tilman
8 Diversity and stability in ecological communities 98

Anthony R. Ives
9 Communities: patterns 111

Robert M. May, Michael J. Crawley, and George Sugihara
10 Dynamics of infectious disease 132

Bryan Grenfell and Matthew Keeling
11 Fisheries 148
John R. Beddington and Geoffrey P. Kirkwood
12 A doubly Green Revolution: ecology and food production 158

Gordon Conway
13 Conservation biology: unsolved problems and their policy implications 172

Andy Dobson, Will R. Turner, and David S. Wilcove
v
vi CONTENTS
14 Climate change and conservation biology 190

Jeremy T. Kerr and Heather M. Kharouba
15 Unanswered questions and why they matter 205

Robert M. May
References 216
Index 249
Acknowledgements
The aims and scope of this book are set out in the bring the authors and others together for a 2-day
beginning of the first chapter (unimaginatively conference, in which the sweep of material in this
labelled Introduction). So these prefacing com- book was exposed to discussion and constructive
ments are confined to acknowledging some of the criticism. It helped shape the book.
help we and the other authors have received in Our thanks are also owed to enthusiastic and
putting this book together. helpful people at Oxford University Press, parti-
The two of us are deeply indebted to the other 21 cularly the commissioning editor, Ian Sherman, the
authors who have contributed to the book, both production editor, Christine Rode, and the copy-
for the work they did and for their exemplary editor, Nik Prowse. R.M.M.’s assistant, Chris
adherence to a rather fast production schedule. Bond, was her usual invaluable self, helping in
Individual authors have wished to thank both every facet of the enterprise with unflappable
funding agencies and helpful colleagues who gave competence.
assistance of various kinds. This would have been Sadly, one of the authors—Geoff Kirkwood—
an impressively long list, but we unkindly decided unexpectedly died the week after the gathering
against including it. in Oxford. Everyone remembers him with affec-
We must, however, recognize the generosity of tion, and his shadow lies on the book. He will be
Merton College and the Zoology Department at missed.
the University of Oxford, and particularly their
respective heads, Dame Jessica Rawson and A.R. McLean and R.M. May
Professor Paul Harvey. They made it possible to 22 September 2006
vii
Contributors
John R. Beddington, Division of Biology, Faculty of Heather Kharouba, Canadian Facility for Ecoinformatics
Natural Sciences, RSM Building, Imperial College Research (CFER), Department of Biology, University of
London, SW7 2BP, UK. E-mail: j.beddington @ Ottawa, Box 450, Station A, Ottawa, ON, K1N 6N5,
imperial.ac.uk Canada. E-mail: hkar075 @ uottawa.ca
Michael B. Bonsall, Department of Zoology, Tinbergen Geoffrey P. Kirkwood, Division of Biology, Faculty of
Building, University of Oxford, Oxford OX1 3PS, UK. Natural Sciences, RSM Building, Imperial College
E-mail: michael.bonsall @ zoo.ox.ac.uk London, SW7 2BP, UK
Gordon Conway, Centre for Environmental Policy, 4th Robert M. May, Department of Zoology, Tinbergen
Floor, RSM Building, Imperial College, South Building, University of Oxford, Oxford OX1 3PS, UK.
Kensington, London SW7 2AZ, UK. E-mail: E-mail: robert.may @ zoo.ox.ac.uk
g.conway @ imperial.ac.uk Angela R. McLean, Department of Zoology, Tinbergen
Tim Coulson, NERC Centre for Population Biology and Building, University of Oxford, Oxford OX1 3PS, UK.
Division of Biology, Imperial College London, Silwood E-mail: angela.mclean @ zoo.ox.ac.uk
Park Campus, Ascot, Berkshire SL5 7PY, UK. E-mail: Sean Nee, Institute of Evolutionary Biology, School of
t.coulson @ imperial.ac.uk Biological Sciences, University of Edinburgh, West
Michael J. Crawley, Department of Biological Sciences, Mains Road, Edinburgh EH9 3JT, UK. E-mail:
Imperial College London, Silwood Park, Ascot, Berk- sean.nee @ ed.ac.uk
shire SL5 7PY, UK. E-mail: m.crawley @ imperial.ac.uk Martin A. Nowak, The Program for Evolutionary
Andy Dobson, Ecology and Evolutionary Biology, Dynamics, Faculty of Arts and Science, One Brattle
Princeton University, Princeton, NJ 08544, USA. Square, Harvard University, Cambridge, MA 02138,
E-mail: dobber @ princeton.edu USA. E-mail: nowak @ fas.harvard.edu
H. Charles J. Godfray, Department of Zoology, Karl Sigmund, Faculty for Mathematics, University of
Tinbergen Building, University of Oxford, Oxford Vienna, Nordbergstrasse 15, A-1090 Vienna, Austria.
OX1 3PS, UK. E-mail: charles.godfray @ zoo.ox.ac.uk E-mail: karl.sigmund @ univie. ac.at
Bryan Grenfell, Biology Department, 208 Mueller George Sugihara, Scripps Institution of Oceanography,
Laboratory, Pennsylvania State University, University University of California, San Diego, 9500 Gilman
Park, PA 16802, USA. E-mail: grenfell @ psu.edu Drive, La Jolla, CA 92093 0202, USA. E-mail:
Michael P. Hassell, Department of Biological Sciences, gsugihara @ ucsd.edu
Imperial College London, Silwood Park, Ascot, Berk- David Tilman, Department of Ecology, Evolution and
shire SL5 7PY, UK. E-mail: m.hassell @ ic.ac.uk Behavior, University of Minnesota, St. Paul, MN 55108,
Anthony R. Ives, Department of Zoology, University of USA. E-mail: tilman @ umn.edu
Wisconsin, Madison, WI 53706, USA. Will R. Turner, Center for Applied Biodiversity
E-mail: arives @ wisc.edu Science, Conservation International, 1919 M St. NW
Matthew Keeling, Department of Biological Sciences Suite 600, Washington, DC 20036, USA. E-mail:
and Mathematics Institute, University of Warwick, w.turner @ conservation.org
Gibbet Hill Road, Coventry CV4 7AL, UK. E-mail: David S. Wilcove, Ecology and Evolutionary Biology and
m.j.keeling @ warwick.ac.uk Princeton Environmental Institute and Woodrow
Jeremy T. Kerr, Canadian Facility for Ecoinformatics Wilson School of Public and International Affairs,
Research (CFER), Department of Biology, University of Princeton University, Princeton, NJ 08544, USA.
Ottawa, Box 450, Station A, Ottawa, ON, K1N 6N5, E-mail: dwilcove @ princeton.edu
Canada. E-mail: jkerr @ uottawa.ca
ix
CHAPTER 1
Introduction
Angela R. McLean and Robert M. May
In this introductory chapter, we indicate the aims church, and the rest in some of the lowest and meanest
and structure of this book. We also indicate some thatched cottages. Now, as these eight pairs—allowance
of the ways in which the book is not synoptic in its being made for accidents—breed yearly eight pairs more,
coverage, but rather offers an interlinked account what becomes annually of this increase? and what
determines every spring, which pairs shall visit us, and
of some major developments in our understand-
re-occupy their ancient haunts?
ing of the dynamics of ecological systems, from
populations to communities, along with practical
This passage is unusual in giving quantitative
applications to important problems.
information about the population of swifts in Sel-
Ecology is a young science. The word ecology itself
borne two centuries ago, a small exception to the
was coined not much more than 100 years ago, and
almost universal absence of population records
the oldest professional society, the British Ecological
going back more than a few decades. It is even
Society, is less than a century old. Arguably the first
more remarkable for its clear articulation of the
published work on ecology was Gilbert White’s The
central question of population biology: what reg-
Natural History of Selborne. This book, published in
ulates populations? Interestingly, the swift popu-
1789, was ahead of its time in seeing plants and
lation of Selborne these days is steadily around 12
animals not as individual objects of wonder—things
pairs, which in ecological terms is not much dif-
to be assembled in a cabinet of curiosities—but as
ferent from eight, even though much of their
parts of a community of living organisms, interacting
environment has changed—entries to the church
with the environment, other organisms, and
tower all wired-off to keep out squirrels, and the
humans. The book has not merely remained in print,
gentrified cottages no longer low and mean with
but has run steadily through well over 200 editions
their thatch, when it remains, neatly wired down
and translations, to attain the status of the fourth
(Lawton and May, 1983). Interpreted generously,
most published book (in the sense of separate edi-
these population data on Selborne’s swifts could
tions) in the English language. The following excerpt
be seen as one of ecology’s longest time series, so it
captures White’s blend of detailed observation and
is sobering to realize there is still no agreed
concern for basic questions.
explanation of what actually regulates the swifts’
numbers.
Among the many singularities attending those amusing Moving on from Gilbert White, the first half of
birds, the swifts, I am now confirmed in the opinion that the twentieth century saw some more explicitly
we have every year the same number of pairs invariably;
mathematical models aimed at understanding the
at least, the result of my inquiry has been exactly the same
dynamical behaviour of populations. Notable
for a long time past. The swallows and martins are so
numerous, and so widely distributed over the village, that
examples include Ross’ work on malaria, with its
it is hardly possible to recount them; while the swifts, first introduction of the basic reproductive num-
though they do not all build in the church, yet so fre- ber, R0, discussed in later chapters of this book,
quently haunt it, and play and rendezvous round it, that and Lotka and Volterra’s indication of the inher-
they are easily enumerated. The number that I constantly ently oscillatory properties of prey–predator sys-
find are eight pairs, about half of which reside in the tems. Despite this, ecology seems to us to have
1
2 THEORETICAL ECOLOGY
remained a largely observational and descriptive ask whether particular factors may be more
subject up to the decade of the 1960s. Witness important than others, and to see if such insight or
two of the most influential texts of that time: guesswork does indeed provide testable explana-
Andrewartha and Birch (1954), an excellent book tions. Mathematical models can be precise tools for
but explicitly antithetic to theory in the form of doing this, helping us to make our assumptions
anything resembling a mathematical model; explicit and unambiguous, and to explore ‘ima-
Odum (1953), arguably foreshadowing aspects of ginary worlds’ as metaphors for such hypothetical
‘systems ecology’ with its insightful focus on simplicity underlying apparent complexity. The
patterns of energy flow in ecosystems, but with 1970s saw much activity of this kind in ecological
the emphasis descriptive rather than conceptual. research, helped in part by basic advances in our
For evolutionary studies as well as for ecological understanding of nonlinear dynamical systems
ones, we think the 1960s saw a change in the zeit- and by the advent of increasingly powerful and
geist. For evolution, much of the stimulus derived user-friendly computers.
from Bill Hamilton’s conceptual advances. For In particular, the phenomenon of deterministic
ecology, it was the reframing by Evelyn Hutchinson chaos received wide recognition in the 1970s. The
(1965) and his student Robert McArthur (1972; see finding that very simple and purely deterministic
also MacArthur and Wilson, 1967) of old questions laws or equations can give rise to dynamical
in more explicitly analytic ways; one could perhaps behaviour that not merely looks like random noise,
say, rephrasing them in the idiom of theoretical but is so sensitive to initial conditions that long-
physics. How similar can species be, yet persist term prediction is effectively impossible, has huge
together? What tends to govern the number of implications. It ends the Newtonian dream that if
species we see on an island, and how does this the system is simple (very few variables) and
number depend on the size and isolation of the orderly (the rules and parameters exactly known),
island? Gilbert White’s question of population then the future is predictable. The ‘law’ can be
abundance was revisited—and expanded beyond as trivial as x(t þ 1) ¼ lx(t) exp[ x(t)], with l a
the sterile controversies of the 1950s about whether known and unvarying constant, but if l is big
populations typically are governed by tight density enough then an error of one part in one million in
dependence or fluctuate greatly under the influence the initial estimate of x(0) will end up producing a
of environmental factors—to ask the more precise completely wrong prediction within a dozen or so
dynamical question of why do some populations time steps. Interestingly, it is often thought that
remain relatively steady, others show regular chaotic phenomena found applications in ecology
cycles, and yet others fluctuate wildly? Given the after others had developed the subject. In fact, one
observed patterns of relative abundance of the of the two streams which brought chaos centre
different species in particular communities, what stage in the 1970s derived directly from ecological
are the underlying causes? What is the relation research on models for a single population
between the complexity of a food web (variously with discrete, non-overlapping generations. These
defined) and its ability to withstand disturbance, models were first-order difference equations; the
natural or human created? other strand was Lorenz’s metaphor for convect-
These more deliberately conceptual or theoret- ive phenomena in meteorology, involving more
ical approaches differed from early work, in our complex—although still relatively simple—three-
view, in that they went beyond the codification of dimensional differential equations.
descriptive material, and the search for patterns Advances in computing have also been of great
within such codification, to ask questions about help in all areas of ecology: statistical design of
underlying mechanisms. To ask questions about experiments; collecting and processing data; and,
why, rather than what. Mathematics enters into coming to the present book, developing and
such studies, essentially as a tool for thinking exploring mathematical models for both simple
clearly. In pursuing a ‘why’ or ‘what if’ question and complicated ecological systems. There are,
about a complicated situation, it can be helpful to however, some associated dangers, which deserve
INTRODUCTION 3
passing mention. The understanding derived from experiments, and theory expressed in mathematical
computer studies of complicated models can terms. The comparison, for example, between the
sometimes be substantially less complete than that first edition of Begon, Townsend and Harper (1986)
gained from the analytic methods of classical and the earlier Andrewartha and Birch (1954) or
applied mathematics and theoretical physics. The Odum (1953) is pronounced. We think this marks a
early days of computers—mechanical calcula- maturation of the subject, although there undeni-
tors—saw them used by theoretical physicists ably remain large and important areas where there
in conjunction with analytic approximations, to are still more questions than answers.
explore previously intractable problems. The
result, however, was that at every step there was
1.1 This book and its predecessors
preserved an intuitive understanding of the rela-
tion between the underlying assumptions and the This book (TEIII) is essentially a greatly transmo-
results. In contrast, many scientists who today use grified version of one first published in 1976 (TEI),
computers to explore increasingly complex math- and followed with substantial changes in 1981
ematical models have little formal background in (TEII; this was not a perfunctory update, but had
mathematics, or have forgotten what they were three chapters completely re-written by different
once taught. Most of this work is interesting and authors, two new chapters added, and all others
excellent. But, absent any degree of intuitive revised; TEI’s 14 chapters involved 11 authors,
understanding of how the input assumptions TEII’s 16 chapters had 13 authors, of whom nine
about the system’s biology relate to the consequent were from TEI). This new version, 25 years on, has
output, we need to be wary (May, 2004). Too often, 15 chapters by 23 authors, only three of whom are
an ‘emergent phenomenon’ means little more than veterans of TEII.
‘I’ve no clue what is going on, but it looks kinda Like the previous two, this book is not a basic
interesting’. Happily, there are very few examples undergraduate ecology text, but equally it is not a
of this in ecology. More particularly, throughout technical tome for the front-line specialist in one or
the present book we aim, wherever possible, to other aspect of theoretical ecology. Rather, the book
provide intuitive understanding of the lessons is aimed at upper-level undergraduate, post-
learned from mathematical models. graduate, and postdoctoral students, and ecological
Be all this as it may, there has been a marked rise researchers interested in broadening aspects of the
in theoretical ecology as a distinct sub-discipline courses they teach, or indeed of their own work. As
over the past three decades or so. Many of the such, we think it fair to claim that TEI and TEII in
practitioners are not to be found in the field or their own time played a part in the above-men-
laboratory; a greater number, however, find their tioned transition in the general subject of ecology,
experimental contributions in field and/or lab- where earlier texts, in which mathematical content
oratory to be inextricably interwoven with their was essentially absent, contrast markedly with
theoretical and mathematical contributions. Ecol- today’s, where theoretical approaches—sometimes
ogy has come a long way from the 1970s, when a explicitly mathematical and sometimes not—play
few empirical ecologists resented outsiders, who an important part, although no more than a part, of
had not paid their dues of years of toil in the field, the presentation of the subject. Some of our
presuming to mathematize their problems (often acquaintances, indeed, still use the earlier volumes
sweeping aside arguably irrelevant, but certainly as supplements to their undergraduate courses.
beloved, details in the process). Others perhaps TEII, although out of print, still trades actively on
welcomed the intrusion too uncritically. the online bookseller Amazon.
The end result, however, is seen clearly by com- This book, on the other hand, differs from the
paring today’s leading ecology texts with those of previous two by virtue of these changes in how the
the 1950s and 1960s. In the latter, you will find very subject of ecology is defined and taught. Much of
few equations. Today, in contrast, you will find a the material in TEI and TEII would now, 25 years
balanced blend of observation, field and laboratory and more on, be seen as a routine part of any basic
ecology text. Other bits, of course, are just out of The past three decades have seen extraordinary
date, overtaken by later advances. advances in our understanding of the behavioural
One essential similarity with its predecessors is ecology and life-history strategies of individuals
that the present book does not aim at synoptic (e.g. Krebs and Davies, 1993). On the one hand,
coverage. Instead, it attempts first (in Chapters this is a formidable field to cover concisely, but on
2–9) to give an account of some of the basic prin- the other hand, only in a relatively few corners does
ciples that govern the structure, function, and this work deal directly with deducing the overall
temporal and spatial dynamics of populations and dynamics of a population from the behavioural
communities. These chapters are not tidily kept to ecology of its constituent individuals. There are
uniform length; we think the dynamics of plant some interesting examples of phenomena whose
populations have probably received less attention understanding unavoidably requires bringing the
than those of animal populations, and so have two together—for instance, odd aspects of brood
encouraged the authors in this area to go into parasitism where you cannot understand the
somewhat greater detail. Conversely, we recognise population dynamics without understanding the
that there are important and interesting areas of evolution of individual’s behaviour, and conversely
theoretical ecology—aspects of macroecology, or (Nee and May, 1993)—but they are few, and seem to
energy flows in ecosystems, for example—which have evoked little interest so far. A good review of
are not covered here. By the same token, the some other open questions at the interface between
‘applied’ chapters are a selection from the larger natural selection and population dynamics is by
universe of interesting and illuminating possibi- Saccheri and Hanski (2006). Resource managers get
lities. In short, advances over the past quarter by, and seem to be content, with treating the para-
century have seen significant growth in field and meters in population models as phenomenological
laboratory studies, along with major theoretical constants, fitted to data.
advances and practical applications. Any book on One really big problem, however, which is in
‘theoretical ecology’ simply has much more many ways as puzzling today as it was to Darwin,
ground to cover—many more subdisciplines and is how large aggregations of cooperating indivi-
specialized areas—than was the case for TEII. The duals (where group benefits are attained for a
result is inevitably that the present book has more relatively small cost to participating individuals,
gaps and omissions than its predecessors; inclu- but where the whole thing is vulnerable to cheats
sions and exclusions are bound to be more quirky. who take the benefits without paying the cost) can
A charitable interpretation would be that, just as evolve and maintain themselves. Relatively early
the gates to Japanese temples, tori, have deliberate work by Hamilton and Trivers pointed the way to
imperfections to avoid angering the gods, so too a solution of this problem for small groups of
we have avoided the sublime. The real reason is a closely related individuals. But much of this work
mixture of our own interests, and a feeling that is so restricted as to defy application to large
enough is enough. aggregations of human or other animals. The past
few years have, however, seen a diverse array of
significant advances in this area, and we thought it
1.2 What is in the book
would be better to begin with a definitive review
Previous editions of this text began with a chapter of this underpinning topic, which is still wide open
on the evolutionary forces which shape the behav- to further advances. Hence Chapter 2, How popu-
iour of individuals on a stage set by specific lations cohere: five rules for cooperation.
environmental and ecological factors, and then
show how such individual behaviour ultimately
1.2.1 Basic ecological principles
determines the demographic parameters—density-
dependent birth and death rates, movement The next two chapters deal with single popula-
patterns, and so on—governing the population’s tions. In Chapter 3 Coulson and Godfray distil the
behaviour in space and over time. essence of several recent monographic treatments
INTRODUCTION 5
of one or other aspect, to discuss how density- much-used term then—the word diversity does not
dependent or nonlinear effects, interacting to var- appear in the index to May’s Stability and Complexity
ious degrees with demographic and environmental in Model Ecosystems (1973a), and although it does
stochasticity, can result in relatively steady, or appear in the indexes for TEI and TEII, it clearly
cyclic, or erratically fluctuating population means simply numbers of species). Ives carefully
dynamics. They also sketch progress that has been enumerates the varied interpretations which have
made in looking at the ‘flipside of chaos’, namely been placed on the terms complexity/diversity and
the question of whether, when we see apparently stability. He goes on to give a thumbnail sketch of
noisy time series, we are looking at ‘environmental the way ideas have evolved in this area, guided by
and other noise’ or a deterministic but chaotic sig- empirical and theoretical advances, and concludes
nal. This survey is woven together with illustrative by presenting models which illustrate how the
accounts of field studies and laboratory experi- answers to questions about community dynamics
ments. In Chapter 4 Nee widens the discussion of can depend on precisely how the questions are
population dynamics to look at some of the com- framed. In Chapter 9 May, Crawley, and Sugihara
plications which arise when a single population is survey a range of recent work on ‘community pat-
spatially distributed over many patches. Fore- terns’: the relative abundance of species; species–
shadowing later chapters on conservation biology area relations; the network structure of food webs;
and on infectious diseases, he emphasizes that you and other things. This survey, which in places is a
do not have to destroy all of a population’s habitat bit telegraphic, seeks to outline both the underlying
to extinguish it. Widening the survey to include two observations and the suggested theoretical expla-
populations interacting as competitors, predator– nations, including null models (old and new) and
prey or mutualists, Nee further indicates other scaling laws.
aspects of the dynamics of such so-called meta-
populations which may seem counter-intuitive.
1.2.2 Applications to practical problems
The next three chapters expand on interacting
populations. Bonsall and Hassell first survey the The next five chapters turn to particular applica-
dynamical behaviour of prey–predator interactions of these theoretical advances. Grenfell and
tions. This chapter takes for granted some of the Keeling (Chapter 10) deal with the dynamics
by-now familiar material presented in TEII, giving and control of infectious diseases of both hum-
more attention to the way spatial complexities ans and other animals. They begin by explaining
contribute to the persistence of such associations how basic aspects of predator–prey theory apply
(and also noting that such spatial heterogeneity here, with particular emphasis on the infection’s
can even be generated by the nonlinear nature of basic reproductive number, R0. Recent applica-
the interactions themselves, even in an homo- tions to the outbreak of foot-and-mouth disease
geneous substrate). Crawley gives an overview of among livestock in the UK are discussed in some
the dynamics of plant populations, interpreting detail, although other examples could equally
‘plants’ broadly to emphasize the range of differ- well have been chosen (HIV/AIDS, SARS, H5N1
ent considerations which arise as we move from avian flu). Grenfell and Keeling emphasize the
diatoms to trees. This chapter also discusses plant– essential interplay between massively detailed
herbivore interactions as an important special case computations (the foot-and-mouth disease out-
of predators and prey. Competitive interactions break was modelled at the level of every farm in
are discussed by Tilman in Chapter 7, drawing Britain, an extreme example of an individual-level
together theoretical advances with long-term and approach to a population-level phenomenon) and
other field studies. basic dynamical understanding of what is going
Chapters 8 and 9 deal with the theoretical ecol- on, based on simple models.
ogy of communities. Ives’ chapter might have In Chapter 11 Beddington and Kirkwood give an
been called Complexity and stability in the 1970s account of the ecology of fisheries and their prac-
(not Diversity and stability; diversity was not a tical management. This chapter explains how the
dynamics of fish populations—as single species or population, and to do so in a way where crops are
in multispecies communities—interacts with prac- adapted to their environment (as distinct from past
tical policy options (quotas, tariffs, licenses, etc.), practice, where too often the environment was
in ways which can be complicated and sometimes wrenched to serve the crops by fossil-fuel energy
counter-intuitive. This is an area in which science- subsidies). Conway stresses that engagement and
based advice can be in conflict with political empowerment of local people is essential if this
considerations, sometimes in ways which have Doubly Green Revolution is to be realized, which
interesting resonance with the problems discussed again harks back to Nowak and Sigmund.
in Nowak and Sigmund’s opening chapter on the Chapter 13 by Dobson, Turner, and Wilcove
evolution of cooperation. In passing, we observe deals directly with conservation biology, survey-
that a vast amount of interesting ecological data, ing some of the factors which threaten species with
and also of excellent theoretical work, is to be extinction, indicating possible remedial actions,
found in the grey literature associated with the but also noting some of the economic and political
work of bodies like the International Council for realities that can impede effective action. Chapter
the Exploration of the Seas (ICES) or the Scientific 14, on Climate Change and Conservation Biology, by
Committee of the International Whaling Commis- Kerr and Kharouba, amplifies one particularly
sion (IWC); it is unfortunate that too little of this important threat to the survival of species, namely
makes its way into mainstream ecological meet- the effects that climate change are likely to have on
ings and scientific journals. We think Chapter 11 is species’ habitats and ranges.
particularly interesting for the way it reaches into The concluding Chapter 15 offers a selective and
this grey literature. opinionated review of some of the major environ-
The term Doubly Green Revolution was coined mental threats that loom for us and other species
by Gordon Conway, one of the three continuing over the coming few centuries. The emphasis is on
authors from TEII (along with Hassell and May). issues where ecological knowledge can provide a
Here, in Chapter 12, he surveys the triumphs and guide to appropriate action, or to areas where
problems of the earlier Green Revolution, which current lack of ecological understanding is a han-
has doubled global food production on only 10% dicap. One thing is sure: the future for other living
additional land area over the past 30 years or so. things on planet Earth, not just humans, depends
Looking to the future, he suggests how new tech- on our understanding and managing ecosystems
nologies offer the potential to feed tomorrow’s better than we have been doing recently.
CHAPTER 2
How populations cohere: five

rules for cooperation
Martin A. Nowak and Karl Sigmund
Subsequent chapters in this volume deal with is a fundamental principle that is required for
populations as dynamic entities in time and space. every level of biological organization. Individual
Populations are, of course, made up of individuals, cells rely on cooperation among their components.
and the parameters which characterize aggregate Multicellular organisms exist because of coopera-
behavior—population growth rate and so on— tion among their cells. Social insects are masters of
ultimately derive from the behavioral ecology and cooperation. Most aspects of human society are
life-history strategies of these constituent indivi- based on mechanisms that promote cooperation.
duals. In evolutionary terms, the properties Whenever evolution constructs something entirely
of populations can only be understood in terms new (such as multicellularity or human language),
of individuals, which comes down to studying cooperation is needed. Evolutionary construction
how life-history choices (and consequent gene- is based on cooperation.
frequency distributions) are shaped by environ- The five rules for cooperation which we examine
mental forces. in this chapter are: kin selection, direct reciprocity,
Many important aspects of group behavior— indirect reciprocity, graph selection, and group
from alarm calls of birds and mammals to the selection. Each of these can promote cooperation if
complex institutions that have enabled human specific conditions are fulfilled.
societies to flourish—pose problems of how coopera-
tive behavior can evolve and be maintained. The
2.1 Kin selection
puzzle was emphasized by Darwin, and remains
the subject of active research today. The heated conversation took place in an unheated
In this book, we leave the large subject of indi- British pub over some pints of warm bitter. Sud-
vidual organisms’ behavioral ecology and life- denly J.B.S. Haldane remarked, ‘I will jump into
history choices to texts in that field (e.g. Krebs and the river to save two brothers or eight cousins.’
Davies, 1997). Instead, we lead with a survey of The founding father of population genetics and
work, much of it very recent, on five different dedicated communist in his spare time never
kinds of mechanism whereby cooperative behavior bothered to develop this insight any further. The
may be maintained in a population, despite the witness of the revelation was Haldane’s eager
inherent difficulty that cheats may prosper by pupil, the young John Maynard Smith. But given
enjoying the benefits of cooperation without pay- John’s high regard for entertaining stories and
ing the associated costs. good beer, can we trust his memory?
Cooperation means that a donor pays a cost, c, The insight that Haldane might have had in the
for a recipient to get a benefit, b. In evolutionary pub was precisely formulated by William Hamilton.
biology, cost and benefit are measured in terms of He wrote a PhD thesis on this topic, submitted a
fitness. While mutation and selection represent the long paper to the Journal of Theoretical Biology, and
main forces of evolutionary dynamics, cooperation spent much of the next decade in the Brazilian
7
jungle. This was one of the most important papers defect it is also better to defect, because P > S.
in evolutionary biology in the second half of the Hence, no matter what the other person will do it
twentieth century (Hamilton, 1964a, 1964b). The is best to defect. If both players analyze the game
theory was termed kin selection by Maynard Smith in this rational way then they will end up defect-
(1964). The crucial equation is the following. ing. The dilemma is that they both could have
Cooperation among relatives can be favored by received a higher payoff if they had chosen to
natural selection if the coefficient of genetic relat- cooperate. But cooperation is irrational.
edness, r, between the donor and the recipient We can also imagine a population of cooperators
exceeds the cost/benefit ratio of the altruistic act: and defectors and assume that the payoff for each
player is determined by many random interactions
r > c=b ð2:1Þ
with others. Let x denote the frequency of coopera-
Kin-selection theory has been tested in numerous tors and 1 x the frequency of defectors. The
experimental studies. Indeed, many cooperative expected payoff for a cooperator is fC ¼ Rx þ
acts among animals occur between close kin S(1 x). The expected payoff for a defector is
(Frank, 1998; Hamilton, 1998). The exact relation- fD ¼ Tx þ P(1 x). Therefore, for any x, defectors
ship between kin selection and other mechanisms have a higher payoff than cooperators. In evolu-
such as group selection and spatial reciprocity, tionary game theory, payoff is interpreted as fit-
however, remains unclear. A recent study even ness. Successful strategies reproduce faster and
suggests that much of cooperation in social insects outcompete less successful ones. Reproduction can
is due to group selection rather than kin selection be cultural or genetic. In the non-repeated Pris-
(Wilson and Hölldobler, 2005). Note that kin oner’s Dilemma, in a well-mixed population,
selection is more likely to work in quite small defectors outcompete cooperators. Natural selec-
groups; in large groups, unless highly inbred, the tion favors defectors.
average value of r will be tiny. Cooperation becomes an option if the game is
repeated. Suppose there are m rounds. Let us
compare two strategies, always defect (ALLD),
and GRIM, which cooperates on the first move,
2.2 Direct reciprocity
then cooperates as long as the opponent coopera-
In 1971, Robert Trivers published a landmark tes, but permanently switches to defection if the
paper entitled ‘The evolution of reciprocal altru- opponent defects once. The expected payoff for
ism’ (Trivers, 1971). Trivers analyzed the question GRIM versus GRIM is nR. The expected payoff for
how natural selection could lead to cooperation ALLD versus GRIM is T þ (m 1)P. If nR > T þ
between unrelated individuals. He discusses three (m 1)P then ALLD cannot spread in a GRIM
biological examples: cleaning symbiosis in fish, population when rare. This is an argument of
warning calls in birds, and human interactions. evolutionary stability. Interestingly, Trivers (1971)
Trivers cites Luce and Raiffa (1957) and Rapoport quotes ‘Hamilton (pers. commun.)’ for this idea.
and Chammah (1965) for the Prisoner’s Dilemma, A small problem with the above analysis is that
which is a game where two players have the given a known number of rounds it is best to
option to cooperate or to defect. If both cooperate defect in the last round and by backwards induc-
they receive the reward, R. If both defect they tion it is also best to defect in the penultimate
receive the punishment, P. If one cooperates and round and so on. Therefore, it is more natural to
the other defects, then the cooperator receives the consider a repeated game with a probability w of
sucker’s payoff, S, while the defector receives the having another round. In this case, the expected
temptation, T. The Prisoner’s Dilemma is defined number of rounds is 1/(1 w), and GRIM is stable
by the ranking T > R > P > S. against invasion by ALLD provided w > (T R)/
Would you cooperate or defect? Assuming the (T P).
other person will cooperate it is better to defect, We can also formulate the Prisoner’s Dilemma
because T > R. Assuming the other person will as follows. The cooperator helps at a cost, c, and
HOW POPULATIONS COHERE 9
the other individual receives a benefit, b. Defectors The tournaments were conducted without strategic
do not help. Therefore we have T ¼ b, R ¼ b c, noise. In a real world, trembling hands and fuzzy
P ¼ 0, and S ¼ c. The family of games that is minds cause erroneous moves. If two TFT players
described by the parameters b and c is a subset of interact with each other, a single mistake leads
all possible Prisoner’s Dilemma games as long as to a long sequence of alternating defection and
b > c. For the repeated Prisoner’s Dilemma, we find cooperation. In the long run two TFT players get
that ALLD cannot invade GRIM if the same low payoff as two players who flip coins
for every move in order to decide whether to
w > c=b ð2:2Þ
cooperate or to defect. Errors destroy TFT.
The probability of having another round must Our own investigations in this area began after
exceed the cost/benefit ratio of the altruistic act reading a News and Views article in Nature where
(Axelrod and Hamilton, 1981; Axelrod, 1984). the author made three important points: first, he
Notice, however, the implicit assumption here that often leaves university meetings with a renewed
the payoff for future rounds is not discounted (i.e. appreciation for the problem of how natural
distant benefits count as much as present ones). In selection can favor cooperative acts given that
evolutionary reality, this is unlikely. We can selfish individuals gain from cheating; second,
address this by incorporating an appropriate dis- strategies in the repeated Prisoner’s Dilemma
count factor in w (May, 1987), but note, from eqn 2, should not be error-free but subjected to noise;
that this makes cooperation less likely. third, evolutionary stability should be tested not
Thus, the repeated Prisoner’s Dilemma allows against single invaders but against heterogeneous
cooperation, but the question arises: what is a good ensembles of invaders (May, 1987). This was the
strategy for playing this game? This question was motivation for the following work.
posed by the political scientist, Robert Axelrod. In In 1989, we conducted evolutionary tourna-
1979, he decided to conduct a tournament of ments. Instead of inviting experts to submit pro-
computer programs playing the repeated Prisoner’s grams, we asked mutation and selection to explore
Dilemma. He received 14 entries, of which the (some portion of) the strategy space of the repe-
surprise winner was tit-for-tat (TFT), the simplest ated Prisoner’s Dilemma in the presence of noise.
of all strategies that were submitted. TFT coopera- The initial random ensemble of strategies was
tes in the first move, and then does whatever quickly dominated by ALLD. If the opposition is
the opponent did in the previous round. TFT random, it is best to defect. A large portion of the
cooperates if you cooperate, TFT defects if you population began to adopt the ALLD strategy and
defect. It was submitted by the game theorist everything seemed lost. But after some time, a
Anatol Rapoport (who is also the co-author of the small cluster of players adopted a strategy very
book Prisoner’s Dilemma; Rapoport and Chammah, close to TFT. If this cluster is sufficiently large,
1965). Axelrod analyzed the events of the tourna- then it can increase in abundance, and the entire
ment, published a detailed account and invited population swings from ALLD to TFT. Reciprocity
people to submit strategies for a second cham- (and therefore cooperation) has emerged. We can
pionship. This time he received 63 entries. John show that TFT is the best catalyst for the emer-
Maynard Smith submitted tit-for-two-tats, a var- gence of cooperation. But TFT’s moment of glory
iant of TFT which defects only after the opponent was brief and fleeting. In all cases, TFT was rapidly
has defected twice in a row. Only one person, replaced by another strategy. On close inspection,
Rapoport, submitted TFT, and it won again. At this this strategy turned out to be generous tit-for-tat
time, TFT was considered to be the undisputed (GTFT), which always cooperates if the opponent
champion in the heroic world of the repeated has cooperated on the previous move, but some-
Prisoner’s Dilemma. times (probabilistically) even cooperates when the
But one weakness became apparent very soon opponent has defected. Natural selection had dis-
(Molander, 1985). TFT cannot correct mistakes. covered forgiveness (Nowak and Sigmund, 1992).
After many generations, however, GTFT is time. Most surprisingly, this strategy is based on
undermined by unconditional cooperators, ALLC. the extremely simple principle of win-stay, lose-
In a society where everybody is nice (using GTFT), shift (WSLS). If my payoff is R or T then I will
there is almost no need to remember how to continue with the same move next round. If I have
retaliate against a defection. A biological trait that cooperated then I will cooperate again, if I have
is not used is likely to be lost by random drift. defected then I will defect again. If my payoff is
Birds that escape to islands without predators lose only S or P then I will switch to the other move
the ability to fly. Similarly, a GTFT population is next round. If I have cooperated then I will defect,
softened and turns into an ALLC population. if I have defected then I will cooperate (Figure 2.2).
Once most people play ALLC, there is an open If two WSLS strategists play each other, they
invitation for ALLD to seize power. This is pre- cooperate most of the time. If a defection occurs
cisely what happens. The evolutionary dynamics accidentally, then in the next move both will
run in cycles: from ALLD to TFT to GTFT to ALLC defect. Hereafter both will cooperate again. WSLS
and back to ALLD. These oscillations of coopera- is a simple deterministic machine to correct sto-
tive and defective societies are a fundamental part chastic noise. While TFT cannot correct mistakes,
of all our observations regarding the evolution of both GTFT and WSLS can. But WSLS has an
cooperation. Most models of cooperation show additional ace in its hand. When WSLS plays
such oscillations. Cooperation is never a final state ALLC it will discover after some time that ALLC
of evolutionary dynamics. Instead it is always lost does not retaliate. After an accidental defection,
to defection after some time and has to be WSLS will switch to permanent defection. There-
re-established. These oscillations are also reminis- fore, a population of WSLS players does not drift to
cent of alternating episodes of war and peace in ALLC. Cooperation based on WSLS is more stable
human history (Figure 2.1). than cooperation based on TFT-like strategies.
A subsequent set of simulations, exploring a
larger strategy space, led to a surprise (Nowak and Win-stay
Sigmund, 1993). The fundamental oscillations were
C (3) …. C D (5) …. D
interrupted by another strategy which seems to be
able to hold its ground for a very long period of C C
Lose-shift
Tit-for-tat Generous tit-for-tat C (0) …. D D (1) …. C (probabilistic)
D D
Figure 2.2 Win-stay, lose-shift (WSLS) embodies a very simple

Always defect Always cooperate
principle. If you do well then continue with what you are doing. If you
are not doing well, then try something else. Here we consider the
Prisoner’s Dilemma payoff values R ¼ 3, T ¼ 5, P ¼ 1, and S ¼ 0. If
both players cooperate, you receive three points, and you continue to
Win-stay, lose-shift
cooperate. If you defect against a cooperator, you receive five points,
Figure 2.1 Evolutionary cycles of cooperation and defection. A and you continue to defect. But if you cooperate with a defector, you
small cluster of tit-for-tat (TFT) players or even a lineage starting from receive no points, and therefore you will switch from cooperation to
a single TFT player in a finite population can invade an always defect defection. If, on the other hand, you defect against a defector, you
(ALLD) population. In fact, TFT is the most efficient catalyst for the receive one point, and you will switch to cooperation. Your aspiration
first emergence of cooperation in an ALLD population. But in a world level is three points. If you get at least three points then you consider
of fuzzy minds and trembling hands, TFT is soon replaced by generous it a win and you will stay with your current choice. If you get less
tit-for-tat (GTFT), which can re-establish cooperation after occasional than three points, you consider it a loss and you will shift to another
mistakes. If everybody uses GTFT, then always cooperate (ALLC) is a move. If R > (T þ P)/2 (or b/c > 2) then WSLS is stable against
neutral variant. Random drift leads to ALLC. An ALLC population invasion by ALLD. If this inequality does not hold, then our evolu-
invites invasion by ALLD. But ALLC is also dominated by win-stay, tionary simulations lead to a stochastic variant of WSLS, which
lose-shift (WSLS), which leads to more stable cooperation than TFT- cooperates after a DD move only with a certain probability. This
like strategies. stochastic variant of WSLS is then stable against invasion by ALLD.
The repeated Prisoner’s Dilemma is mostly known display a large amount of cooperation between
as a story of TFT, but WSLS is a superior strategy non-relatives (Fehr and Fischbacher, 2003). A con-
in an evolutionary scenario with errors, mutation, siderable part of human cooperation is based on
and many generations (Fudenberg and Maskin, moralistic emotions, such as anger directed towards
1990; Nowak and Sigmund, 1993). cheaters or the warm inner glow felt after perform-
In the infinitely repeated game, WSLS is stable ing an altruistic action. Intriguingly, humans not
against invasion by ALLD if b/c > 2. If instead only feel strongly about interactions that involve
1 < b/c < 2 then a stochastic variant of WSLS them directly, they also judge actions between third
dominates the scene; this strategy cooperates after parties as evidenced by the contents of gossip. There
a mutual defection only with a certain probability. are numerous experimental studies of indirect
Of course, all strategies of direct reciprocity, such reciprocity based on reputation (Wedekind and
as TFT, GTFT, or WSLS can only lead to the evo- Milinski, 2000; Milinski et al., 2002; Wedekind and
lution of cooperation if the fundamental inequality Braithwaite, 2002; Seinen and Schram, 2006).
(eqn 2.2) is fulfilled. A simple model of indirect reciprocity (Nowak
and Sigmund, 1998a, 1998b) assumes that within a
well-mixed population, individuals meet ran-
2.3 Indirect reciprocity
domly, one in the role of the potential donor, the
Whereas direct reciprocity embodies the idea of other as potential recipient. Each individual
you scratch my back and I scratch yours, indirect experiences several rounds of this interaction in
reciprocity suggests that you scratch my back and both roles, but never with the same partner twice.
I scratch someone else’s. Why should this work? A player can follow either an unconditional strat-
Presumably I will not get scratched if it becomes egy, such as always cooperate or always defect, or
known that I scratch nobody. Indirect reciprocity, a conditional strategy, which discriminates among
in this view, is based on reputation (Nowak and the potential recipients according to their past
Sigmund, 1998a, 1998b, 2005). But why should you interactions. In a simple example, a discriminating
care about what I do to a third person? donor helps a recipient if her score exceeds a
The main reason why economists and social certain threshold. A player’s score is 0 at birth,
scientists are interested in indirect reciprocity is increases whenever that player helps and decrea-
because one-shot interactions between anonymous ses whenever the player withholds help. Indivi-
partners in a global market become increasingly dual-based simulations and direct calculations
frequent and tend to replace the traditional long- show that cooperation based on indirect reci-
lasting associations and long-term interactions procity can evolve provided the probability, q, of
between relatives, neighbors, or members of the knowing the social score of another person exceeds
same village. Again, as for kin selection, it is a the cost/benefit ratio of the altruistic act:
question of the size of the group. A substantial part
of our life is spent in the company of strangers, q > c=b ð2:3Þ
and many transactions are no longer face to face.
The growth of online auctions and other forms of The role of genetic relatedness that is crucial for
e-commerce is based, to a considerable degree, on kin selection is replaced by social acquaintance-
reputation and trust. The possibility to exploit such ship. In a fluid population, where most inter-
trust raises what economists call moral hazards. actions are anonymous and people have no
How effective is reputation, especially if informa- possibility of monitoring the social score of others,
tion is only partial? indirect reciprocity has no chance. But in a socially
Evolutionary biologists, on the other hand, are viscous population, where people know each other’s
interested in the emergence of human societies, reputation, cooperation by indirect reciprocity
which constitutes the last (up to now) of the major can thrive (Nowak and Sigmund, 1998a).
transitions in evolution. In contrast to other eusocial In a world of binary moral judgements (Nowak
species, such as bees, ants, or termites, humans and Sigmund, 1998b; Leimar and Hammerstein,
2001; Fishman, 2003; Panchanathan and Boyd, defection, D, always leads to a bad reputation, B.
2003; Brandt and Sigmund, 2004, 2005), there are Standing (Sugden, 1986) is like scoring, but it is not
four ways of assessing donors in terms of first- bad if a good donor defects against a bad recipient.
order assessment: always consider them as good, With judging, in addition, it is bad to cooperate
always consider them as bad, consider them as with a bad recipient. For another assessment rule,
good if they refuse to give, or consider them as shunning, all donors who meet a bad recipient
good if they give. Only this last option makes become bad, regardless of what action they choose.
sense. Second-order assessment also depends on Shunning strikes us as grossly unfair, but it
the score of the receiver; for example, it can be emerges as the winner in a computer tournament
deemed good to refuse help to a bad person. There if errors in perception are included and if there are
are 16 second-order rules. Third-order assessment only a few rounds in the game (Takahashi and
also depends on the score of the donor; for Mashima, 2003).
example, a good person refusing to help a bad An action rule for indirect reciprocity prescribes
person may remain good, but a bad person refus- giving or not giving, depending on the scores of
ing to help a bad person remains bad. There are both donor and recipient. For example, you may
256 third-order assessment rules. We display four decide to help if the recipient’s score is good or
of them in Figure 2.3. your own score is bad. Such an action might
With the scoring assessment rule, cooperation, increase your own score and therefore increase
C, always leads to a good reputation, G, whereas the chance of receiving help in the future. There
are 16 action rules.
Reputation of donor and recipient
If we view a strategy as the combination of an
action rule and an assessment rule, we obtain 4096
GG GB BG BB strategies. In a remarkable calculation, Ohtsuki
C G G G G and Iwasa (2004, 2005) analyzed all 4096 strategies
Scoring
D B B B B and proved that only eight of them are evolutio-
narily stable under certain conditions and lead to
C G G G G cooperation (Figure 2.4).
Action of donor
Standing Both standing and judging belong to the leading

D B G B B
eight, but scoring and shunning are not. However,
we expect that scoring has a similar role in indirect
C G B G B
Judging reciprocity to that of TFT in direct reciprocity.
D B G B B Neither strategy is evolutionarily stable, but their
simplicity and their ability to catalyze cooperation
C G B G B in adverse situations constitute their strength. In
Shunning
D B B B B extended versions of indirect reciprocity, in which
donors can sometimes deceive others about the
Reputation of donor
after the action
reputation of the recipient, scoring is the foolproof
concept of ‘I believe what I see’. Scoring judges
Figure 2.3 Four assessment rules. Assessment rules specify how an the action and ignores the stories. There is also
observer judges an interaction between a potential donor and a experimental evidence that humans follow scoring
recipient. Here we show four examples of assessment rules in a world
of binary reputation, good (G) and bad (B). For scoring, cooperation
rather than standing (Milinski et al., 2001).
(C) earns a good reputation and defection (D) earns a bad reputation. In human evolution, there must have been a
Standing is very similar to scoring; the only difference is that a good tendency to move from the simple cooperation
donor can defect against a bad recipient without losing his good promoted by kin or group selection to the strategic
reputation. Note that scoring is associated with costly punishment subtleties of direct and indirect reciprocity. Direct
(Sigmund et al., 2001; Fehr and Gaechter, 2002), whereas for
standing punishment of bad recipients is cost-free. For judging it is
reciprocity requires precise recognition of indivi-
bad to help a bad recipient. Shunning assigns a bad reputation to any dual people, a memory of the various interactions
donor who interacts with a bad recipient. one had with them in the past, and enough brain
GG GB BG BB b b
C G * G * D D
Assessment D
D B G B * C
D C
Action b C
C D C C/D
2b – 5c
2b – 2c C
D C
If a good donor meets a bad recipient, 2b – 3c
the donor must defect, and this action does
not reduce his reputation. Figure 2.5 Games on graphs. The members of a population occupy
the vertices of a graph (or social network). The edges denote who
* can be set as G or B. interacts with whom. Here we consider the specific example of
G cooperators, C, competing with defectors, D. A cooperator pays a
If a column in the assessment module is B cost, c, for every link. Each neighbor of a cooperator receives a
then the action must be C, otherwise D. benefit, b. The payoffs of some individuals are indicated in the figure.
The fitness of each individual is a constant, denoting the baseline
Figure 2.4 Ohtsuki and Iwasa’s leading eight. Ohtsuki and Iwasa fitness, plus the payoff of the game. For evolutionary dynamics, we
(2004, 2005) have analyzed the combination of 28 ¼ 256 assessment assume that in each round a random player is chosen to die, and the
modules with 24 ¼ 16 action modules. This is a total of 4096 neighbors compete for the empty site proportional to their fitness. A
strategies. They have found that eight of these strategies can be simple rule emerges: if b/c > k then selection favors cooperators over
evolutionarily stable and lead to cooperation, provided that everybody defectors. Here k is the average number of neighbors per individual.
agrees on each other’s reputation. (In general, uncertainty and
incomplete information might lead to private lists of the reputation of
others.) The three asterisks in the assessment module indicate a free provided the very selective scenario that led to
choice between G and B. There are therefore 23 ¼ 8 different cerebral expansion in human evolution.
assessment rules which make up the leading eight. The action module
is built as follows: if the column in the assessment module is G and B,
then the corresponding action is C, otherwise the action is D. Note 2.4 Graph selection
that standing and judging are members of the leading eight, but that
scoring and shunning are not. The traditional model of evolutionary game
dynamics assumes that populations are well-mixed
(Taylor and Jonker, 1978; Hofbauer and Sigmund,
1998). This means that interactions between any
power to conduct multiple repeated games two players are equally likely. More realistically,
simultaneously. Indirect reciprocity, in addition, however, the interactions between individuals are
requires the individual to monitor interactions governed by spatial effects or social networks. Let
among other people, possibly judge the intentions us therefore assume that the individuals of a
that occur in such interactions, and keep up with population occupy the vertices of a graph (Nowak
the ever-changing social network of the group. and May, 1992; Nakamaru et al., 1997, 1998; Skyrms
Reputation of players may not only be determined and Pemantle, 2000; Abramson and Kuperman,
by their own actions, but also by their associations 2001; Ebel and Bornholdt, 2002; Lieberman et al.,
with others. 2005; Nakamaru and Iwasa, 2005; Santos et al., 2005;
We expect that indirect reciprocity has Santos and Pacheco, 2005). The edges of the graph
coevolved with human language. On the one hand, determine who interacts with whom (Figure 2.5).
it is helpful to have names for other people and to Consider a population of N individuals consist-
receive information about how a person is per- ing of cooperators and defectors. A cooperator
ceived by others. On the other hand a complex helps all individuals to whom it is connected, and
language is needed, especially if there are intricate pays a cost, c. If a cooperator is connected to k
social interactions. The possibilities for games of other individuals and i of those are cooperators,
manipulation, deceit, cooperation, and defection then its payoff is bi ck. A defector does not pro-
are limitless. It is likely that indirect reciprocity has vide any help, and therefore has no costs, but it
can receive the benefit from neighboring coopera- degree, k, which is given by the average number of
tors. If a defector is connected to k other indivi- links per individual (Ohtsuki et al., 2006). This
duals and j of those are cooperators, then its payoff relationship can be shown with the method of
is bj. Evolutionary dynamics are described by an pair-approximation for regular graphs, where all
extremely simple stochastic process: at each time individuals have exactly the same number of
step, a random individual adopts the strategy of neighbors. Regular graphs include cycles, all kinds
one of its neighbors proportional to their fitness. of spatial lattice, and random regular graphs.
We note that stochastic evolutionary game Moreover, computer simulations suggest that the
dynamics in finite populations are sensitive to the rule b/c > k also holds for non-regular graphs such
intensity of selection. In general, the reproductive as random graphs and scale-free networks. The
success (fitness) of an individual is given by a rule holds in the limit of weak selection and k << N.
constant, denoting the baseline fitness, plus the For the complete graph, k ¼ N, we always have
payoff that arises from the game under con- rD > 1/N > rC. Preliminary studies suggest that
sideration. Strong selection means that the payoff eqn 2.4 also tends to hold for strong selection. The
is large compared with the baseline fitness; weak basic idea is that natural selection on graphs (in
selection means the payoff is small compared with structured populations) can favor unconditional
the baseline fitness. It turns out that many inter- cooperation without any need for strategic com-
esting results can be proven for weak selection, plexity, reputation, or kin selection.
which is an observation also well known in Games on graphs grew out of the earlier tradi-
population genetics. tion of spatial evolutionary game theory (Nowak
The traditional, well-mixed population of evo- and May, 1992; Herz, 1994; Killingback and
lutionary game theory is represented by the com- Doebeli, 1996; Mitteldorf and Wilson, 2000; Hauert
plete graph, where all vertices are connected, et al., 2002; Le Galliard et al., 2003; Hauert and
which means that all individuals interact equally Doebeli, 2004; Szabó and Vukov, 2004) and inves-
often. In this special situation, cooperators are tigations of spatial models in ecology (Durrett and
always opposed by natural selection. This is the Levin, 1994a, 1994b; Hassell et al., 1994; Tilman and
fundamental intuition of classical evolutionary Kareiva, 1997; Neuhauser, 2001) and spatial mod-
game theory. But what happens on other graphs? els in population genetics (Wright, 1931; Fisher
We need to calculate the probability, rC, that a and Ford, 1950; Maruyama, 1970; Slatkin, 1981;
single cooperator starting in a random position Barton, 1993; Pulliam, 1988; Whitlock, 2003).
turns the whole population from defectors into
cooperators. If selection neither favors nor opposes
2.5 Group selection
cooperation, then this probability is 1/N, which is
the fixation probability of a neutral mutant. If the The enthusiastic approach of early group selec-
fixation probability rC is greater than 1/N, then tionists to explain all evolution of cooperation
selection favors the emergence of cooperation. from this one perspective (Wynne-Edwards, 1962)
Similarly, we can calculate the fixation probability has met with vigorous criticism (Williams, 1966)
of defectors, rD. A surprisingly simple rule deter- and even a denial of group selection for decades.
mines whether selection on graphs favors coopera- Only an embattled minority of scientists continued
tion. If to study the approach (Eshel, 1972; Levin and
Kilmer, 1974; Wilson, 1975; Matessi and Jayakar,
b=c > k ð2:4Þ 1976; Wade, 1976; Uyenoyama and Feldman, 1980;
Slatkin, 1981; Leigh, 1983; Szathmary and Demeter,
then cooperators have a fixation probability of 1987). Nowadays it seems clear that group selection
greater than 1/N and defectors have a fixation can be a powerful mechanism to promote coopera-
probability of less than 1/N. Thus, for graph tion (Sober and Wilson, 1998; Keller, 1999; Michod,
selection to favor cooperation, the benefit/cost 1999; Swenson et al., 2000; Kerr and Godfrey-Smith, 2002;
ratio of the altruistic act must exceed the average Paulsson, 2002; Boyd and Richerson, 2002; Bowles
and Gintis, 2004; Traulsen et al., 2005). We only

have to make sure that its basic requirements
are fulfilled in a particular situation (Levin and
Kilmer, 1974; Maynard Smith, 1976). Exactly what D C
these requirements are can be illustrated with a D D
simple model (Traulsen and Nowak, 2006). D
Imagine a population of individuals subdivided
into groups. For simplicity, we assume that the
number of groups is constant and given by m. Each C D
group contains between 1 and n individuals. The CC C
total population size can fluctuate between the
bounds m and nm. Again, there are two types of D C C
individual, cooperators and defectors. Individuals
interact with others in their group and thereby
receive a payoff. At each time step a random indi- Figure 2.6 A simple model of group selection. A population
vidual from the entire population is chosen propor- consists of m groups of maximum size n. Individuals interact with
tional to payoff in order to reproduce. The offspring others in their group in the context of an evolutionary game. Here we
is added to the same group. If the group size is less consider the game between cooperators, C, and defectors, D. For
reproduction, individuals are chosen from the entire population with
than or equal to n then nothing else happens. If the
a probability proportional to their payoff. The offspring is added to
group size, however, exceeds n then with probability the same group. If a group reaches the maximum size, n, then it
q the group splits into two. In this case, a random either splits in two or a random individual from that group is
group is eliminated (in order to maintain a constant eliminated. If a group splits, then a random group dies, in order to
number of groups). With probability 1 q, the group keep the total population size constant. This metapopulation
structure leads to the emergence of two levels of selection, although
does not divide, but instead a random individual
only individuals reproduce.
from that group is eliminated (Figure 2.6)*.
This minimalist model of multilevel selection what we would obtain for a neutral mutant. An
has some interesting features. Note that the evo- analytic calculation is possible in the interesting
lutionary dynamics are entirely driven by indivi- limit q << 1, where individuals reproduce much
dual fitness. Only individuals are assigned payoff more rapidly than groups divide. In this case, most
values. Only individuals reproduce. Groups can of the groups are at their maximum size and hence
stay together or split (divide) when reaching a the total population size is almost constant and
certain size. Groups that contain fitter individuals given by N ¼ nm. We find that selection favors
reach the critical size faster and therefore split cooperators and opposes defectors, rC > 1/N > rD, if
more often. This concept leads to selection among
groups, although only individuals reproduce. The b=c > 1 þ n=ðm 2Þ ð2:5aÞ
higher level selection emerges from lower level
This result holds for weak selection. Smaller group
reproduction. Remarkably, the two levels of
sizes and larger numbers of competing groups
selection can oppose each other.
favor cooperation. We also notice that the number
As before, we can compute the fixation prob-
of groups, m, must exceed 2. There is an intuitive
abilities, rC and rD, of cooperators and defectors to
reason for this threshold. Consider the case of
check whether selection favors one or the other. If
m ¼ 2 groups with n ¼ 2 individuals. In a mixed
we add a single cooperator to a population of
group, the cooperator has payoff c and the
defectors, then this cooperator must first take over
defector has payoff b; the defector/cooperator
a group. Subsequently the group of cooperators
difference is b þ c. In a homogeneous group, two
must take over the entire population. The first step
cooperators have payoff b c, while two defectors
is opposed by selection, the second step is favored
*
by selection. Hence, we need to find out if the This is not the same q as in section 2.3; we have run out of
overall fixation probability is greater to or less than convenient letters.
have a payoff of 0. Thus the disadvantage for groups and migration between groups is not too
cooperators in mixed groups cannot be compen- frequent.
sated for by the advantage they have in homo-
geneous groups. Interestingly, however, for larger
splitting probabilities, q, we find that cooperators 2.6 Conclusion
can be favored even for m ¼ 2 groups. The reason We end by listing the five rules that we mentioned
is the following: for very small q, the initial in the beginning. These rules represent laws
cooperator must reach fixation in a mixed group; of nature governing the natural selection of
but for larger q, a homogeneous cooperator group cooperation.
can also emerge if a mixed group splits, giving rise
to a daughter group that has only cooperators. 1. Kin selection leads to cooperation if b/c > 1/r,
Thus, larger splitting probabilities make it easier where r is the coefficient of genetic relatedness
for cooperation to emerge. between donor and recipient.
Let us also consider the effect of migration 2. Direct reciprocity leads to cooperation if
between groups. The average number of migrants b/c > 1/w, where w is the probability of playing
accepted by a group during its lifetime is denoted another round in the repeated Prisoner’s Dilemma.
by z. We find that selection favors cooperation 3. Indirect reciprocity leads to cooperation if
provided that b/c > 1/q, where q is the probability of knowing
the reputation of a recipient.
b=c > 1 þ z þ n=m ð2:5bÞ 4. Graph selection (or network reciprocity) leads to
cooperation if b/c > k, where k is the degree of the
In order to derive this condition we have assumed
graph; that is, the average number of neighbors.
weak selection and q << 1, as before, but also that
5. Group selection leads to cooperation if
both the numbers of groups, m, and the maximum
b/c > 1 þ z þ n/m, where z is the number of mig-
group size, n, are much larger than 1. For more
rants accepted by a group during its lifetime, n is
information, see Traulsen and Nowak, 2006.
the group size, and m is the number of groups.
Group selection (or multilevel selection) is a
powerful mechanism for the evolution of coop- In all five theories, b is the benefit for the recipient
eration if there is a large number of relatively small and c the cost for the donor of an altruistic act.
CHAPTER 3
Single-species dynamics
Tim Coulson and H. Charles J. Godfray
What determines the densities of the different growth rates for different types of population, and
species of plants, animals, and micro-organisms explore how such calculations, even though they
with which we share the planet, why do their are based on the simplistic assumption of constant
numbers fluctuate and extinctions occur, and how demographic rates, can be very useful for a variety
do different species interact to determine each of problems in applied population biology.
other’s abundance? These are some of the ques- The fact that populations persist over appreci-
tions addressed by the science of ecological able periods of time inescapably means that
population dynamics, the subject that underpins demographic rates—births and deaths, immigra-
all the chapters in this book. In this chapter we tion and emigration—do not remain constant.
introduce some of the basic principles of the In fact, population persistence in the long term
subject by concentrating on the dynamics of single- requires that as populations increase in density the
species systems. These are species whose popula- death rate must rise relative to the birth rate and
tion biology can be studied without also explicitly eventually exceed it. In real populations, such
including the dynamics of other species in the demographic rates are what ecologists call density-
community. The chief justification for this brutal dependent and mathematicians nonlinear,
abstraction is that it allows many of the underlying whereas an engineer might talk about negative
processes to be described simply and more clearly. feedback. It is this nonlinearity that can give rise to
Moreover, arguments based on the analysis of a stable equilibrium, the population density at
single-species population dynamics are often sur- which birth rates precisely equal death rates. But a
prisingly useful in understanding real populations, far more diverse menagerie of dynamic behaviours
especially those in relatively simple environments is possible; the population may not settle on a
such as agro-ecosystems. stable equilibrium but show persistent cycles.
At the core of population dynamics is a simple Stranger still, these cycles may not be regular but
truism: the density or numbers of individuals in a complex and unpredictable in detail—they may
closed population is increased by births, and show mathematical chaos. The possibility of
decreased by deaths. If the population is not closed chaotic dynamics in simple populations was first
then we need also to include immigration and appreciated in the 1970s, and ecological problems
emigration in our calculation. A population in were very significant in the development of this
which births exceed deaths will tend to increase new field of mathematics. The second section of
and one where the reverse is true will tend to this chapter explores the consequences of deter-
decrease. But more significant is the mode of ministic density-dependent demographic rates,
change. If birth and death rates remain constant and explores chaos in ecology.
then the consequent increase or decrease in In the last sentence, by deterministic we mean
population numbers occurs exponentially—popu- that demographic rates are constant or simply
lation dynamics occurs on a geometric rather than determined by density, and do not also vary by
an arithmetic scale. In the first section of this chance. Of course, all real populations are subject
chapter we describe the calculation of exponential to random effects. When the average birth rate is
17
two offspring per year, some individuals will have reproduce immediately). Let the rate at which
fewer or more offspring; and in some years, or in individuals produce female offspring be b (we
some sites within the species’ range, the average assume for now that males have no effect on
may be slightly more or slightly less. More radi- population growth rate, something that is true for
cally, the birth rate may be two, year on year, most but not all organisms) and the rate at which
except for the time the meteorite hit and no-one they die be d. Define the difference between these
reproduced. The last 10 years has seen notable two rates as r ¼ b d. The population increases if
advances in the study of populations that take into r > 0 and decreases if r < 0. Moreover, if the current
account stochastic effects, and these are the subject population size is N0 then the population size t
of our third section. time units into the future is Nt ¼ N0 exp[rt]. The
population increases or decreases at a rate deter-
mined by the power of r. Note that in this simple
3.1 The rate of population growth
model, to predict population growth rates we do
All subjects need their founding myths, with not need to know birth and death rates separately,
appropriate heroes, and while physics has the just their net difference.
giants Newton and Einstein, and evolution the Not all species reproduce continuously. Con-
peerless Darwin, students of population dynamics sider a population of an animal or plant with
are stuck with the far less appetising Thomas discrete generations that produces l female off-
Malthus. Malthus was not the first person to spring before dying. It is straightforward to see
appreciate the geometric nature of population that if population size is now N0 then t generations
growth but he was the first clearly to work through in the future it will be Nt ¼ N0 lt, which can be
its consequences. In his famous pamphlet An Essay written Nt ¼ N0 exp[ln(l) t], the latter expression
on the Principle of Population of 1798, he vividly emphasizing the similarity with the continuous
illustrated the power of geometric growth in terms case, with ln(l) replacing r.
that mirror the modern clichés that if population The parameter r (or ln(l)) is the intrinsic growth
growth is unchecked it would take only a few rate of the population; it allows us to project
years for the total numbers of aphid/cod/elephant population numbers into the future. Of course we
or your favourite animal to weigh more than the do not believe r will stay constant forever—a
Earth (Malthus, 1798). It was the power of this projection should not be confused with a fore-
argument that so influenced Darwin—such great cast—but it tells us something about what will
potential fecundity must be balanced by great happen in the short term, given the current birth
mortality, and any heritable trait that favoured one and death rates. This can be a very important
individual over another would increase in fre- management tool. Suppose for example one is
quency ineluctably. In contrast, the message that trying to assess the potential vulnerability of a
Malthus, an upper-class vicar, drew from his own series of populations of an endangered species.
insight was the need to do something about the Calculating their different population growth
irresponsibly fecund lower classes (as well as rates will not give you a complete answer to this
about other problems such as women and the question, but it will provide an important
French). Type Malthus into Google and you find clue to their different vulnerabilities. Estimations
him a hero to an unpleasant consortium of mod- of population growth rates for more complic-
ern-day social engineers. ated population structures (see below) lie at the
But despite its shady origin, the rate of expo- heart of population viability analysis, a frequently
nential population growth based on current used tool in conservation biology. Epidemiology
demographic rates is an immensely useful quan- provides a rather different example of the impor-
tity. Consider first a simple, unstructured popula- tance of population growth rate. Consider a
tion; by unstructured we mean that birth and population of susceptible hosts exposed to a small
death rates are identical across individuals (clearly number of infectious individuals. From the point of
an approximation, as a newborn individual cannot view of the disease, births consist of new infections
SINGLE-SPECIES DYNAMICS 19
and deaths occur when the host either recovers or while the probabilities of surviving until the new
actually dies. The disease will only spread if season form the lower subdiagonal. The matrix A
r ¼ b d > 0, where b and d are the rates of disease is an example of a population-projection matrix,
‘births’ and ‘deaths’. In the epidemiological litera- and this particular form, where the population is
ture this condition is normally stated as exp(r) ¼ structured by age, is called a Leslie matrix (Leslie,
R0 > 1, which has the simple interpretation that for 1945).
spread to occur every initial infection must leave at The simplest way to explore the growth rate of a
least one secondary infection. As Grenfell and population described by eqn 3.1 is to iterate it on a
Keeling (Chapter 10 in this volume) discuss in more computer, an option not available to the origina-
detail, calculation of R0, usually called the basic tors of these techniques in the 1940s. But there are
reproductive ratio by ecologists or, more correctly, some important mathematical results that allow
the basic reproductive number (it is dimensionless) much greater insight into the population growth
by epidemiologists, lies at the heart of much human process. We do not have the space to derive these
and animal health population analysis. results or explain them in detail, but attempt to
give some flavour of their elegance and import-
ance.
3.1.1 Structured populations The matrix A includes all the information we
need to know about the population’s demographic
The assumption that all populations are made up
parameters (Caswell, 1989, 2001). From this matrix,
of identical individuals with the same demo-
a polynomial equation in an arbitrary variable (say
graphic rates is clearly a gross oversimplification.
Z) can be derived. The order of the polynomial is
How can population growth rates be calculated in
determined by the number of age classes. If there
more complex structured populations?
are five age classes than the equation will be of
We introduce this topic by considering a popu-
order five (terms up to Z5) and if there are 20 age
lation that has discrete breeding seasons so that it
classes then there will be terms up to Z20. Just as
makes sense to census it once a year. We also
the familiar quadratic equation (order two) has
suppose that the population is age-structured:
two roots (values of Z for which the equation
demographic rates vary with age but are constant
equals 0) then these larger polynomials of order x
within an age class. To describe population num-
have exactly x roots, though unlike the quadratic
bers at time t we now need to write down a vector,
they can only be calculated numerically (except for
nðtÞ ¼ fn1 ; n2 ; . . . ; nx gðtÞ, where ni is the number
some special cases). A collection of mathematical
or density of individuals in their ith year at time t
results called the Perron–Frobenius theorem tells
(and x is the oldest age class). To explore how
us that for Leslie matrices (with some minor
population numbers change over time we need to
exceptions that we will return to) there will always
know the probability that an individual of age i
be one root that is larger than all the others.
will survive to the next year (pi) and the number of
Moreover, this root, which is a complicated func-
offspring it produces each year (fi). Then
tion of the different elements of the matrix A,
0 1 0 10 1
n1 f1 f2 f3 fx n1 represents the long-term growth rate of the
B n2 C B p1 0 0 0 C B C
B C B CB n2 C population. In matrix theory the roots are called
B n3 C B 0 C B C
B Cðtþ1Þ ¼ B 0 p2 0 CB n3 CðtÞ eigenvalues and calculation of the largest root, the
B .. C B .. . . . . .. CB .. C
@ . A @ . . . . A@ . A dominant eigenvalue, provides the asymptotic
nx 0 0 px1 0 nx population projection that we require.
ð3:1Þ This powerful result tells us that whatever the
initial distribution of individuals across age classes
which can be written more succinctly n(t þ 1) ¼ the population will eventually grow or decline at a
A n(t). Note that the numbers in the youngest age rate set by the dominant eigenvalue (this inde-
class are given by the numbers in each age class pendence of starting conditions is called ergodi-
the season before multiplied by their fecundity, city). Another ergodic property of these population
models is that the proportions of individuals in plants, which we shall index as stages 1, 2, and 3
different age classes assume constant values (the respectively. The projection matrix for such a
stable age distribution) irrespective of starting species might look like this:
values. These values can be calculated directly
0 1
from the projection matrix: associated with the 0 b2 b3
dominant eigenvalue is a pair of vectors of length x @ a12 a22 a32 A ð3:2Þ
with each element corresponding to an age class. a13 a23 a33
These are the dominant eigenvectors and the rela-
tive magnitude of the elements of one gives us the The top row reflects stage-specific fecundity.
stable age distribution (we note in passing that the Between one time period and the next seedlings do
other dominant eigenvector provides a measure of not reproduce but small plants produce b2 and
Fisher’s reproductive value for each age class). large plants b3 offspring that survive to the seed-
Whereas the population growth rate is an ling stage. The subdiagonal a12 and a23 tell us the
important management tool, applied ecologists are probability that seedlings become small plants,
often also interested in how births and deaths at and small plants become large plants, just as in the
different age classes contribute to the overall pro- Leslie matrix. But we now have further transitions
jection. A conservation biologist may need to know (or lack of transitions): a22 and a33 are the prob-
whether to prioritize efforts on old or young abilities that small and large plants remain the
individuals, while a game manager might need to same size (we assume this option is not open to
know the consequences of allowing animals of seedlings), while a13 allows some plants, perhaps
different ages to be shot. The marginal effects on in extremely favourable microhabitats, to transit
the population growth rate of changing the birth or from seedlings to large plants in one go, and a32
death rate at each age can be calculated, and such allows those unfortunate individuals that encoun-
relationships, which can be defined in different ter a rabbit actually to decrease in size.
ways depending on precisely for what they are All the results that apply to the Leslie matrix
required, are called sensitivities or elasticities, terms transfer to these more complicated structured
borrowed from equivalent problems in economics. populations: we can calculate projected population
The projection matrix can also be used to pro- growth rates, and what we should now call the
vide information on the speed at which the stable stage distribution. Analysis of matrix models
asymptotic growth rate and stable age distribu- of stage-structured populations has proven to be
tions are attained. The key quantity here is called extremely valuable in many fields of ecology, but
the damping ratio, which is defined as the ratio perhaps especially so in plant ecology. However,
of the largest eigenvalue of the projection matrix to there can sometimes be difficulties in placing
the second largest eigenvalue (Caswell, 2001). This individuals that vary in a continuous variable such
root may be a complex number, and this provides as size into the discrete classes that are required of
information about whether there is a smooth or the matrix formulation, and also in choosing the
oscillatory approach to the long-term growth rate appropriate time step for analysis (Easterling et al.,
(Fox and Gurevitch, 2000). 2000). The decision need not be entirely arbitrary,
Classifying individuals in populations by their as there are some theoretical results that suggest
age is perhaps the most common way to relax the optimal choices.
assumption that everyone shares the same demo- We mentioned that there were a few exceptions
graphic parameters. But the matrix formulation is to the simple application of the ergodic results
much more powerful than this and populations of matrix theory, although they are easily dealt
can be classified by size, life-history stage, sex, with by straightforward extensions. These include
geographical location, or essentially any other populations with post-reproductive age classes or
variable. To illustrate this consider a plant whose with single reproductive age classes. For example,
individuals can be classified in one of three life- consider the cicada populations in North America
history stages: seedlings, small plants, and large (Magicicada sp.) that take precisely 17 years to reach
maturity. As long as there is absolutely no mixing we need to know what size classes in the last year
of cohorts then each year class will increase or might give rise to y-sized individuals this year.
decrease in density as determined by the long-run This contribution can occur in two ways: first,
population growth rate. But the ratio of initial individuals of size x may give birth to individuals
frequencies in year 1, 2, . . . , 17 remains the same; that are size y at the next census point (call this
they do not converge on a stable age distribution. b(y, x)); second, individuals of size x may avoid
In America adult Magicicada are abundant only death and grow to become size y (call this p(y, x)).
once in 17 years, and this pattern could be The integral on the right hand side of eqn 3.3
explained by initial conditions (for example all simply sums these contributions to the current y
year classes are wiped out except one) plus lack of class over all possible size classes last year (indexed
cohort mixing. In fact this is highly unlikely, the by x). Analyses of equations such as eqn 3.3 pro-
power of even minor cohort mixing to destroy the duce very similar results to the matrix formulation;
imprint of starting values is so strong that ecolo- most biologically realistic populations increase or
gists have universally rejected this hypothesis and decrease at a growth rate and with an age dis-
sought active processes to maintain the synchro- tribution that is independent of initial starting
nized cohorts. Because the length of the life cycle is values (Easterling et al., 2000).
a prime number (and 11- and 13-year cicada Finally, we can study populations structured by
populations are also found) the classical explana- a continuous variable in continuous time using the
tion is that it is a means of escaping predation, as it famous McKendrick–von Förster equation:
is hard for predator populations with life cycles
qnðy; tÞ qnðy; tÞ
that are not exact divisors of the cicada life cycle to þ ¼ mðyÞnðy; tÞ ð3:4Þ
qt qy
increase in density (Hoppensteadt and Keller,
1976). However, a recent study has suggested that Again n(y,t) represents the density of individuals
although predator satiation and/or competition of size y at time t. This expression simply states
among nymphs can explain why the dynamics of that the numbers in a cohort of individuals born
Magicicada are periodic, they do not explain why at the same time decline with age and time as
the period is a prime number of years (Lehmann- mortality (m(y), which is likely to be age-specific)
Ziebarth et al., 2005). The authors speculate that a inexorably whittles them away. To complete the
physiological or genetic mechanism or constraint model we need a birth process which is introduced
might be responsible. as a boundary condition:
We have dwelt at length on the matrix for- Z 1
mulation partly because it is relatively straight- nð0; tÞ ¼ bðyÞnðy; tÞ dt ð3:5Þ
forward to explain, but we finish this section by 0
briefly describing two alternative approaches.
This states that the numbers of individual of age 0
Suppose first that we are content to census our
(i.e. newborns) are simply the numbers of current
population at discrete time intervals (perhaps
individuals in the population multiplied by their
yearly) but that we are unhappy to shoe-horn
age-specific birth rates (b(y)). These equations can
individuals into discrete classes of size (or other
be generalized to populations structured by size
classifying variable). Instead we want to work with
and by other variables (Wood, 1994).
the more natural continuous size distribution. This
Again, as you would expect from a change of
leads naturally to an integral projection model of
formalism rather than a change in biology, the
the form
behaviour of populations described by this model is
Z 1 similar to those we have discussed above. Most
nðy; tþ1Þ ¼ ½bðy; xÞ þ pðy; xÞnðx; tÞdx ð3:3Þ reasonable assumptions give ergodic population
x¼0
growth and age distributions. Although the
Here n(y, t þ 1) represents the density of indivi- McKendrick–von Förster equation has been
duals of size y at time t þ 1. To calculate this value used extensively in many branches of population
biology, it is notoriously difficult to work with, either a decline in birth rates or an increase in
both analytically and numerically, and it is gen- death rates. An illustration of this is given in
erally a less popular approach than the other Figure 3.2. At low densities populations increase
two described here. from generation to generation but as density
increases the rate of increase slows and then
reverses. At one particular density each individual
3.2 Density dependence
female precisely replaces herself and the popula-
The value of projections is that they tell us some- tion is at equilibrium. In these simple diagrams the
thing about current populations; to move from equilibrium, marked by a star, is easily found as
projections to forecasts we need to know not only the density where the population growth curve
the current values of demographic parameters, crosses the 45 line.
but how they may change in the future. In this Simple though they are, the diagrams can be
section we concentrate on how birth and mortality used to look at dynamic trajectories as well as
rates may be affected by changes in population equilibria through a geometrical trick called
density. cobwebbing. Suppose that the current population
We shall begin to explore this topic using a density is represented by the point 0 in Figure 3.3.
simple, unstructured population model in discrete The density in the next generation is read off the
time which we sample every generation. We will population growth curve at point 1. To find the
plot the population density in the next generation density in the following generation the trick is to
as a function of that in the current generation. draw a line that forms a right angle with the
Figure 3.1 shows the simplest case where popula- (dashed) 45 line and then intercepts the popula-
tion growth rate is independent of density. The tion growth curve (the line 1 ! 2 in Figure 3.3).
dashed line at 45 represents the situation where This process can be repeated indefinitely and by
population density remains the same from gen- noting the population densities given by the series
eration to generation. The solid line illustrates (1, 2, 3, 4, . . . ) we obtain the population trajectory.
density-independent growth, the slope of the line In the case of a population whose biology is
being l, the discrete-time rate of population summarized by Figures 3.2 and 3.3, the dynamics
growth. are a damped approach to a stable equilibrium.
Inevitably, as populations grow they will come In Figure 3.3 there is a single equilibrium point.
to exceed their resource base and this will result in This is also called a globally stable equilibrium
Population density time t + 1
Population density time t +1
Population density time t

Figure 3.2 Mild density-dependent population growth. The star
Figure 3.1 Density-independent population growth. indicates an equilibrium.
Population density time t+ 1

Population density time t +1
T
1
Population density time t Population density time t
Figure 3.3 Cobwebbing with mild density-dependent population Figure 3.4 The Allee effect: populations that start off above the
growth. threshold T move towards the upper equilibrium; those that start
off or fall below the threshold become extinct (a population size of 0
is a stable equilibrium).
because whatever the initial population density the

population trajectory will inevitably home in on B
the same equilibrium point (try cobwebbing from
different starting points). But depending on the

natural history of the population in question more
than one equilibrium can occur. Consider a
population whose growth curve can be repre- T
sented by Figure 3.4. Below the threshold T the
population in the next generation is smaller than in
the current generation. This might occur if indivi-
duals in low-density populations find it difficult to
A
locate mates (or get pollinated), or perhaps in
group-hunting or colonial species if cooperation
breaks down when there only a few individuals
present. Any population that drops below T will Population density time t
continue to fall in density until it becomes extinct.
Figure 3.5 Alternative stable states: there are two non-zero
Moreover, the population will not be able to equilibria (marked by stars); which equilibrium the population moves
increase from low densities. There are thus two to depends on whether it starts at above or below the threshold T.
locally stable equilibria, and which one is attained
depends on whether the initial population density
is above or below the threshold T. The presence of population starts above T it moves to the local
a minimum population density below which the stable equilibrium B and if below to A. But the
population goes extinct is called an Allee effect and lower equilibrium is now not at 0; when very rare
for obvious reasons is something that conservation the population is still able to increase in numbers.
biologists are very concerned about. Note that the population in Figure 3.5 actually has
A species’ population dynamics may include four equilibria: at densities of 0, A, T, and B. But
more than one non-zero stable equilibria. Consider while A and B are stable, 0 and T are unstable: a
Figure 3.5; again we have a threshold (T) which population whose density is precisely 0 or T will
determines the final state of the system. If the remain at that density forever, but the slightest
C between scramble competition where resources are

divided equally among all the individuals in the
population, and contest competition where some
individuals get the resources they require, and

others get none. (Of course, many populations will
show intermediate patterns.) At the population
level, pure contest competition will give growth
S curves that resemble C in Figure 3.6 and scramble
competition curves more like S. With C, as den-
sities get high, a fixed number of individuals sur-
vive or reproduce to form the next generation;
with S, at high densities everyone suffers and
population numbers plummet.
We can explore the dynamic consequences of
moving from contest to scramble competition
Figure 3.6 Possible population-level consequences of contest using cobwebbing. In Figure 3.3 we had a fairly
(C ) and scramble (S) competition. contest form of competition and this resulted in a
smooth approach to a stable equilibrium. In the
three panels of Figure 3.7 we progressively
perturbation will cause the population to move increase the scramble component. In the first panel
away to either A and B. we still have a stable equilibrium but now the
Are there examples of multiple non-zero equili- approach to the equilibrium is not smooth but
bria? There are, but to be strictly accurate most through damped oscillations. In the second panel,
involve multispecies interactions that require a with more scramble, we no longer get a stable
more complex approach than the simple, equilibrium but instead a stable two-point cycle.
unstructured single-species models we are dis- Finally, in the third panel we get curious dynam-
cussing here. A classic example is the spruce ics: oscillations that never exactly repeat them-
budworm moth, Choristoneura fumiferana, which is selves. This is dynamical chaos, which we shall
hypothesized normally to be controlled at a rela- return to again in the following subsection.
tively low population density by a suite of spe- We have derived this series of population-
cialist invertebrate predators (Peterman et al., dynamic behaviours by discussing the spectrum of
1979). If, however, the population is perturbed types of competition from contest to scramble, but
such that control breaks down, the moth increases what is important is not the underlying mechan-
in density to a second equilibrium which is set by ism but the shape of the population growth curve,
bird predation (these predators ignore spruce however this comes about. In particular, the angle
budworm unless it becomes very common). A of the growth curve at the equilibrium where it
further perturbation is needed to switch it back to intersects the 45 line is informative. If the angle is
the low-density equilibrium. The density marked T more than 45 (i.e. the growth curve is below the
is thus a threshold or tipping point separating two 45 line to the left of the equilibrium) we have an
locally stable equilibria. unstable equilibrium, as at T in Figure 3.5. If the
Competition for food and other resources is angle is between 45 and 0 as in Figure 3.3 we have
without doubt the most important process deter- a smooth approach to a stable equilibrium;
mining the dynamics of animals and plants whose between 0 and 45 as in Figure 3.7a we have a
dynamics can be said to be single-species. The damped oscillatory approach to a stable equili-
curves in Figures 3.2–3.5 are population-level brium; and if less than 45 then the equilibrium
emergent phenomena based on complex interac- is unstable but persistent cycles or chaos may
tions occurring at the level of the individual. occur. In analysing population models a common
For example, ecologists have long distinguished practice is to solve for the equilibrium values and
(a) (b)

Population density time t Population density time t
(c)
Figure 3.7 Increasing the nonlinearity of the response of population growth to density can lead from the monotonic approach to
equilibrium in Figure 3.3 to an oscillatory approach (a) and then a two-point cycle (b) and finally chaos (c). Examples of cobwebbing; the
underlying population model is the Ricker equation, xtþ1 ¼ xt expðrð1 xt ÞÞ where x is population density (scaled to be 1 at carrying capacity)
and r is fecundity.
then to conduct a stability analysis to determine the are their multidimensional equivalents can be
equilibrium’s properties. In some cases it is pos- derived (May, 1972).
sible to show that an equilibrium such as that in In this section we have focused on a simple,
Figure 3.3 is globally stable, but such analyses are unstructured population in discrete time. We
mathematically challenging and often there is no could alternatively have studied a population in
global equilibrium. Instead, a local stability ana- continuous time whose dynamics are described by
lysis is performed. For populations described by the equation
the type of growth curve shown in Figures 3.2–3.7
dN
we have been able to derive the local stability ¼ ½bðNÞ dðNÞN ¼ rðNÞN ð3:6Þ
dt
criteria using intuition helped by a little cobweb-
bing. Moving to more complex structured or where birth rate (b), death rate (d), and net popu-
multispecies communities these simple geome- lation growth (r) are now all functions of popula-
trical insights are lost, but algebraic conditions that tion size N. Plotting r(N) against N would similarly
tell us much about the dynamics of the system. generations. Many of these systems show popula-
There is, however, a difference between population cycles with periods shorter than those pre-
tions described by eqn 3.6 and those that we stu- dicted by unstructured models (e.g. Figure 3.7).
died in Figures 3.1–3.7. In eqn 3.6 the effects of The details differ with the natural history of the
population density act instantaneously on popu- different systems but a common pattern is for an
lation growth rate. In our discrete-time examples older cohort of individuals to reduce the numbers
there is an implicit time lag: the level of competi- in a younger cohort by out-competing them for
tion experienced by individuals in the current food or through cannibalism. When the depleted
generation is set by interactions that occurred in younger cohort grow old enough to be dominant
the last generation. Such time lags tend to be competitors or cannibals themselves there are not
destabilizing as they delay the onset of a reduction enough of them to reduce significantly the next
in population growth rates as densities climb, and cohort coming through. This means that the next
make it more likely that any equilibrium is over- group of individuals to mature into the older
shot. Indeed, it is mathematically impossible for a cohort are very numerous and decimate the cur-
population governed by eqn 3.6 to show chaos. We rent younger cohort, and the cycle begins again.
should stress that it is not the difference between
continuous- and discrete-time formalisms that lies
3.2.1 Chaos
behind the contrasting stability, but the presence of
the time lag. Indeed, in continuous time we can get The pioneers of modern mathematical dynamics,
exactly the same dynamics by explicitly making particularly Poincaré at the end of the nineteenth
net population growth rates a function of previous century, realized that the behaviour of highly
population densities, nonlinear systems could be very odd, but in the
absence of computers to help visualize their
dNt dynamics, progress on understanding what was
¼ rðNtt ÞNt ð3:7Þ happening was very slow. When computers began
dt
to become available in the 1960s workers in fields
where t is a time lag of approximately one gen- such as meteorology and ecology were able to see
eration. the complex dynamics produced by beguilingly
Structured population models with density straightforward equations, and this led to a burst
dependence can be built using the same matrix, of interest in both pure and applied mathematics
integral equation, or partial differential equation that laid the foundations of the modern field of
approaches discussed in the section on density chaotic dynamics. In population ecology, the clas-
independence. Naturally they are more complex, sic paper is May’s Simple mathematical models with
and often with a greater potential for destabilizing very complicated dynamics (May, 1976a), which not
time lags. Relaxing the assumption that all indi- only introduced the notion of chaos to the field but
viduals are equal also leads to the possibility of showed that lurking underneath the seeming
more complicated types of interaction than are unpredictability of chaotic dynamics was con-
possible for unstructured populations. Competi- siderable order and pattern. We shall now explore
tion may be asymmetric, typically with smaller a population model of exactly the type that May
individuals suffering disproportionately at the analysed.
hands (or roots) of larger individuals. Moreover, Chaos has already been encountered in this
cannibalism is much more common in the animal chapter as the dynamics that emerge in a simple
kingdom than often realized, and when it occurs it discrete-time population model as the population
is nearly always size-related, with older larger growth curve (or map) becomes sufficiently non-
individuals consuming their smaller conspecifics. linear (the ‘humpiness’ of the curves in Figure 3.7).
Such age-specific interactions have been studied Let us now specify a family of curves that can give
in detail, particularly in insect systems that can rise to the maps in Figures 3.3 and 3.7. For reasons
be maintained in the laboratory for multiple that will be explained in a few paragraphs it does
not particularly matter which family we chose, and cycle: the population oscillates for ever between
we plump for the Ricker equation as it is commonly two densities, one greater and one less than the
used in applied population biology, particularly now unstable equilibrium n ¼ 1. In Figure 3.8 this
for fisheries (Ricker, 1954). appears as two points. The value of this repre-
sentation now becomes clear because instead of
ntþ1 ¼ nt exp½rð1 nt Þ ð3:8Þ
having to try to compare a large number of cobweb
Here nt is population density (scaled to equal 1 at diagrams we can see at one glance how the
equilibrium). When rare the population increases cycles appear at r ¼ 2 and then increase in ampli-
each generation by a factor exp[r] but as densities tude as r gets bigger. The change of behaviour at
approach 1 the increase slows and above 1 it r ¼ 2 is called for obvious reasons a bifurcation,
reverses. If r is high there is the potential for the and the representation itself is a bifurcation dia-
population to overshoot the equilibrium. gram. We can also see that at r ¼ 2.5 a second
We want to picture the dynamics of the whole bifurcation occurs to give a four-point cycle, and
system for different values of the sole adjustable then further bifurcations at increasingly smaller
parameter r. To do this, imagine iterating the intervals of r until a limit is reached. ‘‘What happens
equation by cobwebbing as in Figures 3.3 and 3.7 at the point of accumulation [the limit]?’’ is what May
and then throwing away all the transient dynamics, scrawled on a blackboard in the Theoretical Phy-
perhaps the first 50 generations. For Figure 3.3 sics Department at Sydney University in the early
(corresponding to a value of r ¼ 1) the non-transient 1970s.
dynamics would not be very interesting: it would May showed that what happens is chaos. As the
simply be a population at stable equilibrium, in this third panel in Figure 3.7 illustrates, the trajectory
case n ¼ 1. In Figure 3.8 we plot r along the x axis never converges on a simple cycle but fluctuates
and the non-transient dynamics on the y axis; for aperiodically around very many values of n, never
r ¼ 1 there is a single point at n ¼ 1. The dynamics repeating itself. This is represented in Figure 3.8 by
of the population described by the first panel in a vertical line containing numerous, in fact an
Figure 3.7 (r ¼ 1.9) differ only in their transient infinite number of, points. Cobwebbing can also be
behaviour and so it too would be represented by a used to demonstrate a cardinal property of chaos:
single point at n ¼ 1. Indeed, for the Ricker equa- namely sensitivity to initial conditions. Start two
tion a stable equilibrium occurs for all persistent trajectories very close together and sooner or later
populations with r < 2, which gives the straight line they will diverge. This is not due to a lack of
at n ¼ 1 in the left-hand part of Figure 3.8. computing power: no matter how close the two
The persistent dynamics depicted by the middle initial values they will come to diverge. More
panel of Figure 3.7 (r ¼ 2.3) are a two-point limit accurate estimation of initial values, so that the
2.5
2
Population density
1.5
0.5
0
1 1.5 2 2.5 3
Fecundity parameter r Figure 3.8 The bifurcation diagram for the Ricker
population model (see the legend of Figure 3.7).
measured value is close to the ‘true’ value, can will see a complex pattern of bifurcations, aper-
delay the divergence, but not prevent it, and this iodic and period trajectories; chose part of this
means that there is an absolute limit to our ability picture and enlarge yet again and the same pat-
to predict into the future the behaviour of chaotic terns appear in miniature, and so on ad infinitum.
systems. The beauty of bifurcation diagrams is fragile: add
The dynamics of a system can be described with a little stochastic noise—inescapable in real biolo-
a quantity known as the Lyapunov exponent gical systems—and their more rococo patterns dis-
(named after a Russian mathematician whose appear. However, the extreme sensitivity to initial
name is also transliterated Liapunov or Ljapunov). conditions, the signature of chaos, remains. So
The Lyapunov exponent describes the rate of while it is not mathematically true that the Ricker
separation of infinitesimally close trajectories; a model predicts chaos for all r > 2.69 it might as
positive value means that the trajectories diverge well be for any biological purposes. Another bio-
exponentially, and this extreme sensitivity to initial logically relevant property of chaos is also shown in
conditions is the hallmark of chaos. Algorithms Figure 3.8. Although precise prediction is not
have been derived to estimate Lyapunov expo- possible the different population trajectories are
nents directly from time series (Wolf et al., 1985) bounded, that is they cannot become arbitrarily
and have proved very valuable, especially in the large or small. A pure random walk would not be
physical sciences where relatively long time series bounded (except of course by n ¼ 0). Indeed, it is
are typically easier to obtain. sometimes possible to calculate the probability
Bifurcation diagrams are beautiful objects that distribution of different population states. Depend-
contain a wealth of mathematical detail. They have ing on the system this may be valuable information
quite literally been the subject of tens or possibly for ecologists and population managers.
hundreds of mathematics PhD theses. Our focus Chaos is not just a property of discrete-time
here is on their relevance to biology and we have systems and chaos in continuous time systems has
space to mention only a very few more technical also been extensively studied. Consider the non-
results. First, May and others showed that the transient behaviour of a continuous system. If the
patterns in Figure 3.8 do not just apply to the system is at equilibrium this will be a simple point
Ricker equation but to a very broad class of models but if there are persistent cycles or chaos then it
that all show the same transition through period- will be a continuous line. For single-species
doubling from order to chaos (May, 1973a, 1974c; populations this line can be plotted in a space
Li and Yorke, 1975). There is a limited number of where the coordinates are population densities
routes to chaos and one can derive general results now and at times in the past. For example, on a
that apply to very many systems. For example, the three-dimensional graph the coordinates might be
ratio of the interval of r values in which two-point densities now, 1 month ago, and 2 months ago. In
cycles are found and in which four-point cycles are this space a cycle will be a closed loop while a
found is 4.6692. In fact the same ratio is found for chaotic trajectory will be an object such as that on
every adjacent interval (four-point/eight-point, the left of Figure 3.9. This object looks like a
etc.) not only for the Ricker equation but for every twisted diaphanous sheet and is a fractal: succes-
map that shows this type of transition from order sive magnifications of parts of the sheet show the
to chaos (Feigenbaum, 1978). Second, if you look same self-similar pattern. One point to note is that
closely at the bifurcation diagram to the right of chaos occurs in simple (ordinary) differential
the accumulation point you see that the region of equations only for systems of three or more vari-
chaos contains intervals of simpler dynamics, ables: the dynamics of a two variable-system can
including period-three cycles that undergo their be described in a two-dimensional space which
own transition back into chaos. In fact there is an does not allow for the twisting and mixing of tra-
infinite number of narrow, periodic windows. jectories that are the hallmarks of chaos.
Finally, the bifurcation diagram has fractal struc- There is a close link between chaos and fractals.
ture: enlarge part of the region of chaos and you Objects that represent the non-transient behaviour
of a dynamic system are called attractors (because in Figure 3.9. If the position along the section is
trajectories originating elsewhere in state space are treated as a variable, and if the position in the
attracted to them). In continuous time, points current traverse is plotted against that in the pre-
(stable equilibria) and closed loops (cycles) are vious, one arrives at a map exactly equivalent to
examples of normal attractors whereas fractal the chaotic Ricker map discussed above. Now,
objects such as that in Figure 3.9 are termed strange however, the r parameter is not simply a measure
attractors. All chaotic systems are governed by of single-species fecundity, but a more complex
strange attractors and, as we shall return to amalgam of the life histories of all species or
shortly, determining that a system’s attractor is development stages that influence the dynamics.
fractal is one way of identifying chaos in nature. How might one seek to decide whether natural
The attractor in Figure 3.9 also provides an insight populations are chaotic? Typically this has to be
into why chaos is always associated with extreme done from time-series data, which at least in
sensitivity to initial conditions. The right-hand comparison with data from the physical sciences
panel in Figure 3.9 is a cartoon to illustrate the are inevitably of relatively short duration. There
evolution of a set of initially very similar trajec- are two broad approaches. The first is to try to fit a
tories: the bundle marked 1, which should be flexible population model to the time-series data
imagined as lying flat on the horizontal surface of and then to determine by iterating the model
the attractor in front of the line X. Flow on the whether the dynamics are chaotic. The second is to
attractor occurs in the counter-clockwise direction try directly to reconstruct the attractor governing
and sets of points are first stretched (2, 3) and then the system and determine whether it is fractal.
folded (4, 5). If you imagine this occurring Both approaches are helped by a very important
numerous times it is easy to see how trajectories theorem (Takens, 1981) that states that the attractor
that start off near each other quickly become of a multi-species or complex single-species inter-
separated. The degree of stretching in a system is action can always be reconstructed from single-
quantified by the Lyapunov exponent. variable time-series data in a space made up of
Chaos in continuous- and discrete-time systems a sufficient number of time-lagged dimensions
is intimately related. Consider the section X (i.e. the coordinates are densities at time t, t t,
(called a Poincaré section) through the attractor t 2t . . . where t is a lag). The major proviso is
4
5
3
X
1 2
Figure 3.9 Chaos in continuous time. The object on the left is a strange attractor describing the flow of trajectories of a continuous-time
system in three-dimensional space (the Rössler attractor). X is a Poincaré section discussed in the text. The cartoon on the right describes
how bunches of nearby trajectories become stretched and folded as they move around the attractor. See text for further details.
that you have to have sufficient data, which in exponent. This method has since found wide
practice is usually a very demanding requirement. application beyond biology in econometrics.
The first attempt to fit models to data did not Model-based approaches have also enjoyed
use time series but life-history data on fecundity renewed attention. One strand has sought to
and density-dependent mortality. Hassell et al. develop more accurate mechanistic population
(1976) fitted a two-parameter model to data from models, capitalizing on both the more powerful
24 species of insects with reasonably discrete computing tools now available and statistical
generations and concluded that the vast majority advances in extracting parameter values from
had stable dynamics, indeed not even showing an data. A different strand, with similarities to Sugi-
oscillatory return to equilibrium. Although the hara and May’s approach, fits very flexible non-
authors were at pains to stress the provisional mechanistic population models to time-series data
nature of their conclusions, this paper had a very typically using response surfaces that are opti-
major impact, and to a certain extent inadvertently mized either by traditional least-squares methods
licensed ecologists to treat chaos as a theoretical or more exotic techniques such as thin-plate
curiosity for the next decade. splines or neural nets (Ellner and Turchin, 1995).
The next major attempt to search for chaos used The magnitude of the dominant Lyapunov expo-
model-free approaches and was spurred by the nent is calculated directly from the fitted model. It
growth of empirical chaos studies in the physical is still too early to judge the long-term value of
sciences (Schaffer, 1985; Schaffer and Kot, 1985a, these methods, although they have revealed a
1985b; Olsen and Schaffer, 1990). The basic idea number of systems with apparent chaotic dynam-
was to reconstruct the attractor by embedding the ics, in particularly involving human–disease and
time series in time-lagged coordinates and then predator–prey interactions.
either to take a Poincaré section and look for a one- For single-species interactions, the best examples
dimensional chaotic map, or to estimate the of possible chaos involve laboratory systems,
attractor dimension. In our daily lives we do not including Nicholson’s famous long-term blow fly
normally need tests to tell us whether an object is experiment. A very nice experimental example is
one-, two-, or three-dimensional but mathemati- the work of Costantino et al. (1997) on the flour
cians who often work in much higher dimensional beetle, Tribolium castaneum. Recall we mentioned
space have derived algorithms to estimate arbi- above that strong interactions between different
trary dimensionality. When these are applied to life-history stages can give rise to complex
fractal objects they return a non-integer dimension. dynamics. In Tribolium, adults and larvae canni-
A non-integer dimension implies a fractal and a balize eggs while adults also eat pupae. A popu-
fractal implies chaos. Though clearly worth trying, lation model showed that by varying a single
ultimately this research programme was defeated parameter (pupal mortality) the dynamics of the
by the quality of the data available. To quote system moved from stability to chaos and then to a
Schaffer (2000), ‘Only in the instance of recurrent three-point cycle. Figure 3.10 shows that experi-
outbreaks of measles in human populations, was mentally manipulating pupal mortality leads to
there sufficient data to justify our initial enthu- dynamics that look very like those predicted. It is
siasm’ and, he added, even here the argument true that this is a highly artificial system, yet it is
chiefly rested on the comparison of time-series an impressive demonstration that the dynamics of
data with the output of epidemiological models. these insects have been understood.
In the last 15 years, interest has grown again in
the challenge of detecting chaos from time series.
Sugihara and May (1990) developed a technique 3.3 Randomness
called nonlinear forecasting which measures the
3.3.1 Types of random effect
extent to which predictability decays with time. In
chaotic systems this occurs in a characteristic way Real animals, plants, and micro-organisms are
determined by the magnitude of the Lyapunov continually buffeted by the effects of random
300
(a)
150
0
Number of insects
300
(b)
150
0
300
Figure 3.10 Time series of the number of larval
(c) beetles in laboratory populations for different rates of
150 pupal mortality which were artificially manipulated.
Theoretical models predict that the population in
panel a should have a stable equilibrium, panel b
0 should be chaotic, and panel c should have a three-
point cycle. The experimental data show good
0 20 40 60 80 agreement with the predictions (after Costantino
Time in weeks et al., 1997).
processes and a critical question in population Let us return to the discrete-time model of a
biology is the extent to which insights gained from population with non-overlapping generations,
the analysis of deterministic models survive the Nt þ 1 ¼ Nt l, and for the sake of argument assume
insults thrown at them by stochastic nature. that the value of l is actually constant over time.
There are a variety of different ways in which But this does not mean that every single individual
random or stochastic effects can influence popu- in the population will produce exactly l female
lation dynamics (May, 1973a). Perhaps the most offspring. In the real world there will always be
straightforward is environmental stochasticity, some between-individual variation or demographic
where the value of a demographic parameter stochasticity. For example, consider a parasite that
changes over time. Recall the density-independent, searches randomly for hosts into which it lays a
discrete-time model Ntþ1 ¼ Ntl where l is the single egg; if the average parasite lays l female
annual population growth rate. This model impli- eggs then some will by chance discover more hosts
citly assumes that the value of l is constant, but in and some by chance fewer. This is a Poisson pro-
fact it will almost certainly vary from generation to cess where the variance is the same as the mean.
generation; we might better write the equation One can imagine other natural histories where
Ntþ 1 ¼ Nt lt to emphasize this fact. Note that the variance is much less than a Poisson process
environmental stochasticity affects the demo- (vertebrates that normally produce one offspring
graphic rates of all individuals in a population in a year) and others where the variance is much
the same way, and that this effect is independent greater (organisms living in a highly hetero-
of population size (Lande et al., 2003). Much geneous environment). Now suppose the popula-
research in identifying factors generating envir- tion is small: by chance all individuals in one
onmental stochasticity has focused on climate generation may experience low reproduction and
(Stenseth et al., 2002), although in principle any so the following year the population size would
other factor with unpredictable effects on popula- be significantly less than the expected Nt l. Of
tion parameters can contribute to this process. course, the probability of simultaneous episodes
of good or bad luck become progressively more population size is dominated by very rare, huge
unlikely in larger populations and hence demo- population sizes in the upper tail of the distribu-
graphic stochasticity is most important in small tion. In fact the modal population size, the popu-
populations. In many ways, its action is similar to lation size that will actually be observed in the
drift in population genetics. field, grows not at a rate determined by the simple
A further random process that is sometimes arithmetic mean, l, but the geometric mean
distinguished is catastrophic stochasticity: random ðl0 l1 l2 lt1 Þ1=t .
events that destroy the whole population irre- Several biologically interesting results follow
spective of its size or current demographic from this. First, as long as there is some variance in
parameters. We shall not discuss this type of ran- l the geometric mean will always be lower than
domness further here, although it is particularly the arithmetic mean: poor years have a greater
relevant to studies of metapopulations (see negative effect on population growth than the
Chapter 4 in this volume) and also in conservation positive effect of good years. Second, a single year
biology where populations may be wiped out by with zero net reproduction (l ¼ 0) renders the
human action that can at least be approximated as long-term growth rate 0. This makes intuitive
a random process. sense as the population goes extinct, but note that
this is not what a calculation based on the arith-
metic mean would suggest. Finally, recall that in
3.3.2 Density-independent populations the deterministic case persistence was very
Let us now see how stochasticity affects popula- straightforward: a population would increase if
tion growth rate and population projection. For l > 1 and decrease if l < 1. The situation is now
ease of explanation we shall stick to discrete-time more complicated: populations with geometric
models although the same principles apply to mean growth rates less than one will always ulti-
populations that reproduce in continuous time. mately go extinct, but some may persist for a long
Return once again to the model Nt þ 1 ¼ Nt lt where period of time if by luck they experience a chain of
the subscript to the population growth rate propitious years. Similarly, although populations
emphasizes that it varies between generations, with geometric growth rates greater than 1 will
specifically with mean l and variance sl. This is tend to persist, some will by bad luck go extinct. In
the way that randomness is most frequently dealt fact populations which will, on average, grow to
with in population models, and has been referred infinity also have a probability of extinction of 1 for
to as the equilibrium treatment of noise (Coulson et very long periods of time. This can be seen very
al., 2004). If we take logarithms then we can write simply: if E(Nt) ¼ T2, where t represents time and T
is the length of time since the simulation began,
X
t1
the probability of extinction can be written as
Log½Nt ¼ log½N0 þ log½lx : ð3:9Þ
x¼0
1 1/T. When T gets very large, the expected
population size tends to infinity and the prob-
If the values of l vary independently over time, ability of extinction tends to unity as 1/T approa-
then the right-hand term is the sum of indepen- ches 0. It is possible to calculate the distribution of
dent random variables, which the Central Limit persistence times of populations governed by dif-
Theorem tells us is asymptotically normally dis- ferent distributions of growth rates, and this may
tributed. This implies that population size itself is be helpful in population management.
lognormally distributed. There are some com- In many ways the population effects of demo-
plexities in calculating long-term population graphic stochasticity are similar to its environ-
growth rates in this case (Lewontin and Cohen, mental counterpart. It will increase the variance in
1969). An intuitive procedure might be to see l and so tend to reduce long-term growth rates,
how expected population size grows with time. A and increase the probability of extinction by bad
simple calculation reveals it increases exponen- luck. The major difference is that its effects become
tially at a rate determined by l. But the expected very weak as population size increases. Indeed, the
total variance in reproductive rates can be thought dynamics via two routes. First, stochasticity has a
of as the sum of two components, VE (environ- direct effect on the size and structure of the current
mental stochasticity) and VD/N (demographic population. Second, these changes influence the
stochasticity divided by population size). A rea- future trajectory of the population. This interaction
sonable rule of thumb is that demographic sto- between stochasticity and the deterministic skele-
chasticity can be ignored for populations with ton is sometimes referred to as the active treatment
more than 50 or so female breeders, though note of noise, and is currently an area of considerable
that the population size of large carnivores, even in interest in population biology research. Such
extensive nature reserves, can often be below this effects always reduce the tendency of the popula-
threshold. tion to reach a stable age distribution and, in
We stated above that we were assuming that anticipation of the next section, can also have
stochastic effects were uncorrelated over time. important consequences on population regulation
Often this will not be the case, especially for short- if the strength and action of density dependence is
lived organisms that might, for example, have also influenced by population structure.
several generations in a single summer. Quite
frequently there will be a positive correlation
3.3.3 Density-dependent populations
between the random component of population
growth rates in successive seasons (the term red In a real stochastic environment a population is
noise is sometimes used for these positively corre- highly unlikely to remain at the exact same equi-
lated random effects). The most important effect of librium value from one generation to the next. But
correlated stochasticity is to increase the severity of it is still reasonable to talk about an equilibrium if
poor breeding seasons that now tend to follow one populations above a certain value tend to decline
another. We note in passing that correlated red in numbers, and those below the same value tend
noise may lead to patterns in population dynamics to increase. Conceptually we can think of an
that may be very hard to distinguish from an equilibrium not as a fixed population density, but
underlying deterministic cause, especially in as a probability distribution that remains the same
structured populations. There can also be correla- over time and which determines the likelihood of
tions between environmental and demographic observing the population at any particular level of
stochasticity, in particular the effect of demo- abundance (Turchin, 2003). Of course, we should
graphic stochasticity on population growth may be also consider the possibility that a population,
higher in years when the consequences of envir- even one that tends to increase when rare, goes
onmental stochasticity are most severe, a clear extinct through a run of bad breeding seasons.
concern in conservation biology. More generally, stochastic effects can cause a
The arguments above apply also to structured population to shift from one type of dynamic
populations, though with some complications. behaviour to another. Figure 3.5 depicted the
First, there is no longer a simple relationship dynamics of a species with two locally stable equi-
between arithmetic and geometric population libria; it is possible that a sufficiently large random
growth rates, but a stochastic equivalent to the perturbation can move the population from the
deterministic growth rate can be calculated domain of attraction of one equilibrium to that of
(Tuljapurkar, 1982). As with the unstructured pop- the other. Similarly, where there is an Allee effect a
ulation, adding stochastic effects always reduces species is unable to increase in density when rare so
long-term growth rates. Second, certain age or zero population density is locally stable; random
stage classes may be much more susceptible to effects can push a species density below the critical
stochastic perturbation than others. Random threshold that leads to extinction. It is also possible
effects may thus lead to perturbations that disrupt that a species that for some reason has fallen below
the age-structure of the population (structural the threshold can be rescued by a random set of
variance; Coulson et al., 2001; Lande et al., 2002). good breeding seasons. Of course, even when a
Here, stochasticity influences the population species can increase when rare, stochastic extinction
is permanent if there are no sources of migrants to for a significant period of time. Indeed, this beha-
rescue the population. This treatment of stochasti- viour may go on for ever if stochastic perturba-
city in population models has been called the pas- tions are large enough to prevent the system ever
sive treatment of noise. from settling on the stable cycles. The time series
The shape of the equilibrium probability dis- produced by such a process can be indistinguish-
tribution of abundances will obviously be deter- able from chaos: it can show exactly the same
mined by the magnitude and direction of the extreme sensitivity to initial conditions, and
stochastic perturbations to the demographic para- attempts to reconstruct the attractor would suggest
meters, but also by the dynamic consequences of that it had a non-integer number of dimensions.
the perturbations; that is, the interaction of the In discussing the bifurcation diagram in Figure
noise with the deterministic dynamics. Consider 3.8 we already noted how random effects would
unstructured populations with deterministically interact with the deterministic component of the
stable equilibria which are approached either dynamics to give chaotic population behaviour
smoothly (Figure 3.3) or by damped oscillations throughout the region beyond the ‘point of accu-
(Figure 3.7a). It is very likely that the first popu- mulation’, even though here there are narrow
lation will tend to return towards the equilibrium windows of cyclic behaviour. As with chaotic
faster than the population with damped oscilla- repellers this is another example of the impossi-
tions, and for the same amount of environmental bility of separating the deterministic and stochastic
stochasticity will have a lower variance equili- aspects of population dynamics in general and
brium population density. A population with an chaos in particular.
oscillatory approach to a stable equilibrium can Although it may seem unarguable that we
more easily be prevented from reaching that should seek to develop models with both sto-
equilibrium and thus appear to the observer to be chastic and deterministic components, exactly how
persistently cyclic. This type of dynamic behaviour to do this is not always obvious. For example,
has been termed quasicyclic (Nisbet and Gurney, adding one type of noise to a model with a
1976) and has been seen in several experimental deterministically stable equilibrium and a different
systems, including the flour beetle study described type of noise to a model governed by a chaotic
above as an example of chaos (Costantino et al., attractor can produce dynamics that equally well
1997). match the type of data that ecological field studies
Consider an unstructured dynamic system that produce. Also it is often not clear how stochasticity
is at the edge of chaos, perhaps showing persistent should be introduced into the model, onto which
cycles. If one or more parameters were changed demographic parameters, and with what correla-
slightly, it would move from persistent cycles into tion structure. Nevertheless, we are optimistic
the region of chaos where its dynamics would be about the future. For the analysis of time series and
governed by a strange attractor. Near this thresh- other observational data there are a variety of new
old, the transient behaviour of the population statistical methods and techniques that will help
before it settles into persistent cycles can be very identify the major stochastic drivers, and reveal
complex. Although in this region there is not a how they interact with the underlying biology of
strange attractor, dynamics may be influenced by the species (Coulson et al., 2001; Lande et al., 2003;
an object called a strange repeller (Rand and Wilson, Turchin, 2003; Stenseth et al., 2004). There is also an
1995), which like a strange attractor is a fractal, but increasing willingness of ecologists to experiment,
repels rather attracts dynamic trajectories. One can both in the laboratory and the field, and to inte-
think of the system like the ball in a pinball grate modelling with experimental design and
machine, careering from buffer to buffer, perhaps analysis.
CHAPTER 4
Metapopulations and their

spatial dynamics
Sean Nee
4.1 Introduction (Hanski, 1999). This Scandinavian butterfly lives

in dry meadows which are small and patchily
The study of metapopulation dynamics has had a
distributed. Another reason for local population
profound impact on our understanding of how
extinction is that the habitat patch itself may be
species relate to their habitats. A natural, if naı̈ve,
ephemeral. For example, wood-rotting fungi will
set of assumptions would be that species are to be
find that their patch ultimately rots completely
found wherever there is suitable habitat that they
away (Siitonen et al., 2005) and epiphytic mosses
can get to; that species will rarely, if ever, be found
will ultimately find that their tree falls over (Snall
in unsuitable habitat; that they will be most
et al., 2005).
abundant in their preferred habitat; that species
can be preserved as long as a good-size chunk of
suitable habitat is conserved for them; and that
destruction of a species’ habitat is always detri-
mental for its abundance. We will see that none of
these reasonable-sounding assumptions is neces-
sarily true. Metapopulation biology is a vast field,
so to focus this chapter I will be guided partly by
questions relevant to conservation biology.
There are two important kinds of metapopula-
tion. The so-called Levins metapopulation idea
(Levins, 1970) is illustrated in Figure 4.1. It is
imagined that patches of habitat suitable for a
species are distributed across a landscape. Over
time, there is a dynamical process of colonization
and extinction: the colonization of empty patches
by occupied patches sending out colonizing pro-
pagules and the extinction of local populations on
occupied patches. This extinction can occur for a
Figure 4.1 The Levins metapopulation. Shaded circles are habitat
number of reasons. Small populations are prone to patches with a local population of a species: empty circles are empty
extinction just by the chance vagaries of the habitat patches. An arrow from a filled circle to an empty circle
environment, reproduction, and death—environ- indicates a colonization event: an arrow turning back on itself onto a
mental and demographic stochasticity (May, filled circle indicates an extinction. Two successive periods in the life
1974b; Lande et al., 2003). An example of a species of a metapopulation are illutrated. The long-term dynamics have been
likened to the asymmetrical blinking on and off of Christmas tree
for which this is important is the Glanville fritil- lights (Wilson, 1992). Although extinction is the ultimate fate of
lary butterfly (Melitaea cinxia), which has been any local population, nonetheless the species can persist as a
extensively studied by Hanski and colleagues metapopulation.
35
Source Sink
Figure 4.2 Source and sink populations. In the simplest case, there are two habitat patches of different types, only one of which is of sufficient
quality to sustain a population (the source population). However, a population can be maintained in the suboptimal habitat (the sink population)
if it is topped by immigration from the population in the optimal habitat. Reproduction may occur in the sink habitat, but at a level that is
insufficient to maintain the population in the absence of immigration. The relative sizes of the circles are chosen to make the point that the sink
population may be much larger than the source.
The second type of metapopulation consists of models that are essentially nothing more than
local populations connected by dispersal, but cartoons and (2) the power of graphical models
without the extinction of the local populations. (see also Chapters 3 and 5).
This type has been studied intensively in popu-
lation genetics, the terminology of which refers
to populations and subpopulations rather than
4.2 The Levins metapopulation model
populations and metapopulations. (Perhaps eco- We will now consider Levins’ metapopulation
logists look up, star-wards, whereas population models in more detail. Book-length treatments
geneticists look down, navel-wards?) As an exist (Hanski and Gilpin, 1997; Hanski, 1999).
example of this type of metapopulation we will There are different biological systems that can be
specifically look at source–sink metapopulation studied within the framework of metapopulation
dynamics, illustrated in Figure 4.2. In this model, theory. The patches and local populations may
populations in patches of good habitat sustain correspond to their ordinary meanings: local
populations in poor habitat. populations of butterflies in a meadow or shrews
Both types of metapopulation are, of course, on different islands, for example. Of course, it
abstractions, just like the concept of a population must be established that the dynamical processes
itself. And there are many important related of colonization and extinction are actually occur-
abstractions, like the mainland/island concept of ring. So, for example, Smith and Green (2005)
island biogeography (Macarthur and Wilson, conducted a meta-analysis to see if the metapo-
1967), which can be viewed as a hybrid of the pulation paradigm is appropriate for pond-
above two ideas. For conservation biology, one dwelling amphibians (it largely is). But there are
important way that reality departs from the pure other systems that also belong in the metapopu-
Levins metapopulation is that different local lation framework and some of these are listed in
populations may have different extinction prob- Table 4.1. This more abstract view received pro-
abilities and some patches may be more accessable minence after the work of Lande in his celebrated
for dispersal than others (Harrison and Taylor, study of the Northern spotted owl (Lande, 1987,
1997). Hanski has elaborated metapopulation 1988a, 1988b), a territorial species, although Hast-
theory further to incorporate these and other rea- ings (1980) may have been the first to adopt it in
listic features (Hanski, 1994). the study of coral dynamics.
The structure of this chapter is somewhat unu- This chapter discusses metapopulations at a
sual in that I will split the discussion of the Levins general level that can encompass all these differ-
metapopulation in two and sandwich the source– ent biologies. The disadvantage of this approach
sink section in between. This is done to highlight is that the devil may be in the detail and a
(1) how much can be done with very simple metapopulation analysis that is biology-general
METAPOPULATIONS AND SPATIAL DYNAMICS 37
Table 4.1 Some less obvious examples of Levins metapopulations.
Example Local population Patch Colonization Extinction
Territorial species Breeding pair Territory Occupation of empty Death

(Lande, 1987) territory
Sessile organisms (Hastings, Single individual Sufficient, suitable, Establishing in the Death
1980; Stone, 1995) space for an individual space
Gut flora Bacterial population Individual gut e.g. Neonatal Host death or treatment
ingestion of feces with antibiotics
Infectious disease organisms Individual viral load, Host individual Infection Host death or recovery
(Nee et al., 1997) for example
Persistent infection, e.g. HIV Intracellular agent Suitable cell, Entry into cell Cell death
e.g. CD4 T-cells
(biology-free?) may lack features essential to The curves in Figure 4.3 are the simplest ones
understanding a specific problem at hand. The that satisfy these assumptions—parabolas and a
advantage of the approach is that one can gain straight line. Various simple, analytical results are
insight into general properties of metapopulation derived with these functions below. But we can
dynamics that are not specifically tied to any par- derive some important results from both the graph-
ticular biological detail. An additional advantage is ical model and with simple reasoning.
that results derived in one context can be seen to
be relevant in other contexts. For example, as
4.2.1 Empty habitat
indicated in Table 4.1, epidemiology is a version of
metapopulation biology and, as a vast and mature At equilibrium, not all patches are occupied
field, may have results that can be plundered (Figure 4.3). Furthermore, depending on the nature
for use in other contexts. We will see examples in of the curves, a substantial fraction of the patches of
this chapter. Another recent example of the virtue perfectly suitable habitat may be unoccupied.
of an abstract view of models, if it had been taken, Hanski (1996) discusses the work of Boycott in the
comes from the neutral theory of biodiversity 1920s, who studied the colonization and extinction
(Hubbell, 2001), where much time and effort was of snail populations in ponds over a 10-year period.
spent re-deriving results that are well known in the Boycott demonstrated with transplantation experi-
neutral theory of population genetics (Nee, 2005). ments that ponds without a population of a parti-
The basic Levins metapopulation model is illu- cular snail species were nonetheless perfectly
strated in graphical form in Figure 4.3. The moti- suitable habitat. The large differences between
vation for the colonization curve is simple. If there people in the species composition of their intestinal
are no colonized patches, then there are no pro- flora (Eckburg, 2005) must, at least in part, be a
pagules available to colonize empty patches and consequence of this fundamental fact about
therefore no colonization is occurring. Similarly, metapopulations. There are many examples of
no colonization of empty patches can occur when unoccupied—but perfectly suitable—habitat from
all the patches are occupied. Colonization of empty epidemiology: only a fraction of a population is ever
patches will occur at the highest rate at inter- infected by an infectious, endemic disease organism.
mediate values of patch occupancy, when there are For example, about 25% of us have the bacteria
a lot of patches emitting propagules and a lot of Helicobacter pylori living in our stomachs. A striking
empty patches available for colonization. The demonstration that uninfected people are none-
extinction curve is straightforward: the more pat- theless suitable habitat was provided by Barry
ches there are then the more patches in which Marshall, who showed that Helicobacter causes
extinction can occur. stomach ulcers with true Austalian directness: he
drank a flask of bacterial culture and was success- which is the number of individuals infected by a
fully colonized and ulcerated. For two millennia, single infected individual introduced into a wholly
since the Roman physician Galen, medical ortho- susceptible population (see Chapter 10 in this
doxy has held that excess acid in the diet causes volume). At the equilibrium level of patch occu-
stomach ulcers: one Australian drinking a flask of pancy in the pristine world, y*, which is where the
bacteria consigned this orthodoxy to the dustbin highest colonization curve and extinction curve
of history. intersect in Figure 4.3, 1 – y* patches are unoccu-
pied and so are available to be colonized. At
equilibrium, each patch, over the course of its
‘lifetime’ colonizes, on average, one empty patch;
4.2.2 Eradication threshold
with fewer than 1 – y* empty patches, each
Suppose that habitat destruction occurs, either by occupied patch will colonize less than one other
paving over patches or vaccinating individuals, for patch. Imagine a world with fewer than 1 – y*
example (from the point of view of an infectious patches and a metapopulation consisting of a sin-
disease, vaccination is a wanton act of habitat gle occupied patch in this world. Over the course
destruction). As illustrated in Figure 4.3, habitat of its lifetime, the patch will colonize less than one
destruction has the effect of lowering the coloni- patch and so the metapopulation will go extinct.
zation curve and making it less steep towards the Hence, the threshold level of destruction is simply
origin. This is because colonization is hampered by y*. This argument can be generalized to other
the reduction in the number of patches available to ecological relationships (Nee, 1994): for example, it
actually colonize: for example, visualize seeds can be used to calculate the minimum prey car-
landing on tarmac where before there was a patch, rying capacity needed to sustain a predator
or a person sneezing on someone who is vacci- population.
nated against a disease. If destruction is so exten- The above argument does, of course, make
sive that the slope of the colonization curve at the assumptions. In epidemiology, these are known as
origin becomes smaller than the slope of the weak homogenous mixing (Anderson and May,
extinction curve, extinction is inevitable and this 1991) and assume, for example, that all patches are
does not require all or even necessarily a sub- roughly equivalent in their colonization and
stantial fraction of patches to be destroyed. extinction properties. A particularly important
Epidemiology told us long ago that it is not assumption, from the conservation point of view,
necessary to vaccinate 100% of a population to is that there is no rescue effect.
eradicate an infectious disease (Anderson and The rescue effect (Hanski, 1999) is the exact
May, 1991). So we say there is a threshold level of metapopulation equivalent of the Allee effect in
patch destruction above which extinction will population biology (Stephens and Sutherland,
occur, and this is called the extinction or eradica- 1999). The Allee effect describes a situation in
tion threshold. which either birth rates decline or death rates
Note that metapopulation extinction does not increase at low densities. Many biological features
occur instantly when destruction exceeds the of species can produce an Allee effect. For
threshold. Just like a population in which birth example, plants may have difficulty getting polli-
rates are always, marginally, less than death rates, nated at low densities or animals that rely on social
the metapopulation may dwindle to extinction on defence against enemies may find these breaking
a long time scale. This fact has been called the down at low densities. For example, smaller
extinction debt (Tilman et al., 1994) to recognize colonies of sea birds suffer higher predation (Ser-
that the habitat destruction of today may have to rano et al., 2005). The effect is also known as
be paid for in the future with extinctions. inverse density dependence (Begon et al., 1996a).
What is the threshold? This can be computed In metapopulation biology the rescue effect may
very simply, using an argument from epidemiol- result if an input of colonists into a patch lowers
ogy used to estimate R0 (Anderson and May, 1991), the patch extinction rate. In this case, at lower
METAPOPULATIONS AND SPATIAL DYNAMICS 39
0.3 0.3
Colonization or extinction rate

Colonization or extinction rate
0.2 0.2
0.1 0.1
0.0
y* y* 0.0 0.2 0.4 0.6 0.8 1.0
0.00 0.25 0.50 0.75 1.00
Proportion of patches occupied
Proportion of patches occupied
Figure 4.4 Threshold level of patch occupancy. As in Figure 4.3, the
Figure 4.3 The basic Levins metapopulation model. This illustrates parabola is the colonization curve and here the piece-wise straight
how the metapopulation colonization rate (parabolas) and extinction line is the extinction curve. The lower, left-hand equilibrium is
rate (straight line) vary as a function of the overall level of patch unstable, and if the metapopulation size falls below it, extinction will
occupancy, expressed as a fraction of the total number of habitat result. For a general review of population models with this sort of
patches in a pristine world. Illustrated are two colonization curves behaviour see May (1977a).
corresponding to the situation in a pristine environment (top
parabola) and to the situation in which 50% of the patches, relative
to the total number of patches in the pristine environment, have been
destroyed (bottom parabola). These curves are justified in the text.
The equilibrium level of patch occupancy, y*, is found where the
colonization and extinction curves intersect. As explained in the text, effectively asexual (they are self-fertilizing). May
habitat destruction lowers, crucially, the slope of the colonization (1977a) demonstrated that sexual reproduction can
curve near the origin. Marginally more destruction will result in the produce multiple equilibria as a result of the dif-
colonization curve having a lower slope than the extinction curve at ficulty of finding a mate at low abundance. This
the origin and extinction will be inevitable: destruction will have
suggests that we should find a bimodal distribu-
exceeded the eradication threshold.
tion of patch occupancy (the proportion of indivi-
duals harbouring worms, i.e. the prevalence) in
studies of nematodes, but not cestodes. Data col-
levels of patch occupancy there will be a lighter lected and analysed by Per Arneberg show this is,
rain of colonists over the metapopulation and the indeed, the case: see Figure 4.5.
patch extinction rate may rise. As illustrated
graphically in Figure 4.4, this results in a threshold
level of patch occupancy below which the meta-
4.3 Source–sink metapopulations
population will go extinct. Population biologists have long paid attention
A consequence of this effect is that we may to heterogeneities in populations—in age, for
expect to find in nature a bimodal distribution of example—and to the effects of these hetero-
patch occupancy, with species found in either geneities on population dynamics (see Chapter 3 in
many patches or none (Hanski, 1982; Hanski et al., this volume). Also, the effects of heterogeneity in
1995). We may also find a minority of species at time in the environmental factors affecting birth
intermediate levels of occupancy, either because and death rates have also been considered exten-
the metapopulation is in transit from one equili- sively (Lande et al., 2003). It is a somewhat more
brium to another or because of immigration of the recent interest to consider the effects of hetero-
species from high-occupancy metapopulations geneity in habitat. But as habitat is modified and
outside the study area. destroyed by humans, ecologists are increasingly
An example of this is provided by worms interested in this heterogeneity as well.
inhabiting mammalian intestines (Arneberg and Habitats vary in their suitability for species in
Nee, unpublished work). Nematode worms terms of the individuals’ needs for survival and
have sexual reproduction whereas cestodes are reproduction. Obviously, penguins would be
Another random document with
no related content on Scribd:
Saona, sul punto di entrare, attraversato il territorio sequano, nel
territorio degli Edui. Ma per quanto egli si affrettasse, non riuscì che
ad annientare una piccola retroguardia, rimasta di qua dal fiume.
Passò allora con l’esercito il fiume, e si diede ad inseguire il nemico:
quand’ecco presentarglisi un’ambasceria di Elvezi, alla cui testa era
lo stesso loro capo, Divicone, una vecchia conoscenza dei Romani,
perchè tanti anni prima, giovanissimo, aveva preso parte
all’invasione dei Cimbri e dei Teutoni. Questa ambasceria dichiarò
che gli Elvezi non avevano alcuna intenzione di far la guerra a Roma;
che volevano solo stabilirsi in Gallia e con il consenso dei Romani....
Le richieste non avrebbero potuto essere più discrete: ma chi le
faceva era un antico capo dell’orda cimbrica e Cesare non si fidò.
Respinse le proposte e ricominciò a seguire a qualche distanza gli
Elvezi, che avevano ripresa la marcia, senza tuttavia subito attaccarli,
parte perchè aspettava una buona occasione, parte perchè i suoi
movimenti erano di continuo intralciati dalla mancanza di vettovaglie.
A queste si erano incaricati di provvedere gli Edui: ma troppo spesso
al loro impegno mancavano, allegando ora un pretesto ora un altro.
Cesare volle alla fine mettere le cose in chiaro; fece un’inchiesta; e
allora, per la prima volta, si accorse che inseguendo gli Elvezi era
venuto a cacciarsi nel tremendo ginepraio delle discordie galliche. Se
il governo eduo gli aveva chiesto aiuto contro gli Elvezi, c’era tra gli
Edui un partito — e potentissimo — che considerava gli Elvezi come
amici. Anche tra gli Edui, come presso quasi tutti i popoli gallici, la
vecchia nobiltà, che fin allora aveva tenuto il potere, impoveriva e si
indebitava; arricchiva invece e predominava una piccola plutocrazia,
che accaparrava terre e capitali, monopolizzava la riscossione delle
imposte, esercitava con grande profitto l’usura, e, appoggiandosi
sulla plebe, ch’essa si studiava di favorire e accarezzare, lottava per
spodestare la vecchia aristocrazia, instaurando dei governi personali,
non dissimili da quello, che Cesare, Crasso e Pompeo avevano
costituito in Roma. Questo partito aveva sollecitato gli Elvezi a venire
in Gallia, perchè sperava di servirsene come di una milizia, sia per
scacciare Ariovisto, sia per consolidare il suo potere; ed ora cercava
di favorirli, tagliando i viveri ai Romani.
Entrato in Gallia come un liberatore, Cesare s’accorgeva ad un tratto
che una parte di coloro che egli voleva liberare, erano amici del
nemico e se la intendevano segretamente con lui ai danni del
presunto salvatore. Egli fu così preoccupato di questa strana
posizione, che deliberò di ritornare indietro, per provvedere meglio al
proprio vettovagliamento. Ma allora gli Elvezi attaccarono
d’improvviso i Romani. Lo scontro fu lungo e difficile; e Cesare potè
disimpegnare il grosso delle sue legioni solo a costo di perdite gravi.
Cosicchè, mentre il tanto inseguito nemico poteva tranquillamente
proseguire verso il nord, egli era obbligato a restare tre giorni sul
posto per seppellire i morti e rimediare a tutto lo scompiglio arrecato
da quell’attacco improvviso. Quel che sarebbe successo se Divicone
lo avesse assalito di nuovo il giorno dopo, è difficile dire: ma gli
Elvezi non volevano la guerra a oltranza con Roma; e stanchi della
lunga marcia, impressionati dalle difficoltà dell’emigrazione e dalla
ostilità dei paesi che attraversavano, forse anche atterriti dalle loro
stesse vittorie e dalle prevedute vendette di Roma, offersero
novamente pace. E l’ottennero a condizioni che mostrano quanto
poco Cesare sentisse d’averli vinti. Quelli che vollero tornarono nel
loro antico territorio, ma ottenendo da Cesare il titolo di alleati del
popolo romano: quelli che vollero restare in Gallia, ebbero territori
dagli Edui (58 a. C.).
11. La guerra contro Ariovisto (58 a. C.). — La prima impresa di

Cesare in Gallia era finita senza una grande vittoria e per di più
aveva alienato da Roma il partito nazionale e le moltitudini. Il solo a
cui la guerra di Roma contro gli Elvezi giovava era appunto Ariovisto.
Cesare non tardò ad accorgersene nell’assemblea generale dei Galli,
che venne spontaneamente a radunarsi dopo la pace con gli Elvezi;
e nella quale le città galliche gli domandarono di liberare la Gallia da
Ariovisto. Egli era venuto come un liberatore: quello, non gli Elvezi,
era il nemico della Gallia [9]. Cesare capì che il prestigio suo e di
Roma era finito, se egli non faceva e subito la guerra contro Ariovisto
e non liberava la Gallia dal pericolo germanico. Ma il re degli Svevi,
non era solo un nemico forte; era anche stato dichiarato dal senato,
l’anno prima, «amico ed alleato del popolo romano». Mancava
dunque la ragione legale della guerra: il che accresceva per Cesare il
pericolo di farla, nel caso che fosse vinto. Tuttavia l’impegno era tale,
che Cesare non esitò; e subito prese a cercare un casus belli. Intimò
dunque ad Ariovisto di non condurre altri Germani in Gallia, di
restituire agli Edui gli ostaggi, di smettere da ogni molestia o atto di
guerra contro questo popolo ed i suoi alleati: cercò insomma di
trovare il pretesto di un litigio con Ariovisto nell’incarico di difendere
gli Edui, che il senato gli aveva commesso. Ma Ariovisto rispose — e
non senza acutezza — che il senato, dichiarandolo amico e alleato,
aveva riconosciuto tutte le conquiste da lui fatte in Gallia. Cesare
allora dichiarò la guerra e marciò alla volta del Reno. La battaglia fra
le legioni e le schiere di Ariovisto ebbe luogo nell’Alsazia superiore,
forse non lungi da Mülhouse. Ma questa volta, l’esercito romano
ebbe interamente ragione di un nemico, che la voce pubblica diceva
feroce e crudele. I Germani furono gettati al di là del Reno, e lo
stesso Ariovisto potè a mala pena scampare alla morte con la fuga
(settembre 58).
12. Il richiamo di Cicerone e la prima crisi della triarchia. —

La dominazione germanica in Gallia era caduta per secoli; la Gallia
era restituita a se stessa. Cesare aveva questa volta riportato una
vera e grande vittoria, i cui effetti durerebbero nei secoli. Ma il
giudizio dei contemporanei sugli eventi è spesso fallace; e Roma si
accorse appena di questa vittoria, che apriva alla lingua e allo spirito
latino una delle terre privilegiate dell’Europa. Roma era tutta intenta
a ben altra faccenda: l’ingiusto bando di Cicerone. Esiliando Cicerone
per la condanna dei congiurati, Clodio aveva abusato della sua
potenza. Passato il primo spavento, la pubblica opinione si era
riavuta. Quanti non erano contenti, specie in senato, del governo
della triarchia, avevano visto subito nella prepotenza di Clodio
un’arma per nuocere a Crasso, a Pompeo e a Cesare; e quindi non
avevano risparmiato fatiche per commuovere il pubblico a favore di
Cicerone: il pubblico, che amava e ammirava Cicerone, si era a poco
a poco appassionato: persin dei collegia di artigiani e dei municipi
d’Italia avevano chiesto il richiamo. In breve, mentre Cesare
combatteva Ariovisto, Roma dimenticava ogni altra cosa, non
pensava che a Cicerone e al modo di rendergli giustizia. Pompeo
stesso che, come Cesare del resto, aveva subìto, più che voluto,
questa vendetta del tribuno, non aveva tardato, per non contrariare
troppo il pubblico, a dichiararsi favorevole al ritorno del grande
uomo. Senonchè Clodio, che dell’esilio di Cicerone aveva fatto un
impegno personale, non aveva esitato a rivoltarsi perfino contro
Pompeo, attaccando la sua politica in Oriente, andando alla testa di
bande armate a disturbare i comizi pubblici, agitando Roma con
continue dimostrazioni e tumulti, e cercando di imporsi alla viltà
universale con il terrore. Gli avversari, esasperati, si intestarono
ancor di più a volere a Roma a tutti i costi Cicerone; le elezioni, fatte
sulla disputata questione, furono favorevoli alla nobiltà, anche
perchè Pompeo si era accostato al partito senatorio. La causa di
Cicerone procedeva dunque a gonfie vele; e già i consoli e i tribuni si
preparavano a presentare una legge per il suo richiamo. Ma Clodio
non era uomo da dichiararsi vinto così facilmente e, anche scaduto
da tribuno, tentò impedire con la forza l’approvazione della legge.
Roma era piena di tumulti e di risse.
13. La guerra contro i Belgi (57 a. C.). — Insomma, le cose

volgevano piuttosto male a Roma per Cesare. Roma non badava a
quel che Cesare faceva in Gallia, e il governo personale dei tre
potenti vacillava, dopo un anno, per la discordia nata dalla questione
del richiamo di Cicerone tra Pompeo e Clodio. Il congegno elettorale,
con cui egli si era immaginato di dominare la repubblica, stava per
spezzarsi. Era doppiamente necessario per Cesare di compiere,
l’anno prossimo, in Gallia, qualche impresa anche maggiore. Avendo
appreso nell’inverno dal 58 al 57 che la Gallia settentrionale, il
potente paese dei Belgi e la parte occidentale della Gallia del Nord,
l’Armorica, ossia tutte le popolazioni gallo-germaniche, stendentisi
tra il Reno, la Schelda, l’Atlantico e la Senna, si agitavano, inquiete
per la presenza delle legioni romane, svernanti in Gallia, deliberò di
andare incontro nella prossima primavera al pericolo, interpretando
anche egli l’incarico avuto dal senato così largamente, come aveva
fatto Lucullo nell’ultima guerra mitridatica.
Era una impresa difficile quella a cui Cesare si accingeva.
Cinquant’anni prima, la Belgica aveva resistito strenuamente a quei
Teutoni e Cimbri, che più volte avevano sconfitto gli eserciti romani.
Si diceva che potesse mettere in campo 350.000 guerrieri; ed era
paese ignoto. Nè c’era da far molto assegnamento sulla fresca
amicizia dei Galli del centro. Cesare arrolò due nuove legioni e molti
arcieri e frombolieri in Asia, a Creta, nelle Baleari; persuase gli Edui a
invadere il paese dei Bellovaci, il più forte dei popoli Belgi; ed andò
ad aspettare l’orda belga che si avanzava, sull’Aisne, dove si trincerò.
I Belgi arrivarono e si trincerarono alla loro volta: i due eserciti
stettero parecchi giorni a guardarsi, ciascuno aspettando che l’altro
l’attaccasse; ma nessuno dei due volle fare il giuoco dell’avversario,
e alla fine un giorno, dopo poche scaramucce, l’esercito belga si
ritirò. La meraviglia di Cesare fu così grande, che a tutta prima
dubitò di un inganno. Solo più tardi egli seppe che i Bellovaci,
preoccupati della invasione degli Edui, avevano voluto ritirarsi a
difendere il proprio paese. Quella defezione, l’imperfetto servizio
delle sussistenze, il timore della potenza romana avevano sciolto la
lega. La fortuna aveva dunque mirabilmente servito Cesare; il quale
comprese che quello era proprio il momento di dar addosso al
nemico sparpagliato e di domare una ad una quelle tribù valorose,
ma volubili. Fu questa la seconda e più fortunata fase della sua
guerra belgica. Uno dopo l’altro i Suessioni, i Bellovaci, gli Ambiani,
gli Aduatici, perfino i terribili Nervii, i più bellicosi fra i Belgi, furono
per amore o per forza assoggettati (57 a. C.).
14. L’annessione della Gallia (56 a. C.). — Senonchè mentre

Cesare combatteva con fortuna in Gallia, le cose precipitavano a
Roma. Cicerone era finalmente ritornato, accolto in tutta l’Italia da
entusiastiche dimostrazioni; ma solo dopo che Pompeo aveva trovato
un tribuno della plebe, Tito Annio Milone, il quale aveva raccolto una
banda di gladiatori e di bravi, e con quella, tra tumulti, zuffe e
sangue, fatta approvare, il 4 agosto del 57, la legge che lo
richiamava. Ma Clodio non si era dato per vinto: aveva annunziato la
sua candidatura alla edilità per l’anno prossimo; aveva tentato di
sollevare il popolo contro Pompeo, spargendo la voce che questi
faceva la carestia per essere creato re di Roma; cercava d’impedire,
per mezzo di tribuni amici, che si ripagasse a Cicerone la casa
distruttagli; e infine nelle elezioni per il 56 aveva portato l’inatteso
soccorso delle sue bande elettorali ai candidati del partito senatorio,
avversi alla triarchia, facendo loro conquistare tutti i posti di pretore
e i due consolati. Come se tanti guai non bastassero, la questione
dell’Egitto era inopinatamente risorta. Tolomeo Aulete, che Cesare
aveva ufficialmente fatto riconoscere re d’Egitto, era stato scacciato
da una rivoluzione ed era in quel tempo tornato a richiedere la
protezione della repubblica. Ma troppi erano quelli che ambivano
l’incarico di restituirlo sul trono: primo fra essi Pompeo. Invece il
senato, al solito, non ne voleva sapere, e la sua opposizione era
apertamente o nascostamente aiutata da Clodio, e, pare, anche da
Crasso. Insomma quel governo personale e di clientela, che doveva
sostituire il senato invecchiato e impotente, si sfasciava appena
formato; e verso la metà del 57, potè sembrare ormai spacciato.
Clodio, per vendetta, era passato interamente al partito senatorio e
già aveva tentato di far abrogare le leggi del 59. Fu allora che la
mente vasta e ardita di Cesare concepì un’idea grandiosa, che
doveva aver nella storia effetti immensi, oggi ancora vivi e profondi.
Le vittorie sui Belgi avevano commosso Roma profondamente.
Secondo dirà poco di poi Cicerone, fino a quel momento i generali
della repubblica, Mario compreso, erano stati paghi di respingere i
Galli: Cesare invece era entrato nel loro paese. Che cosa dunque si
sarebbe pensato se Cesare ora, d’un colpo, avesse annunziato a
Roma che la Gallia Transalpina poteva considerarsi soggetta a Roma
tal quale come le Spagne, la Macedonia, la Siria?
Fu questa l’idea che egli venne maturando dopo le vittorie sui Belgi.
Per porla ad effetto, spedì il suo luogotenente P. Crasso, con una
legione, nella Gallia occidentale a ricevere pro forma l’atto di
sommissione delle tribù sparse tra le foci della Loira e della Senna; e
un altro suo luogotenente, Sulpicio Galba, con una legione, nell’alto
Vallese, presso il Gran San Bernardo. Dopo di che tornò rapidamente
nella Cisalpina, annunziando al senato ch’esso poteva deliberare
l’annessione della Transalpina ed inviare, come di regola, i dieci
commissari, che avrebbero dovuto organizzare, insieme con il
proconsole, la nuova provincia. Era questa un’audacia quale nessun
altro generale romano aveva ancora osato: affermare che due guerre
e due anni erano bastati a conquistare un paese così vasto e del
quale tante parti non avevano ancora visto l’elmo d’un legionario o la
toga di un magistrato romano. La temeraria affermazione
scatenerebbe nella Gallia un uragano di guerre, assai più tremende
di quelle che Cesare aveva fin allora sostenute: ma l’impegno
solenne, al quale Cesare incatenava la Repubblica, costringerebbe
questa, e il suo audace generale, a compiere una conquista, che
distraendo Roma dai facili successi orientali, avrebbe spostato l’asse
della civiltà verso Occidente.
15. Il convegno di Lucca e il consolato di Crasso e di Pompeo

(56-55 a. C.). — All’annunzio l’Italia tutta andò in delirio per la
gioia. Roma aveva avuto per tanti secoli tanta paura dei Galli, che
nessuna conquista poteva sembrargli più meravigliosa di questa o
rallegrarla maggiormente. Il popolo deliberò di inviare a Cesare una
deputazione di senatori per felicitarlo della vittoria; il senato,
cedendo all’opinione pubblica, decretò una supplicazione di quindici
giorni, la più lunga ordinata fino ad allora; molti avversari dell’anno
innanzi si convertivano all’ammirazione. Cesare diventò l’idolo del
pubblico; e di questo fugace favore approfittò per ricostituire il
governo del 59: urgente bisogno, perchè dopo Cicerone anche
Catone ritornava da Cipro. Convocò dunque Crasso e Pompeo a
Lucca, dove i suoi amici giunsero, seguiti da una vera corte di
senatori, ed espose loro un vasto piano, con il quale avrebbero
salvato la comune potenza, ormai pericolante sotto i colpi della
ringagliardita opposizione. Crasso riconcilierebbe Clodio con Pompeo;
ed ambedue si proporrebbero candidati al consolato per l’anno 55:
durante il consolato essi farebbero prolungare a lui, per altri cinque
anni, il comando della Gallia e assegnargli i fondi per pagare le
legioni che aveva reclutate, oltre quelle assegnategli dal senato,
dopo il principio della guerra: egli conquisterebbe in quei cinque anni
la Britannia e porterebbe le legioni oltre il Reno; Crasso, dopo il
consolato, avrebbe avuto la provincia della Siria e compiuto la
conquista della Persia; quanto all’Egitto, ambedue ne deporrebbero
l’idea, ma si incaricherebbe Gabinio di ricondurre, senza
autorizzazione del senato, Tolomeo nell’Egitto, a condizione che
pagasse a ciascuno di loro una somma considerevole. Sembra che la
somma chiesta da Cesare fosse di 17 milioni e mezzo di sesterzi. In
cambio, Pompeo avrebbe, dopo il consolato, per cinque anni, le due
Spagne.
Non sappiamo quali discussioni ebbero luogo tra Crasso, Pompeo e
Cesare; ma sappiamo che essi si misero d’accordo e che quella
specie di rivoluzione politica, adombrata da Cesare nell’anno del
consolato, sembrò acquistare forma e corpo di saldo governo. Tre
clientele e tre capi potentissimi governerebbero d’accordo, in luogo
del senato, la repubblica; e invece di temporeggiare con prudenza,
come il senato faceva, conquisterebbero tutto ciò che potevano:
dopo la Gallia, la Britannia e la Persia. E per un istante la repubblica
sembrò davvero, per effetto del rinnovato accordo, riordinarsi ed
agire. Nessuno pensò più ad abrogare la legge agraria di Cesare;
Cicerone rispose trionfalmente, in senato, con la sua storica orazione
de provinciis consularibus, a coloro che, poichè la Gallia era
conquistata, avrebbero voluto scorciare e ridurre i poteri proconsolari
di Cesare; la proposta di ordinare la Gallia Transalpina in provincia e
di inviare all’uopo i dieci legati senatorî fu approvata [10]. I popoli
della Gallia con cui Roma aveva trattato di alleanza, come gli Edui e i
Sequani, i popoli più ricchi e civili del centro, conservarono la
indipendenza con il titolo di alleati; mentre le barbare popolazioni del
settentrione e dell’occidente furono sottomesse al dominio romano.
Per impedire infine che i consoli in carica si valessero dei loro poteri
per fare ostruzione alle candidature di Pompeo e di Crasso, si trovò
modo di rimandare le elezioni sino ai primi giorni del 55, e di farle
sotto la presidenza di un interrex amico. Crasso e Pompeo furono
eletti; e, appena eletti, brigarono affinchè anche la maggioranza
delle restanti magistrature fosse occupata dai loro amici. Così M.
Porcio Catone non fu eletto pretore e riuscì in suo luogo P. Vatinio,
l’autore della legge che nel 59 aveva conferito a Cesare il governo
delle Gallie. A tenore delle leggi vigenti, il senato aveva già in
anticipazione decretato che ai consoli dell’anno toccassero, per il 54,
rispettivamente, la Siria e la Spagna ulteriore. Ma i due consoli
provvidero subito anche a questo: un tribuno, C. Trebonio, propose
che il duplice proconsolato fosse invece quinquennale, e che non
solo la Spagna ulteriore, ma le due Spagne, ulteriore e citeriore,
fossero affidate a Pompeo. Approvata questa, i due consoli fecero
approvare un’altra legge che prorogava di cinque anni a Cesare il
comando delle Gallie e dell’Illiria. Restavano le difficoltà egizie.... Se
non che ad un tratto si apprese a Roma che Tolomeo era stato
ricondotto in Egitto; che sua figlia Berenice, la quale aveva in sua
assenza usurpato il trono, era stata uccisa; che il nodo egiziano era
stato tagliato, e tutto ciò per opera del governatore della Siria, A.
Gabinio. Gabinio aveva agito, senza aspettare gli ordini del senato,
per incarico di Pompeo.
Note al Capitolo Secondo.
8. Sul numero degli Elvezi, cfr. G. Ferrero, Grandezza e Decadenza di

Roma, Milano, 1902, vol. II, p. 3.
9. Sull’interpretazione della complicata questione gallica in questo tempo,

cfr. G. Ferrero, Grandeur et Decadence de Rome, vol. II, App. D; Idem,
Le premier livre des Commentaires et les critiques de M. T. Rice Holmes,
in The Classical Quarterly, 1910, pp. 28 sgg.
10. La Gallia Transalpina fu veramente eretta a provincia nel 56 a. C. Cfr.

Cic., De prov. cons., 12, 29; 13, 32-33; 14, 34; Cic., ad fam., I, 7, 10;
Dio Cass., 39, 25, 1; G. Ferrero, in Revue arch., 1910 (15), pp. 93 sgg.
e, per una più ampia informazione C. Barbagallo, L’opera storica di G.
Ferrero e i suoi critici, Milano, Treves, 1911, pp. 197-209.
CAPITOLO TERZO
LA CRISI DELLA POLITICA CESARIANA
16. L’impresa di Britannia e la disfatta di Crasso in Oriente

(55-53 a. C.). — Ma le cose, per un istante composte, non
tardarono a guastarsi di nuovo. Mentre Pompeo e Crasso portavano
ad effetto in Roma gli accordi del congresso di Lucca, la Gallia
incominciava a dimostrare che era stata annessa a parole, non
conquistata con il braccio. Già nel 56 Cesare aveva dovuto
guerreggiare parecchi mesi per reprimere una insurrezione degli
Armorici e dei Veneti. Era riuscito a domarli; e aveva fatto gran
bottino, incominciando il saccheggio delle Gallie. Aveva anzi
preparato per il 55 l’invasione della Britannia.... Ma al momento di
muoversi, era stato trattenuto da un’invasione germanica, che
irrompeva in Gallia, non più paventata e maledetta, ma sollecitata
dagli indigeni, come l’avanguardia di un esercito liberatore. Erano
queste le orde degli Usipeti e dei Tencteri. Con la rapidità consueta
Cesare mosse contro gli invasori, intimando loro di ripassare il Reno;
con uno strattagemma poco leale, fece prigionieri i loro capi venuti a
lui in ambasceria, e condusse le sue legioni contro l’esercito
germanico rimasto senza duci. La sconfitta, che egli inflisse alle orde,
fu grave; e fu seguita da una sua breve incursione al di là del Reno:
ma questa guerra gli aveva fatto perdere troppo tempo perchè
Cesare potesse ancora, in quell’anno, tentare la divisata invasione
della Britannia. Si contentò dunque di fare nell’isola un rapido sbarco
con due legioni, e rimandò l’impresa all’anno seguente.
L’anno seguente, il 54, Roma tenterebbe dunque di conquistare in
Asia niente meno che l’impero dei Parti; in Europa, la grande isola
che un breve braccio di mare separava dalle coste della Gallia. Come
il Ponto e come l’Armenia, il regno di Parzia era sorto, fin dalla metà
del III secolo a. C., dal disfacimento del grande regno Seleucida. Ma
forse per la lontananza, sino a questo momento Roma e l’impero dei
Parti non erano stati nè amici nè nemici. Indifferente spettatore della
prima guerra mitridatica, l’impero dei Parti aveva accennato, durante
la seconda, ad allearsi con il Ponto e l’Armenia contro Roma, ma
aveva poi desistito; e, quando Pompeo aveva sostituito Lucullo, si
era messo da prima dalla parte di Roma, aiutandolo a conquistare
l’Armenia; poi aveva accennato a rivoltarsi, attaccando l’Armenia
cliente e vassalla di Roma; ma di nuovo aveva desistito. A sua volta
Roma aveva ondeggiato tra opposti propositi: Lucullo aveva pensato
a conquistare anche il regno dei Parti, ma era stato fermato dai
soldati in rivolta; Pompeo si era prima inteso con i Parti, poi, ad un
certo momento, aveva anch’egli inclinato alla guerra; finalmente si
era deciso per l’amicizia. Dopo il ritorno di Pompeo in Italia
l’incertezza aveva perdurato. C’era un partito che voleva vivere in
pace con i Parti; ed era forte assai in senato [11]. C’era un partito,
che voleva far loro la guerra rinnovando le gesta di Alessandro; e
contava i suoi campioni più ardenti tra gli ufficiali, che avevano
accompagnato Lucullo e Pompeo in Oriente, e tra gli amici di Crasso,
di Pompeo e di Cesare. Il convegno di Lucca aveva deliberato di dar
soddisfazione alla parte che voleva la conquista; e Crasso s’era
assunto la difficile impresa. Il suo disegno era, anzi, forse più vasto:
ricalcare le orme di Alessandro, penetrare nell’Iran e portare le
aquile romane fin sulle rive sacre dell’Indo. Partito da Roma sul finire
del 55, egli aveva imbarcate a Brindisi nove legioni di 3500 uomini
l’una, più 5000 cavalieri e 4000 ausiliari; in tutto 40.000 uomini; era
sbarcato a Durazzo e per la via Egnazia aveva attraversato
nell’inverno l’Epiro, la Macedonia, la Tracia; passato il Bosforo, era
entrato, nella primavera del 54, nella Siria settentrionale, aveva
rilevato Gabinio dal suo comando; poi, fatti gli ultimi preparativi,
aveva invaso, durante l’estate del 54, la Mesopotamia indifesa e
occupato parecchie città; allora si era fermato. Egli voleva — e il
piano era ingegnoso — attrarre il nemico in Mesopotamia, per non
essere attaccato troppo lontano dalle sue basi di operazione.
Frattanto Cesare aveva, nell’estate del 54, invasa la Britannia con
una grossa armata e cinque legioni; ma, pur essendo riuscito ad
internarsi nel paese al di là del Tamigi, e a vincere l’esercito nemico,
presto tornò sul continente, pago solo della vana promessa di un
tributo annuo. Le cose pericolavano troppo in Gallia, perchè egli
potesse impegnarsi sul serio a conquistare la Britannia. I Carnuti, i
Senoni, gli Aduatici, i Treviri, erano o in aperta rivolta o
irrequietissimi; cosicchè gli fu forza non solo rinunciare a passar
l’inverno, dal 54 al 53, nella Cisalpina, ma passarlo in Gallia,
combattendo e battagliando, come poteva, nella stagione cattiva:
per ricominciare poi con maggior vigore alla primavera del 53, nel
tempo stesso in cui Romani e Parti ripigliavano le armi in Oriente.
Qui la mossa di Crasso parve da prima riuscire. Nella primavera del
53 le guarnigioni romane, lasciate in Mesopotamia, erano assediate
dai Parti. I Parti venivano dunque a tiro.... Ma in verità il re dei Parti
aveva mandato quasi tutta la cavalleria, leggiera e pesante, sotto il
comando del Surena o generalissimo nella Mesopotamia, mirando
anch’esso ad attirare i Romani più lungi che potesse dalle loro basi di
operazioni. I due avversari impiegavano adunque lo stesso
strattagemma: ma i Parti in un terreno ad essi meglio noto e con un
esercito più addestrato a quel modo di combattere. Per maggior
disgrazia, Crasso si persuase troppo facilmente di avere ingannato il
nemico; e subito varcò l’Eufrate, per correre al soccorso delle città
assediate. Ma il nemico, appena i Romani ebbero varcato l’Eufrate,
levò subito l’assedio, e, come preso da panico, si ritirò
precipitosamente verso l’interno.
Questa ritirata mise in sospetto molti tra gli ufficiali di Crasso, i quali
consigliarono di fermarsi e di vedere più chiaro nelle intenzioni del
nemico. Ma Crasso cadde invece nel tranello; e credendo quello il
mezzo di finire presto la guerra, si slanciò alle calcagna dei Parti. Per
giorni e giorni li inseguì nel deserto, senza raggiungerli: ogni passo
avanti acuiva la smania di agguantare a ogni costo il fuggente
avversario; ma ogni giorno pure cresceva la stanchezza e
l’inquietudine dell’esercito.... Quando, un giorno, appena passata la
città di Carre, mentre i Romani erano per toccare le rive del Belik, i
Parti, fermatisi ad un tratto e voltatisi, offrirono battaglia. L’ora tanto
aspettata era giunta, ma sorprendeva l’esercito romano
stanchissimo. Gli ufficiali volevano rimandare la battaglia e
aspettare; ma Crasso, sempre pauroso che il nemico gli sfuggisse,
ordinò di attaccare. Fu uno strano combattimento, di un genere a cui
l’esercito romano non era adusato. La cavalleria pesante dei Parti
attaccava con impeto le coorti romane, eludendo con le sue manovre
qualunque contrattacco, mentre gli arcieri e i frombolieri a cavallo
riversavano incessantemente sul nemico una grandine di proiettili.
Questi ripetuti assalti non riuscirono a romper le legioni romane, ma
inflissero loro molte perdite; cosicchè, alla sera, Crasso dovette dare
l’ordine della ritirata. Ma l’esercito era scoraggiato, pervaso di strani
terrori, e il mobile nemico non gli dava tregua. Rapidamente la
disciplina si rallentò, i corpi incominciarono a sbandarsi e la ritirata a
volgersi in fuga; finchè un giorno i soldati, sobillati da emissari del
Surena, che prometteva loro di lasciarli ritornare tranquilli in patria,
se acconsentivano a trattar la pace, costrinsero Crasso a recarsi al
colloquio. Crasso diffidava del tranello; ma preferendo perire per
mano nemica che trucidato dai propri soldati, si recò al fatale invito,
e fu ucciso il 9 giugno del 53. Il capo fu inviato alla Corte del re dei
Parti; e le ossa non ebbero sepoltura. Dei soldati una parte si
disperse, o fu fatta prigioniera.
17. L’anarchia a Roma (54-53 a. C.). — Era una sciagura pari

alle maggiori sofferte nei secoli dalle armi romane. E la notizia
giunse a Roma nel mese di luglio, quando solo da poco tempo, dopo
sette mesi di interregno, si era riusciti a eleggere i magistrati per
l’anno stesso, che, secondo le leggi, avrebbero dovuto essere eletti
verso la metà dell’anno precedente. Come era accaduto un così
grande ritardo? Il governo personale di Crasso, di Pompeo e di
Cesare aveva di nuovo, come nel 58, sebbene per altra via, generato
l’anarchia. Soltanto il senato, con la sua autorità, poteva ancora
infrenare un poco e tenere a segno le ambizioni che scendevano
ogni anno a misurarsi nell’agone elettorale. Che pandemonio erano
diventate le elezioni, dopochè il senato era stato esautorato! In
quell’anno, poi, le candidature erano state così numerose; tanti gli
intrighi, le violenze, le corruzioni, le mine e le contromine dei
candidati e i loro accorgimenti d’ostruzione, che non si era potuto
eleggere nessun magistrato. Pompeo, che invece di andare in
Spagna era rimasto nei pressi della città, molle e irresoluto come al
solito, non aveva fatto nulla. Questi scandali avevano non poco
nociuto al governo e ai suoi capi, a Cesare soprattutto, che era il più
discusso e il più bersagliato. La rovina di Crasso non poteva non
accrescere questo malessere. Avevano dunque ragione Catone e i
suoi amici, che si erano sempre opposti alla spedizione e
all’avventurosa politica di tre troppo potenti capi della Repubblica!
Peggio ancora fu, quando le elezioni per l’anno 52 scatenarono di
nuovo l’anarchia. Erano candidati al consolato Milone, Publio Plauzio
Ipseo e Quinto Cecilio Metello Scipione, figlio adottivo di Metello Pio;
alla pretura l’immancabile Clodio; alla questura un ufficiale di Cesare
venuto apposta dalla Gallia, Marco Antonio. La gara delle ambizioni
infuriò di nuovo; Pompeo abbandonò Milone; Clodio per far dispetto
a costui sosteneva gli altri due candidati; e i candidati partigiani degli
uni e degli altri incominciarono a battagliare per le vie. Invano i
consoli tentarono a più riprese di tener i comizi; alla fine il Senato,
non Intendo altro, deliberò di proporre al popolo una legge, per la
quale un magistrato non avrebbe potuto ottenere una provincia, se
non cinque anni dopo esercitata la magistratura. Si sperava di
chetare così un poco la furibonda concorrenza alle magistrature. Ma
intanto si giunse alla fine dell’anno senza aver nominato i magistrati
e non si potè nemmeno nominare l’interrex, perchè un tribuno si
oppose. Quando, al principio del 52, e proprio il 18 gennaio, Clodio,
tornando con il suo seguito da Bovillae, si incontrò sulla via Appia
con Milone che, accompagnato dal suo seguito, andava a Lanuvio.
Le due parti vennero alle mani, e Clodio fu ammazzato. Questa
violenza bastò per scatenare la rivoluzione. Il popolino di Roma,
eccitato dai clienti di Clodio, dopo averne celebrato i funerali con una
pompa quasi selvaggia, appiccò le fiamme del rogo, nel quale il
corpo del suo idolo era scomparso, alla curia stessa del senato.
L’incendio si propagò alla basilica Porzia, e di qui ai maggiori e più
venerati monumenti romani: Roma fu per giorni e giorni piena di
tumulti, di incendi, di risse, di dimostrazioni, di grida....
18. La grande rivolta della Gallia (53-52 a. C.). — Roma non

era proprio la città in cui i governi personali potessero troppo
facilmente imporsi, dopo tanti secoli di governo aristocratico. Di
nuovo il sistema politico, immaginato da Cesare, si dissolveva. Perciò
appunto, sulla fine del 53, Cesare aveva lasciata la Gallia Transalpina
e si era avvicinato all’Italia, comprendendo che era necessario aiutar
Pompeo e la sua fazione a rimettere un po’ d’ordine nella turbata
repubblica. Ma Cesare aveva appena vôlto le spalle alla Transalpina,
ancora irrequieta, che i corrieri delle Gallie lo raggiunsero. L’incendio
riardeva: i Carnuti avevano trucidato i mercanti italiani ed erano
insorti di nuovo; gli Arverni avevano rovesciato il governo amico di
Roma, e condotti da un giovane principe, già amico di Cesare,
Vercingetorige, avevano innalzato la bandiera della rivolta; i Senoni, i
Parisii, i Pictoni, i Cadurchi, i Turoni, gli Aulerci, i Lemovici, gli Andi e
tutti i popoli abitanti sulle rive dell’Oceano si erano sollevati,
riconoscendo Vercingetorige come capo; i Sequani tentennavano; gli
Edui, rimasti fedeli, stavano per esser chiusi, e le legioni romane con
essi, come in un cerchio di ferro; un esercito già si avviava verso la
Gallia Narbonese, mentre un altro invadeva il territorio dei Biturigi,
tributari degli Edui. Cesare non esitò un istante: abbandonò l’Italia,
Roma e Pompeo al loro destino: e volò nella Narbonese. Rinforzò alla
meno peggio la difesa; poi con poche coorti, in pieno inverno,
valicando le Cevenne coperte di neve, si gettò sull’Arvernia, volendo
far credere al nemico che invadeva con grandi forze tutto il paese.
Infatti Vercingetorige, ingannato da questo attacco, accorse con
l’esercito in difesa della sua patria assalita. Allora Cesare, ripassate le
Cevenne e ritornato nella Provincia, con un piccolo corpo di
cavalleria che vi aveva lasciato, cavalcando notte e giorno, mentre gli
insorti lo credevano tra gli Arverni, potè arrivare inaspettato nel
paese dei Lingoni, mettersi a capo delle due legioni che vi
stanziavano, ordinare che le altre legioni, sparse per la Gallia, si
raccogliessero ad Agendicum (Sens). Così, in pochi giorni, si ritrovò a
capo del suo esercito: 35.000 uomini di fanteria, più gli ausiliari
gallici e qualche contingente di cavalleria. Erano queste tutte le
forze, di cui poteva disporre; e non erano molte: ma non c’era da
esitare.... L’audacia sola poteva salvarlo. Difatti, con un sì esiguo
esercito, in mezzo ad un paese in fiamme, Cesare prese una risoluta
offensiva. In pochi giorni attaccò e prese Vellaunodunum, incendiò
Genabum (Orléans), passò la Loira, entrò nel paese dei Biturigi,
assediò Noviodunum.
Vercingetorige, che non era riuscito a fermarlo, immaginò allora un
piano di guerra, che, applicato senza pietà, avrebbe potuto riuscire
rovinoso ai Romani; fare il vuoto intorno al nemico devastando i
paesi; molestarlo e affamarlo ogni giorno con assalti improvvisi di
cavalleria, con catture di convogli e di rifornimenti. Senonchè,
appena fu messo mano ad eseguire il piano, i Biturigi scongiurarono
l’implacabile eroe delle libertà galliche di risparmiare la loro capitale
Avarico, la futura Bourges, ch’essi s’impegnavano a difendere fino
all’estremo. Vercingetorige ebbe la debolezza di cedere; cosicchè
Cesare, invece di smarrirsi nel vuoto, correndo nel deserto dietro un
nemico inafferrabile, ebbe un punto saldo su cui dirigersi e colpire:
Avarico. Vercingetorige non osò soccorrerla; i Biturigi avevano troppo
presunto delle loro forze; Cesare con un vigoroso assedio di poche
settimane la prese, trucidò tutta la popolazione, e si impadronì di
tutte le provvigioni accumulate nella città.
Questa vittoria non permetteva soltanto a Cesare di riposare e
rifornire l’esercito nella ricca e ben provvista città; ma rialzava il
prestigio delle armi romane e scoraggiava gli insorti. Non aveva
Cesare distrutto Avarico, sotto gli occhi di Vercingetorige, senza che
costui osasse soccorrerla? Non era manifesto che Cesare era il più
forte? Cesare dovette illudersi addirittura d’aver vinto la guerra, se si
indusse a dividere le forze. Quattro delle sue dieci legioni
andrebbero, agli ordini di Labieno, contro i Sequani e i Parisii, che da
poco avevano ingrossato le file dell’insurrezione, mentre egli stesso
con sei legioni colpirebbe l’insurrezione al cuore, attaccando il
territorio degli Arverni e obbligando Vercingetorige ad accettare la
battaglia, che terminerebbe la guerra. E così fece. Invano
Vercingetorige tentò d’impedirgli di varcare l’Allier. Cesare eluse con
uno strattagemma la sua sorveglianza; e invase l’Arvernia, ponendo
l’assedio a Gergovia, per farle subire la stessa sorte di Avarico. Ma
egli non aveva più che sei legioni; e Gergovia resistette ostinata....
L’assedio andò per le lunghe; questo suo prolungarsi incominciò a
rianimare il coraggio dei Galli; Cesare volle finirla, e diede l’assalto
alla città; ma fu respinto con tali perdite, che dovette risolversi ad
abbandonare l’impresa e a riprendere la via del nord, per
ricongiungersi con Labieno.
Le conseguenze dell’errore furono assai funeste. Là sconfitta di
Gergovia, annunziata ovunque ed esagerata, scosse le popolazioni
rimaste fedeli; perfino gli Edui passarono al nemico, togliendo ai
Romani la miglior base di rifornimento e tagliando le comunicazioni
di Cesare con Labieno. Cesare capì che occorreva ricongiungersi al
più presto e a qualunque costo con Labieno: per non perdere tempo
a far dei ponti si cacciò nella Loira con l’esercito, e la passò a guado;
poi, risalendo verso il settentrione a marce forzate, raggiunse
Labieno, probabilmente ad Agendicum. Labieno aveva combattuto
con fortuna i Senoni e i Parisii: ma a che servivano queste vittorie?
Tutta la Gallia ormai era insorta; a Bibracte stava per radunarsi una
dieta nazionale, che chiamerebbe alle armi tutti i popoli gallici; che
fare con poco più di 30.000 uomini, in un paese tutto in rivolta?
Cesare deliberò di abbandonare per il momento la Gallia, ritirandosi
nella Provincia: ma il traversar la Gallia in fiamme con quel piccolo
esercito, gli parve impresa così pericolosa, che prima di muoversi
volle accostarsi alla frontiera occidentale della Germania allo scopo di
fare, tra quelle popolazioni germaniche, grandi leve di cavalleria. Il
generale, che sette anni prima era entrato in Gallia per distruggere il
pericolo germanico, intendeva ora servirsi dei Germani contro i Galli,
e pagava i primi con l’oro preso ai secondi.
Ma più che la cavalleria germanica, le discordie, gli errori e le
imprudenze del nemico salverebbero Cesare. Già a Bibracte i Galli
avevano acerbamente discusso sul comando e sul piano: se
continuare la guerriglia o fare la guerra grande. Un po’ per
contentare il partito che voleva la guerra grande e un po’ perchè la
guerriglia richiede poche truppe, ma buone, e Vercingetorige invece
comandava un esercito numeroso, raccogliticcio e scadente, il duce
arverno fu costretto a mutar il modo di guerreggiare, che sino ad
allora gli era riuscito così bene. Quando Cesare, probabilmente nella
prima metà dell’agosto, iniziò la sua ritirata verso la Provenza,
Vercingetorige, abbandonando il suo quartier generale, ch’egli aveva
stabilito in Alesia (Alise St. Reine nel dipartimento della Côte d’Or), e
la guerriglia, seguita fin allora, venne in campo aperto a contrastargli
il passo. Ma sopra un vero campo di battaglia le legioni romane, il
genio del loro duce, l’impeto dei cavalieri germanici ebbero ragione
dell’attacco e della resistenza nemica [12]. E bastò questa battaglia
per mutare le sorti della guerra. Vercingetorige, sconfitto, riparò in
Alesia; Cesare, smessa l’idea di ritirarsi in Provenza, si volse subito
ad investire Alesia; Vercingetorige, chiuso da Cesare entro
giganteschi lavori, chiamò in suo soccorso la Gallia intera.... Si
raccolse così un nuovo grande esercito — oltre 250.000 uomini, dice
Cesare [13] — i quali avrebbero dovuto piombare sull’esercito romano
dal di fuori, mentre da Alesia gli assediati avrebbero fatto l’ultima
sortita della disperazione. Se il piano riusciva, l’esercito romano
sarebbe stato distrutto da un doppio assalto. Non sentendosi la forza
di resistere a questo doppio assalto in campo aperto, Cesare non
esitò a costruire, intorno alla linea delle prime trincee, una seconda
grandiosa opera di fortificazioni, dietro la quale il suo esercito,
assediante e assediato nel tempo stesso, avrebbe potuto resistere al
nuovo nemico. L’espediente era nuovo, ingegnoso, ma temerario; e
non avrebbe sortito alcun buon effetto, se l’esercito di soccorso
avesse assediato con pazienza l’esercito romano, anche a costo di
far morire di fame insieme e nella stessa cerchia Cesare e
Vercingetorige. Ma i duci erano parecchi e discordi fra loro; il
desiderio di salvare Vercingetorige, troppo vivo; il nuovo esercito,
raccogliticcio, impaziente e mal provvisto di materiali e di viveri.
Invece di assediare pazientemente gli assedianti, l’esercito di
soccorso volle far presto, tentò di forzare il campo di Cesare; si
esaurì per sette giorni in furiosi ma vani assalti; poi si sbandò. Allora
Vercingetorige, vinto dalla fame, si consegnò prigioniero nelle mani
del vincitore (settembre 52).
La insurrezione gallica era domata; Cesare era quasi per miracolo
scampato alla sorte di Crasso; la nuova provincia gallica, dopo otto
anni di insurrezioni e di guerre continue, era salva. Il conquistare
l’Occidente era impresa più ardua e di maggior sacrificio dell’Oriente:
ma quanto più duratura e proficua nel lontano avvenire! [14].
Note al Capitolo Terzo.
11. Questo partito aveva riportato una bella vittoria nel 55. Poichè il
governatore della Siria, A. Gabinio, meditava la guerra alla Parzia, il
Senato gli oppose un fermo divieto: cfr. Strab., 12, 3, 34.
12. Il luogo della battaglia è incerto: secondo alcuni, essa seguì sulle rive
della Vingeanne; secondo altri, tra Brevon e l’Ource; secondo altri,
infine, o nei pressi di Montigny, o non lungi da Allofroy.
13. Cfr. Caes., B. G., 7, 75-76. Sebbene la lunga serie di cifre, da cui questo
totale risulta possa in qualcuno dei suoi addendi esserci pervenuta
corrotta, tuttavia il risultato finale non è mai inferiore a 250.000.
14. Su questo periodo di storia romana, cfr. la più particolareggiata

esposizione di G. Ferrero, Grandezza e Decadenza di Roma, vol. II, cap.
IV-VII.
CAPITOLO QUARTO
LA SECONDA GUERRA CIVILE
(49-46 a. C.)
19. Il consolato unico di Pompeo (52 a. C.). — A Roma intanto,

mentre Cesare combatteva in Gallia, Pompeo e il partito del senato,
spaventati dai tumulti che continuavano, dallo sfacelo della
repubblica, dalla catastrofe di Crasso, dalla rivolta della Gallia,
avevano un po’ dimenticato gli odî e i ripicchi antichi. Anche i più
ostinati avversari della triarchia erano stati ammansati dal pericolo;
quanto a Pompeo, era troppo ricco, troppo potente, troppo viziato
dalla fortuna, da non voler primeggiare piuttosto con il favore del
senato, che a suo dispetto. Così, perdurando ed imperversando i
tumulti, la proposta di nominare Pompeo non dittatore — chè il
nome dopo Silla era odioso — ma console unico, fu approvata da
tutti, anche da Catone. Pompeo a sua volta si affrettò a contentare
la parte più autoritaria del partito senatorio, attuando in poche
settimane ciò che quella chiedeva invano da anni. Con una legge de
ambitu e un’altra de vi, abbreviò la durata dei processi, aggravò le
pene ai delitti di corruzione politica, commessi sin dal 70, rinvigorì e
accelerò la procedura contro le violenze commesse nelle elezioni,
diede una spinta vigorosa ai processi. In un batter d’occhio un gran
numero di partigiani di Clodio e di Cesare furono condannati insieme
con qualcuno dei più turbolenti tra i loro avversari. Neanche Milone,
l’antico amico di Pompeo, fu risparmiato. Appropriandosi poi una
proposta fatta l’anno prima, ma inutilmente, dal senato, Pompeo
propose una lex de provinciis, la quale vietava che nessun console o
pretore romano diventasse governatore di una provincia, se non
cinque anni dopo la fine della sua magistratura. Presentò inoltre una
lex de iure magistratuum, la quale riconfermò l’antico divieto di
brigare il consolato a chi fosse assente da Roma; ad eccezione di
coloro che avessero ricevuto o ricevessero dal popolo la dispensa.
Questa eccezione toccava Cesare, che poco prima una legge,
proposta dai suoi amici, aveva autorizzato a presentarsi candidato
per il 48, senza essere presente in Roma.
L’ordine fu ristabilito in Roma; il senato respirò; Pompeo ritornò in
credito, come un secondo Silla, presso quella parte della aristocrazia,
che aveva subìto, ma non accettato, il governo della triarchia. Senza
discussione, gli fu prorogato di cinque anni il governo della Spagna.
Vacillò invece la potenza di Cesare. Il governo da lui fondato
pericolava. La morte di Crasso prima, la lenta conversione di Pompeo
poi, il disastro partico, la rivolta della Gallia, l’anarchia di Roma
avevano prima screditato e poi disciolto la triarchia. Della antica
potenza dei tre capi non restava più che un odio implacabile, tutto
addensato su lui, poichè Pompeo si era riconciliato con i nemici.
Cesare aveva ragione di temere che se, finito il proconsolato, egli
tornasse a Roma semplice cittadino, i suoi nemici gli intenterebbero
qualche processo che, rovinandolo, farebbe scontare a lui, con le
sue, le colpe di Crasso e di Pompeo [15]. Non c’era che uno scampo:
essere rieletto console, e farsi accordare un nuovo e lungo
proconsolato; poichè ogni magistrato era, sinchè copriva la carica,
inviolabile. Ma i suoi poteri proconsolari duravano sino al 1º marzo
del 49 a. C., che era il decimo anniversario del giorno in cui la lex
Vatinia gli aveva assegnato la Gallia. Cesare non poteva dunque
brigare il consolato che nelle elezioni che avrebbero luogo durante
l’anno 49 e farsi nominar console per il 48; onde nei dieci mesi che
correrebbero tra il 1º marzo del 49 e il 1º gennaio del 48, rientrando
nella vita privata, sarebbe stato facile bersaglio ai processi dei
nemici. Una legge gli aveva, sì, concesso di brigare il consolato
assente da Roma: ma che gli serviva, se i suoi poteri spiravano il 1º
marzo?
20. Il conflitto tra Cesare ed il Senato (51-49 a. C.). — Cesare

pensò di chiedere al senato, al principio dell’anno 51, che gli
prolungasse i poteri proconsolari dal 1º marzo del 49 al 1º gennaio
del 48, allegando che questo prolungamento era implicito nella legge
che gli concedeva di postulare il consolato senza essere presente a
Roma. La salvezza di Cesare dipendeva da questa domanda, che a
sua volta dipendeva da Pompeo. Il senato l’approverebbe o la
respingerebbe, se Pompeo l’appoggiasse o la combattesse. Pompeo
quindi fu corteggiato con zelo eguale dagli amici e dai nemici di
Cesare. Ma Pompeo, sebbene ormai fosse avverso a Cesare e
favorevole al partito senatorio [16], allorchè, in aprile, la domanda di
Cesare fu discussa in senato, non si pronunciò, e un tribuno della
plebe lo tolse dall’impaccio di dover dichiararsi, interponendo il veto.
Ma Cesare aveva nemici fanatici; e tra questi c’era il console
Marcello; il quale risollevò il 1º giugno la questione, proponendo
addirittura di richiamare Cesare dalla Gallia. Questa volta Pompeo
non potè più tacere; ma si cavò d’impaccio, dicendo che non si
poteva trattar della successione di Cesare prima del 1º marzo
dell’anno 50. Il senato gli diede ragione, e Marcello ammutolì; ma
per risollevare la questione a suo tempo e proprio il 30 settembre,
proponendo che il 1º marzo dell’anno seguente si discutesse in
senato la successione di Cesare; e che si dichiarasse nullo in
precedenza ogni veto che i tribuni interponessero. Queste proposte
furono occasione di un vivace dibattito; la prima fu approvata e la
seconda sospesa dal veto tribunizio; ma il vero guadagno della
seduta per i nemici di Cesare fu che questa volta Pompeo dovette
aprirsi; e lo fece, sentenziando che, se il 1º marzo i tribuni amici di
Cesare avessero fatto uso dell’intercessione, Cesare doveva
considerarsi e «castigarsi» come ribelle.
La fortuna di Cesare, che frattanto domava le ultime resistenze della
Gallia, pericolava. Pompeo lo abbandonava; e gli aveva ormai quasi
spezzato in mano l’arma del veto. Se avesse cercato di
scaramucciare con il veto dei tribuni amici suoi, dopo quella
dichiarazione, si sarebbe guastato apertamente con Pompeo. Questo
Cesare non voleva; onde immaginò un curioso espediente. Era stato
eletto tribuno per l’anno 50 a. C. Scribonio Curione, un giovane
pieno d’ingegno e di debiti, grande oratore e scrittore, e acerrimo
nemico di Cesare. Promettendo di pagargli i debiti, Cesare ottenne di
trarlo dalla sua parte e di fargli accettare una missione difficilissima:
quella di impedire il 1º marzo la discussione sulla sua provincia,
fingendo di adoperarsi come nemico suo, ma non dell’equità e della
costituzione. Se l’intercessione di un tribuno a lui nemico avesse
impedito la votazione, come avrebbe Pompeo potuto risentirsene
contro di lui? E Curione disimpegnò mirabilmente il suo bizzarro
incarico. Affermando, con affettata imparzialità, che era tempo di
finirla con tutti i poteri straordinari, sia di Cesare che di Pompeo;
presentando leggi opposte, talune di spirito oligarchico, altre a
seconda dell’umore popolare; atteggiandosi a difensore imparziale
della legge e della pubblica pace; attaccando Cesare, ma nello stesso
tempo Pompeo, Curione riuscì a diventar così popolare presso il
pubblico, che voleva la pace, da poter far differire, con il veto e con
altri espedienti, di mese in mese, sino alla fine del 50, ogni dibattito
sulla successione di Cesare. Pompeo, che una malattia aveva
condannato all’inerzia per parecchi mesi, ne approfittò volentieri per
fare le viste di dimenticare le minacce pronunciate nella seduta del
30 settembre; e la maggioranza del senato gli fu grata di differire la
terribile questione. Ma questi abili maneggi, nei quali non si tardò a
sospettare la mano di Cesare, esasperarono i nemici del proconsole
e lo stesso Pompeo. D’altra parte, avvicinandosi la fine dei poteri
proconsolari di Cesare, era necessario definire la questione.
Si venne così alla storica seduta del 1º dicembre 50. Il console
Marcello cominciò a proporre che Cesare cessasse dai suoi poteri
proconsolari il 1º marzo del 49. La proposta fu approvata a grande
maggioranza, e senza che Curione aprisse bocca. Marcello allora,
incalzando, chiese al senato se anche Pompeo dovesse rassegnare il
comando delle Spagne, che, come abbiamo visto, gli era stato
prorogato fino al 45. La nuova proposta fu respinta a grande
maggioranza. Solo allora Curione domandò la parola; e,
introducendola con un discorso abilissimo, fece una terza proposta,
che, a stretto rigore di logica, contradiceva alle deliberazioni già
prese: Pompeo e Cesare abbandonassero insieme il loro governo
proconsolare. La proposta, rispondeva talmente al desiderio di tutti,
— senatori e popolo — che l’assemblea, contradicendosi, l’approvò
con 370 voti contro 22. La deliberazione era savia; ma umiliava
troppo i nemici di Cesare, che non la volevano a nessun costo, e
Pompeo, che non intendeva deporre prima del tempo il potere che il
senato gli aveva prolungato. In fretta e furia Marcello e i più scaldati
nemici di Cesare immaginarono un piano, lo sottoposero a Pompeo,
il quale era ancora a Napoli: Marcello avrebbe proposto al senato di
dichiarare Cesare hostis publicus; se il senato non avesse approvato
o se i tribuni avessero interposto il veto, egli avrebbe di sua autorità
proclamato lo stato d’assedio e affidato a Pompeo la salvezza dello
Stato: il senato allora, intimidito, avrebbe approvato quanto essi
volevano. Non appena giunse da Napoli l’approvazione di Pompeo —
probabilmente il 9 dicembre — Marcello fece il suo colpo di Stato.
Convocò il senato; propose di dichiarare Cesare nemico pubblico, e
di ordinare a Pompeo di prendere il comando delle due legioni, che a
Lucera aspettavano di partire per la Siria; e, quando Curione ebbe
posto il suo veto, uscì da Roma, e si recò a Napoli da Pompeo, per
invitarlo ad assumere, novello Nasica ed Opimio e con gli stessi
mezzi, la difesa della repubblica.
21. Dal Rubicone a Brindisi (10 gennaio-17 marzo 49 a. C.).

— Le cose precipitavano. C’era però ancora una speranza. Cesare
voleva la pace. Voleva la pace perchè sapeva che delitto e che
pericolo sarebbe scatenare una seconda guerra civile, non più
nemmeno per le grandi questioni politiche che avevano preparato la
prima, ma per i miserabili puntigli di due cricche di politicanti.
Deliberò dunque di fare uno sforzo supremo per la pace. Curione,
che era uscito di carica subito dopo la votazione del senato, si era
recato da lui: Cesare lo rimandò con una lettera al suo ex-ufficiale,
ed ora tribuno, Marco Antonio, da leggersi in senato. In questa
lettera egli si dichiarava pronto ad abbandonare il comando della
Gallia e a tornare privatamente a Roma, purchè Pompeo facesse
altrettanto. In caso contrario, soggiungeva, egli avrebbe difeso i suoi
diritti violati. La lettera era scritta con rispettosa fermezza, e Cesare
si riprometteva che farebbe riflettere senza irritare. Non aveva il
senato mostrato, nella seduta del 1º dicembre, che voleva conciliare
il dissidio dei due personaggi e delle due fazioni, con un
provvedimento equo? Senonchè nel frattempo Pompeo aveva
accettato la missione di difendere la repubblica affidatagli da
Marcello, e preso il comando delle legioni di Lucera. Di più i nemici di
Cesare non erano stati inoperosi. Il resultato fu che il senato, nella
seduta del 1º gennaio 49, non si comportò più come un mese prima:
la lettera di Cesare fu accolta da interruzioni e da proteste, come
una minaccia; e Cesare fu dichiarato nemico pubblico, se non avesse
abbandonato il comando entro il luglio. Qualche giorno dopo il
senato dichiarava lo stato di assedio. Il rimedio di Cesare per salvare
la pace era fallito! Non volendo cedere, Cesare non potè che dar di
piglio a un mezzo estremo, un’arme a doppio taglio, l’unica, che
ormai gli restava: dimostrare di essere risoluto a tutto, e far rinsavire
con le minacce il senato, che aveva respinto le proposte concilianti.
Una notte, verso il 10 gennaio, uscì da Ravenna con 1500 uomini, e,
violando la frontiera, che separava l’Italia dalla sua provincia, valicò il
Rubicone, occupò di sorpresa Rimini e nei giorni seguenti Pesaro,
Fano, Ancona e le principali città della costa, spingendo qualche
coorte verso Arezzo.
La seconda guerra civile incominciava, sebbene nessuna delle due
parti l’avesse voluta sul serio; e sebbene l’Italia tutta avesse sempre
e soltanto implorato la pace [17], perchè nessuna jattura poteva
esserle in quel momento più funesta di una guerra civile. Tre anni
prima, nel 52, i mercanti italiani erano riusciti per la prima volta ad
esportare nelle province l’olio fabbricato in Italia. Basta questo fatto
a mostrare che l’Italia non era tutta piena di proprietari rovinati, di
latifondisti o di inquieti e famelici politicanti; ma che c’era anche chi
lavorava — media possidenza i più — e con i capitali, il lavoro e gli
schiavi importati di Grecia e dall’Oriente, tentava di coltivar meglio la
terra sull’esempio dei popoli più esperti in agricoltura. Nel tempo
stesso si affermava l’industria; e anche questa, in parte, grazie agli
schiavi e ai liberti orientali. Nella Cisalpina, da Vercelli a Milano, da
Milano a Modena; nell’Etruria, ad Arezzo, si cominciavano ad aprire
quelle fabbriche di ceramica, di lampade, di anfore, che diverranno
in seguito famose. A Padova e a Verona degli artigiani e dei mercanti
cominciavano a tessere quei tappeti e quelle coperte, di cui tutta
l’Italia dovrà fra non guari fare così largo uso. A Parma e a Modena
si tessevano panni magnifici, con la lana delle numerose greggi
pascolanti nelle campagne circostanti. A Faenza si cominciava a filare
e a tessere il lino, coltivato nei dintorni. Genova, a pie’ delle
montagne selvagge della Liguria, era un emporio di legname, di pelli,
di miele, di bestiame, che i Liguri trasportavano e conducevano dalle
loro valli solitarie. Le miniere di ferro dell’Elba erano sfruttate con
lena vigorosa; e Pozzuoli lavorava il ferro dell’Elba fabbricando ogni
sorta di oggetti. Napoli era la città dei profumi e dei profumieri;
Ancona possedeva fiorenti tintorie di porpora. Le città si ampliavano,
si abbellivano, arricchivano, e in quelle cresceva di numero, di
agiatezza e di potenza un nuovo ceto medio. Con l’agiatezza, con il
nuovo bisogno di pace operosa, con la partecipazione delle classi
minori e degli Italici alla vita pubblica, gli odî di un tempo si erano
placati. Non più la ferocia delle antiche lotte dei plebei contro i
patrizi, dei poveri contro i ricchi, degli Italici contro i Romani. Unica
angustia, i debiti. Senonchè neppure questo universale desiderio di
pace valse contro i rancori e i puntigli dei partiti politici. A furia di
spaventarsi a vicenda con minacce, i partiti resero alla fine la guerra
inevitabile. La mossa di Cesare, che mirava ancora a persuadere i
nemici ad una transazione, fallì il suo effetto, non perchè non
spaventasse abbastanza, ma perchè spaventò troppo. Quando si
seppe a Roma che Cesare aveva occupato Rimini, Ancona, Arezzo,
tutti credettero che volesse marciare con le legioni su Roma; un gran
panico scoppiò; e se qualche spaventato propose di aprire trattative
di pace, Pompeo non ne volle sentir parlare: ordinò anzi che il
senato e i consoli lasciassero Roma e si ritirassero a Capua. Cesare,
che voleva intendersi con il senato e finir presto l’avventura con una
transazione, capì che quella fuga gli accrescerebbe la difficoltà di far
pace; e cercò, con lettere e con quanti mezzi aveva a mano, di
persuadere i senatori a restare in Roma. Ma intanto nel Piceno e nel
Sannio i generali di Pompeo reclutavano soldati: poteva Cesare
lasciarsi crescer sul fianco questa minaccia? Egli richiamò dalle Gallie
le sue legioni, e procedè innanzi: prese Osimo, Cingoli; si impadronì
del Piceno, obbligando i generali di Pompeo che reclutavano soldati a
ripiegare su Corfinio, nel paese degli antichi Peligni, dove si
raccoglieva buon nerbo di milizie, sotto il comando di uno dei più
autorevoli pompeiani, L. Domizio Enobarbo, console nel 54. Ma
poteva Cesare lasciar che Corfinio diventasse un forte punto di
appoggio per Pompeo? Con la consueta rapidità e con le legioni
giuntegli dalle Gallie, alle quali aveva fatto grandi promesse, Cesare
marciò su Corfinio, la assediò, costringendo, dopo soli sette giorni,
Domizio alla resa. Ma voleva intendersi con i nemici, e fu generoso;
mandò liberi Domizio e i nobili pompeiani, ch’erano al suo seguito.
In meno di due mesi con la sua rapida marcia e con la vittoria di
Corfinio, Cesare era riuscito a sconvolgere quella che oggi noi
chiameremmo la mobilitazione del partito avverso, ossia il
reclutamento con cui cercava di levar soldati in Italia. Pompeo in due
mesi di guerra aveva perduto una buona parte della penisola ed era
in pericolo di esser sopraffatto dalle forze soverchianti di Cesare,
perchè aveva in Italia poco più delle due legioni di Lucera e le sue
comunicazioni con la Spagna, dove stavano le sue migliori legioni,
erano minacciate. Tanto più avrebbe dovuto prestare orecchio alle
offerte di pace, che Cesare, spaventato dal precipitar degli eventi,
faceva per differenti canali. Ma ormai era impegnato; e non voleva
parere di aver accettato da Cesare una pace, perchè vinto. Poichè
mezza Italia era perduta; poichè con le forze, di cui disponeva, non
poteva riconquistarla e riaprirsi le comunicazioni con la Spagna,
Pompeo deliberò di abbandonare l’Italia con il senato, i magistrati e
l’esercito, e di salpare da Brindisi alla volta dell’Oriente; dove le
province e i re alleati non avrebbero indugiato ad aiutarlo a rifarsi un
esercito. Ma quando Cesare conobbe questo disegno si spaventò;
capì che una terribile guerra civile avrebbe devastato tutto l’impero,
se egli non riusciva a far la pace con Pompeo in Italia; e a marce
forzate corse su Brindisi, per bloccare il suo avversario e finire la
guerra. Ma non fece a tempo. Pompeo, il senato, l’esercito,
riuscirono ad imbarcarsi, abbandonando a Cesare l’Italia.
22. La guerra di Spagna (marzo-novembre 49 a. C.). — Ormai

il destino si era compiuto. Cesare doveva combattere una immensa
guerra civile — la seconda della storia di Roma. Ma in quale
spaventoso impegno s’era cacciato! Era abbandonato, solo, alla testa
del suo esercito, nell’Italia senza magistrati e separato dalle sue
maggiori province! Cesare non si perdè d’animo, e soprattutto non
perdè tempo; subito spedì quante forze potè ad occupare la
Sardegna, la Sicilia e l’Africa; e senza indugio si recò a Roma per
riorganizzare alla meno peggio il governo e per rifornirsi di danaro.
Ci giunse verso gli ultimi giorni di marzo; racimolò quei pochi
senatori, che erano rimasti, e li considerò come il senato legittimo;
d’accordo con loro provvide alla meglio a sostituire i magistrati che
mancavano; prese diversi provvedimenti a favore del popolo; fece
abrogare la legge di Silla, che escludeva dalle magistrature i
discendenti dei proscritti; e infine si impadronì dell’erario,
minacciando di trucidare un tribuno, L. Cecilio Metello, che voleva
impedirglielo. Poi, dopo un soggiorno di pochi giorni, ripartì per la
Spagna.
Il piano di guerra di Cesare era semplice e ardito: volare in Spagna,
debellare il nucleo maggiore e migliore delle forze pompeiane, poi
recarsi in Grecia a combattere il nuovo esercito che Pompeo
raccoglierebbe. Ma per riuscire, gli occorreva far presto. Invece
subito egli trovò sulla via della Spagna un primo intoppo: Marsiglia.
Città libera, ma devota a Pompeo, Marsiglia intendeva restar neutra
nel conflitto. Cesare richiamò tre legioni dalla Gallia e pose l’assedio

(Ebooks PDF) Download Theoretical Ecology Principles and Applications 3rd Edition Robert May Full Chapters

Uploaded by

Copyright:

Available Formats

(Ebooks PDF) Download Theoretical Ecology Principles and Applications 3rd Edition Robert May Full Chapters

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Ebooks PDF) Download Theoretical Ecology Principles and Applications 3rd Edition Robert May Full Chapters

Uploaded by

Copyright:

Available Formats

Full download ebook at ebookgate.

Theoretical Ecology Principles and

Download more ebook from https://ebookgate.com

Ecology Principles and Applications 2nd Edition Edition

Macroeconomics Principles and Applications 6th Edition

Terrestrial Ecosystem Ecology Principles and

Behavioural Ecology of Insect Parasitoids From

Principles of Tissue Engineering 3rd Edition Robert

Psychological Testing Principles Applications and

Psychological Testing Principles Applications and

Psychological Testing History Principles and

ISBN 978–0–19–920998–9 (Hbk) 978–0–19–920999–6 (Pbk)

2 How populations cohere: ﬁve rules for cooperation 7

4 Metapopulations and their spatial dynamics 35

6 Plant population dynamics 62

7 Interspeciﬁc competition and multispecies coexistence 84

8 Diversity and stability in ecological communities 98

9 Communities: patterns 111

10 Dynamics of infectious disease 132

12 A doubly Green Revolution: ecology and food production 158

13 Conservation biology: unsolved problems and their policy implications 172

14 Climate change and conservation biology 190

15 Unanswered questions and why they matter 205

How populations cohere: ﬁve

Figure 2.2 Win-stay, lose-shift (WSLS) embodies a very simple

Standing Both standing and judging belong to the leading

and Gintis, 2004; Traulsen et al., 2005). We only

Population density time t

Population density time t+ 1

Population density time t Population density time t

because whatever the initial population density the

different starting points). But depending on the

C between scramble competition where resources are

individuals get the resources they require, and

Population density time t + 1

Population density time t Population density time t

Population density time t

Metapopulations and their

4.1 Introduction (Hanski, 1999). This Scandinavian butterﬂy lives

Table 4.1 Some less obvious examples of Levins metapopulations.

Example Local population Patch Colonization Extinction

Territorial species Breeding pair Territory Occupation of empty Death

Colonization or extinction rate

11. La guerra contro Ariovisto (58 a. C.). — La prima impresa di

12. Il richiamo di Cicerone e la prima crisi della triarchia. —

13. La guerra contro i Belgi (57 a. C.). — Insomma, le cose

14. L’annessione della Gallia (56 a. C.). — Senonchè mentre

15. Il convegno di Lucca e il consolato di Crasso e di Pompeo

8. Sul numero degli Elvezi, cfr. G. Ferrero, Grandezza e Decadenza di

9. Sull’interpretazione della complicata questione gallica in questo tempo,

10. La Gallia Transalpina fu veramente eretta a provincia nel 56 a. C. Cfr.

16. L’impresa di Britannia e la disfatta di Crasso in Oriente

17. L’anarchia a Roma (54-53 a. C.). — Era una sciagura pari

18. La grande rivolta della Gallia (53-52 a. C.). — Roma non

14. Su questo periodo di storia romana, cfr. la più particolareggiata

19. Il consolato unico di Pompeo (52 a. C.). — A Roma intanto,

20. Il conflitto tra Cesare ed il Senato (51-49 a. C.). — Cesare

21. Dal Rubicone a Brindisi (10 gennaio-17 marzo 49 a. C.).

22. La guerra di Spagna (marzo-novembre 49 a. C.). — Ormai