Evolution E-Book

Wallace Arthur

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Sprache: Englisch

Seitenzahl: 725

Veröffentlichungsjahr: 2010

Beschreibung

This book is aimed at students taking courses on evolution in universities and colleges. Its approach and its structure are very different from previously-published evolution texts. The core theme in this book is how evolution works by changing the course of embryonic and post-embryonic development. In other words, it is an evolution text that has been very much influenced by the new approach of evolutionary developmental biology, or 'evo-devo'.

Key themes include the following: developmental repatterning; adaptation and coadaptation; gene co-option; developmental plasticity; the origins of evolutionary novelties and body plans; and evolutionary changes in the complexity of organisms. As can be seen from this list, the book includes information across the levels of the gene, the organism, and the population. It also includes the issue of mapping developmental changes onto evolutionary trees. The examples used to illustrate particular points range widely, including animals, plants and fossils.

"I have really enjoyed reading this book. One of the strengths of the book is the almost conversational style. I found the style easy to read, but also feel that it will be invaluable in teaching. One of our tasks in university level teaching is to develop students' critical thinking skills. We need to support them in their intellectual development from a "just the facts" approach to being able to make critical judgements based on available evidence. The openness and honesty with which Arthur speaks to uncertainty in science is refreshing and will be a baseline for discussions with students."
-Professor Patricia Moore, Exeter University 

"This book, written as an undergraduate text, is a really most impressive book. Given the burgeoning interest in the role of developmental change in evolution in recent times, this will be a very timely publication. The book is well structured and, like the author's other books, very well written. He communicates with a clear, lucid style and has the ability to explain even the more difficult concepts in an accessible manner."
---Professor Kenneth McNamara, University of Cambridge

The companion site can be found at www.wiley.com/go/arthur/evolution. Here you download all figures from the book, captions, tables, and table of contents.

Leseprobe

Contents

Preface

PART I Foundations

1 Introduction

1.1 From Darwin to Development

1.2 Development; and Evolutionary Changes in Development

1.3 Development and the Realm of Multicellularity

2 What is Evo-Devo?

2.1 Forerunners of Evo-Devo

2.2 Nineteenth-Century Comparative Embryology

2.3 Diverse Antecedents—1900–1980

2.4 Conclusions from History; Messages for the Present

2.5 The Advent of Evo-Devo in the 1980s

2.6 Broad and Narrow Views of Evo-Devo

2.7 Too Few Laws, Too Many Facts?

3 Development, Cells and Molecules

3.1 Analysing the Developing Organism

3.2 Cells and Development: The Basics

3.3 Genes: Structure, Expression and Developmental Function

3.4 Signalling Pathways Within and Between Cells

3.5 Signalling: From Cell to Embryo

3.6 Long-Range Signalling and Developmental Processes

4 Natural Populations

4.1 The Ecological Theatre and the Evolutionary Play

4.2 Types of Creature; Types of Population

4.3 Spatial Structure

4.4 Age Structure

4.5 Genetic Structure

4.6 Natural Selection

PART II Developmental Repatterning

5 Mutation and Developmental Repatterning

5.1 Mutation in Terms of Altered DNA Sequence

5.2 Mutation in Terms of Proximate Functional Consequences

5.3 Developmental Repatterning at Molecular and Higher Levels

5.4 Developmental Repatterning at the Level of the Whole Organism

5.5 Developmental Repatterning and Fitness

6 Heterochrony

6.1 What is Heterochrony?

6.2 Types and Levels of Heterochrony

6.3 Heterochrony at the Organismic Level

6.4 Heterochrony at the Molecular Level

6.5 Heterochrony and Fitness

7 Heterotopy

7.1 What is Heterotopy?

7.2 Heterotopic Processes Involving Left-Right Asymmetry

7.3 Heterotopic Processes Involving the A-P and D-V Axes

7.4 Other Types of Heterotopy

7.5 Concluding Remarks

8 Heterometry

8.1 What is Heterometry?

8.2 Increasing Relative Size

8.3 Decreasing Relative Size

8.4 Bi-directional Heterometry

8.5 Heterometric Compensation

9 Heterotypy

9.1 What is Heterotypy?

9.2 Altered Products of Developmental Genes

9.3 Altered Pigmentation

9.4 Altered Morphology and the Origin of Novelty

9.5 The Origin of New Cell Types

10 The Integrative Nature of Repatterning

10.1 Repatterning is a Complex Process

10.2 Different Kinds of Repatterning can Produce a Similar Result

10.3 Compound Repatterning at a Single Level of Organisation

10.4 The Kind of Repatterning can Change between Levels of Organisation

10.5 Categories and Subcategories of Repatterning

10.6 The Causes of Repatterning

11 Mapping Repatterning to Trees

11.1 Pattern, Process, Homology and Trees

11.2 The Origin(s) of Animal Segmentation

11.3 The Vertebrate Fin-to-Limb Transition

11.4 The Origin of Flowers

11.5 General Conclusions on Repatterning and Selection

PART III The Direction of Evolution

12 Adaptation, Coadaptation and Exaptation

12.1 Natural Selection on a Continuously Variable Character

12.2 Natural Selection on Two Characters; and the Idea of an Adaptive Landscape

12.3 Developmental and Functional Coadaptation

12.4 Morphological Geometry and Selection

12.5 Long-term Evolution and Exaptation

13 Developmental Bias and Constraint

13.1 A Key Question about Evolution’s Direction

13.2 Making Sure the Question is about Processes, not Terminology

13.3 Dependence versus Independence of Different Characters

13.4 Evo-Devo Meets Quantitative Genetics

13.5 Developmental Bias and ‘Routine’ Evolution

13.6 Developmental Bias and the Origin of Evolutionary Novelties

14 Developmental Genes and Evolution

14.1 The Direction of Evolution at the Developmental/Genetic Level

14.2 Developmental Genes: An Overview

14.3 Developmental Genes: Examples

14.4 The Hox Genes

14.5 Gene-Level Forms of Developmental Bias and Coadaptation

14.6 Changes in Regulatory versus Coding Regions of Genes

15 Gene Co-option as an Evolutionary Mechanism

15.1 What is Gene Co-option?

15.2 Co-option in the Evolution of Segments and Eyes

15.3 Appendage Evolution and Gene Co-option

15.4 Co-option in the Evolution of Zygomorphic Flowers

15.5 Evolution of the ‘Genetic Toolkit’

15.6 Co-option, Exaptation and Developmental Bias

16 Developmental Plasticity and Evolution

16.1 Types of Developmental Plasticity

16.2 Discrete Variants: Winged and Wingless Forms of Insects

16.3 Meristic Variation: the Number of Segments in Centipedes

16.4 Continuous Variation: Plant Growth

16.5 Plasticity and Developmental Genes

16.6 The Evolution of Patterns of Plasticity

17 The Origin of Species, Novelties and Body Plans

17.1 Is Evolution Scale-dependent?

17.2 Speciation

17.3 The Origin of Novelties

17.4 Body Plans I: Overview

17.5 Body Plans II: the Origin of the Vertebrates

17.6 Body Plans III: the ‘Cambrian Explosion’

18 The Evolution of Complexity

18.1 Defining Complexity

18.2 The Lack of a ‘Law of Increasing Complexity’

18.3 Increases in the Complexity of Adults

18.4 Changes in the Complexity of Life-histories

18.5 Complexity at the Molecular Level

PART IV Conclusions

19 Key Concepts and Connections

19.1 Introduction: From Original Idea to Mature Scientific Discipline

19.2 A List of The Book’s Main Points, and the Emergence of Key Concepts

19.3 How do They Inter-Connect?

20 Prospects

20.1 Introduction: From the Present into the Future

20.2 Molecular Evo-Devo

20.3 Integrative Evo-Devo and General Evolutionary Theory

20.4 Wider Challenges

Glossary

Appendix 1: A Little Bit of History

Appendix 2: Naming of Genes and Proteins

Appendix 3: Geological Time

Appendix 4: Inferring Evolutionary Trees from Comparative Data

References

Index

This book has a companion website: www.wiley.com/go/arthur/evolution

“This book, written as an undergraduate text, is a really most impressive book. Given the burgeoning interest in the role of developmental change in evolution in recent times, this will be a very timely publication. The book is well structured and, like the author’s other books, very well written. He communicates with a clear, lucid style and has the ability to explain even the more difficult concepts in an accessible manner.”

Dr Kenneth McNamara, University of Cambridge

“There is much more to evolution than mere gene frequency changes in natural populations. Wallace Arthur was among the first to recognize fully the lack of a developmental dimension from the traditional view of evolution and is among the main actors who have been shaping the emerging agenda of evolutionary developmental biology.

From his research experience, the author has distilled in this book an original approach to the study of evolution, written in his uniquely attractive style where immediateness successfully mixes with conceptual clarity.”

Professor Alessandro Minelli, University of Padova

COMPANION WEBSITE

This book has a companion website:

www.wiley.com/go/arthur/evolution

with Figures and Tables from the book for downloading

This edition first published 2011 © 2011 by Wallace Arthur

Blackwell Publishing was acquired by John Wiley & Sons in February 2007. Blackwell’s publishing program has been merged with Wiley’s global Scientific, Technical and Medical business to form Wiley-Blackwell.

Registered officeJohn Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

Editorialoffices9600 Garsington Road, Oxford, OX4 2DQ, UK

For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book, please see our website at www.wiley.com/wiley-blackwell

The right of the author to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought.

Library of Congress Cataloguing-in-Publication Data

Arthur, Wallace.

Evolution : a developmental approach / Wallace Arthur.

p. cm.

Includes bibliographical references and index.

ISBN 978-1-4051-8658-2 (pbk.) – ISBN 978-1-4443-3720-4 (hardcover) 1. Developmental biology. 2. Evolution (Biology). 3. Comparative embryology. I. Title.

QH491.A769 2011

571.8–dc22

2010040517

A catalogue record for this book is available from the British Library.

This book is published in the following formats: eBook [ISBN]; ePub [ISBN]

“a theory of evolution requires, as some part of it, a theory of development”

Conrad Hal Waddington, The Evolution of an Evolutionist, 1975

Preface

I have written this book first and foremost for students who are taking a course in evolution at a university or college. Although there are many evolution texts ‘out there’, there are none that cover the ground in the same way as this one. This book adopts a very specific approach to the evolution of animals and plants – an approach in which the central theme is how evolution works by altering the course of egg-to-adult development.

When I was a student, I bought a book that defined evolution as ‘a change in the gene frequency of a population’. While evolution does indeed involve change at the population level, it involves changes at other levels too – most importantly at the level of the individual organism, like you and me. What makes us different from our closest living non-human relatives, chimpanzees, lies in our different structures and behavioural capabilities, as well as in the genetics of our populations.

The overly population-based approach to evolution is now giving way to a more integrative approach – one in which the process of development that turns fertilised eggs into adults is seen as being important to the evolutionary process in a variety of ways. However, it is seen as being important as well as, not instead of, changes in gene frequency caused by Darwinian natural selection. This is a crucial point, because some previous developmental approaches to evolution advocated a dismissal of population genetics and a denial that micro-evolutionary changes within species form the basis of most long-term evolution; this denial is now seen to be mistaken.

The recent resurgence of a developmental approach to evolution has several sources, all of them clustered into a short period of a few years around 1980. The most important of these was the discovery that the genetic basis of development in widely different animals is much more similar than was previously thought – a conclusion that was later extended to plants.

This discovery gave birth to the science of evolutionary developmental biology, or ‘evo-devo’ for short, in which comparisons of the developmental roles of homologous genes among different taxa became the focus of attention. But this is not a book devoted solely to evo-devo. There are some good such books already, and thus no need for another one just now. Rather, this is a book about how evo-devo can be integrated with other approaches to evolutionary biology, giving us a more complete view of evolution than has ever been available before.

I have tried to keep the book short and its level in the early chapters introductory, so that it will be useful to undergraduates taking their first university courses in evolution. However, as is appropriate for educational books in general, the level of discussion rises as the book progresses. Some of the later chapters would thus work well as the subject of seminars and tutorials in more advanced courses. This is particularly true from Chapter 10 onwards.

I should now say a few words about the book’s style and structure and, associated with those things, how best to use it.

This book, like other textbooks in science, represents a journey of exploration of a particular field, with the author leading the reader through a sequence of topics. Because of this, I have adopted the ‘we’ style of speech traditionally used by mathematicians but increasingly favoured in science too. So, instead of ‘thus it can be seen that evolution is caused by … ,’ the text reads ‘thus we can see that evolution is caused by…’ I hope that my chosen style will be perceived as it is intended: friendlier and less formal.

With regard to structure, the book’s 20 chapters are grouped into four Parts. At the beginning of each of these, there is a cover-page giving a brief synopsis of what each chapter in the relevant Part deals with. So, while one way to read the book is from start to finish (and of course I would recommend that!), another is to read those four cover-pages first, and to choose a different route. This might be especially useful for readers with enough background in the subject to omit a few chapters, particularly some of the early ones.

At the end of each chapter are a few suggestions for further reading; and at the end of the book there is a list of references. In the text, readers are directed to this reference list (if they wish to be) by superscripts. The rationale behind this split in the pointers to additional reading is that the chapter-end lists are short and usually include only books and review articles, rather than the harder-to-penetrate ‘primary literature’. The latter is grouped together at the end. This way, beginning students can ignore the superscripts in the text and the back-of-the-book reference list, and just look at the short chapter-end lists of more reader-friendly sources. More advanced students can do the opposite, or adopt an intermediate approach.

A word of caution on the references: because this is a student text, I have not attempted to give an exhaustive coverage of the literature. To do so, given the breadth of the field, would result in a reference list of perhaps 5000 publications, which would be inappropriate here. Instead, you will find about 250 publications, the aim being to provide just a sample of original research papers (and some historical texts) spread over the various topics that are covered.

At the back of the book, you will also find additional sources of information. There is a Glossary, which provides brief definitions of many scientific words that are used in the book. The first time a Glossary word is used in each chapter it is highlighted in bold. There are also four appendices: one that gives a little bit of history; one that deals with gene names, which can be inscrutable to newcomers; one that describes the geological time-scale, which is often used in the book to place major evolutionary events in the appropriate temporal context; and one that gives a very introductory account of how to infer patterns of relationship (evolutionary trees) from comparative data on different species.

Finally, a few words of thanks. I am grateful to all of the following for their help. Alec Panchen, Alessandro Minelli, Ken McNamara, and Patricia Moore kindly read and constructively criticised the entire draft manuscript. My sons Michael and Stephen produced much of the original artwork. I am also very grateful to the many colleagues, authors and publishers who granted permission for the use of artwork that was reproduced (in some cases in modified form) from elsewhere. The editorial staff at Wiley-Blackwell have helped in many ways. I would especially like to thank Ward Cooper for his support throughout the project; also Rosie Hayden, Emily Tye, Delia Sandford, Revathy Kaliyamoorthy and Carla Hodge. Much of the writing was done while on sabbatical leave from my University; I am grateful to the University for granting this leave. I am also grateful to Darwin College Cambridge for giving me a Visiting Associate position during my sabbatical. Finally, I would like to thank the various friends, colleagues and students with whom I have discussed the subjects presented herein over many years: you are too numerous to name, but you know who you are! Discussion, including argument between the proponents of different points of view, is a key element of the progress of science.

Wallace ArthurSeptember 2010

PART I

Foundations

‘Science is, I believe, nothing but trained and organised common sense’.

Thomas Henry Huxley (‘Darwin’s bulldog’) Collected Essays, vol. 3, 1893

The exciting challenge that this book sets itself is to explain to you what is known about the ways in which the development of animals and plants evolves; and also, in the final chapter, to explain what is not yet known, and thus what are the key questions for future research. We begin to approach this challenge on the foundations provided by the first four chapters, as follows:

Chapter 1: We start by examining the reason why understanding egg-to-adult development is absolutely crucial for a complete picture of the evolution of multicellular creatures.

Chapter 2: This chapter deals with the forerunners of present-day evolutionary developmental biology (or evo-devo), the way in which evo-devo has arisen, and the nature of evo-devo today, including the range of approaches that can be included within it.

Chapter 3: Here, the focus is on the nature of the developmental process, in terms of the things cells do, which are the results of an interplay between genes, their products, an array of other molecules and, often, environmental effects.

Chapter 4: Our focus now shifts to the structure of natural populations, since it is in these that all evolution happens. We examine their spatial, age and genetic structure. We then look at ways in which the genetic structure of populations can be altered by natural selection.

CHAPTER 1Introduction

1.1 From Darwin to Development

1.2 Development; and Evolutionary Changes in Development

1.3 Development and the Realm of Multicellularity

1.1 From Darwin to Development

The theory of evolution, established by Charles Darwin more than 150 years ago, and still itself evolving, is one of the most impressive products of science. As Darwin said1 in the closing paragraph of The Origin of Species: ‘There is grandeur in this view of life’ – a view in which many diverse creatures, both past and present (Fig. 1.1), have been brought into existence by natural processes, and in particular by the interplay between two such processes – heritable variation and natural selection.

Figure 1.1 A sample of creatures, present and past. All these forms and countless others have been produced from earlier ancestral forms by evolutionary modification of the course of development. This process can in each case be traced back to one of the several origins of multicellularity in the distant evolutionary past. The organisms included here are all discussed as examples later in the book.

Darwin marshalled a wide range of evidence in support of his theory. He drew on information from animal and plant breeding, fossils, behaviour, morphology, embryology and geography, among others. And he used all of these to build a sound basis for his key contribution to evolutionary theory: natural selection.

But evolution is, as noted above, the result of an interplay between two things – heritable variation and natural selection – it is not explicable by either of these on its own. The fact that Darwin was unable to enlighten us as much about the former as about the latter was hardly due to an oversight on his part; rather, it was due to limitations on what was generally known at the time in this area of ‘heritable variation’.

It is worth dissecting this phrase, because it includes both the inheritance of genetic variation and the process of development through which phenotypic variation is produced. Darwin was aware of the problem that there was not, in the 1850s, a clear understanding of how inheritance worked, but proceeded as best he could regardless. He later tried to supply a theory of inheritance – ‘pangenesis’ – but got it wrong. He was doubtless also aware that there was not a clear understanding of how egg-to-adult development worked, in terms of causal mechanisms, but he proceeded to use the information on descriptive and comparative embryology that did exist in the 1850s to good effect. As he remarked (Chapter 13), ‘community in embryonic structure reveals community of descent.’

The two most important things that have happened since Darwin’s synthesis of the evidence for evolution in 1859 have been the incorporation of genetics and developmental biology into the ‘big picture’, with the result that it has even more grandeur than before. The incorporation of genetics, which came first (Appendix 1), had both positive and negative effects on the incorporation of developmental biology that followed, is still in full swing, and is the subject of this book.

But why, actually, is development so important for evolutionary theory? There is a very specific and compelling answer to this question. It relates to the ways in which evolution can and cannot produce one type of animal or plant from another. This point is best made in relation to the type of evolutionary trees typically found in papers and books on the subject, two of which are shown in Fig. 1.2. Notice that in both trees the vertical axis is some measure of time, while the horizontal axis is some measure of difference in the morphology of the animal concerned, in one case ‘generalised’ and hard to quantify, in the other case a very specific measure of a particular structure (the depth of a bird’s beak).

Figure 1.2 Two typical evolutionary trees, as often found in books, articles and posters on the subject. Top: a tree depicting divergence in general form between whales and their sister-group (the extinct mammal Indohyus). Bottom: a tree depicting divergence in a particular character – the depth of a bird’s beak. Note that in the former case the top of the time axis cannot be the present because Indohyus is extinct. (Indohyus redrawn from Thewissen et al. 2007, Nature, 450: 1190–1194.)

In a very important sense, both of these trees represent impossible evolutionary transitions. They both employ the familiar shorthand method of representing an animal by one particular stage of its life-cycle – the adult. But evolution cannot make one kind of adult directly from another. Rather, it can only make a new kind of adult by altering, over a period of generations, the egg-to-adult developmental trajectory (Fig. 1.3).

Figure 1.3 A three-dimensional evolutionary tree (top) including the extra dimension of developmental time. This shows how the value of a character (such as the bird’s beak depth shown in Fig. 1.2) changes during development, and how this pattern of change itself changes in evolution. Note that the beak depth starts at zero because early bird embryos do not have beaks; and that if the developmental time axis is ignored, the tree reduces to two dimensions (bottom) and resembles that shown in Fig. 1.2 (bottom). LCA – last common ancestor; D1 – daughter species 1; D2 – daughter species 2.

So, an adequate theory of evolution must include not only an account of how fitness differences cause changes at the population level, but also an account of how the developmental differences that natural selection acts upon arise in the first place. And these latter differences cannot simply be written off as mutations, because a mutation is merely a change in the DNA sequence of a gene. If the gene that mutates causes the developmental trajectory to alter, then we need to know how this happens. Furthermore, developmental trajectories can in most cases be influenced by environmental factors as well as by genes. This is true not just of extreme cases, such as the production of male or female forms in turtles by egg incubation temperature, but also of more subtle cases, such as slight differences in the amount of left-right body asymmetry (often referred to as fluctuating asymmetry) that can be the result of variation in temperature and other environmental factors.

Such considerations give the environment not just one role in evolution – that of selective ‘sieve’ – but rather two, with the other being a role in the production of the variation in the first place. Of course, non-heritable variation, or phenotypic plasticity, cannot itself contribute to evolution, precisely because of its non-heritable nature. But if different genotypes differ in their pattern of developmental response (or their developmental reaction norm) to environmental variation, as is now widely known to occur, then this provides material for evolutionary change. Indeed, all evolutionary theory that deals with phenotypes that are completely genetically determined can be regarded as a subset of more general evolutionary theory in which the determination of developmental trajectories, and hence of phenotypes, is more complex.

What evolve, therefore, are not just adult animals or plants, but rather complete life-cycles. Furthermore, we should not think of pre-adult stages as evolving ‘in order to enable adult forms to evolve’. This overly ‘adultocentric’ view of things (as recently criticised by the Italian biologist Alessandro Minelli2) is misleading. Instead, what happens is that there are variations at all developmental stages. In each case, some variants may be fitter than others, either because of the advantages they possess at that stage or because of advantages that accrue further downstream in the developmental pathway or, of course, for both reasons. In some cases, particularly in animals with complex life-cycles (or ‘indirect development’ – i.e. development to adult via a larval form), evolution of larval stages may occur quasi-independently of evolution of the adult. This is true, for example, in the case of evolutionary switches between plankton feeding and yolk feeding in echinoderm larvae3, where the plankton-feeders have ‘arms’ that the yolk-feeders do not (Fig. 1.4), but this does not lead to a corresponding difference in the adults.

Figure 1.4 Evolution of early developmental stages, as exemplified by the evolutionary divergence of yolk-feeding and plankton-feeding echinoderm larvae. Top: Comparison of the larvae and adults of the congeneric species Heliocidaris tuberculata and H. erythrogramma (photographs courtesy of E.C. Raff and R.A. Raff). Bottom: Phylogeny showing multiple origins of yolk-feeders (red bars) from ancestral plankton-feeders. I – indirect development; D – direct development. (See Appendix 4 for an introduction to the business of inferring phylogenetic trees from comparative data.) (Reproduced with permission from Int. J. Dev. Biol., 47: 623–632.)

Having now seen that the case for the centrality of development in the evolutionary process is unassailable (but with a caveat to be discussed in Section 1.3), we need to examine development itself, and also to ask about the ways in which it can evolve. This approach (Section 1.2) will reveal several problems, some of which can be easily remedied at our current stage of knowledge, but some of which cannot. These problems include: the absence of some key terms; the previous over-emphasising of some processes (e.g. heterochrony); the need to connect organism-level observations with both molecular and populational ones; and the crucial issue of whether development in some sense guides evolution. This last issue is perhaps the most fascinating of all but is also the hardest to deal with and the most controversial. At stake here is the question of whether the structure of the developmental variation available to natural selection can influence the direction that evolution takes, rather than this direction being entirely set by selection alone; and if the former, then whether the role of development is merely negative (‘constraint’) or both positive and negative (‘bias’). This will be discussed in detail in Chapter 13.

1.2 Development; and Evolutionary Changes in Development

The development of any animal or plant can be thought of as a time-sequence of more or less well-defined stages. The simplest kind of development is ‘direct development’, as in mammals. Indirect development, whether in echinoderms, insects or amphibians, is more complex in that the route to the adult takes what might be thought of as a ‘detour’ via immature stages that are radically different from the adult as opposed to miniature versions of it. Some plants, notably trees, and also some colonial animals, such as bryozoans (‘moss animals’), are of a modular nature, which means that within one individual tree or one bryozoan colony, a major phase of development repeats itself multiple times. This is readily apparent to the casual observer of deciduous trees in successive springs, as the development of leaf modules occurs on a massive scale. It is also apparent, though only readily through a microscope, to observers of the growth of bryozoan colonies (Fig. 1.5), by the development of additional zooid modules around a colony’s periphery.

Figure 1.5 Modular development, as exemplified by a bryozoan colony. Each zooid is genetically identical to every other one in the colony, which is formed by the repeated asexual reproduction and development of zooids from the first one, which is referred to as the ‘ancestrula’. A large colony may consist of many hundreds of zooids (photograph courtesy of Peter Wirtz).

In any animal or plant, both the overall developmental process, and any particular component of it, such as the developmental pathways leading to the appearance of segments, limbs or leaves, can be thought of as a trajectory. Each such trajectory represents a very specific route for a cell population that is different from other possible routes. But equally, each trajectory can vary, both within an individual (the leaves on a single maple tree are not all identical) and, more importantly from an evolutionary perspective, among the various individuals that make up an ecological population. It is here that we move from thinking about development itself to how altered development can arise in evolution.

This is an absolutely crucial moment in our approach to the whole subject of this book. We can now see the nature of what we are dealing with: a change in something that is itself a process of change. This is very different to the old ‘ecological genetics’ approach to evolution, the practitioners of which were usually observing evolutionary change in something that was the fixed outcome of development. An example of the latter is industrial melanism in moths, where the phenotype studied was the external pigmentation of the adult (Chapter 4). Development itself was ignored in most such studies.

Given the more recent ‘change in a process of change’ approach, we need to be very careful that we have a suitable language to use to deal with such a complex situation. The complexity was nicely described by the American developmental biologist Scott Gilbert4 in 2007, as follows: ‘For evolutionary developmental biology, the current challenge is producing a 5-dimensional representation: the four standard dimensions of space and time placed into the context of the paleontological temporal dimension.’

So, now to the issue of a suitable language. Biologists call a change in a gene ‘mutation’. Although this can be extended to the phenotypic level by talking of mutant phenotypes, this usage is not helpful. It connects better with the old ecological genetics approach of ignoring development and concentrating on the adult phenotype. What we need instead is a different term that indicates clearly that we are referring to a change in development. We can’t just call it ‘developmental change’, as that would be ambiguous and more likely to be interpreted just as going along one particular trajectory (in developmental time) rather than switching from one trajectory to another (in evolutionary time). But if ‘developmental change’ won’t do, what will?

One term that definitely will not do as an overall term for evolutionary change in ontogeny (i.e. development) is ‘heterochrony’, despite the book title by McKinney and McNamara5, Heterochrony: The Evolution of Ontogeny, with its implication that the two are synonymous. Rather, heterochrony (evolutionary change in developmental timing) is merely a subset of the overall evolution of development, as indicated by the chapter title ‘It’s not all heterochrony’ in Rudolf Raff’s book6The Shape of Life.

Two phrases that have already been used in the context of evolutionary changes in development may be considered as candidate ‘umbrella-terms’ to cover all such changes, whether heterochronic or other. These are developmental repatterning7 (sometimes in the form ontogenetic repatterning8, which is synonymous) and developmental reprogramming9. There are two reasons why the former phrase is preferable. First, ‘developmental program(me)’, ‘programming’ and ‘reprogramming’ are too philosophically loaded, and are interpreted by some biologists as smacking of ‘genetic imperialism’ (Appendix 1). Second, ‘developmental reprogramming’ has become used in a different, and narrower, way10 in the last few years. So, developmental repatterning now seems the obvious choice as the umbrella-term for all evolutionary changes in development, and we will use it throughout this book.

Any developmental process can be thought of as a pattern in time, space or (usually) both. Already ‘pattern’ is part of the language at different levels of developmental study. For example, at the molecular level we speak of the expression pattern of a gene in an embryo (Fig. 1.6); and at the tissue level we speak of pattern formation, to indicate, for example, the different developing patterns of the five digits of our hand.

Figure 1.6 The expression pattern of a gene in embryos. The example shown is the expression of a gene called distal-less (see Appendix 2 for information on gene names) in crustacean (top) and centipede (bottom) embryos. In the crustacean (Parhyale) the red staining reveals expression of the gene in developing limbs in early-stage (left) and later-stage (right) embryos (photograph courtesy of Nipam Patel); in the centipede (Strigamia) the expression of the gene is shown in green (photograph courtesy of Cornelius Eibner).

If the development of one individual can be thought of as patterning (in a multitude of senses), then the evolution of development can be thought of as repatterning. Logically, considering any aspect of development – for example, the expression pattern of a gene – there are four types of repatterning that can occur: changes in time, place, amount or type. There are well-established terms for the former two – heterochrony and heterotopy; and there are more recently-introduced terms for the latter two – heterometry and heterotypy. This series of terms provides a broad categorisation of the types of evolutionary change that can occur in the developmental process. It will be useful to keep them in mind later (and so they are used as chapter titles in Part II) as the intricacies of particular case studies emerge.

1.3 Development and the Realm of Multicellularity

There is a restriction to the importance of development in evolution that does not apply to the importance of genes; specifically, development is only a necessary part of evolutionary theory when the creatures that are evolving are multicellular. This is because creatures that are unicellular throughout their life-cycle lack ‘development’, at least in the sense in which the word is normally used. Thus evo-devo deals mainly or wholly with what we will call here the realm of multicellularity.

Note that this term – realm – is not part of the taxonomic hierarchy (in the way that domain, kingdom and phylum are). The reason for adopting such a term is that if we take as a group all multicellular creatures, they do not form a single clade. Indeed, far from it: multicellularity has arisen at least five times in evolution, and probably much more often.

The five major origins of multicellularity – in animals, plants, fungi, brown algae and cellular slime moulds – are shown in Fig. 1.7. More minor origins, in the sense that they have led to more restricted invasions of multicellular morphospace, have occurred (inter alia) as follows: in the cyanobacteria (strings and mats of cells); in the diatoms (which have some multicellular forms, despite being very largely a clade of unicells); and in other groups of ‘slime moulds’ unrelated to the group shown in Fig. 1.7. It should be noted that the deep divisions of the living world shown in the figure are different from those that can be seen in comparable trees produced a mere decade ago; and our picture of deep phylogeny may well yet change further. Despite this, the conclusion that multicellularity has arisen several times in evolution is likely to be robust.

Figure 1.7 The five ‘main’ evolutionary origins of multicellularity, each marked with an M on the lineage concerned. In each case the probable sister-group is also indicated. Note that in a typical multicellular creature there is also a developmental origin of its multicellular state in each generation. In some branches, what is shown is ‘exact’ – for example, there was probably only a single origin of multicellularity in animals; but in other cases, for example fungi, what is shown may be a simplification.

How did multicellularity originate? A recent clue has come from the study of choanoflagellates – a clade of unicells that appears to be the sister-group of the animal kingdom. The first choanoflagellate genome project11 has revealed that these creatures possess many genes previously known only from animals and associated with multicellularity, such as genes that make cell adhesion and cell-cell signalling molecules. These proteins may have been used in the unicellular ancestors of animals (and in present-day choanoflagellates) to interact with the environment, including conspecifics (potential mates) and other unicells (potential prey). If so, this would represent an example of exaptation, in which something initially evolved for one selective reason later becomes useful (and hence selected) for another. This fascinating topic will be discussed further in Chapter 12.

As ever, there is a caveat. One main strand of the evidence for a sister-group relationship between choanoflagellates and animals is the similarity in the form of the typical choanoflagellate cell (possessing a collar) and the collar-cells of sponges (Porifera), which have long been regarded as the most primitive animals. But if a recent molecular phylogeny of the animal kingdom12 is to be believed, the ctenophores (comb jellies) are more basal than sponges. However, the authors concerned admit that their placing of the ctenophores as basal should only be regarded as a hypothesis for now. Figure 1.8 shows this hypothesis and its main alternative, along with pictures of collar cells. A complication to both hypotheses is that the sponge group Porifera may be paraphyletic or even polyphyletic.

Figure 1.8 Our understanding of the pattern of relationships among the most basal animal groups is still very imperfect. Top: Conflicting hypotheses as represented by alternative trees. Bottom: Pictures of a ctenophore, a sponge, a collar-cell (choanocyte) of a sponge and a choanoflagellate.

An important general point emerges from this recent dispute about the relationships among the most basal animals. In general, it is best to have an agreed phylogeny of any group of animals or plants before mapping onto it (as in Chapter 11) evolutionary changes in development. For some groups, this is not a problem; but for others, our current view of phylogeny may alter radically in the future. If it does indeed alter, then so too must our views on the nature and temporal sequence of evolutionary changes in development – that is, of developmental repatterning.

SUGGESTIONS FOR FURTHER READING

Two good, but very different, introductory books on the ‘new science of evo-devo’ and on developmental approaches to evolution more generally are:

Minelli, A. 2009 Forms of Becoming: The Evolutionary Biology of Development. Princeton University Press, Princeton, NJ.

Carroll, S. 2005 Endless Forms Most Beautiful: The New Science of Evo-Devo and the Making of the Animal Kingdom. Norton, New York.

CHAPTER 2What is Evo-Devo?

2.1 Forerunners of Evo-Devo

2.2 Nineteenth-Century Comparative Embryology

2.3 Diverse Antecedents—1900–1980

2.4 Conclusions from History; Messages for the Present

2.5 The Advent of Evo-Devo in the 1980s

2.6 Broad and Narrow Views of Evo-Devo

2.7 Too Few Laws, Too Many Facts?

2.1 Forerunners of Evo-Devo

As Winston Churchill said: ‘The farther backward you can look, the farther forward you are likely to see.’ This was intended as a comment on history in general, but it can be applied just as well to the history of science. Any branch of science can only be thoroughly understood in the context of how it has progressed over time.

A good example of how an understanding of present-day evo-devo is informed by a study of its forerunners is the proposed recapitulation by embryos of their evolutionary past. Initially regarded (wrongly) as a law of how development evolves, it was later completely rejected (wrongly) and then eventually regarded (correctly) as one of several possible patterns, all of which we need to be able to explain.

So, looking at the historical background of present-day evo-devo is not optional – it is essential. If you are seriously history-averse, you can skip to the skeletal version in Section 2.4 – but I would advise against doing so.

2.2 Nineteenth-Century Comparative Embryology

In the decades following publication of Darwin’s classic text in 1859, the leading evolutionary embryologist was Ernst Haeckel, whose publications included three major books. While English translations of the first two have never been undertaken (a project for the future, perhaps), the last – a more popular account of Haeckel’s views – was translated13 into English in 1896 as The Evolution of Man. This translation is important because it allows the many English-speaking biologists who cannot read German (alas a large number, and including myself) to understand Haeckel’s key points.

Haeckel’s work has frequently been misunderstood, as has the relationship between his ‘law’ and the ‘laws’ of the pre-Darwinian comparative embryologist Karl Ernst von Baer, whose principle of embryonic divergence is shown in Fig. 2.1. Some leading figures have been guilty of such misunderstanding, including the English biologist Gavin de Beer (see below and Section 2.3). It is important to clarify this matter here, because the evo-devo of today must build on the comparative embryology of the past – but only on those parts of it that we believe to be correct.

Figure 2.1 The pattern known as von Baerian divergence, as illustrated by the embryos of four vertebrates, fish, hen, cow and human, shown at three different stages – early, middle and late. Note the pattern of early similarity giving way to later differences. (Redrawn from A Theory of the Evolution of Development, John Wiley & Sons, Ltd., 1988.)

Von Baerian divergence, as can be seen from Fig. 2.1, is a pattern of early similarity giving way to later differences. Von Baer14 gave four laws to describe this pattern; but it can adequately be described by the first of these: ‘The general features of a large group of animals appear earlier in the embryo than the special features.’ (Source: Alec Panchen’s15Classification, Evolution and the Nature of Biology).

Haeckel famously produced the ‘biogenetic law’ or ‘law of recapitulation’. ‘This fundamental law…’, writes Haeckel (1896: p. 7) states:

that Ontogeny is a recapitulation of Phylogeny; or, somewhat more explicitly: that the series of forms through which the Individual Organism passes during its progress from the egg cell to its fully developed state, is a brief, compressed reproduction of the long series of forms through which the animal ancestors of that organism (or the ancestral forms of its species) have passed from the earliest periods of so-called organic creation down to the present time.

Now, two very important questions arise. First, did Haeckel mean that the ontogenies of descendant species went through stages similar to the adult forms of their ancestors? Second, and related to that, is Haeckel’s law compatible or incompatible with von Baer’s laws?

Regarding the first question, here is the answer provided by de Beer16, who interprets Haeckel’s law as follows (p. 5): ‘The adult stages of the ancestors are repeated during the development of the descendants’ [my italics]. De Beer’s next section is entitled ‘The rejection of the theory of recapitulation’. But should we accept that a thorough student of embryology, such as Haeckel, really believed this? Probably not. Here, again, is Haeckel (1896: p. 18):

The fact is that an examination of the human embryo in the third or fourth week of its evolution [= development] shows it to be altogether different from the fully developed Man, and that it exactly corresponds to the undeveloped embryo-form presented by the Ape, the Dog, the Rabbit, and other Mammals, at the same stage of their Ontogeny [again, the italics are mine].

Clearly, the picture in Haeckel’s mind was like that of von Baer, not like the ‘caricature recapitulation’ involving ancestral adult forms in descendant embryos.

Why has Haeckel frequently been misinterpreted as being both so stupid as to see adult ancestors in descendant embryos and also in some sense anti-von Baer? The answer must lie at least in part in his choice of phrase: recapitulation. At first sight, this would not seem to be compatible with embryonic divergence.

But this is where we must recall that von Baer’s14magnum opus (published in 1828) was pre-Darwinian, and in later life von Baer was anti-Darwinian; in contrast, Haeckel was pro-Darwinian. To a non-evolutionist such as von Baer, the comparisons of vertebrate embryos shown in Fig. 2.1 were not comparisons of animals that were related in an ancestor-descendant way – rather, they were comparisons among a series of independently-created forms. But to an evolutionist such as Haeckel, they were comparisons that related to shared descent – indeed to what Darwin frequently called ‘descent with modification’. Of course ‘the Dog’ is not a human ancestor, and Haeckel could hardly have believed so, but the two embryos show features that derive from their common descent – such as the rudimentary gill-clefts that both dog and human embryos transiently possess as a result of their shared aquatic ancestry. So they both recapitulate (non-adult) features of their ancestors, while also progressively diverging from each other over developmental time. There is simply no conflict between these views.

There may not be a conflict between them, but there is definitely a problem with both views – specifically their description, by their respective authors, as ‘laws’. Generally, in science, ‘law’ is used for something that is either universally true or at least very nearly so. It also tends to be used for something that can be stated quantitatively/mathematically, rather than merely verbally or in pictures. In physical science, Newton’s laws of motion provide a good example. They are very generally true (except at speeds approaching that of light, where Einstein’s relativity takes over); and they can be stated quantitatively. In biological science, Mendel’s laws of inheritance17 provide a good example. They are generally true of diploid organisms, though the second law, of independent assortment, requires that the genes under consideration are on separate chromosomes. And they make quantitative predictions, such as the well-known Mendelian ratios that result from various sorts of cross – for example, a 3:1 phenotypic ratio in the F2 generation from a cross involving homozygous ‘wild-type’ and mutant parents.

It is clear that neither von Baer’s nor Haeckel’s ‘laws’ can be expressed in quantitative form. So they both fail to attract law status in this respect. But it is also the case that they both fail according to the other criterion, that of being universally, or even generally, true.

This second point requires some elaboration. The particular comparisons of developmental trajectories shown in Fig. 2.1 are few in number, and highly non-random in that all involve vertebrate embryos. What of other comparisons? Inter-taxon comparisons could be made within some other major clade than vertebrates (e.g. insects). Also, comparisons could be made between representatives of different major clades (e.g. between a vertebrate and an echinoderm).

Most insects go through a common early developmental stage called the germ band. They are recapitulating this stage, which we believe was present in the development of the ‘stem insect’. But since they go on to be very different adults (e.g. dragonfly and wasp), they are also undergoing von Baerian divergence.

Vertebrates and echinoderms both show radial cleavage as their earliest embryonic process, in contrast to most other animal groups, which show spiral cleavage (Fig. 2.2). We believe this to be the case because they are both descended from a common ancestor (the ancestral deuterostome) that also possessed this feature. The retention of this feature of their ancestor’s development can be thought of as recapitulation. But, since their adult forms are incredibly different (think of a mouse and a sea-urchin, for example), the fact that they are similar in their earliest developmental process, cleavage (albeit starting from very different eggs) means that we can think of the pattern also as being von Baerian divergence.

Figure 2.2 Two types of cleavage – the earliest developmental process in animals. Left – radial cleavage, as found in vertebrates, echinoderms and other deuterostomes. Right – spiral cleavage, as typically found in protostomes, including molluscs. Note that the degree of size difference between micromeres and macromeres varies considerably among different animal groups; in particular this difference is typically smaller (even non-existent in some cases) in radial than spiral cleavage. Also note that these are not the only two types of cleavage found; for example, insects have a type of cleavage referred to as syncytial.

Note, however, that the use of these terms, recapitulation and divergence, is getting somewhat strained. Very few features of the deuterostome ancestor’s development (inasmuch as we can infer that development) are recapitulated. Furthermore, the ontogenies of, say, a mouse and a sea urchin, become so different at such an early developmental stage that ‘divergence’ hardly seems a sensible term for differences in later development.

We must always remember that evolution is a messy process. This means that no single pattern will prevail in all cross-taxon comparisons of developmental trajectories. Thus there are no patterns whose generality of occurrence is sufficient to warrant the invention of a ‘law’. All developmental stages change in evolution. Some stages, in some taxa, are more resistant to change than others, and this can give rise to patterns, both von Baerian/Haeckelian and others. Although these patterns are statistical rather than absolute or law-like, they are nevertheless of considerable interest. One of them, where the very earliest stages in development are less similar than the stages that follow them, giving a specific departure from the pattern suggested by von Baer, will be discussed in Chapter 19.

2.3 Diverse Antecedents—1900–1980

The leading figures working on the relationship between evolution and development during the period from 1900 to the advent of evo-devo in the 1980s were remarkably heterogeneous in their approaches. They certainly do not constitute a school of thought, in the way, for example, that population genetics did over this same period. We will briefly examine the work of six prominent people here, in historical sequence: D’Arcy Thompson; Gavin de Beer; Richard Goldschmidt; C.H. Waddington; Lancelot Law Whyte and Stephen Jay Gould.

Before doing so, it is interesting to note a statement made by J.B.S. Haldane18 (a population geneticist, not a student of development) in 1932. This statement presages an important paper published in 1979 by American evolutionary biologists Stephen J. Gould and Richard Lewontin – we will look at this issue in Chapter 13. For some historical background to the Haldane quote, see Appendix 1. Here is the quote:

D’Arcy Thompson

The most enduring legacy of Thompson’s19On Growth and Form is his ‘theory of transformations’. Thompson pointed out that changes in development over the course of evolutionary time were not piecemeal, in the sense that each of an animal’s characters evolved independently, but rather were co-ordinated changes in body form. These could be represented visually by plotting the morphological outline of one animal on a Cartesian grid and then subjecting that grid to some systematic distortion (Fig. 2.3). Ironically, despite Thompson’s interest in development as well as in evolution, his pictures of evolutionary transformations generally were of adults. Nevertheless, we can easily extend his idea of a transformation to cover the later phases of ontogeny (when allometric growth prevails) rather than just the adults that ultimately result. The take-home message in either case is the same: evolution often involves co-ordinated changes in many aspects of body form.

Figure 2.3 The transformations that can be applied to the form of one species within a group (in this case fish) to produce the forms of others. These transformations were devised by D’Arcy Thompson as early as 1917 (see text). The fact that they are possible indicates that evolution often modifies animal (and plant) forms in a co-ordinated, rather than piecemeal, manner. (From Arthur, 2006, Nature Reviews Genetics, 7: 401–406.)

Thompson’s transformations can be regarded as an attempt to quantify what Darwin referred to as ‘correlation of growth’. However, since there is always a limit to the range of forms that can be connected up by any one series of transformations, such limits can be taken as being against the pan-gradualist spirit of Darwinism, as Thompson pointed out:

Our geometric analogies weigh heavily against Darwin’s concept of endless small continuous variations; they help to show that discontinuous variations are a natural thing, that ‘mutations’ – or sudden changes, greater or less – are bound to have taken place, and new ‘types’ to have arisen, now and then.

Gavin de Beer

De Beer16 published Development and Evolution in 1930; and a revised version entitled Embryos and Ancestors in 1940, with further revised editions (but no further changes of title) in 1951 and 1958. His central theme was evolutionary change in the timing of developmental events, i.e. heterochrony. He emphasised the multitude of types of heterochronic process, highlighting in particular neoteny, the retardation of somatic development relative to the development of the reproductive system, with the result that in a neotenous lineage what was previously a juvenile (or even larval) form becomes reproductively mature. However, he also dealt with many other kinds of heterochrony, and so was essentially pluralist in his approach to the evolution of development. But his pluralism did not extend to recapitulation. Instead of acknowledging that it is one of many patterns that can be found in the evolution of development, he states in his concluding chapter (1958: pp. 170–171) that ‘Recapitulation…does not take place.’ He reached this conclusion due to taking a rather narrow view of what recapitulation means (Fig. 2.4).

Figure 2.4 Two views of recapitulation. Left: ‘perfect’ recapitulation, in which evolution adds new developmental stages at the end of the ancestral ontogeny. This does not occur. Right: ‘imperfect’ recapitulation, in which some features of ancestral ontogenies are repeated (or recapitulated) in the development of descendants, even though they may not lead to functional adult structures in the descendants. This form of recapitulation does occur, and is due to the fact that it is often hard for natural selection to alter early stages. The two forms of recapitulation are shown in terms of abstract developmental stages (A to D) in the top panel; and in terms of a particular example (bottom panel) where A is a vertebrate zygote; B is a vertebrate embryo with gill clefts; C is an adult fish; and D is an adult human.

Richard Goldschmidt

Richard Goldschmidt is remembered most for his saltationist views20. He was much influenced in his overall interpretation of the evolutionary process by his studies of homeotic mutations in Drosophila fruit-flies21 – mutations causing the appearance of ‘the right structure in the wrong place’ – such as the one that causes an extra pair of wings to grow out of the third thoracic segment in place of the small balancing organs that are normally found there (the ‘bithorax’ phenotype: see Fig. 2.5). He argued (i) that since some of these mutations caused a change that was of sufficient magnitude to bridge the gulf (in some characters anyhow) between higher taxa, such as insect orders; and thus (ii) that these mutations might have formed the basis for the origins of the higher taxa (such as orders) concerned.

Figure 2.5 Normal (left) and homeotic mutant (right) forms of the fruit-fly Drosophila melanogaster. The particular mutant shown is ‘bithorax’, in which there is a second pair of wings. This is due to the transformation of the third thoracic segment, which normally produces small flight-balancing structures called halteres, into a repeat of the second thoracic segment, which produces wings. (Reproduced with permission from Carroll, S. et al. 2005, From DNA to Diversity, Blackwell.)

This theory did not end up being accepted, largely because it became clear that all known homeotic mutants of Drosophila are unfit compared to the wild type, probably in all relevant environments. Nevertheless, the fact that less drastic mutations in the same genes have a major role in the evolution of development has now become apparent.

Conrad Hal Waddington

Waddington22 gave us three linked concepts, and new terms to go with them. He pictured developmental trajectories in the form of a ball running down grooves in an epigenetic landscape (Fig. 2.6). As can be seen, the landscape is like a system of river tributaries, but with the overall topography being the reverse of that which characterises real geography: as we go downhill (= later in developmental time), the rivers diverge rather than converge. Such a system involves a degree of canalisation, in that there is local stability of the trajectory – with the amount of stability depending on the depth of the groove. This was Waddington’s abstract way of picturing the fact that real developmental systems are to some extent robust in the face of both genetic (mutational) and environmental perturbations.

Figure 2.6 The ‘epigenetic landscape’. Here, development is pictured in terms of several possible courses (valleys), with the actual course taken in any individual organism of the species concerned being represented by a ball running through particular valleys of the landscape from top (start) to bottom (finish – the adult). (Redrawn from Waddington, 1957, The Strategy of the Genes, Allen & Unwin.)

Some perturbations, however, are too extreme for canalisation to be retained, and so development takes a different course – that is, it becomes repatterned. Waddington showed23 that one particular environmental perturbation – exposure of eggs to ether vapour – could cause Drosophila development to switch from production of a normal fly to production of the bithorax phenotype that can also be produced by genetic perturbations (the mutations in the Bithorax gene-complex whose effects Goldschmidt had studied). Furthermore, Waddington also showed that some genotypes were more sensitive than others to switching their developmental trajectory in response to ether vapour. So, when he performed an artificial selection experiment and bred only those flies that most readily underwent the switch, he could eventually produce bithorax phenotypes without ether treatment. This process, which he called genetic assimilation, looks, but is not, Lamarckian. It is a special case of the evolution of phenotypic plasticity. This subject (Chapter 16) was being studied in parallel by the Russian biologist Schmalhausen, though the two men had little contact.

Lancelot Law Whyte

Whyte produced two papers and a book24 in the 1960s, dealing with what he called first ‘developmental selection’ and later ‘internal selection’. The point he was emphasising is that developmental (and more generally organismic) integration – the compatibility of interacting parts – can act as a selective agent just as can an external, ecological agent such as a predator. Thus he was championing internal coadaptation rather than external adaptation to particular environmental conditions. We will look further at Whyte’s ideas in Chapter 12.

Stephen Jay Gould

We will look in particular at Gould’s famous 1977 book25Ontogeny and Phylogeny. Whether something that is only 40 to 50 years old should be regarded as historical is a moot point. However, we will take the approach here that ‘evo-devo’ originated around 1980, and that everything before then thus constitutes the historical basis for it rather than being part of it. The rationale for this approach is that before 1980 most studies of the evolution of development were at the organismic level. None of them were at the level of the molecular structure of key groups of developmental genes, as this structure did not begin to be revealed until about 1980 (Section 2.5).

Ontogeny and Phylogeny is very much a book of two parts: the first is historical (and entitled ‘recapitulation’); the second is focused on heterochrony. Gould concludes his historical study of recapitulation in a very different way than de Beer did several decades earlier. Instead of concluding that recapitulation does not occur, Gould states:

But recapitulation was not ‘disproved’; it could not be, for too many well-established cases fit its expectations. It was, instead, abandoned as a universal proposition and displayed as but one possible result of a more general process.

The second part of Gould’s book, however, has much in common with de Beer. Both concentrate on, and in fact overemphasise, changes in developmental timing (heterochrony). Gould developed a particular way of looking at these, a ‘clock model’, which is perhaps useful but hardly revolutionary. The historical part of Ontogeny and Phylogeny