Evolutionary Developmental Biology

[Rob]

The fact about science is that everyone who had made a serious contribution to it is aware, or very strongly suspects, that the world is not only queerer than anyone has imagined, but queerer than any one can imagine. This is a most disturbing thought, and one flees from it by stating the exact opposite. (Clark R. 1968, citing a letter from J.B.S. Haldane to Robert Gaves.)

-- Reid, Robert G. B. (1985) Evolutionary Theory: The Unfinished Synthesis. New York: Cornell Univesity Press. p. 117.

The key to development, form, and function in multicellular organisms is differential gene expression, and the most intimate knowledge of the genetic code reveals nothing about the implementation of its information in space and time.

-- Reid, Robert G. B. (1985) Evolutionary Theory: The Unfinished Synthesis. New York: Cornell Univesity Press. p. 46.

A little over 25 years later our "intimate knowledge" has grown significantly, providing insights which have yet to be fully intergrated into evolutionary theory:

Animal body plans, their structures and the functions with which their morphology endows them, are the integrals over time and space of their successive developmental processes.... At the outside, development is mediated by the spatial and temporal regulation of expression of thousands of genes that encode the diverse proteins of the organism, and that catalyze the creation of its nonprotein constituents. Deeper in is a dynamic progression of regulatory states, defined by the presence and state of activity in the cell nuclei of particular sets of DNA-recognizing regulatory proteins (transcription factors), which determine gene expression. At the core is the genomic apparatus that encodes the interpretation of these regulatory states. Physically, the core apparatus consists of the sum of the modular DNA sequence elements that interact with transcription factors. These regulatory sequences "read" the information conveyed by the regulatory state of the cell, "process" that information, and enable it to be transduced into instructions that can be utilized by the biochemical machines for expressing genes that all cells possess. The sequence content, arrangement, and other aspects of the organization of these modular control elements are the heritage of each species. They contain the sequence-specific code for development; and they determine the particular outcome of developmental processes, and thus the form of the animal produced by every embryo. In evolution, the alteration of body plans is caused by changes in the organization of this core genomic code for developmental gene regulation. (Davidson 2006: 1-2)

... [T]he system level organization of the core genomic regulatory apparatus, and how this is the locus of causality underlying the twin phenomena of animal development and animal evolution. Because the sequence of the DNA regulatory elements is the same in every cell of each organism, the regulatory genome can be thought of as hardwired, and genomic sequence may be the only thing in the cell that is. Indeed that is a required property of gene regulatory elements, for they must endow each gene with the information-receiving capacity that enables it to respond properly to every conditional regulatory state to which it might be exposed during all phases of the life cycle, and in all cell types. For development, and therefore for major aspects of evolution, the most important part of the core control system is that which determines the spatial and temporal expression of regulatory genes. As used here, "regulatory genes" are those encoding the transcription factors that interact with the specific DNA sequence elements of the genomic control apparatus. The reason that the regulation of genes encoding transcription factors is central to the whole core system is, of course that these genes generate the determinant regulatory states of development. (Davidson 2006: 2)

There follow several important and general principles of organization of the developmental regulatory apparatus, that is, of the control machinery directing expression of the regulatory genes themselves. First, signaling affects regulatory gene expression: The intercellular signals upon which spatial patterning of gene expression commonly depends in development must affect transcription of regulatory genes, or else they could not affect regulatory state. Therefore, the transcriptional termini of the intracellular signal transduction pathways required in development are located in the genomic regulatory elements that determine expression of genes encoding transcription factors. Second, developmental control systems have the form of gene regulatory networks: Since when they are expressed given transcription factors always affect multiple target genes, and since the control elements of each regulatory gene respond to multiple kinds of incident regulatory factor, the core system has the form of a gene regulatory network. That is, each regulatory gene has both multiple inputs (from other regulatory genes) and multiple outputs (to other regulatory genes), so each can be conceived as a node of the network. Third, the nodes of these genes regulatory networks are unique: Though it is not a priori obvious, each network node performs a unique job in contributing to overall regulatory state, in that its inputs are a distinct set, just as the factor it produces has a distinct set of target genes. Fourth, regulatory genes perform multiple roles in development: The repertoire of regulatory genes is evolutionarily limited, and all animals use more or less the same assemblage of DNA binding domains, which define the classes of transcription factor. However, given factors are frequently required for different processes in different forms of development, and they are often used for multiple unrelated purposes within the life cycle. Thus, both within and among animal species, many regulatory genes must be able to respond to diverse regulatory inputs that are presented in various space/time places in the developing organism. (Davidson 2006: 2-3)

A general character of genomic programs for development is that they progressively regulate their own readout, in contrast, for example, to the way architects' programs (blueprints) are used in constructing buildings. All of the structural characters of an edifice, from its overall form to local aspects such as placement of wiring and windows, are prespecified in an architectural blueprint. At first glance the blueprints for a complex building might seem to provide a good metaphoric image for the developmental regulatory program that is encoded in the DNA. Just as in considering organismal diversity, it can be said that all the specificity is in the blueprints: A railway station and a cathedral can be built of the same stone, and what makes the difference in form is the architectural plan. Furthermore, in bilaterian development, as in an architectural blueprint, the outcome is hardwired, as each kind of organism generates only its own exactly predictable, species-specific body plan. But the metaphor is basically misleading, in the way the regulatory program is used in development, compared to how the blueprint is used in construction. In development it is as if the wall, once erected, must turn around and talk to the ceiling in order to place the windows in the right positions, and the ceiling must use the joint with the wall to decide where its wires will go, etc. The acts of development cannot all be prespecified at once, because animals are multicellular, and different cells do different things with the same encoded program, that is, the DNA regulatory genome. In development, it is only the potentialities for cis-regulatory information processing that are hardwired in the DNA sequence. These are utilized, conditionally, to respond in different ways to the diverse regulatory states encountered (in our metaphor that is actually the role of the human contractor, who uses something outside of the blueprint, his brain, to select the relevant subprogram at each step). The key, very unusual feature of the genomic regulatory program for development is that the inputs it specifies in the cis-regulatory sequences of its own regulatory and signaling genes suffice to determine the creation of new regulatory states. Throughout, the process of development is animated by internally generated inputs. “Internal” here means not only nonenvironmentali.e., from within the animal rather than external to it but also, that the input must operate in the intranuclear compartments as a component of regulatory state, or else it will be irrelevant to the process of development. (Davidson 2006: 16-17)

-- Davidson, Eric H. (2006) The Regulatory Genome. Amsterdam: Academic Press.

SUPERGENES ARE IN COMMAND

We have just seen that the cells of a pluricellular organism all have the some genes. If they differentiate into distinct cell types, it is because they do not express the totality of their genes but practice a selection that varies according to cell type. Otherwise, all the cells of an organism would be identical. Cells thus contain "genetic switches," systems that switch on or off the expression of certain genes. This control is carried out by proteins, called transcription factors, that either stimulate or repress the transcription of the genes involved. These proteins being themselves the products of genes, which are subject in turn to a similar regulation, genomes house a whole complex and hierarchical network of regulatory genes--the term "supergene" is sometimes used--next to those that code for "housekeeping," that is, for enzymes, structural proteins, etc. (de Duve 2002: 155)

Regulatory genes are known in bacteria, in which they are involved, among other things, in the adaptation of metabolism to different nutrients. A historic example, which made the fame of the French investigators Fançois Jacob and Jacques Monod, concerns the manner in which bacteria transferred to a medium containing milk sugar (lactose) as sole food supply switch on the genes coding for enzymes specifically needed to use this sugar. Regulatory genes are, however, much more numerous in eukaryotes, and their number increases with the complexity of the body plan of the species. Such is not the case for housekeeping genes, for which there is hardly any difference among species. Or when there is a difference, impoverishments rather than enrichments most often go together with increasing complexity. Witness the many vitamins we are unable to make, whereas humble bacteria do so without difficulty.... In spite of the advances of biotechnologies, we are still far from mastering evolution. (de Duve 2002: 155-156)

The discovery of regulatory genes has allowed us to discern, at least in principle, the mechanisms that direct and control development. Once fertilized, the egg cell divides into two cells, which similarly divide to produce four, which divide into eight, and so forth. Soon, in the course of this process, the cells cease to be identical. Depending on their location in the ensemble, they start expressing or stifling certain regulatory genes, with the consequence that the proteins translated from those genes create concentration gradients between the areas where they are produced and those where they are not. These gradients influence in unequal fashion the expression of other genes, which in turn influence others, in a cascade whose complexity soon exceeds the limits of our imagination and even anything that can be simulated by the most powerful computer programs. At the end of the game, there is an oak plantule, a jellyfish larva, or a newborn baby, depending on the program written into the genome. (de Duve 2002: 156)

Such a mechanism has long been suspected. Already, in the early part of the twentieth century, the German embryologist Hans Spemann demonstrated, by means of remarkably skillful and ingenious experiments, the existence of what he called morphogenetic--shape creating--gradients in embryos. Modern biology is beginning to flesh out those gradients in terms of genes and their protein products. Particularly important has been the discovery of so-called homeogenes, whose control is so wide-ranging that a single mutation of such a gene may cause a fruitfly to grow an extra pair of wings or to sprout additional antennae on its head. Homeogenes have been recognized throughout the pluricellular world, from simple fungi to the most complex animals. (de Duve 2002: 156)

EVOLUTION OCCURS BY WAY OF DEVELOPMENTAL PROGRAMS

With these elementary notions we may now address the problem of evolution, which is conditioned, as we have seen, by changes in the developmental program of organisms. This fact implies almost necessarily that the underlying genetic changes have as targets regulatory genes. But all depends on the cell type to which the modified gene belongs. Thus, a mutation in a skin, stomach, or brain cell may start a new cell line, for example, a cancerous one. But the individuals concerned do not, if they reproduce, give birth to descendants afflicted with cancer of the skin, stomach, or brain. Only genetic modifications of a germ cell that will eventually be involved in the generation of a new individual can be of significance for evolution. [See Epigenetics] Such modifications are the only ones that can influence the development of the fertilized egg. They are also the only ones that can be hereditarily transmitted, as they affect all the cells of the organism, including those that will become germ cells in turn and will give rise to the next generation. (de Duve 2002: 156-157)

-- de Duve, Christian Nobel Laureate. Life Evolving: Molecules, Mind, and Meaning. Oxford: Oxford University Press; 2002: 155-157.

Chapter 4: Development and Evolution

Constraint and Saltation

Developmental biology has long been a focus for evolutionary theory (Bonner 1982; de Beer 1958; Garstang 1922; Goldschmidt 1938; Gould 1977; Haeckel 1866; Raff 1996; Raff and Kaufman 1983; Waddington 1940). Evolution can be seen as a change in developmental programs that elaborate the phenotype. The effects of genes and the range of genetic variation would best be investigated on a mechanistic basis, yet until the 1990s, we had only a very small window on this enormously important developmental landscape.

Once we can understand the nature of development and how it constructs the phenotype, we confront anew some of the age-old questions of evolutionary biology. Development is legendary for its organization, sometimes appearing to be remarkably automatic and even self-organizing. The strong integration of the developmental process might not easily be breached by a mutant, which would disrupt fundamental and tightly integrated cellular and molecular processes. This would suggest a force for conservatism in evolution. On the other hand, the tremendous organization of developmental processes suggest to many that simple genetic changes might beget enormous salutatory evolutionary change.

The Janus-headed coin of development is illustrated well by the evolutionary change of the tail in ascidian tadpole larva, which has been lost in evolution several times independently (Jeffery 1997). (....) This major switch in morphology is associated with a mundane larval adaptation for reduced dispersal by the tail-less form. (Tadpole larvae are not brilliant dispersers, either.) Tail-less development results from the abbreviation of developmental programs owing to maternal message and gene regulation in the zygote. The zinc-finger gene Manx is expressed in tailed species but is downregulated in tail-less species, which suggests a simple mechanism for a momentous developmental reorganization, dropping some of the lynchpins of the chordate anatomical plan (Swalla and Jeffery 1996).

The message told by the Manx gene is not clear, despite teh elegant experimental results. On the one hand, it tells us that it is rather easy to lose the tail and a host of associated developmental trajectories (e.g., notochord, tail, otolith, and muscle cells). (....) If it is that easy, why is it so uncommon? Again, we face teh two faces of constraint and possibility for major change.

Time and again, the concepts of constraint and saltation have been formulated in terms of development. Developmental constraints are nonrandom channelizations of evolutionary direction due to limitations imposed by complex interactions of gene expression and epigentic interactions, such as tissue inductions, in the developing organism. The disruption of such interactions may strongly influence fitness and therefore restrict evolutionary change. In the context of development, saltations are rapid evolutionary fixations of phenotypic discontinuities guided by developmental controls, which do not permit continuity of form in polymorphic populations.

The Holy Grail: Connecting an Understanding of Genes and Development

(....) We are now at the threshold of a completely new period, in which development and genetics are being connected in great detail. At first, this became apparent from the emerging understanding of a widespread homeobox sequence that united all of the triploblastic animals at least. Now, modern methods of gene sequencing, manipulation of gene expression, and tracing of spatial patterns of gene expression have resulted in an explosion of information that is not leading, as yet, to many useful evolutionary rules. So far, we are seeing the same errors promulgated in lionizing past laws of ontongeny and phylogeny. Beliefs in major genetic revolutions, master switch genes, and other universals are beginning to form a modern version of the ontogentic laws of old, with little consideration for the possibilities of convergence in developmental gene function. Nevertheless, the new tools allow us to better peak through the curtains, and the early flush of enthusiasm will likely be followed by substantial advances in development evolution.

Phylogeneticists and Developmentalists

(….) The developmentalists claim that “the diversity of structures that have been formed through the process of evolution is constrained by the rules which govern pattern formation during development” (Stock and Bryant 1981, p. 432). As such, evolutionary change of necessity is the evolution of developmental sequences. The individual, therefore, is treated in terms of its entire ontogeny, and development is therefore both the constraint and target of selection. There is a developmental toolbox, and certain tools may be used in many contexts, but this does impose a possibly limited set of alternative developmental pathways.

(….) [M]orphological structures often come as complete structures or not at all. Of equal interest is the importance of localization in development. Embryos develop only as the result of a complex series of timing events that bring different cells into contact or place cells or molecules of restricted developmental potency in a proper environment for induction. The spatial position of cell groups seems crucial in the generation of morphological patterns, owing to

· Localized intercellular movement and regional movement of dissolved substances that often set gene expression in motion (Garcia-Bellido, Rippoil, and Morata 1973; Summerbell 1981; Turing 1952; Wolpert 1969)
· Transcellular electric fields (Jaffe and Stern 1979; Nuccitelli 1983)
· Mechanochemical interactions (e.g., Odell, Oster, Alberch, and Burnside 1981; Oster, Murray, and Harris 1983)
· Specialized cell adhesion molecules (Edelman 1986)

Must these not influence the direction of evolution? These two phenomena integrity of structure and topological restriction of development suggest that an embryo can be transformed in only limited number of directions during the process of development and evolution. That is the fundamental message about form that Richard Goldschmidt’s pioneering book Physiological Genetics (1938), derived from Spémann (1938), underscored so well.

Some examples of developmental mutants show the discontinuous and often spectacular nature of possible structural change. Consider the cyclops mutant (Bowen, Hanson, Dowling, and Poon 1966) of brine shrimp males. After the fourth instar, the lateral eyes move forward and fuse together, forming a single large compound eye by the ninth instar. During this fusion, the ganglia and nerves of the two optic stalks fuse; the resultant eye resembles the normal medial eye of the cladoceran Leptodora. Thus, a quirk of development has caused a structure to change from that characteristic of one taxonomic order to another! The development of the vertebrate limb shows similar quantum steps.



(….) A developmental notion of macromutation springs from the nature of development described above. If a simple transplant places toes on wings or replaces scales with feathers, why couldn’t evolution occur in major steps? Some have seen such discontinuities in development as a vehicle for major evolutionary jumps (Goldschmidt 1940; Gould 1980a; Lovtrup 1974; Maderson et al. 1982; Schinderwolf 1936, 1950), or at least see them as possible stuff of major saltations (Alberch, Gould, Oster, and Wake 1979; Frazzetta 1970)

-- Levinton, Jeffrey S. Genetics, Paleontology, and Macroevolution. Cambridge: Cambridge University Press; 2001; pp. 157-162.

[Rob]

[T]he preoccupation with the notion that there was a trend toward becoming increasingly humanlike has been wedded to speculations about the processes of natural selection and the sequence of adaptations that could have brought about this putative transformation from ape to human. While other disciplines in evolutionary science have expanded their appreciation of different possible mechanisms and processes of evolutionary change, the study of human evolution has been firmly planted in the traditional dogmas of Darwinism, incluing the belief that evolutionary change manifests itself only through an insensible series of infinitesimally small modifications. Clearly, the search for ancestors and intermediates that has dominated -- and continues to dominate -- the field of paleoanthropology attests to such beliefs. But what if this isn't the way it happened?

Genetics and Development of the Organism

Fossils provide one kind of information on the evolution of the features of organisms: They can potentially illustrate when, approximately, these features first appeared. Even if exact dates are not decipherable, the relative times at which features arose can be sorted out. For example, we know from the primitive fossil birdlike animal Archaeopteryx that the sequence of events in the evolutionary history of birds was the development of feathers first, then the loss of teeth. But when features first arose in groups of animals and how features actually develop are two entirely different things. As such, in the pursuit of a broad understanding of not only hominid but also animal evolution in general, it is important that we not put all our eggs in the fossil basket. Fortunately, coincident with the recent unearthing of fossils that should force paleoanthropologists to reconsider the very basics of the story of human evolution, there has been an explosion of discoveries in the field of developmental genetics that are demonstrating for the first time, from the finest molecular levels to the whole organism, precisely how anatomical features arise and develop.

-- Schwartz, Jeffrey H. Sudden Origins: Fossils, Genes, and the Emergence of Speicies. New York : John Wiley & Sons; 1999; p. 33.

Darwin had based his theory of evolution by means of natural selection in large part on both the availability of variation and on the correlation between variation and the effects of the environment. But no one, not even Darwin, had actually provided data demonstrating the link between organismal change and environmental change: that, indeed, organisms do track their environment, the former changing in the wake of shifts in the latter. Being an admirer of Darwin's meticulousness in amassing huge amounts of detailed information to support his case that individuals of a species do vary to some degree from one another, Bateson set out to perform a similar Herculean task of data collection. He sought to find support for the proposal that the production of variability is intimately connected to the environment in which organisms find themselves. (Schwartz 1999: 194)

The research plan Bateson outlined for himself to investigate this question took him to Western Central Asia, to the lakes and drying-up lake basins of that region of Russia. There, from the spring of 1886 through the fall of 1887, he studied the fauna of the lakes and dying lakes, recording everything, from locality data to local environmental conditions, water density, water salinity, and lake depth, ad infinitum. While there, he also became fluent enough in Russian and Kirghiz to engage in colloquially accurate conversation. But at the end of his studies and arduous fieldwork, Bateson returned to England not full of examples of organisms varying in synchrony with their environments, as Darwin had claimed was the case, but with a conviction that, to the contrary, a tracking of the environment by the organism was not the rule in nature. There appeared to be no correlation whatsoever between change in environmental conditions -- in his particular study, differences in salinity between viable versus dying lakes -- and organismal change. Sometimes many features seemed to change in a number of individuals. Sometimes there was no obvious pattern. Just as certain shifts in environmental conditions might alter the average representation of certain character states of a population in one direction, so, too, could a change in the opposite direction just as easily return individuals to the previous averages of expressed character states. As far as Bateson could tell, there was little here that supported the Darwinian idea that slight shifts in character states from within the set of character states already present within a species would lead to evolutionary change. (Schwartz 1999: 194)

Bateson's experiences in Western Central Asia, and a similar but shorter study of brackish-water organisms in northern Egypt, although utter failures in his mind, prompted him to pursue an in-depth research program on variability in plants and animals. This he accomplished largely through world-wide correspondence, the ransacking (to paraphrase his wife, Beatrice) of museums, libraries and private collections, and extensive travel in order to see for himself purported examples of variation and abnormality. His excitement about this research was palpable in his correspondence. He had not been engaged in this study for even a year when, in September 1888, he wrote a letter to his sister, Anna, in which his intellectual state almost burst from the page: "My brain boils with Evolution.... It is becoming a perfect nightmare to me." (Schwartz 1999: 195)

(....) Eventually, through a series of short publications, and then in his huge and vastly detailed volume on variation, Materials for the Study of Variation, which was published in 1894 -- in which he compiled 886 examples of discontinuous variation -- Bateson came out clearly and solidly on the side of discontinuous variation among organisms. His work on lake salinity and variation in lake-dwelling organisms had convinced him that "most of the elements of the physical environment are continuous in their gradations, while, as a rule, the forms of life are discontinuous." Not only did he question the nature of organismal variation and its importance in evolution but he also rejected the utilitarian Darwinian notion of adaptation and downplayed the role of natural selection in generating evolutionary change. Like Francis Galton and Hugo de Vries, Bateson relegated the minor variability among individuals that Darwin had promoted as the stuff of evolution to just that: the minor ways in which individuals differ from one another in one feature or another. For Bateson, it was in the recognition of discontinuity between characteristics that one gained insight into the workings of evolution: Evolutionary change is discontinuous in the sense that new features arise rapidly -- a proposition that was wholeheartedly embraced by two of the staunchest advocates of saltation in evolution, Francis Galton and Thomas Henry Huxley. (Schwartz 1999: 196)

-- Schwartz, Jeffrey H. Sudden Origins: Fossils, Genes, and the Emergence of Speicies. New York: John Wiley & Sons; 1999; pp. 194-196.

[Rob]

Stephen J. Gould made some interesting comments regarding a phenomena he called textbook orthodoxy, which I quote (in context) the following:

Professional writing tends to be nuanced and judicious. Even the strongest partisan finesses his commitment and adds at least a footnote or tangential comment, so that any charge of oversimplification or dogmatism may be countered by stating: "but look on page 381 (in the small print); you see, I raised the caveat myself." (Gould 2002: 576)

To learn the unvarnished commitments of an age, one must turn to the textbooks that provide "straight stuff" for introductory students. Yes, textbooks truly oversimplify their subjects, but textbooks also present the central tenents of a field without subtlety or apology -- and we can grasp thereby what each generation of neophytes first imbibes as the essence of a field. Moreover, many textbooks boast authorship by the same professionals who fill their technical writings with exceptions, caveats, and complexities. (Gould 2002: 576)

I have long felt that surveys of textbooks offer our best guide to the central convictions of any era. What single line could be more revealing, more attuned to the core commitment of a profession that bathed in the blessings of Victorian progressivism, and aspired to scientific status in Darwin's century, than the epigram that Alfred Marshall placed on the title page to innumerable editions of his canonical textbook, Principles of Economics: "natura non facit saltum." (Gould 2002: 576)

The changing foci of 20th century textbooks provide direct insight into the history of evolutionary thought and the eventual triumph of Darwinism. In particular, if the Synthesis truly hardened, as I have argued, then texts following the 1959 centennial celebrations -- the apogee of strict selectionism [my emphasis] -- should describe evolution in unambiquously panadaptationist language, and should extol the sufficiency of natural selection to craft the entire range of evolutionary phenomena at all scales, ecological to geological. (Gould 2002: 576)

(....) I have consulted everything I could find, including nearly all major American books for introductory college biology (and several high school textbooks as well). A more complete search, extended back in time to cover the early days of the Synthesis, and the pre-synthetic period as well, would provide a fascinating topic for a dissertation in the history of science .... This field of vernacular expression has been neglected by scholars, though the subject would yield great insight (for such material obviously represents the only formal contact that most students ever receive with any given discipline). (Gould 2002: 576-577)

I appologize for my almost anecdotal approach, but I think that I have identified a robust pattern supporting the hypothesis of hardening. I will focus on the two topics that authors of texts found most congenial in their efforts to explain synthetic evolutionism to introductory audiences: the centrality of adaptation, and the sufficiency of synthetic microevolution to explain events at all scales. (I consider here only the evolution chapters of comprehensive biology texts for introductory courses, not entire textbooks on evolution....) (Gould 2002: 577)

-- Gould, Stephen J. (2002) The Structure of Evolutionary Theory. Belkamp: Harvard.

In a sidebox in the introductory textbook for biology Principles of Human Evolution the following is stated:

Beyond The Facts

Is Lamarkism Dead?

Lamarkism collapsed in the face of the clear evidence that the biological information that an individual acquired during its life could not be passed back to the genes, ans so could not be transmitted to the next generation. There is little to challenge this in the main corpus of biology, and so the Darwinian dogma remains to this day.... However, .... that does not mean that the underlying principles were entirely wrong. In Lamark's case there are two general principles that are important.

The first is the question of whether evolution is driven by internal characteristics of the organism, or by the environment. Darwinism of course showed that the later was of major importance, and Neo-Darwinism emphasized that because the source of variation internal to the organism -- mutation -- was random, therefore this was not an important element. While modern biology has largely supported this view, it is also clear that int he emerging field of developmental molecular genetics, the processes occurring internally are more complex than the term "random" suggests, and that an understanding of these mechanisms, some of which may occur across generations, needs to be taken into account.

The second is the question of acquired inheritance .... [w]here Medelian genetics does not operate -- for example, in cultural inheritance, in certain aspects of immunology, in aspects to do with maternal condition -- it appears there may be a chink in the Darwinian armor through which some element of Lamark may enter evolutionary biology.

Roger Lewin and Robert A. Foley (2004) Principles of Human Evolution. Second Edition; Oxford: Blackwell Publishers. p. 45

For a overview of the field of epigenetics in a popular science journal, see Lamarkianism Revisited. For a more in-depth study see Byran M. Turner's (2001) Chromatin and Gene Regulation: Molecular Mechanisms in Epigenetics. Another source for a historical overview of the issues involved see Epigenetic Inheritance and Evolution: The Lamarckian Dimension by Eva Jablonka and Marion Lamb (2005). The following description is given:

Does the inheritance of acquired characteristics play a significant role in evolution? In this book, Eva Jablonka and Marion J. Lamb attempt to answer that question with an original, provocative exploration of the nature and origin of hereditary variations. Starting with a historical account of Lamarck's ideas and the reasons they have fallen in disrepute, the authors go on to challenge the prevailing assumption that all heritable variation is random and the result of variation in DNA base sequences. They also detail recent breakthroughs in our understanding of the molecular mechanisms underlying inheritance--including several pathways not envisioned by classical population genetics--and argue that these advances need to be more fully incorporated into mainstream evolutionary theory. Throughout, the book offers a new look at the evidence for and against the hereditability of environmentally induced changes, and addresses timely questions about the importance of non-Mendelian inheritance. A glossary and extensive list of references round out the book. Urging a reconsideration of the present DNA-centric view prevalent in the field, Epigentic Inheritance and Evolution will make fascinating and important reading for students and researchers in evolution, genetics, ecology, molecular biology, developmental biology, and the history and philosophy of science.

An interesting review of Jablonka's book raises the question of the nature of chromatin marking, and whether it might not be a form of
regulatory genome.

See the folllowing for an interesting historical review of the evolutionary nature of the very definition of the term epigenetics, including the definition given by the renowned molecular biologist Robin Holliday, who defines "epigenetics to include transmission of information from one generation to the next, other than the DNA sequence itself" (Holliday 1994, page 454).

In 1987, Robin Holliday, renowned by this time for his studies of the molecular mechanism by which chromosomes physically recombine, will write, "The properties of genes in higher organisms can be studied on two levels: first, the mechanism of their transmission from generation to generation, which is the central component of genetics and is well understood, and second, their mode of action during the development of the organism from the fertilized egg to adult, which is very poorly understood. The changes in gene activity during development are generally referred to as epigenetic, a term first introduced by Waddington [with reference given here to Waddington, Symp. Soc. Exp. Biol. 7,186 (1953); Principles of Embryology]. Thus, epigenetic switches turn particular genes on or off during the developmental process, producing either transient changes in gene activity or a permanent pattern of activities" (Holliday 1987, page 163).

Seven years later, Holliday will again consider epigenetics. This time, however, he will develop the idea of epigenetics beyond Waddington’s original definition. He will begin with, "The key feature is the nfolding of the genetic programme, which ultimately depends on the activation or inactivation of specific genes, or the interactions between genes and the products of genes" (Holliday 1994, page 453) and then suggest two variations of this definition with the intention of integrating some intriguing observations of gene function. The point we wish to emphasize is that these two variations will incorporate two new concepts to our understanding of epigenetics.

First, Holliday will point out that changes in gene expression occur not only during development but also during the adult stage of an organism. We believe that it is with this thought in mind that he will propose his first variation, a definition of epigenetics that "is not restricted to development, but to organisms that have several or many types of differentiated cells" (Holliday 1994, page 453). Accordingly, he will suggest epigenetics to be the "study of the changes in gene expression, which occur in organisms with differentiated cells, and the mitotic inheritance of given patterns of gene expression" (Holliday 1994, page 453). Holliday will emphasize that this definition "says nothing about mechanisms, so it can include all types of DNA-protein interactions, as well as changes at the DNA level, as seen in the production of genes coding for immunoglobulins. It could also include the alternative splicing of pre-mRNA transcripts to produce protein isoforms, which can be cell type specific" (Holliday 1994, page 453). (Immunoglobulins are proteins that mediate the ability of organisms to fight infection, and mRNA transcripts are RNA products of genes.)

This new definition will also clearly raise a second issue, which is the notion of inheritance. He will note that as changes in gene activity can be inherited through cell division, the "stable mitotic inheritance of given patterns of gene activity is a key feature of epigenetic controls" (Holliday 1994, page 453). How is this inheritance effected? Holliday will first remind us that DNA can undergo permanent changes in sequence during development and that such changes would be expected to be heritable through cell division. (It’s true. DNA will prove to be quite the dynamic molecule!) Holliday will then move on to heritable changes in gene expression that can be reversed at a later stage, sometimes after meiosis. As most reversible changes in gene regulation are not expected to entail alterations of DNA, it is here that Holliday suggests his second variation, which brings the role of non-DNA elements into the limelight. He proposes a "supplementary definition of epigenetics to include transmission of information from one generation to the next, other than the DNA sequence itself" (Holliday 1994, page 454), in other words, "Nuclear inheritance which is not based on differences in DNA sequence" (Holliday 1994, page 454).

So, here we are, at the brink of, but not quite arrived at, the definition of epigenetics which you have found so puzzling. There remains but one more step to reach this final destination, and that is the simplification, in the form of a fusion, of Holliday’s two definitions. Specifically, the most current interpretation of epigenetics combines the concept of changes in gene expression and the implication of mitotic inheritance (from the first variation) with the use of DNA as a reference point and the implication of generational, including meiotic, inheritance (from the second variation) to give rise to our current definition: the study of changes in gene function that are mitotically and/or meiotically heritable and that do not entail a change in DNA sequence.

-- C.-t. Wu and J.R. Morris. Genes, Genetics, and Epigenetics: A Correspondence. Science, vol. 293.

The idea that the inheritance of acquired characters plays an important role in evolution has been the subject of controversy for over a century. Enthusiasm for the idea, which is usually associated with the name of Lamarck, has sometimes led to charlatanism and fraud, while opposition to it has led to 'Lamarkist' being used as a term of abuse. Nowdays, biologists usually regard ideas about the inheritance of acquired characters as nothing more than an interesting part of the history of biology. Lamarkian evolution is rejected on the grounds that there is no evidence for it, no mechanism that can produce it, and no need for it in evolutionary theory. Some people go even further and argue that the inheritance of acquired characters is theoretically impossible--it is incompatible with what is known about genetics and development.

... [T]here are now well recognized mechanisms by which some acquired characters can be transmitted to the next generation, and that such characters have probably played a significant role in evolution. We want to make it clear right at the outset that although we argue that some types of Lamarkian evolution are possible, there is nothing in what we say that should be construed as being anti-Darwinian.[1] We are firm believers in the power and importance of natural selection. What we do maintain, however, is that some new inherited variations are not quite as random as is generally assumed, but arise as a direct, and sometimes directed, response to environmental challenge, and that the effects of such induced variations deserve more recognition in evolutionary theory.

[1] We feel it necessary to stress our belief in Darwinian evolution because recent history has shown than any argument suggesting that Darwinian evolutionary theory should be modified is liable to be used by Creationists as evidence that the theory of evolution is wrong. Like most Darwinians, we believe that Darwinian evolutionary theory is a flexible theory, quite capable of accommodating modifications and amendments.

[Similarly, some Panselectionists attempt to label any "argument suggesting that Darwinian evolutionary theory should be modified" must be a Creationist or Intelligent Design arguments in disguise, and they therefore distort and twist arguments (like Creationists do with the arguments of scientists, both of which is a dishonest tactic I might add) for their own rhetorical purpose. Indeed, they have a lot in common when it comes to such fallacious rhetorical tactics.]

-- Jablonka, Eva and Lamb, Marion J. (1995) Epigenetic Inheritance and Evolution: The Lamarkian Dimension. Oxford. p.1.

In mammals, and to an unknown extent possibly in other groups, male-female reproductive interdependence extends to embryonic gene expression through the phenomenon of genomic imprinting. Genomic imprinting is parent-specific gene expression, that is, gene expression that depends on which parent contributed the gene (Trivers and Burt, 1999), or the differential modification of genes such that maternal and paternal alleles are distinct (Wei adn Mahowald, 1994). For exmaple, the maternally derived allele of the Igf2r gene of the mouse is methylated at intron region 2, which permits the gene to be expressed, evidently because methylation blocks a silencer of transcription (Wei and Mahowald, 1994). The paternal allele is not methylated and therefore is not expressed. Parental imprints are evidently erased, and then remarked, during gametogenesis, to accord with the sex of the individual. Experimentally constituted mouse embryos with an entirely paternally imprinted genome (androgenomes) or with an entirely maternally imprinted genome (gynogenomes) are incapable of normal development and show severe abnormalities. (West-Ehberhard 2003: 635)

(....) Lyon (1993) called genomic imprinting a kind of "epigenetic inheritance." This brings us full circle from modern molecular studies of gene action back to Darwin, who, with his primitive but molecular gemmular theory of pangenesis, insisted that all inheritance is epigenetic, a product of both the transmission and the development of traits. (West-Ehberhard 2003: 637)

-- West-Ehberhard, Mary Jane. Developmental Plasticity and Evolution. Oxford: Oxford University Press; 2003; pp. 635-637.

[Rob]

The Problem of Continuous vesus Discrete Variation and Change

Darwin placed such great emphasis on gradual change and the importance of small variations that he was willing to stake the validity of his theory on this one point (Darwin, 1859 [1872], p. 135). So it is not surprising that critics of Darwinism have long focused on the evidence against gradualism. In 1883, at the height of controversy over Darwin’s theory, the Royal Society of London established a committee eventually known as "The Evolution Committee," where butter controversy over continuous and discontinuous variation divided the members (see Provine, 1971). Among those active in the debate was William Bateson, who in Materials for the Study of Variation (1894) collected hundreds of examples of developmental anomalies proving the existence of discontinuous variation and, therefore, the feasibility of salutatory phenotypic change. In the fascinating period that followed, the discovery of particulate Mendelian inheritance was hailed by Bateson and others as support for saltation, being evidence of the potential importance of discontinuous variation in evolution, while a prominent school of biometricians defeuded the importance of continuous, quantitative variation (see Provine, 1971)…. [T]his was more than an argument over the kinds of variation: if large variants could instantly produce a new form or species, then Darwin’s selection would be seen as relatively unimportant, compared to developmental innovation, in determining the form of organisms and the course of evolutionary change. (West-Ehberhard 2003: 11)

This was only the beginning of a perennial debate in evolutionary biology about the relative importance of continuous versus discrete variation, selection versus development as the cause of adaptive form. While gradualism has prevailed in descriptions of evolutionary change, many valid points raised by Bateson (1894), Goldschmidt (1940), and others, founded on study of discontinuous developmental phenomena, have been put aside without serious consideration. The gradualism controversy always reappears, like a recurrent nightmare, to haunt Darwinians. It is replayed once more in a recent controversy between modern neo-Darwinians (e.g., see Charlesworth et al., 1982; Simpson’s 1984 introduction to his 1944 book) and punctuationists (e.g., see Gould 1981). Although that discussion really concerns variable rates of evolution, not whether or not it occurs by large or small steps …, the gradualism issue resurfaced because large-step regulatory changes were mentioned as potentially more rapid than small-step gradual change mediated by selection on mutations of small effect. The predominant neo-Darwinian view, as concisely summarized b Charlesworth et al. (1982), is that "selection is regarded as the main guiding force of phenotypic evolution" (p. 474). (West-Ehberhard 2003: 11-12)

The controversy over particulate inheritance and gradualism was in large part resolved by quantitative genetics, which emphasizes the additive qualities of the small particulate contributions corresponding to single genes, and shows how polygenic particulate variation can produce continuous distributions within populations, and gradual change. This resolution is discrete, particulate inheritance with gradualism, though central to the modern neo-Darwinian synthesis (see Charlesworth et al., 1982), did not do away with the existence within species of discrete phenotypic variation not directly corresponding to the discrete particles called genes and not obviously or necessarily explained in terms of cumulative small genetic effects. These two kinds of variation continue to foster two views of evolution and a focus on two kinds of data, both of them real. (West-Ehberhard 2003: 12)

(....) There are hidden gaps in our thinking about the continuous and the discrete. Sometimes a partial resolution of the continuous-discrete problem has been mistaken for a complete one, and this in effect has shelved the real problem and left it unsolved. For example, speciation has been declared by some biologists to be the basis of all phenotypic discontinuities. This implies that discontinuities do not originate within species [i.e., individuals], and leaves continuous, quantitative variation the only kind of variation that could be important for evolution within populations. This line of thought minimizes the importance of divergence within populations in the form of polyphenisms, polymorphisms, and the complex developmental anomalies that some see as potentially adaptive novelties, and it leaves the significance of these things unexplained. It reinforces the unimodal adaptation …, which sees as a difficulty or exception the persistence of polymorphisms and other departures form unimodal equilibria. Meanwhile, developmental biology concerns itself with the discrete products of bifurcating developmental pathways but is little concerned with variation or its adaptive significance. It is no wonder, given this profound duality in approach, that the quantitative genetics of continuous variation and the developmental biology of the discrete have long been estranged. (West-Ehberhard 2003: 12)

The unresolved problem of the continuous and the discrete is hidden like a Freudian complex beneath many of the long-standing controversies of evolutionary biology. I believe that this problem reflects a deeper one, the unperceived tension between compartmentalization and connectedness--subdivision and integration--in biological organization itself. (West-Ehberhard 2003: 12)

-- West-Ehberhard, Mary Jane. Developmental Plasticity and Evolution. Oxford: Oxford University Press; 2003; p. 11-12.

Lamark himself had explicitly rejected a direct effect of the environment on animal structures. He believed that new environmental conditions resulted in new activities and habits, and it was these changes in behavior that caused changes in the body. Even in plants, which do not have 'behavior', Lamark stressed that the response to the environment was mediated by its effect on internal activities, such as those associated with nutrition or transpiration. (Lamb 1995: 6)

-- Jablonka, Eva and Lamb Marion. Epigenetic Inheritance and Evolution: The Lamarkian Dimension. Oxford: Oxford University Press; 1995; p. 6.

The non-DNA systems ... are sometimes referred to as epigenetic inheritance systems, although the distinction between genetics and epigenetics, like the distinction between genotype and phenotype, has become rather blurred. (Jablonka 1995: 25)

(....) In recent years, molecular biology has shown that the genome is far more fluid and responsive to the environment than previously supposed. It has also shown that information can be transmitted to descendents in way other than through the base sequence of DNA.... Critical evaluation of the role of the inheritance of acquired variations in evolutionary change is thus avoided by using arbitrary definitions of heredity and acquired variations. (Jablonka 1995: 26)

(....) The importance of the inheritance systems underlying the variations seen in development has only recently been recognized. Nevertheless, ... there was interest in epigenetics and development, and an awareness that the mechanisms underlying development needed to be integrated into evolutionary theory, long before molecular biology began to uncover the types of mechanisms that may be involved. (Jablonka 1995: 27)

-- Jablonka, Eva and Lamb Marion. Epigenetic Inheritance and Evolution: The Lamarkian Dimension. Oxford: Oxford University Press; 1995; pp. 25-26.

It is often assumed that the only variation that is heritable is variation that is associated with genetic differences. However, ..., this position is empirically inadequate, as we now know of several different ways in which phenotypic variation can be reliably inherited through nongenetic pathways (see, e.g., Sollars et al. 2003; Oyama, Griffiths, and Gray 2001; Jablonka and Lamb 2005).

-- Pigliucci, Massimo and Kaplan Jonathan. Making Sense of Evolution: The Conceptual Foundations of Evolutionary Biology. Chicago: University of Chicago Press; 2006; pp. 15, cf. 1.

1. Sollars, V., X. Lu, X. Wang, M.D. Garfinkel, and D. M. Ruden. 2003. Evidence for epigenetic mechanism by which Hsp90 acts as a capacitor for morphological evolution. Nature Genetics 33: 70-74.
2. Oyama S., P. E. Griffiths, and R. D. Gray, ed. 2001. Cycles of Contingency. Cambridge, MA: MIT Press.
3. Jablonka, E., and M. J. Lamb. 2005. Evolution in Four Dimensions: Genetic, Epigenetic, Behavioral, and Symbolic Variation in the History of Life. Cambridge, MA: MIT Press.
4. Jablonka, E. 2001. The systems of inheritance. In Cycles of Contingency, ed. S. Oyama, P. Griffith, and R. Gray, 99-116. Cambridge, MA: MIT Press.

There are multiple inheritance systems, with several modes of transmission for each system, that have different properties and that interact with each other. The include the genetic inheritance system (GIS), cellular or epigenetic inheritance systems (EISs), the systems underlying the transmission of behavior patterns in animal socieities through social leaning (BISs), and the communication system employing symbolical languages (SIS) (Jablonka, Lamb, Avital 1988). These systems all carry information, which I shall define here as the transmissible organization of an actual or potential state of a system. (Jablonka 2001: 100)

(....) Variations were assumed to be exclusively the consequence of the meiotic reshuffling of genes ... and of several classes of errors in DNA maintenance.... Errors that are not removed or repaired accurately b the DNA maintenance machinery were assumed to be the ultimate raw material for evolution by natural selection. Although there is no doubt that a lot of variation in DNA is indeed random in this sense, the view that all variation is random has been challenged. This challenge has come from several directions.

It has been shown that different nucleotiede sequences [modules, such as "introns" and "exons" and "hotspots" and "gene nurseries"] differ in the likelihood that they will be damaged, invaded by genomic parasites or replicated inaccurately. The rate and type of new variation may thus depend on how teh nucleotides in the sequence are organized, and this organization may be adaptive. (Jablonka 2001: 102)

(....) It seems that through natural selection [we don't really know how the origin of this hierarchical system] the mechanisms that allow selective control of gene expression have been coupled with mechanisms that determine teh fidelity of copying so that the inducible system that turns genes on and off [regulatory genome and epigenetic chromatin marking] also turns the production of mutations on and off. (Jablonka 2001: 102-103)

The "targeted" mutations cannot be said to be random in the classical sense, since adaptively advantageous mutations are preferentially (though not exclusively) induced under the appropriate conditions and in the relevant domains. Randomness has not been eliminated, but it has been restricted and channeled. However, the mutations are not goal-directed in any teleogical sense, and their targeted production is the consequence of natural selection that had acted on random variations. Variation has been targeted by selection to be preferentially generated in a subset of sites, under particular conditions. It is difficult to know how to define such variations. The term patterned variation, which has been suggested by the economist Ekkehart Schlicht with respect to cultural evolution, is the one I choose to use in this paper (Schlicht 1997). It is better than previously suggested terms such as directed, adaptive, induced, and guided variation because it does not carry the teleological connotation of premeditated design, yet does carry the connotation of some degree of preexisting structuring (by past natural selection). Once a system for generating patterned variation has evolved, it channels and guides evolution. (Jablonka 2001: 103)

(....) The ability to generate patterned variations forges direct links between heredity, development, and evolution. The generation of patterned variation is part of the developmental process no less than changes in transcriptional activation of genes, although the effect of changes in DNA may often last longer than changes in transcriptional activity. The process of generating patterned variation is part of both development and evolution. Although there is a certain (short-term) degree of autonomy of heredity and development if mutations are random, if they are patterned, heredity loses this partial independence. (Jablonka 2001: 103)

-- Jablonka, E. 2001. The systems of inheritance. In Cycles of Contingency: Developmental Systems and Evolution, ed. S. Oyama, P. Griffith, and R. Gray, 99-116. Cambridge, MA: MIT Press.