“If facts of the old kind will not help us, let us seek facts of a new kind.” (Bateson, W. 1894. ‘Materials for the Study of Variation’ Preface page vi)
Relationships between the processes of development and evolution were central to biological thought in the latter half of the nineteenth century. Accordingly, comparative morphology and embryology were at the pinnacle of the biological sciences. By the turn of the century, though, comparative morphology had been pushed to its limits, and perhaps beyond. Frustration set in. William Bateson described it well. Trained in the old school, he had been investigating what light the anatomy and development of acorn worms might throw on the origin of the chordates. Reflecting on the outcome of this study, he wrote “From the same facts, opposite conclusions are drawn; facts of the same kind will take us no further. Need we waste more effort in these vain and sophistical disputes” (Bateson, 1894). From then on, the old comparative biology was pushed to the sidelines. Any relationship between development and evolution remained peripheral to the major triumphs of twentieth century biology -to the emergence of molecular cell biology on the one hand, and to the ‘modem synthesis’ in evolutionary biology, with all its subsequent ramifications and revisions, on the other.
Yet even Wilhelm Roux, champion of the experimental approach and polemicist against the old biology, envisaged a time when his ‘developmental mechanics’ might encompass phylogenetic studies. He recognised that this endeavour would have to be adjourned until “Entwickelungsmechanik has been developed so far that we have gained deep insights, not only into the mechanisms forming the individual from the germ plasm, but also into the mechanisms that vary the germ plasm” (Roux, 1892; by germ plasm Roux meant something close to our ‘genome’). Would Roux agree that his criteria are, at least in some measure, fulfilled? We believe he would. This volume charts some of the routes that are bringing new facts to bear on these old problems.
Underpinning all discussion is the need for a secure phylogenetic framework. Unlike Bateson, we do not seek to establish relationships by the comparative analysis of development. We have reason to hope that, as the database for molecular phylogenetic analysis expands, the domain where phylogeny equals mythology will shrink. We believe this, not because molecular data are intrinsically ‘better’ than morphological data, but simply because the extent of the potential database is so great. Already, the systematic analysis of a very few molecular species, notably large and small subunit ribosomal RNA, has confirmed that these molecules conserve phylogenetic information that is useful at many taxonomic levels. However, in their survey of these data, Philippe et al. (pp. 15-25) point out just how extensive sequence data must be, even under ideal conditions, to resolve relationships within explosive radiation events. It will be a long while before molecular taxonomy can approach the temporal resolution of a good stratigraphic series.
Molecular phylogenetics may resolve the relationships between taxa -but they can never tell us what sort of beast a stem group species was. Hypothetical archetypes are no substitute for a few good fossils -a point emphasised by Conway Morris (pp. 1-13), and by Coates (pp. 169-180) in his reconsiderations of tetrapod origins. Palaeontology may not often give us a developmental series (juvenile trilobites not with standing) but it can constrain speculation about the end products of ancestral developmental processes - and provide clues to ancestral states that are no longer evident in extant species.
There would be no need of such clues if ontogeny did recapitulate ancestral developmental stages faithfully, but even Haeckel admitted that it does not. Modem species reflect their phylogenetic history selectively, and even then, the reflections are strangely distorted. In the context of vertebrate gastrulation, De Robertis et al. (pp. 117-124) show how molecular probes can reveal the unity of pattern and process beneath the distortions of morphology. Haeckel would be delighted.
One thing that has not been over-written or replaced during the diversification of the metazoa is the basic tool-kit of development. Since the 1940s, we have been used to the idea that metabolic enzymes and pathways are universal -but only recently has it become clear how extensively conserved are the molecules that regulate development: transcription factors, receptors and extracellular matrix molecules. As little as ten years ago, the idea of searching for insect homologues of such ‘vertebrate’ molecules as fibronectin or Myc seemed almost an irrelevance. (All credit to those few who persevered). Now the very idea of ‘vertebrate’ molecules seems a nonsense, and the relevance of functional studies in tractable model systems is established beyond dispute. Chothia (pp. 27-33) utilises data accumulating in the protein and nucleic databases to survey this metazoan tool-kit. He concludes that life uses rather few of the possible protein structures to generate its diversity -the great majority of proteins may be referred to perhaps no more than a thousand structural families.
Within a few years, genome projects will provide a complete inventory of these proteins for a usefully diverse set of reference species. Such sequences alone will be no sure guide to structure or function, but, without doubt, one striking lesson from these inventories will be the extent to which we are indeed one flesh, mite and man and lowly worm. Yes, there will be new genes -mostly made by recombining old domains (Engel et al., pp. 35-42); duplication of old genes will be rife (Holland et al., pp. 125-133; Ruddle et al., pp. 155-161); some old molecules will be seen to have acquired wholly new functions (the lens crystallins are a good case in point (Piato-gorsky and Wistow, 1989)). Even so, as we see it at present, the history of life since the Cambrian has been dominated by the elaboration of regulatory mechanisms that exploit a common set of genes.
How have these regulatory mechanisms evolved? We cannot yet see ‘the big picture’. Are the same cell types specified by homologous regulatory molecules in different phyla? How conserved is the molecular basis for induction, lateral inhibition or neurogenesis? Several papers in this volume provide glimpses of conserved developmental mechanisms – the hedgehog and TGF(3 family signalling molecules used in analogous ways in vertebrates and invertebrates (Fietz et al., pp. 43-51; Hogan et al., pp. 53-60); transcription factors that seem to specify the same organs in vertebrates and invertebrates -heart, eyes, -despite the most diverse morphology (Manak and Scott, pp. 61-77). Are we seeing homologous mechanisms? If so, at what level does the homology lie? Will downstream and upstream regulatory networks be conserved, or are there constrained steps in cellular differentiation (cytotypic stages?) just as there are during embryogenesis, above and below which regulatory networks are more fluid? There is a new world of evolutionary biology here.
The ‘new facts’ of molecular biology pertain not just to molecular phylogeny and cell biology, but to the questions of organismal form -Bauplan; zootype. The Hox genes provide the outstanding example. The linear deployment of Hox genes along the anteroposterior axis of nematodes, insects and chordates provides a strong argument to establish the primitive homology of this axis in all bilateria (Ruddle et al., pp. 155-161; Manak and Scott, pp. 61-77). The expression of the same genes in echinoderms, in molluscs, even in Cuidaria, now provides a criterion to assess how the body axes of these groups relate to those of other triploblasts. Analogous data may yet place Geoffrey St Hilaire’s classic conjecture (1822) concerning the relation of vertebrate and insect dorsoventral axes in the realm of testable hypothesis.
More immediately, the same Hox genes are being used as molecular labels to indicate homology between specific body regions -between insects and crustaceans (Akam et al., pp. 209-215); between cephalochordates and vertebrates (Holland et al., 1992), even between phyla (Morgan and Tabin, pp. 181-186). It remains to be seen whether this ‘internal representation’ of the genes will reveal relationships where comparative morphology has failed. Whether or not it succeeds, the comparison must provide some indication of the mechanisms underlying morphological change, for the Hox genes and their like are not just passive labels, but tools that sculpt morphology.
This direct link with mechanism is perhaps the most important characteristic of the ‘new facts’. In the past, evolutionary change has been analysed by comparing, not the processes of development, but the static forms generated by these processes. Increasingly, it is becoming possible to compare the processes themselves, at the cellular level (Wray and Bely, pp. 97-106; Sommer et al., pp. 85-95) as well as the molecular (Patel, pp. 201-207; Tautz et al., pp. 193-199; Morgan and Tabin, pp. 181-186). Only when we understand the process of development can we begin to map the relationship between genetic change and morphological effect. It is a commonplace of developmental genetics that minimal genetic change can lead to the most dramatic morphological effect (a single base substitution in the bicoid gene of Drosophila can reverse the axes and symmetry of the embryo (Frohnhôfer and Niisslein-Volhard, 1986; Struhl et al., 1989)). What we do not yet know is the genetic complexity of observed transitions in evolutionary history -of heterochronic changes in rates of growth, of duplications or suppression in segmentation, or inventions of morphological novelty. Papers in this volume provide glimpses of understanding. How did the complex and beautiful patterns on the wings of a butterfly arise - and how have evolutionary pressures moulded them for immediate adaptation? Nijhout (pp. 225-233) sketches, in a formal model, the outlines of a common mechanism that can generate the apparent complexity; Carroll (pp. 217-223) raises the hope that genes we already know, identified in Drosophila, may provide the material basis for part of this complexity.
We do not fully understand butterfly wings, insect segments or vertebrate limbs. Far from it. But we can now pose questions that address the diversity of life in geological time and species space, with some hope of finding answers that are neither trivial nor obvious: answers that go some way towards illuminating that obscure sector on the Venn diagram where genetics, evolution and development intersect.
M. A., P. H., G. W., August 1994
