Heterochrony

Advanced article Article Contents

Tim Horder, University of Oxford, Oxford, UK

. A Basic Definition

Heterochrony is a supposedly important mechanism for macroevolutionary change based on shifts in the timing of developmental events in the embryo.

. The Shifting Meanings of the Term . The Situation Today . An Evaluation

A Basic Definition

doi: 10.1038/npg.els.0004180

This apparently straightforward concept – the term (from Greek heteros different, changed; khronos timing) is almost self-explanatory – has had a remarkably complex history since its introduction by Haeckel in 1875. It is a key concept within evolutionary biology today, but at the same time something of a fig leaf: underneath the seeming simplicity there lies a multiplicity of uncertainties and complications still remaining to be defined and clarified. Heterochrony can be defined as ‘a change in timing of embryonic development through evolution’. The paradigm example is the Mexican axolotl: apparently as a result of slowing and truncation of development during its evolution, the adult of this salamander has acquired the distinctive morphological form of the larval stages of its ancestors. ‘[Heterochrony] seems to best explain most of those traits that make us human: large body size, large brain, long learning stage and lifespan’ (McKinney and McNamara, 1991, p. xi). With origins in Haeckel’s idea of recapitulation (the notion that the stages of ontogeny reflect – and are an accelerated version of – the stages in the phylogenetic evolution of the organism concerned), the concept embodies the central proposition that shifts in timing of events during embryogenesis underlie important changes during evolution. ‘Heterochrony has had a profound effect on macroevolutionary thought’ (Raff, 1996, p. 256), for example, in the study of basic phylogenetic trends in evolution and their underlying mechanisms. In the 1990s, heterochrony became a popular explanation for many forms of evolutionary change. For examples, see McKinney and McNamara (1991), McNamara (1995) and (for botanical cases) Schlichting and Pigliucci (1998). See also: Caudata (salamanders) However, the initial simplicity breaks down when one considers the variety of situations in which such a process might be thought to operate and the diversity of its possible effects. A number of classifications of ‘types’ of heterochrony have emerged and the concept is closely linked with such phenomena as neoteny, paedomorphosis and vestigialization. In its time the term has undergone a series of radical changes in meaning, sometimes amounting to almost total reversals as regards their implications. Many of the historical twists and turns still reverberate in today’s thinking. Gould (1977) remains the main source for more exact definitions and for the history of the subject (see Gould (1990) for a good summary). Subsequent accounts

have largely followed Gould’s historical approach, and much the same will be done here. However, the worst terminological excesses of the past, and some of the more elaborate ongoing discussions, will be avoided in favour of seeking a general perspective on the central concept.

The Shifting Meanings of the Term Because the concept of heterochrony has occupied a crucial position in any discussion of the interrelations of ontogeny (developmental biology) and phylogeny (evolutionary biology) for over 120 years, it is not surprising that its shifting meanings have reflected the advances and vastly changing concerns within biology over that period.

Stage 1 (1866–1875): embryonic development seen as a direct record of the sequence of events in evolution Haeckel (adapting an earlier use by Virchow (Hopwood, 2005)) introduced the term heterochrony as one of several recognizable departures from his basic proposition that the stages of ontogenesis repeat the evolutionary stages of phylogenesis (‘ontogeny recapitulates phylogeny’). He defined heterochrony precisely and specifically as a displacement in time of the ontogenetic appearance and development of one organ with respect to others, causing a disruption of the true overall repetition of phylogenesis during ontogenesis.

Stage 2 (c. 1875–1925): the accumulation of evidence for and against recapitulation During this muddled period, when the idea of recapitulation came under repeated attack but still dominated evolutionary thought, the concept of heterochrony became broadened. It was argued (for example, by Cope) that within each organ the stages of development recapitulate the evolution of that organ as such, despite departing from synchrony with the embryo as a whole. Thus, for some theorists the concept came to refer not so much to exceptions to

ENCYCLOPEDIA OF LIFE SCIENCES & 2006, John Wiley & Sons, Ltd. www.els.net

1

Heterochrony

true recapitulation in the embryo as a whole but also began to include shifts in timing (such as acceleration), illustrative of the recapitulative mechanism itself.

Stage 3 (De Beer, 1930): a landmark statement of the arguments against recapitulation De Beer broadened the concept still further. The arguments against Haeckelian recapitulation were now so significant that de Beer effectively rejected it in all its forms. However, he retained the term ‘heterochrony’ and used it to cover a variety of nonrecapitulatory changes in timing that he thought went a long way towards explaining evolutionary change. De Beer now defined it as follows: [Regarding] alteration and reversal of the sequence of stages (of development) _ the principle of heterochrony will make it possible for any structure to appear later or earlier as well as at the corresponding time, when compared with an earlier ontogeny_ Heterochrony provides eight possibilities of variation in the way in which structures may appear in an ontogeny as compared with the ontogeny of the ancestor (De Beer, 1930, pp. 35–36).

A major difficulty with de Beer’s revamping of the scope and implications of the term was the byzantine complexity of the resulting new subclassification of heterochrony and its associated terminology. But de Beer was spokesman for a viewpoint (deriving from Garstang) that had a considerable impact at the time, focused especially on paedomorphosis. This is a form of heterochrony that involves precisely the opposite mechanism (retardation or slowing of development) to that of recapitulation (which requires compression into the development phase in the descendant of the ancestral phylogenetic stages through acceleration of ontogenetic stages). In paedomorphosis (Figures 1 and 2), the larval phase in an ancestor is slowed (or sexual maturation is brought forward by acceleration into the larval phase), so that the descendant mature adult retains the morphology of the original larval form. Such a

Figure 2 A classic example of the role of paedomorphosis in the origin of new phyla and novel Bauplans. The larval ‘tadpole’ stage of an invertebrate ascidian (a) is here seen as the possible starting point for the origin of the chordates. According to this theory, maturity would be reached in the larval stage and the normal invertebrate adult stage (right) would be abandoned.

mechanism was used to explain large-scale evolutionary trends, i.e. as the origin of the evolution of entirely new adult Bauplans (Figures 2 and 3), and as a means of ‘escape from specialization’ out of the supposed dead-ends implicit in already very specifically adapted phylogenetic lines. See also: Beer, Gavin Rylands De De Beer’s eight categories of heterochrony included caenogenesis, neoteny, retardation and reduction. The latter two categories were used to explain the formation of vestiges and the gradual disappearance of redundant structures during evolution. Neoteny (Figure 1) is one form of paedomorphosis. Caenogenesis (which Haeckel had used in a wider sense as a term for all classes of deviation from strict recapitulation, heterochrony being just one) is for de Beer the insertion of changes early in development, including the acquisition of a distinct intermediate, larval phase in early development in an organism previously lacking one. Even more confusingly, one category (termed ‘acceleration’) involved a process akin to recapitulation. De Beer’s scheme marks a definite change in thinking: now heterochrony is treated as an actual mechanism for evolutionary change, rather than as a form of passive record of

Figure 1 The classic example of neoteny. The neotenous Mexican axolotl, unlike other urodeles, becomes adult and achieves sexual maturity while still retaining typical larval features, including external gills and tail fins (a). However thyroxine can induce metamorphosis artificially, to a typical urodele adult form as shown in (b). After Huxley (1922).

2

Heterochrony

Figure 3 Human evolution has been interpreted as a case of paedomorphic modification and as the retention of the younger developmental stages of an ancestral ape. The young chimpanzee (a) has obviously human-like features compared to the adult ape (b). Reproduced from Naef (1926) # Springer.

previous phylogeny. De Beer supported his mechanistic view with evidence (from Goldschmidt, Ford and Huxley) that the timing of development can be varied by genetic factors. Furthermore, he pointed out how known hormones could mediate these effects (Figure 1).

Stage 4 (Gould, 1977): the initial laying out of present-day perspectives Under the weight of its accumulated obscurities and complexities, heterochrony had largely dropped out of the biologist’s lexicon by 1977, when Gould addressed a lengthy and scholarly tome to the subject (Gould, 1977). With the primary aim of simplification, Gould pointed out that four of de Beer’s categories (including ‘adult variation’ and ‘deviation’) did not really refer to time changes at all – they might best be described as referring to cell differentiation and the substitution of novel forms of molecular and cell expression during evolution. These he therefore discarded. He also argued that the same apparent evolutionary end results in adult morphology could be arrived at by permutations of just two mechanisms. Depending on whether reproductive maturation or the body as a whole was undergoing a change in developmental timing, four categories of heterochrony leading to two basic types of adult morphological change could occur. The first type he termed ‘recapitulation’, the result of either acceleration of body development (de Beer’s ‘acceleration’) or retardation of reproductive maturation (de Beer’s ‘hypermorphosis’); the second was ‘paedomorphosis’, the result of either accelerated maturation (now termed ‘progenesis’) or retardation

of body development (de Beer’s ‘neoteny’). Thus, by reducing the underlying processes (acceleration or retardation of development) to two, Gould was able to simplify and unify the remaining four of de Beer’s categories. Gould went on to represent his scheme diagrammatically in the form of a ‘clock-model’. In effect, Gould single-handedly re-launched the issues and problems that had preoccupied biologists in the 1930s. Gould’s analysis of the term (and many of his arguments and data) derives explicitly from de Beer’s time. Rather than by direct appeal to more recent embryological or molecular genetic evidence, Gould supported his two-way classification of heterochronic trends by showing that they were often associated with different life-history strategies for ecological adaptation. Progenesis favours early reproduction in the life cycle (and r-selection), whereas neoteny favours K-selection. The delayed maturation in the latter case could correlate with the evolution of such features as advanced socialization in higher vertebrates. Without basically changing the way of thinking laid down by de Beer, Gould subtly shifted emphasis on to the whole organism (away from organs) and on to larvae and life cycles (away from embryos as such). Confusingly, Gould retains the word ‘recapitulation’ (whereas de Beer had wanted to substitute the clearer term ‘repetition’ of embryonic stages). Together with the overall process of simplification of the concept of heterochrony undertaken by Gould, there was also a tendency towards formalism and even quantification. This started with the clock model but was soon reinforced (Alberch et al., 1979) by a revised formalization of Gould’s definitions: life histories were now described graphically and algebraically in the form of 3

Heterochrony

‘trajectories’. Whereas Gould’s clock model depicted dissociative changes between the start and end of the life cycle in the relation of size and shape – dimensions reflecting his earlier interest in allometry – Alberch et al. depicted rate changes of broad classes of developmental events at different times through development in two comparison species.

The Situation Today Gould’s analysis coincided with, and to some extent triggered, the renewal of interest in the developmental mechanisms underlying evolution that continues apace today. Heterochrony is a subject that has since attracted considerable attention (McNamara, 1997; Hall, 1998; Klingenberg, 1998; Schlichting and Pigliucci, 1998), perhaps because there are few established alternatives by way of concepts that link development and evolution. Much seems to hang on the term: Heterochrony is the cause of most developmental alterations (McKinney and McNamara, 1991, p. 47). It is the effect of alterations to the timing and rate of development that is one of the cornerstones of evolution (McNamara, 1995, p. ix) The major proposals seeking to account for the mechanistic roots of evolution in ontogeny have centred on heterochrony (Raff and Kaufman, 1983, p. 173). The majority of studies on the evolution of development use models of heterochrony as a framework (Schlichting and Pigliucci, 1998, p. 115).

Approaches to heterochrony have moved somewhat beyond Gould’s position and a few new trends can perhaps be detected. The quantitative formalism initiated by Gould is one possible explanation for the noticeable convergence of the concept of heterochrony with the concept of allometry. The two notions are now often treated together and even interchangeably. The initially comforting simplicity of Gould’s scheme is in danger of collapsing as a new proliferation of terminology is beginning to confuse matters yet again. A recent treatment (McKinney and McNamara, 1991, p. 55) distinguishes ‘global growth heterochronies’, ‘local growth heterochronies’, ‘sequential and organizational heterochronies’ and ‘differentiative heterochronies’. A number of more recent studies seem to aim at extending the use of the concept to the earliest developmental stages, and ultimately to a basis in cellular and molecular terms.

An Evaluation A significant difficulty in pinning down the concept of heterochrony is that the meaning of the term is in a continual 4

state of flux according to the multifarious and sometimes incompatible aims of the evolutionists concerned. The terminological complexity verges on the bizarre, and has often led to inconsistent and confused usages. Particularly problematical is the still unresolved linkage of heterochrony with the equally confused concept of recapitulation. Yet, there has been a consistent thread running through the history of the subject; the concept has stood as a constant reminder of the inescapable point that ontogenesis and phylogenesis are inextricably interrelated. The potential importance of the subject to our understanding of evolution is suggested by the way in which heterochrony has been part of debates on such crucial macroevolutionary themes as saltation, punctuated equilibria, internal factors, orthogenesis and body plans. Evidently, an evolutionary adaptation that is relatively freely open to any species is to curtail the normal full span of embryonic development. Often correlated with miniaturization and rapid reproductive maturation, this seems to be the simplest and most common form of heterochrony and is seen on occasion in most phyla. See also: Punctuated equilibrium and phyletic gradualism Historically, there has been a drift in the definition of heterochrony from one of change within parts or organs of an organism (especially the reproductive organs) towards one of change affecting the organism as a whole, but, despite their different implications and the different methods of measurement required, the two versions are still applied interchangeably. Moreover, there has been a concentration on species with distinctive larval stages and on the process of metamorphosis. As a model of heterochronic processes more generally, the mechanism of metamorphosis (which may be relatively simply controlled genetically, endocrinologically and by environmental triggers; (Tompkins, 1978; Semlitsch and Wilbur, 1989) may be misleading: it is a highly specialized situation and is not necessarily representative of other aspects and stages of embryogenesis. The mechanism of metamorphic switching in itself reveals nothing about how the larval and adult forms themselves acquired their specific characters during evolution (or indeed about embryogenesis in any organism, including species without larval stages). Heterochrony can only vary developmental patterns already established; it can contribute nothing to the more interesting question of how novel features arise in evolution in the first place. Raff (1996) has recently pointed to some of the difficulties in present-day thinking. The interpretation of human evolution in terms of neoteny in particular has been criticized in some detail (reviews in Raff, 1996; Klingenberg, 1998) on the grounds of the arbitrariness of the definition of stages and rates of development when compared across species and also on the grounds that individual morphological structures evolve independently rather than coherently, all at a single rate. Heterochrony is a concept invoked above all by palaeontologists, whose material is effectively limited, inevitably, to adult morphologies. For them, age

Heterochrony

or development cannot usually be assessed in any direct sense: size is often used as proxy for age or developmental stage. Even embryologists – and this perhaps partly explains their usual indifference to heterochrony (see Richardson (1995) for a recent exception) – would regard the notion of a direct measurement of ‘rates of development’ as one of extreme difficulty. Which structure or process should be selected? What does one use as a reference point against which to measure timings when so many parameters in development are known to be open to variation and adjustment in time and space? Even if we stick for the time being with the version of heterochrony proposed by Gould and his immediate successors, there remain some problems of a quite fundamental nature. Has the concept not become too remote from what we already know about the real causal factors in embryology? At present heterochrony seems to be a catch-all notion. How much can heterochrony be expected to explain, even in principle? How can a unidimensional parameter like time conceivably give rise to the multidimensional entity that is an adult organism? What about other quite distinct developmental mechanisms, such as morphogens, cell–cell interactions, local growth factors and the specific inductions that switch on cell differentiation, all of which must change in evolution and must have consequences in changing adult morphologies? Is it appropriate to be looking at such developmental problems specifically from the perspective of time as such: is time really the operative parameter that developing cells respond to when selecting their pathways for adult specialization? Would not the position in embryonic space and growth (which in practice always vary as a function of developmental time) be more relevant parameters? And how is one to decide whether, in a given instance of supposed heterochronic change, timing is the actual cause or merely a correlate or a consequence of other unknown mechanisms? Are heterochronies merely redescriptions of events actually caused in entirely different ways? (Gould (1988, p. 3) has made an eloquent attempt to rebut any doubters). If heterochrony shall be the empirical focus of the organismic biologist’s approach to development as an evolutionary force, then we must ask if this notion of change in developmental timing can shoulder such a burden. Or are its ambiguities too great, its promises too empirically intractable, or its occurrences simply too infrequent? _ [These] problems can be overcome and promises fulfilled.

The concept of heterochrony has in essence changed little from its formulation at a time when understanding of genes, development and evolution was at a primitive stage. Regarding actual mechanisms, our views on how heterochrony might work have largely been dictated by the special case of metamorphosis. In general, heterochrony has been used as a surrogate for detailed unravelling of the

complexities of embryonic mechanisms as such. It can be argued that the concept of heterochrony can only ever be a ‘descriptive’ device – like allometric phenomena, best suited to higher level, morphological aspects of organisms irrespective of the differentiation of constituent cells – which in itself is unlikely to identify causal mechanisms in definitive terms. Whether the concept should survive and whether it can eventually be made more securely based in cellular, genetic and molecular terms remain moot points. The concept sits most uneasily with much of what we now know about the basis on which biological systems change during evolution. Everything points to the fact that evolutionary change can be adequately accounted for in terms of changes in the vast and varied pool of deoxyribonucleic acid (DNA) sequences and their highly specific locally expressed protein products, which together define the characteristics of organisms. Ultimately, it is in these terms that we must expect to find the only realistic way to account for the multidimensionality of embryogenesis and of adult morphology that evolution requires. Heterochrony may have outlived its usefulness. See also: Evolution: history

References Alberch P, Gould SJ, Oster GF and Wake DB (1979) Size and shape in ontogeny and phylogeny. Paleobiology 5: 297–317. De Beer GR (1930) Embryology and Evolution. Oxford: Clarendon Press. [Later editions retitled Embryos and Ancestors]. Gould SJ (1977) Ontogeny and Phylogeny. Cambridge, MA: Harvard University Press. Gould SJ (1988) The uses of heterochrony. In: McKinney ML (ed.) Heterochrony in Evolution: A Multidisciplinary Approach, pp. 1–13. New York: Plenum Press. Gould SJ (1990) Heterochrony. In: Keller EF and Lloyd EA (eds) Keywords in Evolutionary Biology, pp. 158–165. Cambridge, MA: Harvard University Press. Hall BK (1998) Evolutionary Developmental Biology, 2nd edn. London: Chapman & Hall. Hopwood N (2005) Visual standards and disciplinary changes: normal plates, tables and stages in embryology. History of Science 43: 239–303. Huxley J (1922) Ductless glands and development. Amphibian metamorphosis considered as consecutive dimorphism, controlled by the glands of internal secretion. Journal of Heredity 13: 349–358. Klingenberg CP (1998) Heterochrony and allometry: the analysis of evolutionary change in ontogeny. Biological Reviews 73: 79–123. McKinney ML and McNamara KJ (1991) Heterochrony: The Evolution of Ontogeny. New York: Plenum Press. McNamara KJ (ed.) (1995) Evolutionary Change and Heterochrony. Chichester, UK: Wiley. McNamara KJ (1997) Shapes of Time. The Evolution of Growth and Development. Baltimore, MD: Johns Hopkins University Press. Naef A (1926) U¨ber die Urformen der Anthropomorphen und die Stammesgeschichte des Menschenscha¨dels. Naturwissenschaften 14: 445–452. Raff RA (1996) The Shape of Life. Genes, Development, and the Evolution of Animal Form. Chicago: University of Chicago Press.

5

Heterochrony

Raff RA and Kaufman TC (1983) Embryos, Genes, and Evolution. The Developmental-Genetic Basis of Evolutionary Concepts. New York: MacMillan. Richardson MK (1995) Heterochrony and the phylotypic period. Developmental Biology 172: 412–421. Schlichting CD and Pigliucci M (1998) Phenotypic Evolution. A Reaction Norm Perspective. Sunderland, MA: Sinauer Associates. Semlitsch RD and Wilbur HM (1989) Artificial selection for paedomorphosis in the salamander Ambystoma talpoideum. Evolution 43: 105–112. Tompkins R (1978) Genic control of axolotl metamorphosis. American Zoologist 18: 313–319.

6

Further Reading Gould SJ (1990) Heterochrony. In: Keller EF and Lloyd EA (eds), Keywords in Evolutionary Biology, pp. 158–165. Cambridge, MA: Harvard University Press. Minush-Purvis N and McNamara KJ (2002) Human Evolution through Developmental Change. Baltimore: Johns Hopkins University Press. Raff RS and Kaufman TC (1991) Embryos, Genes and Evolution. The Developmental-genetic Basis of Evolution. Bloomington: Indiana University Press. Wilkins AS (2002) The Evolution of Developmental Pathways. Sunderland, MA: Sinauer.