Reconstructing the Evolution of Viviparity and Placentation

Reconstructing the Evolution of Viviparity and Placentation

J. theor. Biol. (1998) 192, 183–190 Reconstructing the Evolution of Viviparity and Placentation D G. B Department of Biology, Life Scien...

151KB Sizes 0 Downloads 21 Views

J. theor. Biol. (1998) 192, 183–190

Reconstructing the Evolution of Viviparity and Placentation D G. B Department of Biology, Life Sciences Center, Trinity College, Hartford, CT 06106, U.S.A. (Received on 18 July 1997, Accepted in revised form on 11 November 1997)

A recent paper in J. theor. Biol. has challenged the proposed application of punctuated equilibrium models to the evolution of reptilian viviparity and placentation. While clarifying some aspects of the models, the paper’s criticisms reflect misinterpretations of the literature and an unnecessary reluctance to apply punctuationist concepts to extant taxa. The punctuated equilibrium model retains potential for clarification of the patterns of stasis and episodic change that may well have characterized squamate reproductive history.

7 1998 Academic Press Limited

Introduction Reptilian viviparity (live-bearing reproduction) and placentation have fascinated biologists for over a century (Giacomini, 1891; Weekes, 1935; Shine, 1985; Stewart, 1989), and commonly are assumed to have evolved incrementally via progressive, gradualistic evolutionary change (Blackburn, 1992). In a recent paper (Blackburn, 1995), I presented two alternative models for their evolution in squamates (lizards, snakes, and amphisbaenians)—one involving saltational change, and the other punctuated equilibria, in which periods of stasis are interspersed with relatively rapid episodic change. Testing these models against empirical data, I suggested that squamates exhibit a dichotomy between two evolutionarily stable strategies: deposition of eggs in early development (typical oviparity) and retention of the young to term (viviparity) with functional placentation. In a response, Qualls et al. (1997) discussed these issues in detail, offering new data and alternative interpretations. Their response has raised issues that deserve careful consideration, because how the evolutionary models are defined and interpreted may direct future research in areas of broad interest to biologists. The purpose of this paper is to reconsider these issues in light of their perspective (including aspects that these authors have helped to clarify), as well as newly available evidence, as a contribution to ongoing dialogues about the evolution of vertebrate 0022–5193/98/100183 + 08 $25.00/0/jt970592

viviparity (e.g. Wourms & Callard, 1992; Shine, 1995). Perhaps the most significant point to be noted is the congruence between our views, for Qualls et al. (1997) agreed with Blackburn (1995) that squamate reproductive modes exhibit long periods of evolutionary stasis; that viviparity can evolve rapidly from oviparity; that the reproductive modes are bimodally distributed; that available evidence is incompatible with the saltationist model; and that viviparity and placentation evolve concurrently. However, their contribution advanced several strong criticisms, notably what they consider to be misapplication of punctuational and gradualistic models; adoption of an invalid predictive test for saltational change; and use of inappropriate scenarios for the evolution of placentation. Although some of these issues represent differences in perspective, I shall show that most reflect misconceptions and misinterpretations, and reliance on incomplete information. These issues appear to be resolvable with data now available, and each, along with recent evidence bearing on the transition to viviparity, deserves careful reconsideration.

Punctuated Equilibrium and Phyletic Gradualism Qualls et al. (1997) accepted the view that squamate reproductive history has been characterized by 7 1998 Academic Press Limited

184

. . 

periods of stasis interspersed with periods of rapid evolution towards viviparity and placentation. Nevertheless, they argued (p. 130) that interpreting this pattern in terms of punctuated equilibria is inappropriate due to historical precedent: that because the concept of punctuated equilibria was first put forward to explain patterns in the fossil record, it is not applicable to non-paleontological data. However, if punctuationism describes patterns of evolutionary change, and not simply the relative likelihood of fossil discoveries, then several reasons exist for not excluding extant species. Living organisms offer far better resolution than does the fossil record for patterns of speciation and change (and incidentally, provide most of our evidence for the evolution of squamate reproductive modes). Consequently, proponents and opponents of the concept of punctuationism have agreed that evidence bearing on its validity is likely to come from populational data, which are difficult to obtain among fossil forms (e.g. Maynard Smith, 1988; also see Stanley, 1979). As for historical precedent, origins of the concept trace directly to work by Ernst Mayr on extant species (e.g. Stanley, 1979; Mayr, 1992; Eldredge, 1993). Moreover, from the outset, the architects of the concept of punctuationism have supported its application outside of the paleontological realm: ‘‘. . . we believe that a general theory of punctuational change is broadly, though by no means exclusively, valid throughout biology’’ (Gould & Eldredge, 1977, p. 45). Accordingly, the concept of punctuated equilibrium is commonly applied to extant species, including by some of its most prominent supporters (Stanley, 1979, 1981). Thus, our chief criterion should be utilitarian: whether punctuational concepts lead us to think differently about a transformation that has long stood as an example of progressive, incremental change. A second objection had to do with my use of the terms ‘‘gradualism’’ and ‘‘gradualistic’’ in reference to phyletic gradualism. Qualls et al. (1997, p. 130) asserted that such usage will mislead people unfamiliar with the subject matter into thinking that ‘‘gradual evolution’’ in general (i.e. non-saltational change) was under attack. Misunderstanding by non-specialists is no small concern, but professional publications require established terminology. My approach was to adopt terms in widespread use in the evolutionary literature (including in work by Gould, Eldredge and Stanley), and to define them operationally to minimize confusion. ‘‘Phyletic gradualism’’ is a cumbersome term that does not lend itself to adjective or to adverb phrases; perhaps for this

reason, the paper by Qualls et al. falls into use of ‘‘gradualism’’ and ‘‘gradualistic’’ identical to that in my paper. Qualls et al. made a useful distinction between the magnitude of evolutionary change and the rate of change, and characterize the evolutionary models according to these parameters. Their characterization has helped clarify the models by emphasizing that phyletic gradualism and stasis differ according to presence or absence of net change in a particular direction. However their application of phyletic gradualism to viviparity yields surprising results: ‘‘through time, a squamate lineage must either continuously flip–flop between oviparous and viviparous reproduction (anagenesis), or continuously give rise to new lineages of the opposite reproductive mode (cladogenesis), which then must either continuously flip–flop or give rise to new lineages of the opposite reproductive mode, which then . . .’’ (p. 131). The authors never explain why they consider it necessary to postulate repeated reversals (‘‘flip– flops’’) between oviparity and viviparity. Phyletic gradualism simply posits net change in some direction, with the possibility of minor fluctuations (whose magnitude is affected by selection) during the period of change. Applied to the transition to viviparity, one would expect a trend involving gradualistic net increases in the degree of egg development in the oviducts prior to oviposition. Once a trait (e.g. viviparity) has evolved, there is no implication that the particular trend can or must continue, and certainly no implication that it undergoes reversal. In fact, viviparity has long been interpreted as a pattern that seldom if ever reverts to oviparity (Fitch, 1970; Blackburn, 1985b; Shine, 1985). I fully agree with the authors that their scenario is restrictive and unrealistic, but not that it follows from (or bears much relationship to) the concept of phyletic gradualism. Reluctance to apply concepts of phyletic gradualism and punctuationism to viviparity seems to reflect these authors’ views on the legitimacy of the former as a model for change. Qualls et al. (1997, p. 130) characterize the concept of phyletic gradualism itself as ‘‘an unrealistic model that is ‘not consistent with modern evolutionary ideas’ (Gould & Eldredge, 1977).’’ In context, the quotation they invoke actually imparts a different idea: Gould & Eldredge (1977) stated ‘‘We merely urged our colleagues to consider seriously the implications for the fossil record of a theory of speciation upheld by nearly all of us, and to recognize the search for phyletic gradualism as a bad

   historical habit not consistent with modern evolutionary ideas’’ (p. 117; emphasis added). Gould & Eldredge (1977, p. 115) were not asserting that phyletic gradualism never occurs, but that it ‘‘is very rare and too slow . . . to produce the major events of evolution’’. Moreover, they devoted nearly 20 pages to testing whether gradualistic or punctuational models offer better explanations of observed patterns, and much of the rest of their monograph to distinguishing between the two models as explanations for change. If phyletic gradualism has not long offered biologists a plausible explanation for observed patterns, the proposal of punctuationism and the ensuing decades of controversy would have been unnecessary. Further, evolutionary biologists would have no reason to continue to test their data against each of the two models, as is done routinely in the literature up through the present (Gould & Eldredge, 1986; Hoffman, 1992; Eldredge, 1993). Although phyletic gradualism and stasis can be distinguished according to presence or absence of net change, there are other potential aspects that distinguish the models that ought not be overlooked. A basic feature of all but the weakest version of punctuationism is that most phenotypic change accompanies speciation, particularly of the cladogenic variety (Stanley, 1979; Somit & Peterson, 1992); in contrast, phyletic gradualism implies anagenetic change that may or may not accompany speciation. Further, although punctuational change can occur at the same pace as that of phyletic gradualism, the former concept allows for periods of highly accelerated change as well. In addition, the idea that speciation occurs in peripheral isolates as a cladogenic process is fundamental to punctuationism (Eldredge & Gould, 1972; Gould & Eldredge, 1977). Moreover, in the context of ‘‘adaptive landscapes’’, the punctuational model for viviparity (Blackburn, 1995) holds that intermediate phenotypes (e.g. prolonged oviparous egg retention) are evolutionarily unstable, and therefore rare. Qualls et al. accept the idea that viviparity can arise rapidly in peripheral isolates via episodic change, and that such change is interspersed between long periods of stasis. For the above reasons, we ought not be reluctant to interpret the pattern as punctuational. In considering application of punctuated equilibrium to the situation, we can estimate the duration of periods of stasis from the fossil record. Major extant squamate clades trace their origins to the Jurassic, and their closest common ancestor may date to the Triassic (Carroll, 1988). If deposition of eggs in the limb bud stage is primitive for extant clades, as parsimony suggests (Blackburn, 1995), then ovipar-

185

ous squamates probably have been in stasis (in this regard) for 150–210 million years. Juxtaposed against the fact that viviparity has often arisen at low taxonomic levels in geologically recent times, the temporal contrast is striking. The majority of the known origins of reptilian viviparity have occurred at subgeneric levels (Blackburn, 1982, 1985b; Shine, 1985), and many probably date from the Pliocene and Pleistocene. With rare exceptions, we have little if any reason to suspect reversions to oviparity (Blackburn, 1985a,b; Shine, 1985); moreover, only three viviparous clades are known to have gone on to develop substantial placentotrophy (Blackburn, 1985a,b, 1993a). Stasis, it appears, is a dominant pattern in the history of squamate reproductive modes. Saltational Change In generating predictive tests, my paper reasoned that if viviparity arises by saltation (macromutation), then intermediate forms cannot exist, even in lineages where viviparity has evolved from oviparity recently and at low taxonomic levels (i.e. Lacerta vivipara). The predictive test is valid, because the absence of such forms follows directly from the definition of saltation. The test also conforms to previous tests of saltational change (Simpson, 1944). Purposefully I did not rule out other explanations for the absence of intermediate forms in such taxa (and elsewhere considered the existence of such forms to refute saltational change as a global explanation). However, in attempting to restate my argument, Qualls et al. (1997, p. 129) mistakenly attributed to my paper the converse of its test of saltationism. They ascribed to my paper the idea that if viviparity arose by a means other than saltation, then intermediate forms should exist in particular lineages (e.g. Lacerta vivipara). Because the predictions and conclusions of this proposition actually are a converse of my own, their refutation of the claim that intermediate forms never existed (not made in my paper) is not necessary. These authors validly point out that intermediate forms could have disappeared through either anagenesis or extinction. However, in proposing that viviparity theoretically could evolve via anagenesis under saltation, their paper does not consider the improbability of such an event. Through anagenesis, an entire lineage undergoes transformation without branching; whatever the frequency of its occurrence, such change is at least possible under conditions of phyletic gradualism. In contrast, a saltational, anagenic origin would require the simultaneous development of viviparity in all individuals contributing to the next generation; the oviparous morphotype

186

. . 

would have to go extinct at the same time, or the result would be cladogenesis. Given the vanishing odds of such an occurrence, little reason exists to consider saltational change under conditions other than cladogenesis. Nevertheless, regardless of whether viviparity arises via cladogenesis or anagenesis, the question remains as to why intermediate phenotypes are so rare among living forms, even those showing evidence of reproductive evolution that is very recent or ongoing. In short, we must explain why intermediate phenotypes undergo differential extinction or transformation. Under a punctuational interpretation, the intermediate pattern tends to be selected against in favor of alternatives that are more stable evolutionarily. Traditional gradualistic accounts did not attempt to explain the absence of the intermediate pattern, under the unquestioned assumption that such a pattern is widespread. Construction and Evaluation of Evolutionary Scenarios The traditional scenario for the evolution of viviparity and placentation in my paper, as Qualls et al. surmised, is constructed from several sources—in fact, most aspects have been widely supported in the literature for decades. For example, as documented below, the view that viviparity and placentation do not evolve simultaneously has predominated for over a century. Contrary to their suggestion, this scenario accurately reflects traditional views about the evolution of viviparity and placentation. A testable scenario need not be explicitly formulated in its entirety by several authors to accurately represent traditional views, if individual aspects have received broad support (Blackburn, 1993b). The very concept of testable scenarios is, after all, relatively new. By implying that every aspect needs to be tested in many lineages, Qualls et al. place demands on the evaluation of evolutionary scenarios that are neither realistic nor necessary. In particular, they argue that because (substantial) placentotrophy has evolved in only three squamate clades, a robust test of the traditional squamate scenario can never be made: ‘‘by linking modes of parity to modes of embryonic nutrition . . . Blackburn has made his gradualistic model for the evolution of viviparity and placentation, as a whole, essentially untestable’’ (p. 133). Therefore, they assert, the evolution of viviparity and the evolution of placentotrophy ought be treated ‘‘as independent or parallel rather than sequential events.’’ This issue seems to reflect different perspectives on the utility of scenarios and how they are to be tested.

I view evolutionary scenarios as heuristic devices that shape and codify our ideas, and direct our investigations. Because they are not experiments to be run lock-step from start to finish, in most contexts, they need not be evaluated ‘‘as a whole’’. In fact, a major strength of evolutionary scenarios with multiple steps is that aspects can be evaluated independently; if any is found to be invalid, then the scenario is falsified and can be modified accordingly. Thus, given that 97% of the viviparous squamate clades are (as far as we know) relatively lecithotrophic (Blackburn, 1992), we can and routinely do (e.g. Shine, 1985) evaluate scenarios for the evolution of viviparity without reference to the specialization of extreme placentotrophy. Yet, whether incipient placentotrophy originates during or long after viviparity, and whether it is adaptive or inconsequential, may have important implications for the evolution of viviparity itself (Stewart & Castillo, 1984; Stewart, 1989). It would be easy to formally split the reproductive scenario in my paper into two, one based on reproductive mode and the other on embryonic nutritional mode. However, such a course gains us little, and obscures ways in which reproductive modes and embryonic nutritional modes are intertwined, functionally and evolutionarily. Viviparity and placentotrophy either evolve sequentially or simultaneously, and their evolution either is or is not causally linked. Determining which is the case demands use of scenarios that incorporate both parameters. Given the scarcity of substantial placentotrophy among squamates (three clades), can we really have little hope of reconstructing its evolution with confidence? It is of course not a weakness of the gradualistic scenario that its final step has occurred only a few times; to consider otherwise conflates epistemological with ontological issues. By coincidence, three also approximates the number of squamate species shown to exhibit intraspecific variation in parity modes, yet these species are providing valuable evidence for the transition to viviparity. In comparison, most of the key adaptations of vertebrates have evolved but once or a few times (e.g. mammary glands, endothermy, jaws, flight, paired fins); yet the literature abounds with scenarios for their evolution and implicit empirical tests of their validity. Given that matrotrophy (which subsumes placentotrophy) has evolved convergently on at least 25 separate occasions among vertebrates (Blackburn, 1992; unpublished data), if general principles govern its evolution, understanding the squamate situation ought to be within our grasp.

   Evolution of Placentation The idea that viviparity evolves separately from placentation has been supported by a long succession of researchers (e.g. Giacomini, 1906; Giersberg, 1923; Jacobi, 1936; Bauchot, 1965; Panigel, 1951; Smith et al., 1972; Angelini & Ghiara, 1991), including, by my count, the majority of individuals who have conducted research on squamate placentation. In the context of this historical tradition, I consider the suggestion that squamate placentas can evolve in association with viviparity (Packard et al., 1977; Guillette, 1982) as a novel insight. The subsequent proposal that incipient placentotrophy is an unselected correlate of viviparity (Blackburn, 1985a, 1992; also see Stewart & Castillo, 1984) was also new. Far from being a ‘‘strawman set up for rejection’’ (Qualls et al., 1997, p. 133), the idea that viviparity evolves separately from placentation reflects a tradition that has dominated the field of comparative reproduction throughout the 20th century. Qualls et al. quoted Weekes (1935, p. 641) as evidence that biologists have long viewed viviparity as evolving simultaneously with placentation. However, the cited quotation says nothing about placentation or the sequence of its evolution vis-a-vis viviparity. In fact, a careful reading of Weekes’ work reveals that she considered squamates not to be truly viviparous until after the shell membrane is entirely lost evolutionarily (e.g. Weekes, 1929, 1930); the rationale for her view was that even a thin shell membrane would limit physiological exchange across a ‘‘placenta’’. By such criteria, most placental squamates [which typically retain a remnant of the shell membrane (Blackburn, 1993a)] are not truly viviparous. Blackburn (1995) cited evidence that placentation and incipient placentotrophy are pervasive among viviparous species and originate during, rather than after, the transition to viviparity. The inference was based on long-standing criteria for the recognition of placentation (Mossman, 1937), that a ‘‘placenta’’ is ‘‘any intimate apposition of the fetal organs to the maternal (or paternal) tissues for physiological exchange.’’ In attempting to extend the concept of placentation to oviparous forms, Qualls et al. (1997, p. 133) substituted the word ‘‘approximation’’ for ‘‘intimate apposition’’ in this definition. Thus, their assertion that all oviparous squamates have placentas appears to follow more directly from Mossman’s definition than is actually the case. On the other hand, as pointed out by these authors, oviparous forms with prolonged egg retention and thin eggshells arguably may verge on a rudimentary form of placentation. As information accumulates, a

187

distinction between placental and non-placental squamates may become difficult to maintain, and I am glad that Qualls et al. have pointed out the implications of recent findings. However, I would caution against scrapping a concept that serves vertebrate zoology so well on the basis of a few lizard species on the verge of viviparity. In any case, the issue does not invalidate my analysis, because my scenario focused on a distinctly different question— whether viviparity evolves in the absence or presence of a placenta, as commonly defined. For a rudimentary form of placentation to evolve just prior to viviparity would require but a minor modification in the punctuational scenario. Implications of New Evidence The view that viviparity and typical oviparity are evolutionarily stable strategies is based on a bimodal distribution of species between these two reproductive patterns, and the paucity of oviparous species that lay eggs in very advanced stages of development. The vast majority of oviparous squamates that have been studied lay eggs at the limb bud stage (Dufaure & Hubert [1961] [‘‘D&H’’] stage 25–33) or give birth to fully-developed young (stage 40). Operationally, the latter category includes species (e.g. Saiphos equalis, Sphenomorphus fragilis) in which the young emerge from the surrounding membranes within a few or several hours of deposition (Shine, 1985; Smith & Shine, 1997); a short delay in emergence of the offspring is well known among viviparous forms. The bimodal distribution is emphasized by the fact D&H stage 27 is equivalent to a 55–60 hour chick embryo (incubation of chicken eggs takes 21 days). My analysis cited one species (Sceloporus scalaris) in which females lay eggs at advanced developmental stages (33–38), and a few others with short incubation times. Qualls et al. (1997) have added a species to this category, based on a paper post-dating acceptance of my own. In the lizard Liolaemus scapularis, females lay eggs at D&H stage 36 (Ramı´ rez Pinilla, 1994), and unpublished work suggests that this pattern occurs among a few other members of the genus (Ramı´ rez Pinilla, 1991). Other recent study has corroborated the figures on S. scalaris and has shown that the developmental stage of the eggs at oviposition differs between populations (Mathies & Andrews, 1995). Interestingly, this species exhibits facultative egg retention (Mathies & Andrews, 1996), a pattern predicted by the punctuational model (Blackburn, 1995). Information is also available on some nominal species with both oviparous and viviparous populations. One such species is Lacerta vivipara, in which

188

. . 

the oviparous females lay eggs at a development stage (Bran˜a et al., 1991) typical of other oviparous squamates (Blackburn, 1995). Because phenotypic intermediates are lacking, this species offers striking support for the punctuational model. In another such species, the Australian skink, Lerista bougainvilli, oviparous forms commonly lay eggs at an average stage of 32.6 (range 31–34) (Qualls et al., 1995; Qualls, 1996); the mean value falls within the range of typical oviparous forms. Interestingly, in small populations with extended egg-retention, females lay eggs with reduced shells and advanced embryos (stages 36–39). However, it is not clear whether the latter populations are evolutionarily intermediate, or have resulted from hybridization between typical oviparous and viviparous forms (Qualls et al., 1995). Hybridization has been produced in captivity between oviparous and viviparous Lacerta vivipara; mating of F1 progeny from such hybrids yield oviparous offspring with thin shells and advanced embryos (stages 35–36) (Heulin et al., 1992, 1993; Arrayago et al., 1996). However, such hybrids have not been found in nature, and their viability is unknown. A third nominal species with reproductive bimodality is the Australian scincid Saiphos equalis. Recent work has shown that the oviparous females deposit eggs at stage 38–39 which hatch in less than a week of deposition (Smith & Shine, 1997). Reproductive bimodality also has been suggested for the lizard Liolaemus alticolor; Ramı´ rez Pinilla & Laurent (1996) indicated both egg-laying and live-bearing modes occur in Argentina. The oviparous females lay eggs with thick eggshells, at a stage equivalent to D&H stage 32–33, a stage lying within the range of typical oviparous squamates (Blackburn, 1995). However, recent taxonomic work indicates that the oviparous and viviparous forms are actually separate species (Espin˜oza et al., 1997; pers. comm.), and a reconstruction of reproductive evolution awaits further study. Qualls et al. (1997) misattributed to my paper a claim that ‘‘no single squamate clade exhibits primitive, intermediate, and advanced stages in the evolution of viviparity’’, and challenge it on the basis of some of the above work. My paper actually stated that ‘‘available evidence has not yet permitted construction of a single, complete phenocline of parity modes and embryonic nutritional patterns out of representatives of a single clade’’, and recent studies have not altered validity of the statement. We still lack, in all the genera cited above, evidence for the existence of two of the four reproductive patterns I cited in this context (oviposition very early in

development; viviparity with substantial placentotrophy), and information on a third of the four patterns (viviparous lecithotrophy and simple placentation) is incomplete or absent. In fact, definitive evidence for oviposition at a very early developmental stage continues to be confined to a single squamate species (Chameleo lateralis) in an unrelated lineage (Blackburn, 1995). Thus, recently published work has revealed an oviparous species in which females deposit developmentally advanced eggs (Liolaemus scapularis), and has demonstrated extended egg-retention in some isolated populations of two unrelated skinks (Lerista bougainvilli and Saiphos equalis). These additions do not substantially alter the bimodal distribution of squamate reproductive stages. Significantly, three nominal ‘‘species’’ with reproductively bimodality (L. bougainvilli, Lacerta vivipara, and Liolaemus alticolor) contain oviparous members that lay eggs with normal eggshells, at a typical (limb bud) stage of embryonic development; this pattern is predictable from the punctuational model. In only one of these three taxa (L. bougainvilli ) do any females deposit developmentally—advanced eggs, and these females possibly are hybrids rather than evolutionary intermediates. Also significant is the fact that a population of squamates with both oviparous and viviparous members has never been discovered. By some criteria, therefore, the oviparous and viviparous ‘‘conspecifics’’ actually represent separate species. An important point to be noted is that the existence of (relatively few) squamates that are phenotypically intermediate between viviparity and typical oviparity is not only consistent with the punctuational model, but is predicted by that model. What is so striking about these forms is their scarcity (very few lizards and no snakes or amphisbaenians), and the fact that each shows strong evidence of very recent (and perhaps ongoing) reproductive evolution. If the ‘‘intermediate’’ pattern is evolutionarily stable, why is it so seldom characteristic of entire species, and why does it not (like viviparity and typical oviparity) characterize entire genera, subfamilies, or families? The situation is especially remarkable, given that squamates date to the Mesozoic, and given that viviparity has evolved on more than 100 separate occasions (Blackburn, 1985b, 1992; Shine, 1985). This extreme scarcity of intermediate forms is to be expected if the intermediate phenotype is less stable than the evolutionary alternatives, and if evolution of viviparity is episodic. The scarcity is also contrary to widespread assumptions in the literature, assumptions fundamental to traditional interpretations based on phyletic gradualism. I concur with Qualls and his

   colleagues that the study of reproductively intermediate forms is likely to provide important information on the evolution of squamate viviparity. However, a focus on these extremely rare forms ought not obscure the likelihood that stasis, punctuated with episodic change, has been a major pattern in the history of squamate reproductive modes. Existence of this historical pattern as a dominant one in squamate history is an important point on which Qualls et al. (1997) and I agree. Not to consider applicability of punctuational models to this pattern arbitrarily marginalizes studies of squamate reproductive evolution, and erects an artificial barrier between research on extinct and extant species. In contrast, labelling this pattern as punctuational is fully consistent with contemporary concepts, and allows us to interpret squamate evolution in light of an important theoretical model that has clarified patterns of change in a wide diversity of other organisms.

I wish to thank Carl Qualls, Robin Andrews, and Tom Mathies for contributing to the dialogue on these issues, Robin Andrews for sending me a preprint of their 1996 paper, and the J. theor. Biol. for the opportunity to reconsider the questions raised. I also thank James Stewart, Robin Andrews, Lora Miller, and anonymous reviewers for comments on the manuscript, and Chris Sidor, William Mace, and Miller Brown for helpful discussions.

REFERENCES A, F. & G, G. (1991). Viviparity in squamates. In: Symposium on the Evolution of Terrestrial Vertebrates. (Ghiara, G., ed), pp. 305–334. Modena: Mucchi. A, M.-J., B, A. & H, B. (1996). Hybridization experiment between oviparous and viviparous strains of Lacerta vivipara: a new insight into the evolution of viviparity in reptiles. Herpetologica 52, 333–342. B, R. (1965). La placentation chez les reptiles. Ann. Biol. 4, 547–575. B, D. G. (1982). Evolutionary origins of viviparity in the Reptilia. I—Sauria. Amphib. Reptilia 3, 185–205. B, D. G. (1985a). The evolution of viviparity and matrotrophy in vertebrates, with special reference to reptiles. Ph.D. Thesis, Cornell Univ., Ithaca, NY. B, D. G. (1985b). Evolutionary origins of viviparity in the Reptilia. II—Serpentes, Amphisbaenia, and Ichthyosauria. Amphib. Reptilia 5, 259–291. B, D. G. (1992). Convergent evolution of viviparity, matrotrophy, and specializations for fetal nutrition in reptiles and other vertebrates. Amer. Zool. 32, 313–321. B, D. G. (1993a). Chorioallantoic placentation in squamate reptiles: structure, function, development, and evolution. J. Exper. Zool. 266, 414–430. B, D. G. (1993b). Lactation: historical patterns and potential for manipulation. J. Dairy Sci. 76, 3195–3212. B, D. G. (1995). Saltationist and punctuated equilibrium models for the evolution of viviparity and placentation. J. theor. Biol. 174, 199–216. B˜, F., B, A. & A, M. J. (1991). Egg retention in lacertid lizards: relationships with reproductive ecology and the evolution of viviparity. Herpetologica 47, 218–226.

189

C, R. L. (1988). Vertebrate Paleontology and Evolution. New York: W. H. Freeman. D, J. P. & H, J. (1961). Table du de´veloppement du le´zard vivipare: Lacerta (Zootoca) vivipara Jacquin. Arch. Anat. Microscop. Morphol. Exp. 50, 309–328. E, N. (1993). Reinventing Darwin. New York: Wiley. E, N. & G, S. J. (1972). Punctuated equilibria: an alternative to phyletic gradualism. In: Models in Paleobiology (Schopf, T. J. M., ed.), pp. 82–115. San Francisco: Freeman, Cooper & Co. E˜, R. E., L, F., C, F. B. & T, C. R. (1997). Viviparity in the Liolaemus alticolor group (Iguania: Tropiduridae): cryptic species or evolution in action? Abstract, Proceedings of the Third World Congress of Herpetology, Prague, Czech Republic, p. 63. F, H. S. (1970). Reproductive cycles in lizards and snakes. Univ. Kansas Mus. Nat. Hist. Misc. Publ. 52, 1–247. G, E. (1891). Materiali per la storia dello sviluppo del Seps chalcides. (Cuv.) Bonap. Monitore Zool. Ital. 2, 179–182, 198–211. G, E. (1906). Sulla maniera di gestazione e sugli annessi embrionali del Gongylus ocellatus. Forsk. Mem. Acad. Sci. Inst. Bologna 3, 401–445. G, H. (1923). Untersuchungen u¨ber Physiologie und Histologie des Eileiters der Reptilien und Vo¨gel; nebst einem Beitrag zue Fasergenese. Zeitschr. wissensch. Zool. 70, 1–97. G, S. J. & E, N. (1977). Punctuated equilibria: the tempo and mode of evolution reconsidered. Paleobiology 3, 115–151. G, S. J. & E, N. (1986). Punctuated equilibrium at the third stage. Paleobiology 35, 143–148. G, L.J., J (1982). The evolution of viviparity and placentation in the high-altitude, Mexican lizard Sceloporus aeneus. Herpetologica 38, 94–103. H, B., A, M. J., B, A. & B˜, F. (1992). Caracte´ristiques de la coquille des oeufs chez la souche hybride (ovipare x vivipare) du le´zard Lacerta vivipara. Can. J. Zool. 70, 2242–2246. H, B., G, C., B, A. & A, M.J. (1993). Interpretation biogeographique de la bimodalite de reproduction de lezard—Lacerta vivipara Jacquin—(Sauria, Lacertidae): un modele pour l’etude de l’evolution de la viviparite. Biogeographica 69, 3–13. H, A. (1992). Twenty years later: punctuated equilibrium in retrospect. In: The Dynamics of Evolution (Somit, A. & Peterson, S. A., eds), pp. 121–138. Ithaca, New York: Cornell University Press. J, L. (1936). Ovoviviparie bei einheimischen Eidechsen. Z. Wiss. Zool. 148, 401–464. M, T. & A, R. M. (1995). Thermal and reproductive biology of high and low elevation populations of the lizard Sceloporus scalaris: implications for the evolution of viviparity. Oecologia 104, 101–111. M, T. & A, R. M. (1996). Extended egg retention and its influence on embryonic development and egg water balance: implications for the evolution of viviparity. Physiol. Zool. 69, 1021–1035. M S, J. (1988). Did Darwin Get it Right? Essays on Games, Sex, and Evolution. New York: Chapman & Hall. M, E. (1992). Speciational evolution or punctuated equilibria. In: The Dynamics of Evolution (Somit, A. & Peterson, S. A., eds), pp. 21–53. Ithaca, New York: Cornell University Press. M, H. W. (1937). Comparative morphogenesis of the fetal membranes and accessory uterine structures. Carnegie Inst. Contrib. Embryol. 26, 129–246. P, G. C., T, C. R. & R, J. J. (1977). The physiological ecology of reptilian eggs and embryos, and the evolution of viviparity within the Class Reptilia. Biol. Rev. 52, 71–105. P, M. (1951). Rapports anatomo-histologiques etablis au cours de la gestation entre l’oeuf et l’oviducte maternal chez le lezard ovovivipare, Zootoca vivipara W. (Lacerta vivipara J.) Bull. Soc. Zool. Fr. 76, 163–170.

190

. . 

Q, C. P. (1996). Influence of the evolution of viviparity on eggshell morphology in the lizard, Lerista bougainvilli. J. Morphol. 228, 119–125. Q, C. P., S, R., D, S. & H, M. (1995). The evolution of viviparity in the Australian scincid lizard, Lerista bougainvilli. J. Zool. (Lond.) 237, 13–26. Q, C. P., A, R. M. & M, T. (1997). The evolution of viviparity and placentation revisted [sic]. J. theor. Biol. 185, 129–135. Rı´  P, M. P. (1991). Estudio histologico de los tractos reproductivos y actividad ciclica anual reproductiva de machos y hembras de dos especes de ge´nero Liolaemus (Reptilia: Sauria: Iguanidae). Ph.D. Diss, Universidad Nacional de Tucuma´n, Argentina. 208 pp. Rı´  P, M. P. (1994). Reproductive and fat body cycles of the oviparous lizard Liolaemus scapularis. J. Herpetol. 28, 521–524. Rı´  P, M. P. & L, R. (1996). Apparent reproductive bimodality in Liolaemus alticolor alticolor (Reptilia: Sauria). Bull. Maryland Herpetol. Soc. 32, 1–13. S, R. (1985). The evolution of viviparity in reptiles: An ecological analysis. In: Biology of the Reptilia (Gans, C. & Billet, F., eds), Vol. 15, pp. 605–694. New York: Wiley. S, R. (1995). A new hypothesis for the evolution of viviparity in reptiles. Amer. Nat. 145, 809–823. S, G. G. (1944). Tempo and Mode in Evolution. New York: Columbia University Press.

S, H. M., S, G., F, J. D. & J, R. E. (1972). A unique reproductive cycle in Anolis and its relatives. Bull. Phila. Herpetol. Soc. 20, 28–30. S, S. A. & S, R. (1997). The evolution of viviparity: intraspecific variation in reproductive mode within the scincid lizard Saiphos equalis. Aust. J. Zool. (in press). S, A. & P, S. A. (1992). (eds) The Dynamics of Evolution. Ithaca, New York: Cornell University Press. S, S. M. (1979). Macroevolution: Pattern and Process. San Francisco: W. H. Freeman. S, S. M. (1981). The New Evolutionary Timetable. New York: Basic Books. S, J. R. (1989). Facultative placentotrophy and the evolution of squamate placentation: quality of eggs and neonates in Virginia striatula. Amer. Nat. 133, 111–137. S, J. R. & C, R. E. (1984). Nutritional provision of the yolk of two species of viviparous reptiles. Physiol. Zool. 57, 377–383. W, H. C. (1929). On placentation in reptiles—I. Proc. Linn. Soc. N.S.W. 54, 34–60. W, H. C. (1930). On placentation in reptiles—II. Proc. Linn. Soc. N.S.W. 55, 550–576. W, C. H. (1935). A review of placentation among reptiles, with particular regard to the function and evolution of the placenta. Proc. Zool. Soc. Lond. 2, 625–645. W, J. P. & C, I. P. (1992). A retrospect to the symposium on evolution of viviparity in vertebrates. Amer. Zool. 32, 251–255.