Paedomorphosis in Newts & Salamanders: Mechanisms & Ecological Impact, Lab Reports of Animal Biology

Download Paedomorphosis in Newts & Salamanders: Mechanisms & Ecological Impact and more Lab Reports Animal Biology in PDF only on Docsity! Evolutionary ecology of facultative paedomorphosis in newts and salamanders Mathieu Denoël1,4, Pierre Joly2 and Howard H. Whiteman3,4 1 Laboratory of Fish and Amphibian Ethology, Behavioural Biology Unit, Department of Environmental Sciences, University of Liège, 22 Quai Van Beneden, 4020 Liège, Belgium (E-mail : Mathieu.Denoel@ulg.ac.be) 2 UMR CNRS 5023 Ecology of Fluvial Hydrosystems, Claude Bernard University of Lyon, 69622 Villeurbanne, France 3 Department of Biological Sciences, Murray State University, Murray, KY 42071, USA 4 Rocky Mountain Biological Laboratory, P.O. Box 519, Crested Butte, CO 81224, USA (Received 14 December 2004; revised 31 May 2005; accepted 8 June 2005) ABSTRACT Facultative paedomorphosis is an environmentally induced polymorphism that results in the coexistence of mature, gilled, and fully aquatic paedomorphic adults and transformed, terrestrial, metamorphic adults in the same population. This polymorphism has been of interest to scientists for decades because it occurs in a large number of caudate amphibian taxa as well as in a large diversity of habitats. Numerous experimental and observational studies have been conducted to explain the proximate and ultimate factors affecting these heterochronic variants in natural populations. The production of each alternative phenotype is based on a genotyperenvironment interaction and research suggests that differences in the environment can produce paedomorphs through several ontogenetic pathways. No single advantage accounts for the maintenance of this polymorphism. Rather, the interplay of different costs and benefits explains the success of the polyphenism across variable environments. Facultative paedomorphosis allows individuals to cope with habitat variation, to take advantage of environmental heterogeneity in the presence of open niches, and to increase their fitness. This process is expected to constitute a first step towards speciation events, and is also an example of biodiversity at the intraspecific level. The facultative paedomorphosis system is thus ripe for future studies encompassing ecology, evolution, behaviour, endocrinology, physiology, and conservation biology. Few other systems have been broad enough to provide varied research opportunities on topics as diverse as phenotypic plasticity, speciation, mating behaviour, and hormonal regulation of morphology. Further research on facultative paedomorphosis will provide needed insight into these and other important questions facing biologists. Key words : adaptation, amphibians, evolution, facultative paedomorphosis, heterochrony, polyphenism. CONTENTS I. Introduction ................................................................................................................................................. 664 II. Proximate mechanisms of facultative paedomorphosis .......................................................................... 665 III. Ultimate mechanisms of facultative paedomorphosis ............................................................................ 665 (1) Resource partitioning ........................................................................................................................... 665 (2) Life-history traits ................................................................................................................................... 666 (3) Sex-specific fitness ................................................................................................................................. 667 IV. Paedomorphosis : a model for speciation? ............................................................................................... 667 V. Models of paedomorphosis ........................................................................................................................ 667 VI. Conclusions and perspectives .................................................................................................................... 668 VII. Acknowledgements ...................................................................................................................................... 669 VIII. References .................................................................................................................................................... 669 Biol. Rev. (2005), 80, pp. 663–671. f 2005 Cambridge Philosophical Society 663 doi:10.1017/S1464793105006858 Printed in the United Kingdom I. INTRODUCTION Facultative paedomorphosis (i.e. the potential for retention of juvenile traits in mature individuals) has roused the curi- osity of biologists for more than a century. This phenom- enon was first suspected by Cuvier (1828) who was reluctant to place the axolotl in the permanently branchiate sala- manders, but its existence was first demonstrated by Dumerill (1866) at the Museum of Paris when he bred branchiate individuals of Ambystoma mexicanum. The meta- morphosis of these axolotls into a form that was known as another salamander species provided the first evidence of developmental heterochrony between the sexual apparatus and the somatic tissues. For decades, such heterochronies in the development of caudate amphibians remained a zoological curiosity until a satisfactory theoretical frame- work was established to explore some explanatory hypoth- eses. The first studies were mechanistic and did not address evolutionary questions (Snyder, 1956; Sprules, 1974a ; Gabrion, 1976). They considered facultative paedomorphosis to be a metamorphic failure caused by the detrimental effects of varied environmental factors. The environmental context of paedomorphosis proved to be so diverse that it became evident that the determinism of this trait was multifactorial (Semlitsch, 1987a ; Breuil, 1992; Whiteman, 1994; Voss, 1995; Denoël, 2003a). Indeed, paedomorphosis occurs in such contrasting habitats as per- manent mountain lakes surrounded by arid grounds and lowland temporary ponds surrounded by wet forests. The rise of modern evolutionary biology has opened new per- spectives on this polymorphism, particularly within the framework of the evolution of phenotypic plasticity. Phenotypic plasticity can result from heterochronic pro- cesses, consisting of a shift in the timing or rate of develop- ment of a functional set of biological traits relative to the development of the whole organism (Schlichting & Pigliucci, 1998). In animals with complex life cycles, heterochronic processes are expected to have profound consequences be- cause of the interplay between larval and postmetamorphic traits that can lead to substantial differences between alternative morphotypes. Subsequently, such polyphenism can have a strong impact on the evolution of ecological communities (Gould, 1977; McKinney & McNamara, 1991). Research suggests that genotypes able to produce alternative phenotypes would occur in heterogeneous environments in which each alternative can experience a higher fitness in the environment where the risks of suboptimal matching between resource availability and morpho-functional abilities are spread through phenotypic differentiation (Kaplan & Cooper, 1984; Gross, 1991; Skulason & Smith, 1995; Schlichting & Pigliucci, 1998). However, the evolution of polyphenism implies that the fitness gain drawn from the ecological specialisation of different phenotypes exceeds the cost of plasticity. Thus, polyphenisms are expected in environments where the scale of variability matches the adaptive potential of the develop- ing phenotype (Levins, 1968). With regard to life history, polyphenisms are particularly expected in species with complex life cycles because the transition between habitats via metamorphosis is risky due to the lack of reliable infor- mation about the quality of the subsequent habitat (Wilbur & Collins, 1973; Werner, 1986). The ontogenetic pathway of species with complex life cycles is indeed characterized by the successive use of contrasting habitats (Nielsen, 1998). As a consequence, these species experience more environmen- tal variation than single-habitat species. Selection for phenotypic plasticity is thus expected to be particularly strong in species with complex life cycles. The duration of each phase of a cycle should depend on the specific cost :benefit ratio (in terms of individual fitness) in each habitat. Newts and salamanders exhibit a wide range of life cycles. At the extremes, some species can totally skip one stage and spend their whole life in only one habitat. This is the case in permanently gilled species (obligate paedomorphosis) that occupy aquatic habitats and of terrestrial species exhibiting direct development (Istock, 1967; Wilbur & Collins, 1973; Wilbur, 1996). Between these extremes, many salamanders exhibit a two-stage ontogenetic pathway with a larval aquatic stage and a terrestrial post-metamorphic stage (obligate metamorphosis) (Wilbur, 1980), while others exhibit facultative paedomorphosis (Duellman & Trueb, 1994). Each of these life cycles can be seen as a result of the interplay between costs and benefits experienced in both aquatic and terrestrial environments. Detrimental aquatic conditions (drying, crowding, low food availability, in- breeding, presence of predators) and unfavourable terres- trial situations (low humidity, absence of shelter, low food availability, high predation) may have driven species towards one of the basic life histories (Wilbur & Collins, 1973; Whiteman, 1994). Paedomorphosis has been observed in fifty-seven species of newts and salamanders, distributed in nine of the ten recognized families. Four of them are only composed of obligatory paedomorphic species (Amphiumidae, Sirenidae, Proteidae, Cryptobranchidae). Such obligate paedomor- phosis also occurs in a few species of Plethodontidae. Paedomorphosis is facultative in the Salamandridae, Ambystomatidae, Dicamptodontidae, Hynobiidae and some Plethodontidae. More than 10% of salamander spe- cies exhibit paedomorphic ontogenetic pathways (Semlitsch & Wilbur, 1989; Duellman & Trueb, 1994; Whiteman, 1994; Denoël, 2003a). Interest in facultative paedomorphosis grew in the 1980s (e.g. Semlitsch & Gibbons, 1985; Harris, 1987; Semlitsch, 1987a ; Semlitsch & Wilbur, 1989) after Wilbur and Collins (1973) published their growth model of metamorphosis timing. Most advances until the early 1990s were sum- marised by Whiteman (1994) who proposed alternative models to account for the variation of ontogenetic pathways among facultative paedomorphic species in natural and experimental conditions. Numerous authors have sub- sequently used facultative paedomorphosis as a model system in which to investigate the genetic bases and benefits of polyphenism in natural populations (Voss, 1995; Kalezic et al., 1996; Whiteman, Wissinger & Brown, 1996; Voss & Shaffer, 1997; Whiteman, 1997; Ryan & Semlitsch, 1998; Denoël & Joly, 2000, 2001; Denoël, Poncin & Ruwet, 2001b ; Boorse & Denver, 2002; Denoël et al., 2002). 664 M. Denoël, P. Joly and H. H. Whiteman reduces fecundity at first reproduction (Semlitsch, 1985). Nonetheless, because salamanders produce dozens to hun- dreds of eggs each year (Semlitsch, 1985; Kalezic et al., 1996), even a small decrease in the age at maturity can have fitness consequences that are similar to a large increase in fecundity (Roff, 1992). By maturing just one year after hatching and more than one year before metamorphic adults, progenetic paedomorphs also benefit from an in- creased probability of breeding that results in rapid increase in fitness (Denoël & Joly, 2000). (3 ) Sex-specific fitness Sex-specific payoffs may influence the maintenance of polymorphism. By studying the time interval between breeding events in tiger salamanders, Whiteman (1997) showed that paedomorphic males bred more often than metamorphic males, while the reverse was true in females. This interaction effect promotes the maintenance of facul- tative paedomorphosis via sex-specific payoffs. The fact that paedomorphic females skip reproduction more often than metamorphic ones may be due to the low energy intake in permanent ponds and the high cost of egg production. The higher reproductive frequency in paedomorphic males is probably affected by their continued presence in breeding ponds, in contrast to metamorphic males, which depend on favourable climatic conditions (rainfall, temperature) during migration events from terrestrial overwintering sites. As predicted, the sex-ratio consistently differed between the two phenotypes with the males dominating in paedomorphs, and the females in metamorphs (Whiteman, 1997). The reverse of these sex ratios has been found in other popu- lations, suggesting that the costs and benefits to each morph and sex vary by species and environmental conditions (Breuil, 1992; Whiteman, 1997; Ryan & Hopkins, 2000; Denoël, 2003a). IV. PAEDOMORPHOSIS: A MODEL FOR SPECIATION? Polymorphisms are suspected to be key steps in species evolution, particularly via sympatric differentiation where they promote reproductive isolation among morphs (Bush, 1994; Skulason & Smith, 1995; West-Eberhard, 2003). In this way, assortative mating within alternative morphs could make offspring of paedomorphs less likely to metamorphose (Semlitsch & Wilbur, 1989; Scott, 1993). Given differences in secondary sexual traits between morphs, sexual isolation might be expected. However, cross-breeding experiments in salamandrids showed that the two morphs interbreed successfully at an identical rate (Denoël et al., 2001b). Additional observations and experiments suggest that this also occurs frequently in ambystomatids (Krenz & Sever, 1995; Whiteman, Gutrich & Moorman, 1999). In several species, the two morphs display similar behavioural patterns at similar frequencies (Whiteman et al., 1999; Denoël et al., 2001b ; Denoël, 2002) or exhibit only slight differences in the content and timing of courtship (Krenz & Verrell, 2002). They also use the same alternative tactics to attract females (Denoël, 2002, 2003b). This mating system differs from other polymorphisms in which morphs display alternative tactics or strategies with equal or different fitness payoffs (Gross, 1991, 1996; Fu, Neff & Gross, 2001). Sexual com- patibility promotes significant gene flow between morphs, and impedes differentiation as long as the environmental pressures do not change (Denoël et al., 2001b ; Krenz & Verrell, 2002). Thus even though facultative paedomor- phosis is adaptive because of the flexibility it affords off- spring, breeding studies suggest that this polymorphism is unlikely to lead to future speciation events. Yet, in some salamander populations sexual isolation might occur because paedomorphs breed weeks or months earlier than metamorphs, thus leading to assortative mating (Whiteman & Semlitsch, 2005). This situation might be reinforced by the low mating activity of females after first insemination (Krenz & Sever, 1995). Spatial segregation of the two morphs in large ponds can also promote assortative mating among morphs (Whiteman & Semlitsch, 2005). In these cases, temporal and spatial differences can lead to partial reproductive isolation, such that gene flow is reduced among certain morph-sex combinations (Whiteman & Semlitsch, 2005). Isolated permanent lakes or springs offer opportunities for divergence by allopatric speciation, particularly if the ter- restrial morph is counter-selected. The spatial distribution of Mexican lakes where paedomorphic species of ambystoma- tid salamanders live provides such an ecological framework because they are isolated from each other by considerable amounts of inhospitable terrestrial habitat. As a conse- quence, paedomorphic populations of the different lakes evolve independently (Shaffer, 1993). A similar situation occurs in Eurycea salamanders (Plethodontids) which inhabit isolated springs on the Edwards Plateau in Texas (Chippindale et al., 2000). Obligatory paedomorphosis in some amphibian families might have arisen in such situ- ations where life on land is counter-selected leading to a fully aquatic lifestyle. Even in less isolated systems, genetic differentiation of developmental mechanism was found in ambystomatid salamanders from South Carolina (Harris et al., 1990; Semlitsch et al., 1990). These results indicate that differentiation of developmental programs can occur quickly when gene flow among populations is low. V. MODELS OF PAEDOMORPHOSIS Both predictive and diversifying models of phenotypic plas- ticity have been proposed to account for the mechanisms leading to environmentally-induced phenotypes (Kaplan & Cooper, 1984; Menu & Debouzie, 1993). According to predictive plasticity models, phenotypes are induced fol- lowing the perception of environmental indices that reliably inform the individual of future environmental conditions (Menu & Debouzie, 1993). Paedomorphic responses to permanent water and low density support this model (Harris, 1987; Semlitsch, 1987a). By undergoing metamor- phosis following a decrease in water level and/or an increase Evolutionary ecology of facultative paedomorphosis in newts and salamanders 667 in density, both larvae and, in some cases, paedomorphic adults escape an aquatic environment that is becoming un- favourable. Such a transition generally occurs just before the pond dries, allowing individuals to optimize the benefit of their aquatic habitat (Semlitsch et al., 1988; Denoël, 2003 c). Food availability can also influence the ontogenetic path- way, but via a more complex mechanism. Whereas the fre- quency of metamorphosis increases in adult paedomorphs as aquatic food resources become scarce (thus supporting the predictive plasticity hypothesis) (Denoël & Poncin, 2001), food limitation can delay metamorphosis in larvae (Sprules, 1974a ; Voss, 1995; Ryan & Semlitsch, 2003). Whiteman (1994) modified the Wilbur–Collins meta- morphosis model (Wilbur & Collins, 1973) to create three hypotheses for the production and maintenance of facultat- ive paedomorphosis according to habitat quality (Fig. 3). The Paedomorph Advantage (PA) mechanism corresponds to the basic model of Wilbur and Collins (1973). It predicts that large, fast-growing animals (larger than P*, the mini- mum size for paedomorphosis in Fig. 3) in good growing habitats (where growth rate is high) become paedomorphic, while animals smaller than P* metamorphose to escape competition with larger paedomorphs. By contrast, the Best of a Bad Lot (BOBL) hypothesis predicts the reverse solution in poor habitats (low growth conditions) : the larger larvae (>M*, the minimum size for metamorphosis) metamor- phose, while the smallest ones (between R, the minimum size for sexual maturity, and M*) keep a larval somatic state and become reproductively mature. Finally, the Dimorphic Paedomorph (DP) hypothesis suggests that paedomorphosis results from both mechanisms according to the local con- ditions experienced by each individual. These predictions assume that growth rate is a key inte- grative parameter, and that body size is related to fitness. For example, the PA hypothesis suggests that fast-growing larvae in rich habitats experience a fitness advantage in this habitat. By contrast, the BOBL hypothesis predicts that larvae that cannot reach a sufficient size for metamorphosis take advantage of early maturity as a stopgap solution. Evidence for both hypotheses in the same population would support the DP. Assuming a strong relationship between fitness par- ameters and body size (Semlitsch, 1985, 1987a), the model predicts that paedomorphs produced through a PA mech- anism would have higher fitness than metamorphs (e.g. in fecundity or mating success), while the reverse would be true when paedomorphs are produced through a BOBL mech- anism (although there might be a decrease in age at first reproduction). The DP suggests that mean fitness might not differ between morphs within a population, but the variance in fitness should be greater in paedomorphs than meta- morphs because paedomorphs are produced through two mechanisms that differ strongly in their fitness payoffs. Although a number of studies have supported both the PA and BOBL hypotheses (Whiteman, 1994; Whiteman et al., 1996; Denoël & Joly, 2000; Ryan & Semlisch, 2003; J. M. Doyle & H. H. Whiteman, unpublished data), by far the most support has been found for PA, perhaps due in part to the good growing conditions in which many studies have been conducted. It is clear from experimental and observational studies of larval growth patterns that both mechanisms can operate to produce paedomorphs in nature; what is less clear is how the fitness consequences of each morph correspond to the hypotheses, if they do so at all. Because current studies have only addressed fitness components and not lifetime reproductive success (LRS), our ability to evaluate this question may have to wait until more complete LRS analyses of facultatively paedomorphic species are performed. VI. CONCLUSIONS AND PERSPECTIVES (1) Facultative paedomorphosis occurs in habitats as diverse as deep oligotrophic permanent alpine lakes and small eutrophic temporary ponds, surrounded by arid areas as well as by wet forests (Healy, 1974; Sprules, 1974b ; Breuil, 1992; Whiteman, 1994; Denoël et al., 2001a). The diversity of these environments, and the animals within them, has made it challenging to create a satisfactory ex- planatory framework. By identifying the multiple pathways that can drive an individual toward paedomorphosis, recent studies now make it possible to draw a conceptual frame- work for explaining the evolutionary success of facultative paedomorphosis in a great diversity of environments (Semlitsch, 1987b ; Whiteman et al., 1996; Voss & Shaffer, 1997; Whiteman, 1997; Ryan & Semlitsch, 1998; Denoël & Joly, 2000, 2001; Denoël et al., 2002). (2) Although many studies have been conducted to explain the existence and maintenance of polymorphisms, and particularly facultative paedomorphosis, the mech- anism(s) of its expression remains poorly understood (Denver et al., 2002; Voss et al., 2003). (3) New methods, such as candidate gene analysis, that have recently been used to determine the alleles involved in the timing of metamorphosis will increase our knowledge in the area (Voss et al., 2003). (4) New research should also focus on the molecular transcription of signals that induce larvae to become R M* M L Body size N um be r of in di vi du al s P P P* Best of a Bad Lot Dimorphic Paedomorph Paedomorph Advantage Fig. 3. Three ecological models of paedomorphosis (modified from Whiteman, 1994) : the Best of a Bad Lot (dark shaded semi-circle), the Paedomorph Advantage (light shaded semi- circle) and the Dimorphic Paedomorph Hypothesis (open semi- circle). R : minimum size for sexual maturity, M* : minimum size for metamorphosis, P* : minimum size for paedomorphosis through the Paedomorph Advantage model, L : larval stage, P : paedomorphic stage, M : metamorphic stage. 668 M. Denoël, P. Joly and H. H. Whiteman paedomorphs and not metamorphs. Stress hormones are obvious key targets for the mediation between environ- mental cues and resulting developmental pathways (Boorse & Denver, 2002), yet little work has focused on the hormonal basis of paedomorphosis. (5) Recent work suggests that sex should also be incor- porated into models to explain the maintenance of poly- morphism (Whiteman, 1997). Further studies, in particular those dealing with hormones, should take sex effects into account as sex hormones act on the processes of metamor- phosis (Denver et al., 2002). (6) Finally, fluctuations in natural populations of paedo- morphic newts and salamanders need to be explained in regard to natural and anthropogenic causes, including how the polymorphism might affect the frequency, amplitude, and metapopulation effects of fluctuation (Denoël et al., 2005; Whiteman & Wissinger, 2005). Rapid environmental changes in systems composed of permanent and drying ponds are of particular interest (Healy, 1974; Kalezic & Dzukic, 1985; Breuil, 1992; Whiteman et al., 1996). (7) The facultative paedomorphosis system is thus ripe for future studies encompassing ecology, evolution, behaviour, endocrinology, physiology, and conservation biology. Few other systems have been broad enough to provide varied research opportunities on topics as diverse as phenotypic plasticity, speciation, mating behaviour, and hormonal regulation of morphology. Further research on facultative paedomorphosis is sure to provide needed insight into these and other important questions facing biologists. VII. ACKNOWLEDGEMENTS We are grateful to N. Gerlanc, B. Kobylarz, J. Boynton, J. Doyle, and C. Eden for their useful comments on the manuscript. M. Denoël is a Research Associate at the FNRS (Belgian National Fundation for Scientific Research) and benefited from a Fulbright grant and FNRS grants ‘credit aux chercheurs ’ 1.5.011.03 and 1.5.120.04, and Crédits pour brefs séjours. H. Whiteman was supported by NSF grant DEB-0109436, a Senior Research Fellowship from the Center for Field Biology, Austin Peay University, and by grants from the Center for Institutional Studies and Research at Murray State University. VIII. REFERENCES BOORSE, G. C. & DENVER, R. J. (2002). Acceleration of Ambystoma tigrinum metamorphosis by corticotropin-releasing hormone. Journal of Experimental Zoology 293, 94–98. BREUIL, M. (1992). La néoténie dans le genre Triturus : mythes et réalités. Bulletin de la Société Herpétologique de France 61, 11–44. BUSH, G. L. (1994). Sympatric speciation in animals : new wine in old bottles. Trends in Ecology and Evolution 9, 285–288. CHIPPINDALE, P. T., PRICE, A. H., WIENS, J. J. & HILLIS, D. M. (2000). Phylogenetic relationships and systematic revision of central texas hemidactyliine plethodontid salamanders. Herpetological monographs 14, 1–80. CUVIER, G. (1828). Le Règne animal distribué d’après son Organisation. Chez Déterville, Paris. DENOËL, M. (2002). Paedomorphosis in the Alpine newt (Triturus alpestris) : decoupling behavioural and morphological change. Behavioral Ecology and Sociobiology 52, 394–399. DENOËL, M. (2003a). Avantages sélectifs d’un phénotype hétéro- chronique. Eco-éthologie des populations pédomorphiques du Triton alpestre, Triturus alpestris (Amphibia, Caudata). Cahiers d’Ethologie 21, 1–327. DENOËL, M. (2003b). Effect of rival males on the courtship of paedomorphic and metamorphic Triturus alpestris. Copeia 2003, 618–623. DENOËL, M. (2003 c). How do paedomorphic newts cope with lake drying? Ecography 26, 405–410. DENOËL, M. (2004a). Feeding performance in heterochronic Alpine newts is consistent with trophic habits and maintenance of polymorphism. Ethology 110, 127–136. DENOËL, M. (2004b). Terrestrial versus aquatic foraging in juvenile Alpine newts (Triturus alpestris). Ecoscience 11, 404–409. DENOËL, M., DUGUET, R., DZUKIC, G., KALEZIC, M. & MAZZOTTI, S. (2001a). Biogeography and ecology of paedomorphosis in Triturus alpestris (Amphibia, Caudata). Journal of Biogeography 28, 1271–1280. DENOËL, M., DZUKIC, G. & KALEZIC, M. (2005). Effect of wide- spread fish introductions on paedomorphic newts in Europe. Conservation Biology 19, 162–170. DENOËL, M., HERVANT, F., SCHABETSBERGER, R. & JOLY, P. (2002). Short- and long term advantages of an alternative ontogenetic pathway. Biological Journal of the Linnean Society 77, 105–112. DENOËL, M. & JOLY, P. (2000). Neoteny and progenesis as two heterochronic processes involved in paedomorphosis in Triturus alpestris (Amphibia : Caudata). Proceedings of the Royal Society of London, Biological Sciences 267, 1481–1485. DENOËL, M. & JOLY, P. (2001). Adaptive significance of facultative paedomorphosis in Triturus alpestris (Amphibia, Caudata) : resource partitioning in an alpine lake. Freshwater Biology 46, 1387–1396. DENOËL, M. & PONCIN, P. (2001). The effect of food on growth and metamorphosis of paedomorphs in Triturus alpestris apuanus. Archiv für Hydrobiologie 152, 661–670. DENOËL, M., PONCIN, P. & RUWET, J. C. (2001b). Sexual compati- bility between two heterochronic morphs in the Alpine newt, Triturus alpestris. Animal Behaviour 62, 559–566. DENOËL, M. & SCHABETSBERGER, R. (2003). Resource partitioning in two heterochronic populations of Greek Alpine newts, Triturus alpestris veluchiensis. Acta Oecologica 24, 55–64. DENOËL, M., SCHABETSBERGER, R. & JOLY, P. (2004). Trophic specializations in alternative heterochronic morphs. Naturwissen- schaften 91, 81–84. DENVER, R. J. (1997). Environmental stress as a developmental cue : corticotropin-releasing hormone is a proximate mediator of adaptive phenotypic plasticity in amphibian metamorphosis. Hormones and Behavior 31, 169–179. DENVER, R. J., GLENNEMEIER, K. S. & BOORSE, G. C. (2002). Endocrinology of complex life cycles : amphibians. In Hormones, Brain and Behaviour, Vol. 2 (eds. D. Pfaff, A. Arnold, A. Etgen, S. Fahrbach and R. Rubin), pp. 469–513. Academic Press. DUELLMAN, W. E. & TRUEB, L. (1994). Biology of Amphibians. The John Hopkins University Press, Baltimore, Maryland. DUMERILL, A. (1866). Observations faites à la ménagerie du muséum d’histoire naturelle sur la reproduction des axolotls batraciens urodèles à branchies extérieures et sur les métamor- phoses qu’ils y ont subies. Bulletin de la Société Impériale d’Acclimatation, 1–11. Evolutionary ecology of facultative paedomorphosis in newts and salamanders 669