The chemical‐speciation hypothesis in Liolaemus: a response to Pincheira‐Donoso

Abstract

With more than 220 species, the South American Liolaemus is one of the most species-rich lizard genera on earth (Lobo, Espinoza & Quinteros, 2010). Strikingly, however, the factors behind this diversification have not been studied much, and hypotheses, such as rapid speciation because of isolation during quaternary glaciations (Fuentes & Jaksic, 1979), have been barely tested (Vidal, Moreno & Poulin, 2012). Recently, I published a study on chemical recognition in Liolaemus species and discussed its role in reproductive isolation (Labra, 2011). I also hypothesized that variation in recognition systems might contribute to rapid speciation in this genus. Pincheira-Donoso (2012) criticized this hypothesis, and I would like to comment upon his criticism. Pincheira-Donoso first questions my premise that Liolaemus has comparatively low morphological and ecological disparity (sensu Losos & Mahler, 2010), relative to its high species diversity. He supports his criticism by citing authors such as Cei (1986, 1993) and some of his own studies (Pincheira-Donoso, Hodgson & Tregenza, 2008a). However, those studies lack a proper control to clarify what ‘high disparity’ really means. Let us illustrate this by comparing Liolaemus with Varanus lizards, a genus with fewer species (<70 sp), but with a wider geographical distribution (Africa, Australia and Asia; Pianka & King, 2004) than Liolaemus, which is restricted to the southern part of South America. The snout–vent lengths of Varanus species range from 7 to 155 cm (Pianka & King, 2004; Collar, Schulte & Losos, 2011), while in Liolaemus this ranges from 3.5 to 11.5 cm (Espinoza, Wiens & Tracy, 2004; Schulte et al., 2004; Pincheira-Donoso et al., 2008a; Labra, Pienaar & Hansen, 2009). Varanus species can be herbivores, carnivores or omnivores, and they can be terrestrial, arboreal or aquatic (Pianka & King, 2004). In contrast, most Liolaemus are insectivorous/omnivorous, very few are strictly herbivores and there are no strict carnivores (Espinoza et al., 2004; Vidal & Labra, 2008; Pincheira-Donoso, Scolaro & Sura, 2008b). In addition, most Liolaemus are saxicolous or ground-dwellers, very few live in trees or shrubs, and there are no aquatic or semiaquatic species (Schulte et al., 2004; Pincheira-Donoso et al., 2009). Finally, the thermal physiology of Liolaemus seems highly conservative across species even considering the wide range of habitats they encounter (Labra et al., 2009). In view of all this information, I cannot agree with Pincheira-Donoso's criticism on this point. However, even if we were to accept the claim of high ecological and morphological disparity in this genus, there are cases of closely related and syntopic Liolaemus species that have similar ecology, morphology and behavior. Certainly, cases like these present a valuable opportunity to investigate whether species recognition plays a role in maintaining reproductive isolation between Liolaemus species. The verification of chemical species recognition in some species (Labra, 2011), together with ample evidence for the importance of chemical communication in the genus (Labra, 2008a, b ), make it plausible that speciation may be facilitated by the fast evolution of chemical sexual signals in the absence of variation in morphology or ecology (Morrison & Witte, 2011; Campagna et al., 2012). I am not implying that sexual speciation would prevent or limit morphological evolution and ecological adaptation, as Pincheira-Donoso assumes. The hypothesis simply predicts that Liolaemus species diversity is higher than what one would expect from ecological adaptation alone, and perhaps that the role of alternative sensory modalities (e.g. vision) in sexual selection would be small. Rapid evolution of chemical communication systems is a key element of my hypothesis. Pincheira-Donoso argues that his finding of a strong phylogenetic signal in the number of precloacal pores (a source of scents) in Liolaemus (Pincheira-Donoso, Hodgson & Tregenza, 2008c) is evidence against rapid evolution of the chemical sensory channel. However, his claim has two problems. First, it is a misconception that a strong phylogenetic signal implies a low evolutionary rate. A strong signal only indicates an association between the trait and the phylogeny, which could be due to similar adaptive responses in related species or to niche tracking, as well as to phylogenetic inertia (Labra et al., 2009 for detailed discussion). This error is perhaps most simply grasped from the fact that the evolutionary rate parameter in the Brownian-motion model used for phylogenetic analyses in Pincheira-Donoso et al. (2008c), is unrelated to the phylogenetic signal predicted by the model. Therefore, Pincheira-Donoso et al. (2008c) present no valid quantitative analysis of evolutionary lability of the chemical channel. Second, even if this source of scents would be an evolutionary constrained character, this does not imply that the chemical composition of precloacal scents, which is a key element in chemical communication (Mason & Parker, 2010), would be constrained. In fact, as I indicated in my study, the chemical composition of the precloacal secretions varies across species, populations and individuals (Escobar, Labra & Niemeyer, 2001; Escobar et al., 2003), which suggests that scents can evolve rapidly. Moreover, the precloacal secretions are just one source of scents used by Liolaemus (Labra, 2008a, b ), implying that these lizards have a huge spectrum of possibilities for scents, and in turn, for signals, to diverge. To summarize, quantitative assessments of the rates of evolution in chemical communication are still lacking for Liolaemus, and phylogenetic analyses of the disparity and variation of the chemical composition of the different secretions can shed some light on the problem. At this point, it is necessary to correct a misrepresentation of my study. Pincheira-Donoso wrote that the study ‘… presents evidence suggesting that these lizards respond more actively to conspecific than to heterospecific scents secreted by male precloacal glands.’ I designed the experiments to include any possible non-volatile secreted scent, not just those of the precloacal glands, because in the studied species, only male lizards have these glands (Labra et al., 2002; Labra, 2008b), as in most Liolaemus species (Pincheira-Donoso et al., 2008c). Therefore, I used a setup that allowed testing the ability of male and female lizards to recognize individuals of their same and different sex. The second major criticism of Pincheira-Donoso is that my study does not present direct evidence for chemically mediated mate choice or intersexual recognition, and so, there is no support for the hypothesis. There is no doubt that mate choice (or more precisely, assortative mating) has to be involved in the origin of reproductive isolation (Ptacek, 2000; Mendelson & Shaw, 2012). However, it is also clear that species recognition may be important in maintaining the cohesion of the species, and the signals used in the establishment of male territories and dominance relations would hardly be completely unrelated to signals involved in mate choice (Uy, Moyle & Filardi, 2009). In fact, coevolution between male signaling and female preferences is an essential part of many sexual selection models and has been documented in some cases (Grace & Shaw, 2011). Then, given that Liolaemus can extract significant information from scents (Labra, 2008a, b ), including the sex of the sender (Labra & Niemeyer, 1999), it is a plausible hypothesis that individuals will also favor the choice of conspecific over heterospecific mates. Therefore, it is not correct to indicate that the hypothesis in discussion has no support. Nevertheless, there is no doubt that we need more studies to have a better understanding of chemical communication and sexual selection in Liolaemus, as well as to test the specific hypothesis under discussion. I hope that the present debate will stimulate further research on these topics, which will surely prove interesting regardless of whether the chemical-speciation hypothesis for Liolaemus is right or wrong. I thank D. Pincheira-Donoso for providing this opportunity to clarify the hypothesis, its theoretical and experimental framework as well as its predictions. I also thank T. F. Hansen, R. Børsjø and V. Hayssen for discussions and comments on the article, and for help with the language. Funds come from FONDECYT 1090251.