ASSORTATIVE MATING IN POISON-DART FROGS BASED ON AN ECOLOGICALLY IMPORTANT TRAIT

Transcription

1 doi: /j x ASSORTATIVE MATING IN POISON-DART FROGS BASED ON AN ECOLOGICALLY IMPORTANT TRAIT R. Graham Reynolds 1,2,3 and Benjamin M. Fitzpatrick 4,5 1 Department of Biology, Duke University, Durham, North Carolina rgraham@utk.edu 4 Department of Ecology and Evolutionary Biology, University of Tennessee Knoxville, Tennessee benfitz@utk.edu Received March 1, 2007 Accepted May 2, 2007 The origin of new species can be influenced by both deterministic and stochastic factors. Mate choice and natural selection may be important deterministic causes of speciation (as opposed to the essentially stochastic factors of geographic isolation and genetic drift). Theoretical models predict that speciation is more likely when mate choice depends on an ecologically important trait that is subject to divergent natural selection, although many authors have considered such mating/ecology pleiotropy, or magictraits to be unlikely. However, phenotypic signals are important in both mate choice and ecological processes such as avoiding predation. In chemically defended species, it may be that the phenotypic characteristics influencing mate choice are the same signals being used to transmit a warning to potential predators, although few studies have demonstrated this in wild populations. We tested for assortative mating between two color morphs of the Strawberry Poison-Dart Frog, Dendrobates pumilio, agroup with striking geographic variation in aposematic color patterns. We found that females significantly prefer individuals of their own morph under two different light treatments, indicating strong assortative mating based on multiple coloration cues that are also important ecological signals. This study provides a rare example of one phenotypic trait affecting both ecological viability and nonrandom mating, indicating that mating/ecology pleiotropy is plausible in wild populations, particularly for organisms that are aposematically colored and visually orienting. KEY WORDS: Aposematism, assortative mating, coloration, Dendrobates pumilio, dendrobatidae, female mate choice, magic trait, speciation. Speciation the splitting of one species into two can be driven by a number of potential mechanisms including genetic drift, natural selection, and sexual selection (Coyne and Orr 2004). For species to persist as separate clusters of phenotypes, reproductive isolation must be strong enough to prevent fusion of those clusters due to interbreeding and production of intermediate offspring. Such isolation could be mediated through positive assortative mat- 3 Present Address: Department of Ecology and Evolutionary Biology, University of Tennessee, 569 Dabney Hall, Knoxville, TN ing among phenotypically similar individuals. Despite growing appreciation of the potential role of hybridization in evolution (Arnold 1997; Rieseberg 1997; Mavarez et al. 2006, Gompert et al. 2006), most individuals of most species mate assortatively (Mayr 1963; Mallet 2005). Thus, the evolution of mate choice is among the most important processes generating and maintaining biological diversity (Boake 2002). Models have shown that when mate choice is based on an ecologically important trait, divergence in that trait can accelerate allopatric speciation and facilitate more controversial processes of divergence-with-gene-flow (e.g., sympatric speciation 2253 C 2007 The Author(s). Journal compilation C 2007 The Society for the Study of Evolution. Evolution 61-9:

2 or reinforcement). However, the frequency with which assortative mating and adaptation are based on the same magic trait (Gavrilets 2004) is unknown and few examples from nature have been documented (Schluter 1996, 1998; McMillan et al. 1997; Jiggins et al. 2001; Podos 2001; Vines and Schluter 2006). In many phytophagous insects, host plant choice by parents simultaneously affects offspring viability and causes assortative mating (Caillaud and Via 2000; Craig et al. 2001; Berlocher and Feder 2002). However, the traits directly affecting performance in such systems (e.g., tolerance of plant secondary chemicals) are not necessarily the same traits as those directly affecting plant choice or mate choice and most models of speciation by host shift assume separate genetic control of preference and performance (Diehl and Bush 1989; Berlocher and Feder 2002; Fry 2003). Gavrilets (2004, p. 297) distinguished models of habitat preference from magic trait models in which a single trait (or set of loci) simultaneously affects ecological components of fitness and nonrandom mating. Here we examine positive assortative mating of poisondart frogs based on their most salient ecological feature: their bright warning-color patterns. Further, we suggest that the evolution of this trait is driving allopatric speciation of Dendrobates in Central America. The Strawberry Poison-Dart Frog, Dendrobates pumilio, is a tropical dart-poison frog inhabiting the Atlantic lowland regions of Nicaragua, Costa Rica, and Panama (Leenders 2001). This species is sexually monomorphic and exhibits bright aposematic coloration that has been shown to serve as an advertisement of sequestered toxic alkaloids in the skin (Duellman and Trueb 1986; Zug et al. 2001). Warning color should be subject to strong purifying selection due to the positive frequency-dependent nature of the benefit to individuals with a particular warning signal (Mallet and Joron 1999; Macedonia et al. 2002; Darst and Cummings 2006). The effectiveness of aposematic coloration depends on a consistent signal-response system between unpalatable prey and potential predators (Endler and Mappes 2004). New or rare variants are more likely to be attacked if predators do not recognize them as unpalatable (Müller 1879; Guilford and Dawkins 1993). Selection against rarity leads to very uniform populations of warning-colored species. On the other hand, warning color variation in toxic butterflies is often dramatic, with sharp distinctions between geographic variants of the same species, divergence between species, and even multiple mimicry rings within local communities (Mallet and Joron 1999). Thus, although aposematic coloration is subject to strong purifying selection, it shows levels of divergence between populations that are rarely seen in other traits. Similarly, color variation among populations of D. pumilio in the Bocas del Toro region of Panama is one of the more dramatic examples of intraspecific morphological diversity yet discovered among vertebrates (Myers and Daly 1983; Summers et al. 2003, 2004). Isolated populations of D. pumilio in this region show distinct color patterns that can vary in dorsal coloration (red, orange, blue, green, black, white, and yellow) and spotting/ speckling pattern (Myers and Daly 1983; Summers et al. 1999, 2003). Despite this diversity, D. pumilio has been considered a single species on the basis of male advertisement call similarity (Myers and Daly 1976) and DNA sequence similarity (Summers et al. 1997). Mitochondrial DNA data indicate that the origin of allopatric color morphs in the Bocas del Toro Archipelago has occurred within the last 6000 years, a relatively recent estimate suggesting a role for strong natural or sexual selection in the diversification of D. pumilio color pattern (Summers et al. 1997). Research on mate choice in frogs has focused almost entirely on the use of acoustic advertisement signals (Ryan 1980; Duellman and Trueb 1986; Schwartz 1987; Ryan and Rand 1990; Tárano 2002; Pröhl 2003). However, many tropical poison frogs are brightly colored and diurnal or crepuscular (Duellmann and Trueb 1986), providing the opportunity to use visual cues in social behavior (Summers et al. 1999). Dendrobates pumilio has color vision (Hailman and Jaeger 1974; Siddiqi et al. 2004) and females do use visual cues in mate choice (Summers et al. 1999). Siddiqi et al. (2004) also demonstrated that different color morphs of D. pumilio present distinct conspicuous signals detectable by both anuran and avian (model predator) visual systems against a variety of spectral backgrounds. These studies raised the intriguing possibility that the same visual signals used to warn predators of the frogs toxicity could also be used as intraspecific sexual signals. If mate choice is influenced by color pattern, divergence in warning color pattern could be driving speciation in the poison frogs of the Bocas del Toro Archipelago. Mating between individuals of D. pumilio from different populations has occurred in captivity; F1 offspring between D. pumilio color morphs are viable and appear intermediate in both color and spotting pattern (Summers et al. 2004). The F1s depicted in figure 1 of Summers et al. (2004) seem less conspicuous than the parentals due to their mixture of parental colors (although this judgment may not predict predator perception). They also exhibit distinct patterns of spectral reflectance (Summers et al. 2004) that are probably discriminable by both avian and anuran visual systems (Siddiqi et al. 2004). This may lead to greater vulnerability to predators (due to ineffective warning coloration) and lower attractiveness to potential mates. If so, color and spotting pattern could be causing reproductive isolation between color morphs by a combination of ecologically based selection and assortative mating. Summers et al. (1999) demonstrated assortative mating based on color between two D. pumilio color morphs (the red Cayo Nancy and the green Pope Island forms). Although they did not emphasize the importance of their results for speciation study, this represents an important example of prezygotic isolation between 2254 EVOLUTION SEPTEMBER 2007

3 Figure 1. Mean association times for female subject frogs from (A) Isla Colon and (B) Cayo Nancy. Darkened symbols represent blue-light trials and gray symbols represent white light trials. Vertical bars illustrate standard errors. allopatric populations with a causal link to divergence in an ecologically important trait, warning coloration. The importance of color was supported by eliminating male behavior (see Methods) and performing a second set of choice tests under blue light in which colors were indistinguishable. Assortative mate choice did not occur under the latter conditions. The morphs studied by Summers et al. (1999) differed in color alone (red vs. green). Here, we extend these results by testing assortative mating between morphs that differ in both overall coloration and the presence or absence of a melanistic spotting pattern. The Isla Colon morph is yellow/green with large black spots (Fig. 1A) and the Cayo Nancy morph is bright red/orange with few or no spots (Fig. 1B). If both background color and spotting pattern are used as cues for assortative mating, some level of prezygotic isolation may exist between all pairs of the 15 distinct D. pumilio forms of the Bocas Del Toro Archipelago in Panama. Materials and Methods We tested for assortative mating using a simple mate choice experiment following the protocol established by Summers et al. (1999). Female subject frogs were placed into a 19-liter terrarium with half of the tank occupied by plants for cover whereas the remainder was left clear. All frogs were left overnight in the terraria to allow for acclimatization before experimental trials were performed. Frogs were not fed during their captive period and were released one to two days after experimentation. Two female object frogs, one Cayo Nancy morph and one Isla Colon morph, were placed inside identical overturned clear glass cups 8 cm from each other at the clear end of the terrarium. Because D. pumilio are sexually monomorphic, females can be effective object males with regard to visual cues (Summers et al. 1999). Females were used as object frogs instead of males because of the males tendency to call during captivity, which could possibly influence subject female choice. As in Summers et al. (1999), object females were matched for snout-vent length to within 1 mm to control for possible size preferences. A previously recorded call of a male from Cayo Nancy of the same snout-vent length as the subject females was played at a consistent volume from a single speaker behind and equidistant to the cups. Calls of D. pumilio do not vary among islands, in fact acoustic signal variation within islands is greater than differences between islands (Myers and Daly 1976; Summers et al. 1999). The preference of subject females was estimated as the amount of time spent facing the object frog while within 4 cm of the cup. An experimental trial was defined as the experimental period (16 min. and 16 sec.) during which a single female was presented with the opportunity to select a potential mate (this time interval was determined by the length of the recorded advertisement call). The influence of dorsal background color versus presence or absence of large black spots was evaluated by comparing results EVOLUTION SEPTEMBER

4 of trials performed under two different light treatments: full white light and filtered blue light. Blue lighting regimes have been previously used to effectively homogenize the coloration of D. pumilio of different morphs, thereby removing any color signal differences between individuals from different populations (Summers et al. 1999). Under full white light, visual information on both color and spotting of the object frog was available to the subject female. Under filtered blue light, the colors of the object frogs were indistinguishable, as they both appeared blue, although any melanistic patterning remained in contrast to the rest of the dorsal coloration, leaving the presence or absence of spotting pattern as the primary visual cue available for the subject females. Filtered blue light was created by fitting blue-green spectrum filters (one BG-12 and one BG-23, Schott filters from BES Optics, Warwick, RT) to two 40-watt incandescent lights placed 50 cm above the terrarium. Totals of 19 Cayo Nancy and 19 Isla Colon females were tested under white light, and 10 of each morph under blue light (small and uneven sample sizes were due to time constraints and practical challenges in collection, housing, and manipulation). The time a subject frog spent associating with each object frog was recorded and the difference between the time spent with the object of the subject s own color morph and the alternative morph was evaluated with a paired t-test. Variation in the degree of preference between the two morphs and between the two light environments was tested with a two-way ANOVA. Here, the difference in individual association time between like and unlike object frogs was the dependent variable and the subject s morph and the light environment were fixed effects. The difference between association times was adequately approximated by a normal distribution for the purposes of parametric t-tests and ANOVAs (Shapiro Wilk test W = 0.92, P = 0.09). Nevertheless, nonparametric statistical tests are also reported below. Results The response of subject females to the object frogs was low (34%), with many trials resulting in no choice. As a consequence, final sample sizes were five Isla Colon females in each light treatment, four Cayo Nancy females in the white light treatment, and six Cayo Nancy females in the blue light treatment. This low response rate and small sample size is consistent with other studies of mating behavior in poison frogs (Summers et al. 1999), yet still allows the detection of strong assortative mating. Many of the 20 (of 58 attempted) females that did respond showed a clear courtship behavior such as stroking or nudging the glass containing the object frog (Limerick 1980; Summers et al. 1999). Overall, female D. pumilio spent over ten times more time associating with their own color morph than the alternative (Fig. 1; means and standard errors: 24.8 ± 5.85 sec vs. 1.7 ± 0.9 sec). The difference was significantly greater than zero according to both parametric and nonparametric tests (Table 1). Females from Cayo Nancy appeared somewhat less responsive than the Isla Colon females, especially under the full light conditions (Fig. 1B), but we could not reject the null hypotheses that preference was equivalent between light treatments and that each color morph was equally choosy (Two-way ANOVA: r 2 = 0.06, F 3,16 = 1.40, P = 0.279). Shirley s (1987) nonparametric two-way ANOVA also found no support for variation in the strength of preference between color morphs and light environments (overall Kruskal Wallis 2 3df = 4.69, P = 0.196). In contrast to results of Summers et al. (1999) for the Cayo Nancy and Pope Island morphs (green with few or no spots), we found that the Cayo Nancy and Isla Colon morphs show strong assortative mating under the monochromatic blue light treatment in addition to the full-spectrum treatment (Fig. 1). Considering only the blue light treatment, subject females spent significantly more time associating with their own morph (mean Table 1. Differences in association times of subject females with like versus unlike object males in Dendrobates pumilio, including all females tested and broken down by morph (Cayo Nancy and Isla Colon, see Fig. 1) and by light treatment within morph. Statistical tests are one-tailed. Group Mean Difference (s) Paired t-test Wilcoxon signed-rank test t df P W P All Cayo Nancy Blue Light Isla Colon Blue Light EVOLUTION SEPTEMBER 2007

5 difference = 25.4 sec, paired t = 3.58, 10 df, P = ; Wilcoxon signed rank test W = 63, P = ). These results hold even when each morph is analyzed separately (Table 1). Discussion Direct links between mate choice and ecologically important traits are expected to contribute to speciation because they bring about divergence in mating behavior as a by-product of ecological divergence. This is an attractive idea, linking sex, ecology, and evolution in a process that could rapidly produce new species as a consequence of natural selection and without the requirement of geographic isolation. However, the probability that the same trait will be involved in both assortative mating and ecological adaptation has been in doubt. Recent studies in Heliconius butterflies (Jiggins et al. 2001) and Darwin s finches (Podos 2001) have linked mate choice to ecologically important traits in comparisons between recently derived species. Together with the results of Summers et al. (1999), our study provides evidence of a direct link between assortative mating and ecologically important warning coloration within a single species. This result suggests that warning pattern evolution could be driving speciation in D. pumilio in the Bocas del Toro archipelago. There are over 15 distinct color morphs of D. pumilio in the Bocas del Toro region (Summers et al. 2003). Morphs are distinguished by color and/or degree of black spotting or striping that stands in contrast to the background color (e.g., see photographs in Siddiqi et al. 2004). Summers et al. (1999) showed assortative mating based on color alone between the Cayo Nancy morph (orange/red with few or no spots) and the Pope Island morph (green with few or no spots). Under the blue-light treatment, both of these unspotted morphs appeared uniformly blue, and females failed to choose their own color morph under these conditions of reduced spectral information (in fact there was a slight tendency to choose the opposite morph). Our results indicate that the contrasting pattern of black spots on a light background was adequate to maintain strong assortative mating between Cayo Nancy and Isla Colon morphs under the monochromatic blue light conditions. Our work and that of Summers et al. (1999) show that background color and melanistic spotting pattern each contribute to assortative mating in D. pumilio. Interestingly, clinal variation in coloration also seems to be a component of D. pumilio population structure on the mainland of the Bocas del Toro region (Summers et al. 2003). We have similarly observed spatial variation in color pattern within a single island, both on Isla Colon and Isla Bastimentos (R. G. Reynolds, pers. obs.). Although assortative mating has not been tested between variants from the same island, reproductive isolation between parapatric color morphs is an intriguing possibility. The reasons for color divergence among D. pumilio populations have yet to be completely understood. Color differences may be ecologically neutral if all color patterns are aposematic, as long as they are conspicuous (Siddiqi et al. 2004). The color differences among islands may represent alternative outcomes of independent processes of runaway (Fisherian) sexual selection whereby female preference drives the evolution of male coloration characteristics within constraints dictated by natural selection on aposematism (Lande 1981; Summers et al. 1997, 1999). The problem is that both mate choice and predator warning should often result in strong purifying selection against rare forms (Mallet and Joron 1999; Gavrilets 2004). If frogs with abnormal color characteristics are discriminated against by potential mates and more likely to be attacked by predators, how can novel color forms become established? Divergent color patterns and preferences may have become established by a drift in small island populations (Gavrilets and Boake 1998). Drift may result in a population passing through an adaptive valley and subsequently climbing a new adaptive peak. Although this is unlikely for very rugged adaptive landscapes (Barton and Charlesworth 1984), it could be facilitated by fluctuations in the intensity of selection (e.g., if predators are occasionally extirpated from small islands), or by the existence of complex high-fitness ridges across adaptive landscapes ( holey adaptive landscapes Gavrilets 1997). Alternatively, divergent coloration may result from divergent natural selection. Different color patterns may be adapted to different light environments (Endler 1993; Boughman 2001) although this seems unlikely in the Bocas del Toro Archipelago due to the similarity of habitats among islands (Summers et at. 2003). Different color patterns may also be adapted to different predator communities or may be involved in alternative Müllerian mimicry rings (Mallet and Joron 1999). Mimicry is unknown in D. pumilio, and we have no empirical basis for speculating that different color patterns signal to different predators. Ironically, the chemical defenses and warning signals of poison frogs are so effective that little is known about specific predators in nature. In short, the causes of warning color diversity in D. pumilio are unknown. However, the consequences for reproductive isolation are clear. Divergent color morphs mate assortatively (Table 1; Fig. 1; Summers et al. 1999), and hybrids tend to be intermediate and distinguishable from both parents (Siddiqi et al. 2004; Summers et al. 2004). Thus, divergence in warning color and pattern result in prezygotic isolation and likely postzygotic isolation due to the same trait. We propose that the strawberry poison frogs of the Bocas del Toro comprise a recent or incipient radiation where the evolution of reproductive isolation is being driven by warning color evolution. ACKNOWLEDGMENTS We wish to thank S. Alberts, Duke University, whose advice, support, contribution, and encouragement were what allowed this project to develop EVOLUTION SEPTEMBER

6 and succeed. Thanks to the Smithsonian Tropical Research Institute and staff; particularly L. Mou and M. Leone. Thanks as well to J. Fordyce, C. Boake, M. Niemiller, S. Gavrilets, and two anonymous reviewers who provided constructive criticism of the manuscript. We are indebted to the Bocas del Toro Smithsonian Tropical Research Institute (STRI) research station for assistance and support, and to the Republic of Panama and the Autoridad Nacional del Ambiente (ANAM) for collection permits. This project was funded by the Duke University Mellon Foundation and Center for Latin American Studies Travel Award Committee, the Duke University Dean s Summer Fellowship, and the Duke University Howard Hughes Fellowship in the Biological Sciences. LITERATURE CITED Arnold, M. L Natural hybridization and evolution. Oxford Univ. Press, Oxford, U.K. Barton, N. H., and B. Charlesworth Genetic revolutions, founder effects, and speciation. Annu. Rev. Ecol. Syst. 15: Berlocher, S. H., and J. L. Feder Sympatric speciation in phytophagous insects: moving beyond controversy? Annu. Rev. Entomol. 47: Boake, C. B Sexual signaling and speciation, a microevolutionary perspective. Genetica. 116: Boughman, J. W Divergent sexual selection enhances reproductive isolation in sticklebacks. Nature. 411: Caillaud, M. C., and S. Via Specialized feeding behavior influences both ecological specialization and assortative mating in sympatric host races of pea aphids. Am. Nat. 156: Coyne, J. A., and H. A. Orr Speciation. Sinauer, Sunderland, MA. Craig, T. P., J. D. Horner, and J. K. Itami Genetics, experience, and host-plant preference in Eurosta solidaginis: implications for host shifts and speciation. Evolution 55: Darst, C. R., and M. E. Cummings Predator learning favours mimicry of a less-toxic model in poison frogs. Nature 440: Diehl, S. R., and G. L. Bush The role of habitat preference in adaptation and speciation. Pp in D. Otte and J. A. Endler, eds. Speciation and its Consequences. Sinauer Associates, Sunderland, MA. Duellmann, W. E., and L. Trueb Biology of Amphibians. McGraw Hill, New York. Endler J. A The color of light in forests and its implications. Ecol. Monogr. 63:1 27. Endler, J. A., and J. Mappes Predator mixes and the conspicuousness of aposematic signals. Am. Nat. 163: Fry J. D Multilocus models of sympatric speciation: Bush vs. Rice vs. Felsenstein. Evolution 57: Gavrilets, S Evolution and speciation on holey adaptive landscapes. Trends Ecol. Evol. 12: Fitness landscapes and the origin of species. Princeton Univ. Press, Princeton, NJ. Gavrilets, S., and C. R. B. Boake On the evolution of premating isolation after a founder event. Am. Nat. 152: Gompert, Z., J. A. Fordyce, M. L. Forister, A. M. Shapiro, and C. C. Nice Homoploid hybrid speciation in an extreme habitat. Science. 314: Guilford, T., and M. S. Dawkins Receiver psychology and the design of animal signals. Trends Neurosci. 16: Hailman, J. P., and R. G. Jaeger Phonotactic responses to spectrally dominant stimuli and use of colour vision by adult anuran amphibians: a comparative survey. Anim. Behav. 22: Jiggins, C. D., R. E. Nalsbit, R. L. Coe, and J. Mallet Reproductive isolation caused by colour pattern mimicry. Nature. 411: Lande, R Models of speciation by sexual selection on polygenic traits. Proc. Natl. Acad. Sci. USA 78: Leenders, T Amphibians and Reptiles of Costa Rica. Distribuidores Zona Tropical, S. A. Miami, FL. Limerick, S Courtship behavior and oviposition of the poison-arrow frog Dendrobates pumilio. Herpetologica 36: Macedonia, J. M., Y. Brandt, and D. L. Clark Sexual dichromatism and differential conspicuousness in two populations of the common collared lizard (Crotaphytus collaris) from Utah and New Mexico, USA Biol. J. Linn. Soc. 77: Mallet, J Hybridization as an invasion of the genome. Trends Ecol. Evol. 20: Mallet, J., and M. Joron Evolution of diversity in warning color and mimicry: polymorphisms, shifting balance, and speciation. Annu. Rev. Ecol. Syst. 30: Mavarez, J., C. A. Salazar, E. Bermingham, C. Salcedo, and C. D. Jiggins Speciation by hybridization in Heliconius butterflies. Nature. 441: Mayr, E Animal species and evolution. Harvard Univ. Press, Cambridge, MA. McMillan, W. O., C. D. Jiggins, and J. Mallet What initiates speciation in passion vine butterflies? Proc. Natl. Acad. Sci. USA 94: Müller, F Ituna and Thyridia: a remarkable case of mimicry in butterflies. Trans. Entomol. Soc. Lond. 1879: Myers, C. W., and J. W. Daly Preliminary evaluation of skin toxins and vocalization in taxonomic and evolutionary studies of Poison-Dart Frogs (Dendrobatidae). Bull. Am. Mus. Nat. Hist. 157: Dart-poison frogs. Sci. Am. 248: Podos, J Correlated evolution of morphology and vocal signal structure in Darwin s finches. Nature 409: Pröhl, H Variation in male calling behavior and relation to male mating success in the strawberry poison frog (Dendrobates pumilio). Ethology 109: Ryan, M. J Female mate choice in a Neotropical frog. Science 209: Ryan, M. J., and A. S. Rand The sensory basis of sexual selection for complex calls in the tungara frog, Physalaemus pustulosus. Evolution 44: Rieseberg, L. H Hybrid origins of plant species. Annu. Rev. Ecol. Syst. 28: Schluter, D Ecological causes of adaptive radiation. Am. Nat. 148:S40 S Ecological causes of speciation. Pp in D. A. Howard and S. Berlocher, eds. Endless forms: species and speciation. Oxford Univ. Press, New York. Schwartz, J. J The function of call alteration in anuran amphibians: a test of three hypotheses. Evolution 41: Shirley E. A. C Applications of ranking methods to multiple comparison procedures and factorial experiments. Appl. Stat. 36: Siddiqi A., T. W. Cronin, E. R. Loew, M. Vorobyev, and K. Summers Interspecific and intraspecific views of color signals in the strawberry poison frog Dendrobates pumilio. J. Exp. Bio. 207: Summers, K., E. Bermingham, L. Weight, S. McCafferty, and L. Dahlstrom Phenotypic and genetic divergence in three species of dart-poison frogs with contrasting parental behavior. J. Hered. 88:8 13. Summers, K., R. Symula, M. Clough, and T. Cronin Visual mate choice in poison frogs. Proc. R. Soc. Lond. 266: Summers, K., T. W. Cronin, and T. Kennedy Variation in spectral reflectance among populations of Dendrobates pumilio, the strawberry 2258 EVOLUTION SEPTEMBER 2007

7 poison frog, in the Bocas del Toro Archipelago, Panama. J. Biogeogr. 30: Cross-breeding of distinct color morphs of the strawberry poison frog(dendrobates pumilio) from the Bocas del Toro Archipelago, Panama. J. Herp. 38:1 8. Tárano, Z Vocal responses to conspecific call variation in the neotropical frog Physalaemus enesefae. J. Herp. 36: Vines, T. H., and D. Schluter Strong assortative mating between allopatric sticklebacks as a by-product of adaptation to different environments. Proc. R. Soc. B. 273: Zug, G. R., L. J. Vitt, and J. P. Caldwell Herpetology, an Introductory Biology of Reptiles and Amphibians. 2nd ed. Academic Press, San Diego, CA. Associate Editor: W. O. McMillan EVOLUTION SEPTEMBER