UNRAVELING the CONTRIBUTION of CLIMATIC HETEROGENEITY on EVOLUTIONARY PROCESSES WITHIN OOPHAGA POISON DART FROGS Diana María G

Home , Brachymyrmex

UNRAVELING THE CONTRIBUTION OF CLIMATIC HETEROGENEITY ON EVOLUTIONARY PROCESSES WITHIN OOPHAGA POISON DART FROGS

Diana María Galindo-Uribe

Grupo de Ecofisiología, Comportamiento y Herpetología-GECOH

Departamento de Ciencias Biológicas

Universidad de los Andes

Bogotá, Colombia

TESIS SOMETIDA AL DEPARTAMENTO

CIENCIAS BIOLÓGICAS EN EL CUMPLIMIENTO PARCIAL DE LOS REQUISITOS

PARA OBTENER EL GRADO DE

DOCTOR EN CIENCIAS - BIOLOGÍA

Tutor

Adolfo Amézquita, PhD

Universidad de los Andes

UNIVERSIDAD DE LOS ANDES

BOGOTÁ, COLOMBIA

Febrero 2017 TABLE OF CONTENTS

ACKNOWLEDGMENTS ...... III

ABSTRACT ...... 1

RESUMEN ...... 2

GENERAL INTRODUCTION ...... 4

COLOUR AND PATTERN VARIATION AND PLEISTOCENE PHYLOGEOGRAPHIC ORIGIN OF THE STRAWBERRY POISON FROG, OOPHAGA PUMILIO, IN NICARAGUA ...... 21

THE ROLE OF CLIMATIC SEASONALITY AND UNPREDICTABILITY IN SHAPING INTERSPECIFIC PATTERNS OF GEOGRAPHIC VARIATION IN WATER BALANCE TRAITS OF DENDROBATES AND OOPHAGA SPECIES ..... 44

THE SENSITIVITY OF LOCOMOTOR PERFORMANCE TO DEHYDRATION IN THE POISON FROG OOPHAGA PUMILIO (ANURA, DENDROBATIDAE) ...... 79

DOES FINER ENVIRONMENTAL HETEROGENEITY PROMOTE MICROEVOLUTIONARY PROCESSES? A CASE STUDY IN THE KÓKÓE-PÁ POISON FROG, OOPHAGA HISTRIONICA, IN COLOMBIA ...... 91

CONCLUSIONS ...... 122

ACKNOWLEDGMENTS

Quisiera Agradecer a mi tutor Adolfo Amézquita por todos estos años de invaluables enseñanzas académicas, pero sobre todo por haberme enseñado otra forma de estudiar la vida y las ranas. También por su apoyo personal sin el cual no habría sido posible terminar mi doctorado. Quisiera agradecer a Heike Pröhl por su gran apoyo logístico, académico y personal en el desarrollo de diferentes capítulos de esta investigación. A Catalina Gonzales por su gran disposición y comentarios personales que contribuyeron enormemente a este trabajo. A Carlos Navas y Daniel Cadena por anteriores evaluaciones y valiosos comentarios a mi proyecto. A los diferentes investigadores que participaron y apoyaron la realización de esta tesis: Miguel Vences, Santiago Merino, Santiago Ron y Susane Hauswaldt. A mis amigos herpetólogos que también participaron de esta investigación y me han apoyado en diferentes etapas: Abel Batista, Jose Alfredo Hernandez, Fernando Vargas, Roberto Marquez e Iliana Medina. A mis amigos de la Universidad de los Andes y del laboratorio GECOH que me han apoyado y aportado en este proceso: Valeria Gonzalez, Javier Mendez, Carolina Esquivel, Mieidy Betancourt, Pablo Palacios, Beatriz Eugenia Velasquez, Catalina Silva, Osvaldo Gil, Vicky Felchas, Nelsy Pinto e Ivan Beltrán. A mis amigos técnicos y estudiantes en el Instituto de Medinica Veterinaria de Hannover que me apoyaron en la pasantía: Sabine Sippel, Johara Brooke, Corinna Dreher y Constantine. A Federico Chinchilla, Alejandro Farji y al curso más Zombie.

A Colciencias, la Facultad de Ciencias de la Universidad de los Andes, la Organización de Estudios tropicales y a la DGTH por financiar mis estudios y proyectos. A la Universidad de Medicina Veterinaria de Hannover, A la Pontificia Universidad Católica de Ecuador, A la Universidad Nacional Autónoma de Nicaragua. Al Ministerio de Medio Ambiente y Desarrollo Sostenible de Colombia, al Ministerio del Ambiente y los Recursos Naturales de Nicaragua, Al Ministerio de Ambiente de Panamá, Al ministerio de Ambiente de Costa Rica por los permisos de investigación otorgados.

iii

Quisiera agradecer a mis profesores de pregrado y anteriores investigaciones que han aportado en mi vida académica y personal: Julio Mario Hoyos, Saúl Prada, Carlos Rivera y Alberto Acosta.

A las ranas, bosques, paisajes y personas que inspiraron este trabajo y me regalaron a mi esposo Javier e hijo Pablo. A mi esposo quiero agradecerle por haberme apoyado, enseñado y participado como herpetólogo, amigo y compañero en todo este proceso. A mi hijo, quien fue el mejor resultado de mi tesis, por darle relevancia y un mejor sentido a mis quehaceres. A todos ellos gracias por su apoyo a paciencia. A mi madre por ser una gran mujer, mi mejor ejemplo a seguir y por su apoyo incondicional a lo largo de mi vida. A todos ellos quisiera finalmente dedicarles este trabajo. Porque este es gracias a ellos y para ellos.

ABSTRACT

Understanding the role of climatic factors on diversification processes remains among the pivotal questions in ecology and evolutionary biology. The main goal of this study was to approach this question at different temporal and spatial scales within a convenient model: the egg-feeding poison frogs in the genus Oophaga and their relatives. I first constructed a phylogeographic hypothesis on the evolution of Oophaga pumilio lineage. Phylogeographic analysis suggest that the Nicaraguan populations originated by a Late Pleistocene range expansion northwards, and reconstructs the origin of the species in southeastern Costa Rica. I investigated geographic variation in physiological traits linked to the regulation of evaporative water loss and tested for correlations with current and past climatic heterogeneity. I further assessed the effect of dehydration on the locomotor performance of Oophaga pumilio. Both temporal and spatial heterogeneity in climate contributed to explain geographic variation in evaporative water loss and body size; and the rates of water loss in frogs had a negative effect on their locomotor performance, eventually affecting the survivorship and fitness of individuals. Finally, to test for the role of microclimatic heterogeneity in shaping dispersal mechanisms and promoting microevolutionary processes, I studied the contribution of local variation in microclimate, topography and availability of reproductive resources in shaping the distributional patterns of individuals, and gene flow and genetic structuring within a population of Ooophaga histrionica. I conclude that climatic heterogeneity was correlated with phenotypic and genotypic divergence in Oophaga frogs at different spatial and temporal scales. Local environmental heterogeneity affected the distribution of individuals and the gene flow among spatial clusters of reproductively active individuals. Combining my data and previous studies, the evidence supports that climatic heterogeneity acted in conjunction with sexual selection and predation processes as the main factors promoting evolutionary divergence within this genus.

RESUMEN

Una de las preguntas más esenciales en ecología y biología evolutiva ha sido entender el rol de factores climáticos en procesos de diversificación. El objetivo principal de este estudio fue abordar esta pregunta a diferentes escalas temporales y espaciales en un modelo convencional: las ranitas comedoras de huevos del género Oophaga y sus parientes. Inicialmente reconstruí una hipótesis filogeográfica acerca de la evolución de Oophaga pumilio incluyendo poblaciones de Nicaragua no incluidas en estudios preliminares. Los análisis filogeográficos sugieren que estas poblaciones se orignaron a través de la expansión de su rango de distribución al norte durante el Pleistoceno tardío y reconstruyen el origen de la especie en el suroccidente de Costa Rica. Posteriormente investigué la variación geográfica en rasgos fisiológicos relacionados con la regulación de pérdida de agua por evapotranspiración y puse a prueba correlaciones entre estos rasgos fisiológicos con las condiciones de heterogeneidad climática temporal. Concluyo que la heterogeneidad climática contribuye a explicar la variación geográfica en la pérdida de agua por evapotranspiración y el tamaño corporal. Posteriormente evalué el efecto de la deshidratación en el desempeño locomotor. Finalmente, para poner a prueba la contribución de la heterogeneidad microclimática en moldear los procesos de dispersión local y promover procesos microevolutivos, estudié la contribución de la heterogeneidad del clima local, la topografía y la disponibilidad del recurso reproductivo sobre la distribución espacial de los individuos, el flujo genético y la estructura genética dentro de una población de Oophaga histrionica. La heterogeneidad ambiental local afectó la distribución de los individuos y el flujo genético entre clusters espaciales, compuestos de individuos activos reproductivamente. A partir de la combinación de mis datos y de estudios previos, la evidencia soporta que la heterogeneidad climática actúa en conjunto con la selección sexual y procesos de depredación, como los principales factores promoviendo divergencia evolutiva al interior del género.

Oophaga pumilio from Río Indio Lodge, Nicaragua.

GENERAL INTRODUCTION

Environmental and historical factors seem to act in conjunction, along current and evolutionary time scales, in shaping the dispersal of lineages and the diversity patterns of assemblages and communities across geographic gradients. Contemporary factors include ecophysiological performance, competition, habitat diversity and habitat choice. Historical factors include evolutionary and biogeographical processes that lead to transformation: the temporal dynamics of the distribution range, speciation, extinction, the geographic history of lineages, the time since the origin of the lineages, their ancestral physiological tolerances, and the opportunities for niche evolution and adaptation. Contemporary and historical factors influence the process of dispersion at different (local vs. regional) spatial scales (Gaston 1998, Wiens & Donoghue 2004, Ricklefs 2006, Hortal et al. 2009).

The contribution of climatic heterogeneity to the patterns of species richness and diversity have been discussed in the light of hypotheses that include deterministic and historical components (Table 1). Some of them predict a positive effect of climatic heterogeneity, energy availability (temperature, water and primary productivity) and the historical climatic stability on species diversity. Some other recall the significance of the climatic conditions and physiological tolerances in the original vs the new (dispersal) distribution areas in constraining species diversity. And other recall the significance of climatic conditions within regions for increasing speciation rates compared to extinction rates (see Currie 1991, Hawkins et al. 2003, Currie et al. 2004, Tews et al. 2004, Wiens and Donoghue 2004, Ricklefs 2006, Araújo et al. 2008, Hortal et al. 2009).

Within the tropics, lineages may be restricted and adapted to narrower and less fluctuating (daily and seasonal) climatic conditions in comparison with temperate species. Therefore, habitat fidelity is expected to be higher and climatic tolerances narrow. Under these conditions, environmental gradients may become relevant barriers for the dispersal of tropical organisms (Janzen 1967). Amphibians are particularly sensitive to environmental conditions, because they are ectothermic,

4 small, and have permeable skins, which make them more prone to desiccation. In the tropics, they should consequently show narrow climatic and geographic ranges as well as strong habitat fidelity (Hawkins et al. 2003, Buckley & Jetz 2008, Hua & Wiens 2010).

Amphibian diversity is considered functionally linked to water availability not only because amphibians are vulnerable to desiccation, but also because water may influence food availability in warm environments (Hawkins et al. 2003). Currie (1991) found that the variability in species diversity in four vertebrate classes (including amphibians and reptiles) is best explained by annual potential evapotranspiration (PET) as an indirect estimate of energy availability. Current patterns of amphibian diversity are also more related to climatic stability between the last maximum glacial (LMG) and the present, than to contemporary climate alone; the distribution of species with narrow ranges seems also constrained by the mean annual freezing conditions in the LGM, whereas widespread species are more constrained by current mean annual freezing conditions (Araújo et al. 2008).

The family Dendrobatidae consists of about 184 frog species distributed from Nicaragua to Bolivia in the west, and to the French Guiana and Brazil in eastern South America (Frost 2016). Some species are known as “Poison-Dart Frogs” in reference to the batrachotoxin from Phyllobates terribilis and allies, used to poison darts by the Colombian aborigines Embera Katio (Myers & Daly 1976). Almost all species are diurnal and all lay terrestrial eggs; they are also territorial and exhibit elaborate courtship and parental care (Myers & Daly 1976, Pröhl 2005). The defense of their territories is mediated by advertisement calls and physical combats (Pröhl 2005). Fertilization and embryo development occur out of water. Once hatched, the tadpoles of most species are transported to small body waters where they complete their development until the metamorphosis (Savage 1968, Crump 1972, Myers & Daly 1976).

Table 1. Hypotheses and mechanisms proposed to understand the contribution of climate to diversity patterns.

Hypothesis Patterns Mechanisms Literature Structurally complex habitats as Bazzaz 1975, A positive effect of successional ecosystems, may provide Hortal et al. Habitat heterogeneity heterogeneity over species more niches and diverse ways of 2009, Tews et diversity exploiting the environmental resources al. 2004 Net primary productivity (NPP) limits A positive effect of energy the number of individuals, the number Brown 1981, Energy availability availability over species of coexisting species and climate Wright 1983 diversity strongly affects NPP Species diversity is limited by Fewer species can physiologically Physiological tolerance the number of species that Root 1988 tolerate lower temperatures can tolerate local conditions A positive effect of energy As ambient energy or temperature Turner et al. Energy-Physiology availability over species declines with an increase in latitude, 1987, Currie diversity species diversity also declines 1991 Low temperature (energy) constraints diversity at higher latitudes, whereas Hawkins et al. water constrains richness in areas with 2003 Species diversity is high energy inputs Water-energy constrained by temperature Broad scale and water availability patterns of richness depend mainly on Francis & Currie the simultaneous availability of heat and 2003 water McGlone 1996, Historic climate Severe climate changes Persistence and speciation are favored Araújo et al. stability constrain diversity by climate stability over time 2008 Species are better adapted to the conditions of the ecological zone of Latham & origin of their lineage and the transition Ricklefs 1993, Ecological zones of The diversity is greater in the to other ecological zones requires Ricklefs 2006, origin zone of origin evolutionary change and may constrain Wiens & the diversity of species. The zone of Donoghue 2004, origin which may be older, more Ricklefs 2006 widespread of less stressful Evolutionary rates may be faster at higher ambient temperatures, due to Rohde 1992; The diversity is greater in shorter generation times, higher mutation Allen et al. 2002 Rates of diversification areas with higher rates, and faster physiological processes diversification rates Speciation rates within warmer and Schemske 2002 humid sites, where biotic interactions are in Currie et al. stronger 2004

Ecology of poison frogs in the genus Oophaga

We focus here on the genus Oophaga and secondarily Dendrobates. Oophaga is present from the west coast of Nicaragua, to the western Andean slopes in Ecuador. It contains O. arborea, O. granulifera, O. pumilio, O. speciosa, O. vicentei in Central America, and O. histrionica, O. lehmanni, O. occultator, and O. sylvatica in South America (Frost 2016). It has been widely studied because of their astonishing intraspecific variability in phenotypic traits but also because of their noticeable and interesting behavior. Dendrobates is present from southern Nicaragua through the French Guiana and adjacent Brazil; it has also been introduced into Hawaii (Frost 2016). It contains D. auratus in Central America and northwestern Colombia; D. truncatus in Colombia; D. leucomelas present in the Amazonian and Guiana shield of Colombia, Venezuela, Brazil and Guiana; D. nubeculosus only known from Guyana; and, D. tinctorius from the Guianas and Brazil (Frost 2016).

I report here on all Oophaga species plus Dendrobates auratus and D. truncatus, which overlap in distribution with Oophaga. Most published studies were conducted on few populations of O. pumilio and O. granulifera; much less is known on O. histrionica, O. lehmanni or O. sylvatica, and almost no information is available on O. arborea, O. speciosa, O. vicente, and O. occultator. Therefore, I describe general aspects of their natural history mostly based on evidence collected in a few populations.

Oophaga, as most other studied dendrobatids, possess an astonishing intraspecific diversity in skin alkaloids (Saporito et al. 2007), body size, escape behavior, color patterns (Daly & Myers 1967, Silverstone 1975, Myers & Daly 1976; Daly et al. 1987, Summers et al. 2003, Batista & Köhler 2008, Mebs et al. 2008, Pröhl & Ostrowski 2011, Willink et al. 2013) courtship behavior (Pröhl & Ostrowski 2011) and call parameters (Lötters et al. 1999, Pröhl et al. 2007), which make them excellent candidates for studies on aposematism and sexual selection. Oophaga species are territorial and within their territories they feed, court, breed and attend their eggs (Pröhl 1997, 2005, Pröhl & Hödl 1999, Haase & Pröhl 2002, Staudt et al.

2010). Males that possess a territory (resident males) defend it from other males by calling and engaging in physical combats (Goodman 1971, Crump 1972, Bunnell 1973, Silverstone 1973, Forester et al. 1993, Pröhl 1997, Pröhl & Berke 2001). They seem to defend their territories for long periods of time from known and new coming intruders (Pröhl & Berke 2001). They do not seem to discriminate between relatives and aggressively react to any calling intruder male within their territories, but not to non-calling males, juveniles or females (Goodman 1971, Pröhl 1997, Gardner & Graves 2005). They also possess a great ability to return to their territories after been experimentally displaced in the field (McVey et al. 1981), possibly on the basis of acoustic and olfactory cues (Forester & Wisnieski 1991). Females in some populations have also been reported to be territorial as well as aggressive (Haase & Pröhl 2002, Meuche et al. 2013) and their territories are smaller than in males but their home ranges are larger (Meuche et al. 2013). In other populations, females do not defend territories but move within specific home range in the search of food, males or tadpole rearing sites, and their home ranges are larger than male territories (Donnelly 1989a, Pröhl 1997, Pröhl & Berke 2001, Gardner & Graves 2005).

In O. pumilio, it has been proposed that space use in both sexes depends upon the distribution of reproductive limiting resources such as tadpole rearing sites (Donnelly 1989a, Donnelly 1989b, Pröhl & Berke 2001). Limiting resources for males appear to be the females, and in the case of females the limiting resources are the tadpole rearing sites (Pröhl & Berke 2001, Pröhl 2005, Meuche et al. 2012). Females also tend to overlap their home ranges around the sites with greater availability of tadpole rearing sites (Donnelly 1989a). Even territorial males tend to “aggregate” and occupy empty spaces in places with high density of females (Donnelly 1989a, Pröhl 1997, Meuche et al. 2012). The immigration rates and survivorship are also higher for adult males and females (Donnelly 1989b, Pröhl & Berke 2001).

These frogs feed mainly on ants and mites taken from the forest floor, and apparently sequester their skin alkaloids from their prey (Daly 1998). Recently, the

8 presence of pumiliotoxins, an alkaloid present in greater abundance in species like O. pumilio, was identified in formicinae ants of the genera Brachymyrmex and Paratrechina, which are part of the diet in these frogs and are very common in the leaf litter (Saporito et al. 2004). Frog territories have more ants of those genera than the unoccupied sites. Within the territories, small areas are used for feeding (Donnelly 1991, Staudt et al. 2010). The frogs and the ants may also share microhabitat requirements (Donnelly 1991). The abundance of these ants may strongly fluctuate temporally in some populations, and therefore alkaloids may represent another limiting and important resource for these frogs (Donnelly 1991, Staudt et al. 2010). Therefore, the establishment of male territories around this feeding and alkaloid resource where the density of females is greater seems very likely (Staudt et al. 2010).

It has been suggested that Oophaga species tend to occur around forest clearings or large fallen logs (Goodman 1971). These clearings have been proposed as sites that may confer greater availability of elevated perches to call for males (Crump 1972) or sites with better transmission of calls to females (Pröhl 2005). It has also been demonstrated, that within their territories, the amount of UV light is used as a cue to avoid perches that might increase the exposure to UV light (Han et al. 2007, Kats et al. 2012). The use of specific perches and the selection of microhabitats appear also to maximize frog conspicuousness against the background and hence predator learning (Pröhl & Ostrowski 2011, reviewed in Rojas 2016). Otherwise, the aggregation of territories and the perches within has also been suggested as a potential mechanism that increases the efficiency of the aposematic signal at high frog densities (Rojas 2016).

In the morning, males call from usually elevated perches within their territories, to attract receptive females and to defend their territories from intruder males (Bunnell 1973, Pröhl 1997, Pröhl & Hödl 1999, Graves 1999, Graves et al. 2005). Females seem to move between territories and select among calling males (Pröhl & Hödl 1999, Haase & Pröhl 2002); it is also possible that females just breed with the closest calling male (Meuche et al. 2013). At the end of a series of calls, courtship

9 postures and tactile interactions, males lead the female to several oviposition sites within the forest litter. Once the female accept an oviposition site, the males leave their sperm in the surface of leaves and the female lays eggs on it in order to complete the fertilization (Crump 1972, Limerick 1980). Then, males participate in the parental care of eggs maintaining them humid with their ventral surface (Pröhl & Berke 2001). Once the eggs hatch, the female takes the tadpoles in her back and transport then to small water bodies in the axils of different plants, like palms, heliconias and bromeliads (Young 1979, Limerick 1980, van Wijngaarden & Bolaños 1992). Then, and this is the trait which explains the genus name (Oophaga sensu Grant), females feed their tadpoles with unfertilized eggs, as their only source of food until they reach the metamorphosis (Brust 1993, van Wijngaarden & Bolaños 1992). Females seem to recognize and only feed their own progeny. During the feeding period, the females appear not to be available for courtship (Pröhl & Hödl 1999, Haase & Pröhl 2002).

Despite the dry season in the wet lowland tropical forest is not pronounced, different authors found a positive association between daytime activity, calling activity, the appearance of juvenile frogs, as well as changes in the size of the ovarian complement in females, with an increase in precipitation or habitat moisture in O. pumilio, suggesting a seasonal reproduction or a reduction in survivorship of eggs and tadpoles along dry periods (Donnelly 1989c, Pröhl 1997, Graves et al. 2005).

Oophaga arborea inhabits cloud forests where nearly constant fog, often accompanied by rain and mist, favours very dense herbaceous and fern layer at the ground level, and a very high density of epiphytes including bromeliads. This species has been characterized as arboreal and has only been associated to these epiphytic plants (Myers et al. 1984). Oophaga granulifera seems to prefer steeped banks along fast moving streams (Goodman 1971, van Wijngaarden & Bolaños 1992, Willink et al. 2013). Within O. pumilio, ecological specialization in some populations has been proposed. For example, near Chiriquí Grande in Panama, red

10 frogs have been associated with the ground adjacent to swamp forests, and green frogs with trees in the swamp forest (Daly & Myers 1967).

Oophaga vicentei from Cerro Narices, Panama.

Oophaga arborea individuals calling from Bromeliads on the top of a tree. Fortuna, Panama.

Along their range, Oophaga species have distributional gaps, as the one in eastern Panama (Myers & Böhme 1996). Also, they inhabit moist to wet evergreen forests. Along the Atlantic coast, this habitat is continuous from Mexico to the Colombian border with Panama, as well as in the western coast of Colombia and northern Ecuador. However, on the pacific Coast of Costa Rica and Panama they occur on islands separated between them by dry habitats. Also, the Cordillera de Talamanca-Chiriquí in Costa Rica and Panama probably represents an effective barrier between the Pacific and Atlantic lowlands. Most species are distributed in elevations corresponding to tropical lowlands (temperature above 24 °C) but some are present in subtropical temperature conditions (between 18 and 25 °C, Savage 1968).

Intraspecific diversity can be distributed geographically in three general patterns. First, different color morphs occur in separated geographic locations; second, a mixture of syntopic color morphs; and third, the same morph is present in geographically disjunct “demes”. In the third case, there can be a continuum gradation in coloration pattern (Myers & Böhme 1996 citing to Myers & Daly 1976 and Daly et al. 1987, Summers et al. 2003).

Dendrobates auratus and Oophaga pumilio at Río Indio Lodge, Nicaragua.

Dendrobates auratus and Oophaga pumilio from Cerro Tébata, Panama.

Oophaga pumilio from Cerro Musún, Nicaragua.

Literature cited

Allen, A. P., Brown, J. H., & Gillooly, J. F. (2002). Global biodiversity, biochemical kinetics, and the energetic-equivalence rule. Science, 297(5586), 1545–1548.

Araújo, M. B., Nogués-Bravo, D., Diniz-Filho, J. A. F., Haywood, A. M., Valdes, P. J., & Rahbek, C. (2008). Quaternary climate changes explain diversity among reptiles and amphibians. Ecography, 31(1), 8–15.

Batista, A., & Köhler, G. (2008). Variation in Oophaga pumilio (Amphibia: Anura: Dendrobatidae) in western Panama. Salamandra, 44(4), 225–234.

Bazzaz, F. A. (1975). Plant species diversity in old‐field successional ecosystems in Southern Illinois. Ecology, 56(2), 485–488.

Brown, J. H. (1981). Two decades of homage to Santa Rosalia: toward a general theory of diversity. American Zoologist, 21(4), 877–888.

Brust, D. G. (1993). Maternal brood care by Dendrobates pumilio: a frog that feeds its young. Journal of Herpetology, 27(1), 96–98.

Bunnell, P. (1973). Vocalizations in the territorial behavior of the frog Dendrobates pumilio. Copeia, 1973(2), 277–284.

Buckley, L. B., & Jetzc, W. 2008. Linking global turnover of species and environments. PNAS, 105(46), 17836–17841.

Crump, M. L. (1972). Territoriality and mating behavior in Dendrobates granuliferus (Anura: Dendrobatidae). Herpetologica, 28(3), 195–198.

Currie, D. J. (1991). Energy and large-scale patterns of animal-and plant-species richness. American Naturalist, 137(1), 27–49.

Currie, D. J., Mittelbach, G. G., Cornell, H. V., Field, R., Guégan, J. F., Hawkins, B. A., Kaufman, D. M., Kerr, J. T., Oberdorff, T., O'Brien, E., & Turner, J. R. G. (2004). Predictions and tests of climate-based hypotheses of broad-scale variation in taxonomic richness. Ecology letters, 7(12), 1121–1134.

Daly, J. W. (1998). Thirty years of discovering arthropod alkaloids in amphibian skin. Journal of natural products, 61(1), 162–172.

Daly, J. W., & Myers, C. W. (1967). Toxicity of Panamanian poison frogs (Dendrobates): some biological and chemical aspects. Science, 156(3777), 970– 973.

Daly, J. W., Myers, C. W., & Whittaker, N. (1987). Further classification of skin alkaloids from Neotropical poison frogs (Dendrobatidae), with a general

survey of toxic/noxious substances in the amphibia. Toxicon, 25(10), 1023– 1095.

Donnelly, M. A. (1989a). Effects of reproductive resource supplementation on space-use patterns in Dendrobates pumilio. Oecologia, 81(2), 212–218.

Donnelly, M. A. (1989b). Demographic effects of reproductive resource supplementation in a territorial frog, Dendrobates pumilio. Ecological Monographs, 59(3), 207–221.

Donnelly, M. A. (1989c). Reproductive phenology and age structure of Dendrobates pumilio in northeastern Costa Rica. Journal of Herpetology, 23(4), 362–367.

Donnelly, M. A. (1991). Feeding patterns of the strawberry poison frog, Dendrobates pumilio (Anura: Dendrobatidae). Copeia, 1991(3), 723–730.

Forester, D. C., Cover, J., & Wisnieski, A. (1993). The influence of time of residency on the tenacity of territorial defense by the dart-poison frog Dendrobates pumilio. Herpetologica, 49(1), 94–99.

Forester, D. C., & Wisnieski, A. (1991). The significance of airborne olfactory cues to the recognition of home area by the dart-poison frog Dendrobates pumilio. Journal of herpetology, 25(4), 502–504.

Francis, A. P., & Currie, D. J. (2003). A globally consistent richness‐climate relationship for angiosperms. The American Naturalist, 161(4), 523–536.

Frost, Darrel R. 2016. Amphibian Species of the World: an Online Reference. Version 6.0 (Date of access). Electronic Database accessible at http://research.amnh.org/herpetology/amphibia/index.html. American Museum of Natural History, New York, USA.

Gardner, E. A., & Graves, B. M. (2005). Responses of resident male Dendrobates pumilio to territory intruders. Journal of herpetology, 39(2), 248–253.

Gaston, K. J. (1998). Species-range size distributions: products of speciation, extinction and transformation. Philosophical Transactions of the Royal Society B: Biological Sciences, 353(1366), 219–230.

Goodman, D. E. (1971). Territorial behavior in a Neotropical frog, Dendrobates granuliferus. Copeia, 1971(2), 365–370.

Graves, B. M. (1999). Diel activity patterns of the sympatric poison dart frogs, Dendrobates auratus and D. pumilio, in Costa Rica. Journal of Herpetology, 33(3), 375–381.

Graves, B. M., Stanley, K. A., & Gardner, E. A. (2005). Correlates of vocal display in a Costa Rican population of strawberry poison-dart frogs, Dendrobates pumilio. Journal of herpetology, 39(1), 101–107.

Haase, A., & Pröhl, H. (2002). Female activity patterns and aggressiveness in the strawberry poison frog Dendrobates pumilio (Anura: Dendrobatidae). Amphibia-Reptilia, 23(2), 129–140.

Han, B. A., Kats, L. B., Pommerening, R. C., Ferrer, R. P., Murry-Ewers, M., & Blaustein, A. R. (2007). Behavioral Avoidance of Ultraviolet-B Radiation by Two Species of Neotropical Poison-Dart Frogs. Biotropica, 39(3), 433–435.

Hawkins, B. A., Field, R., Cornell, H. V., Currie, D. J., Guégan, J. F., Kaufman, D. M., T. Kerr, J. T., Mittelbach, G. G., Oberdorff, T., O'Brien, E. M., Porter, E. E., & Turner, J. R. G. (2003). Energy, water, and broad-scale geographic patterns of species richness. Ecology, 84(12), 3105–3117.

Hortal, J., Triantis, K. A., Meiri, S., Thébault, E., & Sfenthourakis, S. (2009). Island species richness increases with habitat diversity. The American Naturalist, 174(6), E205-E217.

Hua, X., & Wiens, J. J. (2010). Latitudinal variation in speciation mechanisms in frogs. Evolution, 64(2), 429–443.

Janzen, D. H. (1967). Why Mountain Passes are higher in the Tropics. The American Naturalist, 101(919), 233–249.

Kats, L. B., Bucciarelli, G. M., Schlais, D. E., Blaustein, A. R., & Han, B. A. (2012). Ultraviolet radiation influences perch selection by a Neotropical poison-dart frog. PloS one, 7(12), e51364.

Latham, R. E., & Ricklefs, R. E. (1993). Continental comparisons of temperate-zone tree species diversity. In Ricklefs, R. E., & Schluter, D. [eds.] Species diversity in ecological communities: historical and geographical perspectives, 294–314. University of Chicago Press, Chicago, Illinois, USA.

Limerick, S. (1980). Courtship behavior and oviposition of the poison-arrow frog Dendrobates pumilio. Herpetologica, 36(1), 69–71.

Lötters, S., Glaw, F., Köhler, J., & Castro, F. (1999). On the geographic variation of the advertisement call of Dendrobates histrionicus Berthold, 1845 and related forms from north-western South America. Herpetozoa, 12(1/2), 23–38.

McGlone, M. S. (1996). When history matters: scale, time, climate and tree diversity. Global Ecology and Biogeography Letters, 5(6), 309–314.

McVey, M. E., Zahary, R. G., Perry, D., & MacDougal, J. (1981). Territoriality and homing behavior in the poison dart frog (Dendrobates pumilio). Copeia, 1981(1), 1–8.

Mebs, D., Pogoda, W., Batista, A., Ponce, M., Köhler, G., & Kauert, G. (2008). Variability of alkaloid profiles in Oophaga pumilio (Amphibia: Anura:

Dendrobatidae) from western Panama and southern Nicaragua. Salamandra, 44(24), 241–247.

Meuche, I., Brusa, O., Linsenmair, K. E., Keller, A., & Pröhl, H. (2013). Only distance matters–non-choosy females in a poison frog population. Frontiers in zoology, 10(1), 29. DOI: 10.1186/1742-9994-10-29.

Myers, C. W., & Böhme, W. (1996). On the type specimens of two Colombian poison frogs described by Berthold (1845), and their bearing on the locality "Provinz Popayan". American Museum of Natural History, 3185, 1–20.

Myers, C. W., & Daly, J. W. (1976). Preliminary evaluation of skin toxins and vocalizations in taxonomic and evolutionary studies of poison-dart frogs (Dendrobatidae). Bulletin of the American Museum of Natural History, 157(3), 173–262.

Myers, C. W., Daly, J. W., & Martínez, V. (1984). An arboreal poison frog (Dendrobates) from western Panama. American Museum Novitates, (2783), 1–20.

Pröhl, H. (1997). Territorial behaviour of the strawberry poison-dart frog, Dendrobates pumilio. Amphibia-Reptilia, 18(4), 437–442.

Pröhl, H. (2005). Territorial behavior in dendrobatid frogs. Journal of Herpetology, 39(3), 354–365.

Pröhl, H., & Berke, O. (2001). Spatial distributions of male and female strawberry poison frogs and their relation to female reproductive resources. Oecologia, 129(4), 534–542.

Pröhl, H., Hagemann, S., Karsch, J., & Höbel, G. (2007). Geographic variation in male sexual signals in strawberry poison frogs (Dendrobates pumilio). Ethology, 113(9), 825–837.

Pröhl, H., & Hödl, W. (1999). Parental investment, potential reproductive rates, and mating system in the strawberry dart-poison frog, Dendrobates pumilio. Behavioral Ecology and Sociobiology, 46(4), 215–220.

Pröhl, H., & Ostrowski, T. (2011). Behavioural elements reflect phenotypic colour divergence in a poison frog. Evolutionary Ecology, 25(5), 993–1015.

Rohde, K. (1992). Latitudinal gradients in species diversity: the search for the primary cause. Oikos, 65(3), 514–527.

Ricklefs, R. E. (2006). Evolutionary diversification and the origin of the diversity– environment relationship. Ecology, 87(7), S3–S13.

Rojas, B. (2016). Behavioural, ecological, and evolutionary aspects of diversity in frog colour patterns. Biological Reviews. DOI: 10.1111/brv.12269.

Root, T. (1988). Energy constraints on avian distributions and abundances. Ecology, 69(2), 330–339.

Saporito, R. A., Donnelly, M. A., Jain, P., Garraffo, H. M., Spande, T. F., & Daly, J. W. (2007). Spatial and temporal patterns of alkaloid variation in the poison frog Oophaga pumilio in Costa Rica and Panama over 30 years. Toxicon, 50(6), 757– 778.

Saporito, R. A., Garrafo, H. F., Donnelly, M. A., Edwards, A. L., Longino, J. T., & Daly, J. W. (2004). Formicine Ants: An arthropod source for the pumiliotoxin alkaloids of dendrobatid poison frogs. Proceedings of the National Academy of Sciences of the United States of America, 101(21), 8045-8050.

Savage, J.M. (1968). The Dendrobatid frog of Central America. Copeia, 1968(4), 745– 776.

Silverstone, P. A. (1973). Observations on the behavior and ecology of a Colombian poison-arrow frog, the Kõkoé-Pá (Dendrobates histrionicus Berthold). Herpetologica, 29(4), 295–301.

Silverstone, P. A. (1975). A revision of the poison-arrow frogs of the genus Dendrobates Wagler. Natural History, 21, 1–55.

Staudt, K., Ospina, S. M., Mebs, D., & Pröhl, H. (2010). Foraging behaviour and territoriality of the strawberry poison frog (Oophaga pumilio) in dependence of the presence of ants. Amphibia-Reptilia, 31(2), 217–227.

Summers, K., Cronin, T. W., & Kennedy, T. (2003). Variation in spectral reflectance among populations of Dendrobates pumilio, the strawberry poison frog, in the Bocas del Toro Archipelago, Panama. Journal of Biogeography, 30(1), 35–53.

Tews, J., Brose, U., Grimm, V., Tielbörger, K., Wichmann, M. C., Schwager, M., & Jeltsch, F. (2004). Animal species diversity driven by habitat heterogeneity/diversity: the importance of keystone structures. Journal of biogeography, 31(1), 79–92.

Turner, J. R., Gatehouse, C. M., & Corey, C. A. (1987). Does solar energy control organic diversity? Butterflies, moths and the British climate. Oikos, 48(2), 195– 205. van Wijngaarden, R., & Bolaños, F. (1992). Parental care in Dendrobates granuliferus (Anura: Dendrobatidae), with a description of the tadpole. Journal of Herpetology, 26(1), 102–105.

Wiens, J. J., & Donoghue, M. J. (2004). Historical biogeography, ecology and species richness. Trends in Ecology & Evolution, 19(12), 639–644.

Willink, B., Brenes-Mora, E., Bolaños, F., & Pröhl, H. (2013). Not everything is black and white: color and behavioral variation reveal a continuum between cryptic and aposematic strategies in a polymorphic poison frog. Evolution, 67(10), 2783–2794.

Wright, D. H. (1983). Species-energy theory: an extension of species-area theory. Oikos, 41(3), 496–506.

Young, A. M. (1979). Arboreal movement and tadpole-carrying behavior of Dendrobates pumilio Schmidt (Dendrobatidae) in northeastern Costa Rica. Biotropica, 11(3), 238–239.

Río Pataturro, Nicaragua.

COLOUR AND PATTERN VARIATION AND PLEISTOCENE PHYLOGEOGRAPHIC ORIGIN OF THE STRAWBERRY POISON FROG, OOPHAGA PUMILIO, IN NICARAGUA

Diana M. Galindo-Uribe, Javier Sunyer, J. Susanne Hauswaldt, Adolfo Amézquita, Heike Pröhl, & Miguel Vences

Introduction

The Central American strawberry poison frog Oophaga pumilio has become an important model system for studies in ecology and evolutionary biology due to its complex territorial and courtship behaviour (e.g., Donnelly 1989, Pröhl & Hödl 1999, Pröhl & Ostrowski 2011, Saporito et al. 2007, Stynoski 2009), and its intriguing variation in aposematic colour pattern (e.g., Daly & Myers 1967, Summers et al. 1997, Wang & Shaffer 2008, Batista & Köhler 2008, Maan & Cummings 2009). Populations of this species from the Panamanian Bocas del Toro archipelago are extremely polymorphic in colouration and pattern, amongst and in some cases also within populations (Daly & Myers 1967). In contrast, most mainland populations are relatively uniformly coloured, with a mainly red-orange dorsum and at least partly blue limbs (Hagemann & Pröhl 2007).

The phylogeny and phylogeography of O. pumilio are convoluted as well. The genus Oophaga contains 9 species (O. arborea, O. granulifera, O. histrionica, O. lehmanni, O. occultator, O. pumilio, O. speciosa, O. sylvatica, and O. vicentei (Myers et al. 1984, Grant et al. 2006)), and ranges from Colombia to Nicaragua. Molecular phylogenies suggest that O. pumilio does not form a monophyletic group from a mitochondrial perspective (Hagemann & Pröhl 2007, Hauswaldt et al. 2011). The mitochondrial trees typically separate two main lineages within O. pumilio, with uncorrected pairwise divergences among cytochrome b haplotypes of up to 7%. Furthermore, congeneric species (O. speciosa, O. arborea, O. vicentei) are phylogenetically grouped with one of these lineages (Hagemann & Pröhl 2007, Wang & Shaffer 2008, Hauswaldt et al. 2011). These two lineages are roughly distributed in either the northern or the southern part of the range, and microsatellite data

(Hauswaldt et al. 2011) also give a weak signal of a northern (northeastern Costa Rica) and a southern (southeastern Costa Rica and Panama) cluster of populations. However, mitochondrial haplotypes belonging to both groups are also found co- occurring across a wide zone of Panama and Costa Rica. Altogether, this complex phylogeographic pattern suggests an intricate series of various past dispersal and vicariance processes, with allopatric divergence and subsequent admixture, and possible introgression among O. pumilio and other species of Oophaga (Hagemann & Pröhl 2007, Wang & Shaffer 2008, Hauswaldt et al. 2011).

All recent studies on Oophaga pumilio focused on populations from Costa Rica and Panama, omitting populations from the largest portion of the distribution range of the species which extends across much of Nicaragua (Galindo Uribe et al. 2013). Only a few individuals from southernmost Nicaragua (Río San Juan area) were included in general phylogenetic assessments (Grant et al. 2006, Santos & Canatella 2011). At present, knowledge about the Nicaraguan populations is scarce. As far as is known, they have a monomorphic reddish and blue colour pattern (Cope 1874, Noble 1918, Villa 1972, Savage 2002). Given this lack of data, several questions about the intraspecific variation of O. pumilio have remained unanswered: (i) Do all Nicaraguan populations belong to the “northern” mitochondrial lineage, or are additional, strongly different haplotypes found in this part of the range? (ii) Did the Nicaraguan populations originate from a relatively recent northward range expansion? (iii) Do the Nicaraguan populations differ in colour pattern from the populations in northern Costa Rica?

In this study, we aim at addressing these questions by studying newly collected samples of O. pumilio from seven localities across its range in Nicaragua. We analyse the genetic differentiation and phylogeography of these populations based on DNA sequences of the mitochondrial cytochrome b gene and describe their colouration in greater detail.

Materials and methods

Background

Previous knowledge on Oophaga pumilio from Nicaragua dates back to the early 1870s (Belt 1874, Cope 1874). Cope (1874) described Dendrobates ignitus (currently a synonym of O. pumilio) based on specimens collected in April 1873 at “Machuco (=Machuca) on the Río San Juan, Departamento Río San Juan, Nicaragua” (Savage 1973). Belt (1874) recorded this species as being very abundant in the damp forests around Santo Domingo (Chontales Department). He experimented with feeding a young duck with one O. pumilio, who instead of swallowing it, instantly threw it up (Belt 1874). The only other contribution on toxins of O. pumilio from Nicaraguan populations is that of Mebs et al. (2008), who studied alkaloids in populations from Cerro Musún in the central part, and along the Río San Juan and Río Indio in the southeastern portion of the country. The Musún population exhibited a largely different alkaloid profile compared with that of Panamanian specimens, the latter sharing the common alkaloids with those from Río San Juan/Río Indio (Mebs et al. 2008). However, these profiles are of no direct phylogenetic relevance since alkaloids in dendrobatids are of dietary origin (Mebs et al. 2008). Only a few other data are available on the ecology and biology of the species in Nicaragua: Noble (1918) studied the stomach contents of several O. pumilio collected around Cuckra (today in the Atlántico Sur Department), finding mostly small red ants, a single pill bug, and a spider, and Wong et al. (2009) studied heterospecific acoustic interference effects on the calling of adult males O. pumilio at Bartola, Río San Juan Department.

Most of the historical distribution records of this species from Nicaragua still are represented by extant populations, although these are often reduced in size. For example, the populations mentioned by Belt (1874) today persist in a small forest patch close to the Reserva Privada Las Brumas, but are threatened by deforestation. Also the populations referred to by Cope (1874) are still found throughout the southeastern portion of the country. Noble (1918) mentioned populations that are now located in a small forest patch in the Reserva Natural Kahka Creek that was

23 spared the devastating effects of hurricane Felix in 2007. On the other hand, the historical populations around the city of Matagalpa (Matagalpa Department) have not been confirmed since the mid 1960s, and there are other uncertainties regarding the historical distribution of O. pumilio in Nicaragua. As an example, Gaige et al. (1937) did not report O. pumilio from an extensive three-month herpetological collection carried out in 1935 in the area around the Río Escondido and its tributaries in central-eastern Nicaragua (today in the Atlántico Sur Department), an apparently suitable area, while the species is known to occur to the north, south, east and west of this region. Additionally, there have been several questionable records of O. pumilio in Nicaragua from the surroundings of Siuna in the Atlántico Norte Department (Villa 1972) as well as from Managua and Chinandega Departments. It is possible that the voucher specimens for the Atlántico Norte records, as well as several other sites shown in the distribution map of Villa (1972), were lost during the catastrophic earthquake in Managua in 1972, which destroyed all of the alcoholic herpetological specimens housed in the Museo Nacional de Nicaragua (Villa 1981).

Currently, the distribution of O. pumilio in Nicaragua is considerable compared to its entire range. In Nicaragua, O. pumilio occurs across a continuous area in the southeastern corner of the country (Departamento Río San Juan; Köhler 2001, Sunyer et al. 2009) and other remaining lowland forest patches of the Departamento Atlántico Sur. Furthermore, the species is found in a few isolated and threatened forest localities at mid-altitudes in the Departamentos Chontales, Boaco, and Jinotega (Köhler 2001). In Nicaragua, this species is found in association with running waters and forests with different degrees of anthropogenic alteration. Individuals sometimes can be found in high densities, even in grasslands where no substrates for refuge and oviposition are obvious. However, it is unknown how these populations respond to increased habitat degradation and transformation.

Within Nicaragua, O. pumilio is considered a species of medium vulnerability and known to occur in four forest formations and several bioclimatic regions, with a widespread distribution in Lowland Wet and a peripheral distribution in Lowland

Moist, Premontane Wet, and Premontane Moist formations, and can be found from sea level to 960 m a.s.l. (Sunyer & Köhler 2010). According to the IUCN criteria, this species is regarded as a species of Least Concern, although populations are decreasing (Solís et al. 2013), and according to the Convention on the International Trade of Endangered Species (CITES), it has been considered an Appendix II species since 1987 (CITES 2013). Oophaga pumilio is one of the most profitable amphibian species for the pet trade in Nicaragua, and between 1991 and 1996, 95% of the exports of this species were shipped to the USA (Solís et al. 2013). Even though Cerro Kilambé is not included in the IUCN distribution map of the species, this locality was well known to the pet traders in the early 1990s who offered locals one US dollar per live specimen (pers. comm. by local people to J. Sunyer): a large amount in the aftermath of an impoverishing civil war. Nicaragua then fixed export quotas to 1,100 specimens taken from the wild in 1997, 1,200 ranched specimens in 1998, 10,000 ranched specimens in 1999, and 3,450 ranched and captive-bred specimens in 2000 and 2001, respectively (CITES 2013).

Sampling localities and procedures

Specimens of Oophaga pumilio were collected from seven localities in Nicaragua (Figure 1. Table 1) between May and August 2012. In the northernmost distribution area, we sampled two highland populations from the Reservas Naturales Cerro Kilambé (Jinotega region) and Cerro Musún (Matagalpa region), located at higher altitudes in Nicaragua (770 and 572 m a.s.l., respectively). The Kilambé population also represents an extension of the known range and currently is the northernmost known locality of the species. In central Nicaragua, we sampled frogs from the private Reserva Natural Las Brumas at 562 m a.s.l., and from Kahka Creek, a lowland (24 m a.s.l.) population near Laguna de Perlas on the Atlantic coast. In the southern part of Nicaragua, we sampled three populations close to the border with Costa Rica: the Refugio de Vida Silvestre Los Guatuzos close to the Papaturro River, the Reserva Silvestre Privada Refugio Bartola, and the

Río Indio Lodge near the Atlantic coast, at 54, 74 and 14 m a.s.l., respectively. Geographic coordinates were obtained with a Garmin Oregon 550 GPS.

Ventral and dorsal photographs were taken with a digital camera of ten specimens per site. To facilitate body measurements, every frog was placed on a millimetre grid. Colour descriptions were made considering mainly the terminology used by Savage (1968, 2002). Snout–vent length (SVL) measurements were taken based on the photographs using ImageJ 1.46r (Abramoff et al. 2004). Tissue samples were taken by toe-clipping from up to ten specimens per locality and preserved in 96% ethanol.

DNA sequencing

Genomic DNA was extracted from toe clippings using the QIAGEN DNeasy Blood and Tissue kit. Samples and extractions were stored at -20°C. We amplified a fragment of the mitochondrial cytochrome b gene using the primers MVZ 15 (Moritz et al. 1992), and CytbDen1-H (Santos & Cannatella 2011). PCR conditions consisted of an initial 2 min. at 95°C, 35 cycles of 30 sec. each at 95°C, 40 sec. at 42°C, and 60 sec. at 72°C, and a final extension step of 6 min. at 72°C. Amplified fragments were sequenced with the forward primer MVZ15, using dye-terminator chemistry. DNA sequences (784 bp) were subsequently aligned with those from Hauswaldt et al. (2011). All newly determined sequences were deposited in GenBank (accession numbers KF645289–KF645350).

Phylogeography

The Bayesian Information Criterion (BIC) incorporated in jModeltest vs. 2.1.1 (Darriba et al. 2012) was used to determine the best fitting model of nucleotide substitution for two datasets of the cytochrome b sequences. The HKY + Γ model was selected for dataset 1 composed of all haplotypes of O. pumilio as well as of

26 the outgroup species (O. granulifera and O. vicentei), and the TrN + Γ model was selected for the dataset 2 containing sequences of all individuals of O. pumilio.

Dataset 1 was used to reconstruct a phylogenetic tree of haplotypes using Bayesian inference as implemented in Mr. Bayes vs. 3.1.2. (Ronquist & Huelsenbeck 2003) using an unpartitioned approach. Using the default settings of the software, we ran two chains for 10 million generations each, sampling every 100th tree, and checked for convergence. The first 25% of the sampled trees were discarded as burn-in. Bayesian posterior probabilities are based on a majority-rule consensus tree.

Dataset 2 was used with the software Beast 1.7. (Drummond et al. 2012) to simultaneously estimate tree topology and geographic locality of nodes, using the geographic coordinates of the samples as traits and combined this with the phylogeographic method by Lemey et al. (2010). Specifically, we used the Relaxed Random Walk model to estimate the geographic origin of the O. pumilio clade and possible dispersal routes throughout Panama, Costa Rica and Nicaragua. We set the coalescent prior to constant size and used an uncorrelated relaxed clock model (Drummond et al. 2006), the TrN + Γ model of substitution, codon partitioning (1+2, 3), and set the mutation rate to 0.01. We ran a MCMC chain 3 × 108 long, sampling every 30.000th. We ensured that effective sample sizes (ESS) were higher than 200 for all parameters with Tracer 1.5. The first 20% of the trees were discarded as burn-ins. The resulting tree was summarized with Tree Annotator 1.7.x. To generate the KML file that could be plotted with Google Earth (www.googleearth.com), we used the software SPREAD (Bielejec et al. 2011) and a posterior probability setting of 0.05.

Results

Morphology and colour pattern

No substantial variation in colour pattern and body size was detected among Nicaraguan O. pumilio populations. All share the “blue jeans” dorsal colour pattern

27 seen in Costa Rican frogs, with blue or black limbs (Figs 2–3, Table 2). The colour of the body differs slightly amongst populations and ranges from dull red to red- orange. The colour of the limbs varies from blue to grey-blue or black. The presence and amount of black on the dorsum can differ as well and take on the shape of small spots, wavy lines, or reticulated markings. The venter is orange or red, with or without light blue (Figure 3). Specimens from Kilambé and Guatuzos have a particularly bright red body colour and blue limbs with reduced black pattern on the dorsum, and those from Guatuzos have very small light blue markings on the venter. Individuals from Musún have a red–orange dorsum with a black pattern. Specimens from Kahka Creek have a red dorsum with some blue on the cloaca and limbs, and a reduced amount of red on the venter and flanks. Specimens from Bartola have a dull red-coloured dorsum with light blue limbs, and are mainly red on the ventral side. At Las Brumas, the frogs are either red or red-orange with blue legs, and always exhibit a black dorsal pattern. At Río Indio Lodge, two main morphs can be distinguished: red-orange and blue-legged frogs with a black dorsal pattern, and orange-red frogs with or without black legs and a very reduced or even absent dorsal pattern (Figure 2, Table 1). Populations varied in mean body size between 19.7 and 22.7 mm, with the largest frogs occurring in the Río San Juan populations from Guatuzos, and the smallest frogs being found in Bartola (Table 2).

Phylogeography

A total of 62 cytochrome b sequences were obtained from Nicaraguan O. pumilio and compared with previously determined sequences from Costa Rica and Panama (Hauswaldt et al. 2011) to yield a total alignment of sequences from 321 individuals. Because the Costa Rica and Panama sequences were shorter than the newly determined ones from Nicaragua, the alignment used for analysis was only 558 bp in length. 88 haplotypes could be distinguished (Figure 4). The phylogenetic analysis (Figure 4) agrees with previous findings in separating the haplotypes into a northern and a southern clade, and in placing sequences of O.

28 vicentei with the northern clade of O. pumilio. A total of nine haplotypes (numbers as in Figure 4), all grouped in the northern clade, were found among the Nicaraguan populations. Five of these (H3–H7) had not been observed before and were exclusive to Nicaraguan populations. The other four haplotypes (H1, H2, H8, H9) were shared among Nicaraguan and Costa Rican populations. Haplotype sharing with Costa Rican populations was not limited to the populations from Río San Juan (geographically closest to Costa Rica), but was observed in northern Nicaraguan populations as well: e.g., haplotype H1, was found at Musún, Bartola, Río Indio Lodge, and in the Costa Rican population of Tortuguero (Table 1).

The phylogeographic analyses strongly support a centre of origin of O. pumilio in southeastern Costa Rica from where the species spread north and southwards. Dispersal to the island of Escudo de Veraguas likely co-occurred with the spread into Nicaragua (Figure 5). The origin of the populations in Nicaragua is reconstructed as a recent Pleistocene dispersal at approximately 200,000 years before present (Figure 5).

Discussion

Understanding the phylogeographic history of O. pumilio is crucial for making inferences on the evolution of the extraordinary polychromatism of the species. However, given the lack of molecular data from Nicaragua, i.e., from more than half of the species’ range, it used to be impossible to reliably test hypotheses on the geographic origin and directions of range expansion of this species. The data presented herein partly fill this geographical sampling gap, and confirm that the Nicaraguan populations are not strongly differentiated, but rather the result of a recent range expansion from Costa Rica. Our explicit phylogeographic analyses place the origin of the species in southern Costa Rica, and support a radial expansion of this species northwards into Nicaragua and southwards towards the Bocas del Toro archipelago during the Late Pleistocene. However, an important limitation of this analysis is its sole reliance on mitochondrial DNA. If either the

29 northern or southern lineage of O. pumilio were the result of an ancient hybridisation and introgression process with another species of Oophaga, then some of the detailed reconstructions herein might not fully reflect the true evolutionary history of the species. Despite this limitation, given that the northern Nicaraguan populations of O. pumilio share haplotypes with Costa Rican populations, and that Nicaraguan populations overall do not differ strongly in colouration from the ones in Costa Rica, we find it very plausible that Nicaraguan populations indeed originate from a recent range expansion northwards.

The relationships among mitochondrial lineages and species of Oophaga also reveal the need for taxonomic revision. Daly & Myers (1967) and Myers & Daly (1976) suggested recognizing O. pumilio as a single polymorphic species, and this view has either been accepted or controversially discussed (Summers et al. 1997, Hagemann & Pröhl 2007, Batista & Köhler 2008, Hauswaldt et al. 2011). The taxonomic integrity of O. pumilio as a single entity remains uncertain, especially relative to several other species in the genus distributed in Panama (O. arborea, O. speciosa, and O. vicentei). Alternatively, some (or all) of these other taxa could in fact represent morphs of O. pumilio and thus junior synonyms, or some populations considered as O. pumilio could in fact rather belong to one of those closely related species. The strawberry poison frog was originally described as Dendrobates pumilio Schmidt, 1857 from the Bocas del Toro province in Panama. Daly & Myers (1967) applied this name to the small-sized poison frogs occurring from Nicaragua to northwestern Panama, and Grant et al. (2006) suggested classifying this and related species, separate from Dendrobates, in the genus Oophaga. Three other names are currently considered to be junior synonyms of O. pumilio: Dendrobates galindoi Trapido, 1953, described likewise from Bocas del Toro, Dendrobates typographus Keferstein, 1867, and Dendrobates ignitus Cope, 1874, with their type localities in Costa Rica and Nicaragua, respectively (see Savage 1968, Batista & Köhler 2008). However, given the wide co-occurrence of the two haplotype lineages in O. pumilio (Hauswaldt et al. 2011), we consider it to be rather unlikely that they represent distinct and independent evolutionary lineages, and that any of these junior synonyms would require elevation to species level.

Inversely, as mentioned above, the identities, status and distribution ranges, especially of O. arborea and O. speciosa, but also of O. vicentei, require detailed revision. A first step that has to be taken towards such a revision is a comprehensive mitochondrial phylogeny of all Oophaga species, using longer DNA sequences, in order to obtain a strongly supported phylogenetic tree depicting the mitochondrial relationships among these frogs. Using phylogenomic methods such as RAD sequencing (e.g., Rubin et al. 2012) would then allow to reconstruct the true evolutionary relationships among species and populations, and to understand if, when, and where phenomena of mitochondrial introgression might have taken place in this genus.

Figure 1. Location of the Nicaraguan populations of O. pumilio included in this study, and of the Costa Rican and Panamanian populations previously studied by Hauswaldt et al. (2011). Numbers correspond to localities listed in Table 1.

Figure 2. Oophaga pumilio colour patterns observed at Nicaraguan sites. 1 – Kilambé; 2 – Musún; 3 – Kahka Creek (Photo taken from www.monarchzman.deviantart.com); 4 – Las Brumas; 5 – Guatuzos; 6 – Bartola; 7 – Río Indio Lodge.

Figure 3. Representation of the terminology used for describing the dorsal and ventral colours and patterns of Oophaga pumilio populations from Nicaragua.

Figure 4. MrBayes tree based on an unpartitioned dataset of 529 bp cyt-b sequences (all haplotypes are of equal length), using the HKY+G model and O. granulifera as an outgroup. Red branches and numbers indicate haplotypes found in Nicaragua. Grey scales in boxes represent the relative frequency of individuals assigned to each haplotype, from the region of North Nicaragua (NN), Central Nicaragua (CN), South Nicaragua (SN), Costa Rica (CR), and Panama (PA).

Figure 5. Phylogeographic reconstruction for Central American Oophaga pumilio populations using the Relaxed Random Walk Model. Green polygons and red lines indicate early events, whereas black lines and polygons more recent events. The red marking points out the Isla Bastimentos in Panama. MA – million years ago.

Table 1. Locality numbers and names, coordinates, altitudes, and numbers of individuals of O. pumilio sequenced for the cytochrome b fragment, from Nicaraguan (NIC) populations (this study), and from Costa Rican (CR) and Panamanian (P) populations previously studied by Hauswaldt et al. (2011). Populations and numbers of individuals (in parenthesis) assigned to the 93 haplotypes (H) found in 321 individuals of O. pumilio, two of Oophaga vicentei (O.v.), and one of Oophaga granulifera (O.g.).

Locality Locality Province Coordinates [lat. (°N), long. (°W)] Cyt b (N) Haplotypes number 1 NIC: Kilambé Jinotega 13.63 -85.72 8 H5 (7), H6 (1) 2 NIC: Musún Matagalpa 12.96 -85.23 8 H1 (7), H5 (1) 3 NIC: Kahka Creek Atlántico Sur 12.67 -83.72 7 H5 (7) 4 NIC: Las Brumas Chontales 12.28 -85.08 9 H6 (9) 5 NIC: Guatuzos Río San Juan 11.01 -85.06 10 H9 (10) 6 NIC: Bartola Río San Juan 10.97 -84.34 10 H1 (1), H2 (2), H7 (2), H8 (5) 7 NIC: Río Indio Lodge Río San Juan 10.93 -83.73 10 H1 (1), H2 (3), H3 (1), H4 (5) 8 CR: Upala Alajuela 10.91 -85.05 7 H9 (4), H19 (3) 9 CR: Caño Negro Alajuela 10.87 -84.78 11 H20 (5), H21 (1), H22 (2), H23 (3) 10 CR: Tortuguero Limón 10.61 -83.53 9 H1 (4), H8 (2), H10 (3) H2 (2), H8 (1), H33 (1), H34 (2), 11 CR: La Selva Heredia 10.43 -84.00 11 H35 (3), H36 (1), H37 (1) 12 CR: Pueblo Nuevo 10.32 -83.59 11 H8 (2), H24 (7), H25 (1), H26 (1) 13 CR: Guápiles Limón 10.19 -83.82 11 H15 (1), H27 (2), H28 (8) H14 (1), H15 (1), H16 (6), H17 (2), 14 CR: Siquirres Limón 10.10 -83.52 12 H18 (2) H16 (4), H27 (1), H38 (2), H39 (1), 15 CR: Río Reventazón Cartago 10.09 -83.56 9 H40 (1) H8 (1), H27 (1), H58 (6), H89 (4), 16 CR: Hitoy Cerere Limón 9.67 -83.09 14 H90 (2) 17 CR: Puerto Viejo de Talamanca Limón 9.65 -82.76 12 H11 (5), H12 (5), H13 (2) 18 CR: Bribri Limón 9.65 -82.88 9 H29 (3), H30 (3), H31 (1), H32 (2) H41 (8), H42 (1), H43 (1), H44 (1), 19 P: Colón Bocas del Toro 9.39 -82.24 12 H45 (1) 20 P: Solarte Bocas del Toro 9.33 -82.22 12 H42 (1), H50 (11) 21 P: Bastimentos Bocas del Toro 9.30 -82.14 10 H41 (3), H42 (1), H48 (5), H49 (1) 22 P: Almirante Bocas del Toro 9.29 -82.39 9 H50 (3), H51 (2), H52 (3), H53 (1) 23 P: San Cristobal Bocas del Toro 9.27 -82.29 10 H8 (1), H42 (2), H83 (4), H84 (3) H78 (2), H79 (4), H80 (2), H81 (1), 24 P: Pastores Bocas del Toro 9.24 -82.35 10 H82 (1) H46 (1), H50 (1), H69 (7), H70 (1), 25 P: Popa Bocas del Toro 9.22 -82.13 19 H71 (4), H72 (1), H73 (1), H74 (1), H75 (1), H76 (1) 26 P: Tierra Oscura Bocas del Toro 9.18 -82.26 8 H54 (5), H56 (2), H77 (1) 27 P: Punta Alegre Bocas del Toro 9.16 -81.91 8 H85 (3), H86 (3), H87 (1), H88 (1) H50 (7), H59 (1), H60 (1), H61 (1), 28 P: Cayo de Agua Bocas del Toro 9.16 -82.05 12 H62 (1), H63 (1) H53 (3), H54 (2), H55 (1), H56 (4), 29 P: Cauchero Bocas del Toro 9.16 -82.25 12 H57 (1), H58 (1) 30 P: Loma Partida Bocas del Toro 9.14 -82.17 10 H46 (9), H47 (1) H64 (7), H65 (1), H66 (1), H67 (1), 31 P: Escudo de Veraguas Escudo de Veraguas 9.10 -81.55 11 H68 (1) O.v.1 Veraguas 8.51 -81.08 1 H91 (1) O.v.2 Coclé 8.63 -80.58 1 H92 (1) O.g. 1 H93 (1)

Table 2. Dorsal and ventral colours and patterns described for Oophaga pumilio populations in Nicaragua. Colour pattern descriptions and snout–vent length (SVL) measurements (± standard deviation) were taken for 10 individuals.

Dorsal body Ventral Ventral body Population Dorsal limb colour Dorsal body pattern Dorsal leg pattern Ventral body colour SVL colour limb colour and limb pattern

None or very few Blue with or without red Mainly red with light Kilambé Red scattered small spots Reticulated markings Blue Diffused flecks 22.4±0.6 markings on one thigh blue or wavy flecks Blue, with or without red Small spots or wavy Reticulated markings or Mainly orange-red Diffused small Musún Red-orange Blue 22.7±1.1 markings on one thigh flecks diffused spots with light blue spots or flecks Kahka Red and blue near Small spots or wavy Diffused reticulated Mainly light blue with Blue Blue Diffused flecks 22.2±0.8 Creek cloaca flecks markings or small spots red in the flanks Blue, with or without red Small spots or wavy Mainly well- de fined Mainly orange-red Las Brumas Red or orange red Blue Diffused flecks 20.9±0.8 markings on thighs flecks reticulated markings with light blue None or very few Mainly orange-red None or a few Guatuzos Red Blue scattered small spots Small spots with light blue near Blue 23.4±0.7 diffused flecks or wavy flecks cloaca and chin Orange-red with light Grey-blue with or without Very few scattered Bartola Dull red Diffused spots blue or red with light Light blue Diffused flecks 19.7±0.8 red markings on thighs small spots blue near cloaca Río Indio Red-orange or Red-orange, with or without None or small spots or Red, black, Small spots or stripes Orange red Diffused flecks 21.8±1.1 Lodge orange-red black or blue at tips wavy flecks and blue

Literature cited

Abràmoff, M. D., Magalhães, P. J., & Ram, S. J. (2004). Image processing with ImageJ. Biophotonics international, 11(7), 36–42.

Batista, A., & Köhler, G. (2008). Variation in Oophaga pumilio (Amphibia: Anura: Dendrobatidae) in western Panama. Salamandra, 44(4), 225–234.

Belt, T. (1874). The naturalist in Nicaragua.403 pp. London, E. Bumpus.

Bielejec, F., Rambaut, A., Suchard, M. A., & Lemey, P. (2011). SPREAD: spatial phylogenetic reconstruction of evolutionary dynamics. Bioinformatics, 27(20), 2910–2912.

CITES (2013). CITES Species database. http://www.cites.org/eng/resources/species.html/. Accessed 8 September 2013.

Cope, E. D. (1874). Description of some species of reptiles obtained by Dr. John F. Bransford, Assistant Surgeon United States Navy, while attached to the Nicaraguan Surveying Expedition in 1873. Proceedings of the Academy of Natural Sciences of Philadelphia, 64–72.

Daly, J. W., & Myers, C. W. (1967). Toxicity of Panamanian poison frogs (Dendrobates): some biological and chemical aspects. Science, 156(3777), 970– 973.

Darriba, D., Taboada, G. L., Doallo, R., & Posada, D. (2012). jModelTest 2: more models, new heuristics and parallel computing. Nature methods, 9(8), 772.

Donnelly, M. A. (1989). Effects of reproductive resource supplementation on space- use patterns in Dendrobates pumilio. Oecologia, 81(2), 212–218.

Drummond, A. J., Ho, S. Y., Phillips, M. J., & Rambaut, A. (2006). Relaxed phylogenetics and dating with confidence. PLoS Biol, 4(5), e88.

Drummond, A. J., Suchard, M. A., Xie, D., & Rambaut, A. (2012). Bayesian phylogenetics with BEAUti and the BEAST 1.7. Molecular biology and evolution, 29(8), 1969–1973.

Gaige, H. T., Hartweg, N., & Stuart, L. C. (1937). Notes on a collection of amphibians and reptiles from eastern Nicaragua. Occasional Papers Museum of Zoology University of Michigan, (357), 1–18.

Galindo Uribe, D., Sunyer, J., Amézquita A., Köhler, G., Pröhl, H., & Vences. M. (2013). Auf der Suche nach Erdbeerfröschen in Nicaragua. Terraria/Elaphe, (42), 48–53.

Grant, T., Frost, D. R., Caldwell, J. P., Gagliardo, R. O. N., Haddad, C. F., Kok, P. J., Means, B., Noonan, B. P., Schargel, W. E., & Wheeler, W. C. (2006).

Phylogenetic systematics of dart-poison frogs and their relatives (Amphibia: Athesphatanura: Dendrobatidae). Bulletin of the American Museum of natural History, (299), 1–262.

Hagemann, S., & Pröhl, H. (2007). Mitochondrial paraphyly in a polymorphic poison frog species (Dendrobatidae; D. pumilio). Molecular phylogenetics and evolution, 45(2), 740–747.

Hauswaldt, J. S., Ludewig, A. K., Vences, M., & Pröhl, H. (2011). Widespread co- occurrence of divergent mitochondrial haplotype lineages in a Central American species of poison frog (Oophaga pumilio). Journal of Biogeography, 38(4), 711–726.

Keferstein, W. (1867). Ueber einige neue oder seltene Batrachier aus Australien und dem tropischen Amerika. Nachrichten von der Königl. Gesellschaft der Wissenschaften und der Georg-Augusts-Universität zu Göttingen, 1867, 341–361.

Köhler, G. (2001). Anfibios y Reptiles de Nicaragua. Herpeton, Verlag Elke Köhler, Offrenbach, Alemania. 208 págs.

Lemey, P., Rambaut, A., Welch, J. J., & Suchard, M. A. (2010). Phylogeography takes a relaxed random walk in continuous space and time. Molecular biology and evolution, 27(8), 1877–1885.

Maan, M. E., & Cummings, M. E. (2009). Sexual dimorphism and directional sexual selection on aposematic signals in a poison frog. Proceedings of the National Academy of Sciences, 106(45), 19072–19077.

Mebs, D., Pogoda, W., Batista, A., Ponce, M., Köhler, G., & Kauert, G. (2008). Variability of alkaloid profiles in Oophaga pumilio (Amphibia: Anura: Dendrobatidae) from western Panama and southern Nicaragua. Salamandra, 44(24), 241–247.

Moritz, C., Schneider, C. J., & Wake, D. B. (1992). Evolutionary relationships within the Ensatina eschscholtzii complex confirm the ring species interpretation. Systematic biology, 41(3), 273–291.

Myers, C. W., Daly, J. W., & Martinéz, V. (1984). An arboreal poison frog (Dendrobates) from Western Panama. American Museum Novitates, (2783), 1– 20.

Noble, G. K. (1918). The amphibians collected by the American Museum expeditions to Nicaragua in 1916. Bulletin of the American Museum of Natural History, 38(10), 311–347.

Pröhl, H., & Ostrowski, T. (2011). Behavioural elements reflect phenotypic colour divergence in a poison frog. Evolutionary Ecology, 25(5), 993–1015.

Ronquist, F., & Huelsenbeck, J. P. (2003). MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics, 19(12), 1572–1574.

Rubin, B. E., Ree, R. H., & Moreau, C. S. (2012). Inferring phylogenies from RAD sequence data. PLoS One, 7(4), e33394.

Santos, J. C., & Cannatella, D. C. (2011). Phenotypic integration emerges from aposematism and scale in poison frogs. Proceedings of the national academy of sciences, 108(15), 6175–6180.

Savage, J. M. (1968). The dendrobatid frogs of Central America. Copeia, 1968(4), 745–776.

Savage, J. M. (1973). Herpetological collections made by Dr. John F. Bransford, Assistant Surgeon, USN during the Nicaragua and Panama Canal surveys (1872–1885). Journal of Herpetology, 7(1), 35–38.

Savage, J. M. (2002). The amphibians and reptiles of Costa Rica: a herpetofauna between two continents, between two seas. University of Chicago press.

Schmidt, D. (1857). Diagnosen neuer Frösche des zoologischen Cabinets zu Krakau. Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften in Wien, 24, 10– 15.

Solís, F., R. Ibáñez, C. Jaramillo, G. Chaves, J. Savage, G. Köhler & N. Cox (2013): Oophaga pumilio. In: IUCN 2013. IUCN Red List of Threatened Species. Version 2013.1. http://www.iucnredlist.org/. Accessed 01 September 2013.

Stynoski, J. L. (2009). Discrimination of offspring by indirect recognition in an egg- feeding dendrobatid frog, Oophaga pumilio. Animal Behaviour, 78(6), 1351– 1356.

Summers, K., Bermingham, E., Weigt, L., McCafferty, S., & Dahistrom, L. (1997). Phenotypic and genetic divergence in three species of dart-poison frogs with contrasting parental behavior. Journal of Heredity, 88(1), 8–13.

Sunyer, J., & Köhler, G. (2010). Conservation status of the herpetofauna of Nicaragua. Pp. 488–509 In Wilson, L. D., Townsend, J. H., & Johnson J. D. (Eds.), Conservation of Mesoamerican Amphibians and Reptiles. Eagle Mountain Publishing, LC, Eagle Mountain, Utah.

Sunyer, J., Páiz, G., Dehling, D. M., & Köhler, G. (2009). A collection of amphibians from Río San Juan, southeastern Nicaragua. Herpetology Notes, 2, 189–202.

Trapido, H. (1953): A new frog from Panama, Dendrobates galindoi. Fieldiana Zoology, 34(15), 181–187.

Villa, J. D. (1972). Anfibios de Nicaragua. Instituto Geografico Nacional and Banco Central de Nicaragua, Nicaragua.

Villa, J. D. (1981). Three albino snakes. Herpetological Review, 12(3), 81.

Wang, I. J., & Shaffer, H. B. (2008). Rapid color evolution in an aposematic species: a phylogenetic analysis of color variation in the strikingly polymorphic strawberry poison-dart frog. Evolution, 62(11), 2742–2759.

Wong, S., Parada, H., & Narins, P. M. (2009). Heterospecific acoustic interference: effects on calling in the frog Oophaga pumilio in Nicaragua. Biotropica, 41(1), 74–80.

Río San Juan, Nicaragua.

THE ROLE OF CLIMATIC SEASONALITY AND UNPREDICTABILITY IN SHAPING PATTERNS OF GEOGRAPHIC VARIATION IN WATER BALANCE TRAITS OF DENDROBATES AND OOPHAGA SPECIES

Diana M. Galindo-Uribe, Javier Sunyer, Jose Alfredo Hernández, Abel Batista, Andrés Merino, Heike Pröhl, Miguel Vences, Santiago Ron, Catalina González, & Adolfo Amézquita

Introduction

One of the main focuses in evolutionary ecology is to disentangle the mechanisms underlying geographic variation of phenotypic traits in response to spatial and temporal climatic gradients. Variation in body size has long been studied as a relevant trait in the interaction of organisms with the environment (Chown & Gaston 2010). Under selective scenarios, intraspecific variation in body size may result from phenotypic plasticity or adaptation in response to selective pressures imposed by local climatic regimes (Phillimore et al. 2012). Body size is strongly correlated with survival through its pleiotropic connection with temperature regulation and water body content, as well as with behavioral and reproductive processes. Along altitudinal and latitudinal gradients, the complex interaction between climatic conditions (temperature, water availability, and seasonality) and biotic factors (predation, competitors, resource availability, and quality) can shape organism’s life history traits as growing rates, activity times and thereby adult body size (reviewed in Morrison & Hero 2003, Schäuble 2004, Olalla-Tárraga & Rodríguez 2007, Stillwell 2010, Chown & Gaston 2010, Huston & Wolverton 2011). Lastly, body size variation along time can reflect an organism’s response to extreme or progressive climate changes (Gardner et al. 2011, Sheridan & Bickford 2011).

The Bergmann´s rule has been one of the most studied and perhaps most controversial eco-physiological rules in the literature. Originally formulated for among-species comparisons of endotherms, it describes an increase in body size with an increase in latitude, arguing that larger bodies conserve heat better in

44 cooler climates due to a lower surface-to-volume ratio (Olalla-Tárraga & Rodríguez 2007). However, evidence contradicting the Bergmann's rule has been found for specific geographic contexts (Olalla-Tárraga & Rodríguez et al. 2007, Gutiérrez-Pinto et al. 2014), taxonomic groups and species (Olalla-Tárraga & Rodríguez et al. 2007, Boaratti & Rodrigues da Silva 2015), mainly across species of ectothermic animals (Olalla-Tárraga et al. 2009). Also, other processes have been invoked to explain latitudinal variation in body size (reviewed in, Stillwell 2010, Muñoz et al. 2014), namely the regulation of evaporative water loss evaporation (e.g., Le Lagadec et al. 1998, Olalla-Tárraga et al. 2009), as well as other climatic variables apart from mean temperatures: potential evapotranspiration (e.g., Olalla- Tárraga & Rodríguez et al. 2007), seasonality or the length of activity season (e.g., Titon & Gomes 2015), and precipitation or water availability (e.g., Nevo 1973, Olalla-Tárraga et al. 2009, Brandt & Navas 2013, Titon & Gomes 2015).

Within ectotherms, amphibians have permeable skins, used for respiratory and thermoregulatory processes while they must regulate their body water contents (Stinner & Shoemaker 1987, Tracy & Christian 2005, Tracy et al. 2010). This trait alone might restrict amphibians to environments with higher water availability (Adolph 1933, Thorson 1955), lead to the evolution of adaptive mechanisms to reduce water loss in environments with lower water availability (see de Andrade & Abe 1997, Young et al. 2005, Tracy et al. 2010, Wygoda et al. 2011 and references therein), or select for larger amphibians with lower EWL due to a lower surface to volume ratio (Schmid 1965, Nevo 1973, Rogowitz et al. 1999, Olalla-Tárraga et al. 2009).

A relationship between climatic gradients and phenotypic variation in amphibians has been studied mainly within geographically widespread species along latitudinal, longitudinal and altitudinal gradients (e.g., Nevo 1973). Most comparative studies at the interspecific levels have been conducted among species with different levels of terrestriality (e.g., Schmid 1965, Thorson & Svihla 1943, Thorson 1955), different water availability in their habitats (arboreal, terrestrial and aquatic; Schmid 1965, Young et al. 2005, Tracy & Christian 2005) or contrasting

45 habitats (xeric vs. wet areas; e.g., de Andrade & Abe 1997, Titon & Ribeiro 2015, Prates & Navas 2009). Much less is known about this relationship within more subtle climatic gradients such as those present within the tropical regions (Graves 1991, Olalla-Tárraga et al. 2009, Boaratti & Rodrigues da Silva 2015) that may also contain useful biological information (Tieleman et al. 2003). We also know less about the effect of climate change on this relationship in tropical regions, where it has been proposed that species could be more vulnerable to such changes (Huey et al. 2012).

Neotropical frogs belonging to the genus Dendrobates and Oophaga (Dendrobatidae) exhibit large intra- an interspecific variation in phenotypic traits, including body size (Daly & Myers 1967, Savage 1968, Silverstone 1975, Myers & Daly 1976). Previous studies hypothesized a relationship between body size variation within O. histrionica and O. sylvatica and concomitant variation in temperature, humidity and seasonality gradients (Myers & Daly 1976). The contribution of climate on divergence processes within the genus Oophaga has been also studied (Posso-Terranova & Andrés 2016). However, most research to date focused on understanding variation in other phentotypic traits such as toxicity, coloration, and advertisement calls (e.g., Pröhl et al. 2007, Rudh et al. 2007, Brown et al. 2010). The role of body size as a potential adaptation to local climatic conditions has not been tested. The species in these general are diurnal, terrestrial and highly territorial (Silverstone 1975, Graves 1999, Goodman 1971, McVey et al. 1981, Forester et al. 1993), which render them more susceptible to desiccation, in the absence of avoidance or compensatory mechanisms (e.g., Thorson & Svihla 1943, Spotila & Berman 1976, Olalla-Tárraga et al. 2009). Therefore, these frogs represent excellent candidates to study the relationship between climatic gradients and phenotypic variation.

In this context, we expected to find a broad-scale relationship between the interspecific variation in phenotypic traits related to the regulation of water loss and the climatic conditions. Specifically, we hypothesized an increase in body size and/or a reduction in the rates of evaporative water loss as potential adaptations to

46 drier or more seasonal habitats (Schmid 1965, Nevo 1973, Olalla-Tárraga et al. 2009). We extend our prediction to the past by including as predictors variables the climatic rate of change since the last interglacial period (~120,000 - 140,000 years ago, Otto-Bliesner et al. 2006). Our research questions were (1) do Dendrobates and Oophaga species and populations differ in their body weight and rates of evaporative water loss?; if so, (2) are those differences correlated with the spatial and temporal variation in climatic conditions? Because there were variable degrees of phylogentic affinities among the lineages we studied, and closely related lineages tend to be more similar than more distant species (Felsesnstein 1985), we also explored the contribution of phylogenetic signal in our data.

Materials and methods

Study group

The frog genera Dendrobates and Oophaga belong to the Neotropical family Dendrobatidae (Grant et al. 2006). They exhibit complex courtship (Limerick 1980, Crump 1972, Silverstone 1973, Wells 1978) and parental care (Silverstone 1975, van Wijngaarden & Bolaños 1992, Summers et al. 1999, Wells 1978). They are mostly found on the forest floor, where they get their food and oviposition sites, and eventually along trees while they (particularly in Oophaga) transport the tadpoles to temporal water bodies such as bromeliads (Silverstone 1975, Wells 1978, Young 1979, Pröhl & Berke 2001). We studied populations of Dendrobates auratus, D. truncatus, Oophaga granulifera, O. histrionica, O. pumilio North and South clades (according with Hagemann & Pröhl 2007 and Hauswaldt et al. 2011), O. sylvatica, and O. vicentei. Their distribution extends from Nicaragua to Ecuador where macroclimatic gradients in temperature seasonality and precipitation can be evidenced (Figure 1). Species of both genera occur in premontane tropical wet and in tropical cloud forests, as well as in well preserved forests, in transformed secondary forests and in crop plantations (Lötters et al. 2007). Dendrobates truncatus can be found even in dry forests of Colombia. Within Oophaga, some

47 species such as O. histrionica occur in areas with comparatively high and constant precipitations, and reproduce throughout the year, while other species experience more seasonal climates and reproduce only during the rainy season (Lötters et al. 2007): e.g., O. pumilio in the northern limit of its distribution and O. granulifera as well as O. sylvatica in the southern limit of distribution. Most sampled individuals were found on the floor or higher perches (mainly O. vicentei) with forest coverage, near clear running waters. Only individuals captured at Kilambé and Chontales (Nicaragua) were found near the forest and running water, but in areas that had recently been partially cleaned for coffee plantations purposes.

Evaporative water loss

We estimated evaporative water loss on 199 frogs of eight major lineages at 24 localities that differ in climatic conditions (Table 1, Figure 2). Briefly, we measured weight loss of frogs kept within a hand-made portable wind tunnel: a transparent plastic cylinder, 9 cm in diameter and length, covered at the two extremes with plastic mesh. Wind was created from one end of the tunnel with a small ventilator, and kept at 6 m/s with a dimmer. Before entering the tunnel, frogs were placed on a moist paper towel during 60 min. Then, the water excess was removed from the skin with towel paper and the frog entered the tunnel.

The combined weight of the frog and the wind tunnel was recorded with an Ohaus Scout Balance SC4010 at 0.01 g. The tunnel weight was measured at the beginning of each experiment and subtracted from the combined weight, which was recorded every ten minutes, in order to estimate the weight loss by frogs. Because frog posture and behavior might affect water loss (Heatwole et al. 1969), we estimated an ordinal activity score by observing and noting every two minutes whether the frog was: (1) maintaining a conservancy water posture (crouching, feet and hands mostly covered by the body); (2) sitting without movement; (3) walking through the tunnel; and (4) standing up in front of the ventilator, with both hands and feet on the mesh. Each rank (1–4) was used to estimate weighted average

48 activity of the frog throughout the experiment. We estimated the vital limit of the frog by disturbing it with a stick and examining its ability to recover a straight position; we also finished the experiment at the very first sign of lack of activity or skin darking (Thorson & Svihla 1943, Schmid 1965). Then, we recorded the last frog’s weight and immediately placed in in a plastic box with wet towels in order to allow them to recover their body water content. Once fully recovered, each frog was photographed both dorsally and ventrally in a straight position against a millimetric paper; body size (snout-to-urostyle length) was measured on Image J (Schneider et al. 2012).

To describe frog’s physiological performance we used: (1) the total rate of evaporative water loss as the final-initial weight over the time until loss of reaction (TEWL, g x h-1); (2) the rate of water loss by evapotranspiration per surface unit (EWL g x cm-2 x h-1); and (3) the tolerance to desiccation (%WL) as the difference between the initial and final weight, divided by the initial weight, and multiplied by 100 (Schmid 1965, de Andrade & Abe 1997). No frog urinated during the course of the experiments. As necessary covariates, we measured the temperature (0.1°C) and relative humidity (RH %) of the air entering (inflow) and leaving (outflow) the wind tunnel, at the beginning and at the end of each experiment, using a thermohygrometer (Extech RH101). We also recorded frog’s dorsum temperature at the end of the experiment with an Oakton InfraPro 1 infrared thermometer.

Analysis

We summarized the matrix of covariates with principal component analysis and Varimax rotation, and then tested the effect of non-redundant covariates on water loss estimates. Where an effect was statistically evident, we removed it by calculating residuals of the corresponding linear regression and using them as dependent variables in subsequent analyses.

Cross-species and phylogeneticanalyses

To test for differences among species and populations in weight, rTEWL, the rEWL and the r% of Weight loss, we ran linear mixed effect models, with species as fixed factor and populations as random factors nested within species; “lme4” (Bates et al. 2014), was used to build the models and “car” package to obtain significance p- values (Fox & Weisberg 2011) in R software (R Core Team 2015). To calculate the proportion of variance explained by the species and populations as factors (Nakagawa & Schielzeth 2013) we calculated a marginal (for the fixed factor) and a conditional R2 (for both fixed and random factors) with the “piecewiseSEM” package (Lefcheck 2015).

Because of the hierarchical nature of our data and the shared history between related species, our data is prone to the statistical problems caused by the non- independence of observations. We thus explored the contribution of phylogenetic signal to variation in our dependent variables (Blomberg et al. 2003, Revell et al. 2008).

On a phylogenetic hypothesis including the Dendrobates auratus, D. truncatus and the species of Oophaga (Amézquita et al., in Prep.), we eliminated the unused tree terminal tips with “ape” package (Paradis et al. 2004). We then explored the contribution of phylogenetic signal in the lnWeight, TEWL, EWL and %WL, calculating the Blomberg’s K statistic. This analysis provide a way to calculate the amount of phylogenetic signal for each character assuming a random walk in continuous time process, as a Brownian motion (BM) model of character evolution (Blomberg et al. 2003). BM is the simplest model of evolution of continuous traits in which, the branch lengths are assumed to be in units proportional to divergence times and to the variance of character evolution (Garland et al. 2015). Under this model, the change in the value of the character (positive or negative values) is given by random draws from a normal distribution, and more closely related species are expected to have more similar values than more distantly related species (Symonds & Blomberg 2014). K values vary from zero to infinity. K has an expected value of one when the data have evolved by BM model of evolution. K

50 values less than one implies that related species resemble each other less than expected under BM model of evolution. K values greater than one implies a strong phylogenetic signal, with closely related species more similar than expected under the same model (Blomberg et al. 2003, Revell et al. 2008). We conducted this analysis and tested for statistical significance of K values in the “phytools” package (Revell 2012).

Relationships between climate, body size and loss of water by evapotranspiration

We downloaded global current bioclimatic layers (Hijmans et al. 2005) of temperature seasonality (TempSeas, the standard deviation in annual temperature in °C * 100) and the current precipitation (mm) of driest quarter (PDQ) from the Worldclim bioclimatic database (Hijmans et al. 2005). We chose these climatic variables because the Oophaga species differ in these climatic niche dimensions (Amézquita et al., in Prep.). We also downloaded the global paleoclimatic layers reconstructed (downscaled global climate model data from IPPC data, CMIP5) for the Mid-Holocene (6,000 years ago), the Last Glacial Maximum (22,000 years ago), and the last interglacial (~120,000–140,000 years ago, Otto-Bliesner et al. 2006). From each layer or time considered we extracted a value of the conditions in TempSeas and PDQ for the geographic coordinates of this study. To estimate the range of change in these climatic conditions since the last glacial maximum, we first calculated the rate of change, as the difference in climatic conditions between two times divided by the time interval considered, between all possible pairwise comparisons of time. We thus use the short name Temperature Unpredictability and Precipitation Unpredictability to denote the range in the rate of change across four moments in the present, the Mid-Holocene, the Last Glacial Maximum and the last interglacial in temperature seasonality and precipitation of the driest quarter for posterior analyses.

To test for potential relationships between climate and phenotypical traits, we calculated multiple linear regressions between the mean values of climatic

51 conditions and the mean values of frog weight and water loss. To reduce the probability of inflating Type I error rates and thereby generating inaccurate estimates of evolutionary correlations (Martins & Garland 1991, Blomberg et al. 2003) we also ran multiple regressions with the independent contrasts approach (Felsesnstein 1985), forced through the origin, in the “ape” package (Paradis et al. 2004). The independent contrasts (IC), is the most popular method in comparative analyses of continuous data, and has been found to be comparatively unbiased and robust (Diaz-Uriarte & Garland 1996, Martins et al. 2002); it approaches the problem of non-independence of data by considering the differences between closely related species as the independent observations and the outcome of independent evolutionary pathways. By regressing the independent contrasts of the trait against the contrasts of another trait, one gets regression coefficients after accounting for phylogenetic relatedness among species (Symonds & Blomberg 2014). Yet another important factor to consider in comparative analyses is the measurement error (Harmon & Losos 2005). Measurement errors can be the consequence of instrument-related errors, sampling variation, errors in branch lengths and topology, as well as fluctuations or true biological differences in population or individuals due to variation in morphology, physiology and related to location, age, sex, time or season (Blomberg et al. 2003, Ives et al. 2007, Garamszegi 2014). Ignoring measurement error when using IC, increases the probability of inflating type I error, mainly when sample size (individuals per species) is small and the variation within species relative to that between species is large (among species variation is less than half of the total variation in the dataset). The probability of non-significant results or type II error, also increase under those conditions. Therefore, the use of comparative analyses methods that account for statistical error is recommended (Harmon & Losos 2005). We performed a phylogenetic generalized least-squares (PGLS) methods under a BM model of character evolution (Felsenstein 1985, Martins and Hansen 1997), with “ape” (Paradis et al. 2004) and "nlme" (Pinheiro et al. 2017) packages considering measurement error on the predictor variable according to Garamszegi (2014). PGLS control for the expected correlation between species based on their phylogeny and

52 weight for this, in the generalized least squares regression (Symonds & Blomberg 2014).

Results

The species we studied differed in weight (P = <0.001, Figure 2) and total rates of water loss (rTEWL, P = 0.046). The proportion of variance explained by species as a fixed factor was higher compared with the proportion of variance explained by populations in relation to Weight (Table 2, Figure 2). In contrast, the proportion of variance explained by populations was larger compared to species in rTEWL, rEWL, and r%WL. Dendrobates auratus, O. histrionica, and O. sylvatica had the larger weight compared to O. pumilio North, O. pumilio South and O. granulifera. Oophaga vicentei was the one with less weight. The TEWL in O. histrionica and O. sylvatica was lower, followed by D. auratus and D. truncatus. Oophaga granulifera exhibited intermediate rates of TEWL between the last ones and, O. pumilio North and O. pumilio South. Oophaga vicentei differed and had statistically higher rates of TEWLcompared to all other species we studied (Figure 2).

Blomberg’s K values for Weight, rTEWL, rEWL, and %WL were lower than 0.08 supporting the hypothesis of non-phylogenetic signal when ignoring sampling errors (Appendix 2). After including sampling errors, K values increased, but still support the hypothesis of non-phylogenetic signal (all lower than 0.3). However after testing for statistical significance, the rEWL exhibited phylogenetic signal (P = <0.01).

Before controlling for the effect of phylogeny no relationship was statistically supported despite a tendency in the reduction in rTEWL with an increase in TempSeas (Table 3). After controlling for the effects of phylogeny, an increase in Weight with an increase in the RRC TempSeas was supported by the PGLS phylogenetic analysis (P = 0.017). A reduction in the rTEWL with an increase in TempSeas (PIC: P = 0.015; PGLS: P = 0.033) and Weight was also supported after controlling by phylogeny (PIC: P = 0.011; PGLS: P = 0.021). A reduction in rTEWL

53 with an increase in RRC TempSeas was marginally supported (PIC: P = 0.061; PGLS: P = 0.058). Also, a tendency in the reduction in the percentage of weight loss (r%WL) with an increase in the lower precipitation on driest quarter (PDQ) was exhibited by the PGLS (P = 0.021).

Discussion

The species we sampled diverged in weight (Weight) and rates of total evaporative water loss (rTEWL). Dendrobates auratus, D. truncatus, O. histrionica, and O. sylvatica had greater body mass and lower rTEWL (Figure 2), compared to Central American species of Oophaga. Despite the absence of phylogenetic signal in all measured traits with the exception of the rates of evaporative water loss per unit surface area (rEWL), the presence of phylogenetic signal in our data can also be suspected with the increase in type II error when ignoring phylogenies. Also, there is a potential failure to detect phylogenetic signal in small data sets (Revell pers. com.). The absence of phylogenetic signal may in part reflect deviation from the Brownian Motion (BM) model of evolution for some of our traits or potential adaptations in at least some of the species to a particular environmental factor (Blomberg et al. 2003). Posso-Terranova & Andrés (2016) also suggested that Oophaga species climatic tolerances have evolved under a model of evolution different than BM. However, the presence of phylogenetic signal may also reflect the acquisition of similar adaptations under similar conditions experienced by closer relatives (Westoby et al. 1995, Tieleman et al. 2003). It is also important to keep in mind the limitations of using measures of phylogenetic signal to do evolutionary inferences, as different processes produce similar patterns of low phylogenetic signal (Revell et al. 2008, Losos 2011).

The results of multiple regressions correcting by phylogenetic relationships (IC and PGLS) suggest some adaptive hypotheses in our studied species regarding to the regulation of the loss of water by evapotranspiration (EWL): An increase in body weight (and therefore in body size, because both variables are highly correlated)

54 with an increase in temperature unpredictability (RRCTEmpSeas); and, the reduction in the rTEWL in species experiencing more current temperature seasonality (TempSeas) and RRCTempSeas. Therefore, our results suggest different but not exclusive potential adaptive scenarios. In the first one, an increase in body size/weight could represent a potential adaptation to temperature unpredictability (RRCTEmpSeas) but not to current temperature seasonality (TempSeas). In the second one, the reduction of rTEWL with an increase in current temperature seasonality (TempSeas) as well as greater temperature unpredictability (RRCTEmpSeas) could represent another potential adaptation and the potential development of morphological or physiological mechanisms in order to reduce the rates of EWL (reviewed in Navas et al. 2004, Tracy et al. 2010). The contribution of an increase in body size to reduce the rates of rTEWL was also identified, due to a lower surface-area-to-volume ratio (e.g., Thorson & Svihla 1943, Schmid 1965). When considering the rates of EWL per unit surface area (rEWL, therefore excluding the potential contribution of body weight to EWL) some relationship still may emerge between the rEWL and the temperature seasonality (TempSeas) and temperature unpredictability (RRCTEmpSeas), despite they are not statistically significant, also suggesting the presence of mechanism in order to reduce the EWL. Third, the change in body weight as a potential consequence of temperature unpredictability may also contribute to reduce the rates in rTEWL.

The temperature seasonality (TempSeas) estimates variation in temperature conditions that an organism can experience (Hijmans & Graham 2006). Throughout the distribution area of Dendrobates and Oophaga, a higher temperature seasonality implies higher seasonality in precipitation, a wider range of temperature and higher temperatures values in the warmest period along a year period (data not shown). These conditions may well represent a higher risk of desiccation for frogs and thereby reduce the available time for feeding throughout the yearly and daily cycles. It could also represent a greater fluctuation in the quality and quantity of larval deposition sites. A high rate of change in temperature seasonality represents the intensification of these conditions along a

55 historical period as well as the degree of unpredictability of temperature in a given geographical area.

Other studies have associated variation in body size with current seasonality and unpredictability of the Palearctic regions. The seasonality hypothesis predicts an evolutionary increase in body size in more seasonal or unpredictable areas (Boyce 1978), because larger sizes represent an increased ability to withstand periods of low resource availability (see also the Fasting Endurance Hypothesis in Lindstedt & Boyce 1985 or the Starvation Resistance Hypothesis in Cushman et al. (1993). Boyce (1978) links an increase in body size with an increase in climate seasonality within the North America muskrats (Ondatra zibethicus), suggesting that individuals with rapid growth and higher body size could increase its survivorship along periods of resource shortage. Accordingly, Murphy (1985), describes an increase in body size with an increase in temperature seasonality conditions within the House Sparrows (Passer domesticus) across North America and Europe. Cushman et al. (1993) proposed the Migration-Ability hypothesis to explain an increase in body size in the European ant’s assemblages at higher latitudes, in which larger species have higher dispersal abilities (migrate rapidly), and therefore have been able to colonize higher altitude regions.

As an opposite trend, the reduction in body size in response to global climate change has been suggested, because ectothermal organisms increase their metabolic rates with increasing temperatures; high metabolic costs allegedly lead to a reduction in adult body size when metabolic demands cannot be supported (Gardner et al. 2011, Sheridan & Bickford 2011). In the marine iguana Amblyrhynchus cristatus and the tortoise Homopus signatus, reduction in body size has been associated with periods of increased temperatures, low food availability and energetic stress (Wikelski & Thom 2000), as well as with periods of drought (Loehr et al. 2007). However, other species react by increasing body size (Gardner et al. 2011). In desert birds, which face with high temperatures and unpredictable water resources, larger body sizes could represent an advantage in environments with extreme events of temperature increase, because an increase of water

56 requirements under high temperature conditions in a day time can cause more reduced survival times in smaller size species (McKechnie & Wolf 2010). A reduction in body size has also been proposed to be more advantageous under environments with a continuous increase in temperature in comparison with environments where extreme and unpredictable changes may select for larger body sizes (Gardner et al. 2011).

In our study, the potential role of precipitation was not supported by the Water Balance hypothesis (see Olalla-Tárraga et al. 2009), which predict larger body size in species or populations that live under water shortage; by losing comparatively less water due to their low surface-to-volume ratio (e.g., Thorson & Svihla 1943, Schmid 1965). In concordance with this hypothesis, Fattorini et al. (2014) supported an association of larger body size in tenebrionid worms in the hotter and dryer regions of southern Europe. Nevo (1973) found that populations of the North American species Acris crepitans living in low rainfall areas are larger in body size. Olalla-Tárraga et al. (2009) found support to this correlation in the interspecific variation in anuran body size at the Brazilian Cerrado, with larger species present in areas of lower water availability, suggesting the role of water deficit as a determinant factor in shaping body size gradients within tropics. Nevertheless, some authors have elaborated on potential advantages of small frogs in areas of water shortage despite their higher rates of water loss: small frogs may have higher rates of water uptake because through their high surface-to-volume ratio (Thorson 1955, Heatwole et al. 1969). Lastly, behavioral mechanisms that reduce the loss of water loss (e.g., posture or microhabitat selection) may also account for the apparently contradictory patterns (Heatwole et al. 1969).

Different morphological mechanisms have also been identified in frogs to reduce their EWL. The reduction in metabolic rates and the gaseous exchange thought the skin and therefore the rates of EWL through the lungs; the reduction in skin permeability, the reduction in the use urine for nitrogen excretion; the increase the production of xeric cutaneous excretions; the increase in the cutaneous resistance; the increase in the concentration of amino acid in the tissues in order to regulate

57 the amount of water reabsorption; the modification of the morphology of ventral skin (increase in folds and capillaries in the inguinal region) in order to increase the water uptake; and, the ossification of the head as a exposed part of the body are some of the strategies that have been described in order to reduce the EWL in species present in environments with a reduce availability of water like Bufo granulosus (Navas et al. 2007), different species of Corythomantis (Shoemaker & McClanahan 1975, de Andrade & Abe 1997, but see Navas et al. 2002), Litorea (Butternmer et al. 1996), Hyperolius (Withers et al. 1982, Schmuck & Linsenmair 1997) and Phyllomedusa (Stinner & Shoemaker 1987, Shoemaker & McClanahan 1975). Behavioral mechanisms are also diverse but were studied and controlled in our experimental data. However, is interesting to recall some observations made in localities where D. auratus and O. pumilio co-occurr. Oophaga pumilio seems to have seasonal reproduction limited to the rainy season as well as restrict their movement along shaded sites, and their reproductive and parental care activities early in the morning, when the amount of moisture is greater. Dendrobates auratus seem to don’t possess such restrictions (e.g., Young 1979, Donnelly 1989, Graves 1999, Gardner & Graves 2005).

An essential caveat in the adaptive argument for our results is the fact that correlation does not mean causality. To better infer causal relationships, a better understanding of the mechanisms underlying the evolution of body size is required. Other potential ecological scenarios different than the adaptive one, should also be considered (see Gaston et al. 2009, Westoby et al. 1995). It is a very complex trait that often correlates with many other traits affecting ecological performance (Morrison & Hero 2003, Wikelski & Romero 2003). Body size should not be treated as a single isolated trait under a single selection pressure (Gaston et al. 2009) Also, intraspecific variation in body size should also be analyzed in order to estimate the contribution of climatic clines in shaping the emerging patterns we detected at the interspecific level (Gaston et al. 2009). Finally, the potential contribution of phenotypic plasticity to body size variation should be studied (Stillwell 2010, Tieleman et al. 2003), by conducting e.g., reciprocal transplant or common garden experiments. Ideally genetic covariation across climatic clines

(reviewed in Stillwell 2010 and Flat 2016) should provide more conclusive evidence to validate our adaptation hypothesis (Laugen et al. 2005, Wikelski & Romero 2003).

Figure 1. Comparison of two bio climatic parameters (Hijmans et al. 2005), temperature seasonality (TempSeas; left figure) and precipitation of driest quarter (PDQ; right figure), across the distribution range of Dendrobates auratus and D. truncatus (in dark grey in insert graphics); and Oophaga granulifera, O. histrionica, O. pumilio North, O. pumilio South, O. sylvatica, and O. vicentei (in light grey in insert graphics).

Figure 2. Evaporative water loss and body weight in poison frogs with regard to phylogenetic relationships and geographic distribution. Average values and ± 95% confidence intervals are presented for Dendrobates auratus (Da), D. truncatus (Dt), Oophaga granulifera (Og), O. pumilio South (OpS), O. pumilio North (OpN), O. vicentei (Ov), O. histrionica (Oh), and O. sylvatica (Os). r; measurements corrected by the relative humidity conditions of experiments; rTEWL: the residuals of the total rate of evaporative water loss (TEWL g x h-1). Polygons represent the geographic ranges occupied by the species and open circles indicate the localities sampled in this study.

Table 1. Sampled localities of all studied species including their corresponding geographical coordinates (latitude [LAT] and longitude [LON] in decimal degrees) and the number of studied individuals (N) at each population in the following countries: Colombia (COL); Costa Rica (CR); Ecuador (ECU); Nicaragua (NIC); and Panama (PAN).

Species Country Populations LAT LON N

D.auratus NIC Río Indio Lodge 10.9 -83.7 6 D.auratus PAN Cerro Tébata 9.6 -82.9 8 D.auratus PAN Ojo de Agua 9.3 -82.5 7 D.auratus PAN Coquerita 8.6 -77.4 13 D.auratus PAN David 8.4 -82.5 7 D.truncatus COL Mariquita 5.2 -74.9 10 O.granulifera PAN Chorogo 8.3 -83.0 10 O.pumilioSouth PAN Cerro Tébata 9.6 -82.9 10 O.pumilioSouth PAN Isla Colón 9.4 -82.3 7 O.pumilioNorth NIC Kilambé 13.6 -85.7 10 O.pumilioNorth NIC Musún 13.0 -85.2 10 O.pumilioNorth NIC Laguna de Perlas 12.7 -83.7 8 O.pumilioNorth NIC Chontales 12.3 -85.1 10 O.pumilioNorth NIC Guatuzos 11.0 -85.1 8 O.pumilioNorth NIC Bartola 11.0 -84.3 3 O.pumilioNorth NIC Río Indio Lodge 10.9 -83.7 4 O.pumilioNorth CR La Selva 10.4 -84.0 10 O.pumilioNorth PAN Isla Veraguas 9.1 -81.6 10 O.vicentei PAN Cerro Narices 8.6 -81.0 10 O.histrionica COL Arusí 5.6 -77.5 8 O.histrionica COL Delfina 3.8 -76.8 8 O.histrionica COL Naranjo 3.8 -76.7 8 O.sylvatica ECU Guaduales 0.9 -78.5 7 O.sylvatica ECU Durango -0.3 -78.9 7

Table 2. Linear mixed effect model (LMM) with species as fixed factor and populations as random nested factor within species. Response variables are those related to the water loss by evapotranspiration. LnWeight: the natural logarithm of frog’s body weight; r; measurements corrected by the relative humidity conditions of experiments; rTEWL: the residuals of the total rate of evaporative water loss (%WLxh); rEWL: the rate of water loss per unit area (g x cm-2 x h-1); %WL: the tolerance to the loss of water by evapotranspiration.

Weight rTEWL rEWL r%WL Random effects Variance Std.Dev. Variance Std.Dev. Variance Std.Dev. Variance Std.Dev. Population 0.50 0.71 221.40 14.88 0.00 0.01 20.22 4.50 Residual 0.17 0.41 139.20 11.80 0.00 0.01 28.17 5.31 Fixed effects Est. SE t X 2 d.f. P Est. SE t X 2 d.f. P Est. SE t X 2 d.f. P Est. SE t X 2 d.f. P Species 2.75 0.32 8.52 47.90 7 <0.001 -8.90 6.92 -1.29 14.31 7 0.046 0.01 0.01 0.99 10.06 7 0.185 -0.17 2.19 -0.08 1.56 7 0.98 AIC 298.08 1620.12 -1026.84 1285.11 R2 Marginal 0.61 0.29 0.13 0.03 R2 Conditional 0.90 0.73 0.37 0.44

Tabla 3. Multiple regression coefficients between historical and current climatic conditions with the variables related to the loss of water by evapotranspiration (Raw). Correlations were controlled by phylogenetic relationships with phylogenetic independent contrast (PIC) and phylogenetic generalized linear squares (PGLS). TempSeas: temperature seasonality; RRC TempSeas: range in the rate of change in temperature seasonality; PDQ: precipitation of driest quarter; RCC PDQ: range in the rate of change in precipitation of driest quarter; LnWeight: the natural logarithm of frog’s body weight; r; measurements corrected by the relative humidity conditions of experiments; rTEWL: the residuals of the total rate of evaporative water loss (%WLxh); rEWL: the residuals of the rate of water loss per unit area (g x cm-2 x h-1); the residuals of tolerance to the loss of water by evapotranspiration (r%WL). P-values lower than one are present in bold. Signifficant P-vales (<0.05) are in gray.

Multiple regression with estimation of PIC (Multiple Regression through Response PGLS with error Predictor Variable intercept the origin) Variable Coeff. Std_error t-value P-value Coeff. Std_error t-value P-value Value Std.Error t-value P-value Weight Intercept -0.312 2.253 -0.138 0.897 -0.665 0.284 -2.340 0.079 TempSeas -0.001 0.002 -0.440 0.683 -0.001 0.002 -0.226 0.832 0.000 0.000 0.249 0.816 PDQ 0.002 0.002 1.222 0.289 0.002 0.003 0.686 0.530 0.000 0.001 0.367 0.732 RRC TempSeas 0.100 0.075 1.322 0.257 -0.010 0.119 -0.080 0.940 0.079 0.020 3.912 0.017 rTEWL Intercept 84.861 23.221 3.655 0.035 99.094 28.265 3.506 0.039 TempSeas -0.053 0.021 -2.506 0.087 -0.048 0.009 -5.044 0.015 -0.061 0.016 -3.777 0.033 PDQ -0.039 0.024 -1.619 0.204 -0.022 0.014 -1.604 0.207 -0.018 0.016 -1.136 0.339 RRC TempSeas -1.629 0.930 -1.752 0.178 -1.471 0.502 -2.928 0.061 -2.374 0.792 -2.996 0.058 Weight -10.803 5.142 -2.101 0.126 -20.479 3.682 -5.562 0.011 -12.475 2.813 -4.435 0.021 rEWL Intercept 0.042 0.039 1.076 0.361 0.081 0.032 2.486 0.089 TempSeas -2.8E-05 0.000 -0.767 0.499 -4.6E-05 1.7E-05 -2.666 0.076 0.000 0.000 -2.476 0.090 PDQ -1.8E-05 0.000 -0.447 0.685 -4.1E-06 2.6E-05 -0.159 0.884 0.000 0.000 -0.631 0.573 RRC TempSeas -0.001 0.002 -0.683 0.544 -2.3E-03 0.001 -2.479 0.089 -0.002 0.001 -2.594 0.081 Weight -0.003 0.009 -0.329 0.764 -7.3E-03 0.007 -1.08 0.359 -0.003 0.004 -0.644 0.565 r%WL Intercept -0.587 4.507 -0.130 0.905 -3.228 1.426 -2.264 0.109 TempSeas 0.002 0.004 0.447 0.685 0.005 0.002 2.29 0.106 0.000 0.003 0.150 0.890 PDQ -0.004 0.005 -0.906 0.432 -0.007 0.004 -1.909 0.152 -0.007 0.002 -3.049 0.056 RRC TempSeas 0.003 0.180 0.019 0.986 0.154 0.126 1.218 0.310 0.156 0.073 2.128 0.123 Weight 0.433 0.998 0.434 0.694 1.126 0.926 1.216 0.311 0.039 0.564 0.069 0.949

Appendix 1.

S1.1.Pearson (upper right) and Spearman (lower left) correlation coefficients between climatic conditions during experiments, frog behavior, and dependent variables related with the water loss by evapotranspiration and body condition. SVL: snout vent length; Ini.: Initial; Fin.: Final; Weight: body weight; Time (min): The duration of experiments as the time at which the frog lost the ability to straighten up themselves when upset; TEWL: the total rate of evaporative water loss; EWL: the rate of water loss per unit area; the tolerance to the loss of water by evapotranspiration (%WL). Env.: environmental; Temp.: temperature; RH: relative humidity; Inflow: air flow going inside the wind tunnel; Outflow: air flow going outside the wind tunnel.

*: Significant correlation at 0.05 (bilateral). **: Significant correlation at 0.01 (bilateral).

Pearson correlation coefficients/Spearman correlation coefficients 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 SVL (mm) 1 .945** .937** .158* -.171* -.446** .214** .114 -.167* -.167* -.299** -.440** -.425** -.466** -.411** -.443** 2 Ini.Weight (g) .960** 1 .991** .206** -.175* -.463** .150* .111 -.157* -.142* -.268** -.406** -.361** -.403** -.348** -.377** 3 Fin.Weight (g) .955** .988** 1 .162* -.285** -.456** .151* .128 -.143* -.133 -.260** -.410** -.370** -.410** -.356** -.384** 4 Time (min) .295** .371** .327** 1 .344** -.701** -.679** -.148* -.530** -.536** -.471** -.455** .462** .472** .463** .468** 5 % Weight loss -.268** -.247** -.351** .333** 1 .003 -.088 -.124 -.164* -.126 -.093 .027 .208** .202** .199** .188** 6 TEWL (%WLxh) -.428** -.503** -.507** -.900** .056 1 .692** .118 .445** .459** .485** .519** -.257** -.263** -.265** -.270** 7 EWL (gxcm2xh1) .144* .082 .075 -.759** -.101 .764** 1 .167* .443** .468** .339** .297** -.638** -.663** -.634** -.648** 8 Activity score .072 .046 .055 -.118 -.140* .080 .133 1 -.107 -.117 -.072 -.084 .052 .042 .065 .062 9 Ini. Env. Temp (°C) -.199** -.203** -.202** -.499** -.081 .516** .433** -.075 1 .889** .892** .815** -.242** -.202** -.246** -.203** 10 Fin. Env. Temp (°C) -.199** -.191** -.187** -.509** -.084 .530** .477** -.088 .844** 1 .829** .864** -.192** -.228** -.200** -.224** 11 Ini. Body Temp (°C) -.297** -.301** -.299** -.481** -.080 .511** .353** -.020 .886** .790** 1 .861** -.054 -.022 -.058 -.025 12 Fin. Body Temp (°C) -.402** -.397** -.404** -.494** .054 .580** .361** -.062 .827** .876** .850** 1 .064 .067 .052 .062 13 Ini. RH inflow (%) -.386** -.331** -.338** .452** .244** -.382** -.668** .055 -.250** -.200** -.101 -.068 1 .945** .993** .942** 14 Fin. RH inflow (%) -.412** -.353** -.365** .486** .271** -.396** -.706** .021 -.162* -.226** -.033 -.037 .886** 1 .947** .991** 15 Ini. RH outflow (%) -.363** -.310** -.314** .466** .232** -.402** -.670** .073 -.269** -.221** -.121 -.097 .984** .888** 1 .945** ** ** ** ** ** ** ** * ** ** ** ** 16 Fin. RH outflow (%) -.388 -.327 -.337 .482 .247 -.402 -.700 .028 -.171 -.240 -.045 -.060 .870 .984 .873 1

S1.2. Matrix of rotated principal components (PC) across climatic conditions during experiments, frog behavior, dependent variables related with the water loss by evapotranspiration, and body condition. For abbreviations see Appendix 1.

Components Variables Communalities 1 2 3 4 5 6 Ini. RH outflow (%) .964 -.068 -.160 -.113 .059 .015 .976 Ini. RH inflow (%) .959 -.063 -.175 -.117 .068 .004 .973 Fin. RH outflow (%) .944 -.048 -.200 -.168 .050 .038 .966 Fin. RH inflow (%) .938 -.050 -.228 -.181 .060 .021 .971 Ini. Env. Temp (°C) -.180 .937 -.060 .089 -.089 -.032 .932 Ini. Body Temp (°C) .008 .932 -.141 .121 -.041 .000 .904 Fin. Env. Temp (°C) -.144 .926 -.031 .173 -.050 -.066 .915 Ini. Body. Temp (°C) .079 .903 -.269 .160 .056 -.020 .923 Ini.Weight (g) -.216 -.132 .954 -.074 -.033 .043 .982 Fin.Weight (g) -.218 -.130 .943 -.070 -.146 .054 .983 SVL (mm) -.293 -.176 .914 -.053 -.013 .046 .958 TEWL (%WLxh) -.246 .331 -.449 .743 .041 .098 .936 EWL (gxcm2xh1) -.532 .302 .155 .729 .055 .123 .947 Time (min) .422 -.391 .234 -.567 .354 -.145 .854 % Weight loss .110 -.070 -.150 -.005 .967 -.056 .978 Activity score .052 -.082 .095 .123 -.062 .980 .999 % variance 36.479 30.815 11.127 7.865 5.932 2.756

S1.3. Pearson correlation coefficients between the selected climatic conditions in our experiments and the variables related with the water loss by evapotranspiration. For abbreviations see Appendix 1.

Single Linear regression coefficient sig. Ini. RH outflow (%) TEWL (%WLxh) 0.07 0.000 EWL (gxcm2xh1) 0.402 0.000 % Weight loss 0.039 0.005 Ini. Env. Temp (°C) TEWL (%WLxh) 0.198 0.000 EWL (gxcm2xh1) 0.196 0.000 % Weight loss -0.427 0.021 Activity score TEWL (%WLxh) 0.014 0.098 EWL (gxcm2xh1) 0.028 0.018 % Weight loss 0.015 0.080

Appendix 2. Blomberg’s K values ignoring (a) and including (b) sampling errors.

LnWeight rTEWL rEWL r%WL a b a b a b a b K 0.082 0.198 0.079 0.274 0.071 0.154 0.072 0.099 sig2 0.229 67.27 <0.01 1.345

Literature cited

Adolph, E. F. (1933). Exchanges of water in the frog. Biological Reviews, 8(2), 224–240.

Bates, D., Mächler, M., Bolker, B., & Walker, S. (2014). Fitting linear mixed-effects models using lme4. arXiv preprint arXiv:1406.5823.

Blomberg, S. P., Garland, T., & Ives, A. R. (2003). Testing for phylogenetic signal in comparative data: behavioral traits are more labile. Evolution, 57(4), 717–745.

Boaratti, A. Z., & Rodrigues da Silva, F. R. (2015). Relationships between environmental gradients and geographic variation in the intraspecific body size of three species of frogs (Anura). Austral Ecology, 40(8), 869–876.

Boyce, M. S. (1978). Climatic variability and body size variation in the muskrats (Ondatra zibethicus) of North America. Oecologia, 36(1), 1–19.

Brandt, R., & Navas, C. A. (2013). Body size variation across climatic gradients and sexual size dimorphism in Tropidurinae lizards. Journal of Zoology, 290(3), 192–198.

Brown, J. L., Maan, M. E., Cummings, M. E., & Summers, K. (2010). Evidence for selection on coloration in a Panamanian poison frog: a coalescent-based approach. Journal of Biogeography, 37(5), 891–901.

Buttemer, W. A., Van Der Wielen, M., Dain, S., & Christy, M. (1996). Cutaneous properties of the green and golden bell frog Litoria aurea. Australian Zoologist, 30(2), 134–138.

Chown, S. L., & Gaston, K. J. (2010). Body size variation in insects: a macroecological perspective. Biological Reviews, 85(1), 139–169.

Crump, M. L. (1972). Territoriality and mating behavior in Dendrobates granuliferus (Anura: Dendrobatidae). Herpetologica, 28(3), 195–198.

Cushman, J. H., Lawton, J. H., & Manly, B. F. (1993). Latitudinal patterns in European ant assemblages: variation in species richness and body size. Oecologia, 95(1), 30–37.

Daly, J. W., & Myers, C. W. (1967). Toxicity of Panamanian poison frogs (Dendrobates): some biological and chemical aspects. Science, 156(3777), 970– 973. de Andrade, D. V., & Abe, A. S. (1997). Evaporative water loss and oxygen uptake in two casque-headed tree frogs, Aparasphenodon brunoi and Corythomantis greeningi (Anura, Hylidae). Comparative Biochemistry and Physiology Part A: Physiology, 118(3), 685–689.

Diaz-Uriarte, R., & Garland, T. (1996). Testing hypotheses of correlated evolution using phylogenetically independent contrasts: sensitivity to deviations from Brownian motion. Systematic Biology, 45(1), 27–47.

Donnelly, M. A. (1989). Reproductive phenology and age structure of Dendrobates pumilio in northeastern Costa Rica. Journal of Herpetology, 23(4), 362–367.

Fattorini, S., Monaco, R. L., Di Giulio, A., & Ulrich, W. (2014). Climatic correlates of body size in European tenebrionid beetles (Coleoptera: Tenebrionidae). Organisms Diversity & Evolution, 14(2), 215–224.

Felsenstein, J. (1985). Phylogenies and the comparative method. American Naturalist, 125(1), 1–15.

Flatt, T. (2016). Genomics of clinal variation in Drosophila: disentangling the interactions of selection and demography. Molecular Ecology, 25(5), 1023–1026.

Fox, J., & Weisberg, S. (2010). An R Companion to Applied Regression. Second Edition. Thousand Oaks CA: Sage. URL: http://socserv.socsci.mcmaster.ca/jfox/Books/Companion

Garamszegi, L.Z., (2014). Uncertainties Due to Within-Species Variation in Comparative Studies: Measurement Errors and Statistical Weights. Chapter 7. p. 157. Garamszegi, L,Z., (Ed). Modern Phylogenetic Comparative Methods and Their Application in Evolutionary Biology. Concepts and Practice 2014. Springer Heidelberg New York Dordrecht London. 552pp.

Gardner, E. A., & Graves, B. M. (2005). Responses of resident male Dendrobates pumilio to territory intruders. Journal of herpetology, 39(2), 248–253.

Gardner, J. L., Peters, A., Kearney, M. R., Joseph, L., & Heinsohn, R. (2011). Declining body size: a third universal response to warming?. Trends in ecology & evolution, 26(6), 285–291.

Garland, T., Bennett, A. F., & Rezende, E. L. (2005). Phylogenetic approaches in comparative physiology. Journal of experimental Biology, 208(16), 3015–3035.

Gaston K. J., Chown S. L., Calosi P., Bernardo J., Bilton D. T., Clarke A., Clusella- Trullas S., Ghalambor C. K., Konarzewski M., Peck L. S., Porter W. P., Pörtner H. O., Rezende E. L., Schulte P. M., Spicer J. I., Stillman J. H., Terblanche J. S., & van Kleunen M. (2009). Macrophysiology: A Conceptual Reunification. The American Naturalist, 174, 595–612.

Goodman, D. E. (1971). Territorial behavior in a Neotropical frog, Dendrobates granuliferus. Copeia, 1971(2), 365–370.

Grant, T., Frost, D. R., Caldwell, J. P., Gagliardo, R., Haddad, C. F. B., Kok, P. J. R., Means, D. B., Noonan, B. P., Schargel, W. E., & Wheeler, W. C. (2006). Phylogenetic systematics of dart-poison frogs and their relatives (Amphibia: Athesphatanura: Dendrobatidae). Bulletin of the American Museum of natural History, 299, 1–262.

Graves, G. R. (1991). Bergmann's rule near the equator: latitudinal clines in body size of an Andean passerine bird. Proceedings of the National Academy of Sciences, 88(6), 2322–2325.

Graves, B. M. (1999). Diel activity patterns of the sympatric poison dart frogs, Dendrobates auratus and D. pumilio, in Costa Rica. Journal of Herpetology, 33(3), 375–381.

Gutiérrez-Pinto, N., McCracken, K. G., Alza, L., Tubaro, P., Kopuchian, C., Astie, A., & Cadena, C. D. (2014). The validity of ecogeographical rules is context- dependent: testing for Bergmann's and Allen's rules by latitude and elevation in a widespread Andean duck. Biological Journal of the Linnean Society, 111(4), 850–862.

Hagemann, S., & Pröhl, H. (2007). Mitochondrial paraphyly in a polymorphic poison frog species (Dendrobatidae; D. pumilio). Molecular phylogenetics and evolution, 45(2), 740–747.

Harmon, L. J., & Losos, J. B. (2005). The effect of intraspecific sample size on type I and type II error rates in comparative studies. Evolution, 59(12), 2705–2710.

Heatwole, H., Torres, F., De Austin, S. B., & Heatwole, A. (1969). Studies on anuran water balance—I. Dynamics of evaporative water loss by the coqui, Eleutherodactylus portoricensis. Comparative Biochemistry and Physiology, 28(1), 245–269.

Hijmans, R. J., Cameron, S. E., Parra, J. L., Jones, P. G., & Jarvis, A. (2005). Very high resolution interpolated climate surfaces for global land areas. International journal of climatology, 25(15), 1965–1978.

Hijmans, R. J., & Graham, C. H. (2006). The ability of climate envelope models to predict the effect of climate change on species distributions. Global change biology, 12(12), 2272-2281.

Huey, R. B., Kearney, M. R., Krockenberger, A., Holtum, J. A., Jess, M., & Williams, S. E. (2012). Predicting organismal vulnerability to climate warming: roles of behaviour, physiology and adaptation. Phil. Trans. R. Soc. B, 367(1596), 1665– 1679.

Huston, M. A., & Wolverton, S. (2011). Regulation of animal size by eNPP, Bergmann's rule and related phenomena. Ecological Monographs, 81(3), 349– 405.

Ives, A. R., Midford, P. E., & Garland, T. (2007). Within-species variation and measurement error in phylogenetic comparative methods. Systematic Biology, 56(2), 252–270.

Laugen, A. T., Laurila, A., Jönsson, K. I., Söderman, F., & Merilä, J. (2005). Do common frogs (Rana temporaria) follow Bergmann’s rule?. Evolutionary Ecology Research, 7(5), 717–731.

Le Lagadec, M. D., Chown, S. L., & Scholtz, C. H. (1998). Desiccation resistance and water balance in southern African keratin beetles (Coleoptera, Trogidae): the influence of body size and habitat. Journal of Comparative Physiology B, 168(2), 112–122.

Lefcheck, J. S. (2015). piecewiseSEM: Piecewise structural equation modelling in r for ecology, evolution, and systematics. Methods in Ecology and Evolution, 7, 573–579.

Limerick, S. (1980). Courtship behavior and oviposition of the poison-arrow frog Dendrobates pumilio. Herpetologica, 36(1), 69–71.

Lindstedt, S. L., & Boyce, M. S. (1985). Seasonality, fasting endurance, and body size in mammals. The American Naturalist, 125(6), 873–878.

Loehr, V. J., Hofmeyr, M. D., & Henen, B. T. (2007). Growing and shrinking in the smallest tortoise, Homopus signatus signatus: the importance of rain. Oecologia, 153(2), 479–488.

Losos, J. B. (2011). Seeing the forest for the trees: the limitations of phylogenies in comparative biology. The American Naturalist, 177(6), 709–727.

Lötters, S., K.-H. Jungfer, F. W. Henkel, and W. Schmidt. (2007). Poison Frogs. Biology, Species & Captive Maintenance. Frankfurt am Main: Edition Chimaira.

Martins, E. P., Diniz‐Filho, J. A., & Housworth, E. A. (2002). Adaptive constraints and the phylogenetic comparative method: a computer simulation test. Evolution, 56(1), 1–13.

Martins, E. P., & Hansen, T. F. (1997). Phylogenies and the comparative method: a general approach to incorporating phylogenetic information into the analysis of interspecific data. American Naturalist, 149(4), 646–667.

Martins, E. P., & Garland Jr, T. (1991). Phylogenetic analyses of the correlated evolution of continuous characters: a simulation study. Evolution, 45(3), 534– 557.

McKechnie, A. E., & Wolf, B. O. (2009). Climate change increases the likelihood of catastrophic avian mortality events during extreme heat waves. Biology Letters, 6(2), 253–256.

McVey, M. E., Zahary, R. G., Perry, D., & MacDougal, J. (1981). Territoriality and homing behavior in the poison dart frog (Dendrobates pumilio). Copeia, 1981(1), 1–8.

Morrison, C., & Hero, J. M. (2003). Geographic variation in life‐history characteristics of amphibians: a review. Journal of Animal Ecology, 72(2), 270– 279.

Muñoz, M. M., Wegener, J. E., & Algar, A. C. (2014). Untangling intra-and interspecific effects on body size clines reveals divergent processes structuring convergent patterns in Anolis lizards. The American Naturalist, 184(5), 636–646.

Murphy, E. C. (1985). Bergmann's rule, seasonality, and geographic variation in body size of house sparrows. Evolution, 39(6), 1327–1334.

Nakagawa, S., & Schielzeth, H. (2013). A general and simple method for obtaining R2 from generalized linear mixed‐effects models. Methods in Ecology and Evolution, 4(2), 133–142.

Navas, C. A., Antoniazzi, M. M., Carvalho, J. E., Suzuki, H., & Jared, C. (2007). Physiological basis for diurnal activity in dispersing juvenile Bufo granulosus in the Caatinga, a Brazilian semi-arid environment. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology, 147(3), 647–657.

Navas, C. A., Antoniazzi, M. M., & Jared, C. (2004). A preliminary assessment of anuran physiological and morphological adaptation to the Caatinga, a Brazilian semi-arid environment. In International congress series, 1275, 298– 305.

Navas, C. A., Jared, C., & Antoniazzi, M. M. (2002). Water economy in the casque‐ headed tree-frog Corythomantis greeningi (Hylidae): role of behaviour, skin, and skull skin co-ossification. Journal of Zoology, 257(4), 525–532.

Nevo, E. (1973). Adaptive variation in size of cricket frogs. Ecology, 54(6), 1271– 1281.

Olalla-Tárraga, M. Á., & Rodríguez, M. Á. (2007). Energy and interspecific body size patterns of amphibian faunas in Europe and North America: anurans follow Bergmann's rule, urodeles its converse. Global Ecology and Biogeography, 16(5), 606–617.

Olalla-Tárraga, M. Á., Diniz-Filho, J. A. F., Bastos, R. P., & Rodríguez, M. Á. (2009). Geographic body size gradients in tropical regions: water deficit and anuran body size in the Brazilian Cerrado. Ecography, 32(4), 581–590.

Otto-Bliesner, B. L., Marshall, S. J., Overpeck, J. T., Miller, G. H., & Hu, A. (2006). Simulating Arctic climate warmth and icefield retreat in the last interglaciation. Science, 311(5768), 1751–1753.

Paradis, E., Claude, J., & Strimmer, K. (2004). APE: analyses of phylogenetics and evolution in R language. Bioinformatics, 20(2), 289–290.

Phillimore, A. B., Stålhandske, S., Smithers, R. J., & Bernard, R. (2012). Dissecting the contributions of plasticity and local adaptation to the phenology of a butterfly and its host plants. The American Naturalist, 180(5), 655–670.

Pinheiro, J., Bates, D., DebRoy, S., & Sarkar, D., & R Core Team (2017). Nlme: linear and nonlinear mixed effects models. R package version 3.1-117. See http://CRAN. R-project. org/package= nlme.

Posso-Terranova, A., & Andrés, J. A. (2016). Complex niche divergence underlies lineage diversification in Oophaga poison frogs. Journal of Biogeography, 43(10), 2002–2015.

Prates, I., & Navas, C. A. (2009). Cutaneous resistance to evaporative water loss in Brazilian Rhinella (Anura: Bufonidae) from contrasting environments. Copeia, 2009(3), 618–622.

Pröhl, H., & Berke, O. (2001). Spatial distributions of male and female strawberry poison frogs and their relation to female reproductive resources. Oecologia, 129(4), 534–542.

Pröhl, H., Hagemann, S., Karsch, J., & Höbel, G. (2007). Geographic variation in male sexual signals in strawberry poison frogs (Dendrobates pumilio). Ethology, 113(9), 825–837.

R Core Team (2015). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R- project.org/.

Revell, L. J. (2012). Phytools: an R package for phylogenetic comparative biology (and other things). Methods in Ecology and Evolution, 3(2), 217–223.

Revell, L. J., Harmon, L. J., & Collar, D. C. (2008). Phylogenetic signal, evolutionary process, and rate. Systematic Biology, 57(4), 591–601.

Rogowitz, G. L., Cortés-Rivera, M., & Nieves-Puigdoller, K. (1999). Water loss, cutaneous resistance, and effects of dehydration on locomotion of Eleutherodactylus frogs. Journal of Comparative Physiology B, 169(3), 179–186.

Rudh, A., Rogell, B., & Höglund, J. (2007). Non-gradual variation in colour morphs of the strawberry poison frog Dendrobates pumilio: genetic and geographical isolation suggest a role for selection in maintaining polymorphism. Molecular Ecology, 16(20), 4284–4294.

Savage, J. M. (1968). The Dendrobatid Frogs of Central America. Copeia, 1968(4), 745–776.

Schäuble, C. S. (2004). Variation in body size and sexual dimorphism across geographical and environmental space in the frogs Limnodynastes tasmaniensis and L. peronii. Biological Journal of the Linnean Society, 82(1), 39– 56.

Schmid, W. D. (1965). Some aspects of the water economies of nine species of amphibians. Ecology, 46(3), 261–269.

Schmuck, R., & Linsenmair, K. E. (1997). Regulation of body water balance in reedfrogs (superspecies Hyperolius viridiflavus and Hyperolius marmoratus: Amphibia, Anura, Hyperoliidae) living in unpredictably varying savannah environments. Comparative Biochemistry and Physiology Part A: Physiology, 118(4), 1335–1352.

Schneider, C. A., Rasband, W. S., & Eliceiri, K. W. (2012). NIH Image to ImageJ: 25 years of image analysis. Nature methods, 9(7), 671–675.

Sheridan, J. A., & Bickford, D. (2011). Shrinking body size as an ecological response to climate change. Nature climate change, 1(8), 401–406.

Shoemaker, V. H., & McClanahan, L. L. (1975). Evaporative water loss, nitrogen excretion and osmoregulation in phyllomedusine frogs. Journal of Comparative Physiology B: Biochemical, Systemic, and Environmental Physiology, 100(4), 331–345.

Silverstone, P. A. (1973). Observations on the behavior and ecology of a Colombian poison-arrow frog, the Kõkoé-Pá (Dendrobates histrionicus Berthold). Herpetologica, 29(4), 295–301.

Silverstone, P. A. (1975). A revision of the poison-arrow frogs of the genus Dendrobates Wagler. Revisión de las ranas venenosas del género Dendrobates Wagler. Natural History Museum of Los Angeles Country, 21, 1–55.

Spotila, J. R., & Berman, E. N. (1976). Determination of skin resistance and the role of the skin in controlling water loss in amphibians and reptiles. Comparative Biochemistry and Physiology Part A: Physiology, 55(4), 407–411.

Stillwell, R. C. (2010). Are latitudinal clines in body size adaptive?. Oikos, 119(9), 1387–1390.

Stinner, J. N., & Shoemaker, V. H. (1987). Cutaneous gas exchange and low evaporative water loss in the frogs Phyllomedusa sauvagei and Chiromantis xerampelina. Journal of Comparative Physiology B: Biochemical, Systemic, and Environmental Physiology, 157(4), 423–427.

Summers, K., Weigt, L. A., Boag, P., & Bermingham, E. (1999). The evolution of female parental care in poison frogs of the genus Dendrobates: evidence from mitochondrial DNA sequences. Herpetologica, 55(2), 254–270.

Symonds, M. R. E., Blomberg, S. P. (2014). A Primer on Phylogenetic Generalised Least Squares. Chapter 5. p. 105. Garamszegi LZ (Ed). Modern Phylogenetic Comparative Methods and Their Application in Evolutionary Biology. Concepts and Practice 2014. Springer Heidelberg New York Dordrecht London. 552pp.

Thorson, T. B. (1955). The relationship of water economy to terrestrialism in amphibians. Ecology, 36(1), 100–116.

Thorson, T., & Svihla, A. (1943). Correlation of the habitats of amphibians with their ability to survive the loss of body water. Ecology, 24(3), 374–381.

Tieleman, B. I., Williams, J. B., & Bloomer, P. (2003). Adaptation of metabolism and evaporative water loss along an aridity gradient. Proceedings of the Royal Society of London B: Biological Sciences, 270(1511), 207–214.

Titon, B. J., & Gomes, F. R. (2015). Relation between Water Balance and Climatic Variables Associated with the Geographical Distribution of Anurans. PloS one, 10(10), e0140761.

Tracy, C. R., & Christian, K. A. (2005). Preferred temperature correlates with evaporative water loss in hylid frogs from northern Australia. Physiological and Biochemical Zoology, 78(5), 839–846.

Tracy, C. R., Christian, K. A., & Tracy, C. R. (2010). Not just small, wet, and cold: effects of body size and skin resistance on thermoregulation and arboreality of frogs. Ecology, 91(5), 1477–1484. van Wijngaarden, R., & Bolaños, F. (1992). Parental care in Dendrobates granuliferus (Anura: Dendrobatidae), with a description of the tadpole. Journal of Herpetology, 26(1), 102–105.

Wells, K. D. (1978). Courtship and parental behavior in a Panamanian poison- arrow frog (Dendrobates auratus). Herpetologica, 34(2), 148–155.

Westoby, M., Leishman, M. R., & Lord, J. M. (1995). On misinterpreting thephylogenetic correction. Journal of Ecology, 83(3), 531–534.

Wikelski, M., & Romero, L. M. (2003). Body size, performance and fitness in Galapagos marine iguanas. Integrative and Comparative Biology, 43(3), 376–386.

Wikelski, M., & Thom, C. (2000). Marine iguanas shrink to survive El Niño. Nature, 403(6765), 37–38.

Withers, P. C., Hillman, S. S., Drewes, R. C., & Sokol, O. M. (1982). Water loss and nitrogen excretion in sharp-nosed reed frogs (Hyperolius nasutus: Anura, Hyperoliidae). Journal of Experimental Biology, 97(1), 335–343.

Wygoda, M. L., Dabruzzi, T. F., & Bennett, W. A. (2011). Cutaneous resistance to evaporative water loss in the crab-eating frog (Fejervarya cancrivora). Journal of Herpetology, 45(4), 417–420.

Young, A. M. (1979). Arboreal movement and tadpole-carrying behavior of Dendrobates pumilio Schmidt (Dendrobatidae) in northeastern Costa Rica. Biotropica, 11(3), 238–239.

Young, J. E., Christian, K. A., Donnellan, S., Tracy, C. R., & Parry, D. (2005). Comparative analysis of cutaneous evaporative water loss in frogs demonstrates correlation with ecological habits. Physiological and Biochemical Zoology, 78(5), 847–856.

Isla Colón, Panama.

THE SENSITIVITY OF LOCOMOTOR PERFORMANCE TO DEHYDRATION IN THE POISON FROG OOPHAGA PUMILIO (Anura, Dendrobatidae)

Diana M. Galindo Uribe & Adolfo Amézquita

Introduction

Amphibians are ectotherms with a permeable skin that experience fluctuating body temperatures and a constant risk of desiccation when living in environments with low water availability, or when exhibit terrestrial and diurnal habits (Thorson & Svihla 1943, Schmid 1965, Tracy 1976). Therefore, anurans are expected either to develop behavioral, morphological, and physiological mechanisms that reduce the loss of water by evapotranspiration (see Rogowitz et al. 1999, Tracy et al. 2010 and references therein), or to tolerate the variation in environmental conditions and hydration states (Preest & Pough 1989).

Dehydration has a negative effect over locomotor performance of individuals. It reduces exercise capacity by slowing heart rate, cellular metabolism, muscular activity (Pough et al. 2004). Dehydrated individuals move more slowly and fatigue more rapidly when exercising (reviewed in Prates et al. 2013). Through its effect on locomotion, dehydration may profoundly affect escape behavior, feeding, reproduction and social dominance (Rogowitz et al. 1999, Perry et al. 2004, Titon et al. 2010, Köhler et al. 2011, Prates et al. 2013), among others, thereby affecting survival and reproductive success. Anurans might counteract these negative effects and thrive in less humid environments if they reduce the sensitivity of locomotor performance to water loss (e.g., Titon et al. 2010, Prates et al. 2013, Titon & Gomes 2015).

The poison frog Oophaga pumilio (Dendrobatidae) has been widely studied with regard to its complex behaviour: territoriality, space use, courtship, mate choice, and parental care (reviwed in Pröhl 2005, Pröhl & Willink 2015). Because these frogs are terrestrial and diurnal, they are at higher risk of dehydration compared

79 to nocturnal frogs. Previous research stressed the vulnerability of this species to water shortage and described behavioral adjustments that allegedly compensate this risk: courtship and tadpole transport to rearing sites, for instance, appear restricted to the early morning or shade (Young 1979, Pröhl 1997, Graves, 1999, Pröhl & Hödl 1999, Haase & Pröhl 2002, Gardner & Graves 2005, Graves et al. 2005). The pattern contrasts with the co-occurring Dendrobates auratus, which is more active in the late afternoon. Other research hypothesized seasonal reproductive patterns or temporal variation in survival of eggs and tadpoles; calling and breeding activity correlates with the amount of precipitation and the availability of water (Donnelly 1989, Gardner & Graves 2005). We studied here the sensitivity of locomotor performance in Oophaga pumilio to the loss of water by evapotranspiration to low and moderate levels of dehydration.

Materials and methods

At La Selva Biological Station in northern Costa Rica, we captured 36 adult frogs, half during February 2010 and half during August 2012. We compared the locomotor performance of frogs that were fully hydrated (N = 12) and frogs that had lost 10% (N = 12) as well as 15% (N = 12) of their initial weight. Frogs were dehydrated by exposing them to a constant speed (6 m/s) wind current while sitting in a tunnel. They were then placed on a flat surface and covered with a black and small chamber. The relative humidity and the environmental temperature were recorded with a termohygrometer in the chamber. The body temperature was also measured with an infrared thermometer prior just before each experiment: one of us retired the chamber, and prodded the frog urostyle to stimulate it until obtaining 15 successive jumps. A pen was used to mark take-off and landing locations. After the frog completed 15 jumps, we measured the jumping distances and, again, the air relative humidity as well as the environmental and the frog’s body temperature. Following Rogowitz et al. (1999), we estimated the average, maximum, and cumulative distances jumped, as three estimates of locomotor performance. At the end of the experiment, frogs were

80 rehydrated over a water saturated towel. Bivariate correlations and principal component analysis were conducted to reduce the multicollinearity among the climatic variables. The effect of dehydration on locomotor performance was tested using ANCOVA with BCI as a covariate.

Results

During the experiments, relative humidity varied between 68–96%, and environmental and body temperatures between 25–30°C. Both variables were highly correlated (Pearson correlation coefficient > -0.9, P < 0.001 in all cases). Because the snout-vent length and the weight of frogs were highly correlated (Pearson correlation coefficient = 0.883, P = <0.001), we used their residuals as our estimate of relative weight: the body condition index (BCI). BCI and relative humidity were correlated with maximum jumping distance (MJD; Pearson correlation coefficient r = 0.442 and -0.516 respectively, P < 0.01) and cumulative jumping distance (CJD; r = 0.372 and -0.389 respectively, P < 0.05). Therefore, we controlled for the effects of both variables using the residuals of the corresponding linear regressions in the subsequent analyses (Figure 1). The body temperature and the time to reach the target dehydration did not affect the MJD or the CJM.

Hydrated and dehydrated individuals did not show appreciable signals of fatigue (Figure 2). The effects of desiccation treatment and BIC were significant over the residuals (from its relationship with relative humidity) of MJD (ANCOVA F = 6.993, gl = 3, P = 0.001); but were not over the residuals of CJD (ANCOVA F = 3.583, gl = 3, P = 0.024). Our results support a general reduction in the maximum jumping performance and the cumulative jumping performance when frogs are dehydrated to 90% (Figure 2, Figure 3).

Discussion

The reduction in locomotor performance after dehydration has been described in many other frog species such as Eleutherodactylus (Beuchart et al. 1984, Rogowitz et al. 1999), Rhinella (Preest & Pough 1989), and Lithobates (Moore & Gatten 1989). Eleutherodactylus antillensis and E. portoricensis present a reduction of near 20% in the maximum jumping distance after losing a 25% of their body water (Beuchart et al. 1984). In Eleutherodavtylus coqui and E. cooki, there is a decline between 15 – 37 % in the maximum distance per jump and 11 – 36 % in the cumulative distance jumped when frogs are dehydrated between 17 and 29% (Rogowitz et al. 1999). Dehydration of 20% and 30% in Lithobates pipiens and L. clamitans frogs does not have a negative effect over a short term jumping ability, but it reduces their distance traveled after five minutes, as well as in the time to fatigue (Moore & Gatten 1989). Dehydration between 15 and 30% also reduces the distance traveled during 10 minutes in Anaxyrus americanus between almost 20 and 35 m (Preest & Pough 1989).

None of the previous studies, however, found evidence of a negative effect of 10% or lower dehydration on the locomotor performance of these frogs (Moore & Gatten 1989, Preest & Pough 1989, Rogowitz et al. 1999); frogs are indeed known to experience fluctuations of about 10% of hydrated mass under usual conditions (Beuchart et al. 1984, Moore & Gatten 1989, Rogowitz et al. 1999 and references therein).

Our study shows that jumping performance in O. pumilio is negatively affected by moderate levels of dehydration. A reduction in near 7 cm (± 5.6) and 37 cm (± 39.7) in the maximum and cumulative jumping distance were evident at 10% of dehydration. Jumping performance, as we measured it, has a straightforward connection with escaping from predators. Although movement is also required during feeding, courtship, and parental care, those activities hardly require the upper physiological limits of muscle contraction. Also, frogs might simply stop their normal activities and retreat to more humid shelter when exposed to moderate levels of dehydration. That scenario would explain why this species

82 appear to restrict most of their reproductive behaviour to the early morning (Young 1979, Pröhl 1997, Graves, 1999, Phröl & Höld 1999, Haase & Pröhl 2002, Gardner & Graves 2005, Graves et al. 2005). Periods of activity might be indeed very short in this species. Haase & Pröhl (2002) recorded that females O. pumilio from southeastern Costa Rica spent near 85.1% of their daytime without displacement, 9% in locomotion, and barely 1.3% feeding.

Other sources of evidence supports that environmental humidity constrains activity in this species. They disproportionally use perches with lower exposure to UV radiation (Han et al. 2007, Kats et al. 2012), perhaps reducing the risk of desiccation. They also avoid moving through pastures (Robinson et al. 2013); although individuals captured in pastures and forest did not differ in endurance tests, frogs captured in pastures were larger in body size. Large body size is linked to lower evaporative water loss in many other frogs (e.g., Schmid 1965, Nevo 1973), including species of Oophaga (Galindo et al. in Prep.).

Summing up, we found no evidence of physiological adaptations in O. pumilio with regard to jumping performance and dehydration. Instead, the species seem quite sensitive to water loss, which is consistent with field observations of constrained activity in both time and space. We still do not know the extent of geographic variation within and among species in this trait, and whether it has the potential to affect evolutionary processes at different spatial scales, such as gene flow or speciation.

Figure 1. Relationship between body condition index (BCI) and residuals of relative humidity percentage with the maximum jumping distance (MJD) and the cumulative jumping distance (CJD).

Figure 2. Jumping distances during consecutive trials in 10, 20 or 30 individuals of the poison frog Oophaga pumilio, when fully hydrated (above), or after losing 10% (middle) and 15%(lower) of their water content.

Figure 3. Effect of dehydration in the maximum (rMJD) and the cumulative jumping distance (rCJD) in the frog Oophaga pumilio. Points represent the average values and bars denote the 95% confidence intervals. Jumping performance is presented as regression residuals after controlling the effect of relative humidity and the body index condition (BIC).

Literature cited

Beuchat, C. A., Pough, F. H., & Stewart, M. M. (1984). Response to simultaneous dehydration and thermal stress in three species of Puerto Rican frogs. Journal of Comparative Physiology B, 154(6), 579–585.

Donnelly, M. A. (1989). Reproductive phenology and age structure of Dendrobates pumilio in northeastern Costa Rica. Journal of Herpetology, 23(4), 362–367.

Gardner, E. A., & Graves, B. M. (2005). Responses of resident male Dendrobates pumilio to territory intruders. Journal of Herpetology, 39(2), 248–253.

Graves, B. M. (1999). Diel activity patterns of the sympatric poison dart frogs, Dendrobates auratus and D. pumilio, in Costa Rica. Journal of Herpetology, 33(3), 375–381.

Haase, A., & Pröhl, H. (2002). Female activity patterns and aggressiveness in the strawberry poison frog Dendrobates pumilio (Anura: Dendrobatidae). Amphibia-Reptilia, 23(2), 129–140.

Han, B. A., Kats, L. B., Pommerening, R. C., Ferrer, R. P., Murry‐Ewers, M., & Blaustein, A. R. (2007). Behavioral Avoidance of Ultraviolet‐B Radiation by Two Species of Neotropical Poison‐Dart Frogs. Biotropica, 39(3), 433–435.

Kats, L. B., Bucciarelli, G. M., Schlais, D. E., Blaustein, A. R., & Han, B. A. (2012). Ultraviolet radiation influences perch selection by a Neotropical poison-dart frog. PloS one, 7(12), e51364.

Köhler, A., Sadowska, J., Olszewska, J., Trzeciak, P., Berger-Tal, O., & Tracy, C. R. (2011). Staying warm or moist? Operative temperature and thermal preferences of common frogs (Rana temporaria), and effects on locomotion. The Herpetological Journal, 21(1), 17–26.

Moore, F. R., & Gatten Jr, R. E. (1989). Locomotor performance of hydrated, dehydrated, and osmotically stressed anuran amphibians. Herpetologica, 45(1), 101–110.

Nevo, E. (1973). Adaptive variation in size of cricket frogs. Ecology, 54(6), 1271– 1281.

Perry, G., LeVering, K., Girard, I., & Garland, T. (2004). Locomotor performance and social dominance in male Anolis cristatellus. Animal Behaviour, 67(1), 37–47.

Pough, F. H., Andrews, R. M., Cadle, J. E., Crump, M. L., Savitzky, A. H., & Wells, K. D. (2004). Herpetology. Temperature and water relations, 231 – 268. 3rd ed. Prentice Hall, New Jersey.

Prates, I., Angilleta Jr, M. J., Wilson, R. S., Niehaus, A. C., & Navas, C. A. (2013). Dehydration hardly slows hopping toads (Rhinella granulosa) from xeric and mesic environments. Physiological and Biochemical Zoology, 86(4), 451–457.

Preest, M. R., & Pough, F. H. (1989). Interaction of temperature and hydration on locomotion of toads. Functional Ecology, 3(6), 693–699.

Pröhl, H. (1997). Territorial behaviour of the strawberry poison-dart frog, Dendrobates pumilio. Amphibia-Reptilia, 18(4), 437–442.

Pröhl, H. (2005). Territorial behavior in dendrobatid frogs. Journal of Herpetology, 39(3), 354–365.

Pröhl, H., & Willink, B. (2015). Ecología y comportamiento de las ranas venenosas del género Oophaga en Costa Rica y Panamá. Alytes, 32(32), 31-45.

Robinson, D., Warmsley, A., Nowakowski, A. J., Reider, K. E., & Donnelly, M. A. (2013). The value of remnant trees in pastures for a Neotropical poison frog. Journal of Tropical Ecology, 29(04), 345–352.

Schmid, W. D. (1965). Some aspects of the water economies of nine species of amphibians. Ecology, 46(3), 261–269.

Thorson, T., & Svihla, A. (1943). Correlation of the habitats of amphibians with their ability to survive the loss of body water. Ecology, 24(3), 374–381.

Titon Jr, B., Navas, C. A., Jim, J., & Gomes, F. R. (2010). Water balance and locomotor performance in three species of Neotropical toads that differ in geographical distribution. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology, 156(1), 129–135.

Titon, B. J., & Gomes, F. R. (2015). Relation between Water Balance and Climatic Variables Associated with the Geographical Distribution of Anurans. PloS one, 10(10), e0140761.

Tracy, C. R. (1976). A model of the dynamic exchanges of water and energy between a terrestrial amphibian and its environment. Ecological Monographs, 46(3), 293–326.

Young, A. M. (1979). Arboreal movement and tadpole-carrying behavior of Dendrobates pumilio Schmidt (Dendrobatidae) in northeastern Costa Rica. Biotropica, 11(3), 238–239.

Amargal Biological Station, Chocó, Colombia.

DOES FINER ENVIRONMENTAL HETEROGENEITY PROMOTE MICROEVOLUTIONARY PROCESSES? A CASE STUDY IN THE KÓKÓE-PÁ POISON FROG, OOPHAGA HISTRIONICA, IN COLOMBIA

Diana M. Galindo-Uribe, Adolfo Amézquita, & Heike Pröhl

Introduction

Heterogeneity in topography and vegetation structure leads to concomitant variation in the quality and availability of microclimates and limiting resources for organisms. In animals, the differential use of areas with appropriate conditions for physiological processes and ecological performance may shape the dispersal of individuals and result in the spatial structuring of populations (Slatkin 1987, Endler 1992, 1993, Via 1999, Svenning 2001, Théry 2001, Pringle et al. 2003, Mayer et al. 2009). Genetic differentiation within metapopulations is an eventual consequence of spatial structuring and is mediated by random processes, namely founder genetic composition, genetic drift, and gene flow, and by natural selection acting at a local level (Hartl & Clark 2007). Genetic drift in small populations results in the reduction of genetic variation whereas migration tends to increase the genetic homogenization (Avise 1994). High migration rates, in turn, can inhibit local genetic differentiation and reduce the potential of local populations for adaptation (Tallmon et al. 2000). Disentangling the mechanisms underlying the relationship between the environmental heterogeneity and genetic variation is the central focus of research areas such as landscape genetics (Manel et al. 2003).

Amphibians are viewed as highly substructured with regard to their genetic pools (Shaffer et al. 2000). Most species exhibit patchy distributions and strong site fidelity, which is attributable to physiological constrains and to the spatial distribution of reproductive resources. Their permeable skin and ectothermic condition make them susceptible to desiccation and may constrain their ability to disperse (Sinsch 1990, Blaustein et al. 1994) or limit their dispersal through sites with conditions of high humidity (Funk et al. 2005). Their reproductive behaviors and the spatial distribution of the resources used by different life stages arguably

91 affect their distribution, dispersal, and thereby genetic differentiation (Zamudio & Wieczorek, 2007). Therefore, amphibians are ideal subjects for studying the effect of environmental variables on demographic gene flow, and micro evolutionary processes. In spite of this, the contribution of environmental heterogeneity to population structuring has been studied mainly in widespread temperate amphibians with breeding systems associated to ponds or running waters (Newman & Squire 2001). We also ignore the spatial scale at which genetic structuring may occur when habitat heterogeneity constrains dispersal (Zamudio & Wieczorek 2007). Much less is known about Neotropical species with terrestrial or semiterrestrial reproduction.

The Neotropical poison frogs, Dendrobatidae, have their feeding, reproductive and sheltering sites within strongly defended territories (Crump 1972, Donelly 1989a, Pröhl 2005). Compared to other frogs, they also exhibit complex courtship and parental care, and an astonishing intraspecific diversity in aposematic coloration and escape behaviors (Myers & Daly 1976). The spatial heterogeneity in the quality and availability of defended resources may well affect their spatial behavior and promote genetic structuring within a single species (Pröhl 2005). In particular, species in the genus Oophaga are among the most studied. Male adults are highly territorial (Goodman 1971, Crump 1972, Forester et al. 1993, Pröhl 1997, Pröhl & Berke 2001, Haase & Pröhl 2002), and exhibit remarkable ability to home when displaced from their territories (McVey et al. 1981). Both sexes provide parental care (Pröhl & Hödl 1999). After hatching, the tadpoles are transported in the female back to tadpole rearing sites, which can be the small body waters present in the leaf axils of palms and heliconiads (van Wijngaarden & Bolaños 1992), or bromeliads located at high elevations (Young 1979, Myers et a. 1984). Once there, the tadpoles are exclusively fed with unfertilized eggs laid by the mother (van Wijngaarden & Bolaños 1992, Brust 1993). Therefore, both space use and successful reproduction are clearly affected by the spatial distribution of their reproductive resources (Donelly 1989b, 1989c, Pröhl & Berke 2001, Haase & Pröhl 2002).

More recent studies have showed the contribution of environmental factors such as the availability of perches and the light conditions to explain the coexistence of different morphs (Pröhl & Ostrowski 2011, Robinson et al. 2013, Willink et al. 2013). The role of abiotic noise in shaping the call frequency of Oophaga histrionica and Oophaga lehmanni populations living alongside streams has been also addressed (Vargas-Salinas & Amézquita 2013). However, we still ignore whether environmental factors suffices to promote genetic structuring at fine spatial scales. We studied the Kókóe-Pá, Oophaga histrionica, in northwestern Colombian lowlands (see Silverstone 1975, Myer & Daly 1976, Lötters et al. 1999), where these frogs are discontinuously distributed; they form breeding aggregations mostly associated to hill-tops. In our study area, we are aware of the existence of such aggregations or demes, separated each other by less than two kilometers of continuous forest, during more than 10 years (Amézquita pers. obs.). The evident climatic and topographic heterogeneity (Figure 1) allowed to evaluate the role of environmental heterogeneity in shaping demographic and genetic sub-structuring at fine spatial scales (<3 km). We first predicted that the presence and density of individuals within the study area was correlated with concomitant variation in microclimatic and topographic conditions. We then predicted and expected to find evidence of reduced gene flow and genetic structuring associated with the spatial distribution of the demes.

Materials and methods

Study organisms

Oophaga histrionica inhabits lowland rain forests on the Pacific versant of Colombia n west Andes (Myers & Böhme 1996). Our study area is located at the Amargal Biological Station, Municipio de Nuquí, Departamento del Chocó. At the Camino del Yupe in the Chocó region, Oophaga histrionica males are territorial. Individuals seek for refuge in beneath different substrates like leaves, fallen logs or among the roots of stilt palms. Males call from elevated perches (of 56 cm as

93 average and 163 cm as maximum distance above the ground) like tree trunks, fallen logs, and branches or herb leaves. Courtship is complex but amplexus is absent (Silverstome 1973). Adult individuals have been repeatedly observed in the same territories along 2-3 consecutive years (Gómez & Amézquita in Prep.). We sampled three demes (areas with the presence and aggregation of frogs) separated from each other by map distances between 1.5–3 km (Figure 1). We found no evidence of heavy logging or forest fragmentation between consecutive demes.

Microclimate and frog abundance

During a field trip in June 2011, we chose 91 sampling points within and outside each deme; they were selected to include sites with and without frogs, as well as variable elevation and slopes (Figure 2). We fortuitously selected nine of these points to set up MSR145 data loggers at about 1.5 m above the ground. They stored measurements of temperature (°C), relative humidity (%), and intensity of light (counts/lux) every minute during a period of 24 hours, beginning at 13:30 hour, when frog activity is appreciable low, and ending 24 hours later. Within 20 m of each sampling point, we counted the number of calling males, the number of visible frogs, and the number of bromeliads because the spatial distribution of Oophaga females may be conditioned by its spatial distribution, and the distribution of males by the spatial distribution of females (Pröhl & Berke 2001, Pröhl 2002, 2005, Meuche et al. 2013). We discarded night time microclimatic data because these frogs are strictly diurnal and because night measurements were highly homogeneous across sampling points (Figure 3). From daytime data, we calculated average and range values of temperature (T), relative humidity (RH), and light intensity (L) to characterize each site. We used Principal Component Analysis (PCA) with Varimax rotation to reduce potential redundancy between predictor variables. The range in relative humidity (RangeRH) and light conditions (RangeL), the number of bromeliads, and the elevation were selected as the variables that better summarized other predictors and therefore used in subsequent analyses (Appendix 1). To identify the best (if any) subset of variables that predicted the

94 presence and number of frogs we ran generalized linear models (glm) with binomial and poisson link functions. The best model was selected using the Akaike Information (AIC) criterion as implemented in the bestglm r (R Core Team 2015) package.

DNA extraction and microsatellite amplification

We collected buccal swabs from a total of 49 frogs encountered in the three demes (Figure 1). All frogs were immediately released within leaf litter at their capture site. The samples were preserved in 96% ethanol in the field, and then stored at - 20°C in the laboratory. Genomic DNA was extracted using the QIAGEN DNeasy Blood and Tissue kit (Qiagen, Valencia, CA). DNA dilutions were used as templates for amplification of microsatellite loci via the polymerase chain reaction (PCR).

We genotyped individuals for three di-microsatellite loci (Dpum14, Dpum63, Oop- O1) and 12 tetra-microsatellite loci (Dpum12, Dpum13, Dpum110, Oop-B8, Oop- H5, Oop-E3, Oop-G5, Oop-B9, Oop-C11, Oop-C3, Oop-01, Oop-D4), designed by Wang & Summers (2009), and Hauswaldt et al. (2009). Forward primers for each locus were 5’-labelled with fluorescent tags (6-FAM, HEX) for O. pumilio. The thermocycling profile for all microsatellite loci involved an initial denaturation at 95°C for 5 minutes, followed by 35 cycles of denaturation (30 s at 95°C), annealing (90 s at 60°C) and elongation (45 s at 72°C), and a final elongation step at 60°C for 30 min. Amplified PCR products were analyzed on the ABI 3500 Genetic Analyzer. Fragments were sized and analyzed using GeneMapper software (Applied Biosystems, Foster City, CA, USA).

Microsatellite diversity, population substructure, and gene flow

We used GenAlEx 6.5 (Peakall & Smouse 2012) to estimate the effective number of alleles (Ae), the observed (HO), and expected (HE) heterozygosities and the number of private alleles (AP) for each locus across the population. We used GENEPOP on the

95 web (Raymond & Rousset 1995) to calculate one locus F-Statistics (FIS, FIT, FST) following standard ANOVA as in Weir and Cockerham (1984), deviations from Hardy-Weinberg equilibrium (HWE), and the pairwise probability of linkage disequilibrium (LD) for each locus within demes. Significant deviations from HWE were calculated using 1,000 Markov chain steps, 1,000 dememorization steps and 500 batches. LD within each deme and between all pairs of loci was calculated using a likelihood-ratio test with 10,000 permutations.

We calculated pairwise levels of population differentiation in allele frequencies using FST values (Weir and Cockerham 1984) and tested their significance by 10,000 permutations as implemented in Arlequin 3.5.2.2 (Excoffier et al. 2005). We conducted this analysis both using all loci and excluding the loci out of HWE and LD. Because the presence of null alleles can also bias population differentiation estimates, we used FreeNA (Chapuis & Estoup 2007) with the Expectation Maximization (EM) algorithm of Dempster et al. (1977) in order to estimate the null allele frequency and FST values according to Weir’s (1996) with the ENA correction.

We implemented a Bayesian assignment test in STRUCTURE v2.3.1 (Pritchard et al. 2000) to evaluate whether sampled individuals could be assigned to any of three genetic clusters according with the demes observed in the field. We used the polymorphic loci for an admixture model with correlated allele frequencies between populations, with a burn-in of 50,000 generations, a Markov chain of 500,000 and the LOCPRIOR option (Hubisz et al. 2009). We first did 5 runs with 3 predefined K genetic clusters. Because the presence of genetic structure supported the assignment of individuals into two or three K, we determined the most likely genetic population structure following the method of Evanno et al. (2005) over 20 runs with a burn-in of 100,000 generations, a Markov chain of 1,000,000, 1-6 K and the LOCPRIOR option, on the Structure Harvester web service (http://taylor0.biology.ucla.edu/structureHarvester/; Earl & vonHoldt 2012). Because we had violated basic STRUCTURE HWE and LD assumptions we repeated the procedure excluding the polymorphic loci in disequilibrium. We got a mean of

96 permutated matrices across the 20 replicates in the program CLUMPP (Jakobsson & Rosenberg 2007) with the FullSearch algorithm. We developed graphical results with Structure Plot v2.0 (Ramasamy et al. 2014) in the web (http://omicsspeaks.com/strplot2/). We also used the program InsTruct (Gao et al. 2007) which is free of HWE assumptions. We also run the program with 1-6 K, and we used the Deviance Information Criterion to select the optical K value.

We estimated inbreeding coefficients and recent migration rates of one generation of adults between demes using the nonequilibrium Bayesian method BAYESASS 3 (Wilson & Rannala 2003). We performed five independent runs with random seeds from 10-100,000, using a Monte Carlo Markov chain (MCMC) calculation of 10 million iterations (1 million discarded as burn-in) and sampling frequency of 2,000 iterations. We compared the convergence of the different runs. The adjusted delta values used were 0.4 (migration), 0.80 (allele frequencies) and 1 (inbreeding), with acceptance rates of 33%, 30%, and 36%, respectively.

Results

Most variation in microclimatic measurements was observed at midday (Figure 2). The average conditions across the area varied between almost 90 – 99.7% in relative humidity, between 23.7 – 28.3°C in temperature, between 2 – 5,160 counts/lux in intensity of light, between 40 – 169 m in elevation, and between 0 – 43 in the number of available bromeliads. All diurnal climatic conditions were intercorrelated to some degree (Appendix 1). The larger the average (Av) and the range of temperature (AvT) and light (AvL) the lower the average relative humidity (AvRH). A total of 40 frogs were counted in all the sampling areas. 25 % of sampled areas contained frogs. Fourteen percent of the sampling areas contained one individual, 7% two individuals, 2% three individuals, and only two sampled areas contained 4 and 5 individuals (Figure 3). The presence as well as the number of individuals was explained by RangeL, the number of bromeliads and elevation

(R2 = 0.380 and R2 = 0.416 respectively; Figure 4, Table 1).

Microsatellite diversity and population structure

Four loci displayed a single allele (Oop-B8, Oop-D4, Oop-O1 and Dpum14) and were excluded in subsequent analyses. The other 11 microsatellite loci exhibited between two and 11 alleles (A). The number of effective alleles (Ae) varied between almost one and five, the observed heterozygosity (HO) was 0.407 (0.042 – 0.827), and the expected heterozygosity (HE) 0.445 (0.04 – 0.78). Overall population and loci F-statistics were 0.117 (FIS), 0.173 (FIT), and 0.063 and 0.062 respectively for FST with and without null alleles correction (Table 2). The average HO for populations was 0.41 (± 0.05) and the HE of 0.45 (± 0.04). Private alleles varied from one to five among loci. Significant deviations from HWE were identified for loci Dpum109 and OopE3 within Deme1; for OopF1, Dpum110 and OopE3 within Deme2; and OopE3 within Deme3. Linkage disequilibria (LD) was significant for loci Dpum109- OopH5 within Deme1 and Dpum110-OopE3 within Deme3 (Table 3). Most loci (77%) had low frequency of null alleles (r < 0.05) on average across demes, and the other loci had intermediate frequencies (r > 0.05 and < 0.20). Alleles with significant deviation from HWE also showed intermediate or high frequencies of null alleles in at least one deme with the exception of OopC3 (results not shown).

Although the pairwise FST values, obtained with both Arlequin and FreeNA, varied when excluding or not the loci out of HWE expectations, all values were significant and supported some degree of genetic differentiation between demes (Table 3). Both STRUCTURE and the Evanno et al. (2005) method supported two K as the most likely number of genetic clusters within our population (Figure 5, Appendix 2). The inbreeding coefficients obtained with BayesAss were higher for Deme1 (0.380 + 0.134) and Deme3 (0.261 + 0.167) in comparison to Deme2 (0.146 + 0.055). The migration rates (Table 5) were asymmetric. The proportion of individuals coming from the same Deme was higher in Deme 2 (0.95) than in deme 1(0.79) or 3 (0.75).

Discussion

Our evidence suggests that Oophaga histrionica disproportionally use sites with larger variation in light intensity, at higher elevations, and with higher abundance of bromeliads. Light conditions may affect frog distribution by affecting the spatial distribution of reproductive, feeding and sheltering resources. A positive relationship between light availability and the density of tank bromeliads has been suggested (Pittendrigh 1948, Théry 2001). In our study site, the presence and the abundance of frogs was partly explained by the amount of bromeliads, needed in Oophaga species as tadpole rearing sites and indeed found in the core areas of Oophaga pumilio females (see Pröhl & Berke 2001). Bromeliads may also be more abundant at hilltops, where competition for light should be less intense. The light environment may also correlate with the abundance of feeding resources such as ants. Some species of Paratrechina ants, a putative source of frog pumiliotoxins (Saporito et al. 2004), seem to be more abundant at relatively dry and well illuminated places (Mühlenberg et al. 1977). Formicinae ants are also more abundant in forest gaps during the rainy season (Richards & Windsor 2007, but see Feener & Schupp 1998). Lastly, vegetation coverage and light conditions may also affect the demography and dispersal of various Neotropical palm species (Svenning 2001). Individuals of O. histrionica were very often seen to hide within the humid and dark palms roots at our study sites (Galindo-Uribe pers. obs.). To disentangle the direct and indirect effects of light on the physiology and ecology of these frogs probably demands carefully designed manipulative studies, which were beyond the scope of this work.

Microsatellite diversity, population substructure, and gene flow

The microsatellite data analyses support some degree of inbreeding within and appreciable genetic structuring between the demes. The population structure at finer spatial and temporal scales may reveal the contribution of genetic drift in the presence of gene flow and the dynamics of population size (Hartl & Clark 1997).

The asymmetrical gene flow patterns among demes suggest a metapopulation scenario (see Mayer et al. 2009). Alternatively, a reduction in gene flow among demes, as opposed to within them, may be due to the differential survival and reproductive success between frogs occupying high versus low quality habitats.

At the scale of several kilometers, amphibians generally considered to exhibit a high degree of genetic structuring (Newman & Squire 2001). Studies incorporating finer spatial scales in amphibians have been conducted with molecular markers that vary in resolution, such as as allozymes, SSCP and microsatellites; most of them on amphibians associated to breeding ponds or running waters (Zamudio & Wieczorek 2007). Some studies did not find genetic differentiation at distances below 5 km (Tallmon et al. 2000); others found significant but low genetic structuring in the presence of high levels of gene flow (Seppa & Laurila 1999, Rowe et al. 2000, Newman & Squire 2001, Jehle et al. 2005, Zamudio & Wieczorek 2007, Martínez-Solano & González 2008) or high levels of population differentiation with relatively isolated subpopulations in the absence (Routman 1993, Gascon et al. 1996, Driscoll 1998, Shaffer et al. 2000, Kraaijeveld-Smit et al. 2005) or in the presence of apparent dispersal barriers (Hitchings & Beebee 1997). Our data suggest significant but low genetic structuring in the presence of high levels of gene flow. It was comparable to genetic structuring found in Ambystoma maculatum (Zamudio & Wieczorek 2007), Epidalea calamita (Rowe et al. 2000), Rana temporaria (Seppä & Laurila 1999), Rana sylvatica (Newman & Squire 2001), or in Litoria aurea in continuous habitats (Burns et al. 2004).

We found evidence of correlation between habitat heterogeneity and the frogs’ discontinuous distribution; in parallel, we show some degree of genetic differentiation at very low spatial scales. Although inferring a mechanistic connection between both patterns is tempting, studies on the covariation of vital resources for frogs, such as ants, bromeliads and palms, are needed to further support this case. Also, demographic or manipulative studies identifying the effects of habitat heterogeneity on individual survival or reproductive success should

100 provide direct evidence to clarify whether habitat heterogeneity is enough as to structure demes and initiate genetic differentiation in these frogs.

101

1.5 Km

1 Km

Demes

Figure 1. Study area and location of the three demes of Oophaga histrionica in the Colombian Chocó region.

102

Figure 2. Circles denote the spatial distribution of sampling points: white and red indicate the absence or presence of frogs. Red dot size represent the number of individuals, from 1–5. Darker pixels represent areas with lower elevation and lighter pixels areas with higher elevation, between 0–170 m above sea level.

103

° C) 27

Temp ( 25

23 100 )

60 Light ( cnt 20 0 100

96 RH (%) RH 92

0:00 10:00 20:00 Time (Hours)

Figure 3. Temperature (°C), light (cnt) and percentage of relative humidity (% RH) on the sampled points along a 24-hours cycle.

104

Figure 4. Environmental variables that contribute to explain the presence (above) and the abundance (n, below) of Oophaga histrionica frogs among the sampling areas

105

Figure 5. Assignment probability of individual (Oophaga histrionica) frogs after the alignment in Clumpp of 20 STRUCTURE runs with three predefined demes and K2 selected as the best likely genetic cluster according with the Evanno et al. (2005). Upper graph represent the results using all loci. Lower graph represent the results excluding loci out of HWE or LD.

Table 1. Regression coefficients of the best model using the generalized multiple regression (glm) with Range in relative humidity (RangeRH), Range in light amount (RangeRH), Elevation (m) and number of bromeliads (n) as predictor variables and, the presence- absence (Occupation) of frogs and the number of Individuals (n) as response variables. In bold significant p.values.

Response Predictors Estimate Std. Error z value p.value (Intercept) -10.320 3.401 -3.034 0.002 Occupation RangeL 0.001 0.000 2.294 0.022 (presence Elevation 0.060 0.024 2.530 0.011 /absence) BROM 0.106 0.037 2.842 0.004 (Intercept) -5.964 1.572 -3.793 0.000 RangeL 0.000 0.000 3.761 0.000 Individuals (n) Elevation 0.033 0.011 2.920 0.004 BROM 0.052 0.012 4.179 0.000

107

Table 2. Microsatellite statistics at each locus across all demes: number of alleles

(A), number of effective alleles (Ae), observed heterozygosity (HO), expected heterozygosities (HE), unbiased expected heterozygosity (uHE), F statistics (FIS, FIT and

FST), FST values corrected by Null Alleles frequencies (FST-ENA), deviations from Hardy- Weinberg equilibrium (HWE), and the number of private alleles (AP).

Locus A Ae H O H E F IS F IT F ST F ST-ENA AP Dpum109 7 2 0.390 0.481 0.220 0.313 0.119 0.093 1(D2), 2(D3) Dpum110 10 5 0.680 0.780 0.162 0.187 0.030 0.031 3(D1) Dpum13 3 1 0.215 0.247 0.162 0.189 0.032 0.041 1(D3) Dpum44 7 3 0.730 0.696 -0.012 0.062 0.074 0.076 1(D1), 1(D3) Dpum63 2 1 0.249 0.260 0.080 0.124 0.048 0.049 Dpum92 3 1 0.042 0.040 -0.015 -0.002 0.013 0.011 2(D1) OopB9 7 4 0.827 0.716 -0.119 -0.065 0.048 0.048 1(D1) Oop-C11 6 3 0.580 0.625 0.105 0.213 0.121 0.122 1(D1), 1(D3) Oop-C3 2 1 0.042 0.073 0.456 0.499 0.079 0.141 1(D1) Oop-E3 10 2 0.248 0.557 0.592 0.624 0.077 0.058 1(D1), 2(D2), 2(D3) Oop-F1 11 4 0.642 0.722 0.139 0.193 0.063 0.066 2(D2), 2(D3) Oop-G5 2 1 0.071 0.064 -0.086 0.004 0.083 0.080 1(D3) Oop-H5 9 2 0.579 0.529 -0.060 -0.073 -0.013 -0.013 1(D1), 1(D2), 1(D3) All loci 2.386 0.407 0.445 0.117 0.173 0.063 0.062

108

Table 3. Number of individuals sampled per deme, average observed heterozygosity (HO), average expected heterozygosity (HE), and loci with significant departures from Hardy-Weinberg equilibrium (HWE) and linkage equilibrium (LD) within each deme.

Deme n H O H E HW disequilibrium LD 1 18 0.38 0.46 DPUM109, OOPE3 DPUM109, OOPH5 0.08 0.08 2 15 0.39 0.42 OOPF1, DPUM110, OOPE3 - 0.08 0.08 3 16 0.45 0.46 OOPE3 DPUM110, OOPE3 0.09 0.08

109

Table 4. Matrix of pairwise population FST values (lower) between demes of O. histrionica frogs as calculated with Arlequin and FreeNA softwares, including and excluding loci out of HWE in the case of the first one and, with and without correction by null alleles frequencies. * Significance p values at the level of 0.05.

Arlequin wiythout Arlequin with all loci in HWE and FreeNA not using FreeNA using loci LD ENA ENA Demes 1 2 3 1 2 3 1 2 3 1 2 3 1 0.000 * * 0.000 * * 0.000 0.000 2 0.085 0.000 * 0.070 0.000 * 0.069 0.000 0.068 0.000 3 0.082 0.051 0.000 0.119 0.073 0.000 0.047 0.074 0.000 0.048 0.069 0.000

110

Table 5. Means and standard deviations of the posterior distributions of the migration rate into each population of the frog Oophaga histrionica. The demes from which each individual was sampled are listed in the rows. The demes from which they migrated are listed in the columns. Values along the diagonal are the proportions of individuals derived from the source deme each generation. Migration rates > 0.05 are in bold.

Demes 1 2 3 0.7905 0.1380 0.0715 1 (0.0533) (0.0548) (0.0595) 0.0247 0.9474 0.0279 2 (0.0231) (0.0334) (0.0263) 0.0296 0.2183 0.7522 3 (0.0362) (0.0700) (0.0633)

111

Appendix 1.

1.1. PCA correlation matrix between exploratory variables after a Varimax rotated function.

PCA Rotated Components FACTORS Comunalities 1 2 3 4 AvRH -.883 -.261 -.126 .144 .885 RangeRH .844 .397 .180 -.104 .914 AvT .757 -.128 -.448 .189 .827 RangeT .559 .535 .323 -.158 .728 RangeL .063 .917 -.074 -.050 .853 AvL .333 .884 -.131 .018 .910 Bromeliads .082 -.135 .875 .113 .803 Elevation -.076 -.031 .088 .971 .956 % Var.Accum. 43.921 59.675 73.584 85.941

112

1.2. Pearson’s and Spearman’s correlation coefficients and significance values between exploratory variables after 1000 bootstrapping. Significant bilateral correlation at 0.01** and 0.05*.

Pearson\Spearman Correlation Coefficients FACTORS AvRH AvT AvL RangeRH RangeT RangeL BROM Elevation AvRH p\r -.338** -.297* -.889** -.631** -.420** -.169 .314* sig. .006 .017 .000 .000 .001 .182 .011 AvT p\r -.478** -.125 .348** .093 -.133 -.065 -.098 sig. .000 .326 .005 .464 .295 .612 .443 AvL p\r -.467** .250* .264* .457** .732** -.122 -.103 sig. .000 .047 .035 .000 .000 .335 .417 RangeRH p\r -.882** .419** .591** .665** .402** .245 -.346** sig. .000 .001 .000 .000 .001 .051 .005 RangeT p\r -.652** .089 .568** .757** .564** .098 -.233 sig. .000 .486 .000 .000 .000 .439 .063 RangeL p\r -.309* .057 .814** .373** .374** .113 -.235 sig. .013 .652 .000 .002 .002 .373 .061 Bromeliads p\r -.086 -.120 -.140 .100 .073 -.067 -.043 sig. .497 .347 .272 .430 .568 .599 .736 Elevation p\r .173 .008 -.082 -.135 -.116 -.138 .093 sig. .172 .947 .520 .288 .363 .276 .463

113

Appendix 2.

2.1. Results of selection criteria for the number of genetic clusters K in STRUCTURE according with Evanno et al. (2005) criteria and Instruct according with the deviance in formation criteria (DIC).

# K Reps Mean LnP(K) Stdev LnP(K) Ln'(K) |Ln''(K)| Delta K 1 20 -1250.58 0.32 NA NA NA 2 20 -1221.20 1.57 29.39 54.15 34.51 STRUCTURE WITH 3 20 -1245.96 13.53 -24.76 22.47 1.66 HWE AND LD LOCI 4 20 -1248.25 11.20 -2.30 12.75 1.14 5 20 -1237.80 10.30 10.45 9.14 0.89 6 20 -1236.49 11.88 1.31 NA NA 1 20 -675.05 0.31 NA NA NA 2 20 -673.92 5.07 1.14 20.42 4.03 STRUCTURE WITHOUT 3 20 -693.20 17.13 -19.28 26.76 1.56 HWE AND LD LOCI 4 20 -685.72 18.84 7.48 4.89 0.26 5 20 -683.12 21.51 2.60 3.80 0.18 6 20 -684.33 16.53 -1.21 NA NA # K Reps Mean LnP(K) Stdev LnP(K) DIC 1 20 -845.30 0.05 1690.61 2 20 -801.02 0.60 1602.04 INSTRUCT WITH 3 20 -782.41 0.68 1564.81 HWE AND LD LOCI 4 20 -769.97 0.72 1539.93 5 20 -765.11 0.57 1530.22 6 20 -763.49 0.55 1526.99

114

Literature cited

Avise, J. C (1994) Molecular Markers, Natural History and Evolution. Chapman & Hall, New York.

Blaustein, A. R., Wake, D. B., & Sousa, W. P. (1994). Amphibian declines: judging stability, persistence, and susceptibility of populations to local and global extinctions. Conservation Biology, 8(1), 60–71.

Brust, D. G. (1993). Maternal brood care by Dendrobates pumilio: a frog that feeds its young. Journal of Herpetology, 27(1), 96–98.

Burns, E. L., Eldridge, M. D., & Houlden, B. A. (2004). Microsatellite variation and population structure in a declining Australian Hylid Litoria aurea. Molecular Ecology, 13(7), 1745–1757.

Chapuis, M. P., & Estoup, A. (2007). Microsatellite null alleles and estimation of population differentiation. Molecular Biology and Evolution, 24(3), 621–631.

Crump, M. L. (1972). Territoriality and mating behavior in Dendrobates granuliferus (Anura: Dendrobatidae). Herpetologica, 28(3), 195–198.

Dempster, A. P., Laird, N. M., Rubin, D. B. (1977). Maximum Likelihood from Incomplete Data via the EM Algorithm. Journal of the Royal Statistical Society. Series B (Methodological), 39(1), 1-38.

Donnelly, M. A. (1989a). Reproductive phenology and age structure of Dendrobates pumilio in northeastern Costa Rica. Journal of Herpetology, 23(4), 362–367.

Donnelly, M. A. (1989b). Effects of reproductive resource supplementation on space-use patterns in Dendrobates pumilio. Oecologia, 81(2), 212–218.

Donnelly, M. A. (1989c). Demographic effects of reproductive resource supplementation in a territorial frog, Dendrobates pumilio. Ecological Monographs, 59(3), 207–221.

Driscoll, D. A. (1998). Genetic structure of the frogs Geocrinia lutea and Geocrinia rosea reflects extreme population divergence and range changes, not dispersal barriers. Evolution, 52(4), 1147–1157.

Earl, D. A., & vonHoldt, B. M. (2012). Structure harverster: a website and program for visualizing structure output and implementing the Evanno method. Conservation Genetics Resources, 4(2), 359-361.

Endler, J. A. (1992). Signals, signal conditions, and the direction of evolution. American Naturalist, 139, S125–S153.

Endler, J. A. (1993). The color of light in forests and its implications. Ecological Monographs, 63(1), 1–27.

115

Evanno, G., Regnaut, S., & Goudet, J. (2005). Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology, 14(8), 2611–2620.

Excoffier, L., Laval, G., & Schneider, S. (2005). Arlequin (version 3.0): an integrated software package for population genetics data analysis. Evolutionary Bioinformatics Online, 1, 47.

Feener Jr, D. H., & Schupp, E. W. (1998). Effect of treefall gaps on the patchiness and species richness of Neotropical ant assemblages. Oecologia, 116(1–2), 191–201.

Funk, W. C., Blouin, M. S., Corn, P. S., Maxell, B. A., Pilliod, D. S., Amish, S., & Allendorf, F. W. (2005). Population structure of Columbia spotted frogs (Rana luteiventris) is strongly affected by the landscape. Molecular Ecology, 14(2), 483– 496.

Gao, H., Williamson, S., & Bustamante, C. D. (2007). A Markov chain Monte Carlo approach for joint inference of population structure and inbreeding rates from multilocus genotype data. Genetics, 176(3), 1635–1651.

Gascon, C., Lougheed, S. C., & Bogart, J. P. (1996). Genetic and morphological variation in Vanzolinius discodactylus: a test of the river hypothesis of speciation. Biotropica, 28(3), 376–387.

Goodman, D. E. (1971). Territorial behavior in a Neotropical frog, Dendrobates granuliferus. Copeia, 1971(2), 365–370.

Haase, A., & Pröhl, H. (2002). Female activity patterns and aggressiveness in the strawberry poison frog Dendrobates pumilio (Anura: Dendrobatidae). Amphibia-Reptilia, 23(2), 129–140.

Hartl, D. L., & Clark, A. G. (1997). Principles of population genetics (Vol. 116). Sunderland: Sinauer associates.

Hauswaldt, J. S., Ludewig A. K., Hagemann S., Pröhl H., & Vences M. (2009). Ten microsatellite loci for the strawberry poison frog (Oophaga pumilio). Conservation Genetics, 10(6), 1935–1937.

Hitchings, S. P., & Beebee, T. J. (1997). Genetic substructuring as a result of barriers to gene flow in urban Rana temporaria (common frog) populations: implications for biodiversity conservation. Heredity, 79(2), 117–127.

116

Hubisz, M. J., Falush, D., Stephens, M., & Pritchard, J. K. (2009). Inferring weak population structure with the assistance of sample group information. Molecular ecology resources, 9(5), 1322–1332.

Jakobsson, M., & Rosenberg, N. A. (2007). CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics, 23(14), 1801–1806.

Jehle, R., Wilson, G. A., Arntzen, J. W., & Burke, T. (2005). Contemporary gene flow and the spatio-temporal genetic structure of subdivided newt populations (Triturus cristatus, T. marmoratus). Journal of Evolutionary Biology, 18(3), 619– 628.

Kraaijeveld-Smit, F. J. L., Beebee, T. J. C., Griffiths, R. A., Moore, R. D., & Schley, L. (2005). Low gene flow but high genetic diversity in the threatened Mallorcan midwife toad Alytes muletensis. Molecular Ecology, 14(11), 3307–3315.

Lötters, S., Glaw, F., Köhler, J., & Castro, F. (1999). On the geographic variation of the advertisement call of Dendrobates histrionicus BERTHOLD, 1845 and related forms from north-western South America. Herpetozoa, 12(1/2), 23–38.

Manel, S., Schwartz, M. K., Luikart, G., & Taberlet, P. (2003). Landscape genetics: combining landscape ecology and population genetics. Trends in Ecology & Evolution, 18(4), 189–197.

Martínez-Solano, I., & Gonzalez, E. G. (2008). Patterns of gene flow and source-sink dynamics in high altitude populations of the common toad Bufo bufo (Anura: Bufonidae). Biological Journal of the Linnean Society, 95(4), 824–839.

Mayer, C., Schiegg, K., & Pasinelli, G. (2009). Patchy population structure in a short‐ distance migrant: evidence from genetic and demographic data. Molecular Ecology, 18(11), 2353–2364.

McVey, M. E., Zahary, R. G., Perry, D., & MacDougal, J. (1981). Territoriality and homing behavior in the poison dart frog (Dendrobates pumilio). Copeia, 1981(1), 1–8.

Meuche, I., Brusa, O., Linsenmair, K. E., Keller, A., & Pröhl, H. (2013). Only distance matters–non-choosy females in a poison frog population. Frontiers in zoology, 10(1), 1.

Mühlenberg, M., Leipold, D., Mader, H. J., & Steinhauer, B. (1977). Island ecology of arthropods. II. Niches and relative abundances of Seychelles ants (Formicidae) in different habitats. Oecologia, 29(2), 135–144.

Myers, C. W., & Böhme, W. (1996). On the type specimens of two Colombian poison frogs described by A.A. Berthold (1845), and their bearing on the locality" Provinz Popayan". American Museum of Natural History, 3185, 1–20.

117

Myers, C. W., Daly, J. W., & Martínez, V. (1984). An arboreal poison frog (Dendrobates) from western Panama. American Museum Novitates, 2783, 1–20.

Newman, R. A., & Squire, T. (2001). Microsatellite variation and fine-scale population structure in the wood frog (Rana sylvatica). Molecular Ecology, 10(5), 1087–1100.

Peakall, R. O. D., & Smouse, P. E. (2012). GENALEX 6.5: GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research-an update. Bioinformatics, 28(19), 2537–2539.

Pittendrigh, C. S. (1948). The bromeliad-Anopheles-malaria complex in Trinidad. I- The bromeliad flora. Evolution, 2(1), 58–89.

Pringle, R. M., Webb, J. K., & Shine, R. (2003). Canopy structure, microclimate, and habitat selection by a nocturnal snake, Hoplocephalus bungaroides. Ecology, 84(10), 2668–2679.

Pritchard, J. K., Stephens, M., & Donnelly, P. (2000). Inference of population structure using multilocus genotype data. Genetics, 155(2), 945–959.

Pröhl, H. (1997). Territorial behaviour of the strawberry poison-dart frog, Dendrobates pumilio. Amphibia-Reptilia, 18(4), 437–442.

Pröhl, H. (2002). Population differences in female resource abundance, adult sex ratio, and male mating success in Dendrobates pumilio. Behavioral Ecology, 13(2), 175–181.

Pröhl, H. (2005). Territorial behavior in dendrobatid frogs. Journal of Herpetology, 39(3), 354–365.

Pröhl, H., & Berke, O. (2001). Spatial distributions of male and female strawberry poison frogs and their relation to female reproductive resources. Oecologia, 129(4), 534–542.

Pröhl, H., & Ostrowski, T. (2011). Behavioural elements reflect phenotypic colour divergence in a poison frog. Evolutionary Ecology, 25(5), 993–1015.

118

R Core Team (2015). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R- project.org/.

Ramasamy, R. K., Ramasamy, S., Bindroo, B. B., & Naik, V. G. (2014). STRUCTURE PLOT: a program for drawing elegant STRUCTURE bar plots in user friendly interface. SpringerPlus, 3(1), 1.

Raymond, M., & Rousset, F. (1995). GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. Journal of Heredity, 86(3), 248–249.

Richards, L. A., & Windsor, D. M. (2007). Seasonal variation of arthropod abundance in gaps and the understorey of a lowland moist forest in Panama. Journal of Tropical Ecology, 23(02), 169–176.

Routman, E. (1993). Population structure and genetic diversity of metamorphic and paedomorphic populations of the tiger salamander, Ambystoma tigrinum. Journal of Evolutionary Biology, 6(3), 329–357.

Rowe, G., Beebee, T. J. C., & Burke, T. (2000). A microsatellite analysis of natterjack toad, Bufo calamita, metapopulations. Oikos, 88(3), 641–651.

Saporito, R. A., Garraffo, H. M., Donnelly, M. A., Edwards, A. L., Longino, J. T., & Daly, J. W. (2004). Formicine ants: an arthropod source for the pumiliotoxin alkaloids of dendrobatid poison frogs. Proceedings of the National Academy of Sciences of the United States of America, 101(21), 8045–8050.

Seppä, P., & Laurila, A. (1999). Genetic structure of island populations of the anurans Rana temporaria and Bufo bufo. Heredity, 82(3), 309–317.

Shaffer, H. B., Fellers, G. M., Magee, A., & Voss, S. R. (2000). The genetics of amphibian declines: population substructure and molecular differentiation in the Yosemite Toad, Bufo canorus (Anura, Bufonidae) based on single-strand conformation polymorphism analysis (SSCP) and mitochondrial DNA sequence data. Molecular Ecology, 9(3), 245–257.

Silverstone, P. A. (1973). Observations on the behavior and ecology of a Colombian poison-arrow frog, the Kõkoé-Pá (Dendrobates histrionicus Berthold). Herpetologica, 29(4), 295–301.

Silverstone, P. A. (1975). A revision of the poison-arrow frogs of the genus Dendrobates Wagler. Revisión de las ranas venenosas del género Dendrobates Wagler. Natural History, 21, 1–55.

119

Sinsch, U. (1990). Migration and orientation in anuran amphibians. Ethology Ecology & Evolution, 2(1), 65–79.

Slatkin, M. (1987). Gene flow and the geographic structure of natural populations. Science, 236(4803), 787–792.

Svenning, J. C. (2001). On the role of microenvironmental heterogeneity in the ecology and diversification of Neotropical rain-forest palms (Arecaceae). The Botanical Review, 67(1), 1–53.

Tallmon, D. A., Chris Funk, W., Dunlap, W. W., & Allendorf, F. W. (2000). Genetic differentiation among long-toed salamander (Ambystoma macrodactylum) populations. Copeia, 2000(1), 27–35.

Théry, M. (2001). Forest light and its influence on habitat selection. In Tropical Forest Canopies: Ecology and Management (pp. 251–261). Springer Netherlands. van Wijngaarden, R., & Bolaños, F. (1992). Parental care in Dendrobates granuliferus (Anura: Dendrobatidae), with a description of the tadpole. Journal of Herpetology, 26(1), 102–105.

Vargas-Salinas, F., & Amézquita, A. (2013). Stream noise, hybridization, and uncoupled evolution of call traits in two lineages of poison frogs: Oophaga histrionica and Oophaga lehmanni. PloS ONE, 8(10), e77545.

Via, S. (1999). Reproductive isolation between sympatric races of pea aphids. I. Gene flow restriction and habitat choice. Evolution, 53(5), 1446–1457.

Wang I. J., Summers K. (2009). Highly polymorphic microsatellite markers for the highly polymorphic strawberry poison-dart frog and some of its congeners. Conservation Genetics, 10(6), 2033–2036.

Weir, B. S., & Cockerham, C. C. (1984). Estimating F-statistics for the analysis of population structure. Evolution, 38(6), 1358–1370.

Wilson, G. A., & Rannala, B. (2003). Bayesian inference of recent migration rates using multilocus genotypes. Genetics, 163(3), 1177–1191.

Young, A. M. (1979). Arboreal movement and tadpole-carrying behavior of Dendrobates pumilio Schmidt (Dendrobatidae) in northeastern Costa Rica. Biotropica, 11(3), 238–239.

120

Zamudio, K. R., & Wieczorek, A. M. (2007). Fine-scale spatial genetic structure and dispersal among spotted salamander (Ambystoma maculatum) breeding populations. Molecular Ecology, 16(2), 257–274.

121

CONCLUSIONS

The most general conclusions of my research on the role of climate on the physiology and differentiation of poison frogs (Oophaga and Dendrobates) are:

The temporal and spatial heterogeneity in temperature seasonality contribute to explain the geographic variation in the rates of evaporative water loss in frogs; such connection seems to be functionally mediated by frog body size.

Dehydration has detectable effects on frogs’ locomotor performance and might this represent an appropriate proxy to estimate the vulnerability of Oophaga frogs to changing conditions in temperature and water availability.

Local environmental heterogeneity at the scale of few kilometers affects frog dispersal and therefore favours micro evolutionary processes; the mechanism would be the differential use of the habitat by the frogs, based on spatial variation in microclimatic conditions, the abundance of reproductive resources and the elevation.

Oophaga frogs differ in their climatic niche. The temperature seasonality and precipitation of driest quarter seems to be the most important variables in explaining those current climatic niche divergences and represent the major shifts in climatic tolerances along the lineages history.

At large spatial scales, the physiological tolerances of Oophaga frogs may have interacted with climatic heterogeneity in shaping frog dispersal and, ultimately, the process of evolutionary divergence within the genus.

At local spatial scales, climatic heterogeneity also shapes dispersal of the individuals, through its direct effect on individual’s physiological performance and habitat choice, ultimately leading to genetic structuring.

122

Future directions

Since Oophaga are diurnal frogs and coloration appears to be pivotal in sexual selection and anti-predator defenses (aposematism), I wonder whether climatic heterogeneity explains part of geographic variation in coloration. Climate would affect daily activity patterns, microhabitat (e.g., perch) use and, consequently, frog exposure to potential predators. The effectiveness for communication of the visual signal would therefore vary in accordance to these ecological changes.

Other future efforts may focus on understanding other potential mechanisms underlying the relationship between climate and evolutionary divergence within Oophaga. The use of other physiological traits, the study of individual dispersal abilities, and the estimation of the adaptative value of these traits may offer very valuable insights into this relationship. Last but not least, we should devote energy to understand the past and potential role of climatic heterogeneity in extinction processes within Oophaga species, using mechanistic approach in order to feed with hard science the conservationist decisions and programs.

123

Oophaga lehmanni postmetamorphic.

124

Municipio Nuquí, Chocó, Colombia.

125