Trophic and Microhabitat Niche Overlap in Two Sympatric Dendrobatids from La Selva, Costa Rica

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by SEDICI - Repositorio de la UNLP

Cuad. herpetol., 2424 (2):(2): 8181–92,–92, 20102010 81

TROPHIC AND MICROHABITAT NICHE OVERLAP IN TWO SYMPATRIC DENDROBATIDS FROM LA SELVA, COSTA RICA

RODRIGO CAJADE1, EDUARDO FEDERICO SCHAEFER1, MARTA INÉS DURÉ1 & ARTURO IGNACIO KEHR1

1 Centro de Ecología Aplicada del Litoral (CECOAL – CONICET), Ruta 5, Km 2,5, (3400) Corrientes, Argentina. [email protected], [email protected], [email protected], [email protected]

R E S U M E N. — Se estudio la ecología trófica y uso del microhábitat de Dendrobates auratus y Oophaga pumilio en un área de simpatría entre las dos especies ubicada en la Estación Biológica La Selva, Costa Rica. En este sitio ambos dendrobatidos son simpátricos debido a la introducción y naturalización de D. auratus en los últimos 24 años. La dieta de ambas especies se describió a partir del análisis de la técnica del lavado de estómago. Los microhábitas utilizados fueron definidos según el sitio donde cada ejemplar fue capturado. La relación en el uso del microhábitat y la eco- logía trófica entre ambas especies fue evaluada utilizando el índice de solpamiento de Pianka (Ojk) en el análisis de la dieta (proporción de presas y volumen) y en el uso del microhábitat. La dieta de ambos dendrobátidos estuvo caracterizada principalmente por el consumo de himenópteros (hormigas), ácaros, y colémbolos, resultando consecuentemente en un alto índice de solapamiento en la proporción y volumen de las presas, sin embargo, este alto solapamiento no fue significativo y no implicó la presencia de interacciones negativas entre ambas especies. El uso del microhábitat presentó un solapamiento muy bajo y no significativo, indicando una diferenciación en los micro- hábitats utilizados por cada especie. La ausencia de interacciones negativas en cuanto al uso de los recursos tróficos entre ambos dendrobátidos podría deberse a la diferenciación en el uso del microhábitat, y posiblemente, a la abundancia de presas en el área. El gran volumen de formícidos y ácaros en la dieta de estas dos especies son consistentes con la hipótesis del consumo de estos artrópodos como una fuente de alcaloides. PALABRAS CLAVE: Dendrobates auratus, Oophaga pumilio, dieta, microhabitats, comple- mentariedad.

A B S T R A C T. — We studied the trophic ecology of Dendrobates auratus and Oophaga pumilio in La Selva Biological Station, Costa Rica. At this site, both dendrobatids are sympatric due to the introduction and naturalization of D. auratus in the last 24 years. Diets of both dendrobatids were recorded by the analysed stomach flushing technique. Microhabitats uses were defined as the site where each individual was captured. The influence of microhabitat on diets was evaluated by the dietary (prey proportions and volume) and microhabitat overlaps

using Pianka’s (Ojk) overlap index calculated with EcoSim software. Diets of both dendrobatids were principally characterized by the preference of hymenopterans (ants), acarines and collembolans, resulted in a high overlapping in prey proportions and prey volume. However, diets overlaps were not significant, suggesting the absence of negative feeding interactions. Microhabitat use was low overlapped and also not significant, suggesting a differentiation on the use of spatial resource. The absence of negative feeding interactions between Dendrobates auratus and Oophaga pumilio could be due to segregation in microhabitat use and possible by the abundance of trophic resource in the area. The great large volumes of formicids and acarines in the diet of this dendrobatids are in agreement with the hypothesis of these arthropods as a dietary source of alkaloids. KEYWORDS: Dendrobates auratus, Oophaga pumilio, diets, microhabitats, complementarity.

Recibido: 03/09/10 — Revisado: 23/09/10 — Aceptado: 24/11/10 Ed. asoc.: M. Vaira 82 R. CAJADE et al.: Trophic and microhabitat overlap in dendrobatids

INTRODUCTION lishment and naturalization of D. auratus at La Selva, is quite recent (within The differential utilization of resourc- the last 20 years) and due to the intro- es provides essential information to un- duction of this species by a local em- derstand the role of individual species ployee of a tourist lodge in the near lo- in communities and has been indicated cality of Chilamate in 1986 (Guyer and as the key would allow the coexistence Donnelly 2005). However, despite the between sympatric species (Schoener, naturalization of D. auratus at La Sel- 1974; Toft, 1981; Lieberman, 1986; Duré va, its distribution at this site is re- and Kehr, 2004). Several studies constricted only along the margins of Tres ducted in sympatric anurans species, Rios path (R. Cajade, pers. obs). showed that some differences in food Single studies conducted in D. au- resource use may be caused by different rautus and O. pumilio determined that foraging patterns, microhabitat use, sea- these dendrobatids are closely related in sonal and daily activity, and resource diet, reproduction and habitat prefer- availability (Toft, 1980a; Toft 1980b; Toft ence (Toft, 1981; Lieberman, 1986; 1985; Graves, 1999; Almeida-Gómez et Pough and Taigen, 1990; Donnelly, al., 2007; Cuevas and Martori, 2007; 1991). Diet, diurnal activity and apose- Duré et al., 2009). Within the limita- matic coloration of both dendrobatids tions imposed by evolutionary history, are involved in an evolutionary process exploitation of particular prey by a spe- of chemical defense implicating skin cies can influence the interactions of toxins, mainly alkaloids (Caldwell, 1996; that species in a particular environment Santos et al., 2003; Saporito et al., and, hence, may determine activity pe- 2003; Darst et al., 2005; Saporito et al., riods, reproductive features, and preda- 2009).While D. auratus and O. pumilio tor–prey interactions (Polis et al., 1989; are probably the dendrobatids most ex- Caldwell, 1996) tensively studies (Caldwell, 1996; Dendrobates aurautus and O. pumilio Graves, 1999), their feeding ecology and (Anura; Dendrobatidae) are two com- microhabitat use have not been ana- mon leaf litter dendrobatids species lyzed in sympatry. which inhabit the wet tropical and sub- The aim of this study is describe the tropical forest of down lands in Central trophic ecology and microhabitat use of America and North of South America, D. auratus and O. pumilio in a second- being sympatric at many sites in Costa ary forest from La Selva, Costa Rica, Rica, Nicaragua and Panamá (Savage, where these species occurs in a condi- 2002). Both dendrobatids are listed as tion of no natural sympatry. Least Concern by IUCN (Solis et al., 2008a,b) because they are tolerant to a certain degree of habitat modification MATERIALS AND METHODS and their populations appears to be large. The main threats to these spe- Field work was carried out during cies are habitat loss and over-collection dry season between 2-17 March, 2008, in for pet trade. The bright coloration and a plot of secondary forest located in the diurnal activity of D. auratus and O. biological station La Selva, Costa Rica. pumilio make these frogs conspicuous Individuals of D. auratus and O. pumil- and very attractive, playing an impor- io were captured by hand at 7:00 and tant role in attracting tourist along its 18:00 hs, along the margins of Tres distribution. In consequence, D. auratus Rios path between 1200 and 2900 m. of has been introduced in various localities the trailhead (Fig. 1). Immediately stom- to increase tourism in Costa Rica in ach contents of each frog were flushed recent times. For example, the estab- (see Sole et al., 2005 methodology) and Cuad. herpetol., 24 (2): 81–92, 2010 83 fixed in 5% ethanol to identify prey in (3) anthills, and (4) under fallen tree laboratory. This technique allowed us trunks. The number of prey items per obtaining stomach contents quickly stomach for each prey category was re- without sacrificing the frog, and avoid- corded. Volume of each prey item was ing the loss of samples by reducing the estimated using the formula for an el- digestion time occurred since the collec- lipsoid, tion until obtaining its stomach contents (Sole et al., 2005). Body length V= 4/3π (1/2L) (1/2A)2 (mm) and mouth width (mm) were measured for each individual with calipers where V is volume, L is length and to nearest 0.01 mm. Each individual W is width (Dunham, 1983). Prey diver- was identified for its dorsal and ventral sity was calculated used the Shannon pattern registered by a digital photo- index (H´) (Shannon and Weaver, 1949). graph to avoid pseudo-replication. Prey To determine the importance of each items were identified to order (following prey category in the diet, we used the Borror and White, 1970) except ants following formula: that were identified to family and considered as a separate item prey. All I = (F% + N% + V%) / 3 measurements were made using an oc- ular micrometer to the nearest 0.01 where, F% = occurrence percentage, mm under a stereomicroscope. Micro- N% = numeric percentage and V% = habitat categories recorded for each in- volumetric percentage (Biavati et al., dividual frog were: (1) wet ground cov- 2004). We calculated the importance of ered with ferns, (2) leaf litter associated prey categories for pooled stomachs with the base of cacao plants and trees, (IPS). The niche breadth of prey and

Figure 1. Detailed map of the biological station La Selva and location in Costa Rica (box). The black line represents the travel of Tres Ríos path. The study area corresponds to margins of travel between the two black points in the path. 84 R. CAJADE et al.: Trophic and microhabitat overlap in dendrobatids x for (n = 40) from La Selva, Costa Rica. Oophaga pumilio (n = 40) and Dendrobates auratus . Abundance, volume and frequency (Fa = number of stomachs containing prey item) for each item importance dietary inde pooled stomachs (ISP) in the diet of Table 1 Cuad. herpetol., 24 (2): 81–92, 2010 85

prey volume was calculated using Pianka’s (Ojk) overlap index (Pianka, Levins’s index (Levins, 1968), 1973) using EcoSim software 7.72 (Go- telli and Entsminger, 2005) with re- 2 - Nb = (∑pij ) tained niche breadth and reshuffled zero states. The EcoSim program also where pij represents the probability of determines whether measured overlap finding the item i in the sample j. We values differed from what would be ex- calculated dietary (food proportions and pected based on random sampling of volume) and microhabitat overlap using the species data. EcoSim performs Mon-

Figure 2. Importance index of prey categories ingested by Dendrobates aurauts (A) and Oopha- ga pumilio (B) from La Selva Biological Station, Costa Rica. IPS =importance index based on pooled stomachs. Prey categories are as follow: Ac = Acari; Am = Amphipoda; Ar = Aranae; Cl = Collembola; Co = Coleoptera; De = Dermaptera; Di = Diptera; Fo = Formicidae; He = Hemip- tera; Ho = Homoptera; Hy = Hymenoptera (other than ants); Is = Isopoda; La = Larvae (In- sect); Me = Megalloptera; Po = Polydesmida; Si = Siphonaptera; So = Solpugida; Th = Thysa- noptera; Zo = Zoraptera. 86 R. CAJADE et al.: Trophic and microhabitat overlap in dendrobatids te Carlo randomizations to create ‘‘pseu- tively). The most frequently prey were do-communities’’ (Pianka, 1986), and formicids and acarines, (87.5% of adults), then statistically compares the patterns and collembolans (67.5% of adults). Con- in these randomized communities with sidering IPS the most important prey those in the real data matrix. In this were acarines, formicids, collembolans analysis (randomization algorithms RA3; and coleopterans (Figure 2). Prey diver- Winemiller and Pianka, 1990), ‘‘scram- sity was 1.33 and niche breath of prey bled zeros’’, and all values of the origi- and prey volume was 2.86 and 7.57 re- nal matrix were randomized 1000 times, spectively. The mean of body length and and the niche breadth was retained for mouth was 20.41 ± 1.06 mm, and 6.23 each species. In other words, the algo- ± 0.42 mm, respectively. Mean Ln prey rithms retained the amount of special- volume was not correlated with Ln ization for each species (Gotelli and mouth width (r = -0.21; n = 40; P = Entsminger, 2005). 0.19). The proportions of microhabitats Pearson correlations coefficients (Zar, used were: (1) wet ground covered with 1996) were used to establish the relation- ferns (17.5%), (2) leaf litter associated ship between the morphology of preda- with the base of cacao plants and trees tors and prey volume. When assumptions (82.5%), (3) associated with ant nests of normality were broken, the data were (0%), (4) under fallen trunks (0%). natural logarithms (Ln) transformed and normalized. Means were presented as DENDROBATES AURATUS DIET mean ± standard deviation. This anuran consumed 14 prey categories (Table 1), being formicids (69.01 %), acarines (20.8 %), and collembolans RESULTS (5.9%) dominant prey in number. Volu- metrically, the diet was dominated by Fourteen adult specimens of each zorapterans and dermapterans (22.5 % dendrobatids species were captured. All and 14.7 %, respectively). The most fre- individuals had identifiable stomach con- quently prey were formicids (92.5% of tents. In O. pumilio were recorded adults) acarines (85% of adults) and col- three recaptures corresponding to three lembolans (60% of adults). Considering individuals recaptured at different sites IPS the most important prey were, for- respect to the initial capture site. Nine micids, acarines, collembolans and co- recaptures corresponding to three indi- leopterans (Figure 2). Prey diversity viduals were recorded in D. auratus: was 0.96 and niche breath of prey item one individual recaptured six times at and prey volume was 1.90 and 8.09, re- the same site respect to the initial cap- spectively. The mean body length and ture site, other individual recaptured mouth were 25.36 ± 4.25 mm and 7.47 twice at the same site respect to the ± 1.25 mm, respectively. Mean Ln prey initial capture site, and another individ- volume was not correlated with mouth ual recaptured once at a different site width (r = -0.05; n = 40; P = 0.73). The respect to the initial capture site. proportions of microhabitats uses were: (1) wet ground covered with ferns OOPHAGA PUMILIO DIET (55%), (2) leaf litter associated with the The diet of this species consisted of 17 base of cacao plants and trees (12.5%), types of prey (Table 1) and was domi- (3) associated with ant nests (17.5%), (4) nated numerically by acarines (43.9 %), under fallen trunks (15%). Formicids (38.1 %) and collembolans Trophic and microhabitat overlaps. (9.6%). Volumetrically, the diet was Considering prey proportion between composed mainly for polydesmidans and the two dendrobatids the trophic niche coleopterans (26.6% and 15.8%, respec- overlap was high (Ojk = 0.87). Random- Cuad. herpetol., 24 (2): 81–92, 2010 87 izations with all data produced no sig- pumilio in this study, but absent in nificant difference between measured Caldwell (1996). However, the propor- (observed) and simulated (expected) tions and frequency of these prey were overlaps using diet proportion (P [ob- very low in both dendrobatids, respec- served < expected] = 0.83, and P [ob- tively. In fact its diets were predomi- served > expected] = 0.16). Therefore, nantly characterized for the consump- observed mean (0.75) was similar to tion of smaller prey, as was reflected in that expected by chance (0.73 ± 0.019). the lack of correlation between the vol- Moreover prey volumes were highly ume of prey and the widths of the overlaped (Ojk = 0.76). Randomizations mouths of predators suggesting that with all data produced no significant dif- both species select small prey. Accord- ferences between measured (observed) ing to Caldwell (1996), D. auratus and overlaps and simulated (expected) prey O. pumilio eat mainly on smaller prey volume overlaps (P [observed < expect- like ants and mites, who reach a length ed] = 0.82, and P [observed > expected] mean of 2 and 0.6 mm, respectively. = 0.18). Consequently, the observed Considering only the prey type, D. mean (0.76) was similar to that expect- auratus and O. pumilio appears to be as ed by chance (0.75 ± 0.016). generalist predators, because these ani- The spatial niche overlap between mals consumed a wide variety of prey. the two frog species was low (Ojk = However, the prey type proportions 0.38). Randomizations with microhabitat consumed by this species showed that data produced no significant differences they are specialist on formicids and ac- between measured (observed) and simu- arines. The proportionately higher num- lated (expected) (P [observed < expected] ber of fomicids and acarines in the D. = 0.51, and P [observed > expected] = auratus and O. pumilio diet resulted in 0.48) overlaps. The observed mean (0.38) a constricted niche breadth and in a was similar to that expected by chance low diversity index in both species. alone (0.47 ± 0.06). These results are consistent with the results of previous studies (Liberman, 1986; Donnelly, 1991; Caldwell, 1996). DISCUSSION Constricted niche breadth suggests an actively foraging strategy (Perry and We found similarities in prey catego- Pianka, 1997), which has been suggest- ries and diet composition between D. ed for D. auratus and O. pumilio (Toft aurautus and O. pumilio (in term of ,1981; Taigen and Pough, 1983). Howev- number and frequency importance: for- er predators that are specialist in colo- micids, acarines and collembolans). The nial insects are active when they need IPS reflects the four same prey items to encounter the colony. After that, the for both dendrobatids, differing only in predator can act as a sit-and-wait pred- the order of the two most important ator (Pianka, 1973; Donelly, 1991; Duré prey item. Thus, in D. aurautus formic- and Kehr, 2004). In fact, some D. aura- ids were most important prey than acar- tus (recaptures not considered in the ines, while in O. pumulio it was the analysis) were observed returning each opposite. The values of niche breadth of day to the same anthill and forage on prey volume for both dendrobatids were this, as were also documented by Pough wider to those reported by Caldwell and Taigen (1990). Thus, considering the (1996). It may be due to the consump- great proportion of formicids in the D. tion of prey with great volume such as auratus and O. pumilio diets and in dermapterans, and polydesmidans, re- agreement with previous observations corded in D. auratus and, such as poly- (Donnelly, 1991) the foraging strategy desmidans and isopods recorded in O. employed by this species can be consid- 88 R. CAJADE et al.: Trophic and microhabitat overlap in dendrobatids ered intermediate between a sit-and- suggest that differences in microhabitat wait and active predator. utilization seem to be an important fac- Individuals of Dendrobates auratus tor favoring the absence of negative were registered in microhabitats mainly feeding interactions and consequently characterized by high moisture and low the coexistence between D. auratus and sun exposure, as wet ground covered O. pumilio at the study site. Differenti- with ferns, the microhabitat more fre- ation in space use was indicated by quently used. On the other hand, indi- Cunha and Vieira (2004) as a mecha- viduals of O. pumilio were found mostly nism that might counteract complete in the leaf litter associated with the overlap in diet. An optional interpreta- base of cacao plants, a microhabitat tion of the high overlap in trophic with lower moisture and higher degree niche, without implying a negative in- of sun exposure. These microhabitats teraction could be explained by an abun- observations, are consistent with the dance of resources in the area sufficient reports made by Kitasako (1967) and to satisfy demands. Donnelly (1991) re- Guyer and Donnelly (2005) which ob- marked that the inclusion of prey other served a humid and shaded microhabitat than ants or mites in O. pumilio indi- preference in D. aurautus, such as leaf cates that frogs respond to naturally litter associated with nooks and crevic- occurring fluctuations in prey popula- es offered by buttresses of trees and tion sizes by consuming a variety of fallen trunks, and observed a preference prey types. for fairly sunny exposed microhabitats The niche complementarity hypothe- in O. pumilio, such as leaf litter at the sis states that for coexistence to occur, base of cacao plants. We not found O. high overlap in one dimension of the pumilio in association with fallen trunks niche must be compensated by low or prey nests (e.g. anthills). However overlap in another (Schoener 1974). Our Kitasako (1967) and Donnelly (1991), results suggest that in the period of report the occasional use of these mi- our study, D. auratus and O. pumilio crohabitats in O. pumilio. exhibit niche complementarity, with a Niche breadth of prey proportion and high overlap in diet but low overlap in prey volume for the two species over- feeding microhabitat. However, as were lapped extensively (Ojk = 0.75; Ojk = pointed out above, D. auratus and O. 0.76, respectively) indicating a similari- pumilio occurs sympatrically at many ty between the diets. However the sites along its distribution, but at La overlapping for each case was not signif- Selva this condition is not natural. icant. Thus, the overlap in diets does Thus, our results are not extrapolated not imply a negative interaction be- to a real sympatric situation but repre- tween D. auratus and O. pumilio. The sents valuables ecological observations spatial niche overlap between the two where the introduction of D. auratus dendrobatids was low (Ojk = 0.38) and must be evaluated at light of negative no significant, indicating a differentia- possible effects on the ecology of the tion on use of spatial resource. The resident frog O. pumilio. Further inte- traditional opposite interpretation of a grated diet studies that include sex, on- high overlap indicates shared resource togeny, seasonality, microhabitat use, utilization and a lack of competition daily activity, prey availability and in- (Gotelli and Graves, 1996), or a strong teractions are needed to better under- competition that has not yet led to di- stand the interspecific relationships and vergence in resource use (Connell, 1980; coexistence between these two dendro- Sale, 1974). In summary, despite the batids species at La Selva. fact that D. auratus and O. pumilio had Dendrobatids, and members of other a high trophic niche overlap, our results three anuran family (Mantellidae, Bu- Cuad. herpetol., 24 (2): 81–92, 2010 89 fonidae and Myobatrachidae), are known skin extracts of D. aurautus y O. pu- as poison frogs by the presence of alka- milio, and also in ants, mites and milli- loids in dermal skin glands, which act pedes consumed by this frogs. The as chemical defense against predators great volume, frequency of occurrence and/or microorganism. Indeed, more and number of ants and mites reported than 800 alkaloids, organized into 23 in the present study for D. auratus and structural classes, were reported from O. pumilio, is consistent with the hy- the skin of poison frogs mentioned pothesis of arthropods as a dietary above (Saporito et al., 2009). The pres- source of alkaloids. ence of alkaloids in poison frogs was attributed to the three process: 1) biosynthesis, where the frogs synthesize ACKNOWLEDGEMENTS the alkaloids, 2) sequestration, where the frog sequesters the alkaloids from a We thank Daniel Cadena, Federico food source and incorporate at its der- Chinchilla, Alejandro G. Farji-Brener, mal glands, and 3) sequestration and de José M. Serrano and Pamela Medina novo biosynthesis, where the frog se- Díaz for comments and suggestions dur- quester an alkaloid from a food source ing field work and earlier drafts of the an use it then to synthesize a new al- manuscript; two anonymous reviewers kaloid (Saporito et al., 2009). Several and the Associated Editor who provided dendrobatids species are also known as helpful comments on final versions of ant and mite specialist (Toft, 1980a, the manuscript. We thank Ministerio Donnelly, 1991; Caldwell, 1996), and ex- del Ambiente y Energía, Sistema Nacio- perimental evidence indicate that diet nal de Areas de Conservación de Costa specialization in poison frogs is closely Rica and the Directora of Estación Bi- related to presence of alkaloids in der- ológica La Selva, Dra. Deedra Mc- mal skin glands. Certainly, the analysis Clearn, for permission to work at the of skin extracts of some dendrobatids site, and Dirección General de Vida Sil- specimens raised in captivity and fed vestre Ministerio de Recursos Naturales with fruit flies (Drosophila sp.) showed Energía y Minas for export permissions an absence of alkaloids in its skins, of stomach contents. This study was fi- while specimens provided with fresh nancial supported by the Consejo Na- leaf-litter from the frog’s natural habi- cional de Investigaciones Científicas y tat, accumulated a variety of alkaloids Técnicas (CONICET), Argentina and Or- into the skin. In addition, the simulta- ganización para Estudios Tropicales neous occurrence of several structural (OTS), Costa Rica. classes of alkaloids present in arthropods from the frog´s natural habitat and the frog skin extracts reinforce the hy- LITERATURE CITED pothesis of a dietary source of alkaloids (Daly et al., 2000). Recent studies con- ALMEIDA-GOMES, M.; F. H. HATANO; M. ducted in dendrobatids frogs, indicate VAN SLUYS & C. F. D. ROCHA. that principally ants and oribatid mites, 2007. Diet and microhabitat use and secondarily coleopterans and milli- by two Hylodinae species (Anura, pedes, represents a dietary sources of Cycloramphidae) living in sympa- alkaloids (Saporito et al., 2003; Saporito try and syntopy in a Brazilian At- et al., 2004; Saporito et al., 2007). Rep- lantic Rainforest area. Iheringia, resentatives of many of the structural Série Zoologica 97 (1): 27-30. classes of alkaloids, such as pumiliotox- BIAVATI G. M.; H. C. WIEDERHECKER ins, decahydroquinilines, izidines, spiro- & G. R. COLLI. 2004. Diet of Epi- pyrrolizidines, have been detected in pedobates flavopictus (Anura: 90 R. CAJADE et al.: Trophic and microhabitat overlap in dendrobatids

Dendrobatidae) in a Netropical DUNHAM, A. E. 1983. Realized niche savanna. Journal of Herpetology overlap, resource abundance and 38 (4): 510-518. intensity of interspecific competi- BORROR J. D. & R. E. WHITE. 1970. A tion. Pp. 261-280. In: HUEY, R. field guide to insects American D., E. R. PIANKA, AND T. W. North of Mexico. The Peterson SCHOENER (eds.), Lizards Ecology. Field Guide Series. Houghton Mi- Harvard University Press, Cam- fflin Company. New York. 404 pp. bridge, Massachusetts, U.S.A. CALDWELL, J. P. 1996. The evolution of DURÉ, M. I. & A. I. KEHR. 2004. In- myrmecophagy and its correlates fluence of microhabitat on the tro- in poison frogs (Family Dendroba- phic ecology of two leptodactylids tidae). Journal of Zoology 240 (1): from noerthesatern Argentina. 75-101. Herpetologica 60 (3): 295-303. CONNELL, J. H. 1980. Diversity and the DURÉ, M. I.; A. I. KEHR & E. F. coevolution of competitors, or the SCHAEFER. 2009. Niche and re- ghost of competition past. Oikos source partitioning among five 35 (2): 131-138. sympatric bufonids (Anura, Bufoni- CUEVAS, M. F. & R. MARTORI. 2007. dae) from northeastern Argentina. Diversidad trófica de dos especies Phyllomedusa 8 (1): 27-39. sintópicas del género Leptodac- GRAVES, B. M. 1999. Diel activity pat- tylus (Anura: Leptodactylidae) del terns of the sympatric poison dart sudeste de la provincia de Córdo- frogs Dendrobates auratus and D. ba, Argentina. Cuadernos de Her- pumilio, en Costa Rica. Journal of petología 21 (1): 7-19. Herpetology 33 (3): 375-381. CUNHA, A. A. & M. V. VIEIRA. 2004. GOTELLI, N. J., & G. R. GRAVES. 1996. Two bodies cannot occupy the Null Models in Ecology, Smithso- same place at the same time, or nian Institution Press, Washing- the importance of space in the ton, D.C., U.S.A. 369 pp. ecological niche. Bulletin of the GOTELLI, N. J. & G. L. ENTSMINGER. Ecological Society of America 85 2005. EcoSim: Null Models Soft- (1): 25-26. ware for Ecology, Version 7. Ac- DALY, J. W.; H. M. GARRAFFO; P. JAIN; quired Intelligence Inc. & Kesey- T. F. SPANDE; R. R. SNELLING; C. Bear, Burlington, Vermont, U.S.A. JARAMILLO & A. S. RAND. 2000. http://homepages.together.net/ Arthropod-frog connection: deca- ;gentsmin/ecosim.htm. hydroquinoline and pyrrolizidine GUYER, C. & M. A. DONNELLY. 2005. alkaloids common to microsympa- Amphibians and Reptiles of La tric myrmicine ants and dendroba- Selva, Costa Rica, and the Cari- tid frogs. Journal of Chemical bbean slope. University of Califor- Ecology 26 (1): 73-85. nia Press. Berkeley, CA. 367pp. DARST, C. R.; P. A. MENDEZ-GUERRERO; KITASAKO, J. T. 1967. Observations on L. A. COLOMA & D. C. CANNATE- the biology of Dendrobates pumilio LLA. 2005. Evolution of dietary Schmidt and Dendrobates auratus specialization and chemical defense Girard. Unpublished Master’s The- in poison frogs (Dendrobatidae): A sis, University of South. Califor- comparative analysis. American nia, USA. Naturalist 165 (1): 56-69. LEVINS, R. 1968. Evolution in Changing DONNELLY, M. A. 1991. Feeding patterns Environments: Some Theoretical of the strawberry poison frog, Den- Explorations. Princeton University drobates pumilio (Anura: Dendroba- Press, Princeton, New Jersey, tidae). Copeia 1991 (3): 723-730. U.S.A. 132 pp. Cuad. herpetol., 24 (2): 81–92, 2010 91

LIEBERMAN, S. S. 1986. Ecology of the Proceedings of the National Aca- leaf litter herpetofauna of a Neo- demy of Sciences 101 (21): 8045- tropical rain forest: La Selva, Cos- 8050. ta Rica. Acta Zoologica Mexicana SAPORITO, R. A.: M. A. DONNELLY; R. 15: 1-72. A. NORTON; H. M. GARRAFFO; T. PERRY, G. & E. R. PIANKA. 1997. Ani- F. SPANDE & J. W. DALY. 2007. mal foraging: past, present and Oribatid mites as a major dietary future. Trends in Ecology & Evo- source for alkaloids in poison lution 12 (9): 360-363. frogs. Proceedings of the National PIANKA, E. R. 1973. The structure of Academy of Sciences 104 (21): lizard communities. Annual Re- 8885-8890. view of Ecology and Systematics SAPORITO, R. A.; T. F. SPANDE; H. M. 4: 53-74. GARRAFFO & M. A. DONNELLY. PIANKA, E. R. 1986. Ecology and Natu- 2009. Arthropod alkaloids in poi- ral History of Desert Lizards. son frogs: a review of the dietary Princeton University Press, Prin- hypothesis. Heterocycles 79: 277- ceton, New Jersey, U.S.A. 208 pp. 297. POLIS, G. A; C. A. MYERS & R. HOLT. SAVAGE, JM. 2002. The amphibians and 1989. The ecology and evolution reptiles of Costa Rica. The Uni- of intraguild predation: potential versity of Chicago Press. Chicago. competitors that eat each other. 934 pp. Annual Review of Ecology and SCHOENER, T.W. 1974. Resource parti- Systematics 20: 297-330. tioning in ecological communities. POUGH, F. H. & T. L. TAIGEN. 1990. Science 185 (4145): 27-39. Metabolic correlates of the fora- SHANNON, C. E., & W.WEAVER. 1949. ging and social behavior of dart- The mathematical theory of com- poison frogs. Animal Behavior 39 munications. University Illinois (1): 145-155. Press, Urbana, Illinois, U.S.A. 144 SALE, P. F. 1974. Overlap in resource pp. use, and interspecific competition. SOLE, M.; O. BECKMANN; B. PELZ; A. Oecologia 17 (3): 245-256. KWET & E. WOLF. 2005. Stomach- SANTOS, J. C.; L. A. COLOMA & D. C. flushing for diet analysis in anu- CANNATELLA. 2003. Multiple recu- rans: an improved protocol eva- rring origins of aposematism and luated in a case study in Arauca- diet specialization in poison frogs. ria forests, southern Brazil. Stu- Proceedings of the National Aca- dies on Neotropical Fauna and demy of Sciences of the USA 100 Environment 40 (1): 23-28. (22): 12792-12797. SOLÍS, F.; R. IBÁÑEZ; C. JARAMILLO; G. SAPORITO, R. A.; M. A. DONNELLY; R. CHAVES; J. SAVAGE; G. KÖHLER L. HOFFMAN; H. M. GARRAFFO & & N. COX. 2008a. Oophaga pumi- J. W. DALY. 2003. A siphonotid lio. In: IUCN 2010. IUCN Red millipede (Rhinotus) as the source List of Threatened Species. Ver- of spiropyrrolizidine oximes of sion 2010.4. www.iucnredlist.org. dendrobatid frogs. Journal of Che- Downloaded on 08 November mical Ecology 29 (12): 2781-2786. 2010. SAPORITO, R. A.; H. M. GARRAFFO; M. SOLÍS, F.; R. IBÁÑEZ; C. JARAMILLO; G. A. DONNELLY; A. L. EDWARDS; J. CHAVES; J. SAVAGE; G. KÖHLER; T. LONGINO & J. W. DALY. 2004. K.H. JUNGFER; W. BOLÍVAR & F. Formicine ants: An arthropod BOLAÑOS. 2008b. Dendrobates au- source for the pumiliotoxin alka- ratus. In: IUCN 2010. IUCN Red loids of dendrobatid poison frogs. List of Threatened Species. Ver- 92 R. CAJADE et al.: Trophic and microhabitat overlap in dendrobatids

sion 2010.4. www.iucnredlist.org. TOFT, C.A . 1981. Feeding ecology of Downloaded on 08 November Panamanian litter anurans: pat- 2010. terns in diet and foraging mode. TAIGEN, T. L. & F. H. POUGH. 1983. Journal of Herpetology 15 (2): 139- Prey preference, foraging, beha- 144. vior, and metabolic characteristics TOFT, C. A. 1985. Resource partitioning of frogs. American Naturalist. 122 in amphibians and reptiles. Co- (4): 509-520. peia 1985 (1): 1-21. TOFT, C.A. 1980a. Feeding ecology of WINEMILLER, K. O., & E. R. PIANKA. thirteen syntopic species of anu- 1990. Organization in natural ass- rans in a seasonal environment. emblages of desert lizards and tro- Oecologia 45 (1): 131-141. pical fishes. Ecological Monogra- TOFT, C.A. 1980b. Seasonal variation in phs 60: 27–55. populations of Panamanian litter ZAR, J. H. 1996. Biostatistical Analysis, frogs and their prey: A compari- 3rd ed. Prentice Hall, Upper son of wetter and drier sites. Saddle River, New Jersey, U.S.A. Oecologia 47 (1): 34-38. 662 pp.