Population Regulation of Frogs and Lizards in a Lowland Wet Neotropical Forest: Integrating Alternative Hypotheses

Abstract

Populations can be regulated in space and time from the “bottom-up” by resources (e.g., food), from the “top-down” by predation, or from parallel factors, such as interactions with conspecifics or heterospecifics. Amphibians and reptiles are species rich and abundant in lowland wet forests of Central America, and are important components of trophic communities as low or mid-level consumers of arthropods, and as prey to a diverse assemblage of vertebrate and invertebrate species. Theory describing population and community structure of Neotropical amphibians and reptiles has been generated largely from studies of a species-rich assemblage of frogs (Craugastoridae, Dendrobatidae) and anole lizards (Dactyloidae) that occur terrestrially on the forest floor and have generally supported hypotheses of bottom-up regulation mediated by resources, such as leaf-litter and prey abundance. Because trees regulate the input of leaf litter into terrestrial systems, a conceptual model has been suggested where leaf-abscission phenology of trees regulates patch dynamics of terrestrial frogs and lizards through forests in ways consistent with bottom-up regulation (the litter-mosaic hypothesis). However, a comparative review of anole life history in mainland Central American and Caribbean island habitats has described an alternative, top-down hypothesis of population regulation, where anoles in Central America are more greatly influenced by predation than anoles in Caribbean islands (the mainland-island model). Additionally, studies of the island anole Dactyloa aenea have demonstrated that juveniles are attracted to conspecifics when settling habitat, and conspecific attraction may be an additional parallel factor influencing forest dynamics of anoles and frogs in Central America. In an effort to generate a more integrative conceptual model describing population regulation and forest dynamics of frogs and lizards assemblage in Central America, I tested three hypotheses describing bottom-up, top-down, and parallel population regulation of a model assemblage at La Selva Biological Station, Costa Rica. First, I tested the litter-mosaic hypothesis that leaf-drop phenology of trees regulates population cycles and patch dynamics of terrestrial arthropods, frogs (Craugastor bransfordii, Oophaga pumilio), and lizards (Norops humilis). I used a comparative experimental approach, where I (1) compared seasonal abundance cycles of terrestrial litter and fauna beneath two tree species differing in phenology of leaf abscission (Castilla elastica, Dipteryx panamensis), and (2) experimentally manipulated leaf abscission beneath C. elastica with litter supplementation. I observed seasonal patterns of leaf litter, arthropods, and vertebrate abundance, but these patterns did not vary between tree species. Experimental supplementation caused elevated abundance of N. humilis in manipulated C. elastica plots relative to controls, as a result of demographic shifts in either apparent survival, immigration, or recruitment. My results suggest the litter-mosaic hypothesis can be refined and restricted to describing spatiotemporal variance of terrestrial anoles, while excluding frogs. Second, to better understand how predators influence prey frog and lizard ecology, I tested the mainland-island model prediction that predation exerts stronger effects than food resources for anoles and, by extension, frogs in Central America. I modeled the relative contribution of food, microhabitat, and predatory spiders (Ctenidae) to frog and anole occupancy. Frog occupancy was most strongly influenced by predators, an effect which increased at reduced biomass of leaf litter. Anoles occupied sites independent of predators, an observation inconsistent with the mainland-island model. All species were positively associated with leaf-litter depth and had elevated detection when predators were present. Third and last, I used eighteen months of mark-recapture observations of O. pumilio in a homogenous, fallow cacao plantation to test predictions of the conspecific attraction hypothesis. I found that juveniles settled habitat and used space in significant association with adults with previous history in plots, in ways consistent with conspecific attraction of juveniles and strong male territoriality. Together, my results contribute to our understanding of patch dynamics for terrestrial frogs and lizards in Neotropical wet forests. My experimental results suggest that Norops humilis populations respond to pulses of terrestrial litter in ways consistent with the litter-mosaic hypothesis, but the O. pumilio and C. bransfordii do not. Dispersal of O. pumilio and other frogs through the landscape may be driven in part by conspecific attraction. Because predatory spiders exert stronger effects on the occupancy of terrestrial frogs at lower abundance of terrestrial litter, seasonal declines in frog abundance during the wet season may be driven in part by elevated predation pressure when leaf litter is scarce. Predator-prey models reject a large literature invoking predation as the dominant force shaping ecology of diverse mainland anoles, and suggest similar mechanisms may regulate the ecology and evolution of mainland and island anoles.