The Abundance and Diversity of the Herpetofaunas of Tropical Forest Litter

Abstract

Quantitative and qualitative samples of amphibian and lizard faunas were taken from forest litter in Costa Rica and compared with similar samples collected in Borneo, the Philippines, and Panama. Animal abundance is about ten times greater in Costa Rican lowland wet forest than in Borneo. Radically different routes and rates of energy flow are postulated to account for the difference. In a series of upland (1010-1425 m) samples in the Philippines, the number of litter lizards and frogs increases with elevation. A single upland (1200 m) Costa Rican sample contains about three times as many animals per 100 m2 as two lowland sites. This increase with altitude correlates well with hypotheses that overall tropical forest productivity is greatest at intermediate elevations. Herpetofaunal densities are greater in wet areas compared to dry sites and flat terrain compared to slopes. These observations are linked to the greater variation and lesser total litter fall in dry sites and/or slopes. The number of species of amphibians and lizards regularly inhabiting the litter is similar in all wet lowland forests studied. Increasing elevation and decreasing rainfall correlate with faunas having fewer species. The former is seen as the result of the differential ability of lowland species to invade uplands, and the latter as a decrease in the kind and duration of frog-spawning sites. Reduced equitability with increasing elevation is principally due to the much greater proportional increase of the commonest species. In the evolution of American and Bornean-Philippine wet lowland faunas, frogs with direct development (Eleutherodactylus) have radiated in the former in the same fashion that skinks have in the latter. SEVERAL RECENT STUDIES have accumulated a large body of data on the density and composition of the reptile and amphibian faunas inhabiting tropical forest litter. Samples are available from undisturbed forest in the Philippines (Brown and Alcala 1961), Borneo (Lloyd et al. 1968), Panam'a (Heatwole and Sexton 1966; Sexton et al. 1964) and Costa Rica (this study). Analysis of the Bornean samples has raised basic ecological questions (Lloyd et ai. 1968), and a detailed comparison and analysis of all of the available data is now appropriate. These studies give an opportunity to make the first quantitative comparisons between Old and New World tropical vertebrate faunas. The techniques are uniform enough to make rough comparisons between numbers of species, distribution of animals among species, and animal densities. There are enough samples to compare: 1) Old and New World wet lowlands and uplands, 2) different altitudes within a single region, 3) different climates in the lowland Neotropics, and 4) different seasons and topography in the same Neotropical area. Most of the samples are drawn from relatively undisturbed forested areas where the faunas are assumed to have achieved some sort of equilibrium, at least with regard to number of species, animal density, and species-abundance relationships. These kinds of comparisons have not been made for any vertebrate group. The only geographically extensive quantitative comparisons of tropical forest faunas have been those of understory foliage insects by Janzen (1973a, b). Many of his samples were taken near the same sites as the Costa Rican herpetofaunal samples reported here.