The relative importance of habitat characteristics in the maintenance of a larval anuran species assemblage in the tropics / Claude Gascon.

Tese (Ph.D.) Florida State University, 1990

Using descriptive experimental approaches, I determined the relative importance of abiotic factors, predation, and interspecific interactions acting within a species assemblage of tropical larval anurans. Fifty-three independent aquatic sites were sampled across 2 consecutive rainy seasons to establish patterns of species occurrence in natural sites. Transplant experiments were performed to see if differences in habitat alone can explain the observed distribution of the most common species of tadpoles. By exposing different combinations of tadpole prey species to the same suite of common aquatic predators, I could determine the significance and intensity of predation rates on these tadpoles species. The effect of interspecific interactions and how those interactions might vary with variation in breeding phenology were tested in pairwise rearings of tadpole species varying the time of encounter of one species with respect to the other. Twenty-five tadpole species were encountered in 4 distinct types of habitats (terra firme, peccary wallows, stream-side, and stream sites). Rainfall was concentrated between December and May, and phenology of reproduction was strongly affected by the rainfall pattern. For each of the 11 most common tadpole species, I found major abiotic characteristics that differed between sites used and not used by those species. On a broader scale of analysis, there was no concordance among sites in the similarity of either their abiotic characteristics or the actual geographic distances and indices of similarity in their tadpole assemblages. Osteocephalus taurinus, Dendrobates femoralis, and Phyllomedusa tomopterna were the 3 most common tadpole species, widely overlapping in time and space. Osteocephalus taurinus tadpoles occurred in all types of habitats, whereas Dendrobates femoralis and Phyllomedusa tomopterna occurred only in upland-associated natural sites (wallows and terra firme sites). Odonate naiads were the most common predator in upland isolated pools and peccary wallows while fish were most common in lowland, stream-associated habitats. Occurrence patterns for P. romopterna were concordant with the results from transplants that indicated that larvae cannot survive in lowland aquatic sites, although adults are present in lowland areas. Although H. geographica and D. femoralis occur only in certain habitat types, they survive and grow equally well in all habitat types when transplanted there. Osteocephalus taurinos, which occurs in all types of habitats, actually grows and survives better in upland habitats than in lowland ones. Predators reduced tadpole survival in 3 out 4 different experimental prey-species combinations. Aeshnid naiads killed the most tadpoles, followed by Libellulid naiads. Fish were never the most effective predators and their effect was variable. In most cases, growth of tadpoles was inversely related to the number of surviving conspecific tadpoles. Predation rate on each of the 3 prey species varied among predators and sometimes dependend upon which other species were present indicating that predation rate in the system was a function of the particular combination of tadpole species. No significant differences were found in selectivity for any prey species among individual predator types. Predators preferred P. tomopterna, the less abundant species, over O. taurinus whenever the two species were together. The time of introduction of either P. tomopterna or D. femoralis (the second species) had no effect on O. taurinus (the first and more numerous species). Phyllomedusa tomopterna had lower survival and growth when introduced late than when introduced earlier or in the absence of O. taurinus; this result indicates that breeding phenology affects the intensity of competition. In contrast, Dendrobates femoralis did poorly whenever it was reared with O. taurinus. On the whole, individual species respond to habitat characteristics and breed in sites having spcific attributes. No large-scale species assemblages exist that covary in habitat use. Abiotic characteristics suffice to explain the distribution of only one of 4 species; in all other cases some other ecological pressure(s) is acting on the species to produce the observed distribution pattern. Experimental manipulations and occurrence patterns of predators in natural sites suggest that predation is the major force acting on populations of the 3 most common species. Competition will probably be important only in the absence of predators.