Basin building, a rare reproductive strategy used by frogs, consists of constructing a depression in the substrate (a basin) and depositing eggs within the basin. The stream breeding Neotropical hylid frog, Smilisca sordida, constructs several types of basins; specifically, eggs deposited in a circular depression either with eggs floating on the surface of the water or with eggs attached to the substrate (open basins) or eggs buried beneath the substrate (a buried basin). Two hypotheses were tested to explain the selective forces maintaining the basin construction reproductive mode in S. sordida. The microenvironment modification hypothesis proposes that basins modify the environment in ways beneficial to developing young. The predator avoidance hypothesis proposes that basins separate and therefore protect embryos and larvae from predators. The thermal prediction of the microenvironment modification hypothesis is supported but only for open basins: buried basins do not modify the thermal environment relative to streams. Fish consume eggs of S. sordida, suggesting that basins provide refuge from fish predation. Field data on predation rates show that, although basin building may reduce predation by fish, planarians, other invertebrate species, and conspecific tadpoles, all colonize basins and consume eggs of S. sordida. Of 230 clutches, 37.4% experienced total mortality and of clutches that failed, 74.4% were solely caused by predators that colonized basins. The distribution of predators along the stream was nonrandom and the distribution of predators strongly influenced the fate of clutches. Basin desiccation, as well as flooding, caused additional mortality to young and are additional consequences of depositing eggs in basins for S. sordida.