The larval stage of the long-toed salamander (Ambystoma macrodactylum) is the top vertebrate predator in high-elevation fishless lakes in the North Cascades National Park Service Complex, Washington (U.S.A.). Although most of these high-elevation lakes were naturally fishless, trout have been stocked in many of them. We sought to determine the effects of physicochemical factors and introduced trout on abundance and behavior of A. macrodactylum larvae. Larval salamander densities were estimated by snorkeling. Snorkelers carefully searched through substrate materials within 2 m of the shoreline and recorded the number of larvae observed and if larvae were hidden in benthic substrates. Physicochemical factors were measured in each lake on the same day that snorkel surveys were conducted. In fishless lakes, larval salamander densities were positively related to total Kjeldahl-N concentration and negatively related to lake elevation. Crustacean zooplankton, especially cladocerans, were important food resources for larval A. macrodactylum. Crustacean zooplankton and cladoceran densities were positively related to total Kjeldahl-N, suggesting that increased food resources contributed to increased densities of larval A. macrodactylum. Differences in larval salamander densities between fish and fishless lakes were related to total Kjeldahl-N concentrations and the reproductive status of trout. Mean larval salamander densities for fishless lakes with total Kjeldahl-N < 0.045 mg/L were not significantly different from mean larval densities in lakes with reproducing trout or in lakes with nonreproducing trout. In fishless lakes with total Kjeldahl-N a?Y 0.045 mg/L, however, mean larval densities were significantly higher than in lakes with reproducing trout where fish reached high densities. In fishless lakes with total Kjeldahl-N a?Y 0.095 mg/L, mean larval densities were significantly higher than in lakes with nonreproducing trout where trout fry were stocked at low densities. Reduced larval salamander densities in lakes with trout likely resulted from trout predation. There were no significant differences in the percentage of larvae hidden in benthic substrates between fishless lakes and lakes with fish. Our results imply that assessment of the effects of fish on amphibians requires an understanding of natural abiotic and biotic factors and processes influencing amphibian distribution and abundance.