Intentional introductions and biodiversity in fishless waters: The effects of introduced fish on native aquatic species

The introduction of nonnative fishes to naturally fishless waters has been associated with declines in a variety of native aquatic species. This study examines effects of two nonnative fishes in two separate fishless systems. We conducted separate field experiments to test effects of the western mosquitofish (Gambusia affinis) on survival and growth of larval California tiger salamanders (Ambystoma californiense) in vernal pool and permanent pond conditions. In vernal pool conditions, with low initial fish densities, fish did not affect larval survival or size at metamorphosis. At high initial fish densities characteristic of permanent ponds, a four-fold reduction in larval survival and significant reductions in the size of surviving salamanders occurred with fish. Larvae gained a size advantage by delaying metamorphosis in controls, but in fish-ponds there was a trend for metamorphosis at smaller sizes as the experiment progressed. Another experiment showed that Gambusia reduced survival of a vernal pool fairy shrimp, Linderiella occidentalis. Sixty-four percent survived in controls while 17% survived with Gambusia. In mountain lakes and pools, introduced trout (Salmonidae) were associated with loss of long-toed salamanders (Ambystoma macrodactylum sigillatum) from historic sites. Salamander larvae were present in 92% of fishless sites, but only 38% of sites where trout had been introduced. In an enclosure experiment in a fishless mountain lake, introduced trout reduced survival of salamander larvae by 58%. Survivors were smaller in enclosures with trout: larval snout-vent lengths averaged 10.85% shorter than in controls, while larval mass was 25.67% lower than in controls. Finally, in surveys of 144 lakes and pools, A. m. sigillatum were 5.97 times more likely to be found in sites where trout were absent when other habitat variables and spatial associations were considered. The anuran, Rana muscosa was rare within the study area. Depth was a significant predictor of presence of R. muscosa tadpoles; for sites >2 m deep, reproductive R. muscosa populations were more likely to occur in sites lacking trout. In some years Ambystoma larvae overwintered in some deep and shallow high-elevation sites; however, during drought conditions, first year Ambystoma larvae metamorphosed from drying sites at all elevations.