Effectiveness of removal sampling for determining salamander density and biomass: A case study in an Appalachian streamside community

We surveyed a streamside community in a southern Appalachian old-growth forest to determine the feasibility of using removal sampling to estimate salamander density and biomass. We removed salamanders during 21 nightly searches of two 30 x 30 m plots from 29 June to 20 July 1999. Despite the removal of 2433 animals and a lack of evidence for significant trespass onto plots, there was no overall decline in total catch per night Moisture levels associated with recent rainfall history strongly influenced the surface activity of small species (Desmognathus carolinensis, Desmognathus wrighti Eurycea bislineata wilderae) but not large species (Plethodon jordani, Plethodon glutinosus, Plethodon yonahlossee). More refined analyses of individual species using either regression analysis (large species) or covariance analysis (small species adjusted for moisture conditions) indicate that the catch of all species except D. wrighti declined with time. The estimated number of consecutive nights required to remove 70% of the population varied from eight (P. yonahlossee) to 32 (D. carolinensis) and tended to be inversely related to the body size of the species. Conservative estimates of total salamander density and wet biomass in streamside habitats at this site are 18,486 individuals and 16.53 kg per hectare. These values are seven and 14 times greater than those reported for the Hubbard Brook Experimental Forest in New Hampshire and underscore the importance of salamanders in riparian zone communities of southern Appalachian forests.