Guidelines for the use of hyperdiverse taxa in biological monitoring: Change through time in a Southern Appalachian spider assemblage (North Carolina, Tennessee)

This dissertation addresses issues surrounding the inclusion of terrestrial arthropods in conservation planning. I begin the process of developing analytical tools appropriate to a representative arthropod group, the spiders (Order Araneae).; One of the difficulties associated with studying a hyperdiverse taxonomic group is ensuring accurate and timely species identification. Using technology similar to that employed in handwriting analysis and/or face recognition, I present an automated species identification system which makes use of neural networks. This technology, once perfected, will increase the accuracy of identifications by non-specialists and thereby reduce the need to burden systematists with routine identifications of ecological collections.; Using data collected in the Southern Appalachians, I review and evaluate some commonly used analytical techniques in biodiversity studies: rarefaction (scaling down all samples to the size of the smallest), diversity indices, and extrapolation (estimating the total richness of a site based on a sample taken from the site). The diversity indices, Shannon-Weiner and Simpson's, were surprisingly robust to differences in sample size and they detected both successional and disturbance-induced changes in spider assemblages. Directional trends in the diversity indices through time proved to be most informative in conjunction with pair-wise statistical tests. Turnover and Similarity indices were also useful for quantifying change in the assemblages.; Using the same data-sets, I identify and discuss potential sources of error which are characteristic of studies involving hyperdiverse taxa and make recommendations for eliminating or reducing this error. I find that collector bias is a significant problem, in that each collector is not sampling a statistically random subset of the community. But, as expected, some collection techniques are more subject to this kind of bias than others. Also, inclusion of juveniles in diversity estimates and analysis should be avoided as much as possible as their presence appears to obscure ecological trends.; I used the data-sets to investigate the intra-annual (early vs. late summer) and inter-annual (yearly vs. decadal) variability of spider assemblages in six Southern Appalachian habitats. Spider assemblages show greater seasonal variation than yearly variation. The diversity and variability of the spider assemblages I studied were closely tied to gross habitat structure.