Abstract
The ability to partition the variation of faunal assemblages into homogenous units valuable for biomonitoring is referred to as classification strength (CS). In this study, the CSs of three types of geographic classifications: watershed basin, ecological region, and caddisfly region, were compared based on 248 light trap samples of adult caddisflies collected in Minnesota during 1999-2001. The effect on CS of three different levels of taxonomic resolution: family, genus, and species, was also assessed. Primary (broadest possible) a priori classification by watershed basin and ecological region had a lower CS than did secondary classification by these regions. Caddisfly region, an a posteriori classification based directly on caddisfly distribution data, had nearly twice the CS of any a priori classification. CS decreased approximately 20% with a decrease in taxonomic resolution from species to genus, and from genus to family. These results suggest that geographic classification, spatial scale, and taxonomic resolution are all important factors to consider when sampling aquatic insects, and that widely used a priori geographic classifications are not the ideal units for sampling the aquatic biota.