Constructed microhabitat bundles for sampling fishes and crayfishes in coastal plain streams
We investigated fish and crayfish use of standardized, constructed microhabitats (bundles) in three northern Mississippi streams. Cypress Creek and the Little Tallahatchie Canal were channelized and incised and had little woody cover; Puskus Creek was unchannelized and unincised and had abundant woody cover. We constructed three types of bundles (cane, leaf, and string) and deployed replicates of each in winter and spring. Occupancy of bundles by fish and crayfish was high and included 32 fish species representing eight families. Fish abundance did not differ among bundle types or between channel positions (bank or midchannel), but abundance of crayfish showed mixed responses to bundle type and position. Fish and crayfish use of bundles was higher in the channelized streams (89% occupied) than in the unchannelized stream (49% occupied). Furthermore, after a winter storm, fish use increased in the channelized streams but not in the unchannelized stream. Bundles yielded abundance estimates with modest to poor precision (40-73% for fish; 37-125% for crayfish); about 110-140 bundles would be necessary to consistently achieve precision of 30%. Bundles were effective for sampling a subset of fish assemblages (e.g., darters Etheostoma spp. and Percina spp. and madtoms Noturus spp.), but other fish species were conspicuously underrepresented or absent in our samples relative to sampling by electrofishing and seine (e.g., open water species and large individuals). Nevertheless, microhabitat bundles can be effective for sampling small fish and crayfish that associate with woody cover and that are difficult to sample with conventional methods.