Experimental effects of habitat fragmentation on old-field canopy insects: community, guild and species responses.

We examined the effects of habitat fragmentation on the species distributions, guild membership, and community structure of old-field insects using a fine-scale experimental approach. A continuous 1-ha goldenrod field was fragmented into four treatments that varied in both patch size and degree of isolation. Each treatment was replicated four times and arranged in a Latin square design. Canopy insects in fragmented patches were sampled with sweep nets during early and late summer 1995. The species richness of insects was significantly lower in fragmented than in unfragmented treatments during July, but was similar among treatments in September. Overall community abundance showed no treatment effect during either month. We also found significant row and column effects, suggesting there was spatial heterogeneity in species richness and abundance apart from treatment effects. Differences in species richness during July were primarily due to the loss of rare species in highly fragmented plots. Overall abundance was less responsive to community change because deletions of rare species in fragmented areas were not detected in abundance analyses. Four feeding guilds showed different responses to fragmentation: the species richness of sucking herbivores and the abundance of parasitoids were significantly reduced by fragmentation but predators and chewing herbivores were largely unaffected. Analyses of a subset of individual species within guilds suggest that the greater effects of fragmentation on sucking herbivores and parasitoids may be due to the degree of habitat specificity of guild members. The effects of small-scale habitat fragmentation were therefore detectable at the level of community, guild, and individual species. Changes in species richness, guild structure and species distributions were likely due to differential effects of habitat alteration on individual movements and patch selection rather than dispersal or demographic change. Nonetheless, the selective loss of rare species, differential guild effects and changes in species occupancy that we found in this small-scale experiment are also factors that are likely to operate in fragmented habitats over broader spatial scales.