An experimental test of the habitat amount hypothesis reveals little effect of habitat area but transient or indirect effects of fragmentation on local species richness

Context The habitat amount hypothesis (HAH) posits that local species richness is driven more by the amount of habitat in the surrounding landscape than by local patch size or habitat configuration. Habitat amount and configuration influence patch attributes, however, making it difficult to isolate these relative effects on local richness. Objectives We tested the HAH in an experimental system in which the amount (10-80%) and configuration (clumped vs. fragmented) of habitat (red clover) were adjusted independently within individual 'landscapes' (256-m(2) plots). Methods We used generalized linear models and model-selection criteria to evaluate how arthropod richness at two local scales (1 m(2) and 4 m(2)) varied as a function of the amount or fragmentation of habitat and of the size or shape of the local patch. Results Local richness was largely independent of the amount or configuration of habitat, as well as patch size or shape. Local richness was best modeled as a constant for all surveys but the first, in which fragmentation alone best explained variation in species richness, followed by patch shape (which is an indirect effect of fragmentation). Habitat amount had an overwhelmingly positive effect on arthropod richness at the landscape scale, however. Conclusions The HAH was not supported in this experimental system: local richness was generally unaffected by the amount of habitat in the surrounding landscape, whereas habitat configuration was sometimes important. Given that habitat amount affected landscape-wide richness, it may be that the HAH only applies at the 'landscape scale', at least in this system.