Landscape structure should have important and predictable influences on the abundance and distribution patterns of arthropod predators. However, there are few studies that manipulate landscape structure to investigate possible effects on predator density. In response to this gap in our understanding of spatial patterns of predators, we examined the role of landscape structure on the density patterns of spider guilds in an experimental agroecosystem. We created a red clover (Trifolium partense) agroecosystem in which clover patch size, fragmentation of clover, and composition of the surrounding matrix (grass or bare ground) were manipulated according to a factorial design. We measured density of spider guilds and common spider prey in each plot using D-Vac suction sampling. Overall, density was highest in the large patches for all spider guilds: running, ambush, and web building. Ambush and web-building spiders were less common and we uncovered no interactive effects on their abundance. However, there was a significant three-way interaction between size, fragmentation and matrix composition on the density of the running spider guild. Additionally, there was a negative correlation (r2=0.l6, p=0.028) between the density of the dominant leafhoppers (Agalia spp.) and the running spiders, which would indicate that these predators are depressing potential prey numbers and/or that these taxa respond in different ways to landscape structure. The results of this study suggest that landscape structure influences the aggregation of spider guilds and has complex effects on highly mobile predators that appear to be influenced by a combination of landscape attributes and prey habitat use.