Abstract
Using alternating bands of weeds and broccoli I experimentally manipulated vegetation composition and the spatial scale at which the landscape was fragmented in a factorial design. This experimental approach allowed me to distinguish the effect of spatial scale from that of simple crop heterogeneity on crop herbivores. The importance of scale depended on which insect species were examined. Cabbage aphids (Brevicoryne brassicae) were influenced by vegetation composition at all tested scales of fragmentation; cabbage butterflies (Pieris rapae) were not affected by scale or by composition and flea beetles (Phyllotreta cruciferae) revealed a striking dependence on scale of fragmentation as well as an interaction between scale and composition. This approach shows the importance of dissecting out the effects of scale from other aspects of landscape manipulation, and emphasizes the challenge of developing a theory that will enable prediction of species–specific responses to scale.