Barriers to movement and the response of herbivores to alternative cropping patterns.

Does the response of insect herbivores to the spatial dispersion of their food plants depend on whether alternative dispersions are available? To answer this question, four quartets of collard plots (Brassicae oleracae) were planted. Quartets contained two pairs of plots representing alternative planting designs. In one pair the alternatives were separated from one and other by a curtain barrier that effectively prevented plot to plot movement by the flea beetle, Phyllotreta cruciferae. The boundary between the other pair of alternatives was without barriers, and the beetles could range freely between these two plots. We hypothesized that the presence of a choice (i.e. the absence of a barrier) would exaggerate any differences in beetle numbers between the alternative plant arrangements.Beetles were given "choices" between two kinds of alternatives. In two quartets, the alternatives were high density collard monocultures versus low density collard monocultures. In the other two quartets, the alternatives were collard monocultures versus collard-potato discultures. In each quartet, the abundances of P. cruciferae were censused weekly for three weeks.As we hypothesized, the influence of cropping pattern (both diversity and density) was less where there was a barrier between cropping alternatives than where there was no barrier between the same alternatives. This suggests that nonrandom foraging movement, or "patch choice", explains the response of Phyllotreta to cropping treatments. The consequence of this mechanism for interpreting the results of insect-host experiments are substantial; experimental designs allowing insects to move and accumulate in a preferred treatment may have different results from the same treatments in isolation.