Experimental and Mathematical Analyses of Herbivore Movement: Quantifying the Influence of Plant Spacing and Quality on Foraging Discrimination

Using mark—recapture experiments, I studied the foraging movements of two crucifer feeding flea beetles, Phyllotreta cruciferae and P. striolata. To determine the influence of distance between host patches, and their quality, on beetle movement, I released marked beetles in experimental arrays of collards (Brassica oleracea). I quantified and interpreted movement behavior by applying a passive diffusion model and a continuous—time Markov model to recapture distributions. These models provided both a standard protocol for interspecific comparisons and a formal tool for translating movements into steady—state foraging patterns. In homogeneous arrays (all collard patches of similar quality), P. cruciferae and P. striolata moved in accordance with a constant—coefficient diffusion model. The dispersal rates of both beetles increased as the distance between patches declined; thus, patches of collards acted as stepping—stones to accelerate beetle dispersal. Beetle movement was hampered by old—field vegetation; beetles moved along cultivated ground in lieu of entering dense stands of goldenrod. Under all experimental conditions, P. cruciferae was consistently more mobile than P. striolata. By differentially fertilizing and pruning collards, I created "lush" and "stunted" patches of collards, which were arranged alternately along linear arrays. To determine the influence of patch dispersion on foraging selectivity, I varied the distance between these patches of unequal quality. The degree of foraging discrimination exhibited by the beetles was greatest where patches were closest together, and declined as distance between patches increased. At comparable spacings, P. cruciferae consistently exhibited a higher degree of foraging discrimination than did P. striolata. I developed a Markov model of patch—to—patch dispersal in which each flea beetle's probability of movement depends on patch quality and distance to neighboring patches. Using 60—min observations of marked flea beetles to estimate model parameters, I found that the steady states of foraging discrimination predicted by the model matched the discrimination levels exhibited by undisturbed flea beetles in comparable control arrays of collard patches. Thus, simple movement rules were extrapolated into expected foraging behavior; the dependence of that foraging behavior on patch dispersion was shown to be mediated by varying levels of mobility. In particular, proximity between heterogeneous patches affords beetles greater opportunity for nonrandom foraging than when patches are widely scattered. Since movement is the mechanism by which beetles respond to heterogeneous environments, rates and patterns of movement determine each species' ability to concentrate its numbers on the best foodplants. P. cruciferae is more mobile than P. striolata and in turn can better adjust to variation in collard quality. The interplay between beetle movement and plant dispersion may have consequences well beyond the simple distribution of feeding activity among heterogeneous host patches.