Effects of Host Plant Patch Size on Herbivore Density: Underlying Mechanisms

Abstract

This study was designed to examine the mechanisms underlying the response of herbivores to variation in the size of host plant patches. In order to explain why identical changes in patch size affect closely related herbivores in different ways, I separated the effects of patch size alone from the confounded effects of the surrounding plant community. Further tests determined whether herbivore densities were directly affected by surrounding nonhost plants (via changes in movement patterns) and/or indirectly affected by changes in host plant size/quality caused by the surrounding nonhost plants. Population densities of three closely related insect herbivores, Acalymma vittatum (SCB), Diabrotica undecimpunctata howardi (SCR), and Diabrotica virgifera (WCR), were studied in experimental patches of squash host plants (Cucurbita maxima). Patches varied in size (4, 16, or 64 plants per patch) and in the nature of their edges (no tomato plant edge, an edge of tomato plants growing in the ground [=interacting nonhost plants], and an edge of tomato plants growing in pots [=non—interacting nonhost plants]). This experimental design enabled a comparison of herbivore densities in patches with no nonhosts surrounding patches, patches with interacting nonhost neighbors, and patches with non—interacting nonhost neighbors. The three herbivore species differed in their response to both patch size and edge treatment. Densities of first generation SCB were significantly affected by both patch size and edge treatment, being greatest in the smallest patches. For a given patch size, densities were greatest in the patches with non—interacting tomato neighbors (tomatoes in pots). More second generation beetles emerged per plant in the treatment with non—interacting plants, but the total number of second generation individuals per plant was not affected by either patch size or edge treatment. SCR densities were significantly affected by edge treatment only, with individuals being most abundant in patches with non—interacting tomato neighbors. For both species, differences in densities as a function of edge treatment were of much greater magnitude in small patches than in large patches. In contrast, population densities of WCR were not influenced by either patch size or edge treatment. The growth (number of leaves) and reproduction (number of flowers and yield) of squash plants were significantly affected by edge treatment, with the largest plants occurring in patches with non—interacting tomato plants and the smallest plants in patches with interacting tomato plants. Patch size influenced both number of leaves and yield. Plant longevity was not significantly influenced by either patch size or edge treatment. The role of these differences in plant parameters in influencing beetle population dynamics was investigated for SCB. For a given average plant size per patch, beetle densities were significantly greater in the treatment with non—interacting nonhosts than in the treatment with no tomato plant edge. A mark—recapture study with SCB and SCR revealed that rates of recapture were not affected by either size or edge treatment, for either herbivore species. However, a greater proportion of both SCB and SCR individuals colonized neighboring small patches than neighboring large patches on a per plant basis. In order to test the mechanisms underlying these herbivore responses to patch size, experiments were conducted with SCB to examine effects of neighboring nonhost vegetation on herbivore feeding preferences and movement patterns. When given a choice of potted plants inside cages, beetles preferred a squash plant alone compared to a squash plant growing next to a tomato plant in a separate pot. When given a choice of leaves (but no direct contact with plants), beetles did not distinguish between leaves from plants grown with conspecific vs. tomato neighbors, indicating beetles did not perceive differences in leaf quality as a function of neighbor type. Observations of movement patterns of SCB beetles in small patches with and without surrounding tomato plants showed that a significantly greater proportion of beetles left patches without surrounding tomato plants than patches with an edge of tomato plants. Thus, surrounding nonhost plants act as a reflective boundary, retaining insects within patches. Movement behavior was also influenced by patch size. Comparisons of flight patterns in 4—plant and 16—plant patches surrounded by tomatoes showed that a greater proportion of beetles emigrated from the large patch because beetles had a lower probability of landing on a tomato plant in large patches. Results from this study for SCB support both hypotheses explaining patch size effects. Beetle densities are affected by patch size, even with no confounding nonhost neighbors present, but they are also affected by the surrounding nonhost community. The effect of nonhost neighbors on population densities consists of a direct positive effect mediated by movement patterns, as well as an indirect negative effect mediated by changes in plant size. The relevance of these results to general patterns of herbivore response to plant spatial pattern is discussed.