Effects of plant neighborhoods on plant-herbivore interactions: resource dilution and associational effects

• Published in: Ecology (Durham) Vol. 95; no. 5; p. 1370
• Main Authors: Hambäck, Peter A, Inouye, Brian D, Andersson, Petter, Underwood, Nora
• Format: Journal Article
• Language: English
• Published: United States 01.05.2014
• Subjects: Animals; Herbivory - physiology; Models, Biological; Plants - classification; Population Dynamics
• ISSN: 0012-9658; 1939-9170
• DOI: 10.1890/13-0793.1

Effects of neighboring plants on herbivore damage to a focal plant (associational effects) have been documented in many systems and can lead to either increased or decreased herbivore attack. Mechanistic models that explain the observed variety of herbivore responses to local plant community composition have, however, been lacking. We present a model of herbivore responses to patches that consist of two plant types, where herbivore densities on a focal plant are determined by a combination of patch-finding, within-patch redistribution, and patch-leaving. Our analyses show that the effect of plant neighborhood on herbivores depends both on how plant and herbivore traits combine to affect herbivore movement and on how experimental designs reveal the effects of plant density and plant relative frequency. Associational susceptibility should be the dominant pattern when herbivores have biased landing rates within patches. Other behavioral decision rules lead to mixed responses, but a common pattern is that in mixed patches, one plant type experiences associational resistance while the other plant experiences associational susceptibility. In some cases, the associational effect may shift sign along a gradient of plant frequency, suggesting that future empirical studies should include more than two plant frequencies to detect nonlinearities. Finally, we find that associational susceptibility should be commonly observed in experiments using replacement designs, whereas associational resistance will be the dominant pattern when using additive designs. Consequently, outcomes from one experimental design cannot be directly compared to studies with other designs. Our model can also be translated to other systems with foragers searching for multiple resource types.