Induced vs. constitutive resistance and the spatial distribution of insect herbivores among plants

Because herbivores can move away from plants of low quality, induced resistance should affect the spatial distribution of herbivore damage (or herbivores) within or among plants. We used a spatially explicit simulation model and data from a field experiment to examine the effect of induced resistance on the distribution of herbivores and their damage among plants within a population. The model is appropriate only for rapid (as opposed to delayed) induced resistance, assumes that resistance affects herbivore behavior but not performance, and assumes no interplant communication. It is also appropriate only for induced resistance that decays in the absence of herbivore damage. The simulation tracks herbivores on a linear array of plants or plant parts with reflecting boundaries and allows manipulation of the timing of induced resistance and the initial distribution of herbivores. Our model suggests that increasing lags and thresholds for the production of induced resistance causes increasing aggregation of herbivores and their damage, but in the absence of either a lag or a threshold, induced resistance can lead to an even distribution of herbivores. The formation of even or aggregated distributions of herbivores depends not on initial distributions, but on the characteristics of the induced resistance. In the field experiment, we measured the degree of aggregation of Mexican bean beetle larvae among plants within experimental soybean populations consisting of soybean varieties that had no resistance (low constitutive, low inducible resistance), induced resistance (and low constitutive resistance), or constitutive resistance (and low inducible resistance). In general agreement with qualitative predictions of the model, we found that larvae on soybean varieties with induced resistance remained significantly more aggregated than larvae on varieties with no or high constitutive resistance.