Studies of the response of natural enemies to vegetational diversity suggest that movement behavior plays an important role in determining natural enemy abundance in diversified agroecosystems. A simple mathematical model for the response of natural enemies to interplanted strip vegetation is developed based on the assumption that natural enemy movement can be represented as a diffusion process. Attractiveness of interplanted strips, resulting from strip vegetation having an abundance of resources, is represented by natural enemies having lower diffusion rates in the strips. Results of simulations with the model suggest that interplanted vegetation acts as a source of natural enemies when natural enemies colonize strip vegetation before crop germination, but acts as a sink when crop and interplanted vegetation germinate simulataneously. The magnitude of this effect varies with natural enemy mobility. Spatial patterns exhibited by natural enemies will be influenced by mobility and do not reliably indicate whether or not augmentation is occurring. There is a strong interaction between natural enemy mobility and experimental design, suggesting that the results of small-scale studies with agroecosystem diversification must be interpreted with caution. The ability of our model to account for much of the variability in natural enemy responses to diversification suggests that this variability may be caused by an interaction between natural enemy movement and system design rather than by fundamental differences in natural enemy behavior.