Previous studies of the population dynamics of the cereal leaf beetle (CLB), Oulema melanopus, have neglected spatial considerations despite the important role played by dispersal in the life history of this species. In this paper we describe a simulation model of CLB spatiotemporal dynamics. The model was used to examine qualitatively different hypotheses about dispersal in this species. The hypothesis of random diffusion, in which beetle movement is undirected, was contrasted to hypotheses of attraction to host crops and movement away from nonhosts. Model predictions were compared to field observations of beetle distribution and abundance in a 41-km 2 region of southwestern Michigan, U.S.A., in 1976 and 1977. Evaluation criteria were peak and total seasonal abundances of CLB in grain crops in the region, and densities in 69 individual fields. The random model gave satisfactory results on a regional scale, but simulated densities for individual fields were too homogeneous and were generally poorly correlated with actual densities. The alternative models performed better in most respects, although field-to-field variation in density was still too low. Without specific studies of dispersal behavior, there is no basis for selecting one hypothesis over another. The model emphasizes the importance of local features of fields in producing spatial variations in insect density. The between-field environment plays a prominent role in the dynamics of abundance and distribution of beetles in host crops. Factorial sensitivity analysis was used to examine the influence of key parameters for which there are insufficient data. A very high rate of diffusion in nonhost habitats was needed with the random model to obtain simulated peak densities in host crops as high as those observed. To reduce the density of insects in resistant wheat relative to susceptible wheat, as observed, this crop must either repel beetles or cause beetles to move randomly at a much higher rate. Simulated late-season densities in host crops were too high, suggesting either strong seasonality in the dispersal behavior of adults, or an increase in their mortality rate as the season progresses.