Population growth is often affected by environmental variables. How is growth affected by the pattern of variation of an environmental variable? How does this variable interact with the degree of mobility of a population? A numerical model was developed for population growth in a grid of cells. This model includes migration, selection, and mating. Simulations were run for a grid representing the Yucatan Peninsula. One realistic predictor [precipitation] was used to model selection. Simulations were run under several conditions. There were two spatial patterns of variation of the local selective pressures, and three levels of migration. The same initial population, based on a small block of cells, was used for all simulations. Results/Conclusions As expected, the rate at which subpopulations adapt to the local optimum and the degree of adaptation are affected by the rates of migration between cells. They are also affected by the spatial structure of the selective pressures. If selective pressures are patterned on the observed map of precipitation, populations tend to adapt quickly to local conditions. When selective pressures are randomly assigned to cells, there is slower growth and a much higher proportion of subpopulations that are persistently maladapted. We speculate on the meaning of these results for niche conservatism. Our simulations extend previous results by Holt and Gomulkiewicz (1997) and Kirkpatrick and Barton (1997) because we considered an explicit geographic setting with contrasting patterns of selective pressures.