Four mechanisms of coexistence are considered that may contribute to the diversity of desert granivorous rodent communities. In the first, bush/open microhabitat selection, coexistence is possible if there is a trade—off between foraging efficiency in the bush and open microhabitats. In the second, temporal variation in resource abundances, coexistence is possible if there is a trade—off between foraging efficiency and maintenance efficiency. The first species can forage profitably on low resource abundances while the second uses dormancy to travel inexpensively in time between periods of high resource abundances. In the third, spatial variation in resource abundance, coexistence is possible if there is a trade—off between foraging efficiency and the cost of travel. The first species forages patches to a lower giving—up density, (the density of resource at which a forager ceases foraging), while the second can inexpensively travel between patches with high resource abundances. In the fourth, seasonal rotation in foraging efficiencies, coexistence is possible if there is a trade—off between the costs of foraging during different seasons. The species that is the more efficient forager changes seasonally. The first mechanism of coexistence has received much empirical attention and support. The other three have not previously been considered with desert rodents. In a community of four granivorous rodent species, I used artificial seed patches to measure species— and habitat—specific foraging efficiencies and live—trapping to measure population sizes and mean distances between recaptures. Of the four, the fourth mechanism of coexistence best explained the presence of Perognathus amplus, Dipodomys merriami, and Spermophilus tereticaudus in the community. Each species enjoyed a period of the year during which it was the most efficient forager. Furthermore, the annual population densities of these three species fluctuated out of phase. Seasonal changes in species—specific predation risks and body size—dependent metabolic costs may have accounted for these results. The third mechanism of coexistence best explained the presence of Ammospermophilus harrisii in the community. This species preferred to forage a large number of widely spaced patches to a high giving—up density rather than forage a few patches to a low giving—up density.