We tested five mechanisms of coexistence in a community of three common rodent species (two gerbils, Gerbillus allenbyi and G. pyramidum, and one jerboa, Jaculus jaculus) inhabiting sand dunes in the Negev Desert, Israel. The five mechanisms, based on foraging theory, considered various forms of habitat selection in time and space. From November 1986 until January 1988, we live—trapped to census rodent populations, counted rodent spoor in tracking plots to quantify activity, and measured the rodents' giving—up densities (GUDs: the amount of food remaining within a resource patch following exploitation by a forager) in seed trays to determine relative foraging efficiencies. The population sizes of the two gerbil species tended to fluctuate synchronously (unfortunately, we could not live—trap any jerboas). G. allenbyi biased its activity towards the stabilized sand habitat and towards the bush microhabitat. In contrast, G. pyramidum and J. jaculus biased their activity towards semistabilized sand habitats, and J. jaculus also biased its activity towards the open microhabitat. Despite divergent patterns of habitat use among species, G. allenbyi tended to be the most efficient forager and J. jaculus the least efficient forager regardless of microhabitat, sand habitat, or month. G. allenbyi's presence in the community may be assured by its higher foraging efficiency. J. jaculus's presence in the community may either result from its ability to travel greater distances and utilize rare, rich resource patches, or result from an herbivorous diet. G. pyramidum's presence with G. allenbyi in the community appears to require the less stabilized sand habitats and the ability of G. pyramidum to dominate rich patches by interference or nightly temporal partitioning.