Harvester ant foraging and shrub-steppe seeds: interactions of seed resources and seed use

Granivore-seed interactions involve a feedback between granivore seed selectivity and seed availability. We examined this feedback to determine how seed preferences by the western harvester ant, Pogonomyrmex occidentalis, related to seed availability and, in turn, affected the soil seed pool. Preferences were estimated from natural diets as well as from experiments that controlled seed size, relative availability, and distance from ant nests. Seed availability to ants varied with season and over 2 yr. Colony activity and seed intake rates were correlated with seed availability. Seed preference by ants was correlated with the seasonal availability of preferred species, but not with unpreferred seeds. From the soil seed pool, ants preferentially harvested small, sound seeds. They removed 9-26% of the potentially viable seed pool each year, and as much as 100% of available preferred species. Seed densities were lower 2-7 m from nests, where foraging activity was concentrated, than 7-12 m from nests. In controlled preference experiments. P. occidentalis was unselective near nests, but preferred large seeds with higher assimilable energy content in trials 10 m from nests. A relatively low foraging activity > 7 m, however, suggests that this distance-dependent preference is rarely manifested in natural conditions and does not measurably affect soil seed dynamics. Our results point to the importance of studying diet choice in a natural context; preferences measured under unexperimental conditions may not correspond to natural diets. Such discrepancies in food preference measurements will affect predictions about how consumers influence the population dynamics of resource organisms.