Experimental studies of seed predation in old-fields.

In a pair of experiments conducted in old-field habitats in southwestern Michigan (USA), we examined rates of seed loss to post-dispersal predators (ants and rodents). Seeds from 4-6 species of "biennial" plants were tested over a range of seed densities and habitat types. We found that seed removal was significantly higher in vegetated habitats than in areas of disturbed soil (both simulated small-animal diggings and a plowed field). In the undisturbed vegetation, seed losses ranged from 1-20% of seeds removed/day.An exclosure experiment demonstrated that ants and rodents foraged selectively for seeds of the six plant species tested. Rodents (Peromyscus maniculatus) fed preferentially on species producing large seeds (predominantly Tragopogon dubius). Ants (Myrmica lobicornus) foraged on smaller seeds, although their foraging preferences were not based strictly on seed size.Seed density had only a minor effect on predation rate over the range of densities tested. Predators, instead appeared to treat each experimental group of seeds as a single prey patch. Consequently, predation intensity was quite variable over distances of <20 m within a relatively homogeneous section of habitat.These field experiments provide initial estimates of seed losses to post-dispersal predators in old-field habitats. Rates of seed loss were generally less than those reported from desert or semi-arid habitats. However, for some old-field species, seed losses averaged an appreciable 10-20% day. The selective nature of the seed predators, plusthe relative patchiness of predation intensity in space, suggest that postdispersal seed predation can play a role in determining the distribution and/or abundance of old-field herbs.