Abstract

Aim: This study aims to evaluate the relative importance of birds, small mammals, and ants as seed predators at a semi-arid site in northern Chile. Location: Northern Chile, in Parque Nacional Bosque Fray Jorge (30°41? N, 71°40? W, c. 80 m elevation). Methods: We studied the relative abilities of birds, small mammals, and ants to find and remove millet seeds either singly (i.e. background seed removal) or in bulk. Single seeds were set in shallow depressions in Plexiglas trays which were established in long and arbitrary transects, and were available either to birds (diurnally) or small mammals (nocturnally) or were covered by hardware cloth and therefore available only to ants. Bulk removal was evaluated with seeds in Petri dishes that also were established in long and arbitrary transects, and trays were either open diurnally (birds) or nocturnally (small mammals); a third set of trays was covered with hardware cloth cages to excluded vertebrates, and ants were given access to Petri dishes with twigs that were arranged across the edge of the dishes. All experiments lasted four days and nights, and trays and dishes were checked and replenished as needed in the morning and evening. In the former study vertebrate consumption was determined as the mean number of seeds removed from trays (within a given transect) minus the number removed from ant-only trays. Because ants were rarely seen in vertebrate access Petri dishes, however, we did not correct consumption there. Treatments were compared using repeated measures mixed model analysis of variance. In addition to evaluating patterns within this community, we compared our results against those obtained in similar studies in various arid regions. Results: Diurnal seed consumption was significantly greater than nocturnal seed consumption, which in turn was significantly greater than consumption by ants. Diurnal consumption was highly seasonal, evidently corresponding to the seasonal arrival and departure of migratory birds. In general, South American sites exhibit much lower levels of seed predation than sites in the northern hemisphere, but removal at our site appears to be much greater and more strongly avian-dominated than at other sites in South America. Our results are consistent with predictions based on a hypothesis relating precipitation to seed predictability. Main conclusions Both birds and small mammals were much more important seed consumers at our site than elsewhere in South America, whereas ants have been relatively unimportant at all South American sites studied to date. Although the dominant seed consumers differ across sites, overall levels of seed removal appear similar in South America and Australia, and substantially lower than reported from sites in the northern hemisphere and Africa.