The influence of fruit morphology and habitat structure on ant-seed interactions: A study with artificial fruits

Artificial fruits designed to simulate lipid-rich non-myrecochorous diaspores were used to test for the effect of fruit morphology and habitat structure on ant-seed interactions in an Atlantic Forest site in SE Brazil. The outcome of the interaction (i.e., if the fruit was removed, cleaned by ants on the spot or had no interaction with ants) and the time of ant response were the investigated variables. Models simulating drupes and arilate diaspores were used to test for morphological effects and four habitat attributes (litter depth, number of logs, number of trees, and percentage of bromeliad coverage on the forest floor), likely to be correlated with the ant diversity and abundance in the study site, were measured to test for the effect of habitat structure. The proportion of fruits removed or cleaned did not differ between the two morphological models. Sites in which fruits were cleaned had more trees than those in which no interaction occurred. This may be a result of the foraging behavior of arboreal ants that frequently descend to the forest floor to exploit fleshy diaspores. Sites in which model removal occurred had lower litter depth than both those in which models were cleaned and those in which no interaction occurred. A negative correlation was observed between litter depth and ant response time. Accumulation of leaf litter at a given point may have constrained the movements of large ants in general, and ponerine ants (that are important seed removers) in particular. We conclude that that local pattern in litter depth and tree density influence the frequency and outcome of interactions between ants and non-myrmecochorous, fleshy diaspores.