Mice, big mammals, and seeds: it matters who defecates what where

Abstract

Abstract. By placing pseudodefecations of horse dung and cow dung mixed with guanacaste seeds (Enterolobium cyclocarpum) , guapinol seeds (Hymenaea courbaril), black beans (Phaseolus vul- garis), wild lima beans (Phaseolus lunatus), and chaperno seeds (Lonchocarpus costaricensis) (all Leguminosae) in various ages and kinds of forest (and pasture) in seasonally dry Santa Rosa Na- tional Park (northwestern lowland Costa Rica), I determined that an initial seed shadow may be thinned and trimmed differentially by post-disper- sal seed predation; the differences depend at least on what kind of animal defecates the seeds, what kind of seeds they are, and where the defecations occur. Spiny pocket mice Liomys salvini (Hetero- myidae) were the seed predators that removed the seeds from the dung. This was a dry season experi- ment and therefore uncomplicated by seed burial by large dung beetles (Scarabaeidae). The Liomys mice found large seeds more readily than they found small ones, preferred to mine in horse dung rather than in cow dung, removed the seeds more thoroughly from dung with many seeds than from dung with few seeds, learned to reject toxic seeds that they found in the dung, and were more thor- ough in seed removal in forest than in pastures. In the first few days after pseudodefecations were placed out in the forest, there were stronger differ- ences between horse and cow dung, and among habitats, than after the mice had 20-40 days to find all the pseudodefecations and remove as many seeds as they would. In the end, the mice not only removed the majority of the seeds from the dung, but they also removed all of the seeds from the pseudodefecations with high seed density. They also removed all of the seeds from more than half of the 5 liter pseudodefecations that contained a low concentration of seeds (20 guanacaste seeds, five guapinol seeds and 20 black beans). The latter type of trimming of the seed shadow, as opposed to simply thinning it, cannot be compensated for by density-dependent seedling survival.