Contrasting patterns of short-term indirect seed-seed interactions mediated by scatter-hoarding rodents.

It is well known that direct effects of seed predators or dispersers can have strong effects on seedling establishment. However, we have limited knowledge about the indirect species interactions between seeds of different species that are mediated by shared seed predators and/or dispersers and their consequences for plant demography and diversity. Because scatter-hoarding rodents as seed dispersers may leave some hoarded seeds uneaten, scatter hoarding may serve to increase seed survival and dispersal. Consequently, the presence of heterospecific seeds could alter whether the indirect interactions mediated by scatter-hoarding rodents have a net positive effect, creating apparent mutualism between seed species, or a net negative effect, creating apparent competition between seed species. We present a testable framework to measure short-term indirect effects between co-occurring plant species mediated by seed scatter-hoarding rodents. We tested this framework in a subtropical forest in south-west China using a replacement design and tracked the fate of individually tagged seeds in experimental patches. We manipulated the benefits to rodents by using low-tannin dormant chestnuts as palatable food and high-tannin non-dormant acorns as unpalatable food. We found that seed palatability changed the amount of scatter hoarding that occurred when seeds co-occurred either among or within patches. Consistent with our predictions, scatter-hoarding rodents created apparent mutualism through increasing seed removal and seed caching, and enhancing survival, of both plant species in mixed patches compared with monospecific patches. However, if we ignore scatter hoarding and treat all seed harvest as seed predation (and not dispersal), then apparent competition between palatable chestnuts and unpalatable acorns was also observed. This study is the first to demonstrate that foraging decisions by scatter-hoarding animals to scatter hoard seeds for later consumption (or loss) or consume them can influence indirect effects among co-occurring seeds, and rodent-mediated indirect effects vary depending on whether the harvested seeds are hoarded or eaten.