Rodent granivory strengthens relationships between seed size and plant abundance in a desert annual community

Aim: The seed size/number trade-off model explains the co-existence of plants that produce different-sized seeds by assuming an establishment fitness advantage for larger seeds via enhanced competition or environmental tolerance, and a colonization fitness advantage for smaller seeds. The model predicts a negative relationship between seed mass and plant abundance among co-existing species in a community. Here, we evaluate the hypothesis that rodent granivory affects the slope of this relationship, causing it to become less negative when granivore pressure is reduced because rodents selectively harvest large seeds. Location: Chihuahuan Desert, USA. Methods: We analysed long-term data from an annual plant community in southeast Arizona to characterize the relationship between seed mass and mean plant abundance in three types of plots thatmanipulated the rodent community. Some plots excluded all rodents, other selectively excluded only large-bodied kangaroo rats (Dipodomys), while control plots allowed access by all rodents. Results: We observed negative relationships between seed mass and plant abundance on 23 of the 24 experimental plots. The slopes of these relationships were significantly affected by the experimental treatments. Slopes were most negative for control plots and least negative for plots that excluded all rodents. Conclusion: Rodent granivore exclusion caused the slope of the relationship between seed mass and plant abundance to become less negative. These results illustrate how granivory can affect seed mass-plant abundance relationships, and may provide further insight into the co-existence of plants that vary in seed size.