COS 49-4 - Scale dependence of dietary specialization in mammalian herbivores

Abstract: Characterizing niche space is critical for predicting species interactions, extinction risk, and responses to environmental change. While diet is an extensively studied component of niche space, the extent to which animals are dietary specialists remains controversial. Although sampling at different geographic, taxonomic, or temporal scales likely contributes to this debate, it is unclear how scale influences study outcomes. To examine how scale impacts estimates of specialization and whether functional trade-offs constrain individual niche width, we characterized the diets of mammalian herbivores in the genus Neotoma (“woodrats”). These rodents consume a variety of chemically defended plants, with evidence that populations adapt to specific toxins in their diets. Using these widespread herbivores, we examined dietary specialization across temporal, geographic, and taxonomic scales. We first examined intra-individual and temporal variation using >180 samples from N. lepida and N. bryanti collected across four seasons and five years at one site in California, including 44 individuals recaptured 2 to 7 times (mean recaptures = 3). We then examined geographic variation using 20 populations of N. lepida sampled across their range in the western United States. Finally, using data from over 50 Neotoma populations sampled across all 13 woodrat species in the continental USA, we tested whether individual diet breadth increased with increasing population niche breadth. This large data set allowed us to test (1) if niche breadth is constrained by functional trade-offs related to consuming toxic diets, and whether (2) evidence of individual specialization can be detected at multiple scales. Diet components were identified in the feces of wild caught woodrats using plant metabarcoding. For a subset of animals, we also analyzed diet using stable isotopes. This approach underestimated variation as differences in C3 plants strongly influenced diet diversity, suggesting that metabarcoding more accurately characterizes variation in these herbivores. Across populations we found that diet diversity varied, closely located populations often fed on different assemblages, and populations ranged from obligate specialists to generalists. Diet specialization also varied across sampling scales, with range-wide estimates consistently higher than estimates for individual populations. These results emphasize the importance of accounting for sampling scale when measuring specialization and provide insight into the mechanisms underlying diet specialization in free-ranging herbivores.