Species’ partitioning of resources remains one of the most integral components for understanding community assembly. Analysis of stable carbon and nitrogen isotopes in animal tissues has the potential to help resolve patterns of partitioning because these proxies represent the individual’s diet and trophic niche, respectively. Using free-ranging rodents in a southern African savanna as a model community, we find that syntopic species within habitats occupy distinct isotope niches. Moreover, species with strongly overlapping isotope niches did not overlap in their spatial distribution patterns, suggesting an underlying effect of competitive exclusion. Niche conservatism appears to characterize the behaviour of most species in our sample - with little or no observed changes across habitats - with the exception of one species, Mastomys coucha. This species displayed a generalist distribution, being found in similar abundances across a variety of habitats. This spatial pattern was coupled with a generalist isotope niche that shifted across habitats, likely in response to changes in species composition over the same spatial gradient. The case for M. coucha supports contentions that past competition effects played a significant evolutionary role in shaping community structures of today, including the absence of strong interspecific niche overlaps within particular habitats. Our study highlights the value of stable isotope approaches to help resolve key questions in community ecology, and moreover introduces novel analytical approaches to quantifying isotope niche breadths and niche overlaps that are easily comparable with traditional metrices.

Discussions about early hominin diets have generally excluded grass leaves as a staple food resource, despite their ubiquity in most early hominin habitats. In particular, stable carbon isotope studies have shown a prevalent C component in the diets of most taxa, and grass leaves are the single most abundant C resource in African savannas. Grass leaves are typically portrayed as having little nutritional value (e.g., low in protein and high in fiber) for hominins lacking specialized digestive systems. It has also been argued that they present mechanical challenges (i.e., high toughness) for hominins with bunodont dentition. Here, we compare the nutritional and mechanical properties of grass leaves with the plants growing alongside them in African savanna habitats. We also compare grass leaves to the leaves consumed by other hominoids and demonstrate that many, though by no means all, compare favorably with the nutritional and mechanical properties of known primate foods. Our data reveal that grass leaves exhibit tremendous variation and suggest that future reconstructions of hominin dietary ecology take a more nuanced approach when considering grass leaves as a potential hominin dietary resource.

The African savannas that many early hominins occupied likely experienced stark seasonality and contained mosaic habitats (i.e., combinations of woodlands, wetlands, grasslands, etc.). Most would agree that the bulk of dietary calories obtained by taxa such as Australopithecus and Paranthropus came from the consumption of vegetation growing across these landscapes. It is also likely that many early hominins were selective feeders that consumed particular plants/plant parts (e.g., leaves, fruit, storage organs) depending on the habitat and season within which they were foraging. Thus, improving our understanding of how the nutritional properties of potential hominin plant foods growing in modern African savanna ecosystems respond to season and vary by habitat will improve our ability to model early hominin dietary behavior. Here, we present nutritional analyses (crude protein and acid detergent fiber) of plants growing in eastern and southern African savanna habitats across both wet and dry seasons. We find that many assumptions about savanna vegetation are warranted. For instance, plants growing in our woodland habitats have higher average protein/fiber ratios than those growing in our wetland and grassland transects. However, we find that the effects of season and habitat are complex, an example being the higher protein levels we observe in the grasses and sedges growing in our Amboseli wetlands during the dry season. Also, we find significant differences between the vegetation growing in our eastern and southern African field sites, particularly among plants using the C 4 photosynthetic pathway. This may have important implications for the contrasting dietary interpretations that stable

Carbon isotope analysis of fossil micromammal insectivores holds promise for resolving questions about past environments because these animals have restricted home ranges and are generalist feeders, thus their diets likely integrate ecological information about local habitats.In this study, we assessed the degree to which carbon isotope compositions of 35 three sympatric shrew species record spatial and temporal changes in habitat conditions 36 in a mosaic southern African savanna environment. Sampling sites were located within 2 37 km of one another, and microhabitat conditions ranged from very open (< 5% canopy 38 cover) to wooded (~60% canopy cover). We compared shrew hair  13 C values between microhabitat types, and across taxa, in order to test whether these data follow predictable patterns based on local vegetation.Shrew carbon isotope values varied with habitat in a predictable manner within our study area. While taxonomy also influenced  13 C values, this was largely due to 43 differences in habitat preferences of individual taxa and resultant variation in their relative abundance within each environment. Isotopic differences between habitat types were preserved within individual taxa where taxa occurred in multiple habitats. To complement this modern study, we performed isotopic analysis of the enamel of insecteating fossil micromammals from the hominin sites Gladysvale and Sterkfontein in the Cradle of Humankind, South Africa. This subset of fossil micromammals consumed primarily C 4 -derived carbon. 50

No abstract