Ecological interactions between small mammals, large mammals, and vegetation in a tropical savanna of central Kenya

I investigated the indirect ecological interactions between ungulates and small mammals in an East African savanna habitat by (1) monitoring the population and community responses of small mammals to ungulate exclusion on replicated 4 ha plots and (2) quantifying the responses of vegetation to the exclusion of small mammals from replicated 3m by 7m plots. I also quantified patterns of granivory by rodents, ants, and birds. I conducted this research between August 1995 and November 1996 at the Mpala Research Centre in central Kenya. The dominant ungulates in this habitat are elephant, giraffe, Grevy's and common zebra, buffalo, eland, and Grant's gazelle. The dominant small mammal is the pouched mouse, Saccostomus mearnsi. Other small mammals include species in the genera Arvicanthis, Mus, Mastomys, Dendromus, and Crocidura.; I detected pronounced indirect effects of ungulates on small mammals. After one year, S. mearnsi population density had increased two-fold in the absence of ungulates, as a result of increases in survivorship and recruitment. Two other species, Arvicanthis sp. and Mus sp., also had higher abundances in the absence of ungulates. The small mammal community maintained relatively constant species diversity in the ungulate exclusion plots over one year.; Small mammals had a pronounced effect on the quantity and quality of vegetation available to ungulates. When I excluded small mammals for a year, the quantity of aboveground vegetation increased by 40%. Small mammals had an effect on the quality of the vegetation as well. Plots to which small mammals had access were more homogeneous in their relative abundances of two dominant grasses, Themeda triandra and Pennisetum mezianum, than were plots from which small mammals were excluded.; I investigated the ecology and behavior of S. mearnsi. It is a medium-sized, semi-fossorial murid rodent which has never before been reported in high local abundance. On three replicated one hectare grids, S. mearnsi exhibited a 2.5-fold change in abundance over 15 months. Both sexes exhibited great variation in adult body weight. Home ranges of males (0.21 {dollar}pm{dollar} 0.02 ha) were three times the size of those of females (0.06 {dollar}pm{dollar} 0.01 ha). Stomach content analyses showed that S. mearnsi is omnivorous, consuming primarily insects and forbs during the dry season, and seeds following the rains. Adults share burrow entrances in termite mounds; whether they nest individually or communally remains to be determined.; To investigate the taxonomic and spatial patterns of granivory in this habitat, I conducted experiments to determine (1) the relative rates of seed removal by rodents, ants, and birds, and (2) the spatial pattern of seed removal around termite mounds, in which rodents preferentially burrow. Birds removed a large but highly variable quantity of seeds, and rodents removed a more consistent but also considerable amount. Ants removed relatively few seeds. Seeds placed within 6m of termite mounds were 80% more likely to be removed than those placed further away. (Abstract shortened by UMI.)...