Many large mammal species are declining in African savannas, yet we understand relatively little about how these declines influence other species. Previous studies have shown that the removal of large herbivorous mammals from large-scale, replicated experimental plots results in a dramatic increase in the density of small mammals, an increase that has been attributed to the relaxation of competition between rodents and large herbivores for food resources. To assess whether the removal of large herbivores also influenced a predator of small mammals, we measured the abundance of the locally common olive hissing snake, Psammophis mossambicus, over a 19-mo period in plots with and without large herbivores. Psammophis mossambicus was significantly more abundant in plots where large herbivores were removed and rodent numbers were high. Based on results from raptor surveys and measurements of vegetative cover, differences in snake density do not appear to be driven by differences in rates of predation on snakes. Instead, snakes appear to be responding numerically to greater abundances of small-mammal prey in areas from which large herbivores have been excluded. This is the first empirical demonstration of the indirect effects of large herbivores on snake abundance and presents an interesting example of how the influence of dominant and keystone species can move through a food web.