Deserts are, by definition, environmentally similar, and this has lead to hypotheses of convergence in the properties of desert biotic communities as well as the components of these communities. There is considerable evidence for convergence in some characteristics of desert biota, ranging from plant growth forms to the well-known bipedal, nocturnal rodents. One area that has received considerable attention has been granivory by desert rodents, largely because of the effort focused on the North American desert heteromyids, and also because the process of granivory has far-reaching ramifications for desert plant communities. Specific tests for convergence in the impact of rodents as granivores, by means of bait-removal experiments, however, have shown that the high levels of seed removal by rodents in the North American deserts differs from that of rodents in the South American, Australian and South African deserts, where ants are the most important seed harvesters. The only studies to measure the impact of rodents on desert seed fluxes confirm these patterns, with rodents consuming up to 86% of seed production in North American deserts, but less than 1% of seed production in South African deserts. A review of dietary data for desert rodents confirms these trends, with little evidence for the presence of granivores in deserts besides those of North America. A variety of hypotheses have attempted to explain these variations in desert rodent granivory. These include recent extinctions of granivores, that seed burial, low soil nutrients and/or limiting seed production prevented the radiation of granivorous small mammals, and that particular deserts are too young or too recently colonised by rodents for granivorous rodents to have evolved. However, none of these hypotheses are supported by available evidence. Alternative hypotheses suggesting that climate variability may have precluded the development of specialised granivores need to be tested. In particular, more data are needed to confirm these patterns of granivory, and gain an understanding of the effects of Pleistocene and recent desert climate variability on seed production. An alternative perspective suggests that the presence of the heteromyid rodents may explain the high levels of granivory by small mammals in North American deserts. The variability in granivory by small mammals between deserts suggests that deserts will also differ in terms of anti-granivore adaptations of plants, seed fluxes and the mechanisms whereby small mammals coexist.
NORTH AMERICAN DESERT RODENTS: A TEST OF THE MID-DOMAIN EFFECT IN SPECIES RICHNESS