Unlike their plant hosts that are most diverse in the tropics, bees are thought to reach peak diversity in deserts. Among the hypotheses forwarded to explain this counterintuitive species gradient are that deserts are linked to host specificity in bees, and biseasonal rainfall elevates local bee diversity. I discuss recent findings that suggest host specificity predisposes bees to survive highly unpredictable conditions that typify deserts. Host specialization among bees may therefore be considered a form of habitat specialization rather than due to nutrient/ chemical features of their hosts. An additional feature of bee species in deserts that also contributes to diversity may be that they are predominately short-lived species that are active in only one bloom period per year. If true, biseasonal precipitation patterns may contribute to local bee diversity simply because there are two non-overlapping faunas active at different times of the year. This hypothesis is evaluated by comparing long-term studies of bee faunas in xeric and mesic sites in North America.