Bee Biogeography and Distribution
Bees are thought to have arisen sometime during the early to mid-Cretaceous (Michener 2000, Engel 2001, Grimaldi 1999) and are almost certainly a group that arose in the southern continent of Gondwana (Michener 1979, Raven & Axelrod 1974). Bees today are most diverse in arid and semi-arid regions of the world (such as the Mediterranean basin, Central Asia, the southwestern deserts of North America, arid regions of Chile and Argentina, temperate parts of Australia, and southern Africa; Michener 1979; Linsley 1958). Their origins are presumed to be in the arid interior of west Gondwana (Michener 1979) where angiosperms are also thought to have arisen (Raven & Axelrod 1974). As such, early bee diversification may have taken place during the breakup of Gondwana in the Cretaceous. To the extent that historical biogeographic patterns in bees have not been obscured by dispersal, one might expect to see bee higher level phylogeny reflecting Gondwanan fragmentation (Smith et al. 1994). Given the uncertainties associated with resolving family and subfamily level relationships in bees (see above) it is understandable that no previous studies have focused on examining the historical biogeography of the bees at these levels. While Michener (1979) provided a comprehensive and authoritative review of bee biogeography and biogeographic distributions, at the time of his review the higher level phylogeny of the bees was not well undestood, and therefore biogeographic patterns could not be interpreted in a phylogenetic context.
There are, however, intriguing patterns. Each of the major short-tongued bee families appears to have a historical association with the southern hemisphere: Colletidae (Australia, S. America, Africa), Stenotritidae (Australia), Andrenidae (South America), Halictidae (South America and Africa), and Melittidae (Africa). Additional biogeographic patterns which support the southern hemisphere origin of the ST bees are as follows. Within the colletid subfamily Colletinae (excluding Colletes and Hylaeus which both widespread), there are roughly equal numbers of genera and species restricted to Australia on the one hand and S. America on the other. For Australia there are 7 endemic genera and for S. America there are 6. One genus, Leioproctus, occurs on both continents (with a total of over 324 species). In addition, there is a single African genus, Scrapter. The S. American/Australian connection in the Colletinae can most likely be explained by interchange between S. America and Australia via Antarctica, which provided a cool, temperate connection between these continents up until the end of the Eocene (approximately 34 Ma). Likewise, the sister group relationship between the primarily Australian Hylaeinae and the exclusively S. American Xeromelissinae implied by morphology (e.g., the spiculate depression on the prementum; Alexander & Michener 1995) and our preliminary molecular results might also be explained as an ancient connection via Antarctica (Michener 1979).
While intriguing and generally supporting a southern hemisphere origin for the ST bees, these patterns are not rigorously interpretable in the absense of well supported phylogenies at the appropriate levels. Resolving family, subfamily and tribal level relationships in bees should help us better interpret the current distribution of the bees in light of Gondwanan fragmentation.