m m m m tion of the Andes progressively increased the possibility ofintercontinental dispersal by creating cool or arid habitats near the Equator. There is an alternative to long-distance dispersal, however. From time to time in its history, a xeric-adapted genus may give rise to a species able to persist in mesic or humid habitats. For example, Ashmead-iella, a generally xeric-adapted genus, has one species that ranges eastward in North America as far as Indiana and Georgia, and another that occurs from North Carolina to Florida. Should such a species reach another xeric area, it might well speciate there, and if it then disappeared in the intervening mesic area, one could have separate clusters of species in the two xeric areas and a problem in explaining how the genus traversed the mesic area.
In Africa the situation is different, because there is now a savanna corridor (with arid areas here and there) through eastern Africa between the palearctic (Mediterranean) region and the temperate xeric Cape region. Various genera such as Andrena and Nomada are present in this corridor and in southern Africa, to which they appear to have dispersed from the north. Bees with amphitropi-cal disjunct distributions in the Old World exist, however, although much less commonly than in the New World. Examples are Fidelia (in southern Africa and Moroccan deserts) and Aglaoapis (South Africa and Eurasia).
Other disjunct distributions of bees, described by Michener (1979a), include Asiatic groups that range through the islands to Australia, and Malagasy forms with relatives in Africa or in a few cases in Asia. Others are the Paracolletini, which are found principally in temperate parts of Australia and South America, and the Fideliini in desertic parts of Chile, southern Africa, and Morocco, with a relative, the Pararhophitini, in palearctic desert areas in Central and western Asia. Dispersal between Australia and South America through cool-temperate Antarctica might have been possible across moderate water gaps as recently as the beginning of the Oligocene (38 myBP), although land connection was broken in the late Cretaceous, about 80 myBP (Smith, Smith, and Funnell, 1994); see Section 23. This route could have been tra versed by the paracolletines. Connections to Africa were disrupted earlier. By the end of the Cretaceous, while relatively narrow seas separated tropical Africa and South America, the temperate parts of these continents were already well separated. Thus, the Fideliini and perhaps the Paracolletini may have originated as long ago as the Late Cretaceous.
Dispersal among other xeric areas also presents interesting problems. The primarily Australian-North American distribution of the Hackeriapis-Chelostomoides group of Megachile has been discussed elsewhere. (Hack-eriapis is Australian, north to New Guinea savannas; Chelostomoid.es is North American, especially Madrean, south to Colombia.) Long-distance dispersal over water seems extremely unlikely between Australia and North America. The Bering Straits would seem a likely route, but if it were the route used, why is the nearest relative of Chemostomoides in Australia instead of in Asia? Possibly the related Asiatic subgenus Chelostomoda has something to do with this problem. Alternatively, dispersal between Australia and South America across moderate water gaps and through cool-temperate Antarctica might have been possible in the Paleogene, but if this route were used, what happened to the group in most of South America, some of which is climatically similar to the desertic areas of North America and Australia where Chelostomoides and Hackeriapis, respectively, are abundant and diversified?
As noted elsewhere, Hesperapis (southern Africa and western North America) also occurs in deserts of both the Northern and Southern hemispheres; its closest relative is Eremaphanta from deserts of Central Asia. The unusual feature of the Hesperapis group's distribution in the Old World is its absence between Central Asia and South Africa. It probably became extinct between these areas; there is no need to postulate long-distance dispersal. The problem is, how did it get to the New World?
Anthophora of the subgenus Heliophila ofNorth America and those of the Mediterranean-Central Asian area are presumably similar because of common ancestry. Detailed studies have not discovered group differences between them (Brooks, 1988). Some bees may have moved directly between the Old World and New World dry areas. Raven and Axelrod (1974) suggest that eastern North America and Western Europe were in latitudes suitable to warm, seasonally dry climates in the late Cretaceous and early Eocene, and were separated by only moderately broad seas. If so, bees of dry areas could have been exchanged more readily than under present conditions, but probably only by long-distance dispersal. Hesperapis might have moved between the continents at the same time as Heliophila.
The disjunctions noted previously for the Meliponini are unique among bees in that they involve a group that is characteristic of the moist tropics, and with minimal ability to cross water. Yet they occur in all tropical areas of the world except the oceanic islands. The genus Trigona occurs in the neotropical region and from southern Asia to Australia, but is absent from Africa. Speculations on how they might have attained their present distribution can be found in Michener (1979a, 1990a) and Kerr and Maule (1964); all that one can say with certainty is that Meliponini is an old group, probably older than most of the taxa discussed above that do not show transoceanic distributions.
The effectiveness of a barrier, as shown by a lack of disjunct distributions across it, can also tell us interesting things about the taxa concerned. An outstanding biogeo-graphical observation to be made about bees is that, if one ignores cosmopolitan taxa and the Meliponini and Fi-deliini, the bees of Africa and South America are very different. As shown in Table 26-1, both continents have numerous taxa of bees not found in the other. Therefore, it is reasonable to believe that largely tropical taxa like the Augochlorini, Centridini, Diphaglossinae, Ericrocidini, Euglossini, Exomalopsini, and Oxaeinae, widespread in South America but absent from the Old World, including Africa, originated after the separation of Africa and South America. Likewise, taxa like the Nomiini, Nomioid-ini, Ctenoplectrini, and Allodapini, widespread in the Old World, including Africa, but absent from South America, originated or became widespread in Africa after the separation of these two continents. In the south, that separation occurred in the early Cretaceous (about 140 myBP), but the continents were joined in the equatorial region for a long time, i.e., until the middle Cretaceous, about 100 myBP (Smith, Smith, and Funnell, 1994), and must have been close together in that region long after that. Presumably, the tropical taxa listed above originated after the continental separation became too wide and perhaps after the Cretaceous.
Of course, different taxa frequently tell different stories about past continental connections. For example, the Paracolletini indicate past connections or proximities of the southern continents, whereas many Hylaeinae and the Euryglossinae are unique in Australia, with numerous genera being found nowhere else. Although the Eury-glossinae occur only in Australia, the Hylaeinae are widespread; African and South American Hylaeinae are similar to those of the rest of the world but are not similar to the rich Australian fauna. Presumably, this means that the Paracolletini are an older group, and the Hylaeinae and Euryglossinae are enough younger that connections between Australia and other southern continents were fully broken before these bees became common. This idea is consistent with the nonbasal position of Hylaeinae and sometimes Euryglossinae in the phylogenetic analyses of Alexander and Michener (1995); see also Section 20.
Was this article helpful?