Corbiculate Studies
(modified from Ascher et al. 2001)
The corbiculate bees are one of the most thoroughly studied of all bees because they include several economically important species (including honey bees [Apis mellifera]) and because they form some of the most complex eusocial societies known among bees. Much work has been done on this group based both on morphology as well as molecular data. Corbiculate bees are a monophyletic group of long-tongued (LT) bees within the family Apidae and the subfamily Apinae (Roig-Alsina & Michener 1993) and comprise four extant tribes, each of which is monophyletic (Michener 2000 and references therein). These are the Euglossini (orchid bees, 175+ known species, five genera [Cameron 2004]), Bombini (bumble bees, 239 species, all in the genus Bombus [Williams 1998]), Apini (honey bees, at least seven extant species, all in the genus Apis [Engel 1999]), and Meliponini (stingless bees, several hundred species, numerous genera [Michener 1990]). The corbiculate bee tribes are thought to be of Mesozoic age because a fossil meliponine has been described from the late Cretaceous (~80 mya; Michener 2000 and references therein; Grimaldi 1999, Engel 2000b). Corbiculate bees are of particular interest because they include the only advanced eusocial bees, the only bees to store harvestable honey, and the most important pollinators of field and greenhouse crops.
The phylogenetic relationships among corbiculate bee tribes remain extraordinarily controversial. Nine of the 15 theoretically possible rooted trees for these four taxa have been published as potential phylogenies since 1984 (eight are diagrammed by Schultz et al. 1999, Fig. 1; the 9th is Fig. 2c of Koulianos et al. 1999; see also Prentice 1991 and our figure below). However, all recent morphological analyses of extant corbiculates (Prentice 1991; Roig-Alsina and Michener 1993) and extant fossil corbiculates (Engel 2000a, 2001), cladistic analysis of behavior (Noll 2001), the sting aparatus (Cardinal in prep.), morphology of the proventriculus (Serrao 2001), and simultaneous analyses of multiple data sets (Chavarria & Carpenter 1994; Schultz et al. 1999) support the phylogeny proposed by Michener (1944): (Euglossini + (Bombini + (Apini + Meliponini))) (see Fig. a below). This phylogeny is consistent with a single origin of obligate eusociality among corbiculates and advanced eusociality among bees (Michener 1974; Prentice 1991, Fig. 3.6a).
However, phylogenetic studies based on DNA sequence data have supported a different topology (see Fig. b below; Cameron & Mardulyn 2001 and previous papers cited therein). Molecular studies have supported (with reasonably high bootstrap support) the sister group relationship between Bombini and Meliponini with variable placement of Apini and Euglossini . However, these studies have suffered from problems that have called into question the quality of the DNA sequence data (Ascher 2001 et al.). First, the placement of Apini and Euglossini is far from stable (these two groups are sister taxa in the Cameron & Mardulyn 2001 combined analysis of molecular data). Second, the molecular data rarely recovers monophyly of the corbiculate tribes. Monophyly of these tribes is well supported by morphological data and seems incontrovertible. Finally, the branch leading to Apini is often seemingly a long one and there is significant lineage specific base compositional biases (at least in one gene, opsin [see Ascher et al. 2001]). It is not clear which data set is providing the correct picture of corbiculate tribal relationships, although we suspect that the long branch leading to Apis and the highly unstable position of Apis within the different molecular data sets suggest that this branch of the tree may be creating problems. If there is a branch-specific increase in molecular evolution this would effect all genes similarly and might help explain why all genes analyzed so far provide a similarly incongruent tree topology when compared with the morphological data. More genes will probably need to be analyzed in order to resolve this question definitively and the individual morphological and behavioral data sets should be merged into a combined analysis with the molecular data.
We provide below the combined molecular and morphological data set published by Cameron & Mardulyn (2001).
Download
|
Excel spreadsheets: |
Genbank numbers for the numerous corbiculate bee sequences available in Genbank (not yet available) |
Data sets: |
Cameron & Mardulyn (2001) molecular and morphological data set (.txt file) |