Evolution and biogeography of native Hawaiian Hylaeus bees (Hymenoptera: Colletidae)

The only bees native to the Hawaiian Islands form a single clade of 60 species in the genusHylaeus. The group is understudied and relatively poorly known. A data set consisting of 1201 base pairs of the mitochondrial genes cytochrome oxidase I and II and tRNALeucine, and 14 morphological characters was used to construct a phylogenetic tree for 48 of the 60 known species. Genetic variation was high, including amino acid changes, and a number of species showed evidence of heteroplasmy. Tree support was low due to high levels of homoplasy. Biogeographical analysis using DIVA indicates that early radiation took place on the island of Hawaii. This places an upper age limit of only 0.4–0.7 Myr for the group, an unusually short time for such a large radiation. Moreover, it is an unusual biogeographical pattern among the Hawaiian biota. The Willi Hennig Society 2006. The fauna and flora of the Hawaiian Islands are characterized by a small number of introductions followed by extensive speciation within the archipelago (Zimmerman, 1948). The result is a native biota that is depauperate at the order, family, and genus level, but with high species diversity, high levels of endemism, and many extraordinarily large radiations. TheHymenoptera are a prime example of such imbalance. The suborder Symphyta, with 14 families and 8000 species worldwide (Goulet and Huber, 1993), is absent from the native fauna. The Aculeata clade (29 families, about 50 000 described species; Goulet and Huber, 1993) is represented by about 400 species in only four families derived from seven introductions: two Bethylidae, two Crabronidae, two Vespidae, and one Colletidae (Nishida, 2002). Three of these—Sierola (Bethylidae, 184 species plus many more undescribed), the ‘‘Nesodynerus’’ group of Odynerus (Vespidae, 112 species), and Hylaeus (Colletidae, 60 species)—account for 90% of the species. The Hylaeus are the only bees native to the Hawaiian Islands. The Hawaiian Hylaeus belong to the subgenus Nesoprosopis, which is otherwise primarily found in Japan. One species,H. pectoralis, extends across to Europe, and undescribed species have been collected from China (Hirashima, 1977; Ikudome, 1989). In Hawaii, the group has evolved from a single introduction to at least 60 species (Daly and Magnacca, 2003), more than the total number of Hylaeus in America north of Mexico (55 species; Snelling, 1966). The Hawaiian radiation makes the otherwise-minor Nesoprosopis the largest subgenus of Hylaeus aside from Hylaeus sensu strictu (Michener, 2000; the Australian Prosopisteron also has more species but is polyphyletic, T. Houston, pers. comm.). As with many endemic insect groups, the biology of Hylaeus in Hawaii is largely unknown. A taxonomic revision was recently completed (Daly and Magnacca, 2003), but little is known of the bees’ evolution (phylogenetic relationships, biogeographical history), habitat requirements (nesting behavior, pollen usage), or conservation biology (competitive pressure from exotic bees, usage of introduced plants). Most of what is known is based on scattered observations of a few specimens of a single species (Williams, 1927; Swezey, 1954; Daly and Coville, 1982). The Hawaiian radiation includes the only cleptoparasitic colletids *Corresponding author. E-mail address: [email protected] Present address: University of California–Berkeley, Department of Environmental Science, Policy, and Management, Berkeley CA 94720 The Willi Hennig Society 2006 Cladistics www.blackwell-synergy.com Cladistics 22 (2006) 393–411 (Michener, 2000) and a large number of ground-nesting species, a relatively uncommon habit for Hylaeus (Daly and Magnacca, 2003), so there is much fertile ground for future research. Hawaii has recently become a focus of biogeographical work (Wagner and Funk, 1995). The Hawaiian insect fauna contains a number of spectacular radiations that are well-suited to such study. While the Drosophila are best known, with probably over 1000 species, there are several genera with over 100 representatives in the islands, and many with over 50 (Liebherr, 2001). Some, such as Lispocephala (Muscidae) and Sierola (Bethylidae), are widespread in continental areas but the number of species in Hawaii far exceeds those in the rest of the world combined (Evans, 1978; Hardy, 1981). In many others, such as Drosophilidae, Nitidulidae and Hylaeinae, the Hawaiian species form a significant proportion of the world fauna. As a linear hot-spot volcanic chain, the geological history of Hawaii is well-known: as new volcanoes form to the south-east, the islands to the north-west erode, subside, and eventually disappear (Carson and Clague, 1995). Because the ages of the islands are known from K-Ar dating of volcanic rock, biogeographical conclusions can be used to infer the approximate arrival time of a group’s progenitor. Phylogenetic studies of groups that arrived at or before the emergence of the oldest current high island, Kauai, usually find species from older islands to be basal, and those on younger islands apical (Wagner and Funk, 1995). This reflects a pattern of dispersal on to new islands as they arise, followed by within-island radiation that is dependent on time, habitat heterogeneity, and dispersal ability. More recent arrivals tend to show a more random pattern, reflecting the availability of multiple islands as targets for dispersal (Lowrey, 1995). The basic biogeographical questions in an isolated island-chain setting are: (1) when did the progenitor of a group arrive; (2) where did it first become established; and (3) how did diversification progress geographically? Once a phylogeny is established for the species, the last two questions can be answered using biogeographical methods such as component analysis (Page, 1994) and DIVA (Ronquist, 2001). The question of time of arrival is the most difficult to answer, but it is the most interesting in terms of its implications for the evolution of the group. For example, clades of five species that are 3 million years (Myr) old are not unusual; whereas 100species radiations that are only 1 Myr old are rare. The latter case indicates that strong evolutionary pressures are or were being exerted to cause rapid diversification. In an attempt to answer these questions for the Hawaiian Hylaeus, we developed a phylogeny based on DNA sequences and morphology. The relatively recent origin of the group, and the generally low level of morphological diversity in Hylaeus in general, make genetic methods more productive for phylogenetics in the group. The mitochondrial genes cytochrome oxidase I and II were chosen based on their rate of evolution: they are generally considered to be the most conservative of the mitochondrial genes, but still change more rapidly than most nuclear genes (Simon et al., 1994). This relatively fast rate of evolution has made these genes useful in estimating species or species-group level phylogenies of other Hawaiian groups (Wagner and Funk, 1995; C. Ewing, unpubl. data) and other bees (Danforth, 1999). Several nuclear genes (EF-1a, wingless, arginine kinase and opsin), sequenced for a subset of species, had too little variation to be informative (unpubl. data). Materials and methods