Family Colletidae

As noted in Section 21, the Colletidae are morphologically diverse bees, such that one could easily justify recognizing several families among them, as some authors have done. These bees, however, have synapomorphies (as indicated below), and it seems reasonable to retain them as a single, large, worldwide family. It is most abundant and most diversified in temperate parts of Australia and South America. In the holarctic region there are only two common genera, Colletes and Hylaeus; neotropical genera enter the southern USA, especially the Southwest. By contrast, in Australia there are many genera. The family is relatively scarce in the moist tropics, especially so in the Indo-Malayan area.

Nearly all members of the family are easily characterized by glossal features not found in other bees. These features are found in all colletid females; as noted below, some of them are not found in certain males (see Secs. 20, 21). The glossa is short, commonly broader than long, truncate, bilobed (Figs. 19-2b; 20-3a, b; 37-1) or bifid, sometimes drawn out into two long, pointed processes (Fig. 39-12). The disannulate surface is as broad as the annulate surface, the former including an apical zone (beyond the preapical fringe) that is usually expanded into a pair of large apical glossal lobes that bear the conspicuous branched or simple hairs of the glossal brush (Fig. 37-1). The annuli, or those of one area, are fine and close, the annular hairs usually minute and blunt, capitate, or spat-ulate (Fig. 37-2). The distal end of the annulate surface is usually marked by the preapical fringe (Fig. 37-1). The disannulate surface is hairy but lacks seriate hairs (Fig. 20-3b). Some of these features—the apical zone or lobes derived from the disannulate surface, the glossal brush, the fine annuli and minute, blunt or spatulate annular hairs, and the preapical fringe—are unique synapomorphies of the Colletidae. The others may be synapomorphies or may be plesiomorphies derived from sphecoid wasps (see Secs. 20, 21; also Michener and Brooks, 1984, and Mich-ener, 1992c).

The glossal features described above distinguish all female Colletidae from all other bees, and appear to show that the family is a monophyletic unit. Some of them are also found in most males. In males of Hemirhiza, Meroglossa, and Palaeorhiza (Hylaeinae), however, the glossa is pointed (Figs. 20-3c, 47-1i, j), much like that of Andrena, not only in shape but in having alternatives to all the characters listed in the preceding paragraph (Michener and Brooks, 1984). Michener (1992c) further elaborated this point and noted lesser sexual differences in glossal structure in most colletids, although not in certain species of Colletes. Many male colletids with the glossa shaped about as in their female conspecifics nonetheless lack the preapical fringe, apical zone, glossal lobes, and glossal brush, and have coarser annuli with pointed annular hairs. Males of Amphylaeus and Hylaeus (Hylae-orhiza) have an intermediate sort of glossal shape with a small to distinct median apical glossal point, but lack the other Andrena-like glossal features of males of Hemirhiza, etc. For details and relevant illustrations, see Section 47.

Figure 37-1. Diagram of glossa of a Hyleoides (Hylaeinae) female, showing typical structures of a colletid glossa. From Michener, 1992c, after McGinley, 1980. See also Figure 19-4a.

Figure 37-2. Glossa of Niltonia virgiliiMoure (Colletinae), showing details of anterior surface of glossa. a, Basal part of glossa (PS = paraglossal suspensorium); b, Basal or preannular area; c, Apical annular area, showing transverse rows of spatulate annular hairs. (Scale lines = 0.01 mm.) From Laroca, Michener, and Hofmeister, 1989.

Figure 37-2. Glossa of Niltonia virgiliiMoure (Colletinae), showing details of anterior surface of glossa. a, Basal part of glossa (PS = paraglossal suspensorium); b, Basal or preannular area; c, Apical annular area, showing transverse rows of spatulate annular hairs. (Scale lines = 0.01 mm.) From Laroca, Michener, and Hofmeister, 1989.

Less distinctive characters of the family, that is, characters shared with certain other families, can be described as follows: The labrum is usually much broader than long, and the apical margin in both sexes is fringed with bristles. Below each antenna is one subantennal suture, directed toward the inner margin of the antennal socket. [Leioproctus semicyaneus (Spinola) from Chile has two apparent subantennal sutures; in some Australian genera (Xanthesma,, Brachyhesma) the antennal sclerites reach the clypeus, effectively eliminating the subantennal sutures.] The clypeus is usually relatively flat and its lower lateral parts are usually not much bent posteriorly on either side of the labrum. The mentum varies, from absent (i.e., membranous) to sclerotized and tapering from a broad apex to a narrow base. The base of the prementum is not a detached fragmentum. The lorum is either a flat plate (Fig. 33-4e, f) or elevated distally around the base of the mentum (Fig. 33-4b-d), in that case usually forming a strong proboscidial lobe (Fig. 33-4b). The galeal comb is present (Fig. 38-18a, b), but much reduced and consisting of only three or four bristles in Scrapter, and absent in Leioproctus (Excolletes). The episternal groove is usually fully developed (Figs. 28-3a, 37-3) and extends well below the scrobal groove, but in the Diphaglossini it is absent below the scrobal groove, and it is nearly absent in Hesperocolletes (Colletinae). The middle coxa appears shorter than the distance from its summit to the base of the hind wing (Figs. 28-3a, 37-3), because the upper quarter or more is hidden (Michener, 1981b). The propodeal triangle is hairless. The scopa is absent to well developed on the hind leg (trochanter to basitarsus) and sometimes also on the metasomal sterna. The disc of S7 of the male is usually reduced to almost nothing but supports long basolateral apodemes and one to three pairs of long, sometimes elaborate, haired apical lobes (Fig. 132); this feature is lost in a few groups of Hylaeus and Leio-proctus, but is found in Stenotritus (Stenotritidae) and is approached in other families as well. S8 usually has a strong apical process. The volsella is present and free, with recognizable digitus and cuspis. The spatha is absent.

Larvae were described and illustrated in detail by McGinley (1981). Since no combination of characters distinguishes all colletid larvae from those of Andrenidae, Melittidae, etc., an enumeration of the larval characters here seems unwarranted (see McGinley, 1987). Characters of colletid pupae, with emphasis on differences among subfamilies, were tabulated by Torchio and Bur-well (1987).

All colletid species are solitary, although some nest in aggregations. Species in the hairy subfamilies (Colletinae, Diphaglossinae) usually excavate burrows in the soil, although Callomelitta nests in rotting wood and certain Colletes nest in pithy stems. Nests ofthe species in sparsely haired subfamilies may be excavated in the soil (Eury-glossa), but many are in preexisting holes in wood, stems, soil, volcanic rock, etc. Most species that excavate nests in the soil make cells, commonly only one or a few per lateral burrow, that are homomorphic. If more or less horizontal, as is common in Colletinae, the cells are usually bilaterally symmetrical (flatter on the lower side than elsewhere). If the cells are consistently vertical as in Dipha-glossinae, then they are round in cross section and sym-

Figure 37-3. Lateral view of thorax of Colletes fugidus Swenk, showing the long episternal groove and the small exposed part of the middle coxa. From Michener, 1944.

metrical if one ignores the region toward the cell entrance, which curves to one side. Colletes, however, often makes heteromorphous series of cells in subterranean burrows. Species of the sparsely haired subfamilies mostly make heteromorphous cells, often in series, although Euryglossa and allies make homomorphous cells.

A colletid synapomorphy is the cellophane-like cell lining (Pl. 15), a lining different from that of other bees. In spite of physical similarity, the cell linings were reported to be chemically different in Hylaeus and Colletes, in the former being silklike proteins presumably secreted by the salivary glands, in the latter, polyesters (specifically laminesters) from Dufour's gland (Hefetz, Fales, and Ba-tra, 1979; Batra, 1980). The difference is probably not so great as this statement suggests (Espelie, Cane, and Himmelsbach, 1992); in both genera the materials are spread by the glossa before polymerization, and it is likely that in both cases salivary gland components and Dufour's gland secretions are mixed.

As noted by Torchio (1984), the provisions in colletid cells, except for those of Leioproctus and Lonchopria in the Paracolletini, are, so far as is known, liquid (but see Scrapter). The egg floats on the surface of the liquid, or in the case of Colletes, hangs from the top of the cell. The young larva curls on its side on the surface of the liquid. Torchio associated this with a 90° rotation of the late embryo, contrasted with 180° rotation in other Hy-menoptera. The 90° rotation may be an apomorphy of the Colletidae. Unfortunately, the embryology is unknown for Leioproctus and Lonchopria, which have firm subspherical pollen masses and whose larvae may well lie on their ventral surfaces.

The Colletidae are divided into five subfamilies. At least superficially, these subfamilies are very different from one another. The Diphaglossinae (mostly neotropical) are medium-sized to very large, euceriform, densely hairy, fast-flying bees. The cosmopolitan, andreniform Colletinae are also moderately hairy. These subfamilies have large scopae on the hind legs (coxa to basitarsus) of the females and sometimes on the metasomal sterna as well. The remaining subfamilies contain mostly small, sometimes minute, usually slender, hylaeiform bees (Pl. 1), although many Euryglossinae are more andreniform and some are hoplitiform. The pubescence consists of short, generally sparse hairs. Species of the neotropical Xeromelissinae have small scopae on the hind legs and the first three metasomal sterna, but the cosmopolitan Hy-

laeinae and Australian Euryglossinae have no scopa and carry pollen internally, in the crop, instead of externally, on the scopa.

The phylogeny of 20 adult exemplars representing all colletid subfamilies was investigated by Alexander and Michener (1995) as part of a study of S-T bees. The Col-letinae consistently appeared as paraphyletic, although the groups derived from within this subfamily varied. An analysis including many more colletid taxa is needed to settle on the most likely relationships. One group that was often derived from within the Colletinae in the analyses was the Diphaglossinae. I do not believe, however, that this is the correct position for Diphaglossinae, for the larvae of that subfamily spin cocoons. This behavior and the projecting labial region of the larva on which the salivary (silk) opening lies are ancestral features shared with sphe-coid wasps. No other colletid spins cocoons or has such a projecting labium. It is unlikely that the Diphaglossinae arose from a group that does not spin cocoons, like the Colletinae, and re-evolved cocoon spinning and the structures necessary to do so. One of Alexander and Michener's consensus trees (their fig. 13) shows the Diphaglossinae as the sister group of all other colletids, a position supported by these larval and spinning characters. Nonetheless, the phylogenetic position of the Diphaglossinae is not firmly established. The restriction of the Diphaglossinae to the Western Hemisphere is not what one would expect ofthe basal colletid clade. The disjunct panaustral distribution of the Colletinae (ignoring the widespread Colletes) is much more suggestive of an ancient type.

Alexander and Michener's (1995) analyses usually showed the Hylaeinae and Xeromelissinae as sister groups, closely associated sometimes with the Eury-glossinae and sometimes with the tribe Scraptrini of the Colletinae. In some analyses the Euryglossinae were part of this same clade. I tend to accept this relationship as likely, although in other analyses the Euryglossinae fell at the base of the Colletidae or the base of all bees. The analyses seem to establish the relation of Xeromelissinae to Hylaeinae, although the former has a small scopa and the latter lacks a scopa. They do not establish the position of Euryglossinae, but as noted in Section 21, the large, crescentic galeal comb on a curved sclerite in both Eury-glossinae and Hylaeinae seems to be a unique synapo-morphy indicating common ancestry.

The idea that the Euryglossinae might be the sister group to all other bees is supported by such observations as their restriction to Australia, the lack of a scopa, the unusually short proboscis, and, as has been emphasized by John Plant (manuscript, 1991), the lack of a galeal velum in most genera. Except for the association with Australia, these items are wasplike and therefore can be interpreted to support a basal position in bee phylogeny. I interpret these matters differently, however. The scopal loss I consider a probable synapomorphy in common with Hy-laeinae, as is the crescentic sclerite bearing the fused bases of the bristles of the galeal comb, among other characters. In the euryglossine genus Pachyprosopis the galeal velum is present, just as it is in the Hylaeinae and most other bees. Its loss in several other genera could be a derived condition. The reverse hypothesis, that the galeal velum evolved in Pachyprosopis and in other bees, almost certainly would require it to arise twice, for Pachyprosopis is clearly a euryglossine, very different from other bees. The euryglossine galeal blade is not at all similar to that of sphecoid wasps, in spite of the lack of the velum in most genera of both taxa. Features of wasp galeal blades that differ from those of bees, including euryglossines, are the sclerotic plates on the inner galeal surface and the comb, which is not homologous to that of bees. The short proboscis of the Euryglossinae may be a special Australian development. Most native nectar sources in Australia are in the Myrtaceae, whose flowers are wide open like cups of nectar (Michener, 1965b). Long proboscides are thus not needed.

A recent catalog of neotropical species of Colletidae, published in five parts, will greatly facilitate studies, especially on Colletes and Hylaeus in South America. The parts are: 1, Paracolletini, Moure, Graf, and Urban, 1999; 2, Diphaglossinae, Urban and Moure, 2001; 3, Colletini, Moure and Urban, 2002a; 4, Hylaeinae, Urban and Moure, 2002; 5, Xeromelissinae, Moure and Urban, 2002b.

The subfamilies can be distinguished by reference to the following key.

Key to the Subfamilies of the Colletidae

1. Body usually hairy, female with well-formed scopa enclosing corbicula on underside of hind femur; prepy-gidial fimbria (apical hair band of T5) of female much stronger (hairs longer and denser) than hair bands (ifany) of preceding terga, and T6 with abundant hair (pygidial fimbria) lateral to pygidial plate (except in genera that lack both fimbriae and pygidial plate); pygidial plate of female, ifpresent, usually broad and tapering posteriorly; forewing with three submarginal cells or, if two, then second at least two-thirds as long as first, as though second submarginal crossvein is lost; distal submarginal crossvein sinuate, at acute angle to distal part of radial sector forming free part of marginal cell (Fig. 38-17) 2

—. Body with hairs short and relatively sparse, female lacking scopa or with a sparse or short scopa forming corbicula on underside of hind femur; prepygidial and pygidial fimbriae of female nearly always absent; pygidial plate of female absent or, ifpresent, then usually narrow and parallel-sided posteriorly, or spinelike; forewing with two submarginal cells, second usually much shorter than first, as though first submarginal crossvein is lost; second submarginal crossvein usually not sinuate, usually at right or obtuse angle to distal part of radial sector (Fig. 47-2) 3

2(1). Stigma small, shorter than prestigma, as wide as pres-tigma measured to costal wing margin (Figs. 43-1, 44-1, 45-1); glossa deeply bifid with apical lobes commonly directed strongly apicolaterally (Western Hemisphere)

Diphaglossinae (Sec. 42)

—. Stigma usually large, at least longer than prestigma, usually wider than prestigma measured to wing margin (Figs. 39-2a, 39-5, 39-11, 41-1); glossa weakly bilobed to deeply bifid but, ifthe latter, then lobes commonly directed more apicad than laterally (except apicolaterally in Leioproctus tomentosus Houston from Western Australia)

Colletinae (Sec. 38)

3(1). Facial fovea absent or broad, at least one-third as wide as long; female with sparse scopa on S1 to S3 and outlining ventral corbicula on hind femur (Fig. 43-1); longitudinal part of hypostomal carina usually longer than clypeus but, if not, then clypeus protuberant and its lower lateral extremities bent back around ends of labrum

(neotropical) Xeromelissinae (Sec. 46)

—. Facial fovea usually a narrow groove, sometimes a broader area, wider than diameter of scape, absent in a few females and some males (Figs. 47-4, 47-6, 47-7); scopa absent; longitudinal part of hypostomal carina usually not longer than clypeus; clypeus usually not protuberant, not much bent back around ends of labrum 4 4(3). Supraclypeal area elevated abruptly above level of antennal socket (Fig 47-3a); pygidial and basitibial plates usually absent but, fpresent (as in a few Australian and New Guinea species), then pygidial plate of female broad, its margins converging posteriorly; anterior surface of T1 usually without longitudinal median groove; posterior (lower) surface of prementum with longitudinal, usually spiculate depression or fovea (Fig. 41-3a) (weak in a few males) margined by ridges that diverge on basal half of prementum and meet near base of subligu-

lar process Hylaeinae (Sec. 47)

—. Supraclypeal area sloping up from level of antennal socket; apical part of pygidial plate of female slender, sometimes a spine, its margins parallel or converging slightly toward apex or spatulate; basitibial plate usually indicated in female, sometimes only by one or more tubercles; anterior surface of T1 with longitudinal median groove; posterior (lower) surface of prementum lacking longitudinal medial fovea but with comparable spiculate area (Australia) Euryglossinae (Sec. 48)

Was this article helpful?

0 0

Post a comment