No vitates AMERICAN MUSEUM PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY Number 3272, 14 pp., 7 figures, 3 tables 10024 August 10, 1999 The First Fossil Euglossa and Phylogeny of the Orchid Bees (Hymenoptera: Apidae; Euglossin1) MICHAEL S. ENGEL! ABSTRACT The first fossil Euglossa (tribe Euglossini) is described and figured from Miocene Dominican amber as Euglossa moronei new species. The species is differentiated from extant euglossines, and its apparent lack of close affinity to any extant subgenus or species group is discussed. The existing data pertinent to the internal phylogeny of Euglossini are reconsidered, and a new hypothesis of relationships among the genera is proposed. The distribution of the eu- glossines, particularly in the West Indies, and their associations with orchids is briefly consid- ered. INTRODUCTION All the past these mountains saw All the years of toil and strife Lives unknown that went before They gave us this: Our present life. —Theodore D. A. Cockerell (1927) The orchid bees are among the most spec- tacular of the Apoidea owing to their fre- quently robust body size and brilliant metal- lic coloration or banding on the metasoma. The group is sometimes known as the golden bees; however, this is a misnomer since few species exhibit golden coloration and most ' Research Scientist, Department of Entomology, American Museum of Natural History. Copyright © American Museum of Natural History 1999 are brilliant metallic green or blue. Far more notable than their coloration is the extreme modification of the mouthparts. The probos- cis of the orchid bees is elongated such that in many species the end of the glossa extends beyond the metasoma when in repose. Or- chid bees are the most important pollinators of tropical orchids in the Western Hemi- sphere (Dodson, 1967, 1975; Dodson and Frymire, 1961; Dressler, 1982b). Males are the most common visitors of orchid inflores- cences while females are rarely found there. Males collect various fragrant compounds from the flowers using modified tufts of setae on their tarsi; these fragrances are in turn transferred to distinctive foveae on the hind tibiae where they are absorbed. The exact ISSN 0003-0082 / Price $1.80 2 AMERICAN MUSEUM NOVITATES TABLE 1 Hierarchical Classification of Tribe Euglossini Tribe Euglossini Latreille Genus Aglae* Lepeletier de Saint Fargeau & Serville Genus Eufriesea Cockerell Genus Euglossa Latreille’ subgenus Dasystilbe Dressler subgenus Euglossa Latreille subgenus Euglossella Moure subgenus Glossura Cockerell subgenus Glossurella Dressler subgenus Glossuropoda Moure Genus Eulaema Lepeletier de Saint Fargeau Genus Exaerete* Hoffmannsegg # Genus is parasitic. » The subgenera of Euglossa are at present poorly de- fined and much work remains to be done on the char- acterization of natural groups within this genus. function of these fragrances is not entirely understood, but they possibly play a role as pheromones (Vogel, 1966). Females are not attracted to the raw substances on the flowers while males are, and thus the compounds, as they are constituted by the flowers, must be precursors of the final chemical presumably used by males for attracting females (Mich- ener, 1974; Dressler, 1982b). Differing chem- icals and blends attract different bee species (Dodson et al., 1969; Dodson, 1970), and they are believed to play an important role in reinforcing reproductive isolation among floral species (Michener, 1974; Dodson, 1975; Ackerman, 1983). The orchid bees are segregated into five genera and together form the tribe Euglossini (see table 1). Of the 150+ known species, most euglossines are classified in the nomi- nate genus, Euglossa Latreille (1802), six in the parasitic genera Aglae Lepeletier de Saint Fargeau and Serville (1825) and Exaerete Hoffmannsegg (1817), and the remainder in the genera Eulaema Lepeletier de Saint Far- geau (1841) and Eufriesea* Cockerell (1908). The genus Euglossa has been split into sev- eral subgenera and species groups (see table 1; also Dressler, 1978, 1982a; Moure, 1989); however, these are poorly defined and often ? The fossil recently proposed by Poinar (1998a) as Paleoeuglossa, although poorly described, is best placed as an enigmatic species of Eufriesea (new synonymy) and thus results in the new combination Eufriesea mel- issiflora (Poinar). NO. 3272 only are confidently recognized in the males. The euglossines are the most basal tribe of the corbiculate Apinae (Roig-Alsina and Miche- ner, 1993; Chavarria and Carpenter, 1994; Schultz et al., 1999). The corbiculate apines are the most recognized group of bees throughout the world and were, until recently, treated as a single family (Apidae sensu Michener, 1990). The group is most readily identified by the presence of a broad, slightly concave area bordered by long hairs on the hind tibia of females (except in queens and parasitic species). This distinctive morpholog- ical structure is known as the corbicula; hence the vernacular name for this clade. This monophyletic complex of four tribes contains the familiar honey bees (tribe Apini), the bumble bees (tribe Bombini), and the sting- less bees (tribe Meliponini) in addition to the orchid bees. Two of these groups, the honey bees and stingless bees, have played signifi- cant roles in human culture, agriculture, and religion for centuries (particularly the Apini) and, owing to their conspicuous, large, euso- cial colonies, have been the focus of most bi- ological work in bees. Both apiculture and meliponiculture are profitable enterprises in today’s economy. Similarly, the eusocial life- style of the bumble bees, although not as elab- orate as that of the Apini and Meliponini, along with their large, noticeable body size has historically brought them the attention of many biologists. Like their cousins, the eco- nomic value of bombines in agriculture (bom- biculture?) is already becoming quite signifi- cant. Unlike other corbiculate tribes that are eusocial to varying degrees, the euglossines lack eusocial societies altogether and are the least known of the tribes. Most euglossines are completely solitary except for a few spe- cies of Euglossa and Eulaema, which are communal (Dodson, 1966; Zucchi et al., 1969; Gardéfalo, 1985, 1991; Gardfalo et al., 1998; Roberts and Dodson, 1967). Euglossi- nes, Euglossa in particular, frequently collect plant resins for use in the construction of their nests. The resins are used to line the walls of cells that are later provisioned with pollen from a variety of angiosperm families (Dod- son, 1966; Roberts and Dodson, 1967; Saka- gami et al., 1967; Zucchi et al., 1969; Gar6- falo, 1991). Flowers of the families visited by orchid bees for pollen possess poricidal an- 1999 thers and euglossines therefore use “buzz” mechanisms to free the pollen (Michener, 1962), although some other floral types are utilized (Dressler, 1982b). The tribe is restricted to the Western Hemisphere, presumably having arisen sometime after the separation of the African and South American continents. The genera are further restricted to the more tropical re- gions of the New World, reaching only as far north as the southernmost boundaries of Tex- as and as far south as northern Argentina. Of the recognized genera, only Euglossa occurs in the West Indies today but once again in a restricted distribution. Two species occur on the island of Jamaica, these being the endem- ic Euglossa (Euglossa) jamaicensis Moure (1968) and £. (Glossura) ignita Smith (1874), which is more widely distributed on the mainland from Panama to Peru and northern Brazil. Trinidad and Tobago, which are considered to be largely continental areas and faunas, are the only other Caribbean is- lands to host species of this tribe. These in- clude E. (E.) townsendi Cockerell (1904), the type species for the genus, E. (E.) cordata (Linnaeus, 1758), and the widely distributed E. (G.) ignita. On the mainland, E. townsendi ranges from southern Mexico to northern Brazil, while EZ. cordata occurs only in north- ern South America. General reviews of euglossine biology have been presented by Zucchi et al. (1969), Dressler (1982b), Kimsey (1980, 1984, 1987), and Michener (1974, 1990); the par- asitic euglossines have been reviewed by Moure (1964) and Kimsey (1979). The most recent catalog of the Euglossini was pro- duced by Kimsey and Dressler (1986)?; how- ever, several species have been described since then, and a new general catalog is needed. The internal phylogeny of the Eu- glossini has been investigated by Kimsey (1982, 1987) and Michener (1990) presented a critique and modification of her data. A modern, detailed study of the tribe, including the relationships among the subgenera of Euglossa, remains to be undertaken, and at 3Jt should be noted that the name they list as E. stil- bonata Dressler (Kimsey and Dressler, 1986: 233) is an incorrect subsequent spelling for what should have been E. stilbonota Dressler (1982a: 138). ENGEL: FIRST FOSSIL EUGLOSSA 3 present there is no strong evidence for any of the proposed topologies. An attempt is made to reevaluate the current evidence for relationships among the genera and to pre- sent a new hypothesis of tribal phylogeny. The bee fauna of the Dominican amber was, until recently, known only on the basis of the common stingless bee, Proplebeia dominicana (Wille and Chandler, 1964: as Trigona dominicana). Since then, the num- ber of species recognized from the Oligo- cene-Miocene of the West Indies has grown both in number and taxonomic diversity (re- cently summarized by Michener and Poinar, 1996; with some additions by Engel, 1997; Poinar, 1998; Camargo et al., in prep.). At present there are ten Dominican amber spe- cies reported in the literature, representing four families, four subfamilies, and seven tribes of bees, although other unstudied spe- cies exist (Engel, personal obs.; D. A. Gri- maldi, personal commun.). Like most of the Dominican amber insect fauna, these species are closely aligned to present-day taxa, in contrast to some older amber deposits where the species are often difficult to place into recent higher categories (e.g., for bees see Engel, 1998a). The age of Dominican amber is approximately 15—20 Ma, placing it within the Miocene. Grimaldi (1995) argued that Dominican amber must be younger than the Oligocene and Eocene dates that have been stated, while Iturralde-Vinent and MacPhee (1996) provided a sound stratigraphic basis for the Miocene age of these deposits. The orchid bee presented herein is not just the first representative of its genus in amber, but the first Euglossa ever recorded from the geological record. The use of plant resins by orchid bees for the construction of nests makes them suitable for entrapment and preservation in amber. For this reason it is somewhat remarkable that until recently the tribe has been completely unknown from the fossil record. As will be noted below, the bee described herein is easily recognizable as a species of the genus Euglossa but some fea- tures suggest that it represents a lineage of these bees that is no longer extant and there- fore cannot be assigned as to subgenus. Be- cause of the imperfect nature of the specimen and the present difficulties with the infrage- neric classification of Euglossa, I have cho- 4 AMERICAN MUSEUM NOVITATES sen not to erect a new subgenus for it but have provided an epithet for the species. ACKNOWLEDGMENTS I am indebted to the graciousness of Dott. Ettore Morone of Turin, Italy, for allowing me to study this magnificent inclusion from his amber collection. To him I extend my most sincere gratitude. I am also grateful to David A. Grimaldi who acted as courier for the specimens between Morone and myself, for providing the photograph depicted here as figure 1, and for many delightful discus- sions about this study and amber in general. My work is further supported by the gener- osity of Robert G. Goelet, Chairman Emeri- tus of the American Museum of Natural His- tory Board of Trustees, and by Jerome G. Rozen, Jr., both to whom I express my deep- est appreciation. The manuscript received significant improvements through critical re- views from James M. Carpenter, Gabriela Chavarria, David A. Grimaldi, Charles D. Michener, Molly G. Rightmyer, and Jerome G. Rozen, Jr. I thank each of them for their time and kind assistance. TERMINOLOGY The morphological terminology generally follows that proposed by Michener (1944) with additions for mandibular structure pro- vided by Michener and Fraser (1978). Others have used the term “‘lorum’’ for the submen- tum because of the difficulty in determining the homology between this structure and the true submentum of other Hymenoptera or in- sects (e.g., Plant and Paulus, 1987; Michener, 1990). Similarly, the small subapical projec- tion on the submental arms has been termed the “‘loral splinter’? (Plant and Paulus, op. cit.). I prefer to retain the terminology orig- inally used by Michener (1944) and use sub- mentum in place of lorum until the homol- ogy—rather, the lack thereof—has been shown conclusively. Michener (1944), how- ever, did not refer to the structure corre- sponding to the “‘loral splinter’? and so I have provided the term submental spine for this character. Roig-Alsina and Michener (1993) referred to the cleft which demarcates the submental spine as the “longitudinal fis- sure.”” Because the upper paraocular region NO. 3272 separating the lateral ocellus from the com- pound eye is important in apine systematics (e.g., Bombini in particular), I prefer to use a separate term for this area: the ocellorbital area. This small region has been given var- ious Names in apine studies such as the par- ocular area (Sakagami, 1975) and ocellocular area (Sakagami, 1972), but I prefer the term ocellorbital as it was used by Moure and Sakagami (1962). I employ the term intero- cellar furrow as defined by Engel (1998b), the term jugal comb for the series of bristles orchid bees possess in place of the membra- nous region forming the jugal lobe (Kimsey, 1982), and the term anal vein in place of vannal vein. SYSTEMATIC PALEONTOLOGY GENUS EUGLOSSA LATREILLE Euglossa moronei, new species Figures 1, 2, 4, 5 DIAGNOSIS: This species can be separated from extant West Indian species by the mod- erately protuberant clypeus, absence of me- dioclypeal ridge or carina, strongly pubes- cent labrum, apically narrowed labral ovals, entirely fuscous pubescence, absence of plu- mose hairs dorsal to antennal sockets, strong- ly ridged vertex, and interocellar furrow. Many of these same characters differentiate this species from any of the known subgen- era or species groups. For instance the com- plete absence of a clypeal ridge, the only moderately protuberant clypeus, absence of plumose hairs just dorsal to the antennal sockets, and noticeably pubescent labrum are all features not found among other Euglossa. The relatively “‘short’”’ glossa suggests spe- cies of the subgenus Euglossa, while some species of various Euglossa groups have a weakly developed interocellar furrow. Fur- thermore, the long marginal cell is a pecu- liarity of this orchid bee alone. All extant euglossines have a relatively short marginal cell that ends well before the wing apex. This might be an enigmatic species of the subge- nus Euglossa or, more likely, an infrageneric lineage that has since become extinct. DESCRIPTION: Female (male unknown): Head length 5.1 mm, width 4.1 mm. Man- dible tridentate and finely imbricate over 1999 much of its surface, becoming smooth ba- sally, broad along its entire length; imbricate integument most notable on trimmal expan- sion, exposed area of trimma at base with small punctures separated by puncture width, integument otherwise smooth; acetabular ca- rina and groove strong; outer ridges and grooves faint; outer interspace becoming gla- brous; malar space short, much less than bas- al mandibular width. Mentum and submen- tum fused; submental arms reaching nearly to apex of mentum; submental spine present; apex of mentum blunt, not emarginate or gently concave, expanded relative to shaft, about three-quarters wider at apex than along shaft, apex about as wide as base of premen- tum. Stipital comb deeply concave, occupy- ing entire distal half of stipes; maxillary pal- pus two-segmented, second segment slightly longer than first. Proboscis long, probably having extended to sternum 1 or 2 when in repose. Labrum finely imbricate with small areas of rugulose integument along margins; two large, narrow oval areas bordering me- dian area, such areas black and faintly im- bricate; labrum noticeably bends posteriorly at lateral borders. Clypeus not as strongly protuberant as in extant Euglossa (figs. 3, 4); lacking median carina (which runs length- wise in extant species); ‘integument with weak punctures separated by about a punc- ture width or less; integument otherwise mi- nutely imbricate. Supraclypeal area as on clypeus. Epistomal sulcus forming an ap- proximate right angle, slightly bulging near anterior tentorial pit such that clypeus ap- pears to be gently depressed before joining paraocular area. Frontal line strongly im- pressed, running from between antennae to midway between antennal sockets and me- dian ocellus, becoming a faint depression from that point on. Integument surrounding antennal sockets distinctive in a ring approx- imately half of an antennal socket diameter; along inner borders integument of ring with minute punctures separated by 1-2 times puncture width, integument otherwise gla- brous; along outer borders glabrous and im- punctate, small punctures reappearing near border with compound eye; sockets slightly sunken owing to the weak protrusion of the clypeus, supraclypeal area, and suprantennal area; sockets separated by approximately two ENGEL: FIRST FOSSIL EUGLOSSA 5 antennal socket diameters, separated from compound eye by slightly less distance. Lower paraocular area with weak, coarse punctures, a few slightly elongate, separated by about a puncture width, integument be- tween minutely imbricate; punctures becom- ing more well defined and slightly smaller as they near compound eye and in paraocular area at level of antennae, separated by width or less in these areas. Suprantennal area slightly protuberant; punctures along lower margins near antennal sockets and supracly- peal area separated by about a puncture width and more well defined than those of clypeus, supraclypeal area, and lower para- ocular area; integument otherwise minutely imbricate; punctures rapidly becoming closer until separated by much less than a puncture width (but not contiguous), such dense punc- tures covering majority of suprantennal area; punctures becoming slightly smaller and weaker near ocelli and vertex. Small punc- tures between lateral ocelli separated by ap- proximately 1—2 times puncture width and ir- regularly scattered, integument between smooth, eventually giving way to the inter- ocellar furrow (impunctate, gentle depres- sion); median ocellus separated from lateral ocellus by just less than one-half ocellar di- ameter (based on the median ocellus); dis- tance between lateral ocelli approximately one ocellar diameter. Ocellorbital area mi- nutely imbricate and mostly impunctate ex- cept for a few irregularly placed, faint, small punctures on area nearest compound eye; area just slightly more than one ocellar di- ameter wide. Vertex behind interocellar fur- row gently ridged, extending just above up- per tangent of compound eyes; punctures re- appearing on vertex and gena like those on majority of suprantennal area. Forewing with basal vein straight, basad cu-a by about 1.5 times vein width (difficult to determine because of debris on wing); sec- ond submarginal cell about as long as first (but veins demarcating cells are heavily cov- ered with debris and thus hard to determine); 1m-cu entering second submarginal cell along apical quarter of cell; marginal cell reaching nearly to wing apex, cell apex acute, slightly separated from wing margin. Hind wing lack- ing jugal lobe, instead with a jugal comb com- posed of long, stiff bristles (fig. 5). 6 AMERICAN MUSEUM NOVITATES NO. 3272 r ae Dial Fig. 1. Photomicrograph of female head of Euglossa moronei, n. sp. (Photo by D.A. Grimaldi, AMNH). 1999 ENGEL: FIRST FOSSIL EUGLOSSA weenie endail ANTUALCUUMUAAE SUA L/ tr Lees sae NN WZ wqitas I Fig. 2. Illustration depicting scene from figure 1, Euglossa moronei, n. sp., with positions of debris and ant indicated. 8 AMERICAN MUSEUM NOVITATES Figs. 3,4. Female heads of two Euglossa spe- cies, profile. 3. Euglossa (Euglossa) ignita Smith. 4. E. moronei, n. sp., reconstructed profile. Pubescence entirely fuscous and simple. Mandible with minute hairs along faint outer grooves, extending about % to % distance to mandible base; those of ventral surface lon- ger and extending the entire length of man- dible. Labrum with many long hairs, longest and most dense laterally, somewhat shorter medially. Hairs of face longest just below an- tennal sockets and on lateral margins of clyp- eus near epistomal angle; hairs progressively becoming more sparse and shorter on upper half of suprantennal area and around ocelli, once again becoming long on vertex. Upper border of antennal sockets lacking plumose pubescence typical in extant Euglossa; outer border of antennal sockets without pubes- cence. Compound eyes bare. PRESERVATION: The amber piece is tear- drop in shape with the Euglossa head posi- tioned at one side of its broadest end. The tongue is extending away from the face to- wards the broadest border of the amber piece (figs. 1, 2). There is a small ant (worker) po- sitioned near the head (fig. 2) as well as much debris throughout the amber piece. The metasoma is completely lost as is the vast majority of the mesosoma and legs excepting for a few tiny fragments near the forewing which is preserved in about half of total length of the amber piece from the head. The forewing is mostly covered in debris, making it difficult to follow the veins. A fragment of the hind wing, which is mostly torn (fig. 5), is positioned on the opposite side of the piece. The upper left corner of the head is NO. 3272 Ie eis . Bee e Sey e Baad ei Seca : 4 : : Fig. 5. Hind wing fragment of Euglossa mo- ronei, n. sp., depicting the jugal comb at wing base. difficult to view due to the strong curvature of the amber at this point. The posterior sur- face of the head is easily viewed and the pro- boscis, its base in particular, can be examined in quite extraordinary detail. HOLOTYPE: Female (M-1149), Miocene Dominican amber, Morone Amber Collec- tion, Turin, Italy; specimen incomplete. ETyMo_Locy: The specific epithet is a pa- tronym honoring Dott. Ettore Morone who graciously allowed the study of this, and many other, fine fossils from his collection. DISCUSSION The placement of this species within the Euglossini and Euglossa despite the incom- plete specimen can be made with remarkable ease and confidence. Among the corbiculate Apinae the orchid bees and bumble bees lack a jugal lobe in the hind wing. The euglossines are further modified with a series of strong bristles in place of the jugal lobe (the jugal comb); a distinctive synapomorphy of the Euglossini. As was mentioned in the descrip- tion above, a fragment of the hind wing is preserved with the specimen and nicely ex- hibits this unique structure (fig. 5). Another feature unique to the euglossines among cor- biculate bees is the fusion of the mentum and submentum (termed the “‘lorum”’ by Plant and Paulus, 1987). The postmentum in most bee groups exists as a single structure with a well- defined mentum and submentum. In the Ap- ini, Bombini, and Meliponini the mentum and submentum become separated, one of several synapomorphies for this grouping. In the fos- sil, the base of the mouthparts can be plainly seen and, although there is much connective tissue exposed, the entire mentum as well as 1999 TABLE 2 Character Matrix for Euglossine Genera‘ 111111 123456789012345 Aglae 11002000010001 Eufriesea 00011001110001 Euglossa 00111110000001 Eulaema 11102001110001 Exaerete’ azteca Exaerete dentata O0011110000111 Exaerete frontalis 10011100001111 Exaerete smaragdina 1001 1100001111 Exaerete trochanterica 000111000001 11 Apis 00000??700200000 Bombus 000007?700200000 Xylocopa 00000?700200000 “Updated and modified from Kimsey (1987) and Michener (1990) with the inclusion of outgroups (Apis, Bombus, and Xylocopa). Character descriptions are pro- vided in table 3. ’ Because of the variability in some generic-level characters, I have coded each species of this genus in- dividually. 1 1 l 1 OO0OOLTILO0O0O0OI1IILIL 1 1 1 1 submentum can be easily found. As is to be expected for a putative euglossine, there is a clear connection between the base of the men- tum and the medial region of the submentum. The mentum among the Apini, Bombini, and Meliponini tends to be fairly short and any- where from weakly to strongly V-shaped at its apex, while in the fossil and extant Eug- lossini the mentum is long and narrow with a blunt or only weakly concave apex. Similarly, the euglossines are the only corbiculate bees that possess a submental spine (see Terminol- ogy) near the apex of the submental arms and this character can also be clearly distinguished in the fossil. In addition to these characters, the greatly elongate mouthparts and short ma- lar space also attest to the placement of this species in the Euglossini, while its labrum and pubescence position it in Euglossa (refer to Description). The phylogenetic affinities among the gen- era of Euglossini are difficult to ascertain and thus the overall position of Euglossa (and thereby this fossil) is of some uncertainty. In the three phylogenetic analyses that have been undertaken, the genera have undergone considerable rearrangement (fig. 6). As was observed by Michener (1990), many of the characters used in the analyses of Kimsey ENGEL: FIRST FOSSIL EUGLOSSA 9 TABLE 3 Character Descriptions for Data Matrix Presented in Table 2. The two multistate characters (5 and 10) are treated as nonadditive. 1. Labial palpus: four-segmented (0); two-segmented (1). 2. Scutellum in profile: convex, depressed medially (O); flat, bladelike (1). 3. Scutellar tuft: absent (0); present (1). 4. Sternal groove: absent (O); present (1). 5. Male mesotibial carina: absent (0); incomplete (1); complete (2). 6. Male metatibial slit: broad, ovoid (OQ); narrow, curved (1). 7. Male metatibial slit: reaching apex (0); not reaching apex (1). 8. Male metatibial shape: round apical margin (0); pointed apical margin (1). 9. Gonocoxal projection: absent (0); present (1). 10. Volsella: ovoid (0); linear, setose (1); fused (2). 11. Hypoepimeral knob: absent (0); present (1). 12. Male metafemoral dentition: absent (0); present (1). 13. Male mesobasitarsal tooth: absent (0); present (1). 14. Male metatibial slit: absent (O); present (1). 15. Jugal comb: absent (0); present (1). (1982, 1987) are autapomorphies for partic- ular genera and provide no grouping infor- mation. A revised data matrix for the Eu- glossini is presented in table 2 and descrip- tions of characters in table 3. In addition to removing the autapomorphic characters, I have made the following modifications of the studies of Kimsey (1982, 1987) and Miche- ner (1990): included three outgroups (Apis, Bombus, and Xylocopa) in place of assumed character polarizations; included three char- acters as potential synapomorphies for the tribe; broken the genus Exaerete into its con- stituent species because that genus varies; added some characters pertinent to the inter- nal phylogeny of Exaerete. The data were manipulated in DADA (Nixon, 1995) and submitted from there to Hennig86 (Farris, 1988) for analysis using the mh* and bb* commands. Trees were visualized and printed using CLADOS (Nixon, 1993). Three trees resulted of length 22, CI 0.77, and RI 0.85; the strict consensus (i.e., Nelson tree) of which is depicted in figure 7. Only two of the resultant topologies are strictly supported while the third is ambiguously supported (Nixon and Carpenter, 1996). The topology resulting from this analysis 10 AMERICAN MUSEUM NOVITATES NO. 3272 Euglossa Euglossa Euglossa Exaerete Exaerete Exaerete Eufriesea Eufriesea Eufriesea Eulaema Eulaema Eulaema Aglae Aglae Aglae Kimsey, 1982 Kimsey, 1987 Michener, 1990 Fig. 6. Previously proposed phylogenies for tribe Euglossini. Apis 10 Bombus 1>2 Xylocopa 1 2 5 Aglae y 3 8 9 O>10>11>2 La team FUlaema 5 14 15 i 2 O>10>10>1 eee lanl Eufriesea 3 7 j I{-—Euglossa O>10>1 O>1 ; fale j= EX azteca O>1 7 eae f= EX dentata 12 13 O>1 Bia Ex trochanterica ee Ex frontalis I ee Ex smaragdina Fig. 7. Phylogeny of Euglossini derived from cladistic analysis of data presented in table 2 (strict consensus of 3 trees, length 22, CI 0.77, RI 0.85). Black bars indicate unreversed changes, white bars indicate homoplastic character transitions. The character number is indicated above the bar and the state change is indicated below. 1999 differs most notably in the position of Euf- riesea, which is grouped with a Euglossa + Exaerete clade. Both Kimsey (1982, 1987) and Michener (1990) grouped Eufriesea with the genera Eulaema and Aglae. The grouping presented here is supported by the presence of a strong sternal groove into which the elongate proboscis fits when in repose. This character is absent in the outgroup taxa, Eu- laema, and Aglae. Kimsey (1982, 1987) as- sumed that the absence of this character was a derived reversal, the presence of such a groove was plesiomorphic for the tribe, and was thereby unavailable to group Eufriesea with Euglossa and Exaerete. Kimsey’s alter- native grouping of Eufriesea with Eulaema and Aglae was supported by the pointed api- cal margin of the male hind tibia (although absent in Aglae), the presence of the gono- coxal projection (also absent in Aglae), and the linear volsella. Aglae was considered “‘so highly modified’’ (Kimsey, 1987: 68) that the absence of some of these character states were a priori considered reductions and over- looked. Aglae is indeed remarkably unique among euglossine genera as Kimsey de- scribed, yet her position appears unjustified by outgroup comparison and the presence of a sternal groove is here interpreted as derived while the sole character fixed across the three taxa in Kimsey’s arrangement (the linear vol- sella) is considered plesiomorphic for the tribe. The genus Euglossa is believed, once again, to be a fairly derived group of orchid bees and sister to the genus Exaerete. Thus, E. moronei implies that the genus Euglossa is at least 20 Ma, with the origin and more basal diversification of the orchid bees oc- curring sometime before in the early Eocene or perhaps even the Paleocene. It is difficult to make any inferences of zoo- geographical patterns within the Euglossini at the generic level owing to the widespread dis- tribution of the genera. The complete absence of subgeneric or species-level cladograms also prohibits such a discussion. Despite these dif- ficulties the following observations seem ap- parent. The genus (and the tribe for that matter) must have previously been more widely dis- tributed in the Caribbean than it is today with subsequent extinction occurring. Apparently there were at least two dispersal events for euglossines into the West Indies. The earliest ENGEL: FIRST FOSSIL EUGLOSSA 11 one, for which no species survived to the pre- sent day, occurred sometime during or prior to the Miocene. The apparent lack of affinities between E. moronei and any extant subgenus or species group (see Diagnosis above) sug- gests that a lineage of Euglossa had been wide- ly distributed on an ancient Mesoamerican land mass or had at some point invaded the ancient West Indies, perhaps from South America, and been influenced by the tectonic changes in the region. In either case, this species eventually, along with any other orchid bees that may have been with it, perished from the region as the land mass became fragmented and islands be- came submerged and reconstituted. This is also true for other bee genera that are known as amber inclusions but are today extinct in the Greater Antilles [e.g., Chilicola (Colletidae: Michener and Poinar, 1996), Heterosarus (An- drenidae: Rozen, 1996), Neocorynura (Halic- tidae: Engel, 1995)]*. The most recent orchid bee colonization seems to have been from Central America into Jamaica, after the island surfaced from the Gulf sometime in the middle Miocene about 20 Ma (Steineck, 1974; Bus- kirk, 1985), which is supported by the seeming affinities of the endemic FE. jamaicensis with species in southern Mexico and Costa Rica (Dressler, 1978). Trinidad and Tobago are South American in faunal composition and geological origin (a pattern that holds for the euglossine fauna of these islands) and do not represent a dispersal event into the Caribbean. The fixation of orchid specialization across the tribe Euglossini suggests that it is a prim- itive characteristic of the tribe and was al- ready present in the common ancestor of all orchid bees (Schultz et al., 1996). Therefore, given the presumed phylogenetic position of this species within the genus Euglossa, it too was presumably a regular visitor of orchids. If this is indeed the case, then this is the oldest record of the specialized association between orchids and their bee visitors. Moreover, the *Poinar (1998a: 34) enigmatically considers these genera, except Heterosarus, to be entirely extinct when each is actually alive and well in the modern mainland bee faunas. In an accompanying note (Poinar, 1998b) he incorrectly asserts that euglossines do not occur in the West Indies while E. jamaicensis is endemic to at least one of these islands (other species also occur in Trini- dad: see Introduction). He also incorrectly assigns the panurgine genus Heterosarus (under the lapsus Heter- osaurus) to the Halictidae (1998a: 34)! 12 AMERICAN MUSEUM NOVITATES presence of E. moronei and Eufriesea melis- siflora indirectly indicates that the Orchida- ceae was a component of the Oligocene-Mio- cene flora of Hispaniola. A presumed orchid seed exists in a private collection of Domin- ican amber (D. A. Grimaldi, personal com- mun.), which would be the only fossil orchid and would support this notion. In lieu of a definitive fossil orchid, the addition of Orchi- daceae to this paleoflora remains an assump- tion. Similar inferences have indirectly added significant floral components to these ancient forests. For example, the presence of palm bugs (Heteroptera: Thaumastocoridae) and fig wasps (Hymenoptera: Agaonidae) indicates the presence of palm and fig trees, respective- ly (Grimaldi, 1996). Resolution of the relationships among euglossines would potentially provide greater insights into orchid bee biogeography as well as a foundation upon which to examine in greater detail the bee-orchid association and coevolution. As it presently stands we know that males of all of the nonparasitic genera pollinate the orchid tribes Cymbidieae, Epi- dendreae, and Maxillarieae while Exaerete males visit Maxillarieae (the orchid associ- ates of Aglae remain to be discovered). As was also observed by Oliveira (1998), the distribution of these tribes mapped on eu- glossine phylogeny (no matter what topology is preferred) provides little information for understanding bee-orchid coevolution. Fu- ture work should focus on cladistic studies of species-level relationships, particularly in the genus Euglossa, coupled with studies on the relationships among orchids (the foun- dations of which can be found in Dressler, 1993). Such lower-rank analyses will also create the framework for finer discussions of euglossine biogeography, possible morpho- logical specialization, and patterns in spe- cies-specific chemical blends. For the time being, however, we can only marvel at the complexity of orchids and their bee visitors and, like Darwin (1862), speculate on the even greater story of their union that awaits our discovery. REFERENCES Ackerman, J. D. 1983. Specificity and mutual dependency of NO. 3272 the orchid-euglossine bees interaction. Biol. J. Linn. Soc. 20: 301-314. Buskirk, R. E. 1985. Zoogeographic patterns and tectonic history of Jamaica and the northern Ca- ribbean. J. Biogeogr. 12: 445-461. Chavarria, G., and J. M. Carpenter 1994. Total evidence and the evolution of high- ly social bees. Cladistics 10: 229-258. Cockerell, T.D.A. 1904. Descriptions and records of bees. Ann. Mag. Nat. Hist., ser. 7, 14: 21-30. 1908. Notes on the bee-genus Exaerete. Psy- che 15: 41-42. 1927. Zoology of Colorado. Boulder: Univ. Colorado Press. Darwin, C. R. 1862. On the various contrivances by which British and foreign orchids are fertil- ised by insects, and on the good effects of intercrossing. London: John Murray. Dodson, C. H. 1966. Ethology of some bees of the tribe Eu- glossini (Hymenoptera: Apidae). J. Kansas Entomol. Soc. 39: 607-629. Relationships between pollinators and orchid flowers. Atas Simp. Biota Amaz. 5: 1-72. The role of chemical attractants in or- chid pollination. Jn K. L. Chambers (ed.), Biochemical evolution: 83-107. Corvallis: Oregon State Univ. Press. Coevolution of orchids and euglossine bees. Jn L. Gilbert and P. H. Raven (eds.), Coevolution of animals and plants: 91— 99. Austin: Univ. Texas Press. Dodson, C. H., and G. P. Frymire 1961. Natural pollination of orchids. Missouri Bot. Gard. Bull. 69: 133-152. Dodson, C. H., R. L. Dressler, H. G. Hills, R. M. Adams, and N. H. Williams 1969. Biologically active compounds in orchid fragrances. Science 164: 1243-1249. Dressler, R. L. 1978. An infrageneric classification of Eu- glossa, with notes on some features of special taxonomic importance (Hyme- noptera; Apidae). Rev. Biol. Trop. 26: 187-198. 1982a. New species of Euglossa. Ill. The bur- sigera species group (Hymenoptera: Apidae). Ibid. 30: 131-140. 1982b. Biology of the orchid bees (Euglossini). Annu. Rev. Ecol. Syst. 13: 373-394. Phylogeny and classification of the or- chid family. Portland: Dioscorides Press. Engel, M. S. 1995. Neocorynura electra, a new fossil bee 1967. 1970. 1975. 1993. 1999 species from Dominican amber (Hy- menoptera: Halictidae). J. New York Entomol. Soc. 103: 317-323. A new fossil bee from the Oligo-Mio- cene Dominican amber (Hymenoptera: Halictidae). Apidologie 28: 97-102. 1998a. A new species of the Baltic amber bee genus Electrapis (Hymenoptera: Api- dae). J. Hym. Res. 7: 94-101. 1998b. Phylogeny, classification, and evolu- tionary ethology of the bee tribe Au- gochlorini (Hymenoptera: Halictidae). Ph.D. diss., Cornell Univ., Ithaca, NY. Farris, J. S. 1988. Hennig86, Version 1.5. Program and documentation. Port Jefferson Station, NY. Gar6falo, C. A. 1997. 1985. Social structure of Euglossa cordata nests (Hymenoptera: Apidae: Euglos- sini). Entomol. Gen. 11: 77-83. 1991. Comportamento de nidificagdo e estru- tura de ninhos de Euglossa cordata (Hymenoptera: Apidae: Euglossini). Rev. Bras. Biol. 52: 187-198. Gar6éfalo, C. A., E. Camillo, S. C. Augusto, B.M.V. de Jesus, and J. C. Serrano Nest structure and communal nesting in Euglossa (Glossura) annectans Dress- ler (Hymenoptera, Apidae, Euglossini). Rev. Bras. Zool. 15: 589-596. Grimaldi, D. A. 1995. The age of Dominican amber. In K. B. Anderson and J. C. Crelling (eds.), Amber, resinites, and fossil resins: 203-217. Washington DC: Am. Chem. Soc. Symp. Amber: window to the past. New York: Am. Mus. Nat. Hist. Hoffmannsegg, J. C. von 1817. Entomologische Bemerkungen bei Gel- egenheit der Abhandlungen tiber amer- ikanische Insekten. Zool. Mag., Wie- demann 1: 49-110. Iturralde-Vinent, M. A., and R.D.E. MacPhee 1996. Age and paleogeographical origin of Do- minican amber. Science 273: 1850-1852. Kimsey, L. S. 1979. An illustrated key to the genus Exaer- ete with descriptions of male genitalia and biology (Hymenoptera: Euglossini, Apidae). J. Kansas Entomol. Soc. 52: 735-746. 1998. 1996. 1980. The behaviour of male orchid bees (Apidae, Hymenoptera, Insecta) and the question of leks. Anim. Behav. 28: 996-1004. 1982. Systematics of bees of the genus Euf- ENGEL: FIRST FOSSIL EUGLOSSA 13 riesea (Hymenoptera, Apidae). Univ. California Publ. Entomol. 95: 1-125. 1984. The behavioural and structural aspects of grooming and related activities in euglossine bees (Hymenoptera: Api- dae). J. Zool. (London) 204: 541-550. 1987. Generic relationships within the Eu- glossini (Hymenoptera: Apidae). Syst. Entomol. 12: 63-72. Kimsey, L. S., and R. L. Dressler 1986. Synonymic species list of Euglossini. Pan-Pac. Entomol. 62: 229-236. Latreille, P A. 1802. Histoire naturelle des fourmis, et recueil de mémoires et dobservations sur les abeilles, les araignées, les faucheurs, et autres insectes Paris: Crapelet. Lepeletier de Saint Fargeau, A. L. M. 1841. Histoire naturelle des insectes, Hymén- optéres. Paris: Roret. Lepeletier de Saint Fargeau, A. L. M., and J. G. A. Serville Entomologie, ou histoire naturelle des crustaces, des arachnides et des insec- tes. In P. A. Latreille (ed.), Encyclopé- die méthodique, histoire naturelle, In- sectes 10: 1-344. Paris: Diderot. Linnaeus, C. 1758. Systema naturae per regna tria naturae, secundum classes, ordines, genera, spe- cies, cum characteribus, differentiis, synonymis, locis, vol. 1, ed. 10, refor- mata. Holmiae [Stockholm]: Salviae. Michener, C. D. 1944. Comparative external morphology, phylogeny, and a classification of the bees (Hymenoptera). Bull. Am. Mus. Nat. Hist. 82: 151-326. An interesting method of pollen collect- ing by bees from flowers with tubular anthers. Rev. Biol. Trop. 10: 167-175. The social behavior of the bees: a com- parative study. Cambridge MA: Har- vard Univ. Press. Classification of the Apidae (Hyme- noptera). Univ. Kansas Sci. Bull. 54: 75-164. Michener, C. D., and A. Fraser 1978. A comparative anatomical study of mandibular structure in bees. Univ. Kansas Sci. Bull. 51: 463-482. Michener, C. D., and G. O. Poinar, Jr. 1825. 1962. 1974. 1990. 1996. The known bee fauna of the Dominican amber. J. Kansas Entomol. Soc., suppl. 69: 353-361. Moure, J. S. 1964. A key to the parasitic euglossine bees and a new species of Exaerete from 14 AMERICAN MUSEUM NOVITATES Mexico (Hymenoptera Apoidea). Rev. Biol. Trop. 12: 15-18. 1968. Espécies novas de Euglossa da Amér- ica Central (Hymenoptera, Apidae). Bol. Univ. Fed. Parana, Zool. 3: 13-64. 1989. Glossuropoda, novo subgénero de Eu- glossa, e duas espécies novas da Ama- zonia, do mesmo subgénero (Apidae Hy- menoptera). Mem. Inst. Oswaldo Cruz, Rio de Janeiro, supl. 4, 84: 387-389. Moure, J. S., and S. EF Sakagami 1962. As mamangabas sociais do Brasil (Bombus Latr.) (Hym. Apoidea). Stud. Entomol. 5: 65-194. Nixon, K. C. 1993. CLADOS, Version 1.4.88. Program and Documentation. L. H. Bailey Horto- rium, Cornell Univ., Ithaca, NY. 1995. DADA, Version 0.96.1. Program and Documentation. L. H. Bailey Horto- rium, Cornell Univ., Ithaca, NY. Nixon, K. C., and J. M. Carpenter 1996. On consensus, collapsibility, and clade concordance. Cladistics 12: 305-321. Oliveira, M. L. de 1998. Interacdes entre Euglossini e Orchida- ceae: uma mini-reviséo sob uma per- spectiva cladistica e biogeographica. Encontro sobre Abelhas, Ribeirao Preto 3: 144-150. Plant, J. D., and H. FE Paulus 1987. Comparative morphology of the post- mentum of bees (Hymenoptera: Apo- idea) with special remarks on the evo- lution of the lorum. Z. Zool. Syst. Evo- lutions-Forsch. 25: 81-103. Poinar, G., Jr. 1998a. Paleoeuglossa melissiflora gen. n., sp. n. (Euglossinae: Apidae), fossil orchid bees in Dominican amber. J. Kansas En- tomol. Soc. 71: 29-34. [Publication date: 9 February 1999] 1998b. Orchid bees. Mag. Am. Orchid Soc. 67: 578. Roberts, R. B., and C. H. Dodson 1967. Nesting biology of two communal bees, Euglossa imperialis and Euglossa ignita (Hymenoptera: Apidae), includ- ing description of larvae. Ann. Ento- mol. Soc. Am. 60: 1007-1014. Roig-Alsina, A., and C. D. Michener 1993. Studies of the phylogeny and classifi- cation of long-tongued bees (Hyme- noptera: Apoidea). Univ. Kansas Sci. Bull. 55: 123-173. Rozen, J. G., Jr. 1996. A new species of the bee Heterosarus NO. 3272 from Dominican amber (Hymenoptera: Andrenidae; Panurginae). J. Kansas Entomol. Soc. suppl. 69: 346-3572. Sakagami, S. E 1972. Bumble bees collected by the Califor- nia Academy Lingnan Dawn-Redwood Expedition to Central West China, 1948 (Hymenoptera: Apidae). Pan-Pac. En- tomol. 48: 153-174. Some bumblebees from Korea with re- marks on the Japanese fauna (Hyme- noptera, Apidae). Ann. Hist.-Nat. Mus. Natl. Hung. 67: 293-316. Sakagami, S. F, S. Laroca, and J. S. Moure 1967. Two Brazilian apid nests worth record- ing in reference to comparative bee so- ciology, with description of Euglossa melanotricha Moure sp. n. (Hymenop- tera, Apidae). Annot. Zool. Jpn. 40: 45- 54. Schultz, T. R., R. B. Cocroft, and G. A. Churchill 1996. The reconstruction of ancestral char- acter states. Evolution 50: 504-511. Schultz, T. R., M. S. Engel, and M. Prentice 1999. Resolving conflict between morpholog- ical and molecular evidence for the or- igin of eusociality in the corbiculate bees (Hymenoptera: Apidae): a hypoth- esis-testing approach. Univ. Kansas Nat. Hist. Mus. Spec. Publ. 24: 110-123. Smith, E 1874. A-revision of the genera Epicharis, Cen- tris, Eulema [sic], and Euglossa, belong- ing to the family Apidae, section Sco- pulipedes. Ann. Mag. Nat. Hist., ser. 4, 13: 318-322, 357-373, 440-446. Steineick, P. L. 1974. Foraminiferal paleoecology of the Montpelier and lower coastal groups (Eocene-Miocene), Jamaica, West In- dies. Paleogeogr. Paleoclimatol. Pa- leoecol. 16: 217-242. Vogel, S. 1966. Parfiimsammelnde Bienen als Bestéub- er von Orchidaceen und Gloxinia. Os- terr. Bot. Z. 113: 302-361. Wille, A., and L. C. Chandler 1964. A new stingless bee from the Tertiary amber of the Dominican Republic (Hy- menoptera; Meliponini). Rev. Biol. Trop. 12: 187-195. Zucchi, R., S. EF Sakagami, and J. M. EK Camargo 1969. Biological observations on a neotropi- cal parasocia] bee, Eulaema_ nigrita, with a review on the biology of Eu- glossinae. J. Fac. Sci. Hokkaido Univ., Zool., ser. 6, 17: 271-380. 1975. Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates and Bulletin published during the last five years are available at World Wide Web site http://nimidi.amnh.org. Or address mail orders to: American Museum of Natural History Library, Central Park West at 79th St., New York, NY 10024. TEL: (212) 769-5545. FAX: (212) 769- 5009. E-MAIL: scipubs @ amnh.org This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).