Novttates PUBLISHED by the AMERICAN MUSEUM of NATURAL HISTORY CENTRAL PARK WEST at 79TH STREET NEW YORK, N.Y

Home , Ammobatini

AMERICAN MUSEUM Novttates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET NEW YORK, N.Y. 10024 U.S.A. NUMBER 2551 OCTOBER 10, 1974

JEROME G. ROZEN, JR., AND RONALD J. MCGINLEY Systematics of Ammobatine Bees Based on Their Mature Larvae and Pupae (Hymenoptera, Anthophoridae, Nomadinae) ., 1, 'm -; ;f, I1, Systematics of Ammobatine Bees Based on Their Mature Larvae and Pupae (Hymenoptera, Anthophoridae, Nomadinae)

JEROME G. ROZEN, JR. Deputy Director for Research and Curator of Hymenoptera The American Museum of Natural History

RONALD J. MCGINLEY Research Assistant, the American Museum of Natural History Present Address: Department of Entomology and Parasitology University of California, Berkeley

ABSTRACT SYSTEMATICS AND PHYLOGENY Mature larvae of representatives of four genera The Ammobatini were divided by Popov of the Ammobatini are taxonomically de- (1951) into the Ammobatini (Ammobates, Par- scribed-Pseudodichroa, A mmobates, Morgania, ammobatodes, and Oreopasites) and the Pasitini and Oreopasites. A key is provided for their iden- (including among others, Morgania and presum- tification and the tribe is characterized on the ably Pseudodichroa). In reviewing Pseudodi- basis of the mature larvae. Pupae of representatives of two genera are described-Oreopasites chroa, Rozen (1968) used the older, more con- and Morgania. A preliminary key to the pupae of servative classification employed by Michener the subfamily Nomadinae is also presented. (1944) and others because there was some evi- dence that a more detailed study of male geni- The study was undertaken to illuminate the talia might necessitate a revision of Popov's higher classification and phylogeny of antho- classification. Recently Baker (1971) has fol- phorid bees, to provide means for identifying the lowed Popov's division and correctly pointed out tribe Ammobatini and its genera on the basis of that in the Pasitini the labrum of the adult ex- immature characters, and to establish a back- tends only to the mandible, whereas in the Am- ground for a forthcoming evolutionary study of mobatini, sensu stricto, it extends well beyond the ammobatine genus Oreopasites. the mandibles. However, we note that within the The present paper may be considered a "third Pasitini there is a difference among the genera generation" investigation on bee larvae. Miche- with respect to whether the mandibles of adults ner's work on larvae (1953) was the original com- are directed mesiad so that their points overlap in *prehensive treatment of larval forms but did not repose (e.g., Morgania) or whether the mandibles describe any of the Ammobatini. Since its publi- are directed somewhat posteriorly so that their cation various workers, with new and more ex- apexes cross in repose (e.g., Pseudodichroa; see tensive material, readdressed various taxa, and Rozen, 1968, fig. 2). Furthermore, an analysis of specifically Rozen produced a series of papers larval features does not suggest a grouping of treating the larvae of the Anthophoridae. Rozen Morgania and Pseudodichroa on the one hand in 1966 covered the Nomadinae and discussed and Oreopasites and Ammobates on the other. the only genus of ammobatines (Oreopasites) Indeed a phenetic grouping would seem to place then available. The present paper is yet another Oreopasites and Morgania close together, with refinement, made possible by recent acquisitions Ammobates somewhat divergent from them and of larval specimens of three more genera of the with Pseudodichroa apparently farthest away. Ammobatini. Pupae of the tribe have gone un- With respect to phylogeny, such larval features as described until now. the anterior placement of the posterior tentorial All nomadine bees are parasitic, the larvae pits in Pseudodichroa, the navicular shape of the feeding in the cells of various solitary bees. Am- predefecating larva ofPseudodichroa, the swollen mobatine bees even as adults are rare in collec- mandibular bases ofMorgania, and the strong an- tions. The acquisition of immature specimens for terior projection of the head features of Ammo- the study, therefore, would have been virtually bates are probably derived. However, too few lar- impossible without the excellent support of the val characters can be classified as being primitive National Science Foundation (grants GB-5407 or specialized to make possible an analysis of the and GB-32508) which we wish to acknowl- phylogeny of the tribe. We also note the simi- edge with appreciation. We also thank Ms. Liliane larity of the pupae of Morgania and Oreopasites. Floge and Mrs. Barbara Rozen who assisted in For all of these reasons we continue to follow the illustrations and Ms. Phyllis Browne who the more conservative classification, i.e., placing typed the manuscript. All specimens are in the the Ammobatini, sensu stricto, and Pasitini into a collection of the American Museum of Natural single tribe at least for the present, with the hope History. that a broad comprehensive analysis of all avail- 3 4 AMERICAN MUSEUM NOVITATES able characters will lead to a clear understanding and Scrapter seem paradoxical since both feed on of the phylogeny of the group and to an appro- the same semiliquid provisions. Loss or reduction priate classification. of the abductor apodeme is found among certain other colletids with watery nest provisions, e.g., Colletes and Euryglossa. LARVAE Another interesting point concerning Pseudo- The anatomy of bee larvae and techniques for dichroa is that both the anterior and posterior studying larvae have been discussed by Michener points of articulation of the mandible are some- (1953) and more recently by Rozen and McGin- what separate from the articulating points on the ley (1974). The posterior margin of the head head capsule. This fact plus the missing adductor capsule of the Ammobatini (excepting Pseudo- mandibular apodeme suggests that the mandibu- dichroa) and ofNomada possesses two transverse lar mechanism at least of the last instar may be weak intemal ridges, one anterior to the other structurally weak and the mandible may not be ("apparent posterior thickening" and "posterior required to function very actively during feeding. thickening," figs. 17, 24). This characteristic is This deduction would seem to be supported by also weakly developed in Paranomada but has the very short mandibles, the apexes of which not been investigated in Kelita. We presume the hardly reach the mouth area. On the other hand more anterior ridge to be the anatomical poste- larval mandibles of almost all the Nomadinae rior margin of the head capsule because the cap- (this apparently is not true of Paranomada and sule, so defined, would be normal in shape in Isepeolus) have nearly equally short mandibles lateral view for most bee larvae, and because the with adductor apodemes. This situation empha- sclerotization of the capsule ends at the anterior sizes that we know very little about how the ridge in Morgania. If this is the true anatomical short mandibles function, about what role the boundary of the head, then the sclerotized area swollen epipharynx of Nomada, Pseudodichroa, behind the posterior thickening in Ammobates, Ammobates and perhaps other Nomadinae play Pseudodichroa, and Oreopasites represents an ex- in the ingestion of food, and in general about the tension of the functional head capsule. Perhaps it feeding mechanisms of these larvae (or any bee is an ontogenetic holdover of an elongate head larvae). capsule of the first instar, an adaptation that Description of the Ammobatini could accommodate massive mandibular muscula- Based on Mature Larvae ture. In Pseudodichroa (fig. 8) the head exhibits Diagnosis. Mature larvae of the Ammobatini only a single transverse ridge. The placement of key readily to couplet 8 in Rozen (1966, the posterior tentorial pit midway along the hy- pp. 8-10) where Oreopasites and Nomada appear. postomal ridge seems to suggest that the true Because all ammobatines have a nonspiculate hy- posterior thickening has become obliterated but popharynx they can be separated from Nomada that the area from the posterior pits to the appar- which they tend to resemble. After the publica- ent posterior thickening is still homologous with tion of Rozen (1966), mature larvae of two other the extension of the head capsule found in other genera of Nomadini, Kelita and Paranomada, Ammobatini and in Nomadini. were collected. These demonstrate that the tribe We note that the mandible of Pseudodichroa is remarkably diverse (also reflected in the pupa (figs. 4-6) is peculiar because the abductor apo- of Paranomada) and raise intriguing questions re- deme is missing although the adductor apodeme garding the origin and monophyly of both the is reasonably well developed. The mandible of tribe and the subfamily. Although these larvae Scrapter (Colletidae), the host ofPseudodichroa, have not been studied in detail they can be has a well-developed adductor apodeme but the clearly distinguished from the Ammobatini abductor apodeme is absent, although the point because Kelita and Paranomada possess distinct of attachment of the abductor muscle is evident paired dorsal tubercles on most body segments. as a cuticular invagination of the mandibular cor- Furthermore their hypopharnyxes seem to be ium. These differences between Pseudodichroa spiculate as in the hypopharnyx ofNomada. ROZEN AND MCGINLEY: AMMOBATINE BEES 5

Head. Integument usually pigmented; scat- pharynx without distinct lobelike swelling above tered nonsetiform sensilla present; epipharynx, hypopharynx as found in Nomada but more hypopharynx, maxillae, and labium all nonspicu- (Pseudodichroa, fig. 7, Ammobates fig. 16) or late. As seen in frontal view, general outline of less (Oreopasites, Morgania) swollen ventrally. head nearly circular (Pseudodichroa, fig. 7, and Mandible slender (Oreopasites, Morgania, figs. some Oreopasites) to distinctly wider than long 20-22, Pseudodichroa, figs. 4-6) or robust (Am- (Ammobates, fig. 16); as for most of Nomadinae, mobates, figs. 12-14), moderately short, broad at head capsule elongate as seen in lateral view (figs. base when viewed adorally and tapering to simple 8, 17, 24) with frontal area sloping toward pro- apex; enlarged subapical teeth absent; cuspal area duced labroclypeal area; vertex evenly rounded not produced (Morgania, Pseudodichroa) or pro- or with pronounced median swelling (Ammo- duced and poorly defined (Oreopasites, Ammo- bates, fig. 17). Tentorium incomplete; anterior bates); cuspal teeth absent; dorsal and ventral in- arms very short and slender to virtually absent ner edges well formed and serrate or smooth and (Morgania); posterior arms very short and slender nonserrate (Ammobates); apical concavity absent (Morgania, Oreopasites) or more elongate and or moderately well developed (Ammobates). threadlike (Ammobates, Pseudodichroa); dorsal Maxillae recessed, broadly fused with labium; arms absent; anterior pits minute (obscure in palpi robust basally, short (Morgania) to moder- Morgania) and positioned rather low, near pre- ately elongate (Ammobates). Labium greatly re- coila; posterior pits minute to small, situated cessed, not divided into prementum and post- below hypostomal ridge and well anterior to mentum; palpus minute (Morgania) or indicated apparent posterior margin of head (pits equi- by single sensillum. Hypopharynx poorly devel- distant from pleurostomal ridge and apparent oped or large and well defined (Ammobates, fig. posterior margin of head in Pseudodichroa). All 16). Salivary opening small, circular to elliptical internal ridges of head capsule tending to be in shape; salivary lips absent or present as a low weak or absent; as in Nomada, head capsule surrounding rim (Morgania). bordered posteriorly by two faint ridges in most Body. Integument of postdefecating forms cases (see discussion above); hypostomal ridge smooth (Ammobates) to wrinkled (Oreopasites, elongate, tending to be moderately to darkly pig- Morgania), dull (Oreopasites) to shiny (Mor- mented; hypostomal ridge slender to moderately gania); venter faintly to strongly (Oreopasites) well developed anterior to posterior tentorial spiculate; setae absent. Form moderately robust, pits; posteriad of pits, ridge becoming narrower moderately (Oreopasites) to extremely (Pseudo- and fading completely before reaching apparent dichroa, fig. 3) broad in dorsal view, curved an- posterior margin of head; pleurostomal ridge teriorly in postdefecating forms (postdefecating moderately developed to poorly developed and form of Pseudodichroa unknown), tapering pos- indistinct; ridge darkly pigmented to nonpig- teriorly; in postdefecating forms intersegmental mented; epistomal ridge virtually absent (Pseudo- lines well defined, shallowly (Ammobates) to dichroa, Morgania, Ammobates) or weakly devel- deeply (Oreopasites, Morgania) incised; dorsal oped and reaching level of antennae before com- tubercles absent although dorsum may be pro- pletely fading (some Oreopasites); coronal cleav- duced somewhat dorsolaterally in Oreopasites age line not visible (Oreopasites) or evident as a and Morgania; swellings below spiracles present diffused light line extending anteriorly to clypeal and well defined; dorsal intrasegmental lines pres- region; parietal bands faint. Antennal papillae ent and faint (predefecating form of Pseudo- low convexities (best developed in Oreopasites) dichroa, fig. 1) to apparently absent; abdominal each bearing two sensilla; antennal protuberances segment IX not produced ventrally; abdominal slightly developed (Oreopasites, Ammobates, fig. segment X central or perhaps dorsal (Ammo- 17) or absent. Labrum moderately (Oreopasites, bates, fig. 10) in attachment to segment IX; anus Pseudodichroa, fig. 8) to strongly projecting transverse, apical or somewhat dorsal in Pseudo- (Morgania, fig. 24, Ammobates, fig. 17) bearing dichroa (fig. 2); perianal area well defined, elon- two narrowly rounded or pointed tubercles of gate-elliptical and smooth (fig. 2). Spiracles (figs. variable size and distance from each other; epi- 9, 15, 25) not on elevations (Ammobates, Oreo- 6 AMERICAN MUSEUM NOVITATES pasites and postdefecating form of Pseudo- tuberances only slightly developed (fig. dichroa) or on slight elevations, which are lightly 24); maxillary palpi small, smaller than sclerotized and pigmented apically (Morgania); labral tubercles (fig. 24) ...... 3 atrium subglobular to subquadrate (Morgania); 3(2). Body, as seen in dorsal view (fig. 19), projecting above body surface, with rim; atrial conspicuously broad with intersegmental lines deeply incised; spiracular atria wall with (Oreopasites, Pseudodichroa, Morgania) (fig. 25) subquadrate, on slightly pig- or without (Ammobates) rows of small denticles; mented and sclerotized elevations; labrum peritreme present, moderately wide to very wide very strongly projecting (fig. 24); cuspal (Pseudodichroa); primary tracheal opening with area of mandible not produced (figs. 20, moderately long collar; subatrium annulated, 22); venter of abdomen faintly spiculate variable in length...... Morgania histrio transvaalensis Body, as seen in dorsal view, not conspicuously broad, with intersegmental KEY TO THE MATURE LARVAE lines moderately incised; spiracular atria OF THE AMMOBATINI (Rozen, 1954, fig. 2) subglobular, not 1. Posterior tentorial pit nearly equidistant on distinct elevations; labrum only from pleurostomal ridge and apparent moderately projecting; cuspal area of posterior margin of head capsule (fig. 8); mandible moderately produced (Rozen, labral tubercles separated by about 1966, figs. 56-58); venter of abdomen basal diameter of tubercle (fig. 7); strongly spiculate ...... Oreopasites adductor apodeme of mandible (fig. 4) absent ..... Pseudodichroa fumipennis Pseudodichroa fumipennis Bischoff Posterior tentorial pit obviously closer to Figures 1-9 apparent posterior margin of head cap- The following description is based on pre- sule than to pleurostomal ridge (figs. 17, defecating larvae; postdefecating forms are un- 24); labral tubercles usually separated by more than basal diameter of tubercle known. (figs. 16, 23) but, if close set (in Diagnosis. The extreme anterior position of undescribed species of Oreopasites) the posterior tentorial pits (fig. 8), the narrow tubercles somewhat elongate and pointed; separation of labral tubercles (about one basal adductor apodeme of mandible (figs. 12, diameter apart) (fig. 7), and the absence of ad- 20) present ...... 2 ductor apodeme of mandible (figs. 4, 6) distin- 2(1). Body outline in dorsal view (fig. 11) with guish Pseudodichroa fumipennis from all other intersegmental lines not incised, hence, known larval ammobatines. outline smooth; hypopharynx well de- Head (figs. 7, 8). As seen in frontal view (fig. fined and strongly projecting (figs. 16, 7), head nearly circular; median swelling of ver- 17); mandible (figs. 12-14) extremely robust basally with adoral-ventral surface tex absent. Anterior tentorial arms short and of mandible massively produced and slender; posterior arms somewhat elongate and rounded; vertex of head capsule with threadlike; anterior pits minute; posterior pits median, rounded swelling (fig. 17); small, situated below hypostomal ridge and un- antennal protuberances small but well like all other known larval ammobatines, nearly defined and conspicuous (fig. 17); maxil- equidistant from pleurostomal ridge and appar- lary palpi large, larger than labral ent posterior margin of head (fig. 8); posterior tubercles (fig. 17). .A mmobates carinatus thickening of head capsule obliterated (see dis- Body outline in dorsal view (fig. 19) with cussion above); apparent posterior thickening intersegmental lines moderately to deeply well defined but weak; entire area anterior to incised; hypopharynx poorly defined and of weakly projecting (figs. 23, 24); mandible apparent posterior margin head sclerotized; (figs. 20-22) moderately robust basally hypostomal ridge slender and moderately pig- with adoral-ventral surface at most only mented, only slightly curved dorsally near poste- moderately produced and rounded; ver- rior tentorial pit; pleurostomal ridge poorly devel- tex of head capsule without median, oped and indistinct; epistomal ridge virtually ab- rounded swelling (fig. 24); antennal pro- sent; coronal cleavage line present as a poorly ROZEN AND MCGINLEY: AMMOBATINE BEES 7

<~~~~-~~~~~2.0, rmm/ I~~~~~~~~~~~~~ o

/ . )<,ANTERIOR TENTORIAL PIT K)' ~~LABRAL TUBERCLE

8 3 SWELLING

FIGS. 1-9. Mature larva of Pseudodichroa fumipennis. 1. Entire larva, lateral view. 2. Apex of abdomen, lateral view. 3. Entire larva, dorsal view. 4-6. Right mandible, dorsal, inner and ventral views, respectively. 7. Head, anterior view. 8. Head, lateral view. 9. Spiracle, side view. Scale refers to figures 1, 3. 8 AMERICAN MUSEUM NOVITATES defined light line extending to clypeal area. An- The biology of this species was described by tennal protuberances absent. Labrum moderately Rozen and Michener (1968). projecting (not so strongly projecting as those of Morgania, Ammobates); labral tubercles moder- Ammobates carinatus Morawitz ately large, narrowly rounded and narrowly Figures 10-1 7 separated by about one basal diameter of tuber- Diagnosis. This species can be distinguished cle; epipharynx (fig. 7) strongly swollen. Man- from other known larval ammobatines because of dible (figs. 4-6) slender, broad as seen in adoral the smooth lateral outline of its body as seen in view and tapering unevenly to apex; cuspal area dorsal view (fig. 11), the well-defined and not produced; dorsal and ventral inner edges ser- strongly projecting hypopharynx (fig. 16), the rate with serration more extensive on dorsal massively produced adoral-ventral surface of edge; apical concavity not developed; unlike all mandible (figs. 12, 14), the very wide separation other ammobatines studied, the mandibular ad- of labral tubercles (over four basal diameters ductor apodeme is apparently absent. Maxillary apart) (fig. 16), the large maxillary palpi (figs. palpi small, short. Labial palpus indicated by sin- 16, 17), the presence of small antennal protuber- gle sensillum. Hypopharynx poorly defined, with ances and the median swelling of the vertex transverse ridging present, and exceeded by maxil- (fig. 1 7). lae (fig. 8); hypopharyngeal groove virtually ab- Head (figs. 16, 17). As seen in frontal view sent. Salivary opening elongate-elliptical; sur- (fig. 16), head distinctly wider than long, similar rounding rim absent. in overall shape to that of Morgania; vertex pro- Body (figs. 1-3). Integument smooth, not shiny; duced medially as a rounded swelling. Anterior venter faintly spiculate. Form conspicuous- tentorial arms short and slender, better devel- ly broad as seen in dorsal or ventral view (fig. oped than those of Morgania; posterior arms 3), uncurved and gently tapering anteriorly, somewhat elongate and threadlike; anterior pits sharply tapering posteriorly; form navicular in minute but conspicuous; posterior pits small, shape on live larva (Rozen and Michener, 1968); situated well below hypostomal ridge (lower than intersegmental lines not deeply incised; paired those of Morgania) and anterior to apparent pos- dorsal tubercles absent although anterior seg- terior margin of head; posterior thickening of ments produced somewhat dorsolaterally; dorsal head capsule nearly obliterated; apparent posteri- intrasegmental lines present but very faint; unlike or thickening weak but quite evident; area be- other known larval ammobatines, abdominal seg- tween thickenings sclerotized; hypostomal ridge ment X may have conspicuous lateral swellings moderately well developed and darkly pigmented (fig. 3) immediately anterior to perianal area; ab- although fading and incomplete just posterior to dominal segment X centrally attached to segment pleurostomal ridge, curved dorsally posterior to IX; unlike other ammobatines, anterior diameter posterior tentorial pit; pleurostomal ridge poorly of segment X much less than that of segment IX; developed and indistinct; epistomal ridge virtu- anus apical or perhaps slightly dorsal. Spiracles ally absent; coronal cleavage line present as a dif- varying in size; first two pair moderately large, fused light line extending anteriorly to clypeal last small, all others large; spiracles tending to be region. Antennal protuberances small but well directed dorsally because of broad body, not on defined. Labrum very strongly projecting; labral elevations; atrium (fig. 9) subglobular; atrial wall tubercles small, narrowly rounded and widely with numerous rows of small denticles; peritreme separated, about four basal diameters of tubercle very wide; subatrium broad but short, length less apart, epipharynx distinctly swollen. Mandible than two atrial lengths. Imaginal discs of genitalia (figs. 12-14) robust; unlike other ammobatines not visible. treated here, adoral-ventral surface massively pro- Material Studied. Two predefecating larvae, duced and rounded so that mandible appears Kommetjie, Cape Province, Republic of South broader in dorsal and ventral views than in adoral Africa, November 8, 1966, from cell ofScrapter view; tapering unevenly to apex as seen in adoral crassula (J. G. Rozen). Adults determined by view; cuspal area produced but poorly defined; Rozen. dorsal and ventral inner edges apparently without ROZEN AND MCGINLEY: AMMOBATINE BEES 9

APPARENT POSTERIOR THICKENING -POSTERIOR THICKENING MEDIAN SWELLING ANTERIOR TENTORIAL PIT 7/

LABRAL TUBERCLE HYPOPHARYNX -- MAXILLARY PALPUS POSTERIOR TENTORIAL PIT

FIGS. 10-17. Mature larva of Ammobates carinatus. 10, 1 1. Entire larva, lateral and dorsal views, respectively. 12-14. Right mandible, dorsal, inner and ventral views, respectively. 15. Spiracle, side view. 16. Head, frontal view. 17. Head, lateral view. Scale refers to figures 10, 11. 10 AMERICAN MUSEUM NOVITATES teeth or serrations; apical concavity moderately considerably anterior to apparent posterior mar- well developed and adoral in position. Maxillary gin of head; posterior thickening of head capsule palpi moderately large and elongate. Labial pal- virtually absent being indicated by faint line; pus indicated by single sensillum. Hypopharynx apparent posterior thickening present but indefi- well defined and strongly projecting, greatly ex- nite; area between thickenings not so sclerotized ceeding maxillae and labium; hypopharyngeal as head capsule; hypostomal ridge moderately groove weakly defined. Salivary opening nearly well developed (thinner than that of Oreopasites) circular; surrounding rim absent or perhaps only and darkly pigmented, bending dorsally (fig. 24) poorly developed. just anterior to posterior tentorial pit; pleuro- Body (figs. 10, 11). Integument smooth, only stomal ridge poorly developed and indistinct; epi- moderately shiny, venter faintly spiculate. Form stomal ridge virtually absent; coronal cleavage moderately broad in dorsal view (fig. 11); inter- line present as a diffused light line extending an- segmental lines at most only shallowly incised; teriorly to clypeal region. Antennal protuber- these lines not incised laterally, hence, lateral ances absent. Labrum strongly projecting; labral outline of body smooth as seen in dorsal view; tubercles small, narrowly rounded, well separated paired dorsal tubercles absent; lateral swellings by about two basal diameters of tubercle; epi- below spiracles well defined although not so well pharynx only slightly swollen. Mandible (figs. developed and conspicuous as those ofMorgania; 20-22) slender, broad basally as seen in adoral dorsal intrasegmental lines apparently absent; ab- view and evenly tapering to apex; cuspal area not dominal segment X slightly dorsal in attachment produced, possibly indicated by very slight and to segment IX; anus apical. Spiracles (fig. 15) gradual adoral swelling; dorsal and ventral inner uniform in size, not an elevation; atrium subglob- edges well formed and serrate; apical concavity ular; atrial wall without denticles; peritreme not developed. Maxillary palpi small, short. Labi- moderately wide; subatrium moderately long. al palpus indicated by single sensillum. Hypo- Imaginal disc of male genitalia transverse oval, pharynx poorly defined but exceeding maxillae with cuticular scar, situated medially near poste- and labium; hypopharyngeal groove virtually ab- rior margin of sternum IX: discs of female un- sent. Salivary opening nearly circular and sur- known. rounded by low, projecting rim. Material Studied. One postdefecating larva, Body (figs. 18, 19). Integument smooth or 10 km. south of Skhirate, near Rabat, Morocco, distinctly crinkled, venter faintly spiculate. Form April 29, 1968 (J. G. Rozen), from cell of small extremely broad in dorsal or ventral aspect, Eucera. Adult determined by Rozen. sharply curved anteriorly; intersegmental lines deeply incised; intersegmental lines deeply incised laterally so that lateral outline appears Morgania histrio transvaalensis Bischoff strongly ridged when viewed dorsally (fig. 19); Figures 18-25 paired dorsal tubercles absent although most Diagnosis. The combination of deeply incised body segments produced somewhat dorso- intersegmental lines of the body as seen in dorsal laterally; dorsal intrasegmental lines indicated on view (fig. 19), wide separation of labral tubercles most body segments as faint lateral lines, not ex- (about two basal diameters of tubercle apart) tending dorsally, hence, segments not divided (fig. 23), and subquadrate spiracular atria (fig. into cephalic and caudal annulets; segment X 25) on pigmented and slightly sclerotized spi- centrally attached to segment IX; anus apical. racular elevations distinguishes this species from Spiracles (fig. 25) uniform in size, on slight eleva- other known larval ammobatines. tions which are moderately pigmented and scler- Head (figs. 23, 24). As seen in frontal view otized apically; atrium subquadrate; atrial wall (fig. 23), head distinctly wider than long; median with numerous rows of small denticles; peri- swelling of vertex absent. Anterior tentorial arms treme moderately wide; subatrium moderately virtually absent; posterior arms very short and long with approximately eight wide annulations. slender; anterior pits obscure; posterior pits Imaginal disc of male genitalia situated medially small, situated well below hypostomal ridge and immediately in front of ventral intersegmental ROZEN AND MCGINLEY: AMMOBATINE BEES I1I

CUSPAL AREA

0 APPARENT POSTERIOR THICKENING

JIHYPOSTOMAL RIDGE

HYPOPHARYNX FIGS. 18-25. Mature larva of Morgania histrio transvaalensis. 18, 19. Entire larva, lateral and dorsal views, respectively. 20-22. Right mandible, dorsal, inner and ventral views, respectively. 23. Head, frontal view. 24. Head, lateral view. 25. Spiracle, side view. Scale refers to figures 18, 19. line separating abdominal segments IX and X; Material Studied. Two postdefecating larvae, cuticular scar transverse linear, slightly more pig- 3 miles west of Grahamstown, Cape Province, mented than surrounding integument, and almost South Africa, November 28, 1966, from cells of on intersegmental line; characteristics of female Tetralonia (J. G. Rozen). Adults determined by unknown. Rozen. 12 AMERICAN MUSEUM NOVITATES

The biology of this species was discussed by unevenly to apex; dorsal and ventral inner edges Rozen (1969). serrate (serrations possibly absent in one species); apical concavity not developed. Maxillary palpi moderately small, tending to be slightly elongate OREOPASITES COCKERELL but still much less developed than those of The following account is based on larvae of Ammobates. Labial palpus indicated by single several undescribed species as well as of speci- sensillum. Hypopharynx poorly developed al- mens nominally assigned to Oreopasites van- though exceeding maxillae and labium; hypo- duzeei Cockerell, a name that apparently encom- pharyngeal groove weakly indicated. Salivary passes a species complex. opening elongate-oval, not surrounded by low Diagnosis. The moderately wide body, as seen rim. in dorsal view, distinguishes the mature larvae of Body (Rozen, 1954, fig. 1). Integument of Oreopasites from those ofMorgania histrio trans- postdefecating forms finely wrinkled, dull to vaalensis and Pseudodichroa fumipennis. The moderately shiny; venter strongly spiculate. moderately well-incised intersegmental lines of Form moderately broad in dorsal view; interseg- the abdomen, the poorly developed hypophar- mental lines moderately well incised, hence lat- ynx and the small maxillary palpi further distin- eral outline of body, as seen in dorsal view, guish larval Oreopasites from Ammobates carina- ridged (not so pronounced as in Morgania); tus. Unlike other known larval ammobatines, paired dorsal tubercles absent although most Oreopasites usually has moderately large, pointed body segments produced dorsolaterally; dorsal labral tubercles, although some forms of 0. van- intrasegmental lines not evident; abdominal seg- duzeei do have small labral tubercles. ment X central in attachment to segment IX; Head (Rozen, 1954, figs. 5, 6). As seen in anus apical. Spiracles (Rozen, 1954, fig. 2) uni- frontal view, head nearly circular; vertex without form in size, not on distinct elevations; atrium produced, rounded swelling. Anterior and poste- subglobular; atrial wall with rows of small denti- rior tentorial arms very short and slender; anterior cles; peritreme moderately wide; subatrium and posterior pits minute; posterior pit below moderately long to long. Female with paired hypostomal ridge and anterior to apparent pos- imaginal discs on segments VII, VIII, and IX terior margin of head; posterior thickening of (best visible on predefecating forms); discs on head capsule virtually absent, indicated by faint segment IX near midpoint of segment and ap- line; apparent posterior thickening present but proximately one disc diameter apart; discs on indefinite; area between thickenings sclerotized; segments VII and VIII in posterior half of seg- hypostomal ridge moderately developed and ment with those on segment VIII about three darkly pigmented anterior to posterior pits disc diameters apart and those of segment VII (somewhat more slender and lightly pigmented in approximately five disc diameters apart; male one species); posteriad of pits, ridge becomes sexual characters not observed. very thin or virtually absent, curving dorsally, or Material Studied. Oreopasites vanduzeei, I 1 only faintly so; pleurostomal ridge moderately postdefecating larvae, Skeleton Canyon, Arizona, developed and darkly pigmented to indistinct September 3, 1966, from nests of Nomadopsis and lightly pigmented; epistomal ridge virtually puellae (J. Bath, F. Andrews, L. LaPr6). Oreopa- absent or weakly developed and extending dor- sites vanduzeei, one postdefecating and one pre- sally to level of antennae; coronal cleavage line defecating larva, Tuolumne, Tuolumne County, not visible. Antennal protuberances absent. Lab- California, 1973, from a nest of Nomadopsis an- rum moderately projecting; labral tubercles mod- thidia anthidia (Fowler) (J. G. Rozen). Oreopa- erately large (small in some forms of 0. van- sites sp., one postdefecating larva, Fort Robin- duzeei) and pointed, narrowly separated to sepa- son, Dawes County, Nebraska, August 10, 1972, rated by at least two basal diameters of tubercle; from cell of Nomadopsis helianthi (J. G. Rozen epipharynx only slightly swollen. Mandible and R. J. McGinley). Oreopasites species, one (Rozen, 1966, figs. 56-58) slender, broad basally postdefecating and two predefecating larvae, 13 as seen in adoral view and tapering somewhat miles southwest of Apache, Arizona, September ROZEN AND MCGINLEY: AMMOBATINE BEES 13

1, 1971, from cells of Perdita (J. G. Rozen and Michener, 1954, indicated that tubercles were M. Favreau). present on his specimen. PRELIMINARY KEY TO PUPAE PUPAE OF SOME NOMADINE GENERA Although mature larvae of bees have been sub- 1. Apical lateral angle of clypeus with down- jected to considerable evolutionary and system- ward directed tubercle (fig. 29). Ammo- atic study recently, pupae are poorly known. Of batini ...... 2 the Nomadinae, only the pupae of Holcopasites Apical lateral angle of clypeus without (as Neopasites, Michener, 1954) and Isepeolus tubercle ...... 3 (Michener, 1957) have been briefly described and 2(1). Terminal metasomal spine moderately long compared with other bee pupae. Now, however, (fig. 26); mesoscutellum with spined tu- the pupae of a number of genera of nomadines bercles (figs. 26, 28) .... . Oreopasites are preserved in the collections held by the Terminal metasomal tubercle very small and short (figs. 32, 33); mesoscutellum American Museum of Natural History, making without tubercles .Morgania possible the following remarks and key. They in- 3(1). Axillae distinctly produced; if indistinctly clude Isepeolus (one species), Protepeolus (one produced (Protepeolus), then mesoscu- species), Oreopasites (two species), Morgania tum with pair of large, strongly elevated (one species), Nomada (two species), Parano- rounded tubercles. 5 mada (one species), Epeolus (one species), and Axillae inconspicuous, not produced; mes- Neolarra (one species). In addition, a poorly pre- oscutum without tubercles or with nu- served specimen of Odyneropsis and a very poor merous, small, sharp-pointed tubercles. specimen ofHolcopasites are also on hand. Nomadini ...... 4 The pupae of these genera are diverse so that 4(3). Tegula with strongly projecting tubercle characterization of the subfamily is not possible, ...... Paranomada Tegula without tubercle...... Nomada at least at this time; no single feature or set of 5(3). Most metasomal sterna each bearing apical features can be presented to distinguish the No- row of distinct tubercles which may or madinae from other anthophorids or apoids. All may not bear setae. Epeolini...... 6 have rows of small tubercles on the apical mar- Metasomal sterna without distinct tubercles gins of most metasomal terga (e.g., figs. 26, 28, ...... 7 30, 32) as is characteristic of most bee pupae. 6(5). Vertex, scutum, and scutellum with nu- These tubercles are usually sharp-pointed, merous small sharp-pointed spined tu- bearing sclerotized tips that at least in some in- bercles ...... Epeolus stances represent setae. These setae are normally Vertex, scutum, and scutellum without tubercles...... Odyneropsis short and spinelike but in some instances are 7(5). Mesoscutum without rounded tubercles. quite long. Similar spined tubercles are found on Neolarrini ...... Neolarra the vertex, scutum, scutellum, and outer surfaces Mesoscutum with large rounded paired tu- of the tibiae of many but not all noinadines, a bercles. Protopeolini ...... 8 situation that is not widespread among other 8(7). Mesoscutum with single pair of rounded bees. Furthermore, pupae of many of the no- tubercles; mesoscutellum and axillae madine genera have large rounded paired tuber- only faintly produced .... Protepeolus cles arising from the mesoscutellum, but large Mesoscutum with two pairs of rounded tu- tubercles tend to be absent from other parts of bercles; mesoscutellum with pair of large, elevated tubercles; axillae strongly the body, perhaps an indication that the adults produced. tend to have short setae. Hence tubercles are Isepeolus usually absent from coxae, trochanters, femora, A pair of tubercles at the apicolateral angles pronota, and scuta and only Paranomada has a of the clypeus (fig. 29) distinguishes the pupae of tubercle projecting from each tegula. The poorly Morgania and Oreopasites from those of the preserved specimen of Holcopasites lacks tuber- other Nomadinae keyed above. These structures cles on the basal segments of the leg although allow for the development of specialized hair 14 AMERICAN MUSEUM NOVITATES

FIGS. 26-28. Pupa of Oreopasites vanduzeei. 26. Female pupa, lateral view. 27. Left mandible, lateral view. 28. Anterior part of pupa, dorsal view. FIGS. 29-33. Pupa of Morgania histrio transvaalensis. 29. Head, frontal view. 30. Anterior part of pupa, dorsal view. 31. Left mandible, lateral view. 32. Male pupa, lateral view. 33. Apex of metasoma of female pupa, lateral view. Scales refer to figures 26, 28, and 30, 32, 33, respectively. tufts in adult males. Although tufts are absent or are characteristic of all the Ammobatini, the cor- greatly reduced in females, the tubercles are pres- responding tubercles are probably present on all ent on female pupae as well. Because the tufts ammobatine pupae. ROZEN AND MCGINLEY: AMMOBATINE BEES 15

Oreopasites vanduzeei Cockerell ably II-VI in males) with distinct row of moder- Figures 26-28 ately large, spined tubercles, tubercles on tergum Oreopasites vanduzeei is the only named spe- V few in number; tergum VI without tubercles cies for which pupae are known. However, a (in undescribed species of Oreopasites tubercles pupa of an undescribed species is also on hand; on tergum I virtually absent; large strongly pro- some distinct features of it are parenthetically jecting spined tubercles present in distinct rows included in the following description. on terga II-V of females and lI-VI of males; fe- Diagnosis. The pupae of Oreopasites and Mor- male with a few small tubercles present on ter- gania histrio transvaalensis can be distinguished gum VI); sternal tubercles absent; terminal tuber- on the basis of differences in the number and cle moderately elongate, rounded apically. distribution of spined tubercles and in the length Material Studied. One female, Skeleton Can- of the terminal metasomal tubercles. yon, Arizona, September 3, 1966, from cell of Head. Integument microscopically spiculate; Nomadopsis puellae (J. Bath, F. Andrews, L. setae absent. Scape and frons without tubercle; LaPre). vertex just mesiad of compound eye with very small, spined tubercles (in undescribed species these tubercles large and conspicuous); genal tu- Morgania histrio transvaalensis Bischoff bercles absent; each apicolateral angle of clypeus Figures 29-33 with downward-directed and narrowly rounded Diagnosis. See diagnosis of Oreopasites van- tubercle (as in fig. 29); mandible (fig. 27) with duzeei. large tubercle on ventral surface; dorsal surface Head. As discussed for Oreopasites vanduzeei evenly curved, without tubercles or swellings. except for the following: vertex just mesiad of Mesosoma. Integument microscopically spicu- compound eye with minute spine-tipped tuber- late; setae absent. Lateral angles of pronotum cles that are much smaller than those of Oreopa- broadly rounded; posterior lobe of pronotum in- sites vanduzeei; mandible (fig. 31) with low, distinctly produced; mesepisternum without tu- broad swelling on ventral surface and on dorsal bercles; mesoscutum with very small spined tu- surface. bercles (in undescribed species these tubercles Mesosoma. As discussed for Oreopasites van- and their setae larger and conspicuous); axilla duzeei except for the following: Mesoscutum rounded, slightly produced; scutellum with low, without tubercles; axilla not produced; scutellum broadly rounded tubercles (not so strongly pro- with pair of large rounded, moderately well-produced as those of Morgania), bearing small, duced tubercles (perhaps larger in females); these spined tubercles (these tubercles quite pro- tubercles without small spined tubercles as found nounced in undescribed species); metanotum in Oreopasites but with small indistinct swellings without tubercles but with rather distinct, trans- that may be homologous. Wing without distinct verse lateral swellings; propodeum without pro- tubercle although forewing bearing low, medial tuberances. Tegula without tubercles; wing with- swelling. Forecoxae, unlike those of Oreopasites, out tubercle or swelling. Coxa, trochanter, and with low, rounded tubercles; outer apical sur- femur all without tubercles; foretibia without tu- faces of each tibia with one (or sometimes more) bercles (undescribed species with approximately spined swellings. four small spined tubercles on outer surface); Metasoma. Integument microscopically spicu- apex of midtibia with several indistinct spined late; setae absent. Tergum I of male without tu- tubercles (undescribed species with a number of bercles, of female with several small tubercles; spined tubercles on outer surface); hind tibia terga II-V (female) and II-VI (male) with distinct with several minute, indistinct spined tubercles rows of moderately large, spined tubercles; ter- on outer surface (undescribed species with nu- gum VI of female with few tubercles; tergum VII merous distinct spined tubercles). of male without tubercles; sternal tubercles ab- Metasoma. Integument microscopically spicu- sent; terminal tubercle very short and rounded late; setae absent. Tergum I with indistinct row but distinct, in both sexes. of small spined tubercles; terga II-V (and presum- Material Studied. One male, one female, 3 16 AMERICAN MUSEUM NOVITATES miles west of Grahamstown, Cape Province, dae). Trudy Zool. Inst., vol. 9, pp. South Africa, November 28, 1966, from cells of 895-949, figs. 1-19, tables 1, 2. Tetralonia minuta (J. G. Rozen). Rozen, Jerome G., Jr. 1954. Morphological description of the larva of Oreopasites vanduzeei Cockerell LITERATURE CITED (Hymenoptera: Anthophoridae). Pan- Pacific Ent., vol. 30, pp. 203-207, figs. Baker, D. B. 1-6. 1971. A new Pasitomachthes from Rhodesia 1966. The larvae of the Anthophoridae (Hy- (Hymenoptera, Apoidea). Novos Taxa menoptera, Apoidea). Part 2. The No- Ent., no. 98, pp. 1-8, figs. 1-6. madinae. Amer. Mus. Novitates, no. Michener, Charles D. 2244, pp. 1-38, figs. 1-83. 1944. Comparative external morphology, 1968. Review of the South African cuckoo- phylogeny, and a classification of the bee genus Pseudodichroa (Hymenop- bees (Hymenoptera). Bull. Amer. Mus. tera, Apoidea). Ibid., no. 2347, pp. Nat. Hist., vol. 82, art. 6, pp. 151-326, 1-10, figs. 1-10. figs. 1-246, diagrams 1-13. 1969. Biological notes on the bee Tetralonia 1953. Comparative morphological and sys- minuta and its cleptoparasite, Morgania tematic studies of bee larvae with a key histrio transvaalensis (Hymenoptera: to the families of hymenopterous lar- Anthophoridae). Proc. Ent. Soc. Wash- vae. Univ. Kansas Sci. Bull., vol. 35, pt. ington, vol. 71, no. 1, pp. 102-107, figs. 2, no. 8, pp. 987-1102, figs. 1-287, 1-5. table 1. Rozen, Jerome G., Jr., and Ronald J. McGinley 1954. Observations on the pupae of bees (Hy- 1974. Phylogeny and systematics of Melit- menoptera: Apoidea). Pan-Pacific Ent., tidae based on the mature larvae (In- vol. 30, pp. 63-70, fig. 1, table 1. secta, Hymenoptera, Apoidea). Amer. 1957. Notes on the biology of a parasitic bee, Mus. Novitates. Isepeolus viperinus (Hymenoptera, An- Rozen, Jerome G., Jr., and Charles D. Michener thophorinae). Ent. News, vol. 68, no. 6, 1968. The biology of Scrapter and its cuckoo pp. 141-146, figs. 1-5. bee Pseudodichroa (Hymenoptera: Col- Popov, V. V. letidae and Anthophoridae). Amer. 1951. The parasitic bees of the genus Ammo- Mus. Novitates, no. 2335, pp. 1-13, bates Latr. (Hymenoptera, Anthophori- figs. 1-12, tables 1, 2. A4 0