AMERICAN MUSEUM

Nov,tates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2887, pp. 1-1 3, figs. 1-20 July 22, 1987

Nesting Biology and Immature Stages of a New Species in the Genus (: : Melittidae: Dasypodinae)

JEROME G. ROZEN, JR.'

ABSTRACT This paper describes Hesperapis (Hesperapis) described. The food source, Nama hispidum, is trochanterata Snelling, new species, from southern identified, and special adult features used in for- Arizona, and provides data on various aspects of aging and nest excavation are discussed. A brief its nesting biology, including: description of nest note is also presented on the nesting behavior of site, nest architecture, provisioning, and devel- females ofH. (Carinapis) rhodocerata (Cockerell). opment of immatures. Mature larva and pupa are INTRODUCTION This paper presents information on the also northern and central Mexico. It is closely nesting biology and floral preferences ofHes- related to Capicola (southern ) and perapis trochanterata, n. sp., described herein Xeralictoides (western United States) (Mich- by R. R. Snelling. I also illustrate, describe, ener, 1981). A number of papers have been and compare the mature larva with larvae of published on the natural history of one or other known species in the subfamily Dasy- another ofits species, including: Hurd (1957) podinae (Rozen and McGinley, 1974). In ad- (unusual seasonal occurrence of H. fulvipes dition, I present the first published descrip- Crawford); Linsley and MacSwain (1958) tion of any pupa of the Dasypodinae. (sleeping habits of males); Burdick and Tor- The genus Hesperapis is restricted to xeric chio (1959) (nesting biology and floral rela- western North America (except for one species tionships of H. regularis (Cresson)); Mac- in Florida) and is especially abundant in the Swain et al. (1973) (floral relationships, southwestern United States and undoubtedly seasonal activity, mating, nesting biology of

' Curator, Department of Entomology, American Museum of Natural History.

Copyright © American Museum of Natural History 1987 ISSN 0003-0082 / Price $1.70 2 AMERICAN MUSEUM NOVITATES NO. 2887

r lkl. V. .1 .W* II I I.k -:1 .-a,i I.-I .4 C. 1.4. U'6 q I

I.., im-l- I A - ..;. I- 06 .t4 Is -i6..0 ... 1. v0 %?'..- YA. 'k 41-Al" -e 14 w4v -,a Fig. 1. Nesting site of Hesperapis trochanterata. Most nests' entrances were in the barren sand at the periphery of the clearing of the Pogonomermyx nest in center.

H. regularis); and Hurd and Linsley (1975) ACKNOWLEDGMENTS (floral relationships). Rozen (1974) briefly described the nesting biology ofthe southern I thank G. I. Stage and R. R. Snelling for African Capicola braunsiana Friese (Dasy- the opportunity to study their manuscript on podini) and Haplomelitta ogilviei (Cockerell) Hesperapis. The following persons critically (Sambini). reviewed this manuscript: Ronald J. Mc- The mature larvae ofHesperapis have been Ginley, Smithsonian Institution; R. R. Snell- described by Michener (1953) (H. rufipes ing, Los Angeles County Museum ofNatural (Ashmead)), Burdick and Torchio (1959) (H. History; and G. I. Stage, Stafford Springs, regularis), and Rozen and McGinley (1974) Conn. I acknowledge their assistance with (eight species of Hesperapis as well as Capi- gratitude. cola braunsiana). Although there is at present no compre- NATURAL HISTORY hensive account of the biology of the genus, NESTING AREA: This species flew in mod- G. I. Stage undertook a major study of the erate numbers on May 8, 1986, at 4 mi,east nesting biology ofa number ofspecies ofHes- of Willcox, Cochise County, Ariz., several perapis as a part of his Ph.D. thesis. The hundred meters from where nests of Marti- manuscript, currently being revised and ex- napis and Caupolicana had been discovered panded by Stage and Snelling (MS), will in- in August 1985 (Rozen and Rozen, 1986). deed provide a broad understanding of the The area (fig. 1) consisted ofpartly tied down natural history of the genus. The description sand dunes, derived from the Willcox playa ofthe biology ofHesperapis trochanterata has several miles away. Vegetation stabilized the been arranged in this paper to be comparative fine sand except for scattered blowouts re- with information they present. sulting from grazing cattle and the clearing 1987 ROZEN: NESTING BIOLOGY OF HESPERAPIS 3 activities ofharvester (Pogonomyrmex).

Mesquite (Prosopis), the dominant tree, was r .i <*'p,,i> g t* ;uutL4b,%,'e%-ds,,,'$. scattered and much ofthe ground surface was exposed to the sun except for low growing .s,$sh.'M W r; flA,, ,',:Y b*., ^ shrubs and such herbaceous plants as Eri- ogonum, Euphorbia, and Nama hispidum A. : --.5E=4f.... Gray, the pollen plant ofthe bee (kindly iden- tified by James Grimes ofthe New York Bo- tanical Garden). Nama grew abundantly in patches within a radius of several hundred meters of the nesting areas and was visited by Conanthalictus as well as Hesperapis tro- chanterata. I identified a number of nest entrances in a barren, sandy area on May 8, 1986, and excavated them on May 15 and 16 at which time I found entrances in adjoining open areas on the edges ofPogonomyrmex clearings (fig. 1). Subsequent visits to the region in August and early September 1986 revealed activity II in these same areas, and I discovered an ad- ditional site about a mile away. Fig. 2. Hesperapis trochanterata, nest entrance NESTS: Four nest groups were discovered showing characteristic tumulus. in May, all on surfaces of loose, wind-blown sand that were gently sloping (10-200), un- tumulus, generally 2.5-3.0 cm in diameter, shaded, and nearly barren. was low, cratered at the entrance, and grad- At each site the nest entrances were irreg- ually merged with the substrate surface. In ularly distributed and could usually be iden- contrast, nest entrances of other species of tified by their size and by the characteristic consisted ofmounds in sharp reliefwhere tumulus (described below) of darkened sand females piled excavated material just outside (when fresh). Although I excavated only four the entrance. or five active nests (i.e., nests with fresh At the one nest group excavated, the open tumuli), other main burrows became evident burrows (fig. 3), 3-4 mm in diameter, ex- once I had removed the surface layer (3-4 tended obliquely through this upper loose cm thick) of soft sand to expose the more sandy layer at about a 20-30O angle from hor- consolidated substrate of fine sand below. It izontal and were straight to curved. In the is unclear whether these other burrows were consolidated soil below, they descended ver- nests from previous generations, more recent tically in a twisting fashion. Open above, they nests from which the tumuli had been blown branched, with each meandering ramus de- away, or both. However, cells (both recently scending at approximately a 450 angle. The constructed and containing quiescent, post- rami tended to be clogged with loose soil that defecating larvae) subtended them. For every could be blown away with an aspirator at active nest identified on the surface, three to upper levels. However, at lower levels the five burrows were hidden below in the com- substrate became faintly moist as did the fill pact sand, within a radius of 30 cm of the ofthe branches, with a result that I could not fresh nest. Certainly the tumuli of fine, dry remove the fill by blowing and I lost the tun- sand, as well as the entrance holes, would be nels. Although not carefully noted, burrow ephemeral landmarks if the females did not diameters did not seem to be narrower at regularly reestablish them. lower levels than at upper ones, as reported Entrances (fig. 2), which penetrated the sur- for some species of Hesperapis by Stage and face at a low angle, were semicircular (rather Snelling (Ms). than circular) in shape, and the tumuli oc- At various points along the tunnel system curred mostly on the downhill sides. Each arose small chambers, perhaps 5-10 mm long, 4 AMERICAN MUSEUM NOVITATES NO. 2887

T U M U LUS 4.0 mm

APPROXIMATE POSITION OF CELL CLOSURE

U lO.00cm Figs. 3, 4. Hesperapis trochanterata, nest components. 3. Side view of upper section of nest. 4. Cell containing postdefecating larva and feces.

as wide as or slightly wider than the tunnel uncoated, nonwaterproof when tested with itself, and rounded at the blind end. The func- water droplets, and of the same hardness as tion ofthe cul-de-sacs is uncertain in view of the substrate. Simple cavities excavated from the fact that only one female occupied a nest. the substrate, they had a rough texture and Neither these chambers, the main tunnel, nor bore the faint, but characteristic, impressions its branches were lined and waterproofed, and of annular ridges perpendicular to the cell's their walls were generally rough. long axis. A rough wall, unlined either with I encountered cells infrequently in the up- secretions or with material transported from per levels ofnests, the most shallow one being the outside, is uncommon among ground- 9.5 cm deep and two others 13 cm deep. nesting bees, but is known for some other However, most cells ofboth active nests and Hesperapis (Stage and Snelling, MS). nests containing postdefecating larvae oc- Two cells measured approximately 8 and curred at a depth ofapproximately 23-27 cm 9 mm in length and two, 4.0 and 4.5 mm in (N = 8). All were arranged singly, as is char- maximum diameter. The entrance to one was acteristic of the genus, and were nearly hor- 3.0 mm in diameter with the lateral filled izontal, with the front end slightly higher than with loose soil that was blown away. The the rear. In the related Capicola, cells are configuration of the cell closure remains un- arranged in linear series and in Haplomelitta, known. singly (Rozen, 1974). Because tunnels leading PROVISIONING: Females of Hesperapis tro- to lower cells could not be traced, I had to chanterata possess abundant, long setae on chip through large quantities of substrate to the stipes and prementum that are probably chance upon them. The several cells at shal- used to pull pollen from the tubular flowers low depths connected to the main tunnel by of Nama. The stipital setae are wavy to al- descending laterals that curved gradually hor- most kinky apically, and are nearly identical izontally. to long, wavy setae that border the probos- Cells (fig. 4) were elongate ovals with a cidial fossa laterally and posteriorly. The pre- round rear end. Their walls were completely mental setae tend to be shorter, curved api- 1987 ROZEN: NESTING BIOLOGY OF HESPERAPIS 5 cally, and directed somewhat posteriorly. The parent integumentlike material flaked from head ofboth sexes is elongate and has a pro- the white larva beneath. This material evenly jecting clypeus-presumably adaptations that coated the entire integument and showed all permit the bee to gather food from this plant of the contour features of the body wall. It by reaching deeply into the flowers. Males do was fairly rigid, whereas the true integument not have the specialized long wavy or curved beneath was more pliable. Spiracular open- setae on the mouthparts. ings were not coated. I have noticed similar Females transported provisions as moist coatings on larvae of some other noncocoon pollen attached to the anterior surface of the spinning bees (for example, certain noma- hind tibiae. The elongate curved dorsal setae dines). The coating dissolved when boiled in characteristic ofthe hind tibiae ofthis species a solution of potassium hydroxide, an indi- were not involved with the provisions. Fe- cation that it is not chitinous. Its source is males molded the yellow provisions into unknown but it is apparently laid down (at spheres 2.1-2.4 mm in diameter (N = 5) and least in the Nomadinae) at about the time the placed them on the floor of the cells. The larva enters diapause, perhaps as a protection spheres were homogeneously moist, very firm, against desiccation or parasites. and uncoated by waterproofmaterial, in con- Each pupa (fig. 13) was extremely active in trast to the coated spherical food mass of that it rotated its metasoma by flexing it at certain panurgines. One incomplete food the base so that the tip of the metasoma cir- mass, a small ball, was also spherical but more cumscribed a circle. When placed in an ex- uneven than the completed mass. cavated cell, a pupa used the apical meta- DEVELOPMENT: I uncovered a single, somal spine as a posterior pivot against the strongly curved, shiny, translucent white egg rear of the cell and the pronounced vertical (maximum diameter 0.45 mm; length 1.55 tubercles and labral tubercles (see description mm) on top of the provisions in the median of pupa) as anterior pivots against the front sagittal plane of the cell. Its anterior end was cell wall. When the midpart of the dorsum distinctly wider than the more tapering pos- arched upward as a result ofmetasomal flex- terior end. Although I saw several small feed- ing, the dorsal metasomal tergal tubercles ing larvae, they died and desiccated before I came in contact with the cell wall, and when had a chance to study their behavior. Post- the midpart of the body was arched down- defecating larvae (fig. 4), totally quiescent, ward, the sternal metasomal tubercles scraped rested on their dorsa with their heads pointed the cell wall. Of course, in a sealed cell in the toward the cell exits. Of five postdefecating ground the metasoma is not arched just up- larvae kept live, two pupated before May 31 ward and downward, but in all directions as and the rest were preserved. The pupa de- a result ofits rotating action. The result must veloped rapidly, as is characteristic ofall bees, be that the top, bottom, and sides of the wall and were preserved for anatomical descrip- are contacted by the gyrating metasoma. The tion. function of this rotation is not understood, Feces (fig. 4), deposited in a patch in the but would seem to be significant when one upper rear of the cell, consisted of elongate considers the number of special anatomical pellets running generally parallel to the sag- features involved in performing the motion. ittal plane ofthe cell. These bees did not pro- Possibly the metasomal tergal and sternal tu- duce cocoons, and the labiomaxillary region bercles with their pigmented sclerotized tips of the mature larva was greatly recessed, as scrape parasites and predators (such as mites) is characteristic of all Dasypodinae (Rozen against the cell wall. Alternatively, the tergal and McGinley, 1974). and sternal tubercles may merely assist in The integument ofthe postdefecating larva rotating the body in the cell, their sclerotized ofthis species was peculiar in that falling fine tips providing traction against the cell wall. sand from my excavation readily adhered to Rotation of the abdominal areas of pupae of it even though it did not appear moist. When holometabolous is, of course, wide- I removed sand from preserved specimens spread. by gently scraping with forceps, not only the ADULT BEHAVIOR AND SPECIAL FEATURES: sand but patches of thin, tannish, semitrans- Adults were active during midday, first flying 6 AMERICAN MUSEUM NOVITATES NO. 2887 in midmorning. Males commonly flew swift- in sandy situations. Reduction (Neofidelia and ly over the low-growing food plant. They sud- Fidelia) or loss (Hesperapis trochanterata, denly landed on the ground and then just as Parafidelia, and some Fidelia) of basatibial suddenly departed, darting from one clump plates on the hind legs of females may also ofNama to the next. Although I saw no mat- be involved with sand nesting, but H. rho- ings, copulation almost certainly took place docerata possesses well-developed basitibial either at the flowers or at the base of the plates. flower plants. Males did not search for fe- I found two females of Hesperapis tro- males in the nesting areas. chanterata to possess three ovarioles in each At nesting areas I saw females land and ovary (ovariole formula 3:3), as is character- start to enter the substrate by scratching with istic of other Melittidae (Rozen, 1986). their forelegs, only to depart quickly as ifthey I have collected this Hesperapis in the vi- did not find the area suitable. Several females cinity of Willcox, Arizona, in both May and were observed excavating tunnels by digging again in mid-August to mid-September not with their forelegs and flinging the excavated only in 1986, but in previous years. The data sand backwards with the hind legs in such a seem to indicate that the species has at least rapid motion that the hind legs seemed to two generations a year, but the site needs blur. The resulting shower of sand created repeated observations through the growing the tumulus described above. season to provide a clear understanding of The hind legs ofthe females possessed long, voltinism and seasonal activity. curved, thick, simple dark setae that un- PARASITISM: No cuckoo bees visited the doubtedly assisted in sand flinging. Espe- nesting area and no larvae or oth- cially, these setae on the hind basitarsus ar- er nest associates or parasites were found in ranged in two even rows created a trough that the cells. cupped the sand to be flung similar to the adaptations of the hind legs of the unrelated Neofidelia (Rozen, 1973). In Hesperapis tro- DESCRIPTION OF chanterata the setae of one row are directed POSTDEFECATING LARVA posteriorly, those of the other row, dorsally. Figures 5-1 1 In Neofidelia, the hairs in each row are di- DIAGNOSIS: This is a small species (length rectly obliquely dorsally so that the setae of of postdefecating larva 5-6 mm), apparently the two rows diverge apically. closely related to Hesperapis species b (Rozen At 2 mi southeast of Willcox, I observed and McGinley, 1974) as judged by charac- nearly identical sand flinging on the part of teristics of adults (and supported by features nesting females of Hesperapis (Carinapis) of the larvae). Its larva can be distinguished rhodocerata (Cockerell), midday, on Septem- from all known larvae of Hesperapis includ- ber 5, 1986. These females also exhibit the ing Hesperapis species b on the basis of the troughlike arrangements of setae on the hind following combination: Epistomal ridge in- basitarsus, as described for H. trochanterata. complete, fading abruptly medially; area im- Since specimens of some species of Hesper- mediately behind junction of hypostomal apis possess these modifications and others ridge and posterior thickening of head cap- do not, one can test the hypothesis that the sule not greatly swollen; anterior part ofhead troughlike setal arrangement is an adaptation capsule as seen in lateral view (fig. 8) evenly for sand flinging by learning how species curved, not depressed; dorsal inner edge of without the modifications discharge soil from mandible (figs. 9-11) with only a few large nests. teeth in addition to a number ofmuch small- Not only are there similarities between er teeth. The larva of this species can further Neofidelia, Hesperapis trochanterata, and H. be distinguished from that of Hesperapis rhodocerata with respect to modifications of species b (to which it bears a strong resem- the hind tarsi involved with sand flinging, but blance) by its distinctly longer, tapering, also the taxa share enlarged pygidial plates. mandibular apex (figs. 9-11). As pointed out These anatomical similarities may be asso- by Rozen and McGinley (1974), all known ciated functionally in some way with nesting Hesperapis larvae can be distinguished from 1987 ROZEN: NESTING BIOLOGY OF HESPERAPIS 7

12 FEMOAL I LABRAL TUBERCLE TUBERCLE TROCHANTER 1, TUBERCLE COX 111, TUBERCLE Figs. 5-13. Hesperapis trochanterata, immatures. 5. Postdefecating larva, live, lateral view. 6. Perianal area of larva. 7, 8. Head of postdefecating larva, frontal and lateral views. 9-11. Right mandible of postdefecating larva, dorsal, adoral, and ventral views. 12. Head of pupa, frontal view. 13. Pupa, lateral view. Scale refers to figures 5 and 13.

those of Capicola because the former lack erell (Rozen and McGinley, 1974) except for paired dorsal tubercles. Interestingly, the in- the following: Integument strongly crinkled trasegmental lines on abdominal sterna IX so that sensilla and spiculation pattern dif- and X may be absent in H. trochanterata, as ficult to detect. Epistomal ridge well devel- is also the case for Capicola. oped but briefly and completely interrupted HEAD (figs. 7, 8): As described for genus medially; ridge more arched than in Hesper- Hesperapis and Hesperapis pellucida Cock- apis pellucida but not quite reaching level of 8 AMERICAN MUSEUM NOVITATES NO. 2887 antennae. Antennal papilla not strongly pro- the pupa of Hesperapis trochanterata im- duced. Apical section of labrum scarcely mediately distinguish it from the pupae of emarginate if at all. Mandible similar to that , , and Meganomia, all of ofHesperapis species b: cusp moderately well which lack such structures. The paired apical defined and very few teeth extending apically labral tubercles with their elongate darkly along dorsal adoral surface; dorsal apical edge pigmented setiform apices and the peculiar with only a few, enlarged teeth; ventral apical bidentate dorsal metasomal tubercles are also edge without teeth; apex of mandible in gen- unique characteristics of this species. Infor- eral longer and more attenuate than that of mation given by Stage and Snelling (Ms) in- Hesperapis species b. dicate that the pupal features of Hesperapis BODY: As described for the genus Hesper- trochanterata do not apply completely to sev- apis (Rozen and McGinley, 1974) except for eral other species of Hesperapis. the following: Integument of postdefecating HEAD: Integument without setae or spic- forms strongly wrinkled, mostly nonspicu- ules. Vertex (figs. 12, 13) with three elongate, late. Form (fig. 5) moderately slender. Ab- sharp-pointed, apically pigmented tubercles, dominal segment IX with venter essentially each arising from position of ocellus; each normal length, not greatly elongate as in some tubercle with slight irregular constriction at Hesperapis; abdominal sternum IX appar- base. Antennae without tubercles. Genal tu- ently without intrasegmental line; abdominal bercles present behind mandibular base. La- sternum X either without intrasegmental line brum (fig. 12) bearing two apical tubercles, or at least with line obscured in postdefecat- each of which bears elongate setalike, darkly ing form; segment X perhaps somewhat elon- pigmented tapering projections as well as one gate; perianal area (fig. 6) nearly circular and or two subapical somewhat shorter setalike puckered. Spiracular elevations not pro- projections. Mandible with conspicuous ven- nounced. Male with median cuticular scar tral tubercle. Maxilla swollen at base to ac- toward rear ofabdominal sternum IX; female commodate stipital setae of adult. sexual characters unknown. MESOSOMA: Integument without setae or MATERIAL STUDIED: Four postdefecating spicules. Lateral angles ofpronotum not pro- larvae, 4 mi E Willcox, Cochise County, Ariz., duced; posterior lobe of pronotum moder- 16 May 1986 (J. G. Rozen). ately produced and apically acute; mesepi- sternum without tubercles; mesoscutum without tubercles; axilla not produced and DESCRIPTION OF PUPA scarcely discernible; scutellum at most Figures 12, 13 vaguely produced on each side; metanotum The following is believed to be the first slightly produced medially; propodeum with- formal description ofa pupa for any member out tubercles. Tegulae and wings lacking tu- of the Dasypodinae, although Stage and bercles. Fore- and midcoxae each with small Snelling (Ms) give accounts oftwo other Hes- apical tubercle; hind coxa with minute apical perapis pupae. In regard to other melittids, tubercle; each trochanter with moderately Rozen and McGinley (1974) described the small apical tubercle; forefemur with basal pupae ofMelitta leporina (Panzer) and Mac- tubercle; hind tibia with elongate apical tu- ropis europaea Warncke (); Rozen bercle on outer edge; legs of male and female and Jacobson (1980) found the pupa ofMac- approximately same in shape and configu- ropis nuda (Provancher) to be identical to ration in spite of modified hind legs of adult that ofMacropis europaea. Rozen (1977) fig- male. ured and described the pupa of Meganomia METASOMA: Integument without setae. gigas Michener (as M. binghami) (Mega- Terga I-VI (male) and I-V (female) each with nomiinae). subapical row of moderately small, apically DIAGNOSIS: The pattern, size, and shape of pigmented tubercles; larger tubercles curving pupal tubercles ofthe known representatives posteriorly and each bearing subapical tooth; ofthe Melittidae appear distinctive from one smaller ones apically simple; terga without subfamily to the next. The conspicuous ver- median around tubercles as found in Mega- tical tubercles and elongate terminal spine of nomia; sterna IV-V (male) and Ill-V (fe- 1987 ROZEN: NESTING BIOLOGY OF HESPERAPIS 9

17

19

TROCHANTER 15 16 Figs. 14-20. Hesperapis trochanterata, adult male. 14-16. Metasomal sterna 6-8. 17, 18. Genital capsule, ventral and lateral views. 19. Inner view of apex of gonostylus. 20. Metatrochanter and meta- femur. Scale = 0.50 mm, except figure 19 = 0.25 mm. male) each with apical row of small tubercles Stage and R. R. Snelling (Ms), in order that with simple, darkly pigmented, sharp-point- Rozen may describe the nesting biology and ed apices. Terminal spine greatly elongate, immature stages. Stage had seen a few spec- sharp pointed, bearing pigmented apex. imens of this species from Texas (and rec- MINERAL STUDIED: One male pupa, 4 mi ognized that it was new), but did not include E Willcox, Cochise County, Ariz., collected it in his thesis (Stage, 1966). as postdefecating quiescent larva, 16 May The following description is patterned after 1986, pupated 25 May 1986, preserved 28 those in the Hesperapis revision in order to May 1986 (J. G. Rozen); one female pupa, facilitate future comparisons. same data except pupated 30 May 1986, pre- DIAGNOSIS: The male differs from those of served 2 June 1986. all others in the nominate subgenus (sensu Michener, 1981)-including known, unde- APPENDIX scribed species, as well as those of all other subgenera-by the swollen pro- and meta- Description of femora (fig. 20) and especially by the presence Hesperapis (Hesperapis) trochanterata, of a coniform tubercle on the ventral surface new species of the metatrochanter. Within the subgenus Hesperapis, the female may be separated from BY R. R. SNELLING H. elegantula Cockerell and one undescribed Figures 14-20 species by the longer than broad head (length This species is described in advance of a no more than 0.90 times width in those forthcoming revision of Hesperapis, by G. I. species); from the one remaining species (un- 10 AMERICAN MUSEUM NOVITATES NO. 2887 described, from Nevada) the female of H. shortest side of first segment): 6:6:6.5:8; me- trochanterata may be separated by the uni- dian flagellar segments about three-fourths as formly convex clypeal disc, the completely long as wide. Labrum about twice as wide as roughened and dull propodeal triangle, and long, convex, smooth, and glabrous, but api- the dark, long simple hairs of the metatibia. cal margin depressed and with preapical row (In the Nevada species, the clypeal disc is of short, simple, sparse, yellowish hairs, and distinctly impressed along the midline, the marginal fringe of longer, simple, yellowish propodeal triangle is roughened only in the hairs. Preapical tooth of mandible at about middle and very narrowly along the anterior distal two-thirds of mandible length. Galea margin, and the long, simple hairs ofthe hind brown, dorsal surface evenly and conspicu- legs are yellowish.) ously shagreened, slightly shiny; outer mar- FEMALE: Measurements (mm). Head width gin strongly curved at apex and forming near- 1.45; head length 1.52; body length 4-6. ly right angle with straight inner edge; outer Head: Integument shiny black except for edge with fringe of short, straight hairs and light brownish labrum and reddish mandi- a few longer hairs at apex, inner margin with- bles; underside of flagellum yellowish red. out fringe. Labial palpus 1.75 times length of Vestiture entirely silvery whitish, except very glossa, about 0.56 times length of premen- long hairs in ocellar area slightly brownish; tum, about 1.3 times length ofmaxillary pal- vestiture almost entirely appressed and dense, pus; ratio oflengths oflabial palpal segments: nearly concealing integument, thinnest on 16:16:10:11. upper frons and around ocelli; most of clyp- Mesosoma: Integument shiny, dark brown eus and lower three-fourths of gena without to blackish. Appressed vestiture whitish, very appressed hairs; antennal scape, lower frons, short, and more or less obscuring integument ocellar area, and lower three-fourths of gena on pronotum and mesoscutum; longer and with conspicuous long, fully erect, sparse slightly sparser on side of mesosoma, con- hairs. cealing integument across much of metano- Head length 1.00-1.05 times width; vertex tum; appressed hairs ofscutellum very short, in frontal view strongly convex, but slightly but conspicuously sparser than those of flattened behind ocelli; inner eye margins mesoscutum. Long, erect, minutely barbed weakly convergent below, upper interocular hairs ofmesoscutum and scutellum very light distance about 1.1 times lower interocular brownish yellow, with some distinctly distance; interocular distance (measured at brownish near posterior margin ofscutellum; level of lower margin of antennal sockets) those of mesepisternum longer, more abun- about 3.0 times width ofeye; eye length about dant, whitish. 0.71 times head length. Clypeus strongly pro- Mesoscutal and scutellar punctures sub- tuberant and convex, without longitudinal contiguous, mostly about 0.01 mm in di- median impression; punctures ofclypeus mi- ameter, but with scattered larger (about 0.02 nute (0.015 mm in diameter or less), sepa- mm) punctures; mesepisternal punctures rated by 0.50-1.50 times a puncture dia- dense, with more punctures up to 0.02 mm; meter. Supraclypeal prominence largely metepisternum similar, but punctures mostly impunctate; middle offrons subcontiguously about 0.01 mm. Propodeal triangle finely punctate, punctures coarser than those of shagreened and slightly shiny, most ofdorsal clypeus, becoming increasingly finer and face finely irregularly rugulose; posterior face sparser toward eye margins, finer and sparse with dense punctures about 0.02 mm; side on vertocciput, but both areas with scattered with obscure, extremely fine punctures. coarser punctures from which arise long, fully Wings entirely clear, veins and stigma me- erect hairs; punctures of lower three-fourths dium brown; jugal lobe about 0.70 times of gena sparse, coarser than on vertocciput, length of vannal lobe. round to oval. Antennal scape, pedicel, first Legs dark brown. Mesotibial outer vesti- flagellar segment, and dorsal part of rest of ture mostly silvery white, uniformly plu- flagellum dark brownish to blackish; ratio of mose, but anteroapically with scattered am- first four flagellar segments (measured on ber spiniform setae; basitarsal vestiture 1987 ROZEN: NESTING BIOLOGY OF HESPERAPIS I1I similar but less dense, without spiniform se- of glossa, about 0.5 times length of premen- tae and with dense posterior fringe of long, tum, and about 1.5 times length of maxillary plumose hairs; mesotibial spurs strongly bent palpus; ratio of lengths of labial palpal seg- at apex. Metabasitibial plate not visible; pri- ments: 19:17:11:12. mary, appressed vestiture of scopa white and Mesosoma: Integument color and sculp- obscuring surface; secondary, erect vestiture ture as described for female. Vestiture as de- brownish, long, especially along posterior scribed for female, but short, fine, plumose margin of basitarsus. hairs of scutellum almost as dense as those Metasoma: Integument shiny, bright red- of mesoscutum. dish, but with small, dark lateral spots on Wings as described for female. terga 2-4, and large, irregular light brownish Vestiture of legs about as described for fe- areas at base of sterna 2-5. Terga 1-5 with male, but spiniform setae of mesotibia fewer apical pubescent fasciae of whitish, ap- and less conspicuous, mesobasitarsus with pressed, plumose hairs; fasciae slightly longer only sparse, long hairs on posterior margin; at sides; discs of these terga with very short, long hairs fringing posterior margin of suberect, simple, yellowish hairs; tergum 1 metatibia and metabasitarsus in two rows di- with long, erect, minutely barbed, white hairs rected obliquely away from each other (i.e., across basal face and terga 1-5 with similar, analogous to setae on metabasitarsus of fe- but fewer, hairs at each side; terga 2-4 each male) surface between these rows smooth and with single row of long, suberect, simple, polished; fringe hairs dusky, but vestiture brownish hairs arising along base of apical otherwise pale. fascia, dark hairs well separated from one Profemur, in dorsal view, only about twice another and not reaching to apical margin of as long as thick; in anterior view, about twice segment; long erect hairs oftergum 5, and all as long as deep, distal two-thirds of ventral hairs of 6 fuscous. Punctures minute and surface obliquely flattened or slightly con- dense on discs ofterga. Apex ofpygidial plate cave. Metatrochanter with prominent, blunt, narrowly rounded to substruncate; disc more or less conoid, ventral process at apex; smooth and shiny; narrow, preapical median metafemur, in dorsal view, about twice as sulcus present but weak; area on either side long as thick; in anterior view, about twice slightly depressed and weakly reticulate. as long as deep (fig. 20). MALE: Measurements (mm). Head width Metasoma: Integument shiny, dorsally 1.45-1.52; head length 1.45-1.48; body length black or very dark brown, ventral segments 5-6. dark to medium brown. Vestiture about as Head: Integument shiny black; labrum described for female; terga 1-5 with distinct translucent yellowish; tricolored mandible apical fasciae of appressed, plumose, white black, yellowish, red. Vestiture as described hairs. Punctation as in female. Sterna 3-5 for female. each with a mediobasal, triangular swelling, Head length equal to, or slightly less than, its apex directed distad; surface of swollen width; inner eye margins more strongly con- area smooth and shiny, without vestiture. vergent below, upper interocular distance Apical sterna and genitalia as illustrated (figs. about 1.2 times lower interocular distance; 14-19). interocular distance about 3.25 times width TYPE MATERIAL: Holotype female and al- of eye; eye length about 0.67 times head lotype: 4 mi E Willcox, Cochise Co., ARIZONA, length. Clypeus slightly more protuberant 8 May 1986 (J. G. Rozen), in American Mu- than that offemale, otherwise similar, except seum of Natural History. Paratypes (all from that apical one-third has only scattered punc- Cochise Co., ARIZONA): 9Q, 46, same data as tures. Sculpture of head as in female. Ratio holotype; 49, 16, same data except 16 May of lengths of first four flagellar segments: 1986; 76, same data except 17 May 1986; 49, 5:6.5:11:12; middle flagellar segments slight- same data except 5 Sept. 1986 (J. G. and B. ly longer than broad. Labrum as in female L. Rozen); 119, 46, same data except 22 Aug. except pale translucent yellow. Maxilla as in 1986 (R. R. Snelling), on N. hispidum; 19, female. Labial palpus about 1.3 times length same data except 23 Aug. 1985 (J. G. and B. 12 AMERICAN MUSEUM NOVITATES NO. 2887

L. Rozen; 1 , same data except 28 Aug. 1985; vada species the profemur, in dorsal view, is 46, Willcox, 19 May 1985 (J. G. Rozen), on about three times as long as thick and the Nama hallii; 26, Willcox, 19 Aug. 1985 (J. metafemur is a little more than four times as G. and B. L. Rozen), on Nama sp.; 29, same long as thick. Tergum 6 ofthe Nevada species data except 19 Aug. 1983, no host; 16, 5Q, 4 is distinctly pubescent-fasciate, but it is not mi E Willcox, 24 Aug. 1985 (J. G. and B. L. in H. trochanterata. The glabrous basal swell- Rozen), on Nama sp; 19, Willcox, 21 Sept. ings at the base of metasomal sterna 3-5 are 1976 (J. G. Rozen); 16, same data except 16 also unique to this species. May 1985; 16, same data except 18 Aug. 1985, on Nama sp. Paratypes in American Museum ofNatural History and Natural History Mu- seum of Los Angeles County. REFERENCES CITED ETYMOLOGY: The specific name refers to Burdick, D. J., and P. F. Torchio the modified male metatrochanter. 1959. Notes on the biology ofHesperapis reg- DISCUSSION: I have also examined three fe- ularis (Cresson) (Hymenoptera: Melit- males and one male in the Museum of Com- tidae). J. Kans. Entomol. Soc., 32: 83- parative Zoology from 3 mi E Presidio, Pre- 87. Hurd, P. D., Jr. sidio Co., TEXAS, collected by H. E. Evans, 1957. Notes on the autumnal emergence ofthe 23 Apr. 1963. These are the specimens rec- vernal desert bee, Hesperapis fulvipes ognized by G. I. Stage as representing an un- Crawford (Hymenoptera, Apoidea). J. described species. Kans. Entomol. Soc., 30: 10. There is very little variation in the speci- Hurd, P. D., Jr., and E. G. Linsley mens available, other than slight differences 1975. The principal Larrea bees of the South- in size and color as noted in the description western United States (Hymenoptera: above. Apoidea). Smithson. Contrib. Zool., This species is most obviously similar to 193: 74 pp. Nevada Females of Linsley, E. G., and J. W. MacSwain an undescribed species. 1958. Sleeping habits of males of Hesperapis that species are apparent oligoleges on Col- (Hymenoptera: Melittidae). J. Kans. denia. The principal differences between H. Entomol. Soc., 31: 273-274. trochanterata and the Nevada species are in- MacSwain, J. W., P. H. Raven, and R. W. Thorp dicated in the diagnosis above. In addition, 1973. Comparative behavior of bees and On- the appressed vestiture ofthe side ofthe me- agraceae. IV. Clarkia bees of the west- sosoma is much denser in the Nevada species, ern United States, vol. 70, 80 pp. Univ. and almost completely conceals the surface California Press. beneath. Similarly, the fine appressed hairs Michener, C. D. of the scopa of the hind legs cover the entire 1953. Comparative morphological and sys- of the tibia and ex- tematic studies of bee larvae with a key outer surface basitarsus, to the families of hymenopterous lar- cept along the extreme anterior margin. In vae. Univ. Kansas Sci. Bull., 55: 987- females ofH. trochanterata, the anterior one- 1102. third of the outer face of those segments is 1981. Classification of the bee family Melit- without appressed hairs. Finally, females of tidae with a review of species of the Nevada species possess a short carina at Meganomiinae. Contrib. Am. Entomol. the base of the metatibia, the remnant of the Inst., 18(3): 135 pp. posterior margin of the basitibial plate. The Rozen, J. G., Jr. basitibial plate remnant appears to be wholly 1973. Life history and immature stages of the absent in H. trochanterata, or at least it is bee Neofidelia (Hymenoptera, Fideli- the dense idae). Am. Mus. Novitates, 2519: 14 pp. completely concealed by primary 1974. The biology oftwo African melittid bees vestiture of the scopa. (Hymenoptera, Apoidea). J. New York As does the female, the male of the H. Entomol. Soc., 32: 6-13. trochanterata most closely resembles that of 1977. Biology and immature stages of the bee the Nevada species, from which it is imme- genus Meganomia (Hymenoptera, Me- diately separable by the structure of the legs, littidae). Am. Mus. Novitates, 2630: 14 especially of the metatrochanter. In the Ne- PP. 1987 ROZEN: NESTING BIOLOGY OF HESPERAPIS 13

1986. Survey of the number of ovarioles in Rozen, J. G., Jr., and B. L. Rozen various taxa of bees (Hymenoptera: 1986. Bionomics of crepuscular bees associ- Apoidea). Proc. Entomol. Soc. Wash- ated with the plant Psorothamnus sco- ington, 88: 707-710. parius (Hymenoptera: Apoidea). J. New Rozen, J. G., Jr., and Ned R. Jacobson York Entomol. Soc., 94: 472-479. 1980. Biology and immature stages ofMacro- Stage, G. I. pis nuda, including comparisons to re- 1966. Biology and systematics of the Ameri- lated bees (Apoidea, Melittidae). Am. can species of the genus Hesperapis Mus. Novitates, 2702: 11 pp. Cockerell. Unpubl. Ph.D. thesis, Univ. Rozen, J. G., Jr., and R. J. McGinley California, Berkeley, 461 pp. 1974. Phylogeny and systematics of Melitti- Stage, G. I., and R. R. Snelling dae based on the mature larvae (Insecta, Ms. A revision of the Nearctic Melittidae: Hymenoptera, Apoidea). Am. Mus. the subfamily Dasypodinae (Hymenop- Novitates, 2545: 31 pp. tera: Apoidea).

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates, Bulletin, and Anthropological Papers published during the last five years are available free of charge. Address orders to: American Museum of Natural History Library, Department D, Central Park West at 79th St., New York, New York 10024.