Diagnosis

Mature larvae of Dufourea can immediately be distinguished from those of Conanthalictus because the latter have a recessed labiomaxillary region, the salivary opening lacks projecting lips, and the prothoracic dorsal tubercles are absent. The labiomaxillary region of Dufourea is forward projecting with pronounced salivary lips, and the prothorax bears a distinct pair of dorsal tubercles. Unlike all other known rophitines except for Sphecodosoma, mature larvae of Dufourea lack a well-defined articulating arm of the stipes. The large, rounded antennal papilla of Sphecodosoma separates that genus from Dufourea, all known mature larvae of which have smaller antennal papillae, with the sensilla tightly clustered at the apex of the papilla. The characters presented in the key should distinguish the species of Dufourea whose larvae were available for this study. These larvae, however, are very similar one to another.

Diagnosis

The combination of normally broad salivary lips (fig. 3) and moderately large antennal disc (fig. 3) distinguish the mature larva of Dufourea holocyanea from those of D. mulleri, which has a small antennal disc (fig. 28), and D. australis australis, which has narrow salivary lips (fig. 10).

Head (Figs. 3, 4)

Integument of head capsule with scattered small sensilla that are not obviously setiform; those of mouthparts with small setae; spiculation as described below. Integument unpigmented except for mandibular apices and anterior and posterior mandibular articulations.

Head (figs. 1, 2) moderately small compared with body size; head capsule wider than length measured from top of vertex to lower clypeal margin in frontal view. Tentorium complete, including dorsal arms. Anterior tentorial pit close to anterior mandibular articulation (fig. 3); posterior tentorial pits in normal position; postoccipital ridge moderately developed, becoming weaker toward vertex; median longitudinal thickening of head capsule nearly absent; hypostomal ridge moderately developed, its posterior section dividing into dorsally curving ramus that fuses with postoccipital ridge while ventral ramus bending mesad to join postoccipital ridge at posterior tentorial pit (best seen in posteroventral view of cleared head capsule); pleurostomal ridge well developed; epistomal ridge laterad of anterior tentorial pit short because of closeness of pit to anterior mandibular articulation; ridge between pits fading out before reaching midline. Parietal bands evident as integumental scars. Antennal prominence rather weak; antennal disc moderately large, its diameter only slightly less than distance from its lower margin to center of anterior tentorial pit (latter distance measured in maximum profile); papilla projecting moderately, its height no more than one-half basal diameter, each bearing 3 sensilla (figs. 6, 16). Vertex evenly rounded, moderately broad, as seen from side, without projections. Labrum not projecting anteriorly as far as clypeus in lateral view (fig. 4), moderately short in frontal view (figs. 3, 12), apically shallowly bilobed when viewed anterodorsally; labrum spiculate laterally.

Mandible in adoral view tapering to narrowly pointed apex; dorsal apical edge with sharp teeth; ventral apical edge with evenly spaced, sharp teeth; cusp moderately produced in dorsal and ventral views (figs. 7, 9), with numerous teeth; apical concavity shallow, facing hypopharyngeal surface; outer mandibular surface with cluster of three or four conspicuously projecting tubercles, each bearing single short sensillum; apical dorsal surface with series of parallel ridges (some apparently with fine denticles) running in same direction as cuspal teeth; ventral surface with transverse declivity extending from outer tubercles toward cups, but this declivity not as pronounced or long as that of Xeralictus timberlakei and apparently Protodufourea eickworti. Labiomaxillary region strongly projecting in lateral view (fig. 4) with labium projecting well beyond apex of maxilla. Maxilla with dorsal surface spiculate; cardo, stipes evident but not pronounced; articulating arm of stipital sclerite not evident; maxillary palpus moderately large, longer than basal diameter, arising from maxillary apex. Labium strongly divided into prementum and postmentum, moderately long; premental sclerite not evident; labial palpus subequal in size to maxillary palpus. Salivary lips projecting, transverse, moderately narrow in frontal view (figs. 3, 13); as seen with SEM, upper and lower lips with apical fringe of papillae (figs. 14, 15); lower lip with internal surface consisting of parallel longitudinal ridges; upper lip with internal surface without such ridges but with scattered papillae. Hypopharynx protuberant, spiculate toward mouth, not longitudinally grooved; hypopharyngeal groove absent, so that hypopharynx curving onto labial apex.

Body

Integument with fine scattered nonsetiform sensilla and with spiculate areas laterally behind head and toward rear of abdomen; some dorsal body tubercles faintly wrinkled apically, those of abdominal segment 3 and 4 somewhat flattened apically on postdefecating (but not predefecating) larva (see “Discussion of Relationships”). Body form (figs. 1, 2) moderately robust; intersegmental lines moderately weakly incised on predefecating form; on postdefecating form ventral intersegmental lines, particularly of abdomen, deeply incised; paired prothoracic dorsal tubercles present, smaller than those of mesothorax, metathorax, and most abdominal segments (figs. 1, 2); all tubercles conical (not transverse); abdominal segment 9 strongly produced ventrally; abdominal segment 10 positioned dorsally on 9 as seen in lateral view (figs. 1, 2); on predefecating larva (fig. 2), ventral integument of segments 9 extended, causing segment 10 to be partly tilted and making anus appear dorsal on 10; on postdefecating larva (fig. 1), ventral integument of segment 9 strongly retracted. Spiracles small, those of thorax slightly larger than those of abdomen, those of predefecating larva without sclerites, not on tubercles, those of postdefecating larva on low tubercles composed of presumably thick chitin surrounding atrium (these tubercles so low as to be easily overlooked on specimens that are not stained); peritreme (figs. 5, 17) present, narrow; atrium projecting slightly beyond wall, with rim, globose; atrial wall smooth; primary tracheal opening with collar; subatrium normal in length, consisting of about 12 chambers. Male with median, transverse, integumental scar ventrally near apex of protrusion of segment 9 of predefecating larva but closer to posterior margin of segment of postdefecating larva; female sexual characters visible only on some predefecating larva as pair of small, pale, subcutaneous imaginal discs arranged ventrally on each of abdominal segments 7, 8, and 9, with those of 7 farthest apart and those of 9 closest.

Material Studied

3 postdefecating, 6 predefecating larvae, USA: Washington: Kittitas Co.: 21 km N of Yakima, between Yakima River and Highway 821 at 10.5 mi mark, June 2000 (E. Miliczky); 4 last larval instars, same except 15 km S of Ellensburg, along Umtanum Creek, 2 km W Yakima River, July 1996.

Remarks

Eugene Miliczky is preparing a manuscript on the nesting biology of this species, specimens of which he kindly donated for this study.

The mature larva of this species is known only from the postdefecating form.

Diagnosis

Although there are several subtle differences distinguishing the postdefecating larva of this species from that of Dufourea holocyanea (as indicated in the following description) and D. mulleri, the extremely narrow salivary lips (fig. 10), of D. a. australis are the most reliable and easily used diagnostic feature.

Head (Fig. 10)

Integument as described for Dufourea holocyanea.

Head (figs. 10, 11, 18, 19) size and width, and tentorium as described for Dufourea holocyanea. Postoccipital ridge well developed, not becoming weaker toward vertex, in contrast to D. holocyanea and D. mulleri; other internal head ridges as described for D. holocyanea. Parietal bands evident as integumental scars. Antennal features as described for D. holocyanea except papilla (fig. 11) more elongate, its length nearly equal to basal diameter. Vertex and labrum as described for D. holocyanea.

Mandible similar to that of Dufourea holocyanea except tubercles on outer mandibular surface smaller, less conspicuous. Labiomaxillary region as described for D. holocyanea including absence of articulating arm of stipital sclerite and premental sclerite. Salivary lips projecting, transverse, extremely narrow in frontal view (figs. 10, 18), so that their apical width no greater that distance from their base to base of labial palpus. Hypopharynx as described for D. holocyanea.

Body

Integument with fine, scattered, nonsetiform sensilla and with finely spiculate areas laterally behind head, toward rear of abdomen, and on many dorsal body tubercles; some dorsal body tubercles faintly wrinkled apically and somewhat flattened apically, suggesting apices of tubercles subject to deformation when larva resting on them (but see “Remarks” in treatment of Protodufourea eickworti). Body form, intersegmental lines, and dorsal body tubercles as described for postdefecating larva of D. holocyanea. Spiracles small, subequal in size, otherwise similar to those D. holocyanea except subatrial chambers may be numerous. Male with median, transverse integumental scar ventrally close to posterior margin of segment 9 of postdefecating larva; female sexual characters unknown.

Material Studied

2 postdefecating larvae, USA: Arizona Pima County: Desert Station, E. side of Tucson Mts., April 28, 1993 (J.G. Rozen); 12 postdefecating larvae, same except May 1, 1993; 7 postdefecating larvae, same except May 10, 1993.

Remarks

JGR discovered nests of Dufourea a. australis on the steeply sloping (but not vertical) bank of a ravine at Desert Station in 1993. He visited the site on April 28, May 1, and May 10, when the following brief observations were recorded. These observations are augmented with others made during the preparation of the current paper on cells preserved in 1993 and retained in the collection of the AMNH.

One main tunnel and presumably others were open and 4 mm in diameter. Numerous cells were scattered about 15 to 20 cm deep into the bank. When first studied on April 28, they consisted of vacated cells, representing an earlier generation, and cells occupied by feeding and diapausing larvae of both D. a. australis and its cleptoparasite, Neopasites cressoni Crawford (Apidae: Nomadinae: Biastini). When examined on May 10, all larvae were postdefecating, an indication that both D. a. australis and N. cressoni are univoltine. Cell walls (figs. 21, 22) were thick, often 0.5 mm or more on sides and rear, harder than the substrate, and could be extracted intact from the ground as small hard nuggets. Female D. a. australis presumably impregnated the wall with a hardening substance, possibly nectar. When submerged in water, a piece of cell wall started to disassemble immediately and completely fell apart in a minute, as would be expected if the hardening substance were nectar.5 The inner surface of the wall (fig. 21) was dull, even, but with small projecting soil particles, and was unlined, contrary to the reflective cell linings of most bees. When the cell wall disassembled in water, the inner surface may have had slightly more cohesion, possibly resulting from compression of very fine particles by some action (tamping?) by the female. Cell shape, like that of other rophitines, was symmetrical around its long axis, with floor and ceiling equally curved in side view, and wide compared with its length, measure from rear to center of closure (maximum diameter 5.0–5.5 mm, length 6.0–6.5 mm, N  =  8). Cell closures (fig. 22) were even spirals of four coils to the radius, concave on the inside, with a diameter of 2.5–3.0 mm, and on the outside smoothly concave with an even surface. The closures were about 0.5–0.8 mm thick in the middle where the outer and inner surfaces approach one another but much thicker on the periphery.

Completed cocoons of Dufourea a. australis (its cleptoparasite does not spin cocoons) were very thin and adhered tightly to the cell wall, so that the cocoon fabric could not be pulled away from the wall. The cocoon completely filled the cell lumen; its shape corresponded to the internal shape of the cell. The inner surface was slightly shiny, thereby contrasting with the dull cell wall before cocoon construction. In some cells, the cocoon was semitransparent, so that feces and tan substrate indistinctly showed through. In these cases, the cocoon gradually became opaque and paler toward the front end, and at the closure, its inner surface turning white (fig. 23). In other cells the cocoons were more uniformly pale and opaque. The cocoon fabric was a single layer; individual silk strands were usually not clearly identifiable under stereomicroscopic examination. Under SEM examination, white fabric at the closure end seemed to consist of extremely fine, crisscrossing silken strands as well as some coarser strands (fig. 20). In many cells uniformly shaped, elongate, blackish objects (figs. 20, 23), about 0.4 mm long, were scattered on the inner surface of whitish silk. Their flattened shapes suggested that they had desiccated. Although their uniformity of shape suggested that they might have been eggs, they showed no discernable microstructure under high SEM examination (fig. 20). Their identity and source are unknown.

In cells examined before cocoon construction, feces were deposited longitudinally as elongate flattened yellowish stripes6 from the rear of the cell partway toward the closure end, so that the rear 2/3s to 3/4s of the cell wall was evenly coated (fig. 24) but the remainder was uncoated. However, apparently before the feces are so deposited, the larva deposits a thin, brownish, glistening, semitransparent, coating over the anterior end of the cell, with the thickest part covering the cell closure (fig. 25). The source of the material, whether anal or salivary, is unknown, but it is visible as a nearly black mass at the center of the cell closure when the closure particles are removed from the outside, either mechanically or by submersion in water. After the cocoon is spun, this dark material and even the feces toward the rear of the cell are hidden by the cocoon fabric.

Diapausing larvae were invariably found curled in their cocoons with the abdominal apex touching the venter of the head and prothorax, as depicted for Rophites canus (fig. 50).

Vacated cells indicate that fresh adults emerge by chewing through the front end of the cocoon and cell closure.

On each visit, the ratio of recovered host larvae to cleptoparasite larvae was: April 28, 6:8; May 1, 19:9; May 10, 24:6; grand total 49:23. These ratios seem to suggest that Neopasites cressoni may be much more successful finding and parasitizing host cells earlier rather than later in the nesting season.

McGinley (1981) listed some of the important features of the mature larva of Dufourea mulleri and later (McGinley, 1987) presented illustrations of the larva. The following description is based only on predefecating larvae.

Diagnosis

The normally broad salivary lips and small antennal disc, as presented in the key, will best separate the larvae of this species from those of other known Dufourea.

Head (Fig. 28)

Integument as described for Dufourea holocyanea.

Head size and width as described for Dufourea holocyanea. Tentorium and internal head ridges as described for D. holocyanea (including weakening postoccipital ridge toward vertex). Parietal bands and antennal features as described for D. holocyanea except antennal disc small, its diameter almost one-half times distance from it lower margin to center of anterior tentorial pit (latter distance measured in maximum profile); antennal papilla (fig. 30) projecting moderately strongly, its height more than one-half basal diameter. Vertex and labrum as described for D. holocyanea.

Mandible as described for D. holocyanea, except outer mandibular surface with tubercles less pronounced (fig. 29) but clearly present, as in all known Rophitinae, contrary to McGinley (1987: 693, fig. 12.187c). Labiomaxillary region, salivary lips, and hypopharynx as described for D. holocyanea.

Body

Integument as described for Dufourea holocyanea except dorsal body tubercles of postdefecating larva unknown. Body form of predefecating larva (fig. 27) as described for D. holocyanea. Spiracles of predefecating larvae as described for those of D. holocyanea except subatrium slightly longer, consisting of approximately 15 chambers. Male and female sex characters unknown.

Material Studied

12 predefecating larvae, USA: Arizona: Cochise County: 3 mi S of Rodeo, New Mexico, V-4-1966 (J.G. Rozen, M.S. Favreau).

Remarks

Torchio et al. (1967) described the nesting biology of this species (note especially their fig. 5 showing typical rophitine diapausing posture) and that of the following species whose mature larvae were not collected: D. malacothricis Timberlake, D. pulchricornis (Cockerell), and D. trochantera Bohart.

Description

Eickwort et al. (1986) presented an account of the biology and mature larva of this species and compared the larva with others that had been described previously. Although we have been unable to locate the specimens they studied, most of the larval features described and depicted correspond closely with cocoon-spinning rophitine larvae treated here. They stated that the hypopharyngeal groove was present, but their illustration (their fig. 11) does not reveal a groove immediately above the salivary lips where the groove should be. Rather it shows a transverse line much farther toward the mouth, so that we suspect that they misidentified the true groove, which normally is a simple folding of the integument associated with the articulating arms of the maxillae. In their comparisons of larvae, they state that larvae of Dufourea novaeangliae, D. mulleri, Rophites canus, and Systropha punjabensis Batra and Michener are “only slightly lower than the other thoracic tubercles”, contrasting with a statement that the tubercles of these genera were “markedly smaller than the following” tubercles (Rozen and McGinley, 1976). Although we have not been able to examine larvae of D. novaeangliae, larvae of the other cocoon-spinning rophitines examined for the current study have prothoracic tubercles that are noticeably smaller than the other thoracic tubercles, as indicated in figs. 1, 2, 27, 31, and 50. It is inappropriate to compare the description and illustrations of the larva of S. punjabensis (Batra and Michener, 1966: figs. 13–15) with that of the others since that was not a last larval instar.

Remarks

An interesting discovery made by Eickwort et al. (1986) was the presence of a thin, clear, waterproof coating on the provisions of Dufourea novaeangliae, subsequently also noted on the provisions of Sphecodosoma dicksoni (Rozen, 1993). Such covering of provisions had heretofore been observed only among certain Panurginae (Andrenidae) (Rozen, 1958, 1967).

Diagnosis

The large, rounded antennal papilla (Rozen, 1993: fig. 28) in combination with the actually or near absent articulating arm of the stipes distinguishes this species from all known cocoon-spinning larvae of the Rophitinae.

Description

See Rozen (1993) in which the mature larvae are described and illustrated.

Remarks

The nesting of this species was briefly described by Rozen and McGinley (1976), and Rozen (1993) subsequently gave a more complete account of its nesting biology.

Diagnosis

The postdefecating larva of this species and that of Xeralictus timberlakei can be distinguished from those of other known genera in the Rophitinae by pronounced truncate dorsal tubercles on the midbody segments and the ventral transverse declivities on their mandibles. They are quite similar to one another even though Protodufourea eickworti is considerably smaller, has a far broader head when seen in frontal view, and its antennal disc is larger than that of X. timberlakei. Many larval features of these two genera are similar to those of the Old World Systropha planidens.

Head (Figs. 32, 33)

Integument of head as described for Dufourea holocyanea.

Head features as described for Dufourea holocyanea except for following: Postoccipital ridge well developed its entire length; median longitudinal thickening well developed but short; epistomal ridge between anterior tentorial pits weak, scarcely visible, fading out before reaching midline. Diameter of antennal disc about 0.8 times distance from its lower margin to center of anterior tentorial pit; antennal papilla with 3–4 sensilla. Labrum apically broadly rounded when viewed anterodorsally; labral tubercles scarcely evident.

Mandible (figs. 34, 35) in general similar to, and as described for, that of Dufourea holocyanea, except for following: dorsal apical edge distal to dentate cusp apparently without teeth; ventral apical edge with extremely fine teeth; tubercles on outer mandibular surface not strongly projecting, without setiform sensilla, but on dorsal and ventral surfaces giving rise to long, curved declivity that fades before reaching inner mandibular edge, this declivity similar to, and presumably homologous with, that found in Xeralictus timberlakei (Snelling and Stage, 1995: figs. 16, 17). Labiomaxillary region moderately projecting in lateral view. Maxilla with dorsal surface spiculate; cardo, stipes evident; articulating arm of stipital sclerite pronounced even if not strongly pigmented; maxillary palpus moderately large, longer than basal diameter, arising from maxillary apex. Labium strongly divided into prementum and postmentum, moderately long; premental sclerite not evident; labial palpus distinctly smaller than maxillary palpus. Salivary lips projecting, transverse, moderately narrow in frontal view (fig. 32). Hypopharynx protuberant, spiculate toward mouth, without longitudinal groove; hypopharyngeal groove created by articulating arms of stipes.

Body

Integument with fine, scattered, nonsetiform sensilla. Body form moderately robust; on postdefecating form ventral intersegmental lines particularly of abdomen deeply incised; paired prothoracic dorsal tubercles present, smaller than those of mesothorax, metathorax, and most abdominal segments (figs. 31, 36, 37); all tubercles conical (not transverse); paired dorsal body tubercles of meso- and metathorax and of abdominal segments 1–8 tending to be subtruncate to truncate, their flattened surface finely wrinkled, faintly convex to concave (see “Remarks”, below, and “Discussion of Relationships”; abdominal segment 9 strongly produced ventrally; abdominal segment 10 positioned dorsally on 9 as seen in lateral view (fig. 31). Spiracles small, subequal in size; peritreme present, narrow; atrium projecting slightly beyond wall, with rim, globose; atrial wall smooth; primary tracheal opening with collar; subatrium normal in length, consisting of about 12 chambers. Male with median, transverse, integumental scar near posterior margin of segment; female sexual characters unknown.

Material Studied

1 postdefecating larva, USA: Arizona: Pima County: Desert Station, E side of Tucson Mts., March 30, 1995 (J.G. and B.L. Rozen); 1 postdefecating larva, same except collected March 30, 1995, preserved July 8, 2003; 1 postdefecating larva, same except collected March 29, 1995, preserved April 12, 1995.

Remarks

Rozen et al. (1997) described the nesting biology of this species in connection with a study of its cleptoparasite, Rhopalolemma rotundiceps Roig-Alsina (Apidae: Nomadinae: Biastini).

The midbody segments of this species as well as of Xeralictus timberlakei, and very probably Systropha planidens have paired dorsal tubercles that are more or less conspicuously truncate apically. The midbody dorsal tubercles of Rophites canus and even of Dufourea a. australis also tend to be truncate, but the flattened surfaces are smaller and less noticeable. Further, at least in the cases of X. timberlakei, Protodufourea eickworti, and S. planidens, the apical integument of these tubercles is finely wrinkled with the wrinkles forming a nondirectional pattern (e.g., not a transverse or longitudinal pattern) compared with the surrounding integument; thus the surface appears almost pebbled at first viewing (figs. 36, 37). There is some indication that the integument on the sides of these tubercles is soft so that the tubercles can be compressed, accentuating their truncate shape (see figs. 36, 60, 61). It may be that nondirectional wrinkling of the apices provides the tubercles with integumental strength compared with the sides of the tubercles. In the case of predefecating larva of S. planidens the apices, when stained, have a well-defined, thicker integument than the surrounding integument. However, in R. canus and species of Dufourea, the apices of the tubercles seem soft because they become rounded and even truncate by the weight of the larva when lying on its dorsum. The anatomy and functioning of body tubercles, whether truncate or not, in the Rophitinae need further study; dorsal tubercles somehow may enable the larva to move while feeding, defecating, and/or cocoon spinning.

Diagnosis

See “Diagnosis” in the treatment of Protodufourea eickworti, above.

Description

The larva of this species is still known from a single specimen, the original description and illustrations of which appeared in a paper by Snelling and Stage (1995). The diameter of its antennal disc measures about 0.5 times the distance between the lower edge of the disc and the center of the anterior tentorial pit.

Remarks

A seemingly remarkable feature of this unique specimen was the truncate dorsal body tubercles, which at the time of the original description were thought to be an artifact, the result of the larva being confined to a small cell. This conclusion now must be revised in light of similar truncate dorsal body tubercles discovered on the larva of Protodufourea eickworti (fig. 31, 36) and to a lesser extent on larvae of some other species. Please see “Remarks” in the treatment of P. eickworti for a reinterpretation of this anatomical feature shared by these genera.

Diagnosis

Because of its projecting labiomaxillary regions, this species can easily be distinguished from Conanthalictus. The clear presence of the articulating arm of the stipes separates it from all known Dufourea and Sphecodosoma, and characters presented in the key will enable it to be separated from the other taxa treated here.

Head (Figs. 39, 42)

Integument of head capsule as described for Dufourea holocyanea. Integument unpigmented except for mandibular apices.

Other head features as described for Dufourea holocyanea except for following: median longitudinal thickening of head capsule absent; antennal disc moderate in size, its diameter about 0.8 times distance between its lower margin and center of anterior tentorial pit; antennal papilla projecting less than one-half basal diameter; labrum with low labral tubercles (rather than merely bilobed as in D. holocyanea) though not surpassing clypeus in lateral view.

Mandible (figs. 40, 41) as described for Dufourea holocyanea except more sinuate in aboral and adoral views (figs. 40, 44) and tubercle cluster less projecting on outer surface (fig. 41). Labiomaxillary region projecting in lateral view but not as strongly as in Dufourea and Sphecodosoma. Inner surface of lower lip with parallel ridges (figs. 45, 46) similar to those of D. holocyanea; microstructure of inner surface of upper lip uncertain. Other feature of maxilla and labium as described for D. holocyanea, except for following: articulating arm of stipital sclerite evident. Hypopharynx as described for Dufourea holocyanea.

Body

Integument as described for predefecating Dufourea holocyanea. Body shape[?or change “form” 3X in this same clause to “larva”?] of postdefecating form unknown; intersegmental lines moderately weakly incised on predefecating form, unknown on postdefecating form; paired prothoracic dorsal tubercles present, more sharply defined but apparently smaller than those of mesothorax, metathorax, and most abdominal segments (fig. 38); all tubercles conical (not transverse); midbody tubercles appearing apically rounded but with apical surfaces thicker and finely wrinkled, so that those of postdefecating larva possibly truncate; abdominal segment 9 strongly produced ventrally; abdominal segment 10 positioned dorsally on 9 as seen in lateral view (fig. 38); on predefecating larva, ventral integument of segments 9 extended, causing segment 10 to be partly tilted and making anus appear somewhat dorsally on 10; ventral integument of segment 9 on postdefecating larva unknown. Spiracles small; those of thorax slightly larger than those of abdomen, those of predefecating larva without sclerites; those of postdefecating larva unknown; other spiracular features as describe for D. holocyanea except subatrium consisting of about 10 chambers. Male with median, transverse, integumental scar ventrally near apex of protrusion of segment 9 of predefecating larva, unknown on postdefecating larva; female sexual characters visible on predefecating larva as small, paired, ventral, pale, subcutaneous imaginal discs arranged sublaterally on abdominal segments 7, 8, and 9, with those of 7 farthest apart and those of 9 closest.

Material Studied

4 predefecating larva, Turkey: Erzurum: Atatürk University campus, N 39°54′09″ E 41°14′06″, July 4, 2001 (J.G. Rozen); 2 predefecating larvae, same except July 1, 2001 (J.G. Rozen, H. Özbek).

Remarks

We studied two nests of this species July 1, 2001, on the campus of Atatürk University (see fig. 63). Main burrows were open, with diameters of 5.0 and 5.5 mm, and descended more or less vertically. Cells were arranged singly and occurred between the depths of 8 and 14 cm. Because the nests were still early in construction, other cells would likely have been added at lower depths. Cells, symmetrical around their long axis, were nearly horizontal, their rears tilted slightly downward. They appeared ovoid (fig. 47), like the cells of other rophitines, in that their length ranged from 9.5 to 12.0 mm (N  =  3) and their maximum diameter was 8.0 mm (N  =  2). Cell walls were dull, smooth, without any evidence of a shiny lining, but the fine-grained surface was presumably smoothed by some activity of the female. When a water droplet was applied to the wall, the droplet was absorbed almost immediately. A number of cell walls had irregular gouges (fig. 47), very possibly created by the Systropha female removing cleptoparasite eggs with her mandibles. Biastes brevicornis (Panzer) (Apidae: Nomadinae: Biastini), a confirmed parasite of this and another species of Systropha (Warnke, 1982), was collected in the area. One cell closure (fig. 48) was shallowly concave on the inner surface and had 3–4 coils to the radius. Laterals, backfilled after cell closure, were 8–10 mm in length (N  =  2). Provisions were spherical, and two were 5.0 mm in diameter. The above information is consistent with that presented by Malyshev (1925a) concerning the nesting biology of this species, the larva of which, as he points out, “spins a cocoon and ejects excrements” and “hibernates curled up and lying freely in the cocoon.” Grozdanić and Vasić (1968) described nests of this species and reported that one nest was 48 cm deep, far deeper than ours. In most other respects the nest construction was similar to that of our nests including burrow diameter and cell size. However, they reported that inner surface of the cells were shiny (“glänzed”); we think this is unlikely since all other rophitines about which we have information characteristically have dull inner surfaces, including those of S. planidens preserved in the AMNH from the current study.

Figures 62, 63

Rophites canus. 62. SEM micrograph of spiracle. 63. Nesting site (in foreground) at the soccer field on the campus of Atatürk University, Erzurum, Turkey. This is the same soccer field where nests of Systropha planidens were excavated.

Figure 64

Postdefecating larva of Conanthalictus bakeri? (specimen not certainly identified to species; see text).

Fig. 65. Predefecating larva of C. bakeri (specimen certainly identified to species).

Figs. 66, 67. Head of postdefecating larva of C. bakeri?, frontal and lateral views, respectively. Scale refers to figs. 64, 65.

Batra and Michener (1966) described a nest and an immature larva of the closely related Systropha (Systropha) punjabensis Batra and Michener.

Diagnosis

The mature larva of this species can be distinguished from those of Conanthalictus, Dufourea, and Sphecodosoma by the combination of the presence of both projecting labiomaxillary region and the articulating arm of the stipes. Its labral tubercles are forward directed as is also the case with Rophites algirus trispinosus, but unlike in the latter, the salivary lips are moderately wide, their apical width at least equal to the distance between their lateral edges and the base of the maxillary palpi, as seen in frontal view (fig. 52).

Head (Figs. 52, 53, 56)

Integumental setation of head capsule as described for Dufourea holocyanea. Integument unpigmented except for mandibular apices, anterior and posterior mandibular articulations, and anterior and posterior tentorial pits.

Head morphology as described for Dufourea holocyanea except for following: Antennal disc much smaller than that of Dufourea holocyanea, its diameter about 0.6 times distance from its lower margin to center of anterior tentorial pit (latter distance measured in maximum profile). Vertex (fig. 53) narrowly rounded as seen from side because of backward slant of frons. Labrum projecting beyond clypeus in lateral view (fig. 53) because of moderately projecting labral tubercles.

Paired mouthparts as described of Dufourea holocyanea except for following: Mandible (fig. 54) with ventral apical edge lacking teeth. Labiomaxillary region moderately projecting in lateral view, with labium projecting only moderately beyond apex of maxilla (fig. 53). Articulating arm of stipital sclerite evident on cleared specimen, though not pigmented; maxillary palpus moderately slender, as long as twice basal diameter. Labial palpus slender, twice as long as basal diameter, subequal in size to maxillary palpus. Salivary lips projecting, transverse, moderately wide in frontal view (figs. 52, 57); as seen with SEM, upper and lower lips with apical fringe of papillae (figs. 57, 58); lower lip with internal surface covered with dense rows of papillae contrasting with the longitudinal ridges in D. holocyanea; upper lip with internal surface also covered by dense papillae; thus salivary lips of this species differing substantially from those of D. holocyanea. Hypopharynx as described for Dufourea holocyanea.

Body

Integument with fine, scattered, nonsetiform sensilla and with finely spiculate areas ventrally. Body form approximately as described for Dufourea holocyanea including paired prothoracic dorsal tubercles present, smaller than those of mesothorax, metathorax, and most abdominal segments (figs. 50, 60); all tubercles conical (not transverse); midbody tubercles (figs. 60, 61) tending to be truncate with apices finely wrinkled as in Protodufourea eickworti and Xeralictus timberlakei (see discussion under “Remarks” in treatment of P. eickworti); terminal abdominal segments as described for Dufourea holocyanea. Spiracles (fig. 62) as described for Dufourea holocyanea. Male and female sex characters as described for Dufourea holocyanea.

Material Studied

4 postdefecating larvae, Turkey: Erzurum: Atatürk University campus, N 39°54′09″ E 41°14′06″, elevation. 1850 m, August 5–9, 2004 (H. Özbek); 10 last larval instars, same except August 21, 2001.

Remarks

Enslin (1921) presented an extensive account of the life history of Rophites canus and described its mature larva and cocoon. Subsequently, Malyshev (1925b) also described aspects of its nests. More recently Wilkaniec et al. (1985) gave an account of its nesting and also evaluated its pollination potential with respect to seed production of alfalfa. The following refers to the nesting site (fig. 63) where we collected the larvae, described above, on the campus of Atatürk University, Erzurum, Turkey, in the last days of June 2004. We first discovered the site because of the many males searching low over the ground for females at a corner of a seldom-used soccer field. We excavated the site on July 6 and 10, 2004, by which time nesting activity was obvious because of nest-entrance tumuli among the low vegetation of herbs and grasses that partly covered the area. Because the nesting season was early, HÖ returned to the site August 5–9, 2004, to retrieve the postdefecating larvae. We also include here data, inserted parenthetically, from a single nest that we excavated not far from the 2004 site on July 4, 2001.

The site occupied an area of about 1 m² with a horizontal surface. We were able to identify nest entrances by the tumuli of loose, dry soil surrounding them. Main tunnels, approximately 3.0 mm in diameter (3.5 mm in diameter in the 2001 nest), descended vertically in a meandering fashion probably caused by pebbles and roots. Laterals leading to single cells were filled with soil after cell closure. Cells, with the long axis nearly horizontal, were encountered starting at the depth of approximately 10 cm with other cells appearing several centimeters lower (8–13 cm in the 2001 nest). We did not find cells to the depths of 25–30 cm reported by Enslin (1921), to the depth of 20 cm as reported by Malyshev (1925b), or to the depth of 23 cm according to Wilkaniec et al. (1985). However, our nests were in the early stages of construction. Cells were arranged variously: singly, in linear series of two, and in one case two cells each leading to another cell in front of them (two from the 2001 nest may have been in linear series). Cells were ovoid in shape; one measured 4.9 mm in maximum diameter and 5.5 mm in length (a single cell in 2001 measured 5.0 mm in maximum diameter and 6.5 mm long; Wilkaniec et al. (op. cit.) measured one cell 6 mm long and 5 mm in diameter; and Enslin (op. cit.) reported cells 6.0–6.5 mm in diameter). They were symmetrical around their long axis, without a flattened floor. Cell closures (fig. 49) were distinct concave spirals in the inner surface with 5–6 coils to the radius. Cell walls were not noticeably harder than the substrate. Their surface was dull (i.e., without a shiny lining) but somewhat waterproof in that a water droplet placed on the surface took 2 min. to be absorbed, whereas a water droplet placed on the surrounding soil was instantaneously absorbed. The 2001 excavation demonstrated that the cell construction was progressive, with older larvae found above younger ones. Three cells were connected to the main burrow by laterals that were 1.2, 1.5, and 1.6 mm long.

Completed provisions were spherical, pale yellow, with a smooth surface, and without a waterproof coating (they “melted” when placed in water in both the 2001 and 2004 studies). Such a coating has been reported for Sphecodosoma dicksoni (Rozen, 1993) and Dufourea novaeangliae (Eickwort et al., 1986). A sample of eight provisions had a median diameter of 3.3 mm with a range of 2.8–3.6 mm (one from the 2001 study was 3.5 mm in diameter). Malyshev (1925b) measured them at 3.0–3.25 mm in diameter, and Wilkaniec et al. (1985) claimed they ranged 3.0–3.5 mm in diameter. Enslin (1921) stated that the food masses were 6 mm in diameter, a questionable measurement since it is approximately the same as the diameter of the cells that he had measured.

The postdefecating larvae, when extracted from their cocoons, were completely covered with a shiny, clear, dry coating that, on first glance, seemed to (but did not actually) glue the posterior end of the abdomen to the ventral surface of the labium, pro-, and mesothorax. The source of the coating is presumably glandular but from what glands is unknown. It was obviously deposited after cocoon construction. The coating perhaps functions to prevent water loss. JGR has observed similar larval coatings among numerous taxa of solitary and cleptoparasitic bees.

Three cocoons from which live larvae were taken had a median length of 7.5 mm (range of 7.0–8.0 mm) and a median diameter of diameter of 6.1 mm (range of 5.6–6.6 mm). Their shape was roughly the shape of the cell. Each cocoon consisted of an outer layer composed of soil to which the feces and silk adhered closely. The soil was obviously the cell wall to which the silk and feces were glued. A layer of feces was thinly spread over the posterior 2/3 to 3/4 of the inner surface of the cell before silk was applied. The inner layer was composed solely of thin sheets of silk closely but loosely applied to one another. The entire layer was semitransparent, so that one can vaguely see the distribution of the darker fecal layer contrasting with paler cell wall where the feces were not applied. Also seen was a white blotch at the posterior end of the cocoon representing the final pale meconial discharge. The inner surface of the cocoon was pale to medium tan. It appeared to consist of a continuous thin, shiny sheet of silk. Its individual threads, while fused with one another to form the sheet, still maintained their outline, thereby imparting a sheen to the reflective surface. No special external construction, or “nipple”, was evident at the front end of the silken cocoon as observed in many megachilid cocoons. The cocoon described and pictured by Enslin (1921: figs. 2–4) appears to be identical to these.

Diagnosis

The very narrow salivary lips and projecting labral tubercles will enable the larva of this species to be distinguished from that of Rophites canus.

Description

Rozen (1993) described a larva of this species that was not fully mature, although it was a last larval instar. The paired midbody dorsal tubercles were rounded as depicted (ibid.: figs. 21, 22) but only questionably with a finely wrinkled apical surface. The nesting site from which it was collected was being attacked by adults of the cleptoparasite Biastes emarginatus (Schenck) (Apidae: Nomadinae: Biastini). No additional material has become available.

Remarks

Stöckhert (1922) described the larva of the related Rophites (Rophites) quinquespinosus Spinola as being exactly like that of Rophites canus except for size.

Diagnosis

The mature larvae, as non–cocoon spinners, lack projecting labiomaxillary regions, and paired dorsal prothoracic tubercles are absent, characters found in other rophitine taxa. The characters presented in the key can separate the two species treated here.

Diagnosis

The apically more pointed dorsal body tubercles of this species (Rozen, 1993: fig. 37) distinguish it from the more rounded body tubercles of Conanthalictus bakeri (figs. 64, 65).

Description

The reader is referred to the previous description and illustrations of the larva of this species (Rozen, 1993).

Remarks

Rozen and McGinley (1976) briefly described the nesting site of this species, and Rozen (1993) subsequently gave a more complete account of the nesting biology.

Figures 68, 69

SEM micrographs of predefecating larva of Conanthalictus bakeri? 68. Labrum, right side, and right mandible, frontal view, showing tubercle on outer surface and bifid apex of mandible. 69. Right antenna, with sensilla identified by rectangles.

Diagnosis

The rounded body tubercles of this species easily separate it from the more pointed tubercles of Conanthalictus conanthi.

The larval mandibles of the two species of Conanthalictus treated here differ from those of all other rophitines in that the mandibles are interpreted as having a conspicuous subapical tooth in addition to smaller teeth extending along the apical dorsal edge both proximal and distal to the subapical tooth. We have re-examined specimens of both and think that this interpretation is still appropriate since the subapical tooth is substantially larger than all other teeth along the dorsal edge in the two species. Interestingly, when the mandible of the predefecating larva of C. bakeri was re-examined with a compound microscope, it, like that of C. conanthi, appeared to have no teeth along the ventral apical edge. Yet, the SEM micrograph (fig. 68) of the predefecating larva clearly has two teeth along this edge.

In the following, comparisons with Conanthalictus conanthi refer to the description presented by Rozen (1993).

Head (Figs. 66, 67)

Integument of head capsule with scattered small sensilla that are not obviously setiform. Integument unpigmented except for mandibular apices.

Head size (figs. 64, 65) moderately small compared with body size; head capsule much wider than length measured from top of vertex to lower clypeal margin in frontal view. Unlike that described for Conanthalictus conanthi, tentorium complete, including dorsal arms. As in C. conanthi, anterior tentorial pit close to anterior mandibular articulation; posterior tentorial pits in normal position; postoccipital ridge weak, evident only near posterior tentorial pit, vanishing dorsally (perhaps more quickly in postdefecating larva than in predefecating larva); median longitudinal thickening of head capsule absent; hypostomal ridge moderately developed; pleurostomal ridge well developed; epistomal ridge laterad of anterior tentorial pit extremely short because of closeness of pit to anterior mandibular articulation; ridge between pits either absent (postdefecating larva) or present but fading well before median line (predefecating larva). Parietal bands evident both as integumental scars and as subcutaneous white imaginal discs on uncleared specimens, unlike those of C. conanthi. As in C. conanthi, antennal prominence large, globose, occupying much of anterolateral area of parietal, though not strongly projecting; antennal disc and papilla moderately large; papilla projecting slightly, bearing 3–4 sensilla. Vertex evenly rounded as seen from side, without projections. Labrum not projecting as far as clypeus in lateral view (fig. 67), very short in frontal view (figs. 66, 68); labrum and epipharynx apparently without spicules.

Mandible similar to that of Conanthalictus conanthi (Rozen, 1993: figs. 38, 39) with conspicuous, sharply pointed subapical tooth and with series of sharp teeth along dorsal apical edge (see “Remarks” for further explanation of mandibular apex); ventral apical edge with one or two minute teeth; cusp not developed; apical concavity scarcely developed, a flattened area near mandibular apex; outer mandibular surface with one or two, small, inconspicuous tubercles, each bearing single short sensillum. Labiomaxillary region greatly recessed, as in C. conanthi. Maxilla with dorsal surface spiculate; cardo, stipes, and articulating arm of stipital sclerite not evident; maxillary palpus recognizable only by several sensilla at maxillary apex. Labium (unlike that described for C. conanthi7) weakly divided into prementum and postmentum, moderately long; labial palpus not evident except for several sensilla. Salivary lips absent; opening of salivary duct a transverse slit, narrow on postdefecating specimen but wide on predefecating specimen. Hypopharynx nonspiculate, not longitudinally grooved; hypopharyngeal groove absent (alternatively the wide salivary opening may possibly be the median remnant of the hypopharyngeal groove, with the salivary duct opening into it).

Body

Integument with fine scattered nonsetiform sensilla; many areas with fine, widely and evenly spaced spicules; dorsal body tubercles faintly wrinkled apically. Body form (figs. 64, 65) moderately robust; intersegmental lines moderately weakly incised on predefecating form; on postdefecating form ventral intersegmental lines modified as explained under “Remarks”, below; as in Conanthalictus conanthi, paired prothoracic dorsal tubercles absent; mesothorax, metathorax, abdominal segments 1–8 with paired dorsal tubercles (figs. 64, 65), which are conical rather than transverse; venter of abdominal segment 9 scarcely produced ventrally; abdominal segment 10 oriented apically on 9, with dorsal surface shorter than ventral surface, so that anus directed somewhat dorsally; venter of 10 of postdefecating larva with faint transverse line. Spiracles very small, without sclerites, not on tubercles; peritreme present, narrow; atrium projecting slightly beyond wall, with rim, globose; atrial wall smooth; primary tracheal opening with collar; subatrium normal in length, consisting of about 10 chambers. Male with distinct median, transverse, integumental scar ventrally near posterior margin of segment 9; female sexual characters unknown.

Material Studied

1 postdefecating larva, USA: Arizona: Pima Co.: Organ Pipe Cactus National Monument, August 15, 1995 (J.G. Rozen and S. Budick); 1 probably early last larval instar, same, except April 1, 1995 (J.G. and B.L. Rozen).

Remarks

JGR discovered a nest of Conanthalictus bakeri (adult collected and identified) on March 25, 1995, at Organ Pipe Cactus National Monument, Pima County, Arizona, and excavated it the next day. Two immature larvae from it were reared for several days after excavation but were preserved on April 1, 1995, because they were not faring well. At least one of these appeared to be an early-stage last larval instar; it is described here, its identity certain. Later the same year, JGR again visited the National Monument and retrieved a single mature Conanthalictus larva from a nest marked in the spring. No adult voucher specimen was associated with the nest. Although the larva appears to be that of C. bakeri, adults of C. deserticola Timberlake were also common in the spring. Thus, the postdefecating larva described here is possibly C. deserticola. The differences between these two larvae are discussed under “Remarks”, below.

Michener's statement (2000) that Conanthalictus has a fused prementum and postmentum was based on only C. conanthi as reported by Rozen (1993). The prementum is weakly separated from the postmentum in C. bakeri.

Although the postdefecating and predefecating larvae described here are nearly identical, they do differ beyond what one usually sees between these two life phases. The peculiar, accordion-like ridges nearly paralleling the ventral intersegmental lines on many of the abdominal segments of the postdefecating form (fig. 64) are without precedent among bee larvae and probably resulted from a postmortem change, the specimen having expanded in the preservative. Presumably the intersegmental lines would be concealed in life, a coarctation that perhaps serves as a defense mechanism to protect a diapausing individual from attack or dehydration. It seems unlikely that an active, functioning bee larva would have such ridges, and they certainly are not present on the predefecating larva (fig. 65). The ridges are integumental in nature since they persisted even after the specimen was cleared in an aqueous solution of sodium hydroxide.

The apices of the 10th abdominal segments of the postdefecating (fig. 64) and predefecating (fig. 65) larvae appear dissimilar. However, while clearing the predefecating larva in an aqueous solution of sodium hydroxide, we discovered that the anus is very large, and the soft, perhaps rectal integument surround the anus appears capable of everting during defecation. The roughened area surrounding the wide, slitlike anus on the postdefecating larva is thus interpreted to be the shriveled soft rectal integument.

The labiomaxillary regions of the two larvae also appear different, but this may be due to postmortem modifications of the more pliable integument of the feeding larva.

Two differences appear to be unrelated to larval stage. The predefecating larva has three sensilla on each antenna, and the postdefecating larva has four sensilla on each. The predefecating larva shows the internal epistomal ridge advancing part way mesad of the anterior tentorial pit, while the median section of the internal epistomal ridge is completely absent in the postdefecating larva (fig. 66), as in Conanthalictus conanthi (Rozen, 1993: fig. 36). More larval specimens of C. bakeri and C. deserticola are needed to determine whether these differences are intraspecific or interspecific.

Diagnosis

This is the largest rophitine species (length > 8 mm) for which we have pupae, but structurally it is almost identical to pupal Dufourea a. australis (length < 7 mm). In addition to a somewhat larger body, the pupa of D. holocyanea (fig. 70) has a proboscis that is shorter relative to its body size than does D. a. australis (fig. 73). The most reliable pupal character (as opposed to characters listed above that are also features of adults) to distinguish between these two species relate to the bilobed apical edge of S6, which in D. holocyanea (fig. 72) has the lobes slightly more distinctly separated than are those of D. a. australis.

The pupa of Dufourea holocyanea is also similar to that of Sphecodosoma dicksoni (Rozen, 1993: fig. 33), which, however, has a substantially smaller body (length < 4 mm) and much longer mouthparts in relation to its body size. The condition of the apical edge of the male S6 of S. dicksoni is unknown.

The pupa of Rophites canus (length < 6.5 mm) can best be distinguished from others treated here by its much reduced paired vertical tubercles, by its more robust overall appearance, by the sharply pointed, hind tibial tubercle, and by the sharply pointed, forward-directed tubercles on T1 (fig. 72).

Head

Integument without setae. Scape with outer surface bearing several low indistinct swellings and moderately large, rounded apical tubercle; pedicel without tubercle; most basal flagellomeres with small but pronounced tubercle next to eye (not visible in fig. 70); all flagellomeres of both sexes with tuberclelike swelling on outer surface. Vertex with large tubercle immediately mesad of each adult lateral ocellus; each adult ocellus covered with small, low convexity; genal tubercle below eye absent; frons with low varicosities on each side; clypeus without tubercles; labrum unmodified.

Mesosoma

Integument without setae. Lateral angle of pronotum swollen, especially laterally; lateral pronotal lobe swollen; mesoscutum with dorsal surface swollen and indistinctly varicose on each side, leaving a deep longitudinal trough along midline and lesser furrows along parapsidal lines; axilla small but produced dorsally; mesoscutellum with pair of strongly produced, somewhat forward-projecting tubercles; metanotum produced medially, indistinctly bilobed. Tegula with strongly projecting tubercle; forewing with outer surface irregular but without tubercles (including large tubercle one-third distance from apex as in Halictinae, Rozen, 2008). Each coxa with pointed apicoventral tubercle; fore- and midtrochanters each with pointed apicoventral tubercle; hind trochanter with rounded swelling; fore-, mid-, and hind femora each with basal swelling; hind femur with two small rounded apical tubercles; hind tibia with elongate basal tubercle on outer surface; all tibiae each with apical rounded tubercle on outer surface.

Metasoma

Integument without setae; T1–6 (male) or T1–5 (female) with posterior transverse row of moderate-size, acutely rounded, non-seta-bearing tubercles, these rows narrowly interrupted along midline and with largest tubercles in each row found sublaterally; rows on T1 and T6 (male) and T1 and T5 (female) consisting of fewer tubercles relative to other rows; apex of metasoma rounded, not produced as elongate apical spine. Male S1–5 apically flared with sublateral apical edges more or less produced, almost tuberclelike; S6 apically bifurcate; female S1–5 sublaterally apically flared, but questionably not as strongly so as in male.

Material Studied

1 male and 5 female pupae USA: Washington: Kittitas Co.: 17 km NNE of Selah, Yakima River Canyon, Highway 821 at mile 10, April 2005, excavated as prepupae (E. Miliczky).

Diagnosis

See “Diagnosis” of the pupa of Dufourea holocyanea.

Description

As described for male of Dufourea holocyanea.

Material Studied

1 male pupa, USA: Arizona: Pima Co.: Desert Station, E side of Tucson Mts., collected as postdefecating larva May 1, 1993, preserved as pupa June 25, 1993 (J.G. Rozen); 1 male pupa, same except preserved as pupa, August 21, 1998.

Diagnosis

See the “Diagnosis” of the pupa of Dufourea holocyanea.

Description

The reader is referred to the previous description of the pupa of this species (Rozen, 1993). The placements of the major pupal tubercles are the same as described for Dufourea holocyanea, above, although the relative sizes of these tubercles differ from those of D. holocyanea.

Diagnosis

Please see the “Diagnosis” of the pupa of Dufourea holocyanea.

Head

As described for Dufourea holocyanea except for following: Paired vertical tubercles much reduced so as to be easily confused with lateral ocelli.

Mesosoma

Axilla small, only slightly produced; paired mesoscutellar tubercles not forward projecting, smaller than those of Dufourea holocyanea. Tegula with strongly projecting but smaller tubercle than that of D. holocyanea. Each coxa with pointed apicoventral tubercle but that of hind coxa small; fore-, mid-, and hind trochanters each with pointed apicoventral tubercle; fore-, mid-, and hind femora each with basal swelling; hind femur with two, small, rounded, apical tubercles that appear more conspicuous than those of D. holocyanea; hind tibia with elongate basal tubercle on outer surface but this tubercle tapering to pointed apex from broad base.

Metasoma

T1–5 (female) with posterior transverse row of small, acute, non-seta-bearing tubercles, these rows narrowly interrupted along midline; row on T1 with tubercles tending to be forward-directed, longer than those in subsequent rows; each subsequent row tending to have tubercles that are smaller, less acute, and less forward directed than those of previous row.

Material Studied

1 female pupa, Turkey: Erzurum: Atatürk University campus, June 2005 (H. Özbek).