H. E. Evans.
A Study of Bembix u-scripta, a Crepuscular Digger Wasp.
Psyche 67:45-61, 1960.

Full text (searchable PDF, 1480K)
Durable link: http://psyche.entclub.org/67/67-045.html

The following unprocessed text is extracted from the PDF file, and is likely to be both incomplete and full of errors. Please consult the PDF file for the complete article.

PSYCHE
-,
Vol. 67 SEPTEMBER, 1960 No. 3
A STUDY OF BEMBIX U-SCRIPTA,
A CREPUSCULAR DIGGER WASP
BY HOWARD E. EVANS
Museum of Comparative Zoology
One of the most characteristic features of the digger wasp genus Bembix is the almost complete obliteration of the ocelli. Two North American species, arcuata and u-scripta, provide the only important exceptions to this, and these species also possess several other structural features which set them aside from all other species of Bembix. A few years ago I presented some very incomplete data on a colony of u-scripta in southern Texas, suggesting at the same time that the be- havior of the species ought to prove of unusual interest because of its apparently primitive structural features (Evans 1957). I have recently made further studies of this same colony and also of another small colony in southwestern New Mexico, thus providing an unusual opportunity to compare the behavior of two widely separate colonies of an uncommon species of distinctly localized distribution. These studies also demonstrate that u-scr'ipta has one very remarkable feature in its behavior: the females hunt and provision only in the dusk, around and after sunset. Throughout the higher Hymenoptera, large ocelli are distinctly correlated with nocturnal flight. It seems very probable that the preservation of functional ocelli in this species repre- sents an adaptation for flying in the twilight. A second problem concerns the identity of Bembix arcuata: is this in fact a distinct species from u-scriptu? The Texas colony was located not too far from the type locality of arcticita, the New Mexico colony not far from that of u-scripta. The New Mexico wasps had a much more extensively yellow color pattern (more like u-scripts), but there was considerable color variation in wasps from both localities. In fact, on the basis of the Texas locality alone, I have already tenta- tively placed arcuuta in the synonymy of u-scripta (Evans 1959). Clearly it will be of interest to see what light behavior sheds on this matter.
In this paper I shall first consider the structural peculiarities of Bemhix u-scripta and also the differences which are supposed to sepa- Psiche 6745-61 (1960). hup //psyche enlclub org/67/67-045 html



================================================================================

46 Psyche [September
rate u-scrip& and arcuata, since this information lends more significane to the behavioral studies. There follows a discussion of the ethology of the two colonies studied and finally a brief statement of conclusions. The studies of the Texas colony were made with the assistance of a grant from the National Science Foundation, while those of the New Mexico colony were made during tenure of a fellowship from the John Simon Guggenheim Memorial Foundation. I am indebted
to Dr. M. A. Cazier, Director of the Southwestern Research Station of the American Museum of Natural History, for making available the facilities of that station during the summer of 1959. I am also indebted to Drs. Cheng Shan Lin and Oliver S. Flint for as- sisting me in the field work on the Texas colony. The dipterous prey of the Bembix have been identified by the following specialists, to all of whom I am much indebted: W. L. Downes, R. H. Foote, M. T. James, C. Martin, R. H. Painter, L. L. Pechuman, H. J. Reinhard, C. W. Sabrosky, Y. Sedman, H. V. Weems, and W. W. Wirth. MORPHOLOGICAL CONSIDERATIONS
The important structural features of Bembix u-scripta and arcuata are given as follows by Parker ( 1917, 1929) : I. Ocelli unusually well developed and apparently functional (posterior ocelli with lenses clear, convex, circular except truncated laterally; anterior ocellus in form of a fairly wide transverse band). 2. Second transverse cubital vein of front wing only slightly curved (in other species of Bembix it is sharply bent). 3. Male with basal segment of middle tarsus distinctly curved, its inner surface beset with several stout spines; middle femora of male dentate, middle tibiae ending in a spine-like process. 4.
Male with abdominal sternites devoid of processes, but seventh with a pair of carinae which diverge basally. 5.
Labrum with a small median elevation (both sexes). 6. Basal segment of front tarsus with an unusually large number of comb-spines ( 8- 10)
( both sexes) .
7.
Scape of male much thickened.
8.
Pygidium of female slightly rugose and with strong lateral ridges.
9.
Pattern of yellow maculations unusually strongly developed. To these nine characters should be added another : 10.
Apical tergite of male with deep longitudinal lateral incisions which cut off a pair of appendages which much resemble the parameres of the genitalia.




================================================================================

19 601 Evans - Bcrnbix u-scripta 4 7
All of these features are unique or shared with only a very few other species of Bewbix. Some of them, though unique in Bembix, are shared (at least in some measure) with species of the more primi- tive bembicine genus Stictiella; these include characters I, 2, and 3 above. Lohrmann (1948') has studied coloration in Bembix at some length, and has concluded that reduction in color pattern (i.e., less yellow) is correlated with structural advance ; thus, on these grounds, character 9 should also be regarded as primitive. The remaining six characters should be regarded as specializations within the genus. However, those which involve the female (5, 6, 8) are not absolutely unique and the remainder (4, 7, 10) involve male secondary sexual characters, which are remarkably plastic in the bembicine wasps. The more striking features thus suggest that these wasps do, in fact, retain several features which may have been characteristic of the ancestral Bembix, though obviously an assortment of specializations of a less striking sort have been acquired.
In describing Eembix arcuata, Parker ( I 91 7) listed the following characters by which it differed from u-scripta (Fox 1895) : I. The nature of the male "genital stipes" (i.e., parameres), which are much more slender and attenuate than in lu-scripta. 2. Mesosternum marked with black (yellow in u-scr'ipta). 3. Abdominal sternites 2-4 with small lateral yellow spots (in u-scripta these are more extensively yellow and there is some yellow on the sternites behind these).
4.
Apical sternite black (yellow in u-scripts). Since the first of these characters involves an important structural distinction, it should be considered first. Parker based his interpreta- tion of u-scripta on the type specimen, one other male, and four fe- males. I have examined these two males and find that it is true that the parameres are less attenuate than in the type of arcuata - or in males from both colonies studied. However, the difference is slight and relative, and I cannot agree with Parker that the genitalia "differ widely". Parker's figures tend to exaggerate the differences between the two: yet the differences cannot be entirely reasoned away. Study of the color characters supposed to separate these species re- veals that while they may apply to the type specimens, they are by no means absolute when additional specimens are considered. The type of u-scripta is almost wholly yellow, that of arcuata predominantly black. Specimens from the Texas colony which I studied are close to arcuata, but exhibit enough variation to cause confusion in using Parker's key.



================================================================================

48 Psyche [September
Specimens from the New Mexico colony are dose to u-scripfa, but again they do not agree perfectly and there is much variation to be considered. When specimens from other localities are considered also, it becomes obvious that the extent of maculation varies in a complete spectrum all the way from "arcuata" to "n-scrip&'. I have tabulated the variation in two color characters (Tables I and 11), selecting these two merely because Parker emphasized them. I have also studied other characters, such as the very noticeable variation in the develop- ment of the U-shaped markings on the mesoscutum, with very similar results. This species has not been collected in enough localities to en- able one to be certain of the relation of geography to color pattern, but the available evidence suggests a northwest to southeast cline. Locality Number of Specimens
Largely yellow Tip yellow Tip yellow
apically all across on sides
California1 2
Wellton, Ariz. 1
Rodeo, N. Mexico 2
Lordsburg, N. Mexico
Albuquerque, N. Mexico 1
Mexcala, Guerrero
Pt. Isabel, Texas
TABLE 11. COLOR OF MESOSTERNUM OF MALE
Locality Arumber of specimens
Almost Black spots Posterior .6--8 Sternum and pleura wholly posteriorly black, some black except yellow on pleura anterior border
Tucson, Ariz.'
2
Rodeo, N. Mex. 9
Lordsburg, N. Mex. 1
Mesilla, N. M~x.~ 1
Pt. Isabel, Tex. 9 24 5
Cotulla, Tcx4
2 1
Paratypes of u-scripfa, bearing no further locality data. These specimens otherwise have the yellow maculations generally reduced as compared to Arizona specimens, and also have a large black spot on the clypeus. Types of u-scripts
Paratype of arcuata
Type and paratypes of arcuata
OBSERVATIONS ON COLONY AT PORT ISABEL, TEXAS This colony was first discovered June 23-25, 1956, but at this time it was evident that the wasps had nearly completed their activities for



================================================================================

19601 Evans - Bembix u-scr@a 49
the season. Only a few females with badly worn wings and mandibles were still active, and even these ceased activity after June 25 (Evans 1957). I returned to the area May 8-11, 1958, and found the wasps at the peak of their activity. C. S. Lin revisited the area June 3-7 of the same year and found that the males had disappeared and that nest- ing of the females was more advanced. Thus in this area the wasps first become active in early May (probably in late April) and the
females remain active for six to eight weeks. The species is certainly univoltine, as cocoons collected June 23, 1956, gave rise to adult males and females in late March, 1957 (they were overwintered at room temperature at Ithaca, N. Y.) .
The colony was located on the protected beach of Laguna Madre about five miles west of Port Isabel. The soil here is a rather heavy, dark sand containing pieces of shells and bits of organic matter (chiefly bits of decayed seaweed and beach halophytes). This beach is occasion- ally swept by waves from Laguna Madre (which is salt), especially during the fall and winter, but there is little tide action. On certain parts of the beach there are extensive patches of low vegetation, chiefly Salicornia, certain composites, and various grasses. The wasps oc- curred only in bare places on the middle of the beach, that is, from about 8 to 15 meters above the high water mark but below the banks at the upper edge of the beach.
Behavior of males. - Males were observed only during early May, 1958, but at that time they were very abundant. Each morning we entered the area about 0800, and at this time the males were already engaged in their "sun dance". Each male flew rapidly in loops, figure eights, and irregular patterns only 2-6 cm. above the sand surface, much as in Bembix cinerea (Evans 1957). By 1030 some reduction of males could be noted, and by I I I 5 each day the last male had dis- appeared. By 1500 a few males were once again active, and from 1600-1900 each day there was a second though somewhat less popu- lous sun dance. The greater part of this sun dance occurred in a small area somewhat apart from the major nesting area of the females, although males often flew briefly over various parts of the beach before rejoining the sun dance.
During the middle of the day, and also at night, the male's enter short, oblique burrows in the sand and throw up a small barrier of sand behind them. These burrows tend to be grouped in small clus- ters. They are dug by the males, but the males usually re-enter an old burrow rather than digging a new one each day.



================================================================================

50 Psyche [September
During the sun dance the males fly, at least some of the time, with their middle and hind legs hanging down. When they approach a fe- male they attempt to descend upon her from above and grasp her with the legs. During this time the pair rise a meter or two in the air, then, if the female is receptive, descend to the ground, where copulation occurs. During mating, the male rests on top of the female, often buzzing his wings (perhaps to maintain balance), while the female re- mains completely motionless, her wings being clamped in place by the middle legs of the male. The spines of the middle' femora of the male apparently fit over the wing-bases of the female, while the arcuate middle tarsi of the male appear to fit over the coxae of the female. The front legs of the male extend rigidly forward and the hind legs rigidly backward during copulation; neither pair is in con- tact with the female or the substrate. The male genitalia are of course extruded and work up and down rhythmically until they gain entry into the genital chamber of the female. The longest copulation observed lasted two minutes.
"Nesting behavior of females. - Most of the females were concen- trated in one large bare area on the beach. In this area there were at least 100 nests, many of them only a few centimeters apart. Nests on the periphery of the colony tended to be more widely separated. Here and there along the beach there were other smaller groups of nests, so that the total may have approximated 200. Most new nests appear to be started in the morning (0900-1030). The female is able to dig about 10 cm. an hour, but because of fre- quent periods of inactivity, especially at midday, the greater part of a day is generally required to complete a nest. Much use is made of
the mandibles in loosening the soil, and this results in much wear of the mandibles as the season progresses. The soil which is loosened is scraped back with the front legs in the usual manner of wasps of this genus. When the burrow is a few centimeters deep, the wasp allows the soil to clog the entrance, sometimes for Iong periods, before finally coming out and scraping it back from the entrance. The digging movements appear rather slow as compared to other species of the genus.
As digging progresses, a large mound of sand accumulates behind the entrance. This mound measures, on the average, about 18 cm. long, 10 cm. wide, and 1.5 cm. high. When the nest is completed, some individuals level this mound. The leveling movements are very characteristic and unlike those of any other known Bembix. The fe-



================================================================================

19 601 Evans - Be~nbix u-scripts 51
male (without closing the nest entrance) backs out to the middle of the mound of sand, then walks forward in a straight line scraping sand backward. When she arrives at the nest entrance, she turns to one side at a 45-90' angle and continues for another 1-3 centimeters, still scraping sand, which now falls over or slightly beyond the nest entrance. She then flies back to the center of the mound and repeats Fig. 1 -Diagram of leveling movements of Bembix u-scr'ipta. In a is shown the completed mound of earth with the open (blackened) nest en- trance at one end. In b three typical routes of a female leveling the mound are shown; dashed lines indicate movement by flight. In c is shown the final product, with the nest entrance covered (indicated by dashed line) and mound well removed from entrance.
the performance. After a time this results in the soil immediately in front of and around the nest entrance being well scattered and flat- tened. However, the extremity of the mound of soil has not been touched and has in fact been added to. Thus the net result of the leveling movements is to move the bulk of the accumulated soil from the nest entrance to a position well removed from it (by about 8-12 cm.) (fig. I).
In this colony relatively few individuals leveled the mound. We estimated that about 25% leveled, while some others exhibited weak and indefinite leveling movements. The remaining members of the colony left the mound of earth completely intact. As additional cells are added to the nests, some additional soil is added to the mound;



================================================================================

52 Psyche [September
as a result nests containing several cells often have a very large mound of earth (as much as 3 cm. high). It was our impression that indivi- duals that level following completion of the initial cell also level after each additional cell, but our data on this point are weak. When the nest or new cell is completed, and leveling is completed if this is to occur, the female flies oft to obtain her first fly, on which the egg is laid. This invariably occurs in the late afternoon or early evening, and after oviposition the female closes the nest from the in- side for the night. Thereafter the nest entrance is left closed from the inside when the female is inside for long periods of time (at night and during much of the day). During these periods the cell is also closed off from the burrow by a small barrier of sand. The female remains in the bottom of the burrow just outside this barrier or "inner closure". During periods of provisioning, however, both inner and outer close- ures are removed. Nest entrances are occasionally found to be open for brief periods in the morning, probably while the female is out taking nectar from flowers. In this area no wasps were actually ob- served visiting flowers. Final closure of the nest is not especially distinctive and has already been described (Evans 1957). Description of nest. -The burrow measures about 15 mm. in diameter, generally slightly more than this near the entrance. It enters the soil at a 45-70' angle with the horizontal but tends to level off as it approaches the horizontal cell. The burrow is often straight but may have a weak to strong lateral curve part way down. Measure- ments of 36 nests showed the burrow length to vary from 19 to 40 cm., with a mean of 30 cm. Depth of the bottom of the cell, measured vertically from the soil surface, varied from 10 to 22 cm., with a mean of 15 cm. Cells measured about 18 mm. in diameter and from 30 to 40 mm. in length.
Of 16 nests dug out May 8-1 I, 1958, 14 contained one cell and two contained two cells; none of these nests had yet received the final closure and hence all might have eventually had additional cells. Of 20 nests dug out by C. S. Lin June 4-7 of the same year, I I contained one cell, 8 contained two, and one contained three. Only a few of these (including the three-celled nest) had received the final closure. The several nests dug out in late June, 1956, were nearly all multi- cellular (Evans 1957). Two nests were tricellular; one of these was still active while the other was receiving the final closure when dug out. Two other nests appeared to have four cells and another five. However, in these cases all the cells contained cocoons and the burrows



================================================================================

19601 Evans - Bemhix u-scripta 53
leading to them had been filled and hence could not be traced accurate- ly. Because of the close proximity of nests in this colony, it is im- possible to be certain that all of these cells belonged to the nest in question. The evidence now seems strong that the nest is usually closed after the completion of only two or three cells. It is not im- probable that one-celled nests also occur, perhaps also four or five- celled nests rarely. Two or three-celled nests are characteristic of Bemhix nubilipennis (unpublished observations), while B. cinerea and B. belfragei characteristically make one or two-celled nests (Evans 1957).
Fig. 2.- Typical three-celled nest of Bemhix u-scr'ipta (our field note no. 1635, Rodeo, N. Mex.). Cell a contained a completed and hardened cocoon, so was obviously the first cell of the nest. Cell b contained a fresh, still soft cocoon; cell c was a new cell which was empty when the nest was dug out. The figure on the left shows the nest in lateral view, compressed to a single plane, while the sketch on the right shows the proper relationship of the cells in a horizontal plane.
The cells of a given nest are normally separated by several centi- meters of soil. After the first cell is fully provisioned, the bottom .3 to .5 of the burrow is packed tightly with sand. The wasp then con- structs another section of burrow of about equal length at roughly a go0 angle with the main burrow and builds the second cell at the end of it. The third cell-burrow and cell are normally constructed from the opposite side of the main burrow (fig. 2). Provisioning. - During the first two days of study, we were unable to find females provisioning their nests even though the nests which we dug out usually contained fresh flies. We left the area on these
days at 1600-1 700, a time when most digger wasps have completed



================================================================================

54 Psyche [September
[heir activities for the day. On the third day we noted a few females provisioning their nests about 1600, and remained in the area until 1915. We discovered that there was a great emergence of females from their nests just before sunset. During the two-day period of observation, the first female carrying prey was seen at 1520, the last at 1905, with the great bulk of the provisioning occurring between 1700 and 1900. Sunset was at about 1830. Some females and males
were still active at 1915, when it was too dark for us to see them well, but apparently these individuals were merely digging in for the night. In order to determine what types of flies the wasps were taking at this unusual hour, we took 97 flies from females or from the nests. These flies represented 33 species of nine families; that is, there was an average of only three flies of any one kind. The most numerous flies were those of the genus Acrosticta {Otitidae) (26 examples). Bombyliidae, Asilidae, and Tachinidae were represented in roughly equal numbers, each family by several species. A complete list is pre- sented in Table 111. Although many species of Bembix are relatively unselective of their flies, I have never encountered a more remarkable diversity of prey in any one locality. Most of the flies used were rel- atively small, and a great many of them were used per cell (about 40). However, some relatively large flies were found in some nests. It required very little study to determine that virtually all Diptera had ceased activity by the time the Bctnbix began to provision in num- bers. Some of the flies (eg., the otitids, and some of the bornbyliids and sarcophagids) were seen commonly on the beach during the middle of the day, but in the evening they had disappeared. The Bembix ap- parently hunt widely in the vegetation and take virtually any "sleep- ing" flies within a certain size range whkh they encounter there. Females carrying flies enter the nesting area with a rather loud buz- zing only 10-20 centimeters high. They plunge quickly into the nest entrance which, as noted above, is never closed during provisioning. Ill less than a minute, they re-emerge from the burrow and quickly rake wing. Several instances were observed of females attempting to steal flies from one another.
Ovipwitwn and development. - The first fly is placed deep in the cell on its back, head-in, with one wing somewhat extended. The egg is laid erect, glued to the side of the fly near the base of the extended wing, or actually glued to the wing-base. One instance of abnormal oviposition was observed. In this case the fly (Acrosticta sp.) was actually smaller than the egg and was dorsum-up, with the egg of the wasp glued to the base of the abdomen and extending obliquely for-



================================================================================

ward. The egg of u-scriflta measures about 4.5 mm. long. The egg hatches in about two days and the small larva remains at- tached to the fly pedestal by a glutinous thread for several days. Pso-
visioning is fully progsessive; that is, flies are provided each day up until the larva is full-grown and nearly ready to spin its cocoonm The wasp does not clear the fl37 remains from the cell, and they mere- ly accun~ulate in a mat on the bottom of the cell. Apparently about a we~k is required for completion of larval development. TABLE 111.
PREY OF REMBIX U-SCRJPTA -AT PORT ISABEL, TEXAS Family and spc~icf of fly
Number taken
STRATIOMYIDAE
Herm~tia aurota Bell
TABANIDAE
Tabanus te.xanus Hine
BOMBYLIIDAE
.4phoebantus sp. nr. hir~utus Coq.
Lordotus gihbus gihbus Lw.
L. gibbus stridus Fainter
Phthiria sp.
Poecilanthrax lucifm-a Fabr.
Villa parwicon~is Lw,
-ASILIDAE
Erax cressoni Hhe
E. tuberculatus Coq.
PIei.~osomma unicolor Lw.
Psilocurus modestus Will.
P. puellus Bromley
Stenopogon ehyi Bromley
SYRPHIDAE
?~oluceIla fraudulcn~~ Will.
V. unipunctata Curran