PROC. ENTOMOL. SOC. WASH. 97(4), 1995, pp. 151-166 A RECONSIDERATION OF THE CYLINDRICVS GROUP OF THE GENUS ANOPUUS DUFOIJR (HYMENOPTERA: POMPILIDAE) Howard E. Evans Department of Entomology, Colorado State University, Fort Collins, Colorado 80523. — Abstract. Three names are removed from the synonymy of Anopliiis {Pompilinus) cylindricus (Cresson). These are Arachnophila brevihirta Banks, Pompilinus subtruncatus Dreisbach, and P. truncatus Dreisbach. Pompilinus hispidus Dreisbach is regarded as a synonym of P. subtruncatus Dreisbach, while P. clavipes Dreisbach continues to be con- sidered a synonym ofA. {P.) cylindricus (Cresson). Anoplius {Pompilinus) cylindricus is restricted to the southwestern United States and is clearly distinct from other members of this complex. The remaining three species are closely related and show some evidence of intergradation; two of them, A. (P.) subtruncatus and A. (P.) truncatus, occupy more limited ranges within that of the transcontinentally distributed species A. (P.) brevihirta. Some of the variation in A. (P.) brevihirta is analyzed. This species is restricted to broad areas of open sand, and populations in different sandy areas have in some cases diverged slightly in structure and color; there are also broadly geographic trends in structure and color. Key Words: Spider wasps, Pompilidae, Anoplius, synonymy, distribution, variation Retirement should provide time to re- (generotype Psammochares marginalis think some of the problems in one's past Banks). I regard Pompilinus as a probably research in which the solutions proposed monophyletic New World offshoot of the seem naggingly unsatisfactory. Many years widely distributed Arachnophroctonus. In ago (Evans 1951) I placed four names in 1951 I erroneously placed the European An- the synonymy ofAnoplius (Pompilinus) cy- oplius viaticus (L.), under the name A. fus- lindricus (Cresson); a fifth name was later cus (L.), in Pompilinus; more properly it added to the synonymy (Evans 1966). Ev- belongs in Arachnophroctonus. idence to support this lay in the apparent Whether Pompilinus deserves subgeneric continuous variation in features of the male status can be argued, but it is a large group genitalia. Further collecting in the western divisible into at least three species-groups, half of the United States in recent years has so it is convenient to retain it as a subgenus. convinced me that such a simplistic solu- The most useful distinguishing features of tion to this complex is not satisfactory. this group are the fact that the transverse Here I propose that four species be recog- median vein of the fore wing meets the me- nized in this complex, with two other dia beyond the origin of the basal vein, the names being retained in synonymy. third submarginal cell is petiolate, and the Day (1974) placed the subgeneric name spines of the female tarsal comb rarely ex- Pompilinus Ashmead (generotype Pompilus ceed the width of the basitarsus. The spe- cylindricus Cresson) in the synonymy of cies ofthis group are difficult to separate in the subgenus Arachnophroctonus Howard the female sex, and the males can be sep- PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON 758 aratecl with assurance only by study of the forms, so ofcourse nothing can be said con- genilaHa. In 1951 I recognized two species- cerning their predatory or nesting behavior. groups in Pompiliniis, the subcylindricus Males taken away from sandy areas tend to and cylindricus groups. Here I restrict the be larger than A. brevihirta and are all cylindricus group to that species and three black. These males can be sorted into forms previously considered synonyms of groups equivalent to Dreisbach's species The remaining members of the former Pompilinus subtruncatus and P. truncatus it. cylindricus group of Evans (1951) may be and to Cresson's Pompilus cylindricus. On considered to constitute the marginatiis at least one occasion these three forms have group. been taken in one locality on the same date The following are the distinguishing fea- (Tomillo, El Paso Co., Texas). Two ofthem tures of members of the cylindricus group have been taken together on at least six oc- as here defined. Females: eyes widely casions (localities in Texas, New Mexico, spaced, middle interocular distance 0.60 to Colorado, Illinois, and Michigan). Howev- 0.65 times transfacial distance; vertex pass- er, each of these forms (and A. brevihirta) ing straight across between tops of eyes; is subject to variation in features of the ocellar triangle broad and flat; propodeum male genitalia and subgenital plate, varia- convex, with a steep, oblique declivity, and tion that in some cases blurs any lines that bearing at least a few short hairs (may be might be drawn between species. quite hairy); femora usually hirsute; fore An exception is A. cylindricus itself. basitarsus with a comb of three or four Study of the genitalia of the lectotype plac- spines, these spines one to two times as es this as a member ofa population restrict- long as width of basitarsus; one or more ed to parts of the Southwest. The genitalia basal tergites banded with orange. Males: are quite distinctive, and I have seen no subgenital plate slender, acute, hirsute specimens intermediate with other members (Figs. 5-8); genitalia with squamae at base of this complex. These other members, A. of parameres prominent; digiti clothed with brevihirta, A. subtruncatus, and A. trunca- short setae, many of which are bent apical- tus, show evidence of intergradation, the ly, except latero-apically with longer, latter two forming populations (not restrict- straight setae; digiti broadly truncate api- ed to sand dunes) within the broad range of cally (one exception) (Figs. 1-4); basal ter- the dune-inhabiting A. brevihirta. gites with or without orange markings. In 1951 I characterized A. cylindricus as Anoplius (Pompilinus) a sand dune species, and various papers on cylindricus (Cresson) behavior published under this name all ap- Pompilus cylindricus Cresson, 1867, p. 92. ply to sand dune populations. All of these (Lectotype: 6, Texas, ANSP). indicate that females prey on immature Pompilinus clavipes Dreisbach, 1958, pp. Geolycosa spiders occurring in sandy situ- 61-63. (Holotype: <?, Texas: Conlon [? ations, using the spider's burrow in which Conlen, Dallam Co.], 8 Aug. 1952, R.R. to construct their own nest cells (Krombein Dreisbach, MCZ). 1953, Kurczewski and Kurczewski 1968, 1973, Gwynne 1979, Kurczewski 1981). The lectotype of Cresson's Pompilus cy- The name A. brevihirta (Banks) is here ap- lindricus has genitalia and a subgenital plied to these populations. plate virtually identical to those figured for Other elements in the species-group are Pompilinus clavipes by Dreisbach, and as by no means restricted to sandy places, but figured here (Figs. 1, 5). This lectotype was are often taken on plains or prairies. Un- designated many years ago, prior to the fortunately females cannot at present be as- studies of Dreisbach (1949) and Evans sociated with certainty with any of these (1951). The other 12 specimens in Cres- VOLUME 97. NUMBER 4 759 Figs. 1^. Male genitalia (right half, ventral aspect) ofAnopliiis (Pompilinus) species. 1. cylimiricus (Cres- son). 2, suhtnmcatus (Dreisbach). 3, tnmcatus (Dreisbach). 4. hrevihirta (Banks). 760 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON 5 8 7 Figs. 5-8. Male subgenital plates (ventral aspect)ofAnopliiis (Pompilinus)species. 5, cylindricus(Cresson). 6, subtruncatus (Dreisbach). 7, truncatus (Dreisbach). 8, hrevihirta (Banks). VOLUME 97, NUMBER 4 761 son's type series are all assignable to other The type specimens of Pompilinus sub- species. Other males before me are from truncatus and P. hispidus do indeed appear Kansas: Reno Co. (MCZ, NMNH); Texas: quite different. The former has the hairs on El Paso, Jeff Davis, and Pecos counties the subgenital plate short and depressed and (CU, CSU); New Mexico: Chaves Co. the inner apical margins of the digiti not (CSU); and Colorado: Prowers Co. (CSU). angularly extended. The latter has the hairs All have essentially identical genitalia and on the subgenital plate erect and bristling subgenital plates. All are all black, with a and the inner apical angles of the digiti pro- fore wing length from 6.0 to 8.5 mm. The jecting angularly. However, other speci- genitalia are distinctive in that the para- mens combine these two features different- meres are broadly spatulate and heavily se- ly. Dreisbach pointed out two small groups tose along the mesal margin; the digiti are of setae near the midline of the subgenital rounded apically, rather than truncate as in plate of P. hispidus, but these are not pres- other members of this complex. ent in specimens I have seen other than the I have seen several females that very type. probably belong here. All have a small Variation in the condition of the hairs on amount of short erect hair on the propo- the subgenital plate is striking. In some deum, but the femora have only scattered, specimens they are wholly decumbent, in very short hairs. The comb spines on the others only a few are erect, and in still oth- fore basitarsus are short, no longer than ers most are erect. In a male from Hialeah, width of the basitarsus; the basitarsus bears Florida (NMNH), these hairs are especially 3 such spines (in one specimen 4). All have long and bristling. A paratype of P. hispi- the first two tergites orange; fore wing dus from Baldwin, Kansas (NMNH), has a length varies from 7.7 to 10.8 mm. The strongly hirsute subgenital plate but the dig- strongest evidence that this association is iti lack an angulate process on the inner, correct is to be found in a series of 3 fe- apical margin. There are other examples males and 13 males, all taken by myselfon that, to me, suggest that this is a single spe- 7 July 1953 at Medora, Reno Co., Kansas cies subject to discordant variation in seem- (MCZ, NMNH) (the males are all paratypes ingly important features of the male abdo- of Dreisbach's Pompilinus clavipes). Other men. Some specimens have fewer setae females are from Kansas: Riley Co. (CU); along the mesal margin of the parameres Oklahoma: Grant Co. (MCZ); and Texas: than figured, suggesting possible intergra- Bastrop, Bee, Brewster, Dimmit, Travis, dation with A. brevihirta. Smaller speci- Kenedy, and Kleburg counties (BFL, CSU, mens fall within the size range of that spe- CU, MCZ). cies; overall size range (fore wing length) varies from 5.3 to 9.0 mm. I have been un- Anoplius {Pompilinus) subtruncatus able to associate any females with this spe- (Dreisbach) New Status cies with any certainty. Males are all black, but it is probable that females are marked Pompilinus subtruncatus Dreisbach, 1949, with orange as in other members of this pp. 17-18, figs. 58, 59. (Holotype: S, complex. Nebraska, Lincoln, 14 June 1909, C.H. This species is widely distributed Gable, Univ. Nebraska). throughout the United States from the Pompilinus hispidus Dreisbach, 1949, pp. Rockies eastward. I have seen specimens 23-24, figs. 47, 48. (Holotype: 6, Mich- from the following localities: Florida: Hia- igan, Tuscola Co., 20 Aug. 1940, R.R. leah (NMNH); Maryland: Bowie (NMNH); Dreisbach, MCZ). Louisiana (without further data) (NMNH); Anoplius hispidus, Evans, 1970, p. 482. Michigan: Tuscola Co. (MCZ); Iowa: Sioux (Teton Co., Wyoming). City (NMNH); Kansas: Baldwin (NMNH), 762 PROCEEDINGS OFTHE ENTOMOLOGICAL SOCIETY OF WASHINGTON Clay Co. (NMNH); Nebraska: Chadron mm. I have been unable to associate any (Univ. Colorado, Boulder), Lincoln, Neligh females with this species. (Univ. Nebraska); North Dakota: Beach Anoplius truncatus appears to have a (NMNH); Wyoming: Teton Co. (MCZ); more restricted range than A. subtruncatus, Colorado: Bent, Elbert, Larimer, Otero, and occurring from Texas and New Mexico to Weld counties (CSU); Texas: Jeff Davis Illinois and Michigan. Dreisbach's para- (CU), El Paso (MCZ), and Presidio (ANSP) types are from Manistee and Midland coun- counties. ties, Michigan, Carlinville, Illinois, and Roswell, New Mexico. Other males I have Anoplius (Pompilinus) triincatus seen are from Colorado: Bent, Larimer, (Dreisbach) New Status Prowers, and Weld counties (CSU); New Mexico: Hidalgo Co. (CSU); and Texas: El Pompilinus truncatus Dreisbach, 1949, pp. Paso (CSU) and Jeff Davis (ANSP) coun- 15, 16, figs. 40, 41. (Holotype: S, Mich- igan, Gratiot Co., 2 Aug. 1947, R.R. ties. Dreisbach, MCZ). Anoplius (Pompilinus) brevihirta As the species name implies, the para- (Banks) New Status meres of the male genitalia are obliquely Arachnophila brevihirta Banks, 1945, p. truncate apically (Fig. 3). An additional 105. (Holotype: 9, Chicago, Illinois, useful characteris the absence ofsetae from July, C.T Brues, MCZ). the center of the disc of the digitus. The Anoplius (Pompilinus) cylindricus, Evans, subgenital plate (Fig. 7) tends to be more 1951, pp. 294-297, figs. 91, 123. (In slender and pointed than in the preceding part). species, but this is a subtle and not wholly dependable character; the sides of the plate This is the most widely distributed mem- are less sinuate than inAnoplius cylindricus ber of this complex, ranging from coast to (Fig. 5) andA. brevihirta (Fig. 8). The mes- coast and from Canada deep into Mexico. al margin of the paramere is relatively It is the form most often identified as A. sparsely setose, much as inA. brevihirta but cylindricus, under which name there have in contrast toA. cylindricus andA. subtrun- been several reports on its biology (as dis- catus. cussed above). It is characteristic of areas In a series of6 males from Tornillo, Tex- of open sand, and virtually every system of as (CSU), taken on the same day, some dunes I have visited has a population of agree well with A. truncatus and others these wasps. Males average smaller than with A. brevihirta, while one has the left those of the preceding three species, fore paramere truncate (as in A. truncatus), the wing length varying from 4.0 to 7.0 mm. right paramere acute (as in A. brevihirta). The genitalia have the parameres acute, In a series of 4 males from Great Sand their mesal margins sparsely setose; the dig- Dunes National Monument, Colorado iti are fully clothed with setae and the dis- (CSU), taken on the same day, three have tal, mesal angle is prominent (Fig. 4). The genitalia typical of A. brevihirta, while the subgenital plate is slender, with sinuate fourth has truncate parameres (although the margins, and is clothed ventrally with sub- digiti are fully clothed with setae and the recumbent setae (Fig. 8). subgenital plate is sinuate laterally). These Even with the removal of the three pre- specimens, and others, suggest that there is ceding species from synonymy, this re- some interbreeding between these two spe- mains a very variable species. About 75% cies. Males ofA. truncatus are all black and of the males I have seen have one or more average larger than those of A. brevihirta, basal tergites marked with orange. All- fore wing length varying from 5.5 to 7.2 black males cannot always be distinguished VOLUME 97, NUMBER 4 763 from those of other members of this com- Table 1. Geographic variation in male coloration plex without examining the genitaha. Even and the number ofcomb spines on the fore basitarsus so, occasional specimens are puzzling. A of females of Anoplius (Pompilinus) hrevihirta (Banks).' male from Mt. Vernon, Missouri (CSU), has the parameres very slender, almost lin- ear. A few males have the parameres some- what truncate apically; two such specimens were cited under A. tnmcatus. Females vary in fore wing length from 4.5 to 9.5 mm. Those from any one locality often differ greatly in size, suggesting that they have developed on spiders of different sizes. Females also differ in the number of spines on the fore basitarsus and in the amount of erect hair on the femora and the propodeum. Females have at least the sec- ond tergite banded with orange, sometimes part or all of the first and third tergites and some ofthe basal sternites. Since sand dune systems tend to be widely spaced in nature, it is possible that local populations have di- verged slightly in structure and color. Some of the variation also has a broadly geographic component (Table 1). Males from the eastern coastal states invariably have orange coloration on the abdomen, and this is true of males from Kansas, Tex- as, and eastern Colorado. Females from these same areas usually have only 3 comb spines on the fore basitarsus, while those from Michigan and from south central Col- orado usually have 4 comb spines (in these two areas males are usually all black). Fe- males from many western sites tend to have much more hair on the propodeum and femora than those from eastern localities, which may be almost devoid of such pilos- ity. It is not clear why there are resemblances between the population at Great Sand Dunes in Colorado (represented by a series at CSU) and that at Warren Dunes, Michi- gan (represented by a series at MCZ) (Table 1). The latter site is subject to cooling winds from Lake Michigan, such that cer- tain microclimatic conditions may approx- imate those at a much higher elevation (2440 m) in Colorado; however, the Colo- rado site has a much lower annual precipi- 764 PROCEEDINGS OFTHE ENTOMOLOGICAL SOCIETY OF WASHINGTON 20 - 18 - 16 - 14 (/} UJ -<I 12 - UJ 10 - IL. uo. 8 - 6 4 2 4.0- 4.4- 4.8- 5.2- 5.6- 6.0- 6.4- 6.8- 7.2- 7.6- 8.0- 8.4- 8.8- 92- 9.6- 4.3 4.7 5.1 5« 5.9 63 6.7 7.1 7.5 7.9 8.3 8.7 9.1 9.5 9.9 LENGTH OF FORE WING (mm) Fig. 9. A size and color comparison of two generations ofAnopUus (Pompilimis) brevihirta (Banks) from coastal North Carolina and the single generation from southern Colorado at 2440 m elevation. A, Kill Devil Hills, NC, 23 June to 14 Sept.: 2-3 tergites orange. B, Kill Devil Hills, NC, 24 May to 22 June: 1-1.5 tergites orange. C, Great Sand Dunes National Monument, CO, July-August: 1-1.5 tergites orange. er; here the females approximate in size and Key to Males of the coloration members of the spring genera- a. cylindricus group tion in coastal North Carolina. However, these females have much more body hair 1. Digiti rounded apically; parameres broadly and a stronger tarsal comb. It is possible - Dsipgaittuilabtreoa(Fdilgy.t1r)uncateapicallyc;ylpianrdarmiecurses(Clresessson) that the more extensive orange coloration of broadly expanded apically 2 the second generation in North Carolina is 2. Parameres obliquely truncate apically; center the result of summer heat on the pupal of disc of digiti devoid of setae (Fig. 3) stage, as occurs in other Hymenoptera. tnmcatiis (Dreisbach) Males from Texas, New Mexico, and east- - Parameres more or less acute apically; center em Colorado are almost invariably marked ofdigiti covered with short setae 3 3. Mesal margin of parameres heavily setose with orange, and there is a corresponding (Fig. 2); subgenital plate somewhat broader, tendency for females from these areas to be sides gradually tapered, often bearing erect more extensively marked with orange. setae (Fig. 6); all black Clearly we are far from explaining the ob- subtruncatus (Dreisbach) served variation in this species, but it must - Mesal margin of parameres more sparsely se- be recognized when one attempts to cir- twoisteh s(Fiingu.ate4);latseruablgemnairtgalinspla(Ftieg.mo8)r;ebasslaelndteerr,- cumscribe this species as distinct from oth- gites with or without orange coloration ers. brevihirta (Banks) VOLUME 97, NUMBER 4 765 Conclusions Acknowledgments Members of the A. cylindricus species- For the loan of specimens, I am indebted group occur throughout much of North to Arnold Menke, U.S. National Museum America and have commonly been consid- of Natural History (NMNH), Donald Azu- ered to constitute a single variable species. ma. Academy of Natural Sciences of Phil- However, four populations can be distin- adelphia (ANSP), Michael S. Kelley, Mu- guished on the basis of the male genitalia. seum of Comparative Zoology at Harvard While the differences between them are (MCZ), E. Richard Hoebeke, Cornell Uni- slight, they seem sufficiently discrete to versity (CU), and Allan Hook, Bracken- suggest that to a considerable degree they ridge Field Laboratory, University ofTexas breed independently of one another. This is (BFL). Other specimens are located at C.P especially true of A. cylindricus itself, Gillette Insect Biodiversity Museum at Col- which is restricted to the southwestern orado State University (CSU). states. Two others, A. subtruncatus and A. Literature Cited truncatus (both known from males only), occupy limited ranges within that of the Banks, N. 1945. Two new species of Psammochari- very widely distributed A. brevihirta. The dae. Psyche 52: 105-106. last is characteristic of broad areas of open CressNoonr,thE.AmeT.ric1a8,67w.ithNodteesscriopntiotnhseoPfonmepiwlisdpeaceieso.f sand, where females prey on Geolycosa spi- Transactions of the American Entomological So- ders. While two or three of these species ciety 1: 85-150. have sometimes been taken at the same Day, M. C. 1974. A contribution to the taxonomy of time and place and remain recognizable, oc- the genus Anopliiis Dufour (Hymenoptera: Pom- pilidae) including a revision ofthe Palaeotropical casional specimens are somewhat interme- subgenus Orientanoplius Haupt. Bulletin of the diate in characters of the genitalia. Both A. British Museum (Natural History) 30: 373^04. subtruncatus and A. truncatus occur pri- Dreisbach, R. R. 1949. Psammocharini (Hymenop- marily on plains and prairies and may in- tera) of North America and the Antilles: Key to tergrade with A. brevihirta on the periphery genera; new species and key to males of Pompi- linus. Entomologica Americana 29: 1-58. of the sand dune habitats of the latter. Fur- . 1958. Two new species in the genus Pom- ther collecting and data on nesting behavior pilinus (Hymenoptera: Psammocharidae) with are needed to fully clarify the components photomicrographs of the genitalia and subgenital ofthis complex. Molecular studies might be plateofthe male. Entomological News69: 61—65. useful both for associating the sexes and for Evans, H. E. 1951. A taxonomic studyofthe Nearctic spiderwaspsbelongingtothetribePompilini(Hy- better understanding the degree to which in- menoptera: Pompilidae). Part II: genus Anoplius terbreeding occurs. Dufour. Transactionsofthe American Entomolog- A. brevihirta, even as restricted here, is ical Society 76: 207-361. a very variable species. Populations in . 1966. A revision ofthe Mexican and Central American spider wasps of the subfamily Pompi- widely spaced sand dune systems have di- linae (Hymenoptera: Pompilidae). Memoirs ofthe verged slightly in features of structure and American Entomological Society 20: 1^42. color, and there are also broadly geographic . 1970. Ecological-behavioral studies of the tendencies relating to the length and num- wasps ofJack.son Hole, Wyoming. Bulletin ofthe ber of comb spines on the fore basitarsus Museum ofComparative Zoology 140: 451-509. Gwynne, D. T. 1979. Nesting biology of the spider ofthe females, the length and abundance of wasps (Hymenoptera: Pompilidae) which prey on body pilosity, and the amount of orange burrowing wolfwolfspiders (Araneae: Lycosidae, coloration on the abdomen. Further collect- Geolycosa). Journal of Natural History 13: 681- ing of larger series from more sand dune 692. areas may serve to fill out details in the KromPbreoicne,edKi.ngVs. 1o9f53.theKilElntDoemviolloHgiilclaslWaSsopcsi,et1y952o.f spatial distribution of these variable char- Wa.shington 55: 113-135. acters. Kurczweski, E E. 1981. Observations on the nesting PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON 766 behaviors of spider-wasps in southern Florida prey selection. Journal ofthe Kansas Entomolog- (Hymenoptera: Pompilidae). FloridaEntomologist ical Society 41: 1-33. 64: 424-437. 1973. HostrecordsforsomeNorthAmerican Kurczewski, F. E. and E. J. Kurczewski. 1968. Host Pompilidae (Hymenoptera). Third supplement, records for some North American Pompilidae Tribe Pompilini. Journal of the Kansas Entomo- (Hymenoptera) with a discussion of factors in logical Society 46: 65-81.