American Museum Published by the American Museum of Natural History Central Park West at 79Th Street New York, N.Y

Home , Ammobatini, Exomalopsini, Exomalopsis, Leiopodus, Nomada, Nomadinae

NovitatesAMERICAN MUSEUM PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET NEW YORK, N.Y. 10024 U.S.A. NUMBER 2638 OCTOBER 26, 1977

JEROME G. ROZEN, JR. Immature Stages of and Ethological Observations on the Cleptoparasitic Bee Tribe Nomadini (Apoidea, Anthophoridae)

AMERICAN MUSEUM Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2638, pp. 1-16, figs. 1-27 October 26, 1977

Immnature Stages of and Ethological Observations on the Cleptoparasitic Bee Tribe Nomadini (Apoidea, Anthophoridae)

JEROME G. ROZEN, JR.'

ABSTRACT The mature larvae of the North American ses): M. sidaefloris (Exomalopsis near chlorina cleptoparasitic bees Melanomada sidaefloris Cockerell), T. penniger (Exomalopsis compac- (Cockerell), Triopasites penniger (Cockerell), and tula Cockerell), P. velutina (Exomalopsis solani Paranomada velutina Linsley, and the pupae of Cockerell), the South American Brachynomada M. sidaefloris, P. velutina, and Nomada species near argentina Holmberg (Psaenythia annulata are described taxonomically. Cladistic analysis of (Gerstaecker)). Biological data concerning egg the characteristics of these mature larvae demon- deposition, larval habits, search behavior of strates that Melanomada, Triopasites, Para- adults, and other features are given concerning nomada, and probably the South American these parasitic taxa. In general the biological in- Kelita are monophyletic and a sister group of the formation conforms to what is known of other Ammobatini, Nomada Epeolini, Holcopasites, Nomadinae. Neopasites, and Neolarra. Hitherto unreported Triopasites pasitura (Cockerell, 1935) is syn- host associations of various species of nomadine onymized with T. penniger (Cockerell, 1894). parasitic bees are as follows (host in parenthe- INTRODUCTION This paper provides comparative descriptions Ehrenfeld and Rozen (1977) concludes the of the mature larvae of the rare North American paper. cuckoo bee genera Melanomada, Paranomada, The Nomadini, consisting of approximately 10 and Triopasites and also includes an account of genera, most of which contain only a few species, the pupae of Nomada, Melanomada, and Para- have a limited distribution. Nomada, an excep- nomada In addition, ethological notes on these tion, contains hundreds of species and is abun- genera and on the South American genus Brachy- dantly distributed in the Holarctic Region, but nomada are given. An expanded interpretation of sparsely in the Southern Hemisphere. the phylogeny of the tribe and its relation to My research for this paper spanned many others of the subfamily based on the larvae de- years and was centered at the Southwestem Re- scribed here, those treated in Rozen (1966), and search Station of the American Museum of

'Deputy Director for Research and Curator of Hymenoptera, the American Museum of Natural History.

Copyright i The American Museum of Natural History 1977 ISSN 0003-0082 / Price $1.30 2 AMERICAN MUSEUM NOVITATES NO. 2638

Natural History in the southeast corner of Ari- H. Timberlake, Division of Biological Control, zona. While visiting the Station in 1966, my wife University of California, Riverside. and I excavated a nest of Exomalopsis compac- Larvae and adults of hosts and cleptoparasites, tula that had been invaded by several adult Trio- as well as cells and cocoons of host, are in the pasites penniger. We recovered two larvae of the American Museum of Natural History. parasitic bee from it. The following year, ably The research was supported by National Sci- assisted by Ms. Gisela Krueger, I recovered a ence Foundation Grant no. GB32193. number of larvae similar to those of Triopasites from a nest of Exomalopsis solani several miles ETHOLOGICAL OBSERVATIONS from the Triopasites site. We reared these and Most of this section of the paper records in- they turned out to be Paranomada velutina. So formation on Melanomada sidaefloris. In subse- similar were the larvae of these two parasitic bees quent subsections, however, fragments of infor- that I decided that the Triopasites-Exomalopsis mation are included on Paranomada, Triopasites, association of the previous year may have been and Brachynomada. Biological information on spurious, and that possibly I had obtained Para- Kelita, a related Chilean genus, is given by Rozen nomada from both sites. In 1976 I associated (1970) and Ehrenfeld and Rozen (1977). Lit- Melanomada sidaefloris with Exomalopsis near tle other information is available concerning the chlorina and discovered that the mature larva of ethology of the Nomadini, except for Nomada, Melanomada was remarkably similar to the other biological data of which have been published in two lots collected 10 years earlier. This led to a many places and summarized by Bohart (1970). careful re-examination of the anatomical differences among all three lots. As a result I now Melanomada sidaefloris (Cockerell) conclude that each lot does contain a distinct and separate taxon. The type of M. sidaefloris is a male; only fe- In addition to the two persons mentioned males were collected during the course of this above who assisted in the field work, I thank Mr. study. These females agree with the type in every Ronald J. McGinley for his efforts in finding and possible respect (including dense, decumbent plu- excavating the nest containing the Melanomada. mose setae on the venter of mesosoma) although A number of people contributed to the illustra- certain identification will require comparisons of tions presented here: Mrs. Marjorie Favreau, Mr. the same sexes. Further, the type was collected Ronald J. McGinley, and Mrs. Barbara Rozen. at Mesilla, New Mexico, only 130 miles from the The importance of the Southwestern Research nesting site, on the flowers of Sida, which also Station for long-term ethological research on grow at the nesting site. These geographic and natural populations cannot be overemphasized. ecological factors seem to support this identifi- Without such a base of operation our knowledge cation. Females from the nesting site are not con- of many groups of North American bees would specific with the female of M. heleniella (Cock- be far less than it is now. I thank Mr. Vincent erell) or the male type and female specimens Roth, Resident Director of that Station, for his identified by Cockerell of M. grindeliae (Cock- hospitality on my numerous visits there. erell), the other two named species in the genus. A major problem has been the proper identifi- Adult females of this species were first ob- cation of parasites and host bees. The following served flying over the barren, slightly sloping people have contributed either by lending speci- shoulder of a dirt roadway 26 miles south of mens or by providing identifications: Dr. Paul H. Animas, Hidalgo County, New Mexico, on Sep- Arnaud, California Academy of Sciences, San tember 11, 1976. Adults of Zacosmia maculata Francisco; Dr. Paul D. Hurd, Jr., Smithsonian (Cresson), Holcopasites apacheorum Hurd and Institution, Washington, D.C.; Dr. Charles D. Linsley, Nomada species, Triepeolus (two spe- Michener, The University of Kansas, Lawrence; cies), and Sphecodes species were often seen in Dr. Daniel Otte, The Academy of Natural Sci- the same area. ences of Philadelphia, Pennsylvania; Professor P. In addition to Exomalopsis near chlorina 1977 ROZEN: NOMADINI 3

Cockerell, the host of M. sidaefloris, E. sidae Cockerell also nested alongside the roadway1; two nests of E. sidae were found within 2 meters of the nest entrance ofE. near chlorina. The burrow entrances of both species were briefly inves- tigated by females of M. sidaefloris. I excavated the single nest E. chlorina, and recovered the mature larvae of M. sidaefloris described below. McGinley carefully dissected both nests of E. sidae, one of which contained only four or five cells with Exomalopsis larvae; the other nest had no larvae. On the basis of this information, I cannot determine whether M. sidaefloris is restricted to E. near chlorina or whether it is a parasite of both species ofExomalopsis. Nesting Biology of Host. The nest entrance and main tunnel of E. near chlorina were open and no tumulus was present. About 3 mm. in diameter, the main burrow descended vertically with some irregular turns. At a depth of 4 cm. FIG. 1. Cell and provisions of Exomalopsis and again at 10 cm. it branched, each case giving near chlorina, side view. rise to a short (3 and 5 cm.) obliquely descending burrow that ended blindly. The first cell, 27 cm. Over 24 hours, five females were collected deep and connected to a 7 cm. lateral, was open emerging from the burrow and probably repre- and partially filled with soil. All other cells were sent the entire adult population of the nest. No clumped from about 32 to 56 cm. and extended males were found within the nest. outward within a radius of about 9 cm. from the Provisions (fig. 1) were shaped in the charac- imaginary vertical continuation of the main bur- teristic form of those of E. chionura (Rozen and row. All laterals were filled with soil except for MacNeill, 1957, figs. 14) and were approxi- those leading to open cells; open laterals mean- mately 4.0 mm. long and 3.0 mm. wide. A semi- dered considerably. solid, the food mass rested with the rounded end Cells (fig. 1) were approximately 7.5 mm. affixed to the lower rear of the cell by moisture, long and 4.5 mm. in maximum diameter. Tilted with the "foot" touching the cell floor. It should approximately 45 degrees and with the closure be noted that the diagram presented by Stephen, end higher than the rear, the cell had a floor Bohart, and Torchio (1969, fig. 293) has the somewhat flatter than the curving ceiling. food erroneously oriented; Exomalopsis chionura Smooth and shiny, the cell surface was obviously like E. near chlorina had the food mass posi- waterproofed with a thin layer of transparent tioned as shown in figure 1. material, but there was no clear evidence of a The young Exomalopsis larva, as it fed, circled "built-in" cell-lining as reported for Exomalopsis the food mass and left a channel behind it. The chionura Cockerell (Rozen and MacNeill, 1957). "foot" of the mass was quickly consumed so that The closure was an unusually flat spiral being even small larvae were found circling a pillar of only slightly concave on the inside; it exhibited a food attached only at the rear of the cell. depression in the middle, perhaps where the Before finishing feeding, the last instar started tongue of the bee was last removed while form- defecating. Cocoon spinning apparently coming the closure. mences after food is consumed, but while the larva is still defecating, because, in addition to I Both species ofExomalopsis were identified by P. H. smears of fecal materials pressed to the cell wall Timberlake. enclosure feces were incorporated somewhat un- 4 AMERICAN MUSEUM NOVITATES NO. 2638 evenly in the cocoon fabric. The cocoon was predefecating larvae of known Exomalopsini and translucent whitish where no feces occurred, but Melitomini. the cocoon appeared yellow elsewhere because of The nest of E. near chlorina contained 53 the feces. The cocoon fabric was thin, one-lay- cells with contents as follows: ered, fragile, but not obviously fenestrated; it oc- Cells with Exomalopsis near chlorina cupied the entire lumen of the cell. Occupied cocoons 20 Ethology of Melanomada. In searching for Vacated cocoons 2 host nests, female M. sidaefloris flew moderately Feeding larvae 7 swiftly close to the ground. They tended to in- vestigate one area by flying in a meandering Total 29 fashion over it. They would then dart swiftly, in Cells with Melanomada sidaefloris a nearly straight line, without slowing or stop- Postdefecating larvae 3 ping, to a new area which they would explore Defecating larva 1 with a meandering flight. Although essentially Mature, predefecating larvae 5 the same size and color pattern as Holcopasites Feeding larva 1 apacheorum, females of M. sidaefloris could be distinguished in flight by their more rapid flight Total 10 with swift passage to a new area for exploration. Other Cells Holcopasites apacheorum females flew slower, Open cells 4 meandered more or less continuously, often re- Moldy cells 7 traced just-explored areas, and rarely departed to Meloid larva 1 a new section in a straight-line flight. Both Other 2 species stopped briefly at burrow entrances, Total which presumably had the attributes of the en- 14 trances of the appropriate hosts. Because most feces of exomalopsines are de- A female ofM. sidaefloris entered a burrow of posited before or during cocoon spinning, the oc- E. near chlorina and another, the burrow of E. cupied cocoons (20) contained postdefecating sidae, although in neither case did she remain in larvae or at least larvae that were nearly in that the nest sufficiently long to oviposit. condition. At first glance the distributions of de- Although no eggs ofMelanomada were found, velopmental stages of host and parasite seem to two parasitized cells exhibited an oblique slit in be at variance. Specifically, the ratio of 20 quies- the cell wall, similar to egg-deposition slits made cent, postdefecating larvae to seven feeding lar- by Nomada, Kelita (Rozen, 1970), and especially vae of E. near chlorina seems markedly different Holcopasites (Rozen, 1965, fig. 3). Both slits from three postdefecating larvae (one of which were less than 0.5 mm. from the cell closure and was still active) to six defecating or predefecating possessed an elongate flap of earth and cell lining larvae to one feeding larva ofM. sidaefloris. That about 0.5 mm. long attached along one side. In is to say, a large portion of parasites were still each case the hinge or the flap was the part active compared with hosts. This might lead one closest to the cell closure. There is little doubt to conclude that the parasite adult attacks host that Melanomada deposits its eggs as do the other nests for only part of the season that the host is genera mentioned above. provisioning. Such conclusions, however, are ten- The first larval instar, although not dis- uous because nothing is known concerning the covered, is presumed to possess elongate man- comparative rates of development of host and dibles as is characteristic of other Nomadini (and cuckoo bees. indeed most members of the subfamily). How- Melanomada sidaefloris larvae did not start ever, no data are available on the activities of the voiding feces until all food had been consumed, in feeding larvae. Mature, predefecating larvae and contrast to the hosts that started defecating dur- one intermediate-size larva of Melanomada pos- ing the last stadium while they continued to sessed bright yellow Malpighian tubules, a charac- feed. In all cases where the orientation of the teristic of the subfamily and also of at least mature larva of the cleptoparasite was observed, 1977 ROZEN: NOMADINI 5 its head was closest to the cell closure. In two Paranomada velutina Linsley cells, the feces were plastered as elongate, semi- moist pellets to the rear and top of the cell, Both P. velutina and P. nitida Linsley and whereas the rest of the cell lining was free of Michener are known from the vicinity of the such material. In contrast the host applied feces Southwestern Research Station. Paranomada as elongate, moist pellets to the entire cell wall velutina is the far more commonly encountered including the cell closure; as indicated above its species. Its males and females can be recognized feces are also incorporated in the cocoon fabric. by their dark integumental coloring and uninter- Melanomada sidaefloris, like all other known rupted metasomal hair bands. A female reared Nomadinae (except for Protepeolus and Isepeo- from the nest is unquestionably P. velutina even lus-condition for Leiopodus not known), does though the type specimen is a male. not spin a cocoon as reflected in the recessed More than a dozen larvae, including all instars labiomaxillary region of the mature larva. Mela- of this species, were excavated from a single large nomada sidaefloris overwinters as a totally quies- nest of Exomalopsis solani Cockerell, 1 mile cent, white larva. Several live larvae held in the north of Rodeo, Hidalgo County, New Mexico, laboratory pupated in May 1977. on August 28, 1967 by the author and Gisela Krueger. Oviposition incisions were found in the Triopasites penniger (Cockerell) cell walls of six cells: two midway to the rear of Males of the Triopasites collected in the vicin- the cell, two just before the wall curved at the ity of the nest site of the host bee appear con- rear, one two-thirds of the way to the rear of the specific with the male type of T. penniger in the cell, and one just inside the cell closure. In most, collection of The Academy of Natural Sciences the vacated chorion was still inserted in the in- of Philadelphia. cision. Only one of the six had a flap of cell Two nests of Exomalopsis compactula Cock- lining and soil, 0.5 mm. long and 0.3 mm. wide, erell (kindly identified by P. H. Timberlake) were attached along one side to the cell wall as is visited by Triopasites penniger at 5 miles north characteristic of Nomada and Melanomada. As of Rodeo, Hidalgo County, New Mexico, be- these incisions were examined in cells that had tween August 30, 1966 and September 3, 1966. been preserved in the American Museum of Nat- These nests yielded two cells containing the lar- ural History for nine years, I cannot determine vae described in the present paper. Although this whether the flaps had accidentally broken off is the first definite host association of Triopasi- after excavation or whether the cleptoparasite tes, Cockerell had tentatively associated Triopa- does not normally leave a flap when ovipositing. sites pasitura (Cockerell), a junior synonym of First instars possessed a pigmented, progna- penniger, New Synonymy, with E. compactula thous head capsule and elongate, sickle-shaped in 1935. mandibles (as is characteristic of Nomada, Oreo- One cell had a thick covering of fecal material pasites, and indeed most Nomadinae) with which plastered over the cell wall, rear, and closure. In the host egg or early larval instar is killed. One neither cell was there a cocoon. In another Paranomada first instar was found in a cell con- Exomalopsis cell was found an oval oviposition taining a host egg and another was found in a cell incision, 0.5 mm. long by 0.3 mm. wide, in the with a dead, early instar host larva. Three mature cell wall, about two-thirds the distance from the larvae were kept alive: one pupated on June 29, closure to the rear. A shriveled, somewhat moldy 1968, and the adult emerged July 11, 1968; chorion, attached to what might have been a flap another died while trying to pupate and was pre- of cell lining affixed to the feces, was pulled served on May 23, 1969; and the last pupated on away from the incision, a fact suggesting that the July 8, 1968, and was preserved with pigmented egg insertion had been similar to that of other eyes on July 15, 1968. Nomadini. It is not clear whether the cell had This is believed to be the first definite host been inhabited by a postdefecating Triopasites association of any species of Paranomada. How- larva or whether the egg had died and the Exo- ever, Linsley (1945) recorded that P. H. Timber- malopsis had matured. lake suggested that Exomalopsis verbesinae 6 AMERICAN MUSEUM NOVITATES NO. 2638

Cockerell may be the host of Paranomada cali- pigmented, sclerotized, prognathous head cap- fornica Linsley. sule, long slender curved mandibles and elongate The adaptive significance of the dorsoven- labral tubercles. trally flattened body, especially of adult females The first cell in the nest had been filled with of Paranomada, has been a mystery. I suggest soil, suggesting that this behavior pattern may be that this body shape may enable adult clepto- a method by which the host destroys eggs of parasites to appress themselves against burrow cleptoparasites; similarly filled cells have been walls and thereby permit host adults, which nest found in the host nests of other Nomadinae. The communally, to pass without detecting the Para- fact that at least two and very possibly three cells nomada The highly polished integument of the had already been completed and provisioned and vertex, scutum, and other areas as well as the that the Brachynomada female was collected as it posteriorly directed setae on the venter of the was trying to enter the nest opening covered with parasite would further assist passage. Although a drinking glass strongly indicates that a parasite this is probably true, a question remains as to female learns the location of a nest and returns why Melanomada and Triopasites have normally to it time and again to parasitize successive cells convex bodies, as both of these genera are also as they are being provisioned. associated with communally nesting Exoma- lopsis. Conclusions Concerning Ethology of the Nomadini Brachynomada near argentina Holmberg The biological information presented above, These specimens were identified by compari- although previously unrecorded, yields few sur- son with specimens in the collection at the Uni- prises because it is similar to what is known of versidad do Parana, Brazil, with the kind assis- Nomada and of other tribes in the subfamily. tance of Padre J. S. Moure. They are similar to B. All genera about which there is information argentina but are darker and somewhat more (Nomada, Melanomada, Triopasites, Paranomada, densely punctate. Kelita, and Brachynomada) insert their eggs into Three females of this species were collected a notch in the cell wall of the host. This in- flying near several nests of Psaenythia annulata cision is probably made by the terminalia of the (Gerstaecker) (identified by the author by com- female. Some, and perhaps all, place the egg un- parisons with specimens in the collection of the der a flap composed of cell lining and soil which Universidad do Parana) at Fumas, Vila Velha, is attached on one side to the cell wall. The posi- Parana, Brazil, on February 9, 1974. One was tion of the egg in the cell seems variable in the trying to enter an open main nest burrow even few cases where sufficient samples are available though a plastic drinking glass had been inverted (Nomada and Paranomada). In none of the cases over the entrance. A first instar larva, almost cer- presented here was more than one cleptoparasite tainly of this species, was found attacking an egg egg found in a cell. However, Linsley and of the host in one cell of that nest, and another MacSwain (1955), who made a careful study of completed cell possessed an egg inserted in the Nomada opacella Timberlake, indicated that a fe- cell wall. The egg, 0.9 mm. long and 0.3 mm. at male normally deposited two eggs to a cell and the widest part near one end (presumably the that the eggs were characteristically placed one at anterior end) tapered gradually to the other end the base (rear) of the cell and the other near the which was narrowly rounded. Unfortunately the middle, although there was some variability in egg, which was white and had a smooth chorion, position. was dislodged with a flap of earth and cell lining The first instars of most and probably all at the time of excavation and was still attached Nomadini have prognathous, pigmented, sclero- by its tapered end to the flap. The flap had come tized head capsules, with elongate sickle-shaped from a notch in the cell wall just inside the clo- mandibles with which they kill the young host. sure. In all likelihood the egg had been inserted Almost certainly all first instars are able to crawl; in the cell wall very much as are the eggs of after hatching they move from the cell wall onto Nomada or Holcopasites (Rozen, 1965). The first the food mass where they destroy the host. After instar, like that of other nomadines, possessed a feeding on the provisions, they overwinter as a 1977 ROZEN: NOMADINI 7

mature, quiescent larva (not known for Brachy- and lateral portions of maxillae spiculate. Head nomada and for many Nomada). No species is capsule not especially elongate as seen in lateral known to have more than one generation a year view but definitely not so wide as that ofP. velu- although this feature might well be species-spe- tina; vertex as seen in lateral view produced on cific as seems to be the case with some genera in each side above antenna; produced area appear- other nomadine tribes. No species spins a cocoon ing as a wrinkled mound; this production con- (not known for Brachynomada). siderably more pronounced than in Kelita tuber- Host associations are variable. Melanomada, culata Ehrenfeld and Rozen (1977). Ten- Paranomada, and Triopasites attack different torium complete but very thin; anterior tentorial species of Exomalopsis. Brachynomada and Ke- pits very small, best identified by position of an- lita are only associated with panurgine bees. terior tentorial arms; these pits very close to Nomada is often thought to be a parasite of anterior mandibular articulation; posterior pits Andrena, but various species have also been as- small but evident, situated just behind hyposto- sociated with Panurgus (Rozen, 1971), Dasy- mal ridge and posterior thickening of head cap- poda, Nomia, Halictus, and Eucera (Bohart, sule; posterior thickening of head capsule not 1970) and probably other Nomada will be associ- evident except in vicinity of posterior tentorial ated with still other hosts. The hosts of Mela- pits where thickening is curved continuation of nomada, Triopasites, Brachynomada, and some hypostomal ridge; hypostomal ridge itself thin Nomada (Panurgus and Halictus) nest commu- but evident. Parietal bands weak. Antennal papil- nally, whereas the host of Kelita, Brachynomada, la a rather pronounced convexity with four, and and other Nomada are solitary so far as we now in one case five, sensilla; antennal protuberances know. Because females of nomadines probably moderately developed. Labrum projecting mod- always enter and oviposit in cells that are in the erately strongly as seen in lateral view and process of being provisioned, numerous host bearing two sharp-pointed tubercles directed females coming and going in a communal nest anteriorly. Mandibles (figs. 6-8) slender and elon- might require special adaptations on the part of gate, tapering to bladelike apex; mandible seen the female cleptoparasite to avoid attacks by in adoral or aboral view somewhat S-shaped with host adults. With the exception of the flattened apex curving strongly downward; dorsal apical body of the female Paranomada, we are not yet edge finely and evenly serrate; ventral apical edge able to recognize these adaptations. somewhat more coarsely serrate; cusp not produced; dorsal and dorsal adoral surfaces dentate; DESCRIPTIONS OF MATURE LARVAE mandibular adductor apodeme present. Maxillae This section treats mature larvae of available recessed, broadly fused with labrum; maxillary Nomadini. Because of their small size, first in- palpus moderately well developed; galea absent. stars are better studied with a scanning electron Hypopharynx large and well defined, projecting microscope and are therefore reserved for a fur- farther than either maxillae or rest of labium as ther investigation that will deal with first stage seen in lateral view (fig. 4). Labial palpus not forms, not only of Nomadini, but of all the No- defined but apparently represented by a single madinae. Mature larvae of Nomada have been sensilla. Salivary opening a small simple orifice described by Rozen (1966) and those of Kelita without lips. by Ehrenfeld and Rozen (1977), and therefore Body. Integument spiculate ventrally es- are not included here. pecially behind the head; apices of most if not all dorsal tubercles with a few minute setae; abdomi- Melanomada nal segment X with scattered, very minute setae. sidaefloris (Cockerell) Form (fig. 2) moderately robust, with head of Figures 2-8 postdefecating larva notably large compared with Diagnosis. See diagnosis of Paranomada velu- body, as is also the case with Kelita tuberculata; tina and Triopasites pasitura. body shape similar to that of K. tuberculata Head (figs. 3, 4). Integument very faintly pig- (Ehrenfeld and Rozen, 1977), that is, not tapering mented; internal ridges and mandibular tips more posteriorly as seen in lateral view; interseg- darkly pigmented; epipharynx, hypopharynx, mental lines moderately deeply incised; paired 8 AMERICAN MUSEUM NOVITATES NO. 2638

FIGS. 2-8. Mature larva of Melanomada sidaefloris. 2. Entire larva, lateral view. 3. Head, frontal view. 4. Head, lateral view. 5. Spiracle, side view. 6-8. Right mandible, dorsal, inner, and ventral views, respectively. Scale refers to figure 2. conical, rounding dorsal tubercles present on transverse, smooth lips. Spiracles (fig. 5) moder- most body segments; tubercles low on predefe- ately large, not on distinct elevations; atrium sub- cating larva but moderately pronounced on post- globular, projecting above body surface, with defecating larva; dorsal intrasegmental lines rim; atrial wall with rows of indistinct spicules; scarcely evident; abdominal segment X not pro- these rows in some cases vaguely reticulate; peri- duced ventrally, somewhat dorsal in attachment treme present, moderately narrow; primary to segment IX; segment X moderately large and tracheal opening with well-developed collar; rounded ventrally; anus transverse, somewhat subatrium annulate, moderately short and thin, dorsal in position; perianal area appearing as consisting of seven to eight chambers. Imaginal 1977 ROZEN: NOMADINI 9 discs of genitalia not visible, in spite of number 12-14) slender but not so elongate as in M. sidae- of mature larvae available for study. floris; apex as seen in ventral and dorsal view not Material Studied. Four postdefecating larvae, so thin as that of M. sidaefloris; in aboral or four predefecating larvae, 26 miles south of adoral view mandible not S-shaped; both dorsal Animas, Hidalgo Co., New Mexico, September and ventral apical edges bearing slender teeth; 12-13, 1976, from cells of Exomalopsis near apex narrowly rounded. Maxillary palpus larger chlorina (J. G. Rozen). than that of either M. sidaefloris or P. velutina. Labial palpus, though poorly defined, apparently Triopasites penniger (Cockerell) slightly more differentiated than is case with Figures 9-14 other two species. Body (fig. 9). As described for M. sidaefloris This description is based on two specimens except for following: Although dorsal tubercles recovered from the nest of Exomalopsis com- with a few minute setae, presence of setae on pactula in 1966. The illustrations of the live larva abdominal segment X not known. Spiracles with in lateral view and of the head capsule were pre- atrial wall having rows of distinct spicules, but pared in general outline at that time. An attempt without reticulations; subatrium with approxi- was then made to rear the specimens with the mately nine chambers and in general somewhat idea of using cast skins to fill in details, particu- longer and wider than that of M. sidaefloris, larly of the head capsule and spiracles. Unfortu- more similar to that of P. velutina. nately, both specimens died as they were pre- Material Studied. Two postdefecating larvae, 5 paring to pupate. Many anatomical features have miles north of Rodeo, Hidalgo Co., New Mexico, now been recorded from the preserved but dis- August 30, 1966, from cells ofExomalopsis com- torted specimens, but the following description pactula (J. G. and B. L. Rozen). and illustrations lack some information that would have been available if well preserved, Paranomada velutina Linsley unmodified specimens were compared and con- Figures 15-20 trasted with those of Paranomada velutina and Melanomada sidaefloris. Diagnosis. The larva of this species closely re- Diagnosis. Triopasites penniger seems to agree sembles that ofMelanomada sidaefloris. It differs even more closely with M. sidaefloris than does in subtle ways as indicated in the following ac- P. velutina. Like M. sidaefloris, it can be distin- count. The most easily observable diagnostic guished from P. velutina by its distinctly nar- feature concerns the width of the head capsule, rower head capsule and the more pronounced that of P. velutina being considerably greater produced areas on the vertex. Triopasites pen- than that ofM. sidaefloris. Further, the produced niger differs from M. sidaefloris also by the areas on each side of the vertex are less pro- different shape of the mandible and the distinct nounced in P. velutina. rows of spicules on the atrial wall. Head (figs. 16-17). As described for M. sidae- Head (figs. 10, 11). As described for M. sidaefloris except for following: Integument slightly floris except for following: Vertex as seen in more pigmented; spiculations on hypopharynx, lateral view apparently not produced strongly as and lateral portions of maxillae more extensive. is case with M. sidaefloris; wrinkling of integu- Head capsule as seen from front (fig. 16) much ment on produced areas apparently somewhat broader than that of M. sidaefloris; produced less than that of M. sidaefloris. Condition of ten- areas on vertex less pronounced than in M. sidae- torium not known but almost certainly complete floris, but more evident than in Kelita tubercu- and very thin, in keeping with similarities be- lata. Posterior thickening of head capsule faint, tween it and M. sidaefloris and with respect to but somewhat more pronounced than in M. si- internal ridges of head capsule. Parietal bands daefloris. Parietal bands scarcely evident. Anten- weak, but perhaps slightly more evident than nal papilla with four sensilla; antennal protuber- those of M. sidaefloris. Antennal papilla quite ances poorly developed. Labrum on most pronounced, more so than in either M. sidaefloris specimens projecting somewhat less than in M. or P. velutina, with four sensilla. Mandibles (figs. sidaefloris with two sharp-pointed tubercles 10 AMERICAN MUSEUM NOVITATES NO. 2638

FIGS. 9-14. Mature larva of Triopasites penniger. 9. Entire larva, lateral view. 10. Head, frontal view. 11. Head, lateral view. 12-14. Right mandible, dorsal, inner, and ventral views, respectively. Scale refers to figure 9. directed somewhat more downward than in M. atrial wall not evident; subatrium longer than in sidaefloris. Mandible (figs. 18-20) with bladelike M. sidaefloris, usually consisting of 10 to 20 apex somewhat broader and mandible in general chambers; subatrium tending to be somewhat less tapering than in M. sidaefloris; ventral apical wider than that of M. sidaefloris. edge with serrations apparently less distinct than Material Studied. Three postdefecating, eight in M. sidaefioris. predefecating larvae, 1 mile north of Rodeo, Body (fig. 15). As described forM. sidaefloris Hidalgo Co., New Mexico, August 28, 1967, except for following: Head of postdefecating from nest of Exomalopsis solani-nest No. 2 (J. larva not especially large compared with body; G. Rozen and G. Krueger). paired dorsal tubercles somewhat more differentiated from dorsum than is case with M. sidae- DESCRIPTION OF PUPAE flonis. Spiracles having atrial wall with short ridges bearing a number of spicules; reticulations of Pupae of Nomada species, Paranomada velu- 1 977 ROZEN: NOMADINI 11

L.Omm

FIGS. 15-20. Mature larva of Paranomada velutina. 15. Entire larva, lateral view. 16. Head, frontal view. 17. Head, lateral view. 18-20. Right mandible, dorsal, inner, and ventral views, respectively. Scale refers to figure 15. tina and Melanomada sidaefloris are described what more pronounced tubercles on the vertex here. A preliminary key to the pupae of the No- are especially useful in distinguishing its pupa madinae was presented in Rozen and McGinley from that ofParanomada. (1974) which separates the pupae of Nomada and Paranomada in couplet 4. Melanomada also Melanomada sidaefloris (Cockerell) runs to couplet 4. Among all known pupae of Nomadinae only Melanomada and Paranomada Figures 25-26 have a strongly projecting tubercle on each tegu- Head. Scape and frons without distinct tula. As is pointed out in the descriptions the pro- bercles, although integument appearing some- nounced wing tubercle ofMelanomada and somewhat lumpy; vertex with large and small low 12 AMERICAN MUSEUM NOVITATES NO. 2638

FIGS. 21-26. Pupae of Nomadini. 21. Paranomada velutina, male, lateral view. 22. Same, dorsal view. 23. Nomada species, female, dorsal view. 24. Same, lateral view. 25. Melanomada sidaefloris, female, dorsal view. 26. Same, lateral view. Scale refers to all figures. protuberances, all of which are rounded; genal ventral surface; dorsal surface without distinct tubercle apparently absent; each apical, lateral tubercle. angle of clypeus without tubercle; mandible with Mesosoma Setae absent. Lateral angles of only indistinct, somewhat irregular swelling on pronotum moderately produced, perhaps some- 1977 ROZEN: NOMADINI 13 what more pronounced than those ofP. velutina; num without tubercles; mesoscutum without posterior lobes of pronotum strongly produced. either large or small tubercles; axilla rounded, Mesepisternum without tubercles although with faintly produced; scutellum without tubercles; some vague swellings; mesoscutum without dis- metanotum without tubercles; propodeum with- tinct tubercles, but with row of irregular, some- out protuberances. Tegula with conspicuous, dor- what raised striations extending longitudinally on sally projecting tubercle; wings without distinct either side of median line; axilla rounded, not tubercle although anterior wing with vague swell- produced; scutellum with low swelling just lat- ing along anterior margin approximately two- erad of median line; propodeum without protu- thirds of way to apex. All coxae without tuber- berances. Tegula with conspicuous, dorsally cles; trochanters, each with small rounded apical projecting tubercle, similar to that ofP. velutina; tubercle; tibiae without tubercles and with only anterior wing with distinct, acutely rounded vague swellings, mostly at apex; fore-basitarsus tubercle along anterior margin, approximately with conspicuous, rounded tubercle halfway two-thirds of way to apex. Fore-coxa with apical along segment (probably a sex-limited feature, tubercle; mid- and hind coxae each with vague corresponding to special brush found only in apical swelling; all trochanters each with small, male). rounded apical tubercle which is somewhat more Metasoma. Terga I-VI with apical row of small conspicuous than that of P. velutina; fore- and tubercles most of which are sharp-pointed and mid-tibiae each with small, vague, indistinct slightly pigmented at apex; sternal tubercles ab- swellings; hind tibia with moderate sized, low, sent; terminal spine not present. rounded tubercle at base, with somewhat less Material Studied. One male, 1 mile north of distinct one apically, and with number of vari- Rodeo, Hidalgo Co., New Mexico, collected as able-sized, small, rounded tubercles along outer larva, August 28, 1967, pupated July 8, 1968, surface; basitarsi without tubercles. preserved July 15, 1968 (J. G. Rozen and G. Metasoma. Unlike in Nomada and Para- Krueger). nomada, metasoma very large in comparison with head and mesasoma so that pupa appearing NOMADA SPECIES physogastric. Tergum I with vague, apical row of Figures 23-24 indistinct tubercles; terga II-V with apical row of small tubercles, most of which are sharp-pointed; Head. Scape and frons without tubercles; ver- sterna without tubercles; terminal spine not tex just mesiad of compound eye with approxi- present. mately five sharp-pointed, apically pigmented Material Studied. One female, 26 mi. south of tubercles; genal tubercles absent; each apicolat- Animas, Hidalgo Co., New Mexico, collected as eral angle of clypeus without tubercle; mandible larva, Sept. 2, 1976, pupated May 1977 (J. G. without tubercles or swellings. Rozen); this specimen was drawn and described Mesosoma. Setae absent. Lateral angles of while it was still alive. pronotum moderately produced; posterior lobe of pronotum moderately produced; mesepisternum without tubercles; mesoscutum with series Paranomada velutina Linsley of sharp-pointed, apically pigmented tubercles on Figures 21-22 each side of midline; axilla rounded, scarcely pro- Head. Scape and frons without tubercles; ver- duced; scutellum produced on each side and tex in vicinity of lateral ocellus with vague bearing approximately four or five sharp-pointed, mound; genal tubercles absent; each apicolateral apically pigmented tubercles on each side; nei- angle of clypeus without tubercle; mandible with ther metanotum nor propodeum produced. large tubercle on ventral surface; dorsal surface Tegula without tubercles; wings without tuber- evenly curved, without tubercles or swellings. cles or swelling. Each coxa with small nonpig- Mesosoma. Setae absent. Lateral angles of mented apical tubercle; trochanters without pronotum moderately produced; posterior lobe distinct tubercles; each tibia with dorsal, apical, of pronotum not greatly produced. Mesepister- small, sharp-pointed, apically pigmented tubercle; 14 AMERICAN MUSEUM NOVITATES NO. 2638 fore- and mid-tibiae with several small, apically of characters (1) and (3) has been determined pigmented tubercles; hind tibia with more than by out-group comparisons with other families of five tubercles; fore basitarsus without tubercle. bees; that of (2) and (4) by correlation with Metasoma. Tergum I with a single, small, other characters. sharp-pointed, pigmented tubercle; terga II-V Figure 27 is an expanded interpretation of with apical row of sharp-pointed, apically pig- figure 1 in Rozen (1966). The hypothetical mented tubercles; tergum VI with several apically ancestral Nomadinae (fig. 27, 1) possessed the pigmented, sharp-pointed tubercles toward apex; plesiomorphic character states listed above and sterna II-IV with apical row of small, sharp- also those listed in Rozen (1966, p. 6). (The pointed, apically pigmented tubercles best de- synapomorphies defining the Nomadinae will be veloped on each side of midline; terminal spine listed in a forthcoming paper by Rozen, Eick- moderately elongate. wort, and Eickwort, 1978.) The lineage (fig. Material Studied. One pupa, Closter, Bergen other nomodines Co., New Jersey, June 15, 1971, from burrow of Andrena morrisonella Vierick (J. G. Rozen).

CONCLUSIONS AMNMOB ATINI 7 Phylogeny of the Nomadini Based on Mature Larvae EPEOLINI Nomodo Until recently the mature larva of only No- mada among all the genera of Nomadini was known. The phylogenetic significance of the Nomada larva was interpreted by Rozen (1966) 5 in light of other known larvae of the Nomadinae. Poranomada Now, both in the present paper and in a study of Kelita (Ehrenfeld and Rozen, 1977), larvae of 14 other Nomadini are described, permitting a broader interpretation of interrelationships within the tribe and within the subfamily. Of course, such interpretations are tentative as they are Melanomado based upon a limited number of characters whose Isepeolus Trioposites plesiomorphic-apomorphic states can be judged with some degree of reliability. In Rozen (1966, p. 6) the polarity (i.e., plesiomorphic-apomor- 2 phic conditions) of a series of characters was given. To that series I now add the following, listing the plesiomorphic condition first: (1) Normally long mandible of Isepeolus and I Scale Melanomada complex (Melanomada, Paranomada and Triopasites) in contrast to short mandibles found in all other known larval Nomadinae; (2) well-developed dorsal tubercles as found in Isepe- olus and Melanomada complex (including Kelita) as opposed to no dorsal body tubercles; (3) FIG. 27. Phylogenetic interpretation of the setae of Nomadinae based on mature larvae. Numbers minute, widely scattered body Isepeolus refer to text discussion. Sequence of branching and Melanomada complex, including Kelita, as has been determined by cladistic analysis. Length opposed to lack of body setae of all other No- of line approximates relative amount of anatomi- madinae; and (4) vertex recessed as in all Noma- cal change. Direction of line and size of angle dinae except for the Melanomada complex in have no meaning. Scale refers to change in one contrast to vertex with protrusions. The polarity character state. 1 977 ROZEN: NOMADINI 15

27, 2) that gave rise to all the nomadines after that Nomada is in many ways different from the Isepeolus branched off exhibited the following Melanomada complex and therefore that the apomorphic changes: tentorium reduced, maxil- Melanomada complex, Brachynomada, Hespero- lary palpi short, labium recessed, not divided into nomada, and perhaps Kelita should be grouped as prementum and postmentum, labial palpi lost or a separate tribe. In this respect Ehrenfeld and almost so, and hypopharynx exceeding labium. Rozen (1977) have already pointed out, using This lineage then split, one branch (fig. 27, 3) characteristics of adults, that Nomada and Hypo- giving rise to the Melanomada complex (i.e., chrotaenia lineage is a sister group to the rest of Melanomada, Triopasites, and Paranomada) with the Nomadini lineage (i.e., the lineage that gave the development of protrusions on the vertex. rise to Kelita, Brachynomada, Hesperonomada, Unfortunately, this feature is indeterminate in Melanomada, Triopasites, and Paranomada). Kelita, and other characters of its head capsule, Their diagram (Ehrenfeld and Rozen, 1977, especially the shape of the mandible, cannot be fig. 26), though appearing very different from reviewed because the head capsule was acci- figure 27 in the present paper, is consistent with dentally destroyed before such features were it in terms of sequence of branching. Their paper studied. Melanomada, Triopasites, and Parano- did not incorporate an analysis of the adult mada are quite similar to one another, and features of the Ammobatini in relationship to among the differences that they exhibit in the those of the Nomadini. This analysis is now larvae, only the abnormally wide capsule (fig. 27, crucial as a test of the hypothesized relationships 4) of Paranomada can be identified as apomor- presented here. phic. The other branch (fig. 27, 5) underwent con- siderable change. The hypostomal ridge became LITERATURE CITED body setae were weak, mandibles became short, Bohart, George E. lost, and the body tubercles were greatly reduced 1970. The evolution of parasitism among or lost. This branch then split twice, giving rise to bees. 41st Faculty Honor Lecture de- three lines, as described by Rozen (1966). One livered at Utah State University. Utah (fig. 27, 6) produced the Epeolini, another (fig. State Univ., Logan, pp. 1-33, figs. 1-28. 27, 7) the other Nomadinae (Holcopasites, Neo- Cockerell, T. D. A. pasites and Neolarra), and the third (fig. 27, 8) 1935. New records of bees. Amer. Mus. No- Nomada and the Ammobatini (Pseudodichroa, vitates, no. 766, p. 6. Ammobates, Morgania, and Oreopasites) (Rozen Ehrenfeld, Joan, and Jerome G. Rozen, Jr. and McGinley, 1974). Evidence is not on hand to 1977. The cuckoo bee genus Kelita, its sys- indicate which branching occurred first so that tematics, biology and larvae (Antho- Nomadinae). Amer. Mus. are presented as being simultaneous in phoridae, they Novitates, no. 2631, pp. 1-24, 42 figs. figure 27. Linsley, E. Gorton The synapomorphic feature shared by No- 1945. A new species of Paranomada with mada and the Ammobatini (fig. 27, 8) that is, the notes on Melecta thoracica Cresson. double transverse line at the posterior margin of Ent. News, vol. 56, pp. 149-153, 1 fig. the head, is unique among bee larvae and there- Linsley, E. Gorton, and J. W. MacSwain fore is fairly convincing evidence that Nomada 1955. The hiabits of Nomada opacella Tim- and the ammobatines are sister groups. This and berlake with notes on other species the fact that the Melanomada complex (probably (Hymenoptera: Anthophoridae). Re- including Kelita and for that matter also Brachy- printed from Wasmann Jour. Biol., vol. 13, no. 2, pp. 253-276. nomada) is a monophyletic group within the Rozen, Jerome G., Jr. Nomadinae, as shown here, would require the 1965. Biological notes on the cuckoo bee reworking of the tribal classification of the sub- genera Holcopasites and Neolarra (Hy- family if the classifier adheres to strict cladistic menoptera: Apoidea). Jour. New York classification. Even if strict cladistic classification Ent. Soc., vol. 73, pp. 87-91, 4 figs. is not followed a re-examination of features of 1966. The larvae of the Anthophoridae (Hy- adults and immatures may lead us to conclude menoptera, Apoidea). Part 2. The No- 16 AMERICAN MUSEUM NOVITATES NO. 2638

madinae. Amer. Mus. Novitates, no. (A nthophorula) chionura Cockerell, in- 2244, pp. 1-38, 83 figs. cluding a comparison of the biology of 1971. Biology and immature stages of Mo- Exomalopsis with that of other an- roccan panurgine bees (Hymenoptera, thophorid groups. Ann. Ent. Soc. Apoidea). Amer. Mus. Novitates, no. Amer., vol. 50, pp. 522-529, figs. 1-14. 2457, pp. 1-37, 50 figs. Rozen, Jerome G., Jr., and Ronald J. McGinley Rozen, Jerome G., Jr., Kathleen R. Eickwort, 1974. Systematics of ammobatine bees based and George C. Eickwort on their mature larvae and pupae 1978. The bionomics and immature stages (Hymenoptera, Anthophoridae, Noma- of the cleptoparasitic bee genus Pro- dinae). Amer. Mus. Novitates, no. tepeolus (Anthophoridae, Nomadinae). 2551, pp. 1-16, 33 figs. Amer. Mus. Novitates, no. 2640, pp. Stephen, W. P., G. E. Bohart, and P. F. Torchio 1-24, 7 figs. 1969. The biology and external morphology Rozen, Jerome G., Jr., and C. Don MacNeill of bees. Corvallis, Oregon State Univ., 1957. Biological observations on Exomalopsis pp. 1-140, figs. 1-320.

.14