Novitatesamerican MUSEUM PUBLISHED by the AMERICAN MUSEUM of NATURAL HISTORY CENTRAL PARK WEST at 79TH STREET, NEW YORK, N.Y

Home , Brachynomada, Ericrocidini, Ericrocis, Melecta, Melectini, Nomadinae

NovitatesAMERICAN MUSEUM PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 3029, 36 pp., 67 figures, 3 tables November 27, 1991

Evolution of Cleptoparasitism in Anthophorid Bees as Revealed by Their Mode of Parasitism and First Instars (Hymenoptera: Apoidea)

JEROME G. ROZEN, JR.1

CONTENTS

Abstract ...... 2 Introduction ...... 2 Acknowledgments ...... 3 Historical Background ...... 4 Evolution of Cleptoparasitism in the Anthophoridae ...... 6 Systematics of Cleptoparasitic First-Instar Anthophoridae ...... 12 Methods ...... 12 Description of the Nomadinae Based on First Instars ...... 13 Description of the Protepeolini Based on the First Instar ...... 13 Description of the Melectini Based on First Instars ...... 17 Xeromelecta (Melectomorpha) californica (Cresson) ...... 17 Melecta separata callura (Cockerell) ...... 20 Melecta pacifica fulvida Cresson ...... 20 Thyreus lieftincki Rozen ...... 22 Zacosmia maculata (Cresson) ...... 22 Description of the Rhathymini Based on First Instars ...... 24 Rhathymus bicolor Lepeletier ...... 25 Description of the Isepeolini Based on the First Instar ...... 26 Isepeolus viperinus (Holmberg) ...... 27

lCurator, Department of Entomology, American Museum of Natural History.

Description of the Ericrocidini Based on First Instars ...... 29 Aglaomelissa duckei (Friese) ...... 30 Mesoplia rujipes (Perty) ...... 32 Ericrocis lata (Cresson) or pintada Snelling and Zavortink ...... 33 References ...... 34

ABSTRACT The methods by which cleptoparasitic antho- will rest on discovering sister-group relationships phorid bees introduce their eggs into host nests with pollen-carrying anthophorids. and the anatomical and behavioral adaptations of Comparative descriptions of the six cleptoparas- their first instars to kill host immatures are de- itic groups are based on descriptions (or redes- scribed and analyzed. The information is derived criptions) and illustrations of the hospicidal first from the literature, from new field studies, and instars ofthe following species: Xeromelecta (Me- from first instars described here. The study iden- lectomorpha) californica (Cresson), Melecta se- tifies six distinct cleptoparasitic groups: the No- parata callura (Cockerell), Melecta pacificafulvida madinae, Protepeolini, Melectini, Rhathymini, Is- Cresson, Thyreus lieftincki Rozen, Zacosmia ma- epeolini, and Ericrocidini, and each is probably culata (Cresson), Rhathymus bicolor Lepeletier, monophyletic. The lack ofsimilar information re- Isepeolus viperinus (Holmberg), Aglaomelissa garding Coelioxoides and the Osirini (two groups duckei (Friese), Mesoplia rufipes (Perty), Ericrocis recently identified as being unrelated to the No- lata (Cresson) or pintada Snelling and Zavortink. madinae) leaves open the possibility that the An- Illustrations offirst instar Triepeolusgrandis Friese, thophoridae may contain eight cleptoparasitic Townsendiella pulchra Crawford, and Protepeolus groups. It is suggested that each group represents singularis Linsley and Michener are presented to a separate evolutionary development of clepto- exemplify their lineages. parasitism in the family, although some evidence Females of Aglaomelissa duckei and Mesoplia points to a common parasitic ancestor for the Is- rufipes are reported to introduce their eggs into epeolini and Ericrocidini. Proof of separate evo- host cells through small holes they make in the lutionary origin ofcleptoparasitism in each group closures.

INTRODUCTION This paper investigates the evolutionary ism and first instars is derived from previ- relationships of the cleptoparasitic antho- ously published accounts and from new data phorids through an analysis of (1) the mode presented below under Systematics of Clep- of parasitism of cleptoparasitic taxa and (2) toparasitic First-Instar Anthophoridae. the anatomy of their first instars. Mode of Because the apomorphies used here are as- parasitism is the manner in which the clep- sociated only with cleptoparasitism, the toparasitic female introduces her egg into the question as to what pollen-gathering groups host cell, and how the host egg or larva is gave rise to the cleptoparasitic lineages is be- eliminated, allowing the cleptoparasitic off- yond the scope of this study. We as yet know spring to develop on the stored provisions. too little about first instars of nonparasitic The first instars of all known cleptoparasitic bees to recognize other kinds of first-instar anthophorids kill the hosts' young, and the synapomorphies that link cleptoparasites and anatomies ofthese hospicidal first instars re- food-gathering lineages. veal striking modifications to accomplish this Cleptoparasitism is the relationship in purpose. Hence the morphology ofthese first which the young ofone bee species feeds and instars is an integral part ofthe mode ofpar- develops on the food stored for the young of asitism. Knowledge about mode of parasit- another species. This is different from social 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 3

parasitism in which the parasite female lives not subsequent instars, shows the most pro- in the nest of the host species and her off- nounced anatomical modifications involved spring are cared for by the female or females with its hospicidal activities.2 Cleptoparas- of the host species. In the Anthophoridae, ites in other bee families have evolved a wid- social parasitism is found only in the Allo- er range ofmodes ofparasitism thanjust those dapini (see Michener, 1974: 227, for an over- found in the Anthophoridae. Depending upon view). the group, parasitism may be accomplished Eggs of cleptoparasitic anthophorids are (1) by the cleptoparasite female opening the introduced into host cells in one oftwo ways: host cell and killing and removing the host's either the cleptoparasitic female enters the young before she oviposits and reseals the cell open cell while the host is away foraging and (larvae nonhospicidal) (e.g., Sphecodes, Ho- hides her egg by inserting it into the cell wall plostelis), and (2) by the cleptoparasite egg (and/or cell lining), or the cleptoparasite being introduced into an open or closed cell makes a small opening in the closure (or wall) and one or more instars, but not the first of a cell that has already been sealed, inserts instar, being hospicidal (e.g., Coelioxys, Stel- an egg through the hole into the cell lumen, is). and then reseals the opening. Presumably both Unless stated otherwise, specimens de- mechanisms have evolved more than once scribed are in the collection of the American in the family. Museum of Natural History. The hospicidal first instars of the cleptoparasitic anthophorids exhibit behavioral and anatomical modifications that enable them ACKNOWLEDGMENTS to seek out, identify, and kill host eggs or Many ofthe cleptoparasitic first instars were larvae. The universal (but not necessarily ho- collected while I was in residence at biological mologous) feature of these hospicidal larvae field stations. I wish to thank the personnel is attenuate, tapering, sharp-pointed, usually of the following stations for their assistance elongate mandibles used to dispose the hosts' and hospitality: William Beebe Memorial young. Head capsules are modified and Laboratory, Trinidad, West Indies (New York strengthened in various ways in association Zoological Society); Cedar Point Biological with the increased mandibular musculature Station, Ogallala, Nebraska (University of and perhaps with sclerotization that shields Nebraska); and Southwestern Research Sta- the parasitic first instar from host larvae or tion, Portal, Arizona (American Museum of competing cleptoparasitic young. Larval mo- Natural History). bility is important for finding hosts, and in The completeness of this study was aided some lineages modifications of the abdomi- by the loan or donation of specimens from nal apex (such as the development of a py- the following specialists: Frederick D. Ben- gopod) are adaptations for crawling. Other nett, University of Florida, Gainesville; lineages lack obvious anatomical modifica- Charles D. Michener, University of Kansas, tions for crawling but still must crawl, per- Lawrence; Robbin W. Thorp, University of haps by body movement alone. Some larvae California, Davis; Philip F. Torchio, USDA have lateral body extensions, presumably en- Bee Biology and Systematics Laboratory, Lo- abling them to float on liquid provisions while gan, Utah. they seek and kill host young. There is a spec- trum of anatomical features involved with 2 Among the parasitic anthophorids, instars after the identifying host immatures (or immatures of first are much more conventional in anatomy. Hence it other cleptoparasites) in a completely dark would appear that only the first instar is capable ofkilling environment, most notably elongate, sensil- the host and battling other first-instar cleptoparasites. Certainly the first instars are the ones typically encoun- la-bearing head appendages (antennae, palpi, tered assassinating host immatures. However, active and/or labral tubercles). opening and closing of mandibles and agile body move- No female cleptoparasitic anthophorid has ments of some second-stage cleptoparasites suggest that been observed to remove a host immature they too may be able to kill hosts and/or defend them- before ovipositing, and the first instar, and selves. 4 AMERICAN MUSEUM NOVITATES NO. 3029

Several versions of this manuscript were were closely allied to relatives ofNomada and read by my colleagues. Their comments, sug- all had arisen from a common parasitic an- gestions, and criticisms substantially aided cestor. This group included most, if not all, the formulation ofideas expressed in the final those subgroups currently considered (that is, version. I wish to express my appreciation to until the present study and the recent ones them for their efforts on behalfofthis report. by Roig-Alsina, 1989, 1990, in press) to be An early version was reviewed by Byron A. members of the Nomadinae. Alexander, Robert W. Brooks, Charles D. The same year Rozen (1954) published a Michener, Eric Quinter, Arturo Roig-Alsina, brief description of the first instar of Oreo- and Philip F. Torchio. A more recent draft pasites which Michener (1957) later com- was read by Byron A. Alexander, George C. pared with the newly found first instar of Is- Eickwort, Charles D. Michener, and Arturo epeolus. Thus Michener was first to consider Roig-Alsina. first instars in addressing evolutionary relationships ofcleptoparasitic anthophorids, although comparisons of parasitic first instars of a number of families (including Antho- HISTORICAL BACKGROUND phoridae) were made much earlier (Graen- The taxonomic-evolutionary relationships icher, 1905). of cleptoparasitic bees have long been rec- In the 1960s there began a series of inves- ognized as a perplexing problem. Griitte tigations into larvae and life histories ofclep- (1935) provided a historical account of early toparasitic bees by F. D. Bennett, G. E. Bo- studies and then attempted a comprehensive hart, M. Favreau, C. D. Michener, J. G. analysis ofcleptoparasitism (including social Rozen, R. W. Thorp, P. F. Torchio, N. N. parasitism) based on comparative adult anat- Youssef, and others. Bohart (1970) then at- omy. He was at a disadvantage for seveal tempted to sort out the evolutionary rela- reasons: (1) understanding of bee phylogeny tionships. This effort was more successful than at that time was limited, (2) he presupposed that ofGriitte (1935) because it was based on that similar cleptoparasitic forms arose from a more solid understanding ofbee phylogeny different but closely related pollen-collecting derived through studies on adults, and be- ancestors rather than from one another, and cause it incorporated recently discovered life (3) adult features display a confusing array of history information and recent treatments of evolutionary convergences. Michener's (1944) immature stages by others. Bohart concluded early treatment of the cleptoparasitic antho- that there were four origins ofcleptoparasitic phorids, developed within a clearer picture anthophorids: the Nomadinae, the Melectini, ofapoid phylogeny, was also based solely on the Ericrocidini, and the Protepeolini (as rep- adults. He followed Griitte's assumption that resented by Isepeolus and Protepeolus). He many of the nomadine taxa arose indepen- considered the Melectini and Ericrocidini dently from primitive, no longer extant, pol- closely related but followed Rozen's (1 969b) len-collecting anthophorids, but he placed Is- conclusion based on mature larvae that the epeolus, Protepeolus, and also Leiopodus into two had a diphyletic origin. Both he and Roz- a single tribe, the Protepeolini. In contrast to en (1 969b) believed that the Ericrocidini and Griitte, Michener deduced that Melecta and Rhathymini had a common parasitic ances- Thyreus had not arisen from Anthophora but tor. However, in disagreement with Rozen rather from a bee more primitive than those (1966), he concluded that the Protepeolini of any extant Anthophorini. He accorded had a separate evolutionary origin from the separate tribal status to the Melectini, Rhath- Nomadinae, primarily because of the char- ymini, and Ericrocidini, and regarded them acteristics of the mature larva of Isepeolus as being closely related and quite distinct from and its cocoon-spinning ability (the biology the nomadine parasitic bees. He rejected and larva ofProtepeolus were still unknown). Griitte's suggestion that some ofthe parasitic The study of Protepeolus in nests of Dia- anthophorids arose from panurgines. Ten dasia (discovered on the grounds of the Mu- years later Michener (1954) decided that the seum's Southwestern Research Station in Ar- genera included in the Osirini and Epeolini izona by K. R. and G. C. Eickwort) resulted 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 5 in a major rethinking about the relationships However, during the course of the present ofIsepeolus and Protepeolus which until then investigation, Arturo Roig-Alsina indepen- had been placed in a single tribe. Rozen et dently pursued a revealing series of studies al. (1978) retained both genera in the No- on adults of a number of groups of Noma- madinae as early offshoots of the lineage but dinae. He (Roig-Alsina, 1989) concluded that recognized them as distantly related to one Epeoloides, Parepeolus, and Osiris were re- another as well as to the remaining Noma- lated, should all be placed in the Osirini in dinae. Protepeolus was retained in Nomadi- the Anthophorinae, and thus indicated that nae on the basis of seven apomorphies per- their cleptoparasitism had a separate origin taining to biologies and larvae. These will be from that of the Nomadinae. Roig-Alsina discussed below. Isepeolus was also retained (1990) placed Coelioxoides in the Tetrape- in the Nomadinae partly because ofthe num- diini (Anthophorinae), an indication that it ber of similarities (presumably some syna- too was a separate evolutionary excursion into pomorphic) of its first instars to those of the cleptoparasitism. We do not yet know the Nomadinae and partly because comparative mode ofparasitism of either Coelioxoides or biological information was (and still is) un- the Osirini, and their immatures have not available. been collected. When this information is Snelling and Brooks (1985) analyzed the forthcoming it will be a test of the validity cladistic relationships of adult Ericrocidini of the present analysis. and Rhathymini. Their study concluded that In yet another study, Roig-Alsina (in press) either (1) the Rhathymini is a sister group of concluded that the monophyly of Nomadi- the Ericrocidini (suggesting a single origin of nae in the broad sense, that is, including the cleptoparasitism for the tribes) .and the Cen- Protepeolini and Isepeolini, was not sup- tridini is the sister group of their common ported by his analysis of adult features, but, ancestor, or (2) the Rhathymini is the sister exclusive of these two tribes, it was mono- group of the Ericrocidini plus Centridini phyletic. This conclusion coincides with the (Centris, Epicharis, and Ptilotopus) (requir- analysis presented in this paper. ing either a diphyletic origin of cleptopara- Recently Alexander (1990) compared a sitism or an unlikely reversal to pollen gath- cladistic analysis of 15 genera of the No- ering in the Centridini). Both hypotheses madinae including Isepeolus and Protepeolus considered Caenonomada, a pollen-collect- based on mature larvae (from the work of ing South American "primitive" anthophor- Rozen et al., 1978) with a similar analysis id, as the basal sister group of the rhathy- based on adults. One result he found was that mine-ericrocidine-centridine complex. Ofthe Isepeolus and Protepeolus form a monophy- two hypotheses, they favored (1) because of letic group based on adults, but this was not the similarity in the male terminalia of the supported by analysis ofmature larvae. When Ericrocidini and Rhathymini and because the he combined larval and adult characters in a second hypothesis required 11 convergences single analysis, the results closely resembled between the two tribes. The first hypothesis those based on larval characters alone, i.e., was also in agreement with Rozen (1969b) Isepeolus and Protepeolus formed separate and Bohart (1970). However, Snelling and basal clades ofNomadinae. This result is con- Brooks pointed out that there are 12 adult sistent with the conclusion presented below. convergences between melectines and the er- The crux ofthe problem in studies by Rozen icrocidine-rhathymine bees so that 11 is not et al. (1978) on mature larvae and by Alex- an unlikely number ofconvergences between ander (1990) on adults and mature larvae was unrelated cleptoparasitic bees. the initial assumption that the Nomadinae in When I initiated the present project, three the broad sense is monophyletic. The present evolutionary origins ofcleptoparasitism were study reevaluates the presumed biological and recognized among the Anthophoridae: (1) last-instar anatomical synapomorphies used Nomadinae (including Isepeolus, Protepeo- by Rozen et al. and Alexander to define the lus, Coelioxoides, Epeoloides, Parepeolus, and Nomadinae in the broad sense (see treat- Osiris); (2) Melectini; and (3) Ericrocidini- ments of Isepeolini and Protepeolini in next Rhathymini. section). 6 AMERICAN MUSEUM NOVITATES NO. 3029

TABLE 1 Comparison of the Mode of Parasitism and of Anatomical Features of First Instars of Cleptoparasitic Anthophorids About Which Information Is Available (See text for explanation of coding of character states.)

6. Ericrocidini Character 1. Nomadinae 2. Protepeolini 3. Melectini 4. Rhathymini 5. Isepeolini ? Unknown 2 closure after 0. Introduction of 1 Into open cell while 1 Into open cell when host 2 Through closure after 2 Through closure after Through Cleptoparasite Egg host female foraging female foraging cell is closed cell is closed cell is closed into Host Cell ? Unknown 2 Free in cell, 1. Egg Deposition I Embedded in cell wall 1 Embedded in cell wall 2 Free in cell, usually 2 Free in cell, adhering adhering adhering to closure to closure to closure

1 1 2. Head Shape I Prognathous, rarely O More or less hypo- 1 More or less prognathous 0 Hypognathous Prognathous Prognathous (TownLsesdiella, Neolarra) gathous somewhat hypognathous 0 Not 0 Not swollen 3. Parietals 0 Not swollen to somewhat 1 Extending upwards and 0 Normal, unmodified 2 Swollen, globose swollen swollen just in front of backwards posterior head margin 2 fused with 4. Ventral Sclerotized 1 Present, not fused with 0 Absent 2 Present, fused with 1 Present, not fused 2 Present, fused with Present, Postoccipital Bridge labiomaxillary region labiomaxillary region with labiomaxillary sclerotized labiomaxil- sclerotized labiomaxil- region lary region lary region 1 onto 0 at 5. Head Sclerotization 0 Ending at posterior O Ending at posterior 0 Ending at posterior 0 Ending at posterior Extending pro- Ending posterior margin margin margin margin thorax margin

6. Cranial Band of 0 Absent 0 Absent 1 Distinct 0 Absent 0 Absent 0 Absent Spinulae

7. Troughlike External I Present, at least in 0 Absent 1 Present 0 Absent 0 Absent 0 Absent, except in Hypostomal Groove most cases Enicrocis

0 8. Angle of Posterior 0 Right angles or rarely 1 Broadly curved 0 Variable 2 Nearly straight ? Impossible to interpret Right angle Margin of Head to obtuse Hypostomal Groove 0 well 2 9. Antennal Size 1 Almost always small, Low but distinct 2 Large, elongate 1 Scarcely evident Small, defined Large, projecting obscure swelling

10. Antennal Shape 0 Not flattened 0 Not flattened 0 Not flattened 0 Not flattened 1 Dorsoventrally flattened 1 Slightly to moderately dorsoventrally flattened

11. Antennal Fusion I Fused with head capsule I Fused with head 0 Set off by ring from 1 Fused with head I Fused with head capsule 1 Fused with head capsule capsule head capsule capsule

12. Labrum 1 Nonsclerotized, not O Sclerotized, articu- 2 Sclerotized, fused 0 Perhaps faintly sclero- 2 Sclerotized, fused with 2 Sclerotized, fused fused with clypeus lating with clypeus with clypeus tized, not fused with clypeus with clypeus clypeus

13. Labral Tubercle(s) 1 Paired, long to very 2 Three pairs, fLxed 1 Paired, nonarticulating 1 Paired, nonarticulating 3 Single median ? Vague at best long, separately to labrum articulating I 14. Mandible 1 Moderately long to very 0 Not elongate Moderately long 1 Moderately long 1 Very long, slender I Long, robust long

1 0 on 15. Labiomaxillary Present mesad of Absent 1 Somewhat sides and I Only on sides of 2 Complete 2 Complete Sclerotization hypostomal grooves but behind palpi maxillae absent in middle

16. Labium and Maxillae 1 Extensively fused 0 Separate 0 Separate 0 Separate I Extensively fused 1 Extensively fused

17. Maxillary Palpus 1 Well developed, some- 2 Short, large, padlike 0 Small but distinct 3 Evident only because 3 Scarcely evident 3 Scarcely evident times greatly elongate of sensilla

18. Abdominal X 1 forked or 1 Bilobed 0 Rounded Segment Bearing 1 Bilobed 2 Truncate, small, greatly 2 Truncate, trilobed, trilobed, eversible fused to IX fused to IX pygopod

19. I Body Setae 0 Absent Present 0 Absent 0 Absent 1 Present 0 Absent

20. Spiracles of Abdominal 0 Normal in position O Normal in position 0 Normal in position 0 Normal in position 1 Larger than preceding I Usually larger than Segment VIII and size and size and size and size ones and at posterior preceding ones and at margin of segment posterior margin of segment

EVOLUTION OF CLEPTOPARASITISM and the anatomy ofthe first instars, exclusive IN THE ANTHOPHORIDAE of Coelioxoides and the Osirini (for which we This investigation shows that there are six have no information on mode of parasitism distinct groups of cleptoparasitic anthophor- and anatomy of first instars). As indicated in ids as revealed by their mode of parasitism tables 1-3, these are the Nomadinae, Prote- 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 7 peolini, Melectini, Rhathymini, Isepeolini, thophorids also arose in megachilid hospi- and Ericrocidini. If Roig-Alsina's (1989, cidal larvae (though not always first instars). 1990) conclusions are correct regarding Coe- For example, first (and subsequent) instars of lioxoides and the Osirini, then the family may Dioxys have elongate antennae (Rozen, 1967), contain a total ofeight cleptoparasitic groups. suggestive of those of the Melectini; hospi- The Nomadinae, Melectini, and Ericrocidini cidal third instars of Coelioxys are strongly (the taxa for which we have information about prognathous with the head capsule strongly more than one genus) are each believed to be sclerotized both dorsally and ventrally, monophyletic on the basis of the numerous somewhat like those of the Ericrocidini, Is- shared specialized features of the included epeolini, and Nomadinae (Baker, 1971). genera as revealed in table 1 and the system- Hence homoplasy is common among these atics section. The general question to be ad- features ofcleptoparasitic bees, and only when dressed is whether each of the six groups there are numerous suspected synapomor- treated here represent a separate evolutionary phies can we place much confidence in the excursion into cleptoparasitism or whether monophyly of two or more groups. Table 3 two or more of them have a common para- lists the character states of first-instar clep- sitic ancestor. More specifically, can a study toparasitic anthophorids shared by pairs of of their first instars and mode of parasitism taxa. As discussed below, only in the case of shed light on this matter? Isepeolini/Ericrocidini, with 11 shared apo- The only certain way of determining morphies, is there a reasonable (though not whether cleptoparasitism arose independent- necessarily probable) indication that they had ly is to demonstrate through cladistic analysis a common parasitic ancestor. that each cleptoparasitic group has a pollen- As another suggestion that these groups are carrying sister group. This study cannot at- separate cleptoparasitic lineages, I analyzed tempt such an analysis because first instars them cladistically using the data in table 1. ofpollen-carrying taxa are almost completely The character states were identified as ple- unknown, pollen-carrying taxa do not have siomorphic (0) or apomorphic (1, 2, etc.) and behavior patterns (such as searching for host are presented in table 2 as a matrix (also used nests, entering host cells, killing host im- to derive table 3). Primitive character states matures) that can be homologized with those (0) were based on a hypothetical ancestor. ofcleptoparasites, and most (if not all) ofthe This ancestor closely parallels the description apomorphic anatomical features of the de- of the first instar of Exomalopsis chionura scribed first instars (table 1) are believed to Cockerell3 (Rozen, 1957), although the an- be adaptations to hospicidal ways of life and tenna ofE. chionura is somewhat more pro- therefore have no counterpart in nonhospi- nounced than that of some other known first- cidal larvae. instar, pollen-carrying anthophorids. All Nonetheless, there is some reason to be- multistate characters were coded nonaddi- lieve that most (if not all) of these six groups tive, and autapomorphies (characters 5, 6) represent separate evolutionary origins of were deleted. With James S. Farris's Hennig86 cleptoparasitism; namely, there is little evi- program using the implicit enumeration dence to suggest that as a single group they command, the following single most parsi- are monophyletic or that any subgroup of two monious tree topology was produced: (Pro- or more of them is monophyletic (with the tepeolini (Nomadinae (Rhathymini (Melec- possible exception of the Isepeolini and Er- tini (Isepeolini + Ericrocidini))))), length 40, icrocidini). We know that cleptoparasitism evolved independently in three bee families 3At the time the first instar of E. chionura was de- (Halictidae, Megachilidae, Apidae) other than scribed, we were unaware that first instars of many, if the Anthophoridae, so that cleptoparasitism not all, nonparasitic anthophorids are encased in the egg by itself is not a strong synapomorphy. Sim- chorion and that the second instar is the stage the crawls ilarly, hospicidal larvae (with elongate, sharp- away from the chorion. Hence the description may refer to the second instar. However, observations on other pointed mandibles) have also evolved at least primitive anthophorids (including Exomalopsis sidae three times in the Megachilidae. Many ofthe Cockerell andAncyloscelis wheeleri (Cockerell) show that other specialized first-instar anatomical fea- first and second instars of nonparasitic species are quite tures associated with cleptoparasitism in an- similar. 8 AMERICAN MUSEUM NOVITATES NO. 3029

TABLE 2 Data Matrix for Analysis of Relationships of Parasitic Anthoporidae Based on Egg Deposition Features and First Instars (Character coding given in table 1. All multistate characters nonadditive) Character Taxon 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0. Ancestor 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1. Nomadinae 1 1 1 0 1 0 0 1 0 1 0 1 1 1 1 1 1 1 1 0 0 2. Protepeolini 1 1 0 1 0 0 0 0 1 0 0 1 0 2 0 0 0 2 1 1 0 3. Melectini 2 2 1 0 2 0 1 1 0 2 0 0 2 1 1 1 0 0 0 0 0 4. Rhathymini 2 2 0 2 1 0 0 0 2 1 0 1 0 1 1 1 0 3 1 0 0 5. Isepeolini ? ? 1 0 2 1 0 0 ? 0 1 1 2 3 1 2 1 3 2 1 1 6. Ericrocidini 2 2 1 0 2 0 0 0 0 2 1 1 2 ? 1 2 1 3 2 0 1 consistency index 80, retention index 65. This ymini and Ericrocidini were believed related. tree is not intended to indicate that the in- At the same time, no one had suggested that cluded taxa are a single monophyletic clade, the Isepeolini and Ericrocidini might have a but rather to show the relationships of the common ancestor. Hence, relationships sus- cleptoparasitic taxa without considering their pected on the basis of certain life stages are possible pollen-carrying sister taxa. Compar- not supported by those based on other life ing the relationships of the parasitic groups stages. This again suggests that evidence in based on first instars and mode ofparasitism support of monophyly of the cleptoparasitic with the earlier "traditional" relationships anthophorid groups is weak. based on adult and mature larval features As yet another line of reasoning, these six reveals distinct incongruences. Before, the groups deposit their eggs in host cells in dif- Protepeolini and Isepeolini were thought to ferent ways, and each tends to possess dif- be a part of the Nomadinae, and the Rhath- ferent and distinctive suites of first-instar

TABLE 3 Derived Character States of Egg Deposition Features and First Instars of Cleptoparasitic Anthophorids Shared by Pairs of Taxa (For explanation, see text) No. shared Taxa Shared derived states states 1/2 Nomadinae/Protepeolini 0, 1, 11, 18 4 1/3 Nomadinae/Melectini 2, 7, 13, 14, 15 5 1/4 Nomadinae/Rhathymini 4, 9, 11, 13, 14, 15, 18 7 1/5 Nomadinae/Isepeolini 2, 11, 14, 15, 16 5 1/6 Nomadinae/Ericrocidini 2, 11, 14, 15, 16 5 2/3 Protepeolini/Melectini - 0 2/4 Protepeolini/Rhathymini 11, 18 2 2/5 Protepeolini/Isepeolini 11, 19 2 2/6 Protepeolini/Ericrocidini 11 1 3/4 Melectini/Rhathymini 0, 1, 13, 14, 15 5 3/5 Melectini/Isepeolini 2, 4, 12, 14, 15 5 3/6 Melectini/Ericrocidini 0, 1, 2, 4, 9, 12, 14, 15 8 4/5 Rhathymini/Isepeolini 11, 14, 15, 17 4 4/6 Rhathymini/Ericrocidini 0, 1, 11, 14, 15, 17 6 5/6 Isepeolini/Ericrocidini 2, 4, 10, 11, 12, 14, 15, 16, 17, 18, 20 11 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 9 modifications that enable its larvae to search dependently and gives evidence that the fifth for, identify, and kill hosts. Each combina- and sixth groups (the Isepeolini and Ericro- tion ofcharacters represents a different adap- cidini) may have had a common parasitic tive solution that accomplishes the same end ancestor. for all lineages-elimination of host or other (1) Nomadinae: The largest assemblage of larvae that might feed on the food supply. cleptoparasitic anthophorids, this subfamily Once an evolutionary lineage has developed as here constituted (i.e., without Protepeolus, a particular set of features that makes its Isepeolus, and Leiopodus, and presumably members successful cleptoparasites, a differ- without Coelioxoides and Osirini sensu Roig- ent set of features that accomplish the same Alsina, 1989) appears to form a monophy- end would not likely evolve in that lineage. letic group, to the extent that we have knowl- Hence, different sets of characters seem to edge of biologies and first instars of the in- identify separate lineages that became clep- cluded taxa. In all known cases (see Bohart, toparasitic. However, as an argument against 1970, for early references; Ehrenfeld and this hypothesis, two or more of them might Rozen, 1977; Rozen, 1971, 1977, 1984a, have arisen from a common cleptoparasitic 1984b, 1986, 1989a, 1989b, 1989c; Rozen ancestor that was highly plesiomorphic, for and Snelling, 1986; Shanks, 1978; Torchio example, an ancestor that had no special par- and Burdick, 1988), the female enters the open asitic adaptations except for a tendency to cell of the host while the host female is for- deposit its eggs in provisioned cells of other aging, inserts a small egg into the cell wall (or species of bees and a first instar that was cell lining), often in a characteristic genus- slightly more aggressive than that ofthe host. specific way, and then departs. The hospici- During evolution, such a generalized ances- dal larva emerges in time to kill the host egg tor might have led to lineages that became or early instar. The first-instar cleptoparasite more specialized in different ways. possesses elongate, more or less diverging la- As stated above, convincing evidence of bral tubercles that are not fused to one an- the relationships ofthese parasitic groups will other or to the head capsule and that presum- be determined when and if we can identify ably sense the area in front ofthe prognathous, their sister groups. However, the validity of often elongate head. Antennae are scarcely the other lines of reasoning can continue to evident, suggesting that they do not play a be evaluated. For example, if we learn that significant role in host detection. Maxillary the first instar and mode ofparasitism ofCoe- palpi, pronounced to extremely long, must lioxoides (whose pollen-collecting sister tax- acquaint the larva with the substrate beneath on, Tetrapedia, has been identified by adult its head. In contrast, labial palpi are usually synapomorphies [Roig-Alsina, 1990]) is dif- not evident as projections and therefore may ferent from those of the other parasitic an- be oflittle importance. Larval mobility is as- thophorid groups, credence will be given to sisted either by an eversible, forked, poste- the reasoning presented above. Similarly, if riorly directed pygopod arising from the last the pollen-collecting sister group of one (or abdominal segment or by a trilobed process more) of the parasitic groups can be identi- consisting ofa narrow, elongate terminal ab- fied, this too will give weight to the idea that dominal segment subtended on. each side by the other groups also had evolved cleptopar- a lobe (these lobes are perhaps homologous asitism de novo. In both cases, association with the pygopod). of different modes of parasitism with groups Mature larvae have extremely short man- proven to have pollen-carrying sister taxa dibles that contrast to the often very elongate would strengthen the probability that differ- mandibles of the first instar. Their maxillary ent modes of parasitism are an indicator of palpi are small, and labial palpi are not gen- separate origins of cleptoparasitism. erally evident. This is the only group of par- The following is an expansion and discus- asitic anthophorids that does not spin co- sion of the information given in table 1. It coons (although cocoon spinning is also assumes tentatively that four ofthe six groups lacking in the single melectine genus Thyreus evolved their cleptoparasitic way of life in- [Rozen, 1969a] and in nonoverwintering gen- 10 AMERICAN MUSEUM NOVITATES NO. 3029 erations of Protepeolus [Rozen et al., 1978] chio and Trostle, 1986). The first instars of and Xeromelecta californica [Torchio and the Melectini described here reaffirm that the Trostle, 1986]), and the mature larvae have tribe is monophyletic and that it is not closely a recessed labiomaxillary region, character- akin to other parasitic Anthophoridae. The istic of noncocoon-spinning bee larvae. band of spinulae extending across the vertex Although the subfamily is believed to have from the mandibular bases and large conical derived from some primitive nonparasitic antennae appear to be unique apomorphies anthophorid, synapomorphies in mature lar- for the group. The basal membranous ring of vae have not been identified to align the No- the antenna is also unique among parasitic madinae with any group. anthophorids. This feature is probably ple- (2) Protepeolini: Like Isepeolus, Protepeo- siomorphic, so that its absence in other par- lus was considered an early clade in the No- asitic groups must be convergent. madinae by Rozen et al. (1978). Only in the Bohart (1970) hinted that the Melectini and Nomadinae and Protepeolini (Rozen et al., Ericrocidini might be related. Although the 1978) are the eggs known to be embedded in similarities between the two (8, in table 3) the cell wall by the cleptoparasite before the are the second highest number of shared host female seals the cell. This feature was character states among the cleptoparasitic originally considered a strong synapomor- lineages, they are not strong apomorphies. phy, but now is considered a convergence Both Michener (1944) through studies of because it is supported by so few other pos- adults and Rozen (1969b) who compared sible synapomorphies offirst instars and ma- mature larvae concluded that the Melectini ture larvae. This convergence may be ex- and the Anthophorini arose from a common plained as follows: all parasitic anthophorids ancestor. The present study sheds no further introduce their eggs into host cells; two lin- light on this matter. eages (Nomadinae and Protepeolini) inde- (4) Rhathymini: Camargo et al. (1975) con- pendently evolved ways ofhiding eggs in open cluded that the mode ofparasitism ofRhath- cells that are still being provisioned; three ymus was like that of the melectines, based lineages independently found means of in- on finding a cast chorion and first instar ex- troducing eggs into closed cells. uviae in situ in a cell that they also illustrated Similarities ofthe first instars ofProtepeo- (their fig. 4B). This conclusion is supported lus and the Nomadinae can be explained as by new information under Systematics of convergences because these features (as listed Cleptoparasitic First-Instar Anthophoridae, in table 1) tend to be associated with many below. other parasitic lineages. Even the bilobed ab- One of the most surprising discoveries of dominal segment X, which suggests a pygo- the present study was that the first instar of podlike nomadine structure, appears in two Rhathymus is vastly different from all the other lineages. other known cleptoparasitic anthophorids. Only two features ofthe mature larvae sug- The comparatively few shared first-instar gested a relationship between Protepeolus and features of Rhathymus and other cleptopar- the Nomadinae (Rozen et al., 1978): the asites including the Ericrocidini do not sup- modified position of the posterior tentorial port a close relationship between Rhathymus pits and the short mandible with simple apex and the others. Rozen (1969b), Bohart (1970), and reduced cusp. and the favored hypothesis of Snelling and Features are not known that might align Brooks (1985) had concluded that Rhathy- the Protepeolini with one or more extant mus and the Ericrocidini had a common par- groups ofpollen-collecting anthophorids, and asitic origin. However, the greatly swollen the tribe should be assigned to the Antho- hypognathous head (presumably for attach- phorinae for the time being. ment oflarge mandibular muscles) ofthe first (3) Melectini: Females of all known me- instar as well as the other characters pre- lectine genera oviposit through small open- sented in table 1 argue against this hypothesis ings made in the closures of sealed host cells unless, of course, the common ancestor was (see Bohart, 1970, for early references; Tor- highly plesiomorphic. Separate origins of 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES I1I cleptoparasitism in Rhathymini and Ericro- part ofprothorax (extending only to posterior cidini are consistent with the second hypoth- margin ofhead); abdominal apex without py- esis of Snelling and Brooks (1985: fig. 78B). gopod (pygopod well developed, often large). The relatively high number (7, in table 3) The mature larvae ofIsepeolus and the No- of specialized character states shared by the madinae differ in a number of ways: body Nomadinae and Rhathymini are believed to tubercles present (absent; body often stream- be the result of convergence because features lined); labrum with single median, sensilla- of adults and mature larvae do not support bearing apical tubercle (paired tubercles aris- (but do not refute) a close relationships being from disc always present); mandibles tween these two taxa. moderately long (unusually short); larva co- (5) Isepeolini: This tribe has been assigned coon-spinning and therefore possessing pro- to the Nomadinae by most workers, although jecting labiomaxillary region and pro- Bohart (1970) believed it (as Protepeolini) nounced maxillary and labial palpi distinct because ofthe features ofthe mature (noncocoon-spinning; labiomaxillary region larva and its cocoon-spinning habits. Except recessed). for its hospicidal larva, other features con- The large number of striking specializa- cerning mode of parasitism are still un- tions (11, in table 3) of first instars of Ise- known. peolus and Ericrocidini may reflect that the With our increasing knowledge of first in- two taxa had a common cleptoparasitic pro- star parasitic anthophorids (especially the genitor. These features include: a heavily and Nomadinae), it seems obvious that the strik- completely sclerotized ventral head surface ing similarities (only 5, in table 3) of first that is fused to the heavily pigmented head instars of Isepeolus and the Nomadinae are capsule; fusion of the labrum to the clypeus; convergences that have appeared indepen- sclerotization of the area (presumably the dently among many of the lineages. These maxillae) between the mandibles and the include the strongly prognathous, elongate, mouth; distinct antennae; fusion of abdom- heavily sclerotized head capsule, curved inal segments IX and X (more extensive in elongate mandibles, and linear body form. Isepeolus); and enlarged spiracles (and spi- These are unconvincing similarities in the racular tracheae) (except apparently for Eri- presence of the following differences (con- crocis) of abdominal segment VIII that are trasting features of the Nomadinae in paren- positioned at the rear margin ofthe segment. theses): labrum bearing single small median, These enlarged and repositioned terminal probably apical4 tubercle (paired tubercles body spiracles suggest that the larvae are semi- arising from disc rather than apex oflabrum); metapneustic, perhaps an adaptation per- antennae, though small, sharply projecting mitting the larvae to remain mostly sub- (vague swellings); ventral surface of head merged in the semiliquid provisions of their completely sclerotized and all palpi absent or unrelated hosts (Colletes and Centris). at least nonprojecting (labiomaxillary region The features listed above appear to be membranous at least medially and maxillary strong synapomorphies because many (es- palpi conspicuous, often greatly elongate); pecially the nearly completely sclerotized sclerotization ofhead capsule extending over head, the enlarged and posteriorly positioned spiracles, and the fusion of the terminal ab- 4 Reexamination ofthe mature larva ofIsepeolus shows dominal segments) are not known in other that the sensilla-bearing median protuberance of the la- lineages. Furthermore, some of the apparent bium is apical, in contrast to the sensilla-bearing paired differences between the two taxa can be in- tubercles of mature Nomadinae which arise from the terpreted as sequences in a linear transfor- labral disc. This difference in position suggests that the tubercles of the two groups are nonhomologous. Since mation series: small or large but always flat- the labral tubercles of mature larvae are homologous tened and distinct antennae; vague narrowed with those of the first instar of the same taxon, the me- paired labral tubercles or single median labral dian tubercle of the first instar of Isepeolus cannot be a tubercle, but labrum always fused with clyp- homolog ofthe paired labral tubercles offirst-instar No- eus; mandibles robust or slender but always madinae, no matter where they seem to arise. long; greater or less fusion of terminal ab- 12 AMERICAN MUSEUM NOVITATES NO. 3029 dominal segments. Other differences appear and last instars are quite different so that the to be autapomorphies of Isepeolus: head two groups could well be considered separate greatly dorsoventrally flattened, its scleroti- but related tribes in the Anthophorinae. zation extending to anterior part of prothorax; and head and body setae evident. A reexamination ofthe features ofthe ma- SYSTEMATICS OF CLEPTOPARASITIC ture larvae of these taxa (Isepeolus: Rozen, FIRST-INSTAR ANTHOPHORIDAE 1966; Ericrocidini: Rozen, 1 969b, Rozen and METHODS Buchmann, 1990) shows a number of similarities that are plesiomorphic, such as ad- Most specimens used in this study were aptations for cocoon spinning (e.g., project- collected from host nests either as first-stage ing labiomaxillary region and long palpi). larvae or as eggs which then hatched in rear- Autapomorphies exist in both taxa, but no ing containers. To avoid postmortem changes, features that support or contradict a sister- specimens are best studied and drawn when group relationship can be identified. Biolog- freshly killed, but rarely was this convenient. ical attributes ofIsepeolus, such as where and Those collected by me were fixed in Kahle's when its eggs are laid, are unknown, but once solution and later transferred to 70% ethanol discovered may test the suspected relation- for storage. So as to be easily retrieved, very ships of the two taxa. small larvae were placed into microvials (e.g., Roig-Alsina (personal commun.), after 25 x 8 mm) stoppered with cotton and en- reading a preliminary version of this manu- closed in 4-dram storage vials with neoprene script, commented on the possible relation- stoppers. ships of the Isepeolini and Ericrocidini: "If Study procedures for first instars were sim- they are sister taxa, Isepeolines should intro- ilar to those used for mature larvae with cer- duce their eggs into closed cells as Ericroci- tain exceptions dictated by their small size, dines do. But there may be some indirect rarity (some taxa known from single speci- evidence to the contrary. Females of Isepeo- mens), and the often hydrofuge nature oftheir lini have modified sixth sterna which end in body integument, which makes it difficult to a sharp point and, depending on the group, submerge cleared specimense in liquids. Be- may bear short or long spine-like setae. The fore drawing and clearing, the external anat- possession of a modified tip of the abdomen omy ofeach larva was examined. While sub- is convergent with nonhomologous modifi- merged in alcohol, the uncleared specimen cations in the Nomadinae and Protepeolini was then drawn with a camera lucida and and might suggest a similar female behavior: stereoscopic microscope. Next, the head was embedding the egg into the cell wall. On the severed, and both body parts boiled for ap- other hand at least two groups of Isepeolini proximately 5 minutes in a solution of po- have modified mandibles, elongate and sharp tassium or sodium hydroxide until internal in the Melectoides senex group and tridentate tissue became semitransparent. Some speci- in most Isepeolus. I wonder if they might be mens were then kept in an hydroxide solution used to break into closed cells." at room temperature several hours or over- (6) Ericrocidini: Information presented un- night to reduce the internal tissue further. As der Systematics of Cleptoparasitic First-In- required, a specimen was usually briefly boiled star Anthophoridae, below, indicates that the again to further remove tissue. During clear- females of at least two ericrocidine genera ing, a specimen was closely monitored with- insert their eggs through holes made in the out being removed from solution. I did not closures ofhost cells. As explained under the extensively squeeze or mechanically remove Rhathymini, first-instar ericrocidines are very body tissue because manipulation might have different from first-instar Rhathymus, and do altered the body shape, made it difficult to not support the hypothesis that the two had submerge the hydrofuge integument below a common parasitic ancestor. the surface film, and/or in some cases re- For reasons expressed above, the Ericro- sulted in the integument folding and clinging cidini and Isepeolini may have evolved from to itself. a common cleptoparasite. However, their first After clearing, large specimens were trans- 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 13 ferred directly to water, but small ones were of Oreopasites (Rozen, 1954), Kelita (Ehren- placed in a watch glass containing a hydrox- feld and Rozen, 1977), and a number of spe- ide solution. Transferring specimens by for- cies of Triepeolus and Epeolus (Rozen, 1 989b) ceps from one solution of hydroxide to an- were described and illustrated. The following other (even ofdifferent temperature) allowed account is based in part on the previously them to submerge easily rather than to float published descriptions and in part on a sur- and flatten on the surface film as often hap- vey ofrepresentatives ofthe genera listed un- pened if they were placed directly into pure der Genera Studied, in the collections of the water. The hydroxide solution was replaced American Museum ofNatural History. These with pure water by dropper, and the process larvae show considerable variation (com- was repeated several times to eliminate as pare, for example, figs. 2-5 and figs. 6-9) and, much hydroxide as possible. I avoided acid- when studied fully, will probably help clarify ulated water because it might have caused relationships within the subfamily. The nu- residual tissue to coagulate and distort soft merous shared features, mostly derived, sug- exoskeletons. The head and body were then gest that the group is monophyletic and rep- transferred to glycerin in a suitably labeled resents a single evolutionary transformation Wheaton culture slide for study by stereo- into cleptoparasitism. scopic and compound microscope. Details of Illustrations of two taxa are presented to sensilla, integumental sculpturing, and inter- demonstrate the similarities and disparities nal anatomy were added to the pencil draw- in first-instar features in the subfamily: Trie- ings. Inked final illustrations were copied on peolus grandis Friese (figs. 1-5; see also Roz- vellum laid over the pencil drawings. Spec- en, 1 989b) and Townsendiella pulchra Craw- imens were subsequently retained indefinite- ford (figs. 6-9; illustrations based on one ly in glycerin on culture slides. specimen, 11 mi southwest of Congress, Ya- Except for the Nomadinae and Protepeo- vapai Co., Arizona, April 30, 1990 [J. G. lini, one species ofeach ofthe other parasitic Rozen and R. J. McGinley], from nest ofHes- lineages is fully described in the following perapis larreae). sections. Hence descriptions offirst instars of HEAD: Shape prognathous (less so in Neo- related species discovered in the future can larra, Townsendiella (fig. 9), Neopasites), more be prepared by reference to these descriptions (fig. 2) or less (fig. 6) elongate, not strongly and without reference to the tribal/subfa- dorsoventrally flattened (in contrast to Ise- milial descriptions. Tribal/subfamilial de- peolus); parietals not swollen (e.g., Epeolus, scriptions are based on the species descrip- Triepeolus (fig. 5), Oreopasites, Nomada) or tions and identify those features that seem to more or less swollen just in front ofposterior be important for differentiating and assessing margin of head (e.g., Neopasites, Brachyno- the relationships of the higher taxa with our mada, Holcopasites, Neolarra, Townsendiel- present knowledge. The numerous species of la, fig. 9). Integument ofhead capsule heavily Nomadinae are not treated individually, but sclerotized, faintly to deeply pigmented; scle- await study. As indicated below, they can be rotization ending posteriorly at rear of head, described at the tribal/subfamilial level be- not extending over anterior part ofthorax (as cause ofprevious studies of some groups and in Isepeolus); capsule with sensilla either because of a survey of the collections avail- nonsetiform or more rarely represented as able to me. The larva of Protepeolus singu- long, fine setae (e.g., Neolarra, Brachyno- laris Linsley and Michener was described re- mada, Epeolini); head sensilla not borne on cently (Rozen et al., 1978), and is not small tubercles; capsule without ring ofsharp- redescribed here. pointed spinulae across vertex (as found in Melectini). Tentorium complete in some cases but mostly unstudied; posterior tentorial pit DESCIuPTION OF THE NoMADINAE BASED in or below hypostomal groove; conspicuous ON FIRST INSTARS troughlike hypostomal groove externally at Figures 1-9 least on some taxa (figs. 3, 7) (similar to that Although first stage larvae of most No- of Melectini); sclerotization of head capsule madinae have yet to be studied in detail, those often extending mesad of each groove but 14 AMERICAN MUSEUM NOVITATES NO. 3029

Figs. 1-5. Triepeolus grandis, first instar. 1. Entire body, dorsal view. 2-5. Head, dorsal, ventral, frontal, and lateral views. Figs. 6-9. Townsendiella pulchra, head of first instar, dorsal, ventral, frontal, and lateral views. Scale refers to figure 1. Figures 1-5 modified from Rozen, 1989a. these extensions separated by median V- entirely sclerotized; however, head posteri- shaped membranous area (labiomaxillary re- orly often, if not invariably, with ventral gion) or at least labiomaxillary region never sclerotized postoccipital bridge; pleurosto- 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 15 mal and epistomal ridges often not defined I-VIII dorsal, all approximately same size; because of heavy sclerotization of head cap- those ofsegment VIII in normal position, not sule; distinct pale median dorsal ecdysial line at posterior margin of segment as in known evident on those species with pigmented head Ericrocidini. Position of anus unknown. capsules; posterior margin of head capsule GENERA STUDIED: First instars of one or nearly at or at right angle with hypostomal more species of the following genera were groove as seen in lateral view, rarely (e.g., examined: Brachynomada, Caenoprosopis, Brachynomada) at obtuse angle. Each anten- Caenoprosopina, Epeolus, Holcopasites, Kel- na usually a small vague protrusion fused to ita, Neolarra, Neopasites, Nomada, Oreopas- head capsule, rarely more distinct, but never ites, Paranomada, Pasites, Pseudodichroa, a large projection (as in Melectini and Eri- Townsendiella, and Triepeolus. crocidini). Labrum (figs. 4, 8) unsclerotized REMARKS: This diverse subfamily, the larg- except for a single pair of labral tubercles; est cleptoparasitic group ofanthophorids, will these tubercles elongate, usually apically di- be treated in a subsequent study to shed light verging, articulating independently. Mandi- on the interrelationships ofthe included taxa. ble often very long, attenuated, slender apically and either slender or stout basally; inner DESCRIPTION OF THE PROTEPEOLINI edge usually smooth but sometimes (e.g., BASED ON THE FIRST INSTAR Epeolini, figs. 2, 3) bearing teeth or sharp- edged projections. Labiomaxillary region Figures 10-13 fused, membranous except for sclerotized This description is based on the material maxillary palpus in some cases; maxillary used by Rozen et al. (1978) in their treatment palpus pronounced, ranging from being as of the larvae and biology ofProtepeolus sin- long as basal diameter (fig. 9) to being many gularis Linsley and Michener. times longer than basal diameter (e.g., Bra- Head (figs. 11-13; see also Rozen et al., chynomada, Epeolini, fig. 5); labial palpus 1978): Shape more or less hypognathous;5 pa- usually not projecting and therefore not de- rietals extending posterodorsally so that tectable except perhaps for grouping of sen- height ofhead from vertex at posterior mar- silla but present in some groups (e.g., Hol- gin of head to posterior mandibular articu- copasites, Nomada). Hypopharynx (figs. 3, 7) lation nearly twice as long as distance from usually if not invariably a transverse sclero- posterior tentorial pit to apex of clypeus as tized plate that is V-shaped, medially notched, seen in lateral view (fig. 13), hence capsule or curved along its anterior edge. appearing shallow by comparison with all BODY: Form elongate, straight, usually ta- other known cleptoparasitic anthophorids; pering posteriorly, with (fig. 1) or usually capsule not strongly flattened dorsoventrally without (Rozen, 1989b: figs. 23, 27, 29, 30, as in Isepeolus. Integument of head capsule 35, 37) lateral tubercles, occasionally (Town- heavily sclerotized and darkly pigmented; sendiella) with paired ventrolateral tubercles sclerotization ending posteriorly at rear of on abdomen; abdominal segment X bearing head; many sensilla setiform, moderate in eversible bifurcate pygopod; rami ofpygopod length; those on mandibular bases apparently posterolaterally directed, short (e.g., Oreo- borne on small tubercles; capsule without pasites, Pasites, Neolarra) or long (e.g., Pseu- band ofsharp-pointed spinulae across vertex. dodichroa, Epeolini, fig. 1); occasionally (e.g., Tentorium possibly complete, but anterior Townsendiella, Nomada, and apparently arms thin; anterior pits distinct; posterior pits Paranomada) segment X narrow, projecting distinct, arising from hypostomal ridges and posteriorly so that abdomen appearing api- posterior thickening of head; hypostomal cally trilobed (lateral lobes presumably ho- ridge well developed; external groove not mologous with pygopod); prothorax sometimes 5 Interpretation of this feature depends on the orien- protruding ventrally. Integument tation of the head capsule in lateral view. If it is posi- without setae, usually with patches of short, tioned so that the hypostomal groove is horizontal, the fine, inconspicuous spicules. Thoracic spir- head becomes more prognathous. Iforiented so that the acles present (e.g., Pseudodichroa) or absent posterior margin of the head is vertical, the head be- (Epeolini); spiracles on abdominal segments comes more hypognathous. 16 AMERICAN MUSEUM NOVITATES NO. 3029

1.Omm o

10 - g MAXILLARY PALPUS HYPOSTOMAL RIDGE 10 13

Figs. 10-13. Protepeolus singularis, first instar. 10. Entire body, dorsal view. 11-13. Head, dorsal, frontal, and lateral views. Scale refers to figure 10. Figure 13 modified from Rozen et al., 1978.

troughlike (as in some if not all Nomadinae, biomaxillary region (fig. 12; see also Rozen Ericrocis, and Melectini); sclerotization of et al., 1978: fig. 23) completely membranous; head capsule not extending below (mesad of) maxilla distinct from labium (not fused); hypostomal ridges (labiomaxillary region maxilla with large orally directed apical lobe completely unsclerotized); pleurostomal ridge bearing numerous moderately short setiform moderately well developed at least near an- spicules; maxillary palpus borne ventroapi- terior mandibular articulation; epistomal cally as short but large, oval, padlike struc- ridge laterad of anterior tentorial pits mod- ture; labium covered with posteriorly direct- erately well developed but apparently absent ed, moderately short, setiform spicules; palpus mesad of pits; distinct pale median dorsal apparently nonprojecting, flattened, rather ecdysial line present; posterior margin ofhead large, nonspiculated area bearing a few sen- capsule not forming sharp angle with hypos- silla on ventral surface of labium. Hypo- tomal ridge but appearing as a gently curved pharyngeal area apparently not defined be- extension of it, as seen in lateral view (fig. cause salivary opening very close to mouth 13). Each antenna a low but distinct swelling (also not defined in last larval instar, Rozen fused with and gradually arising from front et al., 1978). ofhead. Labrum (fig. 12) a distinct sclerotized BODY: Form (fig. 10; see also Rozen et al., plate bearing three pairs of fixed (nonarti- 1978: figs. 8, 9) moderately robust, straight, culating) tubercles; labral plate distinct from without tubercles; abdominal segment X with and articulating with clypeus; clypeus pre- two lateral posteroventrally directed lobes, sumably very short. Mandible (Rozen et al., possibly but not certainly homologous with 1978: figs. 13, 14) short (by comparison with nomadine pygopod (muscle attachment un- those of Nomadinae and Ericrocidini) but known) and/or with somewhat similar struc- seemingly long because of very slender, tures in Rhathymini. Integument with scat- curved, fanglike apical part that curves into tered setae (those on sides of thoracic and buccal cavity when mandible closed; man- abdominal segments long) as well as exten- dibular base stout; inner edge smooth. La- sive areas oflong setiform spicules (Rozen et 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 17 al., 1978: fig. 24). All spiracles present, in out teeth or projections. Labiomaxillary re- normal position, approximately same size. gion membranous except for sides of max- Anus not detected. illae and base of labium; region not greatly GENUS STUDIED: Protepeolus. fused apically; maxilla with elongate, spiculate (except apparently in Melecta), adorally DESCRIPTION OF THE MELECTINI directed apical lobe; maxillary and labial palpi BASED ON FIRST INSTARS small but distinct papillae. Hypopharynx bilobed, nonspiculate, unsclerotized. HEAD: Shape more or less prognathous, not BODY: Form (figs. 14, 23, 28, 33) moder- as strongly so as in Nomadinae or Ericroci- ately elongate, straight, without lateral body dini, not dorsoventrally flattened; parietals tubercles; abdominal segment X rounded, shaped normally for bee larva. Integument without tubercles or other obvious structures of head capsule sclerotized but at most only for crawling. Integument without setae, usu- faintly pigmented; sclerotization ending at ally with patches offine spicules. All spiracles posterior margin of head; sensilla obscure, present and normal in location. Anus appar- nonsetiform, but at times borne on distinc- ently small transverse slit, apical in position, tive small tubercles; linear band of spinulae except presumably dorsal in position in Me- (small tubercles bearing apical cluster of mi- lecta pacifica fulvida. croscopic projections, figs. 15, 18, 19, 22) GENERA STUDIED: Xeromelecta, Melecta, ringing head from mandible to mandible over Thyreus, and Zacosmia. vertex. Tentorium complete; anterior pits REMARKS: The distinctive hypostomal distinct; posterior tentorial pits arising from groove found throughout this tribe is strik- rear of hypostomal groove; internal hypos- ingly similar to that of the Nomadinae. Pre- tomal ridge not well developed but conspic- sumably it had a separate uous troughlike hypostomal groove (fig. 16) evolutionary origin externally, suggestive of that of Nomadinae; in each group, but its function deserves to be weak ventral sclerotization of head (fig. 16) investigated to shed light on this matter. mesad of(below) hypostomal groove and behind palpus; rest oflabiomaxillary region un- Xeromelecta (Melectomorpha) sclerotized; pleurostomal ridge weakly de- californica (Cresson) veloped; epistomal ridge weakly developed Figures 14-22 laterad ofanterior tentorial pits, absent mes- The first instar ofthis species has not been ad of pits; pale dorsal median ecdysial line described before although Bohart (1970) and scarcely evident because of lack of pigmen- Torchio and Trostle (1986) presented pho- tation of head capsule; posterior margin of tographs of it. head capsule at right angles to hypostomal DiAGNOSIS: This species differs from the groove (Xeromelecta) or at obtuse angle be- first instars of Melecta pacifica fulvida, Za- cause of curving of ridge (Thyreus, Zacos- cosmia maculata, and Thyreus lieftincki in mia). Each antennal papilla (figs. 17, 18, 21) that its head capsule is more elongate ven- separated from head capsule by basal ring, trally as seen in lateral view (fig. 18) and its not fused to head as in other known clepto- labral tubercles are more elongate, more parasitic anthophorids; antenna large, elon- widely spaced apically, and larger (fig. 15). gate forward-curving, tapering apically, However, its head capsule (fig. 15) is less sometimes with sensilla borne on tubercles. elongate and its labral tubercles (fig. 15) and Labrum (fig. 15) sclerotized, fused with clyp- antennae (figs. 17, 18) are shorter than those eus bearing nonarticulating rounded or acute of Melecta separata callura. paired apical tubercles. Mandible moderately LENGTH: 2.2-2.8 mm (N = 4). short by comparison with that ofEricrocidini HEAD (figs. 15-18): Shape prognathous al- and many Nomadinae, but not as short as though not as strongly so as in Epeolini (Roz- that of Protepeolini, gradually tapering api- en, 1 989b), not dorsoventrally flattened, cally from rather stout base, not with apical somewhat elongate especially ventrally, and, part fanglike and abruptly narrower than base as seen from above or below (figs. 15, 16), as in Protepeolini; inner edge smooth, with- quadrate and nearly parallel-sided; foramen 18 AMERICAN MUSEUM NOVITATES NO. 3029

Figs. 14-18. Xeromelecta californica, first instar. 14. Entire body, dorsal view. 15-18. Head, dorsal, ventral, frontal, and lateral views. Scale refers to figure 14.

magnum narrower than width ofcapsule. In- nounced, extending from posterior tentorial tegument of head capsule thickly sclerotized pit to ventral mandibular articulation; this but only faintly pigmented; apices of man- groove formed by protruding ventral surface dible darkly pigmented; ventral integument of head and less protruding parietal; integu- between hypostomal grooves somewhat ment of groove not pebbled as in Epeolini sclerotized and faintly pigmented posterior but irregularly sculptured; sclerotization of to palpi (fig. 16); other areas including hy- head capsule extending mesad ofgrooves, as popharynx unsclerotized. Visible setae ab- discussed above; distance between grooves sent; head capsule with a narrow ring ofsharp- relatively large compared with that of Epeo- pointed spinulae extending from base of lini; pleurostomal ridge moderately weakly mandibles across vertex (figs. 15, 18, 19, 22) developed; epistomal ridge weakly developed (characteristic of known Melectini); integu- laterad of anterior tentorial pits, absent be- ment of antennae, labral tubercles, anterior tween them; pale dorsal ecdysial line only edge of labrum, and base of mandibles with faintly visible because of lack of pigmenta- many sensilla borne on small tubercles. Ten- tion elsewhere. Parietal bands not evident. torium including dorsal arms complete, Each antenna large, projecting, curving for- moderately developed; anterior tentorial pits ward, separated from head by basal ring, and, moderate in size; posterior pits in hyposto- as seen in frontal view (fig. 17), rising above mal grooves; hypostomal ridge (internal) not vertex; antennal sensilla borne on small tu- strongly developed but external groove pro- bercles, giving antenna knurled appearance. 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 19

j Va Figs. 19-22. Head ofXeromelecta californica. 19. Lateral view. Scale = 100 microns. 20. Frontolateral view. Scale = 50 microns. 21. Antenna, lateral view. Scale = 20 microns. 22. Spinulae. Scale = 2 microns.

Labral sclerotization, including that of tu- smooth, without teeth or sharp projections. bercles, continuous with that offront ofhead Labiomaxillary region (fig. 16) not fused, capsule, so that these areas do not articulate semisclerotized at base ofhead capsule; max- with one another; paired labral tubercles large, illa distinct, with elongate, spiculate, orally conical, apically acute, and with numerous directed apical lobe; maxillary palpus small, small sensilla-bearing tubercles apically and about as long as basal diameter, well defined, on mesial sides. Mandible elongate by com- and bearing tight cluster ofnumerous sensil- parison with that of nonhospicidal larva la; palpus situated at base ofapical maxillary (though shorter than that ofEpeolini), sharp- lobe and mesad of hypostomal groove; la- pointed, attenuate and slender apically, stout bium in front oflabial palpus somewhat pro- basally, and with sensilla-bearing small tu- tuberant, as seen in lateral view, nonspicu- bercles on outer aspect at base; inner edge late; labial palpus similar to maxillary palpus, 20 AMERICAN MUSEUM NOVITATES NO. 3029

positioned as in figures 16-18. Hypopharynx Shape more strongly prognathous, more bilobed, nonspiculate, and unsclerotized; hy- elongate. Integument ofhead capsule scarcely popharyngeal groove distinct. Salivary open- pigmented; sclerotization apparently as de- ing contiguous to hypopharyngeal groove (in scribed for X. californica except internal ridg- contrast to opening of mature larva which is es poorly defined. Each antenna enormously borne at apex of labium, [Rozen, 1969b]). long, more so than in any other melectine. BODY: Form (fig. 14) moderately elongate, Labral sclerotization apparently continuous straight; body without pronounced tubercles with front of head capsule; labral tubercles either laterally or dorsally; abdominal seg- longer than in any other known melectine. ments dorsally divided in caudal and cephalic Maxilla with apical lobe very long, nonspi- annulets; abdominal segment X rounded, culate. Labial palpus distinct but smaller than without pygopod or other special ambulatory maxillary palpus. features; venters ofprothorax and abdominal BODY: As described for Xeromelecta cali- segment IX not protruding. Integument with- fornica except for following: Form (fig. 23) out setae, but with patches of fine spicules more elongate. Integument without spicules. particularly ventrally. All spiracles present, Anus not detected. large, of about equal size, and dorsally di- MATERIAL STUDIED: 2 first instars, 14.5 mi rected (as opposed to dorsolaterally directed). N Coalinga, Fresno Co., California, March Anus apparently small, apically transverse 21, 1963, [cell] opened March 27, 1963; pre- slit. served in ethyl alcohol March 29, 1963 (R. MATERIAL STUDIED: 4 first instars, Cedar W. Thorp). In the collection of R. W. Thorp. Point Biological Station, Keith Co., Nebras- REMARKS: Thorp (1969) provided a de- ka, collected as eggs from nests ofAnthophora tailed account of the biology of this species occidentalis July 15, 1988, preserved as first attacking nests ofAnthophora edwardsii Cres- instars July 16, 1988 (J. G. Rozen); 1 first son. These specimens came from his inves- instar, same except collected as dead larva tigations. killed by meloid, July 18, 1988. REMARKS: P. F. Torchio (personal commun.) in reviewing this manuscript gave the Melecta pacifica fulvida Cresson following interesting account of the function Figures 28-32 ofthe spinulae (an autapomorphy ofthe Me- lectini) on the head of this species: "If the DiAGNosIs: This species can be easily dis- splitting of the egg chorion [at time of eclo- tinguished from Xeromelecta californica and sion] does not occur properly as the head Melecta separata callura because ofits short- capsule is moved, the larva shakes its head er antennae and labral tubercles and because from side to side and the spinulae act as would of its more robust head capsule as seen from a can opener in cutting through the egg cho- above or from the sides (figs. 29, 32). Indeed rion so that the necessary splitting ofthe cho- the pronounced differences between the two rion can continue via pressure applied by ex- species of Melecta are surprising. Its thicker pansion of the first 3 body terga. . antennae separate it from Zacosmia maculata. It appears to be most similar to Thyreus lieftincki, but presumably its more pro- Melecta separata callura nounced sensillar tubercles and anteriorly (Cockerell) projecting labium, as seen in lateral view (fig. Figures 23-27 32), will enable it to be recognized. DiAGNosIs: The extremely elongate head, LENGTH: 2.6-3.1 mm. labral tubercles, and antennae distinguish this HEAD (figs. 29-32): As described for Xe- species from all other Melectini including romelecta californica except for following: Melecta pacificafulvida, described below, and Head capsule not so elongate. Antenna short- M. luctuosa (Scopoli) (Giordani-Soika, 1936). er. Paired labral tubercles shorter. Maxillary LENGTH: 3.8-4.3 mm (N = 2). lobes apparently faintly spiculate apically. HEAD (figs. 24-27): As described for Xe- Labial palpus smaller than maxillary palpus. romelecta californica except for following: BODY: As seen from above (fig. 28) most 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 21

Figs. 23-27. Melecta separata callura, first instar. 23. Entire body, dorsal view. 24-27. Head, dorsal, ventral, frontal, and lateral views. Figs. 28-32. Melecta pacificafulvida, first instar. 28. Posterior part ofabdomen, dorsal view. 29-32. Head, dorsal, ventral, frontal, and lateral views. Scales refer to figures 23 and 28 respectively.

body segments with small paired lateral tu- house, USDA, Logan, Utah, June 1967 (P. bercle-like swellings. Anus apparently dorsal. F. Torchio). In the collection ofP. F. Torchio. MATERLAL STUDIED: 2 first instars, green- REMARKS: Torchio and Youssef(1968) de- 22 AMERICAN MUSEUM NOVITATES NO. 3029

scribed the egg deposition habits which re- braunsiana Friese, were treated by Rozen sulted in the larvae described here. (1969a).

Thyreus lieftincki Rozen Zacosmia maculata (Cresson) Figures 33-37 Figures 38-41 No first instar of this genus has been de- This larva was first described by Torchio scribed before. and Youssef (1968). DIAGNosIs: This larva differs from Xero- DiAGNOSIS: This first instar is similar to melecta californica and Melecta separata cal- those ofthe other melectines treated here. Its lura in that the head capsule is shorter in more darkly pigmented head capsule (figs. 38, relation to its width, as seen in dorsal and 39), slender antennae, and near absence of ventral views (figs. 34, 35) and the ventral sensilla-bearing tubercles (especially on man- surface is less elongate as seen from the side dibular bases) help to distinguish it. The head (fig. 37). The reduced pigmentation of the capsule, as seen from above or below (figs. head capsule and stouter antennae separate 38, 39), is more spherical rather than quad- this species from Zacosmia maculata. It is rate and parallel-sided as in the other species. most similar to the first instar ofMelecta pa- Furthermore, its antennae are directed more cifica fulvida but presumably can be recog- anteriorly, so that when seen in frontal view nized by the features presented in the diag- (fig. 40) they do not rise above the upper nosis of the latter. margin of the head. LENGTH: 3.0 mm (N = 2). LENGTH: 2.2 mm (N = 1). HEAD (figs. 34-37): As described for Xe- HEAD (figs. 38-41): As described for Xe- romelecta californica except for following: romelecta californica except for following: Head not so elongate as seen in lateral view Head not so elongate as seen in lateral view (fig. 37). Integument scarcely pigmented ex- (fig. 41), only vaguely quadrate, more spher- cept for mandibular apices; ventral integu- ical as seen from above or below (figs. 38, ment between hypostomal grooves appar- 39). Integumental pigmentation somewhat ently somewhat sclerotized but pigmentation more pronounced than in other two species absent. Sharp-pointed spinulae of head cap- ofmelectines. Sharp-pointed spinulae on head sule generally smaller and less pronounced; capsule smaller, fewer; sensilla-bearing tu- sensilla-bearing tubercles smaller, less pro- bercles nearly absent. Middle part of hypos- nounced. Tentorium weak but complete; tomal groove arching downward as seen in middle part of hypostomal groove arching lateral view, like that of Thyreus lieftincki, downward as see in lateral view rather than rather than X. californica; dorsal ecdysial line nearly straight; dorsal ecdysial line not visible more or less evident because of surrounding because of lack of pigmentation elsewhere. pigmentation. Antennae large but directed Antenna large, but somewhat smaller than in forward, in contrast to other two melectine X. californica. Paired labral tubercles mod- taxa, so that antennae do not arise above head erately large, apically rounded. Maxillary as seen in frontal view (fig. 40). Paired labral lobes shorter, but clearly evident. tubercles moderately large, apically rounded BODY (fig. 33): As described for Xerome- as in T. lieftincki. Maxillary lobes shorter, lecta californica. directed nearly mesially, and apparently non- MATERIAL STUDIED: 2 first instars, 3 mi spiculate; labium more forward projecting so south Avontuur, Cape Province, Republic of that, as seen in ventral view (fig. 39), its apex South Africa, collected as eggs November 15, is almost in line with apices of maxillae and 1966, preserved November 18, 1966 (J. G. hypopharynx is hidden, in contrast to less Rozen). forward-projecting labia of T. lieftincki and REMARECS: The mature larva ofthis species X. californica. was described by Rozen (1 969b). The biology BODY: Essentially identical to postcephalic ofboth this species and its host, Anthophora regions of Xeromelecta californica and Thy- 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 23

1.0 mm

Figs. 33-37. Thyreus lieftincki, first instar. 33. Entire body, first instar, dorsal view. 34-37. Head, dorsal, ventral, frontal, and lateral views. Figs. 38-41. Zacosmia maculata, head of first instar, dorsal, ventral, frontal, and lateral views. Scale refers to figure 33. reus lieftincki; integumental spiculation re- ofRodeo, Hidalgo Co., New Mexico, August duced by comparison with X. californica. 15, 1974 (J. G. Rozen). MATERiAL STUDIED: 1 first instar, 1 mi north REMARxs: Torchio and Youssef(1968) de- 24 AMERICAN MUSEUM NOVITATES NO. 3029

POSTERIOR TENTORIAL PIT

Figs. 42-46. Rhathymus bicolor, first instar. 42. Dorsal view ofbody. 43-46. Head, dorsal, ventral, lateral, and frontal views. Scale refers to figure 42. scribed the first instar, mature larva, and pupa length because of swollen parietals. Integu- of this species. Rozen (1969b) subsequently ment ofhead capsule heavily sclerotized and redescribed the mature larva from their ma- darkly pigmented; sclerotization ending pos- terial. teriorly at rear ofhead; sensilla not setiform, not borne on tubercles; capsule without band DEscRipTIoN OF THE RHATHYmINI BASED ofspinulae across vertex. Tentorium well de- ON FIRST INSTARS veloped (except for dorsal arms); anterior The following account is based the first in- tentorial pits distinct; posterior pits arising stars of Rhathymus bicolor described below from hypostomal ridges and posterior thick- and of an unidentified species previously de- ening of head; hypostomal ridge well devel- scribed by Camargo et al. (1975). oped but external groove not troughlike; scle- HEAD: Shape hypognathous; parietals rotization of head capsule extending below swollen (fig. 43) so that capsule appears glo- ridge but most of labiomaxillary region un- bose, unlike that of any other known clep- sclerotized; ventral sclerotized postoccipital toparasitic anthophorid; capsule with dor- bridge present; pleurostomal ridge well de- soventral distance exaggerated in relation to veloped; epistomal ridge well developed la- 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 25 terad ofanterior tentorial pits, absent mesad ridges somewhat pigmented and sclerotized of pits; distinct pale median dorsal ecdysial but fading before reaching labium except line present; posterior margin ofhead capsule ventral extension ofposterior rim offoramen not forming sharp angle with hypostomal apparently completely sclerotized forming ridge, but appearing as almost straight con- postoccipital bridge (fig. 44) (hence posterior tinuation ofit as seen in lateral view (fig. 45). end ofhead a complete sclerotized ring); api- Each antenna small, scarcely evident, fused ces of labral tubercles vaguely pigmented; with head capsule. Labrum bearing pair of other areas ofhead unpigmented. Visible se- large but not elongate, downward projecting tae absent; head capsule without spinulae (in tubercles and separated from clypeus by ar- contrast to ring ofspinulae in Melectini); sen- ticulating membrane. Mandible moderately silla minute, not borne on small tubercles. long; apex tapering to simple, sharp-pointed Tentorium complete except for dorsal arms, tip; base stout; inner edge smooth. Labiom- well developed; anterior tentorial pits mod- axillary region sclerotized only on sides of erate in size; posterior pits at end of hypos- maxillae; maxilla not fused with labium ex- tomal ridge; hypostomal ridge well devel- cept at base; maxilla with large, apically spi- oped, not associated with external groove, as culated, orally directed apical lobe; maxillary in Epeolini and Melectini, extending down- palpus subapical, not projecting, represented ward almost ventrally from posterior pit and only by cluster of sensilla; labium spiculate bending forward immediately in front ofpos- apically; labial palpus nonprojecting cluster terior mandibular articulation; integument of sensilla. Hypopharynx unsclerotized, bi- near ridge not pebbled, wrinkled, or in other lobed, nonspiculate. ways sculptured; sclerotization of head cap- BODY: Form (fig. 42) elongate, slender, sule extending mesad of(below) ridge, as de- straight, without tubercles; abdominal seg- scribed above; distance between hypostomal ment X with pair of conspicuous, short, pos- ridges great compared with Epeolini; pleu- teroventrally directed lobes, possibly but not rostomal ridge well developed; epistomal certainly homologous with paired structures ridge well developed laterad of anterior ten- on abdominal segment X ofProtepeolini and torial pits, absent mesad ofpits. Parietal bands Nomadinae. Integument without setae but absent. Antenna small, scarcely evident ex- with patches of fine spicules. Spiracles pres- cept for tight cluster ofsensilla which are not ent, of equal size, normal in position. Anus borne on tubercles. Labrum separated from an inconspicuous, apical, transverse slit be- clypeus by articulating membrane; paired la- tween lobe bases. bral tubercles large (fig. 46), conical, down- GENus STUDIED: Rhathymus. ward directed, bearing many small sensilla on anterior surfaces. Mandible elongate, Rhathymus bicolor Lepeletier sharp-pointed, apically attenuate, basally Figures 42-46 stout, without tubercles on outer aspect at base; inner edge smooth, without teeth or DIAGNOsIs: This larva can be immediately sharp projections. Maxillae and labium not distinguished from other known first instars greatly fused, mostly separate as seen in ven- ofcleptoparasitic anthophorids because ofthe tral view (fig. 44); maxilla with large, spicu- spherical appearance of the head capsule as late, orally directed apical lobe; maxillary seen in dorsal view (fig. 43) and its pro- palpus faintly if at all projecting but easily nounced hypognathous conditions as seen in recognized by close cluster ofsensilla; palpus lateral view (fig. 45). subapical in position ofmaxilla; labium pro- LENGTH: 4.5 mm (N = 1). jecting in front oflabial palpi as seen in ven- HEAD (figs. 43-46): Shape distinctly hy- tral view (fig. 44), so that hypopharyngeal pognathous; parietals together globose, un- groove and opening ofsalivary gland not vis- usually wide in relation to horizontal diam- ible from below; labial apex strongly spicu- eter of foramen. Integument of head capsule late; palpus similar to maxillary palpus. Hy- heavily sclerotized, moderately pigmented, popharynx very short so that opening of except apices ofmandibles darkly pigmented; salivary duct and mouth close to each other; ventral integument mesad of hypostomal hypopharynx nonspiculate, unsclerotized, 26 AMERICAN MUSEUM NOVITATES NO. 3029

apparently somewhat bilobed; hypopharyn- pigmented; unlike that in any other known geal groove distinct, but difficult to see be- cleptoparasitic anthophorid, sclerotization cause of being so close to mouth. Salivary extending posteriorly over anterior part of opening immediately in front of hypophar- prothorax; head capsule with sensilla seti- yngeal groove, not visible in views illustrated. form according to Michener (1957) and Oliv- BODY: Form (fig. 42) elongate, straight, eira (1966), but these setae apparently mostly without tubercles laterally or dorsally; ab- lost on specimen examined here; only sensilla dominal segments perhaps divided dorsally at base of mandibles borne on small tuber- into caudal and cephalic annulets, but de- cles; capsule without ring of sharp-pointed marcation vague; abdominal segment X spinulae across vertex. Tentorium incom- bearing pair of short but large, posteroven- plete; anterior tentorial pits small but pres- trally directed lobes (presumably ambulatory ent; small posterior tentorial pits present in features); venters of prothorax and abdomi- vague hypostomal ridges on ventral side of nal segment IX not protruding. Integument head well anterior to posterior margin ofhead without setae but with patches of fine short sclerotization; labiomaxillary region sclero- spicules in many areas; spicules especially well tized and fused with head capsule, unlike that developed on posteroventral lobes ofabdom- in Nomadinae but similar to that in Ericro- inal segment X and parts of abdominal seg- cidini; internal head ridges mostly obscure ment IX. All spiracles present, about equal because ofheavy sclerotization; pale median in size, moderately small, and dorsolaterally dorsal ecdysial line absent; angle formed by directed (rather than dorsally directed). Anus posterior margin of head and hypostomal an inconspicuous, apical, transverse slit be- groove in lateral view unknown because these tween bases of ambulatory lobes. ridges not evident. Each antenna moderately MATERLAL STUDIED: 1 first instar, Maracas small, well defined, anteromesally directed, Valley, Trinidad, April 7, 1965, from nest of and fused with head capsule. Labrum scler- Epicharis rustica (F. D. Bennett). otized, fused with clypeus; small labrum REMARKS: The host cell from which the bearing single conspicuous median tubercle specimen came bore a small hole near the and therefore unlike that ofany known clep- center ofthe closure, and the vacated chorion toparasitic anthophorid. Mandible attenu- remained attached to the inner surface ofthe ated, slender apically, not stout as in Ericro- closure. Further details on the biology ofthis cidini; inner edge smooth. Labiomaxillary species will be presented elsewhere. region completely fused, sclerotized, and Camargo et al. (1975) described and illus- highly modified; maxillary palpus faintly trated the first instar ofRhathymus sp. Their produced; labial palpus not produced, rep- figure 13 ofa recently eclosed larva shows no resented only by sensilla. Hypopharynx ap- differences between their species and R. bi- parently an unpigmented paired attenuated color. However, their figure 10 apparently de- structure immediately below mouth. picts a subsequent instar; compared with a BODY: Form elongate, straight, not taper- first instar, its mandibles are less attenuate, ing posteriorly, most segments laterally its palpi project farther, and the parietal bands rounded; abdominal segments IX and X are evident. fused, somewhat similar to condition in the Ericrocidini but fusion even more complete, DEscIuPrION OF THE ISEPEOLIN BASED without pygopod or other ambulatory mod- ON THE FIRST INSTAR ifications. Integument with a few setae especially on the anterior part, unlike that in Both Michener (1957) and Oliveira (1966) any cleptoparasitic anthophorid except Pro- described the first instar of Isepeolus viperi- tepeolus (which uniquely has elongate, seti- nus (Holmberg). form dorsal body spicules). All spiracles pres- HEAD: Shape prognathous, elongate, and, ent; those of abdominal segment VIII unlike that ofany other Nomadinae or other enlarged, situated at posterior margin of seg- hospicidal anthophorid, strongly dorsoven- ment, as is characteristic ofEricrocidini. Anus trally flattened; parietals not swollen. Integ- not evident. ument ofhead capsule heavily sclerotized and GENUS STUDIED: Isepeolus. 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 27

Isepeolus viperinus (Holmberg) because ofthickening of head capsule except Figures 47-53 in vicinity of posterior tentorial pit; hypostomal groove absent; pleurostomal and hy- The redescription of this species is based postomal ridges obscure; ecdysial line not ev- on the specimen studied by Michener (1957). ident. Parietal bands not evident. Each LENGTH: 4.6 mm (N = 1). antenna (figs. 49, 52) small, projecting sharp- HEAD (figs. 49-53): Shape strongly prog- ly anteromesally; antennae behind anterior nathous, very strongly dorsoventrally flat- edge of head capsule, hence not projecting tened especially anteriorly, so that head seen beyond dorsal mandibular articulation, as in lateral view (fig. 52) wedge-shaped; head seen in lateral view (fig. 52), not arising above capsule elongate both ventrally and dorsally; vertex as seen in frontal view; each antenna sides nearly parallel in front ofposterior con- strongly dorsoventrally flattened, smooth ex- striction, converging slightly anteriorly; fo- cept for sharp-pointed, apical bifurcation as ramen narrowerthan maximum width ofhead seen from above (fig. 49). Labral sclerotiza- capsule, as seen from above or below (figs. tion continuous with that of frons and clyp- 49, 50) but head not constricted posteriorly eus; labral apex produced as single median as seen in lateral view (fig. 52). Integument tubercle bearing sensilla. Mandible slender, of head capsule and also ventral surface of elongate, sharp-pointed, evenly tapering ex- head heavily sclerotized, darkly pigmented, cept for somewhat expanded base bearing i.e., labiomaxillary region sclerotized and small tubercles; inner edge smooth, without completely fused with head capsule; sclero- teeth or projections. Labiomaxillary region tization of ventral surface of head extending (fig. 50) fused and sclerotized; maxillary pal- forward and joining with sclerotization of pus (fig. 5 1) small, slightly produced as weak front of head capsule mesad of mandibular oblique ridge; apex of maxilla (fig. 51, max- bases and laterad of presumed unpigmented illary lobe?) produced anteromedially as flat hypopharyngeal region; integument of head sclerotized continuation ofventral surface of nearly smooth; presumed hypopharynx and head to form serrated-crenulated anterior edge epipharyngeal surface of labrum unpigment- of head between labrum and base of man- ed and unsclerotized (fig. 5 1); unlike in other dible (see Remarks below); hence anterior known hospicidal first instar anthophorids, edges ofmaxillae apparently serving as knife- sclerotization of head extending posteriorly like, destructive edge of head against which onto prothorax both dorsally and especially closing sickle-shaped mandibles force host ventrally (figs. 49, 50). Visible setae mostly immature (analogous to knifelike edge ofhy- absent, although few short setiform sensilla popharynx of Epeolini (Rozen, 1989b); an- obvious on labrum (both Michener, 1957, teromedially apices of maxillae curving and Oliveira, 1966, illustrated the head cap- downward along mesal edges that surround sule as bearing rather long, conspicuous setae; unsclerotized structures (fig. 51, hypophar- these no longer present on specimen studied ynx?) in vicinity of mouth; labial palpus not here, presumably having been lost after pres- produced, represented only by one or more ervation); head capsule without spinulae; in- sensilla behind salivary opening as confirmed tegument of antennae, apex of labrum, and by small but distinctly produced palpus of base of mandibles with numerous fine sen- second instar in similar position. Hypophar- silla; sensilla at base of mandibles borne on ynx presumably unsclerotized, unpigmented, small tubercles; integument of head capsule attenuate paired processes extending forward dorsally and laterally with sensilla, presum- nearly as far as labral apex; hypopharyngeal ably alveolae of missing setiform sensilla. groove not clearly evident. Salivary opening Tentorium incomplete; anterior tentorial pits (fig. 51) very large, directed somewhat ante- small (fig. 53); anterior arms weak and short; riorly, and connecting to very wide salivary posterior pits in faint ridge presumably rep- duct. resenting junction of hypostomal ridge and BODY: Form (fig. 47) elongate, parallel-sid- posterior thickening ofcapsule; these pits far ed, somewhat truncate posteriorly because of forward ofposterior edge of sclerotization of modifications of terminal abdominal seg- head (figs. 50, 52); hypostomal ridge obscure ments and lateral body swellings; body with- 28 AMERICAN MUSEUM NOVITATES NO. 3029

Figs. 47-53. Isepeolus viperinus, first instar. 47. Entire body, dorsal view. 48. Apex of abdomen, lateral view. 49, 50. Head, dorsal and ventral views. 51. Anterior part of head, enlarged, ventral view. 52, 53. Head, lateral and frontal views. Scale refers to figure 47. out dorsal tubercles but most segments swol- X (figs. 47, 48) greatly fused; presumably IX len laterally; abdominal segments perhaps represented laterally by swollen lobes be- divided into weak cephalic and caudal an- tween which is small segment X; distinct in- nulets dorsally; abdominal segments IX and tersegmental line between IX and X not ev- 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 29 ident; pygopod not present; venters of extending over anterior part of thorax (as in prothorax and abdominal segment IX not Isepeolus); sensilla nonsetiform, not borne on protruding. Integument with a few, very fine, tubercles except in Ericrocis; capsule without short setae and mostly without spicules ex- band of spinulae across vertex (as found in cept for scattered indistinct patches. All spir- Melectini) but head tubercles ofEricrocis su- acles present; those ofthorax and abdominal perficially like spinulae. Tentorium, except segments I-VII subequal in size, dorsally di- for dorsal arms, complete (not known for Er- rected, and in normal position; spiracles of icrocis); anterior pits distinct but small; pos- abdominal segment VIII unusually large, sit- terior pits in hypostomal groove anterior to uated near posterior edge of segment, and posterior margin of head; troughlike hypos- posteriorly directed; spiracular tracheae of tomal groove present and granulate in Eri- segments VIII much larger than those of oth- crocis, absent in Aglaomelissa and Mesoplia; er segments and connecting to very large tra- hypostomal ridge only vaguely visible as cheal trunks, hence larvae appearing to be somewhat darker area; entire ventral surface semi-metapneustic. Anus not evident. ofhead (labiomaxillary region) sclerotized and MATERLAL STUDIED: 1 first instar, Arau- darkly pigmented; pleurostomal and episto- caria, Parana, Brazil, January 13, 1956 (C. mal ridges scarcely evident because of head D. Michener) from cell of Colletes kerri sclerotization; distinct pale dorsal ecdysial line Moure. In the collection ofthe University of present; posterior margin of head capsule at Kansas, Lawrence, currently on deposit at the right angle to hypostomal ridge as seen in American Museum of Natural History. lateral view (fig. 59). Each antenna complete- REMARKS: A second instar as well as sev- ly fused to head capsule, large, anterodorsally eral subsequent instars were preserved in the directed, slightly to considerably dorsoven- same vial as this specimen. In sharp contrast trally flattened. Labrum sclerotized, fused to the first instar, head capsules of the other with clypeus, vaguely bilobed apically. Man- specimens were pale and weakly sclerotized, dible robust, elongate, evenly tapering entire and the mandibles were not strongly curved. length from stout base; inner edge without These facts indicate that only the first instar teeth or sharp projections (ventral mandib- is hospicidal. ular surface of Aglaomelissa uniquely with The anatomy of the first instar is highly elongate sulcus). Labiomaxillary region modified and the homologies of some of the greatly fused and sclerotized so that even palpi mouthparts may be open to reinterpretation. scarcely if at all recognizable. Hypopharynx The maxillary palpi are clearly identifiable, bilobed, nonsclerotized, unpigmented. and therefore the sharp-edged, anteromesal BODY: Form moderately elongate, straight, extensions of the anterior ventral head sur- without dorsal or lateral tubercles except: (1) face in front of them appear to be the max- lateral swellings on abdominal segments IX illary apices (lobes). Because the pharynx lies and X of at least Aglaomelissa and Mesoplia immediately above the elongate, paired un- giving larva truncated appearance, and (2) pigmented median processes, they are con- Ericrocis with pair of knoblike lateral pro- sidered to be the hypopharynx. thoracic tubercles; lateral swellings on abdominal segment X possibly assisting in Integument without setae, with OF THE ERICROCIDINI crawling. DESCRIPrION some fine spiculation. All spiracles present; BASED ON FIRST INSTARS those ofthorax ofAglaomelissa and Mesoplia First instars of the Ericrocidini have not somewhat reduced in size; those of abdom- been previously described except for a brief inal segment VIII of Aglaomelissa and Me- account ofMesoplia sp. (Vinson et al., 1987). soplia situated near posterior edge of seg- HEAD: Shape strongly prognathous, elon- ment, directed somewhat posteriorly, and gate, not strongly dorsoventrally flattened al- larger than others; all spiracles of Ericrocis though slightly so in Aglaomelissa; parietals presumably subequal in size and normal in not swollen. Integument of head capsule position. Anus faintly visible at least on one heavily pigmented and sclerotized; scleroti- species, situated apically between lateral lobes zation ending posteriorly at rear ofhead, not of segment X. 30 AMERICAN MUSEUM NOVITATES NO. 3029

GENERA STUDIED: Aglaomelissa, Mesoplia, biomaxillary region) unevenly wrinkled; only and Ericrocis. hypopharynx and epipharyngeal surface of labrum unpigmented and unsclerotized, as seen in ventral view (fig. 57). Visible setae Aglaomelissa duckei (Friese) absent; head capsule without spinulae (as Figures 54-59 found in Melectini); integument of antennae, Several females ofMesoplia rufipes and one anterior edge of labrum, outer surfaces of ofAglaomelissa duckei were collected at the mandibles, and labiomaxillary region with nesting site of Centris carrikeri where this numerous fine sensilla which are generally larva was recovered. This first instar was not borne on small tubercles (those at base identified by the fact that it differs from the ofmandible borne on minute tubercles). Ten- first instars (described below) clearly associ- torium, except for dorsal arms, complete, but ated with adults of M. rufipes from another weak; anterior tentorial pits moderate in size; locality on Trinidad. posterior pits in hypostomal grooves well in DiAGNosIs: First instars ofMesoplia rufipes front of posterior edge of head as seen in and Aglaomelissa duckei are similar in most ventral view (fig. 57); hypostomal ridge ob- respects. They can be distinguished from one scure because of thickness of head integu- another because the mandible of A. duckei ment; external hypostomal groove absent; alone bears an elongate sulcus on its ventral distance between ridges large compared with surface (fig. 57), its head is more dorsoven- that of Epeolini; integument mesad of hy- trally flattened (fig. 58, in contrast to fig. 62), postomal ridge and immediately behind base its mandible is more robust and less curved of mandible produced into transverse ridge, apically (figs. 56, 57, in contrast to figs. 60, suggestive of(but presumably only analogous 61), and its antennal apex extends well be- to) hypostomal process of Epeolini (Rozen, yond the dorsal mandibular articulation (fig. 1 989b); this ridge perhaps part of maxilla as 59, in contrast to fig. 63). The first instar of suggested by counterpart in Isepeolus viper- Ericrocis (figs. 64-67) can immediately be inus (which see); pleurostomal ridge scarcely separated from first-stage larvae ofAglaome- evident because of thickness of head sclero- lissa and Mesoplia because of its conspicu- tization; epistomal ridge absent; pale dorsal ous, sharp-pointed head tubercles, the paired ecdysial line conspicuous from rear of head knoblike prothoracic tubercles, nonfusion of capsule to apex of labrum because of sur- the dorsal and ventral head sclerites mesad rounding pigmentation. Parietal bands not of the mandibular bases, and salivary open- evident. Each antenna large, projecting an- ing not surrounded by the labiomaxillary terodorsally beyond dorsal mandibular artic- sclerotization. ulation as seen in lateral view (fig. 59), rising LENGTH: 3.4 mm (N = 1). above vertex as seen in frontal view (fig. 58); HEAD (figs. 56-58): Strongly prognathous, antenna smooth (not knurled as in Melecti- dorsoventrally flattened (area between anten- ni), slightly ventrodorsally flattened, apically nae and posterior margin of head nearly flat rounded and not conical; antennal sensilla as seen in lateral view, fig. 59, elongate ven- not borne on small tubercles. Labral sclero- trally as well as dorsally, with sides nearly tization continuous with that of clypeus and straight but converging slightly anteriorly; fo- frons, not separated by articulating mem- ramen narrower than width of head capsule. brane or ridge. Labrum without distinct tu- Integument of head capsule heavily sclero- bercles but with apical pair of obtuse projec- tized, darkly pigmented; mandibles darkly tions bearing numerous sensilla, possibly pigmented throughout; ventral surface ofhead homologous with labral tubercles. Mandible between hypostomal ridges sclerotized and robust, elongate, sharp-pointed (although ex- darkly pigmented; this sclerotization extend- treme apex of left mandible slightly expand- ing forward andjoining sclerotization offront ed, fig. 57); mandible evenly tapering its en- of head capsule mesad of mandibular bases tire length, not stout at base and slender and laterad of hypopharynx; integument of apically; inner edge smooth, without teeth or head capsule behind mandibular corium fine- sharp projections (except for extreme apex of ly pitted; that of ventral surface of head (la- left mandible, as described above); ventral 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 31

60 61 Figs. 54-59. Aglaomelissa duckei, first instar. 54. Entire body, dorsal view. 55. Apex of abdomen, lateral view. 56-59. Head, dorsal, ventral, frontal, and lateral views. Figs. 60-63. Mesoplia rufipes, head of first instar, dorsal, ventral, frontal, and lateral views. Scale refers to figure 54. surface ofeach mandible, unlike that ofother the same time). Labiomaxillary region greatly cleptoparasitic anthophorid tribes, with elon- fused and extensively sclerotized, so that even gate sulcus (fig. 57) which may accommodate palpi are nearly indistinguishable; maxillary apex of opposing mandible when it is closed palpus vague, anteriorly declivous area la- (this suggests that lower mandible [the left terad of salivary opening, recognized by nu- one on this specimen] closes more tightly than merous sensilla; labial palpus indistinguish- the upper, for there is no dorsal sulcus on able except apparently for grouping ofsensilla either mandible which would seemingly be behind and mesad ofmaxillary sensilla; palpi necessary if both mandibles closed tightly at fused (not articulating) with labiomaxillary 32 AMERICAN MUSEUM NOVITATES NO. 3029

HYPOSTOMAL GROOVE

Figs. 64-67. Ericrocis lata or pintada, head of first instar, dorsal, ventral, frontal, and lateral views. region. Hypopharynx unsclerotized, unpig- between protruding lateral lobes of abdom- mented, and bilobed; hypopharyngeal groove inal segment X. not evident. Salivary opening in sclerotized MATERIAL STUDIED: 1 first instar, Maracas area, somewhat posterior to hypopharynx. Valley, Trinidad, West Indies, collected as BODY: Form (fig. 54) elongate, parallel-sid- egg on February 23, 1966; preserved as larva ed, posteriorly somewhat truncate because of on February 24, 1966 (F. D. Bennett, J. G. modifications of last two abdominal seg- Rozen), from cell of Centris carrikeri. As- ments; body without tubercles laterally or sociated adult identified by Roy R. Snelling. dorsally except abdominal segment IX pro- REMARKS: As will be reported elsewhere, tuberant on sides and abdominal segment X the egg from which this larva emerged was with posterolateral swelling on each side; ab- attached by one end to the closure, an indi- dominal segments not dorsally divided into cation that the egg was deposited through an caudal and cephalic annulets; abdominal seg- opening in the closure after the cell had been ment X dorsally fusing with segment IX so sealed. that intersegmental line indistinguishable; pygopod, as in Epeolini, not present, but paired lateral swellings may serve ambula- Mesoplia rufipes (Perty) tory function; venters of prothorax and ab- Figures 60-63 dominal segment IX not protruding. Integ- DIAGNOSIS: See Diagnosis ofAglaomelissa ument without setae, mostly nonspiculate duckei. although a few indefinite patches seen dor- LENGTH: 4.8 mm (N = 1). sally toward abdominal apex. All spiracles HEAD (figs. 60-63): As described for Agla- present and most directed dorsally; those of omelissa duckei except for following: Sides of thorax somewhat smaller than those of ab- head capsule subparallel. Integument ofhead domen; those ofabdominal segment VIII un- capsule behind mandibular corium smooth; usual in that they are situated toward pos- ventral surface of head between hypostomal terior edge ofsegment, are larger than others, ridges (labiomaxillary region) smooth; sen- and are directed somewhat posteriorly; spi- silla of mandibles not borne on small tuber- racular tracheae of segment VIII of greater cles. Each antenna large, but smaller than that diameter than those ofother segments. Anus of A. duckei, not projecting beyond dorsal apparently an indistinct scar apically situated mandibular articulation as seen in lateral view 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 33

(fig. 63). Mandible directed anteriorly (fig. 63) DIAGNosIs: See diagnosis of Aglaomelissa (rather than anteroventrally, fig. 59), less ro- duckei. bust than that of A. duckei, slightly more LENGTH: Probably approximately 4.0 mm elongate, and more strongly curved apically (N= 1). than basally but evenly tapering; ventral sur- HEAD (figs. 64-67): As described for Agla- face ofeach mandible smooth, without elon- omelissa duckei except for following: Top of gate sulcus. Maxillary and labial palpi rec- head capsule as seen in lateral profile (fig. 67) ognized only by grouping of sensilla (fig. 61). curved; sides of head capsule diverging an- BODY: As described forAglaomelissa duck- teriorly as seen from above or below (figs. 64, ei except for following (specimens distorted, 65). Although head capsule sclerotized and hence not illustrated): Form perhaps tapering darkly pigmented ventrally and dorsally, pig- slightly posteriorly, more so than in A. duck- mented sclerotization oftwo surfaces not fus- ei. Integument bearing finely spiculated ven- ing anteriorly mesad ofmandibularbases (figs. tral patch on most body segments. Thoracic 65, 66), unlike in other two ericrocidine spe- spiracles somewhat smaller than abdominal cies (figs. 57, 61); integument ofhead capsule ones; abdominal ones small except spiracles behind mandibular corium smooth except for on abdominal segment VIII much larger, sharp-pointed sensilla-bearing tubercles; that about twice diameter of others; these spira- of labiomaxillary region also with sharp- cles on posterior edge of segment, directed pointed sensilla-bearing tubercles (in addi- posteriorly, as in A. duckei, and with enlarged tion to other sensilla). Head capsule without tracheae. Anus not certainly visible on dis- spinulae (but see discussion under Remarks). torted specimen. Condition of tentorium uncertain because MATERIAL STUDIED: 1 first instar, 1 cast only weak posterior and anterior arms extant first instar exuviae still attached to chorion, in cast skin; external hypostomal groove con- Hollis Reservoir, Trinidad, West Indies, spicuous, with granulations; integument im- March 7, 1968, preserved March 10, 1968 mediately behind mandible and mesad of (J. G. and B. L. Rozen) from nest ofEpicharis groove not ridgelike, merely marking ante- albofasciata Smith. Associated adults iden- rior margin ofmaxillary sclerotization; pleu- tified by Roy R. Snelling. rostomal ridge and lateral arms of epistomal REMARKS: Three host cells were parasitized ridge weak but evident in relation to heavily by the Mesoplia at this site. In two cells, a sclerotized head. Antenna not projecting an- small hole was found in the center of the terior to dorsal mandibular articulation as closure, indicating that the Mesoplia female seen in lateral view (fig. 67); antenna ap- made a small hole through which she intro- pearing somewhat similar to that ofMelectini duced her egg into the cell. Further details because sensilla borne on small tubercles, but will be published elsewhere. antenna strongly dorsoventrally flattened; Similar observations regarding egg depo- frons between antennae produced as trans- sition and chorion attachment to first instar verse granulated ridge. Mandibular ventral were reported by Vinson et al. (1987) for an surface without sulcus. Maxillary and labial unidentified Mesoplia. palpi not visible except perhaps for sensilla; labium immediately below (behind) salivary opening projecting as small transverse ridge Ericrocis lata (Cresson) or pintada bearing conspicuous granulations, unlike in Snelling and Zavortink any other known parasitic first instar. Sali- Figures 64-67 vary opening immediately above anterior labial ridge and separated from labiomaxillary Knowledge of the larva of this species sclerotization by short distance (fig. 66). comes from a cast first-instar integument ac- BODY: Form probably elongate and par- companied by the second instar. Because both allel-sided, as judged from cast exuviae; pro- species ofEricrocis were collected at the time thorax, unlike in any other known clepto- that the parasitized host nest was discovered, parasites, with pair of lateral, knoblike the identification of the species of cleptopar- tubercles (figs. 64, 67) which persist in second asite is uncertain. instar; other aspects of body form uncertain 34 AMERICAN MUSEUM NOVITATES NO. 3029 because of telescoping and wrinkling of in- soplia were clearly near to the posterior mar- tegument, but for reasons explained in Re- gin of segment VIII. Hence all spiracles on marks, form of posterior part of body prob- the first instar of Ericrocis may be placed ably similar to shapes ofAglaomelissa duckei normally. and Mesoplia rufipes. Integument spiculate The first instar ofEricrocis superficially re- on venter of most segments. All spiracles sembles that of the Melectini. In addition to present and subequal in size, apparently in- general head shape, the sharp-pointed tuber- cluding those ofabdominal segment VIII; po- cles on the Ericrocis head capsule appear to sition ofspiracles ofabdominal segment VIII be melectine spinulae. However, they do not unknown but for reasons expressed in Re- extend across the vertex as do spinulae in all marks, possibly normal, i.e., not on posterior melectines and microscopically each bears a margin of segment. single sensillum and not an apical rosette of MATERIAL STUDIED: 1 cast integument of points as is the case with spinulae (fig. 22). first instar, 1 mi east Douglas, Cochise Co., Furthermore similar sharp-pointed tubercles Arizona, August 20, 1990 (J. G. Rozen, J. are also present in the labiomaxillary region Krieger), collected with 5 Ericrocis eggs, 1 ofEricrocis, but spinulae never appear in this vacated Ericrocis chorion, 1 Centris egg, and region in melectines. Other features of Eri- 1 live second instar of Ericrocis from single crocis are markedly different from the Me- cell of Centris possibly atripes Mocsary. lectini: antennae dorsoventrally flattened and REMARXS: The host bee, with pale meso- fused with the integument of the head; la- soma and black metasoma, was observed biomaxillary region extensively fused and making the nest on August 16 but was not sclerotized; palpi not evident. collected. Centris atripes was the commonest species present with this color pattern. The shallow nest excavated on August 20 con- REFERENCES sisted of a single cell that was heavily parasitized. The Centris egg as well as 5 Ericrocis Alexander, B. 1990. A cladistic analysis ofthe nomadine bees eggs floating on or near the surface of the (Hymenoptera: Apoidea). Systematic semiliquid provisions presumably had been Entomol. 15: 121-152. killed by the first instar that had emerged Baker, J. R. from the vacated Ericrocis chorion and that 1971. Development and sexual dimorphism then molted to the second instar. of larvae of the bee genus Coelioxys. J. The head capsule of the cast first instar Kansas Entomol. Soc. 44: 225-235. exuviae was as easy to study as that ofa com- Bohart, G. E. plete first instar, and the anterior part of the 1970. The evolution ofparasitism among bees. body, split along the dorsal midline as far the Utah State Univ. 41st Honor Lecture, second abdominal segment, clearly revealed Spring, 1970, The Faculty Assoc., 33 the paired prothoracic PP. lateral, knoblike tu- Camargo, J. M. F., R. Zucchi, and S. F. Sakagami bercles. The rest of the body form was diffi- 1975. Observations on the bionomics of Ep- cult to interpret because oftelescoping ofthe icharis (Epicharana) rusticaflava (Oliv- cast integument. The second instar possessed ier) including notes on it parasite, posterolateral swellings on abdominal seg- Rhathymus sp. (Hymenoptera, Apo- ments IX and X just as did a second instar idea: Anthophoridae). Studia Entomol. of Mesoplia ruJipes, suggesting that the first 18: 313-340. instar ofEricrocis resembles that ofMesoplia Ehrenfeld, J., and J. G. Rozen and Aglaomelissa. 1977. The cuckoo bee genus Kelita, its system- The position ofthe last pair ofspiracles on atics, biology and larvae. Am. Mus. the Ericrocis exuviae could not be ascer- Novitates 2631: 24 pp. tained. These spiracles on the live second in- Giordani-Soika, A. star 1936. Etologia e larva primaria di Melecta luc- were normal in position, approximately tuosa Scop. [Hym. Apidae]. Boll. Soc. halfway between the anterior and posterior Entomol. Italiana 68: 47-48. margin of segment VIII. However, the ter- Graenicher, S. minal spiracles of the second instar of Me- 1905. Some observations on the life history 1991 ROZEN: CLEPTOPARASITISM IN ANTHOPHORID BEES 35

and habits of parasitic bees. Bull. Wis- species ofAfrican Thyreus with life his- consin Nat. Hist. Soc. 3: 153-167. tory notes on two species ofAnthophora Griitte, H. (Hymenoptera: Anthophoridae). J. New 1935. Zur Abstammung der Kuckucksbienen York Entomol. Soc. 77: 51-60. (Hymenopt., Apid.). Archiv. Naturg., 1969b. The larvae of the Anthophoridae (Hy- N. F. 4: 449-534. menoptera, Apoidea). Part 3. The Me- Michener, C. D. lectini, Ericrocini, and Rhathymini. Am. 1944. Comparative external morphology, Mus. Novitates 2382: 24 pp. phylogeny, and a classification of the 1971. Biology and immature stages ofMoroc- bees (Hymenoptera). Bull. Am. Mus. can panurgine bees (Hymenoptera, Nat. Hist. 82: 151-326. Apoidea). Am. Mus. Novitates 2457: 37 1954. Bees of Panamfa. Bull. Am. Mus. Nat. PP. Hist. 104: 175 pp. 1977. Immature stages of and ethological ob- 1957. Notes on the biology of a parasitic bee, servations on the cleptoparasitic bee Isepeolus viperinus (Hymenoptera, An- tribe Nomadini (Apoidea, Anthophor- thophoridae). Entomol. News 68: 141- idae). Am. Mus. Novitates 2638: 16 pp. 146. 1984a. Nesting biology of diphaglossine bees 1974. The social behavior ofthe bees. A com- (Hymenoptera, Colletidae). Am. Mus. parative study. Cambridge, MA: Har- Novitates 2786: 33 pp. vard Univ. Press. 1984b. Comparative nesting biology of the bee Oliveira, B. Lucas de tribe Exomalopsini (Apoidea: Antho- 1966. Descri ao de estidios imaturos de phoridae). Am. Mus. Novitates 2798: Isepeolus viperinus (Holmberg) e con- 37 pp. frontacoes com outras larvas de Antho- 1986. The natural history of the Old World phoridae parasitas conhecidas (Hyme- nomadine parasitic bee Pasites macu- noptera-Apoidea). Bol. Univ. Federal latus (Anthophoridae: Nomadinae) and Parania, Zool. 2: 163-176. its host Pseudapis diversipes (Halictidae: Roig-Alsina, A. Nomiinae). Am. Mus. Novitates 2861: 1989. The tribe Osirini, its scope, classifica- 8 pp. tion, and revision of the genera Pare- 1989a. Two new species and the redescription peolus and Osirinus (Hymenoptera, ofanother species ofthe cleptoparasitic Apoidea, Anthophoridae). Univ. Kan- bee genus Triepeolus with notes on their sas Sci. Bull. 54: 1-23. immature stages (Anthophoridae: No- 1990. Coelioxoides Cresson, a parasitic genus madinae). Am. Mus. Novitates 2956: of Tetrapediini (Hymenoptera: Apo- 18 pp. idea). J. Kansas Entomol. Soc. 63: 279- 1989b. Morphology and systematic significance 287. of first instars of the cleptoparasitic bee In press. Cladistic analysis of the Nomadinae tribe Epeolini (Anthophoridae: Noma- s. str. with description of a new genus dinae). Am. Mus. Novitates 2957: 19 (Hymenoptera: Anthophoridae) PP. Rozen, J. G. 1989c. Life history studies of the "primitive" 1954. Morphological description of the larva panurgine bees (Hymenoptera: Andren- ofOreopasites vanduzeei Cockerell. Pan- idae: Panurginae). Am. Mus. Novitates Pacific Entomol. 30: 203-207. 2962: 27 pp. 1957. External morphological description of Rozen, J. G., and S. L. Buchmann the larva of Exomalopsis chionura 1990. Nesting biology and immature stages of Cockerell, including a comparison with the bees Centris caesalpiniae, C. pallida other anthophorids (Hymenoptera: and the cleptoparasite Ericrocis lata Apoidea). Ann. Entomol. Soc. Am. 50: (Hymenoptera: Apoidea: Anthophori- 469-475. dae). Am. Mus. Novitates 2935: 30 pp. 1966. The larvae of the Anthophoridae (Hy- Rozen, J. G., and A. Roig-Alsina menoptera, Apoidea). Am. Mus. Nov- 1991. Biology, larvae, and oocytes of the par- itates 2244: 38 pp. asitic bee tribe Caenoprosopidini (Hy- 1967. The immature instars of the cleptopar- menoptera: Anthophoridae: Nomadi- asite genus Dioxys (Hymenoptera: nae). Am. Mus. Novitates 3004: 10 pp. Megachilidae). J. New York Entomol. Rozen, J. G., Jr., K. R. Eickwort, and G. C. Eick- Soc. 75: 236-248. wort 1969a. The biology and description of a new 1978. The bionomics and immature stages of 36 AMERICAN MUSEUM NOVITATES NO. 3029

the cleptoparasitic bee genus Protepeo- development of Epeolus compactus lus (Anthophoridae, Nomadinae). Am. Cresson, a cleptoparasite of Colletes Mus. Novitates 2640: 14 pp. kincaidii Cockerell (Hymenoptera: An- Rozen, J. G., and R. R. Snelling thophoridae, Colletidae). Ann. Ento- 1986. Ethology ofthe bee Exomalopsis nitens mol. Soc. Am. 81: 626-636. and its cleptoparasite (Hymenoptera: Torchio, P. F., and G. E. Trostle Anthophoridae). J. New York Entomol. 1986. Biological notes on Anthophora urbana Soc. 94: 480-488. urbana and its parasite, Xeromelecta Shanks, S. S. californica (Hymenoptera: Anthophor- 1978. A revision ofthe cleptoparasitic bee ge- idae), including descriptions oflate em- nus Neolarra (Hymenoptera: Antho- bryogenesis and hatching. Ann. Ento- phoridae). Wasmann J. Biol. 32: 212- mol. Soc. Am. 79: 434-447. 246. Torchio, P. F., and N. N. Youssef Snelling, R. R., and R. W. Brooks 1968. The biology ofAnthophora (Micantho- 1985. A review ofthe genera ofcleptoparasitic phora) flexipes and its cleptoparasite, bees of the tribe Ericrocini (Hymenop- Zacosmia maculata, including a de- tera: Anthophoridae). Contrib. Sci., Los scription of the immature stages of the Angeles County Mus. Nat. Hist. 369: 34 - parasite (Hymenoptera: Apoidea, An- pp- thophoridae). J. Kansas Entomol. Soc. Thorp, R. W. 41: 289-302. 1969. Ecology and behavior of Melecta se- Vinson, S. B., G. W. Frankie, and R. E. Coville parata callura (Hymenoptera: Antho- 1987. Nesting habits of Centris flavofasciata phoridae). Am. Midl. Nat. 82: 338-345. Friese (Hymenoptera: Apoidea: Antho- Torchio, P. F., and D. J. Burdick phoridae) in Costa Rica. J. Kansas En- 1988. Comparative notes on the biology and tomol. Soc. 60: 249-263.

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

THIS PUBLICATION IS PRINTED ON ACID-FREE PAPER. |