A revised genus-level classification for the Neotropical groups of the cleptoparasitic bee tribe Sphecodini Schenck (Hymenoptera, Apidae, Halictinae)

Sphecodini are a clade of obligatory parasitic Halictinae, comprising 332 species and five genera worldwide. Recently, Sphecodes Latreille was considered polyphyletic according to the results of a molecular phylogeny. Morphological analysis of Neotropical groups revealed groups of species in Sphecodes s.l. that can be recognized at genus level to improve the classification of the tribe. The main objective of this study is to review the generic classification of Sphecodini, with emphasis on Neotropical groups. I propose Austrosphecodes Michener as a separate genus from Sphecodes s.s. and describe a new genus, Melissocleptis gen. nov. for eight known species. Both genera comprise most of the Neotropical species. Nesosphecodes depressus sp. nov. is described for Brazil and a revised diagnosis for Nesosphecodes Engel is given to accommodate this new species and to include data on male terminalia. A revised key for the genera from the Western Hemisphere is provided. A R T I C L E I N F O Article history: Received 11 September 2020 Accepted 28 December 2020 Available online 22 February 2021 Associate Editor: Marcel Hermes


Introduction
Sphecodini Schenck are a clade of obligatory parasitic halictine bees related to the tribe Halictini (Danforth et al., 2008). Most studied species behave as cleptoparasites but a social parasite of Dialictus Robertson (Halictini) was studied by . The host range of the tribe includes several bee subfamilies (Michener, 1978). The tribal classification backbone was encompassed by Michener (1978) in his monograph on parasitic halictids and later in his book "Bees of the World" (Michener, 2000(Michener, , 2007. Currently, the tribe comprises 332 species (compiled from Ascher and Pickering, 2020) distributed in five genera: Eupetersia Blüthgen, Microsphecodes Eickwort & Stage, Nesosphecodes Engel, Ptilocleptis Michener and Sphecodes Latreille. According to the only phylogeny for the group (Habermannová et al., 2013), Sphecodes s.l. is polyphyletic in relation to Eupetersia and Microsphecodes.
Eupetersia comprise 34 species from Africa and Oriental region distributed in two subgenera (Michener, 2007;Ascher and Pickering, 2020). This genus resembles Microsphecodes for the weak body punctation, but it is related to the Holarctic species of Sphecodes according to Habermannová et al. (2013). The most reliable diagnostic character for the genus is the relative elongation of F1, being subequal to F2 (Pauly, 2012).
The Neotropical Microsphecodes was originally described by Eickwort & Stage (1972) as a subgenus of Sphecodes. Michener (1978) elevated the group to genus status. Recently, Engel (2013) divided Microsphecodes in two subgenera, Microsphecodes s.s. and Baeosphecodes Engel. Microsphecodes s.s. (five species) occurs from Panamá to Brazil and has striate-foveolate metapostnotum and mesosoma with prominent pubescence, while Baeosphecodes (five species) occurs in the West Indies and has rugoso-striate metapostnotum and mesosoma with sparse pubescence (Engel, 2013). This genus can be separated from the remaining Sphecodini by the long free portion of the marginal cell and the shape of the first tergum (longer than broad). According to Michener's (1978) compilation its hosts are Dialictus and Habralictus. According to the results of Habermannová et al. (2013), Microsphecodes is nested within Austrosphecodes, a subgenus created by Michener (1978) in Sphecodes for a group of Neotropical species.
Nesosphecodes comprises three species from the Antilles (Engel, 2006a). The genus is quite similar to Microsphecodes and was separated by the longer mandibles, broader clypeus, and shorter length of free portion of marginal cell (Engel, 2006a(Engel, , 2013. Nothing is known about its phylogenetic affinities and host species. Similarly, Ptilocleptis also comprises three species from the Neotropical Region (Michener, 1978). This is the most distinctive genus of the tribe due to moderate body punctation and pubescence, head and clypeus shape, transversally carinate pronotum (Michener, 1978(Michener, , 2007. As for the previous genus, nothing is known about its phylogenetic affinities and host species. Sphecodes is the largest genus of the tribe with 282 species and has a cosmopolitan distribution (Ascher and Pickering, 2020). This genus is quite variable and hard to diagnose when compared to the other genera, with its monophyly being frequently questioned (Engel, 2006b;Michener, 2007). Also, a total of 10 genus-group names are available for the group according to the synonymic list updated by Michener (2007). Michener (1978) recognized two subgenera, Austrosphecodes Michener and Sphecodes s.s. Austrosphecodes was originally diagnosed from Sphecodes s.s. by the lack of subapical tooth in the mandible, lack of conspicuous spinelike setae on hind tibiae, and lack of basitibial plate (Michener, 1978). Moure & Hurd (1987) followed this classification and cataloged 32 species of Austrosphecodes. But later, Michener (2000Michener ( , 2007 opted not to consider the subgenera, a position followed by Moure (2007). Hosts of Sphecodes are distributed in several bee subfamilies and tribes according to Michener (1978). Following the phylogeny provided by Habermannová et al. (2013), Austrosphecodes is more related to Microsphecodes and not with the remaining Sphecodes.
The main objective of this study is to provide a revised classification of Sphecodini, with emphasis on Neotropical groups. Eupetersia and non-Neotropical groups of Sphecodes s.s. are beyond the scope of the present revision. This classification is based on morphological characters from females and males, taking into consideration the polyphyly of Sphecodes revealed by molecular data (Habermannová et al., 2013). I recognize Austrosphecodes as a separate genus from Sphecodes and describe a new genus, Melissocleptis gen. nov. A new species of Nesosphecodes is described from Brazil and a revised diagnosis for this genus is given to accommodate this new species and to include data on male terminalia. A revised key for the genera from the Western Hemisphere is provided.

Material and methods
Material mentioned here is deposited in the following institutions: Bee Biology and Systematics Laboratory, Utah State University, Logan, Utah (BBSL), British Natural History Museum, London, England, U.K. The species description was made following Engel (2006a) in order to facilitate the comparison. Structure terminology also follows Engel (2006a) except that here I refer to the basal area of propodeum as the metapostnotum. I use the abbreviations T1, T2, etc., to denote the metasomal terga; S1, S2, etc., to denote metasomal sterna; and F1, F2, etc., to denote the flagellomeres. Generic diagnosis terminology follows Eickwort (1969) and Michener (2007). All measurements are in millimeters. Interspaces among punctures (i) are given in relation to the puncture diameter (pd). Sculpturing terminology follows Harris (1979). For some sculpturing I use two terms to describe transitional patterns, as areolate rugose. For microsculpture, visible only in magnifications above 50x, I use the prefix micro-.
Most examinations were made with an Olympus SZ51 using white LED ring illumination. Measurements were made with Leica Stemi DV4 and a micrometric rule. Photographs were taken with a Nikon D700 with a 105 mm Sigma Macro lens using Helicon Remote for controlling image capture and with a Leica DFC 500 cam coupled with a Leica MZ16 stereomicroscope using the Leica Las 3D Viewer. Illumination followed the system of Kawada and Buffington (2016). Image stacking was made with Helicon Focus (version 6.0.18)-render method based on Method C (Pyramid). Final image adjustments (unsharp mask filtering, level control) were made with GIMP 2.8.16 (©The GIMP Team).

Results
Key to the Genera of the Sphecodini of the Western Hemisphere (modified from Michener, 2007) 1. Inner orbits of eyes strongly converging below; head little wider than long; clypeus about twice as broad as long; carina across pronotum, between dorsolateral angles, continuous; surface of S2 conspicuously convex in profile, its base strongly depressed, suggesting strong constriction between S1 and S2; body tomentum yellowish, usually dense and covering terga . Diagnosis. Females have coarse punctation overall (differently from Eupetersia, Microsphecodes and Nesosphecodes) and sparse body tomentum (differently from Ptilocleptis). The yellow markings are missing (differently from Microsphecodes) and the metasoma frequently has reddish color. Their clypeus is flattened with a longitudinal medial sulcus (differently from Ptilocleptis). Scape is long, reaching mid ocellus (differently from Melissocleptis gen. nov.), and the eyes are subparallel (differently from Ptilocleptis). Mandibles are simple (differently from most Sphecodes) and hypostomal ridge is not raised (differently from Ptilocleptis). F2 is not elongate in relation to F1 (differently from Eupetersia). The preoccipital ridge is carinate (differently from Nesosphecodes). The pronotal transverse carina is incomplete and the lateral ridge present (differently from Ptilocleptis). Mesoscutum anterior surface is abruptly delimited (differently from Nesosphecodes and some Microsphecodes). The free part of marginal cell is short (differently from Microsphecodes and Nesosphecodes). Spinelike setae and hind basitibial plate are not developed (differently from most Sphecodes). The T1 width is broader than long, T5 is fimbriate (differently from Microsphecodes) and S2 is not convex (differently from Ptilocleptis). Males have long flagellum, usually as long as mesosoma; their T7 pygidial plate is narrow (differently from Melissocleptis gen. nov., Ptilocleptis and Nesosphecodes). The spicule of S7 is absent (differently from Ptilocleptis and Nesosphecodes) and the gonocoxites lack striae (differently from Sphecodes, Melissocleptis gen. nov. and Nesosphecodes). The gonapophyses are narrow and the ventral prong is absent (differently from Melissocleptis gen. nov.). The gonostyli have distinct lobes (differently from Melissocleptis gen. nov., Ptilocleptis and Nesosphecodes) and the internal setose patch is absent (differently from most Sphecodes, Melissocleptis gen. nov., Eupetersia and Nesosphecodes).
Comments. For additional comments to separate Austrosphecodes from Sphecodes see Michener (1978). The body size is variable among species, some of them quite small, with about 6 mm and resembling the species of Melissocleptis gen. nov. I did not examine the type species, Sphecodes chilensis, but follow the interpretation of Michener (1978) and unpublished notes of the type made by Moure that this species has the characteristics to diagnose the group. Included species. The 29 valid names transferred to Austrosphecodes follows the subgeneric assignment of Moure & Hurd (1987) (Schrottky, 1906). The later three species have less than 6 mm and the possibility that they belong in Melissocleptis gen. nov. should be further investigated. Additional 11 Neotropical species (Moure, 2007) placed in Sphecodes s.s. according to Moure & Hurd (1987) should also be checked for correct generic placement.
Distribution. Chile to Mexico.
Melissocleptis gen. nov. urn:lsid:zoobank.org:act:0E78B0A3-02FE-4084-BE98-B9494EBCD5F0 Diagnosis. Females have moderate punctation overall (differently from most genera) and body tomentum sparse (differently from Ptilocleptis). Bees are mostly black and the yellow markings are missing (differently from Microsphecodes) and the metasoma frequently has reddish color. Their clypeus is flattened with a longitudinal medial sulcus (differently from Ptilocleptis). Scape is short, not reaching the mid ocellus (differently from all genera) and eyes subparallel (differently from Ptilocleptis). Mandibles are simple (differently from most Sphecodes) and hypostomal lamella is not raised (differently from Ptilocleptis). F2 is not elongate in relation to F1 (differently from Eupetersia and most Nesosphecodes). The preoccipital ridge is weakly carinate (differently from Nesosphecodes). The pronotal transverse carina is incomplete and the lateral ridge present (differently from Ptilocleptis). Mesoscutum anterior surface is abruptly delimited (differently from Nesosphecodes and some Microsphecodes). The free part of marginal cell is short (differently from Microsphecodes and Nesosphecodes). Spinelike setae on posterior margin of hind tibia are as long as nearby plumose setae (differently from most Sphecodes that have thick and short spinelike setae). Hind basitibial plate is not defined along posterior margin (differently from most Sphecodes). The T1 is broader than long, T5 is fimbriate (differently from Microsphecodes) and S2 is not convex (differently from Ptilocleptis). Males have short flagellum, shorter than mesosoma length, the T7 pygidial plate broad (similarly to Ptilocleptis). The spicule of S7 is absent (differently from Ptilocleptis and Nesosphecodes) and the gonocoxite is striate (similarly to some Sphecodes and Nesosphecodes). Gonapophyses are wide and the ventral prong is present (similarly to a few Sphecodes species). The gonostyli have reduced lobes (similarly to Ptilocleptis and Nesosphecodes) and the internal setose patch is present (similarly to most Sphecodes, Eupetersia and Nesosphecodes).
Etymology. Derived from the Greek, melissa, bee, and kleptis, thief, meaning robber bee. The name is feminine.
Comments. Species in this genus have small body size, rarely reaching 6 mm, with metapostnotum usually carinate, being similar to the smallest Austrosphecodes as A. brasiliensis. I include eight described species in this genus but probably more described and undescribed species are waiting to be added. It is interesting to note that the latest species described as Sphecodes from the West Indies belong to Melissocleptis gen. nov. (Engel, 2006b;Gibbs, 2016).
Distribution. Known records of the genus are from Argentina, Brazil, Chile (undetermined species), Costa Rica (identified as Sphecodes costaricensis Friese by Michener), Cuba, Dominica and Saint Vincent (Antilles), but probably the genus has a broader distribution in the Neotropical Region.
Melissocleptis albifacies (Gibbs, 2016) Comments. The females of this species lack a carina separating the posterior surface of the propodeum from its latero-dorsal surfaces and their frons is depressed above the supraclypeal area (also in males). Those characters are shared with other undetermined species from Argentina and Chile.
Comments. This species can be separated from other Melissocleptis gen. nov. from South America by the denser punctation on the mesoscutum, especially along its anterior margin where it is crowded. The species was redescribed by Gonçalves (2017).
Comments. According to the original description and associated images this species should be placed in Melissocleptis gen. nov. Genital capsule as illustrated by Gibbs (2016) is typical of the new genus.
Comments. Based on the images available in BMNH database, the type specimen has size, color and metapostnotum patterns of Melissocleptis gen. nov. This record is important to indicate that most Neotropical lineages of Sphecodini are distributed into the Antilles.
Comments. According to the original description and associated images this species should be placed in Melissocleptis gen. nov. Genital capsule is also typical of the new genus.
Comments. According to the original description and associated images this species should be placed in Melissocleptis gen. nov. Genital capsule is also typical of the new genus.
Comments. The females of this species have a sinuous carina contouring the posterior surface propodeum, giving it a somewhat stylized heart-shape, and their frons is not depressed above the supraclypeal area (also in males). Microsphecodes coriae also shares this condition, but its mesoscutum is more densely punctured. An undescribed species from northern Brazil (Belém, Pará) also has these features.
Diagnosis. Females have a fine punctation overall (differently from Austrosphecodes, Melissocleptis gen. nov. Sphecodes, and Ptilocleptis) and body tomentum sparse (differently from Ptilocleptis). The yellow markings are missing (differently from Microsphecodes) and the metasoma is rarely reddish brown. Their clypeus is flattened with a longitudinal medial sulcus (differently from Ptilocleptis). Scape is long, reaching mid ocellus (differently from Melissocleptis gen. nov.) and eyes subparallel (differently from Ptilocleptis). Mandibles are simple (differently from most Sphecodes) and the hypostomal lamella is not raised (differently from Ptilocleptis). F2 can be elongate in relation to F1, similarly to Eupetersia. The preoccipital ridge is rounded (differently from Austrosphecodes, Eupetersia, Melissocleptis gen. nov., some Microsphecodes and Ptilocleptis). The pronotal transverse carina is incomplete and the lateral ridge is absent (differently from most other genera). Mesoscutum anterior surface is gently convex (differently from most genera, similar to some Microsphecodes). The free part of marginal cell is about two times the portion in contact with the submarginal cells (similar to Eupetersia). Tibial spines and hind basitibial plate are absent (differently from most Sphecodes). The T1 is broader than long, T5 is fimbriate (differently from Microsphecodes) and S2 is not convex (differently from Ptilocleptis). Males have the T7 pygidial plate very weakly developed (differently from all genera). The spicule of S7 is developed (similarly to Ptilocleptis) and the gonocoxite is striate (similarly to some Sphecodes, and Melissocleptis gen. nov.). Gonapophyses are narrow and the ventral prong is present (similarly to Melissocleptis gen. nov.). The gonostyli have distinct lobes (similarly to Ptilocleptis) and the internal setose patch is present (similar to most Sphecodes, Eupetersia and Melissocleptis gen. nov.).

Comments.
A revised diagnosis is necessary to accommodate the new species from Brazil, especially because it does not have an elongate F2 and its clypeus and mandibles are not as wide as those from the previously known species. These characteristics should be considered as distinctive of the Caribbean species. Engel (2006a) did not give information about terminalia and the male diagnosis is based only on the new species. The new species add a substantion range extension to Nesosphecodes, but this can purely be seen as absence of records of the group, since other Neotropical Sphecodini, Austrosphecodes, Microsphecodes, and Melissocleptis gen. nov. have broad distributions.
Specimens, including males, have two or rarely three submarginal cells, while other species have three submarginal cells.
Description. Female. Total body length 5 mm; forewing length 3 mm. Head about as long as wide (width 1.5 mm, length 1.2 mm as measured from clypeal apex to vertex in frontal view). Clypeus apex broadly rounded with oblique lateral margins. Frontal line carinate from just below antennal toruli to half of distance to median ocellus, becoming an impressed line from that point onward. Mandibular base meeting lower border of compound eye. Upper and lower interorbital distances nearly equal but compound eyes slightly converging ventrally; inner margin of compound eye relatively straight. Parocular area depressed near torulus. Gena narrower than compound eye in profile. Scape length 0.6 mm; F1 as long as F2; distal flagellomere with inner apical half glabrous. Intertegular distance 1.0 mm. Basal vein distad cu-a by two times vein width; 1rs-m distad 1m-cu by vein width; 2rs-m distad 2m-cu by ten times vein width; first and second submarginal cells fused due to the absence of 2 nd abscissa of Rs (weakly indicated in one paratype); anterior border of third submarginal cell nearly one-half length of posterior border; hind wing with distal hamuli arranged 3-2. Pygidial plate large and broadly rounded at its apex, margins (lateral and apical) lamellate. Clypeus weakly imbricate with weak punctures separated by a puncture width; remainder of head with minute punctures, i= 1-2 pd, integument between punctures microreticulate and shining, punctures weaker on vertex; postgena microreticulate and impunctate. Pronotum with minute punctures, i=1-3 pd, punctures of anteriolateral borders closer to each other, integument between punctures microreticulate. Mesoscutum with minute punctures, i=1-4 pd (Fig. 3), those of lateral and posterior borders with i =1-2 pd, integument between punctures smooth and shining; tegula translucid, smooth except inner border imbricate and minutely punctured; mesoscutellum nodulose, sculptured as on mesoscutum except that i=1-4 pd over entire surface. Metanotum areolate, integument in between smooth and shining. Anterior surface of mesepisternum rugulose; mesepisternum rugulose anteriorly and almost polished posteriorly; metepisternum rugulose, upper fourth carinate. Metapostnotum with lateral straight and two depressed central areolae, integument between striae smooth and shining; lateral and posterior surfaces of propodeum areolate except for broad, smooth basal triangle just above propodeal pit on posterior surface. Metasomal terga and sterna very sparsely and faintly puncticulate.
Mandible dark reddish brown; labrum and mouthparts dark brown; remainder of head and mesosoma black and shining. Metasoma reddish brown, lighter on first two segments. Legs brown. Wing veins brown; wing membrane hyaline except on costal cell, yellowish.
Pubescence white except when indicated. Face with long, short branched setae on lower half with very few subappressed, highlybranched setae, shorter setae noticeably more dense on lower half of face and clypeus but not obscuring integument; upper half of face and vertex with less numerous, short, simple setae; some setae on vertex slightly fuscous. Setae on gena like those of vertex. Postgena almost glabrous except for few scattered long setae. Pronotum covered in long plumose setae, not obscuring integumental surface, such setae sparse anterior to pronotal lobe. Mesoscutum and mesoscutellum with scattered, short, simple setae, those on mesoscutellum more confined to posterior border where longer with more branches; tegula with minute setae on apical and inner borders. Metanotum with similar setae as to those on mesoscutellum except denser. Anterior surface of mesepisternum covered with long, plumose setae, largely obscuring integumental surface; mesepisternum with long, plumose setae becoming longer, more branched and more numerous ventrally; metepisternum with short, plumose setae dense on lower third and anterior border, sparse on posterior median border. Lateral and posterior surfaces of propodeum densely covered in long, plumose setae that do not obscure integumental surface. Metasomal T1 with simple, yellowish short setae on lateral borders; T2 with yellowish, minute, simple setae scattered on basal two-thirds, such setae longer laterally and denser and longer on marginal areas; T3-5 with similar setae as to those on T2 except longer and denser; T6 with setae with more branches and denser; sterna with scattered simple setae, also yellowish, short to long in length, longer setae generally scattered in apical halves; S6 densely clothed with long branched setae.
Male. As described for the female except total body length 4.8 mm; forewing length 3 mm. Head about as long as wide (width 1.33 mm, length 1.15 mm as measured from clypeal apex to vertex in frontal aspect). Scape length 0.3 mm. Intertegular distance 0.83 mm. On forewing 2 nd abscissa of Rs present or absent Punctures of head denser than on female. Body pubescence shorter, unbranched on terga. Hidden terga and genital capsule as Etymology. The specific epithet refers to the depression on parocular area, from Latin.

Discussion
The monophyly of Sphecodes s.l. was not recovered by the molecular phylogenetic analysis of Habermannová et al. (2013), who found the following topological relationships: ((Austrosphecodes in part+ (Austrosphecodes in part+ Microsphecodes))(Sphecodes s.s. + Eupetersia)). This finding indicates the necessity to propose a new classification for the group. A first decision is to consider the Neotropical group Austrosphecodes at generic status to make Sphecodes s.s. monophyletic. Both groups were recognized by former authors (Michener, 1978, Moure andHurd, 1987) but later Michener (2000Michener ( , 2007 opted to synonymize all genus-group names under Sphecodes, a position followed by subsequent authors. Habermannová et al. (2013) did not include all lineages of Sphecodes and its monophyly should be examined further, especially to investigate the possibility to subdivide the genus: there are 09 genus-group names available for it excluding Austrosphecodes (Michener, 2007); only three names were included on the available phylogeny. I did not have the opportunity to study material from these groups and cannot make statements about them. The second decision is to recognize two separate genera for the 32 species allocated under Austrosphecodes by the catalog of Moure and Hurd (1987). Habermannová et al. (2013) found that Austrosphecodes is paraphyletic with respect to Microsphecodes and I have presented here morphological evidence for recognizing two distinct groups in Austrosphecodes sensu Michener (1978). However, it is important to highlight that I did not examine the vouchers from Habermannová et al. (2013) study (after repeated image requests), so it is not possible to infer properly what lineages correspond to the genera circumscribed here.
Following this decision, the first group, representing Austrosphecodes proper, includes most Neotropical species and is characterized by the female scape reaching mid ocellus (Fig. 1D) and by features of male genital capsule (gonobase outer surface not striate, inner surface without internal setose patch, gonapophysis without ventral prong, gonostylus strongly divided into lobes, Fig. 2A). The second group, Melissocleptis gen. nov., is characterized by the female scape not reaching mid ocellus (Fig. 1A) and features of male genital capsule (gonobase outer surface striate, inner surface with internal setose patch, gonapophysis with ventral prong, gonostylus weakly divided into lobes, Fig. 2C). Despite other features mentioned in the diagnoses, species of Austrosphecodes vary in size and are similar to Sphecodes in body color and sculpturing while the species of Melissocleptis gen. nov. are small, having no more than 6 mm in body length, and being more similar to Nesosphecodes.
The features of the male genital capsule seem to provide important characters for phylogenetic analysis of the group. Some features were already mentioned by Michener (1978) but not properly explored by subsequent authors. The genital capsule of Austrosphecodes ( Fig. 2A) is very similar to that of Microsphecodes (Fig. 2B), suggesting a close relationship. On the other hand, the genital capsule of Melissocleptis   gen. nov. (Fig. 2C) has some attributes shared with Nesosphecodes ( Fig. 5C,D), such as the ventral prong of the gonapophyses. According to these findings it is plausible that these four Neotropical genera form a clade. The most intriguing open question is about the phylogenetic relationships of Ptilocleptis. This genus has a long list of particular features for both sexes (Michener, 1978(Michener, , 2007 and could be sister to all remaining Sphecodini (as suggested by Michener, 2007) or more closely related to the remaining Neotropical genera. Table 1 summarizes the proposed classification of the tribe, currently composed by seven genera and 333 species, which should be tested under a phylogenetic perspective to validate the herein decisions.