Amiseginae and Cleptinae from northeastern Brazil, with the description of four new species (Hymenoptera, Chrysididae)

The diversity of two subfamilies of cuckoo wasps in northeastern Brazil is reviewed. Four new species are described and illustrated: Amisega boyi Lucena, sp. nov., A. sertaneja Lucena, sp. nov., and Duckeia dudui Lucena, sp. nov. (Amiseginae), and Cleptidea nordestina Lucena, sp. nov. (Cleptinae). These new species of Amisega and Duckeia represent the first records of both genera in the core zone of the Caatinga dry region, and they seem to be endemic to this portion of the region. The other two Amisega species previously recorded in northeastern Brazil are restricted to ecotone habitats between Caatinga and Atlantic forest. Cleptidea nordestina Lucena, sp. nov. (Cleptinae) is the most septentrional record for the fasciata species group in South America. With the present contribution, the total recorded diversity of Amiseginae and Cleptinae in northeastern Brazil is represented by the following species: A. boyi Lucena, sp. nov., A. flavipes Kimsey, 1987, A. sertaneja Lucena, sp. nov., A. similis Kimsey, 1987, and D. dudui Lucena, sp. nov. (Amiseginae), and C. nordestina Lucena, sp. nov. (Cleptinae).

as quite heterogeneous region, resulting from the interplay between presenting high spatial and historical complexity ( Fig. 1A-H) (Cardoso da Silva et al. 2017;Queiroz et al. 2017). It has been suggested that its flora displays higher diversity per area than the Amazon rainforest .
Caatinga is mostly characterized by high temperatures throughout the year, with a marked seasonal climate marked by a strong hydric deficit (Fig. 1A, B). The amount of rainfall varies between 240-1500 mm per year (generally below 800 mm per year in the semiarid region), and it is irregularly distributed, usually concentrated in three consecutive months during austral summer (Prado 2003). The adverse weather conditions imposed by the hydric deficit imply that biota in the Caatinga has significant tolerance to dehydration. At least part of the fauna may be unique, potentially with a high number of undiscovered endemic taxa remaining to be found.
Herein, we expand the knowledge of Chrysididae fauna in northeastern Brazil by describing and illustrating four new species in the Amiseginae and Cleptinae. We comment on the new species' habitats and discuss some geographic records of previously known taxa from the northeastern region. Additionally, we discuss and update the list of taxa with occurrence in the core region of Caatinga.

Geographic delimitation
For the analyses of geographic occurrence of species into the semiarid limits, we used delimitation presented by Pereira-Júnior (2007), established considering the isohyet of 800 mm or lower per year, the Thornwaite aridity index equal to 0.50, and hydric deficit with drought risk of 60%. The shapefiles of Morrone's (2014) biogeographic dominions (Löwenberg-Neto 2014) were used as reference for the limits of other major natural areas in South America. The main biogeographic areas shown in the Figure 1 correspond to combinations of units of the complex biogeographical regionalization proposed for the Neotropical region by Morrone (2014): the Amazon rainforest herein corresponds to Boreal Brazilian Dominion + South Brazilian Dominion + Southeastern Amazonian subregion; the Atlantic rainforest corresponds to Parana Dominion; the Cerrado and Caatinga correspond to provinces in the Chacoan Dominion. We used QGIS Development Team (2020) for combining shapefiles and generating new maps for species distribution. Distribution maps were based on locality records taken from specimen labels and complemented by information in the literature.

Material studied
Most specimens were collected using malaise traps standing in the field for at least two consecutive months (Patu and Mossoró-RN), a whole year (Jequié-BA), and four consecutive years (Santa Teresinha-PB) (detailed descriptions of localities and sampling methods were provided in Lucena et al. 2012;Zanella and Lucena 2014;Fernandes et al. 2014Fernandes et al. , 2019Guedes et al. 2019). We also examined specimens housed in the following institutions: BME:  Brazil (Fernando A. Silveira; Alessandro Lima); and USNM: The National Museum of Natural History, Smithsonian Institution, Washington, USA (Seán G. Brady). Part of the type specimens of taxa described by Kimsey (1987) were originally housed at The American Entomological Institute (AEI), Gainesville, Florida, but the AEI collection was recently relocated to Logan, Utah, and became part of the Utah State University, Logan, USA (USU) (James Pitts). Finally, some taxonomic decisions were possible by comparisons of specimens with high-resolution images of the holotypes housed in the Canadian National Collection, Ottawa, Canada (CNC) (Sophie Cardinal).

Photographs
Specimens were photographed at Laboratório de Biologia Comparada e Abelhas, Departamento de Biologia, FFCLRP/USP, using a Leica DFC425 camera attached to a Leica M205C stereomicroscope, equipped with a Leica FlexiDome lighting diffuser, and light system model Leica LED5000 HDI. Images were improved using Leica application suite software-LAS to combine multiple photos, then enhanced with Adobe Photoshop.

Terminology
The general morphological terminology follows primarily Huber and Sharkey (1993), with few modifications intended to incorporate the terms specifically used for Chrysididae, such as: malar space, malar sulcus, propodeal angle, and scapal basin (Kimsey and Bohart 1991: 41-42). The abbreviations F, S, and T are employed for flagellomere, metasomal sternum, and metasomal tergum, respectively. MOD refers to the median ocellus diameter. LID refers to the least interocular distance. PD refers to puncture diameter. OL refers to the minimum distance between the inner margin of the lateral ocellus and median ocellus. OOL refers to the minimum distance from the outer margin of lateral ocellus to compound eye. POL refers to the minimum distance between inner margins of lateral ocelli. Malar space and subantennal distance refer to the minimum distance between the inferior ocular margin and mandible, and between the inferior margin of torulus and dorsal margin of clypeus. Sculpturing terminology follows Harris (1979).

Subfamily Amiseginae
The Amiseginae are specialized parasitoids of walking stick eggs (e.g., Costa Lima 1936;Krombein 1957Krombein , 1960Krombein , 1983Kimsey and Bohart 1991;Baker 2016). Due to their relatively small size and cryptic habits, these wasps are uncommon in collections. They are most frequently collected in forested habitats using Malaise traps and yellow pan-traps. Generic revisions with notes on biology, geographic ranges, list of valid species, and illustrated keys were provided by Krombein (1957Krombein ( , 1960Krombein ( , 1983 and Kimsey and Bohart (1991). There are nineteen species and four genera recorded in Brazil (Lucena and Zanella 2016). Still, we can assume that this number is underestimated for the country, especially considering there are vast geographic areas without any significant sampling. In northeastern Brazil, we found two genera and five species. Three of these species are herein described as new and represent the first records of species of this subfamily for the Caatinga region.
This genus is one of the largest in the Amiseginae. Currently, Amisega includes 24 valid species (Kimsey 1987(Kimsey , 1990(Kimsey , 1993, ranging from southeast Canada, midwestern and southeastern United States, as far west as Arizona, through Mexico, Central America, Chile, and in most South America (e.g., Kimsey and Bohart 1991). Although Amisega is one of the most abundant amisegines in collections, its biology is poorly known. Milliron (1950) reported Amisega kahlii (Ashmead, 1902) (=Mesitiopterous) in North America rearing from eggs of Diapheromera femorata (Say, 1824) (Diapheromeridae), and Kimsey (1990) described the flightless Amisega chilensis Kimsey, 1990 reared from eggs of Heteronemia mexicana Gray, 1835 (=Bacunculus phyllopus) (Heteronemiidae). Kimsey (1987) revised the genus, described twelve new species, and presented the most complete discussion for species distinctions to date, later complemented by Kimsey (1990Kimsey ( , 1993, and Kimsey and Bohart (1991). In Brazil, eight valid species were previously known to occur from the Atlantic rainforest along east coast, as far north as Pernambuco state, through the Cerrado in the central-west and Amazon rainforest on north. The genus is represented in northeastern Brazil by two previously known species, and two new ones are herein described.
Female  Lucena, sp. nov., see below), within the semiarid region. All specimens were collected in a Caatinga native vegetation fragment adjacent to a melon crop Cucumis melo L. (Cucurbitaceae) (Fernandes et al. 2014).
Etymology. The new species is named after Daniell Fernandes (Boy), collector of many specimens used in this study.

Amisega flavipes
Comments. This species can be readily distinguished from other species of the northeastern Brazil by the following combination of characters: scapal basin crossridged; dark brown antenna, clypeus, mandible, and metaleg; and lack of distinct striation on vertex and dorsum of mesosoma ( Fig. 3C-E). In the original description, Kimsey (1987: 67) cited F1 length 4× breadth, with the male interpreted as identical. The analyzed specimen from Jequié-BA has comparatively shorter F1 (length 2.3× breadth), and a distinct medial pit on anterior declivity of pronotum (not cited in the original description). All other diagnostic characters described for A. flavipes are identical with this specimen. This species has been recorded in a transition area between Caatinga and Atlantic rainforest. The Encruzilhada county is in the "Sul-Baiano" highlands, on altitudes above 600 m, with milder climate than usual in the Caatinga (Radambrasil 1983). The new record herein presented for Jequié is only 207 km north of the previous record, and in spite of having a lower altitude and warmer and drier climate than the type locality, still in the transition zone between the Atlantic rainforest and Caatinga. The taxonomic conclusions about this species are based on the original description and by examination of a male paratype from Bahia: Encruzilhada (BME), and a male recently collected in Bahia: Jequié (RPSP).
Vestiture: short sparse pale setation on vertex, gena, and frons, with relatively longer and denser setation on lateral border of face and occiput; mandible, clypeus and labrum with distinct long gold setation; antenna with very short, decumbent, pale setation; eye with sparse microtrichia; dorsum of mesosoma with short sparse light brown setation; lateral pronotum, posterior border of mesopleuron, dorsum of propodeum, and metapleuron-propodeum, glabrous; wing membrane entirely setose; outer surface of metatibia, and pro-and metafemora, with long dark setae; venter of tarsomeres with abundant irregularly-sized spines; T1-T2 mostly glabrous, except for sparse short setae placed marginally; T3-T5 with marginal stripe of long dense dark setae; S2-S5 abundantly setose. Female (Fig. 5). Same as male, except: scape 4.4× longer than its maximum width; F1 length 3.3× breadth, 1.7× longer than F2; LID 0.85× as long as scape; transverse groove on dorsal surface of propodeum barely marked; metasoma dark brown becoming lighter ventrally; T2 with broad impunctate marginal area (Fig. 5C); and denser punctation among striae on dorsum of mesosoma (Fig. 5D, E)  Comments. The holotype is missing F10-F11. This species is only known from localities in the core zone of the Caatinga dry region. It is sympatric with A. boyi Lucena, sp. nov. in Mossoró-RN, which suggests both species are endemic to Caatinga. All specimens of both new species were collected between January to May, which is the period of most intense rainfall in the region.
Variation. Body size: 2.8-3.0 mm; F1 length 2.6-3.3× breadth. Etymology. The name is a Brazilian gentilic adjective for person living in the semiarid region.

Amisega similis Kimsey, 1987
Amisega Comments. This species can be distinguished from other species of the northeastern Brazil by its bicolored legs, blue highlights on dorsum of metasoma, and lack of striation on vertex and dorsum of mesosoma. Kimsey (1987) cited AEI as the repository of the holotype, which is actually housed at CNC. The previous record for Pernambuco is now clarified as it came from Caruaru county (originally misspelled as "Carvary" by Kimsey 1987). Amisega similis is sympatric with A. flavipes in the "Sul-baiano" highlands in southern Bahia, and probably also in northern Minas Gerais state, although A. similis clearly also reaches areas in the Atlantic rainforest in Rio de Janeiro state. Therefore, it is possible that neither species occur in the limits of the Caatinga.

Key to species of Amisega Cameron from northeastern Brazil
1 Vertex and dorsum of mesosoma densely striate (e.g., Fig. 4D (Fig. 2C); antenna, mandible, and clypeus light brown (Fig. 2C); legs light brown; vertex roughly sculptured, with punctures inserted among discrete striae, punctures 1 PD apart or more (Fig. 2B); T2 with medial impunctate polished strip clearly defined (Fig. 2F) (Fig. 3B); scapal basin entirely cross-ridged (Fig. 3D); antenna, mandible, and clypeus dark brown (Fig. 3D); legs dark brown; vertex lacking striae, punctures touching to 2 PD apart; T2 with medial impunctate polished strip not defined (Fig. 3E)  Duckeia is a rarely collected taxon. Currently, the genus comprises three species, two of them were recorded in Rio de Janeiro state, southeastern Brazil, in an area of Atlantic rainforest, and one species is known from Costa Rica (Costa Lima 1936;Kimsey 1993;Lucena and Zanela 2016). According to Costa Lima (1936), the type series of Duckeia cyanea Costa Lima, 1936 was reared from eggs of Prisopus ohrtmanni (Lichtenstein, 1802) (Phasmatidae), in Rio de Janeiro, Brazil. Kimsey (1987) described D. vagabunda based on specimens reared from quarantine material carried by airplane from an unknown locality of Mexico that arrived in Memphis, Tennessee, 27 March, 1962(Kimsey 1987. According to Kimsey (1987), the host phasmatid egg was probably Prisopus berosus Westwood, 1859 (label record). Some years later, Kimsey (1993) recorded D. vagabunda for Costa Rica. For the first time, this genus is reported for the Caatinga in northeastern Brazil, and a new species is described.
Coloration: head dark brown green, with green highlights on frons, vertex and gena, darker basally at malar area; mouthparts light brown; antenna brown, lighter on pedicel and distal flagellomeres; mesosoma dark green brownish, with bluish purple tints dorsally on propodeum; mesopleuron with ventral and posterior light brown stripe; tegula light brown; wing membrane slightly pale to subhyaline, veins brown; legs light brown, becoming yellowish on tibiae and tarsomeres; metasoma dark brown, metallic blue highlights on dorsum, becoming greenish marginally on T3-T4; disc of T1 and T2 brownish; sterna mostly dark brown, except for marginal bluish highlights on S2.
Comments. The holotype is missing part of left metaleg. Two paratypes are poorly conserved, lacking antennae, legs and metasoma, thus, preventing gender determination. This species is only known from Patu county. Specimens were collected in the base of "Serra do Lima" (inselberg), 248 m above sea level , in the core zone of the Caatinga dry region. Interestingly, the specimens were collected in November, which is characterized as part of the dry season in the region.
Etymology. The species is named after Sebastião Antônio de Araújo, Dudu (in memoriam), grandfather of the first author.
This genus is neotropical, occurring mostly in forested warm habitats from southwestern Mexico, through Central America, into South America as far south as northern Argentina (Kimsey 1981(Kimsey , 1986Kimsey and Bohart 1991). Kimsey (1981Kimsey ( , 1986 revised the genus, described eight new species, and provided identification keys and illustrations. Three other species were later described by Móczár (1996a,b) along with an updated key for all species in the genus. Móczár (1996b) also designated a lectotype for Cleptidea fasciata (Dalman, 1823) (a dorsal habitus image of the lectotype specimen is available in Rosa and Vårdal 2015) and synonymized Cleptidea propodealis Kimsey, 1986 under C. fasciata. Currently, the genus includes nineteen valid species, divided into six species groups (sensu Kimsey 1986). Seven species and five species groups are known to occur in Brazil (Lucena and Zanela 2016). The genus was first recorded in northeastern Brazil in a preliminary local survey (Lucena et al. 2012). The morphological distinction between that specimen and members of the fasciata species group is now evident, and this new species is described.
Mesosoma: pronotum crossed antero-and submedially by transverse crenulate grooves, longitudinal sulcus culminating in a deep fovea posteriorly (Fig. 7H); posterior margin of pronotum slightly elevated, posterolateral lobe touching tegula; scutum with notaulus deeply impressed, parapsidal line discrete, faintly marked; disc of metanotum slightly convex, marginal areas depressed, posterior margin with pair of small medial pits (Fig. 7G); M with first abscissa more or less straight, diverging at 1cua; costal cell wider distally, about 2.2× the maximum width of C; dorsum of metacoxa with distinct longitudinal carina; lateral angle of propodeum short, thumb-like.
Vestiture: head with sparse, long, pale setae on vertex; frons, face, gena and scape with relatively shorter and denser setation; flagellomeres with decumbent dense pale setation; eye with tiny, sparse, sub-erect microtrichia; dorsum of mesosoma with long, sparse, pale setation, longer on venter of mesopleuron; marginal depression of metanotum with patch of pale setae posteriorly; wing membrane entirely setose, with some distinctive, long, erect dark setae inserted proximally; legs with abundant long pale setation, comparatively longer on metaleg; inner surface of tarsi with short, thick setation and short spines irregularly distributed; disc of T1 and T2 glabrous, sparse setation only marginally, T3 and T4 with long dense setae posteriorly, S1 and S4 glabrous, S2 and S3 with long sparse setation; gonapophysis with some erect setae apically.
Male Comments. The holotype is missing distal flagellomeres of the left antenna. The new species is only known from Jequié, located in a transition zone between Atlantic rainforest and Caatinga (Fig. 8C). Three unidentified male specimens from Maranhão state (06.vii.1987, Möericke, UFES; but no collector nor locality data) are also new records of the genus Cleptidea for northeastern Brazil. These male specimens are poorly preserved (lacking appendices, altered coloration, and their metasomas are collapsed); thus, we could not assign them to C. nordestina sp. nov. confidently or to interpret them as representing yet another new species.
Etymology. The name refers to the region where the new species was collected, and it is a Brazilian gentile adjective for a person native from the northeastern region.  Fig. 1, and geographical records of species in the genera Amisega and Duckeia (Chrysididae: Amiseginae), and Cleptidea (Chrysididae: Cleptinae). Political division of northeastern Brazilian states is indicated by white lines A distribution records of Amisega species (one paratype of Amisega similis Kimsey is signed with a question mark-see Discussion) B distribution records of Duckeia species C distribution records of Cleptidea species. Some localities harbor more than one species, which are indicated by dashed circles. Cleptidea fasciata and C. xantha Kimsey, 1986 are only known from the same type locality and both are represented by the same symbol in C (see Discussion).

Discussion
Since Kimsey (1987Kimsey ( , 1990Kimsey ( , 1993, further contributions to the taxonomy of Amisega were not published, and consequently, the identities of several taxa remain challenging as new samplings become accessible. Previously to this study, three out of eight valid species known to Brazil had been described by Ducke (1902Ducke ( , 1903 from Amazon rainforest in northern Brazil: A. aeneiceps Ducke, 1903, A. azurescens Ducke, 1903, and A. mocsaryi Ducke, 1902. In contrast, A. semiflava Kimsey, 1987 is known from localities in the Cerrado of central Brazil, and A. flavipes Kimsey, A. rufilateralis Kimsey, 1987, A. similis Kimsey, andA. tenebrae Kimsey, 1987, are known from localities along the eastern coast of South America in the Atlantic rainforest (Kimsey 1987;DAA Lucena unpubl. data). The geographical records cited by Kimsey (1987) for A. similis are the most disparate among these species, ranging from the southernmost limit of the Atlantic rainforest in Rio de Janeiro state (its type locality) to the northern boundary in the transition zone between the Atlantic rainforest and Caatinga in Caruaru-PE (Fig. 8A). The examination of high-resolution pictures of the holotype of A. similis deposited at CNC made it clear that this species is distinct from every new species described in this work. It is possible that the paratypes of A. similis from Caruaru-PE represent yet another undescribed species, which should be temporarily interpreted with caution (we have represented with a question mark in the distribution map: Fig. 8A). Future examinations of the complete type series of A. similis will prove essential to clear any doubts regarding the taxonomic identity of each specimen.
The present records of Duckeia species in quite distant sites in Central America and eastern Brazil without references to intermediate areas suggest a rudimentary knowledge of its diversity. We examined the type material of D. cyanea and D. vagabunda (USNM), and additional material of D. cyanea and D. gracile from localities with Atlantic rainforest collected in Espírito Santo state (Laranja da Terra), Minas Gerais state (Marliéria), and Rio de Janeiro state (Nova Iguaçu) (MZSP, UFES) (Fig. 8B). Duckeia dudui sp. nov. represents the first record of this genus in a habitat characterized by a seasonally dry forest. Kimsey (1986) placed four species in the fasciata species group in Cleptidea, three of them recorded only in Santa Catarina state, southern Brazil (Fig. 8C). The type series of Cleptidea fasciata was collected in an unknown locality in Brazil. Móczár (1996b) synonymized C. propodealis under C. fasciata. The former has only been known from Santa Catarina as well. Previously, Kimsey (1981) had cited C. fasciata for Guatemala, Panamá, Peru, and Argentina (Tucumán), but the distribution was later revised and restricted to Brazil and Argentina (Kimsey 1986: 324;Kimsey and Bohart 1991). We had examined specimens housed in the most important Brazilian entomological collections. Still, we only found new records for C. xanthomelas (Mocsáry, 1889), which seems to be the most common and abundant species in south-eastern Brazil, occurring mostly in localities of Atlantic forest of the following Brazilian states: Espírito Santo (Santa Teresa), Minas Gerais (Marliéria, Santa Bárbara, and São Gonçalo do Rio Abaixo), Paraná (Campina Grande do Sul), Santa Catarina (Blumenau), and São Paulo (Luis Antônio) (Fig. 8C). In this regard, the new species herein proposed and three unidentified male specimens from Maranhão represent now the most septentrional records for the fasciata species group in South America.

Conclusions
The diversity of Amiseginae and Cleptinae is probably the most underestimated in comparison with other groups of cuckoo wasps in South America. Their relatively small size and peculiar parasitoid lifestyles contribute to these wasps being rarely collected and frequently misidentified as other hymenopterans. The new species herein described in Amiseginae and Cleptinae enhance our understanding of chrysidid diversity in a poorly known Brazilian region. These new findings also reinforce the importance of long-term collecting studies and strategic samplings, which provide precious audited data potentially useful for selecting priority areas for conservation.