A new species of the genus Paraxenos Saunders, 1872 (Strepsiptera: Xenidae) from Bembix digger wasps (Hymenoptera: Bembicidae) and a redescription of Paraxenos hungaricus (Székessy, 1955)

. A new species of Strepsiptera of the genus Paraxenos Saunders, 1872 (Xenidae) from the United Arab Emirates is described. It was recorded from the host species Bembix kohli Morice, 1897 and represents the first occurrence of Paraxenos from Bembix Fabricius, 1775 in the Afrotropical region. A detailed redescription of the female cephalothorax of Paraxenos hungaricus (Székessy, 1955) is provided, together with the first description of the male cephalotheca. The holotype of Paraxenos krombeini Kifune & Hirashima, 1987 was redescribed. Additionally, a key for parasites of Bembix among Paraxenos species is provided based on characters of the female cephalothorax and male cephalotheca. The distribution and conservation status of Paraxenos spp. on Bembix are also discussed.


Introduction
Strepsiptera Kirby, 1813 is a highly derived order of endoparasitic insects characterized by a unique and complex life cycle and extreme sexual dimorphism (Pohl & Beutel 2008). Whereas the adult males are free-living and have an excellent flying capacity, adult females of Stylopidia Kinzelbach, 1969 (ca 95% of all strepsipteran species) are neotenic, lacking sensory organs and body appendages such as compound eyes, antennae, wings, and legs (Kinzelbach 1971;Pohl & Beutel 2005). Xenidae Saunders, 1872, a family of Strepsiptera with many derived features, are deeply nested within the clade Stylopiformia Kinzelbach, 1971(Pohl & Beutel 2005McMahon et al. 2011). Traditionally, the family was divided into four genera: Paragioxenos Ogloblin, 1923, Paraxenos Saunders, 1872, Pseudoxenos Saunders, 1872, and Xenos Rossi, 1794. Benda et al. (2022b provided a generic revision and a detailed checklist of Xenidae, and delimited 13 genera based on previous molecular phylogenetic studies (Benda et al. 2019(Benda et al. , 2021. The genus Paraxenos was identified as a lineage with Old World and Australasian distribution, with species parasitizing three genera of digger wasps: Bembecinus Costa, 1859, Bembix Fabricius, 1775, and Stizus Latreille, 1802 (Bembicidae: Bembicinae) (Benda et al. 2022b). Its monophyly was supported in an analysis of molecular data, and it was placed as sister taxon to most other Xenidae (Benda et al. 2021: fig. 1). The genus diagnosis consists of a combination of characters as no morphological synapomorphies were found (Benda et al. 2022b).
The species of the sand wasp genus Bembix are not well known as hosts of Strepsiptera. Stylopization of Bembix (i.e., infestation with strepsipterans) was recorded for the first time by Pierce (1909Pierce ( , 1919 based on correspondence with the British entomologist Robert Cyril Layton Perkins, who mentioned that many members of the family Bembicidae Latreille, 1802 are stylopized in Australia (Perkins 1905). Pierce (1919) also presented a record of stylopized Bembix texana Cresson, 1873 from Louisiana (USA). Another finding was published by Ulrich (1930) who reported parasitized Bembix rostrata (Linnaeus, 1758) from Germany.
Although some other records of stylopized Bembix were mentioned in the literature (Hofeneder & Fulmek 1942a, 1942b, the first species of Strepsiptera from Bembix was described more than ten years later. Székessy (1955) described Paraxenos hungaricus (Székessy, 1955) based on material collected in Central Hungary and reported Bembix oculata Panzer, 1801 as host. In the same year, Beaumont (1955) published findings of stylopized B. oculata and B. rostrata from Spain. Székessy (1959) listed stylopized hosts from the Hungarian Natural History Museum with new data on P. hungaricus from Central Hungary. In his study he also reported Bembix vespiformis F. Smith, 1856 from Australia. Kinzelbach (1978) assigned all published findings of stylopized Bembix from Europe to Paraxenos hungaricus. Previously, he recorded this species only from Hungary (Kinzelbach 1971). Kifune & Hirashima (1987) described three new species infesting Bembix, one from Sri Lanka and two from Australia. Until now, species of Paraxenos from Bembix were almost exclusively described based on the female cephalothorax. First instar larvae are not known, and adult males with a puparium only in the case of P. hungaricus (Székessy 1955).
In the molecular phylogeny of Benda et al. (2021), several strepsipteran samples from Bembix hosts were included, ranging from Italy to Mongolia. All of these samples formed a well-supported monophyletic lineage, and all applied species delimitation methods designated it as a single putative species. Based on these findings and using a morphological approach, Benda et al. (2022b) listed P. hungaricus as a widely distributed (transpalearctic) species, found in many countries but rarely collected. Here we describe a new species of Paraxenos associated with Bembix. We compare the morphology of the female cephalothorax and male cephalotheca with characteristics observed in previously described species. We also describe the male cephalotheca of P. hungaricus for the first time.

Taxon sampling
A total of 41 females and 26 male puparia of Paraxenos were obtained from species of Bembix or investigated directly on the host. Material from the following public and private collections was examined: HNHM = Hungarian Natural History Museum, Budapest, Hungary JSPC = Jakub Straka personal collection, Prague, Czech Republic KUMC = Kyushu University Museum Collection, Fukuoka, Japan NMPC = National Museum, Prague, Czech Republic OLML = Oberösterreichisches Landesmuseum, Linz, Austria

Fixation and preparation
All host individuals were first relaxed in water vapour and then immediately dissected. The endoparasitic females and males were removed from the abdomen of the hosts. Females and male puparia used for the morphological study were cleared using a mixture of lysis buffer ATL and proteinase K (Qiagen) heated to 56°C. The lysis procedure took several hours or overnight. Cleared specimens were washed in distilled water several times and then stored in vials with 96% ethanol. Complete female cephalothoraces and male puparia were air-dried using a micro-pad inserted into the cephalothorax to prevent the cuticle from collapsing during the process. The female body was usually extracted from the cephalothorax before drying. After this step and after the removal of the micro-pad, the dried specimens were glued onto card mounting points, which were pinned.

Measurements
The width and length of the female cephalothorax, the female head capsule, and the male cephalotheca were measured using a Leica S9D stereo microscope with a calibrated ocular micrometer. The cephalothorax length was measured from the apex of the clypeal lobe to the constriction of abdominal segment I; the width of the cephalothorax is the maximum distance between its lateral margins.

Photomicrography
The general habitus of stylopized host specimens and the host abdomen with protruding strepsipterans were documented using microphotography. Multi-focus images were taken using Canon EOS 550D or 70D cameras equipped with EF 50 mm and MP-E 65 mm macro lenses. Lateral lights and a diffuser were used.
For the documentation of the original coloration of the cephalothoraces, cephalothecae, or puparia, we used air-dried specimens glued to card mounting points. The specimens were photographed with a Canon EOS 7D digital SLR equipped with a Canon MP-E 65 mm macro lens (Canon, Krefeld, Germany) fitted with a StackShot macro rail (Cognisys, Traverse City, MI, USA). Each specimen was illuminated with two flashlights (Yongnuo Photographic Equipment, Shenzhen, China) fitted to a transparent cylinder for even and soft light. For the documentation of minute cuticular structures, a Canon EOS 70D camera attached to an Olympus BX40 microscope was used. The microscope was equipped with lateral lights and a diffuser. Zerene Stacker (Zerene Systems LLC, Richland, USA) was used to process stacks of images with different focus.

Scanning electron microscopy (SEM)
Dried female cephalothoraces glued to card points were mounted on a rotatable specimen holder (Pohl 2010). Each specimen was sputter coated with gold with an Emitech K 500 (Sample preparation division, Quorum Technologies Ltd, Ashford, England). The SEM micrographs were taken with an ESEM XL30 (Philips, Amsterdam, Netherlands) equipped with Scandium FIVE (Olympus, Münster, Germany).

Image processing
All images were processed and arranged into plates with Adobe Photoshop ® CS5 (Adobe System Incorporated, San Jose, USA) software. CorelDraw ® X8 (CorelDraw Corporation, Ottawa, ON, Canada) was used for the lettering of the plates.

Morphological terminology and description style
The terminology used for the female cephalothorax and male cephalotheca is adopted from Kinzelbach (1971), Löwe et al. (2016), Richter et al. (2017), and Benda et al. (2022aBenda et al. ( , 2022b. The cephalothorax and cephalotheca are described in a morphological orientation (as in the figures) although their functional orientation in the host's body is inverted.

Diagnosis
Female cephalothorax Differing from P. hungaricus and P. krombeini Kifune & Hirashima, 1987 by the maxilla shaped like an orthogonal triangle; maxillary base wide; maxilla narrowing anteromedially in contrast to anteriorly in P. hungaricus and P. krombeini. Maxillary base about 3× as wide as anterior part of maxilla (mxb, mx, Fig. 2A), very slightly overlapping with mandible proximally. Mandible completely enclosed in mandibular capsule as in P. hungaricus, in contrast to P. krombeini where the mandible overtops the anterior edge of the head capsule. Mandibular base slightly bulging, separated from genal area by furrow. Labial area between maxillae slightly wider than long (lba, Fig. 2A), versus distinctly wider than long in P. hungaricus (lba, Fig. 6C). Dorsal labral field distinctly arcuate (dlf, Fig. 2A) as in P. krombeini, in contrast to slightly arcuate in P. hungaricus. Mouth opening very slightly arcuate to straight and not distinctly sclerotized around margin (os, Fig. 2A), versus conspicuously sclerotized in most specimens of P. hungaricus. Lateral extensions of head capsule dull on ventral side (lehc, Fig. 2A), covered by conspicuous and densely arranged dark papillae, cuticle wrinkled between papillae, in contrast to lateral extensions pale, shiny to dull, without densely arranged papillae in P. hungaricus. Lateral region of abdominal segment I below spiracles only slightly darker than dorsal side (asI, Fig. 1D), and not distinctly contrasting to pale thorax as in P. hungaricus. Anterior head margin distinctly protruding as in P. krombeini, versus only slightly protruding from head capsule in P. hungaricus.

Male cephalotheca
Paraxenos arabicus sp. nov. differs from P. hungaricus by several characters. In lateral view, cephalotheca rounded anteriorly, versus protruding and acute anteriorly in P. hungaricus. Clypeus (clypeal lobe) not projecting in lateral view (cl, Fig. 2D), but prominent in P. hungaricus. Gena around compound eye nearly completely pale (gn, Fig. 2C), with dark area around mandibular base reduced; pale stripe between compound eye and mandibular base wide, about 2× as wide as diameter of compound eye; in P. hungaricus about as long as diameter of compound eye. Maxilla not wide at base, about 1.5× as wide as mandible (mx, Figs 2C, 4F), but about 2× as wide as mandibular width in P. hungaricus. Length of clypeal lobe approximately equal to mandibular length (cll, Fig. 4F), versus clypeal lobe distinctly wider than length of mandible in P. hungaricus.

Etymology
The specific epithet refers to Arabia, the region of origin of the new species. Adjective.
head capSule. Approximately ⅓ as long as entire cephalothorax including lateral extensions. Coloration pale to dark with specific pattern ( Fig. 2A). Clypeal area well separated from labral area, protruding anteriorly and forming clypeal lobe. Surface slightly wrinkled on dorsal side. Lateral extensions of head capsule dull ventrally (lehc, Fig. 2A), covered by conspicuous dark papillae, with cuticle wrinkled between them. Clypeal sensilla mainly concentrated on clypeal lobe and extending to ventral side of clypeal area. Border between clypeal and frontal regions not clearly recognizable, but present. Frontal region distinctly wrinkled, without dark papillae. Segmental border between head and prothorax indistinct on dorsal side, but indicated by change in cuticular surface structure.
antenna. Morphology of antennal vestige not recognizable.
mandible. Anteromedially directed at an angle of 30°, completely enclosed in mandibular capsule or very slightly protruding from it. Mandibular bulge not distinctly raised, with approximately 15 sensilla. Cuticle almost completely smooth, partially sculptured on articulatory area. Mandibular tooth narrow, pointed, sharply curved anteriorly. Mandibular base slightly bulging, divided by furrow from genal area ( Fig. 2A).
Metasternum with large pale area medially without pigmented papillae and width reaching up to ⅔ of metathorax (Fig. 1C). Dorsal side of thorax smooth or slightly reticulated.
abdominal Segment i and SpiracleS. Lateral region of abdominal segment I below spiracles only slightly darker on dorsal side, not distinctly contrasting to pale thorax (asI, Fig. 1D). Spiracles on posterior half of cephalothorax slightly elevated, with lateral or laterodorsal orientation.
Variability of female cephalothorax. Cephalothorax compact, nearly as long as wide or distinctly wider than long. Size slightly variable, length 1.18-1.32 mm, maximum width 1.32-1.68 mm. Dorsal labral field with about 21 to 25 setae inserted in cavities. Mandible anteromedially directed at an angle of 30-45°. Mandibular bulge not distinctly raised, with approximately 12-18 sensilla. Mouth opening very slightly arcuate to straight.
cephalothecal capSule. Compound eyes pale, with darker individual cornea lenses recognizable. Gena around compound eye almost completely pale, dark area around mandibular base missing; pale area between compound eye and mandibular base about 2× as wide as diameter of compound eye (gn, coe, Fig. 2C). Clypeus pale medially and darker laterally. Clypeal lobe straight in frontal view, blunt and not prominent in lateral view. Length of clypeal lobe nearly equal to mandibular length (cll, Fig. 4F). Sensilla mainly concentrated on clypeal lobe, visible. Frontal region with paired furrow of supra-antennal sensillary field, lacking impression or clearly recognizable occipital bulge (fssf, Fig. 2C). Occipital bulge indistinct (ob, Fig. 2C). Diameter of genae between maxillary base and compound eye more than 2× as long as diameter of vestigial antenna.
antenna. Of standard shape, pale or dark, small, with small plates, sensilla and complete torulus (a, Figs 2C, 4F). Periantennal area not clearly delimited from supra-antennal sensillary field, pale or dark.
mandible. Nearly medially directed. Tooth apically pointed, wide basally, but not reaching area of mandibular bulge. Coloration mostly pale, especially middle region, with darker parts apically and basally. Mandibular bulge with sensilla, separated from tooth. Length between mandibles nearly equal to mandibular length. labium and hypopharynx. Labium distinct, located between and below maxillae, darker. Praementum and postmentum more or less distinctly separated by furrow (Fig. 2C). Hypopharyngeal protuberance absent. mouth opening. Visible, not covered by ventral labral field, slightly arcuate, sclerotized around margin.

Female cephalothorax
Shape and coloration. Compact, widened, slightly or distinctly wider than long. Size variable, length 1.28-1.93 mm, maximum width 1.68-2.57 mm. Anterior head margin slightly or scarcely protruding from head capsule. Thorax widening posteriorly. Cephalothorax displaying multiple light brown shades, only around mandible and mouth opening cuticle more sclerotized and darker, but lighter in central region, darker laterally. head capSule. Approximately ⅓ to ½ as long as entire cephalothorax including lateral extensions. Coloration pale to dark, with specific pattern. Clypeal area distinctly or indistinctly separated from labral area, slightly protruding anteriorly or not protruding. Clypeal lobe blunt. Surface slightly wrinkled on dorsal side, reticulated. Lateral extensions of head capsule predominantly dull on ventral side, cuticle wrinkled, but shiny area near submaxillary groove lacking conspicuous dark papillae (lehc, Fig. 3C, lehc, smxg, Fig. 6A). Clypeal sensilla present on ventral side of clypeus, mainly concentrated on clypeal lobe (cls, Fig. 6D). Border between clypeal and frontal region indistinct, but still recognizable (Fig. 5F). Frontal region of head capsule distinctly wrinkled, not covered by dark papillae. Segmental border between head and prothorax indistinct on dorsal side, indicated by change in cuticular surface structure.
labrum. Ventral field distinctly wider than long, elliptic or semicircular. Dorsal field slightly arcuate to nearly straight, > 4× as wide as long in midline. Dorsal field with about 24 setae inserted in deep sockets ( Fig. 6C-D).
labium. Labial area between maxillae distinct, delimited anteriorly by mouth opening and posteriorly by birth opening. Distinctly wider than long in midline, rectangular, flat. Cuticular surface very slightly wrinkled, reticulated. mouth opening. Mouth opening straight or bi-arcuate, sclerotized along margin.
thorax. Pro-mesothoracic and meso-metathoracic borders distinct on ventral side, separated by mesal furrows (sbpm, sbmm, Fig. 3C). Border on dorsal side indistinct, indicated by different cuticular sculpture. Border between metathorax and abdomen formed by ridge in combination with changed cuticular sculpture and coloration. Cuticle of thoracic segments reticulate on ventral side, with small, scattered pigmented papillae. Prosternal extension variable, differentiated by cuticular sculpture and coloration (pst, Fig. 3C). Anterior part usually darker, with more or less distinct pigmented papillae medially. Posterior part usually pale and without conspicuous pigmented papillae. Meso-and metathorax unmodified in shape, transverse. Posteromedial pale area on mesosternum and metasternum variable in shape, in some specimens indistinct (mst, mtst, Fig. 3C). Dorsal side of thorax smooth or slightly reticulated.
abdominal Segment i and SpiracleS. Lateral region of abdominal segment I below spiracles conspicuously darkened on dorsal side, contrasting to pale thorax (asI, Fig. 3D). Spiracles on posterior half of cephalothorax slightly elevated, with lateral or laterodorsal orientation.
cephalothecal capSule. Compound eyes visible, pale to dark, with dark individual cornea lenses. Gena completely pale except dark area around mandibular base; pale area between compound eye and mandibular base narrowed (nearly as wide as diameter of compound eye) (gn, coe, Fig. 4A). Clypeus pale medially (on clypeal lobe) and darker laterally. Clypeus (clypeal lobe) straight in frontal view, prominent in lateral view, but blunt apically (cl, Fig. 4B). Clypeal lobe distinctly wider than mandibular length (cll, Fig. 4E). Clypeal sensilla mainly concentrated on clypeal lobe, visible or indistinct (Fig. 4A, E). Frontal region with paired furrow of supra-antennal sensillary field, lacking impression or occipital bulge (fssf, Fig. 4A-E). Diameter of genae between maxillary base and compound eye approximately 2× as large as diameter of vestigial antenna.
antenna. Of standard shape, dark, small, with small plates or sensilla and complete torulus (a, Fig. 4A, E). Periantennal area dark, not clearly delimited from supra-antennal sensillary field.
mandible. Nearly medially directed. Tooth inconspicuous, apically pointed, wide basally, but not reaching area of mandibular bulge. Coloration pale with darker parts. Mandibular bulge with sensilla, separated from pointed tooth. Distance between mandibles very distinctly exceeding mandibular length (Fig. 4E).

Diagnosis
Female cephalothorax Differing from P. arabicus sp. nov. and P. hungaricus in several characters. Mandible projects beyond anterior edge of head capsule (md, Fig 7A), versus enclosed in mandibular capsule in P. arabicus and P. hungaricus. Maxilla triangular, similar to that of P. hungaricus, but pointed apically (mx, Fig. 7A). Maxilla not overlapping or touching mandible. Labial area between maxillae wide (lba, Fig. 7A). Dorsal labral field distinctly arcuate as in P. arabicus. Clypeal lobe distinctly protruding from head capsule as in P. arabicus versus slightly protruding in P. hungaricus.

Discussion
The species of Paraxenos parasitizing Bembix hosts can easily be distinguished from other species of the genus by very wide cephalothecae and cephalothoraces. Important characters for species identification are the shape of the mandibles and maxillae, the sculpture and coloration of the cuticle, and the shape of the clypeus. These characters are confirmed with important diagnostic features of the female cephalothorax and male cephalotheca used for differentiation of species of Xenos (Benda et al. 2022a). Although we examined the holotype of Paraxenos krombeini, the cuticular sculpture was not visible due to inadequate mounting on a slide (Fig. 7). For future research, we therefore recommend gluing specimens onto the tip of card mounting points for easy examination of the dorsal and ventral sides of the cephalotheca or cephalothorax. This avoids or reduces artefacts caused by the preservation medium and facilitates documentation using scanning electron microscopy and other techniques (Benda et al. 2022b).
The host genus Bembix is composed of more than 300 species of ground-nesting wasps inhabiting sandy substrates (Frank 2022;Pulawski 2022). They are distributed on all continents except for Antarctica, with the greatest diversity in the Afrotropical and Australian regions (Bohart & Menke 1976). Although these hosts are very large and attractive wasps easy to observe and collect in sandy habitats, relatively scarce data on Paraxenos have been available so far. For instance, only three Australian species were recorded and described in a single study (Kifune & Hirashima 1987). One reason may be their inconspicuousness on the host compared to other species of Xenidae. In addition to the generally flattened cephalothorax, species of Paraxenos from Bembix also have flat male cephalothecae (Benda et al. 2022b). Females and male puparia do not project from the host. Remaining concealed below the host abdominal tergites (see Figs 1A, 3A), they can easily escape attention in contrast to species of other genera of Xenidae.
Unfortunately, there is no information on the phylogeography and evolutionary history of Bembix, which could explain the distribution of its Paraxenos parasites. Studies covering the New World (Pierce 1908(Pierce , 1909(Pierce , 1919 tentatively suggested that Paraxenos did not disperse into the New World biogeographic region or alternatively became extinct there. The absence of Paraxenos in the New World can also be explained by the rarity of potential hosts. In this context, the record of stylopized Bembix texana published by Pierce (1919) appears doubtful. This could be due to misidentification or a rare case of host switch by another genus of Xenidae (e.g., Eupathocera Pierce, 1918). Paraxenos arabicus sp. nov. is the first species of Paraxenos recorded in the Afrotropical region as currently recognised biogeographic distribution (Morrone 2002). Intensified screening of stylopized Bembix by researchers in this area would probably reveal more records. The great diversity of Bembix in the Afrotropics suggests that many more undescribed species of Paraxenos may occur in this region.
Although there are still insufficient data concerning the distribution of species of Paraxenos from Bembix, the presently available information suggests possible differences in the size of their distributional areas (Fig. 8, Table 1). Paraxenos hungaricus has a transpalearctic distribution also including Mediterranean islands such as Sicily or Sardinia, which was also supported by molecular data (Benda et al. 2021). In contrast, P. krombeini has only been found on Sri Lanka. Additional data of stylopized Bembix from the Oriental region are needed to confirm it as either endemic or a species with a wider distribution. In some areas of their distribution range, species also may become endangered, especially if they occur only locally in a specific habitat of shifting sands with a high abundance of their hosts. In Central Europe, P. hungaricus is likely to be threatened due to habitat loss and decline of available nesting sites for species of Bembix. The distribution area of its host Bembix rostrata is shrinking, and the number of populations has apparently declined in many parts of Europe, and this probably also applies to other species of Bembix (Blösch 2000;Klein & Lefeber 2004).