Rediscovery of the endemic Afrotropical genus Spathioplites (Hymenoptera, Braconidae, Doryctinae) with major range extension records for Spathioplites phreneticus

The endemic, monotypic Afrotropical genus Spathioplites Fischer, 1962 is rediscovered based on new specimens collected in South Africa and Senegal. Spathioplites phreneticus Fischer, 1962 was previously known from the holotype (male) and 12 paratypes (11 males and a female) collected in Chad in 1959. As part of an ongoing long-term insect inventory survey program in Africa new specimens were recently collected in Tswalu Kalahari Game Reserve in South Africa, extending the distribution range southwards by 4900 km. An additional historical specimen from Senegal was discovered in the collections of the Natural History Museum in Paris, extending the range westwards by 4000 km. Possible reasons for the disjunct distribution exhibited by current locality records for this species are discussed. The holotype male and a paratype female, as well as one of the two newly collected South African females were imaged. These photographs, as well as genus and species re-descriptions, are provided. An identification key to the Old World genera in the doryctine tribe Spathiini s. str. is also presented. All images and interactive identification keys are available on www.waspweb.org. African Invertebrates 62(2): 497–520 (2021) doi: 10.3897/AfrInvertebr.62.74103 https://africaninvertebrates.pensoft.net Copyright Simon van Noort et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. RESEARCH ARTICLE Simon van Noort et al. / African Invertebrates 62(2): 497–520 (2021) 498


Introduction
The generic structure of the tribe Spathiini has undergone extensive changes in composition over the last 20 years. Previously, based only on assessment of morphological characters (defined mainly by the petiolate first metasomal tergite with a strongly elongated acrosternite) this tribe included 20 genera in seven subtribes (Belokobylskij 1992(Belokobylskij , 2002Belokobylskij and Quicke 2000;Belokobylskij et al. 2004). However, more recent molecular phylogenetic studies of the doryctine genera changed the generic composition concept for this tribe (Zaldívar-Riverón et al. 2008;Jasso-Martínez et al. 2019). As a result, Platyspathius Viereck, 1911 and Antespathius Belokobylskij, 1995(former subgenus of Spathius Nees, 1819 were transferred to the tribe Rhaconotini, Hemispathius Belokobylskij and Quicke, 2000 was included in the tribe Doryctini, and the group of the previously studied Neotropical Spathiini genera (including a number of genera from other tribes) were recovered in a well-supported South American clade, unrelated to the Old World spathiines. Currently, the Old World Spathiini includes the following eight genera: Afrospathius Belokobylskij and Quicke, 2000;Ceylonspathius Belokobylskij, 2002; Kalahari Game Reserve and run from October 2015 until May 2017. These surveys comprised deployment of Malaise traps, yellow pan traps, pitfall traps and yellow funnel traps, which operated continuously using propylene glycol as the preservative, and were serviced on a 6-8 weekly basis. The insect traps generated 186 bulk samples of insects, comprising thousands of specimens. The Malaise traps were frequently damaged or destroyed by game animals during this inventory survey, even though white barrier ropes were erected around the traps to make them more visible to animals. The solution would have been to completely surround the Malaise trap with a 2 meter high wire mesh fence, but logistics did not allow for this. Five different vegetation types were sampled: Gordonia Plains Shrubland, Gordonia Duneveld, Kathu Bushveld, Koranna-Langeberg Mountain Bushveld, and Olifantshoek Plains Thornveld in the Kalahari ecosystem, vegetation classification after Mucina and Rutherford (2006).

Photography
Images were acquired at SAMC with a Leica LAS 4.9 imaging system, comprising a Leica Z16 microscope with a Leica DFC450 Camera and 0.63× video objective attached. The imaging process, using an automated Z-stepper, was managed using the Leica Application Suite V 4.9 software installed on a desktop computer. Diffused lighting was achieved using a Leica LED 5000 Dome. Multilayer images were acquired at MNHN using a Canon EOS 6D camera (Tokyo, Japan) with MP-E 65mm Macro f/2.8 lens, associated with a Cognisys Rail macro Stack Shot (Traverse City, MI, USA) driven by the software Helicon Remote (Kharkiv, Ukraine; http://www.heliconsoft. com /). Resulting photographs were combined with Helicon Focus 6. All images included in this paper, as well as additional images are available on WaspWeb: www. waspweb.org (van Noort 2021) and images of the types are available on the MNHN's science collections' website: https://science.mnhn.fr.

Results
Identification key to the Old World Spathiini genera 1 All femora (especially fore femora) with a distinct ventral tooth present subapically (A Postgenal bridge wide. Maxillary palp short, 6-segmented, sixth segment about as long as fifth segment; labial palpi 4-segmented, third segment long. Antenna weakly claviform, all flagellar segments (especially apical ones) distinctly compressed. Scape of antenna wide and short, without apical lobe or basal constriction, its ventral margin (in lateral view) shorter than dorsal margin. First flagellar segment weakly compressed, weakly curved outer, as long as second segment. Apical segment obtuse or very weakly acuminated, without spine on tip.
Legs. Fore tibia with several short and slender spines arranged in rather wide stripe. Middle tibia with distinct, short and thick spines arranged in narrow stripe along its dorsal margin. All tibiae distinctly thickened. Middle tarsal segments rather short. Hind coxa short, wide, subround, without basoventral corner and tooth. All femora wide, without dorsal protuberance. Hind femur elongate-oval, wide. Hind tibia with several short and thick spines along its dorsal margin. Basitarsus of hind tarsus short, about 0.5× as long as second to fifth segments combined.
Metasoma. First tergite semi-petiolate, rather long and wide. Acrosternite of first segment distinctly elongate, 0.45-0.50× as long as first tergite, its anterior margin placed at level of spiracles or behind it. Dorsope and basolateral lobes of first tergite absent or small. Distinct spiracular tubercles situated in basal 0.35-0.40 of first tergite; dorsal carinae absent, but rather distinct in basal half in male. Second and third tergites without any furrows and areas, but in male with rather distinct and almost complete, weakly divergent posteriorly sublateral (started from anterolateral corner) and not high longitudinal carinae. Suture between second and third tergites indistinct or very fine (in male). Laterotergites of second and third tergites not fused laterally with each other. Second to fifth (in female) or second to fourth (in male) tergites with separate laterotergites. Fourth to sixth tergites with relatively wide stripe of short, sparse, semierect pale setae in their posterior halves. Fourth tergite of female weakly enlarged, almost entirely covering following segments; apical segments only weakly protruding behind fourth tergite; but in male, tergites behind third tergite the same length and all protruding. Hypopygium in posterior margin with distinct pointed median process. Ovipositor apically without dorsal nodes. Ovipositor sheath distinctly not widened towards apex, shorter than metasoma.
Diagnosis. The genus Spathioplites is characterized by the combination of the following diagnostic apomorphic characters: antenna apically widened and flattened; frons and vertex anteriorly with lateral protuberances; metanotum with long, flat and distally strongly widened median process; hind tibia with distinct strong spines along the dorsal margin; petiole (1-SR) of discoidal (discal) cell of fore wing in female large and thick; metacarp (1-R1) absent in distal half; submedial cell of hind wing long; acrosternite of first metasomal segment elongate, about as long as half of tergite; prescutellar depression situated around the most part of scutellum. On the basis of these listed diagnostic characters, Spathioplites was placed in the monotypic subtribe Spathioplitina of the tribe Spathiini (Belokobylskij 1992).
Comments. Position of this genus in Spathiini is still questionable, because the acrosternite of the first metasomal segment is not strongly elongated and is only about half as long as the first tergite. This genus is additionally characterized by the presence of distinct spines on the dorsal side of the hind tibia, a character which is also known in the Doryctini group of Doryctophasmus genera (including Doryctophasmus Enderlein, 1912, Esterella Pagliano and Scaramozzino, 1990, Euscelinus Westwood, 1882, and Sonanus Belokobylskij and Konishi, 2001 (Belokobylskij 2015), but Spathioplites is taxonomically not related to the members of this group. Perhaps only molecular data will help to resolve the tribal position of this very peculiar genus, and reveal whether it actually belongs to the tribe Doryctini, Rhaconotini or Spathiini.

Spathioplites phreneticus
Transverse diameter of eye about as long as temple (dorsal view). Ocelli small, in triangle with base 1.1× its sides. POL 1.5-1.8× Od, about 0.4× OOL. Eyes with very fine emargination opposite antennal sockets, distinctly directed forwards (lateral view), 1.4× as high as broad. Malar space 0.8× maximum diameter of eye, 1.2× basal width of mandible. Face convex, its minimum width 1.2× height of eye and 0.9× height of face and clypeus combined. Hypoclypeal depression width 0.4× distance from edge of depression to eye, 0.3× width of face. Head below eyes (front view) strongly and roundly narrowed. Hypostomal flange narrow. Antenna narrow basally and widened apically, 19 to 20 segmented, about 0.6× as long as body. Scape 1.4-1.6× longer than its maximum width. First flagellar segment (lateral view) 4.5-4.8× longer than its apical width, 1.10-1.15× longer than second segment. Penultimate segment (lateral view) 1.4-1.6× longer than wide, 0.5× as long as first segment, 0.8× as long as apical segment.
Metasoma about as long as head and mesosoma combined. First tergite distinctly arched, dorsally and ventrally rather distinctly concave (lateral view), distinctly and almost linearly widened from base to apex (dorsal view). Maximum apical width of first tergite 2.5-2.8× its minimum width; length of tergite 1.4-1.5× its maximum width, about 1.7× length of propodeum. Median length of second tergite 0.9× its basal width, 1.4× length of third tergite. Length of second and third tergites combined 1.4-1.7× basal width of second tergite, 0.8-0.9× their maximum width. Ovipositor sheath 0.6-0.7× as long as metasoma, 0.9× as long as mesosoma, 0.55-0.65× as long as fore wing.
Sculpture and pubescence. Vertex entirely densely and distinctly curvedly transversely striate with reticulation between striae, lateral tubercles densely reticulate-areolate, with fine aciculation in parts; frons entirely curvedly striate and with reticulation between striae; face coarsely and rather sparsely transversely curvedly striate with distinct rugosity between striae; temple subvertically curvedly striate, with subtransverse and fine striation along eye. Median lobe of mesoscutum densely and distinctly reticulate-areolate, lateral lobed finely and densely granulate, with coarse rugosity on rather narrow area in posterior 0.4. Scutellum finely and very densely granulate. Mesopleuron smooth in lower half, transverse striate in upper half. Propodeum densely transversely striate with numerous dense rugosities in anterior 0.7, sparsely or very sparsely longitudinally striate with smooth large areas. Hind coxa densely curvedly transversely striate dorsally or in dorsal half, finely reticulate-coriaceous to smooth in ventral half. Hind femur densely and finely reticulate-coriaceous becoming finer below or smooth in lower half. First tergite rugose-reticulate in basal 0.4, distinctly and densely longitudinally striate with dense rugulosity between striae in apical 0.6. Second tergite densely and finely longitudinally striate with reticulation in basal 0.7, densely and finely reticulate- Figure 5. Spathioplites phreneticus, female paratype (MNHN) A habitus, lateral view B habitus, dorsal view C habitus, dorso-posterior view D head, antenna, anterior view E data labels areolate in apical 0.3. Third tergite densely and finely reticulate-coriaceous becoming very fine posteriorly. Fourth and fifth tergites very finely coriaceous. Vertex with sparse, short and semi-erect pale setae almost entirely. Mesoscutum with very sparse and short semi-erect white setae, glabrous medially. Metapleuron mainly glabrous. Hind tibia dorsally with short, dense and adpressed pale setae; their length much less than maximum width of hind tibia.
Colour. Body light reddish brown, metasoma behind petiole reddish brown becoming dark reddish brown in posterior quarter; sometimes whole body darker. Antenna brownish yellow to light brown in basal 0.3, yellow medially, contrasted dark brown to black in apical quarter. Palpi reddish brown or dark reddish brown. Legs light reddish brown, hind coxa and tibia reddish brown or darker, hind tibia dark basally; rarely middle and hind legs mainly reddish brown. Fore wing strongly maculate, brown spots and stripes formed by brown or dark brown and dense setae, hyaline intermediate areas and stripes without any setae. Pterostigma brown, except for pale yellow to almost white basal quarter.
Biology. According to the original label present on the holotype specimen the host of this species could either be wood-boring Lyctus hipposideros Lesne, 1908, or Enneadesmus forficula (Fairmaire, 1883. A label on the paratype reads "ex A. raddiana" = Acacia raddiana Savi, which is currently considered to be a subspecies of the Umbrella thorn acacia: Vachellia tortilis subsp. raddiana (Savi) Kyal. & Boatwr. The northern African specimens were collected in mid to late northern hemisphere summer (May and August), and the two South African specimens were collected in the latter half of the southern hemisphere summer (between 10 February and 9 May). Seasonal variation in arid habitats can be pronounced as shown in figures 6 and 7 depicting the sampling locality in Tswalu Game Reserve (South Africa). Dead wood was gathered from the surrounding vicinity and stacked under the Malaise trap to increase return on parasitoids of wood-boring hosts (Fig 7 C), and it is likely that the S. phreneticus specimens emanated from this source that probably contained woodboring host beetle larvae.
Comments. The long-term continuous inventory survey sites deployed in Tswalu Kalahari Game Reserve generated 186 bulk samples of insects, comprising thousands of specimens, of which 25 of 42 Malaise trap samples have had the Hymenoptera extracted and sorted to family. Only two female specimens of Spathioplites phreneticus were recovered from the 25 samples processed, and these were both from one (TSW15-DED3) of the five sites sampled with Malaise traps. The habitat of this sampling site is depicted in Figs 6 C-D and Figs 7 A-F. The two females were collected between 10 February and 9 May 2016 during a below average rainfall summer season. The sampling localities across the five different vegetation types are depicted in Fig 6 E. The only site that returned S. phreneticus specimens was situated in Olifantshoek Plains Thornveld.

Discussion
Spathioplites phreneticus is probably an arid-adapted species, based on the current distribution and habitats of the sampled localities. The extreme disparity in the northern Chad and southern South African localities (4900 km apart) is likely determined by the disjunct north-south aridity patterns of the African continent (Masih et al. 2014). Given the additional far western record in Senegal (4000 km west of the type locality), the distribution of the species in northern Africa is likely to be associated with the arid Sahel transverse belt south of the Sahara (Fig. 6 B). The disrupted north-south distribution pattern emulates that of the pycnostigmine figitids (Cynipoidea) (Buffington and van Noort 2007), which also exhibit a similar fragmented distribution pattern, as do a number of other taxa, such as butterflies (Carcasson 1964), masarine wasps (Gess 1992), flies (Bowden 1978), mammals (Grubb 1999;Grubb et al. 1999;Vrba 1992) and plants (White 1983). As for Spathioplites, no pycnostigmine specimens have ever been collected in the mesic intervening areas of central Africa (Buffington and van Noort 2007). Masarine wasps (Vespidae) also exhibit this biogeographical pattern (Gess 1992), suggesting a similar historical evolutionary driving process for these groups of wasps. Historical physiographical change would have driven climate change and hence adaptation of vegetation type and habitat. Aridification of southern Africa commenced around 20 million years ago with the formation of the cold Benguela current upwelling system, which removed the moisture supply to the interior of the continent from the Atlantic Ocean. The process of drying out was further enhanced by upliftment of the continent (the African Super Swell), a process resulting in an increase in the height of the eastern escarpment, further establishing the east-west rainfall gradient (King 1978;McCarthy and Rubidge 2005). The contemporary northern and southern arid areas of the African continent were linked in the past. A period of cooling in the middle Miocene 15.6-12.5 million years ago (Denton 1999), and during repeated interglacial periods of the Pleistocene epoch when forests in tropical Africa contracted considerably, forming refugia with a corresponding expansion of more xeric environments (Axelrod and Raven 1978;Denton 1999) resulted in an eastern swath of aridity on the African continent. Buffington and van Noort (2007) hypothesized that the historical distribution of pycnostigmine wasps would have been affected by these phases of being alternately continuous and disjunct, culminating in the current phase of distributional vicariance. This is a plausible hypothesis of underlying evolutionary drivers that may also have determined the contemporary distribution for Spathioplites phreneticus.
Spathioplites phreneticus is probably not rare and is simply under-sampled due to lack of collecting in arid environments that are often perceived to be depauperate in terms of insect species-richness. Emergence of adult S. phreneticus is likely to occur during a very short window of suitable environmental conditions. Arid areas are usually highly seasonal with sporadic rainfall events. When these precipitation events do occur an explosion of adult invertebrate activity ensues, which may be short-lived. These cycles of population explosion may only occur every couple of years and often are not an annual event, further hindering collection of adult wasps. The invertebrate inventory surveys run by Simon van Noort over the last 30 years have been designed to be continuous and long-term (preferably a minimum of a year), but ideally they are run for at least 5 years to cover annual variations in wasp emergence. This strategy means that periods of short-burst activity do not need to be predicted as traps are already in place to obtain return when these events occur. Continuous sampling also means that during periods of perceived non-optimal environmental conditions (which historically have not been sampled, because of a perception of lack of activity) rare taxa that only emerge during these periods are sampled, such as species of Nixonia Masner, 1958 that emerge as adults during the hot and dry unproductive ecological period in the summer months (van Noort and Johnson 2009). The two Spathioplites females that were collected between February and May 2016 encompassed the mid to end period of the wetter summer season. The collecting of these specimens is likely a result of the dead wood that was gathered from the surrounding vicinity and stacked under the Malaise trap to increase chances of collecting parasitoids of wood-boring hosts (Fig. 7 C). This approach was also used at the other surveyed localities, which were situated in different vegetation types: Gordonia Plains Shrubland, Gordonia Duneveld, Kathu Bushveld and Koranna-Langeberg Mountain Bushveld in the Kalahari ecosystem, vegetation classification after Mucina and Rutherford (2006). The site that returned the S. phreneticus specimens was situated in Olifantshoek Plains Thornveld, and these preliminary results suggest that there may be habitat fidelity, as no specimens were recovered from any of the samples emanating from the other vegetation types.
Implementation of long-term invertebrate inventory surveys are critical to assess seasonal variability and natural longer-term climatic cycles affecting composition of invertebrate faunal assemblages over time. These surveys, however, produce huge numbers of specimens, requiring availability of substantial resources for sorting the bulk samples, and subsequent curation and identification of specimens. Further specimens of S. phreneticus may very well still emerge from the remaining unprocessed samples collected during the invertebrate survey of Tswalu Kalahari Private Game Reserve.