Cryptic Diversity of South African Trapdoor Spiders: Three New Species of Stasimopus Simon, 1892 (Mygalomorphae, Ctenizidae), and Redescription of Stasimopus robertsi Hewitt, 1910

ABSTRACT Three new species of Stasimopus Simon, 1892, are described from the Gauteng and North West provinces of South Africa. They are readily distinguished from all other known Stasimopus species by the presence of spinules in the tarsal scopulae on the first two pairs of legs of adult males. The only described species recorded in the vicinity, Stasimopus robertsi Hewitt, 1910, which appears to be endemic to northern Gauteng Province, is redescribed. The four species are comprehensively illustrated, a key to identify the adult males from those of other new species in the region is provided, and their conservation status discussed. Available evidence suggests that many Stasimopus species remain to be discovered and described, but significant effort will be required to obtain adult males, considered important for accurate species delimitation in the genus.


INTRODUCTION
The trapdoor spider genus Stasimopus Simon, 1892, is a diverse and abundant compo¬ nent of the southern African mygalomorph spider fauna. Forty-five infrageneric taxa have been described to date, mostly from the Eastern Cape, Free State, Northern Cape, and West¬ ern Cape provinces of South Africa (table 1). The genus has also been recorded from Lesotho, and probably occurs more widely in southern Africa. Stasimopus rufidens (Ausserer, 1871) is the largest trapdoor spider species known from the region, with adult females reaching a total length of 50 mm.
Stasimopus occur in dense colonies in some habitats, and are often the most abundant species in mygalomorph spider communities (I. Engelbrecht, unpubl. data). These spiders live in self-constructed, vertical, tubular burrows with a thick, corklike lid (Dippenaar Schoeman, 2002;fig. IE, F). When closed, the lid may be so well camouflaged against the surrounding substrate that it is impossible to detect. The spider will forcefully hold down the lid during attempts to open it, using the front legs and fangs, presumably as a defense against predators.
The genus Stasimopus is extremely conservative morphologically. Females and juveniles are stocky in appearance, with large bodies and short legs, adapted for spending their entire life within their tubular burrow ( fig. 2A, D, F). On reaching sexual maturity, males become more gracile ( fig. 2B, C, E), developing long, slender legs adapted for walking long distances in search of sexually receptive females. Morphological differences among Stasimopus species are often subtle, particularly in females (Hewitt, 1915a;Hendrixson and Bond, 2004). Species diagnosis is based primarily on ocular characters of both sexes, the presence and extent of patches of spinules on the dorsal surfaces of the pedipalps and first pair of legs in females, and the length of the pedipalps and spination of the first two pairs of legs in males. Raven (1985) maintained the placement of Stasimopus in Ctenizidae Tborell, 1887, based on a phylogenetic analysis of morphological characters, according to which Stasimopus shares seven putative synapomorphies with other ctenizid genera. Hedin and Bond (2006) suggested the genus may be sister to the Migidae Simon, 1889, based on a phylogenetic analysis of nuclear DNA sequences, but left its taxonomic placement in Ctenizidae unchanged. The monophyly of Stasimopus remains untested.
The present contribution is part of a greater project to document the diversity, distribution and conservation status of mygalomorph spiders in Gauteng Province, South Africa. Trapdoor spiders were an important component of a spatial biodiversity conservation plan developed for the province in 2001. This plan has been used as the primary decision support tool for envi¬ ronmental impact assessments during the past decade. It became apparent during the course of surveys conducted in the province and adjacent areas that trapdoor spider diversity was underestimated. Several new species were discovered in various genera, including Stasimopus.
In the present contribution, three new species of Stasimopus are described and the only described species recorded in the vicinity, Stasimopus robertsi Hewitt, 1910, is redescribed. The four species are comprehensively illustrated, a key to identify the adult males from those of other new species in the region is provided, and their conservation status discussed. Available evidence suggests that many Stasimopus species remain to be discovered and described, but significant effort will be required to obtain adult males, considered important for accurate spe¬ cies delimitation in the genus.

All Stasimopus material in the National Collection of Arachnida, Plant Protection Research
Institute (NCA), and the Ditsong National Museum of Natural History, formerly Transvaal Museum (TMSA), both in Pretoria, South Africa, was borrowed and examined for this study.
These collections house most of the Stasimopus material collected to date from localities falling within Gauteng Province and adjacent areas. Type specimens of all Stasimopus species and specimens used to describe opposite sexes after the original descriptions were borrowed from the following collections and examined for comparison: Albany Museum, Grahamstown, South Africa (AMGS); Natural History Museum, London, UK (BMNH); South African Museum, Cape Town, South Africa (SAMC); Museum fur Naturkunde, Berlin (ZMHB). Only the types of Stasimopus caffrus (C.L. Koch, 1842) and Stasimopus rufidens (Ausserer, 1871) could not be located. No other material is available for the former species, but the type of P. natalensis, a junior synonym of S. rufidens, and additional material of that species in the TMSA were exam¬ ined. Types of the new species described in this contribution are deposited in the TMSA and the American Museum of Natural History (AMNH).
Additional material was collected using pitfall traps to obtain adult males, and by scraping away the surface layer of soil to locate the burrows of females and juveniles. Specimens were fixed in 96% ethanol for several weeks and then transferred to 70% ethanol for long-term preservation. Selected specimens were photographed alive prior to fixation. Measurements were taken using a pair of digital calipers accurate to 0.1 mm and spine counts recorded using a Nikon SMZl-10 stereomicroscope. Meristic data were recorded from the sinistral appendage unless missing, in which case the dextral appendage was used. Appendage measurements were taken along the dorsal midline of the segment. Carapace length was measured along the mid¬ line from the anterior to the posterior margins, carapace width across the widest points, just above the coxae of the second pair of legs. Images were prepared using a Zeiss SteREO Discovery.V20 and AxioVision V4.8.2.0. Species descriptions are ordered in the text based on morphological similarity. A comprehensive description is provided for the first species. Only characters considered taxonomically informative within the genus are described for each spe-  Simon, 1892(Ctenizidae Thorell, 1887, with countries, provinces, and districts in which recorded. Data from primary taxonomic literature with administrative divisions determined for published locality records.
Color: 8: Carapace and chelicerae black. Sternum, labium, and leg coxae dark brown. Legs I, II, and IV, metatarsus and tarsus, leg III tarsus, and pedipalp tibia orange; rest of legs and pedipalps black. Abdomen, dorsal surface black; pale band across first pair of book lungs, second pair of lung covers pale; spinnerets pale; rest of ventral surface black. 9: Carapace, dorsal surfaces of pedipalps and legs medium to dark brown. Ocular tubercles infuscated.
Sternum, labium, and leg coxae slightly paler than carapace. Pedipalp coxae and legs I and II, ventral surfaces similar in color to dorsal surfaces; legs III and IV, ventral surfaces paler. Abdo¬ men, dorsal surface uniformly medium or dark gray without chevrons or blotches; ventral surface paler to yellowish brown anteriorly and around book lungs; spinnerets similar in color to ventral surface. approximately one-third to half distance to fovea.
The morphological similarity between females of S. robertsi and S. hewitti, n. sp., and the occur¬ rence of both species in the vicinity of Pretoria, casts uncertainty on Hewitt's (1910Hewitt's ( , 1915aHewitt's ( , 1916 identifications of female specimens from this area. Female specimens from Hatfield and Wonderboompoort listed in Additional Material were assumed to be conspecific with S. rob¬ ertsi when associated with adult males bearing the same collection data. However, the identity of female material from localities at which males were not collected is uncertain. The holotype female of S. dubius (TMSA 2745) was also examined. Hewitt (1916)  Distribution: Stasimopus robertsi is known from only four localities in the Pretoria District of Gauteng Province, South Africa, to which this species is presumed to be endemic ( fig. 3).
Habitat: The known locality records of S. robertsi fall within the Moot Plains Bushveld and Marikana Thornveld vegetation types (Mucina and Rutherford, 2006). The locality at Faerie Glen Nature Reserve comprises open woodland with red structured clay soils (Shortlands form), apparently derived from shale ( fig. 1A). NO. 3732 Conservation Status: Approximately 83% of the habitat in the known range of this spe¬ cies has been transformed by urban development, suggesting that this species may be threat¬ ened with extinction, but additional sampling is necessary to determine the full extent of its distribution and habitat requirements before a thorough conservation assessment is possible.   . 3), but additional sampling may reveal that it is more widespread in the northern parts of Gauteng Province and may extend into adjacent areas of neighboring provinces. The predominance of records around Pretoria reflects a peak of sampling by Transvaal Museum staff in the early 1900s.
Habitat: The species occurs in a wide range of habitats and soil types, and appears to be absent only from very sandy soils ( fig. IB). Although most of the known records occur in the Marikana Thornveld vegetation type (Mucina and Rutherford, 2006), it has also been recorded  Distribution: Stasimopus griswoldi, n. sp., appears to be restricted to the valley between the Magaliesberg and Skurweberg mountain ranges, north of the Magalies River and west of the Crocodile River ( fig. 3). It has thus far only been recorded from the Brits District of North West Province, but probably also occurs in adjacent parts of Gauteng Province.
Habitat: The known locality records of this species fall within the Moot Plains Bushveld vegetation type (Mucina and Rutherford, 2006), characterized by mixed woodland with Aca-   (table 2); pedipalps reaching distal end of leg I when both extended forward; higher spine counts on tibia and metatarsus of first two pairs of legs (table 3); tibia of leg I distinctly convex in lateral profile ( fig. 23C). Except for the foveal protuberance, S. filmeri, n. sp., and S. griswoldi, n. sp., appear to be morphologically identical.
Etymology: This species is named in honor of the late Martin Filmer, former chairman of the Spider Club of Southern Africa, who first recognized this species as new to science, and published a photograph of it in his book on southern African spiders (Filmer, 1997: 87), in recognition of his role in fostering arachnology in South Africa.
Description: As for S. hewitti, n. sp., with the following differences. Color: 6: Chelicerae uniformly black. Leg III metatarsus black proximally; leg IV metatar¬ sus mostly black, becoming paler distally. Book lung covers yellow or black; anterior pair pale, posterior pair black in some specimens; genital plate, between anterior book lungs, paler than rest of abdomen or similar in color to pale book lung coverings, creating pale band across anteroventral part of abdomen. 9: Carapace cephalic region darker in very large specimens, rendering infuscation around ocelli indistinct.
Ocelli: 8: Ocelli relatively smaller and more widely separated than in S. hewitti, n. sp., with anteromedian ocelli three diameters apart, anterolateral ocelli one diameter from anteromedian  fig. 24D-F), ventral surface with spines more robust than in S. hewitti, n. sp. Legs III and IV tarsi, paired claws each with two or three long teeth proximally, and smaller teeth extending more than half length of claw. 9: Legs I and II tibia and metatarsus, retroventral spines not elongated; metatarsus, pro dorsal surface with short, erect setae along entire length, retrodorsal surface with patch of spinules extending from proximal margin to beyond midpoint of seg¬ ment, becoming more sparse distally. Leg IV metatarsus with preening comb, comprising four or five elongated slender spines on distal retroventral margin.
Spermathecae: Entire, with inflated terminus (fig. 4D). Distribution: Stasimopus filmeri, n. sp., appears to be restricted to the same valley as S. griswoldi, n. sp., between the Magaliesberg and Skurweberg mountains, west of the Crocodile River ( fig.   3). However, S. filmeri, n. sp., occurs only on the southern side of the Magalies River. The known distribution of this species falls within the Krugersdorp District of Gauteng Province, South Africa, but the species probably occurs throughout the valley in the Gauteng and North West provinces. FIGURE 22. Stasimopus filmeri, n. sp., 9 (TMSA 23998). A. Prosoma, dorsal aspect. B. Prosoma, ventral aspect. C. Carapace, lateral aspect. Scale: 2 mm.
Habitat: The known locality records of this species fall within the Moot Plains Bushveld vegetation type (Mucina and Rutherford, 2006), characterized by mixed woodland with A. karroo, A. nilotica, and R. lancea ( fig. ID). The soils in the area are mostly red structured clays (Shortlands form) and shallow rocky soils (Glenrosa and Mispah forms) with high clay content.
This species occurs on both soil types, but appears to be more abundant on the deeper soils.

DISCUSSION
This study highlighted several important considerations for the systematics and conserva¬ tion of Stasimopus. Dippenaar Schoeman (2002: 31) and Hendrixson and Bond (2004: 2) sug¬ gested that a revision of Stasimopus would result in many new synonyms, but based on the diversity observed in Gauteng Province, we predict that most described species will prove to be valid and many new species await discovery and description.
Approximately 40% of the described species of Stasimopus are known only from females.
However, based on examination of numerous specimens during this study, we conclude that females exhibit limited morphological variation, with few discrete qualitative characters for species delimitation, compared to adult males. The limits of species known only from females should therefore be reevaluated; some may be species complexes and others may need to be synonymized (Engelbrecht and Prendini, 2011). However, species should not be described or synonymized in the absence of adult males. The collection of adult males from the type locali¬ ties of species currently known only from females should be prioritized. Different species have been found in close geographical proximity (sometimes syntopically) and the locations where types were collected is seldom precisely known. Adult males of all species occurring in the vicinity of type localities will therefore need to be collected to accurately determine the identi¬ ties of species previously described from females only, and it may be necessary to undertake scanning electron micrograph analyses of fine structure (e.g. spinneret spigot morphology) or sequence DNA barcodes to associate the sexes (Prendini, 2005).
The extent of habitat transformation caused by agricultural and urban development within the known ranges of all four species discussed in the present contribution is cause for concern.
Stasimopus robertsi may be a significant conservation priority given that it is known from only four localities within an area that has been and continues to be transformed. Further sampling for this species is a high priority to determine the extent and limits of its distribution and the threat of extinction. Extensive habitat transformation has also occurred across the known ranges of the three new species described above and appropriate measures should therefore be implemented to ensure their protection.