Tarsal Organ Morphology and the Phylogeny of Goblin Spiders (Araneae, Oonopidae), with Notes on Basal Genera

ABSTRACT Based on a survey of a wide variety of oonopid genera and outgroups, we hypothesize new synapomorphies uniting the Oonopidae (minus the South African genus Calculus Purcell, which is transferred to the Orsolobidae). The groundplan of the tarsal organ in Oonopidae is hypothesized to be an exposed organ with a distinctive, longitudinal ridge originating from the proximal end of the organ, and a serially dimorphic pattern of 4-4-3-3 raised receptors on legs I–IV, respectively. Such organs typify the diverse, basal, and ancient genus Orchestina Simon. Several other genera whose members resemble Orchestina in retaining two plesiomorphic features (an H-shaped, transverse eye arrangement and a heavily sclerotized, thick-walled sperm duct within the male palp) are united by having tarsal organs that are partly (in the case of Cortestina Knoflach) or fully capsulate (in the case of Sulsula Simon, Xiombarg Brignoli, and Unicorn Platnick and Brescovit). The remaining oonopids are united by the loss of the heavily sclerotized palpal sperm duct, presumably reflecting a significant transformation in palpal mechanics. Within that large assemblage, a 4-4-3-3 tarsal organ receptor pattern and an H-shaped eye arrangement seem to be retained only in the New Zealand genus Kapitia Forster; the remaining genera are apparently united by a reduction in the tarsal organ pattern to 3-3-2-2 raised receptors on legs I–IV and by the acquisition of a clumped eye arrangement. Three subfamilies of oonopids are recognized: Orchestininae Chamberlin and Ivie (containing only Orchestina; Ferchestina Saaristo and Marusik is placed as a junior synonym of Orchestina), Sulsulinae, new subfamily (containing Sulsula, Xiombarg, Unicorn, and Cortestina), and Oonopinae Simon (containing all the remaining genera, including those previously placed in the Gamasomorphinae). The type species of Sulsula and Kapitia, S. pauper (O. P.-Cambridge) and K. obscura Forster, are redescribed, and the female of S. pauper is described for the first time. A new sulsuline genus, Dalmasula, is established for Sulsula parvimana Simon and four new species from Namibia and South Africa.


INTRODUCTION
Goblin spiders (the family Oonopidae) have long been among the most poorly known groups of spiders; the bulk of the species and much of the generic-level diversity of the family have remained undescribed, and the phylogenetic relationships of its members have been poorly understood, at all levels. Thanks to a Planetary Biodiversity Inventory (PBI) project, initiated in September 2006 with support from the U.S. National Science Foundation (NSF), knowledge of these animals has expanded rapidly; at present, the PBI project involves over 45 participants in more than a dozen countries, and almost one-third of the total project budget comes from sources other than NSF, in several nations. Through the efforts of these participants, enough information has now accumulated to allow testing some preliminary hypotheses about the higher-level relationships of oonopids. We present here results based on investigations of the tarsal organ morphology of a wide variety of oonopids and their outgroups.

HISTORICAL BACKGROUND: OUTGROUPS
As treated in the classical literature (e.g., Simon, 1893;Dalmas, 1916;Chickering, 1951;Forster, 1956;Hickman, 1979), oonopids were poorly delimited, and certainly not a monophy-letic group. Some of the major problems were solved by Forster and Platnick (1985), who used scanning electron microscopy of the tarsal organ (a chemosensory structure found near the tips of the legs and palps) to show that many of the austral genera previously assigned to the Oonopidae are actually more closely related to Orsolobus Simon (which was then placed in the Dysderidae) than they are to true oonopids. Forster and Platnick suggested that the monophyly of the superfamily Dysderoidea is supported by a peculiar specialization of the internal female genitalia (the development of a receptaculum associated with the posterior wall of the bursal cavity), and argued that four families of dysderoids should be recognized: the Dysderidae (primarily a Mediterranean group, but with one synanthropic, cosmopolitan species), Segestriidae (a worldwide group of three genera), Orsolobidae (a Gondwanan group, found in Australia, New Zealand, and southern South America, and subsequently discovered in southern Africa by Griswold and Platnick, 1987), and the Oonopidae. This grouping of families was also supported in more recent, matrix-based phylogenetic analyses by Platnick et al. (1991), which incorporated new data obtained by scanning electron microscopy of the spinneret spigots, and by Ramírez (2000), which added new data on respiratory system morphology.
The latter study placed the family Caponiidae as the sister group of dysderoids, based on the shared advancement of the posterior spiracles to a position just behind the epigastric furrow. Resolution within the Dysderoidea was not strongly supported in any of these studies; Platnick et al. (1991: 67) concluded that "familial relationships within the Dysderoidea (and the monophyly of the Oonopidae) remain uncertain" but favored a sister-group relationship between oonopids and orsolobids, and that sister-group relationship was also supported in the later analysis by Ramírez (2000).
More recently, Burger and Michalik (2010) presented the first evidence in support of oonopid monophyly, showing that (unlike all other spiders previously observed) males of a wide variety of oonopid genera have an unpaired, completely fused testis. The single orsolobid species they examined, in contrast, had the paired, unfused testes typical of most other spiders. Interestingly, some dysderids and segestriids have been reported to have partially fused testes, but similar structures also occur in the more distantly related family Scytodidae (Michalik, 2009).

HISTORICAL BACKGROUND: INGROUP
The traditional classification of oonopids stems from the treatment of the family in Simon's (1893) classic volume on the Histoire naturelle des araignées, where he recognized two informal groups, the "Oonopidae molles, " containing soft-bodied species in which the abdomen either lacks scuta entirely or has only a weakly sclerotized epigastric scutum, and the "Oonopidae loricatae, " containing hard-bodied species in which the abdomen has additional (and more heavily sclerotized) scuta. Simon intended these groupings only as artificial aids to identification; he explicitly stated (1893: 292) "Pour en faciliter l' étude, je répartis les Oonopides en deux sections, qui ne correspondent cependant pas à des groupes naturels. " Nevertheless, Petrunkevitch (1923) and subsequent workers recognized these groups formally, as the subfamilies Oonopinae and Gamasomorphinae, respectively. Neither Petrunkevitch nor the other workers who have used the names provided any phylogenetic justification for either of those subfamilies. . Tarsal organ, dorsal view, Ischnothyreus peltifer (Simon), female (16)(17)(18)(19)(20) and male (21)(22)(23)(24)(25), Segestria senoculata (Linnaeus), female (26)(27)(28)(29)(30) Two later papers also attempted to establish formal, subfamilial groupings. Chamberlin and Ivie (1942: 6) erected a monotypic subfamily, the Orchestininae, but provided no relevant evidence, indicating only that "The genus Orchestina is sufficiently distinct from the other genera of the Oonopidae to warrant its separation into a separate subfamily. " Their action seemingly ignored prior work, including that of Simon (1893: 292), who grouped Orchestina Simon with Sulsula Simon, and Dalmas (1916: 205), who added Calculus Purcell to this grouping, commenting that "Les trois genres Orchestina, Calculus et Sulsula sont les seuls de la famille offrant un groupe oculaire complêtement transverse. " Much later, Dumitresco and Georgesco (1983: 103, 114) attempted to establish a subfamily containing only the gamasomorphine genera Triaeris Simon and Ischnothyreus Simon, but as they did not designate a type genus for the group, and did not base its name on either of the included genera, their subfamilial name "Pseudogamasomorphinae" is not available.
Given Simon's intentions, it is hardly surprising that modern workers have found at least the Oonopinae to be paraphyletic. Platnick and Dupérré (2010a) noted that two putatively synapomorphic features, the acquisition of a clumped eye arrangement (rather than a transverse, H-shaped arrangement with a strongly recurved posterior row) and the loss of the heavily sclerotized, thick-walled sperm duct within the male palp, place Oonops Templeton as more closely related to the gamasomorphines than to some of the other genera currently placed as oonopines (including Orchestina and several other basal groups that retain the plesiomorphic states of these characters). Platnick and Dupérré (2010a: 6) also indicated that the limits of the Gamasomorphinae are unclear, and suggested that "gamasomorphy" be treated "as a syndrome of increasing sclerotization that starts, phylogenetically, with the cephalothorax. " Under that view, several genera placed as "molles" by Simon (1893) may be more closely related to the Gamasomorphinae than to Oonops. However, the monophyly of the classical Gamasomorphinae may be supported by at least one synapomorphic character, the presence of a sperm pore on the epigastric scutum of males.

TYPICAL OONOPID TARSAL ORGANS
Study of a wide variety of oonopid genera indicates that the tarsal organ morphology most commonly encountered within the family is that shown by its type species, Oonops pulcher Templeton. In the first comprehensive study of spider tarsal organs, Blumenthal (1935: 669) indicated that in all cases where he succeeded in locating the tarsal organ, the organ occurred on the tarsus of each leg and on the palpal tarsus, always with the same structure (although not always with the same size). In that regard, as is frequently the case, oonopids simply don't play by the same rules as other spiders. As shown here for O. pulcher (figs. 1-10), both sexes typically show serial dimorphism in their tarsal organ morphology; on the anterior legs, the tarsal organ has three raised receptors (figs. 1, 2, 6, 7), whereas on the posterior legs (and palps) the tarsal organ has only two receptors (figs. 3-5, 8-10). In O. pulcher and most other oonopids, the two most proximal receptors on the anterior legs are arranged transversely, whereas the two receptors found on the posterior legs and palps are arranged longitudinally.
In addition to this unusual anterior/posterior dimorphism, most oonopid tarsal organs have a distinctive longitudinal ridge that originates at the proximal end of the tarsal organ (figs. 1, 7, arrows). Such ridges have not been detected, to date, in the relevant outgroup families (Orsolobidae, Dysderidae, Segestriidae, and Caponiidae). We therefore hypothesize that both the anterior/posterior, serial dimorphism in raised receptor number and orientation, and the presence of the proximal, longitudinal ridge, are synapomorphic for the Oonopidae.
This range of taxa constitutes a reasonable sampling of oonopid diversity, particularly as studies in preparation show that tarsal organs of this type occur also in many taxa that are not yet revised or described, including members of the genera Gamasomorpha Karsch (Eichenberger et al., in press), Neoxyphinus Birabén (Abrahim et al., in press), Zyngoonops Benoit (Fannes, in prep.) (Brescovit et al., in press), southern South America (Grismado and Ramírez, in prep.), Madagascar (Álvarez-Padilla et al., in press), and Australia (Baehr et al., in press).
The largest gap in the sampling that has been done to date concerns the two genera, Triaeris and Ischnothyreus, that were treated by Dumitresco and Georgesco (1983) as members of their stillborn subfamily "Pseudogamasomorphinae. " We therefore present here scans of the tarsal organs of the type species of those genera, Triaeris stenaspis  and Ischnothyreus peltifer . Although the late Ray Forster doubted that Ischnothyreus is an oonopid, its type species has tarsal organs that are typical for the family. The longitudinal ridge is relatively short, but the same is true for many other typical oonopids (e.g., Platnick and Dupérré, 2011a: figs. 160-163).

OUTGROUP TARSAL ORGANS
The tarsal organs of the putative sister group of oonopids, the family Orsolobidae, have been documented in detail (see Forster and Platnick, 1985;Griswold and Platnick, 1987;Brescovit et al., 2004;Lise and Almeida, 2006;Baehr and Smith, 2008) because they constitute the best evidence for the monophyly of that family and provide many characters for grouping subsets of its members. Orsolobids resemble oonopids in many respects, but have distinctively elevated tarsal organs, usually accompanied by several cuticular lobes, that are unlike those of any other spiders studied to date (figs. 76-83). Interesting in that regard is one of the basal oonopid genera associated with Orchestina by Dalmas (1916), Calculus. The type (and only known) species of that genus, Calculus bicolor Purcell, is known only from juveniles from South Africa (Purcell, 1910); although they are poorly preserved, their tarsal organ morphology (figs. 84, 85) shows clearly that these juveniles are orsolobids rather than oonopids. Simon (1893: 294) identified similar juveniles from South Africa as members of Sulsula pauper (O. P.-Cambridge), a species otherwise then known only from Egypt. Dalmas (1916: 205) suggested that those South African juveniles actually belong to C. bicolor rather than S. pauper; we have scanned the tarsal organs of one of those juveniles, and can confirm that it does indeed belong to the Orsolobidae rather than the Oonopidae. On the basis of these results, we here transfer Calculus from the Oonopidae to the Orsolobidae.
The tarsal organs of the other dysderoid families, the Segestriidae and Dysderidae, are less well known. Blumenthal (1935) considered that all spider tarsal organs belong either to "der primitive Typus" or "die normale Form. " His studies of the primitive type were based primarily on two species of Segestria Latreille, and his distinction was maintained by Forster (1980), who substituted the more descriptive terms "exposed" and "capsulate" for Blumenthal's primitive and normal types, respectively. Forster and Platnick (1985: 219, figs. 958-962) provided characterizations of the tarsal organs for each of the dysderoid families, as well as a few scans of oonopid, dysderid, and segestriid tarsal organs. Here we present detailed sets of scans for several representative genera, including Segestria (figs. 26-35), Ariadna Audouin (figs. 36-45), Dysdera Latreille (figs. 46-55), Harpactea , and Harpactocrates . The primary relevance of these figures is to show that the typical oonopid tarsal organ morphology detailed above does not occur in those outgroup taxa.
The tarsal organs of the dysderids we have examined offer several features of potential phylogenetic interest. In Dysdera crocata C.L. Koch, there is no longitudinal ridge, and there seem to be two different types of receptors. One type resembles the raised receptors of oonopids, except that they each seem to have a tiny pore (figs. 46, 51-54); the second type are just small pores, surrounded by an elevated rim ( fig. 52). Because the raised receptors are relatively low and the small pores are easily occluded by dirt and debris, it is difficult to determine how many receptors are present on a given tarsal organ. For example, the tarsal organ of leg I seems to have four receptors in females (fig. 46), two raised and two rimmed, whereas in the male only one rimmed and two raised receptors are clearly visible on leg I ( fig. 51). There may be some serial dimorphism as well; females seem to have one, rather than two, rimmed receptors on legs III, IV, and the palps (figs. 48-50).
In Harpactea lepida (C.L. Koch), the tarsal organ of leg I seems to have two rimmed and two raised receptors in males ( fig. 61), and probably in females as well ( fig. 56). Interestingly, in this species both sexes have two raised receptors on legs I and II, but only one on legs III, IV, and the palps (figs. 56-65). It seems likely that this pattern of serial dimorphism is a parallelism with oonopids, but the apparent presence of two rimmed receptors on the female palpal tarsal organ ( fig. 60) suggests that it is also possible that one of the raised receptors has been transformed into a rimmed receptor, rather than lost. Our scans of Harpactocrates drassoides (Simon) are not clean enough to show the rimmed receptors, but they do indicate that a similar serial dimorphism occurs in that species, with two raised receptors on legs I and II, but only one on legs III, IV, and the palps (figs. 66-75).
Interestingly, our scans of Ariadna bicolor (Hentz) indicate that the rimmed receptor type also occurs in that species (figs. 36, 41), and the same may also be true for Segestria senoculata (Linnaeus) (see fig. 31) and Segestria florentina (Rossi) (see Giroti and Brescovit, 2011: figs. 11-14). The rimmed receptors may therefore prove to be a synapomorphy uniting Dysderidae plus Segestriidae, although a much broader survey will be required to test that conjecture. Similarly, more thorough sampling of tarsal organ morphology is also likely to prove useful for future work on orsolobids, as our scans suggest that serial dimorphism occurs in both sexes of at least one species (figs. 76-82).

BASAL OONOPID TARSAL ORGANS
Given that the groundplan for typical oonopids seems well supported by the observations listed above, and clearly different from that of the relevant outgroups, our attention focused on FIGURES 106-120. Tarsal organ, dorsal view, Cortestina thaleri Knoflach,, Xiombarg plaumanni Brignoli,, Unicorn catleyi Platnick and Brescovit, the types of tarsal organs found among the more basal oonopid genera (i.e., Orchestina and similar taxa). Orchestina is a highly diverse, nearly worldwide group of species united by the presence of enlarged femora on leg IV; the enlarged femora enable the animals to jump several times their body length. Its members dominate both the oonopid canopy fauna and the oonopid fossil record (including the oldest known oonopids, from Cretaceous amber, Saupe et al., in press;Marusik and Wunderlich, 2008). The tarsal organs of several undescribed species from South America, Africa, and Madagascar have been examined (see figs. 86-105); they resemble those of typical oonopids in having the longitudinal ridge originating from the proximal edge of the organ, and in showing serial dimorphism between the anterior and posterior legs, but differ in having one additional raised receptor (four on the anterior legs and three on the posterior legs), so that their receptor formula is 4-4-3-3, for legs I-IV, rather than 3-3-2-2 (as in typical oonopids). Forster (1980), based on an extensive survey including most spider families, identified a general trend in tarsal organ evolution: the transformation from an exposed to a capsulate structure. The basal oonopid genera other than Orchestina illustrate this trend well. In the genus Cortestina Knoflach, the tarsal organ is partly capsulate (figs. 106-110; Knoflach et al., 2009, figs. 50, 51). The receptor-bearing portion of the organ is sunken well below a pair of elevated lateral folds, but one can still see at least some of the raised receptors as well as the longitudinal ridge.
In the remaining basal genera, the tarsal organ has become fully capsulate; in dorsal view, the aperture of the tarsal organ is a tiny circle, and the receptors are situated too far below the opening to be visible in scanning electron microscopy. These fully capsulate structures occur in the North African genus Sulsula  and the South American genera Xiombarg  and Unicorn Platnick and Brescovit (figs. 116-120;Platnick and Brescovit, 1995: figs. 9, 10;González-Reyes et al., 2010: fig. 1c). They also occur in the new genus described below as Dalmasula, from Namibia and South Africa, and in an undescribed genus from Argentina (Izquierdo et al., in prep.). Forster's (1980) view that the transformation from exposed to capsulate tarsal organs has occurred repeatedly in spider evolution is born out also within the Oonopidae. Among the typical oonopids, the genus Escaphiella Platnick and Dupérré is notable for having an almost fully capsulate tarsal organ (see Platnick and Dupérré, 2009b: figs. 50-53, 100). The aperture, however, is quite different from that found in the basal genera, forming a long, narrow slit rather than a circle. The structure in Escaphiella is clearly modified from that found in the sister group of that genus, Scaphiella Simon, where the exposed tarsal organ has become elongated and narrowed (see Platnick and Dupérré, 2010a: figs. 510-514, 560-564). A similarly narrowed tarsal organ shape is found in another member of the Scaphiella complex, Pescennina Simon (see Platnick and Dupérré, 2011b: figs. 119-123, 164-168), suggesting that Pescennina may be the sister group of Scaphiella plus Escaphiella. A similarly slit-shaped opening occurs also in the more distantly related Grymeus robertsi Harvey (see Harvey, 1987: fig. 8), and the tarsal organs of Longoonops bicolor Platnick and Dupérré and L. padiscus (Chickering) are notably narrowed (see Platnick and Dupérré, 2010b: figs. 598, 599, 623-625, 670, 671, 692, 693), but all the transformations from exposed to capsulate tarsal organs detected to date within the typical oonopids are morphologically easily distinguishable from the character states found in Cortestina and the fully capsulate basal genera.

PHYLOGENETIC CONCLUSIONS
We hypothesize that the groundplan of oonopids includes an exposed tarsal organ like that of Orchestina, having a distinctive longitudinal ridge originating from the proximal end of the organ, and a serially dimorphic 4-4-3-3 pattern of raised receptors on legs I-IV. So far as we can tell, all oonopids retain the longitudinal ridge, but the ridge (like the receptor nodes) is not visible externally in those taxa that have acquired a fully capsulate tarsal organ morphology. The presence or absence of the ridge (and the number of receptors) in such taxa could be confirmed only by histological sectioning. The majority of oonopid species fit the trend identified within the Orsolobidae by Forster and Platnick (1985: 219), where the number of receptors becomes reduced (from five or six in some species of basal genera such as Tasmanoonops Hickman to only two or three). In the vast majority of oonopids, the tarsal organ is exposed and the receptor pattern is reduced from 4-4-3-3 to 3-3-2-2 (i.e., one raised receptor is lost on each tarsal organ).
The reduction to a 3-3-2-2 pattern of raised receptors occurs in the same set of taxa that show a clustered eye arrangement and have lost the heavily sclerotized, thick-walled sperm duct within the male palp, with a single known exception. The only described oonopid species from New Zealand, Kapitia obscura Forster, appears to retain the plesiomorphic 4-4-3-3 tarsal organ receptor pattern (figs. 319-322) and an H-shaped shaped eye arrangement (figs. 308, 310, 326), even though its male palp clearly lacks a sclerotized sperm duct (compare figs. 323, 324 with 127, 128, 209, 210). These characters suggest that Kapitia Forster is the sister group of the many genera with a 3-3-2-2 tarsal organ receptor pattern and a clustered eye arrangement, and all these taxa, including those previously placed in the Gamasomorphinae, are assigned below to the subfamily Oonopinae. A new subfamily is established below for those more basal genera that share a partly or fully capsulate tarsal organ; that subfamily is here named Sulsulinae, as Sulsula appears to be the oldest available name for any of its members. We thus recognize three subfamilies: the Orchestininae Chamberlin and Ivie, the Sulsulinae, and the Oonopinae Simon.
Although the Oonopinae are united by the "loss" of the heavily sclerotized, thick-walled sperm duct that is typical of other araneomorphs, the absence presumably reflects a major transformation in how sperm are stored within the palp, and possibly also changes in the physiological mechanisms involved in the induction of sperm into the palp, and the expulsion of sperm into the female genitalia.
The question of whether sulsulines are more closely related to oonopines or to Orchestina remains unanswered. Although the tarsal organ receptor patterns of the fully capsulate sulsuline genera are unknown, our scans of Cortestina (figs. 106-110) suggest that in that genus the number of receptors is greatly reduced, probably to 2-2-1-1, so we would expect the fully capsulate basal genera to have a similarly reduced receptor pattern. The apparent retention of a 4-4-3-3 pattern in Kapitia suggests, however, that the reduction in the number of tarsal organ receptors has followed different pathways within the sulsuline and oonopine lineages. Unfortunately, without studies of their ultrastructure or innervation, it doesn't seem possible to homologize individual receptors across the three subfamilies. Thus, other characters should be sought to resolve the basal three-taxon statement within the family. We are far from being able to present a comprehensive, quantitative analysis of oonopid interrelationships. Fully half of the family's generic-level diversity may still be undescribed. Although a few groups of putatively closely related genera have been recognized (e.g., the Scaphiella, Dysderina, Gamasomorpha, Pelicinus, and Stenoonops groups), we are not yet able to assign many of even the currently described genera to such groups, making the choice of exemplars to be included in any such analysis highly problematic. Although our descriptive database includes a substantial amount of character information on a substantial number of species, those characters were chosen primarily for their efficacy at the species level. Many other characters that are important for higher-level relationships will need to be added before such analyses will become realistic; female genitalic characters, for example, are largely uncoded in the descriptive database. Even though a detailed analysis would therefore be premature at this point, the hypotheses we have presented above can effectively be summarized in the following matrix: 1 tarsal organ with proximal longitudinal ridge 2 tarsal organ with raised receptors only, in serially dimorphic pattern (either 4-4-3-3 or a modified, reduced form of that pattern, i.e., 3-3-2-2 or 2-2-1-1) 3 femur IV enlarged 4 tarsal organ at least partly capsulate 5 male palp without heavily sclerotized sperm duct 6 tarsal organ receptor pattern reduced to 3-3-2-2 7 ocular group clumped Computer analysis is not needed to discern that these seven characters support the following groups, respectively: 1

SYSTEMATICS
Our methods follow those of Platnick andDupérré (2009a, 2009b); only differences from the males (beyond the obvious lack of male endite modifications) are mentioned in the descriptions of females. Scans were taken from uncoated right male palps, and the images were flipped for consistency. All measurements are in mm. High-resolution versions of the images, the geocoded locality data, and a distribution map for each species will be available on the goblin spider Planetary Biodiversity Inventory (PBI) project's website (http://research.amnh.org/ oonopidae). Oonopidae Simon, 1890: 80. Diagnosis: Oonopids resemble orsolobids but lack the elevated tarsal organs characteristic of that family (figs. 76-85) and have instead a flat, exposed, or capsulate tarsal organ with a distinctive longitudinal ridge originating at the proximal end of the organ (figs. 1, 7, arrows), raised receptors only, and dimorphism between the anterior and posterior legs, with legs I and II having one more receptor than do legs III, IV, and the palpal tarsi (figs. 1-25). Males of representative studied species have a single, fused testis (Burger and Michalik, 2010); females lack a claw on the palpal tarsus (the claw is typically retained in orsolobids).

Orchestininae Chamberlin and Ivie
Orchestininae Chamberlin and Ivie, 1942: 6. Diagnosis: Orchestinines are easily recognized by their enlarged femora IV. So far as is known, their 4-4-3-3 tarsal organ receptor pattern is shared only with the basal oonopine genus Kapitia. In Orchestina, however, the tarsal organs are typically narrowed at the proximal end, producing a neck-shaped appearance (figs. 86-105) that does not occur in Kapitia. Most species of Orchestina have the posterior median eyes advanced to form a straight row with the anterior lateral eyes (much as in Segestria), but at least some undescribed African species have the posterior median eyes situated more posteriorly, in the more H-shaped pattern shared by the sulsulines and Kapitia.
Included Genera: Only Orchestina Simon; 51 Recent species have been described to date (Platnick, 2012), but as many additional species have already been identified by Matias Izquierdo, Arnaud Henrard, and Natalia Chousou-Polydouri. Saaristo and Marusik (2004) established a monotypic genus, Ferchestina, for a single species from the Russian Far East. Those authors noted (2004: 51) that five of the oonopid genera then recognized are each widely distributed and together constitute more than half of the family's then known species diversity, and somehow concluded from those observations that "it is quite safe to postulate that all these five genera are more or less polyphyletic. " The argument is nonsensical, as neither the number of included species, nor how widely they occur, are evidence of monophyly or polyphyly; only synapomorphic characters, and their distribution among taxa, are relevant to that question. Saaristo and Marusik (2004: 51) noted that Orchestina, in particular, does have at least one putative synapomorphy: "A single key-character used to separate members of Orchestina from other non-scutate oonopids is the markedly swollen tibia [sic, lapsus for femur] IV. " In our view, there are no convincing characters supporting the placement of Ferchestina storozhenkoi Saaristo and Marusik in a genus separate from Orchestina; that type species clearly shares the primary synapomorphy of Orchestina, the enlarged femur IV. Saaristo and Marusik provided no characters suggesting that their species represents the sister group of all other orchestinines (i.e., that all the other Orchestina species form a monophyletic group that excludes Ferchestina). The differences they cited in their generic diagnosis, such as the prominent humps on the male carapace, the projection at the tip of the cheliceral paturon, and the details of the male and female genitalia, are all presumably just species-level autapomorphies, and are therefore irrelevant to the question of what the closest relative(s) of the species may be. Saaristo and Marusik commented that compared to Orchestina pavesii (Simon), the type species, femur IV is "not so thick" but provided no illustration of that supposed difference. The thickness of the fourth femur varies among the many species of Orchestina, possibly in allometric correlation with the total size, which also varies significantly (with some species attaining almost twice the total length of others). The diagnostic character is that the fourth femora are much thicker than femora I-III of the same specimen, not that they are of any particular given thickness (indeed, proportionately much less emphatically enlarged femora IV are also found in some soft-bodied, Neotropical oonopine species currently misplaced in Oonops). Both the male and female genitalia of Ferchestina fit well within the extensive range of variation shown by Orchestina species. We conclude that Ferchestina represents just a highly autapomorphic species of Orchestina, and that its recognition as a separate genus is positively misleading phylogenetically; we therefore place the name as a junior synonym of Orchestina (NEW SYNONYMY).
Although our conclusion is that the recognition of Ferchestina, as currently constituted, renders Orchestina a paraphyletic group and is therefore unacceptable, it is of course possible that future phylogenetic analyses will be able to discern monophyletic subgroups of orchestinines that are each supported by putatively synapomorphic characters. If, at that time, the type species of Ferchestina can be shown to belong to a subgroup that does not also include the type species of Orchestina, then it may be possible to resurrect Ferchestina as a usable name, but it would have to be on the basis of new evidence, not the insufficient data provided by Saaristo and Marusik (2004). Sulsulinae Platnick, new subfamily Type Genus: Sulsula Simon (1882). Diagnosis: This subfamily includes taxa that resemble Orchestina in having a transverse, unclumped eye arrangement and a heavily sclerotized sperm duct within the male palp, but have partly or fully capsulate tarsal organs and a normal, rather than expanded, femur IV.
Included Genera: Sulsula Simon (1882) from North Africa; Xiombarg Brignoli (1979) from Brazil and Argentina; Unicorn Platnick and Brescovit (1995) from Chile, Bolivia, and Argentina; Cortestina Knoflach (see Knoflach et al., 2009) from Austria and Italy, plus two new genera: Dalmasula, described below from Namibia and South Africa, and an undescribed genus from Argentina (Izquierdo et al., in prep.). Based on its partly capsulate tarsal organs, the Laurasian genus Cortestina probably represents the sister group of the other, fully capsulate, Gondwanan genera. The monophyly of the fully capsulate genera may also be supported by the presence of a single row of teeth on the tarsal claws; most other oonopids, and orsolobids, have two rows of teeth on each claw. However, there are typical oonopids in which one of the tooth rows has been lost in males (as in Heteroonops Dalmas; see Platnick and Dupérré. 2009c) or in both sexes (as in Birabenella Grismado; see Grismado, 2010).
Salsula: Simon, 1893: 294. Note: Simon (1893) regarded the original spelling of the genus as a printer's error, but the spelling occurs twice in Simon (1882) and his alternate spelling was rejected by Dalmas (1916: 204) and subsequent authors.
Diagnosis: With the characters of the genus, a male embolus with a relatively short, terminally curved tip (figs. 127-130), and female genitalia with a tubular, sclerotized anterior receptaculum (figs. 137-142) and an oval, unsclerotized posterior receptaculum (collapsed in those figures).
Variation: It is possible that more than one species is represented, but the few available specimens do not provide sufficient evidence to substantiate the description of additional species at this time. Females are available from three sites, but the one from Sudan (figs. 141, 142) has genitalia that seem more similar to that of a female from Algeria (figs. 137, 138) than to the geographically much closer one from Egypt (figs. 139,140). Only two males are available, from Algeria and Egypt. Under a dissecting microscope, the Algerian male appears to have a bulb that extends farther toward the palpal patella (figs. 122-125), but under a compound microscope, that difference is not obvious (figs. 127-130; unfortunately, the bulb of the Egyptian male collapsed in clove oil under the compound microscope). The embolus shows differences under the compound microscope, but at least some of those differences reflect different positioning of the palps and consequent foreshortening in the photographs of the Algerian male.
Synonymy: It appears that Simon (1910Simon ( , 1911 was able to borrow the holotype of Oonops pauper, compared it directly to his material of Sulsula longipes, and concluded that the specimens are conspecific. Although the holotype of O. pauper is a juvenile, there is no evidence that disputes Simon's conclusion. The specimens came from the same area (Ramleh is one of the beaches of Alexandria) and are clearly congeneric; it is unlikely that multiple species of Sulsula occur within Alexandria. Dalmasula Platnick, Szüts, and Ubick, new genus Type Species: Dalmasula lorelei, new species. Etymology: The generic name is a contraction of "Dalmas' Sulsula" and is feminine in gender. It honors Raymond de Dalmas and his pioneering study of Orchestina, in which (after discussing the similarities of Sulsula and Calculus with that genus) he commented (1916: 205) that: "On peut ajouter que Sulsula parvimanus E. Simon (1910: 178), décrit du pays des Namaquas, dans le Sud-Ouest Africain, et dont le type unique est en Allemagne, deviendra peut-être le type d'un quatrième genre de cette série. " The existence of this genus in South Africa was discovered independently by Charles Griswold (in Platnick and Brescovit, 1995: 6).
Diagnosis: Members of this genus resemble those of Sulsula in having a globose abdomen ( fig. 173), but can be distinguished by the presence of a promarginal cheliceral tooth ( fig. 176) and a very wide, hirsute colulus ( fig. 274), and the absence of a distal, semicircular groove and laterally directed ridges on the tarsal organ (figs. 199-202, 237-241). Males typically have both an embolus and a conductor (figs. 275, 277), although they appear to have fused in D. tsumkwe, new species (figs. 254, 255); the female genital area has peculiar, distinctive anterior sclerotizations that probably function as coupling ridges (figs. 215, 259-261, 293, 306). Males also have unusual modifications of the mouthparts; in those of D. lorelei, D. parvimana, and D. griswoldi, the base of the endites bears a triangular projection directed toward the chelicerae (fig. 175). Males of D. tsumkwe apparently lack those projections, but have a similar spur situated distally on the cheliceral paturon (figs. 250, 251).

Dalmasula lorelei
Etymology: The specific name is a noun in apposition taken from the type locality. Diagnosis: Males differ from those of D. parvimana in having a longer embolus (figs. 185, 186), from those of D. tsumkwe in lacking cheliceral apophyses and having an unsclerotized palpal conductor (figs. 209, 210), and from those of D. griswoldi in having a narrow embolus (figs. 185, 186); females have much larger ridges at the anterior end of the genital area ( fig. 215) than do those of the other known females.
Other Material Examined: None. Distribution: Namibia (Otjozondjupa). Etymology: The specific name is a patronym in honor of Charles Griswold, who first discovered the South African members of this genus.

Dalmasula griswoldi
Diagnosis: Males differ from those of the other Dalmasula species in having a more complex embolar region with a dorsal lobe terminating in two ribbonlike lamellae and a broad ventral lobe terminating in a slender spiral prong (figs. 275-288); females differ in having an epigynum with the anteromedian coupling ridges C-shaped and widely separated, and an internal median process with a slender stalk and a rounded head (figs. 293-299).
Other Etymology: The specific name is a patronym in honor of the Belgian arachnologist Domir De Bakker, in recognition of his assistance at the MRAC throughout the visit there during which Tamás Szüts found the specimen described here, formed by combining the first two letters of each of his names.
Diagnosis: Females differ from those of the other Dalmasula species in having an epigynum with the anteromedian coupling ridges straight and posteriorly contiguous, and an internal median process with a thick stalk and an angular head (figs. 306, 307).
Other Material Examined: None. Distribution: South Africa (Northern Cape).
Relationships: So far as is known, a 4-4-3-3 tarsal organ receptor pattern and an H-shaped eye arrangement are retained only in Kapitia, suggesting that this enigmatic, seldom collected New Zealand genus is the sister group of all the other oonopines. Aside from the original description by Forster (1956), information on Kapitia has been supplied only by Paquin et al. (2010), so we present below a redescription of the species. Of particular interest are the teeth on the tarsal claws (figs. 316, 317); it appears that the inner tooth row has been displaced entirely to the tip of the claw, presumably representing a stage in the loss of that tooth row.
Within the massive assemblage of remaining oonopine species united by the reduction to a 3-3-2-2 tarsal organ receptor pattern and a clustered eye arrangement, we can recognize only a few large groupings at this point. Platnick and Dupérré (2010b) suggested that those genera with a distinctly sclerotized cephalothorax might form a monophyletic group; if so, then several genera that were classically placed in the Oonopinae are actually more closely related to the classical gamasomorphines than to Oonops and such similarly soft-bodied taxa as Heteroonops and Oonopoides Bryant. These anteriorly hard-bodied genera include at least Stenoonops Simon, Australoonops Hewitt, Scaphioides Bryant, Khamisia Saaristo and van Harten, and Longoonops Platnick and Dupérré. The classical gamasomorphines might also represent a monophyletic subgroup of this enlarged group. It remains to be seen, for example, whether taxa in the Scaphiella complex, where males have dorsal abdominal scuta that are lacking in females, or the similarly sexually dimorphic taxa in the Dysderina complex, represent independent gains of dorsal scuta in males, or independent losses of dorsal scuta in females. However, even if the presence of a dorsal scutum does not turn out to be a synapomorphy of the classical gamasomorphines, the movement of the male gonopore onto the epigastric scutum may well be synapomorphic for that group. Nevertheless, neither the classical nor the enlarged group could be recognized as a subfamily unless a separate subfamily were to be established for Kapitia and the many other genera that (like Oonops) have both a soft-bodied cephalothorax and a soft-bodied abdomen can also be shown to constitute a monophyletic group (i.e., a smaller Oonopinae). At present, we know of no potentially synapomorphic characters supporting that smaller group. Kapitia Forster, 1956: 166 (type species by original designation Kapitia obscura Forster).

Kapitia Forster
Diagnosis: The lack of a heavily sclerotized sperm duct within the male palp, combined with the presence of a 4-4-3-3 tarsal organ receptor pattern and an H-shaped eye arrangement, is diagnostic for the genus.  . 309), anteriorly narrowed to 0.49 times its maximum width or less, with rounded posterolateral corners, posterolateral edge without pits, posterior margin not bulging below posterior rim, anterolateral corners without extension or projections, posterolateral surface without spikes, surface of elevated portion of pars cephalica smooth, sides smooth, pars thoracica without depressions, fovea absent, without radiating rows of pits; lateral margin straight, smooth, without denticles; plumose setae near posterior margin of pars thoracica absent ( fig. 308). Clypeus margin unmodified, straight in front view, vertical in lateral view, low, ALE separated from edge of carapace by less than their radius, median projection absent. Eyes six, well developed, all subequal, ALE oval, PME oval, PLE circular; posterior eye row recurved from above; ALE separated by their radius to diameter, ALE-PLE touching, PME touching throughout most of their length, PLE-PME touching ( fig. 310). Sternum longer than wide, white, uniform in coloration, not fused to carapace, median concavity absent, without radial furrows between coxae I-II, II-III, III-IV, radial furrow opposite coxae III absent, surface smooth, without pits, microsculpture absent, sickle-shaped structures absent, anterior margin unmodified, posterior margin extending posteriorly beyond anterior edges of coxae IV as single extension, anterior corner unmodified, lateral margin without infracoxal grooves, distance between coxae approximately equal, extensions of precoxal triangles absent, lateral margins unmodified, without posterior hump; setae sparse, light, needlelike, evenly scattered, originating from surface, without hair tufts ( fig. 311). Mouthparts white. Chelicerae slightly divergent, without teeth, anterior face unmodified; fangs without toothlike projections, directed medially, shape normal, without prominent basal process, tip unmodified; setae light, needlelike, evenly scattered. Labium rectangular, not fused to sternum, anterior margin indented at middle, same as sternum in sclerotization; with one or two setae on anterior margin, subdistal portion with unmodified setae (fig. 312). Endites distally not excavated, serrula present as single row of teeth, anteromedian tip unmodified, posteromedian part unmodified, same as sternum in sclerotization. Abdomen: Ovoid, without long posterior extension, rounded posteriorly, interscutal membrane rows of small sclerotized platelets absent posteriorly; dorsum soft portions yellow, without color pattern; book lung covers small; pedicel tube short, unmodified, scutopedicel region unmodified, plumose hairs absent, matted setae on anterior ventral abdomen in pedicel area absent, cuticular outgrowths near pedicel absent; dorsal, epigastric, postepigastric and spinneret scuta absent; setae light, needlelike, epigastric area setae not enlarged at base; dense patch of setae anterior to spinnerets absent. Colulus absent. Legs: Yellow, without color