The Andean Goblin Spiders of the New Genus Scaphidysderina (Araneae, Oonopidae), with Notes on Dysderina

ABSTRACT Dysderina Simon is one of the largest of the classical genera of goblin spiders, containing numerous species that have been associated only because they are heavily scutate gamasomorphines with long, paired spines on the ventral surface of the anterior tibiae and metatarsi. The Old World species that have been assigned to the genus are wildly misplaced, and the New World fauna constitutes a complex of over 225 species belonging to at least nine genera. The northern Andes house a highly diverse fauna, both of Dysderina itself and of closely related genera. The new genus Scaphidysderina is established for one of those related but distinct Andean groups, characterized by a crenulated sternum and by the reduction or loss of the dorsal abdominal scutum in females. Seventeen new species are described from Peru (S. manu, S. pagoreni, S. scutata, S. cajamarca), Ecuador (S. tayos, S. loja, S. molleturo, S. tapiai, S. pinocchio, S. palenque, S. tandapi, S. napo, S. baerti, S. cotopaxi, S. andersoni), and Colombia (S. hormigai, S. iguaque). Males of several species show remarkable modifications of the chilum and chelicerae; the chilum is sometimes enlarged to form a conspicuous snout, and the cheli cerae often bear a heavily sclerotized, dorsally directed spine, A second new genus, Costarina, is established to contain the most commonly encountered species that have been misplaced in Dysderina; Dysderina plena O. P.-Cambridge, from Mexico, is chosen as the type species, and 1.5 additional taxa, all described from Central America by Chickering, are transferred from Dysderina to Costarina: D. abdita, D. belinda, D. concinna, D. dura, D. humphreyi, D. improvisa, D. intempina, D. meridina, D. obtina, D. poteria, D. recondita, D. riigida, D. seclusa, D. silvatica, and D. watina.


INTRODUCTION
The goblin spider genus Dysderina Simon (1891) is one of the largest of the classical oonopid genera, containing heavily scutate species notable for having several pairs of very long spines under the anterior tibiae and metatarsi (as in figs. 155, 156, 193, 194). The group currently includes 44 species (Platnick, 2011) from far-flung regions: the Neotropics, Africa, the Philippines, and the Caroline Islands. This wide distribution, however, is merely an artifact. Examination of the types of most of the described species, and much additional material, indicates that there is a distinctive complex of genera closely allied to Dysderina, but that the Old World taxa are wildly misplaced, and are not even members of that complex, much less of Dysderina itself.
Members of the Dysderina complex share characteristic male genitalia. The palpal bulb is inflated and is usually completely fused to the cymbium, with no trace of a seam (figs. 16, 18, 145, 147); the embolus originates subdistally and is often elaborate in shape (figs. 22-24). The complex appears to include more than 225 species, belonging to at least nine genera, found from Mexico and the West Indies south to Argentina.
Dysderina has been misconstrued throughout its history. Simon (1891) designated Oonops principalis Keyserling (1881), described from Colombia (as "Neu-Granada"), as the type species. In that paper, Simon identified specimens from Saint vincent and from venezuela as belonging to D. principalis, but those specimens are all misidentified (i.e., none of them actually belong to Keyserling's species). For example, Simon's venezuelan specimens of "D. principalis" include representatives of four species, belonging to two different genera, and not a single one of those four species is actually congeneric (much less conspecific) with D. principalis.
Under Article 70.3 of the International Code of Zoological Nomenclature, authors who discover such misidentifications of type species have the option of selecting either the nominal species, or one of those misidentified as that species by the original author of the genus, to be the type. In this case, the problem was addressed first (albeit only implicitly) by Chickering (1968), who recognized that Simon's specimens from Saint vincent do not belong to D. principalis, and described them as a new species, D. soltina. We initially viewed this choice as totally unproblematic, because D. principalis, D. soltina, and most of the other species assigned to Dysderina by Chickering (1968) share a distinctive sternal morphology; in both sexes, the sternum typically bears three conspicuous, elevated, transverse ridges ( fig. 3).
However, although the sternal ridges are presumably synapomorphic (as they are unknown in other spiders), further study of the substantial available collections indicates that the taxa that have such ridges constitute more than one genus. The holotype (and only known specimen) of D. principalis has a distinct groove connecting the two anterior spiracles, and a second distinct groove connecting the two posterior spiracles ( fig. 5; because D. principalis is poorly known, we present here several images of its holotype, as figs. 1-12). There are many species that resemble D. principalis in both sternal and spiracular characters, and also share with that species a characteristic form of the male palp. We therefore consider Dysderina to include only those species that share the sternal, spiracular, and palpal conformation of D. principalis. Under this interpretation, many species currently placed in Dysderina, including some of those most commonly encountered in collections, will have to be transferred to other genera (generic names are already available for some of these misplaced species).
If one were to depart from Chickering's choice of D. principalis as the type species, and regard instead as the type species one of those misidentified as D. principalis by Simon (1891), little nomenclatorial stability could be gained, as no more than six of the specific names considered by Chickering (1968) to belong to Dysderina would remain in the genus, no matter which of Simon's misidentified species was chosen. Moreover, there is already a generic name available for those of Simon's misidentified taxa that do have the sternal ridges (and are therefore closest to the traditional concept of Dysderina); those species actually belong to Simonoonops Harvey (2002), a genus that occurs in venezuela and the Lesser Antilles.
We therefore see no reason to depart from Simon's original designation (and Chickering's subsequent treatment) of D. principalis (Keyserling) as the type species. This means, however, that the large group of Mexican and Central American species that have been assigned to Dysderina in the past are among the taxa that have been misplaced. Because these taxa represent a majority of the specimens that have been identified as Dysderina in collections, we establish below a new genus, Costarina, to contain this assemblage.
The northern Andes house a highly diverse fauna, both of Dysderina itself and of other Dysderina-like taxa; some of these genera seem to be Andean endemics, but others may occur also in venezuela, Bolivia, and/or Brazil. Many of the Andean species are notable because the females lack the dorsal abdominal scutum that is found in the males of those species and in both sexes of all other members of the complex. In this respect, they resemble the species of Scaphiella Simon (1891) and its relatives.
The present paper deals with one of these groups, here described as Scaphidysderina. Its members are united by sternal morphology; in both sexes, the surface of the sternum lacks the transverse ridges found in Dysderina and its closest relatives, but is highly crenulated (figs. 15, 116). In some species, the postepigastric scutum of females extends up the sides of the abdomen (figs. 69, 233), producing a taco-shaped appearance much like that of female Scaphiella.
The 17 species here assigned to Scaphidysderina all appear to be undescribed. Aside from D. principalis, there are six other species from the Andean countries that are currently placed in Dysderina. Three of them are from Colombia: D. globosa (Keyserling, 1877), D. propinqua (Keyserling, 1881), and D. similis (Keyserling, 1881); their types have been examined, and do not belong to Scaphidysderina. Keyserling also described three species from Peru: D. desultrix (Keyserling, 1881), D. machinator (Keyserling, 1881), and D. montana (Keyserling, 1883). The types of those three species were supposed to be deposited in Warsaw but are not currently in that collection (Dominika Mierzwa, in litt.) or with the Keyserling material in London, and are either lost or destroyed. Of those taxa, D. machinator was based on a male; Keyserling's illustration of the palp shows a separate palpal bulb and cymbium, and the species is clearly not a member of the Dysderina complex. Keyserling's figure of D. montana suggests that this female has a small postepigastric scutum confined to the spiracular area, indicating that this species belongs to one of the other Andean genera, rather than to Scaphidysderina. The description of D. desultrix indicates that the female has a dorsal scutum, and that the sternum has lateral elevations, again suggesting that it belongs to a different genus.
Among the Scaphidysderina species known from females, only one, S. scutata from Peru, has a dorsal scutum on the abdomen. That scutum ( fig. 52) is much smaller in size, and much more weakly sclerotized, than those found in females of all the described members of the Dysderina complex, suggesting that the scutum has been lost independently within Scaphidysderina, and that S. scutata may be the sister group of all the remaining species.
Among those remaining species, several closely resemble the female of S. scutata in sternal morphology, with the surface of the sternum highly and irregularly crenulated (figs. 15, 39). However, in a group of 10 species (S. cajamarca, S. tayos, S. molleturo, S. tapiai, S. pinocchio, S. palenque, S. tandapi, S. napo, S. andersoni, and S. hormigai), the raised portions of the sternal crenulations are thickened, forming a system of anastomosing ridges (figs. 126, 139, 175). Two additional groups can be recognized within that distinctive cluster. Two of the three species that are known only from females (S. tayos and S. napo) share a pair of enlarged tubercles situated at the rear of the pars cephalica (figs. 67, 231). These tubercles resemble the carapace spikes found in some of the species of another member of the Dysderina complex, the genus Neoxyphinus Birabén (1953). We presume that the enlarged tubercles will also occur in the males of these two species, but they are not known, in either sex, in any of the other species treated below. Those two species, S. tayos and S. napo, are also united by having the largest, most Scaphiella-like postepigastric scuta within the genus (figs. 69, 233), and by their long, narrow, and distally expanded anterior genitalic process (figs. 73, 237).
Four of the other species with anastomosing sternal ridges (all but S. cajamarca, S. molleturo, S. andersoni, and S. hormigai) share the most elaborate modifications of the cephalothorax within the genus. Scaphidysderina species all seem to have the chilum fused to the clypeus (figs. 58, 173). In males of these four species, the chilum is extraordinarily enlarged, forming a snout that projects over the base of chelicerae (figs. 91, 92). The monophyly of this group of snouted species (S. tapiai, S. pinocchio, S. palenque, and S. tandapi) is also supported by a female genitalic character, the presence of thickened, heavily sclerotized apodemes with anteriorly directed tips (figs. 105, 111, 135, 223).
The males of the snouted species also have elaborately modified chelicerae. The anterior surface of the paturon bears a deep excavation, situated medially; on the inner margin of the excavation, there is a highly sclerotized spine that projects back toward the clypeus (figs. 125, 140). The males of S. cajamarca and S. andersoni also have such a spine on at least one chelicera (figs. 58, 290), and the male of S. hormigai has even more elaborately modified chelicerae (figs. 269, 272); similar modifications might also occur, of course, in the unknown males of S. tayos, S. napo, and S. molleturo.
Interestingly, the males of S. baerti and S. cotopaxi also have the cheliceral spines (figs. 239, 241), even though they lack the anastomosing sternal ridges; those two species plus S. iguaque are united by the presence of prominent, tuberculate setal bases on the sternum (figs. 240, 249, 281) and may thus together represent the sister group of the 10 species with anastomosing sternal ridges. Of the remaining species, S. manu, S. pagoreni, and S. loja have the postepigastric scutum of females entirely fused with the epigastric scutum (just as in all the known males), and share a long, narrow, and distally unexpanded anterior genitalic process (figs. 30, 36, 81). They may together represent the sister group of all the species other than S. scutata (which shows only a slight separation of the postepigastric scutum at its far lateral edge, but is presumably the most basal member of the genus, since it still retains a reduced dorsal abdominal scutum in females).
Our methods follow those of Platnick andDupérré (2009a, 2009b); the species are treated geographically, beginning in southern Peru and proceeding northward. Only differences from the males (beyond the obvious lack of male cheliceral and 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). Etymology: The generic name is a contraction of "Scaphiella-like Dysderina" and is feminine in gender.

COLLECTIONS EXAMINED AMNH
Diagnosis: Members of this genus can easily be recognized by their highly crenulated sternum (figs. 15, 68). They also differ from those of Dysderina in lacking both a spinneret scutum and a groove connecting either the anterior or posterior spiracles (compare figs. 5, 27).
Distribution: Known only from the Andean nations (Peru, Ecuador, and Colombia); found on both slopes of the Andes, at elevations up to 3865 m. Etymology: The specific name is a noun in apposition taken from the type locality. Diagnosis: This seems to be the sister species of S. pagoreni; males share an intricately spined embolus, but those of S. manu have the most ventral of the distal embolar processes both longer and narrower ( fig. 22) than in S. pagoreni; females also have similarly constructed genitalia, but in S. manu the anterior process is wider ( fig. 30)    Etymology: The specific name is a noun in apposition taken from one of the localities at which the species has been collected.
Diagnosis: Males closely resemble those of S. manu but have the most ventral of the distal embolar processes both shorter and wider ( fig. 45); females also closely resemble those of S. manu, but the anterior genitalic process is narrower ( fig. 36).
Male (PBI_OON 525,: Total length 2.00. Chilum triangular, seam present, extending distal about equal to clypeus height. Sternum anterior margin with interrupted transverse groove. Anterior face of paturon with produced ledge, median and lateral edges of ledge forming tubercles. Endites with tip of dorsal projection directed toward ventral projection. Dorsal scutum covering full length of abdomen, no soft tissue visible from above; postepigastric scutum reaching near full length of abdomen. Leg spination: femur I p0-0-1; tibiae: I v4-4-1p; II v2-4-0; metatarsi: I v2-2-1p; II v2-2-0. Embolus originating on strong lobe of cymbium, consisting of narrow basal stalk followed by highly folded extension, dorsal and ventral flanges short, almost touching at tip.  Etymology: The specific name is a noun in apposition taken from the type locality. Diagnosis: Males have anastomosing ridges on the sternum ( fig. 60) but differ from the other known males with such ridges by lacking a greatly elongated chilum and having only a simple, dorsally directed spine on the chelicerae; the rounded dorsal tip of the ventral embolar flange ( fig. 62) is diagnostic.
Other Material Examined: One male taken with the types (QCAZ). Distribution: Ecuador (Azuay). Etymology: The specific name is a noun in apposition taken from the fictional character created by Carlo Collodi, referring to the extraordinary snout at the front of the carapace.

Scaphidysderina pinocchio, new species
Diagnosis: Males resemble those of the other three species with elongated snouts, but can be distinguished by the sigmoid dorsal and ventral embolar flanges (figs. 117, 121); females resemble those of the same species but can be distinguished by the more rectangular anterior genitalic elements (fig. 111).
Other Etymology: The specific name is a noun in apposition taken from the type locality. Diagnosis: Males resemble those of S. tandapi in having a basal process on the dorsal flange of the embolus ( fig. 128, 146), but have the embolar tip shorter and thicker (best seen in lateral view, figs. 127, 129); females have a dorsally directed tip on the anterior genitalic process ( fig. 135).
Etymology: The specific name is a noun in apposition taken from the type locality. Diagnosis: Males are unknown; females resemble those of S. tayos in having a pair of enlarged tubercles at the posterior margin of the pars cephalica ( fig. 231) but can be distinguished by the larger postepigastric scutum, which reaches almost to the top of the abdominal dorsum ( fig. 233). Male: Unknown. Female (PBI_OON 560,: Total length 2.70. Chilum without seam, small, rectangular, extending distance equal to about one-fourth of clypeus height. Sternum anterior margin with interrupted transverse groove, surface depressions separated by anastomosing ridges. Anterior face of paturon with produced ledge, both edges of ledge rounded. Postepigastric scutum covering nearly full length of abdomen, not fused to epigastric scutum, extending up sides of abdomen almost to top. Leg spination: femur I p0-0-2; tibiae I, II v4-4-2; metatarsi: I v2-2-2; II v2-2-1p. Genitalia without conspicuous, unsclerotized atrium, anterior genitalic process with diamond-shaped tip. Other Material Examined: None. Distribution: Ecuador (Napo). Etymology: The specific name is a patronym in honor of Léon Baert (KBIN), who sorted the type specimens and made them available for study.

Scaphidysderina baerti, new species
Diagnosis: Males and females resemble those of S. cotopaxi in having enlarged, tuberculate setal bases on the sides of the sternum ( fig. 240), but can be distinguished by the much narrower embolus of males (figs. 243, 245) and the much shorter apodemes of females ( fig. 229).
Diagnosis: Males resemble those of S. cajamarca but have a distally prolonged embolus tip (figs. 291, 293); females can be recognized by the bell-shaped sclerotization at the anteromedian edge of the genital atrium (figs. 299, 300).
Other Etymology: The specific name is a patronym in honor of Gustavo Hormiga (George Washington University), one of the collectors of the types.
Diagnosis: Males can easily be recognized by the large median lobes on their chelicerae (figs. 269, 272) and the distally expanded embolus (figs. 274, 276), females by the winglike basal extensions on the anterior genitalic process ( fig. 267).
Female (PBI_OON 561,: Total length 2.50. Chilum extending distance equal to about one-fourth of clypeus height. Anterior face of paturon with produced ledge, inner edge of ledge forming tubercle. Postepigastric scutum covering about 1/3 of abdomen length, not fused to epigastric scutum. Leg spination: femur I p0-0-2; tibiae: I v4-4-2; II v3-4-1p; metatarsi: I v3-2-2; II v2-1p-2. Genitalia with atrium reduced to short slit, anterior genitalic process shaped like inverted T. Other Material Examined: Three males and four females taken with the types (ICN). Distribution: Southern Colombia. Etymology: The specific name is a noun in apposition taken from the type locality. Diagnosis: Males resemble those of S. baerti and S. cotopaxi in having prominent, tuberculate bases on the lateral sternal setae, and the embolus is close to that of S. cotopaxi, differing in having a smooth, unserrated ventral edge on the ventral flange (figs. 283, 285).
Costarina, new genus Type Species: Dysderina plena O. P. -Cambridge (1894). Etymology: The generic name is a contraction of "Costa Rican Dysderina, " referring to the extraordinary radiation of species occurring in Costa Rica, and is feminine in gender.
Diagnosis: Members of this genus resemble those of Dysderina and Simonoonops in having three transverse ridges on the sternum, but differ in lacking grooves connecting either the anterior or posterior pairs of spiracles.

ACKNOWLEDGMENTS
This study is part of the oonopid PBI project supported by the U.S. National Science Foundation (grant DEB-0613754) and organizations in several other countries. The assistance of the many participants in that project is immensely appreciated. As always, we thank the many curators of collections that have supplied specimens: Léon Baert (KBIN), Janet Beccaloni ( and Diana Silva (MUSM). We also thank Matthias Burger for translations of Keyserling's descriptions, Steve Thurston for composing the plates, and Antonio Brescovit and Cristina Rheims for their careful reviews of the draft manuscript.