The short-legged Andean cosmetids revisited: the genus Libitia Simon, 1879 with description of two new species (Opiliones, Cosmetidae)

The old genus Libitia Simon, 1879 of small Andean harvestmen is revisited. The monotypic genus Libitiella Roewer, 1947 is herein considered a junior subjective synonym of Libitia. Accordingly, Libitiella bipunctata (Sørensen, 1932) is restored to the combination Libitia bipunctata. The species Libitia cordata and Libitia bipunctata comb. nov. are redescribed and the new species Libitia gandalf sp. nov. and Libitia iguaque sp. nov. are herein described from Cordillera Oriental of Colombian Andes (Chingaza Natural Park and Iguaque Natural Park, respectively). The species Gonyleptes multimaculatus Wood 1869, currently under the synonymy of L. cordata, is revalidated and transferred to Paecilaemella Roewer, 1925 forming Paecilaemella multimaculata comb. nov., and the species Cynorta itacoaiensis H. Soares, 1970 is herein considered a junior subjective synonym of it. In addition, the second current species of the genus, Libitia fusca (Simon, 1879), is transferred to Metalibitia. Libitia is diagnosed based in genital and external morphology, an identification key of the four species and distribution maps are offered.


Introduction
The Northern Andes have a high topographic complexity, recognized as a biodiversity hotspot (Myers et al. 2000). Opilionological fauna is common and abundant in Andean realms but its taxonomy is still poorly understood and its diversity is underestimated. Cosmetid harvestmen of Andes have received better attention in the last few years in Bolivia (Monteiro & Pinto-da-Rocha 2015) and Colombia (García & Kury 2017;Medrano & Kury 2017). Venezuelan species were much favored earlier thanks Separately, in 1869, Wood described Gonyleptes multimaculatus from "Napo and Maravon", based on a single specimen lacking pedipalps and "maxillae". Its description includes the presence of a transverse yellow stripe just posterior to the ocularium and minute whitish spots in dorsal scutum and coxa IV. Although, the lack of pedipalps made diffi cult identifi cation at family level, Butler (1873: 115) recognized it as a cosmetid. Nonetheless, in a lapsus of overconfi dence, he established the synonym with C. cordatus without any detailed explanation, but probably based on the vague similarity of the white spot on the anterior scutum of both species. The name was later combined with the genus Cranaus by Simon (1879: 216) but this combination was not followed by any other author and it remains today a junior synonym of L. cordata.
Libitia currently contains two species described in later 19 th century: L. cordata and L. fusca from the vague locations "Colombia" and "South America", respectively. Recently,  recovered the genus within a group with two other Andean genera of fl at cosmetids: Eulibitia Roewer, 1912 andAmbatoiella Mello-Leitão, 1943, but the status of its species were not studied.
The aim of this work is to review the genus Libitia, which are small cosmetid harvestmen with short legs and a fl at unarmed dorsal scutum. We hereby recognize four valid species in the genus, among them two new species from Meta and Boyacá Departments in the highlands of Colombia. A key for the identifi cation and a map of the known distribution of species are provided. Additional nomenclatural acts are proposed for names mistakenly associated with Libitia.

Material and methods
Descriptions of colors use the standard names of the 267 Color Centroids of the NBS/IBCC Color System (http://people.csail.mit.edu/jaffer/Color/Dictionaries#nbs-iscc) as described in Kury & Orrico (2006). Scanning Electron Microscopy was carried out with JEOL JSM-6390LV at the Center for Scanning Electron Microscopy of Museu Nacional/UFRJ and at Rudolf Barth Electron Microscopy Platform of the Oswaldo Cruz Institute / FIOCRUZ. All measurements are in millimeters unless otherwise indicated. Tarsal formula: numbers of tarsomeres in tarsus I to IV, when an individual count is given, order is from left to right side (fi gures in parentheses denote number of tarsomeres only in the distitarsi I-II).

Repositories
Terminology for outline of dorsal scutum follows . The terminology of the macrosetae of the penis ventral plate follows Kury & Villarreal (2015). The species descriptions do not repeat the generic characteristics. The term coda, here used, refers to the posterior part of the dorsal scutum, recognizable as the part with parallel straight sides, posterior to the main slope-change caused by the widened convex laterals. This term was fi rst used and defi ned by Kury (2012).
The illustrations were made using a stereo microscope with a camera lucida and digitalized using the software Inkscape ver. 0.92. Photographs were taken with a Sony Cybershot DSC-V1 camera attached to the stereo microscope. The software package CombineZP by Alan Hadley, was used to create composite images with extended depth of fi eld, through combination of several images taken at different focal planes. The resulting images were edited with Adobe Photoshop CS5. The plates were prepared in CorelDraw X7 for photographs.
All type material of Opiliones in the Zoologisches Museum of the Georg-August-Universität in Göttingen has been destroyed, according to information from the curator Dr Gert Tröster (June 28, 2011). All types designated here and the examined material were lost in the fi re of September 2 nd 2018 along with the bulk of the arachnological collection of MNRJ   Roewer (1912)]. Libitiella Roewer, 1947: 8 [type species: Libitia bipunctata Sørensen, 1932.

Etymology
Obscure, it sounds to us as a Roman woman's name. Gender feminine.

Diagnosis
Can be differentiated in having shorter legs (femur IV shorter than the total length of the dorsal scutum) contrasting with other short-legged cosmetids as Rhaucus Simon, 1879, Neocynorta or Eulibitia (its femur IV is longer than dorsal scutum length) (Fig. 1B, E-F). Outline of dorsal scutum alpha-type with faint marked constrictions forming a parallel coda in males and divergent coda in females contrasting with Ambatoiella that has well-marked constrictions (Figs 2, 6). Can be separated from other genera by having cheliceral sockets (invaginations in the anterior margin of the carapace where the basichelicerites fi t) shallow, with well-marked lateral quadrangular projections (triangular in Eulibitia, rounded in Rhaucus, bifi d in Metalibitia) (Figs 2,6). Posterior margin of DS without yellow stripe (transversal stripe in Eulibitia and some Rhaucus species) ( Fig. 1A-D). Basitarsomeres I, III and IV remarkably enlarged in males (Figs 1B, 5B, D, 9C-D). Penis with an elongate wattle in stylus forming distally two long serrate membranes and with a medial cleft (Fig. 5), without stylar barbs.
Mesotergum fl at and unarmed. Scutum with yellowish white blots as a 'ladder mask' over dark brown background (see , with intraspecifi c variation from absent ( Fig. 2) to strongly fi lled invading the entire area I and II (Fig. 6) or part of carapace (Fig. 9A).
Coxa IV entirely visible in dorsal view, without groin warts (defi ned in Kury & Barros, 2014) and small prodorsal distal apophysis. Femur III and IV curved and uniformly tuberculate (Fig. 1B). Penis ( Fig. 4): ventral plate (VP) sub-rectangular with latero-apical borders protruding from the ventral plate, stylus with an incrassated wattle forming apically two long serrate membranes creating a medial cleft (Fig. 4D), without stylar barbs. VP laterally with two pairs of MS A, two pairs of MS C (sometimes asymmetrical with 3 MS C on one side, as in L. bipunctata and L. gandalf sp. nov.), two pairs of MS D (D1 three times longer than D2), and ventrally with two pairs of MS E and one pair MS B (Fig. 4C).

Note
As no diagnostic differences have been found among the penes of each of the four species, no specifi c descriptions are made.

Geographic distribution
Central part of Eastern Cordillera of Andes from 2600 to 3700 meters, in Cundinamarca, Boyacá and Meta departments of Colombia. pattern; from L. iguaque sp. nov. by having coda as long as mid-bulge and by possessing penis with MS D1 closer to MS C2 than MS D2.

Etymology
The name is derived from the Latin 'bipunctatus', meaning 'one who has two blots'.

Material examined
Syntypes COLOMBIA • 2 ♀♀; Cundinamarca Department [originally mistakenly reported as "Venezuela"], without further locality data; BMNH • 3 ♂♂, 2 ♀♀; between La Unión and Páramo de Chingaza [originally misspelled as "Chingusa"]; 1000-2400 m a.s.l.; ZMG. VENTER. Free sternites fi nely granular; coxae II-IV fi nely and uniformly granular; coxa I with longitudinal row of tubercles and smooth area corresponding to lace area of pedipalp. Anal operculum fi nely granular CHELICERAE ( Fig. 3B-C). Basichelicerite fi nely and uniformly granular with some rounded protuberances in basal and ectal margins, two ectodistal small setiferous protuberances. Movable fi nger with row of eight teeth; fi xed fi nger with fi ve triangular teeth, decreasing in size from distal to basal part of fi nger. PEDIPALPS (Fig. 3A). Trochanter with strong ventral apophysis. Femur dorsally convex without pronounced dorsal keel or tubercles, with ventral row of ten well-formed setiferous tubercles and a mesodistal process. Patella with mesal keel. Shallow slit along tibia mesal surface, separating dorsal and ventral sides. LEGS (Fig. 1B, E-F). Uniformly granular, femora III and IV slighted curved backwards. Coxae I-IV with prolateral conical apophyses, I and II larger than III and IV. Tarsal claws of legs III and IV smooth.

Female
Dorsal scutum and Fe IV shorter than males, with coda wider. Basitarsomeres I, III and IV not enlarged. Variation (Fig. 2) Pattern of yellow spots varying from an immaculate dorsal scutum to a well-fi lled ladder mask.

Diagnosis
Can be separated from all the other species of the genus in having a heart-shape blot in areas I and II, without invading the carapace (Fig. 5A-B).

Etymology
The name is derived from the Latin adjective 'cordatus', meaning 'heart shaped'.

Remarks
Olivier's species is not an available name according to ICZN because the paper it was described in is not consistent with binominal nomenclature.  VENTER. Free sternites fi nely granular; coxae II-IV fi nely and uniformly granular; coxa I with longitudinal row of tubercles and smooth area corresponding to lace area of pedipalp. Anal operculum fi nely granular. CHELICERAE (Fig. 7F-G). Basichelicerite fi nely and uniformly granular without notable tubercles, ectodistal small setiferous protuberance; movable fi nger with row of nine sharp teeth; fi xed fi nger with six triangular teeth, decreasing in size from distal to basal part of fi nger. PEDIPALPS (Fig. 7D-E). Trochanter with strong ventral apophysis. Femur dorsally convex without pronounced dorsal keel or tubercles, with ventral row of three well-formed setiferous tubercles and a mesodistal process. Patella with mesal keel. Shallow slit along tibia mesal surface, separating dorsal and ventral sides. Female (ICN-AO 1040) Measurements. CL = 1.7, AL = 3.1, CW = 2.6, AW = 3.4, IOD = 0.6, Fe I = 1.6, Ti I = 1.3, Fe II = 2.6, Ti II = 1.9, Fe III = 2.6, Ti III = 1.5, Fe IV = 2.9, Ti IV = 2.4. Dorsal scutum longer than in males, with coda wider. Basitarsomeres I, III and IV not enlarged. Tarsal counts: 5(3)/6(3)/5/5. Variation (Fig. 6) Pattern of yellow spots varying from a complete spot in form of a heart, to a dissociated version of the spot in the middle of areas I and II. In some cases, small spots are present in posterior margin of cephalothorax.

Diagnosis
Can be separated from all the other species of the genus in having the blots pattern invading carapace and occupying entirely area I of mesotergum. Scutal groove between carapace and area I without white marks forming a smile-like pattern immersed in a white 'beard' and one spot without fi lling at each side of that 'smile'.  Paratype COLOMBIA • 1 ♀; same collection data as for holotype; IAvH-I-86.

Variation
No remarkable variation encountered in spots of DS in the examined specimens.

Diagnosis
Can be separated from all the other species of the genus in having the coda shorter than mid-bulge, maximum width of dorsal scutum equal to maximum length (Fig. 11A). Three basitarsomeres in legs I and III (Fig. 11B-C). VP of the penis with MS D1 equally distanced from MS D2 and MS C2 (Fig. 12A-D).

Etymology
Noun in apposition from Iguaque, which in the Chibcha language means 'vigorous mountain' and was the cultural epicenter in the Amerindian period; as sacred territory it represented, for the Muiscas, the universe in continuous regeneration: birth, fecundity, fertility and initiation, symbol of ascent and interior knowledge. Also, it is the name of the Sanctuary that is the type locality of the species.

Variation
No remarkable variation was encountered in spots of DS in the specimens examined.

Remarks
We tentatively propose the new combination based on the external morphology (although examination of penis morphology is imperative to be certain) of the male type (MNHN) by the following features: (1) outline of dorsal scutum gamma-type; (2) eyes poorly spaced and ocularium without depression ("Mamelon oculaire assez grand, ovale transverse, bas, légèrement granuleux." Simon 1879); (3) areas I-III with paramedian tubercles and (4) dorsal scutum and coxa IV without yellow spots. The impossibility to examine the holotype thoroughly and the "doubtful origin" of the specimen described by Simon does not allow us to verify whether this species corresponds to another already described species. It is very diffi cult to conclude whether the records provided by Roewer (1928) and Mello-Leitão (1932) are correct because both localities match with those of various species of Metalibitia (Coronato & Pintoda-Rocha 2017).

Remarks
The synonymy of Gonyleptes multimaculatus with Cosmetus cordatus, established by Butler (1873), is not supported here because the former possess the following features ( Fig. 13): (1) large body (> 6 mm); (2) main solid white spot mostly occupying the cephalothorax; (3) scattered white dots on abdomen, free tergites and coxa IV; (4) long and smooth legs (femur IV longer than the total length of the dorsal scutum); (5) dorsal scutum beta-type and (6) tarsal counts more numerous (6/14/8/9). All those features also show that Gonyleptes multimaculatus does not belong to the genus Libitia, better fi tting the diagnosis of the recently revalidated genus Paecilaemella (see . The specimens of Gonyleptes multimaculatus examined by Wood (1869) were damaged and did not allow an unambiguous diagnosis as a cosmetid, since the pedipalps are missing. Butler (1873), besides his precipitated specifi c synonymy, already showed the relation of the species with cosmetids. After a century, H. Soares (1970) described Cynorta itacoaensis from the Itacoaí River in upper Amazonas with a description and an illustration that entirely match those of Wood's species. Therefore, we propose that Cynorta itacoaiensis H. Soares, 1970 is a junior synonym of Paecilaemella multimaculata (Wood, 1869) comb. nov.

Discussion
Although cosmetids are conspicuous and showy animals with a remarkable diversity (second largest family in Neotropics), its taxonomy is in very bad shape. Little has changed in the conformation of the family as a whole (except for the recent transfer from Gonyleptidae of the genus Platygyndes Roewer, 1943 by Pinto-da-Rocha & Hara 2011) and the subfamilies remain characterized by a single and variable feature since early 20 th century. Nonetheless, the internal conformation is under constant and fundamental changes as re-diagnoses of diverse genera, generic synonyms or new combinations (Kury & Barros 2014;García & Kury 2017;).
Here, we recognize and diagnose the genus Libitia, providing important morphological information such as the size and details of the wattle in the penial stylus, that in cosmetids goes from an extremely reduced membrane with just a few "stylar barbs" (defi ned in  as in Sibambea Roewer, 1917 to the extremely incrassated one without stylar barbs in species of Libitia. Other cosmetids with a similar external appearance to species of Libitia and inhabiting high landscapes in the Andes of Ecuador and Colombia (Fig. 14), such as Ambatoiella, Eulibitia and Rhaucus, may be differentiated from Libitia by their penial morphology which possesses a small wattle with stylar barbs.
Besides the relationship of this genus with Eulibitia and Ambatoiella , Libitia also shows a remarkable resemblance with the genus Oligovonones Caporiacco, 1951, sharing characteristics such as: (1) dorsal scutum alpha-type with slight constrictions, (2) spots in marginal part of the grooves (3) mesotergum unarmed and fl at and (4) legs short and uniformly tuberculated, without spines.
The morphology of tarsomeres was shown in Medrano & Kury (2017) for different Andean cosmetids and seems to be very useful to diagnose some genera. Tarsomeres (counts) were widely used in opilionological taxonomy almost as an exclusive character to defi ne if one species belonged to one genus or another, reaching its extreme with Goodnight & Goodnight (1953). Although an exaggerated splitting vision may carry unjustifi able synonyms and taxonomical chaos (as already happened), the careful use of tarsal counts and other morphological traits may have a phylogenetic signal in suprageneric groups. Here we recognize a reduction in the number of basitarsomeres in legs I, III and IV in most species of Libitia. On the other hand, leg II by having a sensory function and being subject to environmental pressures is different from ambulatory legs (Martens 1978) and shows a remarkable intra-and interspecifi c variation in the number of basitarsomeres, so we recommend mainly the use of legs I, III and IV for taxonomic purposes.
The eastern cordillera in Colombia possesses great abundance of Páramos. Nonetheless, most of material present in biological collections is from areas near Bogotá. The complexity of the territory, low connectivity (roads) and security complexity, make these zones almost inaccessible and it is expected that new species of Libitia or even not yet discovered genera will be encountered in future investigations.