Anomaloglossus meansi sp. n., a new Pantepui species of the Anomaloglossus beebei group (Anura, Aromobatidae)

Abstract Recent extinctions and drastic population declines have been documented in the Guiana Shield endemic frog genus Anomaloglossus, hence the importance to resolve its alpha-taxonomy. Based on molecular phylogenies, the literature has long reported the occurrence of an undescribed species in the Pakaraima Mountains of Guyana in the Pantepui region. We here describe this new taxon and demonstrate that in addition to divergence at the molecular level the new species differs from congeners by a unique combination of morphological characters, notably a small size (maximum SVL in males 18.86 mm, maximum SVL in females 21.26 mm), Finger I = Finger II when fingers adpressed, Finger III swollen in breeding males, fringes on fingers absent, toes basally webbed but lacking fringes, in life presence of a thin dorsolateral stripe from tip of snout to tip of urostyle, and a black throat in preserved males (immaculate cream in females). Virtually nothing is known about the ecology of the new species. We suggest the new species to be considered as Data Deficient according to IUCN standards.


Introduction
In their influential work about bird diversification in the Venezuelan highlands, Mayr and Phelps (1967) coined the term "Pantepui" to describe the high-elevation life zones of the Guiana Shield highlands in north-eastern South America. Pantepui is best known for its numerous isolated vertical-sided sandstone table-top mountains, the iconic Lost World's tepuis, and huge tepuian massifs, last erosional remnants of a vast ancient plateau (see Kok 2013 for details). The number of phylogenetic lineages restricted to Pantepui is remarkable given the relatively reduced size of that bioregion. Pantepui seems to act as a reservoir of endemism at the species level, but also at the genus level and, to a lesser extent, at the family level (see Kok 2013 for summary). Various biogeographical hypotheses have been proposed to explain the origin and drivers of diversification of tepui-summit species/populations (Mayr andPhelps 1967, see Kok 2013 for a summary). Recent phylogeographic studies based on non-flying vertebrates (e.g. Kok et al. 2012, 2018a, b, Leite et al. 2015, Lehmberg et al. 2018) suggested a complex historical biogeography involving the synergy of long distance dispersals, vicariance and habitat shifts.
Vertical isolation makes tepui ecosystems particularly sensitive to global warming (see Vegas-Vilarrúbia 2006, Nogué et al. 2009). Because of their remoteness and the difficulties to access most tepuis and tepuian massifs, sampling in the area has been historically low, hindering the pressing need to evaluate the taxonomic status and accurate distribution of Pantepui endemic species. The situation is particularly critical in some groups in which recent extinctions or drastic population declines have been documented, such as in the Guiana Shield endemic frog genus Anomaloglossus (e.g. Fouquet et al. 2015Fouquet et al. , 2018. The genus currently comprises 28 species (Grant et al. 2017, Fouquet et al. 2018, and likely originated in the Pantepui region (area sensu Kok 2013), where several endemics with restricted distributions are reported; more widespread species are found in the lowlands of the eastern Guiana Shield (Vacher et al. 2017). The beebei species group (sensu Grant et al. 2017) is restricted to the eastern Pantepui region of Venezuela and Guyana and currently contains six species, one of them still undescribed (Figure 1). That unnamed species has previously been reported in the literature as Anomaloglossus sp. Ayanganna (Grant et al. 2006(Grant et al. , 2017, Anomaloglossus cf. praderioi (Kok 2010) and as Anomaloglossus sp. B (Kok et al. 2012) and was recovered sister to A. praderioi (La Marca, 1997) by Grant et al. (2006Grant et al. ( , 2017 and Kok et al. (2012). The new taxon was first collected in October 2000 by AL and RDM during an expedition to Mount Ayanganna in Guyana, then found on the Wokomung Massif, Guyana, in July 2003 by D. Bruce Means and in October 2004 by AL and RDM. There is no additional report of the species since then. Although its status as an undescribed species has never been disputed, no formal description has yet been proposed. It is our aim to resolve the issue and describe this new species based on the eleven collected specimens currently available.  Grant et al. 2017

Nomenclature
Taxonomy and terminology follow Grant et al. (2006Grant et al. ( , 2017. The description format is adapted from the most recent species (re)descriptions in the genus (e.g., Kok et al. 2006, Myers and Donnelly 2008, Kok 2010, Fouquet et al. 2015.

Fieldwork and deposition of specimens
Collecting activities took place on Mount Ayanganna and the Wokomung Massif, west-central Guyana (Figure 2). These two mountains, located in the southern Pakaraima range, are the easternmost high tepuis in the Guiana Shield. Ayanganna and Wokomung are 37 km apart, and some anuran species occur on both mountains (e.g. in the genus Stefania, MacCulloch et al. 2006), although the degree of species overlap is not yet fully known.
Specimens were collected by hand and euthanized by immersion in a solution of MS 222 (ROM specimens) or by immersion in 20% isopropanol (CPI specimen). Tissue (a piece of liver or thigh muscle) was removed from most specimens immediately after euthanasia and preserved in 95-100% ethanol for later molecular analyses. Whole individuals were fixed in 10% formalin and later transferred to 70% ethanol for permanent storage. Type specimens have been deposited in the collections of the Royal Ontario Museum, Canada (ROM) and the Coastal Plains Institute and Land Conservancy, USA (CPI); tissue samples were deposited in the Amphibian Evolution Lab, Biology Department, Vrije Universiteit Brussel (VUB) and ROM. Coordinates and elevations were acquired using Global Positioning System units and referenced to map datum WGS84.

Morphology
All morphometric data were taken from the preserved specimens by the same person (MPJN), to the nearest 0.01 mm, under a Leica stereo dissecting microscope using an electronic digital caliper. Colour pattern in life was taken from field notes and colour photographs. Sex and maturity were determined by the presence/absence of vocal slit(s) and by dissection. Comparisons of external character states are based both on original descriptions and examination of museum specimens (see Appendix for comparative material examined). Abbreviations for measurements are as follows:

SVL
snout-vent length HW head width, at level of angle of jaws HL head length, from angle of jaw to tip of snout IOD inter orbital distance EN eye to naris distance, from anterior edge of eye to centre of naris SL snout length, from anterior edge of eye to tip of snout TSL tip of snout length, from centre of naris to tip of snout IND internarial distance, the distance between the centres of nares EL eye horizontal length TYM tympanum horizontal length HAND I-IV relative lengths of fingers, from the proximal edge of the palmar tubercle to the tip of each finger WFD width of disc on Finger III FAL forearm length, from elbow joint to proximal edge of metacarpal tubercle THL thigh length, from vent opening to flexed knee TIL tibia length, from knee to heel TAL tarsus length, from heel to proximal edge of outer metatarsal tubercle FL foot length, from proximal edge of inner metatarsal tubercle to tip of Toe IV WTD width of disc on Toe IV   (2) skin on dorsum shagreened, venter smooth; (3) tympanic annulus visible anteroventrally; (4) Fingers I and II subequal in length, FI = FII when fingers adpressed; (5) tip of Finger IV not surpassing the base of the distal subarticular tubercle on Finger III when fingers adpressed; (6) distal subarticular tubercle on Finger III and IV present; (7) Finger III swollen in males (conspicuous pre-and postaxial swelling in breeding males); (8) fringes on fingers absent; (9) toes basally webbed, fringes on toes absent; (10) tarsal keel well defined, slightly tubercle-like and weakly curved at proximal end; (11) black arm gland absent, glandular supracarpal pad present in both sexes (larger and more glandular in males); (12) cloacal tubercles absent; (13) pale paracloacal mark present; (14) in life, thin dorsolateral stripe present, from tip of snout to tip of urostyle Comparisons. Anomaloglossus meansi sp. n. can mainly be distinguished from the four described species belonging to the degranvillei group [sensu Vacher et al. 2017and Fouquet et al. 2018, i.e. A. blanci Fouquet, Vacher, Courtois, Villette, Reizine, Gaucher, Jairam, Ouboter & Kok, 2018 (Lescure, 1975), A. dewynteri Fouquet, Vacher, Courtois, Villette, Reizine, Gaucher, Jairam, Ouboter &Kok, 2018 andA. surinamensis Ouboter &Jairam, 2012; characters in parentheses] by having FI = FII when fingers adpressed (FI > FII), the tympanic annulus anteroventrally conspicuous (inconspicuous), and a conspicuous thin dorsolateral stripe from tip of snout to tip of urostyle (absent or inconspicuous).
Compared to the other five species belonging to the beebei group [sensu Grant et al. 2017, i.e. A. beebei (Noble, 1923, A. kaiei (Kok, Sambhu, Roopsind, Lenglet & Bourne, 2006), A. praderioi, A. roraima (La Marca, 1997) and A. rufulus (Gorzula, 1990)], A. meansi sp. n. can easily be distinguished from A. beebei by its larger size in males (maximum SVL 18.86 mm in A. meansi [n = 3,] versus maximum SVL 16.80 mm [n=18] in A. beebei), smooth ventral skin (granular in A. beebei), basal toe webbing (moderate in A. beebei), and in having a conspicuous thin dorsolateral stripe from tip of snout to tip of urostyle (absent or originating from the posterior corner of eye); from A. kaiei in having a conspicuous thin dorsolateral stripe from tip of snout to tip of urostyle (originating from the posterior corner of eye in A. kaiei) and a black throat in preserved males (greyish, never black in A. kaiei); from A. roraima by its larger size in females (maximum SVL 21.26 mm in A. meansi [n = 3,] versus maximum SVL 19.30 mm [n = 18] in A. roraima), smooth ventral skin (granular in A. roraima), and in having a conspicuous thin dorsolateral stripe from tip of snout to tip of urostyle (when present originating from the anterior or posterior corner of eye in A. roraima); from A. rufulus in having a conspicuous thin dorsolateral stripe from tip of snout to tip of urostyle (absent in A. rufulus) and the posterior part of belly unmarked (heavily marbled in A. rufulus). Anomaloglossus meansi sp. n. is most similar to A. praderioi with which it shares a conspicuous thin dorsolateral stripe from tip of snout to tip of urostyle but is immediately distinguished by its smaller size in males (maximum SVL 18.86 mm in A. meansi Compared to the remainder 12 Anomaloglossus species not yet assigned to any group [A. ayarzaguenai (La Marca, 1997), A. breweri (Barrio-Amorós, 2006), A. guanayensis (La Marca, 1997), A. moffetti Barrio-Amorós & Brewer-Carías, 2008, A. murisipanensis (La Marca, 1997, A. parimae (La Marca, 1997), A. parkerae (Meinhardt & Parmelee, 1996), A. shrevei (Rivero, 1961), A. tamacuarensis (Myers & Donnelly, 1997), A. tepequem Fouquet, Souza, Nunes, Kok, Curcio, Carvalho, Grant & Rodrigues, 2015, A. tepuyensis (La Marca, 1997) and A. triunfo (Barrio-Amorós, Fuentes-Ramos & Rivas-Fuenmayor, 2004; characters in parentheses], A. meansi sp. n. mainly differs in having only basal toe webbing (moderate to extensive), and in having a conspicuous thin dorsolateral stripe from tip of snout to tip of urostyle (absent).
Description of the holotype. Adult male (ROM 43896; Figure 3), 18.58 mm SVL, in suboptimal state of preservation (extensive ventral incisions, dorsal skin locally damaged); body robust; head as wide as long, HL 99.7% of HW, HW 32% of SVL; dorsal skin shagreened; ventral skin smooth; snout moderately long, SL 47% of HL, 128% of EL, round in dorsal view, protruding in lateral view, extending past lower jaw; nares located close to tip of snout, directed posterolaterally, visible from front, barely visible in dorsal and ventral views, EN 28% of HL, 77% of ED, EN 60% of SL, TSL 49% of SL; posterior rim of naris bordered posteriorly by an inconspicuous crescent-shaped ridge; IND 40% of HW; canthus rostralis rounded; loreal region concave; IOD 104% of EL, longer than upper eyelid; postrictal tubercles low and inconspicuous; tympanic membrane inconspicuous, round, concealed posterodorsally by a diffuse supratympanic swelling; tympanic annulus visible anteroventrally, TYM 52 % of EL; choanae small, circular, located anterolaterally. Maxillary teeth present, small. Tongue longer than wide, free posteriorly, with rounded margin, small median lingual process longer than wide, tapered. Vocal slits bilateral, large, extending from edge of tongue to angle of jaw.
Forelimb swollen, robust, 94% of FAL. Ulnar fold absent, metacarpal ridge absent; swollen, glandular supracarpal pad present, heavily pigmented with melanophores; hand moderate in size, 24% of SVL, 75% of HW; relative length of fingers III>II=I=IV; pre-and postaxial swelling on third finger (i.e. Finger III swollen); fingers without fringes; tip of Finger IV not reaching distal subarticular tubercle on Finger III when fingers adpressed; finger discs expanded, wider than long, about 1.4 times width of digit; width of disc on Finger III 0.52 mm; palmar tubercle large, egg shaped, 0.72 mm (larger than Finger III disc), thenar tubercle smaller, elliptical; one or two round to ovoid subarticular tubercles (one each on Fingers I and II, two each on Fingers III and IV, with distal tubercle on Finger IV less conspicuous).
Hind limb robust, moderately long, with heel of adpressed leg reaching posterior corner of eye; skin granular with no cloacal tubercles discernible (but this could be an artefact of preservation); TL 46% of SVL, heels not in contact when hind limbs are flexed at right angle to sagittal plane of body; FL 38% of SVL; relative length of adpressed toes IV>III>V>II>I; Toe I very short, its tip barely reaching the base of subarticular tubercle of Toe II when adpressed; toe discs larger than width of toes; disc on Toe I only slightly larger than width of digit; width of disc on Toe IV 0.67 mm; toes basally webbed, lateral fringes absent; one to three round to ovoid subarticular tubercles (one each on Toes I and II, two each on Toes III and V, and three on Toe IV, with distal tubercle on Toe IV the smallest and least conspicuous). Inner metatarsal tubercle protuberant elliptical, 0.47 mm in length, outer metatarsal tubercle round, protuber-ant, pigmented, 0.35 mm in diameter. No medial metatarsal tubercle discernible. Tarsal keel slightly tubercle-like and weakly curved at proximal end, extending distally to preaxial edge of Toe I. Metatarsal fold not visible.
Colour of holotype in life. Dorsal ground colour chestnut brown with a short black middorsal line between shoulders. A black line from snout tip through eye, extending dorsolaterally to groin. A narrow pale brown dorsolateral stripe above this line, blending into the chestnut dorsal ground colour. Upper surface of limbs light brown proximally, becoming dark brown distally. Flanks reddish brown with yellow spots on lower flanks. Venter pale brown with dark brown mottling, throat very dark brown to black. Underside of limbs orange-red, changing to dark reddish brown on distal forearms.
Colour of holotype in preservative. After more than 13 years in preservative, dorsal ground colour became dark chestnut brown with a short middorsal black longitudinal line in the scapular region. No other dorsal marking present. Dorsal surface of arms varies from light brown proximally to dark brown, purplish-black towards the granular supracarpal pads. Dorsal surface of legs light brown with darker brown markings. Flanks dark brown to purplish-black with pale spots on lower flanks. Narrow pale brown dorsolateral stripes indistinguishable from dorsal ground colour, although the black dorsolateral stripe remains visible. Throat black, heavily pigmented with melanophores; belly cream, pigmented with melanophores (less densely distributed than on throat). Pale paracloacal marks are visible. Palms dark brown, soles medium brown (Figure 3). Sexual dimorphism and variation within the type series. Males are usually smaller than females, 18.15-18.86 mm SVL (n = 3) versus 17.66-21.26 mm SVL (n = 5) in females, with Finger III distinctly swollen in breeding males (Figure 3). Supracarpal pads are less extended and less glandular in females and juveniles than in males. Colouration is sexually dichromatic; throat heavily pigmented black in males (immaculate yellowish-orange in females), and belly yellowish-orange pigmented with melanophores in males (immaculate yellowish-orange in females) (Figure 3). Venter immaculate in juveniles, although small pigmented areas on throat may occur (presumably in juvenile males).

Measurements of holotype (in mm
Morphometric variation is summarized in Table 1, illustrations of a male and a female paratype in life are in Figure 4. Snout in dorsal and ventral views varies from round to truncate (the latter more particularly in females, see Figure 3).
There is substantial variation in colour among preserved individuals, obviously due to preservation artefact (CPI11000 for instance is much lighter than all other individuals). Lower lip pigmented in all male and juvenile individuals, but only in three out of five females. The interorbital region is usually darker than the dorsal ground colour. A short middorsal dark brown/black longitudinal line usually present in the scapular region. One female (ROM 43329) has a diffuse diamond shape marking on the anterior dorsum. Upper surface of arms and legs is cream to dark brown, with darker markings on legs. Palms and soles are light to dark brown. Flanks vary from cream to very dark purplish brown.
Distribution Etymology. It is a great pleasure to name this new species after our friend and colleague D. Bruce Means, indefatigable explorer of the "islands in the sky", and who collected one specimen of the new species and contributed with photographs and data. Thanks to his extensive fieldwork, Bruce Means greatly contributed to our understanding of the distribution, ecology, and taxonomy of Pantepui amphibians and reptiles. The specific epithet should be treated as a noun in the genitive case.
Conservation status. Anomaloglossus meansi sp. n. is only known from four localities and the 11 specimens used in the description. Virtually nothing is known about its ecology, breeding behaviour and population density. Given the uncertainty on its population status we suggest Anomaloglossus meansi sp. n. to be listed as Data Deficient according to the IUCN Red List category guidelines (2014).
Discussion. Although Ayanganna and Wokomung are close neighbours, the habitats on their slopes are not exactly similar. The slopes of Ayanganna are a series of relatively flat poorly drained plateaus alternating with steeper slopes. Collecting activities were concentrated on the plateaus, where the vegetation consists of dense, low-canopy high-tepui forest, with a dense understory of woody shrubs and large terrestrial bromeliads (MacCulloch and Lathrop 2009).
The slopes of Wokomung have no large flat plateaus. Habitat at the collecting sites consists of well-drained slopes covered in lower montane cloud forest with some epiphytes and medium density understory, including scattered terrestrial bromeliads. Streams were common on the slopes (MacCulloch et al. 2006). The majority of specimens were found in this habitat, and this may indicate that Anomaloglossus meansi sp. n. prefers this to other habitat types; or is a reflection of collecting effort in these habitats.
Species in the Anomaloglossus beebei group are currently only known from east of the Rio Caroní, in the Eastern Pantepui District. Anomaloglossus rufulus is restricted to the highlands of the eastern part of the Chimantá Massif in Venezuela, whereas A. kaiei has a rather large distribution in the uplands of the Pakaraima Mountains of Guyana ( Figure 2). The sister species A. roraima and A. beebei are allopatric, A. roraima being restricted to the highlands of the eastern tip of the Eastern Tepui Chain, whereas A. beebei is reported further to the east in the uplands of Kaieteur National Park, the Wokomung Massif and Mount Ayanganna (Figure 2). A similar spatial distribution is detected in the sister species A. praderioi and A. meansi sp. n., which are also allopatric, with A. praderioi reported from the uplands of the Eastern Tepui Chain, whereas A. meansi sp. n. is only known further to the east in the uplands of Mount Wokomung and Mount Ayanganna.
As mentioned above, virtually nothing is known about the ecology of A. meansi sp. n. Based on its phylogenetic position it is likely this species has an exotrophic, lentic tadpole ( Figure 1). Comprehensive ecological data are crucial for the assessment of species conservation status, but these assessments are known for a few species only in the Pantepui region and there is a high risk that population declines remain unnoticed in such remote areas.
Two additional phylogenetically distinct species of Anomaloglossus remain to be described in the megacephalus group (see Grant et al. 2017; the authors, in progress), but several locally restricted Anomaloglossus species probably await discovery (Vacher et al. 2017).