Hypogeophis montanus, Maddock & Wilkinson & Gower, 2018

Maddock, Simon T., Wilkinson, Mark & Gower, David J., 2018, A new species of small, long-snouted Hypogeophis Peters, 1880 (Amphibia: Gymnophiona: Indotyphlidae) from the highest elevations of the Seychelles island of Mahé, Zootaxa 4450 (3), pp. 359-375 : 361-369

Hypogeophis montanus sp. nov.

( Figs. 1–7 ; Table 1)

Holotype. BMNH 2005.1824, adult male, collected from Congo Rouge, Mahé island, Seychelles (04°38’43.6” S, 55°26’03.3” E, ca. 718 m /asl) by D.J. Gower, R.G. Kamei, S.T. Maddock and M. Wilkinson on 20 March, 2015.

Paratypes (n = 6). BMNH 2005.1823 , male, same collection data as for holotype . BMNH 2005.1822 , male, collected from Congo Rouge, Mahé island (04°38’43.6” S, 55°26’02.2” E, ca. 729 m /asl) by D.J. Gower, R.G. Kamei, S.T. Maddock and M. Wilkinson on 20 March, 2015 . BMNH 2005.1820 and 2005.1821, males, collected from Congo Rouge, Mahé island (04°38’43.9” S, 55°26’02.9” E, ca. 731 m /asl) by J. Labisko and S.T. Maddock on 24 March, 2013 . BMNH 2005.1926 , male and 2005.1927, sex not determined, collected from Morne Seychellois , Mahé island (04°38’43” S, 55°26’33.2” E, ca. 729 m /asl) by R.M. Bristol, D.J. Gower, S.T. Maddock, J.W. Streicher and M. Wilkinson on 19 September, 2015.

Diagnosis. A Hypogeophis with 76–78 vertebrae. Differs from its most similar congener H. brevis (71–75 vertebrae: Maddock et al. 2017) additionally in having a relatively smaller head (see below). Differs from H. pti (67–69 vertebrae: Maddock et al. 2017) additionally in having tentacular apertures relatively further from the eye (E-TA/E-N 0.77–0.87 versus 0.50–0.63 in known specimens). Differs from H. rostratus (> 95 vertebrae: Parker 1958) most obviously in being much smaller (maximum known total length <120 mm versus> 400 mm), and in having secondary annular grooves on most primary annuli (versus on posteriormost primary annuli only).

Identification. The new taxon is a species of Hypogeophis because (following the generic diagnosis provided by Wilkinson et al. 2011) it is an indotyphlid with eyes not covered by bone, tentacular grooves covered by bone and mesethmoid not exposed dorsally between frontals (data from microCT scans, not shown). In external morphology the new species differs from species of the other Seychelles genera ( Praslinia , Grandisonia ) in having a small pointed head and far anterior tentacular apertures, anterior to the mouth (on that part of the snout projecting beyond the lower jaw).

Description of holotype. Some meristic and morphometric data are given in Table 1. Male; condition good; c. 6 mm ventral incision into coelom c. 25 mm anterior to vent, another c. 3 mm incision c. 50 mm anterior to vent; mouth preserved slightly open; constriction of body c. 50 mm anterior of vent caused by field tag string; shallow, broad midventral groove on most of venter, more apparent on anterior half of body; small sections of few scale pockets opened dorsally.

Slightly dorsoventrally compressed, body width uniform throughout except for narrower anterior third and posteriormost few mm. Head small, distinctly narrow in dorsal view; head length a little less than midbody width, head width much less than body width (less than half of midbody width). In dorsal view head strongly V-shaped, sides straight and converging substantially from back of head to just behind anterior limit of mouth (up to level of TPs), in front of this parallel sided rostrum ends in narrowly rounded snout tip. In ventral view lower jaw and upper lip more broadly rounded than snout, upper jaws visible from CMs forwards. In lateral view upper lip slightly concave (apex closer to eye than TP), lower lip slightly downturned. Snout very prominent in lateral view, projection anterior to mouth approximately two thirds as long as upper lip.

Eyes slightly inset from edges of head in dorsal view (by distance not more than half diameter of eye), approximately halfway between snout tip and back of head; clearly visible; much larger than TAs, larger than TPs. TAs oval, centres slightly below imaginary lines between nares and CMs; slightly closer to imaginary lines between eyes and nares than to underside of snout; distinctly closer to nares than to lip. In lateral view, CMs slightly closer to bottom than top of head. Nares subcircular, approximately equidistant in lateral view from top, bottom and tip of snout; not visible in dorsal or ventral views. TPs visible in dorsal and ventral views, TAs not visible dorsally. TPs approximately equidistant between snout tip and anterior of mouth.

As far as could be determined, all teeth are bicusped. Outer tooth rows longer than inner rows on both upper and lower jaw. PMs extend only a short distance posterior to choanae. No diastema between vomerine and palatine teeth. Palate only slightly concave transversely, generally pale; tongue unattached anteriorly, tip rounded, no plicae, narial plugs large and prominent with clearly delimiting grooves medially. Choanae subcircular to broadly oval, interchoanal distance a little more than the width of each choana; choanal valves not deeply set, visible.

C2 slightly thicker than first PA; C1 thinner than C2, thicker than back of head. In lateral view C1 approximately same length as first PA, distinctly shorter than C2. NG1, NG2 and NG3 clearly marked and complete except for narrow but distinct NG3 midventral gap. NG1 and NG2 largely orthoplicate, NG3 slightly posteriorly bowed ventrally. One TG clearly indicated on C2, halfway between NG2 and NG3, extends across most of dorsum.

AGs inconspicuous to naked eye, slightly clearer posteriorly; conspicuous throughout under microscope. All PAGs complete (except perhaps last few). SAGs on all PAs, dorsolaterally on first PA, ventrolaterally on third PA, SAGs become complete on lateral apex on 20th PA, middorsally complete on 17th PA, midventrally complete on 22nd PA. No substantial regional variation in lengths of PAs. Each AG with irregular row of closely packed, pale, small and larger glands posterior to narrow darker band, darker band thicker posteriorly; larger pale glands on posterior annuli longer, more dense on ventral surface. Posterior edges of annuli increasingly raised far posteriorly, thin whitish line appears between darker band and paler glands on the middle 50% of the body.

Posteriorly (within 15 AGs from TT) four rows of scales in pockets as deep as approximately three quarters the length of a PA at this point; anteriorly (15 PAGs behind collar region) two scale rows in pockets no deeper than one quarter the length of a PA. All observed scales oval, wider than long.

Terminal cap approximately one and a half times length of posteriormost PAs. Posteriormost AG restricted to dorsum, approximately level with front of disc surrounding vent. No AGs on venter behind front of disc; last AG anterior to disc on venter midventrally incomplete. Terminus bluntly rounded, much more so than head; inconspicuous, narrow, shallow, longitudinal left of midline terminal groove behind disc onto posteroventral surface of TT. Denticles of vent slightly everted; disc approximately subcircular, with 11 (six posterior, two lateral, three anterior) approximately bilaterally symmetrical denticulations, those anterior and lateral longest; single large papillus on base of each anterolateral denticulation. Disc boundary difficult to discern, interpreted here as not extending beyond partly everted denticulations surrounding vent.

Body brown to brown-grey in preservative, uniform along length, notably paler ventrally with transition lying more dorsolaterally than laterally. Disc and denticulations entirely pale, whitish. Head greyish brown dorsally, paler posteriorly (than centrally) where no paler than adjacent body. Conspicuous pale patches broadly encircle eyes and extend forwards as eye-tentacle stripe, fading out halfway to TP on right, extending as narrow stripe to TP on left. ST unpigmented, whitish dorsally; underside of rostrum whitish anteriorly, greyish posteriorly. Nares in pale spots; TPs pale. Pale upper and lower lip lines. Broad pale regular stripes on lateral and ventrolateral surfaces of mandibles continuous with pale lips.

In life ( Fig. 4 ), holotype dark brown, darker posteriorly and dorsally. Head paler brown and with whitish patches. Macroscopically, annular grooves marked above by whitish thin line anterior to row of large whitish glands, these glands fewer, smaller laterally.

Variation among paratypes. Six paratypes, condition generally good except for some dehydration of skin causing some wrinkling. Small piece of body wall missing ventrally at approximately midbody in BMNH 2005.1927; head strongly flexed upwards, skin damaged on left upper lip of BMNH 2005.1820; left side of snout strongly squashed in BMNH 2005.1823. Paratypes agree with description of holotype presented above with following exceptions (see Table 1 for variation in quantitative characters).

Sides of rostrum less parallel, more anteriorly converging in dorsal view in BMNH 2005.1927 and 2005.1821. Anterior of upper lip and of lower jaw more broadly rounded than ST in ventral view in BMNH 2005.1821 and especially 2005.1926. Upper lip in lateral view only very slightly concave in some specimens (e.g. BMNH 2005.1821, 2005.1823, and 2005.1926). In dorsal view eyes not separated from sides of head in BMNH 2005.1927, separated by distance approximately diameter of eye in BMNH 2005.1820, other paratypes similar to holotype (eye separated from sides of head by distance less than diameter of eye).

In most paratypes (all but BMNH 2005.1823) C2 not thicker than C1 or back of head. In BMNH 2005.1926 C1 approximately as long as first PA. NG1 and NG2 with slight anterior bow ventrally in 2005.1927; NG3 orthoplicate in 2005.1821 and 2005.1927. NG3 complete in 2005.1823, 2005.1926, and 2005.1927, widely incomplete in 2005.1822. Three TGs on C 2 in BMNH 2005.1821. TG faint in 2005.1822. In 2005.1820 and 2005.1821 SAG on first PA is possibly complete middorsally, though SAGs middorsally incomplete on next few PAs. The conspicuousness of the darker and paler lines associated with AGs is variable across the sample.

Terminal cap variable, a little over one (BMNH 2005.1820 and 2005.1821) or approximately two (2005.1822) times length of posteriormost PAs. Middorsal part of posteriormost AG approximately level with centre of vent in BMNH 2005.1821, otherwise generally level with somewhere on anterior half of disc surrounding vent. No AGs ventrally incomplete immediately anterior to vent in BMNH 2005.1823, two ventrally incomplete in 2005.1822 and 2005.1926, three in 2005.1820.

Colour in preservative generally a little darker and more grey than holotype except paler, browner in BMNH

2005.1820. Only BMNH 2005.1820 resembles holotype in having relatively pale dorsum of head posteriorly. Head more grey than body in most specimens, same colour in BMNH 2005.1927 and 2005.1927, browner and paler than body in 2005.1821–1823. Pale stripe only halfway from eye to TA on left of BMNH 2005.1821. Pale ST with pigmented blotches in BMNH 2005.1926, flecks in 2005.1927. Rostrum more extensively pale in BMNH 2005.1820, 2005.1823. The two Morne Seychellois (and smallest known) specimens were a darker, less reddish brown than the other types in life.

Etymology. The specific epithet is in reference to the restricted, high elevation distribution of the species, known only from above 700 m, on the highest mountains in the Seychelles. For nomenclatural purposes the specific epithet is considered to be a noun in apposition.

Suggested ‘common’ names. Montane Mahé caecilian; montane hypogeophis (English) , leverdter nwanr montanny (Creole).

Distribution, natural history, and conservation. Hypogeophis montanus sp. nov. is known only from Morne Seychellois National Park on the island of Mahé, from elevations of ca. 718–731 m on Morne Seychellois (ca. 729 m) and nearby Congo Rouge (ca. 718–731 m). The former site is ca. 180 m below the highest peak in the Seychelles (Morne Seychellois, 905 m). The two sites are less than 1 km apart. The seven type specimens were collected during approximately 14 person hours of dedicated fieldwork in 2015 (Morne Seychellois) and 2013 and 2015 ( Congo Rouge) above 700 m.

We did not find specimens of H. montanus sp. nov. between 450 and 550 m during approximately 19 person hours of digging between Casse Dents and Congo Rouge in February and March 2013 and February 2014, during approximately two person hours of digging between 460 and 610 m on Morne Seychellois in September 2015, during four person hours of digging at ca. 612 m near the peak of Trois Frères in September 2015, during 100 minutes of digging at ca. 650 m near the peak of Morne Blanc in March 2015, or during approximately 18 person hours of digging at Mares aux Cochon (ca. 430 m) and along the path (ca. 290–410 m) leading up to Mares Aux Cochons from the Chemin le Niol road in March 2013, January and February 2014, and March and September 2015. Other than the H. brevis found at ca. 612 m near Trois Frères reported by Maddock et al. (2017), the only other caecilians we found during fieldwork between 600 and 703 m between 2013 and 2015 were three Grandisonia alternans (Congo Rouge and Morne Blanc) and one G. sechellensis (Congo Rouge). We found no species of caecilians other than H. montanus sp. nov. above 705 m.

At Congo Rouge (loam soil, 3.04 pH) and Morne Seychellois the type specimens were dug from soil <30 cm deep. All but one of the specimens were found very close to wet, moss-covered, subvertical rock faces; most of our digging was restricted to such microhabitats because digging was hindered elsewhere by rocks or large roots. One specimen at Congo Rouge was found close to the base of a tree away from a vertical rock face.

The vegetation at both sites was broadly similar ( Fig. 5 ), disturbed moss/cloud forest with trees and tree ferns ( Cyathea sechellarum Mett. ). Neither site is close to currently cultivated land, but invasive cinnamon ( Cinnamomum verum J.S.Presl ) trees occur at both and guava (Psidium cattleianum Sabine) at least at the Morne Seychellois site. The Morne Seychellois site is more disturbed, here cinnamon dominates and there were none of the native Roscheria melanochaetes (H.Wendl.) H.Wendl. ex Balf. f., Phoenicophorium borsigianum (K.Koch) Stuntz , Nepenthes pervillei Blume or Mapania Aubl. sp seen at the Congo Rouge site.

We do not know how widely H. montanus sp. nov. is distributed, having sampled relatively few localities at higher elevations within Morne Seychellois National Park. We suspect however that H. montanus sp. nov. is restricted to high elevations, almost certainly> 350 m (the lowest known elevation for H. brevis: Maddock et al. 2017 ) and perhaps only occurring> 700 m. There is only ~ 1km 2 of land above 700 m elevation in the Seychelles, and it is possible that H. montanus sp. nov. has one of the smallest distributions of any extant caecilian species. Although the known range of H. montanus sp. nov. lies entirely within a national park, this range is very small and includes probably only a single threat-defined location. Thus, if there is any evidence of decline in extent or quality of habitat or declines in numbers of individuals the new species would qualify for at least Endangered status on the IUCN Red List. Given that there is no immediate prospect of estimating or monitoring population numbers, H. montanus sp. nov. might currently qualify for Near Threatened status.

Paratype BMNH 2005.1821 has a small pit or foramen laterally on the collar region, close to NG2 and on both sides ( Fig. 3d ). This feature is in the approximate position where external embryonic or foetal gills or larval spiracles are found in other caecilian species. Among teresomatan caecilians, only the Ethiopian Sylvacaecilia grandisonae (Taylor, 1970) and some other indotyphlids of the Seychelles are known to have a larval stage (see San Mauro et al. 2014). BMNH 2005.1821 lacks other features typical of larval caecilians such as labial folds, tail fins and lateral line organs (see e.g. Wilkinson 1992). We do not think that the features in BMNH 2005.1821 are genuine spiracles because when gently probed they do not seem to pass through to communicate with the buccal cavity, and they are much narrower than unambiguous spiracles observed in other caecilians. We suggest that these features in BMNH 2005.1821 are instead developmental anomalies, perhaps incompletely ‘healed’ gill scars. The same specimen also has a more upturned and softer snout tip than in the other types, perhaps indicative of additional anomalous developmental features. None of the other known specimens of Hypogeophis montanus sp. nov. has a spiracle-like feature or any other external characteristics of larvae. In preservation BMNH 2005.1821 is 81 mm in total length, no shorter than BMNH 2005.1823 collected sympatrically and at the same time, and substantially longer than BMNH 2005.1926 (66 mm) and 2005.1927 (43 mm). We predict that H. montanus sp. nov. is oviparous, as are all Seychelles caecilians for which reproductive mode is known, but there is little basis for predicting whether it has biphasic or direct development.

Morphometric and genetic differentiation of Hypogeophis montanus . The principal coordinate analysis (PCoA) of all morphometric characters separates all three nominal diminutive species of Seychelles caecilian ( Hypogeophis montanus sp. nov., H. brevis and H. pti ) from one another ( Fig. 6a ). Hypogeophis pti does not overlap in the PCoA with either H. montanus sp. nov. or H. brevis . Although not overlapping in the PCoA plot there is a close proximity between the smallest H. montanus sp. nov. specimen (BMNH 2005.1927) and some of the smallest H. brevis specimens ( Fig. 6a ). There is no correlation between elevation and number of vertebrae within H. brevis ( Fig. 6b ) supporting our interpretation of H. montanus sp. nov. as not simply a higher elevational form of H. brevis . Overall head size (ST-NG1 and WH) in H. montanus sp. nov. is relatively small in comparison to both H. brevis and H. pti , though there is overlap ( Fig. 6c, d ). The position of the TA is in a somewhat intermediate position in Hypogeophis montanus sp. nov. when compared with H. brevis and H. pti ( Fig. 6e ).

Genetic variation is substantial between H. montanus sp. nov. and the two species it most resembles phenotypically, H. brevis and H. pti . Curiously, the lowest mean genetic distance (16s) between H. montanus sp. nov. and any other Seychelles caecilian is with the morphologically distinct (and not obviously closely related) Grandisonia alternans (5.6%). For the partial 16s sequences that we generated there is a mean group p -distance of 14.9% between the sampled H. montanus sp. nov. and H. brevis and a mean of 18.8% between H. montanus sp. nov. and H. pti . The 16s phylogenetic tree supports a sister relationship between H. montanus sp. nov. and H. brevis , though with only moderate support ( Fig. 7a ). Other relationships within this 16s tree are generally weakly supported. Analysis of the nuclear data ( Fig. 7b ) supports the separation of all three of the diminutive, superficially similar species, H. montanus sp. nov., H. brevis , H. pti , and the closer phylogenetic relationship between the two former, Mahé species.

Figures

FIGURE 1. Map showing distribution of Hypogeophis montanus sp. nov. (blue diamonds) and H. brevis (red circles) on Mahé. The dashed pink box on the full map of Mahé indicates the area that has been amplified (representing Morne Seychellois National Park).

FIGURE 2. Holotype of Hypogeophis montanus sp. nov. (BMNH 2005.1824). Scale bar increments in mm.

FIGURE 3. All six paratypes of Hypogeophis montanus sp. nov. from largest (top) to smallest (bottom), showing (from left to right) head and collars in dorsal and right lateral views, and posteriormost part of body in ventral view. (a) BMNH 2005.1820, (b) BMNH 2005.1822, (c) BMNH 2005.1823, (d) BMNH 2005.1821, (e) BMNH 2005.1926, (f) BMNH 2005.1927. Scale bars 2 mm.

FIGURE 4. Hypogeophis montanus sp. nov. holotype (BMNH 2005.1824) in life.

FIGURE 5. Habitat in immediate vicinity of where type specimens of Hypogeophis montanus sp. nov. were found on Mahé. Left: Congo Rouge (type locality) at ca. 725 m elevation photographed on 20th March 2015. Right: Morne Seychellois at ca. 730 m elevation photographed on 19th September 2015.

FIGURE 6. (a) Principal coordinate analysis for all morphometric characters reported in Table 1 for Hypogeophis montanus sp. nov., H. brevis and H. pti, (b) bivariate plot showing total number of vertebrae with respect to collection site elevation for specimens of H. montanus sp. nov. and H. brevis from Mahé, (c) bivariate plot of TL vs ST–NG1 for H. montanus sp. nov., H. brevis and H. pti, (d) bivariate plot of TL vs WH for H. montanus sp. nov., H. brevis and H. pti, (e) bivariate plot of E–TA vs ST-NG1 for H. montanus sp. nov., H. brevis and H. pti. Data for H. brevis and H. pti are from Maddock et al. (2017). Colours + symbols correspond to species as follows: H. montanus sp. nov. = blue squares, H. brevis = red circles, H. pti = green triangles.

FIGURE 7. Phylogenetic relationships of Hypogeophis montanus sp. nov. (a) 16s rRNA maximum likelihood phylogenetic tree of Seychelles caecilians. Numbers on internal branches are ML bootstrap percentages. (b) SplitsTree NeighborNet of the three diminutive, phenotypically superficially similar Hypogeophis species, inferred from standardised p-distances of the multilocus nuclear dataset. Colours refer to species designation used in Fig. 7a: blue = H. montanus sp. nov., red = H. brevis, green = H. pti.