Tytthoscincus martae, Grismer & Wood & Jr. & Quah & Anuar & Ngadi & Izam & Ahmad, 2018

TYTTHOSCINCUS MARTAE View in CoL SP. NOV.

HINDU TEMPLE FOREST SKINK

MENGKARUNG HUTAN KUIL HINDU

( FIG. 13; TABLE 10)

Holotype: Adult male ( LSUHC 12688 ) collected at the Hindu Temple 1.2 km south of the town of Fraser’s Hill on the road to the Gap, Pahang, Peninsular Malaysia (N 03°42.099′, E 101°44.090′; 1078 m in elevation) by Evan S. H. Quah, Marta S. Grismer and L. Lee Grismer on 23 March 2016. GoogleMaps

Paratypes: LSUHC 12517–18 collected on 3 December 2015 and LSUHC 12749 and 12751 collected on 21 June 2016 from the Awana Road , Genting Highlands, Pahang, Peninsular Malaysia (N 3°24.036′, E 101°47.212′; 1052 GoogleMaps

(/), data unobtainable.AXG, axilla–groin length; FL, forelimb length; HDL, hind limb length; HL, head length; SVL, snout–vent length; TD, tympanum diameter.

m in elevation) by Evan S. H. Quah and L. Lee Grismer. ZRC 3.6246 is from the same locality as the holotype and was collected by L. Lee Grismer, Perry L. Wood, Jr. and Timothy M. Youmans on 26 August 2006.

Additional specimen examined: LSUHC 12072 from the Awana Road, Genting Highlands, Pahang, Peninsular Malaysia.

Diagnosis: Tytthoscincus martae sp. nov. can be differentiated from all other species of Tytthoscincus in the montane clade except T. bukitensis by having the combination of 8`1, 9`1 or 10`2 superciliaries; a deep, weakly pigmented tympanum; enlarged pectoral scales; 31–35 midbody scales; 70–74 paravertebral scales; 61–72 ventral scales; keeled, subdigital lamellae; seven subdigital lamellae on the third finger; 12 or 13 subdigital lamellae on the fourth toe; TD/HL = 0.12–0.16; HL/ SVL = 0.18–0.26; AXG/SVL = 0.47–0.55; FL/SVL = 0.21– 0.26; HDL/SVL = 0.33–0.38; and a maximum SVL of 44.0 mm ( Table 6). All non-ratiometric characters are scored across all other Tytthoscincus and species of Sphenomorphus suspected of being Tytthoscincus in Grismer et al. (2016a: 237) and Karin et al. (2016: 416).

Description of holotype: Adult female, SVL 34.9 mm; non-regenerated tail length 48.4 mm; AXG 19.1 mm; HL 7.2 mm; head width 4.5 mm; snout to anterior margin of foreleg 12.4 mm; rostral wider than long, in broad contact with frontonasal; frontonasal wider than long; prefrontals large, in point contact on midline; frontal elongate, diamond-shaped, in contact with first two supraoculars; four supraoculars; frontoparietals in contact posterior to frontal, contacting second, third and fourth supraoculars anterolaterally and parietals and interparietal posteriorly; interparietal diamond-shaped, large, slightly projecting posteriorly, eyespot in posterior projection; parietals large, in medial contact posterior to interparietal, contacting fourth supraocular anteriorly; one slightly enlarged nuchal scale left of midline; nasals small, widely separated, trapezoidal, contacting rostral anteriorly, frontonasal dorsally, first loreal posteriorly, first supralabial ventrally; nostril in centre of nasal; supranasals absent; two loreals, posterior loreal taller than wide; upper and lower preocular present; lower preocular followed by a series of suboculars; eight superciliaries, posterior superciliary elongate and projecting dorsomedially; seven suboculars; six supralabials, third, fourth and fifth below eye; two postsupralabials; three primary temporals; three secondary temporals, uppermost not contacting parietals; lower eyelid transparent, scaly, no enlarged central window; mental twice as wide as long; single, large postmental, contacting first infralabial on each side; two enlarged pairs of chin shields posterior to postmental, anterior pair contacting medially, posterior pair widely separated posteriorly by a single scale; anterior and posterior chinshield pairs contacting first, second and third infralabials; six infralabials; external ear opening 0.9 mm, subcircular, lacking anterior lobules; and tympanum deeply set, non-pigmented.

Body scales smooth, cycloid, imbricate; ventral scales slightly larger than dorsal scales; 31 longitudinal scale rows around midbody; 73 paravertebral scales; 69 ventral scales; slightly enlarged median precloacal scales overlapping outer precloacal scales; tail slightly compressed laterally; subcaudals slightly larger than dorsal caudals; limbs relatively robust, short (FL/SVL = 0.22; HDL/ SVL = 0.33), widely separated when adpressed; scales of dorsal surface slightly larger than those of ventral and posterior surfaces; palmar and plantar scales slightly raised; and digits moderate in length, scales on dorsal surfaces in single row, subdigital lamellae bicarinate, seven on third finger, 12 on fourth toe.

Coloration in life: Overall dorsal ground colour of head, body, limbs and tail dark brown; head speckled with light-coloured markings; supralabials three and four bearing a light central spot; infralabials weakly banded; dorsum, posterior portion of flanks and tail bearing small, light-coloured spots; limbs mottled; anterior of flanks to posterior margin of jaws dull-orange bearing irregularly shaped dark speckles; thick, dark, irregularly margined postorbital stripe extending from eye to groin, merging with dorsal ground colour posterior to hind limbs, bordered dorsally above anterior portion of flanks by a thick, immaculate, dull-orange stripe; all ventral surfaces dull-yellow to beige, generally immaculate; distal portions of forelimbs and palmer surfaces dark; planter surface dark; subcaudal region stippled with dark brown.

Variation: The paratypes closely approach the holotype in all aspects of coloration except for their supralabial scales being distinctly barred ( Fig. 13). The dorsum of LSUHC 12749 is slightly darker overall and the spotting less bold. The juveniles LSUHC 12072 ( SVL 22.4 mm) and 12518 ( SVL 26.5 mm) show no difference in colour pattern or pattern boldness, indicating ontogenetic changes in coloration and pattern are absent. LSUHC 12518 has a damaged neck, LSUHC 12749 has a damaged gular region and LSUHC 12751 has an incomplete tail. Meristic variation is presented in Table 10 .

(/), data unobtainable.AXG, axilla–groin length; FL, forelimb length; HDL, hind limb length; HL, head length; SVL, snout–vent length; TD, tympanum diameter.

Distribution: Tytthoscincus martae sp. nov. is known from the type locality at the Hindu temple south of the town of Fraser’s Hill and Awana Road at Genting Highlands, Pahang, Peninsular Malaysia ( Fig. 9). However, it is likely to range further north of Fraser’ Hill and south of Genting Highlands.

Natural history: Tytthoscincus martae sp. nov. inhabits leaf-litter in hill dipterocarp forests from at least 991–1239 m in elevation ( Fig. 14). Specimens from Awana Road, Genting Highlands were collected by raking leaves that had accumulated at the foot of an earthen bank and by turning logs and rocks buried in leaf-litter. Lizards were most commonly found in areas where moisture from the hillside had seeped down to the edge of the road. During overcast days, lizards were seen running swiftly through small rocks and root tangles on the forest floor beneath the canopy in the nearby forest. However, we do not know if these specimens were T. martae sp. nov. or T. bukitensis . At Fraser’s Hill, specimens were found along the edges of seeps or small streams in damp microhabitats. ZRC 3.6246 was found beneath a log and LSUHC 12688 was found buried in the leaf-litter, both near the edges of small drainages.

Etymology: The specific epithet martae is a matronym honouring my (LLG) wife Marta S. Grismer for her long, non-wavering commitment and support of my work for many years, her efforts in the field on many trips across Southeast Asia and Latin America and for generating the morphological data sets for a number of our studies.

Comparisons: Tytthoscincus martae sp. nov. can be differentiated from T. jaripendek sp. nov. and T. kakikecil sp. nov. by having as opposed to lacking a deeply recessed, weakly pigmented tympanic membrane. Tytthoscincus martae sp. nov. differs discretely from T. perhentianensis and T. jaripendek sp. nov. by having 31–35 vs. 28–30 (collectively) midbody scales and 70–74 vs. 63–66 (collectively) paravertebral scales. It is further discretely differentiated from T. jaripendek sp. nov., T. perhentianensis and T. kakikecil sp. nov. by having seven vs. five or six subdigital lamellae on the third toe. It is further discretely differentiated from T. jaripendek sp. nov., T. perhentianensis , T. kakikecil sp. nov., T. temengorensis and T. ishaki by having 12 or 13 vs. 9–11 (collectively) subdigital lamellae on the fourth toes. Tytthoscincus martae sp. nov. also bears statistically significant mean differences of all ten characters in varying combinations across all other species ( Table 6). Numeric trends of these characters among species in the montane clade are presented in Figures 4 and 5.

Although the molecular phylogeny suggests that Tytthoscincus martae sp. nov. is a separate species being that it is not nested within any other species, it cannot be discretely nor statistically (P <0.05) differentiated morphologically from T. bukitensis and both are sympatric at Fraser’s Hill and syntopic at Awana Road, Genting Highlands. The genetic data infer that the monophyly (exclusivity) of each species is strongly supported (1.00/100 and 1.00/96, respectively; Fig. 2) and they share a modest 5.4–7.4% uncorrected pairwise sequence divergence between them. More importantly, both the BI and ML analyses infer T. martae sp. nov. and T. bukitensis are not sister species – although support for this is moderate (0.89/77) – so this conundrum cannot be solved by considering them conspecific because it would render T. bukitensis paraphyletic. Most notably, however, is that certain individuals from each species are more closely related to conspecifics from the other locality than they are to individuals of the other species found only meters apart at the same locality. For example, T. martae sp. nov. LSUHC 12749 from Awana Road is more closely related to T. martae sp. nov. LSUHC 12688 from Fraser’s Hill, 38 km to the north than it is to T. bukitensis LSUHC 12750 collected within 15 min of T. martae sp. nov. LSUHC 12749 on the same day less than 5 m away in the same pile of leaf-litter ( Fig. 2). Complicating this issue is that the molecular evidence demonstrates that the current paratype of T. bukitensis from Fraser’s Hill (ZRC 3.6246) must be considered part of T. martae sp. nov. and is not closely related to the holotype of T. bukitensis from Fraser’s Hill (ZRC 2.6245). Thus, the type series of T. bukitensis is composed of two different species that are not each other’s closest relatives. Therefore, ZRC 3.6246 is removed here from the type series of T. bukitensis and transferred to the species T. martae sp. nov.

If T. bukitensis and T. martae sp. nov. were reciprocally monophyletic and occurred exclusively at different localities (i.e. they were not sympatric), we would not recognize them as separate species and posit that their genetic distinction represents isolation by distance and individuals from the intervening regions would need to be examined and sequenced. Such is the case with the allopatric sister species Hemiphyllodactylus titiwangsaensis and H. cf. titiwangsaensis – from the same mountain range as T. bukitensis and T. martae sp. nov. – that are morphologically indistinguishable yet have an uncorrected pairwise sequence divergence between them of 12.8% ( Cobos et al., 2016). However, T. bukitensis and T. martae sp. nov. are not sister species and individuals of each occur at both localities and are even syntopic at one – strongly suggesting they are separate, highly convergent, genetically exclusive lineages.

Tytthoscincus martae sp. nov. and T. bukitensis are differentiated in their 50% combined upper and lower quartile ranges of midbody scales that only narrowly approach one another at 32.9 ( Fig. 5). Although the mean values of their respective counts are not statistically significantly different, this is due to a high outlying value of 35 for midbody scales in T. bukitensis (LSUHC 12750). Given that these two species are syntopic in at least one part of their range and the only way to unequivocally separate them is with a molecular analysis, we consider them cryptic species (see below). We are well aware of the potential pitfalls using a single mitochondrial marker in cases such as these and that the use of nuclear markers might resolve the issue of potential conspecificity. However, genomic data would be the best way to test these hypotheses and plans are in progress to do so.

Specimens conferre: We were unsuccessful in obtaining DNA sequences from ZRC 2.5944 from Telom Valley, Cameron Highlands considered by Grismer (2011) to be T. bukitensis and LSUHC 12072 from the Awana Road, Genting Highlands that is newly reported here. Based on retained posteriors from a separate concatenated DAPC analyses, ZRC 2.5944 from Telom Valley clustered with T. kakikecil sp. nov. with only a 43% probability and was closest in morphospace to individuals from Fraser’s Hill. Unfortunately, due to its state of preservation, we could not confirm the presence of reduced tympanae. The left tympanum looks as though it may be deeply inset but the right tympanum did not. We are also concerned with the accuracy of the mensural data. ZRC 2.5944 was collected in 1935 and the specimen is quite limp and in overall poor condition so we cannot conclude that the measurements taken from this specimen are congruous with those from freshly preserved specimens. Some scale counts (MB, PV and VS), however, are outside the range of those of T. jaripendek sp. nov. (the new species from Cameron Highlands) but consistent with those of T. kakikecil sp. nov. ( Tables 6 and 9). We consider this specimen as T. cf. kakikecil sp. nov. but do not include it as part of the type series.

The retained DAPC posteriors indicate that LSUHC 12072 from the Awana Road, Genting Highlands clusters with T. martae sp. nov. with only a 59% probability which is consistent with its meristic and mensural data ( Table 10). Furthermore, it plots within T. martae and outside of T. bukitensis in the concatenated PCA and the DAPC. We consider this specimen as T. cf. martae sp. nov. but do not include it as part of the type series.

Sphenomorphus cophias View in CoL : Boulenger (1908) described Sphenomorphus cophias View in CoL from a single specimen (BMNH 1946.8.3.15) collected on Gunong (=Gunung) Tahan in the Banjaran Timur at ~ 1000 m in elevation and Smith (1922) reported a second specimen (which can no longer be located) from the same mountain at ~ 1100 m in elevation. Both specimens are small (36 mm and 29 mm SVL, respectively) and Boulenger (1908) reported that the ear of BMNH 1946.8.3.15 was hidden and replaced by a depression ( Fig. 15). Lim (1998) and Grismer (2011) re-examined the holotype and found that along with it being less than 45 mm SVL, it had small temporal scales that were the same size and shape as the lateral body scales; and it had short digits with toe IV being equal in length to toe III – characters Linkem et al. (2011) used to diagnose the genus. However, based on aspects of head scalation and lower eyelid morphology, Shea (in literature, 2016) believes it may be an undescribed species of Lipina. Nonetheless, data from the holotype are presented in Table 6. If it belongs in the genus Tytthoscincus View in CoL on the basis of Linkem et al. (2011), its low numbers of digital lamellae (3FL = 5; 4TL = 10), very short limbs (FL/SVL = 0.17; HDL/SVL = 0.28), elongate body (AXG/SVL = 0.58) and lack of a tympanum further extend the trends of these characters shown for T. jaripendek View in CoL sp. nov. and T. kakikecil View in CoL sp. nov. and indicate this species has a semi-fossorial ecomorphology as demonstrated by its placement in the PCA and DAPC analyses ( Fig. 15). This, coupled with its upland locality in hill dipterocarp forests, suggests it may belong to the montane clade allied to the semi-fossorial lineage. However, given that there are at least two clades of distantly related species of Tytthoscincus View in CoL in Peninsular Malaysia and the propensity for convergence of semi-fossorial ecomorphology in skinks in general (see below), we do not propose a formal placement in either clade in the absence of molecular data.