Hemidactylus alkiyumii, Carranza & Arnold, 2012

Carranza, Salvador & Arnold, Edwin Nicholas, 2012, 3378, Zootaxa 3378, pp. 1-95 : 35-40

Hemidactylus alkiyumii sp. nov.

( Figs. 2, 5C, 7, 14–16, Table 1; Appendix I; Appendix IIIC)

MorphoBank M95099 – M95289 M99609 – M99718

Hemidactylus yerburii Arnold, 1977: 101 (part.); Arnold, 1986: 283 (part.); Arnold, 1986: 420 (part.); Schätti and Desvoignes, 1999: 52 (part.); van der Kooij, 2000: 113 (part.); Sindaco and Jeremcenko, 2008: 117 (part.).

Holotype

BMNH2005.1662 , male from Tawi Atair , 610 m, Dhofar region (South Oman), 17.11639’ N 54.54861 ’E WGS84, collected in October 2005 by S. Carranza, E.N. Arnold and D. Donaire (MorphoBank M95264–M95275; Fig. 13A). Paratypes: ONHM3707 , male, same collecting data as Holotype (MorphoBank M95290 –M95304); BMNH2005.1663 , female, same collecting data as Holotype (MorphoBank M95276–M95289); IBES8078 , female from Tawi Atair, 610 m, Dhofar region (South Oman) 17.11639’ N 54.54861 ’E WGS84, collected in October 2010 by S. Carranza and F. Amat (MorphoBank M99654–M99660); IBES8079 , female, same collecting data as IBES 8078 (MorphoBank M99661–M99667); IBES8080 , female, same collecting data as IBES 8078 (MorphoBank M99668–M99674).

Other material examined

Twenty-three vouchers listed in Appendix I under H. alkiyumii sp. nov. and not mentioned above. Specimen CAS 227519, IBES 7666, IBES 7740 and samples S3337, S3472, S7789, JS2, JS3, JS4, JS7, JS62, JS63, JS64, JS77, JS78, JS79, JS80, JS87, JS88, JS89, JS90, JS91, JS92, JS93, JS94, JS95, JS96, JS97, and S7194 were included in the molecular analyses only (Table 1).

Diagnosis

A medium-sized Hemidactylus with a maximum recorded SVL of 74.5 mm; with a mean of 12.9 (11–14) longitudinal rows of dorsal tubercles at mid-body; adhesive pads medium-sized; lamellae under the 1 st toe of pes mean 7.0 (6–9); lamellae under the 4 th toe of pes mean 10.8 (10–12); preanal pores mean 7.3 (6–10); expanded subcaudal scales usually beginning 1–4 verticils behind vent (average about 2); dorsum somber, sometimes with a pattern of irregular spots or dark transverse crosses with approximately one on neck, three on body and one or two on anterior sacrum ( Fig. 15A), not diffused with yellow in life; tubercles on body sometimes with opaque white pigment, which may be on medial side of tubercles while lateral sides are dark; tail not light distally, with pattern of 11–14 dark bands that are not especially widely separated and only extend ventrally towards the tail tip where they are not very conspicuous.

Hemidactylus alkiyumii differs from H. yerburii in its larger size (SVL max. 74.5 mm, compared with max. 67.6 mm), in having fewer longitudinal rows of dorsal tubercles at mid-body (mean 12.9, 11–14, compared with mean 16.7, 16–17), fewer preanal pores in males (mean 7.3, 6–10, compared with mean 12.8, 10–15), and in having enlarged tubercles on tail that are not spinose. It differs from H. yerburii montanus , endemic to the highlands of Yemen, in its larger size (SVL max. 74.5 mm, compared with max. 68 mm), in having higher number of lamellae under the 1 st toe of pes (mean. 7.0, 6–9, compared with mean 6.2 in both males and females, 5–7), in having fewer longitudinal rows of dorsal tubercles at mid-body (mean 12.9, 11–14, compared with mean 15.1 in males and 15.4 in females, 14–16), and in having fewer preanal pores in males (mean 7.3, 6–10, compared with mean 10.1). It differs from H. jumailiae from Yemen (formerly H. yerburii ) in its larger size (SVL max. 74.5 mm, compared with max. 47 mm), in having higher number of lamellae under the 1 st toe of pes (mean. 7.0, 6–9, compared with mean 6.3, 6–7), and in having large trihedral tubercles present on back (small cycloid tubercles in H. jumailiae ). It differs from H. shihraensis from Yemen in its larger size (SVL max. 74.5 mm, compared with max. 48.2 mm), in having higher number of lamellae under the 1 st toe of pes (mean 7.0, 6–9, compared with 6), and more preanal pores in males (mean 7.3, 6–10, compared with 6). It differs from H. saba in its larger size (SVL max. 74.5 mm, compared with max. 59 mm), in having more preanal pores in males (mean 7.3, 6–10, compared with 6), and more supralabials (mean 10.0, 9–12, compared with 8–9). For differences from the second species of Dhofar Hemidactylus see below.

Etymology

The species epithet “ alkiyumii ” is a genitive Latin noun to honor Ali bin Amer Al Kiyumi, Director General of Nature Conservation of the Sultanate of Oman, for his knowledge and interest in the preservation of the biodiversity of Oman and for his help and support towards our ongoing studies on the reptile fauna of Oman.

Genetic and phylogeographic remarks

Hemidactylus alkiyumii is monophyletic in the phylogenetic analyses of Dataset 1 ( Fig. 5C) and Dataset 3 (Appendix IIIC). In both trees it is not closely related to any specific taxa. According to Fig. 5, it is sister to a well- supported clade formed by H. robustus , H. sp. 1, lineages D, E, F, G and H. homoeolepis . In Appendix III it is also sister to a similar assemblage but in this latter case the sister clade of H. alkiyumii includes also H. sp. (OTU 7 in Busais & Joger 2011a), H. shihraensis and H. saba ; specimens for which only the 12S gene was available. Based on this molecular evidence, it is clear from Fig. 5 and Appendix III that H. alkiyumii is not phylogenetically closely related to H. y. yerburii , H. y. montanus , or to former members of H. yerburii : H. jumailiae and the other Dhofar Hemidactylus described below ( Fig. 5D) or the two new species of Hemidactylus similar to H. yerburii described from Yemen by Busais and Joger (2011a) ( H. shihraensis and H. saba ). According to the results of the dating analysis inferred with Dataset 2, H. alkiyumii split from its sister clade about 11.1 mya (95% HPD: 7.1–15.9). Uncorrected genetic distances between H. alkiyumii and H. y. yerburii are 19.4% in the cytb and 9.7% in the 12S; between H. alkiyumii and H. y. montanus 9.3% in the 12S; between H. alkiyumii and H. jumailiae 9.3% in the 12S; between H. alkiyumii and the other Dhofar Hemidactylus described below ( Fig. 5D) 14.1% in the cytb and 6.5% in the 12S; between H. alkiyumii and H. shihraensis 7.3% in the 12S; and between H. alkiyumii and H. saba 9.5% in the 12S.

The level of genetic variability within H. alkiyumii is very high: 8.8% in the cytb and 2.5% in the 12S. As shown in Fig. 5C and Appendix IIIC, H. alkiyumii consists of three very well differentiated and well-supported clades, C1, C2 and C3; being in all the analyses C1 sister to a clade formed by C2 and C3. The uncorrected genetic distances between these three clades are 10.6%, 13.5% and 11% in the cytb (C1 vs. C2, C1 vs. C3 and C2 vs. C3, respectively) and 4.3%, 4.2% and 4.2% in the 12S (C1 vs. C2, C1 vs. C3 and C2 vs. C3, respectively). As shown in Fig. 2, clades C1, C2 and C3 are clearly delimited geographically from East to West. According to the calibrations, clade C1 split from the ancestor of C2 and C3 approximately 4.4 mya (95% HPD: 2.5–6.5), while clades C2 and C3 split about 3.1 mya (95% HPD: 1.6–4.8) (see Fig. 5C). The results of the nuclear networks presented in Fig 7 and a network analysis including all specimens from Dataset 1 (data not shown) indicate that all alleles of H. alkiyumii for all three independent loci analyzed (c -mos, mc1r and rag2) are private (not shared with any other species included in the present analyses). It is also interesting to notice that, although clades C1, C2 and C3 share alleles of the c -mos and rag2 genes, specimens of clade C1 do not share alleles of mc1r with C2 and C3 and the latter two clades only share a single allele (AO129a) of the mc1r gene. Despite the genetic differentiation between clades C1, C2 and C3 of H. alkiyumii , which suggests long separation of these three units, the absence of clear morphological differences between these three clades and the relatively low number of available vouchers to carry out a thorough morphological analysis (Appendix I), prevents us from reaching any taxonomic conclusions at present. Future studies should clarify the taxonomic status of clades C1, C2 and C3 of H. alkiyumii (work in progress).

As shown in Fig. 5C, Appendix IIIC and Fig. 7, the mitochondrial and nuclear DNA of specimens AO128 and AO129 turn out to be virtually identical with that of a specimen originally identified as H. macropholis from 11 km NW of Bargal, Bari region, Somalia (CAS227519) that was included in the phylogeny by Carranza and Arnold (2006; Fig. 1). This makes the original determination of the Bargal individual doubtful, especially as a second specimen of H. macropholis , from the Bari region, 11 km SE of Bosasso (CAS227511) also included in Carranza and Arnold (2006) and in the present study (Table 1; specimen CAS227511), has quite different mitochondrial and nuclear DNA with a genetic divergence from H. alkiyumii of 19% in cytb and 9% in 12S. As a result of that, and that the collector of the two specimens was the same and visited Tawi Atair (South Oman) and Somalia within the same trip, we consider specimen CAS227511 most probably from Tawi Atair (South Oman) and belonging to H. alkiyumii

Distribution

This species inhabits the forested seaward face of the mountains of Dhofar and Eastern Yemen, from Damqawt ( Yemen) in the West to North of Wadi Hasik in the East ( Fig. 2). Across its distribution range it has been recorded from sea level (8 m in Salalah City, South Oman) up to 800 m in Taiq Cave (South Oman) (Table 1).

Habits

Often in relatively mesic forested areas, though also in more open wadis in East Dhofar and in the gardens within Salalah City (South Oman). Found on rock faces, in shallow caves and on buildings. Mainly nocturnal, several specimens were out during the day in the shadow in densely forested areas in Dalkut (South Oman) and hiding in the caves in Tawi Atair (South Oman) ( Fig. 16). It can be locally abundant inside large caves. It is relatively quick and losses its skin when handled. Therefore, sometimes specimens have scars of regenerated skin on the back, probably as a result of fights with conspecifics or attacks from predators ( Fig. 15F). Hemidactylus alkiyumii shares habitat with Ptyodactylus hasselquistii and small Hemidacytlus of the H. homoeolepis group, although the latter Hemidactylus are mainly ground-dwelling, while H. alkiyumii is rock-dwelling.

Description

Males up to 74 mm SVL. Head and body markedly depressed (but less so than in H. luqueorum , H. hajarensis and H. persicu s and head less broad than in these species). Head length about 25–29% of SVL (mean males and females 27%), head width 68–85% of head length (mean males 78%, mean females 75%), head height 39–53% of length (mean males 47%, mean females 46%). Adhesive pads on digits quite broad; in adults maximum width of pad on fourth hind toe about half its length.

Nostril between rostral, supranasal and two superposed postnasals, with the first supralabial scale usually also entering narrowly into its border. Usually one scale separating the supranasals on midline. About 10–15 scales in a straight line from postnasal to edge of orbit. Small conical tubercles scattered on orbital area, on crown of head and larger on temporal area above the level of ear opening and immediately in front the upper part of this. Ear opening often elongated, its longest axis running upwards and backwards, smooth-edged, about half or more of eye diameter. Supralabial scales mean 10.0 (9–12), infralabials mean 7.9 (7–9). Mental scale broadly triangular, posteriorly bordered by two large postmentals making contact behind it, a second pair of more lateral postmentals also present, all four with a fairly smooth common transverse posterior border, which may be concave posteriorly, the postmentals contacting the first and second supralabials. Second and more posterior infralabials bordered by more irregular and smaller enlarged scales. Gulars fine.

Enlarged tubercles present on back, arranged in obliquely diagonal rows from near midline to flank, mean 12.9 (11–14) across mid-body and 13–15 in a paraventral row from the level of the axilla to that of the groin, where they are separated by spaces usually less than their own length. Tubercles strongly keeled, trihedral and striated, largest on the upper flanks but becoming smaller and more projecting and rounded lower down. Ventrals small, but larger than dorsals and imbricate, about 44–48 in a transverse row at mid-body between lateral folds (when these are discernible). Males with 6–10 preanal pores (mean 7.3), sometimes separated by one or two scales giving a formula of 3+3, 4+3, 4+4, or 5+5. Scales on upper forelimb small and imbricate, interspersed with enlarged tubercles on distal section that are smaller in the East. Scales beneath hind leg about same size as belly scales and imbricate, rather larger on front surface of thigh, enlarged tubercles present on upper surface of both femur and tibia where may be in contact, but smaller in East; also on posterior edge of foot. Lamellae under the toes of pes: 1 st toe mean 7.0 (6–9), 4 th toe mean 10.8 (10–12).

Tail relatively slender, although sometimes clearly swollen at base; six enlarged, keeled and pointed tubercles on each whorl proximally, dropping to four around whorl 9–12. Tubercles about one half the length of basal whorls, becoming smaller and placed more posteriorly on whorls distally. Small dorsal scales on tail may be muticarinate, 8–9 in longitudinal row on fourth whorl after vent, around 2–5 small scales between tubercles on fourth and fifth whorls. Subcaudal scales enlarged and broad, extending proximally as far as whorls 1–4 after the vent (average 2), and starting just after the hemipenial bulge in males.

Color varying from brown-grey in the West to pale buff in the East; sometimes a dark stripe from the nostril, through the eye, on to the cheek above ear and often on to neck; body sometimes with irregular spots; occasionally dark transverse crosses on mid-back (one on neck, three on body and one or two on anterior sacrum; Fig. 15A). Some opaque white pigment on tubercles in the East, where it may occur on one side of tubercles while the other is dark. Belly pale. Tail with numerous transverse dark bands more distally, initially on every other whorl and then on each one, the total number being around 11–14. Underside of tail pale but large subcaudals grey, the color increasing in intensity distally and made up of dark chromatophores. Underside of toe pads also grayish.

Distinctive features of Holotype

Half grown male; 56.4 mm SVL; tail 48 mm long with tip missing; some skin missing from mid-belly. Supralabial scales 11/12, infralabials 8/7; about 13 rows of enlarged tubercles at mid-back; 8 (4+4) preanal pores; lamellae under the 1 st toe of pes 8/8, 4 th toe of pes 12/12.