ZOOTAXA

ISSN 1175-5326 (print edition)

ISSN 1175-5334 (online edition)Copyright © 2014 Magnolia Press

Zootaxa 3835 (4): 501–527

www.mapress.com/zootaxa/Article

http://dx.doi.org/10.11646/zootaxa.3835.4.4

http://zoobank.org/urn:lsid:zoobank.org:pub:ADE617DC-B33C-428A-BD2B-9B2BE6D991C6

Systematics and biogeography of Hemidactylus homoeolepis Blanford, 1881

(Squamata: Gekkonidae), with the description of a new species from Arabia

RAQUEL VASCONCELOS1,2 & SALVADOR CARRANZA2,3

1CIBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos, InBIO Laboratório Associado, Universidade do Porto,

Campus Agrário de Vairão, R. Padre Armando Quintas, 4485-661 Vairão, Portugal2Institute of Evolutionary Biology (CSIC-Universitat Pompeu Fabra). Passeig Marítim de la Barceloneta 37-49, E-08003 Barcelona,

Spain3Corresponding author. E-mail: salvador.carranza@ibe.upf-csic.es

Abstract

A new species of gecko of the genus Hemidactylus (Squamata: Gekkonidae) is described from Oman and extreme

eastern Yemen. Hemidactylus minutus sp. nov. is characterized morphologically by its very small size, being the smallest

Hemidactylus in mainland Arabia, absence of enlarged tubercles anywhere on the body, expanded subcaudal scales

beginning some way from tail base, number of preanal pores, number of lamellae under the first and fourth toes, and

weakly contrasted black and white banded pattern on the ventral part of tail. It is also genetically distinct from H.

homoeolepis to which it has previously been referred, and from all other closely related Hemidactylus from the arid clade

in DNA sequence data for mitochondrial (12S, cyt b, ND4) and three nuclear (RAG1, MC1R, c-mos) markers. An adult

female from southern Yemen and a badly preserved juvenile from southwestern Saudi Arabia previously assigned to H.

homoeolepis are morphologically differentiated from this species and from H. minutus sp. nov. and temporarily referred

to as Hemidactylus sp. 12 and Hemidactylus sp. 13, respectively until more specimens are collected and analyzed.

Up to now, H. homoeolepis was the only non-endemic native species of the Socotra Archipelago. With the

description of H. minutus sp. nov., all native reptile species of Socotra are now endemic, such that this archipelago has

one with the highest number of endemic reptiles in relation to its small size. In addition, as a result of our taxonomic

change, the area of occupancy and extent of occurrence of H. homoeolepis have changed dramatically and thus its

conservation status should be updated. Although H. minutus sp. nov. seems widely distributed and relatively abundant,

its conservation status should also be re-evaluated.

Key words: gecko, DNA, morphology, taxonomic revision, Socotra, Oman

Introduction

The genus Hemidactylus Oken, 1817 currently consists of 124 named species distributed across all tropical and subtropical continental landmasses, including hundreds of intervening oceanic and continental islands (Brogard 2005, Sindaco & Jeremčenko 2008, Uetz 2013). Although a complete phylogeny of the genus is still lacking, partial molecular phylogenies indicate that all the species analyzed to date can be assigned to four phylogenetically divergent clades: 1) the African-Atlantic clade; 2) the H. angulatus clade; 3) the tropical clade; and 4) the arid clade (Arnold et al. 2008; Bansal & Karanth 2010; Bauer et al. 2010; Carranza & Arnold 2006, 2012; Šmíd et al. 2013a,b). Recently, the arid clade has been the subject of several taxonomic revisions, which have resulted in the description of 13 new species within the last three years (Busais & Joger 2011a,b; Carranza & Arnold 2012; Moravec et al. 2011; Šmíd et al. 2013b). Now, with more than a third of all species, the arid clade is currently the most speciose of the four main Hemidactylus clades. Recent phylogenetic studies (Gómez-Díaz et al. 2012; Moraveck et al. 2011; Šmíd et al. 2013a,b), as well as field investigations carried out in previously unsampled regions of the Arabian Peninsula, including the Socotra Archipelago (which comprises four islands of continental origin situated in the Arabian Sea, near the Gulf of Aden, Fig. 1), and the Horn of Africa indicate that the diversity of the arid clade is still largely underestimated and that many new species will likely be described in the next few years (Šmíd et al. 2013a).

Accepted by D. Gower: 26 May 2014; published: 14 Jul. 2014 501

One of the smallest species of Hemidactylus that belong to the arid clade is H. homoeolepis Blanford, 1881. Until Arnold (1977) reported it from mainland Arabia, this species was believed to be endemic to the Socotra Archipelago. A recent revision of the genus Hemidactylus from Oman using multilocus molecular data and morphology (Carranza & Arnold 2012) showed that Arabian H. homoeolepis were very variable and, as a result of that, three new species were described: H. paucituberculatus Carranza and Arnold, 2012, endemic to the Dhofar region of Oman; H. masirahensis Carranza and Arnold, 2012, endemic to Masirah Island, Oman; and H.

inexpectatus Carranza and Arnold, 2012, currently known only from a single locality in costal central Oman and from one specimen from the Jazirat Hamar an Nafur Island (south of Masirah Island), previously assigned to this species based on morphology only (Fig. 1). Recent surveys carried out in 2013 (unpublished data) have shown that H. inexpectatus is widely distributed on coastal areas of Al Wusta region in central Oman. After the taxonomic revision by Carranza and Arnold (2012), the distribution range of H. homoeolepis was restricted to Socotra, Samha and Darsa Islands, Socotra Archipelago, Yemen, and the Dhofar region and coastal areas in central and north Oman (Fig. 1). Two specimens deposited in the Natural History Museum, London (BMNH): BMNH1953.1.6.99 from Shaqra in Southeast Yemen, and BMNH1992.168 from Khiyat, Saudi Arabia, have also been assigned to this species despite the existence of some morphological differences (Carranza & Arnold 2012). Unfortunately, no material is available for molecular analyses (Fig. 1).

Multilocus phylogenetic analyses including specimens of H. homoeolepis from the Socotra Archipelago and mainland Arabia (Carranza & Arnold 2012; Gómez-Díaz et al. 2012; Šmíd et al. 2013a,b) indicate that there is a high level of genetic differentiation between H. homoeolepis populations from these two areas (uncorrected genetic distances of 10.4% in cyt b and 5.7% in 12S; see Carranza & Arnold 2012). This genetic differentiation was also matched by preliminary morphological data, which indicated that specimens from Socotra and mainland Arabia differed in some morphological characters (Carranza & Arnold 2012), although only six specimens from Socotra were included in the analyses. According to the phylogenetic results, H. homoeolepis arrived to the Socotra Archipelago between 2.5 and 6.4 million years ago (mya) by transmarine dispersal, because at that time Socotra was already close to its current position (Bosworth et al. 2005; Laughton 1966; Samuel et al. 1997). After the initial arrival, populations from Samha and Darsa Islands split from populations from Socotra Island approximately 0.9 – 2.5 mya (Carranza & Arnold 2012; Gómez-Díaz et al. 2012; Šmíd et al. 2013a) when the so called “Greater Socotra” island (that includes also Samha and Darsa) was partially submerged during the repeated marine transgressions.

In the present work, we use morphology, mtDNA and nDNA data to revise the taxonomy of the H.

homoeolepis species complex. The results of the analyses including specimens from its entire distribution range, suggest that up to three new species may be present in mainland Arabia, one of which is described herein.

Material and methods

Molecular analyses.

Molecular samples, DNA extraction and amplification. A total of 72 individuals of Hemidactylus were included in the molecular study. A list of all individuals included in the molecular analyses with their taxonomic identifications, sample codes, voucher references, corresponding geographical distribution data and GenBank accession numbers for all sequenced genes is presented in Appendix I. A map indicating the geographical distribution of all samples used in the genetic study is shown in Fig. 1.

Genomic DNA was extracted from ethanol-preserved tissue samples using the Qiagen DNeasy Blood & Tissue Kit. Samples from 72 specimens were amplified and sequenced for up to three mitochondrial genes encoding the ribosomal 12S rRNA (12S), cytochrome b (cyt b), NADH deshidrogenase 4 (ND4) and three nuclear markers encoding the recombination activating gene 1 (RAG1), the melano-cortin 1 receptor (MC1R), and the oocyte maturation factor Mos (c-mos). Primers, PCR conditions and source references for the amplification of all mitochondrial and nuclear markers are the same as listed in Table 2 of Carranza & Arnold (2012) and in Table S2 of Šmíd et al. (2013a).

Sequence analysis. Chromatographs were checked manually, assembled and edited using Geneious v. 6.1.6 (Biomatters Ltd.). DNA sequences were aligned using MAFFT v.6 (Katoh & Toh 2008) with the options

VASCONCELOS & CARRANZA502 · Zootaxa 3835 (4) © 2014 Magnolia Press

maxiterate 1000 and localpair. Coding mtDNA and nDNA gene fragments were translated into amino acids and no stop codons were observed. For nuclear loci, c-mos, MC1R and RAG2, heterozygous individuals were identified based on the presence of two peaks of approximately equal height at a single nucleotide site. Phased nuclear sequences were used for the network analyses. SEQPHASE (Flot 2010) (http://www.mnhn.fr/jfflot/seqphase/) was used to convert the input files, and the software PHASE v. 2.1.1 implemented in DNAsp (Librado & Rozas 2009) to resolve phased haplotypes (Stephens et al. 2001). Default settings of PHASE were used except for phase probabilities that were set as ≥ 0.7 (see Harrigan et al. 2008). All sites that remained heterozygotic after phasing the data were excluded from the network analyses. Uncorrected between and within groups mean genetic distances were calculated using MEGA 5 (Tamura et al. 2011), using p-distances and 1000 bootstraps. Gaps and missing data were removed for each sequence pair.

Phylogenetic and network analyses and estimation of divergence times. The dataset used for the phylogenetic analyses consisted of an alignment of 3318 base pairs (bp) of concatenated mitochondrial and unphased nuclear DNA for 76 Hemidactylus specimens, including 52 H. homoeolepis, 20 specimens of the new species described herein, and two H. masirahensis and two H. inexpectatus that were used as outgroups based on published evidence (Carranza & Arnold 2012). Best-fitting models were inferred for each gene independently using jModeltest v.0.1.1 (Posada 2008) under the Akaike information criterion (AIC). Clock-like behavior was tested for each gene independently with a Likelihood-ratio test implemented in MEGA 5 (Tamura et al. 2011). All information related to each partition including alignment length, model selected, number of variable and parsimony-informative positions and the results of the likelihood-ratio test are presented in Table 1.

Phylogenetic analyses were performed using Maximum Likelihood (ML) and Bayesian (BI) methods. Separate ML and BI analyses were also performed on all six independent partitions (12S, cyt b, ND4, RAG1, MC1R, and c-mos,). Alignment gaps were treated as missing data. Maximum Likelihood analyses were performed with RAxMLv.7.0.3 (Stamatakis 2006) and included 100 addition random replicates. A GTR+I+G model was used with parameters estimated independently for each partition. Reliability of the ML tree was assessed by bootstrap analysis (Felsenstein 1985) with 1000 replications. Bayesian analyses were performed with BEAST v.1.6.1 (Drummond & Rambaut 2007) using the same dataset used in the ML analysis but without the outgroups. The BEAST analysis using the concatenated alignment was used to infer the phylogenetic relationships and to estimate simultaneously the timing of the cladogenetic events. As already highlighted by Carranza & Arnold (2012), the lack of internal calibration points in Hemidactylus precludes the direct estimation of the time of the cladogenetic events in our phylogeny. Therefore, the mean substitution rate of the same 12S and cyt b mitochondrial regionscalculated for other lizard groups was used for this purpose (Carranza & Arnold 2012). Specifically, a normal distribution prior for the ucld.mean parameter of these partitions were set based on the combined meanRate posteriors 0.007556 ± 0.00247 for the 12S and 0.02286 ± 0.00806 for cyt b (see Carranza & Arnold 2012 for details). Information about the clock-like behavior of the different partitions (Table 1) was used to choose between the strict clock and the relaxed uncorrelated lognormal clock priors implemented in BEAST. Analyses were run

three times for 5x107 generations with a sampling frequency of 10000. Evolutionary models and clock priors are presented in Table 1. Other prior specifications applied were as follows (otherwise by default): coalescent constant size process of speciation; random starting tree; alpha Uniform (0, 10); ucld.mean of 12S Normal (initial value: 0.00755, mean: 0.00755, Standard deviation, SD: 0.00247); ucld.mean of cyt b Normal (initial value: 0.0228, mean: 0.0228, SD: 0.00806). Internal branches were considered strongly supported if they received ML bootstrap values ≥ 70% and posterior probability (pp) support values ≥ 0.95 (Huelsenbeck & Rannala 2004; Wilcox et al.

2002).The genealogical relationships among taxa were assessed with haplotype networks constructed using statistical

parsimony (Templeton et al. 1992), implemented in the program TCS v.1.21 (Clement et al. 2000), using phased sequences (see above) with a connection limit of 95% and deletions treated as a fifth state.

Morphological analyses.

Morphological samples, variables and museums acronyms. A total of 89 voucher specimens of the H.

homoeolepis clade were examined, and 81 of those were included in the morphological analyses (juveniles and poorly preserved specimens were excluded; Appendix II). These specimens included representatives of all known populations from mainland Arabia and the Socotra Archipelago. The specimens compared were from material in

Zootaxa 3835 (4) © 2014 Magnolia Press · 503SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS

the extensive collection of the Natural History Museum, London, UK (BMNH and NHMUK), Museo Civico di Storia Naturale di Carmagnola, Italy (MCCI), Museo di Storia Naturale of the University of Pavia, Italy (MSNPV), S. Carranza’s field series housed at the Institute of Evolutionary Biology (IBE), Barcelona, Spain, and some specimens from the National Museum Prague of Prague, Czech Republic (NMP6V), and the Oman Natural History Museum, Muscat, Oman (ONHM). The following metric variables were measured by the same person (S.C.) using a digital caliper with accuracy to the nearest 0.1 mm and were expressed in millimeters: snout-vent length (SVL), measured from tip of snout to vent; trunk length (TRL), measured from posterior edge of forelimb insertion to anterior edge of hindlimb insertion; tail length (TL), from vent to tip of tail – excluded from the multivariate analyses because many individuals did not have tail; head length (HL), distance between retroarticular process of jaw and tip of snout; head width (HW), measured at its widest part, usually at the level of temporal region; head height (HH), maximum height of head, measured from occiput to underside of jaws; orbital diameter (OD), considered as the greatest diameter of the left orbit; nares to eye distance (NE) of the left side, distance between tip of snout and anteriormost point of the left eye; anterior interorbital distance (IO1), distance between left and right supraciliary scale rows at anteriormost point of eyes; posterior interorbital distance (IO2), distance between left and right supraciliary scale rows at posteriormost point of eyes. In addition to the metric dimension measured, the following pholidotic (meristic) variables were also collected by the same person (S.C.) using a dissecting microscope: number of preanal pores (PAP) – excluded from the multivariate analyses because it is only valid for males; number of supralabial (SL) on the left and right side and infralabial (IL) scales on the left and right side; number of lamellae under the first and fourth toes of pes on the left and right side (LP 1st and LP 4th).

The morphological characteristics of the 84 specimens were carefully photographed using a Nikon 300 camera with a 60 mm macro-lens, in order to make all the data available to the scientific community. The complete collection of these high-resolution photographs has been deposited in Morphobank (http://www.morphobank.org/), Project 991. Some photographs from Project 483 (Carranza & Arnold 2012) are also used in the present work.

A list of all studied specimens with their Museum accession codes and corresponding sample code, geographical distribution data, sex, metric and meristic information and Morphobank accession numbers is presented in Appendix II. A map indicating the geographical distribution of all samples used in the morphological study is shown in Fig. 1.

Multivariate analyses. Statistical analyses were used to investigate if there were differences in size and shape between the new species described herein and H. homoeolepis from the Socotra Archipelago. Specimens BMNH1953.1.6.99 from Shaqra, southwest Yemen, and BMNH1992.168 from Khiyat, southwest Saudi Arabia were not included in the multivariate analyses because, as explained below, although they may constitute new species, there is only a single representative of each one (and in the case of BMNH1992.168 a juvenile) and no material for molecular analyses is available. Juvenile and poorly preserved specimens from which it was impossible to extract more than three morphological variables were eliminated from the multivariate analyses, resulting in a dataset of 81 specimens (Appendix II). For the multivariate analyses all variables were log-transformed to obtain data normality and increase the homogeneity of variance. Since Carranza & Arnold (2012) already showed that there was no significant sexual dimorphism in H. homoeolepis for all 14 characters included in their study (excluding the number of preanal pores, which are only present in males), both sexes were pooled together in the analyses. Although it is not always the case, size and shape might have independent evolutionary signals; and therefore we treated them independently. In order to extract the body-size effect of the dataset, as linear body measurements are generally correlated with it, all eight linear measures (TRL, HL, HW, HH, OD, NE, IO1, and IO2; see Appendix II) were regressed to SVL and we used the residues as the shape proxies. A principal component analysis (PCA) was then performed on the correlation matrix of the residuals to describe the shape variation. Regarding size, differences between specimens belonging to the new species and H. homoeolepis from Socotra were tested using a one-way ANOVA on the log-transformed values of SVL. Regarding shape, differences between these two groups were tested by means of a MANOVA on the PCA scores, using a Euclidean distance matrix with 1000 permutations to assess significance. Only seven of the 12 components were included in the analysis because their cumulative proportion to explain the total morphological variation was already above 80 percent. In order to detect which traits contributed to separate the two species a one-way ANOVA on each principal component was performed. All analyses were performed in the R environment using the packages Stats (R development Core Team 2013) and vegan (Oksanen 2013). The Fisher exact probability test on some categorical variable was performed using the web application VassarStats (www.vassarstats.net).

VASCONCELOS & CARRANZA504 · Zootaxa 3835 (4) © 2014 Magnolia Press

FIGURE 1. Distribution map of the samples used in this study. Color dots indicate specimens included in the molecular analyses and listed in Appendix I. Color stars indicate voucher specimens examined and listed in Appendix II, most of which were included in the morphological analyses. The specimens’ location colors match the colors used in the following figures. Localities of the undescribed specimens Hemidactylus sp. 12 (BMNH1953.1.6.99 from Shaqra, Yemen) and Hemidactylus sp.13 (BMNH1992.168 from Khiyat, Saudi Arabia) are also shown in the map.

Zootaxa 3835 (4) © 2014 Magnolia Press · 505SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS

TA

BL

E 1. Inform

atio

n on the genes and len

gth (in base pairs) of all partitions used in the ph

ylo

genetic analyses, m

odel of sequence evo

lution selected

by jM

odelT

est

(P

osada 2008), the clock-like behaviour, nu

mber of variable and parsim

on

y-in

form

ative sites calculated w

ith and w

itho

ut the outgroup, and the results of the L

ikelihood

Ratio T

est (L

RT

) run in M

EG

A 5 (T

am

ura et al. 2011) (see m

aterial and m

ethods). R

esults of the u

ncorrected betw

een and w

ithin groups m

ean genetic distances and

respective variance are also given. T

hese w

ere calcu

lated w

ith M

EG

A 5 (T

am

ura et al. 2

01

1), using the p-distance m

odel and 1000 bootstraps to estim

ate the variance. �

�

Gene

length

(bp)

model

clock

variable sites

parsim

on

y-inform

ative

sites

rate hom

ogeneity

test (L

TR

)

p-d

ist

with

outgroup

with

ou

t

outgroup

with

outgroup

with

ou

t

outgroup

betw

een

species

with

in

H. m

inutus sp

. n

ov.

with

in

H.hom

oeolepis

12S

395

GT

R+

I+

G

strict

71

47

62

38

not rejected

(P

>0.450)

0.065 ±

0.011

0.007 ±

0.002

0.014 ±

0.003

cyt b

303

HK

Y+

G

relaxed uncorrelated

lognorm

al

100

80

86

65

rejected

(P

<0.003)

0.110 ±

0.015

0.024 ±

0.004

0.040 ±

0.006

ND

4

585

GT

R+

I+

G

strict

161

114

150

103

not rejected

(P

>0.999)

0.115 ±

0.013

0.012 ±

0.003

0.032 ±

0.005

RA

G1

967

TrN

strict

24

11

22

9

not rejected

(P

>0.256)

0.004 ±

0.001

0.001 ±

0.001

0.001 ±

0.000

MC

1R

666

HK

Y

strict

18

9

14

5

not rejected

(P

=1.000)

0.007 ±

0.003

0.001 ±

0.000

0.000 ±

0.000

c-m

os

402

TrN

strict

12

5

11

4

not rejected

(P

=1.000)

0.004 ±

0.002

0.003 ±

0.001

0.000 ±

0.000

�

VASCONCELOS & CARRANZA506 · Zootaxa 3835 (4) © 2014 Magnolia Press

Results

Molecular analyses

Phylogenetic and network analyses and estimation of divergence times. The results of the ML and BI analyses are presented in Fig. 2. The two trees are almost identical, differing only in the position of the clade comprising samples from Samha and Darsa islands. This clade is sister to all Socotran samples in the ML tree (see Fig. 2) and nested within Socotran samples (sister to the clade including all the samples derived from the most recent common ancestor of samples S5624 and S10687) in the BI tree (data not shown). The two species are reciprocally monophyletic and well supported by high bootstrap and pp values (Fig. 2). The uncorrected genetic distances (p-distances) between them are 6.5 ± 1.1% for the 12S and 11.0 ± 1.5% for the cyt b (Table 1). The level of genetic variability in the mitochondrial markers within the new species described herein is 0.7 ± 0.2% for the 12S and 2.4 ± 0.4% for the cyt b, and within H. homoeolepis 1.4 ± 0.3% for the 12S and 4.0 ± 0.6% for the cyt b (Table 1). The results of the haplotype network analyses are presented in Fig. 3 and clearly show that all eight RAG1 and five MC1R haplotypes of the new species are private as well as all 14 RAG1 and six MC1R haplotypes of H.

homoeolepis. Only for c-mos is one haplotype is shared between the two species. The remaining haplotypes are private (four for the new species and six for H. homoeolepis).

Morphological analyses

Multivariate analyses. Size and shape differences between the new species described herein from mainland Arabia and H. homoeolepis from Socotra are presented in Figs. 4 and 5. Descriptive statistics for all 15 measured variables are presented in Table 2. Size differences between species were tested using a one-way ANOVA on the log-transformed values of SVL and this test result was significant (F=88.41; d.f.=1;�P<0.001). Shape differences between species were tested with a MANOVA on the PCA scores of the seven components and this test result was significant (F=5.7088; d.f.=7;�P<0.001). The first test revealed that the first PCA component was significant (F=26.90;�d.f.=1;�P<0.001), although the proportion of variance explained by it was low (34%; Appendix III). In this component the highest loadings are for head width and height, nostril–eye, anterior interorbital and posterior interorbital distances, and for mean number of lamellae under the fourth toe (Appendix III). These results indicate that the�new species described herein mainly presents a combination of lower values of snout-vent length, head width and height, nostril–eye, anterior interorbital and posterior interorbital distances, and of lamellae under the fourth toe than H. homoeolepis (Table 2; Appendix II). In fact, the differences in the number of lamellae under the first and fourth toe are remarkable. While more than half of the first toes of the individuals of the new species had four lamellae (43/79), none of the H. homoeolepis counted had four lamellae, (77/79 had five and 2/77 had six lamellae). Differences in the fourth toe were even more marked. While most of the fourth toes of individuals of the new species counted had less than nine lamellae (4/75 had seven and 69/75 had eight; only 2/75 had nine), all 77 fourth toes of the individuals of H. homoeolepis counted had nine or more lamellae (14/77 had nine, 58/77 had 10, and 5/77 had 11); so there is very little overlap between the two species in this character. Even though it was excluded from the multivariate analyses, according to the values presented in Appendix II, it is also clear that the number of preanal pores (PAP) in males of the new species described herein is significantly higher than in H.

homoeolepis (Fisher Exact Probability, d.f.=3, P<0.001). Most males of the new species had six (21/26 specimens) or five (4/26 specimens) PAP (Fig. 5F), and only one specimen out of the 26 included in the analyses had four PAP. In H. homoeolepis, most of the specimens (14/21) had four PAP (Fig. 5M) and only five specimens had either five (2/21) or six (3/21) PAP. Two specimens of H. homoeolepis had only three PAP, a condition not found in the new species described herein.Although not included in the genetic or multivariate analyses, the specimens from Yemen (BMNH1953.1.6.99) and Saudi Arabia (BMNH1992.168) seem morphologically different from the new species from mainland Arabia and, to a certain extent, to H. homoeolepis from the Socotra Archipelago (Fig. 6). The specimen from Shugra, Yemen (BMNH1953.1.6.99) is larger than the new species described herein (see Table 1 and Appendix II; SVL of specimen BMNH1953.1.6.99 is 36.4 mm, whereas the range of SVL in the new species is 20.5–34.6 mm), and it has fewer lamellae under the fourth toe than H. homoeolepis (eight lamellae in specimen BMNH1953.1.6.99 compared with 9–11 in H. homoeolepis). Moreover, it clearly differs from both the new species

Zootaxa 3835 (4) © 2014 Magnolia Press · 507SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS

and H. homoeolepis in the presence of enlarged scales on the posterior part of the body and especially at the tail base (Fig. 6C). The specimen from Saudi Arabia (BMNH1992.168) is a very poorly preserved juvenile and therefore it does not have PAP (Fig. 6I). However, even in this poor preservation state, it is evident that it differs from the new species from mainland Arabia described herein in the number of lamellae under the first toe (five or less in the new species and six in BMNH1992.168). Six lamellae under the first toe is also a very rare condition in H. homoeolepis (only present in 2/79 toes counted; Appendix II). As for the fourth toe, specimen BMNH1992.168 has nine lamellae, a condition present in only 2/75 fourth toes analyzed of the new species (Fig. 6J).

Taxonomic account

Although the morphological features of specimens BMNH1953.1.6.99 (Yemen) and BMNH1992.168 (Saudi Arabia) suggest that they are most probably new species of Hemidactylus, we prefer not to describe them until new material (including some males) can be thoroughly analyzed using both molecular and morphological methods. However, until this happens, we prefer to highlight their distinctiveness by applying the same sequential code system as in Šmíd et al. (2013a) and refer to them as Hemidactylus sp. 12 (BMNH1953.1.6.99), and Hemidactylus

sp. 13 (BMNH1992.168).Given the genetic distinctiveness of the populations from mainland eastern Yemen and Oman from H.

homoeolepis from the Socotra Archipelago in three mitochondrial and three nuclear gene fragments analyzed (Figs. 1–3), plus several morphological traits (see Results of the multivariate analyses above and diagnosis below), we therefore describe the populations of H. homoeolepis from eastern mainland Yemen and Oman as a new species.

Hemidactylus minutus sp. nov.

(Figs. 2–5, Table 1,2)

Hemidactylus homoeolepis Arnold, 1977: 103 (part.); Arnold, 1980: 279 (part.); Arnold, 1986: 419 (part.); Schätti & Desvoignes, 1999: 50 (part.); Rösler & Wranik, 2004: 515 (part.); Sindaco & Jeremčenko, 2008: 115 (part.); van der Kooij, 2000: 111 (part.); Razzetti et al. 2011: 8 (part.); Carranza & Arnold, 2012: 47 (part.); Šmíd et al. 2013a: 7 (part.); Gardner, 2013: 134 (part.).

Holotype.�NHMUK2013.901 (sample code�S7676), male, from Asylah, Oman, 21.95181N, 59.60820E WGS84, collected on the 9th October 2010 by Salvador Carranza and Fèlix Amat (Fig. 5B, D-E, Morphobank M999494-M9999).

Paratypes. NHMUK2013.902 (sample code�S7657), NHMUK2013.903 (sample code S7673), ONHM3716 (sample code�S7668) and NMP6V 74885 (sample code�S7664), All paratypes are males, except the latter one (female), and have the same collection data as the holotype (Morphobank M100008-M100011, M100012-M100017, M100000-M100007, M100039-M100043, respectively).

Referred material. Twenty specimens for genetic analyses (five of them not included in the morphological analyses; Appendix I) and 40 voucher specimens for morphological analyses (Appendix II), including the holotype and paratypes.

Etymology. The species epithet “minutus” is a Latin adjective that refers to the small size of this species, the smallest Hemidactylus in mainland Arabia.

Diagnosis. A small Hemidactylus characterized by the following combination of morphological characters: (1) maximum recorded snout-vent length, SVL 34.6 mm (mean 29.2 ± 3.4 mm); (2) absence of enlarged tubercles anywhere on the body; (3) expanded subcaudal scales beginning some way from tail base; (4) head narrow and low (4.2–6.7 mm in width and 2.4–3.7 mm in height); (5) relatively short snout (2.0–2.9 mm nostril–eye); (4) 4–6 preanal pores, PAP (mean 5.8 ± 0.5); (6) four or five lamellae under the first posterior toe, LP1st (4.5 ± 0.5); (7) seven to nine, but most usually eight lamellae under the fourth posterior toe, LP4th (mean LP4th 8.0 ± 0.2); (8) weakly contrasted black and white banded pattern on the ventral part of tail.

Hemidactylus minutus sp. nov. is morphologically similar to its sister taxon H. homoeolepis but is differentiated from it by its smaller adult size (mean SVL 29.2 mm, maximum 34.6 mm, compared with mean 37.1 mm, maximum 46.8 mm; Table 2), higher number of PAP (80.8% with six PAP and none with three, compared

VASCONCELOS & CARRANZA508 · Zootaxa 3835 (4) © 2014 Magnolia Press

with 66.7% with four PAP and 3.8% with three; Fig. 5F versus Fig. 5M, Table 2), fewer lamellae under the first toe (55.1% with four lamellae, 44.9% with five lamellae, compared with 0.0% with four lamellae, 97.5% with five lamellae and 2.5% with six lamellae), and fewer lamellae under the fourth toe (97.3% with either seven or eight lamellae, compared with 100% with nine to eleven lamellae; Fig. 5G, versus Fig. 5N; Table 2), and the less contrasted black and white banded pattern on the tail, especially on the ventral part (Fig. 5B, E compared with Fig. 5I, L).

TABLE 2. Mean, standard deviation (SD) and range of the morphological variables examined for H. minutus sp. nov.

from mainland Arabia and H. homoeolepis from Socotra.

Hemidactylus minutus sp. nov. differs morphologically from neighboring populations of H. paucituberculatus

(formerly H. homoeolepis) from South Oman in its rather smaller adult size (SVL mean 29.2 mm, maximum 34.6 mm, compared with mean 32.2 mm, maximum 38.4), absence of enlarged tubercles anywhere on the body (some enlarged tubercles present on the lateral sides of the second half of body), and expanded subcaudal scales beginning some way from tail base (expanded subcaudal scales usually extending almost to tail base). It differs from H. masirahensis (formerly H. homoeolepis) endemic from Masirah Island, Oman, by its smaller adult body size (SVL mean 29.2 mm, maximum 34.6 mm, compared with mean 32.2 mm, maximum 45 mm), higher number of PAP in males (96.1% with five or six pores, compared with 100% with four pores), fewer lamellae under the first toe of pes (four or five lamellae, compared with six lamellae), and under the fourth toe of pes (from seven to nine lamellae, compared with 10-11 lamellae), absence of enlarged tubercles anywhere in the body (enlarged tubercles present on lateral sides of body), expanded subcaudal scales beginning some way from tail base (expanded subcaudal scales usually extending almost to tail base), and less contrasted black and white banded pattern on the tail, especially on the ventral part (dark bands of the tail very conspicuous and marked, especially on the underside of tail). It differs from H. inexpectatus (formerly H. homoeolepis) endemic to Oman, by its smaller adult body size (SVL mean 29.2 mm, maximum 34.6 mm, compared with mean 37.5 mm, maximum 44.1 mm), higher number of PAP in males (96.1% with five or six pores, compared with 100% with four PAP), fewer lamellae under the first toe of pes (four or five lamellae, compared with six lamellae), and under the fourth toe of pes (from seven to nine lamellae, compared with 10–11 lamellae), absence of enlarged tubercles anywhere on the body (presence of conical or weakly keeled and extensive tubercles on the body, nape, hind legs and tail).

H. minutus sp. nov. H. homoeolepis

Variable code mean ± SD range mean ± SD range

Snot-vent length SVL 29.2 ± 3.4 20.5–34.6 37.1 ± 4.0 29.2–46.8

Trunk length TRL 12.48 ± 1.7 8.8–14.9 15.9 ± 2.0 12.1–19.8

Tail length TL 28.5 ± 4.4 23.5–36.0 32.1 ± 8.9 17.0–44.0

Head length HL 7.6 ± 0.8 6.2–9.6 9.4 ± 1.1 7.4–12.0

Head width HW 5.2 ± 0.6 4.2–6.7 6.9 ± 0.9 5.3–9.0

Head height HH 3.0 ± 0.4 2.4–3.7 4.2 ± 0.5 3.4–5.0

Orbital diameter OD 1.9 ± 0.2 1.4–2.3 2.5 ± 0.3 1.7–3.1

Nostril–eye distance NE 2.4 ± 0.2 2.0–2.9 3.2 ± 0.4 2.7–3.9

Anterior interorbital distance IO1 2.6 ± 0.4 2.0–3.5 3.6 ± 0.4 2.9–4.5

Posterior interorbital distance IO2 3.6 ± 0.4 2.8–4.5 4.8 ± 0.7 3.3–6.0

Preanal pores PAP 5.8 ± 0.5 4.0–6.0 4.3 ± 0.9 3.0–6.0

Mean number of supra-labial scales SL 9.2 ± 0.5 8.0–10.0 8.7 ± 0.7 8.0–10.5

Mean number of infra-labial scales IL 7.6 ± 0.5 6.5–8.5 7.9 ± 0.5 6.0–9.0

Mean number of lamellae under the first toe

LP1st 4.5 ± 0.5 4.0–5.0 5.0 ± 0.2 5.0–6.0

Mean number of lamellae under the fourth toe

LP4th 8.0 ± 0.2 7.5–9.0 9.9 ± 0.5 9.0–11.0

Zootaxa 3835 (4) © 2014 Magnolia Press · 509SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS

FIGURE 2. Maximum likelihood phylogenetic tree rooted with H. masirahensis and H. inexpectatus (outgroups not shown). Black dots on the nodes indicate ML bootstrap values ≥ 70 and BI posterior probabilities ≥ 0.95. Mean ages and 95% highest posterior density intervals of the nodes estimated with BEAST dating analysis are respectively indicated above and below the nodes. Clades are colored according to their geographic location marked on the map depicted in Fig. 1.

H. masirahensis

DarsaIsland

H. minutussp. nov

H. homoeolepis

H. inexpectatus

SamhaIsland

Yemen(mainland)

Oman

SocotraIsland

0.02

S5306

S10573

S5624

S3398

SPM001511

S7673

AO81

S10579

AO85

S3422 S3308

JS6

S7668

24Hhomo

S2166

S3562

S10577

JS5

S5353

S7707

S7909

S7091

S5154

S3807

S5060

S7710

S7664

S10745

S3399

S3444

S3419

S3380

S5305

S10574

S3257

AO119

S7924 JS75

S10687

S10572

S7929

S3351

S5373

S2831

SPM002057

S5106

S5648

S3281

S10630

JS8

S5234

S5189

S3264

S10629

S10575

S5004

S4209

S10576

S3289

S3314

S7871

S7966

S7735

S5097

S3430

S7676

S10803

S10578

S7657

S3260

S5430

S10631

S7893

AO136

S5142 S3347

2.43 - 6.64

4.37

0.85 - 2.73

1.71

0.75 - 2.08

1.35

0.10 - 0.39

0.23

0.22 - 0.83

0.50

0.29 - 0.82

0.52

Figure 2

VASCONCELOS & CARRANZA510 · Zootaxa 3835 (4) © 2014 Magnolia Press

FIGURE 3. Parsimony networks corresponding to RAG1, MC1R and c-mos nDNA sequence variation. Lines represent a mutational step, circles haplotypes and dots unsampled haplotypes. The size of circles is proportional to the number of individuals. Samples from the same species are similarly colored but with different tones for different geographic locations. For correspondences of sample and location codes see Appendix I.

S5106 bS5306 bS5353 b

S3257 aS3257 bS3260 aS3260 b S3264 a S3281 aS3281 b

S10629 aS10629 b

S3264 b

S3314 aS3314 bS3351 aS3351 bS3398 aS3399 aS3399 b

S5648 aS5648 b

S10576 aS10576 bS10577 aS10577 bS10579 aS10579 bS10630 aS10630 bS10631 aS10687 aS10687 bS10745 aS10745 b

24Hhomo a 24Hhomo b

SPM001511 aSPM001511 b

S3419 aS3419 bS3422 aS3807 aS4209 aS4209 bS5004 aS5004 bS5060 aS5060 bS5097 aS5097 bS5142 aS5142 bS5234 aS5234 bS5430 aS5624 aS5624 b

S3398 bS5430 b

S10631 b

S3422 b

S3807 b

S7909 aS7909 bS7924 aS7929 aS7929 bS7966 a

AO119 a AO119 b AO136 a AO136 b

AO81 a AO81 b AO85 a AO85 b

JS75 a JS75 b

S7091 aS7091 bS7871 aS7893 aS7893 b

S7871 b

S7966 b

S7657 aS7657 bS7664 aS7664 bS7668 aS7668 bS7673 aS7673 bS7676 aS7676 bS7924 b

c-mos

JS5 bJS6 bJS8 aJS8 b

JS5 a JS6 a

S5106 aS5154 aS5154 bS5306 aS5353 aS5373 aS5373 bS2831 a

S2831 bS3562 aS3562 bS5305 aS5305 bS5189 aS5189 b

OmanYemen (mainland)Yemen (Darsa Island)Yemen (Samha Island)Yemen (Socotra Island)

Legend

AO136 aAO136 b S7871 b

S2831 a S2831 b S3562 a S3562 b S5189 a S5189 b S5305 a S5305 b

S5353 aS5353 b

S5004 b

S3257 a S3257 b S3260 a S3264 a S3264 b S3281 a S3281 b S3308 a

S5097 b S5142 a S5142 b S5234 a S5234 b S5430 a S5430 b S5624 a S5624 b S5648 a S5648 b

S10629 a S10629 b S10687 a

S3308 b S3314 a S3347 a S3347 b S3351 a S3351 b S3398 a S3398 b

S3314 b

S7673 a S7673 b S7676 a S7676 b S7871 a S7893 a S7909 a S7909 b S7924 a S7924 b S7929 a S7929 b S7966 a S7966 b

AO119 a AO119 b

AO81 a AO81 b AO85 a AO85 b

JS75 a S7091 a S7091 b S7657 a S7657 b S7664 a S7664 bS7668 a

S7668 b

S7893 b

MC1R

S5106 a S5106 b S5154 a S5154 b S5306 a S5306 b S5373 a S5373 b S3399 a

S3399 b S3419 a S3422 a S3422 b S3444 a S3807 a S3807 b S4209 a S4209 b S5004 a S5060 a S5060 b S5097 a

S3260 bS3419 b

S10687 b

JS6 aJS6 b

JS75 b

RAG1

S7664 aS7664 b S7668 a S7668 b S7673 a S7673 b S7924 a S7966 a

S3281 aS5305 a

S5189 aS5189 b

S3257 a S3257 b S3314 a S3314 b S3399 a S3399 b S3807 a S5004 a S5004 b S5060 a S5060 bS5097 a

S5306 b

JS5 aJS5 b

JS75 b

S5624 a

S5154 aS5154 bS5306 aS5353 aS5353 b

S3398 a

S3260 a S3260 b S3419 a S3419 b

S3807 b

S5624 b

S3562 a

S3281 bS3562 bS5305 b

S4209 aS4209 bS5234 a

S5234 b

S5097 b

S3264 aS3398 b

S3264 b

AO81 aAO85 aAO85 b

JS75 aS7657 aS7657 b

JS8 a

JS8 b

S7871 a S7909 a S7924 b S7966 b

AO136 a AO136 b S7091 a S7909 b

AO119 b AO81 b S7871 b

AO119 a S7091 b

Figure 3

Zootaxa 3835 (4) © 2014 Magnolia Press · 511SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS

FIGURE 4. Size and shape differences between Hemidactylus minutus sp. nov. (blue lines and squares) from mainland Arabia and H. homoeolepis from Socotra (red lines and rectangles). The total contribution of the first three components of the PCA analysis to explain the total morphological variation is also given. See material and methods for details.

Hemidactylus minutus sp. nov. is clearly differentiated from sympatric or geographically close populations of H. hajarensis Carranza & Arnold, 2012, H. alkiyumii Carranza & Arnold, 2012, H. festivus Carranza & Arnold, 2012, and H. robustus Heyden, 1827 by its smaller adult body size and absence of enlarged tubercles anywhere on the body (all four species with maximum SVL larger than 34.6 mm, and with conical or strongly keeled and extensive tubercles on the body, nape hind legs and tail); from H. lemurinus Arnold, 1980, and from H. flaviviridis

Rüppell, 1835 by its smaller adult body size (maximum SVL larger than 34.6 mm in both species), absence of femoral pores (4–14 femoral pores on the underside of each thigh in H. flaviviridis), and fewer lamellae under the first toe of pes (counts higher than four or five in both species).

Genetic and phylogeographic remarks. The phylogenetic analyses by Carranza & Arnold (2012), Gómez-Díaz et al. (2012), and Šmíd et al. (2013a) clearly indicate that H. minutus sp. nov. and H. homoeolepis are two genetically well differentiated sister taxa. The phylogenetic analyses performed in this study support the hypothesis that these species are reciprocally monophyletic (Fig. 2) and present a high level of genetic divergence between them in the mitochondrial markers: p-distance (12S, cyt b, ND4) = 6.5 ± 1.1, 11.0 ± 1.5, and 11.5 ± 1.3%, respectively (Table 1). In addition, the network analyses also depict no haplotype sharing between the two species in two of the three nuclear markers and only one haplotype shared in c-mos (Fig. 3).

Description of the holotype. NHMUK2013.901. Data on 10 morphometric, and five meristic variables (see Material and Methods) are provided in Appendix II.

VASCONCELOS & CARRANZA512 · Zootaxa 3835 (4) © 2014 Magnolia Press

FIGURE 5. Type localities, general dorsal views and details of Hemidactylus minutus sp. nov. from mainland Arabia and H.

homoeolepis from Socotra. A) and C) type locality (Asylah, Oman) of H. minutus sp. nov.; B) dorsal view of Hemidactylus

minutus sp. nov. holotype (voucher code: NHMUK2013.901; sample code: S7676); D), E), F) and G) details of the head, tail underside pattern, preanal pores, and lamellae of the 1st and 4th toes of the holotype; H) dorsal view of one of the two H.

homoeolepis syntypes (voucher code: original BMNH81.7.22.6 and after the II World War BMNH1946.9.6.99), and I) of an unvouchered specimen (photograph by Edoardo Razzetti); J) habitat type of H. homoeolepis, generally dry places with rocky substrate; K), M) and N) details of the head, preanal pores, and lamellae of the 1st and 4th toes of the�syntype, respectively; L) detail of the pattern of underside the tail of and unvouchered specimen.

Zootaxa 3835 (4) © 2014 Magnolia Press · 513SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS

FIGURE 6. General dorsal views and details of two possibly undescribed new species from mainland Arabia. A) dorsal view of the female Hemidactylus sp. 12 specimen from Shaqra, Yemen (voucher code: BMNH1953.1.6.99); B), C), D) and E) details of the head, posterior dorsal part with large tubercles, absence of preanal pores, and lamellae of the 1st and 4th toes of the same specimen, respectively; F) dorsal view of the juvenile Hemidactylus sp. 13 specimen from Khiyat, Saudi Arabia (voucher code: BMNH1992.168); G), H), I), and J) details of the head, posterior dorsal part with no tubercles, absence of preanal pores, and lamellae of the 1st and 4th toes of the same specimen, respectively.

Adult male, small (SVL 32.8 mm), body distinctly depressed (Fig. 5B). Right thigh muscle partially removed for tissue sample; small patch of skin missing on the dorsal surface of the right posterior limb (Morphobank M99996-–M99997). Absence of enlarged tubercles; homogeneous small, granular, and generally round, not imbricate dorsal scaling on head, body, limbs, and tail (Fig. 5B; Morphobank M99994). Heterogeneous, roughly hexagonal ventral scales; gular and anteriormost ventral scales small, juxtaposed or slightly imbricate, increasing in size, height, and degree of imbrication towards the pelvic area; scales on the sides of the body with dark spot in the middle (Morphobank M99995). Supralabials 9/9, infralabials 8/8 (Morphobank M99998–M99999). Head narrow (HW/SVL=0.2), pointed and elongated (HW/SVL=0.3). Round nostrils not protuberant, defined by rostral (entering broadly into lower nostril borders on both sides), first supralabial and three supranasals in contact with upper borders. Rostral approximately square with a median notch above. Inner supranasals in contact with rostral and separated from each other by one vertical column of two polygonal scales. Chin shield rhomboid, followed by two pairs of postmentals. Anterior postmentals large and very elongated, in wide contact behind mental, both in contact with the first and second infralabials; posterior postmentals smaller, both in contact with the second infralabials (Morphobank M99995). Eye relatively large (ED/HL=0.3); ear opening oval to round (Fig. 5D; Morphobank M99998–M99999). Parietal and temporal region covered with scales of similar size and shape to others parts of the head. Lamellae under the first toe 5/5 (Fig. 5G); lamellae under the fourth toe 8/8 (Fig. 5G). Tail complete, longer than body (TL/SVL=1.1); the continuous row of enlarged tile-like subcaudal scales begins about a quarter of the tail length behind the vent (Fig. 5E; Morphobank M99996). Hemipenial swellings very obvious, with two visible elliptical openings below the vent; six pre-anal pores, PAP, in a V-shaped single row, no femoral pores or enlarged femoral scales (Fig. 5F; Morphobank M99997).

VASCONCELOS & CARRANZA514 · Zootaxa 3835 (4) © 2014 Magnolia Press

Color in life much richer than in the fixed specimen; very pale magenta-ochre, almost transparent dorsally, white-yellowish on tail (Fig. 5B). Iris in life colorful, golden with brown venations and an orange central vertical stripe.

Coloration in alcohol pale grey-buff or buff; a broad dark stripe from the nostril, through the eye, on to cheek and interrupted above ear; dorsal part of body, head, and limbs with irregular dark spots and streaks; belly pale (Morphobank M99994, M99998–M99999, M99995, respectively). Tail with ten dark bands, being narrower than pale intervening areas and increasing in intensity distally; ventral surface of tail pale with the three most distal dorsal bands extending on to it (Morphobank M99996).

Variation. Data for 10 morphometric, and five meristic characters (see Material and Methods) for all four paratypes, NHMUK2013.902, NHMUK2013.903, NMP6V 74885 and ONHM3716, are provided in Appendix II. All the specimens are very similar to each other, varying slightly in size related measurements, number of supralabials, infralabials and lamellae under the fourth toe (see Appendix II). Tails of paratypes NHMUK2013.902, and NHMUK2013.903 broken at the base, and of NMP6V 74885, and ONHM3716 almost at the base, but pieces are preserved intact together with the specimens (e.g. Morphobank M100008, M100006). A piece of tail of paratype NHMUK2013.903 and left forearms of paratypes NMP6V 74885 and ONHM3716 removed for genetic analyses. Paratypes NHMUK2013.902 and NHMUK2013.903 with ventral side of tail irregularly stippled or blotched with undefined banding. Paratype ONHM3716 with regenerated tail. Paratypes NHMUK2013.902 and NHMUK2013.903 with lateral dark stripe from the nostril continuous on to neck (Morphobank M100010, M100016). Tail of paratype NMP6V 74885 with the four most distal dorsal bands extending on to ventral side (Morphobank M323685). Anterior right postmental of paratype NMP6V 74885 only in contact with first infralabial; posterior left postmental of paratype NHMUK2013.903 in contact with the second and third infralabials (Morphobank M100040, M100013).

Main coloration very similar to the holotype, with NHMUK2013.902 and NHMUK2013.903 paratypes with darker spots and streaks.

Distribution. Distributed along the Arabian Sea coast, from northeastern Oman to extreme eastern Yemen (Fig. 1; Appendix I, II). In Dhofar it is found more than 70 km inland, as far as Thumrait. Although the population of Hemidactylus “homoeolepis” from the Hasikaya Island in the Hallaniyat Archipelago may belong to this species, a detailed genetic and morphological analysis is needed to assess whether this population is H. minutus sp.

nov., H. paucituberculatus, (also present in coastal Dhofar) or a new species. Previous reports of Hemidactylus

“homoeolepis” from Masirah Island and from Jazirat Hamar an Nafur Island have been recently assigned to H.

masirahensis and H. inexpectatus, respectively (Carranza & Arnold 2012; pers. observ.). Hemidactylus minutus sp.

nov. can be considered nearly endemic to Oman. Natural history. The new species is a ground dwelling strictly nocturnal gecko, usually found in dry places

with stony, gravely or even sandy substrates with rocky outcrops (Fig. 5A, C). It is abundant in many parts of its distribution range. Observations of specimens of H. minutus sp. nov. carried out by Arnold (1980) at Wadi Ayoun and Thumrait indicated that almost all the specimens (63/64) were first sighted either on the ground (38/64) or lower than 60 cm from it (25/64), and only one was found on a rock above this height. At Wadi Sayq, eighteen individuals reported also by Arnold (1980) were at heights of between 50 cm and 2 m on rock faces, but this may

have been because the ground here was covered by dense vegetation following the monsoon. Hemidacytlus minutus

sp. nov. is very agile, often proceeding in a series of leaps when pursued. Gravid females, each carrying a single egg, have been recorded in late September at Khawr Sawli (Arnold 1980).

Conservation status. Not evaluated.

Discussion

According to the analyses of the present work, the ancestor of Hemidactylus minutus sp. nov. and H. homoeolepis, split approximately 4.4 mya, when one lineage colonized either Socotra, Samha or Dharsa island (Fig. 2). These results are congruent with those of Carranza & Arnold (2012), Gómez-Díaz et al. (2012), and Šmíd et al. (2013a) and clearly indicate that H. homoeolepis arrived to the Socotra Archipelago by transmarine dispersal from mainland Arabia, because at that time the islands were already close to their current position (Bosworth et al. 2005; Laughton 1966; Samuel et al. 1997). This was also the case of the ancestor of the endemic Hemidactylus from Abd Al Kuri (Carranza & Arnold 2012, Gómez-Díaz et al. 2012), as well as other Socotra reptiles, such as the skinks of the genus Trachylepis (Sindaco et al. 2012a).

Zootaxa 3835 (4) © 2014 Magnolia Press · 515SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS

All previous works and the present one show that H. minutus sp. nov. is distinct from H. homoeolepis both at the mitochondrial and nuclear levels (Figs. 2–3). The high level of agreement between these independent markers ensures that no misleading effects of introgression or hybridization are compromising the interpretation of an independent species. The lack of shared haplotypes in two nuclear genes (Fig. 3) suggests that there is no gene flow between the mainland and island species, as would be expected taking into account that they are geographically very isolated, with the Arabian Sea separating them by at least more than 400 km (Fig. 1). The fact that there is one haplotype shared between these two species in the c-mos gene is most probably due to incomplete lineage sortingof ancestral polymorphism. In fact, it is indeed an ancestral and not a derived haplotype (located in a central position in the haplotype network, Fig. 3) that is shared. It is interesting to notice that the three samples of H.

minutus sp. nov. from Yemen are genetically different from the other samples included in the phylogenetic analyses (Fig. 2), and that this cannot be explained by isolation by distance. More sampling in continental Yemen is needed to clarify if this lineage extends to the west getting in contact with the Oman lineage only in the border, but currently political instability is limiting access to that area. According to our calibration, the populations from Yemen split approximately 1.71 mya. This pattern, with nearby populations from Yemen being somewhat distinct from populations across the border in Oman is also found in H. festivus but it is not so apparent in other taxa such as H. alkiyumii (Carranza & Arnold 2012). Within H. homoeolepis, Gómez-Díaz et al. (2012) indicate the presence of three evolutionarily significant units (ESUs), one in Samha and Dharsa islands, and another two in Socotra Island (Fig. 2). All these units should be considered in conservation management and future reserve design studies of the Socotra Archipelago. The new species forms together with H. homoeolepis, H. forbesii, H. masirahensis, H.

oxirhinus, H. paucituberculatus, H. inexpectatus a monophyletic group, which could be called the Hemidactylus

homoeolepis species group. Although no DNA is available for Hemidactylus sp. 12 and Hemidactylus sp. 13 from Yemen, morphology suggests that most probably they also belong to this clade.

Differentiation between H. minutus sp. nov. and H. homoeolepis is also supported by morphology. Regarding size, there are significant differences between them (Fig. 4). The fact that H. homoeolepis from Socotra is larger than H. minutus sp. nov. fits the general trend on islands that usually present faunas much larger or smaller than their sister taxa from mainland (Arnold 2000). In this case, the larger size of H. homoeolepis might be explained by the fact that, according to Gómez-Díaz et al. (2012) and Šmíd et al. (2013a), when the ancestor of H. minutus sp.

nov. and H. homoeolepis arrived to the Socotra Archipelago there was already a ground-dwelling Hemidactylus (H.

pumilio Boulenger, 1899) that is now sympatric with H. homoeolepis. So, size variation in both ground-dwelling species might have been favored by character displacement to reduce competition with H. pumilio, presently the smallest gecko of the arid clade. The situation in mainland Oman and extreme eastern Yemen is completely different, because the only ground dwelling species are the former members of H. homoeolepis (H. masirahensis, H. inexpectatus, H. paucituberculatus and H. minutus sp. nov.), which have allopatric ranges, therefore avoiding competition, most likely due to competitive exclusion. Although H. minutus sp. nov. is sympatric with other native Hemidactylus, such as H. festivus, H. alkiyumii, H. lemurinus, these other species are all much larger and occupy a different microhabitat (rock pavements and large boulders) than the ground-dwelling H. minutus sp. nov.

Considering that H. minutus sp. nov. and H. homoeolepis occupy the same general niche, occurring in similar habitats and microhabitats in mainland Arabia and Socotra, respectively (Fig. 5C, J), it is no surprise that shape differences are not very marked between these species (Fig. 4). Nevertheless, the morphological analyses indicate that they are distinct in a combination of traits apart from snout-vent length.

The Socotra Archipelago is considered one of the most remote and most biodiversity rich and distinct archipelagos in the world. Its complex geological history, which includes a long period of isolation that dates back to the origin of the Gulf of Aden approximately 30 mya (Austin et al. 2013), together with its topography and the presence of many different microclimates, are considered the main cause of its biodiversity, which includes a high number of endemic genera and species. For example, 37% of its 825 plant species and 95% of its more than 100 land snail species are endemic (Van Damme 2009). Reptiles are one of the most prominent vertebrate groups of the archipelago and include 31 species of which nine (29%) are Hemidactylus. The level of endemicity in reptiles is very high and, in fact, if one excludes the recently introduced taxa H. robustus and H. flaviviridis, H. homoeolepis

was considered the only native non-endemic species of the archipelago (Sindaco et al. 2009, 2012a; Razzetti et al.

2011). Regarding the impact of the taxonomic changes presented in this study on distribution, it should be stressed that now all native reptile species of Socotra are endemic, such that this archipelago has one with the highest number of endemic reptiles (Kier et al. 2009). In addition, the area of occupancy and extent of occurrence of H.

VASCONCELOS & CARRANZA516 · Zootaxa 3835 (4) © 2014 Magnolia Press

homoeolepis have changed dramatically following our taxonomic revision and thus its conservation status should be updated (Cox et al. 2012, Sindaco et al. 2012b). The conservation status of the newly described species H.

minutus sp. nov. should also be re-evaluated.

Acknowledgements

We wish to thank Fèlix Amat, Elena Gómez-Díaz, Margarita Metallinou, Philip de Pous, Marc Simó, Roberto Sindaco, Jiří Šmíd, Ali Alghafri and Sultan Khalifa for assisting in sample collection in the field and lending specimens and tissue samples; Joan Garcia-Porta for help with the statistical analyses. Special thanks are due to Saleh Al Saadi, Mohammed Al Shariani, Thuraya Alsariri, Ali Alkiyumi, and the other members of the Nature Conservation Department of the Ministry of Environment and Climate, Sultanate of Oman for their help and support and for issuing all the necessary permits (Refs: 08/2005; 16/2008; 38/2010; 12/2011; 13/2013; 21/2013). We also appreciate the help of Michael Robinson in Oman. This work was supported by the project “Field study for the conservation of reptiles in Oman” funded by the Ministry of Environment and Climate Affairs (Ref: 22412027), and grant CGL2012-36970 from the Ministerio de Economía y Competitividad, Spain (co-funded by FEDER). Authors are members of the Grup de Recerca Emergent of the Generalitat de Catalunya: 2009SGR1462. Raquel Vasconcelos is supported by a Postdoctoral grant from the Fundação para a Ciência e Tecnologia (FCT) (SFRH/BPD/79913/2011). We thank Roberto Sindaco and Jiři Moravec for their comments that helped to improve the manuscript. Finally, special thanks to the Environmental Protection Agency, Socotra branch from the Ministry of Water & Environment of Yemen (EPA) and to Ahmed Saeed Suleiman for providing logistic support in Socotra.

References

Arnold, E.N. (1977) The scientific results of the Oman flora and fauna survey 1975. Little-known geckoes (Reptilia: Gekkonidae) from Arabia with descriptions of two new species from the Sultanate of Oman. Journal of Oman Studies

Special Report, 1, 81–110.Arnold, E.N. (1980) The scientific results of the Oman flora and fauna survey 1977 (Dhofar). The reptiles and amphibians of

Dhofar, southern Arabia. Journal of Oman Studies Special Report, 2, 273–332.Arnold, E.N. (1986) A key and annotated checklist to the lizards and amphisbaenians of Arabia. Fauna of Saudi Arabia, 8,

385–435.Arnold, E.N. (2000) Using fossils and phylogenies to understand evolution of reptile communities on islands. In: Rheinwald, G.

(Ed.), Isolated vertebrate communities in the tropics. Bonner Zoological Monographs, 46, 309–324.Arnold, E.N., Vasconcelos, R., Harris, D.J., Mateo, J.A. & Carranza, S. (2008) Systematics, biogeography and evolution of the

endemic Hemidactylus geckos (Reptilia, Squamata, Gekkonidae) of the Cape Verde Islands: based on morphology and mitochondrial and nuclear DNA sequences. Zoologica Scripta, 37, 619–636. http://dx.doi.org/10.1111/j.1463-6409.2008.00351.x

Autin, J., Bellahsen, N., Leroy, S., Husson, L., Beslier, M.-O. & Acremont, E. (2013) The role of structural inheritance in oblique rifting: Insights from analogue models and application to the Gulf of Aden. Tectonophysics, 607, 51–64. http://dx.doi.org/10.1016/j.tecto.2013.05.041

Bansal, R. & Karanth, K.P. (2010) Molecular phylogeny of Hemidactylus geckos (Squamata: Gekkonidae) of the Indian subcontinent reveals a unique Indian radiation and an Indian origin of Asian house geckos. Molecular Phylogenetics and

Evolution, 57, 459–465. http://dx.doi.org/10.1016/j.ympev.2010.06.008

Bauer, A.M., Jackman, T.R., Greenbaum, E., Giri, V.B. & de Silva, A. (2010) South Asia supports a major endemic radiation of Hemidactylus geckos. Molecular Phylogenetics and Evolution, 57, 343–352. http://dx.doi.org/10.1016/j.ympev.2010.06.014

Blanford, W.T. (1881) Notes on the lizards collected in Socotra by Prof. I. Bayley Balfour. Proceedings of the Zoological

Society of London, 1881, 464–469.http://dx.doi.org/10.1111/j.1096-3642.1881.tb01304.x

Bosworth, W., Huchon, P. & McClay, K. (2005) The Red Sea and Gulf of Aden Basins. Journal of African Earth Sciences, 43, 334–378. http://dx.doi.org/10.1016/j.jafrearsci.2005.07.020

Boulenger, G.A. (1889) Second account of the fishes obtained by Surgeon-Major A. S. G. Jayakar at Muscat, east coast of Arabia. Proceedings of the Zoological Society of London, 1889 (pt 2), 236–246.

Brogard, J. (2005) Inventaire zoogéographique des Reptiles. Zoogeographical checklist of Reptiles. Vol. 1. Région afrotropicale

Zootaxa 3835 (4) © 2014 Magnolia Press · 517SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS

et région paléartique. Afrotropical and paleartic realms. Dominique editions, Condé sur Noireau, 301 pp. Busais, S.M. & Joger, U. (2011a) Three new species and one new subspecies of Hemidactylus Oken, 1817 from Yemen

(Squamata, Gekkonidae). Vertebrate Zoology, 61, 267–280.Busais, S.M. & Joger, U. (2011b) Molecular phylogeny of the gecko genus Hemidactylus Oken, 1817 on the mainland of

Yemen. Zoology in the Middle East, 53, 25–34. http://dx.doi.org/10.1080/09397140.2011.10648859

Carranza, S. & Arnold, E.N. (2006) Systematics, biogeography, and evolution of Hemidactylus geckos (Reptilia: Gekkonidae) elucidated using mitochondrial DNA sequences. Molecular Phylogenetics and Evolution, 38, 531–545. http://dx.doi.org/10.1016/j.ympev.2005.07.012

Carranza, S., Arnold, E.N. (2012) A review of the geckos of the genus Hemidactylus (Squamata: Gekkonidae) from Oman based on morphology, mitochondrial and nuclear data, with descriptions of eight new species. Zootaxa, 3378, 1–95.

Clement, M., Posada, D. & Crandall, K.A. (2000) TCS: a computer program to estimate gene genealogies. Molecular Ecology, 9, 1657. ttp://dx.doi.org/10.1046/j.1365-294x.2000.01020.x

Cox, N.A., Mallon, D., Bowles, P., Els, J. & Tognelli, M.F. (compilers) (2012) The Conservation Status and Distribution of

Reptiles of the Arabian Peninsula. Cambridge, UK and Gland, Switzerland: IUCN, and Sharjah, UAE: Environment and Protected Areas Authority.

Drummond, A. & Rambaut, A. (2007) BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology, 7, 214. http://dx.doi.org/10.1186/1471-2148-7-214

Felsenstein, J. (1985) Confidence-limits on phylogenies – an approach using the bootstrap. Evolution, 39, 783. http://dx.doi.org/10.2307/2408678

Flot, J.F. (2010) Seqphase: a web tool for interconverting phase input/output files and fasta sequence alignments. Molecular

Ecology Resources, 10, 162–166. http://dx.doi.org/10.1111/j.1755-0998.2009.02732.x

Gardner, A.S. (2013) The amphibians and reptiles of Oman and the UAE. Chimaira, Frankfurt am Main, Germany, 480 pp.Gómez-Díaz, E., Sindaco, R., Pupin, F., Fasola, M. & Carranza, S. (2012) Origin and in situ diversification in Hemidactylus

geckos of the Socotra Archipelago. Molecular Ecology, 21, 4074–4092. http://dx.doi.org/10.1111/j.1365-294x.2012.05672.x

Harrigan, R.J., Mazza, M.E. & Sorenson, M.D. (2008) Computation vs. cloning: evaluation of two methods for haplotype determination. Molecular Ecology Resources, 8, 1239–1248. http://dx.doi.org/10.1111/j.1755-0998.2008.02241.x

Heyden, C.H.G. von (1827) Atlas zu der Reise im nördlichen Afrika von Eduard Rüppel. Reptilien. Brönner, Frankfurt am Main, 24 pp.

Huelsenbeck, J.P. & Rannala, B. (2004) Frequentist properties of Bayesian posterior probabilities of phylogenetic trees under simple and complex substitution models. Systematic Biology, 53, 904–913. http://dx.doi.org/10.1080/10635150490522629

Katoh, K. & Toh, H. (2008) Recent developments in the MAFFT multiple sequence alignment program. Briefings in

Bioinformatics, 9, 286–298. http://dx.doi.org/10.1093/bib/bbn013

Kier, G., Kreft, H., Lee, T.M., Jetz, W., Ibisch, P.L., Nowicki, C., Mutke, J. & Barthlott, W. (2009) A global assessment of endemism and species richness across island and mainland regions. Proceedings of the National Academy of Sciences, 106 (23), 9322–9327. http://dx.doi.org/10.1073/pnas.0810306106

Laughton, A.S. (1966) The Gulf of Aden. Philosophical Transactions of the Royal Society of London Series A, Mathematical

and Physical Sciences, 259, 150–171.Librado, P. & Rozas, J. (2009) DnaSP v5: a software for comphensive analysis of DNA polymorphism data. Bioinformatics, 25,

1451–1452. http://dx.doi.org/10.1093/bioinformatics/btp187

Moravec, J., Kratochvíl, L., Amr, Z.S., Jandzik, D., Šmíd, J. & Gvoždík, V. (2011) High genetic differentiation within the Hemidactylus turcicus complex (Reptilia: Gekkonidae) in the Levant, with comments on the phylogeny and systematics of the genus. Zootaxa, 2894, 21–38.

Oksanen, J., Blanchet, F.G., Kindt, R, Legendre, P., Minchin, P.R., O'Hara, R.B., Simpson, G. L., Solymos, P., Stevens, M.H.H. & Wagner, H. (2013) Vegan: Community Ecology Package. R package version 2.0-7. Available from: http://CRAN.R-project.org/package=vegan (accessed 9 June 2014)

Posada, D. (2008) jModelTest: Phylogenetic model averaging. Molecular Biology and Evolution, 25, 1253–1256. http://dx.doi.org/10.1093/molbev/msn083

R Core Team (2013) R: a language and environment for statistical computing. R foundation for statistical computing, Vienna, Austria. Available from: http://www.R-project.org/ (accessed 9 June 2014)

Razzetti, E., Sindaco, R., Grieco, C., Pella, F., Ziliani, U., Pupin, F., Riservato, E., Pellitteri-Rosa, D., Suleiman, A.S., Al-Aseily, B.A., Carugati, C., Boncompagni, E. & Fasola, M. (2011) Annotated checklist and distribution of the Socotran

VASCONCELOS & CARRANZA518 · Zootaxa 3835 (4) © 2014 Magnolia Press

Archipelago Herpetofauna (Reptilia). Zootaxa, 2826, 1–44.Rösler, H. & Wranik, W. (2004) A key and annotated checklist to the reptiles of the Socotra Archipelago. Fauna of Saudi

Arabia, 20, 505–534.Rüppell, E. (1835) Neue Wirbelthiere zu der Fauna von Abyssinien gehörig. Vol. 3. Amphibien. S. Schmerber, Frankfurt am

Main, 18 pp.Samuel, M.A., Harbury, N., Bott, R. & Manan, A. (1997) Field observations from the Socotran platform: Their interpretation

and correlation to Southern Oman. Marine and Petroleum Geology, 14, 661–673. http://dx.doi.org/10.1016/s0264-8172(96)00033-5

Schätti, B. & Desvoignes, A. (1999) The herpetofauna of southern Yemen and the Sokotra Archipelago. Gilbert-E. Huguet, Genève, 179 pp.

Stamatakis, A. (2006) RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics, 22, 2688–2690. http://dx.doi.org/10.1093/bioinformatics/btl446

Sindaco, R. & Jeremčenko, V.K. (2008) The reptiles of the Western Palearctic. Annotated checklist and distributional atlas of

the turtles, crocodiles, amphisbaenians and lizards of Europe, North Africa, Middle East and Central Asia. Monografie della Societas Herpetologica Italica – I. 579 pp.

Sindaco, R., Ziliani, U., Razzetti, E., Carugati, C., Grieco, C., Pupin, F., Al-Aseily, B.A., Pella, F. & Fasola, M. (2009) A misunderstood new gecko of the genus Hemidactylus from Socotra Island, Yemen (Reptilia: Squamata: Gekkonidae). Acta

Herpetologica, 4, 83–98.Sindaco, R., Metallinou, M., Pupin, F., Fasola, M. & Carranza, S. (2012a) Forgotten in the ocean: systematics, biogeography

and evolution of the Trachylepis skinks of the Socotra Archipelago. Zoologica Scripta, 41, 346–362. http://dx.doi.org/10.1111/j.1463-6409.2012.00540.x

Sindaco, R., Grieco, C. & Riservato, E. (2012b) Hemidactylus homoeolepis. In: IUCN 2013. IUCN Red List of Threatened Species. Version 2013.2. Available from: http://www.iucnredlist.org (accessed 28 April 2014)

Šmíd, J., Carranza, S., Kratochvíl, L., Gvoždík, V., Nasher, A.K. & Moravec, J. (2013a) Out of Arabia: A complex biogeographic history of multiple vicariance and dispersal events in the gecko genus Hemidactylus (Reptilia: Gekkonidae). PloS one, 8 (5), e64018. http://dx.doi.org/10.1371/journal.pone.0064018

Šmíd, J., Moravec, J., Kratochvíl, L., Gvoždík, V., Nasher, A.K., Busais, S.M., Wilms, T., Shobrak, M.Y. & Carranza, S. (2013b) Two newly recognized species of Hemidactylus (Squamata, Gekkonidae) from the Arabian Peninsula and Sinai, Egypt. ZooKeys, 355, 79–107. http://dx.doi.org/10.3897/zookeys.355.6190

Stephens, M., Smith, N.J. & Donnelly, P. (2001) A new statistical method for haplotype reconstruction from population data. The American Journal of Human Genetics, 68, 978–989.

Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M. & Kumar, S. (2011) MEGA5: Molecular evolutionary genetics analysis using Maximum Likelihood, evolutionary distance, and Maximum Parsimony methods. Molecular Biology and

Evolution, 28, 2731–2739. http://dx.doi.org/10.1093/molbev/msr121

Templeton, A.R., Crandall, K.A. & Sing, C.F. (1992) A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III. Cladogram estimation. Genetics, 132, 619–633.

Uetz, P. (2013) The Reptile Database. Available from: http://wwwreptile-databaseorg (accessed 10 December 2013)Van Damme, K. (2009) Socotra Archipelago. In: Gillespie, R.G. & Clague, D.A. (Eds.), Encyclopedia of Islands. University

California Press, Berkeley, CA, pp. 846–851. van der Kooij, J. (2000) The herpetofauna of the Sultanate of Oman. Part 2: the geckos. Podarcis, 1, 105–120.Wilcox, T.P., Zwickl, D.J., Heath, T.A. & Hillis, D.M. (2002) Phylogenetic relationships of the dwarf boas and a comparison of

Bayesian and bootstrap measures of phylogenetic support. Molecular Phylogenetics and Evolution, 25, 361–371.

Zootaxa 3835 (4) © 2014 Magnolia Press · 519SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS

VASCONCELOS & CARRANZA520 · Zootaxa 3835 (4) © 2014 Magnolia Press

Zootaxa 3835 (4) © 2014 Magnolia Press · 521SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS

VASCONCELOS & CARRANZA522 · Zootaxa 3835 (4) © 2014 Magnolia Press

Zootaxa 3835 (4) © 2014 Magnolia Press · 523SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS

VASCONCELOS & CARRANZA524 · Zootaxa 3835 (4) © 2014 Magnolia Press

Zootaxa 3835 (4) © 2014 Magnolia Press · 525SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS

VASCONCELOS & CARRANZA526 · Zootaxa 3835 (4) © 2014 Magnolia Press

APPENDIX III. Loading, standard deviation, proportion and cumulative proportion of variance explained of the seven

components selected from the Principal Components Analysis for each morphological variable. The asterisk (*)

signalling the first component indicates that it is significantly different between the two species from mainland Arabia

and Socotra, and its loadings in bold indicates the most important variables (absolute values above 0.30).

Principal Components

Variable code 1* 2 3 4 5 6 7

Trunk length TRL 0.07 0.30 0.58 0.46 -0.07 0.30 0.33

Head length HL -0.24 0.10 0.17 -0.71 -0.23 0.00 0.29

Head width HW -0.37 0.20 0.16 -0.16 0.17 -0.28 -0.22

Head height HH -0.33 -0.01 -0.27 0.36 0.16 -0.34 -0.09

Orbital diameter OD -0.30 0.03 -0.36 0.15 -0.01 -0.22 0.70

Nostril–eye distance NE -0.33 -0.27 0.10 0.22 -0.07 0.03 -0.05

Anterior interorbital distance IO1 -0.36 0.18 0.12 0.06 0.41 0.20 -0.25

Posterior interorbital distance IO2 -0.35 0.22 0.36 -0.06 0.12 -0.15 0.09

Mean number of supra-labial scales SL 0.12 -0.60 0.26 -0.09 0.59 -0.04 0.34

Mean number of infra-labial scales IL -0.17 -0.50 0.37 0.12 -0.50 -0.32 -0.18

Mean number of lamellae under the first posterior toe LP1st -0.30 -0.29 -0.14 -0.14 0.10 0.55 -0.14

Mean number of lamellae under the fourth posterior toe LP4th -0.35 -0.10 -0.15 0.11 -0.30 0.45 0.14

Standard deviation of components 2.01 1.12 1.11 1.02 0.88 0.87 0.83

Proportion of variance of components 0.34 0.10 0.10 0.09 0.07 0.06 0.06

Cumulative proportion of components 0.34 0.44 0.55 0.63 0.70 0.76 0.82

Zootaxa 3835 (4) © 2014 Magnolia Press · 527SYSTEMATICS OF HEMIDACTYLUS HOMOEOLEPIS