Content uploaded by Akshay Khandekar
Author content
All content in this area was uploaded by Akshay Khandekar on Mar 05, 2019
Content may be subject to copyright.
ORIGINAL ARTICLE
The hills are alive with geckos! A radiation of a dozen species on sky
islands across peninsular India (Squamata: Gekkonidae,
Hemiphyllodactylus) with the description of three new species
Ishan Agarwal
1,2
&Akshay Khandekar
1,2
&Varad B. Giri
1,3
&Uma Ramakrishnan
1
&K. Praveen Karanth
2
Received: 6 August 2018 /Revised: 9 January 2019 /Accepted: 12 February 2019
#Gesellschaft für Biologische Systematik 2019
Abstract
Sky Islands are high-elevation environments that are separated by warmer, low elevations, forming natural patches of unique
montane habitat that often persist through changing climates. Peninsular India was ancestrally forested and has gradually become
more arid since at least the Oligocene, and open landscapes have dominated since the middle-late Miocene. Mesic forests today
are largely restricted to coastal mountains and some other montane habitats. A mitochondrial phylogeny and fossil-calibrated
timetree of Indian Hemiphyllodactylus reveal an Indochinese origin and an endemic radiation with 12 species-level lineages,
where a singlespecies was known,that diversified inthe Oligocene-Miocene across montane forest habitatsin the Eastern Ghats
and south India. The phylogeny also suggests the discontinuous Eastern Ghats mountain range encompasses two distinct
biogeographic entities: north and south of the Pennar/Krishna-Godavari River basins. This study highlights the deep history of
the region and the importance of montane habitats as islands of unique biodiversity that have persisted through millions of years
of changing climates. We describe three new species: Hemiphyllodactylus arakuensis sp. nov., H. jnana sp.nov.andH. kolliensis
sp. nov. from montane habitats above 1000 m. The montane habitats of these species are emerging hotspots of reptile endemism,
and this study emphasizes the need for systematic biodiversity inventory across India to uncover basic patterns of diversity and
distribution.
Keywords Biogeography .Divergence dating .Eastern Ghats .Systematics .Western Ghats
Introduction
The mountains in the tropics have more narrowly distributed
species than at temperate latitudes was first proposed by
Janzen (1967), in part based on the observations that lower
latitudes have less climatic variability in space and time
(McCain 2009). Species on tropical mountains do in fact tend
to have narrower elevational ranges and thermal tolerances,
amongst the drivers of the high richness and endemism in
tropical montane habitats (Ghalambor et al. 2006; Deutsch
et al. 2008;McCain2009). Montane habitats form a natural
system of ‘sky islands’, climatically similar regions that are
physically isolated from each other by warmer, lower eleva-
tions. These upland habitats can act as refugia for cool-
adapted or forest-restricted species through cycling climates
and often harbour lineages with a deep evolutionary history
(e.g. Joshi and Karanth 2010; Tolley et al. 2011;Belletal.
2010; Grismer et al. 2015).
Temperature has an even more overriding effect on tropical
squamates than habitat changes (Frishkoff et al. 2015), and
upslope displacement has been documented with recent
warming (Raxworthy et al. 2008). Tropical forest lizards,
thermoconformers in particular, are especially sensitive to
warm temperatures (Addo-Bediako et al. 2003; Deutsch
et al. 2008; Tewksbury et al. 2008;Hueyetal.2009). Most
gekkonids are nocturnal thermoconformers, one genus with a
largely montane distribution separating deeply divergent
Electronic supplementary material The online version of this article
(https://doi.org/10.1007/s13127-019-00392-5) contains supplementary
material, which is available to authorized users.
*Ishan Agarwal
ishan.agarwal@gmail.com
1
National Centre for Biological Sciences, Tata Institute of
Fundamental Research, Bangalore 560065, India
2
Centre for Ecological Sciences, Indian Institute of Science,
Bangalore 560012, India
3
Foundation for Biodiversity Conservation, A1903, Shubh Kalyan,
Nanded City, Pune 411041, India
Organisms Diversity & Evolution
https://doi.org/10.1007/s13127-019-00392-5
lineages (paleoendemics; sensu Tolley et al. 2011; Grismer
et al. 2015)isHemiphyllodactylus Bleeker, 1860; Agarwal
2016. The genus includes 28 described species patchily dis-
tributed across a vast area including peninsular India and Sri
Lanka, extending through the Andaman and Nicobar Islands
and mainland Southeast Asia east from Myanmar into
Vietnam, Laos, Thailand and peninsular Malaysia up until
the South Pacific (Zug 2010; Chandramouli et al. 2012;
Grismer et al. 2018a,b;Sungetal.2018). Excluding six
insular species in Hong Kong, Palau, the Philippines and
Sumatra, the remaining 22 Hemiphyllodactylus species are
from montane (~ 600–1700 m), forested habitats in mainland
Southeast Asia and India-Sri Lanka (Zug 2010; Grismer et al.
2013,2015,2018a,b;Nguyenetal.2014; Tri et al. 2014;
Sukprasert et al. 2018;Sungetal.2018) including a number
of paleoendemics (Tolley et al. 2011; Grismer et al. 2015).
Hemiphyllodactylus are relatively small geckos (snout to vent
length 31–62 mm, only six species exceeding 50 mm), and as
previously noted, are relatively inconspicuous in colour pat-
tern and often occur at low densities with local distributions
and are consequently poorly known and rarely collected (Zug
2010; Grismer et al. 2013,2015,2018a,b). Zug (2010)pro-
vided a comprehensive review of the taxonomy and nomen-
clature of the genus and described two species, taking the
number of valid Hemiphyllodactylus species to nine.
Including the elevation of three subspecies to species level,
19 new species have been named since Zug (2010)—more
than tripling the species diversity in Hemiphyllodactylus.
This resurgence in species descriptions is a combination of
extensive sampling in mainland Southeast Asia and mitochon-
drial sequences for all described species (Heinicke et al. 2011;
Grismer et al. 2013,2015,2018a,b;Sukprasert et al. 2018).
Hemiphyllodactylus aurantiacus (Beddome, 1870), the so-
le representative of the genus in peninsular India (Bauer &
Das 1999; Zug 2010), was described from Yercaud, Tamil
Nadu at an elevation of ~ 1200 m in the Shevaroys (Fig. 1;
Beddome 1870). The species was considered a subspecies of
the widely distributed Hemiphyllodactylus typus Bleeker,
1860 (Smith 1935), though relatively recent topotypical col-
lections and molecular sequence data confirm its specific sta-
tus (Bauer and Das 1999; Heinicke et al. 2011).
Hemiphyllodactylus aurantiacus is poorly represented in col-
lections, with scattered distributional records from mid-high
elevation forests (~ 600–1400 m) across south India including
the Shevaroys and other isolated massifs, the southern
Western Ghats and the southern Mysore Plateau (Smith
1935; Daniels 1994; Daniels and Kumar 1998;Aengals
2013; Ganesh and Arumugam 2016), as well as the northern
Eastern Ghats from both the Mahendragiri and Nallamala
Hills (Sanyal et al. 1993; Chetri and Bhupathy 2010;Javed
et al. 2010)(Fig.1). The single Sri Lankan record isbased on a
specimen labelled as originating from Hambonota, Sri Lanka,
which may not have originated on the island (Zug 2010).
As part of herpetological surveys and ongoing work on
lizards across the Eastern Ghats and south India from 2009
to 2018, the research community based out of the Centre
for Ecological Sciences and National Centre for Biological
Sciences in Bangalore, and colleagues studying Indian bio-
diversity have accumulated a small collection of
Hemiphyllodactylus from across montane habitats broadly
spanning the known Indian range of the group. Given the
high diversity and endemicity of the genus in other parts of
its range (e.g. see Grismer et al. 2013,2015,2018a,b)and
the isolated nature of the Indian montane localities, we
expected to see some genetic structure. We sequenced the
mitochondrial gene ND2 and used a fossil-calibrated
timetree and biogeographic analyses to establish the timing
and geographic origin of Indian Hemiphyllodactylus and to
evaluate species diversity. Sequence data reveals a deep
split between populations in the northern Eastern Ghats
and those in south India, with a dozen species level line-
ages distributed between the two clades. Here, we describe
three new species, Hemiphyllodactylus jnana sp. nov. and
Hemiphyllodactylus kolliensis sp. nov. from south India
and Hemiphyllodactylus arakuensis sp. nov. from the
northern Eastern Ghats. We also discuss the implications
for Indian biogeography based on Hemiphyllodactylus di-
versity and distribution.
Methods
Phylogenetic analysis
Genomic DNA was extracted from liver or tail tissues stored
in ethanol using the Qiagen DNeasy extraction kit for 18 sam-
ples from 16 localities across peninsular India, including the
type locality of Hemiphyllodactylus aurantiacus (Table S1;
Fig. 1). A partial fragment of the mitochondrial ND2 gene
(756–1041 nucleotides) was amplified with the primers
MetF1 and H5934 and sequenced using MetF1 (Macey et al.
1997). Sequences were obtained commercially from the
NCBS sequencing facility and Medauxin, Bangalore.
Published ND2 sequences for all described and a few un-
named divergent lineages of Hemiphyllodactylus were obtain-
ed from GenBank and trees rooted with Gehyra spp.
(following the results of Heinicke et al. 2011;TableS1).
MEGA 5.2.2 (Tamura et al. 2011) was used to align sequences
and calculate uncorrected percentage pairwise sequence diver-
gence (Table S2). A maximum likelihood (ML) phylogeny
was constructed using the codon-partitioned ND2 data and
the GTR + G model in RA × ML HPC 7.4.2 (Stamatakis
2006) implemented through the raxmlGUI 1.3 (Silvestro and
Michalak 2012), with 10 independent ML runs and support
assessed with 1000 rapid bootstraps.
Agarwal I. et al.
Species delimitation and divergence dating
We did not use tree-based species delimitation as we had
sequences for one or two individuals per population, pre-
cluding a thorough evaluation of within and between spe-
cies branching patterns. Instead, we used an uncorrected
ND2 sequence divergence value of 5% as suggestive of
species level divergence, as has been previously applied
for Hemiphyllodactylus and other geckos (e.g. Grismer
et al. 2013). The alignment with lineages reduced to one
per putative species was included in a larger gekkotan
alignment for divergence dating analyses (Table S1).
Starting xml files were created in BEAUTI 1.8.4
(Drummond et al. 2016) with the HKY + G model applied
to the codon-partitioned ND2 data, a Yule speciation tree pri-
or, relaxed uncorrelated lognormal clock models for each par-
tition and the Diplodactylidae constrained to be sister to the
rest of Gekkota (matching the ML tree). Four fossils were
used as calibration points with exponential distributions and
an arbitrary mean of 5—amber fossils from Myanmar (crown
Gekkota, offset 99; Daza et al. 2016), New Zealand
diplodactylid material (most recent common ancestor (mrca)
New Zealand Diplodactylidae; offset 19; Lee et al. 2009a,b;
Skipwith et al. 2016), Pygopus hortulanus (stem calibration
for mrca Pygopus Merrem, offset 23; Hutchinson 1997;
Jennings et al. 2003; Lee et al. 2009b)andSphaerodactylus
dommeli (mrca Sphaerodactylus Wagler, offset 20; Kluge
1995; Iturralde-Vinent and MacPhee 1996; Heinicke et al.
Fig. 1 Distribution of Hemiphyllodactylus in peninsular India, and inset,
timetree for Indian Hemiphyllodactylus with 95% HPD indicated by bars
at nodes. Lines on the map indicate localities sampled in this study, ‘?’
indicates unsampled published records (from Sanyal et al. 1993;Daniels
1994; Daniels and Kumar 1998; Javed et al. 2010; Ganesh and
Arumugam 2016), major east-flowing rivers marked in blue. Mountain
ranges are marked by two capital letters, AG, Agasthyamalai; AN,
Anaiamalai; BR, Biligirirangana Hills; CM, Cardamamom; JW,
Jawadhi Hills; KC, Kalrayan-Chitteri; KM, Kollimalai; MG,
Meghamalai; MH, Mahendragiri; ML, Malaygiri; MM, Male
Mahadeshwara; MP, Mysore Plateau; NG, Nagari; NL, Nilgiris; NM,
Nallamala; PL, Palani; PM, Pachamalai; SC, Seshachalam; SH,
Shevaroy; SM, Sirumalai; VK, Velikonda; YG, Yelagiri
The hills are alive with geckos! A radiation of a dozen species on sky islands across peninsular India...
2011). An additional geological calibration was used, the di-
vergence of Phelsuma inexpectata Mertens, 1966 on Reunion
from its closest relative on Mauritius P. ornata Gray, 1825
(uniform prior, 0.05–5; Agarwal 2011;Gillotetal.1994;
Rocha et al. 2010). Analyses were run in BEAST 1.8.4
(Drummond et al. 2012,2016) for 50 million generations,
sampling every 5000 generations, log files examined in
Tracer 1.6 (Rambaut et al. 2014) for convergence (ESS >
200) and a maximum clade credibility (MCC) tree was gen-
erated in TreeAnnotator 1.8.4 (Drummond et al. 2016)after
conservatively discarding the first 25% of trees as burn-in.
Original alignments, BEAST xml files or output tree files
are available on request.
Biogeographic analysis
The BEAST MCC tree from the previous analysis was
trimmed to include only Hemiphyllodactylus species, which
were assigned to three subregions of the Indo-Malayan region:
peninsular India, Indochina (mainland Southeast Asia, south-
ern China) and Sundaland (Thai-Malay Peninsula south of the
Isthmus of Kra, through Singapore to Borneo). We used the
dispersal-extinction-cladogenesis (DEC) ML model in
Lagrange 20130526 (Ree and Smith 2008)toestimaterange
evolution in Hemiphyllodactylus. The analyses were set up
using the Lagrange configurator (http://www.reelab.net/
lagrange/configurator), with a maximum of two ancestral
areas allowed and no dispersal constraints. We also treated
distribution as a categorical variable and conducted an ML
ancestral state reconstruction using the Mk1 model in
Mesquite 3.51 (Maddison & Maddison 2018).
Morphological analysis
A total of 36 Hemiphyllodactylus specimens, including 11
topotypical Hemiphyllodactylus aurantiacus and the three
new species, were examined in this study for morphological
data. Museum abbreviations are as follows—Centre for
Ecological Sciences, Bangalore (CES); the Bombay Natural
History Society, Mumbai (BNHS) and the National Centre for
Biological Sciences, Bangalore (NCBS/AQ/AK field series).
Morphological data for the lectotype of H. aurantiacus were
obtained from Zug (2010) and published data for other
Hemiphyllodactylus species from Zug 2010; Grismer et al.
2013,2014a,b,2015,2018a,b; Agarwal et al. 2012;Tri
et al. 2014; Nguyen et al. 2014; Guo et al. 2015;Cobos
et al. 2016; Yan et al. 2016; Sukprasert et al. 2018;Sung
et al. 2018. Photographs were taken with an SLR camera in
life and after preservation; counts and measurements were
taken under a Wild M5 dissecting microscope and on the
right side of the body, where appropriate. The terminology
and methodology used to define characters follow Zug
(2010) and Grismer et al. (2013,2018a). The following
measurements were recorded using a Digimax digital calliper
(rounded off to the nearest 0.1 mm)—snout vent length (SVL,
from tip of snout to vent), tail length (TL, from vent to tip of
tail), tail width (measured at widest point of tail), trunk length
(TRL, from posterior margin of forelimb insertion to anterior
margin of hindlimb insertion), body height (maximum body
height), body width (maximum body width), forearm length
(FL, from elbow to distal end of wrist), crus length (CL, from
knee to heel), head length (HL, distance between posterior
margin of retroarticular process of jaw and snout-tip), head
width (HW, maximum width of head), head height (HH, max-
imum head depth at occiput), eye diameter (ED, greatest hor-
izontal diameter of eye), nares to eye distance (EN, distance
between anterior margin of eye and posterior edge of nostril),
snout to eye distance (SE, distance between anteriormargin of
eye and tip of snout), ear to eye distance (EE, distance from
anterior edge of ear opening to posterior margin of eye), ear
length (maximum length of ear opening), internarial distance
(distance between nares), interorbital distance (shortest dis-
tance between the left and right supraciliary scale rows).
Meristic characters recorded included internasal scales (the
number of scales touching rostral between the supranasals);
number of supralabials (SL) and infralabials (IL), from rostral
and mental, respectively to posterior-most enlarged scale at
angle of the jaw; chin scales (number of scales touching me-
dial edge of infralabials and mental between juncture of 2nd
and 3rd infralabials on the left and right); dorsal scales (ABS,
the number of longitudinal dorsal scales at mid-body
contained within one eye diameter); ventral scales (VS, the
number of longitudinal ventral scales at mid-body contained
within one eye diameter); the number of pore-bearing femoral
scales; the number of pore-bearing precloacal scales; the num-
ber of poreless scales between the precloacal and femoral
pores on each side; the number of subdigital lamellae wider
than long on the first finger and toe; the lamellar formula,
determined as the number of U-shaped, subdigital lamellae
(split and single) on digital pads of digits II–Vand the
basal lamellar formula, the number of subdigital lamellae
wider than long except U-shaped lamellae on the digital
pads on digits II–V.
We carried out principal components analysis (PCA) to
summarise multivariate variation on the meristic and mensural
characters independently, including only adult specimens
(SVL > 29 mm). We excluded pore counts as they are absent
in females and TL and TW due to missing data. Shape was
obtained from untransformed measurements by using resid-
uals from regression against SVL, and SVL was subsequently
dropped from the analysis. These data were then log trans-
formed and scaled to their standard deviation before
performing PCA using the prcomp function in R 3.5.1
(https://www.r-project.org/). This was to demonstrate the
three divergent lineages named in this paper are separated
from each other and H. aurantiacus in morphospace and to
Agarwal I. et al.
confirm the placement of the lectotype of H. aurantiacus.
MissingdataforthelectotypeofH. aurantiacus was
replaced with median values for each character from the
entire dataset.
Results
Phylogeny, species diversity, divergence times
and biogeography
The overall topology matched previously published phyloge-
nies (Grismer et al. 2013,2018a), recovering a basal split
separating the harterti and typus groups, and the seven clades
within the typus group with slightly different interrelation-
ships amongst each other (Fig. 2). Species described subse-
quent to the most recently published comprehensive phylog-
eny (Grismer et al. 2018a) include H. hongkongensis Sung
et al. 2018 from Hong Kong placed as a member of clade 6
(clade names follow Grismer et al. 2018a and are marked in
Fig. 2); H. khlonglanensis Sukprasert et al. 2018 from
Thailand recovered as sister to clade 4; and H. flaviventris
Sukprasert et al. 2018 from Thailand sister to H. sp. nov. 9,
the two sisters to the H. aurantiacus clade (together forming
clade 5). All sampled Indian Hemiphyllodactylus were placed
in a strongly supported H. aurantiacus clade, within which a
deep basal split (up to 22.1% sequence divergence) separates a
clade distributed across the hills of south India (south India
clade) and a clade distributed in the northern Eastern Ghats
(Eastern Ghats clade; Fig. 2). Applying the 5% sequence di-
vergence cutoff, the south India clade includes seven line-
ages—Hemiphyllodactylus aurantiacus,Hemiphyllodactylus
kolliensis sp. nov., H. sp. nov. IN 6; Hemiphyllodactylus jnana
sp. nov. sister to H. sp. nov. IN 5, the two sister to H.sp.nov.
IN 7 and H. sp. nov. IN 8 from the Western Ghats, uncorrected
sequence divergence within the clade 5.7–15.8% (Table S2). The
0.1
Sunda +
Indochina
Indochina
India
Sunda
Hemiphyllodactylus tonywhitteni
Hemiphyllodactylus sp. nov. IN 4
Hemiphyllodactylus yunnanensis (N3)
Hemiphyllodactylus jinpingensis
Hemiphyllodactylus titiwangsaensis
Hemiphyllodactylus sp. nov. IN 8
Hemiphyllodactylus insularis
Hemiphyllodactylus sp. nov. IN 3
Hemiphyllodactylus sp. nov. IN 2
Hemiphyllodactylus flaviventris
Hemiphyllodactylus engganoensis
Hemiphyllodactylus linnwayensis
Hemiphyllodactylus sp. nov. IN 1
Hemiphyllodactylus khlonglanensis
Hemiphyllodactylus larutensis
Hemiphyllodactylus sp. nov. IN 6
Hemiphyllodactylus changningensis
Hemiphyllodactylus sp. nov. 5
Hemiphyllodactylus hongkongensis
Hemiphyllodactylus jnana sp. nov. (BNHS 1937)
Hemiphyllodactylus ganoklonis
Hemiphyllodactylus banaensis
Hemiphyllodactylus montawaensis
Hemiphyllodactylus typus
Hemiphyllodactylus sp. nov. 9
Hemiphyllodactylus cf. titiwangsaensis
Hemiphyllodactylus dushanensis
Hemiphyllodactylus aurantiacus (AK237)
Hemiphyllodactylus bintik
Hemiphyllodactylus huishuiensis
Hemiphyllodactylus kolliensis sp. nov. (CESG068)
Hemiphyllodactylus kolliensis sp. nov. (AK276)
Hemiphyllodactylus sp. nov. IN 7
Hemiphyllodactylus jnana sp. nov. (BNHS 1936)
Hemiphyllodactylus arakuensis sp. nov.
Hemiphyllodactylus sp. nov. 4
Hemiphyllodactylus yunnanensis (N29)
Hemiphyllodactylus jnana sp. nov. (AQ191)
Hemiphyllodactylus chiangmaiensis
Hemiphyllodactylus sp. nov. 2
Hemiphyllodactylus sp. nov. IN 7
Hemiphyllodactylus sp. nov. IN 7
Hemiphyllodactylus aurantiacus (AMB)
Hemiphyllodactylus jnana sp. nov. (CYL01)
Hemiphyllodactylus cicak
Hemiphyllodactylus zugi
Hemiphyllodactylus harterti
Hemiphyllodactylus sp. nov. IN 5
Hemiphyllodactylus kiziriani
Hemiphyllodactylus uga
Hemiphyllodactylus longlingensis
Hemiphyllodactylus tehtarik
harterti group
clades 1 + 2
clades 6 + 7
clade 5
clades 3 + 4
aurantiacus group
typus group
Eastern
Ghats clade
south India
clade
Fig. 2 Maximum likelihood
phylogeny of Hemiphyllodactylus
(ND2) showing ancestral ranges
at key nodes. Nodes with > 90%
bootstrap support marked with
solid black circles, clade names
follow Grismer et al. (2013,2015,
2017), outgroups not shown
The hills are alive with geckos! A radiation of a dozen species on sky islands across peninsular India...
Eastern Ghats clade includes Hemiphyllodactylus arakuensis sp.
nov. sister to H. sp. nov. IN 4, the two sister to an Odisha clade
containing H. sp. nov. IN 2 from Devagiri sister to H. sp. nov. IN
3 from Mahendragiri, the two sister to H. sp. nov. IN 1 from
Daringbadi; uncorrected sequence divergence 5.9–12.5% within
the Eastern Ghats clade (Table S2). The elevational range of
sampled lineages is between 500 m (H. sp. nov. IN 7 from
Coimbatore) and 1400 m (H. aurantiacus from Yercaud).
The divergence times we recovered had similar median
ages to and broadly overlapping 95% HPDs with published
estimates (Grismer et al. 2015), with the basal split within
Hemiphyllodactylus estimated at 50 (56–44) millions of years
ago (mya) (Fig. S1 for complete BEAST tree). The split be-
tween Indian Hemiphyllodactylus and H. sp. nov. 9 +
H. flaviventris was dated to 33 (38–28) mya. The divergence
between the south India clade and Eastern Ghats clade was
estimated at 24 (28–20) mya, diversification within the former
clade beginning at 13 (16–10) mya and within the latter at 8
(10–7) mya (Fig. 1).
The mrca of Hemiphyllodactylus was reconstructed as being
distributed in Indochina and Sundaland (DEC, relative probabil-
ity 0.80) or Sundaland (Mk1, relative probability 0.99), as was
the mrca of the typus group (DEC, Indochina + Sundaland rela-
tive probability 0.85; Mk1, Sundaland relative probability 0.80);
the mrca of clades 3–7 in Indochina (DEC & Mk1 relative prob-
ability 1.00); the mrca of clade 5 as Indochina + India (DEC,
relative probability 1.00) or Indochina (Mk1, relative probability
0.95) and the Indian lineage derived through a single dispersal
event from Indochina around 38–28 mya (Fig. 2).
Morphology
Sampled members of the Hemiphyllodactylus aurantiacus
clade can be diagnosed against all other Hemiphyllodactylus
species by the subdigital lamellar formula of 2222 on the
manus (vs. > 3 lamellae below fingers two to five in all
species, except H. bintik Grismer et al. 2015 and H. cicak
Cobos et al. 2016 that have > 3 subdigital lamellae
only below fingers three and four). The first two principal
components of the shape data (MoPC1–MoPC2) explained a
cumulative76% of variation in the original data, with MoPC1
accounting for 66% variation and MoPC2 an additional 10%
variation, with low factor loadings across a number of char-
acters in each principal component. The first two principal
components of the shape data (MePC1–MePC2) explained a
cumulative 61% of variation in the original data, with MePC1
accounting for 43% variation and MePC2 an additional 18%
variation, with dorsal, ventral and snout scales loading most
strongly on MePC1 and chin scales and 4th toe lamellae on
MePC2. Most of the variation in MoPC1–PC2 is not discrim-
inatory, while MePC1–PC2 provide slightly more separation
but are unable to separate Hemiphyllodactylus aurantiacus
and the three new species completely (Fig. 3). We provide
an expanded diagnosis for Hemiphyllodactylus aurantiacus
and describe three genetically and morphologically divergent
lineages from Bangalore, the Kolli Hills and Araku as new
species below.
Systematics
Hemiphyllodactylus aurantiacus
Hemidactylus aurantiacus (Beddome 1870,p.33)
Shevaroy slender gecko
Lectotype
Adult male (BMNH 74.4.29.1333), from ‘Shevaroys, un-
der stones about Yercaud and elsewhere at an elevation
of 4000 feet’.
−2
−1
0
1
2
−2024
MePC1 (43%)
MePC2 (18%)
−2
−1
0
1
2
3
−5.0 −2.5 0.0 2.5 5.0
MoPC1 (66%)
MoPC2 (10%)
arakuensis
aurantiacus
jnana
kolliensis
Fig. 3 Plot of first two principal
components (PC1–2) for
morphometric (Mo) and meristic
(Me) data showing the position in
morphospace of
Hemiphyllodactylus aurantiacus,
H. arakuensis sp. nov., H. jnana
sp.nov.andH. kolliensis sp. nov.
The holotype of H. aurantiacus is
marked by an asterisk (*)
Agarwal I. et al.
Paralectotypes
BMNH 74.429.1332, 1334–1337, ZMB 10233 (one adult
male, three females and four juveniles), same collection local-
ity as lectotype.
Referred material
Adult males (AK 240–241, AK 244–247) and adult females
(AK 237–239 AK 242–243) from near Grange resort,
Yercaud, (11.775° N 78.219° E; ca. 1390 masl.), Salem dis-
trict, Tamil Nadu state, India, collected by A. Khandekar, I.
Agarwal and N. Gaitonde on 12 September 2018.
Diagnosis
Hemiphyllodactylus aurantiacus can be distinguished from
other members of the genus by a combination of maximum
SVL of 40.9 mm; 9–12 chin scales, postmentals not enlarged,
9–13 SL, 10–13 IL, 13–16 dorsal scales and 9–12 ventral
scales at mid-body contained within one longitudinal eye
diameter; 3–5 subdigital lamellae on the first finger and toe;
lamellar formula of manus 2222; lamellar formula of pes
2333; males with six or seven precloacal scales separated by
8–11unpored scales from a series of six or seven pore-bearing
femoral scales on each thigh; no plate-like enlarged
subcaudals; dark postorbital stripe and longitudinal markings
on nape extending to or beyond forelimb insertions; dorsal
pattern of irregular dark bands or strong reticulations, fine
light paravertebral spots present; postsacral marking with
light-coloured anteriorly projecting arms, belly stippled with
black.
Note
Zug (2010) designated a lectotype and provided a redescrip-
tionofthetypeseries,whileBauer&Das(1999) provided
a diagnosis based on topotypic material. We provide a
summary of the characters in H. aurantiacus based on
our material from the type locality and published data on
the lectotype in Table 1and data for individual specimens
in Table 2.
Table 1 Major diagnostic characters across all Indian Hemiphyllodactylus species. Mean given in parentheses
Species H. aurantiacus H. arakuensis
sp. nov.
H. jnana
sp. nov.
H. kolliensis
sp. nov.
Maximum SVL 41 39 41 39
Chin scales 9–12 8–11 10–12 10–12
Supralabials 9–12 9–12 10–12 9–11
Infralabials 10–13 9–12 9–13 10–11
Internasals 3–42–33–53–4
Dorsal scales 13–16 (14.8) 13–16 (15) 16–20 (18.3) 16
Ventral scales 9–13 (11.1) 7–9(7.9) 11–15 (13.5) 10–13 (11.3)
Lamellae under first finger 3–5 4 or 5 4 or 5 3 or 4
Lamellae under first toe 4 or 5 5 or 6 4 or 5 4 or 5
Lamellar formula of manus 2222 2222 2222 2222
Lamellar formula of pes 2233, 2333 or 3333 2333 or 3333 2222 2222
Precloacal pores 6–78–99–10 9
Scales between precloacal and femoral pores 9–11 11–14 10–12 5
Femoral pores 6–82–36–78
TRL/SVL 0.47–0.51 0.48–0.56 0.50–0.55 0.51–0.55
HL/SVL 0.20–0.22 0.20–0.23 0.19–0.24 0.19–0.21
HW/HL 0.80–0.97 0.75–0.86 0.68–0.85 0.78–0.86
Dark postorbital stripe Y Y Y Y
Light paravertebral spots on trunk Y N Y Y
Dark dorsal transverse bands/blotches Reticulations, sometimes
forming irregular bands
Regular or irregular
bands
Fine paravertebral
reticulations
Paravertebral
reticulations
Extent of longitudinal markings on nape Usually beyond forelimb
insertion
Up to forelimb
insertion
Up to forelimb
insertion
Up to forelimb
insertion
Postsacral marking with light-coloured
anteriorly projecting arms
Y Indistinct Y Y
The hills are alive with geckos! A radiation of a dozen species on sky islands across peninsular India...
Table 2 Data for Hemiphyllodactylus aurantiacus. Asterisk (*) indicates broken tail or measurement/count could not be taken, data for lectotype from Zug (2010)
Hemiphyllodactylus aurantiacus
Lectotype Referred material
Specimen no. BMNH 74.4.29.1333 AK 240 AK 241 AK 244 AK 245 AK 246 AK 247 AK 237 AK 238 AK 239 AK 242 AK 243
Sex M MMMMMMF F F F F
Snout to vent length 34.5 29.2 29.3 32.9 37.9 32.6 32.2 40.9 33 41 40.1 39.1
Tail length * 19.5 19.33 25.4 31.2 20.8 19 11* 2.5* 24.1 24.2 29.9
Tailwidth * 2.22.52.63 3 2.33.63.33.43.72.8
Trunk length 17.5 14.1 14 15.7 18.7 16.6 15.5 21.1 16.8 19.2 20.7 18.8
Body height * 2.5 2.5 2.8 3.6 2.7 3.1 4.8 3.76 5.3 5.7 5.1
Body width * 6 5.66 5.7 7.1 5.6 5.6 8.1 5.6 7.9 7.4 7.5
Head length 7.8 6.5 6.1 6.9 7.8 6.9 6.8 8.1 6.8 8.2 8.1 8
Head width 5.6 5.3 5.9 5.6 7.1 5.7 5.9 7 5.7 6.9 6.5 6.5
Head height * 3.2 3.2 3.2 3.7 3.6 3.6 4.5 3.4 4 4.7 3.9
Forearmlength * 2.82.93.33.73.23.24.13 4 4.13.8
Crus length * 3.5 3.3 3.8 4.4 3.8 3.4 4.6 3.9 4.5 4.7 4.3
Eye diameter 2.0 1.7 1.6 1.7 1.9 1.6 1.6 2 1.8 1.9 1.9 1.9
Nostril to eye distance 2.8 2.1 2.2 2.4 2.7 2.4 2.4 2.8 2.5 2.8 2.8 3
Snout to eye distance 3.7 2.9 3.1 3.2 3.6 2.9 3 3.8 3.2 3.8 3.7 3.9
Eye to ear distance * 2.6 2.5 3 3.3 2.9 3.1 3.3 3 3.4 3.3 3.2
Earlength * 0.50.30.50.40.40.60.70.60.70.70.5
Internarialdistance * 0.91 0.91.11 1.11.10.91.21.21.3
Interorbital distance * 1.8 1.7 1.6 2.4 2 1.6 2 2.1 2.2 2 2
Supralabials (L/R) 13 11/9 12/12 11/10 10/9 12/10 12/11 10/10 11/10 11/11 10/10 10/11
Infralabials (L/R) 11 11/11 12/11 10/12 12/10 13/10 12/12 10/11 12/10 11/11 11/12 12/11
Internasals * 34334433434
Chin scales * 10 9 10 10 10 12 10 10 10 12 10
Dorsal scales 16 16 16 16 14 16 15 13 16 13 14 14
Ventral scales 12 9 13 12 11 12 12 11 12 9 10 11
Femoral pores (L/R) 7 /8 6/6 7/7 7/8 7/7 8/7 7/6 0 6/6 0 0 0
Precloacalpores 7 67767707000
Scales between femoral and precloacal pores (L/R) * 10/10 9/9 10/10 8/6* 10/10 10/10 0 11/9 0 0 0
Lamellae under first finger * 4/4 4/4 4/4 4/4 4/3 4/4 4/5 4/4 4/4 5/4 4/4
Lamellae under first toe * 4/4 4/4 5/5 5/5 4/3 4/4 5/5 4/4 4/4 5/5 5/5
Lamellar formula of manus 2222 2222 2222 2222 2222 2222 2222 2222 2222 2222 2222 2232
Lamellar formula of pes 2333 2333 2333 2333 2333 2333 2333 2333 2333 2233 2233 3333
Basal lamellar formula of manus * 3554 3554 3464 3455 3354 3454 4465 3554 3465 4564 3555
Basal lamellar formula of pes * 3554 4454 3464 3455 4454 3453 3564 3554 3575 3575 3554
Agarwal I. et al.
Distribution
Hemiphyllodactylus aurantiacus is known only from the vi-
cinity of Yercaud, Salem District, Tamil Nadu.
Natural history
Hemiphyllodactylus aurantiacus was described from an ele-
vation of ~ 1200 m (Beddome 1870) and other localities are
also in and around Yercaud (~ 1200–1600 m) (Beddome
1870; Bauer & Das 1999; Ganesh & Arumugam 2015). We
encountered the species after dark along low stone walls in
Grange Resort, Yercaud (~ 1390 m) in high abundance, with
over 25 specimens encountered in about an hour of search.
The vegetation in the Shevaroys varies from scrub at the foot-
hills, transitioning to deciduous forests at middle elevations
with evergreen forests on parts of the plateau (Jayakumar et al.
2009).
Hemiphyllodactylus jnana sp. nov.
Bangalore slender gecko
Holotype
Adult male (AQ 191) collected 26 October 2015 by Rajesh
Sanap and Zeeshan Mirza from the National Centre for
Biological Sciences campus, Bangalore, Bangalore Urban
District, Karnataka, India (13.072° N, 77.581° E, 930 m).
Paratypes
Adult males (AQ 186, AQ 187) and adult females (AQ 188–
190, AQ 192–195) have same collection data as holotype;
adult female (CESL 014) collected on 10 April 2010 by
Saunak Pal and Mrugank Prabhu, from Indian Institute of
Science, Bangalore, Bangalore Urban District, Karnataka
(13.019° N, 77.567° E, 950 m); adult female (G 470) collected
on 30 November 2014 by Aparna Lajmi and Tarun Khichi
from near Kolar, Kolar District, Karnataka (13.134° N,
78.09° E, 1020 m) and adult female (CYL 01) collected on
5 December 2016 by R Krishna Chaitanya from Yelagiri,
Vellore District, Tamil Nadu (12.576° N, 78.654° E, 960 m).
Referred material
Adult male (BNHS 1936) collected on 17 September 2008 by
Jahnavi Joshi from Indian Institute of Science, Bangalore,
Bangalore Urban District, Karnataka (13.019° N, 77.567° E,
950 m) and adult male (BNHS 1937) collected on 14
December 2008 by Varun Torsekar, Henrik Krehenwinkel
and Ishan Agarwal from Aiyur, Hosur District, Tamil Nadu
(12.415° N, 77.841° E, 980 m).
Diagnosis
Hemiphyllodactylus jnana sp. nov. can be distinguished from
other members of the genus by a combination of maximum
SVL of 39.6 mm; 10–12 chin scales; postmentals not en-
larged; 10–12 SL; 9–13 IL; 16–20 dorsal scales and 11–15
ventral scales at mid-body contained within one longitudinal
eye diameter; four or five subdigital lamellae on the first finger
and toe; lamellar formula of manus and pes 2222; males with
nine or ten pored precloacal scales separated by 10–12
unpored scales from a series of six or seven pore-bearing
femoral scales on each thigh; no plate-like subcaudals; fine,
broken, dark postorbital stripe just extending onto trunk and
sometimes extending to tail base; longitudinal markings on
nape usually not extending beyond forelimb insertion; dorsal
colour pattern of indistinct darker reticulations or longitudinal
paravertebral markings, light paravertebral spots present;
postsacral marking without distinct light-coloured anteriorly
projecting arms and belly stippled with black.
Etymology
The specific epithet, jnana (jñāna or nyaa-na), is the Kannada
word for knowledge, derived from the same root in Pali/
Sanskrit and is used as a noun in apposition. The name is
given in honour of two scientific institutions in Bangalore,
the Indian Institute of Science (IISc) and the National Centre
for Biological Sciences (NCBS), within the grounds of which
the species was first found. The Centre for Ecological
Sciences (CES) at IISc and NCBS are at the centre of research
in ecology and evolution in India, and the authors have all
either worked or studied at these institutions.
Description of holotype
Adult male fixed in a slightly sigmoid shape, small vertical
incision on thorax (2.4 mm) for tissue collection. SVL
33.7 mm. Head short (HL/SVL 0.20), slightly elongate
(HW/HL 0.80), slightly depressed (HH/HL 0.52) and distinct
from neck. Prefrontal region flat; canthus rostralis rounded,
snout rounded in dorsal profile. Snout short (SE/HL 0.46),
slightly longer than eye diameter (ED/SE 0.59); scales on
snout, canthus rostralis, prefrontal, inter-orbital and occipital
region homogenous, granular; scales on the snout and canthus
rostralis much larger and pronounced than those on occipital,
prefrontal and inter-orbital regions, two or three rows border-
ing supralabials flat and slightly elongate. Eye small (ED/HL
0.27); pupil vertical with crenulated margins; supraciliaries
small, slightly mucronate,gradually increasing in size towards
front of the orbit, those at the anterior end of orbit larger. Ear
opening sub-circular (greatest diameter 0.5 mm); eye diameter
three-fourth eye to ear distance (ED/EE 0.76). Rostral undi-
vided, wider than deep (length/width 0.50); single large
The hills are alive with geckos! A radiation of a dozen species on sky islands across peninsular India...
supranasal on each side, separated by a slightly smaller single
scale; single large postnasal, subequal in size to supranasals on
each side; rostral in contact with nostril, supralabial I,
supranasals and small scale separating supranasals; nostrils
small, oval; external nares surrounded by rostral, supralabial
I, single postnasal, scale between postnasal and supranasal
(slightly separated but no intervening scales). Mental triangu-
lar, bordered laterally by infralabial 1 on either side and pos-
teriorly by three chin scales; twelve scales along internal edge
of infralabials and mental between juncture of 2nd and 3rd
infralabials on either side. Labials gradually decrease in size
posteriorly, supralabial I and infralabial I largest; supralabials
(to midorbital position) 8 (right)—8 (left) on both sides;
supralabials (to angle of jaw) 11 (right)—11 (left); infralabials
(to angle of jaw) 10 (right)—11 (left).
Body moderately elongate (TRL/SVL 0.50), ventrolateral
folds indistinct. Scales on dorsum of head and neck granular,
smaller than those on the snout and prefrontal region; scales
on body dorsum larger, flat and rounded and 20 scales
contained within one eye diameter at mid-body. Ventral scales
much larger than dorsals, smooth, imbricate and sub-circular,
gradually increasing in size and becoming strongly imbricate
and pointed posteriorly, except five or six rows above cloaca
that are much smaller; 14 scales contained within one eye
diameter at mid-body. Gular scales granular, smaller than ven-
trals and similar in size to dorsals, slightly larger, flat and
juxtaposed on anterior aspect. Scales on palm and sole flat
and rounded; scales on dorsal and ventral aspect of limbs flat
and subimbricate, those on anterolateral aspect of thigh larg-
est. Fore and hind limbs stout, forearm short (FL/SVL 0.09)
and tibia short (CL/SVL 0.11). Digits with well-developed
lamellar pad; first digit on manus and pes vestigial, slender
partially sheathed claw visible under first toe; digits II–Vwell
developed, with free terminal phalange arising from within
lamellar pad, ending in an unsheathed, recurved claw; lamellar
pads of digits II–V with basal series of undivided, transverse
lamellae, and apical series of large U-shaped lamellae, which
are divided/deeply notched, with a single wedge-shaped ter-
minal lamella; four transversely expanded lamellae on digit I
of manus, proximal lamellar formula 2222 (right, left), basal
lamellar formula 3554 (right, left); five transversely expanded
lamellae on digit I of pes, proximal lamellar formula 2222
(right, left) and basal 4355 (right, left).
Tail original except for one-fifth which is regenerated,
not segmented; tail slightly shorter than snout vent length
(TL/SVL 0.85); scales on tail flat, subimbricate, smooth
and rounded; scales on dorsal aspect of tail larger than
those on body dorsum, gradually increasing in size ven-
trally. Two enlarged, subequal postcloacal spurs on both
sides. Nine pored precloacal scales in an angular series
flanked by seven pore-bearing femoral scales on each
side, both series separated by a diastema of 12 (right)
and 10 (left) poreless scales.
Colouration in life (based on BNHS 1936; Fig. S3)
Ground colour of dorsum, head and limbs brown. Indistinct
dark pre-orbital stripe, two broken, fine, dark post-orbital
stripes that extend to tail base, more broken from mid-body
to tail base; a pair of indistinct paravertebral stripes on nape
terminating between forelimb insertions. Dorsum with light
paravertebral spots and indistinct dark longitudinal markings,
head dirty brown without distinct markings. Limbs with indis-
tinct fine dark reticulations. Postsacral marking light brown to
cream, chevron shaped, flanked anteriorly by narrow black
chevron that divides lighter marking; light-coloured anteriorly
projecting arms indistinct. Original tail light brown with irreg-
ular crossbars formed by dark scales flanked by a few light
scales, orange ventrally. Ventral surfaces off-white, with some
dark scales and fine dark spots on lighter scales.
Colouration in preservative (based on holotype NCBS
AQ191; Fig. S4)
Ground colour of dorsum, head and limbs pale brown.
Indistinct dark pre-orbital stripe, two distinct dark post-
orbital stripes that terminate posterior to forelimb inser-
tion, being replaced by indistinct dark reticulations
enclosing white spots along the dorsolateral aspect of
body, a pair of indistinct paravertebral stripes on nape.
Head mottled with few darker markings. Limbs with in-
distinct reticulations. Postsacral marking pale buff,
consisting of a crossbar divided anteriorly by black chev-
ron; no anteriorly projecting arms. Original tail brown
with ~ 5 pairs of indistinct black crossbars posterior to
postsacral marking, pale buff ventrally, regenerated por-
tion spotted. Ventral surfaces off-white, stippled with fine
dark spots on lighter scales and with few darker scales.
Variation (Fig. S5)
Dorsal colour pattern varies from almost uniform (AQ 193–
194) to strong reticulations with numerous lighter spots (AQ
187); 10–12 black markings on original tail, regenerated tail
mottled brown; amount of stippling on the ventrum variable.
Female paratypes without pores, with prominently swollen
white endolymphatic sacs, most prominent in AQ 188–190,
AQ 192–194. Variation in important scalation and measure-
ments is given in Table 3.
Distribution
Hemiphyllodactylus jnana sp. nov. is known from Bangalore,
Bangalore Urban District and Kolar, Kolar District in
Karnataka, and Aiyur, Hosur District and Yelagiri, Vellore
District in Tamil Nadu. These naturally occurring populations
were previously misidentified as H. aurantiacus and their
Agarwal I. et al.
disjunct distribution in urban areas around Bangalore led to the
speculation that these were accidentally introduced by humans
(Daniels 1994; Bauer & Das 1999;Zug2010). The species is still
seen in Bangalore City, and an uncollected adult male was ob-
servedonMcPhersonRoad,CookeTownon22February2019.
Natural history
Hemiphyllodactylus jnana sp. nov. has been recorded from
scattered localities on the southern edge of the Mysore
Plateau at elevations between 930 and 1020 m, across a range
of habitats that include a mixture of scrub and degraded decid-
uous forest, besides plantations and other human modified hab-
itats (also see Giri et al. 2009;Agarwal2016). The species has
been encountered from an hour after dark to ~ 6 h after sunset,
mainly on trees and walls in and around human habitation as
well as on the ground. The species has been seen during the day
beneath tree bark and in crevices in trees at the type locality
(Rajesh Sanap, pers obs), besides in leaf litter and twice on a
shaded road. The new species is common at its type locality, in
Table 3 Type series of Hemiphyllodactylus jnana sp. nov. Asterisk (*) indicates broken tail or measurement/count could not be taken
Hemiphyllodactylus jnana sp. nov.
Holotype Paratypes
Specimen no. AQ 191 AQ 186 AQ 187 AQ 188 AQ 189 AQ 190 AQ 192 AQ 193 AQ 194 AQ 195 CYL 01
Sex M MMFFFFFFFF
Snout to vent length 33.7 33.7 30.4 39.6 38.1 37 38.2 34.9 38.3 31.5 38.4
Tail length 28.8 24.5 10.7* 26.7 38.3 23.8 25.4 25.0 3.2* 4.4* 34
Tail width 2.9 2.7 2.5 3.1 3.2 3.2 2.8 2.6 2.7 2.4 2.6
Trunk length 15.2 18.1 15.3 20.8 19.6 19.2 19.5 18.8 19.7 16.7 21.3
Body height 2.6 3.3 2.4 3.8 2.8 3.8 2.8 3.6 3 3.6 3.2
Body width 5.7 6.3 5.3 8.5 6.8 7 6.4 6.2 7.3 5.7 7.2
Head length 6.9 7.3 7 7.5 7.7 7.8 7.8 7.3 7.8 7.6 7.7
Head width 5.5 5.7 5.6 6.2 6.3 6.2 6.6 5.6 6.2 5.2 5.8
Head height 3.6 3.2 2.7 3.7 3.6 3.9 3.9 3.5 3.9 3.3 3.6
Forearm length 3.3 3.4 3 3.9 3.7 3.3 3.8 3.5 3.6 2.9 3
Crus length 4.0 3.8 3.5 4.9 4.4 4.0 4.2 3.8 4.1 3.8 4.1
Eye diameter 1.9 1.8 1.7 1.9 1.8 2.0 2.0 1.8 2.0 1.7 2.0
Nostril to eye distance 2.9 2.5 2.2 2.8 2.5 2.7 3.0 2.5 2.9 2.4 2.4
Snout to eye distance 3.2 3.1 2.9 3.5 3.5 3.5 3.7 3.3 3.5 3.5 3.3
Eye to ear distance 2.5 2.7 2.5 2.9 3.1 2.9 3.2 2.9 2.9 2.6 3.1
Ear length 0.5 0.4 0.4 0.4 0.5 0.6 0.7 0.5 0.3 0.3 0.4
Internarial distance 1.0 1.0 1.1 1.1 1.2 1.2 1.3 1.1 1.2 1.0 1.2
Interorbital distance 2 2.3 2 2.6 2.5 2.5 2.4 2.2 2.7 2.2 1.8
Supralabials (L/R) 11/11 11/11 11/11 12/11 11/12 10/11 11/11 10/11 10/11 12/11 11/11
Infralabials (L/R) 11/10 10/10 11/11 11/11 11/11 9/11 11/12 11/11 11/10 12/11 10/9
Internasals 4 4444354543
Chin scales 6/6 5/5 6/5 6/6 6/5 5/5 5/6 5/5 5/5 5/5 5/6
Dorsal scales 20 18 19 17 18 16 20 20 19 18 17
Ventralscales 14 13151214141413141413
Femoralpores(L/R) 7/77/76/600000000
Precloacalpores 9 91000000000
Scales between femoral and
precloacal pores (L/R)
10/1212/1210/1100000000
Lamellae under first finger 4/4 4/4 4/4 4/4 4/4 4/4 4/4 4/4 4/4 4/4 4/4
Lamellae under first toe 5/5 4/4 5/4 4/4 5/4 4/4 5/4 5/5 4/4 5/5 4/4
Lamellar formula of manus 2222 2222 2222 2222 2222 2222 2222 2222 2222 2222 2222
Lamellar formula of pes 2222 2222 2222 2222 2222 2222 2222 2222 2222 2222 2222
Basal lamellar formula of manus 3554 3455 3453 4565 4555 3554 4564 *465 3555 3465 4454
Basal lamellar formula of pes 4355 3545 3554 3554 4554 4554 4565 3555 3545 4455 4454
The hills are alive with geckos! A radiation of a dozen species on sky islands across peninsular India...
the NCBS lawns and adjacent GKVK campus which include
mainly exotic tree species, and we have encountered 10–12
individuals in about two man hours of search.
Comparisons
Hemiphyllodactylus jnana sp. nov. can be distinguished from
H. aurantiacus by the presence of two lamellae under toes
four and five (vs. three lamellae under toes four and five),
more precloacal pores in males (mean 9.3, range nine or 10
vs. 6.7, six orseven) and more dorsal scales(mean 18.3, range
16–20 vs. 14.8, 13–16) and ventral scales (mean 13.5, range
11–15 vs. 11.1, 9–13). Hemiphyllodactylus jnana sp.nov.is
compared against H. kolliensis sp. nov. and H. arakuensis sp.
nov. after their descriptions. Major diagnostic characters
across all Indian Hemiphyllodactylus species are scored in
Table 1.
Hemiphyllodactylus kolliensis sp. nov.
Kolli slender gecko
Holotype
Adultmale(CESG138)collectedon23March2011by
Aniruddha Datta-Roy and Ishan Agarwal from near Selur
Nadu, Kolli Hills (known locally as Kollimalai), Namakkal
District, Tamil Nadu, India (11.219° N, 78.354° E, 1170 m).
Paratypes
Adult females (AK 276–278) have same collection data as
holotype except, collected on 20 April 2011, AK 277 from
near Semmedu (11.268° N, 78.321° E, 1180 m) and AK 278
from near Thinnanurnadu (11.242° N 78.330° E, 1100 m).
Diagnosis
Hemiphyllodactylus kolliensis sp. nov. can be distinguished
from other members of the genus by a combination of maxi-
mum SVL of 39.3 mm; 10–12 chin scales; postmentals not
enlarged; 9–11 SL; 10 or 11 IL; 16 dorsal scales and 10–13
ventral scales at mid-body contained within one longitudinal
eye diameter; three or four subdigital lamellae on the first
finger and four or five on toes; lamellar formula of manus
and pes 2222; males with 10 pored precloacal scales separated
by 5 unpored scales from a series of 8 pore-bearing femoral
scales on each thigh; no plate-like subcaudals; dark postorbital
stripe just extending onto trunk; longitudinal markings on
nape extending beyond forelimb insertion; dorsal pattern of
few scattered light spots and indistinct reticulations or dark
paravertebral reticulations enclosing light paravertebral spots;
postsacral marking with light-coloured anteriorly projecting
arms and belly stippled with black.
Etymology
The specific epithet is a toponym for the type locality of the
species, the Kolli Hills (known locally as Kollimalai).
Description of holotype
Adult male fixed with the head bent towards the left, tail
missing. SVL 32.7 mm. Head short (HL/SVL 0.21), slightly
elongate (HW/HL 0.78), slightly depressed (HH/HL 0.47) and
distinct from neck. Prefrontal region flat; canthus rostralis
rounded, snout rounded in dorsal profile. Snout short (SE/
HL 0.45), slightly longer than eye diameter (ED/SE 0.54);
scales on snout, canthus rostralis, prefrontal, inter-orbital and
occipital region homogenous, granular; scales on the snout
and canthus rostralis much larger than those on occipital, pre-
frontal and inter-orbital regions, two or three rows bordering
supralabials flat and slightly elongate. Eye small (ED/HL
0.25); pupil vertical with crenulated margins; supraciliaries
small, slightly mucronate,gradually increasing in size towards
front of the orbit, those at the anterior end of orbit larger. Ear
opening sub-circular (greatest diameter 0.3 mm); eye diameter
~ two-thirds eye to ear distance (ED/EE 0.68). Rostral undi-
vided, wider than deep (length/width 0.42); a single large
supranasal on each side, separated by three slightly smaller
scales; single large postnasal, slightly smaller than
supranasals; rostral in contact with nostril, supralabial I,
supranasals and three small scales separating supranasals;
nostrils small, oval; external nares surrounded by rostral,
supralabial I, single postnasal, scale between postnasal and
supranasal. Mental triangular, bordered laterally by infralabial
I on either side and posteriorly by three chin scales; twelve
scales along internal edge of infralabials and mental between
juncture of 2nd and 3rd infralabials on either side. Labials
gradually decrease in size posteriorly, supralabial I and
infralabial I largest; supralabials (to midorbital position) 7
(right)—8 (left); supralabials (to angle of jaw) 10 (right)—9
(left); infralabials (to angle of jaw) 11 (right)—10 (left).
Body moderately elongate (TRL/SVL 0.51), ventrolateral
folds indistinct. Scales on dorsum of head and neck granular,
smaller than those on the snout and prefrontal region; scales
on body dorsum larger, flat and rounded; 16 scales contained
within one eye diameter at mid-body. Ventral scales much
larger than dorsals, smooth, imbricate and sub-circular in out-
line, gradually increasing in size and becoming strongly im-
bricate and pointed posteriorly, except 4–5 rows above cloaca
that are much smaller; 11 scales contained within one eye
diameter at mid-body. Gular scales granular, smaller than ven-
trals and similar in size to dorsals, slightly larger, flat and
juxtaposed on anterior aspect. Scales on palm and sole flat
and rounded; scales on dorsal and ventral aspect of limbs flat
and subimbricate, those on anterolateral aspect of thigh larg-
est. Fore and hind limbs stout; forearm short (FL/SVL ratio 0.
Agarwal I. et al.
09); tibia short (CL/SVL ratio 0.11). Digits with well-
developed lamellar pad; first digit on manus and pes vestigial,
slender partially sheathed claw visible under first toe; digits
II–V well developed, with free terminal phalange arising from
within lamellar pad, ending in an unsheathed, recurved claw;
lamellar pads of digits II–V with basal series of undivided,
transverse lamellae and apical series of large U-shaped lamel-
lae, which are divided/deeply notched, with a single wedge-
shaped terminal lamella; four transversely expanded lamellae
on digit I of manus, proximal lamellar formula 2222 (right,
left), basal lamellar formula 3453 (right, left); five(right) and
four (left) transversely expanded lamellae on digit I of pes,
proximal lamellar formula 2222 (right, left) and basal 3565
(right, left).
Tail original, two-thirds missing, rounded, not forming
caudal segments; scales on dorsal aspect of tail larger
than those on body dorsum, gradually increasing in size
ventrally. Two enlarged, subequal postcloacal spurs on
the left and three on the right. Ten precloacal scales with
pores in an angular series flanked by eight femoral pore-
bearing scales on each side, both series separated by a
diastema of five poreless scales.
Colouration in life (Fig. S6)
Ground colour dark coffee, faintly suffused with dark rust
orange, more prominently on the head. Faint orange stripe
on each side running from snout through eye, along dorsolat-
eral margin of body and coming together at tail base; a pair of
indistinct paravertebral stripes on nape extending to forebody.
Limbs and body with small orange spots, most formed by 2–4
orange scales. Postsacral marking saffron, with saffron anteri-
orly projecting arms, dark medial spot and two broad mark-
ings coveringtail dorsum. Original tail with alternating orange
and dark colouration dorsally, dark markings laterally and
orange white ventrally, regenerated portion uniformly mottled
brown black. Throat strongly stippled with dark spots and
belly with numerous dark scales, underside of limbs and
precloacal region strongly pigmented.
Colouration in preservative (Fig. S7)
Ground colour brown. Dull light stripe on each side running
from snout and through eye, until beyond forelimb insertions,
dark, indistinct pre-orbital and two postorbital streaks, a pair
of indistinct paravertebral stripes on nape. Limbs with small
light spots. Postsacral marking off-white, with prominent
light-coloured anteriorly projecting arms and dark medial
spot. Throat strongly stippled with dark spots and belly with
numerous dark scales, underside of limbs and precloacal re-
gion strongly pigmented.
Variation (Fig. S8)
The three female paratypes lack precloacal or femoral
pores but are otherwise similar to the holotype in
scalation and measurements (Table 4). Paratypes lighter
in ground colouration and with more prominent dark
dorsal markings than holotype, nuchal stripes extending
past forelimb insertions, indistinct dark reticulations
enclosing lighter spots on flanks, amount of stippling
on ventrum variable.
Distribution
Hemiphyllodactylus kolliensis sp. nov. is known from three
localities within 10 km of each other in the Kolli Hills,
Namakkal District, Tamil Nadu, India.
Natural history
The Kolli Hills are an isolated massif about 40–60 km
south of the Shevaroys in southern peninsular India,
which rise steeply from a base elevation of ~ 300 to ~
1400 m, forming an undulating plateau above ~ 1000 m.
The vegetation varies from scrub at the foothills,
transitioning to deciduous forests at middle elevations
with evergreen forests on parts of the plateau
(Jayakumar et al. 2009). Hemiphyllodactylus kolliensis
sp. nov. has been collected between 1930 and 2230 h
at elevations of 1100–1180 m. The holotype was collect-
ed on the low wall of a small culvert, one paratype was
on the ground near a stone wall, one was in leaf litter
and another about 5 cm off the ground on a small shrub.
The three paratypes were collected on a single night. All
the habitats we collected this species from are human
modified, a mix of coffee plantations, farmlands and
small towns and villages interspersed with small patches
of degraded evergreen and moist deciduous forests. We
did not see any Hemiphyllodactylus on the walls of the
resortatwhichwestayedin~6hofsearchingacross5–
6 nights though there were numerous Cnemaspis sp. both
during the day and night.
Comparisons
Hemiphyllodactylus kolliensis sp. nov. can be distinguished
from Hemiphyllodactylus aurantiacus by the presence of 2
lamellae under toes four and five (vs. 3 lamellae under toes
four and five), greater number of pore-bearing precloacal
scales in males (9 vs. 6–7) and fewer unpored scales separat-
ing the pore-bearing precloacal and femoral scales on each
side (5 vs. 9–11) and a slightly more elongate trunk (TRL/
SVL 0.51–0.55 vs. 0.47–0.51). Hemiphyllodactylus kolliensis
sp. nov. differs from Hemiphyllodactylus jnana sp. nov. in
The hills are alive with geckos! A radiation of a dozen species on sky islands across peninsular India...
having more pore-bearing femoral scales in males (8 vs. 6–
7) and fewer unpored scales separating the pore-bearing
precloacal and femoral scales on each side (5 vs. 10–12),
fewer dorsals (16 vs. mean 18.3, range 16–20) and fewer
ventrals (mean 13.5, range 11–15 vs. 11.3, 10–13).
Hemiphyllodactylus kolliensis sp. nov. is compared against
Hemiphyllodactylus arakuensis sp. nov. after its descrip-
tion. Major diagnostic characters across all Indian
Hemiphyllodactylus species are recorded in Table 1.
Hemiphyllodactylus arakuensis sp. nov.
Araku slender gecko
Holotype
Adult male (G 446) collected on 2 March 2014 by Aparna
Lajmi, V. Deepak and Aniruddha Datta-Roy from Forest Rest
House, Araku, VisakhapatnamDistrict, Andhra Pradesh,India
(18.326° N, 82.877° E, 940 m).
Paratypes
Five adult females (G 442–G 444, G 448–449), two adult
males (G 445, G 447), have same collection data as holotype.
Table 4 Type series of
Hemiphyllodactylus kolliensis sp.
nov. Asterisk (*) indicates broken
tail
Hemiphyllodactylus kolliensis sp. nov.
Holotype Paratypes
Specimen no. CES G138 AK 276 AK 277 AK 278
Sex M F F F
Snout to vent length 32.7 37.3 39.3 38.3
Tail length 4.6* 11 16.9* 32.4
Tail width 2.6 2.7 2.8 2.7
Trunk length 16.8 19.6 21.7 19.8
Body height 2.7 2.8 4.5 4.7
Body width 6.6 6.4 7.3 7.4
Head length 6.8 7.4 7.6 7.6
Head width 5.3 6.0 6.1 6.6
Head height 3.2 3.7 3.7 4
Forearm length 3.1 3.3 3.3 3.4
Crus length 3.7 4.1 4.1 4.2
Eye diameter 1.7 1.8 2.0 2.0
Nostril to eye distance 2.4 2.4 2.5 2.7
Snout to eye distance 3.1 3.2 3.3 3.5
Eye to ear distance 2.5 2.9 3.1 2.9
Ear length 0.3 0.5 0.5 0.4
Internarial distance 1.0 1.1 1.1 1.1
Interorbital distance 2.0 2.1 1.8 2.2
Supralabials (L/R) 9/10 10/10 10/10 11/10
Infralabials (L/R) 10/11 10/10 10/11 11/11
Internasals 3 3 4 3
Chin scales 6/6 5/5 5/5 5/5
Dorsal scales 16 16 16 16
Ventral scales 11 10 13 11
Femoral pores (L/R) 8/8 0 0 0
Precloacal pores 9 0 0 0
Scales between femoral and precloacal pores (L/R) 5/5 0 0 0
Lamellae under first finger 4/4 3/3 4/3 4/4
Lamellae under first toe 4/5 4/4 5/4 5/4
Lamellar formula of manus 2222 2222 2222 2222
Lamellar formula of pes 2222 2222 2222 2222
Basal lamellar formula of manus 3453 2353 3554 3453
Basal lamellar formula of pes 3565 3545 4555 3555
Agarwal I. et al.
Referred material
Adult male (BNHS 2275) collected on 18 September 2014 by
Aniruddha Datta-Roy, Tarun Khichi and Ishan Agarwal from
near Araku (18.255° N, 82.991° E, 1170 m).
Diagnosis
Hemiphyllodactylus arakuensis sp. nov. can be distinguished
from other members of the genus by a combination of maxi-
mum SVL of 39.0 mm; 8–11 chin scales; postmentals not
enlarged; 9–12 SL; 9–12 IL; 13–16 dorsal scales and 7–9
ventral scales at mid-body contained within one longitudinal
eye diameter; four or five subdigital lamellae on the first finger
and five (rarely six) on toes; lamellar formula of manus 2222;
lamellar formula of pes 2333/3333; males with eight or nine
precloacal scales separated by 11–14 unpored scales from a
series of two or three pore-bearing femoral scales on each
thigh; no plate-like enlarged subcaudals; dark postorbital
stripe and longitudinal markings on nape extending to fore-
limb insertions; dorsal pattern of 7–9 irregular dark bands or
strong reticulations, light paravertebral spots very fine if pres-
ent; postsacral marking with light-coloured anteriorly
projecting arms and belly stippled with black.
Etymology
The specific epithet is a toponym for the type locality of the
species, Araku.
Description of holotype
Adult male fixed with tail-tip bent towards the left and
everted hemipenes. SVL 30.1 mm. Head short (HL/SVL
0.23), slightly elongate (HW/HL 0.77), slightly depressed
(HH/HL 0.49) and distinct from neck. Prefrontal region
flat; canthus rostralis rounded, snout rounded in dorsal
profile. Snout short (SE/HL 0.45), slightly longer than
eye diameter (ED/SE 0.52); scales on snout, canthus
rostralis, prefrontal, inter-orbital and occipital region ho-
mogenous, granular; scales on the snout and canthus
rostralis much larger than those on prefrontal, occipital
and inter-orbital regions, two or three rows bordering
supralabials flat and slightly elongate. Eye small (ED/HL
ratio 0.23); pupil vertical with crenulated margins;
supraciliaries small, slightly mucronate, gradually increas-
ing in size towards front of the orbit, those at the anterior
end of orbit larger. Ear opening sub-circular (greatest di-
ameter 0.6 mm); diameter of eye greater than half eye to
ear distance (ED/EE 0.57). Rostral undivided, wider than
deep (length/width 0.54), a single large supranasal on each
side, separated by two slightly smaller scales; single large
postnasal on each side, slightly smaller in size than
supranasals; rostral in contact with nostril, supralabial I,
supranasals and two small scales separating supranasals;
nostrils small, oval; external nares surrounded by rostral,
supralabial I, single postnasal, scale between postnasal and
supranasal. Mental triangular, bordered laterally by
infralabial I on either side and posteriorly by two slightly
larger chin scales; ten scales touching internal edge of
infralabials and mental between juncture of 2nd and 3rd
infralabials on either side. Labials gradually decrease in
size posteriorly towards angle of jaw, supralabial I and
infralabial I largest; supralabials (to midorbital position) 7
(right)—7 (left); supralabials (to angle of jaw) 10 (right)—
10 (left); infralabials (to angle of jaw) 10 (right)—9 (left).
Body moderately elongate (TRL/SVL ratio 0.49), ven-
trolateral folds indistinct. Dorsal scales of head and neck
granular, smaller than those on snout and prefrontal re-
gion, those on the body dorsum slightly larger than the
rest, flat and rounded; 15 scales contained within one eye
diameter at mid-body. Ventral scales much larger than
dorsals, smooth, imbricate and sub-circular, gradually in-
creasing in size and becoming strongly imbricate and
pointed posteriorly, except five or six rows above cloaca
that are much smaller; eight scales contained within one
eye diameter at mid-body. Gular scales granular, smaller
than ventrals and similar in size to dorsals, slightly larg-
er, flat and juxtaposed on anterior aspect. Scales on palm
and sole flat and rounded; scales on dorsal and ventral
aspect of limbs flat and subimbricate, those on anterolat-
eral aspect of thigh largest. Fore and hind limbs stout;
forearm short (FL/SVL ratio 0.10); tibia short (CL/SVL
ratio 0.12). Digits with well-developed lamellar pad; first
digit on manus and pes vestigial, slender partially
sheathed claw visible under first toe; digits II–Vwell
developed, with free terminal phalange arising from
within lamellar pad, ending in an unsheathed, recurved
claw; lamellar pads of digits II–V with basal series of
undivided, transverse lamellae and apical series of large
U-shaped lamellae, which are divided/deeply notched,
with a single wedge-shaped terminal lamella; four trans-
versely expanded lamellae on digit I of manus, proximal
lamellar formula 2222 (right, left), basal lamellar formula
3564 (right, left); five transversely expanded lamellae on
digit I of pes, proximal lamellar formula 2333 (right,
left) and basal 4552 (right, left).
Tail original, rounded, no segments; tail slightly longer
than snout vent length (TL/SVL ratio 1.17); scales on tail flat,
subimbricate, smooth and rounded; scales on dorsal aspect of
tail larger than those on body dorsum, gradually increasing in
size ventrally. Two enlarged, subequal postcloacal spurs on
the left and three on right. Nine precloacal scales bearing pores
in an angular series flanked by three (right) and two (left)
pore-bearing femoral scales, both series separated by a diaste-
ma of 13 (right)—14 (left) poreless scales.
The hills are alive with geckos! A radiation of a dozen species on sky islands across peninsular India...
Colouration in life (Fig. S9)
Ground colour of dorsum, head and limbs coffee. Indistinct
dark pre-orbital stripe, two distinct dark post-orbital stripes
that terminate at forelimb insertions. Dorsum with 9 narrow
black crossbars with a fine posterior margin of light scales,
head mottled with darker markings. Limbs with indistinct
crossbars, reticulations and small orange spots. Postsacral
marking saffron, chevron shaped, flanked anteriorly by black
chevron that just divides saffron marking, no anteriorly
projecting arms. Original tail orange with ~ 10 pairs of indis-
tinct black crossbars posterior to postsacral marking, bright
orange ventrally. Ventral surfaces off-white, stippled with fine
dark spots on lighter scales.
Colouration in preservative (Fig. S10)
Ground colour of dorsum, head and limbs brown. Indistinct
dark pre-orbital stripe, two distinct dark post-orbital stripes
that terminate posterior to forelimb insertion. Dorsum with 9
narrow black crossbars, head mottled with darker markings.
Limbs with indistinct crossbars and reticulations and small
pale orange spots. Postsacral marking pale buff, chevron
shaped, flanked anteriorly by black chevron; no anteriorly
projecting arms. Original tail off-white with ~ 10 pairs of in-
distinct black crossbars posterior to postsacral marking, pale
buff ventrally. Ventral surfaces off-white, stippled with fine
dark spots on lighter scales.
Variation (Fig. S11)
Dorsal pattern varies from more or less well-formed crossbars
to strong reticulations, regenerated tail with few black spots.
There is little variation in scale counts, the number of
infralabials and supralabials varies from nine (G 443) to 12
(G 445); precloacal pores are absent in females and the num-
ber in males varies between eight and nine (Table 5).
Distribution
Hemiphyllodactylus arakuensis sp. nov. is only known from
in and around its type locality Araku, Visakhapatnam District,
Andhra Pradesh.
Natural history
Araku is a small town at ~ 940-m elevation in the
Mahendragiri Hill Range, just 80 km from the east coast in
Visakhapatnam District, Andhra Pradesh. The elevation
ranges from < 100 m at the foothills to > 1500 m, with mainly
deciduous forests and some evergreen species at higher eleva-
tions. Hemiphyllodactylus arakuensis sp. nov. has been col-
lected from two localitiesat 940 m and 1170 m around Araku.
The collection locality in Araku was the compound of the
Forest Department Rest House, with mainly non-native
vegetation including many large Eucalyptus trees, while
the other locality was near a small riparian patch of forest
in a coffee plantation (also see Agarwal et al. 2012;Datta-
Roy et al. 2013). Geckos were spotted after dark on the
base of trees (< 1 m high), rocks and walls and a few were
on the ground. This species was locally abundant at its
typelocalitywith~10specimensfoundbyIAin1hof
searching after dark.
Comparisons
Hemiphyllodactylus arakuensis sp. nov. can be distinguished
from Indian congeners by males having two or three pore-
bearing femoral scales on each thigh (vs. 6–8in
Hemiphyllodactylus aurantiacus,sixorseveninH. jnana and
8inH. kolliensis), 7–9 ventral scales at mid-body contained
within one longitudinal eye diameter (9–13 in H. aurantiacus,
11–15 in H. jnana and 10–13 in H. kolliensis) and lamellar
formula of toes 2333 or 3333 (vs. 2222 in H. jnana and
H. kolliensis). Major diagnostic characters across all Indian
Hemiphyllodactylus species are scored in Table 1.
Discussion
Species diversity
The discovery of three new species and as many as eight
additional unnamed deeply divergent Hemiphyllodactylus lin-
eages from south India and the Eastern Ghats (descriptions of
additional species in prep) adds to the list of Indian endemic
radiations (Karanth 2015), highlighting the deep diversity of
peninsular India and how poorly this is known. The land-
scapes within which Hemiphyllodactylus are distributed are
emerging as centres of reptile endemism—the type locality
for three geckos are in the vicinity of Bangalore,
Hemidactylus graniticolus Agarwal et al., 2011,
Cyrtodactylus (Geckoella)srilekhae Agarwal, 2016 and
Cnemaspis mysoriensis (Jerdon, 1853; Agarwal et al. 2013)
(see Giri et al. 2009), the latter two restricted to nearly the
same region as Hemiphyllodactylus jnana sp. nov. The Kolli
Hills has a few undescribed Cnemaspis Strauch, a
Cyrtodactylus Gray and a Hemidactylus Oken (Agarwal,
Khandekar and Giri, in prep) besides an endemic uropeltid
snake Rhinophis goweri Aengals and Ganesh, 2013;Addo-
Bediako et al. 2003, and Yercaud is the type locality for
Cnemaspis yercaudensis Das and Bauer, 2000;Agarwal
et al. 2015, an endemic uropeltid snake Uropeltis shorttii
(Beddome, 1863; Aengals 2013) (Ganesh et al. 2014) and
has a number of undescribed Cnemaspis and a
Cyrtodactylus (Agarwal unpubl. data; Ganesh and
Agarwal I. et al.