ArticlePDF Available

Abstract

A known haven of amphibian diversity, South Asia is also a hotspot of taxonomic confusions. Vastly distributed from Saudi Arabia to Myanmar, the dicroglossid genus Euphlyctis (“skittering” or “skipper” frogs) is a representative example. Combining phylogenetic analyses with 16S barcoding and genome size variation of 403 frogs from 136 localities, we examined genetic diversity and distributions across the whole range of Euphlyctis, with a particular focus on taxonomic and nomenclatural issues. We recovered two deeply divergent mitochondrial clades totalling ten lineages that we considered as species, and eight could be attributed valid taxonomic names and junior synonyms. The first clade (subgenus Phrynoderma) is confirmed in South India, Bangladesh and Sri Lanka, and comprises six species: E. karaavali, E. hexadactyla, E. aloysii, E. kerala and two undescribed taxa. Five are endemic to the Western Ghats biodiversity hotspot and four of them form the E. aloysii species complex. The second clade (subgenus Euphlyctis) extends across South Asia and neighbouring regions, and comprises four species: E. ehrenbergii, E. jaladhara, and two widespread lineages erroneously called “E. mudigere” and “E. kalasgramensis” in recent literature, while their oldest valid names appear to be E. cyanophlyctis and E. adolfi, respectively. Additional analyses on this pair of taxa highlighted strong phenotypic resemblance, notable intraspecific phylogeographic structure, and an extensive contact zone along the southern slopes of the Himalaya, with putative signs of genetic introgression. Through an independent investigation of the historical literature, we identified overlooked issues and misconceptions regarding the status of many old and recent taxa, and proposed solutions, such as transferring “E. ghoshi” to the genus Limnonectes. Our study illustrates how range-wide genetic barcoding can clarify taxonomic confusions, and we call to solve remaining issues prior to the description of new taxa
Research Article
Shedding light on taxonomic chaos: Diversity and distribution of
South Asian skipper frogs (Anura, Dicroglossidae, Euphlyctis)
CHRISTOPHE DUFRESNES
1
, STEPHEN MAHONY
2,3
, VISHAL KUMAR PRASAD
4,5
,
RACHUNLIU G. KAMEI
3,6
, RAFAQAT MASROOR
7
, MUAZZAM ALI KHAN
8
,
AWADH M. AL-JOHANY
9
, KUMUDANI BALA GAUTAM
5
, SANDEEP KUMAR GUPTA
5
,
LEO J. BORKIN
10
, DANIEL A. MELNIKOV
10
, JURIY M. ROSANOV
11
, DMITRIY V. SKORINOV
11
,
AMAËL BORZÉE
4
, DANIEL JABLONSKI
12
& SPARTAK N. LITVINCHUK
11,13
1
LASER, College of Biology and the Environment, Nanjing Forestry University, Nanjing, Peoples Republic of China
2
School of Biology and Environmental Science, University College Dublin, Dublin 4, Ireland
3
Department of Life Sciences, The Natural History Museum, London, UK
4
Laboratory of Animal Behaviour and Conservation, College of Biology and the Environment, Nanjing Forestry University,
Peoples Republic of China
5
Wildlife Institute of India, Chandrabani, Uttarakhand, India
6
566 Kohima-Meriema Road (AH1), Meriema, Kohima, 797001, Nagaland, India
7
Zoological Sciences Division, Pakistan Museum of Natural History, Garden Avenue, Shakarparian, 44000, Islamabad Pakistan
8
Department of Botany, Bacha Khan University Charsadda, Charsadda, KP, Pakistan
9
Department of Zoology, King Saud University, Riyadh, 11451, Kingdom of Saudi Arabia
10
Laboratory of Herpetology, Zoological Institute, Russian Academy of Sciences, Saint Petersburg, 199034, Russia
11
Institute of Cytology, Russian Academy of Sciences, Saint Petersburg, Russia
12
Department of Zoology, Comenius University, in Bratislava, Mlynsk
a dolina, Ilkovi
cova 6, Bratislava, 842 15, Slovakia
13
Dagestan State University, Makhachkala, Russia
(Received 31 March 2022; accepted 13 July 2022)
A known haven of amphibian diversity, South Asia is also a hotspot of taxonomic confusions. Vastly distributed from
Saudi Arabia to Myanmar, the dicroglossid genus Euphlyctis (skitteringor skipperfrogs) is a representative
example. Combining phylogenetic analyses with 16S barcoding and genome size variation of 403 frogs from 136
localities, we examined genetic diversity and distributions across the whole range of Euphlyctis, with a particular focus
on taxonomic and nomenclatural issues. We recovered two deeply divergent mitochondrial clades totalling ten lineages
that we considered as species, and eight could be attributed valid taxonomic names and junior synonyms. The first clade
(subgenus Phrynoderma) is confirmed in South India, Bangladesh and Sri Lanka, and comprises six species: E.
karaavali,E. hexadactyla,E. aloysii,E. kerala and two undescribed taxa. Five are endemic to the Western Ghats
biodiversity hotspot and four of them form the E. aloysii species complex. The second clade (subgenus Euphlyctis)
extends across South Asia and neighbouring regions, and comprises four species: E. ehrenbergii,E. jaladhara, and two
widespread lineages erroneously called E. mudigereand E. kalasgramensisin recent literature, while their oldest
valid names appear to be E. cyanophlyctis and E. adolfi, respectively. Additional analyses on this pair of taxa
highlighted strong phenotypic resemblance, notable intraspecific phylogeographic structure, and an extensive contact
zone along the southern slopes of the Himalaya, with putative signs of genetic introgression. Through an independent
investigation of the historical literature, we identified overlooked issues and misconceptions regarding the status of many
old and recent taxa, and proposed solutions, such as transferring E. ghoshito the genus Limnonectes. Our study
illustrates how range-wide genetic barcoding can clarify taxonomic confusions, and we call to solve remaining issues
prior to the description of new taxa.
Key words: Euphlyctis adolfi,Euphlyctis cyanophlyctis, genetic barcoding, genome size, phylogeography, systematics,
Phrynoderma, taxonomy, Western Ghats biodiversity hotspot
Correspondence to: Christophe Dufresnes. E-mail:
Christophe.Dufresnes@hotmail.fr
ISSN 1477-2000 print / 1478-0933 online
#The Trustees of the Natural History Museum, London 2022. All Rights Reserved.
https://dx.doi.org/10.1080/14772000.2022.2102686
Systematics and Biodiversity (2022), 20(1): 2102686
Published online 30 Aug 2022
Introduction
Assessing the geographic extent of newly discovered
species is necessary to provide first-hand evaluations of
their conservation status and ascertain their taxonomic
identity. As genuine evolutionary lineages are increas-
ingly considered to be valid species by taxonomists,
including when external differences are not obvious
(cryptic taxa), species identification using molecular
methods (genetic barcoding) has become an important
part of biodiversity surveys (Borkin et al., 2004;
Mahony et al., 2020; Padial et al., 2010; Vences et al.,
2005a). In turn, reliable knowledge of lineage distribu-
tions and divergence can provide key insights onto the
biogeography of species and regions.
In this respect, the diversity and evolution of a sub-
stantial part of the herpetofauna of South Asia (India
and neighbouring countries), which encompasses several
of Earths biodiversity hotspots (Myers et al., 2000:
Meegaskumbura et al., 2002; Roelants et al., 2004: Das
et al., 2006), remains poorly understood. In particular,
amphibian taxonomy has been studied by naturalists
since colonial times (Chanda et al., 2001;G
unther,
1860,1864), but genetic studies have highlighted strong
discordances with previous taxonomic arrangements and
assumptions regarding the distributions of many species
(e.g. Chen et al., 2005). This taxonomic chaosstems
from various causes. First, as the phylogeographic litera-
ture increasingly suggests, the genetic diversity of South
Asian amphibians remains insufficiently documented
(e.g. Vijayakumar et al., 2016). Second, amphibian sur-
veys in South Asia still essentially rely on phenotypic
identification in the field. However, some species appear
more polymorphic than previously assumed, in turn
causing frequent misidentification and thus erroneous
conclusions on distribution ranges (Dinesh et al., 2021).
Third, when genetic barcoding is implemented, it often
targets amphibian communities of a given region (e.g.
Akram et al., 2021; Hasan et al., 2012,2014), not entire
species ranges, and relies only on mitochondrial DNA
(mtDNA). Yet, clarifying the taxonomic identity of spe-
cies sometimes requires range-wide sampling that
encompasses the type localities of all available nomina.
In addition, due to the high propensity of cyto-nuclear
discordance among amphibian species complexes (e.g.
ghost mitochondrial lineages or super-cryptic species;
Dufresnes et al., 2019,2020), barcoding analyses are
more informative when combining mitochondrial with
nuclear-encoded markers.
In respect to the above issues, we aimed to clarify lin-
eage distribution and taxonomy in a complex group of
South Asian anurans, the dicroglossid genus Euphlyctis
Fitzinger, 1843, also known as skittering frogsor
skipper frogs.Widespread from the Arabian
Peninsula, across desert areas of Iran and Afghanistan,
to tropical Myanmar and Sri Lanka (Akram et al., 2021;
Dinesh et al., 2021; Khatiwada et al., 2021; Wagner
et al., 2016; Zug, 2022), most populations have long
been attributed to two species groups, namely the
Euphlyctis cyanophlyctis (Schneider, 1799) group and
the Euphlyctis hexadactyla (Lesson, 1834) group. Both
groups have been reported to differ externally, form
monophyla that diverged since the Mid- or Early-
Miocene, and occur in sympatry over large parts of their
ranges (Alam et al., 2008; Dinesh et al., 2021; Dubois
et al., 2021; Dufresnes & Litvinchuk, 2022; Howlader
et al., 2015). In their attempt to name every supraspe-
cific amphibian clade, Dubois et al. (2021) raised the
two groups as separate genera, respectively Euphlyctis
(E. cyanophlyctis group) and Phrynoderma Fitzinger,
1843 (E. hexadactyla group). Using molecular
approaches, especially 16S ribosomal DNA sequences, a
myriad of lineages was revealed within both groups
(Akram et al., 2021; Alam et al., 2008; Dinesh et al.,
2021; Khajeh et al., 2014) and new allopatric species
were formally described from Bangladesh and India
(Dinesh et al., 2021,2022; Howlader et al., 2015; Joshy
et al., 2009). Taxonomists now list nine species, more
than half reported within the last 15 years (Frost, 2022).
An inherent challenge in Euphlyctis taxonomic and
conservation research is that allopatric species are mor-
phologically very similar (Joshy et al., 2009), especially
for sexually immature individuals, which complicates
species identification without genetic barcoding. As a
consequence, the respective distributions of taxa remain
largely elusive, which casts doubts on the identity of
type populations. For instance, several cases of syn-
onymy were suspected in the E. cyanophlyctis group,
such as Euphlyctis mudigere Joshy, Alam, Kurabayashi,
Sumida, and Kuramoto, 2009 with E. cyanophlyctis
(Dinesh et al. 2021); Euphlyctis kalasgramensis
Howlader, Nair, Gopalan, and Meril
a, 2015 with
Euphlyctis cyanophlyctis seistanica (Nikolskii, 1899)
(Dinesh et al., 2021,asEuphlyctis seistanica) or with
Euphlyctis adolfi (G
unther, 1860) (Borkin et al., 2020;
Dufresnes & Litvinchuk, 2022; Litvinchuk et al., 2017).
Some presumed species are yet to receive a molecular
assessment, like Euphlyctis ghoshi (Chanda, 1991),
while the strong mitochondrial divergence documented
within others are perhaps indicative of species com-
plexes in need of taxonomic attention, like Euphlyctis
aloysii Joshy, Alam, Kurabayashi, Sumida, and
Kuramoto, 2009 (Dinesh et al., 2021). In parallel, the
status of historical taxa such as Rana bengalensis Gray,
1830 and Rana leschenaultii Dum
eril and Bibron, 1841
remains unsolved (Dinesh et al., 2021), and these may
2 C. Dufresnes et al.
be found to have priority over several of the recently
discovered and described lineages.
To get a more comprehensive picture of the
Euphlyctis diversity, distribution and taxonomy across
South Asia, here we integrate new and available
molecular barcoding data (16S) from the ranges of all
known lineages. We further conduct phylogeographic
and morphometric analyses of the two most widespread
species, which have been the sources of much confusion
in recent publications. Finally, we provide an independ-
ent examination of the historical literature in order to
identify and propose solutions to overlooked taxonomic
issues that are presently threatening the stability of sev-
eral species.
Methods
16s rRNA barcoding, phylogeny
and taxonomy
For mtDNA barcoding, 81 tissue samples from 39 local-
ities across India, Nepal, Iran, Pakistan and Saudi Arabia,
were obtained from voucher specimens (thigh muscles)
or live individuals (toe clips, blood, buccal swabs), pre-
served in 7096% ethanol. DNA was isolated using
DNAeasy Blood and Tissue kits (Qiagen). For 74 sam-
ples, a 550bp fragment of the 16S rRNA gene was
amplified using universal primers 16SALand16SBH
(Palumbi et al., 2002;Vencesetal.,2005b), and
sequenced in one direction (16SAL). For seven samples
(series RGKin Supplemental material - File S1a), a
15002000 bp fragment encompassing 12S-tVal16S
was obtained using the primers pairs LX12SN1(f) and
LX16S1R(r), and 12SAL(f) and 16S2000H(r) (Zhang
et al., 2008), and sequenced in both directions. Sequences
were quality-checked in MEGA X (Kumar et al., 2018)
and uploaded in GenBank under accession numbers
ON818920ON818993 (16S) and ON814172ON814178
(12S-tVal16S). We then aligned and trimmed the
sequences manually in Seaview 5 (Gouy et al., 2010),
which involved adding 13 bp gaps at a few positions.
To complement this dataset, we searched for
Euphlyctis 16S sequences available on GenBank (Sayers
et al. 2022). A total of 182 sequences were retrieved as
of March 2022 (Supplemental material - File S1a), orig-
inating from 35 sources 19 associated to published
studies (Akram et al., 2021; Alam et al., 2008,2010;
Ali et al., 2020; Al-Qahtani & Amer, 2019; Anoop
et al., 2017; Bossuyt & Milinkovitch, 2000; Dinesh
et al., 2021; Hasan et al., 2012,2014; Howlader et al.,
2015; Joshy et al., 2009; Khajeh et al., 2014; Khatiwada
et al., 2021; Kosuch et al., 2001; Kotaki et al., 2010;
Kurabayashi et al., 2005; Priti et al., 2016; Singh &
Prakash, 2006). Among these, 165 sequences could be
accurately or approximately georeferenced. The remain-
ing 17 were also included in order to consider all the
genetic variation documented in Euphlyctis, and to
crosscheck their taxonomic identity. The final alignment
(480 bp) thus consisted of 263 Euphlyctis sequences,
including 246 georeferenced sequences from 104 local-
ities (Supplemental material - File S2). We first identi-
fied the main lineages from a preliminary tree estimated
in PhyML 3.2 (Guindon et al., 2010) and collapsed
identical sequences into unique haplotypes. Net pairwise
divergences between and within lineages were computed
in MEGA, with default settings. Sampled localities were
then mapped in QGIS 3.4 (QGIS Development Team,
2018) and visualized according to their 16S lineages.
In a second step, we reconstructed and dated phylo-
genetic divergences among 16S haplotypes, adding a
sequence of the closely related dicroglossid species
Fejervarya cancrivora (Gravenhorst, 1829) as the out-
group (AB070738, Sumida et al., 2002). We performed
a Bayesian analysis with BEAST 2.5 (Bouckaert et al.,
2019), using a strict clock and a birth-death tree prior.
Although our phylogenetic dataset is limited to a single
gene, we implemented a time calibration prior in order
to get general insights about divergence times in
Euphlyctis, and to retrieve a rough estimation of the 16S
clock rate for downstream demographic analyses.
Specifically, we used a normally distributed prior for
the split between the E. hexadactyla and E. cyanophlyc-
tis groups, estimated at 20 ± 3 million years (My) in
the dicroglossid phylogeny of Alam et al. (2008). The
chain was run for 10 million iterations, which was
enough to reach stationarity and independently sampled
trees, as confirmed by inspection of the log file in
Tracer 1.7 (Rambaut et al., 2018).
In addition, to explore the relevance of the identified
lineages in species delimitation, we applied the parti-
tioning algorithm of ASAP (Puillandre et al., 2021)
included in the iTaxoTools software collection (Vences
et al., 2021). We explored the solutions obtained with
(Jukes-Cantor, Kimura K80) or without substitution
models to compute genetic distances, and retained the
ones with the lowest ASAP score.
Finally, we provisionally assigned 16S lineages to
taxonomic names based on an investigation of the taxo-
nomic literature, especially the historical publications,
museum records (notably online registries), and on the
examination of relevant scientific collections and curated
specimens by one of us (SM). Type localities and
nomenclatural validity of older and junior synonyms
were then critically assessed to update and discuss taxo-
nomic arrangements. As part of this effort, the following
museum catalogue number prefixes were used. BMNH:
Diversity and distribution of South Asian skipper frogs 3
Natural History Museum, London; BNHS: Bombay
Natural History Society, Mumbai; CAS: California
Academy of Sciences; MASB: Museum of the Asiatic
Society of Bengal; MNHNP, MNHN and NMNHP:
Museum National dHistoire Naturelle, Paris; NMSL:
Sri Lanka National Museum: ZMB: Universit
at
Humboldt, Zoologisches Museum, Berlin; ZSI:
Zoological Survey of India, Kolkata.
Ranges and diversity of widespread taxa
Additional analyses were performed to get preliminary
insights into the phylogeography of the two most wide-
spread taxa, identified as E. adolfi and E. cyanophlyctis
(see Results). First, to complement the mtDNA barcod-
ing, we assessed range limits based on genome sizes. To
this end, 151 individuals from 37 localities were anaes-
thetized (immersion 1% MS222) and sampled for blood.
Genome size (nuclear DNA content) was then measured
by DNA flow cytometry following the procedures in
Borkin et al. (2001). All the sampled specimens are cura-
ted at the Institute of Cytology of the Russian Academy
of Science (INCRAS). Their origins and taxonomic iden-
tity (based on 16S barcoding for several samples) are
detailed in Supplemental material - File S1a.Variationin
genome size, expressed in picograms (pg), was compared
between species, averaged over populations, and mapped
using a diagnostic threshold (see Results).
Second, the 16S dataset was used to build haplotype
networks for the species identified as E. adolfi (n¼154
sequences) and E. cyanophlyctis (n¼51 sequences) with
the R package pegas, under default settings (Paradis,
2010). For each species, haplogroups were identified,
and mapped across geographic ranges with QGIS. In
addition, we inferred coalescent Bayesian Skyline Plots
(BSP) in BEAST (Drummond et al., 2005) to model
demographic fluctuations, in which we applied the clock
rate (0.005 substitution per site per My) obtained from
the time-calibrated phylogeny (see above). The chain
was run for 10 million iterations, and skyline plots were
produced with Tracer.
Finally, building up on the newly inferred ranges of
E. adolfi and E. cyanophlyctis, we compared their
morphology based on 11 characters typically used in the
Euphlyctis literature (Joshy et al., 2009). To this end,
one of us (SNL) measured 48 specimens (representative
of 17 localities) from the Zoological Institute of St.
Petersburg (ZISP) and INCRAS collections, using a
digital caliper to the nearest 0.1 mm. Sexual maturity
and sex were determined by direct examination of
gonads. Measurements comprised snout-vent length
(SVL), head length, from the tip of the snout until the
posterior edge of the mandibles (HL), head width,
between the posterior edges of the mandibles (HW),
snout to nostril distance, from the snout tip to the anter-
ior edge of the nostril (SN), inter-nostril distance (NN),
nostril to eye distance, from the posterior edge of the
nostril to the anterior bony orbital border (NE), eye
diameter, measured horizontally between the bony
orbital borders (ED), inter-orbital distance, as the min-
imum distance between orbits (EE), tympanum diameter
(TD), femur length, from mid cloacal opening to the
knee (FEL) and shank length (TIL). We then computed
body ratios for the latter ten characters (character/SVL),
and included five additional ratios, namely HL/HW, SN/
NE, TD/ED, NN/EE and TIL/FEL (as in Joshy et al.,
2009). A PCA and a MANOVA were performed on
these 15 values in R 3.6, with the specific aims of
assessing how the morphology varied with respect to
sex and species, identifying the contributing variables,
and eventually designing diagnostic criteria.
Furthermore, we combined the SVL measured for the
48 ZISP/INCRAS specimens with an additional 32
adults of known origin and sex (16 newly measured by
SLN and RGK, 16 taken from Howlader et al., 2015;
Joshy et al., 2009; Prasad et al., 2020), and performed
ANOVAs to assess the effect of sex and species. Raw
measurements and specimen origins are detailed in
Supplemental material - File S1b.
Results
Genetic barcoding and taxonomic
arrangements
Across the 263 Euphlyctis 16S sequences analyzed
(480 bp), we identified 60 unique haplotypes forming 10
clear lineages, which we mapped across 104 localities
(Fig. 1;Supplemental material - Files S1a). The haplo-
type alignment and full tree with terminal branches are
available in Supplemental material - Files S2S3,
respectively. The best species delimitation partition
obtained with ASAP distinguished eight species (Fig. 1)
with an ASAP score of 3.0 (ranked 2
nd
and 4
th
for the
P-val and W indices, respectively), while the second
best partition (10 species) received a score of 5.0. The
same results were obtained with different models of
sequence evolution.
Combining lineage distributions (circles in Fig. 1)
with georeferencing of known type localities (stars in
Fig. 1) and an examination of the historical literature,
we critically assigned available taxonomic names to
most of the lineages by applying the International Code
of Zoological Nomenclatures principle of priority rule
(ICZN, 1999)(Table 1). As developed in the following,
we uncovered a multitude of taxonomic and
4 C. Dufresnes et al.
nomenclatural issues that appear to have not been previ-
ously reported in the Euphlyctis literature.
Gender of Euphlyctis Fitzinger, 1843.Euphlyctis taxa
alternatively appear feminine (e.g. Euphlyctis cyano-
phlyctis seistanica) or masculine (e.g. Euphlyctis hexa-
dactylus) in the literature, and it seems that the
etymology and gender of the name have never been
clarified. Among the first modern uses of Euphlyctis as
a genus, Dubois (1992) listed Euphlyctis hexadactylus,
thus assuming that it was masculine, a conjugation sub-
sequently followed until nowadays.
In Browns(1954)Composition of Scientific
Words,the prefix euand suffix phlyctisare seem-
ingly derived from Greek, and are defined as follows:
eu- <Gr. eu, good, well, agreeable, easy, very, true,
original, primitive(p. 309).
L. phlyctaena (Gr. phlyktaina;phlyktis,-idos), f.
blister, pustule; phlyzakion, n. dim.(p. 169).
Fitzinger (1843) did not provide any indication of an
etymology for Euphlyctis, and his description may have
largely been based on the description of the type species
Rana leschenaultii Dum
eril and Bibron, 1841. Although
this taxon (considered as a synonym of E. cyanophlyctis,
but see below) can hardly be considered a very tubercu-
lated frog, Dum
eril & Bibron (1841) did mention small
warts/tubercles in the description and elsewhere, for
instance Corps sem
e de petites
eminences coniques, et
perc
e de pores dispos
es en lignes parcourant le cou, les
c^
ot
es du dos et le ventre [Body strewn with small con-
ical eminences, and pierced with pores arranged in lines
running along the neck, the sides of the back and the
belly](p. 342). Based on these observations, and from
the variety of meanings given above for the prefix eu,
we thus interpret that the Greek name Euphlyctis was
probably intended as a translation for very pustular.
Moreover, as indicated by the abbreviation f.in
Brown (1954)s definition, the suffix phlyktis
is feminine.
In conclusion, Euphlyctis should be considered femin-
ine, and we hereby implement the proper spelling to the
relevant taxa.
Genus-level split within Euphlyctis.The 16S phylogeny
recovered the E. cyanophlyctis and E. hexadactyla
groups as two strongly supported clades, estimated to be
20 My old in Alam et al. (2008), an estimation re-
used here for the molecular calibration (Fig. 1). Dinesh
et al. (2021) provided a morphological diagnosis of the
two groups that comprised broadly overlapping size
ranges and the presence or absence of longitudinal
Fig. 1. Diversity and distribution of Euphlyctis. The tree shows phylogenetic relationships among 16S lineages, with statistical
support (posterior probability) for internal branches (:>0.95) and 95% highest probability density of divergence times (grey bars).
Species delimitation with ASAP are shown, based on the best partition (eight species). The maps illustrate lineage distribution, as
inferred from 16S barcoding (circles) and genome size (squares). Note that the Arabian E. ehrenbergii is displayed on the small top
inset (yellow circles). Type localities of taxa are represented by stars: filled stars labelled with upper case letters for valid names;
empty stars labelled with lower case letters for junior synonyms, as follows. Euphlyctis adolfi (A), with junior synonym E.
kalasgramensis (b); E. cyanophlyctis (C, presumed location), with junior synonym E. mudigere (d) and subspecies E. c. flavens (e),
E. c. fulvus (f) and E. c. typicus (g); E. hexadactyla (H), with junior synonym Rana robusta (i). The affiliation of subspecies E. c.
microspinulata (j) and E. c. seistanica (k), as well as the type localities (and the corresponding species) of the nomina R. leschenautii
and R. bengalensis remain unsettled (see text). See Supplemental material - File S1a for locality and haplotype information, and
Supplemental material - File S3 for the full 16S tree. Photo credit: SNL.
Diversity and distribution of South Asian skipper frogs 5
stripes on the dorsum and flanks. We find the presence
or absence of stripes unreliable based on observations of
variation within several species, e.g. E. hexadactyla and
E. adolfi individuals with no visible stripes, and exam-
ples of all species with stripes on the flanks (SM pers.
obs.). Recently, Dubois et al. (2021) recognized these
two groups as separate genera, Euphlyctis (¼E. cyano-
phlyctis group) and Phrynoderma (¼E. hexadactyla
group), solely on the basis that their sister relationship
was poorly supported among other closely related
Table 1. Summary of the taxonomy and distribution of Euphlyctis species.
Lineage Taxon Synonymsor subspecies Distribution
Subgenus Euphlyctis
yellow Rana ehrenbergii Peters, 1863 none Yemen and Saudi Arabia
brown Euphlyctis jaladhara Dinesh,
Channakeshavamurthy,
Deepak, Shabnam, Ghosh, and
Deuti, 2022
none Western Ghats
red Dicroglossus adolfi, 1860 Euphlyctis kalasgramensis
Howlader, Nair, Gopalan, and
Meril
a, 2015
Iran, southern Afghanistan,
Pakistan, Nepal, northern
India, southern Bhutan,
Bangladesh,
Western Myanmar
orange Rana cyanophlyctis
Schneider, 1799
Rana cyanophlyctis var.
typicus De Silva, 1958
Rana cyanophlyctis var. fulvus
De Silva, 1958
Rana cyanophlyctis var.
flavens De Silva, 1958
Euphlyctis mudigere Joshy,
Alam, Kurabayashi, Sumida,
and Kuramoto, 2009
most of India, Sri Lanka,
Eastern Pakistan
Subgenus Phrynoderma
purple Euphlyctis karaavali Priti,
Naik, Seshadri, Singal,
Vidisha, Ravikanth, and
Gururaja, 2016
none Western Ghats
blue Rana hexadactyla
Lesson, 1834
Dactylethra bengalensis
Lesson, 1835
Rana cutipora Dum
eril &
Bibron, 1841
Rana robusta Blyth, 1855
eastern India, Sri
Lanka, Bangladesh
dark green Euphlyctis aloysii Joshy,
Alam, Kurabayashi, Sumida,
and Kuramoto, 2009
none Western Ghats (Karnataka)
pale green Euphlyctis kerala Dinesh,
Channakeshavamurthy,
Deepak, Ghosh, and
Deuti, 2021
none Western Ghats (Kerala)
dashed dark green none (cf. aloysii)Western Ghats (Karnataka)
dashed pale green none (cf. aloysii)Western Ghats (Goa)
Unsettled taxa
red / orange / purple / dark
green / pale green / dashed
dark green / dashed pale green
Rana leschenaultii Dum
eril
and Bibron, 1841
Bengal and perhaps
Pondicherry and/or
Malabar Coast
red / orange Rana bengalensis Gray, 1830
Rana cyanophlyctis var.
seistanica Nikolskii, 1899
Euphlyctis cyanophlyctis
microspinulata Khan, 1997
Bengal
Neizar in Seistano, Iran
southwest of Khuzdar in
Pakistanese Balochistan
Color codes refers to the Fig. 1.
6 C. Dufresnes et al.
dicroglossids. However, their phylogeny did not provide
evidence of paraphyly, included only five Euphlyctis
and Phrynoderma species, and the authors warned that
our terminals are composite, often consisting of gene
sequences from more than one specimenand While
we have attempted to minimise incorrectly labelled
sequences from Genbank (see 4.3.1.2 below), there is
no doubt that not all our speciesare holophyletic, as a
few of them are likely to be hybrid populations or clus-
ters of cryptic species, and that some specimens have
been misidentified.(p. 294). Due to the widespread
misidentification of GenBank sequences in Euphlyctis
(see below) and the lack of resolution of this part of
Dubois et al.s(2021) tree, the proposed taxonomic rear-
rangements appear based on little evidence, which we
consider potentially highly destabilizing. Accordingly,
their split does not meet two out of the three priority
taxon naming criteria (TNCs) formalized by Vences
et al. (2013), namely phenotypic diagnosability and
clade stability. For the latter, at least four subcriteria are
violated: robust clade support (ii), absence of evidence
for non-monophyly (iii), dense taxon sampling (iv) and
support by independent data sets (v). We therefore do
not recognize the two genera within the long-established
genus Euphlyctis, but instead suggest that Euphlyctis
and Phrynoderma may be used as subgenera for the
simplification of discussion in the scientific literature, as
implemented below.
Subgenus Euphlyctis (E. cyanophlyctis group).The
subgenus Euphlyctis forms a strongly supported clade
composed of four notable lineages, all partitioned as dif-
ferent species in the ASAP analysis (Fig. 1,Table 1):
1. An Arabian lineage barcoded in Yemen and Saudi
Arabia (yellow-coded in Fig. 1). It corresponds to
Rana ehrenbergii Peters, 1863, the only name
available in the region, now Euphlyctis ehrenbergii
(Peters, 1863).
2. A southern lineage restricted to the southwestern
coastal plains of India (brown-coded in Fig. 1). This
lineage has been associated with the recently
described Euphlyctis jaladhara Dinesh,
Channakeshavamurthy, Deepak, Shabnam, Ghosh,
and Deuti, 2022, which shares the same type locality
(Thattekad Bird Sanctuary)asEuphlyctis kerala
Dinesh, Channakeshavamurthy, Deepak, Ghosh, and
Deuti, 2021, a more distantly related lineage from
subgenus Phrynoderma (see below). The brown-
coded lineage was accordingly barcoded in
Thattekad (Dinesh et al., 2021), but given the
information provided in the GenBank records and
the description of E. jaladhara (Dinesh et al. 2022),
it is unclear whether these sequences correspond to
the type specimens.
3. A northern lineage extending from Iran to
Bangladesh along a narrow corridor flanking the
Himalayas (red-coded in Fig. 1), which requires
taxonomic revisions. As shown by the barcoding of
three topotypes (Supplemental material - File S1a),
this lineage inhabits the type locality of
Dicroglossus adolfi G
unther, 1860 (Kulu and
Simla, Himalaya (24004200 feet above the level of
the sea),now situated in Himachal Pradesh State,
India). The original description (G
unther, 1860)
mentioned considerable variation within the species,
based on a type series that contained multiple
specimens, of which four syntypes are currently
available (BMNH 1947.2.28.46 [previously BMNH
60.3.19.1365], BMNH 1947.2.28.47 [previously
BMNH 60.3.19.1367], BMNH 1947.2.4.60 and
1947.2.4.61 [both previously BMNH 60.11.6.5]; SM
pers. obs.). Boulenger (1882) has been credited with
synonymizing this name under Rana cyanophlyctis
Schneider, 1799 (e.g. Frost, 2022), although he
provided no explanation for this action. The name
was subsequently forgotten or considered invalid by
many authors (e.g. Baig et al., 2008; Howlader
et al., 2015; Masroor, 2011), and the same
biological species was recently described again as
Euphlyctis kalasgramensis Howlader, Nair, Gopalan,
and Meril
a, 2015, from a type series collected in
Bangladesh (Howlader et al., 2015).
As previously proposed (Borkin et al., 2020;
Dufresnes & Litvinchuk 2022; Litvinchuk et al.,
2017), the oldest available nomen that can be
assigned with certainty to the red-coded lineage is
Euphlyctis adolfi (G
unther, 1860). Given the lack of
genetic differentiation between E. adolfi topotypes
and the E. kalasgramensis type series, we formally
consider Euphlyctis kalasgramensis Howlader, Nair,
Gopalan, and Meril
a, 2015 as a junior subjective
synonym of E. adolfi.
4. A widespread lineage distributed in India, Sri Lanka
and Western Pakistan (orange-coded in Fig. 1),
which occurs at the currently accepted type localities
of Rana cyanophlyctis Schneider, 1799,Rana
bengalensis Gray, 1830,Rana leschenaultii Dum
eril
and Bibron, 1841 and Euphlyctis mudigere Joshy,
Alam, Kurabayashi, Sumida, and Kuramoto, 2009.
Several nomenclatural and taxonomic ambiguities
require clarification because recently described
species have largely relied on questionable
assumptions associated with these names.
First, the type locality of R. cyanophlyctis was
given by Bauer (1998)asSyntypes: ZMB
Diversity and distribution of South Asian skipper frogs 7
319798,ex. India orientali,two specimens sent
by [Christoph Samuel] John, probably from
Tranquebar.Though Bauer (1998) provided a rea-
sonable hypothesis for why Johns collections
probablycame from the historical region of
Tranquebar (now Tharangambad in Tamil Nadu
State), he did not make an explicit statement restrict-
ing the type locality. The type specimens must be re-
examined in order to verify whether they actually cor-
respond to the orange-coded lineage now widely
regarded as E. cyanophlyctis. Until this is verified,
the type locality for the species remains ex. India
orientali,which in 1799 included most of South and
Southeast Asia (and not Kerala, India,as stated in
Howlader et al., 2015).
Second, R. bengalensis was described from Bengal
in 1830, which at that time comprised modern-day
Bangladesh and several states in neighbouring India
(West Bengal, Jharkhand, Bihar and the Cachar region
of Assam). The type specimen is the animal depicted
in Gray (1830:fig. 2), the whereabouts of which is
currently unknown: it is present in neither the BMNH
(SM pers obs.) nor the ZSI collections (Sclater, 1892).
Thus, this specimen might have been lost, Gray (1830)
may have provided an inaccurate drawing, or he
described the species based on an uncollected animal.
Third, R. leschenaultii was described from a syntype
series involving at least four collectors and two local-
ities, Pondich
ery(collector Leschenault) and
Bengale(collectors M. Duvaucel, M. Dussumier
and M. Roux) (Dum
eril & Bibron, 1841). By 1841,
Bengal had further expanded to include modern day
Assam, Nagaland State and Meghalaya State.
Boulenger (1920)statedtohaveexaminedninetype
specimens in MNHNP, and according to Frost (2022),
Guib
e(1950) identified MNHNP 436670 as syn-
types. However, the online collection database for
MNHNP lists as many as 20 specimens as syntypes:
MNHNP 0.43660.4368, 0.4370, 0.6431, 0.768,
1999.79227924, 1999.79277934 collected by
Rouxfrom Bengale,and MNHNP 0.4369,
1999.7925 and 1999.7926 collected by Jean-Jacques
Dussumierfrom C^
ote de Malabar(Malabar Coast).
This conflicts with historical literature (including the
description) with respect to the size, collectors and
locality details of the syntype series. Importantly, if
the MNHNP details are correct, the lectotypification
of a Malabar Coast type would likely threaten the sta-
tus of one of the species recently described from the
Western Ghats.
Cantor (1847) suggested that R. bengalensis and R.
leschenaultii might represent the same species but did
not propose a synonymy. Peters (1863)formallysyn-
onymized R. leschenaultii with R. cyanophlyctis,stat-
ing that the former matches the description and the
two type specimens of the latter. Both R. bengalensis
and R. leschenaultii were later included in the syn-
onymy of R. cyanophlyctis (although without explana-
tions) by G
unther (1864) and Steindachner (1867),
where they have remained until today. However,
based on lineage distributions (Fig. 1), R. bengalensis
and R. leschenaultii could each represent either senior
subjective synonyms of E. adolfi (red-coded lineage),
or junior subjective synonyms of E. cyanophlyctis
(orange-coded lineage), even more so because these
species lack clear morphological distinction (see
below), especially in former Bengal (SM pers. obs.).
Hence, R. bengalensis requires neotypification and R.
leschenaultii requires lectotypification (after verifying
its actual syntype series), in order to resolve their
nomenclatural and taxonomic status.
Fourth, De Silva (1958) provided brief descriptions
for three colour varietiesof Rana cyanophlyctis
from Sri Lanka, given as Variety typicus
(Schneider),Variety fulvus (De Silva)and Variety
flavens, the first one having been completely ignored
in subsequent literature. De Silva's taxonomic inten-
tions are somewhat ambiguous because he did not
provide the new names as trinomials, hence the need
for clarification. Of particular confusion is that an
authorship was assigned to two of the names pro-
posed, although they may not necessarily convey taxo-
nomic authorities. First, the Variety fulvus (De
Silva)makes reference to De Silva (1957), but this
publication reports on zoological observations (involv-
ing Rana cyanophlyctis) on the northern Sri Lankan
coastal islands and lacks any new nomenclatural
Fig. 2. Genome size variation in Euphlyctis adolfi and E.
cyanophlyctis among 159 individuals. See Supplemental
material - File S1a for the raw data.
8 C. Dufresnes et al.
content. Second, Variety typicus (Schneider)makes
reference and actually copies parts of Boulengers
(1890) account of Rana cyanophlyctis Schneider, 1799
in Sri Lanka (p. 442), which tends to indicate that this
variety is meant as the nominal form, i.e. Rana cya-
nophlyctis cyanophlyctis.Because it is impossible to
disentangle whether these citations correspond to
authorships of the names or simply mentions of
authors who had previously discussed the colour vari-
eties, we refer to Article 45.6 of the Code (ICZN,
1999) to settle on their status:
"The rank denoted by a species-group name
following a binomen is subspecific, except that [ ]
45.6.4. it is subspecific if first published before 1961
and its author expressly used one of the terms
"variety" or "form" (including use of the terms
"var.," "forma," "v." and "f."), [ ]."
All three varieties of De Silva (1958) are nomenclatur-
ally available subspecies names, given that they are
explicitly referred to as varieties, followingabino-
men (Rana cyanophlyctis), and published before 1961.
We must thus consider Euphlyctis cyanophlyctis typi-
cus (De Silva, 1958), with type locality Ceylon [¼
Sri Lanka],and unstated type specimens, which may
include specimens available to De Silva in the
National Museum of Columbo (now NMSL), or speci-
mens examined by Boulenger (1890); Euphlyctis cya-
nophlyctis fulvus (De Silva, 1958), with type locality
from brackish water pools on the islands of
Karaitivu, Leyden [¼Velenai] and Punkudutiva [¼
Pungudutivu] off Jaffna (N.P.)in Jaffna District,
Northern Province, Sri Lanka, and unstated syntypes
collected in 1956, presumably deposited in NMSL
(De Silva, 1957); Euphlyctis cyanophlyctis flavens (De
Silva, 1958), with type locality from clear pools from
Ratnapura, Ceylonin Ratnapura City, Ratnapura
District, Sabaragamuwa Province, Sri Lanka, and
unstated syntypes, two collected in 1952 and one in
1957 or 1958, presumably deposited in NMSL (De
Silva, 1958). Attempts to locate and examine the puta-
tive type specimens would help clarify the taxonomic
status of these three nomina. Given the variable color-
ation of E. cyanophlyctis throughout its range, and
because our phylogeographic analyses suggested little
genetic differentiation between mainland and Sri
Lankan populations (Fig. 1,Supplemental material -
File S3), these taxa most likely represent junior syno-
nyms of E. cyanophlyctis. However, these three names
have not been formally synonymized, and without
direct evidence of the contrary, they remain valid sub-
species of E. cyanophlyctis.
According to current knowledge, we thus tentatively
maintain application of the oldest nomen Euphlyctis cya-
nophlyctis (Schneider, 1799) to the orange-coded lineage
and provisionally retain the other nomina as synonyms
or subspecies.
In the E. cyanophlyctis group, we could further infer
the ranges of E. adolfi and E. cyanophlyctis using gen-
ome-size estimates of 151 individuals. Genome sizes
varied from 6.5 to 7.6 pg, and the variation was mostly
diagnostic between the two taxa (Fig. 2). The species
means differed by nearly 6% (7.2 pg in E. adolfi and
6.8 pg in E. cyanophlyctis), and frogs assigned to E.
adolfi and E. cyanophlyctis had genome sizes above and
below 7.0 pg, respectively. Genome size could thus
serve as a nuclear barcoding tool to distinguish these
species. Specifically, it confirmed the respective distri-
butions of E. adolfi and E. cyanophlyctis in northern
India and Nepal. We did however notice some regional
discordances: in parapatric ranges, genome sizes typical
of E. cyanophlyctis were measured near populations that
featured only E. adolfi mtDNA haplotypes, which could
indicate genetic introgression.
Subgenus Phrynoderma (E. hexadactyla group).The
subgenus Phrynoderma forms a strongly supported clade
composed of six notable lineages, assembled into four
species in the ASAP analysis (Fig. 1,Table 1):
1. A southwestern Indian lineage (purple-coded in Fig.
1) that corresponds to Euphlyctis karaavali Priti,
Naik, Seshadri, Singal, Vidisha, Ravikanth, and
Gururaja, 2016, as confirmed by barcoding of the
holotype. Furthermore, the purple-coded lineage is
the only one detected at the type locality, and no
other nomina seem to correspond to this taxon.
2. A widespread coastal lineage found in Sri Lanka,
eastern India and neighbouring Bangladesh (blue-
coded in Fig. 1). The oldest nomen for this lineage has
long been considered to be Rana hexadactyla Lesson,
1834, for which topotypes were accordingly barcoded.
The nomenclatural priority of E. hexadactyla
requires some clarification because Lesson published
a second name for the species, Dactylethra benga-
lensis, which has been the source of a lot of confu-
sion. Although D. bengalensis is sometimes credited
to Dumeril & Bibron (1841), these authors clearly
referred to LessonsIllustrations de zoologie [¼
Illustrations de zoologie, ou, Recueil de figures
d'animaux peintes d'apr
es nature]and not Lesson
(1834) (as suggested by Frost, 2022). Various, usu-
ally reliable online sources (e.g. British Library,
Biodiversity Heritage Library) give a variety of
Diversity and distribution of South Asian skipper frogs 9
publication dates for this book (e.g. 18311835?,
1832,”“18321835,”“1835), thus it is necessary
to determine which one is correct to infer whether
hexadactylaor bengalensishas priority.
LessonsIllustrations de zoologiecomprises 60
plates and was issued in 20 livraisons of 3 plates
each. The date Juin 1831 [¼June 1831]appears
at the end of the text description of the species;
however, the published prospectus for the books
stated that the first livraison was due to be published
on 1
st
March 1832. Mathews (1911) obtained the
publication dates and noted that the first livraison
was actually published on 14
th
July 1832, while the
16th to 20th livraisons were all issued some time
between 17
th
January 1835 and December 1835. The
figure and associated description of Dactylethra ben-
galensis was published in plate XLVII [¼plate
47]and thus was published in 16
th
livraison.
Therefore, the correct year of publication for
Dactylethra bengalensis is 1835.
Moreover, Lesson (1835) erroneously mention R.
hexadactyla as a synonym of D. bengalensis.Theori-
ginal descriptions of these taxa are identical, suggest-
ingthatbotharereferringtothesamespecimen(s),
collected by Belanger from Pondich
ery [Puducherry,
southeastern India].Hence, we consider Dactylethra
bengalensis Lesson, 1835 to be a junior objective
synonym of Rana hexadactyla Lesson, 1834.The
paratypespecimen reported by Guib
e(1950),
NMNHP 4636 (now catalogued as MNHN-RA-
0.4363) from Bengaleappears to have a blotched
dorsal pattern with no midvertebral stripe, which does
not resemble the frog illustrated in Lesson (1835)
a mostly plain brown frog with a pale vertebral stripe.
Since there is no mention of any specimens from
Bengalein either of Lessons descriptions (Lesson,
1834,1835), MNHN-RA-0.4363 should not be
regarded as a type (contra Frost, 2022). Instead, the
lack of holotype designation for R. hexadactyla
(Lesson 1834) suggests a syntype series; it is conceiv-
able that historical specimens in the MNHNP collec-
tion from Pondich
ery(e.g. MNHN-RA-0.4361,
0.4364) and Indes Orientales(e.g. MNHN-RA-
0.4365) could have been a part of this series, provid-
ing that they were made available to Lesson.
Other junior synonyms include the names Rana
cutipora Dum
eril & Bibron, 1841 and Rana sapar-
oua Dum
eril & Bibron, 1841, which have both been
treated as substitute names (objective synonyms) for
R. hexadactyla (e.g. Dinesh et al., 2021; Frost,
2022). Though Dum
eril & Bibron (1841) clearly cre-
ated R. cutipora to apply to R. hexadactyla, they
made no indication as to why they rejected Lessons
older valid name. Rana cutipora was even adopted
for the species until G
unther (1859), who as far as
we are aware of was the first to relegate it as a
synonym, where it remains today. In contrast, R.
saparoua seems to have been listed in the synonymy
of E. hexadactyla only recently the earliest men-
tion we have noticed is Dinesh et al. (2009), appar-
ently following Frosts website. This name thus
requires some clarification.
Dum
eril & Bibron (1841) published the following
synonymy in their species account for Rana cuti-
pora. Nobis,where a Rana saparouais
first mentioned:
SYNONYMIE. Rana saparoua. Mus. Lugd. Bat.
Rana hexadactyla Less. Voy. Indes orient. Bel. zool.
Rept. pag. 331, tom. VI.
Dactylethra Bengalensis. Id. Illustrat, zool.
PI. XLVII.
Rana hexadactyla. Tschudi, Classif. Batrach. M
em.
Soci
et. sciene. nat. Neuch^
at. tom. II, pag. 80.
The name saparouaappears nowhere else in
Dum
eril & Bibron (1841), and this single mention
assigned neither an author citation for the name, nor
was the name followed by nobis,so they were not
intentionally creating a new name. In addition, the
abbreviation Mus. Lugd. Bat.that follows the
name corresponds to the Musei Lugduno-Batavae,
Leiden, so it probably indicated a specimen label.
The word saparouarefers to the Indonesian island
known today as Saparua, which in the 19
th
century
had various spelling (e.g. Ile Saparoua,”“Saparna,
Saparnae,”“Saparoea; Dum
eril & Dum
eril, 1851;
Balfour, 1885; Tschudi, 1838). In his discussion of
Rana hydromedusa Kuhl,Tschudi (1838) made a
similar reference to the Saparua specimen(s), using
one of the alternative spellings of the island:
Hyla histrionica Boje ist das junge, und Rana
Saparnae Mus. Lugd. das ganz alte Their davon.
In what appears to be the earliest mention of a
Rana Saparnae,Tschudi (1838) also did not asso-
ciate an author to the name, and neither did he indi-
cate that he was coining a new name. While Rana
Saparnaereminds of a binomen, it could also sim-
ply be a reference to a specimen or specimens iden-
tified as Rana from SaparnaeIsland, in which
case it would not have any nomenclatural status. If
10 C. Dufresnes et al.
it was actually meant as a binomen, it would then
have first appeared as a synonym of R. hydromedusa
Tschudi (1838), and thus be nomenclaturally unavail-
able (Article 11.6 of the Code). In the case of R.
saparoua, the use of italics by Dum
eril and Bibron
(1841) suggests that they intended it as a binomen,
but because it was likewise first published in a syn-
onymy, R. saparoua is also nomenclaturally unavail-
able. Dum
eril and Bibrons(1841) inclusion of this
name in the synonymy of their R. hexadactyla (as R.
cutipora) appears to have been accidental because:
(1) they cited the distribution of R. cutipora as east-
ern India, and specifically as Bengaleand
Pondich
ery(not Indonesia); (2) they made no refer-
ence to Tschudis(1838) synonymization (or updated
identification) of the Saparua population under Rana
hydromedusa Kuhl;(3)theydidtreatRana grun-
niens (Latreille, 1801) as a valid species (now
Limnonectes grunniens), under which they included
Rana hydromedusa Kuhlin the synonymy (where it
remains today); and (4) they recognized that the type
series of R. grunniens comprised two specimens (now
in MNHNP), one of which represented R. hexadac-
tyla and the other R. grunniens sensu stricto (noted
earlier by Tschudi, 1838, see also Stejneger, 1940).
Limnonectes grunniens has accordingly been reported
from Saparna Island (Evans et al., 2003). In conclu-
sion, both Rana Saparnae Tschudi 1838 and Rana
saparoua Dum
eril & Bibron 1841 should be placed
in the synonymy of L. grunniens (rather than E. hexa-
dactyla) and be considered nomina nuda.
The Sri Lankan population was described as Rana
robusta Blyth, 1855, based on two male syntypes
currently reported as ZSI 9123 and ZSI 9124
(Chanda et al., 2001; Sclater, 1892). Rana robusta
was then synonymized shortly after by its descriptor
in a brief footnote (Blyth, 1856).
R. robusta, nobis, J. A.S. XXIII, 298, is (we are
now satisfied) a phase of R. CUTIPORA, D. and B.
The original description of the two specimens gener-
ally conforms with Euphlyctis morphology, and the
size provided for the species (SVL 3 inches ¼
76.2mm) places it within the adult size range of E.
hexadactyla.Theobald(1868), however, referred to
two specimens of Rana cutipora collected by Kelaart
from Ceylon in MASB (where the types were origin-
ally deposited), as an adult and young.Because
these two specimens were the only ones presented by
Kelaart in the collection (Sclater, 1892), they are
likelytobetheR. robusta types, in which case Blyth
(1855) may have reported only the size for the larger
male (the one identified as an adult by Theobald,
1868). The young specimen mentioned by Theobald
(1868) particularly requires re-examination because
MASB scientists (e.g. Blyth, Theobald, Anderson)
usually determined the sex and maturity of frogs by
the observation of obvious sexually dimorphic charac-
ters (e.g. external vocal sacs) and relative size differ-
ences between individuals thought to represent the
same species, rather than dissection and gonad identi-
fication (SM pers obs.). This practice led to consider-
able taxonomic confusions when those authors
erroneously synonymized superficially similar species
despite size differences (Mahony et al., 2018). Hence,
the syntype of R. robusta, alternatively mentioned as
amale(Blyth,1855) or a young (Theobald, 1868)
could as likely be an adult male E. cyanophlyctis
instead of a young R. robusta.
Given these ambiguities, the two syntypes of R.
robusta must be re-examined to verify that they do
not represent different species. Without further infor-
mation, here we provisionally maintain R. robusta as
a synonym of E. hexadactyla, also because our bar-
coding data confirmed that Sri Lankan and mainland
populations lack notable divergence (Fig. 1,
Supplemental material - File S3).
In summary, the oldest valid nomen available for
the blue-coded lineage is Euphlyctis hexadactyla
(Lesson, 1834). Junior synonyms include the object-
ive synonyms Dactylethra bengalensis Lesson, 1835,
Rana cutipora Dum
eril & Bibron, 1841, and Rana
robusta Blyth, 1855 as a subjective synonym.
36. Four closely related lineages found in southwestern
India, allied to E. aloysii (Fig. 1).
One unambiguously represents Euphlyctis aloysii
Joshy, Alam, Kurabayashi, Sumida, and Kuramoto,
2009 sensu stricto (s. s.), given barcoding of the
holotype (dark green-coded in Fig. 1).
The second one putatively corresponds to
Euphlyctis kerala Dinesh, Channakeshavamurthy,
Deepak, Ghosh, and Deuti, 2021 (pale green-coded
in Fig. 1). This association is not straightforward,
however, because Dinesh et al. (2021,2022)
remained vague about the origin of the three
sequences attributed to E. kerala: GU136102
(India,but labelled as being part of a Western
Ghats unpublished barcoding project), MH423736
and MH423738 (both NA)(Supplemental
material - File S1a). According to the metadata
submitted to GenBank, the latter two were obtained
from vouchers (labeled EU1Sand EU2S), but
which cannot be related to the type series of the
species (holotype ZSI/WGRC/A/973, paratypes
ZSI/WGRC/A/974 and ZSI/WGRC/A/975). It thus
Diversity and distribution of South Asian skipper frogs 11
remains to be clarified whether the type series was
actually barcoded.
The two remaining lineages do not seem to corres-
pond to any known taxon, and we refer to them as
Euphlyctis cf. aloysii. One was detected only in
Mudigere, Karnataka (in syntopy with E. cyano-
phlyctis) and appears to be splitting from the most
basal node of the complex (dashed dark green-
coded in Fig. 1). The last one was discovered with
our sample from Baga in Goa State (dashed pale
green-coded in Fig. 1), which groups with E. aloysii
s. s. in our 16S tree.
The E. aloysii complex is the only Euphlyctis phy-
logroup where some currently recognized species are
not delimited as such in our 16S-based ASAP ana-
lysis (Fig. 1). The best partition only distinguished
two species for the complex: E. cf. aloysii from
Karnataka and a set of lineages including E. aloysii,
E. kerala and E. cf. aloysii from Goa (Fig. 1).
Additional taxa.From our barcoding data, we could not
definitely settle on which lineage the subspecies
Euphlyctis cyanophlyctis microspinulata Khan, 1997 is
affiliated with. This taxon was described from
Balochistan in western Pakistan (type locality in south-
west of Khuzdar, southeast Kalat Division; Khan, 1997),
where E. adolfi (present in northern and southern
Balochistan) or E. cyanophlyctis (present in eastern
Pakistan) could potentially occur (Fig. 1). The type
locality is located above 1,200 m a.s.l., and because only
E. adolfi has been recorded at such elevations (E. cya-
nophlyctis is always found below 1,100 m, SNL pers.
obs.), it would a priori make a more plausible candidate.
In addition, it remains to be confirmed that Balochistan
skipper frogs actually represent a distinct taxon: no add-
itional lineage was discovered among our samples from
southern and northern parts of this mountain range
(Figs. 1,3). Without more decisive information (barcod-
ing of the type locality), we must maintain this taxon
where it was originally described, as the subspecies
Euphlyctis cyanophlyctis microspinulata Khan, 1997.
Regardless of its status and affiliation, the type series of
E. c. microspinulata is a matter of issue. It putatively com-
prised the holotype, BMNH 1990.8,and multiple para-
types: CAS 170531, four specimens grouped under
BMNH 1990.9,and16or17specimenswithonlyMSK
numbers (the authors private collection; Khan, 1997).
Wagner et al. (2016) confirmed that the holotype was never
inventoried in the BMNH, being probably still in a private
collection or lost,and although the BMNH specimen
accessions register does include the lot 1990.11990.17
(presented by M.S. Khan), the numbers 1990.8, 1990.9 and
1990.10 are indeed marked as vacant(¼no specimen).
In addition, one of us (SM) was unable to find Euphlyctis
specimens in the BMNH with collection data that would
match the missing types. Hence, the holotype and pre-
sumed BMNH paratypes must be considered lost, and the
only extant type specimen known to be accessioned into a
museum collection is the paratype CAS 170351, which
would be a potential candidate for neotypification.
Fig. 3. Mitochondrial phylogeography of the two most widespread Euphlyctis lineages, E. adolfi and E. cyanophlyctis, illustrated by
16S haplotype networks, Bayesian skyline plots and sampling localities. Both lineages are genetically structured and show signals of
population expansion for E. adolfi, but not for E. cyanophlyctis. See Supplemental material - File S3 for a phylogenetic
representation.
12 C. Dufresnes et al.