ArticlePDF Available

Abstract and Figures

We review morphology and systematics of Phoxophrys using new specimens of previously rare species. In addition to external characters, we relied heavily on skull morphology visualized using computed tomography data of all currently recognized species in this genus. Phylogenetic analysis of ND4, 12S, and 16S mDNA sequences reveal that Ph. tuberculatais sister to a clade containing Dendragama, Lophocalotes, and insular Pseudocalotes. Phoxophrys tuberculatais only distantly related to Bornean congeners. Phylogenetic analysis of 29 morphological characters scored for all the species of Phoxophry sand a diverse set of 22 outgroup taxa found four well-supported lineages: Ph. tuberculata, a clade of all Bornean congeners, Ph. nigrilabris, and a clade of the four large Bornean species. To resolve paraphyly of Phoxophrys, we revalidate Pelturagonia Mocquard for all Bornean species of this genus. As redefined, Phoxophrys contains a single species: Ph. tuberculata of Sumatra. We describe a new species of Pelturagonia from the Meratus Range of southeastern Kalimantan. The new species is the sister species of Pe. spiniceps. Like that species, it differs from congeners in having large, dorsally projecting scales between the dorsolateral caudal crests and apostorbital process of the frontal bone reaching the postciliary ornament.
Content may be subject to copyright.
Herpetological Monographs, 33, 2019, 71–107
Ó2019 by The Herpetologists’ League, Inc.
Phoxophrys After 60 Years: Review of Morphology, Phylogeny, Status of Pelturagonia, and a
New Species from Southeastern Kalimantan
MICHAEL B. HARVEY
1,7
,THORTON R. LARSON
2
,JUSTIN L. JACOBS
2
,KYLE SHANEY
2,6
,JEFFREY W. STREICHER
3
,AMIR HAMIDY
4
,
NIA KURNIAWAN
5
,AND ERIC N. SMITH
2
1
Department of Biological Sciences, Broward College, 3501 Southwest Davie Road, Davie, FL 33314, USA
2
The Amphibian and Reptile Diversity Research Center and Department of Biology, University of Texas at Arlington, 501 South Nedderman Drive,
Arlington, TX 76010, USA
3
Department of Life Sciences, The Natural History Museum, London SW7 5BD, UK
4
Laboratory of Herpetology, Museum Zoologicum Bogoriense, Research Center for Biology, Indonesian Institute of Sciences–LIPI,
Jl Raya Jakarta Bogor km 46, Cibinong, West Java, 16911, Indonesia
5
Department of Biology, Universitas Brawijaya, Jl Veteran, Malang, East Java, 65145, Indonesia
ABSTRACT: We review morphology and systematics of Phoxophrys using new specimens of previously rare species. In addition to external
characters, we relied heavily on skull morphology visualized using computed tomography data of all currently recognized species in this genus.
Phylogenetic analysis of ND4, 12S, and 16S mDNA sequences reveal that Ph. tuberculata is sister to a clade containing Dendragama,
Lophocalotes, and insular Pseudocalotes. Phoxophrys tuberculata is only distantly related to Bornean congeners. Phylogenetic analysis of 29
morphological characters scored for all the species of Phoxophrys and a diverse set of 22 outgroup taxa found four well-supported lineages: Ph.
tuberculata, a clade of all Bornean congeners, Ph. nigrilabris, and a clade of the four large Bornean species. To resolve paraphyly of Phoxophrys,
we revalidate Pelturagonia Mocquard for all Bornean species of this genus. As redefined, Phoxophrys contains a single species: Ph. tuberculata of
Sumatra. We describe a new species of Pelturagonia from the Meratus Range of southeastern Kalimantan. The new species is the sister species of
Pe. spiniceps. Like that species, it differs from congeners in having large, dorsally projecting scales between the dorsolateral caudal crests and a
postorbital process of the frontal bone reaching the postciliary ornament.
Key words: Agamidae; Borneo; Computed tomography; Osteology; Pelturagonia anolophium sp. nov.; Skull; Sumatra; Systematics
EARLY in a long and prolific career studying systematics of
Bornean reptiles and amphibians, Inger (1960) revised the
small agamid genus Phoxophrys Hubrecht (Fig. 1). Prior to
his research, most herpetologists referred Bornean species of
these lizards to Japalura (e.g., Peters 1864; Boulenger 1885,
1891; De Rooij 1915), a genus now confined to mainland
Asia (Wang et al. 2018). Four of the five species of
Phoxophrys known to Inger occur only on Borneo, whereas
Ph. tuberculata is endemic to Sumatra. Sixty years later these
lizards remain poorly known, leading Manthey and Denzer
(2019:9) to remark recently that ‘‘ hardly any other draconine
lizard genus has been studied less in recent decades than
Phoxophrys.’’ Monophyly of the genus has never been
tested, and some characters thought to be diagnostic of
Phoxophrys occur in only some of the species or are
widespread in other draconines. In recent years, herpetol-
ogists obtained sizable series of Ph. nigrilabris (e.g., Lloyd et
al. 1968). However, few additional specimens of the other
species reached museum collections. At least on Borneo,
these small- to medium-sized lizards (maximum snout-to-
vent length [SVL] ¼84 mm) can be readily distinguished
from other agamids by distinctive dorso-lateral crests at the
base of the tail, strongly heterogenous dorsal scalation,
relatively wide heads, bluish buccal epithelia, and no visible
tympana. No aspect of their ecology has been systematically
studied. Nonetheless, they lay only 2–4 eggs, are cryptically
colored, and occur in the understory of humid forests of
Borneo and Sumatra (Malkmus et al. 2002; Das 2010).
Phoxophrys nigrilabris appears to be restricted to the
lowlands and foothills, whereas Ph. borneensis and Ph.
cephalum reach 2,000 m on Mount Kinabalu (Malkmus et al.
2002).
In November 2016, T.R. Larson and M. Munir collected a
small number of reptiles and amphibians on Gunung Lumut,
a remote mountain of the Meratus Range in southern East
Kalimantan Province. The samples included four specimens
of Phoxophrys. Our attempts to identify and then properly
diagnose these lizards required a review of external
morphology of the genus. Our survey of morphology
includes new data for the skull and raised questions about
the monophyly of the genus. Herein, we present the results
of our morphological review before evaluating relationships
among species of Phoxophrys and formally describing the
new lizards from Kalimantan.
MATERIALS AND METHODS
Fieldwork
T.R. Larson and M. Munir surveyed reptiles and amphib-
ians on Gunung (¼Mountain) Lumut, Kalimantan Timur,
Indonesia on 12–14 November 2016. They conducted most
surveys at night and recorded GPS coordinates using the
WGS-84 geodetic system. They photographed specimens in
life, euthanized them with benzocaine, fixed the lizards in 10%
formalin, and transferred them to 70% ethanol for permanent
storage at the Museum Zoologicum Bogoriense (MZB) and
Amphibian and Reptile Diversity Research Center, University
of Texas at Arlington (UTA). We sexed the specimens by
examination of the gonads. Before fixing, Larson and Munir
weighed each specimen to the nearest 1 g using an electronic
6
PRESENT ADDRESS:DepartamentodeEcolog
´
ıa de la Biodiversidad,
Instituto de Ecolog´
ıa, Universidad Nacional Aut´
onoma de M ´
exico,
Apartado Postal 70-275, Ciudad Universitaria, Ciudad de M ´
exico,
04510, M´
exico
7
CORRESPONDENCE: e-mail, mharvey@broward.edu
71
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
balance. They placed each specimen on a flat surface next to a
ruler and took photographs of the dorsal, lateral, and ventral
sides. For selected specimens, they also made multiple photos
of live animals in natural settings. We deposited photos of all
specimens at UTA.
Collection of Morphological Data and Definition of Characters
We compared museum specimens of our new species to
62 specimens of all currently recognized congeners (Appen-
dix I). Our comparative series includes type specimens of all
species except Phoxophrys nigrilabris. We use the museum
acronyms of Sabaj (2016). ZRC(IMG).2.187a-f are image
vouchers deposited in the Lee Kong Chian Natural History
Museum, Singapore. The specimen of Ph. borneensis
appearing in these photos was not deposited at the museum
(K. Lim Kok Peng, personal communication). Asad et al.
(2015) misidentified this specimen as Ph. spiniceps. This
female specimen lacks the various diagnostic spinose scales
FIG. 1.—Female (A, UTA 21844, SVL ¼44.5 mm) and male (B, MZB 14993, SVL ¼42.8 mm, photos by T.R. Larson) paratypes of Pelturagonia
anolophium from Mount Jumut, Kalimantan Timur, Indonesia, 1,090 m; male Pe. cephalum (C, not collected, photo by Elijah Wostl) from Mount Kinabalu,
Sabah, Malaysia; male Pe. nigrilabris (D, not collected, photo by Chien C. Lee; http://www.chienclee.com) from Sarawak, Malaysia; male Pe. spiniceps from
Mount Murud, Sarawak, Malaysia (E, not collected, photo by Andrej Masonzravky); and male Phoxophrys tuberculata (F, UTA 65254, SVL 36.5 mm, photo
by E.N. Smith) from Batang Gadis, Sumatera Utara, Indonesia, 1,223 m.
72 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
of Ph. spiniceps and cannot be referred to Ph. cephalum,
because it has 3/3 (count on left side/right side) sublabial
tubercular scales, has a lorilabial separating the enlarged
suprarictal scale from the scales of the rictal fold, and lacks a
row of paired enlarged subcaudals. We assign ZMB 57237
and ZMB 57238 to Ph. cephalum (Malkmus et al. 2002
initially identified these specimens as Ph. borneensis;
however, Manthey 2010:Fig. RA03421-2 later reidentified
the larger specimen as Phoxophrys cf. cephalum), because
they have two (vs. four) rows of enlarged subcaudals; a single
large sublabial tubercular scale below the rictus without a
row of tubercular or heavily keeled scales anterior to it;
gulars with rounded low keels; 25 and 23 (respectively)
gulars down the midline from the chin shields to the chest;
and an enlarged suprarictal scale in broad contact with the
first scale of the rictal fold.
We scored each specimen for a suite of meristic,
mensural, and qualitative characters that have proven useful
for diagnosing and describing lizards in other draconine
genera (Harvey et al. 2014, 2017a,b, 2018). To the nearest
0.1 mm with a digital caliper, we measured SVL (from the tip
of the snout to the anterior lip of the cloaca), body length
(from the posterior insertion of the arm to the anterior
insertion of the leg), pectoral width (axilla to axilla), length of
the tail (from the posterior lip of the cloaca), head length
(from the occiput to the center of the snout), head width (at
rictus), maximum diameter of the orbit (bony edge to bony
edge; we took this measurement along a line parallel to the
ocular aperture, because the orbit is widest along this line),
width of the snout (between the upper margins of the
nostrils), greatest width and height of the medial rostral (if
present), length of the shank (from the center of the knee to
the preaxial base of Toe I), height of the longest nuchal crest
scale (a straight-line measure from the anteriormost edge of
the scale to its apex), and height of the longest dorsal crest
scale (using the same technique as for nuchal crest scales).
We measured Fingers III and IV and Toes IV and V by
pressing the digits to a flexible ruler and measuring from the
interdigital skin to the base of the claw.
For characters that could be scored on either side of the
body (i.e., bilateral characters) such as length of the shank
and counts of labials, we scored meristic characters on the
left side and mensural characters on the right side of each
specimen. When a bilateral character could not be scored on
one side, we switched sides.
Our counts of labials differ from those of previous revisors
of Phoxophrys (notably Inger 1960). Confusion may also
arise when identifying the first supralabial, because Bornean
Phoxophrys have vertically divided rostrals (Fig. 2). We
consider the first supralabial to be the scale behind the
lateralmost rostral. A postrostral separates the lateralmost
rostral from the nasal, and the nasal either contacts the first
supralabial or is separated from it by a lorilabial. As in
Lophocalotes (Harvey et al. 2018), Phoxophrys has an
enlarged lorilabial just below and behind the eye. This scale
is positioned above the last supralabial. The last infralabial is
the scale positioned directly below the last supralabial.
Herein, we use the term supercilium to refer to the flap of
skin that roofs the orbit. We identified the last canthal as a
scale ending directly above or extending beyond a straight
vertical line at the anterior margin of the orbit. We counted
circumorbitals using the method of Harvey et al. (2014) and
transorbitals as the number of scales in a transverse line
between, but not including, the last supraciliary scales. Our
transorbital count is identical to the ‘‘ Head Scales: HeadStr’’
character defined by Zug et al. (2006). Counts of subdigital
lamellae include all scales from the interdigital skin to the
claw and include the elongate ungual scale. We use some
special terminology defined by Harvey et al. (2014, 2017a,b,
2018) for scales of the head, neck, and flanks. Names of
scales and skin folds associated with the corner of the mouth
frequently include the suffix ‘‘rictal.’’ In our experience, a
rictal fold (sensu Cadle 1991) is invariably present in
draconines.
Vertebral crests of Phoxophrys often contain a complex
mix of enlarged flat or projecting scales, small vertebrals, and
paravertebral scales in medial contact with one another.
When counting vertebral scales, we use the same methods as
in our recent revisions of Dendragama and Lophocalotes
(Harvey et al. 2017a, 2018). However, we could not
confidently distinguish between tiny vertebrals and paraver-
tebrals in medial contact within the dorsal crests of Ph.
anolophium,Ph. borneensis, and Ph. cephalum. Thus, we
caution readers that nuchal crest counts exclude paraverte-
brals in medial contact (as in Harvey et al. 2014, 2017a,b,
2018), but dorsal crest counts include these scales, unlike
counts in our earlier studies. The nuchal crest counts and
pectoral gap counts can be directly compared to similar
FIG. 2.—Cephalic morphology of Pelturagonia anolophium (adult female
holotype, MZB 14992).
73
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
counts in other draconines described by us; however dorsal
crest counts and total vertebral counts cannot.
We struggled to find a repeatable method of counting
caudal dorsolateral crest scales. Distally on the tail, the
degree to which these scales project laterally diminishes
continuously. To count these scales, we viewed the tail in
dorsal aspect and stopped the count with the last scale
having a slightly pointed distal tip of the keel.
To study the skull and pectoral girdle, we collected
computed tomography (CT) data from each of the six known
species of Phoxophrys and 22 additional species in 16
draconine genera (Appendix I). We immobilized each
specimen, then scanned it for 1545 min per scan. We
scanned the holotypes of Ph. spiniceps and Japalura
robinsoni at the Natural History Museum, London (formerly
BMNH) using a Nikon Metrology HMX ST 225 scanner at
80 kV and 80 lA. We scanned all other specimens at the
UTA Shimadzu Center for Environmental Forensics and
Material Sciences, using the Shimadzu inspeXio SMX-100
CT scanner at 65 kV and 40 lA. We used Shimadzu’s
inspeXio software to reconstruct raw X-ray data and exported
them as stacks of 1,024 31,024 16-bit tif images. We then
rotated and cropped each stack in ImageJ (National
Institutes of Health, Bethesda, MD) and imported it into
the open-source program Drishti 2.0 (available at http://sf.
anu.edu.au/Vizlab/drishti/). We then generated one surface
per skeletal structure and recorded characteristics from
these three-dimensional surfaces.
With a few exceptions (defined on first use), we use the
terminology of Evans (2008) in our description of the skull.
In our description, we reference character states of Moody
(1980) and Stilson et al. (2017) in brackets, and we
developed a set of measurements and ratios that quantify
some descriptive terms used by these authors. We measured
width (between lateral margins of quadrates at their articular
facets with the mandible) and length (from posterior edge of
lateral articular surface of quadrate to middle of premaxilla)
of the skull. For the dermal roofing bones, we measured the
midline length (points placed at anterior tip of nasal where it
curves laterally to surround premaxilla and a midpoint
between posterior tips of nasals) and maximum width (from
tip of lateral process to tip of lateral process of paired nasals
divided by two) of the nasal, length (anterior tip where
midline suture of nasal ends to frontoparietal suture) and
width (at prefrontal–frontal suture) of the frontal, and
midline length (from frontoparietal suture to bone’s
transverse ridge) and width (lateral edge of bone’s orbital
process) of the parietal.
For the circumorbital bones and temporal arcade, we
measured the greatest length and width of the supratemporal
fossa (from ridge or crest to ridge or crest. The measure
captures the dimensions of the dorsolateral aperture. Length
of this fossa is the longest diameter and width is the longest
diameter perpendicular to the length.), width and greatest
length of the suborbital fenestra (in orbital view), width and
greatest height of the lacrimal canal, and length of the
anterior process of the supratemporal extending along the
medial edge of the supratemporal fossa (reported as a ratio
to length of fossa and measured as a straight line from its
anterior tip to top of squamosal at about the same point as
that used for measuring length of fossa).
For the palatoquadrate derivatives, we measured height
(from posterolateral edge of dorsal condyle immediately
below shallow groove in squamosal to lateral edge of ventral
condyle) and width (in posterior aspect, widest point just
above ventral condyle) of the quadrate, width of the conch
(from base of pillar to widest point of conch), width of the
medial lamina of the quadrate (from base of pillar), widths of
the lateral and medial portions of the ventral condyle (taken
in posterior aspect when measuring width of quadrate. After
recording width of the quadrate, we measured from the
lateral point to the middle of a vertical groove separating the
lateral and medial portions of the condyle. We then
subtracted the width of the lateral portion from the quadrate
width to obtain width of the medial portion of the condyle),
length of the epipterygoid (from the fossa columella to its
dorsal tip), and distance from the fossa columella to the
pointed ventral process in line with the epipterygoid on the
supratemporal process of the parietal.
For the palatal bones, we measured the length (from
anterior tip to posterior tip of nasal process. We located this
process in ventral aspect, because it is usually overlapped by
the nasal dorsally) and alveolar width (widest point below
premaxillary process of maxilla) of the premaxilla, greatest
width of the nasal process of the premaxilla (usually just
above premaxillary process of maxilla), angle of the maxilla–
palatine suture to the maxillary teeth row (in lingual aspect,
points placed at dorsal end of suture where it curves into
lacrimal canal and at base of teeth with apex placed at base of
teeth directly below posterior tip of suture), length of the
maxilla (from just above first maxillary tooth to posterior tip
of bone), length and width of the vomer in palatal view,
length (from anterior tip of bone, usually on palatal ridge but
sometimes in dorsal flange edging interpterygoid vacuity, to
medial tip of its posterior process) and width (palatal ridge to
infraorbital canal at point where palatine–pterygoid suture
enters canal) of pterygoid, anterior separation of pterygoids
in palatal view (at palatine–pterygoid suture from palatal
ridge to palatal ridge), and separation of pterygoids at
posterior inflection (from palatal ridge to palatal ridge at
inflection point where posterior process begins, above
anterior end of basipterygoid process).
In the otoccipital region, we measured the length (from
foramen magnum to base of processus ascendens) and width
(at dorsal constriction, lateral to common crus of anterior
and posterior semicircular ducts) of the supraoccipital,
midline distance from foramen magnum to dorsal constric-
tion of supraoccipital, greatest diameter of the stapedial
footplate, smallest diameter of the stapedial rod; length and
smallest diameter of the basipterygoid process; greatest
diameter of the articular facet of the basipterygoid; distance
from the center of a tuber of the basioccipital to a point in
the middle of the medial side of the facet of the
basipterygoid process; midline distance from the midpoint
between the tubera to the basisphenoid–basioccipital suture;
distance between the ventral tips of tubera (points placed in
center of distal surface of each tuber); length of the left tuber
(base of tuber to its distal point. We established the base of
the tuber when placing the apex of the angle measuring this
process’s orientation in sagittal view); greatest diameter of
the foramen magnum; and distance between the articular
facets (at their midpoints) of the paroccipital processes of the
otoccipitals. To compare relative sizes of selected fenestrae,
74 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
we first measured the greatest diameter (a) of each opening
and then measured a smaller diameter (b) perpendicular to
this diameter. We then calculated the area of each opening
using the formula for an ellipse (¼abp).
We measured the angle separating the supratemporal
processes of the parietal (points placed at the posterior tips
with apex above the fossa for the processus ascendens. For
this process of the parietal, we prefer Oelrich’s 1956 name
‘‘supratemporal process’’ to ‘‘ postparietal process’’ as used
by Evans 2008); the angle formed by the tubera of the
basioccipital with one another (by placing points distally in
center of each tuber then affixing apex at a point halfway
between tubera on surface of basioccipital); the angle
formed by a tuber in sagittal view (by affixing first point on
alar process of sphenoid and second point distally in center
of tuber, then affixing apex to center of base of tuber in line
with first point); and the angle made by a paroccipital
process to horizontal. (We established the horizonal plane by
affixing the first point and apex of the angle to the bases of
the process, at the margin of each recessus scali tympani. We
placed the second point in the center of the articular face of
the paroccipital process. Finally, we subtract the obtuse
angle formed by the points from 1808.) In a parasagittal
plane we measured the downward curve of the supra-
temporal process of the parietal to horizontal (skull
positioned so that maxillary tooth row was horizontal).
For the teeth, in alveolar aspect, we measured the length
(from edge of tooth’s socket to its distal tip) of pleurodont
teeth on the premaxilla, maxilla, and dentary. In lingual view,
we measured height (from gutter to tip of tooth) and width
(where tooth abuts its neighbors) of the second and
penultimate acrodont maxillary and dental teeth.
Herein, we approximate Moody’s (1980) measurements of
the mandible by defining three measurements that abut one
another. We measured the precoronoid length by affixing
anchor points at the anteriormost tip of the mandible and on
the labial side of the mandible in line with the tooth row and
the dorsal tip of the coronoid process. From this point, we
then measured the coronoid–articular length to the posterior
margin of the lateral condyle of the quadrate (this position is
close to the center of the articular, used as a landmark by
Moody 1980) and length of the retroarticular process from
the lateral condyle to the tip of the retroarticular process.
Length of the posterolateral processes of the coronoid also
abuts the precoronoid length, but ends at the distal tip of the
posterolateral process. Additionally, we measured the total
length of the mandible from its anterior tip to the posterior
tip of the retroarticular process, greatest length of the
exposed splenial in lingual view, greatest length of the
adductor fossa in dorsal aspect, height of the dorsal process
of the coronoid in labial view (minimal distance from its
dorsal tip to the dentary), and height of the coronoid in
lingual view (straight-line measurement from dorsal tip of
dorsal process of coronoid to ventral tip of the posteromedial
process of the coronoid. This measure is not usually vertical,
but slightly angled forward, because the posteromedial
process projects behind the dorsal process).
Inferring Species Boundaries
We view species as separately evolving metapopulations
and agree with de Queiroz’s (1998) observation that many, if
not all, contemporary species concepts are special cases of
the general lineage concept. Operationally, we recognize
allopatric populations as species when they are diagnosable
by concordance of multiple apparently independent charac-
ters (Avise and Ball 1990; Wiens and Penkrot 2002).
For interspecific comparisons of meristic data, we used
Tukey’s test to compare means when samples satisfied
assumptions of normality (verified with the Shapiro–Wilk
test) and homogeneity of variance (verified by Levene’s test).
We used the Mann–Whitney U-test as a nonparametric
alternative to Tukey’s test. To test for sexual size dimorphism
in meristic characters, we verified that our meristic
characters satisfied assumptions of homogeneity and nor-
mality, then compared sexes using a Student’s t-test. We
used Welch’s nonparametric t-test when data violated
assumptions. To compare species for mensural traits, we
used analysis of covariance (ANCOVA), treating SVL as a
covariate. For the ANCOVAs, we verified the assumption of
parallel slopes with an F-test. We log-transformed characters
that violated the assumption of parallel slopes. To avoid
inflation of the Type I error rate, we adjusted probabilities of
ANCOVAs using the Bonferroni method whenever the same
set of measurements were used in more than one set of
comparisons. We used the PAST 3.20 statistical software
program (Hammer et al. 2001) for all statistical tests.
In this study, we compared means or size-adjusted means
for 20 meristic and 8 mensural characters. Our comparisons
of means serve two purposes: (1) to identify interspecific
differences of potential diagnostic value and (2) to uncover
evidence of interrupted gene flow. Tukey’s test and the
Bonferroni method avoid inflation of the Type 1 error rate
when using the same data set in multiple comparisons. They
are suitable for testing null hypotheses of the form: Means of
character X are the same for the various species. However,
one might argue that each test of whether a character differs
among species is being used as a proxy for a different null
hypothesis: Population A and population B are the same
species. If the Type 1 error rate is set at 5%, then these tests
become repeated trials with a fixed rate of success. Using the
binomial equation, there is a 35.0% probability that at least 1
of our 28 tests will reject the null hypothesis even if it is true.
This probability falls precipitously when multiple test have
low probabilities. For 28 tests, the probability drops to 4.9%
that four or more tests will be significant. Therefore, when
comparing means, we accept four or more significant results
as supporting the alternative hypothesis that two or more
species are present. We also approached this problem in a
way analogous to the Bonferroni correction. For individual
characters, we multiplied each probability by the number of
characters tested. When a probability is still less than 5%
after making this calculation, we add an asterisk (*)
immediately after the probability.
Phylogenetic Analysis
Based on our survey of morphology, we define 29
characters that we later use to estimate phylogeny of
Phoxophrys. We surveyed each character across a diverse
set of outgroup taxa (Table 4). Then, we polarized characters
using parsimony (method of Maddison et al. 1984) by
mapping the character states on a recent phylogenetic
hypothesis of draconine relationships (Grismer et al. 2016, as
further refined by Wang et al. 2018). These studies
recovered a clade containing Diploderma,Otocryptis,
75
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
mainland Pseudocalotes,andSitana as sister to Phoxophrys
with Acanthosaura sister to this larger clade. We chose to use
this phylogenetic hypothesis to polarize characters, because
it is the most comprehensive to date, both in terms of
number of species and genes sampled.
Our survey includes all known species of the Sunda Shelf
genera Dendragama and Lophocalotes. Neither Grismer et
al. (2016) nor Wang et al. (2018) include these two genera,
however. To err on the side of caution, we treated
Dendragama and Lophocalotes as potential sister groups to
Phoxophrys when polarizing characters.
We analyzed our morphological data matrix with the
branch-and-bound algorithm (Implicit Enumeration) with
the program Tree Analysis Using New Technology (TNT;
Goloboff et al. 2008). We evaluate branch support by (1)
standard bootstrapping using 1,000 replicates and (2) the
implicit enumeration setting in TNT and Bremer support on
suboptimal trees up to five steps longer than our phylogeny.
To test the monophyly of Phoxophrys, we generated a
data set of sequences from the 12S, 16S, and ND4
mitochondrial genes. This data set included 12S and 16S
mitochondrial DNA sequences of Ph. nigrilabris download-
ed from GenBank, new 16S and ND4 sequences from Ph.
anolophium, sequences previously generated by Harvey et
al. (2017b, 2018) and Shaney (2017) for Ph. tuberculata and
19 other draconines. We failed to amplify the 12S gene for
Ph. anolophium, and, for Ph. nigrilabris, we lack the ND4
gene. In the description of Ph. anolophium, we provide
GenBank accession numbers for new sequences generated
in this study.
To sequence the 16S gene fragment of the new species,
we used the forward primer 50CGCCTGTT TAACAAAAA-
CAT-30(16Sf) and reverse primer 50CCGGTCTGAAC
TCAGATCACGT (16Sr; Leach ´
e et al. 2009). To sequence
the ND4 fragment, we used the forward primer
50CACCTATGACTACCAAAAGCTCATGTAGAAGC-
30(ND4) and reverse primer 50CATTACTTTTACTT
GGATTTGCACCA-30(LEU), which targeted an 892–base
pair (bp) region (Harvey et al. 2017b). For these genes, the
thermal cycling profiles consisted of an initial denaturation at
948C for 3 min, followed by 30 cycles of denaturation at 948C
for 30 s, a 508C annealing phase for 45 s, and a 728C
extension for 1 min, followed by a 728C extension for 7 min,
then a holding phase at 48C. We cleaned the products of
amplification using Sera-Mag Speedbeads (Fisher Scientif-
ic), following the procedure outlined by Rohland and Reich
(2012).
We sequenced PCR products in both directions with
Sanger sequencing technology. To edit sequences, we used
Geneious Prime v2019.2.1 (available at https://www.
geneious.com; Kearse et al. 2012). We aligned all sequences
using the MAFFT alignment setting (Katoh et al. 2002;
Katoh and Standley 2013) implemented within Geneious
Prime. We checked the ND4 alignment for stop codons by
eye and trimmed them to equal lengths to remove missing
data.
We conducted maximum-likelihood analyses using raxml-
GUI (Stamatakis 2006). In this analysis, we utilized the
thorough bootstrapping setting, sampling over 10 runs of
10,000 repetitions, and carried out 20 searches for the best
tree. We selected the most likely model of evolution for each
codon position using Bayesian information criteria imple-
mented in jModelTest v2.1.10 (Darriba et al. 2012). Based
on the results of jModelTest, we partitioned by each codon
position using GTR þCþI. We carried out Bayesian
phylogenetic analysis using MrBayes v3.2.6 (Ronquist et al.
2011), and used four independent runs (nruns ¼4) and four
chains for 10 million generations, sampling every 1,000
generations on the CIPRES Science Gateway (Miller et al.
2010). We used default temperatures for chains. We
assessed the appropriate amount of burn-in and convergence
by inspecting the log files in the program Tracer (v1.6;
BEAST, Bayesian Evolutionary Analysis Sampling Trees)
and discarded the first 25% of trees in TreeAnnotator
(v2.4.6; BEAST, Bayesian Evolutionary Analysis Sampling
Trees).
REVIEW OF EXTERNAL MORPHOLOGY AND LIST OF CHARACTERS
In this section, we discuss characters found to vary among
the six species of Phoxophrys. We number characters that we
code for phylogenetic analysis and omit numbers from
characters used for diagnostic and descriptive purposes only.
Many of the excluded characters are continuous or
polymorphic. Although methods for coding polymorphic
and continuous data exist (e.g., Smith and Gutberlet 2001;
Lawing et al. 2008), sample sizes of only Ph. nigrilabris are
adequate, and three of the six species are known from four or
fewer specimens. Small sample sizes would likely lead to
spurious inferences about relationships and unnecessarily
obscure relationships revealed by the various qualitative
characters.
External Morphology
Lorilabial and infraorbital series.—Number and con-
dition of scales below the eye have considerable diagnostic
value among draconines (e.g., Wood et al. 2009; Mahony
2010; Harvey et al. 2017a,b). Between the orbital margin and
supralabials, species of Phoxophrys invariably have an
enlarged row of infraorbitals (Fig. 3). The infraorbitals may
contact the supralabials or be separated from them by a row
of small lorilabials. Like Inger (1960), we use the term
infraorbitals for both types of scales and use the term
lorilabial when we need to draw a distinction between these
two types of scales. Inger’s (1960) key uses variation in these
scales at two places. First, in his Couplet 2, he distinguishes
Ph. nigrilabris from congeners based on a lorilabial
separating the nasal from the supralabials. Then, in Couplet
4, he uses a count of infraorbital rows to distinguish Ph.
borneensis (one row) from Ph. cephalum (two rows). A
continuous series of small lorilabials separates enlarged
infraorbitals from the supralabials in his ‘‘two rows’’
character state, whereas one or more of the enlarged
infraorbitals contacts the supralabials in his ‘‘ one row’’
character state.
With larger samples than those available to Inger (1960),
number of infraorbital rows proves to be polymorphic in
Phoxophrys cephalum. Among our sample of Ph. cephalum,
40% have one row (Table 1) and this character should not be
used to distinguish Ph. cephalum from Ph. borneensis. Inger
(1960:224) also noted some variation in this character, when
he remarked ‘‘ a second syntype [i.e., referring to Ph.
cephalum] has only one continuous row of infraorbitals.’’
76 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
Consequently, we do not use this character to distinguish
between Ph. cephalum and Ph. borneensis in our revised key.
Polymorphism of a character in one species should not
negate its usefulness as a diagnostic character in other
species, because most multistate characters likely pass
through a period of polymorphism during their evolution
(Harvey 2008). Each of the four specimens of Phoxophrys
anolophium has a continuous row of lorilabials separating the
enlarged infraorbitals from the supralabials, whereas one or
more infraorbitals contact the supralabials in the three types
of Ph. borneensis (Inger 1960; this study) in
ZRC(IMG).2.187, on both sides of the two MCZ specimens,
and in all but one of the eight NMW specimens. Thus, this
character will usually distinguish Ph. anolophium from Ph.
borneensis. The holotype of Ph. spiniceps and a specimen
illustrated in our Fig. 1 both lack continuous rows of
lorilabials below the enlarged infraorbitals.
Nasal–supralabial contact.—The nasal usually contacts
one supralabial in Phoxophrys anolophium,Ph. borneensis,
Ph. cephalum, and Ph. spiniceps, whereas a lorilabial usually
separates the nasal from the supralabials in Ph. nigrilabris.
The nasal contacts a supralabial in only 1 of our 20
specimens (5%) of Ph. nigrilabris. Interestingly, the nasal
contacts supralabials 1 and 2 on all eight sides of the four
specimens of Ph. tuberculata. Among the other species, we
found this character state in only one specimen of Ph.
borneensis and one specimen of Ph. cephalum (Table 1).
Contact between suprarictal scale and first elongate
scale above rictal fold.—Lorilabials of Phoxophrys ano-
lophium,Ph. borneensis, and Ph. nigrilabris usually separate
the enlarged suprarictal scale from the first elongate scale of
the rictal fold (Table 1). In contrast, these two scales are in
broad contact in Ph. cephalum.InPh. spiniceps (both sides
of all three specimens), the enlarged scale contacts the last
supralabial, and a small scale behind and slightly below it
separates the enlarged scale from the first scale of the rictal
fold. Lorilabials separate the scales on both sides of the
holotypes of Ph. tuberculata and Japalura robinsoni.
However, a large projecting suprarictal scale broadly
contacts scales of the rictal fold on both sides of the new
male specimen of Ph. tuberculata.
1. Shape of supraciliaries.—The supraciliaries are sub-
rectangular and juxtaposed to imbricating (0) or arched
dorsally, producing a serrate edge to the supercilium (1).
Phoxophrys nigrilabris and Ph. tuberculata have distinctly
arched supraciliaries forming a serrate series, whereas Ph.
anolophium,Ph. borneensis,Ph. cephalum,andPh. spiniceps
have low, subrectangular supraciliaries.
Among outgroups, supraciliaries may be subrectangular to
elongate and overlapping to varying degrees, but other
draconines examined by us did not have arched supraciliaries
like those in Ph. nigrilabris and Ph. tuberculata.
2. Supraciliary spine.—Subrectangular or serrate supra-
ciliaries extend to the postciliary scale or ciliary notch (0) or
they terminate in an elongate spine in the middle of the
supracilium (1).
Phoxophrys spiniceps has a prominent superciliary spine,
positioned above the center of the eye (Fig. 1). This scale is a
modified scale of the supraciliary series. In front of the spine,
the supraciliary scales are subrectangular. Interestingly,
small granular scales extend from the spine to the postciliary
scale. Several scales surrounding the base of the spine are
somewhat elongate and project upward, evidently supporting
the spinose scale.
Among other draconines, postciliary, temporal, and/or
posttemporal scales may be elongated into spines. For
example, the prominent spine at the corner of the eye of
Acanthosaura is a modified postciliary scale. A single spinose
scale above the center of the eye appears to be unique to
Phoxophrys spiniceps.Ceratophora karu has several elon-
gate scales in the center of the supercilium, but they are not
long spines.
FIG. 3.—Snout morphology of Pelturagonia anolophium (female holotype,
MZB 14992) and Phoxophrys tuberculata (UTA 65254). Rostral scales
shaded gray.
77
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
3. Projecting and triangular supracilium.—The supra-
cilium is rounded laterally, unmodified, and roofs the orbit
(0) or it projects upward and outward as a pointed, triangular
flap (1; Fig. 4).
Several draconines have modified supercilia. For exam-
ple, many species of Gonocephalus have vertically arched
supercilia that are sometimes supported by bony pre- and
postorbital processes. None of the draconines examined in
this study have a distinctive triangular supracilium like
Phoxophrys tuberculata. Nonetheless, the supercilium of Ph.
tuberculata resembles that of Ceratophora aspera and C.
karu. Like Ph. tuberculata, the supercilia of these Sri Lankan
endemics projects upward and outward. In all three species a
patch of distally attenuate, flattened, and multicarinate
superciliary scales comprise a triangular patch above the
center of the eye. Pethiyagoda and Manamendra-Arachchi
(1998:Fig. 22) illustrate this structure.
4. Vertical division of rostral scale.—The rostral is entire
(0) or divided into two or more scales (1).
The Bornean species of Phoxophrys usually have three
rostrals (Table 1; Fig. 2) with a postrostral separating the
nasal from the rostral series on each side. Phoxophrys
borneensis exhibits more variation than congeners in the
rostral series with 39% (n¼13) of specimens having four or
five rostrals. In contrast, Ph. tuberculata has a single rostral.
Among outgroups, the three Sri Lankan endemics all have
divided rostrals (four in Ceratophora stoddartii, three in
Cophotis ceylanica,andfiveorsixinLyriocephalus
scutatus). Like its congeners, Bronchocoela hayeki has a
wide medial rostral. As defined here, we consider a small
scale on either side of the medial rostral to be a rostral rather
than the first supralabial, because the lateralmost postrostral
separates this scale from the nasal.
Reduction of the mental scale and its width relative
to the rostral.—Along with vertical division of the rostral,
the mental also appears to have undergone vertical division
in all Bornean Phoxophrys, except Ph. cephalum. In this
species, the mental is wider than or of equal width to the
medial rostral. This character is not applicable to specimens
with even numbers of rostrals.
FIG. 4.—Cephalic morphology of Phoxophrys tuberculata (adult male,
UTA 65254).
TABLE 1.—Frequencies of selected diagnostic characters among Phoxophrys tuberculata and five species of Pelturagonia. Data for Pe. borneensis includes
ZRC(IMG) 2.187 for characters visible in the photos.
Diagnostic character Pe. anolophium Pe. borneensis Pe. cephalum Pe. nigrilabris Pe. spiniceps Ph. tuberculata
Infraorbital rows 1 (0%, n¼4)
2 (100%)
1 (92%, n¼13)
2 (8%)
1 (40%, n¼15)
2 (60%)
1 (10%, n¼20)
2 (90%)
1 (100%, n¼3)
2 (0%)
1 (50%, n¼4)
2 (50%)
Nasal–supralabial
contact
0 (0%, n¼4)
1 (100%)
1–2 (0%)
0 (8%, n¼13)
1 (69%)
1–2 (8%)
2 only (15%)
0 (16%, n¼19)
1 (79%)
1–2 (5%)
0 (95%, n¼20)
1 (5%)
1–2 (0%)
0 (0%, n¼3)
1 (100%)
1–2 (0%)
0 (0%, n¼4)
1 (0%)
1–2 (100%)
Enlarged suprarictal
scale of rictal fold
Separated
(100%, n¼4)
In contact (0%)
Separated
(92%, n¼12)
In contact (8%)
Separated
(0%, n¼14)
In contact (100%)
Separated
(100%, n¼20)
In contact (0%)
Separated
(100%, n¼3)
In contact (0%)
Separated
(67%, n¼3)
In contact (33%)
Rostrals –
3 (100%, n¼4)
2 (8%, n¼13)
3 (53%)
4 (31%)
5 (8%)
2 (15%, n¼13)
3 (85%)
3 (89%, n¼18)
4 (11%)
3(n¼3)
1 (100%, n¼3)
Medial rostral
compared to
mental
Wider
(100%, n¼317)
Wider
(92%, n¼13)
Subequal (8%)
Narrower
(85%, n¼13)
Subequal (15%)
Wider
(100%, n¼13)
Wider
(100%, n¼3)
Wider
(100%, n¼3)
Chin shields
contacting
infralabials
1 (50%, n¼4)
2 (25%)
3 (25%)
1 (8%, n¼13)
2 (31%)
3 (61%)
2 (67%, n¼15)
3 or 4 (33%)
1 (74%, n¼19)
2 (26%)
1 (33%, n¼3)
2 (67%)
1 (25%, n¼4)
2 (25%)
3 (50%)
Tubercular sublabials 4–7 (5 61, n¼4) 2–6 (4 61, n¼13) 0 (n¼14) 2–6 (4 61, n¼19) 2 (n¼3) 2 or 3 (n¼3)
Maximum snout-to-
vent length of
females (mm)
80 62 68 (74 mm, Inger
1960)
53 59 (60, Manthey
2010)
43
78 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
Reduction of the mental is particularly pronounced in
Phoxophrys nigrilabris. Arrangement of the mental and chin
shields resembles other draconines in most species of
Phoxophrys. Because of the mental’s large size, the first pair
of chin shields have relatively short medial contact behind it
or the pair may be separated by one or more gulars. In
contrast, the mental of Ph. nigrilabris is substantially
reduced in size relative to the first pair of chin shields, and
the chin shields have a relatively long point of medial
contact. Unlike any other Phoxophrys, 26% of Ph. nigrilabris
(n¼19) have a single azygous postmental positioned behind
the mental and between the first pair of chin shields. We
refer to this scale as a postmental instead of a gular, because
of its shape and because the first pair of chin shields usually
contact one another behind it. In our sample of Ph.
nigrilabris, a gular and an azygous postmental completely
separate the first pair of chin shields in a single specimen
(5%; n¼20). Finally, we note that a single chin shield
contacts the infralabials in all but one specimen of Ph.
nigrilabris, whereas two or more chin shields contact the
infralabials in Ph. cephalum and all but one Ph. borneensis
(Table 1).
Chin shields contacting infralabials.—In most Bor-
nean Phoxophrys two or more chin shields contact the
infralabials, whereas only the first chin shield contacts the
infralabials in most Ph. nigrilabris.
5. Sublabial tubercular scales.—Sublabial scales are
unmodified (A) or tubercular (B) anterior to a large
tubercular scale below the rictal fold.
Harvey et al. (2017a) recently used counts of tubercular
sublabial scales to diagnose species of Dendragama. These
same scales have diagnostic value in Phoxophrys.All
Phoxophrys have a prominent tubercular scale positioned
below the rictal fold (Fig. 3). Most species also have 2–7
similar tubercular scales on sublabials anterior to this
subrictal scale. However, Ph. cephalum lacks the additional
scales and only has the single subrictal tubercular scale.
Tubercular sublabials may be related to condition of the
gulars. Most Phoxophrys have spinose gulars with prominent
keels and projecting mucrons. In contrast, Ph. cephalum
lacks spinose gulars. Its gulars bear low rounded keels and
lack mucrons.
Among outgroups, most species lack both enlarged
subrictal and sublabial tubercular scales (Table 4). None-
theless, presence of tubercular sublabials in Dendragama
(Harvey et al. 2017a) makes it impossible to polarize this
character without further resolution of draconine phylogeny.
Each of the species of Diploderma examined in this study
have swollen to tubercular sublabials; Cristidorsa planidor-
sata,D. slowinskii,D. splendidum, and D. yunnanese also
have subrictal tubercular scales behind the tubercular
sublabial scales. Lyriocephalus has a continuous row of
enlarged, keeled scales in the same anatomical location as
the sublabial and subrictal tubercles of Phoxophrys; howev-
er, its scales are not tubercular and not separated from one
another by smaller, unmodified scales.
6. Dorsal crests.—A dorsal crest of continuous projecting
scales extends from the pectoral gap to the base of the tail (0)
or is replaced by widely spaced tubercular scales frequently
in pairs or short transverse series (1).
All known Phoxophrys have a nuchal crest of relatively
low triangular scales. Phoxophrys anolophium,Ph. borneen-
sis,Ph. cephalum, and Ph. spiniceps lack a continuous dorsal
crest and instead have widely spaced tubercular (spinose in
Ph. spiniceps) scales along the vertebral midline, behind the
pectoral gap. Inger (1960) noted that Ph. nigrilabris has a
continuous dorsal crest and used this character in his key.
With discovery of male specimens of Ph. tuberculata,we
now know that condition of the dorsal crest is sexually
dimorphic in this species. As Inger (1960) observed, females
lack a crest or, at least, have a poorly defined one. Males
have a continuous, projecting dorsal crest of small triangular
scales extending from the pectoral gap to the base of the tail
(Fig. 1; see also Manthey 2010:Fig. RA03453-4, who
illustrates an additional male specimen from ‘‘ near Solok’’
with a clearly visible dorsal crest extending onto the tail).
Thus, when identifying male Phoxophrys, this character
causes confusion at Couplet 2 of Inger’s key, potentially
leading some to misidentify Ph. tuberculata as Ph. nigrilab-
ris.
Draco sumatranus lacks a dorsal crest, but also lacks the
widely spaced tubercles forming a broken dorsal crest in
Phoxophrys. Widely spaced projecting dorsals occur in
Ceratophora stoddarti and Lyriocephalus scutatus. Howev-
er, these species lack the distinctive paired or transverse
rows of tubercles found in Bornean Phoxophrys.
7. Orientation of middorsal scales on the neck and in the
pectoral gap.—Middorsal scales on the neck point dorsally
and posteriorly (0) or they point anteriorly and undergo a
1808rotation across the pectoral gap (1).
Unusual orientation of scales on the neck and in the
pectoral gap of Phoxophrys nigrilabris is arguably one of this
species’ most distinctive characteristics. With only two
specimens at hand in 1960, Inger did not mention this
character. In this species, middorsal scales on the neck point
anteriorly, whereas they point posteriorly in all congeners.
The middorsal scales then undergo an abrupt 1808
counterclockwise rotation across the pectoral gap to point
dorsally and posteriorly along the dorsal crest. In the center
of the gap, some scales point transversely to the left in 58%
of specimens (n¼33; Fig. 5), whereas in the remaining 42%
of specimens the scales abruptly switch direction somewhere
in the middle of the gap. In all other Phoxophrys, scales
across the entire gap point backward.
Among outgroup species, scales on the sides of the neck
point dorsally and anteriorly and undergo a 1808reorienta-
tion across the pectoral gap in Acanthosaura armata,A.
crucigera, and Gonocephalus grandis.
8. Dorsolateral caudal crests.—The tail lacks dorsolateral
crests (0) or a pair of enlarged, heavily keeled scales form
dorsolateral crests at the base of the tail (1).
Bornean Phoxophrys have prominent dorsolateral crests
(Figs. 1 and 6) at the base of their tails. A row of enlarged
scales, thickened medially and projecting laterally, forms
each of the crests. Inger (1960) illustrated this distinctive
caudal morphology in males of Ph. borneensis,Ph. cephalum,
and Ph. nigrilabris. In his diagnoses, he included counts of
crest scales and used the counts to differentiate Ph.
borneensis from Ph. cephalum. Inger (1960) described
similar crests in Diploderma flaviceps.Diploderma varcoae
and Cristidorsa planidorsata have dorsolateral caudal crests,
but also have a vertebral crest at the base of the tail. All three
crests in these two species are continuous with dorsolateral
79
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
and vertebral crests on the body, whereas the bodies of
Bornean Phoxophrys lack dorsolateral crests.
In Phoxophrys, sexual dimorphism in tail morphology has
been overstated somewhat. Both sexes of all Bornean
Phoxophrys have dorsolateral crests at the base of the tail,
and we did not find intersexual differences in counts of crest
scales for any of the species. Keratinized layers of the crests
thicken substantially in larger male Ph. borneensis and Ph.
cephalum, whereas thickness of the keratinized layers
resembles that of adjacent scales in females.
Counts of crest scales prove more variable than originally
thought (Table 1), and we did not find differences in counts
among Bornean Phoxophrys.
Phoxophrys tuberculata lacks dorsolateral crests and its
vertebral crest extends onto the base of the tail. Our male
specimen has 2/2 multicarinate tubercular scales positioned
dorsolaterally on the proximal tail. These scales are a
continuation of the dorsolateral series of widely spaced
multicarinate scales found on the body. Five/seven small
scales laterally separate each pair of multicarinate scales.
Dorsally, the tail of Ph. tuberculata is somewhat rounded
rather than flat; medially, five scales separate the first pair of
multicarinate scales and three scales separate the second
pair.
9. Projecting scales between dorsolateral crests.—Dorsal-
ly, the tail is flat between the crests (A), or dorsally projecting
scales separate the crests medially (B). This character is not
applicable to any outgroups examined by us, all of which
either lack dorsolateral caudal crests or have a vertebral crest
between the caudal crests.
In Phoxophrys borneensis,Ph. cephalum,andPh.
nigrilabris, the area between the crests is flattened dorsally,
so that the proximal tail is subtriangular in cross section. In
these species, the area between the crests is flat and covered
by scales that are mostly small and keeled, but not
projecting. In contrast, Ph. anolophium and Ph. spiniceps
have tubercular scales between the crests. These scales are
larger than the crest scales in both species, pointed and
projecting in Ph. anolophium and spinose in Ph. spiniceps.In
the other species, a row of flat scales supports the crests
medially; although enlarged, these supporting scales are
FIG. 5.—(A) Relative length of the shank in Pelturagonia and Phoxophrys.
(B) Relative head width in males (solid symbols) and females (open symbols)
of Pelturagonia anolophium,Pe. borneensis, and Pe. cephalum (key to
symbols in A also applies to B).
FIG. 6.—Pectoral gap of Pelturagonia anolophium (A, holotype, MBZ
14992) and Pe. nigrilabris (B, FMNH 138483), illustrating 1808counter-
clockwise rotation of vertebral scales in Pe. nigrilabris.
80 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
nonetheless smaller than scales of the crest. Scales
separating the crests reduce from 4–10 at the beginning of
the crests to 3–4 distally in Ph. borneensis and Ph. cephalum.
In Ph. anolophium scales separating the crests reduce from 6
to 2 (75%, n¼4) or 3 (25%); they reduce to 2 large scales in
the holotype of Ph. spiniceps.
10. Enlarged subcaudals.—Enlarged subcaudals are
absent or in four rows (0) or in two rows (1) near the base
of the tail.
In ventral aspect, the tail tapers distally to the end of the
retracted hemipenis, then bears a subtle swelling extending
about as far as the dorsolateral crests (Fig. 7). This swollen
portion of the tail bears enlarged subcaudals in Bornean
Phoxophrys. As noted by Inger (1960), Ph. cephalum has two
enlarged subcaudals, whereas other Bornean species have
four enlarged subcaudals. The scales are poorly differenti-
ated in females of Ph. anolophium, but countable and
evident in both sexes of the other Bornean species. Eight to
thirteen scales separate the enlarged subcaudals from the
vent. However, these scales increase in size distally and the
last row is almost as large as the first row of enlarged
subcaudals. Phoxophrys tuberculata lacks enlarged subcau-
dals.
Interestingly, in the six males of Phoxophrys borneensis in
the NMW, a wash of greenish to greenish-orange pigment
encircles the tail over the section with enlarged subcaudals
(Fig. 7). We noted an identical band overlying enlarged
subcaudals of Lophocalotes ludekingi in the same collection.
We cannot explain the presence of these bands of pigment.
They warrant histological study and likely represent an
undescribed glandular product or chemical compound
deposited in the stratum corneum of this portion of the tail.
The keys of Inger (1960) and Malkmus et al. (2002) do not
mention the difference in subcaudals; instead, they distin-
guish Phoxophrys borneensis from Ph. cephalum based on
keeling of scales on the sides of the tail. Use of keeling on
sides of the tail is unfortunate for three reasons. First,
enlarged supracloacal scales have heavy keels and these
scales are on the sides of the tail in both species, potentially
confusing some novice users of these keys. Second, distal to
the supracloacal scales some small male (e.g., FMNH
152166, SVL ¼56 mm) and female (e.g., FMNH 251007,
SVL 45 mm) Ph. cephalum have low keels on scales covering
sides of the tail. Possibly, thickening of the beta-keratin–
containing layers obliterates the keels in adults, but this
hypothesis requires further study. Third, the enlarged pair of
subcaudals is by far the most distinctive caudal character of
Ph. cephalum. Accordingly, we omit any mention of keeling
on the sides of the tail in our revised key.
Among outgroups, we found two rows of enlarged, heavily
keeled subcaudals only in Gonocephalus grandis and
Ceratophora stoddartii, whereas four rows occur in several
genera. Bronchocoela hayeki has eight rows of enlarged
subcaudals at the base of its tail. In many Diploderma and
mainland Pseudocalotes, scales on sides of the tail are
relatively large and usually not or only slightly smaller than
subcaudals. When distinguishable (e.g., in D. slowinskii,D.
splendidum,Ps. kakhienensis, and Ps. kingdonwardi) four
medial rows of subcaudals are present in these genera.
Scale organs.—Comparing Phoxophrys to Japalura,
Inger (1960:221) wrote ‘‘all species of Japalura seen have
FIG. 7.—Proximal tails of female Pelturagonia anolophium (A, holotype, MZB 14992) and Pe. borneensis (B, MCZ 43487) in dorsal and lateral aspect.
Dorsolateral crest scales shaded gray.
81
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
hair-like sense organs on the cephalic scales similar to those
described and figured by Scortecci (1937, pl. 1) and Cherchi
(1958) for species of Agama; this type of sense organ is
absent in Phoxophrys.’’ Later, Moody (1980:data matrix in
his Appendix D) coded Phoxophrys as having hair-like sense
organs. Nonetheless, Inger’s remarks misled some subse-
quent authors (e.g., Ananjeva and Stuart 2001), who used
absence of hairlike sense organs to distinguish Phoxophrys
from all other Draconinae.
We confirm Moody’s (1980) coding of this character: all
species of Phoxophrys have hairlike sense organs like those
of other Draconinae. In old specimens, the hairs may be lost
along with patches of stratum corneum; however, we found
at least some remaining sensillae on all specimens examined
in this study.
External meristic characters.Phoxophrys tuberculata
can be described as a large-scaled species (Tables 2 and 3). It
has fewer circumorbitals, transorbitals, subdigital lamellae,
infralabials, ventrals, and scales around midbody than the
Bornean species. Moreover, ranges for the last three of these
characters do not even overlap ranges of the other species.
In contrast, Phoxophrys nigrilabris can be described as a
small-scaled species. Compared to congeners, it has higher
counts of circumorbitals, transorbitals, parietals, supralabials,
infraorbitals from the nasal to the posterior orbital border,
and scales between the nasal and postciliary scale. Ranges for
each of these characters overlap those of at least one of the
other species, but barely so.
Most meristic characters of Phoxophrys cephalum,Ph.
anolophium, and Ph. borneensis overlap. Nonetheless, Ph.
cephalum has fewer gulars than the other two species, more
loreals and subdigital lamellae than Ph. borneensis, and more
supralabials and fourth-toe lamellae within the span of Toe V
than Ph. anolophium. Phoxophrys anolophium has more
nuchal crest scales than either Ph. cephalum or Ph.
borneensis, and more loreals than Ph. borneensis. This
difference in nuchal crest scale count can be attributed to a
higher number of paravertebrals and small vertebrals
interrupting the nuchal crests of these latter two species.
More scales comprise the dorsolateral crests of Ph.
borneensis than in either Ph. anolophium or Ph. cephalum.
We did not find interspecific differences for counts of
scales separating the circumorbital series, total vertebrals,
postrostrals, and combined canthals and supraciliaries (P.
0.10).
External mensural characters.—Treating SVL as a
covariate, Phoxophrys nigrilabris has a relatively longer
shank than the other Bornean species (F
3,39
¼39.28, P,
0.001*; Table 2). Excluding Ph. nigrilabris, the shank of Ph.
tuberculata is relatively shorter than the large Bornean
species (F
3,24
¼3.10, P¼0.046) but not after correcting the
probability value for multiple tests. Phoxophrys nigrilabris
has a relatively wider head (F
4,42
¼3.73, P¼0.01) and larger
eye (orbital length; F
4,42
¼11.24, P,0.001*)than
congeners. Both Ph. nigrilabris and Ph. tuberculata have
relatively narrower snouts (log-transformed character, F
4,42
¼3.819, P¼0.010) and shorter fifth toes (F
3,38
¼7.02, P¼
0.001*. We excluded Ph. tuberculata from this analysis but
our conclusion is supported by the low ratios presented in
Table 2) than Ph. borneensis and Ph. cephalum. We did not
find interspecific differences in head length, length of Finger
IV, or length of Toe IV.
11. Maximum size.—Adults attain SVL greater than 52
mm (0) or do not exceed SVL of 43 mm (1).
The six species of Phoxophrys can be categorized into
three groups based on maximum SVL. Phoxophrys ano-
lophium,Ph. borneensis,Ph. cephalum,andPh. spiniceps
have SVLs exceeding 59 mm, Ph. tuberculata does not
exceed 43 mm SVL, and Ph. nigrilabris is intermediate in
size, attaining a SVL of 53 mm. With a SVL of 80 mm, the
holotype of Phoxophrys anolophium is the largest known
female Phoxophrys exceeding by 5 mm the maximum known
SVL of female Ph. cephalum (Inger 1960; Malkmus et al.
2002). At least in Ph. borneensis and Ph. cephalum, males
reach larger sizes than females. Some adult male Ph.
cephalum are the longest specimens of Phoxophrys, reaching
a SVL of 84 mm and total length of 140 mm (Inger 1960;
Malkmus et al. 2002). Apparent sexual dimorphism of
maximum size may be an artifact of inadequate sampling:
adult female (up to 53.3 mm, n¼9) and male (up to 52.2, n
¼11) Ph. nigrilabris have similar maximum SVLs and the
largest male Ph. borneensis is only 6 mm longer than the
longest female.
Although still known from very few specimens, Phox-
ophrys tuberculata appears to be a tiny species. MZB 9454
with a SVL of only 29.2 mm is gravid, containing a single
shelled egg in its right oviduct. Similarly, the male specimen
(SVL 36.5 mm) has convoluted vasa efferentia and appears to
be sexually mature. The holotype (RMNH 4140) is the
largest known specimen of Ph. tuberculata and has a SVL of
43 mm and total length of 101 mm.
Among near outgroups, species of Acanthosaura,Den-
dragama,Diploderma,Lophocalotes,Pseudocalotes,and
Sitana all attain maximum SVLs greater than Ph. tuber-
culata. Nonetheless, the SVLs of some Sitana such as S.
bahiri (45 mm), S. devakai (46 mm), and S. ponticeriana (49
mm) only just exceed those of Ph. tuberculata (43 mm;
Amarasinghe et al. 2015; Deepak et al. 2016). The Sri
Lankan endemics Ceratophora aspera and C. karu do not
exceed SVLs of 35 mm (Pethiyagoda and Manamendra-
Arachchi 1998) and may be the only other draconines that
are smaller than Ph. tuberculata.
Sexual dimorphism.—Large males of Phoxophrys ceph-
alum, with their robust, pale heads, at first appear to have
proportionally wider heads than females. For example, in his
description of the type series of Ph. cephalum, Mocquard
(1890:130) remarked ‘‘la t ˆ
ete est proportionnellement tr `
es
grosse chez les m`
ales’’ and Malkmus et al. (2002:247, their
key to Phoxophrys) described the heads of male Ph.
cephalum as ‘‘conspicuously large’’ and the heads of male
Ph. borneensis as ‘‘‘normal’ in size.’’ However, these
assertions have never been tested, and available data suggest
that they may well be illusory (Table 2; Fig. 8).
The holotype of Phoxophrys anolophium has a relatively
wide head, 21.3% as wide as SVL. This value falls within our
range for adult male Ph. cephalum (20.1–24.0%, 21.9 6
1.8%, n¼5), leading us to question past assertions of sexual
dimorphism in head width. Constrained by small sample
sizes, this hypothesis can only be tested rigorously for Ph.
nigrilabris and Ph. borneensis. Nonetheless, we also
combined our samples of Ph. anolophium,Ph. borneensis,
and Ph. cephalum because of phenotypic similarity among
these three species.
82 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
TABLE 2.—Variation in selected meristic and mensural characters among four species of Pelturagonia and Phoxophrys tuberculata.
Character Pe. anolophium Pe. borneensis Pe. cephalum Pe. nigrilabris Ph. tuberculata
Circumorbitals 12 or 13 (n¼4) 11–14 (13 61, n¼11) 10–13 (12 61, n¼16) 14–19 (16 61, n¼18) 8–10 (9 61, n¼4)
Transorbitals 17–22 (20 61, n¼4) 15–20 (17 62, n¼12) 15–18 (16 61, n¼15) 20–26 (23 62, n¼20) 13–15 (14 61, n¼4)
Parietals 7–9 (8 61, n¼4) 7–11 (9 61, n¼12) 6–12 (9 62, n¼12) 8–13 (11 62, n¼19) 7 in male
Palpebrals 12–14 (14 61, n¼4) 12–15 (13 61, n¼12) 8–14 (12 62, n¼12) 12–15 (13 61, n¼20) 9–11 (10 61, n¼4)
Loreals between last canthal and supralabials 6–9 (7 61, n¼4) 4–6 (5 61, n¼13) 5–8 (6 61, n¼16) 4–7 (8 61, n¼20) 4–7 (6 61, n¼4)
Supranasals 1 (100%, n¼4) 1 (58%, n¼12)
2 or 3 (42%)
1 (86%, n¼14)
2 (14%)
1 (100%, n¼19) 1/1 in male
Nasal to postciliary 16 or 17 (16 61, n¼4) 13–18 (15 61, n¼12) 12–19 (15 62, n¼14) 15–20 (17 62, n¼19) 14/14 in male
Infraorbital series 10–12 (11 61, n¼4) 10–12 (11 61, n¼13) 9–13 (11 61, n¼13) 12–16 (15 61, n¼19) 10/11 in male
Supralabials 8 or 9 (8 61, n¼4) 7–11 (9 61, n¼13) 9–12 (10 61, n¼1, n¼15) 9–12 (11 61, n¼20) 8–10 (9 61, n¼4)
Infralabials 11 or 12 (12 61, n¼4) 10–13 (11 61, n¼13) 10–13 (11 61, n¼15) 10–13 (12 61, n¼20) 8 (n¼4)
Gulars 31–36 (33 62, n¼4) 26–36 (31 63, n¼12) 23–30 (26 62, n¼15) 29–40 (35 63, n¼20) 25–30 (28 62, n¼4)
Ventrals 51–58 (56 63, n¼4) 48–63 (57 65, n¼12) 52–62 (57 63, n¼13) 56–68 (62 64, n¼20) 40–49 (45 64, n¼4)
Scales around midbody 92–109 (100 68, n¼4) 80–113 (96 69, n¼11) 81–115 (94 68, n¼11) 87–107 (97 66, n¼20) 56–67 (62 65, n¼4)
Nuchal crest 10–12 (11 61,n¼4) 3–8 (5 62, n¼13) 4–8 (5 61, n¼14) 10–17 (13 62, n¼19) 6 (n¼2)
Total vertebrals 73–90 (81 69, n¼4) 61–85 (73 67, n¼10) 60–78 (70 66, n¼9) 60–75 (69 64, n¼19) 61 in male
Lamellae under Finger IV 18 or 19 (n¼4) 15–19 (18 61, n¼4) 17–21 (20 61, n¼13) 16–19 (17 61, n¼20) 13–17 (16 62, n¼4)
Lamellae under Toe IV 18–24 (22 63, n¼4) 19–26 (22 62, n¼12) 23–29 (26 62, n¼14) 19–23 (21 61, n¼20) 17–19 (18 61, n¼4)
Fourth-toe lamellae within span of Toe V 1–4 (3 61, n¼4) 2–5 (4 61, n¼12) 3–7 (4 61, n¼14) 1 or 2 (1 60, n¼19) 0–2 (n¼4)
Maximum snout-to-vent length (SVL; mm) Males: 42.9
Females: 79.5
Males: 68
Females: 62
Males: 76 (84, Inger 1960)
Females 68
Males: 41.5
Females: 41.5
Males: 36.5
Females: 43
Head width/SVL Male: 22.3%
Females: 21.3–23.6%
(n¼3)
Male: 20.3–23.3%
(22.3 61.0, n¼7)
Females: 19.6–22.7%
(21.0 61.2, n¼5)
Males: 20.1–24.0%
(21.9 61.8, n¼5)
Females: 20.0–21.4%
(n¼3)
Males: 22.0–25.0%
(23.4 60.9, n¼10)
Females: 21.7–24.2%
(23.1 60.7, n¼9)
Male: 22.8%
Females: 20.6–22.5%
(n¼3)
Head length/SVL Male: 25.4%
Females: 22.5–27.3%
(n¼3)
Males: 24.3–27.1%
(25.7 61.0, n¼7)
Females: 23.1–25.5%
(24.0 60.9, n¼5)
Males: 23.7–26.9%
(25.6 61.1, n¼5)
Females: 23.0–24.8%
(n¼3)
Males: 24.6–27.6%
(25.9 60.9, n¼10)
Females: 24.6–28.4%
(25.9 61.1, n¼9)
Male: 27.8%
Females: 27.4–28.5%
(n¼3)
Tail length/total length Male: 60.3%
Females: 58.7–61.6%
(n¼3)
Males: 58.3–63.1%
(60.1 61.7, n¼7)
Females: 54.5–63.8%
(59.4 63.4, n¼5)
Males: 62.9–65.5%
(64.2 61.1, n¼4)
Females: 61.1–64.6%
(n¼3)
Males: 59.0–63.5%
(61.8 61.3, n¼10)
Females: 60.0–63.5%
(62.1 61.3, n¼9)
Male: 58.8%
Females: 57.4–61.5%
(n¼3)
Body length/SVL Male: 46.2%
Females 47.2–48.9%
(n¼3)
Males: 36.5–52.0%
(45.5 66.3, n¼6)
Females: 47.6–52.0%
(49.7 62.0, n¼4)
Males: 38.5–49.6%
(46.5 65.3, n¼4)
Females: 48.8–53.0
(n¼3)
Males: 41.8–57.4%
(48.4 64.2, n¼10)
Females: 40.5–54.6%
(46.8 64.6, n¼9)
Males: 52.1%
Females: 47.7–51.4%
(n¼2)
Length of shank/SVL Male: 21.4%
Females: 22.4–24.3%
(n¼3)
Males: 22.2–26.1%
(24.3 61.4, n¼7)
Females: 21.7–27.4%
(23.4 62.3, n¼5)
Males: 20.6–24.2%
(22.6 61.4, n¼5)
Females: 18.3–23.3
(n¼3)
Males: 27.8–31.3%
(n¼29.6 61.2, n¼10)
Females: 27.8–31.6%
(29.6 61.5, n¼9)
Male: 19.4%
Females: 19.4–23.1%
(n¼10
Length of orbit/SVL 10.3–13.7%
(12.3 61.4, n¼4)
10.5–13.1%
(11.5 61.0, n¼12)
9.6–12.9%
(11.4 61.0, n¼9)
10.8–15.7%
(13.3 61.1, n¼20)
9.7–11.7%
(10.8 60.9, n¼4)
Width of snout/SVL 6.8–8.8%
(8.2 61.0, n¼4)
7.5–9.6
(8.8 60.6, n¼12)
8.3–9.7%
(8.8 60.4, n¼9)
7.3–9.1%
(8.3 60.5, n¼20)
7.4–8.7%
(7.8 60.6, n¼4)
Toe V/SVL 10.3–13.0
(12.0 61.2, n¼4)
9.9–13.0%
(11.5 61.0, n¼12)
9.7–13.3
(11.6 61.1, n¼9)
9.6–12.0%
(10.8 67.8, n¼18)
10.3–10.5%
(n¼3)
83
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
TABLE 3.—Summary of all possible comparisons among samples of four species of Pelturagonia and Phoxophrys tuberculata. In the upper half of the table,
we report Tukey statistics and probabilities for comparisons that satisfied assumptions of normality and homoscedasticity. In the lower half, we report
nonparametric comparisons: a Mann–Whitney Ustatistic and Bonferroni-corrected probabilities for samples of each species. Sample sizes of each test appear
in Table 2. NS ¼not significant. For emphasis, we set significant results in bold font, and we add an asterisk to probabilities that are still significant when
multiplied by 28, the number of mensural and meristic characters used in this study.
Results of Tukey’s test (Q,P)
Character Pe. anolophium Pe. borneensis Pe. cephalum Pe. nigrilabris
Infraorbitals
Pe. borneensis 0.57, NS
Pe. cephalum 0.76, NS 1.93, NS
Pe. nigrilabris 9.45, 0.000*13.54, 0.000*15.65, 0.000*
Supralabials
Pe. borneensis 2.92, NS
Pe. cephalum 4.35, 0.03 2.05, NS
Pe. nigrilabris 7.79, 0.000*7.28, 0.000*5.32, 0.004
Ph. tuberculata 1.68, NS 0.85, NS 2.25, NS 5.62, 0.002
Nasal to postciliary
Pe. borneensis 2.20, NS
Pe. cephalum 4.52, NS 0.40, NS
Pe. nigrilabris 0.95, NS 4.86, 0.007 5.54, 0.002
Gulars
Pe. borneensis 2.06, NS
Pe. cephalum 6.22, 0.001*5.97, 0.001*
Pe. nigrilabris 1.44, NS 5.41, 0.003 12.55, 0.000*
Ph. tuberculata 4.00, NS 2.84, NS 1.19, NS 6.60, 0.000*
Ventrals
Pe. borneensis 1.12, NS
Pe. cephalum 1.07, NS 0.09, NS
Pe. nigrilabris 4.53, 0.019 5.03, 0.007 5.26, 0.005
Ph. tuberculata 5.36, 0.004 7.68, 0.000*7.67, 0.000*11.45, 0.000*
Nuchal crest
Pe. borneensis 9.08, 0.000*
Pe. cephalum 9.44, 0.000*0.409, NS
Pe. nigrilabris 2.83, NS 18.76, 0.000*19.62, 0.000*
Scales around midbody
Pe. borneensis 1.35, NS
Pe. cephalum 2.28 NS 1.32, NS
Pe. nigrilabris 1.07, NS 0.54, NS 1.91, NS
Ph. tuberculata 11.59, 0.000*12.69, 0.000*10.8, 0.000*13.89, 0.000*
Dorsolateral caudal crests
Pe. borneensis 4.09, 0.029
Pe. cephalum 0.77, NS 6.99, 0.000*
Pe. nigrilabris 1.98, NS 3.45, NS 4.24, 0.022
Lamellae under Finger IV
Pe. borneensis 1.88, NS
Pe. cephalum 2.69 NS 6.63, 0.000*
Pe. nigrilabris 2.92, NS 1.41, NS 8.96, 0.000*
Ph. tuberculata 5.03, 0.007 4.28, 0.031 8.96, 0.000*3.57, NS
Fourth-toe lamellae within span of Toe V
Pe. borneensis 2.21, NS
Pe. cephalum 4.18, 0.025 2.78, NS
Pe. nigrilabris 4.59, 0.011 10.31, 0.000*13.9, 0.000*
Results of Mann–Whitney test (U, Bonferroni-corrected P)
Character Pe. anolophium Pe. borneensis Pe. cephalum Pe. nigrilabris
Transorbitals
Pe. borneensis 9.5, NS
Pe. cephalum 4.5, NS 49, NS
Pe. nigrilabris 14, NS 0.5, 0.000*0, 0.000*
Ph. tuberculata 0, NS 1, NS 2.5, 0.048 0, 0.020
Circumorbitals
Pe. borneensis 21.5, NS
Pe. cephalum 20, NS 60.5, NS
Pe. nigrilabris 0, 0.022 4, 0.000*0, 0.000*
Ph. tuberculata 0, NS 0, 0.032 1, 0.038 0, 0.021
Parietals
Pe. borneensis 15, NS
Pe. cephalum 21.5, NS 64.5, NS
Pe. nigrilabris 4.5, 0.040 34.5, 0.007 44.5, 0.028
Loreals (last canthal–supralabial)
Pe. borneensis 1, 0.018
Pe. cephalum 22, NS 32, 0.006
Pe. nigrilabris 26.5, NS 1.5, 0.000*58, 0.008
Ph. tuberculata 4.5, NS 13, NS 26, NS 9.5, NS
84 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
We did not find head-width sexual dimorphism in Ph.
nigrilabris (F
1,17
¼0.69, P¼0.421), Ph. borneensis (F
1,9
¼
3.08, P¼0.113), or in our combined samples of Ph.
anolophium,Ph. borneensis, and Ph. cephalum (F
1,23
¼2.98,
P¼0.098). As for head width, we did not detect sexual
dimorphism in tail length, body length, or length of the
shank (ANCOVA P.0.20), when treating SVL as a
covariate in Ph. borneensis or Ph. nigrilabris. Male Ph.
borneensis have proportionately longer heads (F
1,9
¼14.24,
P¼0.004) than females, but we did not find a similar
difference in Ph. nigrilabris.
Male (n¼7) Phoxophrys borneensis appear to have more
ventrals than females (n¼5), although our comparison was
just nonsignificant (t¼2.15, P¼0.057). We did not detect
intersexual differences in counts of nuchal crest scales or
total vertebrals in this species (P.0.12). Comparing males
(n¼10) to females (n¼9) of Ph. nigrilabris, we did not find
intersexual differences in numbers of ventrals, nuchal crest
scales, dorsal crest scales, or total vertebrals (P.0.2).
Coloration.—The species of Phoxophrys are cryptically
colored: primarily brown and green with transverse bands on
the body, limbs, and tail (Fig. 1). Usually, the most
prominent band crosses the neck, is widest paravertebrally,
and extends ventrally behind the antehumeral fold. Facial
pattern invariably includes a white, yellow, or pale saffron
subocular stripe or blotch, extending from the eye to the first
elongate scale edging the rictal fold. This subocular stripe
and scales of the orbito-tympanic series usually abut a darkly
pigmented area on the lower side of the head. However, this
region is pale green like the rest of the head in adult male Ph.
cephalum. All species have pale blue buccal epithelia and
lack the bright yellow or orange buccal pigment diagnostic of
some species of Dendragama,Lophocalotes,andPseudoca-
lotes (Harvey et al. 2014, 2017a, 2018). Phoxophrys also lacks
the extensive black peritoneal pigmentation characteristic of
most draconines (Harvey et al. 2014, 2017a, 2018).
Both sexes of Phoxophrys borneensis have prominent
charcoal stripes extending posteriorly and medially across
the throat (Fig. 7). In contrast, the holotype and one
TABLE 3.—Continued.
Results of Mann–Whitney test (U, Bonferroni-corrected P)
Character Pe. anolophium Pe. borneensis Pe. cephalum Pe. nigrilabris
Palpebrals
Pe. borneensis 16, NS
Pe. cephalum 6.5, NS 28.5, NS
Pe. nigrilabris 27.5, NS 113.5, NS 54.5, NS
Ph. tuberculata 0, NS 0, 0.035 7.5, NS 0, 0.015
Lamellae under Toe IV
Pe. borneensis 21, NS
Pe. cephalum 5, NS 21.5, 0.013
Pe. nigrilabris 31.5, NS 89, NS 3, 0.000*
Ph. tuberculata 2.5, NS 1, NS 0, 0.032 1, 0.023
TABLE 4.—Matrix of character states for five species of Pelturagonia,Phoxophrys tuberculata, and outgroup species. ‘‘Near Outgroups’’ appear in first half
of table. na ¼not applicable.
Species
Characters
1 2 3 4 5 6 7 8 9 1011 12 1314151617 18 19 20212223242526272829
Pe. anolophium 0001B101 B 0 0 1 0 BB B A 1 1 0 0 0 1 1 1 1B 1 1
Pe. borneensis 0001B101 A 0 0 1 0BA A A 1 0 0 0 0 0 1 1 1B 1 1
Pe. cephalum 0001A101 A 1 0 1 0 BB A A 1 0 0 0 0 1 1 1 1 B 1 1
Pe. nigrilabris 1001B011 A 0 0 1 0 BB A A 2 0 1 1 1 0 1 1 0 A 1 1
Pe. spiniceps 0101B101 B 0 0 1 0 B B B A 1 0 0 0 0 1 1 B 1 B 1 1
Ph. tuberculata 1010B000na 0 1 1 1B 1 A B 2 0 0 0 0 0 1A 0A 0A
Outgroup Node 0 0 0 0 ? 0 0 0 ? 0 0 0 0 ? ? ? ? 0 0 0 0 0 0 0 ? 0 ? 0 0
Lophocalotes 0000A000na 0 0 0 0 A A A B 0 0 0 0 0 0 0B 0 B 0 B
Dendragama 0 0 0 0 B 0 0 0 na 0 0 0/1 0 A B B B 0/1 0/1 0 0 0 0 0 B 0 A 0 A
Diploderma splendidum 0000B000na 0 0 0 0BB B A 1 0 0 0 0 1 0 A 0A 1 B
Pseudocalotes floweri 0000A000na0 0 0 0 AB A A 0 0 0 0 0 0 0 A0 A0 B
Pseudocalotes kingdonwardi 0000A000na0 0 0 0 AB A A 0 0 0 0 0 0 0 B 0 A 0 1
Sitana 00000 000na0 0 0 0 BB A A 0 0 0 0 0 0 0 A1 A0 B
Acanthosaura armata 0000A010na 0 0 0 0ABnaA 0 0 0 0 0 0 1 A0 A0 B
Acanthosaura crucigera 0000A010na 0 0 0 0A 1naA 0 0 0 0 0 0 1 B 0 A 0 B
Calotes mystaceus 0000A000na 0 0 0 1 B Bna A 0 0 0 0 0 0 0 B 0 A 0 B
Aphaniotis acutirostris 0000A000na 0 0 0 0AB A A 0 0 0 1 0 0 0 A 0 A 0 1
Draco sumatranus 0000A100na 0 0 0 0AA A A 0 0 0 0 1 1 0 A 0 A 1 B
Bronchocoela hayeki 0001A000na 0 0 0 0AA B A 0 0 0 0 0 1 0B 0A 0B
Gonocephalus grandis 0000A010na1 0 0 0AB A A 1 0 0 2 0 1 0 B 0A 0 B
Ptyctolaemus collicristatus 0000A000na 0 0 0 0 B B A A 0 0 0 0 0 0 0 A 0 B 0 B
Salea anamallayana 0000A000na 0 0 1 0 B B A A 0 0 0 0 0 0 0B 0 A 0 B
Lyriocephalus scutatus 0001A100na 0 0 0 0AB B A 1 0 0 0 0 1 0B 0A 1 1
Ceratophora stoddartii 0001A100na 1 0 1 0B B B A 2 0 0 0 0 0 0 B 0 A 1 B
Cophotis ceylanica 0001A000na 0 0 1 0BB B A 0 0 0 0 0 0 0 A 0A 1 B
85
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
paratype of Ph. anolophium (MZB 14993) have some pale
saffron tubercles laterally, but their throats are otherwise
immaculate. Throats of the other two paratypes have heavy
charcoal mottling, but the dark pigment does not form
regular stripes.
Dorsal coloration will immediately distinguish adult male
Phoxophrys borneensis from male Ph. cephalum. Male Ph.
borneensis are very dark green (dark blue in preservative)
with narrow yellow bands (white in preservative; 1–2 dorsals
long and including enlarged tubercles), whereas male Ph.
cephalum are light green (pale blue or gray in preservative)
with narrow, dark brown chevrons (black edged, dark blue,
or brown scales in preservative). Male Ph. cephalum have
pale green or pale gray heads, immaculate except for few
scattered dark green spots and lacking obvious lines radiating
across the palpebrals. In contrast, male Ph. borneensis have
dark green heads with large yellow blotches and bands edged
in black (white in preservative) and black-edged bands
radiating out from the ocular aperture across the palpebrals
and sides of the head. We have not seen enough female
specimens in life to propose color characters to distinguish
these two species.
All Phoxophrys have a distinct stripe extending from an
outward bulge produced by the epiphyseal tuberosity of the
ilium down to the insertion of the leg. This stripe is white,
yellow, or pale saffron and edged in black in all species
except Ph. tuberculata, where it is entirely black or absent
altogether. The stripe covers enlarged scales that armor the
ilium.
Dermatocranial Roofing Bones
12. Pineal foramen.—The pineal foramen is present (0) or
completely absent (1).
The species of Phoxophrys lack a pineal foramen and have
extensive sculpturing of the frontal and parietal bones,
including along their sutures. Among other Sunda Shelf
draconines, we observed loss of the pineal foramen in
Dendragama where this character varies within the genus.
Dendragama schneideri has a small foramen, whereas D.
australis,D. boulengeri, and D. dioidema lack a pineal
foramen.
Our coding of this character differs from that of Moody
(1980). For his Character 16, Moody (1980) assigned State 1
to both Dendragama and Phoxophrys; however, he defined
this state as ‘‘tiny pinhole or occluded.’’ Thus, defined more
broadly, Moody (1980) also reported a pinhole or occluded
pineal foramen in Ceratophora,Cophotis,Coryphophylax,
Draco volans,Gonocephalus chamaeleontinus,Lophocalotes,
Lyriocephalus, and Pseudocalotes. Our specimens of Cera-
tophora,Cophotis, and Salea lack a pineal foramen. Our
female specimen of Lyriocephalus has a tiny foramen and
the male has a relatively large foramen. All remaining
outgroup taxa have a pineal foramen.
13. Transverse shortening of the prefrontal and frontal–
nasal contact.—In dorsal aspect, the prefrontal bone
separates the nasal and frontal (0) or a narrow anterior
process of the frontal extends medial to the prefrontal to
contact the nasal (1).
The prefrontal of Phoxophrys tuberculata is modified in
two ways relative to congeners. It forms a prominent
triangular projection dorsolaterally, evidently to support
the modified supercilium of this species. In congeners, a
small knoblike process of the prefrontal extends the canthus
for a short distance into the orbit. Medially, the prefrontal
appears to be somewhat shortened relative to congeners. In
our two skulls of Ph. tuberculata, a narrow process of the
frontal separates the nasal from the prefrontal and contacts
the facial process of the maxilla, whereas the nasal and
prefrontal contact one another in all other congeners.
Among other Draconinae examined, the frontal separates
the nasal and prefrontal only in Calotes mystaceus.
The frontal of Phoxophrys resembles that of most
draconines. Compared to congeners, Ph. nigrilabris has a
narrower frontal, as reflected by its low frontal width/length
ratio (34.3–34.5% vs. 45.0–57.8% in other Phoxophrys). We
cannot say whether the narrow frontal of this species is
apomorphic, however, because relative frontal width varies
widely among draconines (26.5–61.1%). Acanthosaura,
Diploderma,Gonocephalus,Pseudocalotes,andSitana all
contain species with frontals as narrow as or narrower than
that of Ph. nigrilabris. Bony extensions of the frontal along
the ocular margin support the highly modified supercilium of
this species, and presence of these extensions makes the
frontal of Ph. tuberculata substantially more concave than
the frontals of congeners.
Circumorbital Bones and Temporal Arcade
Relative width of supratemporal fossa.—Compared to
other agamids, the species of Phoxophrys have relatively
narrow supratemporal fossae (ANCOVA of width vs. length
of fossa F
33
¼9.06, P¼0.005), although relative dimensions
of the fossa overlap substantially with species in other
draconine genera. Width of the supratemporal fossa is 51.2–
70.9% of its length in Phoxophrys compared to 53.5–84.3%
in other genera examined.
14. Erosion of the prefrontal below the lacrimal canal.—
The prefrontal forms the ventral border of the lacrimal canal
(A) or it is eroded below the canal so that the palatal shelf of
the maxilla forms the canal’s ventral border (B).
FIG. 8.—Ventral coloration of adult male Pelturagonia borneensis,
illustrating black lines in gular region and greenish coloration of enlarged
subcaudals. From left to right, specimens are NMW 20854-3, 20854-1,
20854-7, and 20854-5 (SVL of NMW 20854-5 is 59 mm). A color version of
this figure is available online.
86 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
Like many agamids (Siebenrock 1895; Moody 1980), the
species of Phoxophrys lack a lacrimal bone and have a
relatively large lacrimal canal. Moody (1980) included this
character in his study of agamid relationships. Complete
erosion of the prefrontal further enlarges the lacrimal canal
in Phoxophrys, so that the canal extends to the palatal shelf
of the maxilla. Among outgroups, we observed this condition
in Calotes,Cophotis,Ceratophora,Diplodema,Ptycholae-
mus,Salea, and Sitana, whereas the prefrontal forms the
ventral portion of the canal in other genera examined in this
study. Notably, the prefrontal forms the ventral border of the
lacrimal canal in Sunda Shelf species other than Phoxophrys
and Calotes. Although this character may contain phyloge-
netic signal within the Draconinae, it is invariate within
Phoxophrys and both states occur in near outgroups of the
genus.
15. Anterior extent of squamosal in lateral aspect.—In
lateral aspect, the squamosal abuts the jugal (A), extends for
a short distance horizontally below the jugal (B) or the
squamosal extends anteriorly below the jugal and curves
ventrally to brace against the vertical posterior margin of the
jugal (1).
Posteriorly, the jugal is nearly vertical before it curves
backward to contribute to the temporal arch (Fig. 9). In the
temporal arch, this bone laterally overlaps the anterior end of
the squamosal. The squamosal extends furthest anteriorly on
the medial side of the jugal; nonetheless, a portion of the
squamosal braces the jugal ventrally in most draconines (Fig.
9). This bracing portion is short, narrow, and horizontal in
Phoxophrys anolophium,Ph. cephalum,Ph. nigrilabris, and
Ph. spiniceps. Phoxophrys tuberculata has a substantially
broader and longer bracing portion, extending below the
jugal and curving ventrally to brace against the vertical
posterior margin of the jugal. At the other extreme, Ph.
borneensis lacks a bracing portion of the squamosal below
the jugal, and the squamosal and jugal abut one another in
lateral aspect.
Most outgroup species have short, narrow bracing
portions, resembling those of most Bornean Phoxophrys.
Like Ph. borneensis,Bronchocoela hayeki,Draco sumatra-
nus, and both species of Lophocalotes lack bracing portions,
and the jugal and squamosal abut one another in these
species. Phoxophrys tuberculata has the most robust
bracing portion of any draconine. Among the various
outgroup species, only the bracing portion of Acanthosaura
crucigera is comparable. As in Ph. tuberculata the bracing
portion of this species reaches the vertical portion of the
jugal.
16. Postorbital process of frontal and postciliary orna-
ment.—A lateral process of the frontal is separated from the
postciliary ornament (A) or reaches the postciliary ornament
(B).
A rounded knob on the postorbital bone, hereafter
referred to as the postciliary ornament, lies at the dorso-
posterior border of the orbit in all Phoxophrys (Fig. 10). The
postciliary scale named by Harvey et al. (2017a) overlays this
ornament. In Ph. anolophium,Ph. spiniceps,andPh.
tuberculata a postorbital process of the frontal reaches this
ornament laterally, whereas a sizable gap separates the
process and ornament in Ph. borneensis,Ph. cephalum, and
Ph. nigrilabris.
Among outgroups, the postorbital process of the frontal
reaches the postciliary ornament in Bronchocoela,Dendra-
gama,Diploderma, and the Sri Lankan endemics, whereas a
gap separates the distal end of the process from the
postciliary ornament in the remaining species (this character
is not applicable to Acanthosaura armata,A. crucigera, and
Calotes mystaceus, all of which lack postciliary ornaments).
17. Orientation of postorbital process of frontal.—The
postorbital process of the frontal is oriented transversely (A)
or angled antero-laterally because of substantial expansion of
the postorbital above the orbit (B).
The postorbital extends substantially above the orbit in
Phoxophrys tuberculata so that the postorbital process of the
frontal is angled antero-laterally, whereas this process is
transversely oriented in the remaining species and the
postorbital does not extend above the orbit. Like Ph.
tuberculata, the two species of Lophocalotes and five species
FIG. 9.—Morphology of the temporal arcade and conch of the quadrate in
Pelturagonia anolophium (A, MZB 14992) and Phoxophrys tuberculata (B,
UTA 65254). Skulls illustrated in lateral aspect and rotated slightly ventrally.
87
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
of Dendragama have antero-laterally angled postero-lateral
processes of the frontal, whereas this process is straight in
the remaining outgroups.
18. Postciliary bony ridge.—A bony ridge on the
postorbital is absent, inconspicuous, or very short (0);
extends from the postciliary ornament to the ventral margin
of the supratemporal fossa (1); or curves ventrad to extend
vertically down the postorbital (2).
Phoxophrys anolophium,Ph. borneensis,Ph. cephalum,
and Ph. spiniceps have a relatively short ridge of bone,
extending from the postciliary ornament to the temporal
fossa and contributing to the ventral border of this fossa. In
contrast, the ridge curves ventrad and is relatively long in Ph.
nigrilabris and Ph. tuberculata. In these two species, the
ridge does not reach the temporal fossa, although it closely
approximates the fossa in the holotype of Japalura robinsoni.
Most draconines either lack postciliary ridges or have
ridges that reach the temporal fossa. Among our outgroups,
only Ceratophora has a prominent postciliary ridge that
curves ventrad like the ridges of Ph. nigrilabris and Ph.
tuberculata.
Otoccipital Region of Braincase
19. Ossification of parasphenoid.—The parasphenoid is
ossified (0) or the floor and cultriform process of the
parasphenoid are membranous (1).
The large holotype of Phoxophrys anolophium has ossified
trabeculae communis but a membranous hypophyseal floor
and cultriform process of the parasphenoid, whereas all
three elements are ossified in congeners. Among outgroups,
we observed this trait in both specimens of Dendragama
australis and in the male of D. boulengeri. All four specimens
with this character are large adults, and the membranous
portions of the parasphenoid cannot simply be ascribed to
ontogenetic variation.
20. Reduction of lateral aperture of recessus scali
tympani.—The area of the lateral aperture of the recessus
scali tympani is more than 21% as large as the fenestra ovalis
(0) or it is less than 19% as large (1).
Phoxophrys nigrilabris (Fig. 11) has a noticeably reduced
lateral aperture to the recessus scali tympani. In congeners,
the lateral aperture is 42–94% as large as the fenestra ovalis,
whereas it is only 17–18% as large in Ph. nigrilabris. On the
other hand, the medial aperture of the recessus has a size
comparable to other agamids. Consequently, the ratio of the
area of the medial to the lateral aperture is 74–75% in Ph.
nigrilabris, but only 11–27% in congeners and 22–63% in
other agamid genera.
21. Erosion of crista interfenestralis and base of tuber of
the otoccipital.—The crista interfenestralis is eroded below
the recessus scali tympani and does not extend onto the base
of the tuber of the otoccipital (0), or it extends onto the tuber
to merge with the crista tuberalis (1).
In all Phoxophrys, the crista tuberalis extends onto to the
posterolateral face of the tuber. However, the crista
interfenestralis does not extend to the tuber in any
Phoxophrys except Ph. nigrilabris.InPh. nigrilabris, the
crista interfenestralis extends as a narrow lamina to the
anterior base of the tuber, and the crista tuberalis crosses the
lateral face of the tuber to merge with the crista
interfenestralis. Between the confluence of these two cristae
and the ventral margin of the aperture of the recessus scali
tympani, the basioccipital is excavated in Ph. nigrilabris,
leaving an elongate depression extending from the aperture
of the recessus antero-ventrally to the base of the tuber.
Among outgroups, the crista interfenestralis extends onto
the tuber in Aphaniotis acutirostris; however, this species
has a relatively large lateral aperture of the recessus scali
FIG. 10.—Shape of the postorbital process of the frontal and morphology
of the postciliary ornament in Pelturagonia anolophium (A, MZB 14992), Pe.
borneensis (B, FMNH 71856), and Phoxophrys tuberculata (C, UTA 65254).
Skulls shown in dorsal aspect.
88 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
tympani and lacks the distinctive excavation of the basioc-
cipital. In Lyriocephalus scutatus, the crista interfenestralis
extends to the crista tuberalis, but these cristae join just
below the recessus scali tympani rather than at the base of
the tuber as in Ph. nigrilabris. As for A. acutirostris, this
species also lacks the excavation of the basioccipital. Below a
relatively elongate recessus in Gonocephalus grandis,
shallow excavation of the basioccipital extends onto the
tuber proximally; however, the crista interfenestralis is
eroded in this species as in Phoxophrys other than Ph.
nigrilabris.
22. Merger of crista alaris and crista prootica.—A distinct
gap separates the cristae alaris and prootica (0) or these two
cristae merge with one another (1).
In Phoxophrys nigrilabris, the crista alaris merges
seamlessly with the crista prootica (Fig. 11), whereas in
most other agamids, the crista alaris approaches the crista
prootica, but does not merge with it. Among outgroups
examined in this study, the cristae alaris and prootica merge
only in Draco sumatranus.
23. Lateral flaring of crista prootica behind crista alaris.—
Except for a broad excavation below the crista alaris, the
crista prootica is straight (0) or flared behind the crista alaris
(1).
The crista prootica of Phoxophrys nigrilabris is noticeably
reduced relative to other agamids, consisting of merely a low,
straight ridge. The crista prootica is shelflike and roofs a
distinct groove (i.e., the recessus vena jugularis, sensu
Oehrich 1956) in all other agamids examined by us. Below
the crista alaris, the crista prootica bears a broad excavation
in most agamids. All Phoxophrys except Ph. nigrilabris have
the excavation below the crista alaris, but development of the
shelf varies somewhat among specimens. The holotype of Ph.
anolophium, male specimen of Ph. cephalum, and holotype
of Ph. spiniceps have conspicuous latero-ventral flaring (i.e.,
expansion) of the crista prootica both behind and in front of
the excavation, whereas the female Ph. cephalum only has
conspicuous flaring behind the crista alaris. The males of Ph.
borneensis and Ph. tuberculata lack conspicuous expansions
and their cristae prootica are straight except for the
excavation.
Among outgroups, the crista prootica flares outward
behind the crista alaris in Diploderma spendidum,Draco
sumatranus,Bronchocoela hayeki,Gonocephalus grandis,
and Lyriocephalus scutatus.
Palatal Bones and Teeth
Maxillary foramina.—The Bornean Phoxophrys have
three maxillary foramina in the facial process, whereas Ph.
tuberculata has only two. Nonetheless, we chose not to code
this character for the following reasons. Among other
draconines, our counts of these foramina range from one
to five and show considerable variation both among and
within genera. Counts are also polymorphic in some species.
For example, our specimens of Dendragama australis have
one foramen, whereas D. boulengeri and D. schneideri have
three foramina. The trait is polymorphic in D. dioidema: the
male specimen has two foramina and the female has one.
Finally, small foramina are not always well resolved in CT
scans, and some counts in our data may be artifacts of poor
scan quality.
Elongation and narrowing of pterygoid.—Compared
to other draconines, the species of Phoxophrys have
relatively long and narrow pterygoids (F
1,31
¼28.11, P¼
0.000*; Fig. 12). In Phoxophrys the pterygoids are 56.1–
64.7% of skull length (the female Ph. cephalum is the only
specimen with pterygoids less than 60% as long as the skull).
Among the various species we examined, relative pterygoid
lengths of four species fell within the range of Phoxophrys:
Acanthosaura crucigera (56.3%), Draco sumatranus (58.0%),
Sitana ponticeriana (58.6%), and Lyriocephalus scutatus
FIG. 11.—Several characters of the otoccipital region of the braincase of Pelturagonia nigrilabris (A, FMNH 269084) and Pe. anolophium (B, MZB 14992).
Skulls oriented slightly ventro-posterior to lateral with bones of maxillary arcade removed.
89
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
(61.6%), whereas the remaining draconines had pterygoids
47.1–55.8% as long as head length.
24. Narrowing of alveolar portion of premaxilla.—The
premaxilla is more than 11% of head width and bears three
or more teeth (0), or it is less than 10% of head width and
bears a single tooth or a large single tooth flanked by a pair of
much smaller teeth (1).
In Phoxophrys, the alveolar portion of the premaxilla is
reduced relative to other draconines. Measured in alveolar
view, the premaxilla is 2.8–9.6% of head width. Acantho-
saura has a comparable ratio of 7.8–11.0%, whereas
remaining draconines have premaxillae 11.1–21.7% of skull
width.
The narrow premaxillae of Phoxophrys frequently have a
single large pleurodont tooth. Nonetheless, the holotypes of
both Ph. spiniceps and Japalura robinsoni have a very small
tooth on either side of the large medial tooth. The female of
Ph. cephalum has two small teeth on its premaxilla, but lacks
a large medial pleurodont tooth.
After returning most specimens to museums, we attempt-
ed to count premaxillary teeth on whole specimens. Many of
the FMNH specimens of Ph. nigrilabris were preserved with
their mouths open and the premaxilla could be examined by
gently pressing on the upper lip of others. Among 12 of these
Ph. nigrilabris, the premaxillae of 10 have three teeth, one
has a single large pleurodont tooth, and one has two very
small teeth separated by a gap where the large medial tooth
would normally be. Of the 10 with three teeth, all except 1
(FMNH 147657) resemble the holotypes of Ph. spiniceps
and Japalura robinsoni in having a large medial pleurodont
tooth flanked on either side by a single tiny tooth. FMNH
147657 has three premaxillary teeth that are subequal in size
but somewhat smaller than the medial pleurodont tooth of
other specimens of this species. In all of these Ph. nigrilabris
the medial pleurodont tooth of the premaxilla is substantially
smaller than the largest pleurodont tooth on the premaxilla.
Among other draconines examined, Acanthosaura armata
and A. crucigera have single pleurodont teeth and all other
species examined have 3–5. Counts of acrodont and pleuro-
dont teeth on other bones of Phoxophrys are comparable to
the draconines examined.
25. Suturing and fusion of premaxillary processes of
maxillae.—The premaxillary process of the mandible abuts
or barely extend onto the premaxilla (A), extends onto the
premaxilla to closely approximate the contralateral element
(B), or sutures or fuses medially with the contralateral
element (1).
The premaxillary processes of the maxillae abut the
premaxilla in Phoxophrys tuberculata; whereas, in most
Bornean congeners, the premaxillary process of the maxilla
extends onto the face of the premaxilla to suture (female
specimen of Ph. cephalum) or fuse with the contralateral
element. In the holotype of Ph. spiniceps, the premaxillary
processes closely approximate one another, but do not form a
suture. Moody (1980:Character 43.2, Appendix D) found
suturing or fusion of the premaxillary processes to be
relatively rare in agamids, observing this character state only
in Phoxophrys,Bronchocoela,Ceratophora,Gonocephalus
interruptus, and Lyriocephalus. Except for Phoxophrys,a
gap separates the premaxillary processes [Moody’s Character
43.1] in specimens of these genera examined by us. We have
not examined skulls of G. interruptus. Nonetheless, the
processes only partially overlap the premaxilla in our large
male G. grandis [43.1]. The processes almost meet in
Ceratophora stoddartii [43.1].
26. Absence of palatal ridge from palatine bone.—The
palatal ridges extend anteriorly beyond the palatine–ptery-
goid suture and terminate on the palatine (0), or they
terminate posterior to the suture on the pterygoid (1).
The palatal ridges of Phoxophrys tuberculata and Ph.
nigrilabris cross the pterygoid–palatine suture as in all other
draconines examined except Sitana, whereas these ridges
stop short of the suture in Ph. anolophium,Ph. borneensis,
Ph. cephalum, and Ph. spiniceps (Fig. 12).
27. Condition of ventral flange on posterior pterygoid.—
Near its posterior end, the pterygoid bears a distinct, high
ventral flange (A) or lacks a flange (B).
Stilson et al. (2017) defined two character states for the
ventral flange of the posterior process of the pterygoid. Near
the pterygoid–quadrate articulation, the flange may be
hooked (their Character 14.1) or not (14.0). Referring to
illustrations provided in their supplementary material, skulls
with either state have a high ventral flange that abruptly
stops near the quadrate.
FIG. 12.—Palatal morphology of Pelturagonia anolophium (A, MZB
14992) and Phoxophrys tuberculata (UTA 65254) illustrating anterior extent
of the palatal ridge relative to the palatine–pterygoid suture. Skulls shown in
ventral aspect with mandible and hyoid apparatus removed.
90 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
The pterygoids of Phoxophrys anolophium,Ph. borneen-
sis,Ph. cephalum, and Ph. spiniceps lack the flange or have a
low ridge that ends in an inconspicuous, shallow dip near the
quadrate (Fig. 13). In a broader analysis of agamids, the new
character defined here could be treated as a third character
state for Character 14 of Stilson et al. (2017). Among other
Draconinae, the pterygoids of only Lophocalotes and
Ptycholaemus lack the ventral flange. Pterygoids of Phox-
ophrys nigrilabris and Ph. tuberculata have well-developed
ventral flanges, but lack terminal hooks.
Palatoquadrate Derivatives
Shape of quadrate.—Ventrally, the quadrate of Phox-
ophrys is relatively wider than that of other draconines (F
1,33
¼4.46, P¼0.043), treating height of the quadrate as a
covariate. Nonetheless, relative widths overlap substantially,
especially for smaller species. Treating quadrate height as a
covariate, we did not find a difference in width of the conch
or medial lamina between Phoxophrys and other draconines.
Moreover, we did not find substantial intrageneric differ-
ences in width of these structures. Though divergent in other
aspects of quadrate morphology, our specimen of Ph.
tuberculata actually fell on the regression line for width of
the conch versus quadrate length and was very close to the
line for width of the medial lamina versus quadrate length.
28. Lateral edge of conch of quadrate.—The conch is
ovoid and its antero-lateral edge is outwardly convex, lacking
a wide dorsolateral triangular portion (0), or the conch is
rudimentary (¼subtriangular, not too weakly concave, its
antero-lateral edge straight) with a wide dorsolateral
triangular portion (1).
Quadrates of the Bornean Phoxophrys are relatively
similar to one another and differ in three ways from the
quadrate of Ph. tuberculata. (1) The conchs of Bornean
Phoxophrys are subtriangular and weakly to not concave with
straight antero-lateral borders, sloping obliquely to the
lateral condyle of the quadrate. In contrast, Ph. tuberculata
has a concave conch with an outwardly bowed border (Fig.
9). (2) All members of the genus have a thickened tab of
bone at the dorsolateral border of the conch for articulation
with the squamosal. The tab is large and subtriangular in
Bornean Phoxophrys, and less pronounced and positioned
slightly lower along the lateral border of the conch in Ph.
tuberculata. (3) On the medial face of the quadrate, a
shallow facet for the pterygoid lies just posterior to the
medial lamina and the posterior tip of the pterygoid
articulates with the medial lamina and column of the
quadrate. Above this articulation, the medial lamina of
Bornean Phoxophrys is flat and subrectangular. In postero-
medial aspect, the medial lamina merges seamlessly with the
column of the quadrate. However, in anterior aspect, a
rectangular patch of conspicuous and thick bony reinforce-
ment demarcates the merger of the medial lamina and
column in each of the Bornean species. Phoxophrys tuber-
culata lacks this bony reinforcement and the medial lamina
of Ph. tuberculata is triangular and weakly concave postero-
medially. This species has a rounded thickening of the
medial lamina about one-third of its length down from the
top of the quadrate, but based on its position and shape, we
do not consider this thickening to be homologous with the
rectangular reinforcement of Bornean species.
We have no reason to believe that the three differences
are functionally linked and note that they do not always co-
occur in outgroups. Nonetheless, we choose to err on the
side of caution and code only one of these traits for
phylogenetic analysis.
Moody (1980) coded Phoxophrys as having a rudimentary
conch (his Character 24.2), describing this character state as
FIG. 13.—Medial aspect of mandible and pterygoid of Pelturagonia
anolophium (A, MZB 14992), Pe. borneensis (FMNH 71856), and
Phoxophrys tuberculata (UTA 65254).
91
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
‘‘conch rudimentary, only a small flat flange or absent’’
(Moody 1980:100). He also recognized an intermediate
condition [24.1] described as ‘‘concave conch present but
lateral margin straight and without a beaded edge.’’ In his
matrix, Moody reported rudimentary conchs only for
Phoxophrys and the Sri Lankan endemics Ceratophora,
Cophotis, and Lyriocephalus. We confirm his observations
for the Sri Lankan endemics and note that the conch is even
‘‘more rudimentary’’ in these genera than in Bornean
Phoxophrys. Nonetheless, the Sri Lankan genera have
triangular medial laminae, shaped more like the lamina of
Ph. tuberculata than the laminae of the Bornean species. The
conchs of the female specimen of Ph. cephalum and holotype
of Ph. anolophium are slightly concave, though not as
concave as various species coded as 24.1 by Moody.
Among outgroups, the quadrate only Diploderma splen-
didum has all three characters found in Bornean Phox-
ophrys. Draco sumatranus has a rudimentary conch with a
wide triangular portion dorsally. Though rectangular and
relatively narrow, the medial lamina of this species lacks
rectangular bony reinforcement on its anterior face.
Mandible
Mensural characters of the mandible.—Treating
length of the mandible as a covariate, Phoxophrys has a
shorter precoronoid length (F
1,32
¼10.01, P
Bonferroni
¼
0.009) and longer retroarticular process (F
1.32
¼29.19,
P
Bonferroni
¼0.000*) than other draconines. Treating height
of the coronoid as a covariate, the species of Phoxophrys
have shorter dorsal processes of the coronoid than other
draconines (F
1,32
¼8.55, P¼0.006). Nonetheless, in
bivariate plots we found considerable overlap for each
character and do not consider any of the three traits to be
diagnostic of the genus or of the Bornean group.
29. Angular process of mandible.—The angular process is
absent (A), developed horizontally (B), or developed
vertically in a transverse plane (1).
The Bornean Phoxophrys have large tablike angular
processes, with high vertical development in a transverse
plane. In these species, the anterior face of the process is
concave. In sharp contrast, Ph. tuberculata lacks an angular
process (Fig. 13). Neither of these character states is
common among other draconines. Among the species we
examined, only the four species of Dendragama entirely lack
angular processes, whereas only Aphaniotis acutirostris,
Pseudocalotes kingdonwardi,andLyriocephalus scutatus
have transverse, vertical processes. The remaining species
have horizontal processes that widen anteriorly. In Cerato-
phora and Lophocalotes, the angular processes have some
vertical development anteromedially, but in a plane slightly
oblique to longitudinal. Our interpretation of this character
differs from that of Moody (1980), who assigned his State 2
(¼process ‘‘tuberculate shaped or absent’’ )toAphaniotis,
Bronchocoela,Ceratophora,Cophotis,Lophocalotes, and
Sitana. Species of these genera examined by us have
horizontal processes; however, development of the angular
process appears to exhibit sexual dimorphism in some
species. The process is horizontal and well developed in
our male specimens of Ceratophora stoddartii,Lophocalotes
achlios, and L. ludekingi, but reduced to a rounded mound
in the female of C. stoddartii and a low thickening of the
prearticular in the female of L. achlios.
SYSTEMATICS
Phylogeny
Our Bayesian and maximum-likelihood analyses of
mitochondrial DNA sequences recovered trees with near-
identical topologies and with strong nodal support (Fig. 14).
The maximum-likelihood and Bayesian trees differed only in
placement of Calotes versicolor. With a posterior probability
of 100%, the Bayesian analysis recovered C. versicolor as
sister to all other draconines except for Draco and Bornean
Phoxophrys, whereas the maximum-likelihood analysis
placed C. versicolor above mainland Pseudocalotes, albeit
with a bootstrap value of 65. Figure 13 does not show
intergeneric relationships among the various species of
Dendragama,Lophocalotes,andPseudocalotes. With the
exception of the three species of Dendragama, Harvey et al.
(2017b, 2018) published these intergeneric relationships.
Interested readers may also consult our supplementary
material to view the unedited maximum-likelihood and
Bayesian trees used to construct the simplified phylogeny
in Fig. 13.
Molecular data did not support monophyly of Phoxophrys
as defined by Inger (1960). Although we lack DNA
sequences for half of the species of Phoxophrys, we found
Ph. tuberculata to be more closely related to Dendragama,
Lophocalotes, and insular Pseudocalotes than to the Bornean
species Ph. anolophium and Ph. nigrilabris (Fig. 14). The
two Bornean species form a clade that is sister to all other
draconines except Draco in our analysis.
Phylogenetic analysis of our morphological characters
found a single tree (Fig. 14) with strong support for
Phoxophrys tuberculata, a clade containing the five Bornean
species, Ph. nigrilabris, and a clade containing the four large
Bornean species. Our survey of morphology found only one
potential synapomorphy of Phoxophrys sensu lato: narrowing
of the alveolar portion of the premaxilla (24.1), a trait that
also occurs in Acanthosaura. Although Ph. tuberculata and
the Bornean species have narrow premaxillae, the Bornean
species brace the premaxilla with premaxillary processes of
the mandible that fuse or suture over the premaxilla dorsally,
whereas the premaxilla abuts short premaxillary processes in
Ph. tuberculata.
This single putative synapomorphy of Phoxophrys tuber-
culata þthe Bornean species stands in sharp contrast to the
numerous differences between these lineages. Autapomor-
phies of Ph. tuberculata are a supercilium projecting upward
and outward as a pointed, triangular flap (Character 3.1),
small size with SVLs not exceeding 43 mm (11.1), transverse
shorting of the prefrontal allowing nasal–frontal contact
(13.1), and the squamosal extending anteriorly below the
jugal to curve ventrally and brace against the vertical
posterior margin of the jugal (15.1). Unlike the Bornean
species, Ph. tuberculata lacks characters 4.1, 8.1, 25.1, 28.1,
and 29.1. Unlike Ph. tuberculata, the Bornean species share
a pale ischial stripe, nasals contacting one or no supralabials
rather than broadly contacting two supralabials, multiple
high scale counts with ranges not overlapping those of Ph.
tuberculata, rectangular bony reinforcement on the anterior
face of the quadrate, and a rectangular medial lamina of the
quadrate. Finally, evolution of an antero-laterally angled
postorbital process of the frontal (17.B, found only in Ph.
tuberculata,Dendragama, and Lophocalotes) and loss of the
92 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
angular process of the mandible (29.A, found only in Ph.
tuberculata and Dendragama) are consistent with the
molecular phylogeny, since the Bornean species lack these
characters.
Diagnostic characters proposed by Hubrecht (1881) and
Inger (1960:221) are not synapomorphies of Phoxophrys þ
the Bornean species, because they do not occur in both
clades or they are either plesiomorphic or have equivocal
FIG. 14.—(A) Phylogeny of Phoxophrys and Pelturagonia based on morphology. Apomorphies identified by the TNT program are mapped on branches in
bold font. Numbers in regular font immediately above and behind each node are bootstrap/Bremer support values. (B) Intergeneric relationships of
Draconinae recovered by maximum-likelihood analysis of 16S and ND4 genes. Numbers below branches are bootstrap support values above percent
posterior probabilities of clades in the Bayesian tree. (Supplementary data for this publication includes intrageneric topologies in both the maximum-
likelihood and Bayesian trees.)
93
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
polarity in the ancestor to Phoxophrys. Inger’s (1960) study
provided good reasons for removing Bornean Phoxophrys
from Japalura, but did not make a strong case for placing the
Bornean species in Phoxophrys.Withonlyweakand
contradictory evidence that the Bornean species share
common ancestry with Ph. tuberculata, we remove Peltur-
agonia Mocquard from the synonymy of Phoxophrys
Hubrecht.
Synapomorphies of Pelturagonia are vertical division of
the rostral into two or more scales (4.1), presence of
dorsolateral caudal crests (8.1), premaxillary processes of the
mandible suturing or fusing with one another (25.1), a
rudimentary conch of the quadrate with a wide dorsolateral
triangular portion (28.1), and a vertical angular process of
the mandible (29.1). Synapomorphies of the clade of large
Pelturagonia are loss of a continuous dorsal crest on the body
(6.1), a bony ridge extending from the postciliary ornament
of the postorbital bone to the ventral margin of the
supratemporal fossa (18.1), palatal ridges ending behind
the palatine–pterygoid suture (26.1), and loss of the ventral
flange from the posterior end of the pterygoid (27.B).
Relationships among the four large Bornean species
require further study. Based on available data, Pelturagonia
anolophium is the sister species of Pe. spiniceps. Synapo-
morphies of this clade are enlarged dorsally, projecting
scales between the caudal crests (Character 9.B) and the
postorbital process of the frontal reaching the postciliary
ornament of the postorbital bone (16.B). Evolution of lateral
flaring of the crista prootica behind the crista alaris (23.1)
supports a sister relationship between Pe. cephalum and the
Pe. anolophium þPe. spiniceps clade; however, this group is
only weakly supported by resampling values.
Special similarity between Phoxophrys tuberculata and
Pelturagonia nigrilabris contradicts our phylogenetic hy-
potheses. In addition to both being relatively small species,
they share two derived characters not found in the large
species of Pelturagonia: arched supraciliaries producing a
serrate edge to the supercilium (1.1) and a bony ridge that
curves ventrad from the postciliary ornament to extend
vertically down the postorbital bone (18.2). Pelturagonia
nigrilabris is the most distinctive member of Pelturagonia
and has evolved two to four times as many apomorphic
characters as its congeners (Fig. 13).
REVISED TAXONOMY AND DESCRIPTION OF NEW SPECIES
Phoxophrys Hubrecht
Phoxophrys Hubrecht 1881:51. Type species Phoxophrys
tuberculata Hubrecht, by monotypy.
Diagnosis.—Small draconine agamids reaching a SVL of
at least 43 mm and distinguished from all other Agamidae by
the following combination of characters: (1) tympanum and
external auditory meatus absent, extrastapes attaching to
skin; (2) head robust, 77.4–92.1% as wide as long, comprising
27.4–28.5% of SVL; (3) rostral single; (4) prominent
tubercular scale positioned below corner of jaw; (5) super-
cilium projecting upward and outward as pointed, triangular
flap; (6) dorsal scales heterogenous with larger tubercular
scales interspersed among smaller keeled scales, 56–67
scales around midbody; (7) premaxillary processes of
mandible abutting premaxilla; (8) frontal and parietal
extensively sculptured, not containing pineal foramen; (9)
transverse shorting of prefrontal bone allowing nasal–frontal
contact; (10) squamosal extending anteriorly below jugal to
curve ventrally and brace against vertical posterior margin of
jugal; (11) angular process of mandible absent.
Content.Phoxophrys tuberculata Hubrecht (1881).
Etymology.—Hubrecht (1881) did not discuss the
derivation of the name Phoxophrys. This feminine Latin
noun in the nominative singular contains the Greek prefix
phoxos meaning pointed or peaked and noun ophrys
meaning eyebrow. The name likely refers to the arched,
triangular supercilium of Ph. tuberculata and inspired
coining of the common English name ‘‘ eyebrow lizard,’’ as
frequently used by various websites and in some recent
treatments of the Agamidae (Murphy and Hanken 2018).
Distribution.Phoxophrys occurs only on Sumatra in
Jambi, Sumatera Barat, and Sumatera Utara Provinces.
Remarks.—Although we currently consider Phoxophrys
to be monotypic, differences among our male specimen and
the two female types give us pause. The male has a vertebral
crest, whereas females only have widely separated vertebral
tubercles. The male was collected in lowlands near Medan
on the eastern side of Sumatra’s Bukit Barisan Range,
whereas the female types were collected west of this range.
Nonetheless, male specimens from west of the range have
continuous vertebral crests, as illustrated by a specimen
obtained near Solok and illustrated by Manthey (2010).
Species boundaries should be investigated within Phox-
ophrys once series of both sexes can be obtained from both
the eastern and western populations.
Pelturagonia Mocquard
Pelturagonia Mocquard 1890:130. Type species Pelturagonia
cephalum Mocquard, by monotypy.
Diagnosis.—Medium-sized draconine agamids reaching
a SVL of at least 84 mm and distinguished from all other
Agamidae by the following combination of characters: (1)
tympanum and external auditory meatus absent, extrastapes
attaching to skin; (2) head robust, 77.5–98.3% as wide as
long, comprising 22.5–28.4% of SVL; (3) rostral vertically
divided into two or more scales; (4) prominent tubercular
scale positioned below corner of jaw; (5) pair of dorsolateral
caudal crests, not separated medially by vertebral crest; (6)
dorsal scales heterogenous with larger tubercular scales
interspersed among smaller-keeled scales, 81–115 scales
around midbody; (7) alveolar portion of premaxilla relatively
narrow, usually with single tooth and with premaxillary
processes of mandible suturing or fusing over top of it; (8)
frontal and parietal extensively sculptured, not containing
pineal foramen; (9) conch of quadrate rudimentary, with
wide dorsolateral triangular portion; (10) angular process of
mandible large, tab-like, and vertically developed.
Content.Pelturagonia anolophium Harvey et al. (this
study), Pelturagonia borneensis (Inger 1960) comb. nov.,
Pelturagonia cephalum Mocquard (1890), Pelturagonia
nigrilabris (Peters 1864) comb. nov., Pelturagonia spiniceps
(Smith 1925) comb. nov.
Etymology.—Explaining the derivation of his new name,
Mocquard (1890:130) described the distinctive angled aspect
of the caudal crests and remarked ‘‘ pour cette raison, nous
imposons le nom g´
en´
erique de Pelturagonia.’’ Pelturagonia
94 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
is a feminine Latin noun in the nominative singular derived
from the feminine Greek words pelte meaning shield and
gonia meaning angle.
Distribution.—The genus occurs throughout Borneo and
in the Natuna Achipelago (Fig. 15; Das 2010; Manthey
2010).
Remarks.—When Mocquard (1890) erected Pelturagonia
for Pe. cephalum, he did not comment on Pe. nigrilabris, the
only other member of the genus described at that time.
Peters (1864) assigned Pe. nigrilabris to subgenus Japalura
within Otocryptis as Otocryptis (Japalura) nigrilabris.
Boulenger (1885) retained nigrilabris in Japalura, which
he recognized as a genus. Later (1891), Boulenger referred
Pelturagonia cephalum to the synonymy of Japalura
nigrilabris. De Rooij (1915) accepted Boulenger’s synonymy,
and her description of Japalura nigrilabris is actually based
on Pe. cephalum as previously pointed out by Inger (1960).
Smith (1925) thought Pe. spiniceps was more closely related
to Phoxophrys tuberculata and referred this species to
Phoxophrys. Finally, Inger (1960) revalidated Pe. cephalum
and transferred this species and Pe. nigrilabris to Phox-
ophrys.
Pelturagonia anolophium sp. nov.
Holotype.—An adult female (MZB 14992, Field tag
21838) collected by Thorton R. Larson and M. Munir on
Gunung Lumut, Paser, Kalimantan Timur, Indonesia, 832 m
elevation, 1.410788S, 115.977718E at 21:01 h. ND4 GenBank
accession number MN548383; 16S GenBank accession
number MN537800.
Paratypes (3).—Two subadult females (UTA 65255,
65256) and one subadult male (MZB 14993), collected by
T.R. Larson and M. Munir from 19:10 to 19:34 h at the type
locality, 1,090 m elevation, 1.402748S, 115.985918E. ND4
GenBank accession number MN548384 for UTA 65255 and
MN548385 for UTA 65256; 16S GenBank accession number
MN537801 for UTA 65255 and MN537802 for UTA 65256.
Diagnosis.—A large species of Pelturagonia reaching at
least 193 mm (80 mm SVL) in length, distinguished from all
congeners by the following combination of characters: (1)
FIG. 15.—Skull of Pelturagonia anolophium (holotype, MZB 14992, skull width 15.9 mm) in dorsal, lateral, ventral, and posterior views.
95
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
supracilium lacking long spinose scale; (2) tubercular
sublabials 4–7 anterior to enlarged subrictal tubercle; (3)
enlarged middorsal scales widely spaced, not forming
continuous dorsal crest; (4) nuchal crest consisting of 10–
12 triangular scales mostly in continuous series; (5)
paravertebrals on neck and scales of pectoral gap pointing
backward; (6) base of tail with dorsolateral crest of 7–10
enlarged scales; (7) area between caudal dorsolateral crests
not flat, with enlarged projecting scales; (8) four rows of
enlarged subcaudals near base of tail (poorly differentiated
in females); (9) gulars sharply keeled and mucronate, 31–36;
(10) scales around midbody 92–109; (11) dorsum green with
contrasting pale, saffron tubercles and reddish brown to
brick red bands (at least in females and juvenile males; adult
male unknown); (12) gular region without black pigment or
with black reticulation but lacking sharp black lines
extending postero-medially from labials.
Comparisons.Pelturagonia anolophium is most likely
to be confused with Pe. borneensis and Pe. cephalum
(characters in parentheses). Unlike these species, Pe.
anolophium attains a larger size (80 mm SVL in females
compared to 62–74 mm in the other two species) and has
enlarged projecting, tubercular scales dorsal to the caudal
crests (area between crests flat with small keeled scales), 10–
12 nuchal crest scales (4–8), nuchal crest mostly not
interrupted by much smaller scales (2–5 small scales
interrupting crests of Pe. borneensis and 3–13 small scales
interrupting crests of Pe. cephalum), 2–3 enlarged post-
temporal modified scales (single in all specimens except one
Pe. cephalum, FMNH 152165), and the postorbital process
of the frontal reaching the postciliary ornament (process and
ornament separated by a large gap). All three species have
strongly heterogeneous dorsal scales, but Pe. anolophium has
scales that are distinctly more spinose. Additionally, unlike
Pe. borneensis,Pe. anolophium has a continuous row of small
lorilabials between the infraorbitals and supralabials (usually
one or more large infraorbitals usually contacting supra-
labials) and lateral flaring of the crista prootica behind the
crista alaris (crista prootica straight behind crista alaris); Pe.
anolophium lacks oblique lines across the throat (prominent
black to charcoal lines extend posteriorly and medially).
Unlike Pe. cephalum,Pe. anolophium has spinose gulars
(bluntly keeled), 4–7 tubercular sublabial scales anterior the
subrictal tubercle (0 tubercular sublabials anterior to the
subrictal tubercle), and four rows of enlarged subcaudals
(two).
External morphology.—Characters of the female holo-
type appear in brackets after ranges of all the females.
Females reaching 193 mm (SVL 80 mm) in length; two
female paratypes 102 and 111 mm (SVL 63 and 66 mm); one
male paratype 108 mm (SVL 65 mm); SVL 39.7% and tail
length 60.3% of total length in male; SVL 38.4–41.3%
[41.3%] and tail length 58.7–61.6% [58.7%] of total length in
females; tail length 152 times as long as SVL in males and
142–161 [142] times as long as SVL in females; distance from
axilla to groin accounting for 46.2–48.9% [48.9%] of SVL;
distance between axillae 18.0–19.3% [19.3%] of SVL; head
robust, 85.9–95.1% [95.1%] as wide as long, its length
accounting for 22.5–27.3% [22.5%] of SVL; snout round in
dorsal view, subacuminate in profile, sloping upward at
about 808to horizontal (Fig. 16); dorsal head scales weakly
imbricate, keeled; rostrals three (100%); small medial rostral
squarish but somewhat wider than tall [21.1% as wide as
internarial distance], wider than mental; postrostrals 5–7 [6];
lateral postrostral contacting nasal (100%); noticeably
enlarged azygous scales on snout arranged as inverted Y;
each arm of Y consisting of one or two [2/2] large scales
contacting base; arms of Y separated from orbital margin by
one small scale with heavy keels directed postero-laterally;
base of Y consisting of three or four [4] enlarged azygous
scales in medial row with keels oriented longitudinally;
slightly depressed prefrontal, frontal, and parietal regions
bound by arms of Y, medial borders of orbit, and transverse
parietal ridge; scales between orbits similar in size to
supraoculars, distinctly smaller than circumorbitals, 5–7 [7]
scales minimally separating circumorbital series; low, obtuse
V-shaped parietal ridge covered by 7–9 [8] somewhat
elongate parietals with low keels; discrete interparietal
absent; no visible parietal eye; area in front of nuchal crest
depressed, flanked on either side by elevated, rounded
supratemporal areas.
One (100%) [1/1] supranasal scale separating postrostral
series from first canthal; circumorbitals enlarged, angulate,
12 or 13 [13/13] from last canthal to postciliary scale;
supraoculars somewhat heterogenous in size (¼1–4 supra-
oculars in center of posterior half of supraocular region
larger than rest), smaller than circumorbitals; transorbitals
17–22 [19]; canthals four or five [4/4]; supraciliaries 7–10
[10/10], resembling canthals and first best identified by
probing orbital margin; supraciliaries subrectangular and not
projecting, imbricating slightly (almost juxtaposed); com-
bined count of canthals and supraciliaries 12–14 [14/14];
supraciliary notch prominent, containing three or four [3/3]
small scales between last supraciliary and postciliary scale; 16
or 17 [17/17] scales between nasal and postciliary scale;
postciliary scale prominent, enlarged, swollen and subpyr-
amidal, abutting slightly smaller though still noticeably
enlarged scale of temporal region; scalation of temporal
region heterogenous, consisting of large angulate scales with
heavier keels separated by small angulate keeled scales; 2–4
[4/3] noticeably enlarged scales between postciliary and
posttemporal modified scales, with 2–4 scales [3/3] separat-
ing last enlarged temporal from largest posttemporal
modified scale; posttemporal modified scales two or three
[2/3] positioned on raised skin overlying posterior margin of
skull; lower posttemporal modified scale as large as largest
scales of nuchal crest, spinose and projecting dorsolaterally,
surrounded by 7–10 [10/9] scales (including an enlarged
spinose to pyramidal posttemporal dorsal to it, but somewhat
smaller).
Nasal oval to trapezoidal, its dorsal border contributing to
canthus; nasal contacting lateral rostral and first supralabial;
nostril oval, directed laterally, occupying about one-fourth to
one-third of nasal; scales of loreal region smooth to feebly
keeled, 6–9 [7/6] scales in vertical row between last canthal
and supralabials, 4–6 [4/4] scales in horizontal row from
anterior border of orbit to nasal; orbit 45.8–50.0% [45.8%] of
head length; palpebrals mostly granular; anterior and
posterior scales of first row of palpebrals above eye evidently
fused to second; heavily keeled palpebrals in center of eye
not fused to first row; 12–15 [14/15] palpebrals between
ocular angles; row of enlarged suboculars separated from
supralabials by continuous row of lorilabials; distinctly
enlarged scale positioned above first elongate scale of rictal
96 Herpetological Monographs 33, 2019
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
fold (except in UTA 65255) and separated from it by small
lorilabial; when enlarged scale present, scale in front of it
divided longitudinally 100% of time; 10–12 [11/10] enlarged
infraorbital scales from nasal to posterior border of orbit;
tympanum and external auditory meatus absent; however,
auditory region recognizable as subcircular, depressed area;
4–6 [5/6] scales comprising orbito-tympanic series, three or
four [3/4] of them enlarged and somewhat rugose; distance
from anterior border of auditory region to orbit 29.8–34.8%
[34.8%] of head length; scales below and behind orbito-
tympanic series heterogenous in size with numerous
tubercular scales scattered among much smaller keeled to
spinose scales.
Supralabials smooth, eight or nine [8/9]; rictal fold short,
bordered dorsally by two or three [2] elongate scales; first
scale bordering rictal fold longer than supralabial abutting it;
infralabials smooth to feebly keeled, 11 or 12 [12/11]; mental
entire, pentagonal; first pair of chin shields in medial contact
behind mental and first 1–3 [2/3] contacting infralabials;
sublabial tubercular scales 5–8 [5/6]; last sublabial tubercular
scale much larger than rest, subpyrimidal and positioned
below rictus; large spinose tubercular scales separated from
one another by much smaller spinose scales in area behind
rictus, below auditory region, and extending onto postero-
lateral throat; gulars heavily keeled and terminating in
downward-pointing mucrons, 31–36 [31] from point of
medial contact between first pair of chin shields to preaxial
margin of arm; 8–12 [9] enlarged and more pointed gular
scales down midline in front of gular pouch, but these scales
not forming obvious crest; transverse gular fold incomplete
medially.
Nuchal crest prominent, consisting of 10–12 [11]
triangular scales contacting one another except for single
instance of paravertebrals in medial contact in MZB 14993;
nuchal crest beginning immediately behind last occipital
scale (50%, n¼4) or separated from it by single flat scale
(50%) [in contact]; pectoral gap of 11–16 [12] small-keeled
scales separating nuchal crest from transverse band of three
or four [3] enlarged, heavily keeled and projecting scales;
dorsal crest of widely spaced, single, enlarged, heavily
keeled, projecting scales or short transverse bands of 2–4
similarly enlarged scales; 10–17 [17] enlarged scales from
pectoral gap to posterior border of leg (count including
single enlarged vertebrals, pairs of enlarged paravertebrals,
and short transverse bands of enlarged vertebrals and
paravertebrals), 50–62 [50] scales total in dorsal crest (count
includes all scales along vertebral midline of dorsal crest,
because we could not confidently distinguish between
paravertebrals in medial contact and small vertebrals); total
number of scales along vertebral line from posterior border
of leg to occiput 73–90 [73].
Imbrication pattern on neck complex: wide band of
paravertebrals pointing upward and backward anteriorly
then reorienting to point backward by level of pectoral gap;
FIG. 16.—Distribution of Pelturagonia and Phoxophrys (areas above 500 m shaded gray).
97
HARVEY ET AL.—REVISION OF PHOXOPHRYS
Downloaded From: https://bioone.org/journals/Herpetological-Monographs on 16 Mar 2020
Terms of Use: https://bioone.org/terms-of-useAccess provided by University of Texas at Arlington
scales on lateral surface of neck pointing forward above
antehumeral fold, then reorienting to point upward and
backward at level of scapula; scales also reorienting below
antehumeral fold to point downward and backward;
paravertebrals on neck not enlarged relative to dorsals below
them.
Dorsals much smaller than ventrals, weakly imbricate,
strongly heterogenous with enlarged, heavily keeled spinose
to pyramidal tubercles interspersed among smaller scales;
4–6 [5/5] irregular oblique lines one tubercular scale wide,
directed anteriorly and ventrally on flanks; tubercles in
oblique lines mostly separated from one another; smaller
dorsals mostly with upturned tips and usually projecting
somewhat but not as strongly as dorsal tubercles; dorsal
tubercles extending onto ventro-lateral regions from
posterior throat to groin; wide band of paravertebral scales
pointing backward; sharp transition from backward-point-
ing paravertebrals to dorsals that point upward and
backward below dorsolateral series of 11–14 [13/14] large
tubercular scales (the abrupt reorientation is particularly
evident in a few places where a backward pointing
dorsolateral tubercle abuts a dorsal flank tubercle); first
scale of dorsolateral series separated from first scale of
dorsal crest by 6–8 [6] small paravertebrals; first scale of
dorsolateral series spinose, larger than other enlarged
scales of series and somewhat larger than other tubercular
scales on flanks, directed laterally, whereas other tubercular
scales of series directed posteriorly; scales around midbody
92–109 (100 68, n¼4) [109]; four enlarged and heavily
keeled scales in vertical line over ilium (uppermost
positioned just above epiphyseal tuberosity), upper two of
these as large as femoral armor and much larger than scales
on flank, lower two of these about as large as tubercular
scales on flank; ventrals imbricate, heavily keeled, 51–58
(56 63, n¼4) [51] from preaxial edge of arm to vent, in
21–23 [22] longitudinal rows at midbody.
Base of tail rounded in cross-section; dorsolateral crest of
7–10 (8 61, n¼8 sides of four specimens) [10/8] enlarged,
heavily keeled scales; dorsally, scales between crests
somewhat heterogenous in size, heavily keeled, projecting;
minimally two scales between crests (100%, n¼4),
maximally six scales between crests (100%, n¼4) dorsally;
below caudal crests, scales subequal, keeled; scales above
cloaca and immediately behind thigh small except for two or
three (both sides with two except for three on right side only
of MZB 14993) [2/2] distinctly enlarged scales and more
heavily keeled scales.
Scales on limbs much larger than those on body,
imbricate, heavily keeled, and pointed distally; in center of
antebrachium, two or three noticeably much larger scales
with high keels and pointed tips on postaxial edge; scales of
palm keeled, imbricate, mucronate, often with three
mucrons (largest mucron at end of central keel and a single
small mucron on either side of it); Finger IV 71.4–73.7%
(72.9 61.1, n¼4) [71.4%] as long as Toe IV, slightly shorter
than Finger III (length of Finger III 85.7–90.0%, 87.5 62.2,
n¼4 [90.0%] of length of Finger IV); lamellae under fingers
bicarinate (some multicarinate or divided at base), 18 or 19
[18/18] under Finger IV; scales on dorsal surfaces of legs
heterogeneous in size; 4–6 [5/5] enlarged scales postaxial to
dorsal midline with noticeably higher, keels with prominent
points directed more postaxial than adjacent scales; shank
21.4–24.3% (22.7 61.1, n¼4) [22.5%] as long as SVL;
scales on dorsal surface of shank somewhat heterogeneous:
enlarged scales with higher keels in center of shank; scales of
sole keeled with mucron directed distally and downward;
Toe V relatively short (1–4 [1/2] fourth-toe lamellae within
span of fifth toe), 53.7–61.1% (58.0 63.1, n¼4) [58.6%] as
long as Toe IV; 18–24 (21 62, n¼4) [18/20] lamellae under
Toe IV; subdigital lamellae of toes bicarinate, including those
at base of Toe III; lamellae unicarinate under proximal
phalanx of Toe IV (100%, n¼4).
Bristled sense organs (Scortecci 1937; Ananjeva et al.
1991) widespread on dorsal and ventral surfaces; single
bristled sense organs usually positioned just below apex of
keels and projecting beyond apex; callous glands absent;
femoral and precloacal pores absent.
Coloration.—We base our description on photos of the
type series including photos of live paratypes and the
holotype immediately after being euthanized. In life,
Pelturagonia anolophium is green with contrasting pale
saffron tubercles and reddish brown to brick red markings.
In dorsal aspect, five (UTA 65255, 65256) or six (MZB
14992, 14993) wide W-shaped bands cross the body, the
most prominent overlapping the scapulae. These bands
narrow on the flanks and have somewhat paler centers.
The paratypes had green pigment in interspaces verte-
brally; however, the entire area above the dorsolateral
series of tubercles was brown (dark brown bands, lighter
brown