Smithsonian Institution
Scholarly Press
smithsonian contributions to zoology • number 631
Speciation
and
Dispersal
in
A Chronology of
a Low Diversity Taxon:
Middle
Missouri
Plains
The Slender Geckos
Village
Sites
Hemiphyllodactylus
(Reptilia, Gekkonidae)
By Craig M. Johnson
with contributions by
George R. Zug
Stanley A. Ahler, Herbert Haas, and Georges Bonani
SERIES PUBLICATIONS OF THE SMITHSONIAN INSTITUTION
Emphasis upon publication as a means of “diffusing knowledge” was
expressed by the irst Secretary of the Smithsonian. In his formal plan
for the Institution, Joseph Henry outlined a program that included
the following statement: “It is proposed to publish a series of reports,
giving an account of the new discoveries in science, and of the changes
made from year to year in all branches of knowledge.” This theme of
basic research has been adhered to through the years by thousands of
titles issued in series publications under the Smithsonian imprint, commencing with Smithsonian Contributions to Knowledge in 1848 and
continuing with the following active series:
Smithsonian Contributions to Anthropology
Smithsonian Contributions to Botany
Smithsonian Contributions to History and Technology
Smithsonian Contributions to the Marine Sciences
Smithsonian Contributions to Museum Conservation
Smithsonian Contributions to Paleobiology
Smithsonian Contributions to Zoology
In these series, the Institution publishes small papers and full-scale
monographs that report on the research and collections of its various
museums and bureaus. The Smithsonian Contributions Series are
distributed via mailing lists to libraries, universities, and similar institutions throughout the world.
Manuscripts submitted for series publication are received by the
Smithsonian Institution Scholarly Press from authors with direct afiliation with the various Smithsonian museums or bureaus and are subject
to peer review and review for compliance with manuscript preparation
guidelines. General requirements for manuscript preparation are on the
inside back cover of printed volumes. For detailed submissions requirements and to review the “Manuscript Preparation and Style Guide for
Authors,” visit the Submissions page at www.scholarlypress.si.edu.
smithsonian contributions to zoology • number 631
Speciation and Dispersal in
a Low Diversity Taxon:
The Slender Geckos
Hemiphyllodactylus
(Reptilia, Gekkonidae)
George R. Zug
washington D.C.
2010
ABSTRACT
Zug, George R. Speciation and Dispersal in a Low Diversity Taxon: The Slender Geckos Hemiphyllodactylus
(Reptilia, Gekkonidae). Smithsonian Contributions to Zoology, number 631, xi + 70 pages, 25 igures, 7
tables, 2010.—Hemiphyllodactylus is a genus of small geckos occurring widely, although uncommonly seen,
throughout the Indo-Paciic islands and South Asia. These geckos consist of both bisexual and unisexual species. The unisexual species, Hemiphyllodactylus typus, the most widespread of these geckos, apparently attained its Polynesian to Mascarene distribution (invasion) through accidental human transport. The bisexual
species have much smaller distributions, geographically restricted to island groups or limited continental
areas. Until the early 1990s, most bisexual populations were considered subspecies of H. typus. In the last
two decades, herpetologists have regularly used species epithets proposed for the region under their investigation. This resurrection of species names has occurred largely without explanation or taxonomic study.
This study examines the morphology of Hemiphyllodactylus throughout its known range, using 13 regional
samples, irst examining the differentiation of unisexual and bisexual populations and individuals, then the
possibility of regional differentiation among the different bisexual populations. Variation and consistency in
morphology in and among the regional sample identify the existence of a wide-ranging unisexual species,
H. typus, and at least eight geographically restricted bisexual species. Available museum specimens for some
regions are adequate to characterize eight bisexual species, H. aurantiacus, H. ganoklonis n. sp., H. harterti,
H. insularis, H. larutensis, H. margarethae, H. titiwangsaensis n. sp., and H. yunnanensis. Potentially unique
bisexual populations occur in Hong Kong, southern Indochina, Borneo, and Sri Lanka, but samples are too
small to adequately characterize these populations. The origins and evolution of the species are examined,
and the study concludes with a taxonomy for the identiied species.
Cover image: Palauan slender gecko Hemiphyllodactylus ganoklonis. (Drawing by Molly Dwyer Grifin.)
Published by SMITHSONIAN INSTITUTION SCHOLARLY PRESS
P.O. Box 37012, MRC 957
Washington, D.C. 20013-7012
www.scholarlypress.si.edu
Text and images in this publication may be protected by copyright and other restrictions or owned by individuals and entities other than, and in addition to, the Smithsonian Institution. Fair use of copyrighted material includes the use of protected materials for personal, educational, or noncommercial purposes. Users must
cite author and source of content, must not alter or modify content, and must comply with all other terms
or restrictions that may be applicable. Users are responsible for securing permission from a rights holder for
any other use.
Library of Congress Cataloging-in-Publication Data
Zug, George R., 1938–
Speciation and dispersal in a low diversity taxon : the slender geckos Hemiphyllodactylus
(Reptilia:Gekkonidae) / George R. Zug.
p. cm. — (Smithsonian contributions to zoology ; no. 631)
Includes bibliographical references and index.
1. Hemiphyllodactylus. I. Title.
QL666.L245Z84 2010
597.95'2—dc22
2010042310
ISSN: 0081-0282 (print); 1943-6696 (online)
The paper used in this publication meets the minimum requirements of the American National Standard
for Permanence of Paper for Printed Library Materials Z39.48–1992.
Dedication
I dedicate this taxonomic study to Jay M. Savage
for the excellence of his half-century of biogeographic
and systematic research and in appreciation for his
professional support—often “behind the scenes”—and
friendship throughout my herpetological career.
Contents
LIST OF FIGURES
vii
LIST OF TABLES
ix
PREFACE
xi
INTRODUCTION
Nomenclatural History
1
MATERIALS AND METHODS
6
CHARACTER ANALYSIS: RESULTS AND DISCUSSION
Baseline Estimate of Intra-Observer Variation
Recognition of Unisexual and Bisexual Populations
Unisexual—Visceral Anatomy
Unisexual—Morphometry
Unisexual—Scalation
Unisexual—Coloration
Regional Variation among Bisexual Populations
Bisexual—Visceral Anatomy
Bisexual—Morphometry
Bisexual—Scalation
Bisexual—Coloration
1
7
7
8
8
9
12
14
15
15
16
20
25
GEOGRAPHY AND TAXONOMY
Regional Patterns of Morphology and Speciation
General Observations
Morphological Differentiation
Taxonomic Decisions and Geography
Species Accounts
Key to the Species of Hemiphyllodactylus
52
ACKNOWLEDGMENTS
55
29
29
29
29
32
35
vi
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
APPENDIX 1: CHARACTER DEFINITIONS
57
APPENDIX 2: SPECIMENS EXAMINED
59
APPENDIX 3: STATISTICAL ANALYSES
63
REFERENCES
65
INDEX
69
Figures
1. Type localities for the available names of species currently
assigned to genus Hemiphyllodactylus
2. Visceral pigmentation of Hemiphyllodactylus species
3. Contrasting habitus of adult Hemiphyllodactylus
4. Principle component graphs of unisexual and bisexual adult
females of sunda Hemiphyllodactylus
5. Frequency distribution of precloacal–femoral pores of
unisexual Hemiphyllodactylus typus
6. Dark and light phases of coloration in Hemiphyllodactylus
typus
7. Morphology of the chin scales in various populations of
Hemiphyllodactylus
8. Cloacal spur morphology in Hemiphyllodactylus
9. Digital lamellae morphology in select species of
Hemiphyllodactylus
10. Precloacal–femoral pore morphology of Hemiphyllodactylus
11. Coloration of select Hemiphyllodactylus taxa
12. Types of Bingtang slender gecko
13. Holotype of Hemiphyllodactylus typus Bleeker
14. Geographic occurrence of Hemiphyllodactylus typus
15. Syntypes of Hemiphyllodactylus aurantiacus Beddome
16. Geographic occurrence of Hemiphyllodactylus aurantiacus
and H. yunnanensis
17. Hemiphyllodactylus ganoklonis from Ulebsechel Island, Palau
18. Holotype of Hemiphyllodactylus ganoklonis from Ulebsechel
Island, Palau
19. Geographic occurrence of Hemiphyllodactylus ganoklonis
20. Geographic occurrence of Hemiphyllodactylus harterti,
H. margarethae, H. titiwangsaensis, and Borneo bisexuals
21. Holotype of Hemiphyllodactylus insularis Taylor
3
8
11
11
13
15
22
23
23
24
27
28
36
38
39
40
41
42
43
45
46
viii
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
22. Geographic occurrence of Hemiphyllodactylus insularis
23. Types of Hemiphyllodactylus margarethae Brongersma
24. Types of Hemiphyllodactylus titiwangsaensis
25. Lectotype of Gehyra yunnanensis Boulenger
47
48
50
51
Tables
1. Available names for populations and species of
Hemiphyllodactylus geckos
2. Summary statistics on select characters of unisexual
Hemiphyllodactylus samples
3. Summary statistics on select mensural characters of adult
females of the bisexual Hemiphyllodactylus samples
4. Summary statistics on select metric characters of adults of
the bisexual Hemiphyllodactylus from southern Asia
5. Comparison of some mensural characters of adult females
among the Chinese populations of Hemiphyllodactylus
yunnanensis
6. Summary statistics on select coloration and scalation
characters of juveniles and adults of the bisexual
Hemiphyllodactylus samples
6
10
17
19
20
21
Appendix 1
A1.1. Abbreviations and deinitions for characters examined
58
Preface
M
y fascination with Hemiphyllodactylus began during a study of
Fijian lizards. Often, the presence or absence of secreting precloacal and/or femoral pores is used to determine the sex of adult
geckos: males possess them, females do not. Although a reliable
assumption for some gekkonid and iguanid lizards, a chance observation on an
adult Fijian Hemiphyllodactylus typus showed its potential for incorrect sex
determination. The Fijian specimen had well-developed pores, yet I remained
uncertain of sex even though its gonads appeared to be ovaries. Histology of
a gonad revealed developing follicles, thus the specimen was an adult female.
Other adult H. typus from Oceania had pores, and examination of their gonads
revealed that all were females. This evidence suggested that all Paciic H. typus
populations are unisexual (Zug, 1991). Further, this discovery caused me to continue my examination of Hemiphyllodactylus specimens and led to an observation that all individuals from coastal localities from Hawaii and Tahiti westward
to New Guinea and those of the Mascarenes share the typus morphotype and
are females.
Not all Hemiphyllodactylus populations, however, are unisexual. The bisexual populations typically occur inland in forested situations from Palau and
the Philippines to Sri Lanka and the Eastern Ghats of India. There are a variety
of names available for these populations (Kluge, 2001): insularis, harterti, yunnanensis, aurantiacus (east to west); and other available names not listed by
Kluge. My primary goal here is to examine morphological variation among all
populations of Hemiphyllodactylus and to address the systematics issues that
arise from this study.
Speciation and Dispersal in a Low
Diversity Taxon: The Slender Geckos
Hemiphyllodactylus (Reptilia, Gekkonidae)
INTRODUCTION
George R. Zug, Department of Vertebrate
Zoology, National Museum of Natural History,
Smithsonian Institution, P.O. Box 37012, MRC
162, Washington, D.C. 20013-7012, USA;
zugg@si.edu. Manuscript received 31 August
2009; accepted 5 May 2010.
Hemiphyllodactylus are small, inconspicuous geckos but incredibly wideranging in the Indo-Paciic realm. The H. typus morphotype occurs from the
Mascarenes eastward through southern Asia to eastern Polynesia and Hawaii.
Throughout this broad distribution, these geckos are not commonly seen, even
by biologists looking for them; thus they have attracted little attention by herpetologists and hobbyists.
How does such an inconspicuous gecko attain such a broad occurrence?
Human transportation seems the obvious answer, although the subsequent questions of how, why, and when are largely unanswered. The search for a datasupported answer is one of the goals of this study. The major goal is to uncover
the diversity of this taxon and to place this diversity in a irm taxonomic setting.
NOMENCLATURAL HISTORY
Bleeker (1860) was the irst naturalist to recognize the uniqueness of this
small gecko. He described his Sumatran gecko as a new species and genus. His
description is adequate, and the survival of the type specimen assures that Hemiphyllodactylus typus Bleeker is associated correctly with a gecko population
today. Although the assignment of the name to a speciic taxon is unambiguous, confusion exists about the type locality and the nomenclatural status of
Ptyodactylus gracilis. These dificulties arise from the last sentence in Bleeker’s
description (1860:237): “Ik bezit eene afbeelding dezer soort; afkomstig van
de voormalige Natuurkundige kommissie, voorzien van den naam Ptyodactylus gracilis en naast welke aangeteekend is, dat de soort ook op den Goenong
Parong (Java) leeft.” My interpretation (based on a translation by T. Ulber, in
litt.) is that Bleeker is telling the reader that the yellow underside of the tail
2
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
and other characteristics (from the preceding descriptive
sentence) are seen in an illustration of his new species
that is labeled P. gracilis; the source of this illustration is
presumably an unpublished report of the Natuurkundig
Commission. As interpreted by Wermuth (1965) in his
gekkonid checklist, Bleeker was not offering a substitute
name. Kluge (1968:342) noted this error of interpretation,
and he proposed correctly that “a picture (drawing) of his
[Bleeker’s] new species . . . bore the name Ptyodactylus
gracilis.” Ulber’s translation shows that Bleeker was noting that H. typus also occurred at Goenong Parong (Java)
based on an unpublished illustration. Ptyodactylus gracilis
is thus a nomen nudum and unavailable.
Kluge (1968:342) noted that the association of Goenong Parong with P. gracilis “led Smith (1935) and Wermuth (1965) to incorrectly consider the type locality of
typus as Goenong Parong, Java.” Kluge’s identiication of
the type locality (Figure 1) with the title of Bleeker’s article
is correct, that is, Agam, a locality in central Sumatra (see
gazetteer in David and Vogel, 1996).
Bleeker was not the only naturalist to recognize the
uniqueness of this gecko. Bavay discovered this gecko on
buildings during his ield work in New Caledonia. He recognized it as a new species, Platydactylus crepuscularis, in
his catalog of New Caledonian reptiles (Bavay, 1869), apparently unaware of Bleeker’s description. While he noted
this gecko’s similarity to Lepidodactylus lugubris, at least
to the description of that taxon provided by Duméril and
Bibron, Bavay’s description explicitly characterized P. crepuscularis as a Hemiphyllodactylus typus Bleeker; thus P.
crepuscularis is a junior subjective synonym. The history
of this name and its type specimens is detailed in Bauer’s
(1994) comment sections for H. typus and L. lugubris. I
wish to examine only one aspect—the holotype or syntypes
of P. crepuscularis. Bavay typically gave the dimension of
a single specimen in his species accounts, whether describing a new or established taxon. Boulenger (1883:123),
however, suggested Bavay had two specimens: “and two
others, male and young, the types of the species, communicated to me by M. Bavay.” Was the “communicated”
a letter with data on the specimens or actual specimens
sent to Boulenger? If the latter, they were not cataloged in
the British Museum, because Boulenger (1885) listed only
the two Benchley specimens that he had mentioned in his
1883 description. There is no evidence that Bavay deposited the type(s) in the Paris Museum, because Sauvage’s
(1879) subsequent description of a type was based on a
specimen of L. lugubris (see Bauer, 1994). Hence the type
of P. crepuscularis is lost, but fortunately Bavay’s description clearly refers to H. typus Bleeker.
Major Beddome, a forestry oficer in Madras (present-day Chennai, India), collected a variety of reptiles
and described them in 1870. One of them, Hemidactylus aurantiacus Beddome, was a Hemiphyllodactylus species from mid-elevation in the Shevaroy Hills. Nothing
in Beddome’s characterization identiies the new species
unequivocally as Hemiphyllodactylus. Boulenger’s (1885)
description is more detailed, and his placement with Lepidodactylus was a better assessment of aurantiacus’ afinities. Boulenger also noted that the type series consisted of
many adult males, females, and juveniles.
Günther (1872) described a typus gecko from the
“East-Indian archipelago” as Spathodactylus mutilatus.
The generic and speciic descriptions and the illustration
of the fore- and hindfeet readily identify the holotype as
Hemiphyllodactylus and likely H. typus. Although Günther did not identify the source of the specimen, Boulenger
(1885) did—Dr. Bleeker. This information suggests that
the types of H. typus Bleeker and S. mutilatus Günther
are the same specimen; thus the latter name is a junior
objective synonym of the former. Because Boulenger included neither Hemiphyllodactylus nor typus as names in
his catalog, it indicates that neither he nor Günther was
aware of Bleeker’s description. Boulenger (1885) did recognize that Günther’s Spathodactylus was a homonym for
a ish and provided a new generic name Spathoscalabotes.
Subsequently, Malcolm Smith (1935) listed the type locality of S. mutilatus as Agam, Sumatra, in his synonymy
of H. typus; this restriction is correct owing to Bleeker’s
original source of the specimen, although in the same synonymy, Smith incorrectly gave Java as the type locality of
H. typus Bleeker.
As noted above, Boulenger (1885) was apparently unaware of Bleeker’s description of Hemiphyllodactylus and
H. typus, because these names are absent from his catalog. He placed crepuscularis, ceylonensis, and aurantiacus
in the genus Lepidodactylus and continued to recognize
Günther’s mutilatus as a monotypic taxon although correcting the generic homonymy.
Boulenger followed the species account of Lepidodactylus crepuscularis (=Platydactylus crepuscularis Bavay)
with an exceedingly brief description of Lepidodactylus ceylonensis. The description identiies the specimen
as Hemiphyllodactylus only by Boulenger’s (1885:164)
statement: “This species resembles exactly the preceding
[L. crepuscularis] in proportions, pholidosis, and coloration.” Boulenger’s illustration is suggestive of Hemiphyllodactylus, but it would it other geckos as well.
Stejneger (1899) provided the irst review of the Hawaiian terrestrial reptiles. Of the seven lizard species then
number 631
FIGURE 1. Type localities for the available names (see Table 1) for the species currently assigned to genus Hemiphyllodactylus. Abbreviations: d, H. yunnanensis dushanensis; j,
H. y. jinpingensis; l, H. y. longlingensis.
•
3
4
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
occurring on these islands, only one appeared to represent a
new species, Hemiphyllodactylus leucostictus Stejneger. Stejneger gave a thorough description and included illustrations
of the head, pelvic area, and hindfoot of his new species. He
stated that the differences of the Hawaiian specimens to H.
crepuscularis, H. ceylonensis, and H. typus were slight but
real, hence appropriate to recognize a new species.
Werner (1900) received a single small gecko from
Malakka (now Malaysia) and noted its distinctiveness. He
named it Lepidodactylus Harterti after the collector and
ornithologist Ernest Johann Otto Hartert. He highlighted
the presence of strongly V-shaped lamellae (11 on fourth
toe) and proposed that it was closely related to Lepidodactylus lugubris, although the inner digit was less well
developed than in L. lugubris.
Boulenger examined a small collection of amphibians
and reptiles from the Larut Hills, Perak, Malaysia, many
of which were new. His descriptions (Boulenger, 1900) of
the new species included a new gecko, Gehyra larutensis.
It is unclear why he had shifted his generic placement of
typus-like geckos. Using the same generic assignment three
years later, Boulenger (1903) described another typus-like
gecko from China, Gehyra yunnanensis. In both descriptions, he mentioned the low number of chevron-like lamellae on the digits characteristic of Hemiphyllodactylus
geckos. He continued to use Gehyra for these geckos in
his book on the Malayan herpetofauna (Boulenger, 1912).
More importantly, Boulenger (1912:48) noted that G.
larutensis “may prove to be identical with this species
[H. harterti Werner, 1900].” He further noted that Werner’s Malakka type locality should be replaced by Gunong
Inas, a site where Hartert collected birds in 1888 (Hartert,
1901, 1902). The nomenclatural signiicance of this tentative correction [reassignment] is addressed in the Taxonomic Decisions and Geography section.
In Reptiles of the Indo-Australian Archipelago, de
Rooij (1915) recognized Hemiphyllodactylus typus and
Lepidodactylus ceylonensis. She used the digital lamellar
morphology and rudimentary and clawless irst digits as the
main diagnostic features for Hemiphyllodactylus. De Rooij
was the irst researcher to recognize a broad distribution for
H. typus and reported its occurrence widely in Sumatra and
Java and on Borneo. She used Lepidodactylus ceylonensis
as an all-encompassing taxon deined by a rudimentary irst
inger and with a distribution from Borneo through Java,
Sumatra, and Nicobar to Myanmar and Sri Lanka.
Taylor (1918) recognized that the Philippine typus
gecko as Hemiphyllodactylus insularis. He compared it
only to H. leucostictus Stejneger, noting that he was uncertain that the Hawaiian gecko was “actually distinct.”
Presumably, he meant distinct from H. typus; however,
he does not mention Bleeker’s taxon or any of the other
typus-like taxa. Taylor was the irst systematist to discuss
variation within a type series (Mindoro) of a typus-like
taxon as well as presenting geographic variation of specimens from other islands in the Philippines. He was also the
irst author to provide natural history observations, noting
that females lay two adhesive eggs beneath tree bark and
that all specimens were beneath bark on seaside trees.
In 1924, Barbour received specimens of Hemiphyllodactylus yunnanensis from The Reverend John Graham,
who had provided the original series of specimens to
Boulenger. Barbour was so struck by the morphology of
foot lamellae that he proposed a new generic name for
this taxon, Cainodactylus. Barbour stated that Dr. Stejneger agreed with him on the uniqueness of the foot
morphology, but Barbour seemingly was so focused on
the differences of his specimens to species of Gehyra and
Hemidactylus that he overlooked Stejneger’s description
and illustrations of H. leucostictus and consequently created a synonym.
Brongersma (1931) described Hemiphyllodactylus
margarethae from four Sumatran specimens representing
two montane localities. He recognized this species’ afinity and differentiated it from H. typus and H. ceylonensis.
Subsequently, Brongersma (1932) published an evaluation
of the nomenclature and characterization of Hemiphyllodactylus and its species. He recognized only two species:
H. aurantiacus and H. typus. Although he recognized H.
aurantiacus, Brongersma examined only two specimens
and purposefully kept his remarks brief. Thus his “Notes”
refer mainly to H. typus, in which he synonymizes H. ceylonensis, H. crepuscularis, H. leucostictus, H. insularis, and
H. margarethae. He noted that M. A. Smith had independently reached the same conclusions. Brongersma reached
his conclusion through the evaluation of three characters
regularly used to diagnose Hemiphyllodactylus species: (1)
denticulate digits, (2) number of precloacal pores, and (3)
if present, number of femoral pores. He concluded that
denticulation was slight in most specimens and “purely individual” (Brongersma, 1932:214). He recognized the dificulty of distinguishing pits and pores, noted the absence
of femoral pores in some males, and concluded that the
number of pores (precloacal and femoral) “seems to be of
no value in this genus” (Brongersma, 1932:216).
M. A. Smith’s conclusions, revealed to Brongersma in
a letter, were subsequently promulgated in his work on
the lizards of British India (Smith, 1935). Therein, Smith
recognized two species of Hemiphyllodactylus: H. typus
typus, H. typus aurantiacus; and H. yunnanensis. The
number 631
nominate subspecies included as synonyms all the species
mentioned in the preceding paragraphs except H. larutensis, thereby giving H. typus a distribution from Sri Lanka
eastward into Oceania. Hemiphyllodactylus t. aurantiacus
retained a southern India distribution. Hemiphyllodactylus
yunnanensis was identiied with a Yunnan, northern Laos,
and northern Myanmar distribution. In a footnote, Smith
(1935:109) proposed that H. yunnanensis was “perhaps a
northern representative of the Malayan Hemiphyllodactylus larutensis (Boulenger).” Smith (1933) in an article that
likely was preparatory to his 1935 catalog examined the
confusion of species assignment to Hemiphyllodactylus.
Therein he provided a concise deinition of the genus and
a list of three included species: typus, yunnanensis, and
harterti. He noted that the latter name had appeared three
months before Gehyra larutensis Boulenger.
Bourret (1937) described Hemiphyllodactylus typus
chapaensis from northern Vietnam (Chapa, Tongking).
He noted that it resembled H. yunnanensis but that his
taxon was not greatly different from H. typus, hence his
assignment to subspeciic status. His description included
ive detailed illustrations of the type.
After Bourret, Hemiphyllodactylus occurred irregularly in the herpetological literature until the 1960s, typically in regional lists, reappearing in Taylor’s lizards of
Thailand (Taylor, 1963) with a full description and Wermuth’s (1965) checklist of all gekkonid lizards. Wermuth
recognized three species (larutensis, typus, and yunnanensis) and three subspecies of H. typus (nominate form,
aurantiacus, and chapaensis). Kluge (1968) addressed
the relationships of Hemiphyllodactylus as well as commenting on nomenclatural matters; these matters were
discussed above. He considered typus and yunnanensis as
full species of Hemiphyllodactylus and left the status of
aurantiacus, chapaensis, and harterti for additional investigation. Kluge considered Hemiphyllodactylus as a sister
group of Lepidodactylus. This hypothesis returns conceptually to Boulenger’s catalog treatment, although retaining
typus and its congeners as a separate genus (lineage).
Wermuth (1966) reexamined a gecko, Platydactylus
minutus Giebel 1862, captured in Baltic amber. He proposed that the specimen was a Hemiphyllodactylus typus.
His Figure 2 of the right forefoot shows subdigital lamellae similar to those of H. typus; however, the dorsal view
of the entire gecko (Wermuth, 1966: ig. 1) is not typuslike. The head, neck, and body are robust and not elongated. The fore- and hindlimbs are large, long, and would
overlap one another if laid along the trunk. With this habitus, Platydactylus minutus Giebel is not a synonym of H.
typus or vice versa.
•
5
Through the 1970s and 1980s, Hemiphyllodactylus
species, mostly H. typus, appeared in assorted publications on regional herpetofaunas. For example, Brown and
Alcala (1978) continued the interpretation of H. insularis
as a synonym of H. typus in their Philippine gecko catalog.
Auffenberg (1980) reviewed the herpetofauna of Komodo
and observed that the Komodo H. typus were lightly colored and nearly patternless. He described the Komodo
population as the subspecies H. t. pallidus. Zhou et al.
(1981) examined a large collection of H. yunnanensis
from Yunnan, Guizhou, and Guangxi Provinces, China,
and observed regional variation in digital lamellae patterns. Because the variation was concordant within four
regions, they recognized three new subspecies: H. y. dushanensis, H. y. jinpingensis, and H. y. longlingensis.
Lazell (1989) made an unexplained alteration (1989)
of leucostictus Stejneger to albostictus in a magazine article.
Zug (1991) revealed the unisexual aspect of Fijian and other
Oceania populations of H. typus. Bauer’s (1994) checklist
of Australian and Oceania gekkonids provided a full synonymy of Hemiphyllodactylus typus and an abbreviated
review of the various nomenclatural usage and alterations.
Manthey and Grossmann (1997) recognized two species
(larutensis, typus) of Hemiphyllodactylus in the Sunda area.
The two have strikingly different coloration in their illustrations and descriptions, conirming the presence of two species in this area. Their concept of H. typus, however, was
as a bisexual species, with males deined by the presence
of femoral–precloacal pores. Soon thereafter, Chan-ard et
al. (1999) listed four species (harterti, larutensis, typus,
and yunnanensis) from Thailand and peninsular Malaysia.
Their photographs show variable coloration among the
specimens identiied as H. harterti and H. larutensis from
the Cameron Highlands, Pahang State, Malaysia.
Subsequently, Bauer and Das (1999) visited the type
locality of Hemidactylus aurantiacus Beddome and captured three adult specimens. Their examination of these
specimens and specimens from Malaysia, Philippines, and
elsewhere demonstrated that the Shevaroyan geckos had
several diagnostic traits that clearly distinguished this
midmontane population from other Hemiphyllodactylus
typus. On the basis of these consistent differences, they
recognized Hemiphyllodactylus aurantiacus as a full species. Kluge’s most recent gekkonid checklist (Kluge, 2001)
similarly returned Hemiphyllodactylus insularis Taylor to
speciic status but without explanation. Gaulke (2003)
briely examined the nomenclatural history of Philippine
Hemiphyllodactylus and, presumably because of the presence of males and females, accepted H. insularis as a distinct taxon from H. typus. The available names for the
6
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
TABLE 1. Available names for populations and species of Hemiphyllodactylus geckos. Type localities presented are from the original
descriptions; because many of these localities may not be obvious or known to some readers, the country in which the locality occurs
is included in brackets.
Date
Name
Author
Type locality
1860
1869
1870
1872
1885
1899
1900
1900
1903
1918
1931
1937
1980
1981
Hemiphyllodactylus typus
Platydactylus crepuscularis
Hemidactylus aurantiacus
Spathodactylus mutilatus
Lepidodactylus ceylonensis
Hemiphyllodactylus leucostictus
Lepidodactylus Harterti1
Gehyra larutensis
Gehyra yunnanensis
Hemiphyllodactylus insularis
Hemiphyllodactylus margarethae
Hemiphyllodactylus typus chapaensis
Hemiphyllodactylus typus pallidus
Hemiphyllodactylus yunnanensis dushanensis
Bleeker
Bavay
Beddome
Günther
Boulenger
Stejneger
Werner
Boulenger
Boulenger
Taylor
Brongserma
Bourret
Auffenberg
Zhou et al.
Agam [Sumatra]
Nouvelle-Calédonia
Shevaroys . . . about Yercaud [India]
East Indies archipelago
Ceylon
Kauai, Hawaiian Islands
Malakka [Malaysia]
Larut Hills [Malaysia]
Yunnan Fu [China]
Sumagui, Mindoro [Philippine Islands]
Fort de Kock, Sumatra
Chapa [Vietnam]
Vai Nggulung, Loho Liang, Komodo
Dushan County, Guizhou Province, China2
1981
1981
Hemiphyllodactylus yunnanensis jinpingensis
Hemiphyllodactylus yunnanensis longlingensis
Zhou et al.
Zhou et al.
Jinping County, Yunnan Province, China2
Longling County, Yunnan Province, China2
1
2
Harterti is capitalized as it appears in the original description.
Type locality presented in Chinese.
various populations of Hemiphyllodactylus are summarized chronologically in Table 1.
MATERIALS AND METHODS
Despite the broad distribution of Hemiphyllodactylus, the availability of voucher material in museum collections is relatively poor. The assembly of an adequate
quantity of specimens required access to many collections;
the collection names are abbreviated here for subsequent
mention in the text.
MCZ
NMB
NMW
QM
RMNH
SAM
SDMNH
SMF
THNHM
AMNH
AMS
BMNH
BPBM
CAS
CM
FMNH
IRSNB
KUZ
American Museum of Natural History
Australian Museum, Sydney
The Natural History Museum, London
Bernice P. Bishop Museum
California Academy of Sciences
Carnegie Museum of Natural History
Field Museum of Natural History
Institut royal des Sciences naturelles de
Belgique
Kyoto University, Department of Zoology
UF
USNM
WAM
WmBeckon
ZMA
Museum of Comparative Zoology,
Harvard University
Naturhistorisches Museum, Basal
Naturhistorisches Museum, Wien
Queensland Museum
Nationaal Natuurhistorisch Museum
(formerly Rijkmuseum van
Natuurlijke Historie)
South Australian Museum
San Diego Museum of Natural History
Natur-Museum u. Forschungs Institut
Senckenberg
Thailand Museum of Natural History,
National Science Museums
Florida Museum of Natural History,
University of Florida
U.S. National Museum (National Museum of Natural History,
Smithsonian Institution)
Western Australian Museum
William N. Beckon, private collection
Zoölogische Museum, Universiteit van
Amsterdam
number 631
ZMB
ZMFK
ZRC
ZSM
Museum für Naturkunde, Universität zu
Humboldt
Zoologische Forschungsinstitut u.
Museum Alexander Koening
Zoological Reference Collection,
National University Singapore
Zoologisches Sammlung des Bayerischen
Staates
I grouped the specimens into 13 regional samples,
each representing a putative biogeographic area abbreviated in small capitals and deined as follows:
China
Fiji
Hawai
India
Mascar
NCal
NGuin
Palau
Philip
Polyn
sEasia
Sunda
Taiwan
China and northeastern Myanmar
Fiji and Tonga
Hawaiian Islands
India and Sri Lanka
Mascarenes
New Caledonia and Vanuatu
New Guinea and Solomon Islands
Republic of Palau
Philippine Islands
Polynesia
Thailand (north of Isthmus of Kra) to
Vietnam and Hong Kong
Malaysia and Indonesia
Taiwan and Japan
These regional samples vary in size (n = 9–85) and
geographic extent. In the latter case, a sample can be examined as two or more subsamples of restricted localities
if adequate specimens are available or if intrasample variation indicates a mixed sample. Further, I combined samples and repartitioned specimens when an initial analysis
suggested the presence of multiple bisexual taxa in one or
more of geographically adjacent samples.
Morphological data consist of a combination of morphometric and meristic characters. These characters are
identiied and their abbreviations deined in Appendix
1. Sex and maturity were determined by dissection and
examination of the gonads. Maturity criteria were those
outlined in Zug et al. (2003). The small size of this taxon
seems to have resulted in a high level of inattentiveness to
the preservation and positioning of specimens. Contorted
specimens, commonly with clenched fore- and hindfeet,
made data-gathering challenging and certainly increased
the variation in most measurements and scale counts.
The data were analyzed by a variety of univariate and
multivariate statistics using SYSAT version 11. My goal has
been to examine and describe intra- and interpopulational
variation as thoroughly as possible considering the variable
•
7
preservation state of many specimens. The multivariate
models were used as exploratory techniques to compare
populations and possibly reveal differentiation within and
among samples. Explanation of the use of the multivariate
analyses and the results are available in Appendix 3.
In addition to the abbreviations deined above and in
Appendix 1, other symbols and abbreviations used in this
publication are deined as follows:
alt.
CV
DFA
GPD
IDH
MPI
PCA
SD
r2
altitude
coeficient of variation
discriminant function analysis
glycerosphophate dehydrogenase
isocitrate dehydrogenase
mannose phosphate isomerase
principal components analysis
standard deviation
coeficient of determination
CHARACTER ANALYSIS:
RESULTS AND DISCUSSION
BASELINE ESTIMATE OF INTRA-OBSERVER VARIATION
How much of the variation observed in each sample results from the researcher’s data-gathering behavior? Hayek
et al. (2001) addressed that question and recommended a
repeated measuring protocol to obtain an estimate of this
portion of a character’s and a sample’s variation. A single
specimen (in this case USNM 563683, female from Palau)
was measured and scalation recorded 10 times, each time on
a separate day over a period of 6 weeks. Central tendency
statistics reveal that for measurements, the coeficients of
variation (CVs) range from 0.7% (mean snout–vent length
[SVL] = 32.5 mm ± SD 0.24) to 9.3% (mean SnW = 1.1 mm
± SD 0.10). The larger measurements (SVL, TrunkL, HeadL,
HeadW—see Appendix 1 for deinitions) have the least variation (CV, 0.7–1.9%) relative to the smaller ones (SnEye,
NarEye, EyeD, SnW; CV, 3.2–9.3%). The coloration traits
were invariant, as were the majority of the scalation characters. Ventral (mean = 9.3 mm ± SD 1.16) and 2ToeLm
(3.7 mm ± SD 0.48) had the highest CVs (12.5%, 13.1%,
respectively); in all other variable traits, CV was <8%.
These results provide a baseline for assessing the
variation observed in regional samples. Further, these estimates probably represent the lowest variation for the
Hemiphyllodactylus data as they were recorded from a
well-preserved and well-positioned specimen and collected
under optimal laboratory conditions. While the data
8
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
reported throughout this study were gathered by a single
individual (me), they were gathered over two decades
(1989–2008) and in a variety of museum situations. As
noted above, Hemiphyllodactylus specimens infrequently
receive adequate preparation attention when collected. In
spite of the inattentive preparations, the subsequent results reveal that intrasample variation is surprisingly low
in most characters and samples; nonetheless, the reader is
advised to be cautious in over-interpreting reported differences, particularly in small samples and/or where differences are less than 2 times a character’s standard deviation.
RECOGNITION OF UNISEXUAL AND BISEXUAL POPULATIONS
Unisexual populations contain only females. Although it may be a statement of the obvious, how does
one conirm the unisexuality of a population? Because neither data on reproduction of virgin females nor mitochondrial DNA were available to me, I relied on the absence of
males in samples as a hypothesis of unisexuality for populations or sets of populations. On that basis, the Hawaiian
(hawai), Polynesian (Polyn), Fijian–Tongan (Fiji), New
Caledonian–Vanuatuan (nCal), New Guinean–Solomons
(Nguin), Taiwan (taiwan), and Mascarene (Mascar)
samples are considered unisexual populations. All other
samples contain males, but owing to their manner of assembly, some samples, especially the Sundaland (sunda)
and Indian–Sri Lankan (india) ones, are likely mixtures
of unisexual and bisexual individuals.
This situation requires an assumption: Hemiphyllodactylus typus Bleeker (BMNH 1946.8.30.83) represents
(is) a unisexual species. The holotype is an adult female
with precloacal and femoral pore series separated. To
test this assumption, the initial character analyses examined variation in and among the Paciic samples (hawai,
Polyn, Fiji, nCal, nguin, taiwan) and then when they
proved geographically homogeneous compared them to
the type specimen. Assuming that this comparison yielded
accurate “diagnostic” data, H. typus specimens were removed from the mixed unisexual–bisexual samples prior
to examination of bisexual’s within sample and between
(interregional) sample variation.
Unisexual—Visceral Anatomy
Because sex and maturity were determined by dissection, I observed a consistent pattern of differential pigmentation in the viscera of Paciic Hemiphyllodactylus, although I
did not report this observation in my study of Fijian lizards
(Zug, 1991). All Paciic specimens have the caecum and
oviducts heavily pigmented (melanin) (Figure 2A,B). For
FIGURE 2. Visceral pigmentation of Hemiphyllodactylus species.
All images in ventral view. (A) Pigmented caecum visible externally
between pair of oviducal eggs, H. typus, Hawaii (U.S. National Museum [USNM] 570747); (B) pigmented caecum and oviduct, large
intestine inlected anteriorly (left side of image), H. typus, Hawaii
(USNM 570742); (C) unpigmented caecum and oviduct, H. yunnanensis, Myanmar (USNM 570734).
number 631
the oviduct the peritoneum sheath likely bears the melanin.
The pigmented caecum and oviducts, however, are not conined to the unisexual samples. This pigmentation pattern is
widespread, but not global, in bisexual samples.
The black caecum is often visible through the ventral
body wall and skin (Figure 2A). The pigmentation of the
oviducts is darkest in virgin females. It is invisible in the
greatly stretched oviducts of gravid females. The oviducts
become dusky brown once the eggs are expelled. The duct
walls retain a laccid and stretched morphology following
the irst egg production cycle.
The type of Hemiphyllodactylus typus Bleeker was
not dissected, so I am unable to conirm the pigmentation
of its oviduct or its sex by a direct examination of the
reproductive tract. It does not show any thickness at the
base of the tail, thereby indicating the absence of hemipenes. The black caecum is visible through the body wall.
Précis.
Unisexual Hemiphyllodactylus typus
possess darkly pigmented caeca and oviducts. Bisexual
Hemiphyllodactylus are variable in this pair of traits.
Unisexual—Morphometry
hawai is the largest of the Paciic samples and serves
as a base to examine levels of variation and differentiation
within and among the Paciic samples. Hawai has neither
the largest nor smallest adult females. Its mean and median SVL (36.7 and 36.3 mm) match those of Polyn and
taiwan and is less than those of the other samples. nguin
and Fiji are similar, with mean/median of 40 mm SVL. The
smallest adult (29.2 mm SVL) is from Taiwan; the largest
(46.1 mm) is from Fiji. The other eight mensural traits
show the same pattern of similarity among the samples.
Intrasample variance is also similar within and among the
Paciic samples, as seen by a comparison of the ranges of
the coeficient of variation (CV) with the CV for the hawai
datum: SVL 6.5% (hawai datum), 3.0–10.4% (range for
the Paciic samples); TrunkL 11.3%, 5.2–11.7%; HeadL
5.0%, 4.6–8.3%; SnEye, 7.0%, 5.0–13.9%; NarEye
7.9%, 6.9–12.8%; EyeD 7.2%, 6.5–11.3%; SnW 10.5%,
5.9–15.5%; HeadW 9.2%, 5.1–12.1%. The higher CVs
are associated mainly with taiwan (n = 9), which has the
greatest range of adult SVL (29.2–43.6 mm). Mascar is
also an all-female sample and presumably represents a
unisexual population. Its CVs match those of the Paciic
samples (Table 2). Variation and means of the samples,
either individually or combined (i.e., total Paciic unisexual sample; Pacif), are equivalent. There is no evidence
of mensural differentiation among the Paciic insular samples or the distant Mascar sample. Principal components
analysis (PCA) and discriminant function analysis (DFA)
•
9
of the combined Paciic samples, Mascar sample, and the
holotype of H. typus similarly reveal a uniform morphology among these geckos. (See synopsis of PCA and DFA
results in Appendix 3.)
The Paciic unisexuals and typus holotype are slender,
elongate geckos (Figure 3). The proportionately short limbs
accentuate the trunk elongation. A proportion of hindlimb
length to trunk length would demonstrate this morphology; however, my preliminary measurements of hindlimb
length were extremely variable owing to the dificulty of
measuring accurately tiny twisted limbs with ist-like preserved feet in many specimens. Thus I excluded this trait
from subsequent data gathering. The proportion TrunkL/
SVL provides a metric, although a less satisfactory one, for
portraying the relative elongation of the trunk. Linear regression of the two preceding traits also reveals the degree
of elongation through the depression of the slope; however, in contrast to the proportion, variation from linearity
(as measured by coeficient of determination, r2) was high,
thereby reducing the reliability of the slope as indicator
of trunk elongation. The means and standard deviations
for TrunkL/SVL (as percentage) among the Paciic typus
samples are 52.0% ± 4.0 (hawai), 54.4% ± 3.5 (Polyn),
55.4% ± 4.2 (Fiji), 52.6% ± 1.6 (nCal), 54.0% ± 2.4
(nguin), 52.9% ± 3.9 (Pacif). To place these proportions
in the context of other geckos, the mean ± SD TrunkL/SVL
for Gehyra oceanica is 42.7% ± 3.3 (n = 113), Hemidactylus frenatus is 42.0% ± 2.8 (n = 21), and Lepidodactylus
lugubris is 44.1% ± 3.2 (n = 149). Trunk length is over
50% of total SVL in the H. typus samples and distinctly
less than half of SVL in other Paciic geckos. Part of this
proportional difference derives from a smaller head in H.
typus, that is, 21.3% ± 0.9 HeadL/SVL (Pacif), as compared to 26.7% ± 1.6 for G. oceanica, 26.4% ± 1.6 for H.
frenatus, and 24.2% ± 1.7 for L. lugubris. Combining the
head and trunk proportions shows that neck length is proportionately shorter in H. typus in contrast to the visual
impression of the attenuate habitus of H. typus.
Do these proportions permit the segregation of unisexual H. typus from bisexual Hemiphyllodactylus? Any
of the regional samples might contain both unisexual
and bisexual species. Because the Sunda most certainly
is mixed, it is the appropriate sample on which to test the
morphometrics for differentiation.
The sunda sample (n = 83) contained 55 adult females
of which 39 were identiied as H. typus during data collection (based on coloration and pigmentation of caecum
and oviducts), 15 males and 7 females as unknowns or
uncertains, the holotypes each of H. larutensis (Boulenger)
(adult male) and H. harterti (Werner) (adult female), holotype (adult male) and three paratypes (2 adult females,
10
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
TABLE 2. Summary statistics on select characters of unisexual Hemiphyllodactylus samples. Statistical values are mean (mensural traits)
or median (scalation) ± standard deviation (SD), range of minimum to maximum, and coeficient of variation (CV); and modes and
frequency (%) of occurrence (inger and toe lamellae). Sample sizes (n) are mature females and total sample, respectively; statistics derive
solely from adults for mensural traits, and from juveniles and adults for scalation ones. The n value is the total number of specimens
examined for each locality sample; the actual statistic may have derived from fewer individuals because not all characters could be measured or counted in all specimens. Character abbreviations are deined in Appendix 1.
Sample (n)
Character and statistic
SVL
Mean ± SD
Range
CV
TrunkL
Mean ± SD
Range
CV
HeadL
Mean ± SD
Range
CV
SnEye
Mean ± SD
Range
CV
PostocSpt
Median ± SD
Range
CV
SnS
Median ± SD
Range
CV
Suplab
Median ± SD
Range
CV
Chin
Median ± SD
Range
CV
Dorsal
Median ± SD
Range
CV
CloacS
Median ± SD
Range
CV
4ToeLm
Median ± SD
Range
CV
FingerLm b
Modal values
Frequency
ToeLm b
Modal values
Frequency
a
typus Holotype
(1)
Hawaiian Islands
(37, 42)
Total Paciic unisexual sample
(99, 118)
Mascarene group
(8, 11)
43.3
36.7 ± 2.28
32.4–42.8
6.2%
38.0 ± 2.91
29.2–46.1
7.6%
38.1 ± 1.00
38.1–40.9
2.5%
22.4
19.4 ± 1.94
15.1–23.9
10.0%
20.2 ± 2.37
14.0–28.0
11.8%
20.7 ± 1.31
19.2–23.1
6.3%
9.1
7.9 ± 0.38
7.1–8.8
4.8%
8.1 ± 0.55
6.6–9.9
6.8%
8.3 ± 0.26
7.8–8.6
3.2%
3.7
3.3 ± 0.22
2.7–3.7
6.8%
3.3 ± 0.29
2.3–4.1
8.7%
3.5 ± 0.16
3.3–3.8
3.5%
a
4 ± 0.80
1–5
22.6%
3 ± 0.95
1–5
29.7%
3.2 ± 1.10
2–5
33.7%
3
2 ± 1.01
1–5
40.7%
2 ± 0.81
1–5
37.7%
2 ± 0.50
2–3
22.1%
11
11 ± 0.85
10–14
7.6%
11 ± 0.94
9–14
8.3%
11 ± 0.82
10–13
7.3%
13
12 ± 0.99
10–14
8.4%
11 ± 1.08
9–14
9.4%
10 ± 0.71
10–12
6.7%
13
14 ± 1.53
12–18
10.6%
15 ± 1.64
12–19
11.0%
15.0 ± 1.48
13–17
10.0%
2
3 ± 0.80
1–5
29.6%
2 ± 0.79
0–5
34.0%
2 ± 1.00
1–4
50.0%
4
5 ± 0.33
4–5
6.8%
5 ± 0.33
4–5
6.7%
4 ± 0.47
4–5
8.4%
3-3-4-3
3-4-4-4
42.5%
3-4-4-4
47.7%
3-4-4-3 c
27.3%
4-4-4-4
4-4-5-4
51.2%
4-4-5-4
50%
4-4-5-4
36.4%
This value/character is unknown in the holotype because of fading.
Lamellae formulae represent the most frequent formula (mode) for each sample and the percent of the sample with this formula (frequency).
c
Two inger formulae share 27%; the second is 3-4-4-4.
b
number 631
•
11
FIGURE 3. Contrasting habitus of adult Hemiphyllodactylus: (A)
elongate morphotype, H. typus, Hawaii (USNM 27924); (B) robust
morphotype, H. yunnanensis, China (British Museum of Natural
History [BMNH] 1904.11.29.10D). (Illustration by J. Kilby.)
immature male) of H. margarethae Brongersma, and 11 (5
adult females, 6 adult males) other Malaysian “harterti.”
The sunda H. typus included Bleeker’s holotype. A PCA of
the body proportions of all sunda females yielded a clustering of H. typus in the bottom quadrant of the PCA scores
graph (Figure 4). EyeD/NarEye, NarEye/HeadL, and SnW/
HeadW had the strongest loading (0.63–0.88) on the irst
component, HeadW/SVL and HeadL/SVL strongest loading (0.95, 0.81) on the second component, and SnW/HeadL
and SnW/HeadW on third component (0.88, 0.66). The
irst three components accounted for 68% of the total variance (80%, with the inclusion of the fourth component).
These results also place the H. typus holotype (BMNH
1946.8.3) within the H. typus cluster (Figure 4). The bisexuals lie principally outside the typus cluster in the upper
left quadrant. One of the H. margarethae paratypes (ZMA
11096) is a distant outlier (upper right quadrant) from all
other Hemiphyllodactylus females. The other H. margarethae paratype (IRNS9338B) is on the outer edge of the
typus cluster. Overall, these results indicate a difference in
head shape (PCA 1) and relative size (PCA 2) between the
unisexual H. typus and the bisexual species. I note that I
identiied all Bornean females as H. typus and all Bornean
males as unknown bisexuals. All Bornean females lie within
the typus cluster. The holotype of H. typus (0.288, −1.362)
lies within the lower half of the H. typus cluster (Figure 4B).
FIGURE 4. Distribution of unisexual and bisexual adult females of
sunda Hemiphyllodactylus in multivariate (principal components)
morphometric space: (A) Females identiied by locality (LOC) and
(B) females identiied by species (ID). Abbreviation by locality: BA,
Bali; BO, Borneo; JV, Java; KO, Komodo; MO, Mollucas; MY,
Malaysia; SN, Singapore; SU, Sumatra. Abbreviations by species: ha
(circle), H. harterti; mg (plus sign), H. margarethae; ty (square), H.
typus; un (diamond), unknown/unassigned species.
The assignment of sunda specimens to H. typus was
based on the presence of pigmented caeca and oviducts
and, to a lesser extent, on dorsal coloration. The unknowns
and bisexual type specimens had either unpigmented caeca
or both oviducts and caeca unpigmented (Types were not
dissected so oviducal pigmentation is usually unknown for
12
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
them.). Re-examining the proportions of the outliers and
the types versus the typus specimens revealed that most
(37 of 39 individuals) sunda H. typus had mean TrunkL/
SVL values ≥50% (52.3% ± SD 2.3, range of 46–56%)
and a more ambiguous result in the other group. H. margarethae paratypes had TrunkL/SVL of 52% and 54%;
the outliers were ≤50%, although not greatly so except for
one Sumatran specimen (RMNH 7371, 40%). sunda H.
typus HeadL/SVL matched well the Pacif sample (mean
21.5% ± SD 1.0, range 19–23.8%), showing the smallheaded condition; the bisexuals ranged 22.2–26.8%. The
three highest loading proportions showed a greater degree
of overlap between unisexuals and bisexuals, for example,
EyeD/HeadL (highest loading on irst component) 75.1%
± 8.6, 61–106% for unisexuals versus 49–81% for bisexuals; both the 49% and 106% values were extreme outliers
of their samples and likely data collection errors.
The segregation of the unisexuals (H. typus) and bisexuals in the sunda sample allowed a summary of the
major statistics for the Sundan H. typus (n = 40, including
the holotype): SVL (mean length ± SD, range): 38.7 mm ±
3.19, 32.3–44.1 mm; TrunkL: 20.3 mm ± 1.96, 16.9–24.2
mm; HeadL: 8.3 mm ± 0.62, 7.1–9.4 mm; TrunkL/SVL
(mean proportion ± SD, range): 0.52 ± 0.02, 0.46–0.56;
HeadL/SVL: 0.21 ± 0.01, 0.20–0.24; HeadW/HeadL: 0.65
± 0.04, 0.55–0.63.
Philip is predominantly a bisexual sample, containing
eight unisexuals from Palawan and two from Mindanao.
The summary statistics of unisexuals are SVL: 39.0 mm ±
SD 2.75 with range of 35.3–41.3 mm; TrunkL: 21.1 mm ±
1.84, 18.4–22.4 mm; HeadL: 8.3 mm ± 0.77, 7.2–9.0 mm;
TrunkL/SVL (mean proportion ± SD, range): 0.54 ± 0.02,
0.52–0.57; HeadL/SVL: 0.21 ± 0.01, 0.20–0.22; HeadW/
HeadL 0.70 ± 0.08, 0.66–0.82. The sEasia sample includes only one unisexual (SVL 38.3 mm), conirmed by
pigmented oviducts and caecum; this adult female, however, lacks secreting precloacal–femoral pores, thereby
questioning its assignment to H. typus. india has four unisexual specimens but only in the Sri Lankan component of
the sample; two other Sri Lankan specimens are bisexuals.
One of the unisexuals is the holotype (BMNH 74.4.1326)
of Lepidodactylus ceylonensis Boulenger. Summary statistics for the india H. typus are SVL (mean ± SD, range):
34.0 mm ± 10.43, 18.5–40.5 mm; TrunkL: 17.4 mm ±
5.34 9.4–20.5 mm; HeadL: 7.4 mm ± 1.98, 4.5–8.9 mm;
TrunkL/SVL (mean proportion ± SD, range): 0.51 ± 0.02,
0.50–0.54; HeadL/SVL: 0.22 ± 0.02, 0.20–0.24; HeadW/
HeadL: 0.66 ± 0.02, 0.63–0.67.
Précis.
(1) The Oceania H. typus samples are
homogeneous within and among samples, representing
a single genetic entity and henceforth treated as a single
sample (Pacif). (2) Morphometrics weakly differentiate
unisexual individuals (Hemiphyllodactylus typus) from
bisexual ones. Two proportions, TrunkL/SVL and HeadL/
SVL, appear most useful in this differentiation.
Unisexual—Scalation
An overview of scalation variation within unambiguous unisexual samples is presented in Table 2. Intrasample variation, as estimated by CV, is nearly identical to
intersample variation (i.e., CVs of Pacif), resulting from
the similarity of means, medians, and ranges of the scalation traits; thus the subsequent character survey focuses
on the Pacif sample and examines the regional samples
only when one or more of the latter samples deviates from
Pacif. As noted in the morphometric section, many museum vouchers were poorly positioned at time of preservation. Also, Hemiphyllodactylus are small geckos, and
many were examined in circumstance of poor lighting and/
or optics. I assumed that such circumstances would cause
high variation, but fortunately, “measurement error” from
these data-gathering dificulties is low, and even modestsized samples (≥10 individuals) provide reliable estimates
of population parameters.
Five traits (CircNa, SnS, CloacS, TotPore, PreclPor)
show high variation (CV ≥ 16%). The high variation for
CircNa, SnS, and CloacS results from a high frequency of
one character state and a lesser occurrence of the other
states. For Pacif, CircNa is three scales for 78% of the
sample, and one, two, four, and ive scales for the remainder of the sample. The situation is not as extreme for SnS
and CloacS, but with two scales (range 1–5) representing
more than 50% of the sample for SnS, and two (~50%)
and three (~30%) spurs (range 0–5) for CloacS. TotPore
and PreclPor variation results from a broad range (0–26
[median 14], 0–13 [10], respectively) of pore numbers
(Figure 5). The majority of adult H. typus have femoral
pores, although this trait was not recorded during data
collection; it can be calculated by subtracting PreclPor
from TotPore.
Because the presence of secreting pores in adult females characterizes the unisexual H. typus, PreclPor and
TotPore require further examination. Of the 92 Pacif
adults examined, 88 individuals possessed PreclPor. The
four females without pores ranged in SVL from 35.6 to
42.2 mm, well above the minimum SVL (29.2 mm) at
sexual maturity. The condition of the ovaries was not recorded for the two largest individuals (40, 42 mm SVL).
The 39-mm individual had small ovarian follicles, and a
number 631
FIGURE 5. Frequency distribution of precloacal–femoral pores in the Paciic sample of unisexual Hemiphyllodactylus typus. Size classes are length from snout to vent (SVL) at 2-mm intervals for PreclPor and TotPore
with midpoints plotted on x axis: (A) Precloacal pores and (B) total precloacal–femoral pores.
•
13
14
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
35-mm female had mid-vitellogenic follicles. These data
indicate that all poreless individuals were mature, although the pores may have closed owing to reproductive
quiescence or senescence. Both PreclPor and TotPore have
broader ranges and greater variation (Figure 5) than for
most other gecko species. While gathering data on Hemiphyllodactylus, I developed the impression that the number of pores associated with body size/maturity. Such an
association is not supported by regression or correlation
analyses: for example, Pearson R = 0.074, PreclPor to
SVL; 0.27, femoral pores to SVL; 0.22, TotPore to SVL.
Secreting pores occur only in adult H. typus; perhaps new
pores develop with age and/or each successive egg development event. Precloacal and femoral pores are not continuous in any H. typus. Another pore trait of H. typus
is the absence of femoral pores in some sexually mature
individuals. There is no obvious pattern to their absence.
Femoral pores (≥1) occur in the majority of adults, more
or less symmetrically on left and right: hawai 37 individuals with pores, 10 of these lack femoral pores; Polyn 6,
3; Fiji, 13, 3; nCal 9, 1; nguin 15, 5; and taiwan 10, 1.
Suplab, Inlab, and Chin display normal levels of variation (i.e., CV < 10%). Suplab has a median of 11 scales
(range, 10–12), Inlab 11 (9–11), and Chin 11 (11–12).
Dorsal (15, 13–18) and Ventral (11, 10–14) are somewhat
more variable (CV = 9.8%, 10.4%, respectively). Subcaud
are invariably equal-sized to adjacent caudal scales.
The typus digital formulae using the median value of
the individual digits are 3-4-4-4 (forefoot) and 4-4-5-4
(hindfoot). Both 4FingLm and 4ToeLm are invariant, 4
and 5, respectively. Digital lamellae vary by only one scale
from the median value, either 3 or 4 for 2Fing- to 5FingLm
and 2ToeLm, or 4 or 5 for 3Toe- to 5ToeLm. 1FingLm
and 1ToeLm have median values of 5 (4–5 1FingLm, 4–6
1ToeLm).
The Mascar sample, although considerably smaller
than Pacif, shares similar medians and ranges for the
scalation traits (Table 2). PreclPor differs slightly (range
4–12) and TotPore has a maximum of 17. These differences likely result from vagaries of sampling. The other
difference, probably of similar origin, is the median forefoot formula of 3-4-4-3.
sunda has a much larger H. typus component (n =
40), although as for Mascar, the pattern of variation and
ranges match those of Pacif. TotPore and PreclPor share
similar ranges, although the TotPore median is less, 6.5
pores. The digital formulae for the fore- and hindfoot
are the same. The much smaller H. typus component of
Philip (n = 4) and Sri Lanka (n = 4) also largely matches
the medians of the Pacif sample.
Importantly, pooling adult H. typus from all localities
(n = 143) yields coeficients of variation for measurements
and scalation (excluding PreclPor and TotPore) that are
nearly identical to CVs for the repeated measures sample.
This result argues strongly for the genetic homogeneity of
H. typus populations across the entire Indo-Paciic distribution of this unisexual taxon.
Précis.
(1) The island and island group samples are homogeneous within and among samples, thereby
representing a single genetic entity, Hemiphyllodactylus
typus (Pacif). This homogeneity is shared among the unisexual components of the other regional samples – Philip,
sunda, india, and Mascar.
Unisexual—Coloration
Hemiphyllodactylus typus are not brightly colored
geckos (Figure 6). Their background color ranges from a
dusky tan in the light color phase to a reddish brown in the
dark phase. In both phases, a series of narrow, dark brown,
transverse bars or blotches lies middorsally from neck to
base of tail. These bars are commonly irregularly edged,
and they either extend entirely across the back or are broken middorsally. A series of moderately spaced, small light
(white to creamy beige) spots lie dorsolaterally on each side
of the trunk from the neck to hindlimbs. These spots are
a continuation of the pre-orbital light stripe running from
the naris to the anterior temporal area, ending in a brighter
spot and then continuing as spots on the neck. In most instances, a dark brown stripe lies below the light stripe on the
lip, fading and disappearing beyond the ear opening. The
venter from chin to vent and onto tail is a dusky light tan to
yellowish tan; the dusky appearance is created by a multitude of tiny dark brown spots, a few in each scale. Dorsally
at the base of the tail (postsacral), a dark brown bar and
abutting white to beige bar begin the irregular dark–light
banding of the tail; this banding quickly becomes progressively diffuse and indistinct in most individuals.
In life, a few individuals will appear uniformly brown
except for faded head markings. This situation is the common condition for preserved specimens and limited the development of quantitative coloration coding only to two
characters (OrbStrp and PostocS; Appendix 1).
Neither of these two traits shows any striking variation within or among the Paciic regional samples (Table
2). Similarly there was no variation in coloration among
the non-Paciic samples that was not encompassed by the
description above.
Précis.
Hemiphyllodactylus typus are predominantly dull-colored geckos. The most striking coloration
number 631
•
15
FIGURE 6. Dark and light phases of coloration in Hemiphyllodactylus typus from the Paciic
population: (A) USNM 310814, Hawaii, Oahu (photograph by G. Zug) and (B) USNM 267979,
Fiji, Viti Levu (photograph by J. R. H. Gibbons).
features are the lateral head stripe ending in a bright light
spot, a series of small light spots dorsolaterally on the
trunk, and bright double bar of dark and light at the base
of the tail.
REGIONAL VARIATION AMONG BISEXUAL POPULATIONS
The preceding analysis of regional samples identiied
six samples containing bisexual individuals and populations and revealed the possibility of eight or more different
bisexual populations: Palau, Philip, sunda (potentially
four bisexuals), sEasia, China, and india (potentially
two bisexuals).
Bisexual—Visceral Anatomy
In the bisexual samples, three (Palau, Philip, india)
have females with pigmented caeca and oviducts (see Figure 2B). Females of sEasia and China commonly lack
pigmentation on these organs; however, an adult sEasia female (BMNH 1931.11.21.1) from Thailand has pigmented
oviducts and caecum. Within the Sunda sample, some females have the pigmented condition and others do not. I
observed no differences in extent or intensity of pigmentation among the unisexual and bisexual populations. Occasionally, a female of a “pigmented” population will lack
pigmentation on the caecum or oviduct. One each nguin
and a sunda H. typus has unpigmented oviducts, and a
Philip male and a Sri Lankan female have unpigmented
caeca. In pigmented populations, commonly the peritoneum over the testes and epididymides is pigmented, although not as densely as oviducal pigmentation in females.
Most individuals identiied as bisexuals in the sunda
sample lack organ pigmentation. All sunda bisexuals have
unpigmented caeca except three adult males from Borneo
(FMNH 158734, 196268A, 213665), and the last male
has pigmented epididymides. Four Sumatran females have
pigmented oviducts and two of these are paratypes (IRSN
9338B, ZMA 11096) of Hemiphyllodactylus margarethae
Brongersma.
Précis.
(1) Pigmentation of caecum and oviducts
is not conined to unisexual H. typus; this pigmentation
16
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
trait also occurs in females of the bisexual Palau, Philippines, and Indian populations, and some Sumatran females. (2) Generally, this pigmentation is absent in bisexual
females of mainland Southeast Asia and the Greater Sunda
Islands.
Bisexual—Morphometry
In bisexual populations, sexual dimorphism is the irst
morphometric issue to examine. A comparison of adult
females and males in each regional sample (excluding
sunda) reveals size dimorphism in Palau, Philip, SEasia,
and China. In all four samples, females average larger
than males: 32.8, 35.1, 39.0, 43.3 mm SVL, respectively;
30.6, 31.4, 35.7, 39.9 mm (Student t test, P < 0.025 for
signiicant difference between means of females and males;
see Table 3 for female ranges). Disparities in average body
size result in signiicant differences (P < 0.05) for most
measurements (not EyeD and SnW) in China and (not
SnEye and SnW) in Palau. Philip shows dimorphism in
TrunkL, HeadL, and SnEye. sEasia shows dimorphism
only for SVL. None of the proportions displays dimorphism in any sample.
Among all samples, average adult male SVL ranges
from 30.3 mm (range 28.3–31.6 mm, Palau) to 39.9
mm (35.1–46.4, China). Regional variation in female
morphometrics is detailed in Table 3. The two eastern
mainland Asian populations (Vietnam and Hong Kong)
average strikingly larger than the other mainland bisexual populations. The Palauan population has the smallest adults of any bisexual or unisexual populations. It is
noteworthy that Palau, Philip, and india adult females
average smaller than Hemiphyllodactylus typus (38.1 mm
SVL; Table 3). Subsequently, I will provide some evidence
for the hybrid origin of H. typus from the Palauan and
Philippine bisexual populations.
Variation within the bisexual samples is similar to
that for the regional H. typus samples. The CVs for measurement data (females of Palau, Philip, sEasia, China,
india) are: SVL 3.1–11.7%; TrunkL 4.3–15.2%; HeadL
3.6–12.1%; HeadW 4.1–21.4%; SnEye 5.6–13.5%; NarEye 4.3–14.9%; EyeD 5.0–13.6%; SnW 6.2–21.4%. Relative to the repeated-measure CVs, the preceding CVs are
greater but only about 3% higher than observed in the
repeats. Of these CV ranges, Palau females have the least
variation (except SnW), Philip the lowest and SEAsia the
greatest. The samples are clustered with one group (Palau,
Philip, india) at the low end of the range and SEAsia
and China at the high end; this clustering is consistent
across all measurements. The high variation of SEAsia
and China suggests the possibility of samples containing
two or more taxa. Such mixing has a probability for SEasia owing to its geographic composition extending over
18° of latitude (Hong Kong to Chiang Mai) and 10° of
longitude (length of Thailand); mixing will be examined
below. india in subsequent mensural discussions consists
of only Indian specimens from the vicinity of the type locality. Both Sri Lankan bisexual specimens are adults, a
male (26.3 mm SVL) and a female (36.0 mm).
This disparity in size is striking. The Sri Lankan female lies within the range of other bisexual female samples
(Table 3). In contrast, the male (BMNH 1910.3.16.4; 26.3
mm SVL) is nearly the smallest adult in all bisexual samples. Only one each, Indian and sEasia males (27.2, 25.5
mm SVL, respectively), share small adult size. Even the
Palau and Philip males are larger, and these two populations average the smallest in adult size of all Hemiphyllodactylus populations. The small size is not a mistake in
the recognition of maturity, as this Sri Lankan male has
23 TotPore (16–24, range for Indian males). Although
not statistically dimorphic, Indian adult females average
slightly larger (35.3 mm SVL, range 33.1–37.9 mm, n = 6)
than males (33.5 mm SVL, range 27.2–36.9 mm, n = 8).
The proportions are nearly identical in the two sexes,
thereby reinforcing the absence of sexual dimorphism
within H. aurantiacus. The absence of morphometric
differentiation between the Indian and Sri Lankan specimens was evident in the high frequency of misclassiication of Indian males and females analyzed (separately) in
DF analysis (a synopsis of results is in Appendix 3). The
recently collected adult female and two males of H. aurantiacus (Bauer and Das, 1999) lie within the size ranges
noted above. These authors emphasized scalation and coloration; their indings will be discussed later.
sunda has 32 bisexual individuals (29 adults) amidst
the total sample (n = 83). This sample visually consists
of at least two bisexual taxa, and these bisexuals derive
from three geographic areas (adults from peninsular Malaysia, n = 17; Sumatra, 8; Borneo, 3; no precise locality,
1). These samples sizes are inadequate to address mensural variation in detail; however, owing to the existence of
three names (harterti, larutensis, margarethae), statistical
examination of the mensural data is necessary. The Malaysian sample consists of nine adult females (mean 48.3,
40.9–62.1 mm SVL) and eight adult males (47.2, 35.3–
56.9 mm); the smallest male (BMNH 1901.3.20.2) is the
holotype of Gehyra larutensis Boulenger. This BMNH
male is smaller than the other males (36.5–56.9 mm). The
smallest female (40.9 mm) is the holotype of Lepidodactylus Harterti Werner (ZMB 15360). A series of 17 adults
number 631
•
17
TABLE 3. Summary statistics on select mensural characters of adult females of the bisexual Hemiphyllodactylus samples. The values
are mean ± standard deviation (SD) and range of minimum to maximum. Organization as in Table 2, except sample sizes (n) are adult
females, adult males, and total sample, respectively. sunda was excluded because it was a mixed sample of two or more taxa. Character
abbreviations are deined in Appendix 1.
Sample (n)
Character and
statistic
SVL
Mean ± SD
Range
TrunkL
Mean ± SD
Range
HeadL
Mean ± SD
Range
SnEye
Mean ± SD
Range
SnW
Mean ± SD
Range
TrunkL/SVL
Mean ± SD
Range
HeadL/SVL
Mean ± SD
Range
HeadW/SVL
Mean ± SD
Range
HeadW/HeadL
Mean ± SD
Range
SnW/HeadL
Mean ± SD
Range
OrbD/NarEye
Mean ± SD
Range
PALAU
(11, 12, 24)
PHILIP
(19, 17, 36)
SEASIA a
(32, 22, 61)
CHINA
(17, 18, 38)
INDIA b
(6, 8, 17)
32.8 ± 1.01
31.1−34.2
35.1 ± 2.87
29.6−41.3
39.0 ± 4.55
31.9−50.5
43.4 ± 3.89
34.9−49.3
35.3 ± 2.22
33.1−37.9
17.8 ± 0.76
16.8−18.9
18.5 ± 1.87
15.5−22.4
18.5 ± 2.82
14.9−25.6
20.9 ± 2.89
16.1−26.5
17.4 ± 1.00
16.5–18.8
7.1 ± 0.26
6.6−7.5
7.7 ± 0.60
6.6−9.0
9.0 ± 1.09
7.4−12.1
10.1 ± 1.01
7.6−11.5
8.1 ± 0.36
7.7−8.7
2.8 ± 0.16
2.6−3.1
3.1 ± 0.34
2.6−3.7
3.7 ± 0.50
3.0−5.0
4.3 ± 0.55
3.0−5.2
3.2 ± 0.37
2.7−3.7
1.1 ± 0.11
1.0−1.3
1.3 ± 0.08
1.2−1.4
1.5 ± 0.33
1.0−2.3
1.7 ± 0.22
1.3−2.2
1.4 ± 0.06
1.3−1.4
0.54 ± 0.023
0.50−0.57
0.53 ± 0.019
0.49−0.57
0.47 ± 0.033
0.40−0.54
0.48 ± 0.030
0.43−0.55
0.49 ± 0.020
0.46−0.50
0.22 ± 0.006
0.20−0.22
0.22 ± 0.009
0.20−0.24
0.23 ± 0.013
0.21−0.27
0.23 ± 0.009
0.21−0.24
0.23 ± 0.009
0.21−0.24
0.13 ± 0.006
0.12−0.14
0.14 ± 0.011
0.13−0.17
0.16 ± 0.018
0.12−0.19
0.17 ± 0.017
0.14−0.20
0.16 ± 0.018
0.14−0.18
0.59 ± 0.028
0.55−0.64
0.65 ± 0.053
0.58−0.82
0.71 ± 0.060
0.55−0.80
0.72 ± 0.063
0.59−0.82
0.71 ± 0.078
0.58−0.79
0.15 ± 0.017
0.13−0.19
0.17 ± 0.011
0.15−0.19
0.17 ± 0.026
0.11−0.22
0.17 ± 0.012
0.15−0.19
0.17 ± 0.009
0.16−0.18
0.83 ± 0.049
0.74−0.90
0.83 ± 0.073
0.66−0.95
0.81 ± 0.075
0.70−1.00
0.78 ± 0.075
0.63−0.90
0.81 ± 0.067
0.71−0.91
a
The sEasia sample includes Hong Kong specimens because of their geographic proximity to the Southeast Asian specimens and their greater geographic
separation from the Yunnan specimens largely composing the China sample.
b
The india sample’s statistics derive exclusively from mainland Indian specimens.
from the Cameron Highlands and Fraser Hill contains
the largest adult Hemiphyllodactylus encountered in my
survey of museum collections, females ranging from 42.2
to 62.1 mm (mean ± SD, 50.2 mm ± 6.52; n = 8), males
from 36.5 to 56.9 mm (48.1 mm ± 6.44, n = 7). Even
though the means of the two sexes differ, there is no sexual
dimorphism. This latter sample from the central mountain
range (Titiwangsa) attracts attention because the named
Malaysian taxa (harterti and larutensis) derive from the
western mountain range (Bintang). An exploratory DFA
of the two Bintang specimens and the 15 Titiwangsa ones
yields 100% overall accuracy in the original classiication
18
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
but only 76% accuracy in the jackknifed one (synopsis of
results in Appendix 3). Classiication function assignments
indicate differences in head shape by identifying HeadL,
HeadW, NarEye, Eye D, and SnW as the major variables
in the classiication function. A DFA of mensural proportions yielded lower classiication accuracy, indicating no
differentiation value.
The Sumatran specimens are one adult male (38.8
mm) and seven adult females (41.2, 36.0–46.9 mm). This
sample includes the four type specimens of Hemiphyllodactylus margarethae Brongersma: two adult females
(42.0 mm, ISRN 9338B; 46.9 mm, ZMA 11096), a juvenile male (38.0 mm, IRSN 9338A), and an adult male
(38.8 mm, holotype ZMA 11095). The Bornean sample
consists of three adult males (33.2 mm, 27.8–37.3). The
inal sunda bisexual is an adult male (38.6 mm) of uncertain provenance in the East Indies.
sunda adults do not display sexual dimorphism in
any mensural or proportional trait. The preceding description of size differences in the three regional subsamples
suggests the possibility of size dimorphism in the Malaysian sample (i.e., females larger but not statistically signiicantly so) and regional differences (smaller adults in
Borneo, although the H. larutensis holotype is small also).
The sEasia sample shows dimorphism in only SVL
of the mensural and proportional traits. In compiling this
sample, I assumed that specimens from Thailand eastward
through Vietnam and South China represented a single
population or taxon, hence potentially recognizable as
H. chapaensis (Bourret). The initial analysis of morphometrics and the general similarity (means differ but strong
overlap of ranges occurs) of the China and sEasia samples (Table 3). indicated that my assumption was incorrect
and required testing. Another question on the uniformity
of the sEasia sample arises from the broad longitudinal
breadth (99°–115° E) of the sample, extending from Chiang Mai in the west to Hong Kong in the east. The latitudinal depth of the sample is much less, approximately 7°
(19°–12° N). The China sample (assumed to represent H.
yunnanensis) is broad but nonetheless not as geographically expansive as sEasia and limited to the “highlands”
(Shan and Yunnan plateaus).
To test the preceding assumption of homogeneity between and within these two samples, I examined
them in several ways: (1) variation within a combined
China-SEAsia sample and (2) the subdivision of these
two samples into various subsamples and comparison
of the subsamples. The combined China-sEasia sample
contains 48 adult females and 39 adult males. Student t
tests of all mensural and proportional traits show that
size dimorphism persists in most dimorphic traits (SVL,
TrunkL, HeadL, HeadW, SnEye, NarEye) identiied in the
China sample and additionally the proportions EyeD/
HeadL and EyeD/NarEye. Typically, the p values for the t
test are slightly higher in the combined sample (e.g., SVL
P = 0.008, 0.019, China versus combined). Even though
the SEAsia sample is the larger of the two, its absence of
dimorphism did not swamp the average size difference of
females larger than males.
The combination of these two samples, however, results in a near doubling or more of CV of most mensural
traits as compared with the China sample. This increased
variation relects a mixing of two or more phenotypes,
presumed here to represent distinct genetic entities. A
combination based on topography (China [contains only
Yunnan and Shan plateau specimens] and the highland
areas of northwestern Thailand [Tak to Mae Hong Song
and Chiang Rai]) did not alter combined sample variation
(i.e., CVs equivalent to those of China alone). The mean
SVL of this China-Thai sample decreased from 43.3 mm
(China) to 40.7 mm, but the mean values of all proportions of the two samples are less than 1% different, supporting the homogeneity of Hemiphyllodactylus from this
region. Addition of the northern Vietnam (Chapa area)
specimens to the China-Thai sample did not alter means
or CVs; similarly, the addition of Hong Kong specimens
did not alter the level of variation. These additions of a
few individuals to a large (n > 30) sample are, however,
unlikely to alter CV unless a striking disparity exists in the
added sample.
The realignment of China and sEasia specimens
suggests the existence of a northern (upland) population and a southern one (“lowland” Thailand only in
the present sample). The latter averages smaller than the
former (Table 4). The situation (afinities) of other Southeast Asian specimens is unclear owing to small samples
(Table 4). The Vietnam chapaensis sample averages larger
(SVL) than the four other samples, but its range is within
that of the China–NW Thailand sample. Its proportions
similarly match the latter sample’s proportions. The Laos
sample is the smallest (n = 2) of this comparison, hence
dificult to interpret, and a single male from Cambodia
is immature. Presently, I note only the low TrunkL/SVL
proportion representing a shorter trunk length than in
the other Oriental samples. Interestingly, they appear less
robust than the Hong Kong Hemiphyllodactylus, whose
general appearance matches the stouter habitus (Figure 3)
of large adult female H. yunnanensis; yet they have a longer trunk (53% TrunkL/SVL, Table 4) similar to the taxa
with the slender habitus. A DFA of Oriental adults using
number 631
•
19
TABLE 4. Summary statistics on select metric characteristics of adults of the bisexual Hemiphyllodactylus from southern Asia. The values are mean ± standard deviation (SD) and range of minimum to maximum. Realigned and restricted regional samples from the China
and sEasia samples: China–northwestern Thailand; Thailand, area exclusive of the northwest; Laos, Phong Saly; Vietnam, chapaensis
type locality; and Hong Kong. Sample sizes (n) are total adults, females, and males, respectively. Character abbreviations are deined in
Appendix 1.
Sample (n)
Character and
statistic
CHINA–NW Thailand
(61, 33, 28)
Thailand
(16, 9, 7)
Laos
(2, 1, 1)
Vietnam
(3, 3, 0)
Hong Kong
(4, 2, 2)
39.4 ± 4.68
25.5−49.3
35.1 ± 2.77
30.0–39.9
37.1 ± 0.78
36.5–37.6
46.2 ± 3.16
42.7–48.8
43.0 ± 6.13
35.3–50.3
0.47 ± 0.031
0.40–0.55
0.47 ± 0.036
0.42–0.52
0.42 ± 003
0.42–0.43
0.49 ± 0.011
0.48–0.51
0.53 ± 0.032
0.51–0.58
0.23 ± 0.010
0.21–0.26
0.23 ± 0.014
0.21–0.26
0.24 ± 0.005
0.23–0.24
0.23 ± 0.004
0.23–0.24
0.22 ± 0.007
0.21–0.23
0.17 ± 0.018
0.14–0.22
0.17 ± 0.017
0.14–0.20
0.17 ± 0.002
0.16–0.17
0.17 ± 0.002
0.17
0.16 ± 0.010
0.15–0.17
0.41 ± 0.025
0.34–0.46
0.41 ± 0.026
0.38–0.46
0.39 ± 0.024
0.37–0.42
0.42 ± 0.023
0.40–0.44
0.41 ± 0.0.37
0.38–0.46
0.17 ± 0.020
0.11–0.22
0.16 ± 0.025
0.12–0.20
0.21 ± 0.008
0.20–0.22
0.17 ± 0.002
0.16–0.17
0.16 ± 0.015
0.15–0.18
0.24 ± 0.036
0.15–0.36
0.22 ± 0.035
0.16–0.29
0.30 ± 0.001
0.30
0.24 ± 0.007
0.23–0.24
0.22 ± 0.014
0.21–0.24
SVL
Mean ± SD
Range
TrunkL/SVL
Mean ± SD
Range
HeadL/SVL
Mean ± SD
Range
HeadW/SVL
Mean ± SD
Range
SnEye/HeadL
Mean ± SD
Range
SnW/HeadL
Mean ± SD
Range
SnW/HeadW
Mean ± SD
Range
all 10 proportions yields a poor classiication (jackknifed)
of these subsamples; all were classiied at ≤50% (synopsis
of results in Appendix 3). Larger samples are essential to
address these confusing morphometrics.
Zhou et al. (1981) described three subspecies of H.
yunnanensis from China without reference to any other
species or populations of Hemiphyllodactylus. They had
good samples (≥20 individuals) for each subspecies, and
although they reported some measurements (Table 5), they
presented no statistical analysis and no mensural data for
the nominate subspecies. The westernmost subspecies (H.
y. longlingensis) is the smallest of three forms and shares
the range of my Yunnan sample. The two eastern forms
(H. y. dushanensis and H. y. jinpingensis) are distinctly
larger geckos than the western ones. Their adult SVLs
(Table 5) do not overlap with either H. y. longlingensis
or the Yunnan sample. TrunkL/SVLs of the east and west
forms also do not overlap, but surprisingly HeadL and
SnEye do. This overlap suggests that either these forms/
populations have proportionally smaller heads or there
was a lack of precision in recording these measurements.
All three of these subspecies show strong sexual dimorphism in SVL with only H. y. longlingensis showing a
slight overlap of the largest male and smallest female. The
absence of overlap suggests that the specimens examined
(and reported in the tables) underwent selection prior to
analysis because of the strong overlapping ranges of adult
females and males in my China sample, even though the
average SVLs are signiicantly different.
As an aside, one Thailand adult or near-adult Hemiphyllodactylus (BPBM 3502, Sakaerat) is hermaphroditic,
with large testes and a pair of vitellogenic follicles (maximum diameter 3.2 mm). All other unisexual and bisexual
specimens examined had gonads of only one sex.
20
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
TABLE 5. Comparison of some character measurements of adult females among the Chinese populations of Hemiphyllodactylus yunnanensis. Data for the subspeciic samples of H. yunnanensis were derived from Zhou et al. (1981: tbls. 2–4). Character abbreviations
are deined in Appendix 1; a dash (-) means no data were available.
Length (mm) range for sample or subspecies (n)
Character
Yunnan
(12)
yunnanensis a
(0)
longlingensis
(19)
jinpingensis
(10)
dushanensis
(10)
SVL
TrunkL
HeadL
37.2–48.8
17.4–23.4
8.8–11.4
-
39.0–46.0
20.0–23.5
9.0–10.0
49.0–53.5
24.0–28.5
10.5–11.0
48.0–51.0
25.0–27.0
10.0–11.0
3.7–5.1
-
-
4.0–5.0
10.0–12.0
13.0–16.5
4.5–5.0
11.0–12.0
16.0–17.5
5.0–5.5
11.0–12.0
16.0–17.5
SnEye
ForelimbL
HindlimbL
a
Despite a sample of 249 specimens, no measurements were reported by Zhou et al. (1981) in the description of H. y. yunnanensis nor was a table of
measurements provided. Measurements also were not included in the descriptions of the new subspecies.
Précis.
(1) Morphometry is only modestly useful in the differentiation of intra- and interregional samples
of the bisexual populations. (2) Palau, Philip, and China
display size dimorphism between adult females and males.
Females average signiicantly larger than males. (3) The
Palauan population has the smallest average SVL of all
populations. The Philippine and Indian-Sri Lankan populations are also small but average signiicantly larger than
the Palauan one. Average size of the China and sEasia
samples is signiicantly larger than the preceding three
samples. (4) The original composition of india, sEasia,
and sunda each probably includes representatives of at
least two taxa/populations. india contains an Indian component (H. aurantiacus) and a Sri Lankan one (unnamed).
Inclusion of northwestern Thailand specimens in the
China population does not increase variation, thus suggesting genetic homogeneity among these “highland” populations. The remainder of the sEasia sample potentially
remains a mixed sample. Presently, there are insuficient
museum specimens for a ine-scale geographic analysis.
This problem is also shared by sunda, which clearly has
multiple taxa therein.
Bisexual—Scalation
Sexual dimorphism in PreclPor and TotPore is shared
by all samples. Adult males have secretory pores; females
do not or do so uncommonly. Females with secreting
pores occur in all Asian samples. The sample india-india contains a single adult female (BMNH 74.4.1332, a
syntype of H. aurantiacus) with three PreclPor, about half
the number found in males (Table 6). China has three
females with pores, one with only seven precloacal pores,
another with nine precloacal pores separated from one
femoral pore on each side, and a inal individual with a
continuous precloacal–femoral series of 19 pores. sEasia
also has three pore-bearing females; the holotype of H. t.
chapaensis has nine PreclPor, and two Thai females have
continuous series of 14 and 15 pores. In sunda, pores
occur only in Sumatran females, two of seven individuals
(12 PreclPor in one and separate series of 9 and 20 in the
other). Generally, but not always, TotPore is distinctly less
in females than in males. No other scalation traits show
dimorphism in Palau, Philip, sEasia, China, or india.
The composite nature of sunda does not permit rigorous
testing for dimorphic traits.
Most scalation traits are fairly uniform across the
six bisexual samples (see Table 6, although Sunda not
included). I present medians because the data for scalation traits are discontinuous values and integers better
portray the actual number of scales than does a decimal
value. The uniformity among samples is most evident in
the overlapping ranges of minimum and maximum values.
Among the six samples, CircNa is identical (three scales)
in all samples. SnS ranges strongly overlap amidst the six
samples with sEasia individuals typically possessing only
two “internasal” scales, and Palau and india each with
four SnS. This latter condition and the three-SnS one represents a large “supranasal” above the naris on each side
separated medially by two or one small scales. Inlab are
similar with most samples having ten scales, Palau with
nine and india-India with 11. Most samples have 10 Suplab, with 8 for Palau and 11 for sunda. Although these
differences (for Suplab, but other traits as well) are slight
number 631
•
21
TABLE 6. Summary statistics on select coloration and scalation characters of juveniles and adults of the bisexual Hemiphyllodactylus
samples. Organization as in Tables 2–4. Sample sizes (n) are juveniles, adults, and total sample, respectively. Statistical values are either
median ± standard deviation (SD) and range of minimum to maximum (for coloration and scalation characters) or modes and frequency
(%) of occurrence (for inger and toe lamellae). Character abbreviations are deined in Appendix 1.
Sample (n)
Character and
statistic
PALAU
(11, 12, 24)
PHILIP
(15, 19, 35)
SEASIA a
(32, 22, 61)
CHINA
(17, 18, 38)
INDIA b
(6, 8, 17)
PostocS
Median ± SD
4 ± 0.88
3.5 ± 1.18
1 ± 1.38
2 ± 1.41
0 ± 2.17
2−6
0−6
0−4
0−5
0−7
4 ± 0.65
3−5
3 ± 0.48
2−4
2 ± 0.68
1−4
3 ± 0.72
2−5
4 ± 0.89
3−6
8 ± 0.78
8−11
10 ± 1.15
8−13
10 ± 0.91
8−12
10 ± 0.70
9−12
10 ± 0.89
10−13
11 ± 0.90
9−12
11 ± 1.52
8−14
8 ± 1.59
6−18
8 ± 1.38
6−11
11 ± 1.12
10−14
15 ± 1.44
11−18
15 ± 1.50
12−18
14 ± 1.78
9−18
13.5 ± 1.80
12−17
13 ± 1.87
11−17
2 ± 1.01
1−4
2 ± 1.89
0−3
1 ± 0.69
0−4
1 ± 0.38
1−2
2 ± 0.63
1−3
Median ± SD
Range
4ToeLm
Median ± SD
Range
FingerLm d
22.5 ± 8.15
16−28
27 ± 5.35
17−38
21.5 ± 4.28
11−26
19 ± 3.31
11−24
21.5 ± 2.91
16−25
4 ± 0.38
4−5
4 ± 0.41
4−5
3 ± 0.80
3−5
4 ± 0.42
3−5
3 ± 0.46
2−3
Modal values
Frequency
ToeLm d
3-4-4-3
55.0%
3-3-3-3
65.7%
3-3-3-3
72.2%
3-3-3-3
36.8%
2-2-2-2
100%
Modal values
Frequency
3-4-4-4
45.8%
3-4-4-4
50.0%
3-3-3-3
50.1%
3-4-4-4
57.9%
2-2-3-2
37.5%
Range
SnS
Median ± SD
Range
Suplab
Median ± SD
Range
Chin
Median ± SD
Range
Dorsal
Median ± SD
Range
CloacS
Median ± SD
Range
TotPore c
a
The original composition of China and sEasia samples is retained for consistency of comparison with Table 4 and mensural results described and
discussed earlier in the text and tables.
b
The values are only for the mainland India portion of the India sample.
c
These values are only for the adult male portion of each sample.
d
Lamellae formulae represent the most frequent formula (mode) for each sample and the percent of the sample with this formula (frequency).
and usually not statistically signiicant, I suggest the differences relect genetic differentiation among the regional
populations and are not the result of sampling or measurement error. This proposition derives from the relatively
low variation (CV) observed for most scalation traits
(see the earlier Baseline Estimates section). The re-aligned
China–NW Thai sample usually has the same medians as
China and similar standard deviations
Chin ranges strongly overlap; however, the low medians (8) of sunda, sEasia, and China result from six or
seven scales in most individuals and the higher medians
of Palau, Philip, and india from most individuals having
22
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
FIGURE 7. Scale morphology of the chin in various populations
of Hemiphyllodactylus: (A) H. typus USNM 570742 Hawaii; (B)
H. ganoklonis USNM 563682 holotype, Palau; (C) H. margarethae
ZMA 11095 holotype, Sumatra (redrawn from Brongersma 1931:
ig. 4); (D) H. y. yunnanensis China (from Zhou et al., 1981: ig.4
left); (E) H. y. longlingensis China (from Zhou et al., 1981: ig.4
right); (F) H. chapaensis NMHN 1948.43 holotype, Vietnam (redrawn from Bourret, 1937: ig. 1b).
10 or more chin scales. The difference in total number of
scales in the chin scale arc results from the enlargement
of the two median scales posteriorly bordering the mental
(i.e., postmentals). Postmental shape and size are variable
within each Asian sample; nonetheless differences in postmental size serve to delimit southern Asian populations
and denote populational differentiation.
Beginning in the north with populations identiied as
H. yunnanensis (China), each postmental is one-third to
half the area of the irst supralabial (Figure 7D,E); the laterally abutting chin scale is commonly one-third to twothirds the area of the postmental and distinctly larger than
chin scales behind and beside it. This anteriormost row,
including the postmental, is an arch of larger scales, and
posteriorly the chin scales quickly decrease in size to granular scales. In China, the mental is modestly large and
roughly an equilateral triangle to a lattened pentagonal
shape. In SEAsia, the chin scale condition is as described
above, although I have the impression that the next posterior scale row to the “large” scale arc has proportionately
larger scales than my China specimens and then a rapid
diminution to granular chin scales. I note no substantial
difference from the preceding pattern for individuals from
Hong Kong to northwestern Thailand.
Interestingly, my examination of the holotype of H.
chapaensis revealed a low Chin (seven scales), hence a
modestly enlarged postmental, but Bourret’s (1937:60)
description noted the postmentals as “une paire de très
petites plaques.” His illustration (reproduced in Figure
7F) shows an arc of enlarged chin scales with postmentals not much larger than the posterior chin scales. This
morphology contrasts to Hong Kong H. chapaensis (so
taxonomically labeled by Lazell, 2002), which possess
large postmentals followed by a second arc of enlarged
(although not as greatly) chin scales. A southern Vietnamese specimen (USNM 146161) has very large, rectangular
postmentals similar to those depicted for H. margarethae
(Figure 7C). The Cambodian male (FMNH 270569) has
an arc of nearly equal-sized scales (Chin = 10). All sunda
specimens have large, rectangular postmentals. For most
and the above-mentioned Vietnamese specimen, the postmentals broadly contact medially, and posteriorly chin
scales rapidly decrease in size (Figure 7).
Dorsal and Ventral, each, have overlapping ranges
among the six samples. The pattern of variation between
these two traits is different. Note that the number of scales
for each derives from the EyeD distance; thus the values
represent the number of scales within the same area for
each specimen, although there may be an increase in variance within a sample owing to the manual measurement
of EyeD. The CV is 12–15% and in the same range as
Chin and most inger and toe lamellae counts. Dorsal values have identical (or nearly so) medians (14% and 15%)
for all samples except india-India, which is distinctly
lower (12%; Table 6). For Ventral, the medians are more
variable. sunda, sEasia, and China have eight scales,
Palau and india 10, and Philip 11. The Sundan to Chinese populations have proportionally larger ventral scales
compared both to dorsal scales within individuals and to
the ventral scales of Paciic and Indian specimens.
CloacS ranges from absent to four spurs in the total
bisexual sample (Figure 8). Part of this variation is due to
sexual dimorphism, with females tending to have fewer
or no spurs (statistically signiicant only in Palau and
Philip); nonetheless, there are populational differences,
with sEasia and China individuals averaging a single
number 631
FIGURE 8. Cloacal spur morphology in Hemiphyllodactylus:
(A) Pinted state, H. yunnanensis, Thailand (USNM 310807); (B)
rounded stated, H. typus, Hawaii (USNM 279240). Illustration by
J. Kilby.
spur. This average or median value is not driven by a larger
number of females in the samples, as the number of adult
females nearly equals the number of males in all samples.
In most populations the median row of ventral scales
on the tail (Subcaud) equal the size of adjacent scales. Only
in a few individuals and only in the sunda, SEAsia, and
india samples are Subcaud slightly enlarged over adjacent
scales. No Hemiphyllodactylus has enlarged subcaudal
scales like those of many Gehyra species.
Finger and toe lamellae counts (Table 6; Figure 9) are
relatively uniform among all samples except for india.
Individuals of the latter population have distinctly fewer
enlarged U-shaped subdigital lamellae on the fore- and
hindfeet. This difference is strikingly apparent; the foreand hindfoot lamellae formulae (second through ifth digit
of each foot) of india is 2-2-2-2, 2-2-3-2, contrasting to
the modal 3-4-4-3, 3-4-4-4 of the other Asian samples.
The modality of these formulae, however, is 3-4-4-4 and
4-5-5-4 in the Vietnamese, Hong Kong, and Cambodian
specimens and 3-4-4-4 and 4-5-5-5 for the Sunda Titiwangsa specimens.
As with the mensural characters, it is necessary to
comment on the Zhou et al. (1981) study of Chinese
•
23
FIGURE 9. Digital lamellae morphology in select species of
Hemiphyllodactylus: (A) Right hindfoot, H. yunnanensis, Thailand (USNM 310807); (B) right forefoot, H. yunnanensis China
(BMNH 1904.11.29.10D); (C) right hindfoot, H. typus Hawaii
(USNM 27924); (D) right hindfoot, H. yunnanensis China (BMNH
1904.11.29.10D). Illustrations by J. Kilby.
Hemiphyllodactylus populations. Their scalation data examined seven traits: rostral notched posteromedially, enlarged scale posteriorly bordering “supranasal,” Suplab,
Inlab, chin scale bordering mental posteriorly (postmentals)], digital lamellae formulae, and precloacal–femoral
pores. The pore condition is examined later. I did not score/
record either of the irst two traits. My initial research in
unisexual Hemiphyllodactylus populations indicated that
a rostral notch was always present although of variable
development (i.e., indistinctly notched to cleft one-third
length [rarely] of rostral scale); hence this trait was not included in my character set. The variation in the relative size
of the post “supranasal” scale completely escaped my attention. Zhou et al. used the relative size of the postmental
scales, in part, to differentiate H. y. yunnanensis and H.
y. longlingensis. I have not seen specimens from the range
of H. y. longlingensis; however, four Hemiphyllodactylus
(USNM 570732–570735) from the western edge of the
Shan Plateau (Myanmar) match the enlarged second arc
of chin scales shown for H. y. longlingensis (Figure 7) with
the exception that in the four Burmese specimens, the chin
scales are more uniform in size and the median one is no
24
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
larger than its laterally abutting scales in the arc. Because
my China sample derives largely from Kunming, my results match the Zhou et al. concept of H. y .yunnanensis.
My results for Suplab and Inlab similarly match their data.
Their data for digital lamellae, however, contrast
sharply to mine. Foremost is low variation in their counts
compared with mine. Deciding which lamellae is the proximalmost one on each pad can be challenging owing to a
gradual reduction in size proximally. My protocol to increase the consistency (precision) of my counts is/was to
count only the distinctly U-shaped lamellae. Comparison
of my China formulae to their H. y. yunnanensis formulae
yields ive forefoot formulae (2-3-3-3, 3-3-3-3, 3-4-4-3,
3-4-3-3, and 3-3-4-3; the second and third formulae represent 37 and 21%, respectively, of the sample) versus
two formulae in Zhou et al. (3-4-4-3, 3-4-4-4; they do not
present the frequency of occurrence of each), I recorded
nine different hindfoot formulae (ranging from 3-3-3-3 to
4-5-5-5) with 3-4-4-4 in 60% of the sample, occurrence
of others is ≤10%, usually much less, versus an invariant 3-4-4-4 reported by Zhou et al. I, obviously, believe
that my data more accurately relect the natural variation
of the Yunnan population. My data cause me to question
the discriminatory potential that Zhou et al. attributed
to these formulae for the differentiation of populations
within their broader China sample. Digital formulae do
discriminate populations (e.g., compare india to the other
ive bisexual samples [Table 6]), but the validity of the
Zhou et al. data requires an independent study.
Precloacal–femoral pore morphology is similar in medians (19–22.5) and range of (11–28) of TotPore (Table 6)
for Palau, sEasia, China, and india. Philip has a distinctly higher median TotPore and a partially overlapping
range, although extending well beyond the upper limits
of the four former samples. The medians and ranges of
PreclPor for the preceding ive samples are more similar
to one another (median range 7.5–11), presumably owing
to all geckos having proportionally the same pubic or escutcheon area for pore development. Higher or lower TotPore counts derive mainly from an increase or decrease in
the length of the femoral pore series.
Number of pores, however, disguises two pore morphologies (Figure 10): (1) continuous series of precloacal
and femoral pores and (2) separate longitudinal patch of
precloacal and femoral pores. The continuous pore morphology occurs in all males of SEAsia and China samples
and the separate-patch morphology in all males of Palau,
Philip, and india (including a single Sri Lankan male)
samples. The relative size of the pore scales appears identical in the two morphologies. In some sEasia males (e.g.,
USNM 146161, southern Vietnam), the pore scales are
separated by one (usually) or two smaller (one-third or
less pore scale area) scales. From the low frequency of this
condition and the very few specimens from southern Vietnam through southern Thailand, it is uncertain whether
this condition is a variant or populational state.
Pore morphology in sunda shows regional differentiation with the occurrence of both continuous-patch and
FIGURE 10. Schemata of precloacal–femoral pore morphology of Hemiphyllodactylus: (A) Separate pore series and
(B) continuous pore series. (A modiied from and B reproduced from Taylor, 1918: ig. 4; H. insularis.)
number 631
separate-patch states, each state apparently occurring at a
speciic location. Even though sample sizes of the various
islands or sites are small, the differences among localities
are suggestive of genetic differentiation of populations.
Sumatran and Bornean adult males (n = 1, 3, respectively;
median TotPore 26, 29) have the separate-patch state; further as noted in discussion of sexual dimorphism, some
adult females from these two islands possess precloacal
pores and occasionally femoral pores. Malaysian males
have continuous precloacal–femoral series with a lower
median TotPore (21, 17–39; Titiwangsa specimens only,
n = 7). The presence of these two states in SEAsia is nomenclaturally important because of three available names.
The H. margarethae holotype has the separate-patch state,
the male G. larutensis holotype a continuous state, and the
female L. Harterti holotype lacks pores.
Précis.
(1) Sexual dimorphism of scalation
occurs only in the presences (males) and absences (females) of precloacal and femoral pores. (2) Most scalation traits are uniform throughout the Asian and Paciic
Rim populations. Difference among samples is evident in
the characters of Chin, Dorsal, Ventral, digital lamellae,
and precloacal–femoral pore. (3) Chin shows two states:
enlarged chin scales (postmentals) in sunda, sEasia, and
China and no postmentals in Palau, Philip, and india.
(4) india has large dorsal trunk scales, only slightly
smaller than the ventral ones; all other samples display a
greater size difference between dorsal and ventral scales
and typically more scales per unit area. (5) Digital lamellae
formula is strikingly lower in india than in the other ive
regional samples, and the formulae in the latter samples
are the same. (6) Pore morphology differs with China,
sEasia, and some sunda populations with continuous series of precloacal–femoral pores and Palau, Philip, some
sunda, and india with separate patches of pores.
Bisexual—Coloration
The drab, presumably cryptic, coloration of Hemiphyllodactylus makes characterization of coloration dificult. I recorded two external coloration characters:
PostocS and OrbStrp. The results show regional differences. For example, Palau and Philip have higher median
PostocS (Table 6) in contrast to the median of one or two
postocular spots for SEAsia, China, and india; however,
the ranges for these ive localities are the same. The “identical” ranges relect the dificulties of seeing these spots
in older or poorly preserved specimens. If an actual difference exists, it will require documentation with living
specimens or recently preserved ones. OrbStrp is less likely
•
25
to disappear with the age of a specimen owing to its dark
pigmentation; nonetheless, old specimens fade to unicolor
and this trait can be lost also. My data (Table 5) suggest
the near universal presence of OrbStrp in bisexual specimens, at least to the level of the ear.
With my attempt at quantiication of coloration illfounded for museum specimens, I offer a comparative
descriptive approach using the major pattern features
identiied in the H. typus coloration section. These descriptive data derive from my notes on museum specimens, descriptions and illustrations from herpetological
literature, and ield notes of mine and others sources are
noted in brackets.
Coloration is variable locally within populations and
broadly throughout the distribution of bisexual populations. Dorsal ground color for all populations is variously
described as tan to grayish brown. The intensity of the
“brown” background depends on the color phase, light
or dark, assumed at death and preservation or when photographed. The phase also affects the conspicuousness
of markings, whether they are light (usually off-shade
whites to tans) or dark (brown to nearly black-brown).
The venter also lightens and darkens, hence ranges from
near-white to dusky; because each scale normally shows
tiny spots (melanophores), the venter is never white. The
following descriptions will emphasize ive areas: (1) loreal
or snout area, (2) side of head and neck (ignoring the lightening resulting from accumulation of calcium carbonate
in the endolymphatic system), (3) dorsal and lateral trunk,
(4) sacrum and base of tail, and (5) sides and top of tail.
Several traits are shared by most individuals of the
bisexual populations. Intensity and color vary from population to population. A dark brown lateral stripe extending from the loreal area to at least the anterior quarter
of the neck occurs widely, although its thickness and loreal development are variable. For example, in a series
of northwestern Thai specimens, the loreal stripe ranges
from a dark spot immediately in front of the eye through a
well-deined stripe from naris to eye to a fully dark brown
loreal area. Similarly behind the eye, the lateral stripe
(OrbStrp) ranges from narrow to broad (always above the
ear opening), terminating at head–neck juncture or anywhere from this juncture to the shoulder. A dark lateral
stripe is variously evident on the trunk, and if present, it
is typically moderately broad but irregularly edged above
and below, occasionally bordered above by a broad lighter
area. Often this stripe breaks into a series of dark blotches.
Dorsally, the head is unicolor or nearly so with a faded
and diffuse mottling. The back is similar, ranging from
unicolor through small, paired parasagittal dark spots to
26
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
broad, irregular dark brown bars (continuous across the
midline or interrupted). Most individuals display a signature mark posteriorly from the sacrum onto base of the
tail. This mark contains a (supracloacal) middorsal dark
blotch bordered posteriorly by a (supravent) light mark
that can have anterior extension on each side of the middorsal dark blotch with the anterior arm usually ending
at the top of the ilium, hence a U-shaped light mark. This
contrasting dark and light postsacral mark is often the
brightest area of coloring of a Hemiphyllodactylus. Posteriorly the top and sides of the tail range from light to medium brown, nearly unicolor to distinctly banded. In life,
the underside of the tail is variously reported as shades of
red or orange.
Palau Hemiphyllodactylus are generally drab lizards
with the exception of the bright postsacral signature mark.
Ground color is light gray to tan. The dorsal pattern ranges
from a sparse speckling of dark brown spots to small, irregular V- or trident-shaped bars. A dorsolateral row of
widely spaced pinkish yellow spots extends from the rear
of the head to the postsacral mark; the spots on the posterior third of the trunk are the most prominent ones. The
postsacral mark has a small dark middorsal spot (roughly
triangular) encompassed by a broad-based U-shaped light
pink mark, which is bordered laterally on each side by a
diffuse dark dorsolateral stripe (Figure 11). The tail ranges
from yellow to tan with diffuse midline dark marks and
narrow light pink transverse bands. In the loreal region, a
preorbital stripe is always present but generally only moderately developed, ranging from a dark preorbital spot to
a narrow stripe extending midway to the naris. The postorbital stripe is narrow, always fragmented, and often
only with a few fragments on the side of head, somewhat
more extensive on the neck to the shoulder. (Coloration
is reported from Ronald I. Crombie’s [RIC] unpublished
1993–2002 ield data notes and my personal observation
of color in RIC’s images and preserved specimens.)
Philip Hemiphyllodactylus are similar to Palau ones
with light to medium brown backgrounds. The dorsal
pattern ranges from a faint dark lecking or reticulation
through paired, parasagittal dark longitudinal dashes to
small dark blotches. Dorsolaterally, a row of brick red,
dark-edged spots occurs on each side from behind eye to
and onto tail. The postsacral mark has a brown V-shaped
anterior border to the broader red U-shaped portion; the
latter is usually strongly edged in brown laterally and posteriorly. Tail is lightly colored, presumably tannish and
commonly with small dark paired spots to tip. A lateral
stripe, variously fragmented, extends from the loreal area
to anterior neck; the preorbital portion is usually sharply
deined (Taylor, 1918; Brown and Alcala, 1978; G. R. Zug
[GRZ], personal observation of museum specimens).
China and sEasia geckos tend to have darker backgrounds than individuals from the Paciic Rim samples.
Although Asian geckos have a light phase, the background
in this phase is gray to brownish gray. The overall impression of a darker background is heightened by more
extensive dorsal and lateral markings in many individuals
(Figure 11). Dorsally, the neck and trunk vary from faint
reticulations of medium brown through diffuse wavy and
brown fragmented crossbars or chevrons to broad, dark
brown transverse blotches (usually paired); a moderate
pattern seems to be the average condition. Dorsolaterally,
a row of light tan to nearly white spots occurs on each
side from behind the head to and seemingly forming the
anterolateral arms of the postsacral mark. This mark is
variable in contrast intensity with the anteromedial dark
brown portion, ranging from a large triangular blotch
to a moderate transverse bar. The lighter portion of the
postsacral mark is typically shades of cream to light yellow. The postsacral mark is broad and deep and extends
well on to the tail base; the anterolateral arms are usually weakly developed and short (inconspicuous). The tail
is strikingly lighter than trunk with some dark transverse
spots or bars (Boulenger, 1903; Bourret, 1937; GRZ, pers.
obs. of color images and specimens).
Hong Kong is a geographic outlier from other specimens comprising the China and SEAsia samples. Their
coloration differs from the preceding in two major ways.
First, the dorsolateral spotting is commonly so faded that
it is nearly invisible. The immediate postsacral area is light
brown and a median dark edge or bar lies above the vent,
and the lighter area is the irst band of the diffuse light and
dark tail banding (Chan et al., 2006; GRZ, pers. obs. of
specimens).
india H. aurantiacus is a boldly colored gecko of
dark brown bars and blotches on a light to medium brown
background (Figure 11A). A dorsal pattern of dark, narrow, wavy crossbars begins on the shoulders and continues to the sacral area; these crossbars can be variously
fragmented and/or narrowly connected to anterior or posterior crossbars. A dorsolateral series of light (presumably
white to light tan) spots extend from the neck to the inguen; these spots are not evident in all individuals. The tail
base pattern is similar to one described above for Hong
Kong Hemiphyllodactylus. The remainder of the tail is
distinctly banded in brown and dark brown; the lighter
bands are about twice the width of the dark ones. The
loreal area usually has a dark stripe from naris to eye. A
dark supraorbitial stripe extends from the rear edge of the
number 631
•
27
FIGURE 11. Coloration of select Hemiphyllodactylus taxa: (A) H. aurantiacus (India, Yercaud; AMBauer 5749) (photograph by I. Das);
(B), H. ganoklonis (Palau; no ield/museum number) (photograph by R. I. Crombie); (C) H. harterti (Malaysia, Bukit Larut; KUZ 21264)
(photograph by H. Ota); (D) H. typus (Tonga, ′Eua; USNM 322119) (photograph by G. Zug); (E) H. typus (Fiji, Viti Levu; USNM 267978)
(photograph by G. Zug); (F) H. titiwangsaensis (Malaysia, Cameron Highlands, Tanah Rata; KUZ, no ield/museum number) (photograph by
H. Ota); (G) H. yunnanensis (Thailand, Loei Province; no ield/museum number) (photograph by P.-P. van Dijk); (H) H. yunnanensis (Myanmar, Pyin Oo Lin; USNM 570734) (photograph by G. Zug).
eye to the anterior trunk. A dark postorbital stripe extends
from the eye to the axilla. Both of these stripes are usually unfragmented (Beddome, 1870; Bauer and Das, 1999;
GRZ, pers. obs of specimens).
Coloration for the sunda gecko “species” cannot be
unequivocally resolved with the data at hand. My notes
and Boulenger’s (1900) description of the H. larutensis
type are minimal. Its ground color was originally grayish brown, although now (in preserved state) it is faded
to light tan. A paired series of small dark spots extend
from the neck onto the base of the tail. A few dark spots
occur on the sides of the trunk. A dark loreal and a postorbital stripe are present. My notes do not indicate the
presence or absence of a dorsolateral series of light spots
28
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
or a postsacral mark. Werner’s (1900:196) description of
H. harterti reports “gleicht in Färbung and Zeichnung
ganz den L. lugubris.” I interpret this coloration as a
medium brown background with dark wavy crossbars.
Photographs (Figure 12) of the type show it faded to near
unicolor tan. Brongersma (1931) did not include a description of coloration for H. margarethae. Although I
examined his type series, I have no notes on their coloration. Presumably, all four type specimens are faded.
That is certainly the condition of the two ZMA syntypes
photographed (Figure 12).
Excellent color illustrations of Sundan Hemiphyllodactylus are available in some recent ield guides. Additionally, color photographs were provided by colleagues
and an examination of recently preserved specimens from
the Malaysian Cameron Highlands. These images and
observations allow me to conclude that the three named
populations have different colorations.
Manthey and Grossmann (1997, hereinafter M&G)
depicted a H. larutensis (ig. 172; adult male, Cameron
Highlands, Palang, western Malaysia) and a H. typus (ig.
173; adult male, Berastagi, northern Sumatra) (In this paragraph, I am using the species identiications provided by
the authors for their illustrated specimens.) The same individual (my assumption) of H. typus is portrayed in the Cox
et al. (1998:85) ield guide and in Malkmus et al. (2002: ig.
275). A different individual of M&G’s H. typus is shown in
Chan-ard’s et al. (1999:130) (photographic checklist; adult
female, Bukit Larut, Perak, Malaysia); another H. typus
(Chan-ard et al., 1999:130) (adult male, KhaoYai Natl.
Park, Nakhon Rachasima, Thailand) is darker. Chan-ard
et al. (1999:128–130) also included six images (Cameron
Highlands, Malaysia) of H. larutensis (Chan-ard’s species
identiication); the male (Chan-ard et al., 1999:130) is the
same individual of H. larutensis as presented in M&G’s ield
guide. Chan-ard et al. (1999:128) also provided an image
of H. harterti (adult male, Cameron Highlands). Without
specimens available for examination, I cannot conirm the
identiication of the depicted specimens; nonetheless, a few
comments on their speciic identiication and coloration
seem useful although speculative.
First, the Berastagi “H. typus” does not have a typical H. typus coloration, and, indeed, it is quite striking
from other populations described above. Because it lacks
the dark dorsal crossbars and the dorsolateral series of
light spots, I do not believe the specimen was a H. typus.
Further, if it was really an adult male, it could not be a
H. typus, because H. typus is a unisexual species. Could
it be a H. margarethae? Possibly, because it derived from
the same mountain range as the latter, although about 3°
latitude northward of the Fort de Koch (now Bukittinggi)
FIGURE 12. Types of Bingtang slender gecko. (A–C) Lepidodactylus Harterti Werner 1900 (ZMB 15360): (A) dorsal view of whole
body, (B) ventral view of throat and chin, and (C) ventral view of
pelvic area (photographs by M.-O. Rödel); (D–F) Gehyra larutensis Boulenger 1900 (BMNH 1901.3.20.2): (D) dorsal view of whole
body, (E) ventral view of throat and chin, and (F) ventral view of
pelvic area (photographs by G. Zug).
type locality. The two Chan-ard igured H. typus possess
the near uniform mid-dorsa of the Berastagi one, and the
Bukit Larut one has the same overall coloration, although
darker, including the continuous dark dorsolateral lateral
number 631
stripes and bright anterolateral arms of the postsacral
mark of the Berastagi gecko. Chan-ard’s Thai H. typus
has a dorsolateral series of small light spots and well-developed postorbital stripe; nevertheless, I do not believe it
is a H. typus, owing to light anterolateral arms of a postsacral mark. It is labeled as an adult male and appears to
have a hemipenial bulge on the base of the tail. The images
labeled H. harterti and H. larutensis represent individuals from the Cameron Highlands, Malaysia. They do not
have the typus-style coloration and might represent one or
two bisexual species. Certainly, the pattern of several of
the specimens is like Lepidodactylus lugubris, as noted in
Werner’s Lepidodactylus Harterti description.
Using these images, recent specimens, and colleagues’
images, I propose that the coloration of three Sundan
populations is distinct for each and offer the following
hypotheses. H. margarethae is a lightly marked gecko
with a striking contrast between body and tail color. The
break between the darker trunk and head coloration is at
the postsacral “mark.” Dorsally, the head and body are
a light brown with faded darker brown dorsolateral and
lateral stripes. The nape and neck can have a brown middorsal stripe. A light canthal stripe extends from the naris
to eye and across the temporal area. The dark postsacral
mark is small to absent, but the light arms extend on the
posterior trunk. These beige colored arms are continuous with a nearly unicolor tail. (The preceding description
was based on Manthey and Grossman, 1997: ig. 173.)
Because Larut Hills (Bukit Bintang range) and Cameron
Highlands (including Fraser’s Hill; both of Banjaran Titiwangsa range) specimens do not share a common coloration in any images available to me, I propose them as
separate genetic entities; the former equals H. harterti
and H. larutensis, and the latter represents an unnamed
species. Hemiphyllodactylus harterti has contrasting
body–tail coloration as in H. margarethae but is overall darker. The center of the postsacral mark is a narrow
dark brown V, which is continuous with a dark brown
dorsolateral stripe on each side of the trunk, extending
anteriorly to mid neck. The light colored arms (border)
extend to above the hips and posteriorly rapidly fade into
the unicolor olive of the tail. Light canthal stripes are not
evident. (The preceding coloration was based on Chanard et al. (1999:130) Bukit Larut “typus.”) The Titiwangsa gecko is strongly marked. Its background ranges from
light tan to olive brown, overlain by a dense dark brown
reticulation. This reticulation lacks the regularity of the
dark ladder pattern of Lepidodactylus lugubris, although
it is reminiscent of that pattern. A narrow dark postsacral
V with light-colored (to over hips) is usually present. Tail
background color generally matches the light color of the
•
29
postsacral V; additionally, the tail bears light and dark
transverse bars, commonly equal sized. A dorsolateral
dark stripe extends from mid-loreal area to axilla. One
coloration variant is scattered dark blotches dorsally and
laterally on neck and trunk.
Précis.
(1) Although generally drab, bisexual
Hemiphyllodactylus display some bright markings that
seemingly are unique for different populations. (2) Colorful postsacral marks occur in Palau and Philip specimens.
The postsacral marks are distinctly U- or V-shaped in the
two preceding samples, whereas the dorsolateral arms
are weakly developed in most individuals of sEasia and
China. The postsacral mark is the beginning of a lighter
(than trunk) tail in Sumatran and Larut Hills individuals
and less distinct in Titiwangsa specimens. The postsacral
mark for india and Hong Kong geckos is the irst contrasting band of the tail.
GEOGRAPHY AND TAXONOMY
REGIONAL PATTERNS OF MORPHOLOGY AND SPECIATION
General Observations
The present study does not address the generic status of the various species currently assigned to the genus
Hemiphyllodactylus. I recognize the presence of two adult
body types among the species currently assigned to this
genus (e.g., attenuate body in Hemiphyllodactylus typus
and robust body in H. yunnanensis). I, however, weigh the
lamellar morphology of the fore- and hindfeet much more
strongly and am comfortable with Hemiphyllodactylus as
a small monophyletic clade, at least until a robust molecular data set demonstrates otherwise.
Morphological Differentiation
Several morphological features allow the recognition
of unique populations (samples) and sets of populations.
Foremost among these features is the absence of males
in the Hemiphyllodactylus populations that are the most
geographically widespread. Another striking trait is the
presence or absence of gonadal, peritoneal, and caecal
pigmentation among both unisexual and bisexual populations. Two body forms are displayed among populations,
at least in the older and/or mature individuals. Further, average adult SVL differs among populations, even though
size at maturation appears similar among all populations.
Three features of scalation show populational differences. Chin scales in contact with mental and anterior
30
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
infralabial scales show two patterns: small (subequal) or
slightly larger than the posteriorly adjacent chin scales; or
distinctly enlarged with median pair largest (= postmental scales). Secreting precloacal and femoral pores occur
consistently in adult females of some Hemiphyllodactylus
populations and rarely or not at all in other populations.
Also, two precloacal–femoral pore arrangements exist
(each state uniform within a population): continuous pore
series or discontinuous with precloacal pores separated
from left and right femoral pores series. In all populations the shape of the digital lamellae is distinctly U- or
lyre-shaped. These U-shaped lamellae match Russell’s
morphological deinition (Bergmann and Russell, 2003)
of scansors, and they distinctly show a dense microvillous
surface as the pad surface dries. The number of lamellae
on the digits and hence the digital lamellae formulae varies
among populations.
Although these geckos are mostly dull in coloration
and this somberness is accentuated by their small size,
color and pattern also offer evidence of populational differentiation. The overall dorsal pattern of dark blotches
and stripes ranges from nearly nonexistent or diffuse to
bold stripes and bars. Intensity and color of the dorsolateral spots varies among populations, as does the presence
and coloration of the postsacral mark.
Beginning with reproductive biology, unisexuality is
universal among most Paciic-area populations and coastal
ones from southern Melanesia through Sunda and Southeast
Asia to Sri Lanka. These populations and the Mascarene
ones represent the nominate H. typus. All these populations
share a densely pigmented intestinal caecum, which is frequently visible through the body wall and the pigmentation
of the peritoneum covering the gonads and gonadal ducts
(speciically the oviducts and epididymides). This pigmentation is not unique to the unisexual populations but also
occurs in bisexual ones: Palau, Philip, and india.
I propose that this unique visceral pigmentation (Figure 2) relects close phylogenetic relatedness among the
unisexual H. typus and the three bisexual species (justiication of the speciic status for each Palau, Philip, and
india below). I further propose that H. typus arose from
the hybridization of individuals from the Philippines and
Palau. An electrophoretic study (T. D. Schwaner, unpublished data, 1990) examined individuals from northern
Thailand (n = 13), Philippines (3), Fiji (5), and Hawaii
(9) using nine isozymic loci. The Philippine and two Paciic samples share allozymes for all nine loci; although
for GPD and IDH, the Philippine sample has two alleles
for each of these two isozymes and MPI of the Paciic
samples has two alleles. The Thai sample shares alleles
for only three isozyme loci for the Philippine and Paciic
samples. The similarity of shared alleles supports the Philippine species as one of the parental species of H. typus.
Similarities in morphology of Palau and Philippine individuals suggest the former as the other parental species.
When and where the hybridization occurred is far more
speculative, although a few reasonable conjectures are
possible based on the present observations. First, the hybridization event(s) occurred somewhere other than on the
parental island groups. Hemiphyllodactylus typus occurs
infrequently in the Philippine sample and not at all in the
Palauan one. Additionally, survival of hybrids without a
competitive (numerical) swamping by bisexuals and the
opportunity to develop suficient population density for
subsequent dispersal seem more likely to occur off the parental islands than on them. This supposition derives from
my observation on rarity of H. typus on the islands of
the south central Paciic and their rarity on human-made
structures in the presence of other gecko species. When
other geckos occur with H. typus, it occurs in the darkest area, well removed from Gehyra and Lepidodactylus
lugubris, and this segregation occurred before the Hemidactylus frenatus invasion of the late twentieth century.
Unlike L. lugubris in which the unisexuals have displaced
the bisexuals to marginal habitats (Ineich, 1999), H. typus
has not displaced bisexuals or has displaced them only in
human habitats. I further speculate that the hybridization
event occurred in a location outside the distribution of a
bisexual Hemiphyllodactylus and likely subsequent to the
arrival of humans to the Paciic islands. The preceding origin hypothesis derives from my assumption that the entire
distribution of H. typus derives from human transport—
shipboard, possibly during the eighteenth and early nineteenth centuries.
Phenetic similarities (i.e., primitive advance state polarities not determined) suggest the relatedness of the Palauan and Philippine populations. They share the elongate
habitus, small body size (medians, 33 and 35 mm SVL,
respectively; Table 3), separated precloacal and femoral
pore series, no enlarged chin scales (Table 6), usually two
cloacal spurs on each side, and similar coloration traits
of a U-shaped postsacral mark and series of small, but
bright, spots dorsolaterally on the trunk. Of the preceding
traits, the two Paciic bisexual populations share habitus,
body size, pore morphology, and chin scale and cloacal
spur morphologies with the Indian sample. The presumed
hybridization of the Paciic bisexuals yielded a unisexual
hybrid (H. typus Bleeker) with a larger average body
size (>36 mm SVL) but otherwise identical or very similar to the bisexual parental species. H. typus is the only
number 631
Hemiphyllodactylus taxon in which most (>90%) adult
females have secreting precloacal pores. The pores of the
unisexual adults appear generally smaller than those of
the males of the bisexual taxa. Dorsolateral spots occur
in unisexual individuals, although the spots are white to
light tan, and the postsacral mark has lost the anterolateral arms.
Even though similar, Palauan and Philippine populations have morphological differences that I interpret to
demonstrate speciation. The dorsolateral spots and postsacral mark are red in the Philippine population (H. insularis Taylor) in contrast to pink spots and yellow suffused
with pink postsacral mark of the Palauan population (unnamed). These two populations also differ in average size,
with H. insularis slightly larger; more than half the Philip
sample exceeds the maximum SVL of Palau specimens.
The size distribution of Palau seems unlikely to result
from sampling owing to the number of individuals in the
sample and the thoroughness, recency, and duration of the
Palauan inventory effort (Crombie and Pregill, 1999). Palauan geckos have a proportionately broader head than
H. insularis ones (65 versus 59% HeadW/HeadL; Table
3, and see also SnW/HeadL therein). In scalation, Palauan
geckos average more TotPore than did the Philippine population (27 versus 22.5; Table 6); as noted in the Results
section, an increase in TotPore relects the addition of
more femoral pores to the left and right series.
The Indian sample (H. aurantiacus Beddome) is a distant geographic outlier to the Paciic Hemiphyllodactylus
taxa. The Indian geckos also differ strikingly from them by
a bolder coloration. The dark transverse dorsal markings
are broader and more continuous across the dorsum, and
the pre-postorbital stripe is also broader and continuous
from naris to shoulder. The postsacral mark is dark yellow to gold and the anterolateral arms are variously developed. The average size of Indian geckos (Table 3) matches
that of H. insularis, but the relative trunk length (TrunkL/
SVL) is shorter (49%) in H. aurantiacus than the Paciic
taxa (53% Philip, 54% Palau). H. aurantiacus has a signiicantly broader head (76 versus 59 and 65% HeadW/
HeadL) than the Paciic taxa, although relative snout width
(SnW/HeadL) is the same for all three. The most striking
difference of H. aurantiacus is its digital lamellae formulae
of 2-2-2-2 forefoot and 2-2-3-2 hindfoot. No other population of Hemiphyllodactylus has such a low number of
lamellae on its digits. The two Sri Lankan non-typus specimens examined have very different formulae: 2-2-2-2 and
?-3-3-2 adult male (BMNH 1910.3.16.4); 3-4-4-4 and 5-56-5 adult female ([NMB 8552). The male is potentially a
H. aurantiacus. Its locality data are Hambonota, Ceylon;
•
31
however, given that other early BMNH specimens with
Ceylon locality data are not members of the Sri Lankan
herpetofauna (R. Somaweera, personal communication,
March 2007), its status is questionable until a broader
sampling of Sri Lankan Hemiphyllodactylus is available.
The NMB female’s locality is only Ceylon; it is possibly a
H. typus with an unusual hindfoot lamellae formula.
The broader geographic ranges of the eastern Asian
samples and prior recognition of multiple taxa makes the
resolution of populational and taxonomic boundaries dificult. Bisexual Hemiphyllodactylus populations occur
from west-central Myanmar (CAS 231030, Chin State)
to eastern Guangxi (Zhou et al., 1981: ig. 7). Eastward,
a gap of ~1000 km exists from the Guangxi occurrence
to the population in the Hong Kong area (Lazell, 2002;
Chan et al., 2006). Bisexual populations occur southward
from Yunnan-Guangxi through Malayan Peninsula and
the Greater Sunda Islands. The occurrence of populations
in Indo-China/SE Asia appears very spotty. It is uncertain
whether this spottiness is actual occurrence or simply a
result of few inventory surveys and/or inadequate vouchering of specimens. I must note that inding Hemiphyllodactylus even with intense ield surveys is uncommon.
For example, in the Pyin-Oo-Lwin area (Myanmar), six
researchers inventoried a site multiple times over 2 weeks
(mid-monsoon, August 2003) and found only two individuals, one each on two consecutive evenings.
The Chinese populations of H. yunnanensis seemingly
contrast sharply with the preceding statement of rarity.
Late nineteenth century sampling in the Kunming (=Yunnan-fu) area resulted in series of H. yunnanensis in many
European and North American museums. More recently,
Zhou et al. (1981) amassed a collection of 640 specimens
of this taxon from nine localities within Yunnan and Guizhou. This number does not indicate rarity within this area
and possibly relects the absences of microsympatry with
other geckos, particularly Hemidactylus (see Zhao and
Adler (1993:304) for list of Yunnan and Guizhou geckos).
This abundance also supports the low competitiveness hypothesized for H. typus earlier in this section.
Hemiphyllodactylus yunnanensis and the other samples and populations of mainland eastern Asia share the
adult robust body form, absence of caecal and gonadal pigmentation, a pair of enlarged postmental chin scales, and
a continuous series of precloacal–femoral pores in adult
males. The robust morphology (Figure 3) is most apparent
in larger individuals and perhaps in adult females, although
my measurements and proportions are inadequate to test
this proposition. Juveniles and subadults retain a slender
trunk, hence a more elongate appearance than adults. The
32
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
enlargement of chin scales is more variable within and
among populations. None of the eastern Asian populations
lack the enlargement of the pair of chin scales abutting the
mental scale; however, the degree of enlargement ranges
from barely to strongly among populations. The relative
enlargement is consistent within each population. Zhou
et al. (1981) showed two patterns of postmental enlargement (Figure 7D,E) in their H. yunnanensis sample. In their
H. y. yunnanensis, the postmentals are large with a rapid
decrease in size of the arc of chin scales bordering the infralabials; the next arc of chin scales are distinctly smaller
and nearly subequal to the remainder of chin scales. The
postmentals of H. y. longlingensis are moderately large,
the infralabial abutting arc of chin scales slightly enlarged
as also are the scales in the second arc, inwardly the chin
scales quickly become smaller and “typical.” Are these two
patterns discrete or a continuum? I favor the latter interpretation, although individuals from single-site subsamples
that I have examined can usually be assigned to a single
state, a few individuals grade into the opposite state. As
noted in the Results section, my chin character does not
discriminate between the two postmental states of Zhou
et al., but all eastern Asian samples share the low count
state, including the holotype of Hemiphyllodactylus typus
chapaensis Bourret. I single out this latter specimen for two
reasons: First, Bourret noted small chin scales abutting the
mental and infralabials, and his illustration (reproduced
in Figure 7F) is ambiguous on this condition; second, if
the chapaensis population has no enlarged postmental, it
differs from the other eastern Asian populations. In this
respect, Lazell (2002; also later publications) has identiied
the Hong Kong population as H. chapaensis. Hong Kong
geckos match the general eastern Asian morphology, perhaps with a proportionately longer trunk (53 versus <49%
for all other Asian subsamples). They have enlarged, although not proportionately larger, postmentals than other
Asian mainland population; however, the second arc and
some of third arc of chin scales seem proportionately larger
than in the other populations. In a contrasting situation, a
southern Vietnamese specimen (USNM 146161) has strikingly large postmentals, similar to those depicted for H.
margarethae in Figure 7C. In contrast, the single southern
Cambodian specimen has an arc of small scales.
There is no pattern/differentiation in precloacal–
femoral pores visible within the populations and samples
from mainland southeastern Asia. These populations
share the continuous pattern of precloacal and femoral
pore series in contrast to discontinuous pattern of unisexual H. typus and the bisexual populations of India,
Philippines, and Palau. The continuous pattern continues through the peninsular populations of Thailand and
Malaysia. In contrast, the bisexual males of Sumatra and
Borneo have discontinuous (separated) pore series.
Taxonomic Decisions and Geography
Which came irst, the gecko or the egg? This twisted
banality highlights the necessity of addressing the geography of unisexual Hemiphyllodactylus separately from that
of the bisexual species. Because squamate unisexuality appears to be universally derived from hybridization (Zug
et al., 2001), I examine the geographic patterns of occurrence and taxonomy of the bisexual species irst.
Among the bisexual species, H. aurantiacus (Beddome) was the irst to be described and is the easiest of
the Hemiphyllodactylus populations to be addressed taxonomically. Small adult body size and uniquely low foreand hindfoot lamellar formulae readily identify it as a
distinct genetic entity and phylogenetic lineage. Assuming
habitus, visceral pigmentation, and precloacal–femoral
pore morphology relect phylogenetic afinity, H. aurantiacus is a member of the typus clade.
The two other bisexual members of the typus
clade are geographically distant—Palau and Philippine
islands—from H. aurantiacus. Taylor (1918) recognized
the distinctiveness of the Philippine populations by several
minor differences in scalation from Stejneger’s description
of Hemiphyllodactylus leucostictus. He even noted that
it might not be distinct from the latter species, although
he highlighted some differences in coloration. Crombie
and Pregill (1999) were the irst biologists to document
the occurrence of Hemiphyllodactylus in Palau. Early in
their herpetofaunal survey of this island group, they recognized the differentiation of the Palauan population from
H. typus, irst because of the presence of males and second
because of its distinct coloration. The latter trait and a few
others demonstrate its uniqueness to me, and I describe it
as a new species in the following Species Account section.
The bisexual populations of mainland and islands
Asia are less readily delimited. This dificulty results
largely from inadequate sampling in number of individuals from most localities, too few localities, and localities
geographically distant. The uniqueness of H. yunnanensis
(Boulenger) is unquestionable. Its traits (robust body, unpigmented caeca and gonads, and continuous precloacal–
femoral pores) serve to identify a China–Indochina–Sunda
clade. I propose that the populations from north central
Burma through south central China and the adjacent
northern half of Thailand, Laos, and northern Vietnam
represent a single taxon H. yunnanensis. Some regional
populational differentiation occurs in this widely distributed species, although morphological data suggest minimal
number 631
genetic differentiation. This hypothesis thus relegates the
Zhou et al. subspecies to the synonymy of a monotypic H.
yunnanensis. Hemiphyllodactylus typus chapaensis Bourret similarly becomes a synonym of H. yunnanensis (Boulenger). Geographically, the H. t. chapaensis type locality
is only 300 km south of Kunming (type locality of H. yunnanensis) at the southern terminus of a continuous range
of mountains. This mountain range also contains the type
locality of H. y. jinpingensis Zhou et al., and this locality
is less than 100 km north of Chapa. Zhou et al. (1981)
did not compare the morphology of their samples to H. t.
chapaensis Bourret.
I noted earlier Lazell’s use of H. chapaensis for Hong
Kong Hemiphyllodactylus. This usage must be discontinued because the morphology of the Hong Kong slender
geckos is distinct from that of H. typus chapaensis Bourret.
Does the Hong Kong population represent an outlier of H.
yunnanensis or is it a unique Hemiphyllodactylus population? My present data suggest the latter interpretation, but
they are insuficient to address this hypothesis rigorously.
Presently, I am unaware of vouchered records of Hemiphyllodactylus in Guangxi and Guangdong provinces of
southern China. A broad distributional gap also exists for
geckos of the Hemidactylus bowringii group. Initially, we
(Zug et al., 2007) proposed that Hong Kong “bowringii”
was an exotic species, accidentally introduced from Burma
or India. A more detailed study (McMahan and Zug, 2007)
subsequently demonstrated the uniqueness of the Hong
Kong population. This latter study urges caution, so I am
hesitant to postulate the speciic status of the Hong Kong
Hemiphyllodactylus without enlarging my sample and locating vouchers from the Guangxi–Guangdong corridor.
The preceding restriction (in Morphological Differentiation subsection) of H. yunnanensis to slender gecko
populations from southern China and adjacent northern
Indochina leaves the taxonomy of southern Indochina
populations unresolved. Again, the inadequacy of museum vouchers does not permit a satisfactory resolution.
Presently, the data are adequate to declare that Malay
peninsular populations are distinct; however, it is uncertain whether they are conined to the peninsula or extend
into southern Indochina, likely the former owing to their
restricted occurrence to montane rainforests. The single
specimen (USNM 146161) from southern Vietnam suggests this possibility, although morphologically, it appears
more similar to the Sumatran population. Also, a sample
(THNHM 4910-4917) from Kaeng Krachan National
Park, Thailand, just north of the Isthmus of Kra is more
similar to other southern Thailand geckos than to Malayan
ones. For the present, I recommend labeling southern Indochina and Hong Kong specimens as “H. yunnanensis.”
•
33
Two names are available for Malayan Hemiphyllodactylus populations. H. larutensis derives from Bukit Larut
(=Maxwell’s Hill or Larut Hills), a mountain at the southern end of the Bintang range. Although the type locality
of H. harterti was given as “Malakka” by Werner (1900),
Boulenger (1912) noted that this locality name referred to
a general locality and not a speciic site (i.e., Malakkahalbinsel [German] equals Malaya [English]). Furthermore,
because Ernest Hartert (the donor of the specimen to the
Berlin museum) collected birds at Gunong Inas in 1888
(Hartert, 1901–1902; AMNH Department of Ornithology records), Boulenger (1912) tentatively recommended
changing the type locality to Gunong Inas. Gunong Inas
is ~40 km north of Bukit Larut, and both are mountains
within the western Malayan mountain range (Bukit Bintang). I have examined personally only one specimen
(G. larutensis type) and have data and images from a second specimen (L. Harterti type), and there are no traits
suggesting that these two represent different taxa. Thus I
accept Boulenger’s restriction of the type locality to Gunong Inas. I note only that Werner’s (1900:196) “gleicht
in Färbung und Zeichnung ganz dem L. lugubris” better matches the pattern of geckos of the central Malayan
mountain range (Bunjaran Titiwangsa); nevertheless, I can
ind no evidence that Hartert collected in the latter mountain range or received specimens from there. The similarity
of morphology of the two type specimens and the type restriction of H. harterti to the Bintang range makes H. harterti the senior synonym and valid name for the Bintang
taxon. The preceding nomenclatural decision results in the
populations of Bunjaran Titiwangsa, at least those from
the Cameron Highlands southward to the Fraser Hill area,
being nameless. These populations have coloration and
minor morphological differences from H. harterti, their
occurrence in a mountain range with a different geological
history and long separation by a lowland valley leads me
to propose them as a separate phylogenetic lineage.
The sample of Sumatran bisexual geckos is modest
but adequate to demonstrate the morphological distinctiveness of these geckos. Thus, in spite of Brongersma’s
belated change of mind, H. margarethae is a unique lineage and presumably Sumatran endemic species. Like H.
harterti and the central mountain taxon, it appears to be a
montane species, conined to cooler and moister habitats
than the invasive H. typus, whose type locality occurs in
Sumatra as well.
I identiied most Bornean Hemiphyllodactylus specimens as unisexual typus. As I analyzed and reexamined
my data, I discovered that my identiications resulted in
all female specimens as H. typus and the three males as
unknowns. All Bornean slender geckos have pigmented
34
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
caeca; all males have pigmented testes-epididymides peritoneum, all females except one with pigmented oviducal
peritoneum. Were my identiications driven by knowledge
of sex? The evidence deinitely suggests a bias. A reexamination of the specimens is not possible at this stage of the
analysis; however, a reexamination of the data suggests
that most females were correctly identiied as H. typus.
Tentatively, bisexual individuals can be differentiated
from unisexual ones by a higher lamellar formula. The
presence of males with normal appearing gonads indicates
the presence of a bisexual species as well as the unisexual
H. typus on Borneo (entire island, not just Indonesian Borneo). Formal recognition of this bisexual species requires a
reexamination of all Bornean specimens and comparison
with the Philippine H. insularis.
Elsewhere in the Sundas, all Hemiphyllodactylus are
unisexual typus. A peculiar situation occurs in Komodo
where all individuals are pale, hence Auffenberg’s (1980)
decision to describe them as a subspecies (pallidus) of H.
typus. Pale H. typus are observed elsewhere as one end
of a color-phase shift. Because Auffenberg had only two
specimens and presumably observed no additional ones,
I suggest that by chance he collected and preserved both
in their light phase. Presumably, selection for lighter individuals can occur in clonal vertebrates, but the Komodo
environment does not seem greatly different from other
island habitats with resident populations of H. typus.
Phylogeny and biogeography: speculations.
This study was not designed to address phylogenetic relationships within Hemiphyllodactylus or its
relationships with other gekkonid genera. Morphological
similarities and differences among the recognized species
taxa led to my hypothesis of two subclades—the typus and
the yunnanensis species groups—and the monophyly of
Hemiphyllodactylus. Accepting these two items suggests an
ancient origin (?Miocene or earlier) of Hemiphyllodactylus
and an early divergence of the elongate and robust-bodied
species groups. Kluge (1968) implicitly proposed Hemiphyllodactylus as a member of the Lepidodactylus clade,
although in his study, he examined nomenclatural issues,
not phylogenetic ones. His subsequent studies of gekkonid
relationships did not address directly either the interspeciic
or intergeneric relationships of Hemiphyllodactylus. Similarly, this genus has been absent from most other gekkonid
phylogenetic studies, although the broad taxon-based studies of A. Bauer and colleagues likely will provide information on the relationships of Hemiphyllodactylus and some
of the species in this genus. Their initial evaluation (Han et
al., 2004) leaves its relationships unresolved.
Having proposed the existence of two sister groups
(subclades) of Hemiphyllodactylus, I wish to delineate and
deine these species groups. The typus species group contains four, and possibly ive, species: H. typus Bleeker, H.
aurantiacus (Beddome), H. insularis Taylor, the Palauan
population described in the following species accounts,
and possibly a ifth, the populations in Borneo. The yunnanensis species group contains four, and likely more, species: H. yunnanensis (Boulenger), H. harterti (Werner),
a central Malaysian taxon, H. margarethae Brongersma,
and possibly a separate species in South China–Hong
Kong and southern Indochina. The typus species group
members have an elongated habitus, accentuated by a
long trunk (TrunkL/SVL >50%) with short limbs and
proportionately small head, caecum and gonadal ducts
darkly pigmented, and precloacal and femoral pore series
separated. The yunnanensis species group members have
a stouter adult habitus although still elongated (TrunkL/
SVL ~50%) and proportionately larger head, caecum not
pigmented and gonadal ducts rarely so, and precloacal
and femoral pore series separated or continuous.
Continuing my speculations on origins and biogeography of the bisexual populations, I suggest that the elongate clade derived from the more robust clade and that its
origin occurred in northern precursor-Sunda with dispersal eastward into the Philippines and hence into the western Paciic, and westward to peninsular India, now with
only relict populations remaining. Hemiphyllodactylus
aurantiacus represents a surviving member of this early
“dispersal” to India with a long period of isolation. Presently, the situation for the Sri Lankan bisexual population
is undecipherable owing to paucity of data. The robustbodied Hemiphyllodactylus has remained and differentiated within the area of the group’s origin and northward
in southwestern China.
I suggested previously (Morphological Differentiation section) that the origin of the hybrid H. typus was an
extralimital hybridization of H. insularis and the Palauan
species. When the hybridization occurred is unknown.
A human-induced event seems most likely, as also does
an “off the native island” event. The event could have
occurred during the early human migrations into and
through the Paciic islands or, my preference, with the
Euro-American exploration and commercialization of the
Paciic. “Off-island” hybridization seems likely owing to
presumed low interspeciic competitiveness of the unisexual H. typus in contrast to the displacement of parental
bisexuals by clones of the unisexual Lepidodactylus lugubris (Ineich, 1999). I postulate the possibility of onboard
ship hybridization then colonization and population expansion in central Paciic. While such an origin could have
occurred during the initial human migration into the Paciic, onboard hybridization seems more likely aboard the
number 631
larger European sailing ships. Whose ships? The broad
distribution and the early occurrence in the Mascarenes
hint at an association with French exploration; however,
the impact of whaling vessels for the dispersal of Paciic
lizards has been overlooked. Whaling vessels were notorious for their catholic (trashy) cargo and for their regular
and widespread shore leave. Thus whaling ships provide
a “safe-haven” habitat for hybridization and the initial
survival of the hybrids; then they could serve for the transfer of hybrids to diverse islands and to other whalers to
broaden the dispersal of the hybrids.
Other phylogenetic and biogeographic questions remain for this low diversity taxon. First and foremost is why
are there so few species in this gekkonid clade? Its putative
sister clade, Lepidodactylus, has at least 4 times as many
species, and its unisexual species L. lugubris (multiple
clones, multiple origins) has “outcompeted” its bisexual
parent in many Paciic island ecosystems. The Hemiphyllodactylus clade displays several distributional anomalies.
The widespread occurrence of unisexual H. typus contrasts sharply with the bisexual species. In one aspect, this
feature is shared with L. lugubris, although its low density
at most invasive sites differs greatly from mourning gecko
occurrences. Also, in contrast to L. lugubris, it dispersed
more widely. Another peculiarity is the dominance of the
stout-bodied Hemiphyllodactylus in southeastern Asia and
the division of the slender body clade into extreme eastern and western outposts. The stout-bodied geckos show
a southern China distributional hiatus, absent from the
China–Indochina border then occurring in the Hong Kong
area. I earlier noted a similar pattern for the Hemidactylus
bowringii species group, hence my reluctance to attribute
the Hong Kong occurrence to human introduction. The
documentation of the distributional patterns of tropical
Asian amphibians and reptiles is rudimentary. This situation derives from our poor knowledge of tropical Asian
diversity, particularly among mainland taxa, and this poor
knowledge results from the continuing recognition of panAsian species, when few such species exist (Stuart et al.,
2006; Zug, In press).
SPECIES ACCOUNTS
Hemiphyllodactylus typus Bleeker
Indo-Paciic slender gecko
Hemiphyllodactylus typus Bleeker, 1860:327 [type locality: “Agam” (Sumatra); holotype, BMNH 1946.8.30.83].
Platydactylus crepuscularis Bavay, 1869:8 [type locality: “Nouvelle-Calédonie” (locality implied from title of publication); holotype lost (Brygoo,
1990:49)].
•
35
Spathodactylus mutilatus Günther, 1872:594 [type locality: “East-Indies archipelago”; holotype, BMNH 1946.8.30.83].
Lepidodactylus ceylonensis Boulenger, 1885:164, Pl. XIII, ig. 3 [type locality: “Gampola” (Ceylon); holotype, BMNH 74.4.29.1326].
Hemiphyllodactylus leucostictus Stejneger, 1899:800 [type locality: “Kauai,
Hawaiian Islands”; holotype, USNM 23500].
Hemiphyllodactylus typus pallidus Auffenberg, 1980:72 [type locality:
“along Vai Nggulung, Loho Liang, Komodo, 30 mm”; holotype, UF
28985].
Hemiphyllodactylus albostictus Lazell, 1989:126 [spelling error].
Comments.
“Bleeker’s types were sold to the
British Museum in 1863 (or at least years before 1879, . . .).
The type of H. typus is the same as that of Spathodactylus
mutilatus Gthr. (Boulenger l.c).” The preceding statement
is from Brongersma’s review (1932:212[footnote 2]) of
Hemiphyllodactylus.
Taylor’s (1963) description of H. typus was based on a
male from Fraser’s Hill, Malaysia. In addition to the specimen being a male, it had a continuous precloacal–femoral
pore series; hence the description is not representative of
H. typus.
Description.
An all-female taxon of geckos
(Gekkoninae) with elongate, slender habitus, slightly
compressed, elongated appearance accentuated by short
limbs and small head (see Figures 3, 6, 13), tail round in
cross section and commonly shorter than SVL. Adults
29.4–46.1 mm SVL (mean ± SD, 38.4 mm ± 2.91; n =
143), 14–36 mm TailL (28.5 mm ± 5.03), 15.0–28.0 mm
TrunkL (20.3 mm ± 2.11), 6.6–9.9 mm HeadL (8.2 mm
± 0.56), 3.7–6.6 mm HeadW (5.2 mm ± 0.52), 2.3–4.1
mm SnEye (3.4 mm ± 0.30), 1.8–3.4 mm NarEye (2.6
mm ± 0.28), 1.5–2.4 mm EyeD (2.1 mm ± 0.17), 0.9–1.7
mm SnW (1.3 mm ± 0.17). Adult proportions 40–65%
TrunkL/SVL (mean ± SD, 52.9% ± 3.2), 18–24% HeadL/
SVL (21.3% ± 0.9), 10–16% HeadW/SVL (13.7% ± 1.1),
51–77% HeadW/HeadL (64.1% ± 4.8), 34–48% SnEye/
HeadL (41.1% ± 2.4), 24–40% NarEye/HeadL (32.2% ±
2.8), 20–32% EyeD/HeadL (25.3% ± 1.9), 11–21% SnW/
HeadL (16.2% ± 1.9), 61–106% EyeD/NarEye (79.0% ±
7.9), 16–34% SnW/HeadW (25.3% ± 3.3).
Scalation predominantly granular from head onto tail,
both dorsally and ventrally; ventral trunk scales slightly
larger than dorsal ones, 12–19 Dorsal (median ± SD, 15
± 1.6) and 8–14 Ventral (10 ± 1.3); similarly, subcaudal
scales slightly larger than dorsal caudal scales but not platelike. Cloacal spurs present, modest sized, 1–5 CloacS (2 ±
0.8). Larger scales on lips and snout, rostral largest, rectangular to pentagonal, often slightly concave on dorsomedial edge with slight cleft; 1–5 CircNa (3 ± 0.6), 1–5 SnS
(2 ± 0.7); labial scales enlarged from rostral to below eye,
36
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
becoming progressively smaller in subocular rictus, 9–14
Suplab (11 ± 1.0), 7–13 Inlab (10 ± 0.9); 9–14 Chin (11 ±
1.1), those behind mental slightly or not enlarged; ear
opening distinct with no bordering enlarged scales. Each
digit with expanded pad, terminal two phalanges free, arising from within pad on second to ifth digits of fore- and
hindfoot and each clawed; pads of these digits each with
large triangular apical lamella bordered proximally by lyreshaped lamellae (scansors); modal digital formulae 3-4-4-4
(forefoot) and 4-4-5-4 (hindfoot) for scansors; irst digit of
fore- and hindfeet compressed, usually 5 rectangular lamellae (4-5 fore, 5-6 hind) ventrally, terminal phalanx not free
with or without minute claw. Adults usually with precloacal pore series (0–14 PreclPor, median ± SD = 10 ± 3.9)
always separated from femoral pore series (0–12 combined
left and right femoral series), 0–26 TotPore (12 ± 7.2).
Dusky tan to reddish brown ground color dorsally
and laterally from head to tail, usually with dark ocular
stripe from loreal area to anterior trunk; series of widely
spaced small white spots, often darkly edged, dorsolaterally from temporal area to inguina; dorsal postsacral dark
brown blotch bordered posteriorly by transverse bar of
white or beige. Underside dusky from chin to vent, pale
yellowish orange on tail.
Major diagnostic features are as follows: all-female
taxon; pigmented caecum and gonadal ducts; if present
(uncommonly), femoral pore series separate from precloacal pore series; chin scales bordering mental and irst
infralabial not greatly enlarged; digital lamellae formulae
3-4-4-4 (forefoot) and 4-4-5-4 (hindfoot); average adult
SVL ~38 mm; series of white spots dorsolaterally on trunk
and bright postsacral bar of white and dark brown.
Description of holotype:
An adult female (Figure
13), 43.3 mm SVL, 40.0 mm TailL, 22.4 mm TrunkL,
9.1 mm HeadL, 5.8 mm HeadW, 3.7 mm SnEye, 2.9 mm
NarEye, 2.1 mm EyeD, and 1.1 mm SnW. Proportions:
52% TrunkL/SVL, 21% HeadL/SVL, 13% HeadW/SVL,
64% HeadW/HeadL, 41% SnEye/HeadL, 32% NarEye/
HeadL, 23% EyeD/HeadL, 12% SnW/HeadL, 72% EyeD/
NarEye, 19% SnW/HeadW. Scalation: 3 CircNa, 3 SnS,
11 Suplab, 11 Inlab, 13 Chin (anteromedial ones only
slightly larger than adjacent ones), 13 Dorsal, 8 Ventral,
2 CloacS, Subcaud not enlarged, 15 PreclPor, 23 TotPore
with no contact between precloacal and femoral, digital
formulae 3-3-4-3 fore and 3-4-4-4 hind. Pigmented caecum, pigmentation unknown for oviducts.
Body ground color faded to a uniform orangish tan,
no lateral spotting evident, dark lateral stripe from in
front of eye to axilla, broken dark reticulations on rear of
head and nape, dark chevron at tail base.
FIGURE 13. Holotype of Hemiphyllodactylus typus Bleeker, 1860
(BMNH 1946.8.30.83): (A) dorsal view of whole body, (B) ventral
view of throat and chin, and (C) ventral view of pelvic area.
Etymology.
Bleeker offered no explanation
of his selection of the name typus. The name is a Latin
noun for impression, shape, Figure, or example. I assume
that he chose typus because this species represented the
type species for his new genus Hemiphyllodactylus.
I propose a standard English name at variance to the
commonly used Indo-Paciic tree gecko. Whereas this
gecko is occasionally found on trees, it more commonly
occurs in the leaf whorls of Pandanus and on humanmade structures. Slender refers to the elongate and attenuate appearance of H. typus and most of its congeners.
Variation.
The means or medians and ranges
are detailed in the preceding Description section. No signiicant or striking variation was seen in any of the characters among the samples from throughout the range of
this taxon. The variation observed within each character is
equivalent to that observed in the repeated measures data
of a Palauan bisexual female. Where character variation
is greater, the variation is likely attributed to data-gathering variance arising from the poor preservation quality of many specimens. I suggest that this low variation
among the widely separated population is evidence that
all populations of H. typus derive from a single hybridization event and subsequent dispersion of this single clonal
number 631
population. The lightness of Auffenberg’s Komodo specimens is attributed to chance.
Distribution.
Broadly, if somewhat spottily,
distributed from Hawaii and French Polynesia in the central Paciic westward to Paciic Rim islands and coastally
through New Guinea, Sunda, and Indochina to Sri Lanka;
also occurs in Mascarene Islands (Figure 14).
Hemiphyllodactylus aurantiacus Beddome
Southern Ghats slender gecko
Hemidactylus aurantiacus Beddome, 1870:33 [Type locality: “Shevaroys, under stones about Yercaud and elsewhere at an elevation of 4000 feet”
(Tamil Nuda, India); syntypes: BMNH 74.4.29.1332–1337, ZMB
10233; lectotype, BMNH 74.4.29.1333].
Comments.
Beddome’s description is ambiguous on the number of specimens available to him as
he composed the description. His description is based on
a single specimen (~30–31 mm SVL), presumably immature, sex uncertain (no precloacal pores). The syntypic
series (BMNH 74.429.1332–1337) consists of nine individuals, two adult males, three females (all with adult
SVLs but two with immature ovaries), and four juveniles.
I designate the male BMNH 74.429.1333 as the lectotype
of Hemidactylus aurantiacus Beddome, owing to my reluctance to use an immature specimen as a type and the
absence of an immature specimen matching Beddome’s
dimensions. Additionally, I accept Bauer and Günther’s
(1991) assessment that ZMB 10233 is a syntype; it is a
mature male (32-mm SVL) with precloacal and femoral
pores, hence also not the specimen on which Beddome
based his description.
Description.
A bisexual taxon of geckos
(Gekkoninae) with elongate, slender habitus, slightly compressed, elongated appearance accentuated by short limbs
and modest head (see Figures 3, 11, 15), tail somewhat
elliptical in cross section and regularly shorter than SVL.
Adults 27.2–37.9 mm SVL (mean ± SD, 34.3 mm ± 2.80;
n = 14), 26–33 mm TailL (29.2 mm ± 2.87), 13.8–18.8
mm TrunkL (16.7 mm ± 1.27), 6.0–8.7 mm HeadL (7.9
mm ± 0.64), 4.3–6.5 mm HeadW (5.6 mm ± 0.72), 2.3–
3.7 mm SnEye (3.1 mm ± 0.36), 1.9–2.8 mm NarEye (2.4
mm ± 0.25), 1.7–2.1 mm EyeD (2.0 mm ± 0.13), 1.2–1.5
mm SnW (1.3 mm ± 0.10). Adult proportions 44–51%
TrunkL/SVL (mean ± SD, 48.7% ± 2.0), 21–26% HeadL/
SVL (23.0% ± 1.0), 14–19% HeadW/SVL (16.5% ± 2.9),
57–79% HeadW/HeadL (71.3% ± 7.1), 34–42% SnEye/
HeadL (39.2% ± 2.6), 27–33% NarEye/HeadL (30.6% ±
1.8), 22–28% EyeD/HeadL (25.0% ± 1.7), 14–20% SnW/
•
37
HeadL (16.7% ± 1.4), 69–96% EyeD/NarEye (82.0% ±
7.7), 21–30% SnW/HeadW (23.5% ± 2.7%).
Scalation is predominantly granular from head onto
tail, both dorsally and ventrally; ventral trunk scales
slightly larger than dorsal ones, 11–17 Dorsal (median ±
SD, 13.0 ± 1.87) and 8–12 Ventral (10.0 ± 1.51); similarly,
subcaudal scales slightly larger than dorsal caudal scales
but not plate-like. Cloacal spurs present, modest sized,
1–3 CloacS (2 ± 0.6). Larger scales on lips and snout, rostral largest, rectangular to pentagonal, often slightly concave on dorsomedial edge with slight cleft; 2–4 CircNa
(3 ± 0.5), 3–6 SnS (4 ± 0.9); labial scales enlarged from
rostral to below eye, becoming progressively smaller in
subocular rictus, 10–13 Suplab (10 ± 1.0), 8–12 Inlab (11
± 1.0); 10–14 Chin (11 ± 0.8), those behind mental slightly
or not enlarged; ear opening distinct with no bordering enlarged scales. Each digit with expanded pad, terminal two
phalanges free, arising from within pad on second to ifth
digits of fore- and hindfoot and each clawed; pads of these
digits each with large triangular apical lamella bordered
proximally by lyre-shaped lamellae (scansors); modal digital formulae 2-2-2-2 (forefoot) and 2-2-3-2 or 3 (hindfoot)
for scansors; irst digit of fore- and hindfeet compressed,
usually 4 rectangular lamellae (3-4 fore, 4-5 hind) ventrally, terminal phalanx not free with or without minute
claw. Adult females rarely with precloacal pore series (0–3
PreclPor), males always with precloacal pores (median ±
SD, 7 ± 1.6; range, 6–11) always separated from femoral
pore series, 16–25 TotPore (21.5 ± 2.92).
Dusky brown ground color dorsally and laterally
from head to tail, dark ocular stripe from loreal area to
axilla thereafter interrupted and part of zigzag dorsal
markings; also narrow dorsolateral dark stripe from rear
of eye to axilla, where it also breaks into pieces of the
dorsal zigzag marks; small white spots dorsolaterally on
trunk but overwhelmed by dark trunk markings; dorsal
postsacral mark, anteriormost broad dark brown traverse
bar bordered behind by light golden bar then tan and subsequently by irregular edged dark brown bars separated
by tan interspaces.
Major diagnostic features are as follows: bisexual
taxon; pigmented caecum and gonadal ducts; in adult
males femoral pore series separated from precloacal pore
series (TotPore typically ≥20-≤25), always absent in females; chin scales bordering mental and irst infralabial not
greatly enlarged; digital lamellae formulae 2-2-2-2 (forefoot) and 2-2-3-2 or 2-2-3-3 (hindfoot); average adult SVL
~33–35 mm; bold body pattern of contrasting dark brown
and dusky brown background and dorsal postsacral double bar of dark brown and light orange (Figure 15).
38
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
FIGURE 14. Geographic occurrence of Hemiphyllodactylus typus. Not all localities in the same area are plotted. (A) Asia records [Asia
and Islands] and (B) Oceania records. Symbols: solid circle = specimen(s) represented by museum vouchers and speciic identity conirmed; open square = locality from published records or museum records but specimen not examined.
number 631
•
39
FIGURE 15. Syntypes of Hemiphyllodactylus aurantiacus Beddome, 1870: (A) lectotype (BMNH 74.4.29.1333) and (B) syntypic series (from
top row, left to right, BMNH 74.4.29.1332–1337 and three unnumbered juveniles). (Photographs by G. Zug.)
Description of lectotype:
An adult male: 34.5 mm
SVL, broken TailL, 17.5 mm TrunkL, 7.8 mm HeadL, 5.6
mm HeadW, 3.7 mm SnEye, 2.8 mm NarEye, 2.0 mm
EyeD, and 1.2 mm SnW. Proportions: 51% TrunkL/SVL,
23% HeadL/SVL, 16% HeadW/SVL, 72% HeadW/HeadL,
40% SnEye/HeadL, 30% NarEye/HeadL, 26% EyeD/
HeadL, 15% SnW/HeadL, 87% EyeD/NarEye, 21% SnW/
HeadW. Scalation: 3 CircNa, 5 SnS, 13 Suplab, 11 Inlab,
12 Chin (anteromedial ones only slightly larger than adjacent ones), 16 Dorsal, 12 Ventral, 1 CloacS, Subcaud not
enlarged, 7 PreclPor, 22 TotPore with no contact between
precloacal and femoral, digital formulae 2-2-2-2 (forefoot)
and 2-3-3-3 (hindfoot). Pigmented caecum, pigmentation
unknown for testis epididymis.
Body ground color brown, no lateral light spotting evident, dark dorsolateral stripe from eye to shoulder, lateral
stripe from in front of eye to axilla, these stripes broken on
trunk and form lateral parts of dorsal chevron or zigzag
markings of trunk, dark chevron at tail base.
Etymology.
Beddome offered no explanation
for his choice of aurantiacus as the epithet for his new
species. He did mention the orange color of the tail base,
and because aurantium is a new Latin noun for the orange
(fruit), I suggest his choice derived from the association
between the color of the orange and the gecko’s tail.
Variation.
The means or medians and ranges
are detailed in the preceding Description subsection.
Males are somewhat smaller than females, but the difference in average size is slight and not statistically signiicant. None of the mensural traits shows signiicant sexual
dimorphism and neither do any meristic traits other than
precloacal and femoral pores. One large female possesses
three secreting precloacal pores; no other females have secreting pores. All adult males possess both precloacal and
femoral pore series.
Distribution.
This gecko is an endemic
of the southern tip of India (Tamil Nadu) (Figure 16).
It occurs mainly in association with evergreen forest at
40
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
FIGURE 16. Geographic occurrence of Hemiphyllodactylus aurantiacus and H. yunnanensis. Not
all localities in the same area are plotted. Symbols: circle, specimens represented by museum vouchers and speciic identity conirmed; square, localities from published records or museum records but
specimen not examined; diamond or square below mid-Indochina boundary (dotted line), specimens and records tentatively assigned to H. yunnanensis.
mid-elevations in the Nilgiri and Anaimalai Hills (Western
Ghats) and Shevaroy and Kolli Hills (Eastern Ghats). Its
occurrence in Banglore is ediicarian and appears to represent a recent accidental transport. (Distributional summary largely from Bauer and Das, 1999.)
Hemiphyllodactylus ganoklonis Zug,
new species
Palauan slender gecko
holotype.
USNM 563682, adult male from
Palau Islands, extreme northern tip of Ulebsechel Island,
50 m south of channel between Ngermalk and Ulebsechel,
collected by Ronald I. Crombie, 2 August 1998.
Paratypes.
All subsequent specimens are from
Republic of Palau (= Palau Islands); that datum is removed from each subsequent locality for brevity. SAM
R47715, Babeldaob Island, 4 mi [6.4 km] north on west
coast road to Aimeliik, 7°24′N 134°30′E, collected by
Christopher C. Austin, 2 June 1996. USNM 495065,
Babeldaob Island, east of Nekkeng, 0.5 km west of Tabecheding River (Aimeliik State) 7°27′25″N 134°30′40″E,
collected by Gordon H. Rodda, 17 January 1993; USNM
495066, Babeldaob Island, 1 km east of Ngerchaech
Mountain, Koksai Radio Station (Ngetpang State) 25 m
7°26′42″N 134°31′48″E, collected by Gordon H. Rodda
and Renee J. Rondeau, 18 January 1993; USNM 563663,
Babeldaob Island, access road to Palau airport, near turnoff into airport itself (Airai State), collected by Ronald I.
Crombie and Christopher C. Austin, 7 July 1996; USNM
563664–665, Babeldaob Island, just north of Ulimang village on road to Galap village (Ngaraard State), collected
by Ronald I. Crombie and Gregory K. Pregill, 14 January 1995; USNM 563666, Babeldaob Island, south of
Ulimang village (Ngaraard State), collected by Ronald I.
Crombie, 6 April 1995; USNM 563667, Ngeanges Island, 7°12′26″N 134°22′21″E, collected by Ronald I.
Crombie, 9 January 1998; USNM 563668, Ngeaur Island, Ngaramasch village, collected by Ronald I. Crombie, Gregory K. Pregill, and G. Ken Creighton, 30 January
number 631
1993; USNM 563669–71, Ngercheu Island group, Carp
Island, 7°05′36″N 134°16′44″E, collected by Ronald I.
Crombie, 18 July 2001; USNM 563672–673 same as preceding, except 13 August 2001; USNM 563674 same as
preceding, except 15 August 2001; USNM 563675 Ngerekebesang Island, southwest (by road) of Meyungs village,
just northeast of turnoff to Echang village, collected by
Ronald I. Crombie and Gregory K. Pregill, 9 June 1994;
USNM 563676 Ngerektabel Island, approx. 0.5 km (air)
northwest of Rael Di, at a sandy beach labelled Oimaderuul on topographic map, collected by Ronald I. Crombie,
4 August 1998; USNM 563677, Oreor Island, southwest
of Ngermid village at Ngerunguikl, Hotel Nikko Palau,
collected by Ronald I. Crombie, 22 April 1992; USNM
563678, Ulebsechel Island, Snake Dick Point, midpoint of
east coast, 7°19′19″N 134°29′15″E, collected by Ronald I.
Crombie and Artemio B. Asis, 28 December 1997; USNM
563679–681, Ulebsechel Island, same data as holotype;
USNM 563683, Ulebsechel Island, same data as holotype,
except 12 February 2002.
Adult females: USNM 495066, 563663–664, -666–
667, -672–673, -675, -679, -681, -684; adult males: SAM
R47713, USNM 495065, 563665, -668–671, -674, -676,
-678, -682; juveniles: USNM 563677.
Description.
A bisexual taxon of geckos
(Gekkoninae) with elongate, slender habitus, slightly compressed, elongated appearance accentuated by short limbs
and small head (see Figures 3, 11, 17, 18), tail round in
cross section and usually shorter than SVL. Adults dimorphic, females larger than males: 31.1–34.2 mm (mean ±
SD, 32.8 ± 1.01; n = 11 females), 28.3–31.6 mm (30.3 mm
± 1.11, n = 12 males) SVL; TailL ~2/3–3/4 of SVL; 16.8–
18.9 mm (17.8 mm ± 0.76), 14.4–17.4 mm (16.0 mm ±
0.95) TrunkL; 6.6–7.5 mm (7.1 mm ± 0.26), 6.3–7.0 mm
(6.7 mm ± 0.24) HeadL; 3.8–4.4 mm (4.2 mm ± 0.17),
3.5–4.1 mm (3.8 mm ± 0.20) HeadW; 2.6–3.1 mm (2.8 mm
± 0.16), 2.4–3.0 mm (2.6 mm ± 0.18) SnEye; 2.1–2.3 mm
(2.2 mm ± 0.09), 1.8–2.3 mm (2.1 mm ± 0.17) NarEye;
1.7–1.9 mm (1.8 mm ± 0.09), 1.5–1.9 mm (1.7 mm ± 0.15)
EyeD; 1.0–1.3 mm (1.1 mm ± 0.11), 1.0–1.1 mm (1.0 mm
± 0.05) SnW. Adult proportions not dimorphic, 49–57%
TrunkL/SVL (mean % ± SD, 53.5% ± 2.3), 20–23% HeadL/
SVL (21.8% ± 0.6), 12–15% HeadW/SVL (12.6% ± 0.7),
53–65% HeadW/HeadL (58.0% ± 3.2), 36–45% SnEye/
HeadL (39.2% ± 2.0), 28–34% NarEye/HeadL (31.0% ±
1.4), 23–28% EyeD/HeadL (25.5% ± 1.5), 13–19% SnW/
HeadL (15.6% ± 1.2), 73–95% EyeD/NarEye (82.4% ±
5.6), 23–32% SnW/HeadW (27.0% ± 2.0%).
Scalation is predominantly granular from head onto
tail, both dorsally and ventrally; ventral trunk scales slightly
•
41
FIGURE 17. Hemiphyllodactylus ganoklonis from Ulebsechel Island, Palau (USNM 563680). (Illustration by M. D. Grifin.)
larger than dorsal ones, 11–18 Dorsal (median ± SD, 15 ±
1.4) and 9–12 Ventral (10 ± 1.0); similarly, subcaudal scales
slightly larger than dorsal caudal scales but not plate-like.
Cloacal spurs present, modest sized, 1–4 CloacS (2 ± 1.0).
Larger scales on lips and snout, rostral largest, rectangular
to pentagonal, often slightly concave on dorsomedial edge
with slight cleft; 2–4 CircNa (3 ± 0.6), 3–5 SnS (4 ± 0.7);
labial scales enlarged from rostral to below eye, becoming progressively smaller in subocular rictus, 8–11 Suplab
42
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
(8 ± 0.8), 8–10 Inlab (9 ± 0.6); 9–12 Chin (11 ± 0.9),
those behind mental slightly or not enlarged; ear opening
distinct with no bordering enlarged scales. Each digit with
expanded pad, terminal two phalanges free, arising from
within pad on second to ifth digits of fore- and hindfoot
and each clawed; pads of these digits each with large triangular apical lamella bordered proximally by lyre-shaped
lamellae (scansors); modal digital formulae 3-4-4-3 (forefoot) and 3-4-4-4 (hindfoot) for scansors; irst digit of foreand hindfeet compressed, usually 4 rectangular lamellae
(3-5 fore, 3-5 hind) ventrally, terminal phalanx not free
with or without minute claw. Adult females lack precloacal
pore series, males always with precloacal pores (median ±
SD, 8 ± 0.9; range, 6–9) always separated from femoral
pore series, 16–28 TotPore (22.5 ± 4.01).
In life, light dusky tan to reddish beige ground color
dorsally and laterally from head to hips, narrow dark
brown ocular stripe on posterior edge of loreal area and
continuing behind eye either slightly or to edge of temporal
region (jowl), a narrow dark lateral stripe on neck broken
and reduced to complete stripe, no dark lateral stripe but
in some individuals widely spaced series of dark spots or
lines; dorsally on head some dark brown bilateral spotting,
continuing onto trunk as parasagittal row of small dark
marks; dorsolateral series of yellow spots on each side from
jowl to anterior arm of postsacral mark; bright postsacral
mark with middorsal dark brown spot bordered behind
and laterally by light orange to pinkish yellow V, arms of
which extending to posterior edge of abdomen. Tail distinctly lighter (yellowish) than trunk, widely spaced narrow
orangish rings and middorsal diffuse dark spots in brown
interspaces. Venter similar to dorsal ground color, a shade
lighter; underside of tail light to bold yellow.
In alcohol, pattern same as above, generally ground
color darker from light to medium brown; dark stripes
and other markings remain distinct, lighter marks lose
color becoming white to light tan, including the orange
border of postsacral mark; venter strikingly lighter than
dorsum, retaining a dusky appearance owing to one or
two small brown spots in most ventral scales.
Major diagnostic features are as follows: bisexual
taxon; pigmented caecum and gonadal ducts; in adult
males femoral pore series separate from precloacal pore
series, absent in females; chin scales bordering mental and
irst infralabial not greatly enlarged; digital lamellae formulae usually 3-4-4-3 (forefoot) and 3-4-4-4 (hindfoot);
average adult SVL ~31–32 mm; mute coloration of light
brown background with small scattering of dark brown
spots, contrasting with a bright postsacral marking with
small dark brown center with light orange V-shaped border and series of widely spaced dorsolateral orange spots.
Description of holotype:
An adult male (Figure
18), 31.2 mm SVL, 23 mm TailL (regenerated), 16.9 mm
TrunkL, 7.0 mm HeadL, 4.0 mm HeadW, 2.8 mm SnEye,
1.9 mm NarEye, 1.7 mm EyeD, and 1.2 mm SnW. Proportions: 53% TrunkL/SVL, 22% HeadL/SVL, 12% HeadW/
SVL, 57% HeadW/HeadL, 38% SnEye/HeadL, 30% NarEye/HeadL, 25% EyeD/HeadL, 16% SnW/HeadL, 84%
EyeD/NarEye, 28% SnW/HeadW. Scalation: 3 CircNa, 5
SnS, 8 Suplab, 9 Inlab, 12 Chin (anteromedial ones only
slightly larger than adjacent ones), 16 Dorsal, 11 Ventral, 4
CloacS, Subcaud not enlarged, 9 PreclPor, 25 TotPore with
FIGURE 18. Holotype of Hemiphyllodactylus ganoklonis (USNM 563682) from Ulebsechel Island, Palau. (Photograph by G. Zug.)
number 631
no contact between precloacal and femoral, digital formulae
3-4-4-4 forefoot and 4-4-5-4 hindfoot. Pigmented caecum,
testes lightly pigmented, no pigment on anterior portion of
epididymis, heavily pigmented posterior two-thirds.
In alcohol, body ground color light brown with medium to dark brown markings, dorsolateral light spotting
from eye to inguina (4 between eye and axilla, 9 between
axilla and inguina (right side), partial dark dorsolateral
stripe from eye to neck, no lateral stripes elsewhere, series
of widely spaced small dark dashes and spots parasagittally
and fewer laterally, postsacral dark brown chevron middorsally at tail base edged laterally by broad white border.
Tail background slightly darker than trunk with widely
scattered dark lecks. Ventrally dusky cream from chin to
vent because many ventral scales with central dark spot.
Etymology.
The name ganoklonis derives
from the Greek adjective and noun, ganos for bright or
brightness and klonis for buttock or rump. The bright
rump refers to the bright yellow chevron (postsacral)
mark at the base of the tail. The name is proposed as an
adjective.
Variation.
The means or medians and ranges
are detailed in the preceding Description section. Hemiphyllodactylus ganoklonis is the smallest Hemiphyllodactylus taxon, yet it possesses a slight, but signiicant,
size dimorphism with females averaging 2 mm larger than
males. This larger size in females also is relected in other
mensural traits: TrunkL, HeadL, HeadW, NarEye, and
EyeD. Body proportions and scalation are not dimorphic.
Overall, variation within each trait is low and often
the lowest of all regional samples. This low variation relects the high quality of preservation of the Palauan sample and is not due to a limited geographic sampling. The
Palauan sample derives from seven different islands and
multiple sites on a few islands.
Distribution.
Hemiphyllodactylus ganoklonis
occurs throughout the major islands of Palau (Figure 19).
Hemiphyllodactylus harterti Werner
Bintang slender gecko
Lepidodactylus Harterti Werner, 1900:196 [type locality: “Malakka” (Malaysia), restricted to “Gunong Inas” (Perak, Malaysia); holotype, ZMB
15360].
Gehyra larutensis Boulenger, 1900:188 [type locality: “Larut Hills, . . . , at
3500 feet altitude” (Malaysia); holotype, BMNH 1901.3.20.2].
Comment.
The type locality of H. harterti was
tentatively restricted by Boulenger in a footnote (1912: 48):
“Dr. Hartert collected on Gunong Inas, the type locality of
•
43
FIGURE 19. Geographic occurrence of Hemiphyllodactylus ganoklonis in Palau Islands. Circles indicate specimens represented by museum vouchers. (Image modiied from map by D. Dalet.)
G. larutensis, and this should, perhaps be substituted for
“Malacca,” over 200 miles distant.” Following my review
of E. Hartert’s publications and the locality data from his
bird collection, it is evident that Boulenger’s assessment is
correct and his restriction should be followed.
The type of Gehyra larutensis Boulenger and other
specimens from Boulenger’s descriptions of new frogs
and reptiles from the Larut Hills, Perak, are listed as in/
from the Selangor Museum. Presumably, that is their origin, but it appears that Boulenger retained them because
the catalog number (1901.3.20.2) of G. larutensis indicates that it became part of the British Museum collection
in 1901. My identiication of the holotype relies on the
44
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
BMNH’s labeling of the specimen as type, and in most
features my measurements and counts match those of
Boulenger. My count of the continuous precloacal–femoral pore series, however, is six pores less than Boulenger’s
count of 42.
I have seen only a single specimen of H. harterti, that
is, the holotype of H. larutensis. In spite of the insuficiency of my observations, I recognize this taxon owing
to a unique coloration (from published images) and the
values of several mensural and meristic traits lying on the
edges of the ranges of those traits for geckos of the Titiwangsa mountain range.
Description.
A bisexual taxon of geckos
(Gekkoninae) with robust habitus, slightly compressed
trunk and moderately large head (Figures 3, 11, 12), tail
round to elliptical in cross section and somewhat shorter
than SVL. An adult male (holotype of H. larutensis; Figure
12), 35.3 mm SVL, 31 mm TailL, 17.8 mm TrunkL, 8.3
mm HeadL, 6.3 mm HeadW, 3.5 mm SnEye, 2.6 mm NarEye, 2.1 mm EyeD, and 1.8 mm SnW. Proportions: 50%
TrunkL/SVL, 24 % HeadL/SVL, 18% HeadW/SVL, 75%
HeadW/HeadL, 42% SnEye/HeadL, 31% NarEye/HeadL,
25% EyeD/HeadL, 22% SnW/HeadL, 81% EyeD/NarEye,
29% SnW/HeadW. Scalation: 2 CircNa, 3 SnS, 10 Suplab,
10 Inlab, 9 Chin (anteromedial ones strongly enlarged,
≥2× larger than adjacent ones that are also enlarged), 15
Dorsal, 6 Ventral, 2 CloacS, Subcaud not enlarged, 36
TotPore precloacal and femoral pores continuous with
demarcation between two series, digital formulae 3-3-3-3
(forefoot) and 3-3-4-3 (hindfoot). Caecum not pigmented,
pigmentation unknown for testis epididymis. Aside from
coloration, female morphological traits are detailed in the
holotype description below.
Head to trunk dorsally and laterally a yellowish to
dusky tan ground color; tail distinctly brighter (lighter)
than body and limbs. Light spots in a dorsolateral series
from neck to postsacral mark or dark brown dorsolateral
stripe from neck merging into dark median border of postsacral mark. Tail uniformly colored.
Major diagnostic features are as follows: bisexual
taxon; caecum and (likely) gonadal ducts not pigmented;
precloacal–femoral pore series continuous; chin scales
bordering mental and irst infralabial distinctly enlarged;
digital lamellae formulae 3-3-3-3 (forefoot) and 3-3-4-3
(hindfoot); adult SVL < 40 mm; dorsum of head and trunk
either nearly uniform tan or with narrow dark dorsolateral stripes and contrasting with lighter tail, outer edge of
postsacral mark continuous with caudal color.
Description of holotype:
An adult female (Figure
12) , 40.9 mm SVL, 39 mm TailL, 21.5 mm TrunkL, 9.7
mm HeadL, 6.3 mm HeadW, 4.0 mm SnEye, 3.2 mm NarEye, 2.9 mm EyeD, and 2.0 mm SnW. Proportions: 53%
TrunkL/SVL, 24% HeadL/SVL, 15% HeadW/SVL, 65%
HeadW/HeadL, 41% SnEye/HeadL, 33% NarEye/HeadL,
30% EyeD/HeadL, 21% SnW/HeadL, 91% EyeD/NarEye, 32% SnW/HeadW. Scalation: 3 CircNa, 2 SnS, 10 Suplab, 11 Inlab, 6 Chin (anteromedial ones enlarged, twice
as large as adjacent ones that are also enlarged), ? Dorsal,
? Ventral, 1 CloacS, Subcaud not enlarged, 0 PreclPor, 0
TotPore, digital formulae not known. No pigmentation on
caecum or oviducts.
Specimen faded to uniform beige dorsally and laterally, somewhat lighter ventrally. Evidence of a dark postsacral mark.
Etymology.
Werner (1900) noted that a
single specimen (holotype) of this gecko was collected in
Malakka by Hartert and deposited in the Berlin collection. Presumably, the Hartert referred to by Werner is Ernest Johann Otto Hartert, an ornithologist who served as
the bird curator in L. W. Rothschild’s private museum at
Tring, UK, between 1892 and 1929. Prior to his employment at Tring, Hartert visited Asia and elsewhere and collected birds, insects, and other animals. He reported his
research travels in a popular book, Aus den Wanderjahren
eines Naturforschers (Hartert, 1901–1902).
Distribution.
Presently known from Larut
Hills and Gunong Inas, Perak (Figure 20); presumably,
it occurs throughout the forest of the Bukit Bintang
mountains.
Hemiphyllodactylus insularis Taylor
Philippine slender gecko
Hemiphyllodactylus insularis Taylor, 1918:237 [type locality: “Sumagui,
Mindoro” (Philippines); holotype, CM 2052].
Description.
Adults dimorphic, females
larger than males: 29.6–37.3 mm (mean ± SD, 33.9 ±
1.82; n = 15 females), 28.8–34.4 mm (31.3 mm ± 1.86,
n = 19 males) SVL; 15.5–19.3 mm (17.8 mm ± 1.08),
14.3–17.7 mm (15.8 mm ± 1.05) TrunkL; 6.6–8.6 mm
(7.6 mm ± 0.51), 6.2–8.2 mm (7.2 mm ± 0.51) HeadL;
3.8–5.7 mm (4.8 mm ± 0.48), 4.0–5.8 mm (4.8 mm ±
0.46) HeadW; 2.6–3.7 mm (3.0 mm ± 0.28), 2.1–3.4 mm
(2.8 mm ± 0.28) SnEye; 2.0–2.6 mm (2.3 mm ± 0.20),
1.9–2.6 mm (2.2 mm ± 0.17) NarEye; 1.8–2.1 mm (2.0
mm ± 0.11), 1.6–2.1 mm (1.9 mm ± 0.14) EyeD; 1.2–1.4
mm (1.3 mm ± 0.08), 1.0–1.5 mm (1.3 mm ± 0.16) SnW.
Adult proportions not dimorphic: 45–58% TrunkL/SVL
(mean ± SD, 50.1% ± 2.9), 21–24% HeadL/SVL (22.7%
number 631
•
45
FIGURE 20. Geographic occurrence of Hemiphyllodactylus harterti, H. margarethae, H. titiwangsaensis, and bisexual specimens from Borneo. Symbols: circle, H. harterti; diamond, H. titiwangsaensis; solid square, H. margarethae; open square, bisexuals.
± 0.1), 12–18% HeadW/SVL (14.9% ± 0.1), 54–82%
HeadW/HeadL (65.7% ± 5.9), 26–44% SnEye/HeadL
(39.4% ± 3.1), 27–42% NarEye/HeadL (31.2% ± 2.5),
23–32% EyeD/HeadL (26.1% ± 1.6), 14–21% SnW/
HeadL (17.2% ± 1.8), 74–95% EyeD/NarEye (84.0% ±
5.2), 20–37% SnW/HeadW (26.3% ± 3.2%).
Scalation is predominantly granular from head onto
tail, both dorsally and ventrally; ventral trunk scales
slightly larger than dorsal ones, 13–18 Dorsal (median ±
SD, 16 ± 1.4) and 8–14 Ventral (11 ± 1.6); similarly, subcaudal scales slightly larger than dorsal caudal scales but
not plate-like. Cloacal spurs present, modest sized, 0–3
CloacS (1 ± 0.9). Larger scales on lips and snout, rostral
largest, rectangular to pentagonal, often slightly concave
on dorsomedial edge with slight cleft; 1–4 CircNa (3 ±
0.8), 2–4 SnS (3 ± 0.6); labial scales enlarged from rostral
to below eye, becoming progressively smaller in subocular
rictus, 9–13 Suplab (10 ± 1.2), 9–11 Inlab (10 ± 0.6);
8–14 Chin (11 ± 1.5), those behind mental slightly or not
enlarged; ear opening distinct with no bordering enlarged
scales. Each digit with expanded pad, terminal two phalanges free, arising from within pad on second to ifth digits of fore- and hindfoot and each clawed; pads of these
digits each with large triangular apical lamella bordered
proximally by lyre-shaped lamellae (scansors); modal
digital formulae 3-3-3-3 (forefoot) and 3-4-4-4 (hindfoot)
for scansors; irst digit of fore- and hindfeet compressed,
usually 4 or 5 rectangular lamellae (2–5 fore, 3–6 hind)
ventrally, terminal phalanx not free with or without minute claw. Adult females never with precloacal pores, males
always with precloacal pores (median ± SD, 9 ± 1.5; range,
6–13) always separated from femoral pore series, 17–38
TotPore (27 ± 5.35).
In alcohol, light to medium brown ground color dorsally and laterally from head to tail; top of head with scattering of small dark brown blotches, lateral dark brown
stripe from loreal to shoulder variously developed (barely
visible to sharply deined); dorsally on trunk, dark blotches
46
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
variously arranged from series of parasagittal elongate
spots through randomly arranged spots creating irregular
narrow transverse bars to nearly absent; dark lateral stripe
on trunk typically series of lateral dark spots or blotches;
dorsolateral series of light spots from temporal area to inguina, posteriormost one enlarged and forming anterior
end of lateral light arm of postsacral mark; well-deined
postsacral mark in all individuals with large pentagonal
middorsal dark brown mark bordered behind and on sides
by light (whitish) base and arms, which are edged laterally
and caudally in dark brown. Tail usually lighter brown
than trunk with amorphous dark smudges middorsally.
Venter, chin to tail, dusky cream owing to dark spot in
most ventral scales.
In life, light dusky tan to reddish brown ground color
dorsally and laterally from head to hips; pattern of markings as described for preserved individuals. Light dorsolateral spots and light area of postsacral mark brick red.
Major diagnostic features are as follows: bisexual
taxon; pigmented caecum and oviducts; no precloacal–
femoral pores in females, present in males, precloacal and
femoral pore series separated; chin scales bordering mental
and irst infralabial not greatly enlarged; digital lamellae
formulae 3-3-3-3 (forefoot) and 3-4-4-4 (hindfoot); average adult SVL ~34, 31 mm (females, males); series of red
spots dorsolaterally on trunk and bright postsacral bar of
red and dark brown.
Description of holotype:
An adult male (Figure
21), 30.2 mm SVL, 14.7 mm TrunkL, 6.9 mm HeadL,
2.6 mm HeadW, 4.0 mm SnEye, 1.9 mm NarEye, 1.8
mm EyeD, and 1.2 mm SnW. Proportions: 49% TrunkL/
SVL, 23% HeadL/SVL, 13% HeadW/SVL, 58% HeadW/
HeadL, 38% SnEye/HeadL, 28% NarEye/HeadL, 26%
EyeD/HeadL, 17% SnW/HeadL, 95% EyeD/NarEye, 30%
SnW/HeadW. Scalation: 3 CircNa, 3 SnS, 10 Suplab, 11 Inlab, 11 Chin (anteromedial ones only slightly larger than
adjacent ones), 15 Dorsal, 11 Ventral, 3 CloacS, Subcaud
not enlarged, 9 PreclPor, 27 TotPore precloacal and femoral series separated, digital formulae 3-3-3-3 (forefoot) and
3-4-4-4 (hindfoot). Pigmented caecum, pigmentation unknown for testis epididymis.
Faded, body ground color brown.
Etymology.
Taylor (1918) did not explain
his choice of the epithet insularis, presumably because
he assumed the name was self-explanatory; insularis is a
Latin adjective for of islands.
FIGURE 21. Holotype of Hemiphyllodactylus insularis Taylor, 1918 (CM 2052): (A) dorsal view
of whole body, (B) ventral view of head, and (C) ventral view of posterior half of trunk. (Photographs by M. McNaugher.)
number 631
•
47
Variation.
The means or medians and ranges
are detailed in the preceding Description section. Males
are smaller, statistically signiicantly so, than females, but
the difference in average size is slight (~2.5 mm). None
of the meristic traits shows signiicant dimorphism among
adults, other than presence of precloacal and femoral
pores in males and their absence in females. Variation of
most scalation traits is modest to low with the values of
most traits equaling the median. Chin scales are usually
small, although a few individuals have a modest enlargement of those touching the mental and irst supralabial.
There are six digital formulae each for forefoot and hindfoot. Forefoot formulae range from 3-3-3-3 (66%) to 3-44-4 with only 3-3-4-3 (17%) also occurring in more than
two individuals. Hindfoot formulae range from 3-3-4-3 to
4-5-5-4; 3-4-4-4 is the most frequent (49%), followed by
3-4-4-3 (17%) and 4-4-5-4 (14%).
Distribution.
Hemiphyllodactylus insularis
occurs throughout the Philippine Islands (Figure 22) from
Mindoro to Mindanao and westward on both the Palawan
and Sulu Archipelago arcs. The presence on both these latter island groups recommends a reexamination of the bisexual Bornean Hemiphyllodactylus.
Hemiphyllodactylus margarethae Brongersma
Sumatran slender gecko
Hemiphyllodactylus margarethae Brongersma, 1931:11 [type locality: “Fort
de Kock, Sumatra” (Bukittinggi, Sumatera Barat); holotype, ZMA
11095].
Comment.
Brongersma (1932:218 [footnote])
noted that while the H. margarethae description was in
press and after he had examined additional Hemiphyllodactylus specimens, he attempted to suppress the new
name in page proofs, but his recommended changes were
not made.
Description.
A bisexual taxon of geckos
(Gekkoninae) with robust habitus, slightly compressed
trunk and moderately large head (see Figures 3, 23), tail
round in cross section and subequal to SVL. Adults not dimorphic: 36.0–46.9 mm SVL (mean ± SD, 40.8 mm ± 3.6],
n = 8), 14.5–25.4 mm TrunkL (20.6 mm ± 3.3), 8.2–10.4
mm HeadL (9.6 mm ± 0.87), 5.8–8.2 mm HeadW (6.8
mm ± 0.86), 3.3–4.7 mm SnEye (4.1 mm ± 0.56), 2.7–3.6
mm NarEye (3.2 mm ± 0.40), 1.6–2.5 mm EyeD (2.1 mm
± 0.35), 1.4–1.9 mm SnW (1.7 mm ± 0.23). Adult proportions: 40–54% TrunkL/SVL (mean ± SD, 50.3% ± 4.8),
21–26% HeadL/SVL (23.5% ± 1.6), 15–19% HeadW/
SVL (16.7% ± 1.3), 66–79% HeadW/HeadL (71.2% ±
FIGURE 22. Geographic occurrence of Hemiphyllodactylus insularis in the Philippine Islands. Not all localities in the same area are
plotted. Circles indicate specimens represented by museum vouchers
and whose speciic identity is conirmed.
4.4), 39–46% SnEye/HeadL (42.6% ± 2.4), 29–35% NarEye/HeadL (33.0% ± 2.0), 16–25% EyeD/HeadL (22.0%
± 3.6), 15–19% SnW/HeadL (17.3% ± 1.4), 47–81%
EyeD/NarEye (67.0% ± 12.6), 22–28% SnW/HeadW
(24.3% ± 1.9%).
Scalation is predominantly granular from head onto
tail, both dorsally and ventrally; ventral trunk scales slightly
larger than dorsal ones, 11–17 Dorsal (median ± SD, 12.5
± 2.3) and 6–12 Ventral (7.5 ± 2.3); similarly, subcaudal
scales slightly larger than dorsal caudal scales but not platelike. Cloacal spurs present, modest sized, 1–2 CloacS (2 ±
0.5). Larger scales on lips and snout, rostral largest, rectangular to pentagonal, often slightly concave on dorsomedial
48
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
edge with slight cleft; 2–3 CircNa (3 ± 0.4), 2–4 SnS (3 ±
0.6); labial scales enlarged from rostral to below eye, becoming progressively smaller in subocular rictus, 10–13
Suplab (11 ± 1.2), 9–12 Inlab (10 ± 0.9); 6–11 Chin (7.5 ±
2.0), those behind mental distinctly enlarged; ear opening
distinct with no bordering enlarged scales. Each digit with
expanded pad, terminal two phalanges free, arising from
within pad on second to ifth digits of fore- and hindfoot
and each clawed; pads of these digits each with large triangular apical lamella bordered proximally by lyre-shaped
lamellae (scansors); modal digital formulae likely 4-4-4-4
(forefoot) and 4-5-5-5 (hindfoot) for scansors; irst digit
of fore- and hindfeet compressed, usually 5 rectangular lamellae (4–8 fore, 4–7 hind) ventrally, terminal phalanx not
free with or without minute claw. Adult females often with
precloacal pores series (0–12 PreclPor), males (only one
adult in sample) always with precloacal pores (median 11)
always separated from femoral pore series, 0–29 (female)
26 (male) TotPore.
Coloration, no images of living specimens and preserved specimens pattern indistinct owing to fading.
Major diagnostic features are as follows: bisexual
taxon; unpigmented caecum, oviducts pigmented or not; if
present, femoral pore series separate from precloacal pore
series; chin scales bordering mental and irst infralabial
distinctly enlarged; digital lamellae formulae 4-4-4-4 (forefoot) and 4-5-5-5 (hindfoot); average adult SVL ~41 mm.
Description of holotype:
An adult male (Figure
23), 38.8 mm SVL, 40 TailL, 18.7 mm TrunkL, 9.9 mm
HeadL, 7.5 mm HeadW, 4.3 mm SnEye, 3.4 mm NarEye, 1.6 mm EyeD, and 1.9 mm SnW. Proportions: 48%
TrunkL/SVL, 25% HeadL/SVL, 19% HeadW/SVL, 76%
HeadW/HeadL, 43% SnEye/HeadL, 34% NarEye/HeadL,
16% EyeD/HeadL, 19% SnW/HeadL, 47% EyeD/NarEye,
25% SnW/HeadW. Scalation: 3 CircNa, 3 SnS, 10 Suplab,
11 Inlab, 6 Chin (anteromedial ones distinctly enlarged,
see Figure 7C), 13 Dorsal, 7 Ventral, 2 CloacS, Subcaud
not enlarged, 11 PreclPor, 26 TotPore, separate precloacal
and femoral series, digital formulae 4-5-5-5 (forefoot) and
5-5-6-5 (hindfoot). Caecum and testis epididymis presumably unpigmented.
Body ground color faded to pinkish beige, small
paired dark blotches dorsally on trunk; paratype (ZMA
11096) has a hint of a postsacral mark.
Etymology.
The origin of the epithet margarethae is unknown.
Variation.
The means or medians and ranges
are detailed in the preceding Description section. Sample
is small and specimens old and generally poorly preserved;
hence comments on variation are inappropriate.
FIGURE 23. Types of Hemiphyllodactylus margarethae Brongersma, 1931: left, ZMA 11095 holotype, male; right, 11096 paratype, female. (Photograph by G. Zug.)
Distribution.
With a single exception, the
H. margarethae specimens derive from the central mountain ranges of central and northern Sumatra (Figure 20),
extending from Bukittinggi in the south to Takengon in
the north. The Nias island locality is the exception and
requires conirmation.
Hemiphyllodactylus titiwangsaensis
Zug, new species
Titiwangsan slender gecko
holotype.
ZRC 2.4782, adult male from
Malaysia, Pahang Province, Cameron Highlands, Gunong
Brinchang (= Berincang) summit area, collected by H. H.
Tan and others, 25 June 2000.
Paratypes.
All subsequent specimens are from
Malaysia, Pahang Province; that datum is removed from
each subsequent locality for brevity. ZFMK 32284–286,
Cameron Highlands, Tana Ratah (=Tanah Rata), collected by Dietmar Kiehlmann, July 1980; ZRC 2.4780–
781, 2.4783–785, collecting data same as holotype; ZRC
2.4832, Cameron Highlands, Tanah Rata, Bala’s Lodge,
collected by H. H. Tan, 4 May 2000; ZRC 2.5165, Cameron Highlands, Parit Falls, T. M. Leong, and L. J. Lim, 30
July 2001; ZRC 2.5419, Cameron Highlands, Parit Falls,
collected by B. L. Lim and K. K. P. Lim, 10 May 2002;
number 631
ZRC 2.5942, Cameron Highlands, Telom Valley, Kuala
Terla 4000–4500′, collector unknown, March 1935; ZRC
2.5943, Cameron Highlands, Telom Valley, Gunong Siku
at ~4500′, collector unknown, March 1935. Adult females: ZFMK 32284, 32286; ZRC 2.4780–4781, 2.4785,
2.4832, 2.5165, 2.5943; adult males: ZFMK 32385; ZRC
2.4783–4784, 2.5419, 2.5942.
Description.
Bisexual taxon of geckos (Gekkoninae) with robust habitus, slightly compressed trunk
and moderately large head (see Figures 3, 11, 24), tail
round to elliptical in cross section and somewhat shorter
than SVL. Adults not dimorphic: 36.5–62.1 mm (mean ±
SD, 49.2 mm ± 6.34; n = 15) SVL; 18.1–32.5 mm (23.3
mm ± 3.69) TrunkL; 8.9–13.8 mm (12.0 mm ± 1.18)
HeadL; 6.1–10.2 mm (8.2 mm ± 0.98) HeadW; 3.4–5.8
mm (4.9 mm ± 0.60) SnEye; 2.6–4.4 mm (3.6 mm ± 0.42)
NarEye; 2.4–3.6 mm (3.0 mm ± 0.31) EyeD; 1.4–2.6
mm (2.0 mm ± 0.32) SnW. Adult proportions: 42–50%
TrunkL/SVL (mean ± SD, 46.4% ± 3.4), 23–27% HeadL/
SVL (24.6% ± 1.4), 16–19% HeadW/SVL (16.9% ± 1.0),
64–73% HeadW/HeadL (68.5% ± 3.5), 37–43% SnEye/
HeadL (40.7% ± 2.1), 26–32% NarEye/HeadL (29.4% ±
2.1), 21–30% EyeD/HeadL (25.6% ± 2.5), 14–21% SnW/
HeadL (16.5% ± 2.5), 68–94% EyeD/NarEye (85.4% ±
9.6), 21–31% SnW/HeadW (23.8% ± 3.2%).
Scalation is predominantly granular from head onto
tail, both dorsally and ventrally; ventral trunk scales slightly
larger than dorsal caudal scales, 14–19 Dorsal (median ±
SD, 16 ± 1.9) and 7–9 Ventral (7 ± 1.0); similarly, subcaudal scales slightly larger than dorsal ones but not plate-like.
Cloacal spurs present, modest sized, 1–4 CloacS (3 ± 1.0).
Larger scales on lips and snout, rostral largest, rectangular
to pentagonal, often slightly concave on dorsomedial edge
with slight cleft; 3 CircNa (3 ± 0.0), 1–3 SnS (3 ± 0.8); labial scales enlarged from rostral to below eye, becoming
progressively smaller in subocular rictus, 9–11 Suplab (10 ±
0.7), 8–10 Inlab (9 ± 0.7); 8–9 Chin (9 ± 0.5), those behind mental distinctly enlarged; ear opening distinct with
no bordering enlarged scales. Each digit with expanded
pad, terminal two phalanges free, arising from within pad
on second to ifth digits of fore- and hindfoot and each
clawed; pads of these digits each with large triangular apical lamella bordered proximally by lyre-shaped lamellae
(scansors); modal digital formulae 3-4-4-4 (forefoot) and
4-5-5-5 (hindfoot) for scansors; irst digit of fore- and hindfeet compressed, usually 5 or 7 rectangular lamellae (4–6
fore, 5–8 hind) ventrally, terminal phalanx not free with or
without minute claw. Adult females never with precloacal
pores; males always with continuous precloacal–femoral
pore series 17–39 TotPore (median ± SD, 21 ± 7.95).
•
49
In life, dorsal and lateral ground color ranges from
light grayish tan to medium brown, head to tail occasionally distinctly lighter than neck and trunk (Figure 11).
This lightness is emphasized by absence or diffuseness of
dark markings on head. Neck and trunk bear numerous
transverse dark brown irregularly shaped bars, lighter interspaces typically narrower than dark bars. Bars extend
onto sides; dorsolaterally in shoulder area bars are darker,
creating an impression of dark dorsolateral stripe. Dark
lateral stripe from loreal to neck, occasionally to midneck.
Dark stripe bordered above by cream to beige stripe from
canthus rostralis to shoulder, often continuing as series of
spots or dashes on trunk and at inguina becoming narrow
arm of postsacral mark; center dark spot of mark absent to
small. Tail usually lighter than trunk and distinctly banded
in light and dark, relative size of which very variable. (Coloration from images by H. Ota and Chan-ard et al., 1999.)
Coloration in alcohol is muted, although dark and
light pattern usually persists. Ventrally from chin onto tail,
uniform light cream in most individuals, brown in a few.
Females seem to be more boldly patterned than males.
Major diagnostic features are as follows: bisexual
taxon; caecum and gonadal ducts not pigmented; precloacal–femoral pore series continuous in males (TotPore
17–39), absent in females; chin scales bordering mental
and irst infralabial distinctly enlarged; digital lamellae
formulae usually 3-4-4-4 (forefoot) and 4-4-5-5 or 4-5-5-5
(hindfoot); average adult SVL ~49 mm; dorsal and lateral
trunk pattern of dark brown irregular transverse bands,
muted postsacral bar of narrow white arms onto hips.
Description of holotype:
An adult male (Figure
24), 56.9 mm SVL, 48 mm TailL (regenerated), 24.1 mm
TrunkL, 13.2 mm HeadL, 9.6 mm HeadW, 5.4 mm SnEye,
3.9 mm NarEye, 3.0 mm EyeD, and 2.4 mm SnW. Proportions: 42% TrunkL/SVL, 23% HeadL/SVL, 17% HeadW/
SVL, 73% HeadW/HeadL, 41% SnEye/HeadL, 28% NarEye/HeadL, 23% EyeD/HeadL, 18% SnW/HeadL, 81%
EyeD/NarEye, 25% SnW/HeadW. Scalation: 3 CircNa, 3
SnS, 10 Suplab, 9 Inlab, 8 Chin (anteromedial ones enlarged), 16 Dorsal, 7 Ventral, 3 CloacS, Subcaud not enlarged, precloacal and pore series continuous, 30 TotPore,
digital formulae 4-4-5-4 (forefoot) and 4-5-5-5 (hindfoot).
No pigmentation on caecum or oviducts.
Specimen brown dorsally and laterally with scattered
indistinct dark brown markings, somewhat lighter ventrally. Postsacral mark indistinct, small median dark spot
on irst tail segment, anterior arms muted. In life, the ventral surface of the tails (type series) were orangish pink.
Etymology.
These geckos occur in the south
central region of the Banjaran Titiwangsa; hence the
50
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
FIGURE 24. Types of Hemiphyllodactylus titiwangsaensis: left to right, ZRC 2.4780–2.4785; 2.4782 is the holotype. (Photograph by
G. Zug.)
taxon is a resident (likely endemic) of Titiwangsa and so
named.
Variation.
The means or medians and ranges
are detailed in the preceding Description section. Adults
are not dimorphic, and both sexes have broad ranges
of adult sizes, females 42.2–62.1 mm SVL and males
36.5–56.9 mm. This broad range yields a modest increase
in variation among the mensural traits (CV = 8–17%).
Larger (≥46 mm SVL) adults are distinctly robust-bodied
geckos.
Scalation has low variation with the exception of phalangeal formulae. Seven forefoot formulae range from 2-33-3 to 4-5-5-5; however, 3-4-4-4 is the mode and median
with a uniform and sharp drop in number of individuals on each side of the mode. Hindfoot formulae (8) are
dominated by 4-4-5-5 (n = 3) and 4-5-5-5 (7); the other six
formulae observed have low (n = 1 each) representation;
3-3-3-3 is the lowest.
Coloration has two patterns. The transverse bar pattern (somewhat Lepidodactylus lugubris-like) described
above occurs in the majority of individuals, and a few
individuals show gradation from this pattern into a ragged
spotted one.
Distribution.
All specimens derive from localities within Banjaran Titiwangsa (Figure 20) and particularly from Fraser Hill and Cameron Highland areas.
At both sites, H. titiwangsaensis occurs in forest and on
and around buildings. I assume that these geckos occur
more broadly in moist evergreen forest of the Titiwangsa
mountain range and that the present vouchers represent
the most accessible areas.
Hemiphyllodactylus yunnanensis (Boulenger)
Asian slender gecko
Gehyra yunnanensis Boulenger, 1903:429 [type locality: “Yunnan Fu”
(= Kumning, Yunnan Province, China); lectotype, BMNH 1904.1.26.1].
Hemiphyllodactylus typus chapaensis Bourret, 1937:60 [type locality:
“Chapa” (Sa Pá [also Lao Cai], Vietnam); holotype, MNHN 1948.43].
Hemiphyllodactylus yunnanensis longlingensis Zhou and Yang in Zhou et
al., 1981:203 [type locality: Longling Junior High School in Longling
County, Yunnan Province (alt. 1530 m) (China) (original in Chinese)].
number 631
Hemiphyllodactylus yunnanensis jinpingensis Zhou and Yang in Zhou et al.,
1981:204 [type locality: Jinpling Junior High School in Jinping County,
Yunnan Province (alt. 1260 m) (China) (original in Chinese)].
Hemiphyllodactylus yunnanensis dushanensis Zhou and Yang in Zhou et al.,
1981:206 [type locality: Dushan Junior High School in Dushan County,
Guizhou Province (alt. 970 m) (China) (original in Chinese)].
Comments.
Readers are reminded that
H. yunnanensis is considered here to include the highland
populations of Hemiphyllodactylus across southern China
and adjacent northern Southeast Asia from Myanmar to
Vietnam. The taxonomic status of the “lowland” populations of Southeast Asia and Hong Kong remains unresolved owing to the sparsity of vouchers in numbers and
geography.
Boulenger (1903) identiied two syntypes “male and
young.” Both specimens (BMNH 1904.1.26.1–2; Figure 25)
are extant. I designate the adult male (BMNH 1904.1.26.1)
as the lectotype of Gehyra yunnanensis Boulenger.
Bourret said in the type description: “J’ai pris à Chapa
une femelle à queue reconstituée (S130) . . .” Brygoo
(1990:44) and I interpret this statement as a description
based on a single female specimen. This interpretation conlicts with Guibe’s (1954) type catalog listing of two specimens; the sex of neither is identiied by Guibe, although
FIGURE 25. Lectotype of Gehyra yunnanensis Boulenger, 1903
(BMNH 1904.1.26.1): (A) dorsal view of whole body, (B) ventral
view of throat and chin, and (C) ventral view of pelvic area. (Photographs by G. Zug.)
•
51
one is noted to be damaged and 56 mm long. Brygoo
noted that of the two specimens labeled syntypes, both
possess Bourret’s registration numbers and one (S130) is
unambiguously the holotype of H. typus chapaensis Bourret. Also, the holotype is 43 mm SVL (Brygoo 1990:44)
and 42.7 mm (my measurement), and 33 mm (Bourret,
1937:60). This disparity, yet similarity, suggests that Bourret accidentally entered 33 instead of 43.
Description.
A bisexual taxon of geckos
(Gekkoninae) with robust habitus, slightly compressed
trunk and moderately large head (Figures 3, 11, 25),
tail round in cross section and typically shorter than
SVL. Adults dimorphic, females larger than males: 31.9–
49.3 mm (mean ± SD, 40.7 mm ± 4.44; n = 33), 25.5–
46.4 mm (37.9 mm ± 4.58, n = 28) SVL; 15.4–26.5 mm
(19.5 mm ± 2.93), 12.8–22.5 mm (17.9 mm ± 2.30)
TrunkL; 7.6–11.5 mm (9.5 mm ± 0.98), 6.7–10.3 mm
(8.8 mm ± 0.89) HeadL; 5.4–8.4 mm, (6.8 mm ± 0.95), 4.8–
7.4 mm (6.3 mm ± 0.76) HeadW; 3.0–5.2 mm (3.9 mm ±
0.55), 2.4–4.7 mm (3.7 mm ± 0.50) SnEye; 2.1–4.0 mm
(3.0 mm ± 0.44), 1.9–3.4 mm (2.8 mm ± 0.37) NarEye;
1.9–3.2 mm (2.4 mm ± 0.31), 1.6–3.0 mm (2.2 mm ± 0.27)
EyeD; 1.0–2.2 mm (1.7 mm ± 0.25), 0.9–2.0 mm (1.5 mm
± 0.25) SnW. Adult proportions not dimorphic: 40–55%
TrunkL/SVL (mean ± SD, 47.4% ± 3.1), 21–26% HeadL/
SVL (23.3% ± 1.0), 14–22% HeadW/SVL (16.7% ± 1.8),
59–83% HeadW/HeadL (71.9% ± 6.5), 34–46% SnEye/
HeadL (41.4% ± 2.5), 26–35% NarEye/HeadL (31.4% ±
2.1), 22–29% EyeD/HeadL (25.1% ± 1.7), 11–22% SnW/
HeadL (17.3% ± 2.0), 63–100% EyeD/NarEye (80.3% ±
6.8), 15–36% SnW/HeadW (24.3 ± 3.6%).
Scalation is predominantly granular from head onto
tail, both dorsally and ventrally; ventral trunk scales
slightly larger than dorsal ones, 9–18 Dorsal (median ±
SD, 13 ± 1.8) and 6–12 Ventral (8 ± 1.1); similarly, subcaudal scales slightly larger than dorsal caudal scales
but not plate-like. Cloacal spurs usually present, modest sized, 0–2 CloacS (1 ± 0.3). Larger scales on lips and
snout, rostral largest, rectangular to pentagonal, often
slightly concave on dorsomedial edge with slight cleft;
2–4 CircNa (3 ± 0.2), 2–5 SnS (3 ± 0.7); labial scales
enlarged from rostral to below eye, becoming progressively smaller in subocular rictus, 8–13 Suplab (10 ± 1.0),
8–12 Inlab (10 ± 1.1); 6–11 Chin (8 ± 1.1), those behind mental moderately to distinctly enlarged; ear opening distinct with no bordering enlarged scales. Each digit
with expanded pad, terminal two phalanges free, arising
from within pad on second to ifth digits of fore- and
hindfoot and each clawed; pads of these digits each with
large triangular apical lamella bordered proximally by
52
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
lyre-shaped lamellae (scansors); modal digital formulae
3-3-3-3 (forefoot) and 3-4-4-4 (hindfoot) for scansors;
irst digit of fore- and hindfeet compressed, usually 4
and 5 rectangular lamellae (4–6 fore, 4–7 hind) ventrally,
terminal phalanx not free with or without minute claw.
Adult females rarely with precloacal–femoral pore series
(7–9 PreclPor, n = 2; 0–19 TotPore), males always with
continuous precloacal–femoral pore series 11–25 TotPore
(median ± SD, 20.0 ± 3.26).
In life, light grayish brown to medium reddish brown
ground color dorsally and laterally from head to tail
base; top of head with scattering of small dark brown
marks, lateral dark brown stripe from loreal to shoulder
bordered above by white to tan stripe to end of head;
dorsally on trunk, narrow dark brown transverse lines
to squiggles, dorsolateral medium-sized light spots from
neck to above hindlimb and laterally dark lateral stripe
occasionally across neck to anterior trunk and thereafter fragmented to dark dashes or diffuse brown marks;
typically, area between dorsolateral light spots and dark
lateral stripe lighter than dorsal ground color; postsacral
mark small to large blotch bordered behind by rectangular white spot, no anterior extensions; tail ground color
lighter than trunk with series of transverse blotches of
narrow dark brown bordered behind by broader area of
light tan. Venter dusky, tail base pinkish to light orange
blush.
In alcohol, pattern as above and fading toward a uniform light to medium brown with brown marks.
Major diagnostic features are as follows: bisexual
taxon; caecum and gonadal ducts not pigmented; precloacal–femoral pore series continuous in males (TotPore
11–25), usually absent in females; chin scales bordering
mental and irst infralabial distinctly enlarged; digital lamellae formulae usually 3-3-3-3 (forefoot) and 3-4-4-4
(hindfoot); average adult female SVL ~41 mm, males ~39
mm; dorsal trunk pattern of narrow dark brown irregular
transverse bands bordered dorsolaterally by longitudinal
series of light spots, postsacral bar of dark and light with
no anterior extensions dorsolaterally.
Description of lectotype:
An adult male, 40.3 mm
SVL, ~41 regenerated TailL, 20.1 mm TrunkL, 9.9 mm
HeadL, 7.8 mm HeadW, ~4.3 mm SnEye, ~3.3 mm NarEye, ~2.6 mm EyeD, and ~1.8 mm SnW. Proportions: 50%
TrunkL/SVL, 25% HeadL/SVL, 19% HeadW/SVL, 79%
HeadW/HeadL, ~44% SnEye/HeadL, ~34% NarEye/
HeadL, ~26% EyeD/HeadL, ~18% SnW/HeadL, ~79%
EyeD/NarEye, ~26% SnW/HeadW. Scalation: 3 CircNa,
3+ SnS, 9 Suplab, 9 Inlab, ±10 Chin (anteromedial ones
distinctly larger than adjacent ones), ~16 Dorsal, ~8
Ventral, 2 CloacS, Subcaud not enlarged, precloacal and
femoral pore series continuous 36 TotPore, digital formulae (estimate) 3-3-3-3 (forefoot) and 3-4-4-4 (hindfoot).
Pigmentation of caecum and testis epididymis unknown,
likely no pigmentation.
Body ground color grayish brown above and below,
scattered dark spots and small dark spots middorsally, no
lateral spotting on trunk; dark dorsolateral stripe from eye
to shoulder, lateral stripe from in front of eye to end of head.
Caecum not visible through body wall; not dissected
so unable to conirm gonadal pigmentation.
Etymology.
The name yunnanensis identiies
this species as the gecko from Yunnan, the type locality of
Boulenger’s new species.
Variation.
The means or medians and ranges
are detailed in the preceding Description section. Adults
are dimorphic in size, and both sexes have broad ranges of
adult sizes. The variation in mensural and meristic traits is
examined in the bisexual portion of the Character Analysis section.
Distribution.
Highlands of southwestern
China and adjacent uplands from the western edge of
the Shan Plateau in Myanmar, across northern Thailand,
Laos, and Vietnam (Figure 16). The southern limits of the
distribution are ill-deined owing to limited sampling of
these geckos through much of Asia.
KEY TO THE SPECIES OF HEMIPHYLLODACTYLUS
1.
Chin scales bordering mental scale posteriorly distinctly enlarged [Figure 7C,D] appear as a pair of scales labeled postmentals in other geckos; caecum and gonadal peritoneum white [Figure 2C] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1′. Chin scales bordering mental scale posteriorly slightly or not enlarged [Figure 7A,B], their size nearly same as more medial
chin scales; caecum and gonadal-duct peritoneum pigmentation usually black [Figure 2B] . . . . . . . . . . . . . . . . . . . . . . 2
2. Adult females with actively secreting precloacal and femoral pores; unisexual species, all individuals are females; adult size
often >36 mm SVL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. typus
2′. Adult females with no or fewer than ive secreting precloacal pores; populations of females and males; adult size seldom
>38 mm SVL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
number 631
3.
3′.
4.
4′.
5.
5′.
6.
6′.
7.
7′.
8.
8′.
•
Usually two U-shaped digital lamellae under fourth digit of forefoot; dorsal trunk pattern bold, transverse dark blotches,
longitudinal series of white dorsolateral spots and postsacral mark of dark brown and orange [Figure 11A] . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. aurantiacus
Three or four U-shaped digital lamellae under fourth digit of forefoot; dorsal trunk pattern muted, faded and small dark
blotches or widely separated dark spots [Figure 11F]; postsacral mark with U- or V-shaped outer edge of yellow or red;
dorsolateral spots yellow or red . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Total number precloacal and femoral pores (TotPore) usually <24 (16–28) in males; forefoot digital lamellar formula usually 3-4-4-3; postsacral mark outer edge yellow to pinkish yellow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. ganoklonis
Total number precloacal and femoral pores (TotPore) usually >24 (17–38) in males; forefoot digital lamellar formula usually 3-3-3-3; postsacral mark outer edge red . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. insularis
Precloacal and femoral pore series separate; females commonly with precloacal pores; forefoot digital lamellar formula
usually 4-4-4-4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. margarethae
Precloacal and femoral pore series continuous; females usually lack precloacal pores.: forefoot digital lamellar formula
3-3-3-3 or 3-4-4-4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Adults large, usually >45 mm SVL; precloacal–femoral pore series usually >22 (17–39) pores; hindfoot digital lamellar
formula usually 4-4-5-5 or 4-5-5-5; postsacral mark with anterior arms . . . . . . . . . . . . . . . . . . . . . . H. titiwangsaensis
Adults moderate size, usually <42 mm SVL; hindfoot digital lamellar formula usually 3-3-3-3 or 3-4-4-4, occasionally
higher; postsacral mark without anterior arms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Trunk usually with distinct dark dorsolateral stripe; precloacal and femoral pore series continuous with >30 pores . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. harterti
Trunk usually without dark dorsolateral stripe; precloacal and femoral pore series continuous with <26 pores . . . . . . 8
Precloacal and femoral pore series usually >18 pores; hindfoot digital lamellar formula usually 3-4-4-4; postsacral mark
of anterior dark blotch and posterior larger light bar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. yunnanensis
Precloacal and femoral pore series usually <18 pores; hindfoot digital lamellar formula usually 4-5-5-4; postsacral mark
absent or muted dark transverse bar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hong Kong population
53
Acknowledgments
I
offer my appreciation to numerous colleagues who have “kept the faith”
in my dogged effort to study the relationships of Oceania lizards. Thank
you all. There are more taxonomic studies in progress. I express my highest
appreciation to another group of colleagues and collections management
staffs worldwide. They have been generous in their assistance and courtesies
during my visits to their collections and have permitted me to borrow specimens
and retain some for an inordinately long time. I thank D. Kizirian, R. Pascocello, R. Zweifel (AMNH); R. Sadlier, A. Greer (AMS); C. McCarthy (BNMH);
C. Kishinami, K. Imada (BPBM); J. Vindum, R. Lucas, H. Brignall (CAS);
M. McNaugher and S. Rogers (CM); A. Resetar, C. Redhed, H. Voris (FNMH);
G. Lenglet, G. Coulon, M. Lang (IRSN); T. Hikida, M. Matsui, H. Ota (KUZ);
J. Rosado (MCZ); I. Ineich (MNHN); E. Kramer (NMB); F. Tiedemann,
R. Gemel (NMW); J. Covacevich, P. Couper (QM); M. Hoogmoed, J. Arntzen
(RMNH); M. Hutchinson (SAM); G. Pregill (SDMNH); A. Schulter, K. Kramer
(SMF); T. Chan-ard, S. Mekchai (THNMH); D. Auth, K. Krysko, W. King (UF);
R. How, L. Smith (WAM); private collection of William Beckon (WmBeckon);
L. van Tuijl (ZMA); R. Günther, M.-O. Rödel (ZMB); W. Böhme (ZMFK);
K. Lim (ZRC); U. Gruber (ZSM). Additionally, I thank the collection staff of
the National Museum of Natural History, Smithsonian Institution, who regularly support my specimen-based research: F. Blasdell, R. Crombie, S. Gotte,
T. Hartsell, K. Tighe, and R. Wilson.
Numerous individuals have aided my Hemiphyllodactylus research. R. Wilson did the initial georeferencing of Hemiphyllodactylus specimens and type localities. B. Trimmer assisted with an early phase of data veriication. T. Schwaner,
performed an electrophoretic analysis of Thai, Philippine, and Hawaiian Hemiphyllodactylus. A. Bauer, H. Ota, and I. Das have kept my interest in mind and
regularly apprised me of newly collected specimens and observations. R. Crombie made a special effort to obtain specimens from Palau and Hawai’i and provided ield notes and photographs of Palauan specimens. P.-P. van Dijk provided
notes and images of northern Thailand H. yunnanensis. H. Ota gave sketches,
notes, and images of Asian Hemiphyllodactylus. I. Das provided Hemiphyllodactylus images from India and Borneo, and J. R. H. Gibbons gave me images
56
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
of Fijian H. typus. M.-O. Rödel generously took data and
photographs of the type of Lepidodactylus Harterti Werner. L. Grismer challenged me to examine more closely the
nomenclatural status of harterti and larutensis. S. Rodgers
arranged for the photography of the type of H. insularis.
Jennifer Kilby [né Westhoff] of Luray Zoo and Molly Grifin of Redlands lent their artistic skills. T. Ulber translated
Bleeker’s type description of H. typus, and a volunteer
from Smithsonian Volunteer Services translated the Hemiphyllodactylus section from V. V. Bobrov and D. V. Semenov’s (2008) Lizard Fauna of Vietnam. My wife, Pat,
contributed to this research by her regular assistance in
data gathering, entry, and organization and by her review
of an early draft of the manuscript. Other colleagues—
A. Bauer, M. Cota, R. Crombie, R. Fisher, F. Kraus,
H. Ota, J. Vindum—read and commented on drafts of the
manuscript; their comments signiicantly improved the
quality of the current monograph. I am responsible for
remaining errors and misinterpretations. I thank all of the
above for their time and effort on my behalf.
Appendix 1: Character Deinitions
M
any of the characters examined and recorded in this study are
used broadly in other systematics studies of geckos. I use the
abbreviations proposed by me previously (Zug, 1998) for conciseness and for permitting quick identiication of the characters. Most characters are deined in Zug et al. (2003). Any not deined there or
that are deined differently for Hemiphyllodactylus are presented in Table A1.1
below. All measurements were recorded in millimeters to the nearest 0.1 mm
and from the right side; bilateral meristic characters were also recorded from
right side.
58
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
TABLE A1.1. Abbreviations and deinitions for characters examined.
Character class and abbreviation
Mensural characters
EyeD
HeadL
HeadW
NarEye
SnEye
SnW
SVL
TailL
TrunkL
Meristic characters of scalation
Chin
CircNa
CloacS
Dorsal
Inlab
PoreC
PreclPor
SnS
Character name
Deinition
Orbit diameter
Head length
Head width
Nares–eye length
Snout–eye length
Snout (internarial) width
Snout–vent length
Tail length
Trunk length
Maximum horizontal diameter of exposed eyeball
Chin scales
Number of scales touching internal edge of infralabials and
mental from juncture of 2nd and 3rd infralabials on left
and right
Circumnasal scales
Cloacal spurs
Dorsal scales
Femoral and precloacal pores series
Scales between supranasals
Subcaud
Suplab
Supralabial scales
TotPore
Ventral
Total number of secreting pores
Ventral scales
1FingLm
1ToeLm
2-5FingLm, 2-5ToeLm
First digit lamellae
First digit lamellae
Second to ifth digit lamellae
Meristic characters of coloration
CaecMel
OrbStrp
OvidMel
PostocS
Pigmentation of caecum
Postorbital stripe
Pigmentation of oviduct
Postocular spots
Number of scales longitudinally at midbody on dorsum
contained within one EyeD
As for Suplab
FemPor and PreclPor series continuous or separated
Number of scales touching rostral scale between left and right
supranasals
Scales subequal to dorsal scales or enlarged into plates
Number of enlarged scales from rostral to top of mouth
curve, usually equivalent to end of orbit
Total number of left and right femoral pores and PreclPor
Number of scales longitudinally at midbody on venter
contained within one EyeD
Number of lamellae (wider than long) on 1st digit of forefoot
Number of lamellae (wider than long) on 1st digit of hindfoot
Number of entire, U-shaped subdigital lamellae (=scansors)
on enlarged pad of 2nd to 5th digit, single apical lamella not
counted, only large U-shaped lamellae touching edge of pad
Caecum pigmented or not
Dark lateral stripe from eye to mid-neck or beyond, absent or
present
Oviduct pigmented or not
Number of light spots above OrbStrp from behind eye to
front of shoulder
Appendix 2: Specimens Examined
LOCALITY SAMPLES
The specimens are segregated by the geographic-speciic samples. These
samples are arranged from east to west and, for those within similar longitudinal bands, from north to south. The number of specimens listed below for a
locality can exceed the number of individuals in a locality sample [small cap
name in brackets] because data were not collected on all specimens due to size
or state or preservation, although the specimen’s speciic identity was conirmed.
Catalog numbers for primary type specimens are in bold.
Hemiphyllodactylus typus
HAWAIIAN ISLANDS [HAWAI].
No island given: AMNH 22340, MCZ
R20268, R154043, USNM 21220. Hawaii: USNM 23459–460, 310815–816,
518722, 570745–748. Kauai: USNM 163573, 23485, 23499, 23500, 279241.
Lanai: USNM 570736–744. Molokai: BPBM 1576, 6595, 6715–17. Maui: BPBM
11557–560, MCZ R1093, R174988. Oahu: BMNH 1903.2.21.5–7, BPBM
0863–64, 6158, 6567, FMNH 42251, 212245, USNM 23509, 58969, 59482,
59493–496, 59722–723, 279238–240.
POLYNESIA [POLYN].
Cook Islands, Mangaia: SDNHM 67822–824. French
Polynesia, Marquesas: BMNH 1926.1.20.38, 1926.1.20.50, FMNH 17914,
MNHN 1988.3034; Society Islands: MNHN without number, USNM 68047.
Henderson Island: BMNH 1913.1.17.1–17.3.
FIJI AND TONGA [FIJI].
Vanua Levu: AMS R107894, USNM 322442;
Viti Levu: AMNH 41689, BMNH 1938.8.2.7, QM J048853, J048898, USNM
230185, 267928, 267978–979, 310810–814, 345104, WmBeckon 80–82, 148,
169, 173, 175. Samoa: USNM 345102. ‘Eua: USNM 268045–046, 322119; Tongatapu: CAS 49971, USNM 268044, 322120; Vava’u: USNM 333617; TongaAta: BMNH 91.11.13.1.
NEW CALEDONIA AND VANUATU [NCAL].
New Caledonia: AMS R125697,
R125699, R125787–788, BMNH 71.4.16.30A–B, 85.11.16.8, CAS 172739,
MNHN 1887.270, NMB 6978. Vanuatu: FMNH 69613, ZSM R110.
60
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
taiwan and Japan [taiwan].
Ryukyu Islands: KUZ 018095–096. Taiwan: CM 118859, KUZ
009612, USNM/ield 123689, USNM 291710–712,
291807–810.
Philippines [Philip].
Palawan: CAS-SUR
28692–696, FMNH 52003, MCZ R150339. Saub: MCZ
R26082.
new guinea and solomon islands
[nguin].
New Guinea: AMNH 59045, 95887,
100200–206, CAS 192984, 192986–987, MCZ R49273,
R140954–140955, R145981, USNM 119246, 203865.
Solomon Islands: MCZ R79198, USNM 287441.
islands of indonesia and Malaysia
[sunda].
Malaysia: KUZ 18094, MCZ 43480,
140968. Singapore: BMNH 96.6.25.11, ZRC 2.2291,
2.3282, 2.3378, 2.3469, 2.5361, 2.5385–86, 2.5415–16,
2.5622, 2.6021–22, 2.6596. Borneo: BMNH 912287,
1959.115A–B, FMNH 63661–662, 138545, 158734,
196268A, 213665, 239661, 243789, KUZ 8732, 8746,
MCZ R43478, USNM 313965, ZMB 11355, ZRC
2.5671–72, 2.5675–78, 2.5955. Sumatra: BMNH
91.10.27.2, 1931.5.5.2, 1946.8.30.83, MCZ R38971,
NMW 179171, RMNH 4172 (3), 7161, SMF 23125,
30326, ZMA no number (3). Java: BMNH 85.12.3.17,
NMW 179172, RMNH 28008–014, SMF 8930, 22611,
USNM 44202, ZMB 31280, ZRC 2.1392. Bali: SMF
23126, WAM 109012. Komodo: UF 28878, 28985. East
Indies: RMNH 3991.
southeast asia [sEasia].
Thailand: ZRC
2.5367.
india and sri Lanka [india].
Sri
Lanka: BMNH 74.4.29.1326, 90.11.8.5, 1908.7.2.1,
1910.3.16.4, 1972.2108, NMB 8552.
Mascarenes [Mascar].
Mauritius: BMNH
1926.1.20.38, 1926.1.20.50, IRSN 24309, MCZ R51642–
643, USNM 149760, 565090–093, ZFMK 25350, ZMA
14717, 14766. Rodriquez: BMNH 1975.416.
Hemiphyllodactylus aurantiacus
india and sri Lanka [india].
India:
BMNH 74.4.29.1332, 74.4.29.1333, 74.4.29.1334–
1337, 91.11.27.1–3, 94.8.30.2, NMB 2900, 9682, NMW
14753, ZMB 10233, ZRC 2.4601, 2.4678–680.
Hemiphyllodactylus ganoklonis
Republic of Palau [Palau].
Babeldaob: SAM R47715, USNM 495065–066, 563663–666.
Ngeanges: USNM 563667. Ngeaur: USNM 563668.
Ngercheu: USNM 563669–674. Ngerekebesang: USNM
563675. Ngeruktabel: USNM 563676. Oreor: USNM
563677. Ulebsechel: USNM 563678–681, 563682,
563683.
Hemiphyllodactylus harterti
islands of indonesia and Malaysia
[sunda].
Malaysia: BMNH 1901.3.20.2, ZMB
15360.
Hemiphyllodactylus insularis
Philippines
[Philip].
Bohol:
CAS-SU
25107. Borocay: CAS 127889, 127965–971. Cancuman:
MCZ R26600. Cebu: CAS-SU 27310, 125228, 132632,
136844, 138320, 145922–929, CAS-SU 28451, 28602.
Great Govenen: CAS 60605. Mantique: CAS-SU 28987.
Mindoro: BMNH 26085, CAS 62065, CM 2052, 2053,
MCZ R26084, R26601, ZMA no number. Negros:
AMNH 86598–599, 115512, BMNH 1976.1681, CAS
131795, 137652, 137654, 137659, 156017, 156019,
185989, CAS-SU 19373–374, 24832, MCZ 37700,
RMNH 18009, USNM 310791–793. Palawan: CAS
139142, MNHN A951. Panay: CAS 137581–583. Poro:
CAS 124517–518. Semirara: CAS 127855–857. Siquijor:
CAS-SU 26450, 26597–607. Tabalas: CAS 137203–206,
MCZ 26083.
Hemiphyllodactylus margarethae
islands of indonesia and Malaysia
[sunda].
Sumatra: AMS R129492, BMNH
91.9.24.9, IRSN 2375A–B, RMNH 7341 ZMA no number (1), 11095, 11096.
Hemiphyllodactylus titiwangsaensis
islands of indonesia and Malaysia [sunda]:
Malaysia: AMS R135270, MCZ
R166921, ZFMK 32284–286, ZRC 2.4780–81, 2.4782,
2.4783–85, 2.4832, 2.1565, 2.5419, 2.5942–5943.
Hemiphyllodactylus yunnanensis
China
[China].
Myanmar:
BMNH
1933.7.8.11, USNM 310819, 570732–735. Yunnan:
BMNH 1904.1.26.1, 1904.11.29.1–9, 1904.11.29.10A–
N, CMS 8153, FMNH 7716–17, MCZ R18967, MNHN
1912.293, 1912.295A–B, 1912.296, NMB 9541. Laos:
number 631
•
61
FMNH 14451–452. Thailand: BMNH 1931.11.21.1,
BPBM 3502, FMNH 178328, 180867, 215988–994, QM
4820, THNHM 0153–54, 5943–949, USNM10621–622,
310798–808. Vietnam: MNHN 1948.43–44, RMNH
28007, USNM 310797.
Hemiphyllodactylus titiwangsaensis Zug: ZFMK 32284–286, ZRC 2.4780–
Hemiphyllodactylus “yunnanensis”
ADDITIONAL LOCALITY RECORDS
southeast asia [sEasia].
Cambodia:
FMNH 270569. Thailand: THNHM 075, 4714–715,
4910–17, 8620, ZRC 2.3567. Thailand, country only:
BPBM 3502 hermaphrodite with large testes and pair of
vitellogenic follicles (diameter 3.2 mm). Vietnam: USNM
146161.
The following localities derive from distributional records appearing in publications and from museum specimen records for specimens that I did not examine directly.
The speciic identity provided by the museum or in the
publication is the one usually followed; however, where
information was adequate and contrary to author’s species
determination, I have re-identiied the specimen.
Hemiphyllodactylus ganoklonis Zug: SAM R47715, USNM 495065–066,
563663–683.
4785, 2.4832, 2.1565, 2.5419, 2.5942–5943.
Hemiphyllodactylus [species indeterminate]
Hemiphyllodactylus typus
islands of indonesia and Malaysia
[sunda].
Borneo: Brunei: ZRC 2.5672, 2.5675–78;
Kalimatan: KUZ R8723, R8746, USNM 313965; Sabah:
BMNH 95.9.11.5A&B, 1929.12.22.87, FMNH 63661–
662, 239661, 243789, MCZ R43478; Sarawak: FMNH
138545, 158734, 196268A, 213665, ZRC 2.5671,
2.5955.
China [China].
Hong Kong: MCZ R182874–
876, MNHN 1912.293.
india and sri Lanka [india].
Sri Lanka:
BMNH 91.03.16.4, NMB 8552.
TYPE SPECIMENS
Catalog numbers for the primary type specimens here
are identiied in bold in the preceding locality samples.
Hemiphyllodactylus typus Bleeker: BMNH 1946.8.30.83.
Hemidactylus aurantiacus Beddome: BMNH 74.4.29.1332–1337, ZMB
10233.
Spathodactylus mutilatus Günther: BMNH 1946.8.30.83.
Lepidodactylus ceylonensis Boulenger: BMNH 74.4.29.1326.
Hemiphyllodactylus leucostictus Stejneger: USNM 21220, 23459–460,
23485, 23499–500, 23509.
Lepidodactylus Harterti Werner: ZMB 15360.
Gehyra larutensis Boulenger: BMNH 1901.3.20.2.
Gehyra yunnanensis Boulenger: BMNH 1904.1.26.1–26.2.
Mascarene islands.
La Réunion: Déso et
al. (2007); Rodriques: Schröder and Röll (2004).
nicobar islands.
Great Nicobar Island:
Biswas and Sanyal (1980).
sumatra.
Pulau Enggano: MVZ 39345–
39346, 239586.
thailand.
Kanchanaburi, Nakhon Ratchasima, Narathiwat, Phang-Nga, Phuket, Ranong, Trang,
and Trat provinces: Pauwels and Sumontha (2007).
Vietnam.
Southernmost mapped locality: Bobrov and Semenov (2008).
China.
Hainan Island: MVZ 42817–42818.
Ryukyu.
Iriomotejima: Ota (1990).
taiwan.
Main island: Ota (1989).
Papua new guinea.
Milne Bay Province,
Pini Range: Kraus and Allison (2004).
solomon islands.
Guadalcanal: McCoy
(2006).
hawaiian islands.
All major islands: McKeown (1996).
Marshall islands.
Enewetak: R. I. Crombie (unpublished manuscript, “Paciic amphibian and reptile distributions,” 1994).
Cook islands.
Rarotonga: Gill (1998).
French Polynesia.
Marquesas: Elao, Hivo
Oa, Mohotani: Ineich and Blanc (1989); Pitcairn: Ineich
(1992); R. I. Crombie (unpublished manuscript, 1994).
Hemiphyllodactylus insularis Taylor: CAS 62065, CM 2052–53.
Hemiphyllodactylus margarethae Brongersma: ZMA 11095–096, IRSN
Hemiphyllodactylus aurantiacus
2375A–B.
Hemiphyllodactylus typus chapaensis Bourret: MNHN 1948.43–44.
Hemiphyllodactylus typus pallidus Auffenberg: UF 28878, 28985.
india.
Anaimalai Hills, Bangalore, Kolli Hills,
Nilgiri Hills, Shevaroy Hills: Bauer and Das (1999);
62
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
Bangalore: Daniels (1994); Andhra Pradesh, Visakhapatnam District, Araku Valley and Madhygulu: Sanyal et al.
(1993).
Hemiphyllodactylus yunnanensis
China.
Yunnan–Changyuan,
Chengjiang,
Chuxiong, Gejiu, Jinping, Lijang, Longling,Yao’an; Guizhou–Anlong, Dushan, Huishui, Xingyi; Guangxi–Dayaoshan: Zhou et al. (1981).
Vietnam.
All mapped localities N of 15°N:
Bobrov and Semenov (2008, op. cit.); Tam Dao: MVZ
226500.
Hemiphyllodactylus “yunnanensis”
Cambodia.
Phnom Tumpor: Grismer et al.,
(2008); Koh Rongnieur and Koh Khlee-Ay islands, Megong R.: Bezuijen et al. (2009).
thailand.
Khao Yai National Park: Chan-ard
et al. (1999).
Hemiphyllodactylus [species indeterminate]
(THESE LOCALITIES NOT MAPPED)
sri Lanka.
Southwestern Sri Lanka: Somaweera and Somaweera (2009).
Appendix 3: Statistical Analyses
I
used standard univariate statistics to summarize variation of the characters within each sample. These data are the main ones presented for the
comparison and description of samples, the means for mensural data, and
the medians for meristic data. I tested all samples with adequate numbers
of adult females and males for sexual dimorphism with the Student t test, signiicance at P ≤ 0.05. A repeats protocol provided an estimate of the variation
derived from my data gathering (see detailed explanation in Baseline Estimate of
Intra-Observer Variation subsection).
Multivariate statistics (discriminant function analysis [DFA] and principal
components analysis [PCA]) were used mainly to explore the homogeneity of
samples. My goal was to discover which mensural traits best differentiated between unisexual and bisexual individuals within large regional samples. I did
not use scalation traits in these analyses, although I did use proportions in some
analyses but did not mix measurements. Some authors (e.g., Atchley et al.,
1976) have argued against the use of proportions in multivariate tests and demonstrated problems with proportional data through simulation studies. Other
authors have shown that proportions and nontransformed measurements do not
yield signiicantly different results in data sets from museum specimens such as
frogs (Heyer, 1978) and turtles (Iverson, 1981). I justify my use of proportions
herein because I was neither testing differences between or among groups or
relying on statistical signiicance in the assignment of specimens to taxonomic
group.
The following paragraphs present a synopsis of the statistical results from
the comparison using PCA and DFA. They are arranged in the same sequence
as in the text.
Uniformity among the unisexual samples:
A DFA was employed to test
uniformity (homogeneity) of the adults (n = 119) of the Paciic samples including
the holotype of H. typus using the eight mensural traits (EyeD, HeadL, HeadW,
NarEye, SnEye, SnW, SVL, TrunkL; TailL was not examined in this test nor any
of the subsequent ones). The adjusted classiication matrix (jackknifed) yielded
an average accuracy of 23% assignment. The irst three predictor variables (SVL,
TrunkL, HeadL) had eigenvalues of 0.373, 0.317, and 0.180 and accounted for
64
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
36.5%, 31.0%, and 17.6% of the total variance, respectively. A PCA (covariance matrix) of the same combined
samples and traits produced a compact clustering on the
irst two components. The irst three components had eigenvalues of 24.72, 1.064, and 0.156, respectively, with the
irst component accounting for 94.6% and the second component 4.1% of total variance. SVL, TrunkL, and HeadL
had the strongest loading (eigenvalues of 0.826, 0.534,
and 0.131).
Principal components analysis of body proportions in
SUNDA sample:
Principal components analysis (correlation matrix) results for adult females (n = 72) are summarized in the text with the exception of eigenvalues for
the irst four components: 2.712, 2.400, 1.639, 1.206;
these components accounted for 79.7% of total variance.
A PCA of adult males (n = 9) identiied HeadL/SVL and
SnW/HeadW as the highest loading variables on the irst
component, NarEye/HeadL on the second, TrunkL/SVL
on the third, and SnW/HeadL on the fourth. In total, these
four components accounted for 91% of total variance,
with 48.1% for the irst component. Eigenvalues for the
irst four components were 4.826, 2.567, 0.948, and 0.826.
Dispersion on a plot of irst and second components was
broad, although a regional clustering occurred with Sumatran males in the upper left quadrant, Bornean males in the
lower third of the upper right quadrant, and Malaysian
males in the lower left quadrant.
Discriminant function analysis examination of INDIA
sample:
A DFA of india males (n = 6 India, 1 Sri Lanka)
yielded 71% accuracy (jackknifed classiication) using eight
mensural traits and a considerably lower accuracy (14%)
with the 10 proportional traits. The accuracy for the unadjusted classiication was 100% for both character sets. The
eigenvalues for the two analyses were 13.073 and 7.029,
respectively. Using the 10 proportional traits, classiication
accuracy was better (80% jackknifed) in the adult females
(n = 4 India, 1 Sri Lanka); mensural traits were not examined in females.
Discriminant function analysis examination of Malaysian Hemiphyllodactylus samples:
The distinctiveness
of the H. harterti sample (n = 2 adults) and a central mountain range sample (15 adults) was explored with two data
sets, the eight mensural traits, and a subset of eight meristic
traits (CircNa, SnS, Suplab, Inlab, Chin, CloacS, Subcaud,
TotPore). The latter subset had been identiied as the best
set of discriminators by an earlier DFA of all 22 scalation
characters. Of the mensural set, HeadL, SnEye, and OrbD
were assigned the largest classiication functions. The unadjusted classiication matrix yielded 100% accuracy of group
assignment and the jackknifed matrix only 76% total accuracy, with a single eigenvalue (3.045) reported. The meristic
set yielded 100% classiication accuracy in the unadjusted
matrix and 94% in the jackknifed matrix, with a single eigenvalue (13.241) reported.
References
Atchley, W. R., C. T. Gaskins, and D. Anderson. 1976. Statistical Properties of Ratios.
I. Empirical Results. Systematic Zoology, 25:137–148.
auffenberg, W. 1980. The Herpetofauna of Komodo, with Notes on Adjacent Areas.
Bulletin of the Florida State Museum, Biological Sciences, 25:39–156.
Barbour, T. 1924. A Yunnan Gecko. Occasional Papers of the Boston Society of Natural
History, 5:133–135.
Bauer, A. M. 1994. Familia Gekkonidae (Reptilia, Sauria). Part I Australia and Oceania.
Das Tierreich, 109:i–xiii,1–306.
Bauer, A. M., and I. Das. 1999. The Systematic Status of the Endemic South Indian
Gecko Hemiphyllodactylus aurantiacus (Beddome, 1870). Journal of South Asian
Natural History, 4:213–218.
Bavay, A. 1869. Catalogue des Reptiles de la Nouvelle-Calédonie et Description d’Espèces
Nouvelles. Mémoires de la Société Linnénne de Normandie, 15:1–37.
Beddome, R. H. 1870. Description of Some New Lizards from the Madras Presidency,
with 2 Plates. Madras Monthly Journal of Medical Science, 1:30–35.
Bergmann, P. J., and A. P. Russell. 2003. Lamella and Scansor Numbers in Thecadactylus
rapidcauda (Gekkonidae): Pattern Revealed through Correlational Analysis and Implications for Systematic and Functional Studies. Amphibia-Reptilia, 24:379–385.
Bezuijen, M. R., B. Vinn, and L. Seng. 2009. A Collection of Amphibians and Reptiles
from the Mekong River, North-Eastern Cambodia. Hamadryad, 34:135–164.
Biswas, S., and D. P. Sanyal, 1980. A Report on the Reptilia Fauna of Andaman and
Nicobar Islands in the Collection of Zoological Survey of India. Record of Zoological Survey, India, 77:255–292.
Bleeker. 1860. Reptiliën van Agam. Natuurkundig Tijdschrift voor Nederlandsch Indie,
ser. 4, 20:325–329.
Bobrov, V. V., and D. V. Semenov. 2008. Yashcheritsy Vbetnama [Lizard Fauna of Vietnam]. Moscow: Scientiic Publications, KKM (in Russian).
Boulenger, G. A. 1883. On the Geckos of New Caledonia. Proceedings of the Zoological
Society, London, 1883:116–131.
———. 1885. Catalogue of the Lizards in the British Museum (Natural History) Second
Edition. Volume 1. Geckonidae, Eublepharidae, Uroplatidae, Pygopodidae, Agamidae. London: Taylor and Francis.
———. 1900. Description of New Batrachians and Reptiles from the Larut Hills, Perak.
Annals and Magazine of Natural History, ser. 7, 6:186–193.
———. 1903. Description of New Lizards in the Collection of the British Museum. Annals and Magazine of Natural History, ser. 7, 12:429–435.
———. 1912. A Vertebrate Fauna of the Malay Peninsula from the Isthmus of Kra to
Singapore Including the Adjacent Islands. London: Taylor and Francis.
66
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
Bourret, R. 1937. Lézards et Serpents Reçus au Laboratoire
des Sciences Naturelles de l’Univeristé au Cours de l’Année
1937. Descriptions de Deux Espèces et de Deux Variétés
Nouvelles. Bulletin Général de l’Instruction Publique, Gouvernment Général de l’Indochiné, 4:57–80.
Brongersma, L. D. 1931. “Reptilia.” In Résultats Scientiiques
du Voyage aux Indes Orientales Néerlandaises de LL. AA.
RR. le Prince et la Princesse Léopold de Belgique, ed. V. van
Straelen, 4 pls. Mémoires du Musée Royal d’Histoire Naturelle de Belgique, 5(2):1–30.
———. 1932. Some Notes on the Genus Hemiphyllodactylus
Bleeker. Zoologische Mededeelingen, 14:211–223.
Brown, W. C., and A. C. Alcala. 1978. Philippine Lizards of the
Family Gekkonidae. Dumaguete City, Philippines: Silliman
University Press.
Brygoo, É. R. 1990. Les Types de Gekkonidés (Reptiles, Sauriens) du Muséum national d’Histoire naturelle Catalogue
Critique. Bulletin du Muséum national d’Histoire naturelle,
Paris, ser. 4, sect. A, 12(3–4, suppl.):19–141.
Chan, S. K. F., K. Chaung, C. Ho, F. Lam, and W. Jong. 2006. The
Geckos of Hong Kong. Hong Kong Biodiversity, 13:1–8.
Chan-ard, T., W. Grossmann, A. Gumprecht, and K.-D. Schultz.
1999. Amphibians and Reptiles of Peninsular Malaysia and
Thailand. An Illustrated Checklist. Wuerselen, Germany:
Bushmaster Publications.
Cox, M. J., P. P. van Dijk, J. Nabhitabhata, and K. Thirakhupt.
1998. A Photographic Guide to Snakes and Other Reptiles
of Peninsular Malaysia, Singapore and Thailand. Sanibel Island, Fla.: Ralph Curtis Books.
Crombie, R. I., and G. K. Pregill. 1999. A Checklist of the Herpetofauna of the Palau Islands (Republic of Belau), Oceania.
Herpetological Monographs, 13:29–80.
Daniels, R. J. R. 1994. Notes on a Rare South Indian Gecko, Hemiphyllodactylus typus Beddome [sic]. Dactylus, 2:132–133.
David, P., and G. Vogel. 1996. The Snakes of Sumatra. An Annotated Checklist and Key with Natural History Notes. Frankfurt am Main: Edition Chimaira.
de Rooij, N. 1915. The Reptiles of the Indo-Australian Archipelago. I. Lacertilia, Chelonia, Emydosauria. Leiden, Netherlands: E. J. Brill.
Déso, G., J. M. Probst, and I. Ineich. 2007. Hemiphyllodactylus typus Bleeker, 1860 (Sauria: Gekkonidae) sur l’île de La
Réunion: Écologie et repartition. Bulletin de la Societé Francoise, 124:31–48.
Gaulke, M. 2003. Übersicht der philippinischen Gecko-Fauna.
Gekkota, 4:3–24.
Gill, B. J. 1998. Records of Cook Islands Lizards. Tuhinga,
10:151–157.
Grismer, L. L., T. Neang, T. Chav, P. L. Wood, Jr., J. R. Oaks,
J. Holden, J. L. Grismer, T. R. Szutz, R. Thomas, and T. M.
Youmans. 2008. Additional Amphibians and Reptiles from
the Phnom Samkos Wildlife Sanctuary in Northwestern
Cardamom Mountains, Cambodia, with Comments on
Their Taxonomy and the Discovery of Three New Species.
Rafles Bulletin of Zoology, 56(1):161–175.
Guibe, J. 1954. Catalogue des Types de Lézards. Paris: Imprimerie Nationale.
Günther, A. 1872. On the Reptiles and Amphibians of Borneo. Proceedings of the Zoological Society, London, 1872:586–600.
Han, D., K. Zhou, and A. M. Bauer. 2004. Phylogenetic Relationships among Gekkotan Lizards inferred from C-mos
Nuclear DNA Sequences and a New Classiication of
the Gekkota. Biological Journal of the Linnean Society,
83:353–368.
Hartert, E. 1901–1902. Aus den Wanderjahren eines Naturforschers. Reisen und Forschungen in Afrika, Asien und
Amerika, nebst daran Anknüpfden, meist ornithologischen
Studien. Berlin: R. Friedländer & Sohn.
Hayek, L.-A. C., W. R. Heyer, and C. Gascon. 2001. Frog Morphometrics: A Cautionary Tale. Alytes, 18:153–177.
Heyer, W. R. 1978. Systematics of the Fuscus Group of the Frogs
Genus Leptodactylus (Amphibia, Leptodactylidae). Natural
History Museum Los Angeles County, Scientiic Bulletin,
29:1–85.
Ineich, I. 1992. Hemiphyllodactylus typus (Indo-Paciic tree
gecko). Herpetological Review, 23(4):123.
———. 1999. “Spatio-temporal Analysis of the UnisexualBisexual Lepidodactylus lugubris Complex (Reptilia,
Gekkonidae).” In Tropical Island Herpetofauna: Origin,
Current Diversity, and Conservation, ed. H. Ota, pp. 199–
228. Amsterdam: Elsevier Science B.V.
Ineich, I., and C. P. Blanc. 1989. Distribution des Reptiles Terrestres et Polynesie Orientale. Atoll Research Bulletin,
[1988](318):1–75.
Iverson, J. B. 1981. Biosystematics of the Kinosternon hirtipes
Species Group (Testudines: Kinosternidae). Tulane Studies
in Zoology and Botany, 23:1–74.
Kluge, a. g. 1968. Phylogenetic Relationships of the Gekkonid
Lizard Genera Lepidodactylus Fitzinger, Hemiphyllodactylus Bleeker, and Pseudogekko Taylor. Philippine Journal of
Science, 95(3):331–352.
———. 2001. Gekkotan Lizard Taxonomy. Hamadryad, 26:1–
209.
Kraus, F., and A. Allison. 2004. New Records for Reptiles and
Amphibians from Milne Bay Province, Papua New Guinea.
Herpetological Review, 35(4):413–418.
Lazell, J. 2002. The Herpetofauna of Shek Kwu Chau, South China
Sea, with Description of Two New Colubrid Snakes. Memoirs
of the Hong Kong Natural History Society, 25:1–82.
Lazell, J. D., Jr. 1989. Kaho’olawe: Hawaii’s Largest Uninhabited Island. Explorers Journal, 64:123–126.
Malkmus, R., U. Manthey, G. Vogel, P. Hoffmann, and J. Kosuch.
2002. Amphibians & Reptiles of Mount Kinabalu (North
Borneo). Ruggell, Germany: A.R.G. Gantner Verlag K.G.
Manthey, U., and W. Grossmann. 1997. Amphibien & Reptilien
Südostasiens. Münster, Germany: Natur u. Tier Verlag.
McCoy, M. 2006. Reptiles of the Solomon Islands. Moscow:
Pensoft.
McKeown, S. 1996. A Field Guide to Reptiles and Amphibians
in the Hawaiian Islands. Los Osos Calif.: Diamond Head
Publications.
McMahan, C. D., and G. R. Zug. 2007. Burmese Hemidactylus
(Reptilia, Squamata, Gekkonidae): Geographic Variation in
the Morphology of Hemidactylus bowringii in Myanmar
and Yunnan, China. Proceedings of the California Academy
of Sciences, 58(24):485–509.
Ota, H. 1889. “A Review of the Geckos (Lacertilia: Reptilia) of the
Ryukyu Archipelago and Taiwan.” In Current Herpetology
number 631
in East Asia, ed. M. Matsui, T. Hikida, and R. C. Goris, pp.
222–261. Kyoto: Herpetological Society of Japan.
———. 1990. The Tree Gecko, Hemiphyllodactylus typus typus
(Lacertilia: Gekkonidae): An addition to the Herpetofauna
of Japan. Japanese Journal of Herpetology, 13(3):87–90.
Pauwels, O. S. G., and M. Sumontha. 2007. Hemiphyllodactylus
typus typus (Common Indopaciic Tree Gecko). Herpetological Review, 38(2):218–219.
Sanyal, D. P., B. Dattagupta, and N. C. Gayen. 1993. “Reptilia.”
State Fauna Series 5. Fauna of Andhra Pradesh, Part I, pp.
1–63. Calcutta: Zoological Survey of India.
Sauvage, H. E. 1879. Note sur les Geckotiens de la NouvelleCaldonie. Bulletin de la Société philomathique de Paris,
3(1878):63–73.
Schröder, E., and B. Röll. 2004. Hemiphyllodactylus typus
Bleeker. Sauria, suppl. 26(3):617–622.
Smith, M. A. 1933. Remarks on Some Old World Geckos. Records of the Indian Museum, 35:9–19.
Smith, M. A. 1935. The Fauna of British India, including Ceylon
and Burma. Reptilia and Amphibia. Vol. II—Sauria. London: Taylor and Francis.
Somaweera, R., and N. Somaweera. 2009. Lizards of Sri Lanka:
A Colour Guide with Field Keys. Frankfurt am Main, Germany: Edition Chimaira.
Stejneger, L. 1899. The Land Reptiles of the Hawaiian Islands. Proceedings of the United States National Museum,
21:783–813.
Stuart, B. L., R. F. Inger, and H. K. Voris. 2006. High Level
of Cryptic Species Diversity Revealed by Sympatric Lineages of Southeast Asian Forest Frogs. Biology Letters,
2006.0505:1–5.
Taylor, E. H. 1918. Reptiles of Sulu Archipelago. Philippine
Journal of Science, 13:233–267, 3 pl.
———. 1963. The Lizards of Thailand. University of Kansas Science Bulletin, 44:687–1077.
Wermuth, H. 1965. Liste der rezenten Amphibien und Reptilien. Gekkonidae, Pygopodidae, Xantusiidae. Das Tierreich,
80:i–xxii, 1–246.
•
67
———. 1966. Der Status von Platydactylus minutus Giebel
1862 (Reptilia, Sauria, Gekkonidae). Stuttgarter Beiträge
zur Naturkunde, 163:1–6.
Werner, F. 1900. Beschreibung Einiger noch unbekannter neotropischer und indischer Reptilien. Zoologischer Anzieger,
23:196–198.
Zhao, E., and K. Adler. 1993. Herpetology of China. Oxford,
Ohio: Society for the Study of Amphibians and Reptiles.
Zhou, K., Y. Liu, and G. Yang. 1981. Three New Subspecies of
Hemiphyllodactylus yunnanensis (Boulenger) from China
(Lacertiformes: Gekkonidae). Acta Zootaxonomica Sinica,
6:202–209. [Original in Chinese; see Smithsonian Herpetological Information Services no. 110 (1996) for English
translation by H. Ota.]
Zug, g. R. 1991. The Lizards of Fiji: Natural History and Systematics. Bishop Museum Bulletin in Zoology, 2:i–xii,
1–136.
———. 1998. Australian Populations of the Nactus pelagicus Complex (Reptilia: Gekkonidae). Memoirs of the
Queensland Museum, 42:613–626.
———. In press. “Tropical Asian Dry-Forest Amphibians and
Reptiles: A Regional Comparison of Ecological Communities.” In The Ecology and Conservation of Seasonally Dry
Forests in Asia, ed. W. J. McShea et al. Washington, D.C.:
Smithsonian Institution Scholarly Press.
Zug, G. R., J. V. Vindum, and M. S. Koo. 2007. Burmese Hemidactylus (Reptilia, Squamata, Gekkonidae): Taxonomic
Notes on Tropical Asian Hemidactylus. Proceedings of the
California Academy of Sciences, 58(19):387–405.
Zug, G. R., L. J. Vitt, and J. P. Caldwell. 2001. Herpetology.
Introductory Biology of Amphibians and Reptiles. 2nd ed.
San Diego, Calif.: Academic Press.
Zug, G. R., D. Watling, T. Alefaio, S. Alefaio, and C. Ludescher.
2003. A New Gecko (Reptilia: Squamata: Genus Lepidodactylus) from Tuvalu, South-Central Paciic. Proceedings
of the Biological Society of Washington, 116:38–46.
Index of Taxa Names
Page numbers in bold italic font indicate the start of the taxonomic account.
Cainodactylus, 4
Gehyra, 4, 23, 30
larutensis, 4–6, 16, 25, 28, 33, 43, 50, 61
oceanica, 9
yunnanensi, 4, 6, 51, 61
Hemidactylus,
aurantiacus, 6
bowringii, 33, 35
frenatus, 9, 30
Hemiphyllodactylus, ii, v, vii, xi, 1–6, 8–12, 14, 16–19, 21, 23–32, 34–35, 47
albostictus, 5, 35
aurantiacus, ii, vii, xi, 2–5, 16, 20, 26–27, 31–32, 34, 37, 39–40, 53, 60–61
ceylonensis, 4
chapaensis, 18–20, 22, 32–33
crepuscularis, 4
ganoklonis, ii, vii, 22, 27, 40, 41–43, 53, 60–61
harterti, ii, vii, xi, 3–5, 9, 11, 16–17, 27–29, 33–34, 43, 44–45, 53, 60, 64
insularis, ii, vii, viii, xi, 3–6, 24, 31, 34, 44, 46–47, 53 60–61
larutensis, ii, 3, 5, 9, 16–18, 28–29, 33, 44
leucostictus, 4, 6, 32, 35, 61
margarethae, ii, vii, viii, 3–4, 6, 11–12, 15–16, 18, 22, 25, 28–29, 32–34, 45, 47, 48,
53, 60–61
titiwangsaensis, ii, vii, viii, 27, 45, 48, 50, 53, 60–61
typus, ii, vii, xi, 1–6, 8–16, 22, 23, 25, 27–34, 35, 36–37, 52, 59, 61, 63
aurantiacus, 4–5
chapaensis, 5–6, 32–33, 50–51, 61
pallidus, 3, 5–6, 34–35, 61
typus, 4–5
yunnanensis, ii, vii, viii, xi, 3–5, 8, 11, 18–20, 22–23, 27, 29, 31–34, 40, 50, 51, 53,
60, 62
dushanensis, 3, 5–6, 19–20, 51
jinpingensis, 3, 5–6, 19–20, 33, 51
longlingensis, 3, 5–6, 19–20, 22–23, 32, 51
yunnanensis, 20, 22–24, 50
70
•
SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY
Lepidodactylus, 5, 34–35
aurantiacus, 2
ceylonensis, 2, 4, 6, 12, 35, 61
crepuscularis, 2
Harterti, 4, 6, 16, 25, 28–29, 33, 43,
61
lugubris, 2, 4, 9, 28–30, 33–35, 50
Platydactylus,
crepuscularis, 2, 6, 35
minutus, 5
Pytodactylus,
gracilis, 1, 2
Spathodactylus, 2
mutilatus, 2, 6, 35, 61
Spathoscalabotes, 2
REQUIREMENTS FOR SMITHSONIAN SERIES PUBLICATION
ALL MANUSCRIPTS ARE REVIEWED FOR ADHERENCE TO THE SISP MANUSCRIPT PREPARATION AND
STYLE GUIDE FOR AUTHORS (available on the “Submissions” page at www.scholarlypress.si.edu). Manuscripts not
in compliance will be returned to the author. Manuscripts intended for publication in the Contributions Series are evaluated by a content review board and undergo substantive peer
review. Accepted manuscripts are submitted for funding approval and scheduling to the Publications Oversight Board.
MINIMUM MANUSCRIPT LENGTH is thirty manuscript
pages. If a manuscript is longer than average, an appropriate
length will be determined during peer review and evaluation
by the Content Review Board. Authors may be asked to edit
manuscripts that are determined to be too long.
TEXT must be prepared in a recent version of Microsoft
Word; use a Times font in 12 point for regular text; be double
spaced; and have 1" margins. Each chapter/section must be
saved in a separate ile.
REQUIRED ELEMENTS are title page, abstract page, table
of contents, main text, and reference section. See the SISP
Manuscript Preparation and Style Guide for Authors for the
order of all elements.
HEADINGS should be styled so different levels of headings are
distinct from each other and so the organization of the manuscript is clear. Insert one line space above and one line space
below all headings.
FRONT MATTER should include title page, abstract page,
and table of contents. All other sections are optional. Abstracts must not exceed 300 words. Table of contents should
include A-, B-, and C-level headings.
TABLES (numbered, with captions, stubs, rules) should be
submitted in separate MS Word iles; should include footnotes, if appropriate; should have rules only at top, bottom,
and beneath column heads. Print outs of each table should
accompany the manuscript to ensure correct layout of data.
Tabulations within running text should not be numbered or
formatted like formal tables, and should be included in the
text of the manuscript.
FIGURE CAPTIONS should be provided in a separate MS
Word ile.
FIGURES (e.g., photographs, line art, maps) should be numbered sequentially (1, 2, 3, etc.) in the order called out; be
placed throughout text, not at end of manuscript; have all
components of composites lettered with lowercase letters and
described in the caption; include a scale bar or scale description, if appropriate; include any legends in or on the igure
rather than in a caption.
ART must not be embedded in the main text.
Figures must be original and submitted as individual TIFF
or EPS iles. Resolution for art iles must be at least 300 dpi
for grayscale and color images and at least 1200 dpi for line
art. Electronic images should measure no more than 100%
and no less than 75% of inal size when published. JPG iles
will not be accepted. Color images signiicantly increase
costs so should be included only if required. Funding for
color art is subject to approval by SISP and the Publications
Oversight Board.
TAXONOMIC KEYS in natural history papers should use
the aligned-couplet form for zoology. If cross referencing is
required between key and text, do not include page references
within the key but number the keyed-out taxa, using the same
numbers with their corresponding heads in the text.
SYNONOMY IN ZOOLOGY must use the short form (taxon, author, year:page), with full reference at the end of the
paper under “References.”
IN-TEXT REFERENCES should be used rather than bibliographic notes and should follow the author-date system in the
following format: “(author last name, year)” or “. . . author
(year)”; “(author, year:page used within the text)” or “. . .
author (year:page).” A full citation should be included in a
“References” section.
ENDNOTES are to be used in lieu of footnotes and should
be keyed manually into a separate MS Word ile, in a section titled “Notes”. Notes should not contain bibliographic
information. Manually type superscript numerals in text and
use full-sized numerals at the beginning of each note in the
“Notes” section. SISP will determine the best placement of the
notes section, either at the end of each chapter or at the end
of the main text.
REFERENCES should be in alphabetical order, and in chronological order for same-author entries. Each reference should
be cited at least once in main text. Complete bibliographic
information must be included in all citations (e.g., author/editor, title, subtitle, edition, volume, issue, pages, igures). For
books, place of publication and publisher are required. For
journals, use the parentheses system for volume(number):pagination [e.g., “10(2):5–9”]. Do not use “et al.”; all authors/
editors should be included in reference citations. In titles,
capitalize irst word, last word, irst word after colon, and all
other words except articles, conjunctions, and prepositions.
Examples of the most common types of citations are provided
in the SISP Manuscript Preparation and Author Style Guide.
For questions regarding the guidelines, please email SISP at
schol.press@si.edu.
15
Tropical Asian Dry Forest
Amphibians and Reptiles
A Regional Comparison of
Ecological Communities
George R. Zug
A
sian seasonally dry tropical forests extend the breadth of South Asia from the
edge of the Thar desert (India, approximately lat. 72° E) to the Vietnamese
coast (approximately 109° E) of Indochina (Wikramanayake et al. 2002).1 The
forests occur in a subtropical-tropical band from about 9° N to 24° N, typically at low
elevations (approximately 5–300 m). Numerous physiographic and climatic factors
within this large geographic area divide the dry deciduous or semi-deciduous forest
into six Indian and three Indochinese forest categories.
The diversity of seasonally dry tropical forests suggests a concomitant diversity of
amphibian and reptilian communities. I realized the potential difficulty of reviewing
the herpetofaunas of these forests, owing to the existence of few rigorous site-specific
herpetofaunal inventories of Asian dry forests. This challenge and other ones to a
meaningful comparison of different dry forest herpetofaunas are detailed below. In
spite of these challenges, three objectives can be realized: (1) Identify the species composition of the herpetofauna at several sites throughout the breadth of the Asian dry
forests; (2) Examine each herpetofauna for a semblance of community structure; and
(3) Compare species composition and community structure between sites to test for
similarities/differences arising from shared/different environmental features of sites.
MATERIALS, METHODS, AND LIMITATIONS
Definition of Terms
Because my usage of ecological terms likely differs from other naturalists, I offer a few definitions to enhance communication of my results and interpretations.
275
276
CHAPTER 15
A community is an assemblage of organisms living in the same place, interacting
through mutualism, predation, competition. I accept Drury’s (1998) community
concept as the organisms of a locality assembled by chance and organized similarly
by the stochastic interactions of physical and biotic environment, hence I use community and assemblage interchangeably. A herpetofauna is a community of amphibians and reptiles living together in the same area, habitat, or microhabitat. Generally,
most frequently herein, herpetofauna refers to the amphibian and reptilian populations occupying the same habitat—another ecological term or entity used variously
by me and most other naturalists.
My use of taxon is restricted to a species in both the biological and phylogenetic
species concept sense of a population or populations of phylogenetically related and
potentially interbreeding individuals. Guilds are groups of animals sharing similar prey
and prey-capture behaviors.
Dry Forests
The dry forests referred to herein are those labeled as “dry broadleaf forests” by
Wikramanayake et al. (2002, Figure 1.3, to which the ecoregion numbering scheme
below refers) and specifically as “dry forests” or “dry deciduous forests” (peninsular
Indian forests [16, 18–22]; Indochina forests [58, 71–72]). The primary herpetofaunas
discussed below derive from four dry forest types: Khathiarbar-Gir dry deciduous
forests (Gir Forest National Park), northwestern India (16); northern dry deciduous
forests (central Nallamala Hills), west-central India (19); Irrawaddy dry forests (Chatthin Wildlife Sanctuary, Shwe-Settaw Wildlife Sanctuary), central Myanmar (71); and
central Indochina dry forests (Sakaerat Biosphere Reserve of central Thailand and hilly
eastern Cambodia; 72).
There are no sets of distributional maps of South Asian amphibians and reptiles
that pinpoint the precise occurrence of individual species. I recognized this lack of
a consolidated data source as the first challenge in a search for patterns of species
distributions that would be concordant with the maps of the dry forest ecoregions
(Wikramanayake et al. 2002). Hence, the preceding assignment of a herpetofauna to a
particular forest (Figure 1) depends upon the general habitat description in the report
and requires that the site lies within the mapped boundaries of a dry forest ecosystem.
Herpetofaunal Surveys of Dry Forest
Lists of herpetofaunas for dry forest sites in tropical Asia are of variable quality.
Only two sites have received rigorous year-round inventorying using several sampling
techniques: Sakaerat Environmental Research Station (Nakhon Ratchasima Province,
Thailand, 14°30′ N 101°55′ E; Inger and Colwell 1977) and Chatthin Wildlife Sanctuary, Sagaing Division, Myanmar, 23°34.46′ N 95°44.26′ E; Zug et al. 1998, 2004;
G. Zug, unpublished data). These two sites had weekly or more-frequent inventories
for one year or longer. The inventory techniques included random quadrat searches,
cruise collecting along transects, general-random collecting, and additionally at Chat-
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
277
Figure 1. Location of sites for the dry forest herpetofaunas compared in this chapter.
thin, a week of drift-fence pitfall trapping each month. Thus, these herpetofaunas serve
as touchstones for the sites less-thoroughly inventoried.
The herpetofaunal surveys at the other sites derived largely from general-random
collecting. In India, the inventory of the Gir Forest National Park (Gujarat State,
21°08′ N 70°48′ E) derived from intermittent fieldwork over a one-year interval
(Bhatt et al. 1999; Vyas 2000); the fieldwork was supplemented by interviews with
local people and park personnel and a review of unpublished reports. The Nallamala
Hills (Andhra Pradesh State, 15°30′ N 79°15′ E) inventory encompasses a large area
(approximately 13,000 km2), and species occurrence data drew on over a century of
assorted fieldwork, with some recent (1995–2004) fieldwork of more-intensive sampling of a few areas. To make the herpetofaunal listing of Nallamala more comparable
in geographic scope to those of other sites, I use only the herpetofauna of the central
portion of the hills (see Appendix A) for comparison.
Additional Burmese herpetofaunas are available for comparison. These herpetofaunas derived from moderately thorough herpetological inventories of variable periods,
some for more than 30 days. Two such sites provide intra-Burma comparison: ShweSettaw Wildlife Sanctuary (Magway Division, 20°06 N 94°44 E); Min-Gon-Taung
278
CHAPTER 15
Wildlife Sanctuary (Mandalay Division, 21°24 N 95°47 E). The Shwe-Settaw site
has experienced 44 survey days in both dry and wet seasons, and its herpetofauna is
enumerated in Appendix A. Min-Gon-Taung had 38 survey days, but all fieldwork
occurred in the latter third of the wet season.
I have been unable to locate a site-specific inventory of a dry forest area in northeastern Cambodia or southern Vietnam. Site-specific inventories in these two countries
have focused on evergreen and moist forests. A sequential survey (Stuart et al. 2006)
at five sites in hilly eastern Cambodia comes closest to matching the preceding herpetofaunal surveys, and its results are listed (Appendix A) although the duration at all
sites totaled 37 days and much of the fieldwork appears to have been concentrated in
evergreen forests (those taxa are excluded from the site species list). The areas surveyed
included Seima Biodiversity Conservation Area, Phnom Nam Lyr Wildlife Sanctuary,
and Virachey National Park (three areas centered roughly at 13° N 107° E).
Other sites throughout the dry forest band have had surveys of varying intensity
and duration, often concentrating on either amphibians or reptiles, or a subset of one
these groups. Although their inventories are incomplete and of variable quality (e.g.,
inadequate confirmation of species identification), they are useful for comparative
purposes but not for numerical analysis.
Taxonomic Inequalities
Comparisons of faunas require the accurate identification of component species comprising each fauna. In Asia and especially in Burma, such accuracy has become increasingly
difficult, not because we are unable to recognize each component species within each
site’s herpetofauna and readily assign a scientific name to most specimens, but rather
because the names we assign to populations in northern or southern Burma and those of
central Thailand, for example, likely do not represent the same taxon. These taxonomic
differences are becoming increasingly evident as geographic coverage and voucher-sample sizes increase permitting detailed examination of morphological variation.
Two examples are sufficient to demonstrate the unreliability of depending upon
our current taxonomy of Asian amphibians and reptiles to determine the genetic sameness of populations in India, Burma, and Indochina. The classic example (Wüster and
Thorpe 1989, 1990, 1992; Wüster 1996) of regional differentiation in Asian cobras is
now nearly two decades old, and yet their model of analysis of population variation
across the breadth of Asia has been little followed, and then only recently so. Until
Wüster’s studies, Asian cobras were considered a single, wide-ranging species, Naja
naja, although several subspecies were recognized. By 1996, Wüster had recognized
ten species in his review article for toxicologists and medical doctors. Notably, the
geographic distributions of formerly recognized subspecies are not concordant with the
currently delimited ranges of Naja species.
In frogs, we still lack a breadth-of-Asia study. The necessity of such studies is
evident in widespread species, such as Polypedates leucomystax and its putative sister
taxa (Matsui et al. 1986; Orlov et al. 2001), but the best anuran example of high regional differentiation is the Fejervarya limnocharis complex. The details of speciation
in this group of frogs remain largely unresolved, but the taxonomic history outlined
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
279
in Table 1 gives a hint at the complexity of the situation. This table contains fifteen
taxa; more than twenty species had been recognized (published) as of June 2008. Our
herpetological surveys in Burma have identified a minimum of six species in the north
and central regions; two sympatric species were identified during my initial fieldwork
at Chatthin, based on two sizes of gravid females (30–39 mm and 47–67 mm svl
[snout–vent length]; Zug et al. 1998).
It is essential to remember in the subsequent text that the same species from distant
locations, e.g., Microhyla ornata in western India and eastern Vietnam, are not the
same genetic entity and possibly not even near relatives.
RESULTS
Species Inventories and Completeness
Species Accumulation Rates
The herpetofaunas of Gir, Nallamala, Chatthin, and Sakaerat (Appendix A) are assumed to be nearly fully inventoried. Confirmation of an inventory’s completeness
Table 1. Chronological summary of the recognition of speciation in the paddy or rice frog Fejervarya
limnocharis complex.a Inger (1954) was the first researcher to examine geographic variation of morphology
within these frogs. He conclusively demonstrated that the Philippine populations (vittigera) differed
substantially from those (limnocharis) elsewhere in Southeast Asia. He chose to recognize the two
populational groups as subspecies. His taxonomy was followed into the 1980s when the subspecies
concept was largely abandoned in herpetology.b
Fejervarya
Species Name
General Distribution
Source
l. limnocharis
l. vittigera
nepalensis
pierrei
syhadrensis
andamensis
nilgiris
teraiensis
vittigera
mysorensis
orissaensis
iskandari
limnocharis
sakishimensis
mudduraja
kudremukhensis
caperata
Tropical Asia, Pakistan to the Philippines
Philippines
Central and Eastern Nepal
Central and Eastern Nepal
Eastern and Western India, adjacent Nepal
Andaman Islands
Kerala and Tamil Nadu, India
Southern Nepal, adjacent India
Philippines
Karnataka, India
Orissa, India
Java
Java (implicit restriction of occurrence)
Southern Ryukyu Islands
Central Western Ghats, India
Central Western Ghats, India
Central Western Ghats, India
Inger 1954
Inger 1954
Dubois 1975
Dubois 1975
Dubois 1975
Dubois 1984
Dubois 1984
Dubois 1984
Dubois 1984
Dutta and Singh 1996
Dutta 1997
Veith et al. 2001
Veith et al. 2001
Matsui et al. 2007
Kuramoto et al. 2007
Kuramoto et al. 2007
Kuramoto et al. 2007
a
This listing of F. limnocharis complex species is intentionally incomplete (selective). The goal is to demonstrate the ongoing recognition of diversity (speciation) within paddy frogs. As of June 1, 2008, more than twenty species of paddy frogs had been recognized.
b
For a broader perspective on the worldwide recognition of increasing species diversity in amphibians, see Köhler et al. 2005.
280
CHAPTER 15
can be assessed by examination of species accumulation rates. Time-of-capture data
are available for the intensely surveyed Chatthin and Sakaerat herpetofaunas, and for
the less-intensively surveyed Shwe-Settaw and Min-Gon-Taung faunas. The rate of
species accumulation for Chatthin and Sakaerat (Figure 2a) are similar (and typical)
with an initial rapid vouchering of the sites’ herpetofaunas and then a gradual slowing
of the acquisition of new taxa. For these two sites, the near-total herpetofaunas were
discovered in 45 and 40 weeks, respectively. Although the discovery rates were similar
between these two sites, survey design and available man-power yielded different rates.
The Sakaerat survey obtained 50 percent of the herpetofauna in 5 weeks and 90 percent in 19 weeks; Chatthin sampling, in contrast, obtained 50 percent in 2 weeks and
90 percent after 43 weeks. Note that collecting effort, as measured in man-hours, differed between the two sites, and this effort is impossible to quantify precisely owing to
many staff and visitors capturing specimens and giving them to the survey. There was
also wide variation in the abilities of collectors to see and capture animals. Additionally, the surveys started in different seasons: the middle of the dry season at Sakaerat
(February), and the middle of the wet season at Chatthin (late July). The near-total
species diversity is not known for either Shwe-Settaw or Min-Gon-Taung; nevertheless,
a total similar to Chatthin is probable for each. The 8 and 7 week surveys, respectively,
were able to inventory 50 percent of the presumed herpetofauna at each site by the
fifth week. The accumulation curves of Shwe-Settaw and Min-Gon-Taung (Figure 2b)
do not rise as steeply as Chatthin’s, although they attain the same level by the end of
two months of inventory.
Although not directly evident from the curves but extractable from the dateof-first-capture data, there was a different sequence of captures for Chatthin and
Sakaerat. Both show the typical rapid rise in species discovery; however, at Chatthin,
frogs dominated early captures because of an early rainy season start. Within the first
week, 81 percent of the frog fauna and 46 percent of the lizards were vouchered. At
Sakaerat, the survey began in the late dry season and lizards dominated; 56 percent
of the lizard species were captured in the first week, contrasting to 32 percent of
the frogs. At Sakaerat, all frog species were documented by the end of the twentyfirst week and most lizards (90 percent) by the end of the fifteenth week. Similarly
at Chatthin, most frogs (94 percent) were vouchered by the end of the third week,
but to obtain most lizards (90 percent) required 43 weeks. Snake species were documented more slowly, most (94 percent) by the end of week 25 at Sakaerat, and 91
percent by the end of week 45 at Chatthin. In neither site was the herpetofauna fully
inventoried at the end of the first year. Two additional snake species were discovered
at Chatthin in the second year of the survey. A lizard (Varanus bengalensis) and
a snake (Cylindrophis ruffus) were vouchered prior to initiation of the inventory
(mid-July 1997) and were not documented during the subsequent three years of the
Chatthin project.
Even though the duration of the Shwe-Settaw inventory was only eight weeks (not
continuous and occurring over four years, including visits in both wet and dry seasons), the inventory attained the same species numbers (80 percent) as at Chatthin by
the eighth week (Figure 2B). The rate of capture of new frog, lizard, and snake species
was similar at Shwe-Settaw and Min-Gon-Taung.
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
281
Figure 2. Rate of discovery of
new species at several dry forest
sites. (A) Comparison of weekly
cumulative totals of new species
inventoried at a Burmese dry forest site (Chatthin Wildlife Sanctuary) and a Thailand site (Sakaerat
Environment Research Station).
(B) Comparison of cumulative
results for three Burmese dry
forest surveys (Chatthin Wildlife
Sanctuary, Shwe-Settaw Wildlife
Sanctuary, and Min-Gon-Taung
Wildlife Sanctuary).
Species Composition and Abundance
Species Composition— Diversity
The total and component compositions of the six major dry forest herpetofaunas are
summarized in Table 2. Of the two best-studied sites, Sakaerat has 22 (greater than 40
percent) more species than Chatthin. This -diversity is nearly matched by Nallamala.
The remaining three sites were less-intensely surveyed and, not unexpectedly, their species count is lower (Table 2). Shwe-Settaw at the southern end of the Burmese central dry
zone with only eight survey-weeks has 81 percent of the Chatthin herpetofaunal total,
and Min-Gon-Taung with six survey-weeks is 65 percent (35 species) of the Chatthin
total. Cambodia has the lowest diversity (Table 2), but also the fewest survey-weeks.
282
CHAPTER 15
Table 2. Numerical summary of the amphibian and reptilian components of some Asian dry forest
herpetofaunas (Appendix A).
Taxon
Gir
Caecilians
Frogs
Turtles
Lizards
Snakes
Total
0
7
2
14
17
40
Nallamala
0
19
4
22
30
75
Chatthin
0
16
3
13
22
54
Shwe-Settaw
0
11
5
15
13
44
Sakaerat
1
20
2
19
36
78
Cambodia
0
13
0
8
9
30
In the Indian dry forest sites (Table 2), Gir’s 40 species seems unrealistically low,
especially considering that Vansda National Park, also in Gujarat State although
mainly with moist deciduous forest habitats, has 54 species (Vyas 2004). The diversity
at Nallamala is nearly 90 percent greater than that at Gir and equivalent to the diversity at Sakaerat.
Although the actual numbers of species differ among the sites, the relative diversity
of each herpetofaunal component is similar (Figure 3; Table 2). At all sites, snakes
have the highest diversity, and for the three best-surveyed sites, snakes comprise 40–41
percent (Nallamala and Chatthin) to 48 percent (Sakaerat) of the herpetofaunal assemblage. As detailed in the next section, this diversity associates with low abundance.
Frogs and lizards share a similar diversity (approximately 25 percent each) to one
another at each site and between sites (Figure 3).
The similarity in component diversity at the three well-inventoried sites across a
broad latitudinal distance highlights the incompleteness of the inventory data from the
other sites. The low diversity of frogs (18 percent) in the Gir forest contrasts sharply
with the other sites and suggests insufficient attention to nocturnal surveys; however,
in the moist deciduous forests of “nearby” Vansda National Park (Vyas 2004), frog
diversity is only 22 percent. This comparison of these two Gujarat state forests is
problematic because their faunal lists derive from inventories by the same researcher
and presumably share the same bias or limitation for nocturnal surveys. A contrasting
situation is evident for the eastern hills of Cambodia where frog diversity represents 43
percent of the herpetofauna, indicating a bias toward an amphibian inventory.
As noted previously, dry forest herpetofaunal inventories are rare. A survey of
the Kalakad-Mundanthurai Tiger Reserve (Tamil Nadu State, India; Vijayakumar et
al. 2006) describes the lower hill forest as a mixture of scrub, dry deciduous thicket,
dry deciduous savanna, dry evergreen forest, and riparian forest. This “dry forest”
Kalakad assemblage consists of 17 frogs, 2 turtles, 22 lizards, and 10 snakes (total
herpetofauna: 51). An inventory of the entire reserve (Cherian et al. 2000) including
higher elevations and the evergreen sholas yielded 0 caecilians, 32 frogs, 2 turtles, 13
lizards, and 15 snakes (total herpetofauna: 65). The differences in number of species
and kind reflect more than the inclusion of the shola habitat but a difference in the
manner of surveying and in the habitats surveyed. In both, the snakes are poorly inventoried, and in combination, the two surveys still inventoried only 22. Although the
total diversity is less than for central Nallamala, the frog diversity is nearly double that
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
283
Figure 3. A portrait of relative species numbers (diversity) within each herpetofauna at three intensely
surveyed sites: Nallamala Hills, India; Chatthin, Burma; and Sakaerat, Thailand. The site percentages derive
from the total number of in-site species. Actual species occurrence is listed in Appendix A.
of Nallamala, and snake diversity is about two-thirds the latter (likely inadequately
surveyed; a minimum of 30 snake species is predicted).
Species Composition—Taxonomy
Among the major dry forest sites, a caecilian occurs only at Sakaerat. Frogs occur at all
sites (Appendix A) and represent a substantial component of the herpetofaunal community. The six sites share five anuran families. The endemic Ranixalidae is represented
by a single species (Indirana leithii) at Nallamala. Two anurans, the black-spiny toad
(Duttaphrynus melanostictus) and ornate narrow-mouthed frog (Microhyla ornata)
occur at all sites. Both of these species occur geographically and ecologically beyond
the dry forest zone. Neither is a human commensal, although both regularly populate
anthropogenic habitats; both occur in natural forest from sea level to low montane (up
to 300 m). As emphasized in Materials, Methods, and Limitations above, pan-Asian
species are suspect, that is, are they the same species throughout South Asia? For M.
ornata, the answer is no (Matsui et al. 2005) but no nomenclatural change has been
proposed. The population genetics for the spiny toad is unstudied. A similar pattern
is evident among other frogs where some resolution of regional and sympatric genetic
differentiation among pan-Asian species has been identified and taxonomically formalized. In these taxa (e.g., the Fejervarya limnocharis species group [SG], or complex;
284
CHAPTER 15
the Hoplobatrachus tigerinus SG; the Polypedates leucomystax SG), the SG has a
pan-Asian distribution and its members have similar appearances, and our largely
anecdotal knowledge of their ecology and behavior indicates “ecological equivalence”
throughout the SG’s distribution.
Another pattern among the non–pan-Asian species is dual geography of some
species, that is, a species occurs in the dry forest of two of three geographic pairs:
India-Burma, Burma–Southeast Asia, or India–Southeast Asia. The India-Burma taxa
are Sphaerotheca breviceps and Microhyla rubra; the Burma–Southeast Asia taxa are
Glyphoglossus molossus, Kaloula pulchra, Microhyla pulchra, Pelophylax lateralis,
and Chiromantis nongkhorensis.
Among turtles, there are no pan-Asian species. This absence results from better
taxonomic studies. A single dual geography exists for two geographic pairs: IndiaBurma, Melanochelys trijuga; and Burma–Southeast Asia, Indotestudo elongata. There
are no dry forest endemics.
Lizards display the same distributional patterns and share the same taxonomic
difficulties as anurans. The latter problem is highlighted by the Calotes versicolor SG,
which until recently (Zug et al. 2006) was considered a pan-Asian species, but the recognition of two sympatric Burmese species (C. htunwini, and C. irawadi, the former
potentially a dry forest endemic) revealed a multiplicity of distinct populations.
Only one lizard species, the house gecko (Hemidactylus frenatus), occurs in all
three areas. Because of its near-total anthropogenic association, it cannot be considered a true dry forest resident. There is no lizard species occurring in India-Burma. The
Burma–Southeast Asia species are Calotes mystaceus, Gekko gecko, Eutropis multifasciata, and Sphenomorphus maculatus. A number of taxa occur in one area only, such
as Calotes rouxi and Hemidactylus triedus in India or Bronchocoela smaragdina and
Lygosoma bowringii in Southeast Asia. None of these appears to be endemic to dry
forest, although such endemicity might occur for the two Leiolepis species.
Snakes exhibit the same patterns and taxonomic difficulties as noted for frogs and
lizards. A few snakes (Ahaetulla nasuta, Chrysopelea ornata, Lycodon aulicus, Ptyas
mucosus, Amphiesma stolatum, Ramphotyphlops braminus) occur in India, Burma,
and Southeast Asia dry forests. All these species also occur in a variety of habitats,
including anthropogenic ones. No snake species occurs at all sites, although Ahaetulla
nasuta and Ptyas mucosus likely occur in southeastern Cambodia as they occur in
southern Vietnam (Campden-Main 1970).
Geographic pairs exist: India-Burma, Xenochrophis piscator and Python molurus;
Burma–Southeast Asia, Boiga multomaculata, Boiga ocellata, Coelognathus radiatus,
Dendrelaphis subocularis, Sibynophis collaris, Rhabdophis subminiatus, Bungarus fasciatus, Crypteltropis albolabris, and Xenopeltis unicolor. With the exception of Naja
mandalayensis and potentially some of the Oligodon species, dry forest endemicity is
lacking, and even Oligodon species likely overlap into drier or moister habitats.
Species Composition—Size Structure
Size influences what a species eats, where it lives, and what eats it. Thus, a comparison of amphibian and reptilian body sizes among the six dry forest sites offers
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
285
another means to examine similarities and differences of the six selected communities. To make such a comparison, I categorized body size into five classes for frogs,
three for turtles, six for lizards, and five for snakes (see Appendix B for categorization procedure).
For frogs (Figure 4), only Chatthin possessed a miniature species (less than 20 mm
adult svl). Small species (22–44 mm svl) were generally the most frequent size class
among all sites, ranging from a proportional frequency of 29 percent (Gir) to 58 percent (Cambodia), although medium-sized frogs (46–70 mm svl) are the most frequent
class at the Gir forest (43 percent) and its small frogs have the same frequency as Chatthin (33 percent). Medium-sized species typically represent the second most frequent
size class. Large (72–96 mm svl) and big (≥ 98 mm svl) species often occur with the
same frequency (Figure 4).
The few turtle species occurring in dry forests do not permit a meaningful comparison among sites.
Lizards have about the same diversity at a site as frogs (Table 2), yet lizards partition into more size classes. The smallest (less than 42 mm svl) and largest (greater than
150 mm svl) size classes have the lowest frequency of occurrence at all sites (Figure
4): 6–8 percent, which equals one species. For the largest class, that taxon is Varanus
bengalensis, which is three to four times larger than the next-largest class (120–148
mm svl). For the smallest class, the lizard is either a gecko (Cnemaspis, Dixonius, or
Hemidactylus) or a lacertid (Ophisops). Chatthin and Cambodia have no small species. Nallamala has three small species, whereas a single species each occurs at all other
sites. The medium-small class (22–44 mm svl) dominates at Shwe-Settaw and Sakaerat,
the medium class (64–94 mm svl) dominates at Gir, and Chatthin has equal frequency
of medium-small and medium class taxa. Moderately large (98–118 mm svl) and large
(120–148 mm svl) taxa represent about a quarter of the lizard taxa at the Indian and
Burmese sites, and more than a third of the taxa at the Southeast Asian ones.
Snake sizes clustered into five discrete classes (Appendix B). The frequency of the
size classes differs strikingly among the six sites (Figure 4), although the medium-small
(300–590 mm svl) and medium (600–990 mm svl) classes comprise at least 60 percent
of the taxa present at all sites. The disparity in sampling effort is evident for ShweSettaw and Cambodia compared to the other sites. Both latter sites lack the large class
(greater than 1,500 mm svl) taxa, and the small class (less than 300 mm svl) is also
absent for Cambodia. These absences skew the frequency of the other classes and make
their comparison to other sites suspect. The other four sites have near-equal frequencies of the medium-large (1,000–1,500 mm svl) class taxa. At Nallamala, mediumsmall taxa are most abundant, which is striking in comparison to the other sites. Gir
and Sakaerat have roughly equal frequency of taxa in the medium and medium-small
classes. At both Burmese sites, medium-sized snakes dominate.
Behavioral Preferences
Behavioral preferences (categories) for this review include activity pattern (diurnal or
nocturnal), habitat choice or use (fossorial, arboreal, aquatic), and diet (herbivory,
omnivory, and several prey-classes of carnivory) at a gross level (Appendix B).
286
CHAPTER 15
Figure 4. A portrait of frequency
of size classes for frogs (A), lizards
(B), and snakes (C) at three
intensely surveyed sites: Nallamala
Hills, India; (Chatthin, Burma; and
Sakaerat, Thailand. The frequencies
(%) derive from the total number
of in-site species. Size classes are
defined in Appendix B.
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
287
In activity patterns, frogs are largely nocturnal, although a few semiaquatic and
aquatic taxa, e.g., Euphlyctis cyanophlyctis and Occidozyga lima, occur regularly at
water’s edge during the day; reproduction and feeding are at night. Turtles appear
predominantly diurnal, and with the exception of the nocturnal geckos, lizards are
diurnal. The relative homogeneity of the preceding activity patterns within a group
shows that this biological aspect is “evenly” distributed among the six sites.
Snakes are the most variable in time of activity. The pythons and viperids are
nocturnal, seemingly associated with the capture of sleeping birds or mammals and
the ambush-capture of nocturnally active rodents. This apparent foraging strategy is
shared also with the colubrids Boiga and Lycodon taxa. Most natricines are nocturnal,
although Amphiesma stolatum and Atretium schistosum are diurnal species. A similar
clade dichotomy occurs among the elapids with the kraits (Bungarus, Calliophis) being nocturnal and cobras (Naja) being diurnal/crepuscular. Most other colubrids are
diurnal, and although the Oligodon species are commonly labeled as nocturnal, our
Chatthin surveys suggest otherwise. The homalopsids and slug-eaters (Pareas) are nocturnal. No differences in community organization is evident among the sites.
With their nocturnal behavior, frogs have adapted to the full spectrum of subaerial
habitats. Dry forest toads (Bufonidae) are terrestrial as are the microhylids with fossorial (Glyphoglossus, Uperodon) to semifossorial (some Microhyla) species. The other
terrestrial-semifossorial frogs are Sphaerotheca. The arboreal frogs are Chiromantis
and Polypedates. The ranids and dicroglossids are semiaquatic to aquatic, and this
segregation is often equivocal, depending upon moisture level. The frog communities
have three organizational patterns, and at all sites, terrestrial species are the most numerous. Gir and Shwe-Settaw lack arboreal and fossorial taxa; terrestrial species are
two or more times as numerous as the aquatic and semiaquatic ones. Fossorial frogs
are not documented at Cambodia; semiaquatic and aquatic species are equal in number
and nearly so to the terrestrial taxa; arboreal frogs are the fewest. The pattern at Nallamala, Chatthin, and Sakaerat is similar. Terrestrial species are two to three times more
numerous than any of the other behavior classes; arboreal, semiaquatic, and aquatic
species are roughly equal in number, and a single fossorial species occurs at each site.
For turtles, the behavioral differences between semiaquatic (Cyclemys, Melanochelys)
and aquatic (Pangshura, all trionychids) species are distinct. Aquatic taxa emerge from the
water only to bask and lay eggs. Tortoises (Geochelone, Indotestudo) are terrestrial.
Again, turtles are too few to show differences in testudine habitat use among sites.
Among lizard taxa, only terrestrial and arboreal behaviors occur. Some of the small
skinks (e.g., Lygosoma) and lacertids (Ophisops) forage beneath the surface litter, but
this behavior is not truly fossorial. All skinks are terrestrial, and most geckos are arboreal. The agamids are also predominantly arboreal with two terrestrial taxa (Leiolepis,
Psammophilus). Monitors (Varanus) forage mainly on the ground but are adept climbers. With only two habitat behaviors, community organization is simple. At both Indian
sites, the number of arboreal and terrestrial taxa is the same. Terrestrial lizards outnumber arboreal ones by about a third in Burma, and the converse in Southeast Asia, although the numbers of species exhibiting the two behaviors are nearly equal at Sakaerat.
Snakes occupy the full range of habitat use, although most dry forest species
are either arboreal or terrestrial. The preference is often clade-related, e.g., arboreal:
288
CHAPTER 15
Ahaetulla, Boiga, Dendrelaphisor; and terrestrial: Coelognathus, Oligodon, Ptyas, Sandboas (Eryx), Cylindrophis, Xenopeltis, and blindsnakes (Typhlopidae), which are fossorial. The natricines are regularly semiaquatic, and the truly aquatic homalopsid snakes
occur in dry forest only where some ponds or streams have water year-round. Comparison of snake community habitat use is best done with the two Indian sites, Chatthin, and
Sakaerat, both of which have complete snake inventories. Terrestrial snakes dominate at
all four sites. The number of arboreal snakes is about one-half to two-thirds that of the
terrestrial species at Chatthin and Sakaerat, and roughly a third of the terrestrial ones
at Gir and Nallamala. Aquatic snakes are absent at Chatthin and exist as two to three
species at the other sites. All sites have a few (1–3) fossorial snakes.
The diets of frogs and most lizards are insects and other invertebrate prey; however, focused dietary studies for Asian amphibians and lizards are lacking. Indeed,
dietary data for all Asian amphibians and reptiles are largely anecdotal.
Dietary diversity among lizards is herbivory in Leiolepis (Burma and Southeast
Asia) and vertebrate carnivory in adult Gekko gecko and Varanus bengalensis. Among
turtles, the testudinids are herbivores; the aquatic species are typically identified as
carnivores, but most species tend toward omnivory. Because of the relative uniformity
among these groups, there is no evident community organization.
Snakes display the greatest diversity of prey and prey-specialization. None is an
herbivore or omnivore; all eat living or recently dead prey that the individual snake
has killed. A few specialize on invertebrates: the blindsnakes (Typhlopidae) on termites
and ants, and Pareas on terrestrial molluscans. Pareas occurs mainly in Southeast
Asian moist evergreen forest; blindsnakes likely occur at all sites although are not yet
vouchered at all. Most snakes eat vertebrates, and my prey categories (Appendix B)
emphasize this dietary preference.
The diversity of prey preference in snakes permits an examination of possible
community structure, and as in habitat choice, Nallamala, Chatthin, and Sakaerat are
the only sites examined owing to the completeness of their inventorying. The relative
frequency of the prey classes is remarkably similar (Table 3) across the three sites. The
differences occur between fish-eaters (none at Chatthin) and strictly bird predators
(only at Chatthin). Overall, snake community structure is the same among the three
sites, i.e., much more similar than different.
Relative Abundance
Availability of data on population densities of different species in the different forest
sites, and restrictions on chapter size, allow only a few observations. First and perhaps foremost is the effect of seasonality and multiyear populational fluctuation on
the visibility of species. Obviously, the size of frog populations cannot be assessed in
the dry season or for that matter for a few days or even a couple of weeks within a
single wet season. Only Inger’s Sakaerat survey (Inger and Colwell 1977; Inger 1980)
provides data allowing a quantitative assessment of relative abundance, and those data
are for a single calendar year, 1969. The Sakaerat data used herein derives not from
Inger’s published tables but from my tabulation of his specimen voucher collection
at the Field Museum of Natural History (FMNH). This reexamination of abundance
could not segregate the counts of species occurring in both the dry and the evergreen
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
289
Table 3. The relative frequency (percentage) of prey preference among the snake assemblages at Nallamala,
Chatthin, and Sakaerat. The prey classes are defined in Appendix B.
Prey Preference
Nallamala
Insects
Fish
Anurans
Reptiles
Ectotherms
Birds
Mammals
Endotherms
Vertebrates
7
10
7
27
6
0
7
7
30
Chatthin
4
0
9
23
8
4
4
9
31
Sakaerat
6
6
9
26
9
0
3
11
31
forest, hence I might have slightly over-estimated the abundance for some species. The
voucher-based data yield a total herpetofauna of 2,904 individuals (tadpole numbers
are not included): 2 (less than 0.01 percent) caecilians; 1,410 (48.6 percent) frogs;
1,242 (42.8 percent) lizards; 230 (7.9 percent) snakes; and 20 (less than 0.01 percent)
turtles. Within the frogs, Microhyla heymonsi was the most abundant taxon (235
individuals), then Polypedates leucomystax (200 individuals), Fejervarya “limnocharis” (157 individuals), and M. butleri and M. ornata (122 individuals each). Relative
abundances of the frogs decline in a smooth curve to the least abundant, Kaloula
mediolineata (5 individuals). Among the lizards, three species (Dixonius siamensis
[198 individuals], Eutropis macularia [190 individuals], and Calotes “versicolor” [189
individuals]) have near-equal abundance. Abundance declines in a step-like fashion to
a single individual each of Eutropis longicaudata and Varanus bengalensis. Gongylosoma scripta is the most abundant snake (27 individuals); thereafter snake abundance
declines sharply, but smoothly, to seven species represented by 2 individuals and six
species by 1 individual.
Our inventory vouchering at Chatthin was less intense and provides no quantitative data for comparison with Sakaerat. Frogs were definitely the most abundant component of the herpetofauna, and because of the scarcity of all lizard species, except the
commensal Hemidactylus frenatus, I estimate that frogs constitute 80 percent or more
of the total herpetofaunal abundance. Calotes “versicolor” is one of the more-abundant lizard taxa at Chatthin (where it is actually two taxa [C. htunwini, C. irawadi]),
yet transect surveys in the forest during the late dry season yielded less than one “versicolor” sighting per kilometer. Amphiesma stolatum and Lycodon aulicus were the
most common snakes, about one of each seen each week of surveying.
DISCUSSION
Species Inventories and Completeness
Species Accumulation Rates
How long does it take to inventory completely a herpetofauna of a well-demarcated
site? The species accumulation curves from Sakaerat and Chatthin demonstrated the
290
CHAPTER 15
majority (90 percent) of a herpetofauna is vouchered within one year. Our Chatthin
survey continued to find “new” species after the first year, that is, two snakes in the
second year and another one in the third year. This continual discovery phenomenon is
not unexpected and occurs even at the most intensely inventoried sites. The best confirmation of continual discovery derives from the Savannah River Ecological Laboratory
(SREL) reserve in the Piedmont of South Carolina. This reserve of more than 800 km2
of mixed forests and aquatic habitats hosts 95 species of amphibians and reptiles (Gibbons and Semlitsch 1991), determined by weekly monitoring for four decades. Yet, it
required 21 years and the capture (and release) of over 6,000 snakes before the first
pine-wood snake (Rhadinaea flavilata) was captured (Whiteman et al. 1995). Even
more striking is the rediscovery (Luhring 2008) of Brimley’s chorus frog (Pseudacris
brimleyi) at SREL after 41 years of regular surveys and over 50 years subsequent to
this species’ previous vouchering in the reserve area. These discoveries highlight the
difficulty of obtaining a total herpetofaunal inventory and the necessity of continuous
and rigorous monitoring to obtain such an inventory.
In spite of the protracted time and high man-power requirements for a total inventory, the use of species accumulation curves remains a valuable and relatively accurate
tool to assess an inventory’s success. The shared similarity of the accumulation rate of
new species (Figure 2) supports the probability of equal accessibility to the dry forest
herpetofauna in different areas. Flattening (plateauing) of the curve at different species
densities, but within a similar time-frame, indicates the robustness of the technique
and the ability to obtain a reliable assessment of a dry forest site’s herpetofauna with
a year of intense survey.
Species accumulation curves are available for only two other tropical Asian sites. A
dry forest site in the Western Ghats (Vijayakumar et al. 2006) yielded 13 amphibians
and reptiles over a four-month survey, suggesting either a depauperate herpetofauna,
or that the diurnal transect protocol used was a poor inventory strategy, hence the accumulative curve is uninformative. Bain and Nguyen (2004) examined two disturbed
moist forests in northern Vietnam and documented 36 amphibian and 16 reptile species in 26 days of intense and varied surveying. Their accumulation curve began to
flatten at the end of their third week, which might indicate about 90 percent of the local amphibian species had been inventoried; however, only 16 reptile, and particularly
only 12 snake species, suggests an incomplete inventory.
Species Composition and Abundance
Species Composition— Diversity
Of the data presented (Figure 2), the number of species for each site (Nallamala,
Chatthin, Sakaerat) represents their near-total diversity. Nallamala and Sakaerat have
near-equal diversity. Chatthin has significantly fewer species than these two sites,
although all three sites share a similar proportion of frogs, lizards, and snakes—the
main components of dry forest herpetofaunas. Diversity at the other three sites is low.
Shwe-Settaw and Cambodia were knowingly incomplete faunas when presented (Appendix A), and I anticipate that the diversity of the former will match Chatthin and
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
291
the latter Sakaerat when fully inventoried. Gir has an unrealistically low total herpetofauna of 40 species. Frog diversity (7 species; 18 percent) is a strong component of this
low diversity, and snake diversity also is low. Whether Gir is depauperate or incompletely sampled remains unclear. Another Gujarat state reserve (Vansda National Park)
has only 12 frog species (22 percent of total herpetofauna), nearly double the Gir frog
diversity. The data from both Gujarat dry forests show low frog diversity. Perhaps the
Gir herpetofauna is actually depauperate; however, I suspect that the low number of
frog species indicates an incomplete inventory.
Other Indian areas have been inventoried. Some of these inventories are of dry
forest, but often other forests are included, and discrimination of which species derive
from the dry forest is not explicit in the reports. The Agasthyamalai Hills (Tamil Nadu
State) is a dry forest mosaic, and a survey (Vijayakumar et al. 2006) at several sites
during the dry season revealed 10 species of frogs, 10 species of lizards, and 7 species
of snakes. As is common for many Indian inventories, the survey provides an initial
faunal assessment but is inadequate for reliable comparison of dry forest herpetofaunas
across tropical Asia. Aside from Sakaerat, no other Southeast Asian dry forest site has
been inventoried.The similarity of Nallamala and Sakaerat diversity suggests a shared
environment factor. That factor may be an adjacent evergreen forest, which serves as a
population reservoir for some less arid-adapted species. In contrast, Chatthin is distant
from a moist forest habitat, and this isolation yields a fauna that can persist under
alternating extremes of wet and dry. This association of decreasing diversity with increasing aridity is a general phenomenon of tropical forest (Heatwole 1982), although
the concept does not appear to have been linked with differential regional diversity.
Finally, in spite of differences in the number of total species and the broad geographic distances between sites, the proportional similarities among Nallamala, Chatthin, and Sakaerat are striking and unexpected. An explanation is not readily apparent.
Species Composition—Taxonomy
The proportional similarities of the herpetofaunal components might obtain from
numerous shared species. I earlier noted, however, the low likelihood of shared species among India, Burma, and Southeast Asia. The sharing is mainly of membership in
species groups, hence similarity in appearance (morphology), behavior, and ecology.
Using the concept of ecological or niche taxon-equivalents permits interregional examination of these “taxonomic-ecological” equivalent taxa to interpret the numerous
similarities in community taxonomic organization.
Frogs and their inability to prevent and tolerate dehydration are, surprisingly, a
major component of the dry forest community and, herpetologically, the component
that becomes numerically dominant when the monsoon begins. The most abundant
paddy or pond-side frogs consist of two taxon-groups, Fejervarya and Hoplobatrachus
(medium and large species, respectively). These two groups occur at all sites (occasionally a dry forest site has two differently sized species of one of them), and at all sites,
Fejervarya is abundant throughout the first half of the monsoon, while Hoplobatrachus is most visible during the initial weeks and its own reproductive splurge. A waterside skittering species (Euphlyctis [India], Occidozyga [Burma, Southeast Asia]) occurs
292
CHAPTER 15
abundantly at all sites. Typically two or more Microhyla occur at each site; these are
small semifossorial taxa. A much larger microhylid (Glyphoglossus [Burma, Southeast
Asia], Uperodon [India]) is present at all sites. There is a conundrum in the dominance
of terrestrial frogs in dry forest sites, considering the duration and intensity of aridity
at these sites, but the smaller number of aquatic and semiaquatic frogs correlates with
the presence of less-permanent water.
Similar interregional taxon-equivalents exist among lizards and snakes. Within
lizards, the arboreal, and often upside-down on a tree trunk, Calotes “versicolor” occurs broadly in Asian dry forest. Hemidactylus geckos, typically two or more species,
are present; often one is predominantly terrestrial in spite of the enlarged digital pads,
and the other occurs on tree trunks. Several skinks are dry forest residents. Lygosoma
is always present, and this elongate lizard lives beneath the floor litter.
Snakes are represented in Asian dry forests by a variety of taxon-equivalents,
ranging from diurnal (Ahaetulla, Dendrelaphis) and nocturnal (Boiga) tree snakes to
semifossorial hognose snakes (Oligodon). A Python species occurs throughout the dry
forest zone, as also do large ratsnakes (Ptyas, Coelognathus) and a modest-sized terrestrial ambush predator (Echis [India], Daboia [Burma, Southeast Asia], Calloselasma
[Southeast Asia]).
Even with the uncertainty of species boundaries and distributions that encompassed more than one geographic area, I identified geographic pairs of frogs, lizards,
and snakes for India-Burma and Burma–Southeast Asia. Numerically, there are more
Burma–Southeast Asia pairs in all three groups. No India–Southeast Asia pair exists,
although a few “species” occur in all three areas.
Finally, I note the rarity of dry forest endemics. A few exist, such as Naja mandalayensis, but presently it is impossible to predict whether this low endemicity results from
inadequate inventory of dry forest herpetofaunas, hence poor sampling and inadequate
taxonomic study, or the broad ecological tolerance of dry forest species and their reproductive success in other habitats—ones less physiologically stressful although potentially
with more predators and competitors. It is also noteworthy that many of the widespread
taxon-equivalent species live successfully in anthropogenic habitats.
Species Composition—Size Structure
Comparison of communities by body size shows a more-variable structure. Part of this
variation arises from the use of proportional representation of community components.
Proportional representation is quasi-quantification, and the smaller the fauna, the
greater is the effect of the presence or absence of a single species or size class in altering
the depiction of organization. Again, the focus is on the three well-surveyed sites.
Small and medium-sized frogs are the dominant classes at all three sites, particularly so at Sakaerat with nearly 50 percent of the species in the small class (Figure 4).
The proportion of large and big frogs declines from India to Southeast Asia. For lizards,
Chatthin lacks the smallest class; the medium-small and medium classes dominate and
are proportionately equal. This pattern contrasts with Nallamala and Sakaerat. In the
former, the small and medium-large classes are well represented, although there are
somewhat fewer medium-small and medium classes. At Sakaerat, the medium-small
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
293
classes numerically dominate, and combined with the medium-large class, these lizards comprise approximately 65 percent of the lizard fauna. The two largest classes
are similar for the three sites. Snakes at the major sites are represented by one or two
small and one large species (Python). The middle three size classes dominate (over 80
percent) at the three sites, although in a slightly different way at each. Medium-large
snakes have the lowest representation at all sites; medium-large snakes are about onehalf the snake species at Nallamala, medium-sized ones the majority at Chatthin, and
medium-small and medium snakes roughly equal at Sakaerat. These different frequencies might reflect different dietary cohorts at the three sites; however, diet structuring
of the snake assemblages does not match the size organization.
Behavioral Preferences
Of the three behaviors, only diets suggest community organization. There is a distinct
taxon/clade association of diurnal and nocturnal behavior. Habitat selection displays
more variety in choice, and within all three herp groups, proportional representation
among the habitat classes is similar at Nallamala, Chatthin, and Sakaerat.
Diet is relatively uniform among frogs and lizards. As in the previous two behaviors, snakes utilize a broad variety of prey and, depending upon the species, can be
highly specialized or generalists within a broad prey class. The snake communities of
the three intensively surveyed sites are similarly structured (Table 3).
Cadle and Greene (1993) examined community structure among neotropical rain
forest snakes. Although their emphasis was phylogenetic, the general structuring of
their broad community possesses organization similarities in each behavioral category
of this study. These similarities, in spite of the snake communities’ occupancy of different continents and habitat types, support Cadle and Greene’s advocacy of phylogenetic
relationships in community organization. The similarity also demonstrates how snake
diversity enables more species to occupy more niches and, as a group, to dominate species representation in a herpetological community but at the “cost” of low abundance.
Relative Abundance
The abundance numbers are derived from the Sakaerat voucher collection (FMNH);
some of these data are also presented in Inger’s analyses (Inger and Colwell 1977, Tables
6, 7; Inger 1980, Tables 1, 2) but partitioned differently. The two data tabulations show
the same pattern for lizards and frogs (all species of each) and show similar relative abundance. In both lizards and frogs, a few species have high abundance, but most species
occur at much lower densities. Lizards numerically dominate at Sakaerat, although they
are much less abundant at Chatthin. This latter observation is not supported yet by data
analysis, as abundance data for other Asian dry forest sites are lacking.
In Other Forests
The similarity of community structure among tropical Asian dry forest herpetofaunas was unexpected, and immediately generated questions on the level of similarity
294
CHAPTER 15
between faunas of dry forest and nearby evergreen forest, and between Asian dry
forests and those of other continents. Some tentative answers are possible; a more
rigorous examination will follow.
Again relying on the Sakaerat data of Inger and Colwell (1977) and examining
only species diversity, we find the adjacent evergreen forest has 87 species, 32 percent
each for frogs and lizards, 34 percent for snakes, and 1 percent for turtles. This moister
forest has proportionately more frogs and lizards, and fewer snakes. A dry forest area
(approximately 10°28′ N 85°22′ W) in Guanacaste Province, Costa Rica (Scott et al.
1983) has 78 species composed of 1 (1 percent) caecilian, 22 (28 percent) frogs, 3 (4
percent) turtles, 17 (22 percent) lizards, and 35 (45 percent) snakes. The total species
matches the Sakaerat herpetofauna (Table 2), but the community has fewer lizard species and more frogs. Proportionately, the pattern is similar to that for other tropical
Asian dry forests.
PRÉCIS
1. The near-total herpetofauna (greater than 90 percent) of Asian dry forest sites
can be determined in less than one year by intensive weekly inventories using a
variety of sampling techniques.
2. Three sites, Nallamala, Chatthin, and Sakaerat (representing India, Burma, and
Southeast Asia, respectively), serve as the major herpetofaunas for examining
community structure in tropical Asian dry forest habitats owing to the intensity
of survey effort at these sites.
3. The total Asian dry forest herpetofauna contains more than 45 species of amphibians and 130 species of reptiles, more than one-half of which are snakes.
4. The α-diversity of Nallamala and Sakaerat are nearly equal; the Chatthin herpetofauna is roughly two-thirds of the Nallamala and Sakaerat faunas.
5. Although the total number of species differs between sites, the relative frequency
of frogs, lizards, and snakes is quite similar in each dry forest community.
6. A few taxa occur throughout the breadth of Asia’s dry forest, but the majority
(greater than 50 percent) of the dry forest taxa in each region (India, Burma,
Southeast Asia) represents different species, although they are representatives of
shared species groups.
7. Size-class organization of the various communities differs slightly among the
regional dry forests, especially in the frequencies of the mid-range size classes.
8. There is no apparent community organization in activity patterns or habitat use
among frogs, lizards, and snakes.
9. Only snakes show community organization in diet, and the same organization
occurs among the three well-surveyed forests.
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
295
ACKNOWLEDGMENTS
The Chatthin herpetological team (1997–1999: Htun Win† and Thin Thin† [team
leaders], Win Zaw Lhon, Than Zaw Min, and Kyaw Kyaw†; 1999–2000: Mya Than
Da Nyeine [team leader], Kyaw Kyaw†, Than Zaw Min, Win Zaw Lhon, Kyi Aung,
and Kyaw Zin Tun) was responsible and essential for year-round monitoring of the
Chatthin herpetofauna. I greatly appreciate and thank them for their enthusiastic assistance. Special thanks go to Alan Resetar (Field Museum of Natural History) and
Jens Vindum (California Academy of Sciences), who queried their respective catalog
databases to provide me with date-of-first-capture data for examining species-accumulation rates for the Sakaerat, Shwe-Settaw, and Min-Gon-Taung herpetofaunas.
The Biodiversity Survey and Inventory Program of the Smithsonian’s National
Museum of Natural History provided support (1997–2000) for my travel to Myanmar and for the Chatthin monitoring team’s monthly monitoring and inventory
work. The National Science Foundation Biodiversity Surveys and Inventories program
(DEB-9971861 and DEB-0451832) has supported our (the herpetological staffs of
the California Academy of Sciences, the National Museum of Natural History, and
the Myanmar Nature and Wildlife Conservation Division) all-country survey of the
Burmese herpetofauna. The preceding survey has provided the data on the Burmese
dry forest sites reported herein. These data would have been unattainable without the
logistic support and encouragement of directors of the Myanmar Nature and Wildlife
Conservation Division: U Uga, U Khin Maung Zaw, and U Tin Tun. Our survey team
(1999–2004) was exceptionally diligent, and I am most appreciative of the excellence
of their survey and inventory work, often under difficult environmental and climatic
conditions. They are Awan Kien Shain, Hla Tun, Htun Win†, Kyi Soe Lwin, Sai
Wunna Kyi, San Lwin Oo, and Thin Thin†.
As always, I have received assistance from colleagues and readily thank them for
their help. Raoul Bain, Sayantan Biswas, and Whit Gibbons provided prompt and
detailed assistance with references on the herpetofaunas of Indochina, India, and the
Savannah River Reserve, respectively. Several colleagues (Raoul Bain, Steve Busack,
Ron Crombie, Al Leviton, Rom Whitaker, and Pat Zug) reviewed drafts of this chapter and eliminated a variety of errors; I am responsible for those that remain and any
failure to accept their good advice.
296
CHAPTER 15
APPENDIX A
Asian Dry Forest Herpetofaunas
The species are arranged alphabetically within general group by family, then by genus, and finally by species. Taxonomy has been updated from original sources1 to conform to Frost’s amphibian and Uetz’s
and Hallerman’s reptile websites (June 2008).
Taxon
Gir1
Nallam1
Chatthin1
Shwe-S1
Sakaerat1,2 Cambodia1,3
–
–
–
–
+
–
–
+
+
+
+
+
–
–
+
–
–
+
–
–
+
–
–
+
+
+
–
–
+
–
–
–
–
–
–
+
–
+
+
–
–
+
–
+
–
+
–
–
+
+
–
–
+
+
–
–
–
+
+
+
–
+
–
–
–
+
+
–
–
–
+
–
+
–
+
–
–
–
–
–
–
+
–
+
–
+
+
–
–
–
–
–
–
–
+
–
+
–
+
+
–
–
–
–
–
–
–
–
–
–
+
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
+
–
+
–
–
–
+
+
+
–
+
+
–
+
–
–
–
+
–
–
–
+
–
–
–
–
–
–
–
–
+
–
–
–
+
+
+
+
–
–
–
–
–
+
+
–
+
+
+
+
+
+
+
–
–
–
+
–
–
–
–
–
–
–
–
–
–
+
+
+
–
–
–
–
–
–
–
–
–
–
–
–
+
–
–
+
–
–
–
CAECILIANS
Ichthyophiidae
Ichthyophis kohtaoensis
FROGS
Bufonidae
“Bufo” scaber
“Bufo” stomaticus
Duttaphrynus melanostictus
Dicroglossidae
Euphlyctis cyanophlyctis
Euphlyctis hexadactyla
Fejervarya “limnocharis”/std
Fejervarya “limnocharis”/small
Fejervarya “limnocharis”/India
Fejervarya “limnocharis”/Thai
Hoplobatrachus crassus
Hoplobatrachus rugulosus
Hoplobatrachus tigerinus
Occidozyga lima
Occidozyga martenseii
Sphaerotheca breviceps
Sphaerotheca dobsoni
Microhylidae
Calluella guttulata
Glyphoglossus molossus
Kalophrynus interlineatus
Kaloula mediolineata
Kaloula pulchra
Microhyla berdmorei
Microhyla butleri
Microhyla heymonsi
Microhyla ornata
Microhyla pulchra
Microhyla rubra
Microhyla sp.
Microhyla sp.-mini
Micryletta inornata
Ramanella variegata
Uperodon globulosus
Uperodon systoma
Ranidae
Hylarana erythraea
Hylarana macrodactyla
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
Taxon
Hylarana taipehensis
Hylarana sp.
Pelophylax lateralis
Rana johnsi
Ranixalidae
Indirana leithii
Rhacophoridae
Chiromantis nongkhorensis
Chiromantis vittatus
Polypedates leucomystax
Polypedates maculatus
297
Gir1
Nallam1
Chatthin1
Shwe-S1
Sakaerat1,2 Cambodia1,3
–
–
–
–
–
+
–
–
–
–
+
–
–
–
–
–
–
–
+
–
+
–
–
+
–
+
–
–
–
–
–
–
–
–
–
–
–
+
+
–
+
–
–
–
–
–
+
+
+
–
+
–
+
–
–
–
–
–
–
–
+
+
–
–
–
–
+
–
+
–
–
+
–
–
–
–
–
–
+
–
–
+
–
–
–
–
+
–
+
+
–
–
+
–
–
–
–
+
–
–
+
–
+
–
+
–
–
+
–
–
–
–
–
–
–
–
–
–
–
+
–
–
–
–
–
+
–
–
–
–
+
+
–
–
–
+
+
+
–
+
+
+
–
–
+
–
–
–
–
–
–
+
+
+
–
–
+
–
–
–
–
–
–
–
–
+
–
+
–
+
+
–
–
–
+
–
–
+
–
+
–
–
–
–
–
–
+
+
–
–
–
–
–
–
–
+
–
+
–
–
–
–
–
–
–
–
–
–
+
–
–
–
–
–
+
–
+
–
–
–
–
–
TURTLES
Geoemydidae
Cyclemys fusca
Cyclemys oldhamii
Melanochelys trijuga
Pangshura tentoria
Testudinidae
Geochelone elegans
Geochelone platynota
Indotestudo elongata
Trionychidae
Amyda cartilaginea
Lissemys punctata
Lissemys scutata
LIZARDS
Agamidae
Bronchocoela smaragdina
Calotes htunwini
Calotes irawadi
Calotes mystaceus
Calotes rouxii
Calotes “versicolor”
Leiolepis belliana
Leiolepis reevesii
Physignathus cocincinus
Psammophilus blanfordanus
Psammophilus dorsalis
Sitana ponticeriana
Chamaeleonidae
Chamaeleo zeylanicus
Gekkonidae
Cnemaspis sp.
Cyrtodactylus sp.
Dixonius siamensis
Geckoella collegalensis
Gehyra lacerata
(Continued)
298
CHAPTER 15
Taxon
Gehyra mutilata
Gekko gecko
Hemidactylus aquilonius
Hemidactylus brookii
Hemidactylus flaviviridis
Hemidactylus frenatus
Hemidactylus giganteus
Hemidactylus karenorum
Hemidactylus leschenaultii
Hemidactylus platyurus
Hemidactylus reticulatus
Hemidactylus thayene
Hemidactylus triedrus
Lacertidae
Ophisops jerdonii
Ophisops leschenaultii
Ophisops minor
Takydromus sexlineatus
Scincidae
Eutropis carinata
Eutropis dissimilis
Eutropis longicaudata
Eutropis macularia
Eutropis multifasciata
Eutropis novemcarinata
Eutropis quadricarinata
Lygosoma albopunctatum
Lygosoma bowringii
Lygosoma guentheri
Lygosoma lineolatum
Lygosoma punctatum
Lygosoma quadrupes
Scincella reevesii
Sphenomorphus indicus
Sphenomorphus maculatus
Varanidae
Varanus bengalensis
Gir1
Nallam1
Chatthin1
Shwe-S1
Sakaerat1,2 Cambodia1,3
–
–
–
+
+
–
–
–
–
–
–
–
+
–
–
–
+
+
+
+
–
+
–
+
–
+
–
+
+
–
–
+
–
–
–
–
–
–
–
–
+
–
–
–
+
–
+
–
–
–
+
–
+
+
–
–
–
+
–
–
–
+
–
–
–
+
+
–
–
–
–
–
–
–
+
–
–
–
+
–
–
–
+
+
+
–
–
–
–
–
–
–
–
–
–
–
–
+
–
–
–
–
+
–
–
+
–
–
–
+
–
+
–
+
–
–
–
–
+
–
–
+
–
–
–
–
–
+
–
+
–
–
–
–
–
+
–
–
–
+
+
–
–
–
+
–
–
–
–
+
–
–
–
–
+
+
–
–
–
–
+
–
–
–
+
+
–
–
+
+
+
–
–
–
+
–
–
–
–
+
–
+
–
–
–
+
+
–
–
–
–
–
–
–
+
–
–
–
+
+
+
+
+
–
–
–
+
+
–
–
–
–
–
–
–
–
+
–
+
+
–
–
–
–
–
+
–
–
+
–
+
–
–
–
+
–
–
–
–
–
+
+
+
+
–
–
–
–
–
+
–
–
+
+
–
–
+
–
+
–
+
–
–
–
–
–
–
–
–
–
SNAKES
Boidae
Eryx conicus
Eryx johnii
Colubridae-Colubrinae
Ahaetulla nasuta
Ahaetulla prasina
Ahaetulla pulverulenta
Argyrogena fasciolata
Boiga cyanea
Boiga forsteni
Boiga multomaculata
Boiga ochracea
Boiga ocellata
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
Taxon
Boiga trigonata
Chrysopelea ornata
Coelognathus flavolineata
Coelognathus helena
Coelognathus radiatus
Dendrelaphis pictus
Dendrelaphis subocularis
Dendrelaphis tristis
Dryocalamus davisonii
Gonyosoma oxycephalum
Liopeltis calamaria
Liopeltis stoliczkae
Lycodon aulicus
Lycodon laoensis
Lycodon striatus
Lycodon travancoricus
Oligodon arnensis
Oligodon cinereus
Oligodon planiceps
Oligodon quadrilineatus
Oligodon splendidus
Oligodon taeniatus
Oligodon taeniolatus
Oligodon theobaldi
Oligodon travancoricus
Ptyas korros
Ptyas mucosus
Sibynophis collaris
Sibynophis subpunctatus
Colubridae-Natricinae
Amphiesma stolatum
Atretium schistosum
Macropisthodon plumbicolor
Rhabdophis chrysargos
Rhabdophis nigrocinctus
Rhabdophis subminiatus
Xenochrophis flavipunctatus
Xenochrophis piscator
Cylindrophiidae
Cylindrophis ruffus
Elapidae-Elapinae
Bungarus caeruleus
Bungarus fasciatus
Calliophis maculiceps
Calliophis melanurus
Naja naja
Naja kaouthia
Naja mandalayensis
Elapidae-Psammophiinae
299
Gir1
Nallam1
Chatthin1
Shwe-S1
Sakaerat1,2 Cambodia1,3
+
+
–
–
–
–
–
–
–
–
–
–
–
–
+
–
+
–
–
–
–
–
+
–
–
–
+
–
+
+
–
–
+
–
–
–
+
–
–
+
–
+
–
+
+
+
–
–
–
–
–
+
–
+
–
+
–
+
–
+
–
–
+
–
+
–
–
–
–
–
+
–
–
–
–
–
–
–
+
–
–
+
–
–
+
+
–
–
+
–
–
+
–
–
–
–
–
–
–
+
–
–
–
–
–
+
–
+
–
–
+
–
–
+
–
–
–
+
+
–
+
+
+
–
+
+
–
–
+
+
–
–
–
+
–
+
–
+
–
–
–
+
+
+
–
–
–
–
–
–
–
–
–
–
–
–
+
–
–
–
–
–
+
–
–
–
–
–
–
–
+
–
–
–
–
–
–
–
–
–
–
+
+
+
+
–
–
–
–
+
+
–
–
–
–
+
–
–
+
–
–
–
–
–
–
+
+
–
–
+
+
+
+
–
–
–
–
–
–
+
–
–
–
–
+
–
–
–
–
–
–
+
+
–
–
+
–
–
–
+
–
–
–
+
–
–
–
–
+
–
–
–
–
–
–
–
–
+
+
–
–
+
–
–
+
–
–
–
–
–
(Continued)
300
CHAPTER 15
Taxon
Psammophis condanarus
Psammophis leithii
Homalopsidae
Enhydris enhydris
Enhydris plumbea
Homalopsis nigroventralis
Pareatidae
Pareas carinatus
Pareas margaritophorus
Pythonidae
Python molurus
Python reticulatus
Typhlopidae
Grypotyphlops acutus
Ramphotyphlops braminus
Typhlops porrectus
Viperidae
Calloselasma rhodostoma
Crypteltropis albolabris
Daboia russelii
Daboia siamensis
Echis carinatus
Trimeresurus gramineus
Viridovipera vogeli
Xenopeltidae
Xenopeltis unicolor
Gir1
Nallam1
Chatthin1
Shwe-S1
Sakaerat1,2 Cambodia1,3
–
+
–
–
–
–
+
–
–
–
–
–
–
–
–
+
–
–
–
–
–
–
–
–
–
+
–
–
+
+
–
–
–
–
–
–
–
–
+
–
–
+
+
–
+
–
+
–
–
–
–
+
–
–
–
+
+
+
+
–
–
+
–
–
–
–
–
+
–
–
–
–
–
–
–
–
+
–
–
–
–
+
–
+
+
–
–
+
–
+
–
–
–
–
–
–
+
–
–
–
+
+
–
–
–
–
+
+
–
–
–
–
–
–
–
–
+
–
+
–
1
Species occurrence sources: Gir—Bhatt et al. 1999, and Vyas 2000; Nallamala—Srinivasulu and Das 2008; Chatthin—Zug et al.
1998, personal data, CAS, MBM-NWCD, and USNM; Shwe-Settaw—Zug et al. 2004, CAS, and USNM; Sakaerat—Inger and Colwell
1977 (Table 1), taxonomy updated (June 2008) from FMNH database; hilly eastern Cambodia—Stuart et al. 2006.
2
This list includes the species occurrences for both the deciduous forest and agricultural lands listed separately in Table 1 of Inger and Colwell (1977). Our sampling at Chatthin and Shwe-Settaw was not as precisely recorded as Inger’s more rigorous data
gathering; additionally he notes, “Though collected only in agricultural land, species almost certainly occur[s] in deciduous forest.”
3
This list of species occurrences (Stuart et al. 2006) includes records from anthropogenic habitats (Table 2 in Stuart et al.) and dry
deciduous forest. Species from evergreen forest were purposefully excluded, even if deciduous forest occurs at the site, because
data did not allow discrimination of precise habitat origin of voucher specimens.
APPENDIX B
Size and Ecological Coding for Asian Dry Forest Amphibians and Reptiles
Size—Each of the four groups (frogs, turtles, lizards, snakes) is individually categorized
for body size (carapace length [cl] for turtles; snout-vent length [svl] for the other three
groups). All size data are in millimeters (mm). Adult size data derive from numerous
sources (literature; G. Zug, unpublished data) and include ranges for females, ranges
for males, median/mean for females, and maximum size. These data and the subsequent ecological coding-data are available from the author, as space is not available
for presentation in this book. From the size data, incomplete for many taxa, I selected
a midpoint size representative for each taxon using the median or mean when available or an estimate of the midpoint when median or mean was not available. These
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
301
midpoints for each herp group were plotted (bar graphs) to identify size classes by
clustering of midpoints. This strategy yielded five size classes for frogs (less than 20,
22–44, 46–70, 72–96, and greater than 96 mm); three classes for turtles (less than
230, 250–300, and greater than 400 mm); six classes for lizards (less than 42, 44–62,
64–94, 98–118, 120–148, and greater than 150 mm); and five classes for snakes (less
than 300, 300–590, 600–990, 1,000–1,500, and greater than 1,500 mm).
Habitat Preference—These ecological categories are at a gross level, in part because of
our incomplete knowledge for many taxa but also because broader categories allow a
more even comparison of community structure between dry-forest sites. The categories
are fossorial-semifossorial, terrestrial, arboreal (usually found off the ground in shrubs
and trees), semiaquatic (typically waterside, often feeding there), and aquatic (uncommonly found outside of water).
Activity Time—This ecological parameter identifies whether a taxon pursues most of
its life-history activities during daylight (diurnal) or at night (nocturnal). Observations
are insufficient to recognize any truly crepuscular taxa, and many diurnal reptile taxa
shift to dawn and twilight behavior when daily temperatures soar.
Diet—Most amphibians and reptiles are carnivorous as juveniles and adults, hence herbivory is not subdivided. A few reptiles, mainly turtles, are omnivores. The carnivores’
prey are partitioned into nine categories: insects and other invertebrates, fish, amphibians, reptiles, ectothermic vertebrates, birds, mammals, endothermic vertebrates, and
vertebrates (amphibians to mammals). Some taxa specialize in one life-history stage,
for example, reptile eggs for Oligodon; such specializations are not coded, and the
taxon’s diet is considered as one of the preceding nine categories.
NOTE
1. This study and chapter are dedicated to Robert F. Inger and the staff of the Chatthin
Wildlife Sanctuary 1997–2000. Dr. Inger’s studies of the Asian herpetofauna span the full spectrum of systematics and ecology, and are a hallmark of thoughtful design, intensive fieldwork,
and rigorous analysis. His Sakaerat community research encouraged and guided my research at
Chatthin Wildlife Sanctuary. The friendship and enthusiastic support of the Chatthin staff made
my research visits productive and enjoyable.
REFERENCES
Bain, R. H., and N. Q. Truong. 2004. Herpetofaunal Diversity of Ha Giang Province in
Northeastern Vietnam, with Descriptions of Two New Species. American Museum Novitate
3454:1–42.
Bhatt, K., R. Vyas, and M. Singh. 1999. Herpetofauna of Gir Protected Area. Zoos’ Print Journal 14 (5): 27–30.
Cadle, J. E., and H. W. Greene. 1993. Phylogenetic Patterns, Biogeography, and the Ecological
Structure of Neotropical Snake Assemblages. In Species Diversity in Ecological Communities:
302
CHAPTER 15
Historical and Geographical Perspectives, ed. R. E. Ricklefs and D. Schluter, 281–93. Chicago: University of Chicago Press.
Campden-Main, S. 1970. A Field Guide to the Snakes of South Vietnam, i–iv, 1–114. Washington, DC: Division of Reptiles and Amphibians, National Museum of Natural History,
Smithsonian Institution.
Cherian, P. T., K. Rema Devi, and M. S. Ravichandran. 2000. Ichthyo and Herpetofaunal Diversity of Kalakad Wildlife Sanctuary. Zoos’ Print Journal 15 (2): 203–6.
Drury, W. H., Jr. 1998. Chance and Change: Ecology for Conservationists, i–xxiii, 1–223.
Berkeley: University of California Press.
Dubois, A. 1975. Un nouveau complexe d’espèces jumelles distinguées par le chant: Les grenouilles du Népal voisines de Rana limnocharis Boie (amphibiens, anoures). Comptes rendus
de l’Académie des Sciences, Paris, séries D 281, 1717–20.
———. 1984. Note préliminaire sur le groupe de Rana limnocharis Gravenhorst, 1829 (amphibiens, anoures). Alytes 3 (4): 143–59.
Dutta, S. K. 1997. A New Species of Limnonectes (Anura: Ranidae) from Orissa, India. Hamadryad 22 (1): 1–8.
Dutta, S. K., and N. Singh. 1996. Status of Limnonectes limnocharis (Anura: Ranidae) Species
Complex in Asia. Zoos’ Print Journal 11 (8): 15, 21.
Frost, D. R. 2007. Amphibian Species of the World: An Online Reference. Version 5.1, http://
research.amnh.org/herpetology/amphibia/index.php (accessed October 2007).
Gibbons, J. W., and R. D. Semlitsch. 1991. Guide to the Reptiles and Amphibians of the Savannah River Site. Athens: University of Georgia Press.
Heatwole, H. 1982. A Review of Structuring in Herpetofaunal Assemblages. In Herpetological
Communities, ed. N. J. Scott, Jr., 1–19. Wildlife Research Report 13. U.S. Department of the
Interior Fish and Wildlife Service.
Inger, R. F. 1954. Systematics and Zoogeography of Philippine Amphibia. Fieldiana: Zoology
33 (4): 183–531.
———. 1980. Relative Abundance of Frogs and Lizards in Forests of Southeast Asia. Biotropica
12 (1): 14–22.
Inger, R. F., and R. K. Colwell. 1977. Organization of Contiguous Communities of Amphibians
and Reptiles in Thailand. Ecological Monographs 47:229–53.
Köhler, J., D. R. Vietes, R. M. Bonett, F. H. García, F. Glaw, D. Steinke, and M. Vences. 2005.
New Amphibians and Global Conservation: A Boost in Species Discoveries in a Highly Endangered Vertebrate Group. BioScience 55 (8): 693–96.
Kuramoto, M., S. H. Joshy, A. Kurabayashi, and M. Sumida. 2007. The Genus Fejervarya
(Anura: Ranidae) in Central Western Ghats, India, with Descriptions of Four New Cryptic
Species. Current Herpetology 26 (2): 81–105.
Luhring, T. M. 2008. “Problem Species” of the Savannah River Site, Such as Brimley’s Chorus
Frog (Pseudacris brimleyi), Demonstrate the Hidden Biodiversity Concept of an Intensively
Studied Government Reserve. Southeastern Naturalist 7 (2): 371–73.
Matsui, M., H. Ito, T. Shimada, H. Ota, S. K. Saidapur, W. Khonsue, T. Tananka-Ueno, and
G.-F. Wu. 2005. Taxonomic Relationships within the Pan-Oriental Narrow-Mouth Toad Microhyla ornata as Revealed by mtDNA Analysis (Amphibia: Anura, Microhylidae). Zoological
Science 22:489–95.
Matsui, M., T. Seto, and T. Utsunomiya. 1986. Acoustic and Karyotypic Evidence for Specific
Separation of Polypedates megacephalus from P. leucomystax. Journal of Herpetology 20
(4): 483–89.
Matsui, M., M. Toda, and H. Ota. 2007. A New Species of Frog Allied to Fejervarya limnocharis from the Southern Ryukyus, Japan (Amphibia: Ranidae). Current Herpetology 26 (2):
65–79.
TROPICAL ASIAN DRY FOREST AMPHIBIANS AND REPTILES
303
Orlov, N. L., A. Lathrop, R. W. Murphy, and Ho Thu Cuc. 2001. Frogs of the Family Rhacophoridae (Anura: Amphibia) in the Northern Hoang Lien Mountains (Mount Fan Si Pan, Sa
Pa District, Lao Cai Province), Vietnam. Russian Journal of Herpetology 8 (1): 17–44.
Scott, N. J., J. M. Savage, and D. C. Robinson. 1983. Checklist of Reptiles and Amphibians. In
Costa Rican Natural History, ed. D. H. Janzen, 367–74. Chicago: University of Chicago Press.
Srinivasulu, C., and I. Das. 2008. The Herpetofauna of Nallamala Hills, Eastern Ghats, India:
An Annotated Checklist, with Remarks on Nomenclature, Taxonomy, Habitat Use, Adaptive
Types and Biogeography. Asiatic Herpetological Research 11:110–31.
Stuart, B. L., K. Sok, and T. Neaung. 2006. A Collection of Amphibians and Reptiles from
Eastern Hilly Cambodia. Raffles Bulletin of Zoology 54 (1): 129–55.
Uetz, P., and J. Hallermann. 2008. The Tiger Reptile Database. www.reptile-database.org.
Veith, M., J. Kosuch, A. Ohler, and A. Dubois. 2001. Systematics of Fejervarya limnocharis
(Gravenhorst, 1829) (Amphibia, Anura, Ranidae) and Related Species. 2. Morphological and
Molecular Variation in Frogs from the Greater Sunda Islands (Sumatra, Java, Borneo) with
the Definition of Two Species. Alytes 19 (1): 5–28.
Vijayakumar, S. P., A. Ragavendran, and B. C. Choudhury. 2006. Herpetofaunal Assemblage in
a Tropical Dry Forest Mosaic of Western Ghats, India: Preliminary Analysis of Species Composition and Abundance during the Dry Season. Hamadryad 30 (1, 2): 41–54.
Vyas, R. 2000. Supplementary Notes on Herpetofauna of Gir Forets [sic]. Zoos’ Print Journal
15 (5): 263–64.
———. 2004. Herpetofauna of Vansda National Park, Gujarat. Zoos’ Print Journal 19 (6):
1512–14.
Whiteman, H. H., T. M. Mills, D. E. Scott, and J. W. Gibbons. 1995. Confirmation of Range
Extension for the Pine Woods Snake (Rhadinaea flavilata). Herpetological Review 26 (3): 158.
Wikramanayake, E., E. Dinerstein, C. J. Loucks, D. M. Olson, J. Morrison, J. Lamoreux, M.
McKnight, and P. Hedao. 2002. Terrestrial Ecoregions of the Indo-Pacific: A Conservation
Assessment. Washington, DC: Island Press.
Wüster, W. 1996. Taxonomic Changes and Toxicology: Systematic Revisions of the Asiatic
Cobras (Naja naja species complex). Toxicon 34 (4): 399–406.
Wüster, W., and R. S. Thorpe. 1989. Population Affinities of the Asiatic Cobra (Naja naja) Species Complex in South-East Asia: Reliability and Random Resampling. Biological Journal of
the Linnean Society 36:391–409.
———. 1990. Systematics and Biogeography of the Asiatic Cobra (Naja naja) Species Complex
in the Philippine Islands. In Vertebrates in the Tropics, ed. G. Peters and R. Hutterer, 333–44.
Bonn: Museum Alexander Koenig.
———. 1992. Asiatic Cobra: Population Systematics of the Naja naja Species Complex (Serpentes: Elapidae) in India and Central Asia. Herpetologica 48 (1): 69–85.
Zug, G. R., H. H. K. Brown, J. A. Schulte II, and J. V. Vindum. 2006. Systematics of the Garden
Lizards, Calotes versicolor Group (Reptilia, Squamata, Agamidae), in Myanmar: Central Dry
Zone Populations. Proceedings of the Californian Academy of Sciences 57 (1): 1–33.
Zug, G. R., Htun Win, Thin Thin, Than Zaw Min, Win Zaw Lhon, and Kyaw Kyaw. 1998.
Herpetofauna of the Chatthin Wildlife Sanctuary, North-Central Myanmar with Preliminary
Observations of Their Natural History. Hamadryad 23:111–20.
Zug, G. R., Sai Wunna Kyi, and Htun Win. 2004. Shwesettaw Wildlife Sanctuary (Myanmar)
Turtles. Turtle & Tortoise Newsletter 8:4.
George Zug