Content uploaded by Akio Takahashi
Author content
All content in this area was uploaded by Akio Takahashi on Jan 09, 2024
Content may be subject to copyright.
Content uploaded by Akio Takahashi
Author content
All content in this area was uploaded by Akio Takahashi on Jan 21, 2014
Content may be subject to copyright.
Current Herpetology 26(1): 1–11, June 2007
2007 by The Herpetological Society of Japan
A New Species of the Genus Geoemyda (Chelonii:
Geoemydidae) from the Upper Pleistocene of Tokunoshima
Island, the Central Ryukyus, Japan
AKIO TAKAHASHI1, TAKAFUMI KATO2, AND HIDETOSHI OTA1*
1Tropical Biosphere Research Center, University of the Ryukyus, Nishihara, Okinawa
903–0213, JAPAN
2College of Life Science, Department of Health Science, Kurashiki University of Science
and the Arts, Tsurajima, Kurashiki, Okayama 712–8505, JAPAN
Abstract: Turtle fossils discovered from the putative Upper Pleistocene fissure-
filling deposit on Tokunoshima Island of the Amami Group is described as a
new species, Geoemyda amamiensis (Chelonii: Geoemydidae). These fossils
include a carapace exclusive of peripherals and a few other elements, fragments
of a neural, a costal, eight peripherals, an anterior half of the plastron, two
epiplastra, and an incomplete right humerus. Geoemyda amamiensis most
resembles G. japonica, an extant species endemic to three islands of the
Okinawa Group, but is distinguished from the latter by the presence in dorsal
view of a short anterior projection in the entoplastron. These two species seem
to have split from each other through vicariance within the central Ryukyus.
Key words: Geoemydidae; New species; Pleistocene; Ryukyu Archipelago;
Tokunoshima
INTRODUCTION
The Ryukyu Archipelago is a chain of conti-
nental islands extending in subtropical East
Asia between Kyushu of Japan and Taiwan.
Within this archipelago, the region often referred
to as the central Ryukyus (i.e., the Okinawa
and Amami Island Groups, and a few south-
ern Tokara islands: Fig. 1) is noted for the
occurrences in particularly high frequencies
of endemic taxa and relict lineages of non-
volant terrestrial vertebrates (Ota, 1998, 2000;
Motokawa, 2000). Such faunal and phylogeo-
graphical characteristics of the central Ryukyus
are generally considered as reflecting longer
isolation of this region as an island or a group
of islands than the other island groups of the
Ryukyu Archipelago (i.e., since the Late Miocene
or Early Pliocene vs. the Middle or Late Pleis-
tocene: Hikida and Ota, 1997; Ota, 1998; Iryu
et al., 2006: but see, Ujiié [1990] and Kimura
[1996, 2002] for different views).
Within the central Ryukyus, the extant endemic
taxa also show several patterns of geographic
distributions with the primary faunal boundary
being located between Tokunoshima Island and
more southwestern islands of the Amami Group
(Ota, 1998, 2000; Motokawa, 2000), and Ota
(1998) suspected that this boundary played a
major role in relatively recent vicariance events
* Corresponding author. Tel: +81–98–895–8937;
Fax: +81–98–895–8966;
E-mail address: ota@sci.u-ryukyu.ac.jp
2Current Herpetol. 26(1) 2007
in a number of terrestrial lineages. To exam-
ine the validity of this view, it should be
beneficial to investigate not only the extant
taxa, but also fossil taxa, as was pointed out
in more general biogeographical context by
several authors (Ota, 2002; Lieberman, 2003;
Jackson and Erwin, 2006). This is particularly
true when considering the Late Pleistocene
mass extinction of terrestrial vertebrates in the
Ryukyus, a view that has been substantiated by
discoveries of a number of extinct Pleistocene
species from several Okinawa Group and
southern Ryukyu islands (Ota, 2003).
The uppermost Pleistocene fissure-fillings
and cave deposits of the central and southern
Ryukyus contain numerous terrestrial vertebrate
fossils (e.g., see Oshiro, 2002; Otsuka, 2002;
Takahashi et al., 2003), offering a good oppor-
tunity of testing and revising the historical
biogeographical hypotheses of the Ryukyus
that were mostly or exclusively organized on
the basis of neontological data. To date, how-
ever, such an advantage has not yet been
effectively taken due to insufficiency of paleon-
tological explorations on many islands including
Tokunoshima Island, and of appropriate com-
parative studies that are essential to correct
identifications of fossils collected (Ota, 2003).
Recently, we investigated a fossiliferous
fissure-filling deposit on Tokunoshima Island
and found 15 pieces of fossils representing an
extinct geoemydid turtle. This turtle is ver y
similar to Geoemyda japonica, an extant spe-
cies currently endemic to Okinawajima Island
and two adjacent islands (Tokashikijima Island
and Kumejima Island) of the Okinawa Group
(Yasukawa et al., 1992). However, the fossil
Tokunoshima turtle differs from G. japonica
in shape and relative position of the entoplas-
tron. We thus describe this turtle as a new
species and discuss biogeographical implica-
tions of this finding.
MATERIALS AND METHODS
The geoemydid fossils were collected in
FIG. 1. Map showing locations of the central Ryukyus and Tokunoshima Island.
TAKAHASHI ET AL.—NEW PLEISTOCENE GEOEMYDA 3
March 2005 from a fissure-filling deposit
developed in the Pleistocene Tokunoshima
Formation of the Ryukyu Group (Yamada et
al., 2003) on a sea cliff at Kojima, Isen Town,
Tokunoshima Island, along with fossils of
other vertebrates (Cervus astylodon, Man-
ouria oyamai, Pentalagus furnessi, etc.). The
collection site was approximately 5 m south of,
and stratigraphically continuous to the fossil
site reported by Otsuka et al. (1980). Although
the present fossiliferous deposit has not yet
been dated, occurrences of C. astylodon and
M. oyamai (Otsuka, 1990; Takahashi et al.,
2003) strongly suggest that it was formed
during the Late Pleistocene.
All Tokunoshima geoemydid specimens col-
lected by us were deposited in The University
Museum, University of the Ryukyus (Fujukan:
RUMF). Morphological comparisons with cur-
rently recognized species were made using skel-
etal specimens in the private collections of the
senior author (AT) and of Ren Hirayama (RH),
and the collections of The Kumejima Island
Firefly Museum (KFM) and of The Kyoto
University Museum (KUZ). Relevant informa-
tion was also taken from Yasukawa et al.
(2001). Classification largely followed Yasukawa
et al. (2001), but with slight modifications on
the basis of findings from recent molecular
studies (Parham et al., 2004; Diesmos et al.,
2005; etc.). Shell terminology followed Zangerl
(1969). The shape of neural bones was expressed
by the neural formula, which corresponded to
the neural configuration in Pritchard (1988).
Measurements were taken to the nearest 0.1 mm
using digital slide calipers.
SYSTEMATICS
Order CHELONII Brongniart, 1800
(Latreille, 1800)
Suborder Cryptodira Cope, 1868
Superfamily Testudinoidea Batsch, 1788
Family Geoemydidae Theobald, 1868
Genus Geoemyda Gray, 1834
Geoemyda amamiensis sp. nov.
(Figs. 2, 3, 4A, and 5A)
Holotype
RUMF-GF-5011, an incomplete anterior half
of the plastron (38.0 mm long and 45.4 mm
wide) consisting of complete epiplastra, com-
plete entoplastron, and incomplete hyoplastra.
Paratype
RUMF-GF-5000, an incomplete carapace
(59.8 mm long and 40.6 mm wide) consisting
of the second to eight neurals, first suprapygal,
and medial parts of the right second to eight
and the left third to eight costal plates.
Referred specimens
RUMF-GF-5001, neural (8.6 mm long and
13.3 mm wide); RUMF-GF-5002, costal
(26.9 mm long and 11.2 mm wide); RUMF-
GF-5003 and 5004, right first peripherals
(RUMF-GF-5003: 12.3 mm long and 14.3 mm
wide; RUMF-GF-5004: 11.2 mm long and
12.2 mm wide); RUMF-GF-5005, left second
peripheral (16.3 mm long and 13.6 mm wide);
RUMF-GF-5006, left fourth or sixth peripheral
(15.2 mm long and 11.5 mm wide); RUMF-GF-
5007, left sixth peripheral (15.2 mm long and
13.3 mm wide); RUMF-GF-5008–5010, three
isolated peripherals of the posterolateral cara-
pace (RUMF-GF-5008: 12.8 mm long and
12.3 mm wide; RUMF-GF-5009: 11.5 mm
long and 12.1 mm wide; RUMF-GF-5010:
11.0 mm long and 11.8 mm wide); RUMF-GF-
5012 and 5013, two right epiplastra (RUMF-
GF-5012: 20.7 mm long and 22.7 mm wide;
RUMF-GF-5013: 18.2 mm long and 17.8 mm
wide); RUMF-GF-5014, proximal portion of
the right humeral shaft (22.0 mm long and
2.6 mm wide).
Locality
Sea cliff at Kojima, Isen Town, Tokunoshima
Island, the Amami Group (Kagoshima Prefec-
ture), the Ryukyu Archipelago, Japan.
Horizon
Fissure-filling deposit developed in the Pleis-
tocene Tokunoshima Formation of the Ryukyu
Group.
4Current Herpetol. 26(1) 2007
Diagnosis
This is a medium sized geoemydid species
and most closely resembles Geoemyda japon-
ica by a combination of following character
states: neurals reaching the first suprapygal;
posterior neurals with short sides posterolater-
ally; carapace tricarinate, with median keel
stronger and wider than lateral keels; forelobe
of plastron emarginate anteriorly; strong serra-
tions in anterolateral and posterolateral margins
of carapace; entoplastron intersected by gulo-
humeral; small concavity at anterior corner of
posterior vertebrals; and laterally rounded
forelobe of plastron. However, this new species
differs from G. japonica in the shape of the
entoplastron, which is pyriform with a short
FIG. 2. Carapacial bones of Geoemyda amamiensis. A–B, medial part of the carapace (paratype:
RUMF-GF-5000) in dorsal (A) and ventral (B) views. C–D, a neural (RUMF-GF-5001) in dorsal (C) and
ventral (D) views. E–F, a costal (RUMF-GF-5002) in dorsal (E) and ventral (F) views. G–J, two right first
peripherals (RUMF-GF-5003 [left] and RUMF-GF-5004 [right]) in dorsal (G and H) and ventral (I and J)
views. K–L, left second peripheral (RUMF-GF-5005) in dorsal (K) and ventral (L) views. M–N, left fourth
or fifth peripheral (RUMF-GF-5006) in dorsal (M) and ventral (N) views. O–P, left sixth peripheral (RUMF-
GF-5007) in dorsal (O) and ventral (P) views. Q–V, three posterolateral peripherals (RUMF-GF-5008 [left],
RUMF-GF-5009 [middle], and RUMF-GF-5010 [right]) in dorsal (Q, S and U) and ventral (R, T and V)
views. Scale bars equal 10 mm.
TAKAHASHI ET AL.—NEW PLEISTOCENE GEOEMYDA 5
FIG. 3. Plastral and limb bones of Geoemyda amamiensis. A–B, anterior half of the plastron (holo-
type: RUMF-GF-5011) in ventral (A) and dorsal (B) views. C–F, nearly complete (RUMF-GF-5012) and
incomplete (RUMF-GF-5013) right epiplastra in ventral (C and E) and dorsal (D and F) views. Right
humerus exclusive of the distal end (RUMF-GF-5014) in dorsal (G), ventral (H), and medial (I) views. Scale
bars equal 10 mm.
FIG. 4. Carapaces in dorsal views of Geoemyda amamiensis and the two extant congeneric species. A,
G. amamiensis (RUMF-GF-5000, paratype). B, G. japonica (KUZR R62453). C, G. spengleri (KUZ
R62351). Abbreviations are: N, neural; C, costal; SP, suprapygal; V, vertebral scute; PL, pleural scute. Scale
bars equal 10 mm.
6Current Herpetol. 26(1) 2007
anterior projection in dorsal view.
Etymology
Specific epithet, amamiensis, refers to the
Amami Group of the Ryukyu Archipelago, to
which its type locality (Tokunoshima Island)
belongs.
Description
Carapace.—The paratype (RUMF-GF-5000)
exhibits normal thickness (ca. 1.5–3.0 mm),
and is considerably depressed dorsally with
prominent discontinuous obtuse medial keel. The
growth annuli do not occur on any element.
The carapace is overlapped by the second to
fifth vertebral scutes, and the right second to
fourth and the left third and fourth pleural
scutes. The neurals are variable in shape. The
second one preserves its original right postero-
lateral fringe including a short side contacting
the third costal. The third neural, lacking its
left lateral portion, has short sides facing
posteriorly to the fourth costals. The fourth
neural is tetragonal and articulates exclusively
with the fourth costals. The fifth neural,
octagonal in shape, has four short facets
touching the fourth costals anterolaterally and
the sixth costals posterolaterally. The sixth
neural is hexagonal with short sutural connec-
tions to the seventh costals posterolaterally.
The seventh neural is rectangular, and the
eighth hexagonal with short anterolateral
sides articulating with the seventh costals. These
shapes and arrangement are expressed by the
neural formula as: ?-6P-6P-4-8-6P-4-6A (Fig.
4). Ventrally, the fourth to ninth dorsal verte-
brae are preserved along with the fourth to
ninth costal ribs on the right side and the fifth
to ninth costal ribs on the left side. The first
suprapygal presents. The third and fourth
vertebral scutes are hexagonal and wider than
long, with the latter being slightly wider than
the former. Each of the third and fourth
vertebrals is notched posteriorly with a very
small and short projection medially.
Of the referred specimens, the isolated
incomplete left costal (RUMF-GF-5002: 0.9–
1.7 mm thick) has a posteriorly notched sul-
cus, indicating that this specimen represents
the second or fourth or sixth costal. It also has
a longitudinal obtuse ridge, which obviously
corresponds to the left lateral keel. The periph-
erals exclusive of bridge components (RUMF-
GF-5003–5005 and 5008–5010) show that the
margin of the carapace is strongly serrated
anterolaterally and posterolaterally. The bridge
components (RUMF-GF-5006 and 5007) pre-
serve the sutural connection to the plastron,
and one of them (RUMF-GF-5007: Fig. 1P)
has a socket in distal fringe for the costal
process as seen in other geoemydids (Auffen-
berg, 1974).
Plastron.—The holotype (RUMF-GF-5011)
consists of the epiplastra without the right
posterolateral portion, the entoplastron, and
the anterior portion of the hyoplastra. The
high degree of ossification and shallow large
ventral concavity in its posteromedial portion
FIG. 5. Plastra in dorsal views of Geoemyda amamiensis and the two extant congeneric species. A, G.
amamiensis (RUMF-GF-5011, holotype). B, G. japon ica (KUZR R62453). C, G. spengleri (KUZ R62351).
Abbreviations are: EPI, epiplastron; ENT, entoplastron; HYO, hyoplastron. Scale bars equal 10 mm.
TAKAHASHI ET AL.—NEW PLEISTOCENE GEOEMYDA 7
indicate that this individual is an adult male
(Yasukawa et al., 1992). Of the remaining
plastral materials (two right epiplastra), one
(RUMF-GF-5012) is nearly completely pre-
served, whereas the other (RUMF-GF-5013)
lacks its posteromedial portion. These speci-
mens indicate that the anterior portion of
the plastron of this species is weakly rounded
laterally, moderately upturned, and emargin-
ate anteriorly with a distinct gulohumeral
notch. Dorsally, the forelobe lip, overlaid by
the gular, humeral, and possibly also by the
abdominal, is thin in the humeral region and
thick in the gular region.
The epiplastron in holotype (RUMF-GF-
5011; Fig. 3A and B) is almost as long as wide
in both ventral (21.7 mm long and 21.1 mm
wide on the left side) and dorsal views
(20.6 mm long and 21.1 mm wide on the left
side). The median lengths of the epiplastron
in dorsal and ventral views, 9.6 mm and
8.3 mm, respectively, are much shorter than
those of the entoplastron (15.5 mm and 18.2 mm,
respectively). Ventral surface of the gular scute is
wide and trapezoidal, overlapping with the
entoplastron posteriorly.
In ventral view, the entoplastron is relatively
large, pyriform (18.6 mm wide), and is inter-
sected by the gulohumeral and humeropec-
toral sulci. Dorsal surface of the entoplastron
is relatively small (14.2 mm wide) and subcir-
cular in shape with distinct convexity to the
posteromedial portion of the epiplastra.
The hyoplastra in holotype preserve the
anteromedial portion on the right side and
anterior half with the anterobasal portion of
the axillary buttress on the left side. The
baseline of the posterior forelobe on the hyo-
plastra is free from the entoplastron.
Humerus.—Our specimen (RUMF-GF-5014;
Fig. 3G–I) lacks the distal portion, central part
of the caput humeral, and distal end of the
medial process. The shaft is slender and arcu-
ate dorsally with the ectepicondylar groove in
its mediodistal portion. The caput humeral is
in a nearly right angle to the humeral shaft
laterally.
COMPARISONS
Geoemyda amamiensis shares the following
combination of character states only with spe-
cies of the genera Geoemyda, Vijayachelys,
Leucocephalon and Melanochelys, and Heose-
mys depressa within the family Geoemydidae:
the posterior neurals contacting the suprapy-
TABLE 1. Comparisons of Geoemyda amamiensis with other geoemydid species. Data include those
taken from Smith (1931), Wermuth and Mertens (1961), Moll et al. (1986), Ernst and Barbour (1989),
Yasukawa et al. (1992, 2001), Iverson and McCord (1997), McCord et al. (1995), Ernst et al. (2000), and
Praschag et al. (2006).
Serration in
anterolateral
margin of
carapace
Strong
serration in
posterolateral
margine of
carapace
Distinct and
wide medial
keel on
carapace
Small
concavity at
anterior corner
of posterior
vertebrals
Plastron
rounded
anterolater-
ally
Entoplastron
with short
anterodorsal
projection
Geoemyda amamiensis +++ + ++
Geoemyda japonica +++ + +−
Geoemyda spengleri +++ − −+
Vijayachelys silvatica −−+ − +−
Heosemys depressa −−+ − +−
Leucocephalon
yuwonoi ++− − − +
Melanochelys trijuga −−− − + −
Melanochelys tricarinata −−− − +?
8Current Herpetol. 26(1) 2007
gal series (derived state of character 27 in
Claude and Tong [2004]=character 26 in
Yasukawa et al. [2001]); the anterior neurals
with short posterolateral sides (derived state of
character G in Hirayama [1985]=G16 in
Gaffney and Meylan [1988]); carapace tricari-
nate (derived state of character F of Hirayama
[1985]=character 24 of Yasukawa et al.
[2001]=character 32 of Claude and Tong
[2004]: also see McCord et al. [1995], Iverson
and McCord [1997], and Praschag et al.
[2006]); peripherals in the bridge region with
sutural connection to the plastron (primitive
state of character 21 of Yasukawa et al. [2001]:
also see McCord et al. [1995], Iverson and
McCord [1997], and Praschag et al. [2006]).
The two extant Geoemyda species (G. japon-
ica and G. spengleri) and Leucocephalon
share with G. amamiensis distinct serrations
in the anterolateral and posterolateral margins
of the carapace (Ernst and Barbour, 1989;
Yasukawa et al., 2001). Moreover, the medial
keel in G. amamiensis is wide and developed
as in G. japonica, G. spen gleri, H. depressa,
and V. si l v a t i c a (Yasukawa et al., 1992; Moll
et al., 1986; Ernst and Barbour, 1989; Pra-
schag et al. 2006). In contrast, the keel is
undeveloped in Melanochelys and developed
but less prominent in Leucocephalon (Smith,
1931; Ernst and Barbour, 1989; Takahashi,
unpublished observations). Thus, G.
amamiensis most resembles G. japonica and
G. spengleri, and this justifies its assignment
to the genus Geoemyda. (Table 1).
In G. amamiensis, the entoplastron is inter-
sected ventrally by the gulohumeral. Yasukawa
et al. (2001) considered that the entoplastron
is isolated from the gulohumeral in the genus
Geoemyda (including Vijayach ely s then).
Recently, however, Hirayama et al. (2007)
demonstrated that the state of this character is
actually variable in G. japonica. Indeed, 13
out of 47 G. japonica examined by us had the
entoplastron intersected by the gulohumeral.
The entoplastron seems to be consistently iso-
lated from the gulohumeral in Vijayachelys
and G. spengleri (Yasukawa et al., 2001),
although stability of this character state in
these taxa also needs verification on the basis
of larger samples.
Geoemyda amamiensis and G. japonica
exclusively share two character states: pres-
ence of a very small notch at the anterior
corner in each of the fourth and fifth verte-
brals (Fig. 4); and laterally rounded forelobe
(Fig. 5). In G. spengleri, there are no such
small notches in vertebrals at al, and the
forelobe is trapezoidal in outline. On the
other hand, G. amamiensis shares only with
G. spengleri the shape of the entoplastron
which is characterized by the presence of a
short anterior projection: G. japonica lacks
such projection in the entoplastron (Fig. 5A).
The results of these comparisons suggest that
G. amamiensis is a distinct species, most
closely allied to G. japonica.
DISCUSSION
The genus Geoemyda had once accommo-
dated various extant and extinct lineages as a
“catchall taxon” (Wermuth and Mertens, 1961),
but recent taxonomic and phylogenetic studies
established the consensus that this genus
actually consists of only two extant species,
G. spengleri from Vietnam and southern
China, and G. ja ponica from three islands
(Okinawajima Island, Kumejima Island, and
Tokashikijima Island) of the Okinawa Group,
the Ryukyu Archipelago (Ernst and Barbour,
1989; Iverson, 1992; Yasukawa et al., 1992,
2001; Praschag et al., 2006). Thus, G. amamiensis
described here is the first fossil taxon of the
genus in strict sense.
Tokunoshima Island consists of the pre-
Quaternary basement rocks and the Middle
Pleistocene coral reef complex deposits (so-
called the Ryukyu Group comprising the
Tokunoshima Formation and the Metegu
Formation in ascending order: Yamada et al.,
2003). These geographic and stratigraphic data
indicate that Tokunoshima Island had already
been exposed above sea level when partial
deposition of the Tokunoshima Formation
initiated in the Middle Pleistocene (transgres-
sion phase). Presently Tokunoshima Island is
TAKAHASHI ET AL.—NEW PLEISTOCENE GEOEMYDA 9
separated from the Okinawa Group where G.
japonica occurs by deep sea (ca 250 m at
maximum depth), and this indicates that they
have been isolated for a fairly long period,
having no land-bridge connections even during
the Late Pleistocene Glaciations when the
sea level dropped by approximately 120 m at
maximum (e.g., Fairbanks, 1989; Nakada et
al., 1991; Sagawa et al., 2001). It is unlikely
that the geoemydid turtles have strong saltwa-
ter tolerance for oversea dispersals, because
their natural occurrences on oceanic islands
have never been reported (see Iverson [1992]
for review of their geographic distributions).
These suggest that G. amamiensis and G.
japonica split from each other through vicari-
ance, which started before the Middle Pleis-
tocene.
There are two small and low islands (Yor-
onto Island and Okinoerabujima Island), both
belonging to the Amami Group, between
Tokunoshima Island and the Okinawa Group
(Fig. 1). These islands mostly consist of the
Ryukyu Group, which is comparable to that on
Tokunoshima Island in the date of formation
(Kizaki, 1985; Yamada et al., 2003). It is there-
fore likely that populations of G. amamiensis
or G. j aponi ca or their allies also occurred on
the two islands. Current absence of Geoemyda
populations on Yoronto and Okinoerabujima
seems to reflect their extinction, presumably
through submergence of the whole or a large
part of each island during the Middle Pleis-
tocene transgression (Kimura, 2002). This pre-
diction may possibly be verified through an
effort to find turtle fossils from the middle
Pleistocene or earlier deposits on these two
islands.
ACKNOWLEDGMENTS
We thank R. Hirayama (Waseda University)
and Y. Yasukawa (University of the Ryukyus)
for allowing us to examine skeletal specimens
of geoemydid turtles in their care, and for a
number of helpful comments on the present
subject. We are also much grateful to G.
Masunaga, H. Sato, M. Kishaba, T. Sasaki
and Y. Yamamoto (University of the Ryukyus),
F. Sato (Kume Island Firefly Museum), K.
Maja (Ginowan City, Okinawa), T. Otani
(Okinawa City, Okinawa), Y. Kadota (Kyoto
University), and E. Setiyabudi (Kagoshima
University) for provision of specimens includ-
ing road-kills of Geoemyda japonica and
other geoemydid species for comparisons.
Special thanks are due N. Gi (Isen Town,
Kagoshima), who kindly permitted us to
conduct fieldworks and T. Ikeda (Kagoshima
University), who helped us with excavation
of fossils. This study was supported by a
grant for the 21st Century COE program “The
Comprehensive Analyses on Biodiversity in
Coral Reef and Island Ecosystems in Asian
and Pacific Regions” at University of the
Ryukyus from the Ministry of Education,
Culture, Sports, Science and Technology, Japan
(Monbukagakusho).
LITERATURE CITED
AUFFENBERG, W. 1974. Checklist of fossil land
tortoise (Testudinidae). Bull. Florida State Mus.,
Biol. Sci. 20: 47–110.
CLAUDE, J. AND H. TONG. 2004. Early Eocene
testudinoid turtles from Saint-Papoul, France,
with comments on the early evolution of modern
Testudinoidea. Oryctos 5: 3–45.
DIESMOS, A. C., J. F. PARHAM, B. L. STUAR T, AND
R. M. BROUN. 2005. The phylogenetic position
of the recently rediscovered Philippine forest
turtle (Bataguridae: Heosemys leytensis). Proc.
Calif. Acad. Sci. 56: 31–41.
ERNST, C. H. AND R. W. BARBOUR. 1989. Turtles
of the World. Smithsonian Institution Press,
Washington, D. C.
ERNST, C. H., R. G. M. ALTENBURG, AND R. W.
BARBOUR. 2000. Turtles of the World. World
Biodiversity Database, CD-ROM Series, Windows
Version 1.2. Biodiversity Center of ETI, Amster-
dam.
FAIRBANKS, R. G. 1989. A 17,000-year glacio-
eustatic sea level record: influence of glacial
melting rates on the Younger Dr yas event and
deep-ocean circulations. Nature 342: 637–642.
GAFFNEY, E. S. AND P. A . M EYLAN. 1988. A
10 Current Herpetol. 26(1) 2007
phylogeny of turtles. p. 157–219. In: M. J.
Benton (ed.), The Phylogeny and Classification
of the Tetrapods, Volume 1: Amphibians, Rep-
tiles, Birds. Systematics Association Special Vol-
ume, 35A. Oxford University Press, New York.
HIKIDA, T. AND H. OTA. 1997. Biogeography of
reptiles in the subtropical East Asian islands. p.
11–28. In: K. Y. Lue and T.-H. Chen (eds.),
Proceedings of the symposium on the phylogeny,
biogeography and conservation of fauna and
flora of East Asian region, National Science
Council, Taipei.
HIRAYAMA, R. 1985. Cladistic analysis of Batagu-
rine turtles (Batagurinae: Emydidae: Testudi-
noidea); A preliminary result. Stvd. Geol.
Salmanticensia, vol. esp. 1 (Stvdia Palaeochelo-
niologica I): 141–157.
HIRAYAMA, R., N. KANEKO, AND H. OKAZAKI.
2007. Ocadia nipponica, a new species of aquatic
turtle (Testudines: Testudinoidea: Geoemydidae)
from the Middle Pleistocene of Chiba Prefecture,
central Japan. Paleontol. Res. 11: 1–19.
IRYU, Y., H. MATSU DA, H. MACHIYAMA, W. E.
PILLER, T. M. QUINN, AND M. MUTTI. 2006.
Introductory perspective on the COREF Project.
Island Arc 15: 393–406.
IVERSON, J. B. 1992. A Revised Checklist with
Distribution Maps of the Turtles of the World.
Privately Printed, Richmond, Indiana.
IVERSON, J. B. AND W. P. McCORD. 1997.
Redescription of the Arakan forest turtle Geoe-
myda depressa Anderson 1875 (Testudines:
Bataguridae). Chelonian Conserv. Biol. 2: 384–
389.
JACKSON, J. B. C. AND D. H. ERWIN. 2006. What
can we learn about ecology and evolution from
the fossil record? Trends Ecol. Evol. 21: 321–
328.
KIMURA, M. 1996. Quaternary paleogeography of
the Ryukyu Arc. J. Geogr. 105: 259–285. (in
Japanese, with English abstract)
KIMURA, M. 2002. Establishment and paleogeog-
raphy of the Ryukyu Arc. p. 19–54. In: M.
Kimura (ed.), The Formation of the Ryukyu Arc
and Migration of Biota to the Arc. Okinawa
Times, Naha. (in Japanese)
KIZAKI, K. (ed.) 1985. Geology of the Ryukyu Arc.
Okinawa Times, Naha. (in Japanese)
LIEBERMAN, B. S. 2003. Paleogeography: The
relevance of fossils to biogeography. Ann. Rev.
Ecol. Evol. Syst. 34: 51–69.
McCORD, W. P., J. B. IVERSON, AND BOEADI. 1995.
A new batagurid turtle from northern Sulawesi,
Indonesia. Chelonian Conserv. Biol. 1: 311–316.
MOLL, E. O., B. GROOMBRIDGE, AND J. VIJAYA.
1986. Redescription of the cane turtle with notes
on its natural history and classification. J.
Bombay Nat. Hist. Soc., Suppl. 83: 112–126.
MOTOKAWA, M. 2000. Biogeography of living
mammals in the Ryukyu Islands. Tropics 10: 63–
71.
NAKADA, M., N. YONEKURA, AND K. LAMBECK.
1991. Late Pleistocene and Holocene sea-level
changes in Japan: Implications for tectonic
histories and mantle rheology. Palaeogeogr.
Palaeoclimatol. Palaeoecol. 85: 107–122.
OSHIRO, I. 2002. Quaternary geohistor y of the
Ryukyu Archipelago and the dispersal of ani-
mals. p. 129–140. In: M. Kimura (ed.), Establish-
ment of the Ryukyu Arc and Colonization of
Biota. Okinawa Times, Naha. (in Japanese)
OTA, H. 1998. Geographic patterns of endemism
and speciation in amphibians and reptiles of
the Ryukyu Archipelago, Japan, with special
reference to their paleogeographic implications.
Res. Popul. Ecol. 40: 189–204.
OTA, H. 2000. The current geographic faunal
pattern of reptiles and amphibians of the Ryukyu
Archipelago and adjacent regions. Tropics 10:
51–62.
OTA, H. 2002. Advantages and disadvantages of
neontological approaches to the reconstruction
of paleogeography, with an example of herpeto-
fauna and the Ryukyu Archipelago. p. 175–186.
In: M. Kimura (ed.), The Formation of the
Ryukyu Arc and Migration of Biota to the Arc.
Okinawa Times, Naha. (in Japanese)
OTA, H. 2003. Toward a synthesis of paleontologi-
cal and neontological information on the terres-
trial vertebrates of the Ryukyu Archipelago. I.
Systematic biogeographic review. J. Fossil Res.
36: 43–59.
OTSUKA, H. 1990. Fossil deer from the Tokuno-
shima, Ryukyu Islands, southwest Japan. Mem.
Natn. Sci. Mus. 23: 185–195 plus 1 plate. (in
Japanese, with English summary)
TAKAHASHI ET AL.—NEW PLEISTOCENE GEOEMYDA 11
OTSUKA, H. 2002. Paleovertebrate fauna of the
Ryukyu Archipelago and its origin. p. 111–127.
In: M. Kimura (ed.), The Formation of the
Ryukyu Arc and Migration of Biota to the Arc.
Okinawa Times, Naha. (in Japanese)
OTSUKA, H., T. HORIGUCHI, AND H. NAKAGAWA.
1980. On the fossil deer discovered from
Tokunoshima, Ryukyu Islands. p. 49–54. In: K.
Kizaki (ed.), Geological Studies of the Ryukyu
Islands, vol., 5. University of the Ryukyus. (in
Japanese, with English abstract)
PARHAM, J. F., B. L STUART, R. BOUR, AND U.
FRITZ. 2004. Evolutionary distinctiveness of the
extinct Yunnan box turtle (Cuora yunnanensis)
revealed by DNA from an old museum specimen.
Proc. Royal Soc. London, B. 271, Biol. Lett.
Suppl. 6: S391–394.
PRASCHAG, P., C. SCHMIDT, G. FRIRTZSCH, A.
MÜLLER, R. GEMEL, AND U. FRITZ. 2006.
Geoemyda silvatica, an enigmatic turtle of the
Geoemydidae (Reptilia: Testudines), represents a
distinct genus. Org. Divers. Evol. 6: 151–162.
PRITCHARD, P. C. H. 1988. A survey of neural
bone variation among recent chelonian species,
with functional interpretations. Acta Zool.
Cracov. 31: 625–686.
SAGAWA, N., T. NAKAMORI, AND Y. I RYU. 2001.
Pleistocene reef development in the southwest
Ryukyu Islands, Japan. Palaeogeogr. Palaeocli-
matol. Palaeoecol. 175: 303–323.
SMITH, M. A. 1931. The Fauna of British India,
Including Ceylon and Burma. Reptilia and
Amphibia. Vol. 1—Loricata, Testudines. Taylor
and Francis, London.
TAKAHASHI, A., H. OTSUKA, AND R. HIRAYAMA.
2003. A new species of the genus Manouria
(Testudines: Testudinidae) from the Upper Pleis-
tocene of the Ryukyu Islands, Japan. Paleontol.
Res. 7: 195–217.
UJIIÉ, H. 1990. Geological History of the Ryukyu
Island Arc. p. 251–255. In: H., Ujiié (ed.), Nature
of Okinawa—Geomorphology and Geology.
Hirugisha, Naha. (in Japanese)
WERMUTH, H. AND R. MERTENS. 1961. Schild-
kröten, Krokodile, Brückenechsen. Gustav Ficher
Verlag, Jena.
YAMADA, T., K. FUJITA, AND Y. I RYU. 2003. The
Ryukyu Group (Pleistocene coral reef deposits)
on Toku-no-shima, Kagoshima Prefecture,
Japan. J. Geol. Soc. Japan 109: 495–517.
YASUKAWA, Y., H. OTA, AND T. H IKIDA. 1992.
Taxonomic re-evaluation of the two subspecies
of Geoemyda spengleri (Gmelin, 1789) (Rep-
tilia: Emydidae). Jpn. J. Herpetol. 14: 143–159.
YASUKAWA, Y., R. HIRAYAMA, AND T. H IKIDA.
2001. Phylogenetic relationships of geoemydine
turtles (Reptilia: Bataguridae). Cur. Herpetol.
20: 105–133.
ZANGERL, R. 1969. The turtle shell. p. 311–339. In:
C. Gans and T. Parsons (eds.), Biology of the
Reptilia, Vol. 1, Academic Press, New York.
APPENDIX
Specimens examined
Geoemyda japonica (47 specimens): KUZ
R62042–045, 62047–048, 62050–051, 62053,
62055, 62161–163, 62165–166, 62185–188,
62190, 62192–200, 62203–204, 62206–207,
62322–324, 62453; AT 15, 130, 153, 158, 161,
169; KFM 1–4. G. spengleri (four specimens):
KUZ R62351, 59219, 62455–456. Heosemys
depressa (one specimen): RH1047. Leucoceph-
alon yuwonoi (one specimen): KUZ R62454.
Melanochelys trijuga (one specimen): AT 87.
Accepted: 22 February 2007
