Botanical Studies (2007) 48: 97-116.
*
Corresponding author: E-mail: chtsou@gate.sinica.edu.tw.
INTRODUCTION
Taiwan is situated at the junction of the Ryukyu and
the Luzon Arcs, and is separated from Fujian Province,
China by the 150 km wide Taiwan Strait (Figure 1). When
the overall flora of this region became explored in the
early 20
th
century, it became clear that the flora of Taiwan
mainly originated from continental China (Wilson, 1920;
Masamune, 1934; Kanehira, 1936; Li,
1957; Hosokawa,
1958) and the elements from the Philippines were limited
and restricted to the very south (Merrill, 1926; Li and
Keng, 1950; Li, 1953); while the flora of the Ryukyus (up
to the Amami group) was considered as an extension of the
flora of Taiwan,
separated from the flora of Japan (Kyushu
Island and to the north) (Masamune, 1934; Sonohara et
al., 1952; Hosokawa, 1958). These earlier viewpoints on
the floristic affinities have been supported by later floristic
analyses (Liu and Teruya, 1980; Hsieh et al., 1994; Shen,
1997; Hsieh, 2002). A well accepted explanation based
on geological/geographical information is that the Taiwan
Strait is mostly only 60-80 m deep and served as a bridge
connecting continental China and Taiwan several times
during the Pleistocene glacial episodes, whereas the Bashi
Strait between Taiwan and Luzon Island is deeper than
2,000 m and these islands have never been connected
(Shen, 1996). Our comparison on the native genera of
the Ryukyus listed in the "Flora of the Ryukyus, South of
Amami Island" (Hatusima and Amano, 1994) also reveals
that 96% (650 out of 672) of its native genera are also
native to Taiwan.
Phylogeny and taxonomy of Eurya (Ternstroemiaceae)
from Taiwan, as inferred from ITS sequence data
Chi-Chih WU, Zhi-Fu HSU, and Chih-Hua TSOU*
Institute of Plant and Microbial Biology, Academia Sinica, Taipei, Taiwan, R.O.C.
(Received May, 19 2006; Accepted August 24, 2006)
ABSTRACT. Eurya is the largest genus in the Ternstroemiaceae. The 13 Eurya species in Taiwan constitute
an important element in various habitats from low to high elevated forests in the island. In an attempt to
understand the interspecies relationships and geographical history of these 13 species, we sequenced 123
samples representing 32 Eurya species from Taiwan, the Ryukyus, continental China, and Southeast Asia
at the nuclear ITS region. Sixteen species of five other genera of Ternstroemiaceae were also incorporated.
Molecular phylogenetic trees show that the Eurya species studied form a monophyletic group, organized into
five clades with geographical correlations. All the 17 Eurya species from Taiwan and the Ryukyus studied,
except E. loquaiana, together constitute two coherent, but distantly related clades: one clade stands at a basal
and isolated position, with four out of its five species endemic to these two regions; they are probably Tertiary
elements of northern China or their direct descendants. The other clade is much more advanced and has strong
affinity to the Eurya in southeast China, which suggests that the members were probably originated from the
Quaternary flora of southeast China. Ecological partitioning is thought to be the major mechanism for the
speciation of four Taiwan-endemic species from E. chinensis.
Keywords: Eurya; ITS; Ternstroemiaceae; Phylogeny; Taiwan; The Ryukyus.
TAXONOMY
Figure 1. Map of SE Asia and E Asia.
pg_0002
98
Botanical Studies, Vol. 48, 2007
These phytogeographical studies of Taiwan were large-
scaled analyses of the entire floras or ligneous floras of
these regions. In recent years, adopting molecular tools
in phytogeographical studies on the plants of Taiwan has
become prevalent; however, the majority were focused
on the inter-polulation relationships within a single
species, e.g., Cycas taitungensis (Huang et al., 2001),
Cyclobalanopsis glauca (Huang et al., 2002); Kandelia
candel (Chiang et al., 2001), Michelia formosana (Lin,
2001; Lu et al., 2002), Myrica japonica (Cheng et al.,
2000), Trochodendron aralioides (Wu et al., 2001), and
Tritanotrichun oldhamii (Wang et al., 2004). Congeneric
species have rarely been investigated, e.g., Gentiana
(Chen, 2001).
The goal of our study is to approach the
phylogeography of a genus from Taiwan by analyzing it
together with congeneric species occurring in the Ryukyus
and Fujian Province, China. The genus Eurya was chosen
for three reasons. First, we have been working on a
revision of Eurya in Taiwan and have acquired an essential
species concept of the genus in this region. Second, the
total 19 Eurya species occurring in the Ryukyus and
Taiwan exhibit complex distribution patterns. Five are
endemic to the Ryukyus (E. osimensis, E. ryukyuensis,
E. sakishimensis, E. yaeyamensis, and E. zigzag), eight
to Taiwan (E. crenatifolia, E. glaberrima, E. hayatai, E.
leptophylla, E. nanjenshanensis, E. rengechinensis, E.
septata, and E. strigillosa). Three occur in both Taiwan
and SE China, but not in Japan/the Ryukyus (E. chinensis,
E. loquaiana, E. nitida). One is present in Taiwan and
probably in the Philippines as well (E. gnaphalocarpa)
(Ling, 1998), one is present in the Ryukyus (and Japan)
and China, but not in Taiwan (E. japonica), and one occurs
in all three regions (E. emarginata). Thirdly, the ratio of
endemic Eurya species is high in Taiwan (8/13=62%) and
the Ryukyus (5/7=71%); thus, Eurya appears to be a good
subject for phytogeographical research.
Eurya, comprising c. 130 species (Ling, 1998),
was the largest of the nine genera in the subfamily
Ternstroemioideae of Theaceae; but in the recent
angiosperm phylogeny system, Ternstroemioideae is
removed from Theaceae and combined with Pentaphyllax
(Pentaphyllacaceae) to form Ternstroemiaceae (Stevens,
2006). The genus Eurya distributes from SE Asia, E Asia,
to the Pacifics, with continental China as the center of
diversity (Ling, 1998). The species of Eurya are often
common or even dominant in various habitats in Taiwan,
such as riverbanks, coastal forests, windy mountain slops,
or cold and humid forests. In this study, for the taxonomic
treatment of Eurya, we followed the Manual of Taiwan
Vascular Plants (Liu et al., 1998) for the species in
Taiwan, with the addition of one new species, E. septata
(Wu et al., 2003). For Eurya in the Ryukyus, besides the
six species recognized in the Flora of Okinawa and the
Southern Ryukyu Islands (Walker, 1976) and the Check
List Vascular Floras of Ryukyus Islands (Shimabaku,
1997), we also accepted E. ryukyuensis as an independent
species. For Eurya in continental China, the Flora of
China (Ling, 1998) was followed.
MATERIALS AND METHODS
Sampling: All the 13 Eurya species distributed in
Taiwan, five out of the seven occurring in the Ryukyus,
and 12 out of the 18 occurring in Fujian, China were
collected in the field. (The 12 Fujian species were
collected from Fujian, Guangdong, and/or Hong Kong).
Eurya species from other areas and five other genera
in Ternstroemioideae, including Adinandra, Cleyera,
and Euryodendron of tribe Freziereae and Anneslea
and Ternstroemia of tribe Ternstroemieae, were also
incorporated (Table 1, Appendix 1). Leaf materials were
immediately put in silica gel after detachment in the
field; while voucher specimens were deposited in the
herbaria listed in Table 1. Sequencing for chloroplast
DNA regions psbA-trnH and trnL-trnF was carried out
in 24 and 4 species, respectively (Appendix 2); because
their respective sequence variations were too low for
interspecies comparison, no more sequencing was
continued (details provided in the Results). Sequencing
for the nuclear ITS (internal transcribed spacer) DNA
region was carried out in 123 samples of 32 Eurya species
and 23 samples of 14 species from the other five genera
(Table 1). The rather extensive sampling was intended
to have a preliminary survey on the ITS intraspecific
variation of Eurya species and to ascertain the suitability
of inferring ITS sequences in the present phylogenetic
study. The design of the sampling was that
for each Eurya
species occurring in Taiwan and the Ryukyus at least
three samples should be sequenced. When two or more
con-specific samples exhibited identical sequences, no
more sequencing for that species was required and that
sequence type would be chosen to represent the species
and registered in the GenBank; otherwise, more samples
should be sequenced until the requirement was fulfilled.
As for the species from other regions and of other genera,
the amount of samples was not very many, thus most of
the samples were sequenced. The number of con-specific
samples in this study ranged from one to 27. For those
Eurya species with wide distribution ranges, multiple
con-specific samples were usually chosen from different
countries/islands.
DNA extraction, amplification, and sequencing.
Total DNA was extracted from silica gel-dried or
herbarium material following the protocols described in
Struwe et al. (1998).
ITS.
Amplification of the internal transcribed spacer
(ITS) region which is composed of ITS1, the 5.8 gene, and
ITS2 was done by using primers ITSleu1 (Urbatsch et al.,
2000) and ITS4 (White et al., 1990). The PCR program
was set as 94¢XC for 5 min followed by 35 cycles of 94¢XC
for 1 min, 56¢XC for 1 min, 72¢XC 1 min, and a single cycle
of 72¢XC for 7 min. Sequencing was done by using ITSleu1
and ITS4 primers.
trnL-trnF.
Amplification of the trnL-trnF which
is composed of trnL intron, trnL3¡¦ gene and trnL-trnF
intergenic spacer was done by using primers TabC and
pg_0003
WU et al. ¡X Phylogeny and taxonomy of
Eurya
from Taiwan
99
pg_0004
100
Botanical Studies, Vol. 48, 2007
pg_0005
WU et al. ¡X Phylogeny and taxonomy of
Eurya
from Taiwan
101
pg_0006
102
Botanical Studies, Vol. 48, 2007
TabF (Taberlet et al., 1991). The PCR program was set as
94¢XC for 5 min followed by 30 cycles of 94¢XC for 1 min,
49¢XC for 1 min, 72¢XC min for 1 min 30 sec, and a single
cycle of 72¢XC for 7 min. Sequencing was done by using
primer TabC and TabF, when necessary, TabD and TabE
primers were used in addition (Taberlet et al., 1991).
trnH-psbA.
Amplification of the trnH-psbA
intergenetic spacer was done by using primers trnH (Tate
and Simpson, 2003) and psbA (Sang et al., 1997). The
PCR program was set as 96¢XC for 5 min followed by 35
cycles of 96¢XC for 50 sec, 53¢XC for 1 min, 72¢XC for 30
sec, and a single cycle of 72¢XC for 5 min. Sequencing was
done by using trnH and psbA primers.
PCR products were cleaned with the QIAquick PCR
purification kit (Qiagen, Valencia, California, USA).
Sequencing was done on an ABI 377 automated DNA
sequencer.
Sequence alignment.
Sequences were assembled
from both directions and ambiguous sites were checked
against the electrophenograms. All the sequence data
were aligned with the GCG program (Wisconsin Package
Version 10.3) and then adjusted manually by using Se-Al
(Rambaut, 1996).
Phylogenetic relationship analysis
Cladistic analysis was performed with Maximum
Parsimony by using PAUP* 4.0b10 (Swofford, 2002)
and Bayesian analysis by using MrBayes version 3.0b4
(Huelsenbeck and Ronquist, 2001). In the parsimony
analysis all characters were unordered and weighted
equally (Fitch, 1971). The data matrix was analyzed by
employing an heuristic tree search with 1000 replicates
with stepwise to create the initial trees, and asis sequence
addition and tree-bisection-reconnection (TBR) branch
swapping were set. The MaxTrees was set to 10000.
Relative support was estimated with the bootstrap option
in PAUP* employing a heuristic search with 1000
replicates. In the Bayesian analysis, MrModel test version
3.6 (Posada and Crandall, 1998) was adopted to estimate
the parameters, the Markov chain Monte Carlo algorithm
was set as four simultaneous chains, and a tree was saved
every 5000 generations. Two million generations were
performed in each analysis. Trees from the burn-in period
were discarded, and a 50% majority rule consensus tree
was constructed from the remaining trees.
RESULTS
Chloroplast DNA sequences
The psbA-trnH sequences of the 20 Eurya species
examined were uniformly 416 bp long and 16 of them
showed exactly the same sequence (Table 2); as for the
other four species, E. acuminata and E. groffii had two
single substitution mutations each, and E. glaberrima
and E. distichophylla one substitution mutation (Table
2). The total six variations were singletons. Sequencing
of the trnL-trnF region was carried out in four distantly
distributed Eurya species (Table 2). Three of them showed
identical trnL-trnF sequences. As for the fourth species,
it had a deletion of 27 bp. Both psbA-trnH and trnL-trnF
regions seemed unsuitable for the phylogenetic analysis of
Eurya, this part of sequencing work was thus intermitted.
Nuclear ITS sequences
The length of ITS in the six genera of Ternstroemiaceae
ranged from 632 to 648 bases before alignment and was
673 after alignment. Within Eurya the length was fairly
uniform, with 637 bases long in 27 species, 638 bases in
three species, and 632 in one; this segment included the
ITS1 247-248 bases long, 5.8S 164 bases long, and ITS2
221-227.
ITS sequence variation
The number of con-specific samples of Eurya species
varied from one to
27. For the species with two or more
samples, limited intraspecific variation did exist in most
of them; nevertheless, the representative sequence of
each species, which was generated from at least two con-
specific samples, usually emerged within a small sample
size, and the other con-specific sequence types differed
from the representative at only one to three sites (Table 1).
Most of these intraspecific variations were independent,
single substitution mutations within the intron and a small
percentage were insertions or deletions. The exceptions
were found in E. crenatifolia and E. leoptophylla where
their ITS sequences were much variable. It is worthy
to mention that Eurya chinensis was much extensively
surveyed (20 samples) because its morphological
variations were great and many individuals with different
features were sampled (Table 1); but it turned out that ITS
sequence was highly conserved among the con-specific
samples of E. chinensis that seven out of the 10 samples
from Taiwan exhibited the same type (E. chinensis-T)
and all nine samples from Hong Kong and Guangdong,
China showed the same type (E. chinensis-C). These
two dominant sequence types, E. chinensis-T and E.
chinensis-C differ at only one site that the former has a
C whereas the latter has a T at base number 200 in the
aligned matrix (Table 2). It is important to note that a C
at base 200 characterizes all the samples of Taiwanese E.
chinensis, E. emarginata, and E. septata and the Ryukyu¡¦s
E. sakishimensis; whereas a T is found in the remaining
species. Further discussions on this point will be given
in Discussion. In the combined data matrix including the
representative sequences of the 32 Eurya species, the total
variation site is 62 bases in the aligned matrix (Table 2).
ITS distance within Eurya
The distance between any two of the 32 Eurya species
studied ranges from 0 to 16 bases, under the condition
that the representative sequences are compared (Tables
1, 2). Identical ITS sequences were found in two pairs of
species, E. chinensis-T and E. septata, and E. chinensis-C
and E. nanjenshanensis. These two pairs actually differ
pg_0007
WU et al. ¡X Phylogeny and taxonomy of
Eurya
from Taiwan
103
Table 2. ITS sequence variations among the 32 Eurya species, the numbers of the sites are based on the aligned matrix.
at a single site, site 200 in the matrix (Table 2), as just
mentioned. Minor ITS variation is commonly found
between species occurring in the same geographical
region. Among the 13 Eurya species from Taiwan, 10 of
them form two coherent groups. Group one including E.
glaberrima, E. gnaphalocarpa, E. rengechiensis, and E.
strigillosa is mainly distributed in
the Central Mountain
Range of Taiwan, from 650 m to 2,000 m elevation.
With E. glaberrima as the center, the other three differ
from it at only one or two sites, and the total variation in
this group is three sites (Table 2). Group B includes six
species, viz., E. chinensis, E. crenatifolia, E. emarginata,
E. leptophylla, E. nanjenshanensis, and E. septata, with
interspecies ITS distance from zero to two sites except that
E. leptophylla differs from E. chinensis-T and E. septata
at three sites (Table 2). Seven species from SE China form
a coherent group as well and with E. amplexifolia as the
center, E. hebeclados, E. loquaiana, E. macartneyi, E.
muricata, E. subintegra, and E. weissiae differ from the
center at only one or two sites (Table 2). Although minor
ITS variation is mostly found between species occurring
in the same geographical area; the reserve is not true.
Eurya loquaiana, E. glaberrima, and E. hayatai co-occur
in many mid-elevated forests in Taiwan, but their pairwise
distance is 12 (E.l. vs. E.g.), 10 (E.g. vs. E.h.), and 13
(E.l. vs. E.h) bases, respectively. Eurya yaeyamensis and
E. ryukyuensis can be found in the same locality in the
Iriomote island, Japan; but their ITS sequences differ at 10
sites (Table 2).
The greatest ITS distance, 16 variation sites, was found
in three species pairs, i.e., E. laotica vs. E. sakishimensis;
E. acuminata var. acuminata vs. E. ryukyuensis; and
E. acuminata var. acuminata vs. E. sakishimensis.
Coincidentally, these three pairs have the greatest
geographical distance that E. laotica and E. acuminata are
SE Asian whereas E. ryukyuensi and E. sakishimensis are
Ryukyu endemics. The next level, 15 variation sites, was
found in species pairs between these two SE Asian species
and many Taiwan-Ryukyus endemic species (Table 2).
Gene tree and species tree
In order to evaluate the level of paralogy of the ITS
intraspecific variation revealed in this study, all the
sequence types, 92 types generated from 146 samples of
the 46 species from six genera (Table 1) and three other
sequences from GenBank, were analyzed to produce
the gene tree (Figure 2). The most parsimony (MP) was
adopted for analysis with Anneslea and Ternstroemia
from the tribe Ternstroemieae as the outgroups. In the
gene tree, the con-specific sequence types are usually
grouped in a clade of their own, such as in cases of E.
yaeyamensis, E. distichophylla, E. muricata, E. loquaiana,
E. acuminata, E. japonica, E. hayatai, E. nitida, E .
ryukyuensis, E. sakishimensis, and E. emarginata. In many
other species they form polytomous branches in a larger
clade, such as those of E. rengenchiensis, E. strigillosa,
E. glaberrima, E. gnaphalocarpa, E. hebeclados, E.
macartneyi, E. nanjeshanensis, and E. septata. Only those
of E. crenatifolia, E. chinensis, and E. leptophylla spread
somewhat widely in the lowest clade and are occasionally
mixed with some other species to form a smaller clade
(Figure 2). On the basis of the gene tree (Figure 2),
the intraspecific sequence variation of most species
shows shallow paralogy and the selected representative
sequences do group with their con-species samples except
that sequences of E. crenatifolia, E. chinensis, an d E.
leptophylla are widely spread in the lowest clade. Thus, we
consider that the propriety of the representative sequence
pg_0008
104
Botanical Studies, Vol. 48, 2007
Figure 2. The gene tree based on 92 ITS sequence types produced from 49 species of six genera. This is the MajRule tree generated
from the parsimony analysis, with total characters=673, informative characters=148, tree length=347, CI=0.6945, RI=0.9132, and
10000 trees retained. Numbers showing on the branches are percentage of the presence of the clade among the 10000 trees (numerator)
and bootstrap values larger than 50 (%) (denominator). The bootstrap values were based on 800 replicates.
pg_0009
WU et al. ¡X Phylogeny and taxonomy of
Eurya
from Taiwan
105
is supported. The species trees (Figures 3, 4) were then
produced based on the 52 representative sequences which
representing 49 species (Table 1) by using the most
parsimony and Bayesian analyses. In general, the MajRule
gene tree and the two species trees are highly congruent
in their topologies (Figures 2, 3, 4). Their common and
important points revealed in these trees are:
1. Among the ingroups, Adinandra and Cleyera together
form a clade with 100% bootstrap support, this clade
is sister to Euryodendron and Eurya. The three species
of Cleyera are embedded within the seven Adinandra
species, further consideration on the unification/
separation of these two genera is necessary. The two
Adinandra species endemic to Taiwan, A. formosana
and A. lasiostyla, differing at 19 sites, are not the
closest to each other. The lowland A. formosana forms
a clade with the SE China distributed A. millettii,
with only one site in difference; the highland A.
lasiostyla forms a clade with the Ryukyu endemic A.
yaeyamensis, with six sites different, but sharing seven
substitution mutations of their own.
2. Euryodendron, a monotypic genus, is sister to and the
closest to the Eurya clade. The status of Euryodendron
as a monotypic genus is strongly supported in this
study.
3. Eurya, with 32 species studied, appears as a
monophyletic genus. The 32 species unanimously form
five clades and the allocation of these species in five
clades is exactly the same in all these MajRule trees
(Figures 2, 3, 4). The relationships of these five clades
are better revealed in MP trees than in Bayesian trees
since the five clades are arranged in three levels in the
former, but are parallel in the latter. These five clades
possess strong geographical constraints and are thus
defined:
(1)
Ryukyu-Taiwan (RT) endemic group ¡X this clade
is sister to the other four clades in MP MajRule
trees (Figures 3, 4). Five species are included, with
Eurya yaeyamensis endemic to the Ryukyus, and
E. reneichiensis, E. strigillosa, and E. glaberrima
endemic to Taiwan. Eurya gnaphalocarpa is
reported as also occurring in the Philippines and
Taiwan (Ling, 1998). Among these five species, the
former two are sister to the latter three.
(2)
SE Asia-SW China group ¡X four species are
included, E. acuminata, E. laotica, and E. groffii
with a distribution range in SE Asia and E.
quinquelocularis in SW China. It is interesting
that E. acuminata wallichiana shows intermediate
sequencing between E. groffii and E. acuminata
(Table 2).
(3)
S China grou p ¡X E. acuminatissima, E. disticha
and E. distichophylla form a small clade, they
mainly distributed in the southern provinces of
China.
(4)
SE China group ¡X eight species, viz., E .
amplexifolia, E. herbeclados, E. impressinervis,
E. loquaiana, E. marcartneyi, E. muricata, E.
subintegrifolia, and E. weissiae form a well
supported clade. All these species are confined to
SE China except that E. loquaiana also extends to
Taiwan.
(5)
Ryukyu-Taiwan (RT) dominant group ¡X twelve
species are included in this clade. The most basal
one, E. rubiginosa, is only present in Guangdong,
China. At the next level is E. nitida, occurring
in continental China and Taiwan. Among the
remaining 10 species, a clade consisting of E.
hayatai and E. japonica is sister to the other eight
species. Eurya japonica occurs in continental China
and Japan and E. hayatai is endemic to Taiwan.
The remaining eight species are endemic to either
Taiwan or the Ryukyus except for the widely
distributed E. chinensis. Since E. chinensis stands
as the center of these eight species, as the sequences
are concerned, these eight are collectively termed as
an E. chinensis branch. This E. chinensis branch is
successfully established in Taiwan and the Ryukyus.
The sequence variations among them are so low
that the resolution of this branch is very poor. It is
worthy to note that in this ladder-like arrangement
of this big clade, species distribution range shifts
stepwise from SE China (E. rubiginosa) to China
+ Taiwan (E. nitida), to E China-Taiwan-Japan
(E. chinensis, E. hayatai, E. japonica), and then
to Taiwan-Ryukyu endemism (E. crenatifolia, E.
leptophylla, E. nanjenshanensis, E. ryukyuensis, E.
sakishimensis, and E. septata).
In summary, the 32 Eurya species studied form five
well supported clades (Figures 2, 3, 4). Seventeen out of
the 18 Eurya species occurring in Taiwan and the Ryukyus
are confined to two clades, the R-T endemic and the R-T
dominant clades; these two clades are situated respectively
at the most basal and the most derived positions in the MP
MajRule trees. The remaining Taiwanese E. loquaiana is
grouped with seven SE China distributed species. The 12
Eurya species occurring in Fujian, on the other hand, are
spread in four out of the five clades.
DISCUSSION
The nuclear ITS gene (18S-5.8S-26S) has hundreds
to thousands copies arrayed as tandem repeats on the
chromosomes. Its phylogenetic inference has been
extensively employed by plant taxonomists (Alvarez
and Wendel, 2003), but more questions on the impact
of its sequence polymorphism on the phylogenetic
reliability have been raised (Alvarez and Wendel, 2003;
Andreasen and Baldwin, 2003; Bailey et al., 2003). In
order to ascertain the suitability of ITS being used in
the present study, a lot of efforts were made to collect
and sequence multiple samples for each Eurya species
in Taiwan and neighboring regions. We sequenced 3-27
pg_0010
106
Botanical Studies, Vol. 48, 2007
Figure 3. Species tree one based on 52 ITS sequence types representing the total 49 species. This is the Majrule tree generated from
the parsimony analysis, with total characters=673, informative characters=141, tree length=309, CI=0.6958, RI=0.9035, and 121 trees
retained. Bootstrap supports higher than 50% are shown beneath the branches. Numbers showing on the branches are percentage of the
presence of the clade among the 121 trees (numerator) and bootstrap values larger than 50 (%) (denominator). The bootstrap values
were based on 1000 replicates. The five clades of the Eurya species are defined, and the Arabic numerals 1, 2, and 3 are designated to
the distribution range covering Fujian, Taiwan, and the Ryukyus, respectively.
pg_0011
WU et al. ¡X Phylogeny and taxonomy of
Eurya
from Taiwan
107
Figure 4. Species tree two with the same dataset as species tree two. This is the MajRule tree generated from the Bayesian analysis.
The posterior probabilities higher than 50% are shown beneath the branches. Asterisks (*) indicate the presence of anther septation,
two species are unknown in this regard and marked with a ".", the remaining species have no anther septa.
pg_0012
108
Botanical Studies, Vol. 48, 2007
samples for every Eurya species in Taiwan and most
species in the Ryukyus; in general, for most of these
species a specific and dominant sequence type was found
after a small sampling. Identical ITS sequences are
commonly obtained from con-specific samples collected
from different countries/islands (e.g., E. chinensis, E.
emarginata, E. japonica, E. macartneyi, E. nitida, E.
ryukyuensis, and Cleyera japonica) or from localities
hundreds kilometers away (e.g., E. glaberrima, E.
loquaiana, E. septata, E. strigillosa, and Adinandra
formosana) (Table 1, Appendix 1). As for con-specific
samples collected from the same population, their ITS
sequences are even more homogeneous. The tendency of
retaining a specific ITS sequence type within a species
resulting from the concerted evolution has been reported
in many plant groups (Wendel et al., 1995; Sang et al.,
1995); we believe that concerted evolution is affecting
Eurya as well. In addition, in our integrated analysis,
all the 92 sequence types generated in this study were
incorporated to produce the gene tree (Figure 2), the
intraspecific sequence variations of most species represent
shallow paralogy which indicates that they would not
adversely affect species tree reconstruction (Figures 2, 3,
4). In conclusion, we believe that ITS sequence is reliable
for this phylogenetic study of Eurya. In the E. chinensis
lineage the interspecific sequence variations are usually
low or even none; whereas the intraspecific variations of
E. chinensis, E. crenatifolia, and E. leptophylla are greater
than the interspecific variations; thus the sequence types
of these three species may mix with a few other species
in several small clades. The much greater intraspecific
sequence variation of two Taiwan-endemic species, E.
crenatifolia and E. leptophylla, might be due to their
slower concerted evolution relative to mutation rates and
incomplete lineage sorting, etc. (Andreasen and Baldwin,
2003). It is noteworthy to point out that the dominant
sequence type of the species with multiple samples is
also the consensus sequence for each species. They were
chosen to be registered in the GenBank and used for
interspecies interpretations (Table 1).
Taxonomic implications
The genus Eurya has a rather large number of species,
c. 130; and criteria used for discriminating species are
poor and often microscopical. The present sequencing
work shows that the ITS sequences of the 32 Eurya
species differ at 62 sites (Table 2) and the greatest distance
between two species is 16 mutations, which is potentially
useful for taxonomic treatment of Eurya at theoretical
and practical levels. During the course of this study, a
few taxonomic problems were resolved by using the ITS
sequencing data:
1. Taxonomic treatments of E. chinensis in Hong
Kong. Eurya chinensis and E. nitida are the most
common Eurya species in Hong Kong and sympatric
in numerous sites. Usually they can be easily
distinguished by the thick pubescence and the smaller
leaves in E. chinensis and the lack of pubescence
and slightly larger leaves in E. nitida; however, in
many windward mountainous sites, abundant typical
(hairy) E. chinensis individuals are mixed with a small
percentage of glabrous individuals and the latter are
not readily distinguishable from E. chinensis or E.
nitida. Sometimes, there are hairy individuals, but with
much smaller (c. 1/2 of that of E. chinensis) and thicker
leaves, which are similar to the Taiwan-endemic E.
crenatifolia in gross morphology. Local taxonomists
use the name, E. chinensis var. glabra to the glabrous
individuals and E. chinensis var. chinensis to the small-
leaved ones (Lai et al., 2004). Owing to the lack of
any further studies, we sequenced eight individuals
of E. chinensis from Hong Kong including the three
phenotypes (typical, glabrous, and small-leaved).
Interestingly, their ITS sequences came out the same
as the E. chinensis-C type collected from Guangdong,
China (Table 1). The results strongly support the
taxonomic treatment of these phenotypes under E.
chinensis. In addition, this preliminary study suggests
that E. chinensis possesses great morphological
plasticity and may well sustain severe environments
where other Eurya species can not. The populations
surviving in the harsh conditions may express greater
morphological diversity.
2. D o E. acuminata and E. japonica occur in Taiwan.
In the most recent treatments of Eurya in Taiwan,
i.e., Flora of Taiwan (Hsieh et al., 1996), Manual of
Taiwan Vascular Plants (Liu et al., 1998), and Flora of
China (Ling, 1998), the recognition of E. acuminata
and E. japonica is controversial. The distribution of
Eurya acuminata in Taiwan is reported in the first and
the third treatments; and that of E. japonica in the
third. Nonetheless, our samples of E. acuminata var.
acuminata collected from Singapore and E. acuminata
var. wallichiana from N Thailand are well separated
from Taiwanese Eurya in both morphological and
ITS sequence aspects (Table 2). Plants in Taiwan
previously determined as E. acuminata were mostly of
E. loquaiana, probably due to their similar acuminate
leaf apices. For clarifying the issue of Eurya japonica,
samples were collected from Zhe-Jiang Province,
China and Okinawa Island, Japan and they showed the
same morphological characteristics and identical ITS
sequences (Table 1, Appendix 1); but in the sequence
aspect, E. japonica differs from those Eurya in Taiwan
at five sites (E. hayatai and E. nitida) or more. Among
the Taiwanese Eurya, the endemic E. hayatai is often
misidentified as E. japonica because they are similar
morphologically; with the major distinction on the leaf
lower surface where the venation is clear in E. japonica,
but nearly invisible in E. hayatai. These two species are
so far closely related as suggested by the species trees
(Figures 3, 4). In this study, ITS sequencing work helps
to delete the presence of E. acuminata and E. japonica
in the current flora of Taiwan.
3. Taxonomic status of Eurya ryukyuensis, E .
nanjenshanensis, a nd Cleyera morii. Eurya
pg_0013
WU et al. ¡X Phylogeny and taxonomy of
Eurya
from Taiwan
109
ryukyuensis, a Ryukyu-endemic taxon, was first
published as an independent species by Masamune in
1935, but was transferred to a variety of E. emarginata
by Hatushima in 1956 which was followed in all the
important floras of that area, such as Flora of Okinawa
and the Southern Ryukyu Islands (Walker, 1976) and
Check List of the Vascular Flora of the Ryukyu Islands
(Shimabuku, 1997). ITS sequencing data shows that
E. ryukyuensis is different from all the other Eurya
species studied. The ITS distance from E. ryukyuensis
to E. emarginata is four sites and the shortest
distance is found in E. chinensis (sequence type of E.
chinensis-C), with two different bases (Table 2). The
treatment of E. emarginata var. ryukyuensis is thus
not supported. Eurya nanjenshanensis was published
as E. nitida var. nanjenshanensis in Flora of Taiwan
(Hsieh et al., 1996), mainly due to the glabrous feature;
but was then recognized as a distinct species in the
Manual of Taiwan Vascular Plants (Liu et al., 1998).
The ITS sequence of E. nanjenshanensis differs from
that of E. nitida at four sites, but is the same as the E.
chinensis-C sequence type. Again, the treatment of E.
nitida var. nanjenshanensis is not supported. Cleyera
morii, a Taiwan endemic, was first published as a
variety of Eurya, Eurya ochnacea Szyzy. var. morii
Yama. by Yamamoto in 1927. It was then transferred
to Cleyera japonica Thunb. var. morii (Yama.) Masa.
by Masamune (1935) and then changed to C. morii
(Yama.) Masa. in 1939. Nonetheless, C. japonica var.
morii was preferred by Kobuski (1938) and adopted in
later important treatments, such as Flora of Taiwan, the
First Edition (Li, 1976) and the Second Edition (Hsieh
et al., 1996), the Manual of Taiwan Vascular Plants (Liu
et al., 1998), and Flora of China (Ling, 1998). Our three
samples of C. japonica collected from the Ryukyus
and Taiwan have identical ITS sequence and the two
samples of C. morii collected from two populations in
Taiwan have identical sequences as well. These two
species differ at five sites, which strongly supports an
independent species status of C. morii, in addition to
their morphological distinctions (Hsieh et al., 1996).
In summary, the present sequencing work supports the
independent species status of E. ryukyuensis and E.
nanjenshanensis. Also, Cleyera morii, differing from
C. japonica at five sites, also deserves the status of a
species.
4. Taxonomic value of anther septation in Eurya. Large
scaled taxonomic studies on Eurya have been poor; the
most recent is by Ling (1998) in the Flora of China.
Since his earlier revision (1966) on Eurya, Ling has
strongly stressed the importance of anther septa and
even classified the genus as two subgenera based on this
single criterion. This work, though the sampling covers
only one fourth of the total number of species, shows
that anther septa could have evolved from nonseptate
ancestors in different lineages independently (Figure 4).
We assume that the evolution of anther septa in Eurya
is possibly associated with pollination advantages, i.e.,
retaining pollen grains in the anthers in the up-side-
down floral orientation. Taxonomic applications of this
character in Eurya classification are better restricted to
species discrimination.
Phylogeography of
Eurya of Taiwan
Our study included all the 13 species of Eurya
occurring in Taiwan, five of the seven (71%) occurring in
the Ryukyus, and 12 of the 18 (67%) in Fujian Province,
China, and several others from SW China and SE Asia.
Increasing the number of species and the regions of
sampling will certainly give better resolution of the
interspecies relationships. Nonetheless, the relationships
of the species occurring in Taiwan and the Ryukyus that
they constitute two coherent but distantly related groups
plus one isolated species (E. loquaiana) is strongly
indicated. First, the intimacy among the five species in
the RT endemic clade and that among the eight in the RT
dominant clade are obvious by the high ITS sequence
similarity and the possession of several synapomorphic
sites within each clade, which is also reflected in the high
bootstrap value. Second, the distant relationship between
these two clades as shown in Table 2 and Figures 3 and 4
can not be refuted because members of these two groups
constantly have 10 or more different sites at the ITS
region. Thirdly, E. loquaiana is closely associated with a
few members of SE China clade with a strong support of
92% bootstrap value (Figure 4). Its difference from other
Eurya species in Taiwan ranges from 10 to 15 sites (Table
2) and its isolated position within the Eurya in Taiwan
is definite. Phylogenetically, Eurya species of Taiwan
and the Ryukyus form two coherent, but distantly related
groups, plus one isolated species, is distinct and unlikely
to be altered even when more samples are analyzed.
On the basis of the species relationships of Eurya,
which we here revealed, a few phylogeographical points
can be deduced:
1. Among the seven Eurya species occurring in the
Ryukyus, five are here studied. They are either
embedded in the RT endemic clade (e.g., E .
yaeyamensis) or in the RT dominant clade (e.g.,
E. emarginata, E. japonica, E. ryukyuensis, and E.
sakishimensis). The remaining two Eurya species
yet studied are rare and endemic E. osimensis and E.
zigzag. Morphologically, E. osimensis is very similar
to the Taiwan-endemic E. strigillosa in general features
and especially in possessing yellowish, long (c. 0.6-0.8
mm) hairs on terminal buds and branches; it has even
been treated as a variety of the latter by Masamune
(1955). Eurya zigzag is very similar to another
Ryukyu-endemic E. yaeyamensis which is embedded
in RT endemic clade grouped with four species from
Taiwan. Generally speaking, the majority of Eurya in
the Ryukyus show close affiliations with the Eurya
in Taiwan. But there are other facts to be considered.
First, the endemic rate of the Eurya in the Ryukyus is
very high (5/7); second, E. japonica distributes in E.
pg_0014
110
Botanical Studies, Vol. 48, 2007
China, Japan, and Korea, but not in Taiwan. Thirdly,
the only Eurya species shared by the Ryukyus and
Taiwan, i.e., E. emarginata, is probably of avian seed
dispersal because it is distributed in the costal regions
of many islands in SE China, Taiwan and the Ryukyus;
thus its coexistence in these two regions does not imply
any pre-existing land bridges between the two regions.
And fourthly, no Eurya species is currently distributed
to the Ryukyus and Taiwan only. In conclusion, our
preliminary analysis suggests that the rather close
linkage between the Eurya of the Ryukyus and Taiwan
was based on ancient interactions; the prolonged
isolation of the Ryukyus probably since the middle of
Pleistocene (Kimura, 2000) results in the very high
endemism of Eurya. No recent interactions through
land bridges between the Ryukyus and Taiwan can be
identified in the case of Eurya.
2. The phylogeography of the Eurya in Taiwan is a little
more complicate. The four species in RT-endemic
clade show comparatively close associations with
the Ryukyu members, which is probably caused by
ancient interactions. Some species (e.g., E. hayatai, E.
loquaiana, E. nitida) show affiliations with extant SE
China or E China members. And, the remaining six
species, including four endemic species, the SE China-
Taiwan distributed E. chinensis, an d E. emarginata,
form an intimate lineage, with E. chinensis as the
center (the E. chinensis lineage) (Figures 3, 4). These
nine species, as a whole, are connected with the current
continental progenitors. In this connection, the origin
of the four endemic species in the E. chinensis lineage
can be discussed. Eurya crenatifolia, E. leptophylla,
E. nanjenshanensis, and E. septata possess identical
ITS sequences to that of E. chinensis, either to the
representative of SE China (E. chinensis-C) or Taiwan
(E. chinensis-T). As we suggested in the Results, E.
chinensis has great ability in living in diverse habitats
and developing great morphological diversities. It is
possible that during the several Pleistocene glacial
periods when Taiwan was connected with Fujian
(Zhao, 1982; Shi et al., 1986; Hsieh and Shen, 1994;
Voris, 2000), E. chinensis as a member of SE China
flora could have migrated to Taiwan several times.
Speciation originated from E. chinensis might have
taken place in different ecological habitats and at
different geological times, and then gave rise to
these four Taiwan-endemic taxa. Eurya crenatifolia,
characterized by tiny and thick leaves, is adapted to
windy and humid environment, mainly in northern
Taiwan; E. leptophylla, characterized by small and
thin leaves, is restricted to higher elevations; E .
nanjenshanensis, with glabrous and slightly obovate
leaves, is found in the tropical southern lowlands;
and E. septata, differing from E. chinensis in having
septate anthers and slightly larger leaves, can well
survive in disturbed habitats (Wu et al., 2003). As for
the other three Taiwan-endemic species, E. glaberrima,
E. hayatai, an d E. strigillosa, since their closest
relatives are not yet evident, no interpretations on their
endemism can be provided.
In summary, this study shows that the current
composition of the Eurya in Taiwan is derived from
various sources, which may well reflect the complicate
phytogeographical history of Taiwan. It has long been
noticed that the highlands of Taiwan not only possess
much higher rate of endemic species (Hsieh, 2002), but
also retain a great number of Tertiary relics from N China
(Li, 1957; Liu, 1988; Shen, 1996, 1997), whereas the
lowlands of Taiwan have strong affinity with the lowlands
of SE China (Hsieh et al., 1994). The rationale is that
during the Tertiary the climate of N China was warm
and humid. By the end of Tertiary and the beginning of
Quaternary when global temperature decreased drastically
and the cycles of glacial/interglacial episodes proceeded,
the Tertiary flora of N China migrated southwards (Liu,
1988). A considerable percentage of this flora completely
vanished in N China and SE China during the middle
Pleistocene, but a great number of species were retained
in the highlands in Taiwan and Japan to the east, and
Yunnan, Guanxie, Sichuan Provinces of China to the
west (Yang and Hsu, 1980; Liu, 1988, Shen, 1996). Such
floristic similarity between SW China and the highlands
of Taiwan was noticed as early as 1920 by Wilson. In our
study, the RT endemic group of Eurya stands at the basal
most position and is distant from the remaining Eurya in
Taiwan, and it is not close to Eurya species from other
regions, either. This group is likely to be an isolated
and ancient group based on the information available.
The four Taiwanese species in the RT endemic clade are
restricted to the middle and upper elevated forests in the
Central Mountain Range where many well known Tertiary
elements are accommodated, for examples Amentotaxus,
Chamaecyparis, Fagus, Hedyderia, Juglans, Kalopanax,
Keteleria, Taiwania, Taxus, etc. (Shen, 1996). The other
group of Eurya, the RT dominant clade, shows high
affinities with Eurya species currently present in SE China
(Fujian Province), and most of them are distributed in
lowlands. Therefore, we propose that the five species of
RT endemic clade are most likely Tertiary elements from
N China or their direct descendants which evolved in
Taiwan. The remaining nine Eurya species in Taiwan are
mostly members or derivatives of the Quaternary flora of
the SE China. Along with the uplifting of Taiwan during
the late Tertiary and Quaterany, earlier immigrants tend to
be found in higher elevated areas in Taiwan. Ecological
partitioning is suggested as an important mechanism for
the speciation of several Taiwan-endemic species from
different populations of E. chinensis at probably different
geological times.
Acknowledgements. Great thanks are extended to the
following directors: Tokushiro Takaso of Iriomote Station,
the Ryukyus, Kwok-leung Yip of HK Herbarium, and
Ruth Kiew of Herbarium of Singapore Botanical Garden
for the help with local collecting and the use of herbaria.
We also appreciate the following people providing us
pg_0015
WU et al. ¡X Phylogeny and taxonomy of
Eurya
from Taiwan
111
valuable materials for sequencing: Hua-Gu Ye (South
China Institute of Botany, Academy of Science, China),
Wen-Ju Zhang (Fudan University, China), Tetsuo Denda
(University of the Ryukyus), Jenn-Che Wang (Normal
Taiwan University, Taiwan), Mong-Huai Su and Chun-
Neng Wang (National Taiwan University, Taiwan),
Yu-pin Cheng (Forestry Research Institute, Taiwan),
David Johnson (Ohio Wesleyan Univ., USA), Yin-Wei
Lin (HK Herbarium) and Suei-Ya Liu (Herbarium of
Academia Sinica, Taiwan). And, valuable comments
from Lucia Kawasaki, J. F. Maxwell, and Susan Renner
on the manuscript were much appreciated. The project
was partially supported by National Science Council
(NSC 88-2311-B-001-037) and the Institute of Plant and
Microbial Biology, Academia Sinica, Taiwan, Republic of
China.
LITERATURE CITED
Alvarez, I. and J.F. Wendel. 2003. Ribosomal ITS sequences and
plant phylogenetic inference. Mol. Phy. Evol. 29: 417-434.
Andreas en, K. and B.G. Baldwin. 2003. Nuclear ribos omal
DNA sequence polymorphism and hybridization in checker
mallows (Sidalcea, Ma lvac eae). Mol. P hy. Evol . 29:
563-581.
Bailey, C.D., T.G. Carr, S.A. Harris, and C.E. Hughes. 2003.
Characterization of angiosperm nrDNA polymorphis m,
paralogy, and pseudogene. Mol. Phy. Evol. 29: 435-455.
Chen, C.H. 2001. The phylogenetic study of Gentiana sect.
Chondrophyl lae B unge (Gent iana cea e). P h D t hes is ,
National Taiwan Normal Univerity, Taipei, Taiwan, R.O.C.
(in Chinese, with English abstract)
Cheng, Y.P., C.T. Chien, and T.P. Lin. 2000. Population genetics
of ge ographica lly re strict e and wi de spread s pe cies of
Myrica (Myricaceae). J. Hered. 91: 61-66.
Chiang, T.Y., Y.C. Chiang, Y.J. Chen, C.H. Chou, S. Havanond,
T.N. Hong, a nd S . Huang. 200 1. Ph yloge ograph y of
Kandeli a c andel in E as t As iati c ma ngrove s ba se d on
nuc leotide variation of chloropla st and m itochondrial
DNAs. Mol. Ecol. 10: 2697-2710.
Fitch, W.M. 1971. Toward defining the cours e of evolution:
minimal change for a specific tree topology. Syst. Zool. 20:
406-416.
Hatusima, S. and T. Amano. 1994. Flora of the Ryukyus, South
of Amami Island. Second Edition. The Biological Society
of Okinawa.
Hsieh, C.F. 2002. Composition, endemism and
phytogeographical affinities of the Taiwan Flora. Taiwania
47: 298-310.
Hsieh, C.F. and C.F. Shen. 1994. Introduction to the flora of
Taiwan, 1: geography, geology, climate, and soils. In T.C.
Huang et al. (eds.), Flora of Taiwan, Second Edition, Vol. 1,
pp. 1-3.
Hsieh, C.F., C.F. Shen, and K.C. Yang. 1994. Introduction to
the flora of Taiwan, 3: floristics , phytogeogra phy, and
vegetation. In T.C. Huang et al. (eds.), Flora of Taiwan,
Second Edition, Vol. 1, pp. 7-18.
Hsieh, C.F., L.K. Ling, K.C. Yang, S.Z. Yang, and C.H. Tsou.
1996. Thea cea e. In T.C. Hua ng et al . (e ds.), F lora of
Taiwan, Second Edition, Vol. II., pp. 662-693.
Hosokawa, T. 1958. On the synchorological and floristic trends
and discontinuities in regard to the Japan-Liukiu-Formosa
area. Vegetatio 8: 65-92.
Huang, S ., Y.C. Chiang, B.A. Schaal, C.H. Chou, and T.Y.
Chiang. 2001. Organelle DNA phylogeography of Cycas
taitungensis, a relic species in Taiwan. Mol. Ecol. 10:
2669-2681.
Huang, S.S.F., S.Y. Hwang, and T.P. Lin. 2002. Spatial pattern
of chloroplast DNA variation of Cyclobalanopsis glauca in
Taiwan and East Asia. Mol. Ecol. 11: 2349-2358.
Heulsenbeck, J.P. and F. Ronquist. 2001. MRBAYES: Bayesian
inference of phylogenetic trees. Bioinformatics 17:
754-755.
Kanehira, R. 1936. Formosan Trees. Department of F orestry,
Government Research Institute, F ormosa, pp. 5-20. (in
Japanese)
Kimura, M. 2000. Pale ogeography of t he Ryukyu Is lands.
Tropics 10: 5-24.
Kobuski, C.E. 1938. Studies in the Theaceae. III. Eurya
subgenera Eur yodes and Penteurya. Ann. Missouri Bot.
Gard. 25: 99-359.
Lai, C.C., Y.W. Lam, P.S. So, K.Y. Tam, Y.M. Wan, and K.L.
Yip. 2004. Checklist of Hong Kong Plants. Hong Kong
Herba rium , Agricu lture , F i she ries and C onse rvat ion
Department, Hong Kong.
Li, H.L. 1953. Floristic interchanges between Formosa and the
Philippines. Pacific Sci. 2: 179-186.
Li, H.L. 1957. The genetic affinities of the Formosa flora. Eighth
Pacific Science Congress, pp. 189-195.
Li, H.L. 1976. F lora of Taiwan, 1
st
Ed. Epoch Press , Taipei,
Taiwan.
Li, H.L. and H. Keng. 1950. Phytogeographical affinities of
southern Taiwan. Taiwania 1: 103-128.
Lin, T.P. 2001. Allozyme variations in Michelia formos ana
(Kanehira) Masamune (Magnoliaceae), and the inference
of a glacial refugium in Taiwan. Theor. Appl. Genet. 102:
450-457.
Ling, L.K. 1966. A revis ion of genus Eurya in China. Acta
Phytotaxonomica Sin. 11: 263-342. (in Chinese)
Ling, L.K. 1998. Theaceae, Ternstroemioideae. In Flora
Reipublicae Popularis Sinicae, edited by Delectis Florae
Reipublicae Popularis Sinicae Agendae Academiae Sinicae
Edita, Tomus 50 (1), Science Press. (in Chinese)
Liu, H.Y., Y.P. Yang, S.Y. Lu, and B.L. Shih. 1998. Manual of
Taiwan Vascular Plants. Vol. III. The Council of Agriculture,
Excutive Yuan, Taipei, Taiwan, R. O. C. (in Chinese)
Liu, K.B. 1988. Quaternary history of the temperate forests of
China. Quat. Sci. Rev. 7: 1-20.
L iu, T.S . and Z. Teruya. 1980. The phytogeography of the
pg_0016
112
Botanical Studies, Vol. 48, 2007
Ryukyus and Taiwan from viewpoint of woody plants. Ann.
Taiwan Mus. 23: 1-65. (in Chinese, with Japanese abstract)
L u, S .Y., K.H. Hong, S .L. Liu, Y.P. Che ng, W.L. Wu, and
T.Y. Chi ang. 2002. Geneti c vari ation a nd populat ion
differentiation of Michelia formosana (Magnoliac eae)
ba se d o n cp DNA va ria ti on a nd R APD fing erpr int s :
relevance to post-Pleistocene recolonization. J. Plant Res.
115: 203-216.
Masamune, G. 1934. P hytogeographic pos ition of Form os a
when her indigenous genera are concerned. Trans. Nat.
Hist. Soc. Formosa 22: 164-194. (in Japanese, with English
abstract)
Masamune, G. 1935. Beitrage zur kenntnisse der Flora von Sud-
japan V. Trans. Nat. Hist. Soc. Formosa 25: 248-252.
Masamune, G. 1939. Miscellaneous notes on the flora of the
eas tern Asi a XIV. Trans. Nat. Hist. S oc. Formos a 29:
339-345.
Masamune, G. 1955. tracheophytarum Ryukyu Insularum. Sci.
Rep. Kanazawa Univ. (VII) 4: 45-134.
Merrill, E.D. 1926. Enumeration of the Philippine flowering
plants. The Government of the Philippine Islands, Manila.
Nylander, J.A.A. 2004. MrModeltest v2. Program distributed
by the aut hor. E volutionary Biol ogy Ce ntre, Upps ala
University.
Rambaut, A. 1996. Se-Al: Sequence Alignment Editor. Available
at
http://evolve.zoo.ox.ac.uk/
.
Posada, D. and K.A. Cradal. 1998. MODELTES T: testing the
model of DNA substitution. Bioinformatic 14: 817-818.
Sang, T., D.J. Crawford, and T.F. Stuessy. 1995. Documentation
of reticulate evolution in peonies (Paeonia) using internal
transcribed spacer sequences of nuclear ribosomal DNA:
Implications for biogeography and concerted evolution.
Proc. Natl. Acad. Sci. USA 92: 6813-6817.
Sang, T., D.J. Crawford, and T.F. Stuessy. 1997. Chloroplast
DNA phylogeny, reticulate evolution, and biogeography of
Paeonia (Paeoniaceae). Amer. J. Bot. 84: 1120-1136.
Shen, C.F. 1996. The biogeography of Taiwan: 1. Background.
Ann Taiwan Mus 39: 387-428. (in Chinese, with English
abstract)
Shen, C.F. 1997. The biogeography of Taiwan: 2. Some
preliminary thoughts and studies. Ann. Taiwan Mus. 40:
361-450. (in Chinese, with English abstract)
Shi , Y.F., B.H. Ren, J .T. Wang, and E . De rbys hire. 1986.
Quaternary glaciation in China. Quat. Sci. Rev. 5: 503-507.
S himabuku, K.I. 1997. Check lis t of Vas cular flora of t he
Ryukyu islands, Revised Edition. Univ Kyusyu Publishing,
Japan.
Sonohara, S., S. Tawada, and T. Amano. 1952. Flora of Okinawa.
United States Civil Administrations of the Ryukyu Islands.
Stevens, P.F. 2006. Angiosperm P hylogeny Website. Version
7, May 2006. http://www.mobot.org/MOBOT/research/
APweb/.
Struwe, L., M. Thiv, J. Kadereit, T. Motley, A.S.R. Pepper,
J . Rova, K. Potgieter, P. W hite, and V.A. Albert. 1998.
Saccifolium (S accifoliaceae), an endemic of Sirra de la
Neblina on the Brazilian-Venezuelan frontier, is related to a
temperate-alpine lineage of Gentianaceae. Harvard Papers
Bot. 3: 199-214.
Swofford, D.L. 2002. PAUP *: phylogenetic analys is using
p ar s im o ny (* a nd ot he r m et h ods ) , v e r 4 . S i na u er,
Sunderland, Massachusetts, USA.
Taberlet, P., L. Gielly, G. Pautou, and J. Bouvet. 1991. Universal
primers for amplification of three non-coding regions of
chloroplast DNA. Pl. Mol. Biol. 17: 1105-1109.
Tate, J.A. and B.B. Simpson. 2003. Paraphyly of Tarasa
(Malvaceae) and diverse origins of the polyploid species.
Syst. Bot. 28: 723-737.
Urba tsc h, L .E ., B.G. Ba ldwin, and M.J. Donoghue. 2000.
P hyl ogeny of coneflowers a nd relat ive s (He liantheae:
Asteraceae) based on nuclear rDNA internal transcribed
spacer (ITS) sequences and chloroplast DNA restriction site
data. Syst. Bot. 25: 539-565.
Voris, H.K. 2000. Maps of Pleistocene sea levels in southeast
As ia: s horelines , ri ve r sys tems and tim e dura tions. J .
Biogeography 27: 1153-1167.
Walker, E.H. 1976. Flora of Okinawa and the southern Ryukyu
Islands. Smithsonian Instuitute P ress, Washington DC.
U.S.A.
Wang, C.N., M. Moller, and Q.C.B. Cronk. 2004. Polupation
genetic structure of Titanotrichum oldhamii (Gesneriaceae),
a subtropical bulbiliferous plant mixed sexual and asexual
reproduction. Ann. Bot. 93: 201-209.
Webb, T. and P.J. Bartlein. 1992. Global changes during the last
3 million years: climate controls and biotic responses. Ann.
Rev. Ecol. Syst. 23: 141-173.
Wendel, J.F., A. Schnabel, and T. Seelanan. 1995. Bidirectional
interlocus concerted evolution following allopolyploid
speciation in cotton (Gossypium). Proc. Natl. Acad. Sci.
USA 92: 280-284.
White, T.J., T. Bruns, S. Lee, and J. Taylor. 1990. Amplification
and direct sequencing of fungal ribosomal RNA genes for
phylogenetics. In M.A. Innis et al. (eds.), PCR protocols,
Academic Press, San Diego, California, USA, pp. 315-322.
Wilson, E.H. 1920. A phytogeographical sketch of the ligneous
flora of Formosa. J. Arn. Arbor. 2: 25-41.
Wu, J.E., S. Huang, J.C. Wang, and W.F. Tong. 2001. Allozyme
va riat ion and the gene tic s truc ture of popul ati ons of
Trochodendron aralioided, a mon otypi c and narrow
geographic genus. J. Plant Res. 114: 45-57.
Wu, C.C., Z.F. Hsu, and C.H. Tsou. 2003. S tudies of Eurya
(Ternstroemiaceae) in Taiwan (1), a new endemic species,
Eurya septata. Bot. Bull. Acad. Sin. 44: 67-72.
Yang, H.R. and X. Hs u. 1980. Qua ternary e nvironment al
changes in eastern China. J Nanjing Univ (Natr Sci Series)
1: 121-144. (in Chinese, with English abstract)
Zhao, Z.B. 1982. A preliminary study on the evolution of Taiwan
Strait. Taiwan Strait 1: 20-24. (in Chinese, with English
abstract).
pg_0017
WU et al. ¡X Phylogeny and taxonomy of
Eurya
from Taiwan
113
Appendix 1. Information of collection site of the 146 vouchers listed in Table 1.
Species name
Collection site
Eurya acuminata DC. var. acuminata
Singapore: Tsou 1538*.
Eurya acuminata var. wallichiana Dyer
N. Thailand: Maxwell 2004-233*; Maxwell & Euder 1.
Eurya acuminatissima Merr. & Chun
Hong Kong: Tsou et al. 1813*, 1814.
Eurya amplexifolia Dunn
China - Guangdong: Tsou 1578*.
Eurya chinensis R. Br. var. chinensis
E. chinensis-T
Taiwan - Taipei: Tsou et al. 1269, 1327, 1330*, 1378, 1379;
Nantou: Tsou et al. 1258; Pingdong: Tsou et al. 1250, 1423;
Yilan, Turtle Isl.: Tsou et al. 1720.
Ryukyu - Iriomote Isl.: Tsou et al. 1425.
E. chinensis-C
China ¡V Guangdong: Tsou et al. 1583*; Hong Kong: Tsou et al.
1792, 1794, 1800, 1836, 1837, 1891.
Eurya chinensis var. glabra Hu et L. K. Ling
Taiwan - Yilan, Turtle Isl., Chen, C.C. 7403; Hong Kong: Tsou
et al. 1802*, 1892.
Eurya crenatifolia (Yamamoto) Kobuski
Taiwan - Taipei: Tsou et al. 1253*, 1319, 1381, 1512; Hsinchu:
Tsou et al. 1485, 1487; Taichung: Tsou et al. 1333; Yilan: Tsou
et al. 1604, 1614.
Eurya disticha Chun
China - Guangdong: Tsou et al. 1973*, 1975.
Eurya distichophylla Hemsl
China - Guangdong: Tsou et al. 1577*, 2016; Fujian: Tsou et al.
2113.
Eurya emarginata Makino
Taiwan - Kinmen Isl.: Tsou et al. 1399, 1400*; Taipei: Tsou et al.
1252, 1495.
Eurya glaberrima Hayata
Taiwan - Taoyuan: Tsou et al. 1370, 1374, 1376; Taichung: Tsou
et al. 1414*; Kaohsiung: Tsou et al. 1652; Yilan: Tsou et al.
1601, 1603.
Eurya gnaphalocarpa Hayata
Taiwan - Taichung: Tsou et al. 1331*; Nantou: Tsou et al. 1260,
1864; Yilan: Tsou et al. 1314, 1764.
Eurya groffii Merr.
Hong Kong: Tsou et al. 1890*, 1894.
Eurya hayatai Yamamota
Taiwan - Hsinchu: Su, Mong Huai 227, 282; Taichung: Tsou et
al. 1626*, 1627; Kaohsiung: Su, Mong Huai 109.
Eurya hebeclados Ling
China - Guangdong: Tsou et al. 1944*; Zhejiang: Zhang, Wen Ju,
2000-12-18-c.
Eurya impressinervis Kobuski
China - Guangdong: Ye, Hua Gu 2003-5-30-a*.
Eurya japonica Thunb.
Japan - Nago: Tetsuo Denda 2000-12-6-a*, 2000-12-6-b; China -
Shanghai: Zhang, Wen Ju 2001-9-19-a, 2001-9-19-b.
Eurya laotica Gagnep.
Vietnam - Sapa: Tsou et al. 1620*.
Eurya leptophylla Hayata
Taiwan - Hsinchu: Tsou et al. 1355*, 1356; Nantou: Lin, Chia
Hua 523; Tsou et al. 1456, 1634; Kaohsiung: Tsou 1647;
Taidong: Tsou 1503.
Eurya loquaiana Dunn
Taiwan - Hsinchu: Tsou et al. 1352; Nantou: Tsou et al. 1341;
Yilan: Tsou et al. 1316*.
Eurya macartneyi Charmp.
China - Guangdong: Tsou et al. 1579*; Hong Kong: Tsou et al.
1831, 1839.
pg_0018
114
Botanical Studies, Vol. 48, 2007
Species name
Collection site
Eurya muricata Dunn
China - Zhejiang: Zhang, Wen Ju 2000-12-18-a*, 2000-12-18-b.
Eurya nanjenshanensis (Hsieh, Ling, & Yang) Yang & Lu Taiwan - Pingdong: Tsou et al., 1419*, 1421, 1422.
Eurya nitida Korthals
Taiwan - Taipei: Tsou et al. 1279, 1328.
China - Guangdong: Tsou et al. 1582.
Hong Kong: Tsou et al. 1798, 1806, 1826*, 1868.
Eurya quinquelocularis Kobuski
China - Guangdong: Ye, Hua Gu 2002-6-29-b*, 2002-6-29-e.
Eurya rengechiensis Yamamota
Taiwan - Nantou: Tsou et al. 1342*, 1367; Su, Mong Huai 241,
242.
Eurya rubiginosa Chang
China - Guangdong: Tsou et al. 1580*.
Eurya ryukyuensis Masamune
Ryukyu - Iriomote Isl.: Tsou et al. 1436*; Wang, Chun Neng
1520; Wang, Jenn Che 2003-10-6; Okinawa: Wang, Jenn Che
2003-10-5.
Eurya sakishimensis Hatusima
Ryukyu - Iriomote Isl.: Tsou et al. 1301*, 1442; Wang, Chun
Neng, s.n.
Eurya septata Wu, Hsu, & Tsou
Taiwan - Taipei: Tsou et al. 1470, 1471; Nantou: Tsou et al.
1476, 1477*; Taidong: Tsou et al. 1499.
Eurya strigillosa Hayata
Taiwan - Nantou: Tsou et al. 1335*; Kaohsiung: Tsou et al. 1650,
1623.
Eurya subintegra Kobuski
China - Guangdong: Ye, Hua Gu 2002-6-29-c*; Tsou et al. 1987.
Eurya weissiae Chun
China - Guangdong: Ye, Hua Gu 2002-6-26-b*, Ye, Hua Gu s.n.
Eurya yaeyamensis Masamune
Ryukyu - Iriomote Isl.: Tsou et al. 1282, 1439, 1540*.
Adinandra dumosa Jack
Singapore: Tsou 1539*.
Adinandra elegans How & Ko ex Chang
China - Guangdong: Tsou 1584*.
Adinandra formosana Hayata
Taiwan - Hsinchu: Tsou et al. 1493; Taipei: Tsou 827
(AF089713), Tsou 1546*.
Adinandra lasiostyla Hayata
Taiwan - Nantou: Tsou et al. 1334, 1346*.
Adinanda millettii (Hook. & Arn.) Benth. & Hook. Ex Hance Hong Kong: Tsou et al. 1799*, 1889.
Adinandra yaeyamensis Ohwi
Ryukyu - Iriomote Isl.: Tsou et al. 1303*.
Anneslea fragrance Wall. var. lanceolata Hayata
Taiwan - Pingdong: Tsou et al. 1099*.
Cleyera japonica Thunb.
Ryukyu - Iriomote Isl.: Tsou et al. 1296*, 1446.
Taiwan - Kaohsiung: Tsou et al. 1631.
Cleyera morii Masamune
Taiwan - Taipei: Tsou et al. 858*, 1272.
Euryodendron exclesum Chang
China - Guangdong: Chung-Shan Univ., Tsou et al. 1590*.
Ternstroemia gymnanthera (Wight et Arn.) Sprague
Taiwan - Taipei: Tsou et al. 1098*, 1304.
Ternstroemia kwangtungensis Merr.
China - Guangdong: Tsou et al. 1585*.
Ternstroemia luteoflora Ling
China - Guangdong: Tsou et al. 1587*; Hong Kong: Tsou et al.
1842.
Ternstroemia microphylla Merr.
China - Guangdong: Tsou et al. 1586*.
Appendix 1. (Continued)
pg_0019
WU et al. ¡X Phylogeny and taxonomy of
Eurya
from Taiwan
115
Appendix 2. List of psbA-trnH and trnL-trnF samples and their accession numbers. Vouchers are deposited at HAST (Herabrium,
Academian Sinica, Taiwan)
Species
Accession
number Voucher information
psbA-trnH
Eurya acuminata DC. var. acuminata
AY943244 Tsou 1538, Singapore
Eurya acuminata var. wallichiana Dyer
AY943245 Maxwell 2003-6-13, N Thailand
Eurya acuminatissima Merr. & Chun
AY943246 Tsou et al. 1813, Hong Kong
Eurya amplexifolia Dunn
AY943247 Tsou 1578, Guangdong, China
Eurya disticha Chun
AY943248 Ye, Hua Gu 2002-6-26, Guangdong, China
Eurya distichophylla Hemsl
AY943249 Tsou et al. 1577, Guangdong, China
Eurya emarginata Makino
AY943250 Tsou et al. 1400, Kinmen Isl., Taiwan
Eurya glaberrima Hayata
AY943251 Tsou et al. 1414, Taichung, Taiwan
Eurya gnaphalocarpa Hayata
AY943252 Tsou et al. 1331, Taichung, Taiwan
Eurya groffii Merr
AY943253 Tsou et al. 1890, Hong Kong
Eurya hayatai Yamamota
AY943254 Tsou et al. 1626, Taichung, Taiwan
Eurya hebeclados Ling
AY943261 Zhang, Wen Ju, 2000-12-18-c, Zhejiang, China
Eurya impressinervis Kobuski
AY943255 Ye, Hua Gu 2003-5-30-a, Guangdong, China
Eurya laotica Gagnep
AY943256 Tsou et al. 1620, Sapa, Vietnam
Eurya macartneyi Charmp
AY943257 Tsou et al. 1579, Guangdong, China
Eurya nitida Korthals
AY943258 Tsou et al. 1826, Hong Kong
Eurya quinquelocularis Kobuski
AY943259 Ye, Hua Gu 2002-6-29-b, Guangdong, China
Eurya rubiginosa Chang
AY943260 Tsou et al. 1580, Guangdong, China
Eurya strigillosa Hayata
AY943263 Tsou et al. 1335, Nantou, Taiwan
Eurya subintegra Kobuski
AY943264 Ye, Hua Gu 2002-6-29-c, Guangdong, China
Eurya weissiae Chun
AY943262 Ye, Hua Gu s.n., Guangdong, China
Euryodendron exclesum Chang
AY943269 Tsou et al. 1590, Guangdong, China
Ternstroemia gymnanthera (Wight et Arn.) Sprague AY943270 Tsou et al. 1098, Taipei, Taiwan
Ternstroemia kwangtungensis Merr.
AY943271 Tsou et al. 1585, Guangdong, China
Ternstroemia luteoflora Ling
AY943272 Tsou et al. 1587, Guangdong, China
TrnL-trnF
Eurya emarginata Makino
AY943273 Tsou et al. 1716, Turtle Is., Taiwan
Eurya glaberrima Hayata
AY943274 Tsou et al. 1603, Yilan, Taiwan
Eurya strigillosa Hayata
AY943275 Tsou et al. 1650, Pingdung, Taiwan
Eurya acuminata var. wallichiana Dyer
AY943276 Maxwell 2004-233, N Thailand
pg_0020
116
Botanical Studies, Vol. 48, 2007