Molecular Phylogenetics and Evolution
Vol. 14, No. 1, January, pp. 11–19, 2000
Article ID mpev.1999.0672, available online at http://www.idealibrary.com on
Molecular Phylogeny of Coriaria, with Special Emphasis
on the Disjunct Distribution
Jun Yokoyama,* Mitsuo Suzuki,† Kunio Iwatsuki,‡ and Mitsuyasu Hasebe§,1
*Biological Institute, Tohoku University, Aoba, Aramaki, Aoba-ku, Sendai, Miyagi 980-8578, Japan; †Botanical Garden, Tohoku
University, Kawauchi, Aoba-ku, Sendai, Miyagi 980-8576, Japan; ‡Faculty of Science, Rikkyo University, 3-34-1 Nishi-Ikebukuro,
Toshima-ku, Tokyo 171-0021, Japan; and §National Institute for Basic Biology, 38 Nishigonaka, Myodaiji-cho, Okazaki 444-8585, Japan
Received October 12, 1998
Coriaria, which has the most conspicuously disjunct
distribution of the flowering plants, is distributed in
four separate areas of the world. The phylogenetic
relationships of 12 Coriaria species collected from the
representative disjunct areas were inferred by compar-
ing 2416 bp of the combined data set of rbcL (a large
subunit of ribulose 1,5-bisphosphate carboxylase/
oxygenase) and matK (maturase K) genes. The phyloge-
netic tree shows that the Chile–Papua New Guinea–
New Zealand–Pacific islands species and the Central
America–northern South America species form a sister
group, and the Eurasian clade is more basal to them.
The divergence time between the Eurasian group and
the other species was estimated as 63 or 59 million
years ago using rbcL and matK molecular clocks,
respectively. These results do not support previously
proposed hypotheses which explain the disjunct distri-
bution on the basis of continental drift but suggest that
the distribution pattern was formed by several geo-
graphical migrations and separations in the Cenozoic.
r 2000 Academic Press
Key Words: Coriaria; angiosperm; biogeography, mo-
lecular systematics; rbcL; matK.
INTRODUCTION
Genus Coriaria is one of the more interesting taxa of
angiosperms because of its isolated geographical distri-
bution. Coriaria is widely distributed in four separate
areas of the world (Fig. 1): the Mediterranean; continen-
tal and insular eastern Asia (the Himalayas of southern
China, Japan, Taiwan, and the Philippines); Papua
New Guinea to New Zealand, including the adjacent
Pacific islands; and Northern Mexico to southern Chile
through Central America (Good, 1930; Allan, 1961;
Skog, 1972). Good (1930) speculated that the progres-
sion of glaciers in the Tertiary period divided the
1 To whom correspondence should be addressed. Fax: ⫹81-564-55-
7546. E-mail: mhasebe@nibb.ac.jp.
11
habitat and caused the present discontinuous distribu-
tion in each hemisphere. He also proposed that the
discontinuous distribution between the two hemi-
spheres was a result of continental displacement. Mel-
ville (1966, 1981) suggested that Coriaria enlarged its
distribution from the Gondwanic home in the Southern
Hemisphere to the Northern Hemisphere, passing onto
the former Pacific continent in the early Cretaceous,
although the existence of that continent has not been
demonstrated. Croizat (1952) hypothesized that Cori-
aria migrated from the Southern Hemisphere to the
Northern Hemisphere via the Pacific islands. Schuster
(1976) speculated that Coriaria originated in Gond-
wanaland and migrated to Asia by rafting on the Indian
plate.
The phylogenetic position of the Coriariaceae is
unsettled. Many authors placed the Coriariaceae near
to or within a sapindalean–rutalean group, although as
an isolated group (Scholz, 1964; Takhtajan, 1980; Dahl-
gren, 1983; Thorne, 1983). This view was supported by
results from embryology (Sharma, 1968), pollen struc-
ture (Garg, 1981), and chemical compounds (Bohm and
Ornduff, 1981). Alternatively, Croizat (1952) and Cron-
quist (1981) suggested that the most reasonable posi-
tion for the family is in the Ranunculales because floral
characters and anatomy of the stem seem more compat-
ible with Ranunculales than with any other orders.
Corner (1976) showed that the seed coat structures of
Coriaria resemble those of the Ranunculaceae. On the
other hand, based on the large subunit of ribulose-1,5-
bisphosphate carboxylase/oxygenase gene (rbcL) se-
quences, Chase et al. (1993), Swensen et al. (1994), and
Swensen (1996) produced the unexpected result that
Coriaria was closely related to the Begoniaceae, Cucur-
bitaceae, and Datiscaceaea and it did not have close
relationships to either the sapindalean–rutalean group
or the Ranunculales sensu Cronquist (1981).
Species delimitation in Coriaria is controversial, and
5 to 20 species have been recognized, depending on the
authors (5 species: Skog, 1972; Cronquist, 1981; Thorne,
1983; 10 species: Scholz, 1964; 15 species: Airy Shaw,
1055-7903/00 $35.00
Copyright r 2000 by Academic Press
All rights of reproduction in any form reserved.
 12
YOKOYAMA ET AL.

FIG. 1. Distribution of Coriaria. This figure is based on Good (1930), Melville (1981), Hotta (1974), and Dr. Hideaki Ohba’s unpublished study. Species used in this study are
indicated by asterisks.
 MOLECULAR PHYLOGENY OF Coriaria 13

1973; and 20 species: Good, 1930), probably because of
the large morphological variation caused by natural
hybridization (Good, 1930; Allan, 1961). Although Good
(1930) recognized three groups in the genus based on
some conspicuous and constant vegetative and reproduc-
tive characters, the interspecific relationships of Cori-
aria have not been studied because of the lack of
informative morphological characters and appropriate
outgroups for cladistic analysis. The inferred phyloge-
netic tree is expected to be useful to hypotheses about
the evolutionary processes which caused the disjunct
distribution of Coriaria.
In cases in which morphological characters useful for
phylogenetic inference are scarce, DNA sequence data
are often helpful for phylogenetic analyses and produce
useful insights in biogeographical studies. The purpose
of this study is (1) to elucidate the interspecific relation-
ships within Coriaria using DNA sequence data de-
rived from two plastid genes, rbcL and matK, and (2) to
evaluate the previously proposed hypotheses about the
disjunct distribution of this genus.
MATERIALS AND METHODS
Plant materials used in this study are listed in Table
1. The Coriaria sp. collected in Fiji is different from any
other known Coriaria species, and further taxonomic
analysis will be published elsewhere. The species were
selected to represent all areas of the disjunct distribu-
tion of this genus. Leaves (2–3 g) pulverized with liquid
nitrogen were used for isolations of total DNAs by the
modified CTAB method (Hasebe and Iwatsuki, 1990).
The PCR mix contained 100 to 200 ng of total DNA, 1
µM each primer, 200 mM each deoxynucleotide, 10 mM
Tris–HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, and 1.25
units of Taq polymerase.
We used primers described by Hasebe et al. (1994) to
amplify the rbcL gene. For the matK, we used a pair of
primers: matK-AF (Ooi et al., 1995) and matK-8R
(Steele and Vilgalys, 1994) to amplify an approx 1.2-kb
region of matK. Then, we used the two primers and an
additional four internal primers listed in Table 2 for
sequencing. The thermal cycle profile for amplifications
was 1 cycle of 94°C (2 min), 45°C (2 min), 60°C (3 min),
30 cycles of 94°C (1.5 min), 45°C (2 min), 60°C (3 min),
followed by 15 min of extension profile at 72°C. The
PCR products were separated on 1% agarose gels and
purified using the GENECLEAN II kit (BIO 101). Cycle
sequencing reactions were performed using about 80–
100 ng of purified PCR products and the PRISM Ready
Reaction DyeDeoxy Terminator Cycle sequencing kit
(Applied Biosystems) with the manufacturer’s instruc-
tions. DNA sequences were obtained by using the 373A
DNA Sequencer or the PRISM 310 Genetic Analyzer
(Applied Biosystems).
Sequences obtained in this study were aligned manu-
ally. Phylogenetic analyses were performed by three
methods: the maximum parsimony (MP) method, the
neighbor-joining (NJ) method, and the maximum likeli-
hood (ML) method. For MP analyses, we used PAUP
ver. 3.1.1 (Swofford, 1993). We assigned equal weight to
each codon position. The heuristic search option using
100 replications of RANDUM addition with TBR branch
swapping was employed to search for multiple islands
of the most parsimonious trees (Maddison, 1991). Boot-
strap (Felsenstein, 1985) and decay analyses (Bremer,

TABLE 1

List of Coriaria, Begonia, and Datisca Used in this Study

GenBank GenBank
Accession Accession
Taxon Locality Specimen a No. of rbcL No. of matK

C. arborea Lindsay New Zealand: Wakarewarea Forest Park, North Suzuki et al., 19871 AB016440 AB016454
Island
C. intermedia Matsum. Taiwan: Taroko National Park Suzuki and Suzuki, 19891 AB016441 AB016455
C. japonica A. Gray Japan: Tedori River, Ishikawa Pref. Suzuki et al., 19891 AB016442 AB016456
C. lurida Kirk New Zealand: Waimea, South Island Suzuki et al., 19871 AB016443 AB016457
C. microphylia Poir. Mexico Tobe and Oginuma, 19921 AB016444 AB016458
C. myrthifolia L. Morocco: Marruecos, Chefchaouen Ron, 19891 AB016445 AB016459
C. napalensis Wall. Nepal: Ghasa-Lete, Gandaki Zone Suzuki, 19891 AB016446 AB016460
C. papuana Warb. Papua New Guinea: Dauro Pass, Eastern High- Watanabe, 19891 AB016447 AB016461
lands Prov.
C. ruscifolia L. Chile: La Junta, Prov. Aisen Uemura et al., 19891 AB016448 AB016462
C. sp. Fiji: Wainimala, Viti Levu Isl. Setoguchi, 19901 AB016449 AB016463
C. sarmentosa Forster f. New Zealand: Near Kikiwa, Waimea, South Island Suzuki et al., 19871 AB016450 AB016464
C. terminalis Hemsl. Nepal: Wolanchung Gola, Taplejung Suzuki et al., 19921 AB016451 AB016465
Begonia evansiana Andr. Japan: Aramaki, Aoba-ku, Sendai, Miyagi Pref. Yokoyama 9500012 AB016452 AB016466
Datisca cannabina L. Cultivated in Botanical Gardens, University of Hasebe 9504012 AB016453 AB016467
Tokyo
a Voucher specimens are deposited in TI(1) or TUS(2).
14 YOKOYAMA ET AL.

TABLE 2 base pair positions 57 and 1280 from the initial methio-
nine codon of Nicotiana tobaccum (Shinozaki et al.,
List of matK Primer Sequence and Location
1986) was used for phylogenetic analyses. There are 97
Primer name Sequence (58 = 38) Position Reference variable nucleotide positions, of which 34 nucleotides
are potentially informative. The obtained matK se-
CTATATCCACTTATCTTTCAG- quences for phylogenetic analyses are between base
matK-AF GAGT 55–79 Ooi et al., 1995
GGTTTGCACTCATTGTGGAA-
pair positions 91 and 1291 from the initial methionine
CRmatK-F1 ATTCC 296–320 This study codon of N. tobaccum (Shinozaki et al., 1986). Com-
TCCTATATAATTCTCATGTAT- pared to the N. tobaccum matK, two deletions (115–120
CRmatK-F2 GTGA 641–665 This study and 340–348 of N. tobaccum) are present in Coriaria,
GCCTCTTCTGATGAATAAGT- Begonia, and Datisca matK and a 6-nucleotide inser-
CRmatK-F3 GGAA 882–905 This study
Steele and tion (774–775 of N. tobaccum) exists in Coriaria. Thus,
AAAGTTCTAGCACAAGAAAG- Vilgalys, the 1186-bp region shared among Coriaria, Begonia,
matK-8R TCGA 1331–1208 1994 and Datisca was used for phylogenetic analyses. One
TACCACTGAAGGATTTAGTC- hundred and eighty-four variable nucleotide sites are
CRmatK-R1 GCACA 1044–1020 This study
AAGATGTTAATCGTAAATG-
present, of which 110 are potentially informative.
CRmatK-R2 AGAAG 724–701 This study The intrageneric relationship of Coriaria was in-
GGTATTAATACATCTGACA- ferred using the rbcL or matK sequence data set with
CRmatK-R3 CATAAT 467–443 This study the MP method, and both trees obtained are almost
congruent (data not shown). To assess the congruence
Note. Position indicates base pair from the initial methionine codon
of Nicotiana tobaccum (Sugita et al., 1985). between the trees based on the different data sets, the
Mickevich–Farris incongruence index (IMF; Mickevich
and Farris, 1981) was calculated. The lengths of the
1988) were used to obtain a measure of confidence for most parsimonious trees for the rbcL, the matK, and
each branch. One thousand bootstrap replications were their combined data sets are 103, 244, and 348, respec-
carried out with equal weighting, simple sequence tively, and the IMF value is 0.0029, indicating that
addition, and NNI branch swapping. 0.29% of the total character incongruence is due to the
For NJ analyses, we used the NEIGHBOR program combining of the rbcL and matK data sets. Because of
of PHYLIP version 3.572 (Felsenstein, 1995). In calcu- this low value and the congruency of the tree topologies
lating the distance matrix, the Kimura two-parameter inferred by two different data sets, we used the com-
model of nucleotide substitutions (Kimura, 1980) and bined data set in further analyses.
the DNAML model of Felsenstein (1981) were imple- In the MP analysis with equal weighting using the
mented with the DNADIST program of the PHYLIP combined data set, 6 equally most parsimonious trees
program package. The transition/transversion ratio of 351 steps were found on one island. These trees have
was fixed to 2.0. One thousand bootstrap replications a consistency index (excluding uninformative charac-
were conducted with the SEQBOOT program of the ters) of 0.84, a retention index of 0.90, and a rescaled
same package. consistency index of 0.82. The strict consensus tree
We used the fastDNAml version 1.0 (Felsenstein, with bootstrap values for nodes supported in ⱖ50% of
1981; Olsen et al., 1994) for the ML analysis. The transition/ 100 bootstrap replicates is shown in Fig. 2A. The NJ
transversion ratio was fixed to 2.0. Ten random se- tree using distances calculated by the DNAML model of
quence addition searches by the Jumble option with Felsenstein (1981) is shown in Fig. 2B. The topology of
global branch swapping (the Global option) were em- this tree is the same as that of a tree obtained with the
ployed to obtain a ML tree with the best log-likelihood. Kimura two-parameter method (Kimura, 1980). In the
One hundred bootstrap replications were carried out by ML analysis, each random sequence addition search
the Bootstrap option of the same program. produced 714 to 1242 different trees, and the ML tree
The number of synonymous substitutions was esti- with the best log-likelihood (⫺5524.81) is shown in Fig.
mated by method 2 of Ina (1995) to calculate the diver- 2C.
gence time between species of Coriaria. The constancy These three different methods generate mostly con-
of evolutionary rate was assessed with the relative rate gruent trees. The clade which includes the Eurasian
test (Wu and Li, 1985) in which Begonia evansian and species is basally branched off from other species.
Datisca cannabina were used as reference taxa. Monophyly of the Eurasian clade is well supported by
the bootstrap analyses. Two distinct clades are recog-
RESULTS nized in the Eurasian clade. One clade consists of
species in continental Asia (C. terminalis and C.
Phylogenetic Relationships in Coriaria nepalensis) and the Mediterranean (C. myrtifolia), and
The rbcL sequences could be aligned without any the other clade is composed of species in islands of
insertions or deletions. The 1224-bp region between eastern and south eastern Asia (C. japonica and C.
 MOLECULAR PHYLOGENY OF Coriaria
intermedia). The phylogenetic relationship within the
former clade is incongruent among the three methods,
and none of them are supported with high bootstrap
values. C. microphylla is a sister taxon to all Coriaria of
the South Hemisphere, in which C. papuana of Papua
New Guinea and C. sp. of Fiji form a basal clade,
although the monophyletic relationship of the clade is
not well supported with the bootstrap analyses. Species
of New Zealand and Chile form a clade, in which C.
arborea of New Zealand is basal. Monophyly of a
Chilean and two New Zealand species (C. ruscifolia, C.
sarmentosa, and C. lurida) is well supported by the
bootstrap analyses in all methods.
Estimation of Divergence Time
Nucleotide substitution rates are not perfectly clock-
like (Bousquet et al., 1992) and are considered to be
influenced by life history factors and/or generation
times in seed plants (Bousquet et al., 1992; Li, 1993;
Hasebe et al., 1998). The relative rate test (Wu and Li,
1985) was used to determine the constancy of evolution-
ary rates among species in Coriaria using the esti-
mated number of synonymous substitutions of each
rbcL and matK gene. Datisca cannabina, inferred to be
closely related to Coriaria (Swensen et al., 1996), was
selected as a reference species, and its rbcL and matK
genes were sequenced (Accession No. of each gene is
AB016453 and AB016467, respectively). Although Oc-
tomeles, Tetrameles, and Corynocarpus are inferred to
be more closely related to Coriaria by Swensen et al.
(1996), we could not obtain the materials. When we
compared the evolutionary rate of each Eurasian spe-
cies (C. intermedia, C. japonica, C. myrtifolia, C.
nepalensis, and C. terminalis) to one of the other
species (C. microphylla, C. papuana, C. sp., C. arborea,
C. ruscifolia, C. sarmentosa, and C. lurida), the null
hypothesis of constancy in the evolutionary rate was
not rejected at the 5% significance level in both lin-
eages, except in C. intermedia and C. sp. (data not
shown). The evolutionary rates of both rbcL and matK
genes are accelerated in the C. intermedia lineage but
decelerated in the C. sp. lineage. Therefore, we ex-
cluded C. intermedia and C. sp. from the following
analyses. When Begonia evansiana (Accession Nos.
AB016452 and AB016466 for rbcL and matK genes,
respectively), which is also inferred to be closely related
to Coriaria (Swensen et al., 1996), was used as a
reference species, the results were not changed (data
not shown).
Several fossils have been reported as Coriaria but
most of them are detached leaves (Ozaki, 1991), pollens
(Praglowski, 1970; Muller, 1981), and seeds (Gregor,
1980), which are difficult to relate to extant species.
However, one specimen is a well-preserved fossil of a
shoot containing inflorescences and leaves, C. longaeba
(Saporta, 1865), from the Armissan beds of the Oli-
gocene age, about 30 million years ago (MYA). The
15
inflorescence is a long terminal raceme characteristic of
C. terminalis in Coriaria. All other extant Coriaria
have lateral and axial inflorescences, and it is apparent
that the fossil is closely related to C. terminalis (Sa-
porta, 1865; Good, 1930). We therefore assume that the
terminal inflorescence is a synapomorphic character of
C. terminalis and C. longaeba and that C. longaeba
forms a sister group with C. terminalis, assigning the
divergence time between the two lineages as at most 30
MYA. The synonymous nucleotide substitution rates
(Ks) of the Coriaria rbcL and matK genes were esti-
mated as 1.7 ⫻ 10⫺10 ⫾ 0.2 ⫻ 10⫺10 and 3.6 ⫻ 10⫺10 ⫾
0.9 ⫻ 10⫺10 nucleotides per site per year, respectively.
The evolutionary rate of the Coriaria rbcL gene is
similar to previously reported rates of woody plants
(Albert et al., 1993) but slower than those of herbaceous
taxa (Wolfe et al., 1989). The rate difference between
the rbcL and the matK genes is about double, which is
greater than that of Steele and Vilgalys (1994) but
smaller than that of Johnson and Soltis (1995). Based
on the evolutionary rates of the Coriaria genes, the
divergence time between the Eurasian group and the
other species was estimated to be 63.1 ⫾ 14.7 MYA
(rbcL) and 59 ⫾ 9.7 MYA (matK ) on average (Fig. 2C).
DISCUSSION
Intrageneric Relationships of Coriaria
Good (1930) recognized three groups (A, B, and C) in
Coriaria, based on the geographic distribution, inflores-
cence type, shape of winter bud scales, and sexuality of
flowers. Group A consists of only one species, C. termina-
lis, which has unusual terminal inflorescences instead
of the lateral and axial inflorescences of other Coriaria.
Group B includes all Eurasian species except C. termi-
nalis, C. intermedia, C. japonica, C. myrtifolia, and C.
nepalensis. The other species used in this study are all
included in group C, which is characterized by rounded
leafy bracts subtended at the base of an axial inflores-
cence in contrast to the bud scales shared by the
members of group B. Our phylogenetic analyses using
the combined nucleotide sequence data of the rbcL and
matK genes showed that groups A and B should be
combined into one group because separation of C.
terminalis makes group B polyphyletic (Fig. 2). Mono-
phyly of group C is well supported by the bootstrap
analyses (Fig. 2). Therefore, combining our results
together with Good’s system, we propose to divide
Coriaria into two groups: group 1, including C. termina-
lis and the species in group B of Good (1930), and group
2, corresponding to group C of Good (1930) (Fig. 2C).
Biogeographic Implications
Species in Coriaria are scattered into separate areas
(Fig. 1). This scattering has been noted as the most
conspicuous disjunct distribution in angiosperms (Good,
1930). The processes leading to such a disjunct distribu-
16 YOKOYAMA ET AL.

FIG. 2. Phylogenetic trees of Coriaria species based on the combined data set of 1224-bp rbcL and 1186-bp matK sequences. Bootstrap
values are indicated for nodes supported in ⱖ50% boot strap replicates. The trees are rooted by Datisca cannabina and Begonia evansiana. (A)
Strict consensus of 6 equally most parsimonious trees obtained using maximum parsimony with equal weighting, simple sequence addition,
and NNI branch swapping with MULPARS and STEEPEST DESCENT selected. Branch lengths are arbitrary. The value of the decay index
for each node is indicated under the bootstrap value. (B) Neighbor-joining tree (Saitou and Nei, 1987) with the DNAML distance (Felsenstein,
1981). Branch lengths are proportional to the expected mean number of substitutions per site along the branch, as quantified by the scale bar.
(C) The best maximum likelihood tree with a log-likelihood of ⫺5524.81 obtained after 10 random sequence addition searches with local and
 MOLECULAR PHYLOGENY OF Coriaria
tion have been widely discussed. Croizat (1952), Mel-
ville (1966, 1981), and Schuster (1976) speculated that
Coriaria originated in the eastern part of Gondowanal-
and, which was composed of present-day Papua New
Guinea, New Zealand, the Pacific islands between the
former two big islands, and South America. Croizat
(1952) believed the Pacific islands to be tracks of
dispersal from the Southern Hemisphere to Eurasia.
Melville (1966, 1981) hypothesized that C. papuana of
New Guinea and C. intermedia of Taiwan and the
northern Philippines form a sister pair and that the
latter species migrated to the Northern Hemisphere
from the Southern Hemisphere. The parental species of
these two species was suggested to be ‘‘C. ruscifolia-like
species’’ corresponding to the ancestor of the extant
species in the Southern Hemisphere. He also specu-
lated that C. intermedia expanded its distribution to
China, the Himalayas, Japan, and the Mediterranean
area and diverged into the different species presently
observed. Schuster (1976) proposed the migration of
Coriaria to the Northern Hemisphere by the movement
of India from Gondowanaland to Eurasia. Both Melville
(1966, 1981) and Schuster (1976) speculated that Cori-
aria expanded its distribution from South America to
Central America, suggesting C. microphylla and C.
ruscifolia form a sister group.
The phylogenetic tree of the combined rbcL–matK
nucleotide sequences indicates that the Eurasian spe-
cies (group 1 in Fig. 2C) are the most basal in Coriaria,
while the species of the South Hemisphere (C. papuana,
C. sp. C. arborea, C. ruscifolia, C. sarmentosa, and C.
lurida) form a clade after the divergence of C. micro-
phylla (Fig. 2C). The data suggest that Coriaria did not
expand its distribution from the South Hemisphere via
the Pacific islands. C. ruscifolia, regarded as an ances-
tral form of extant Coriaria by Melville (1966, 1981), is
not located in a basal lineage but is rather close to the
end of the tree. Although C. intermedia and C. papuana
are adjacently distributed, as Melville (1966, 1981)
pointed out, these species are not closely related to each
other. Therefore, the proposed migration via the Pacific
islands which accompanied the speciation of C. interme-
dia and C. papuana does not seem to have occurred. In
the Eurasian species, C. intermedia is not basal, elimi-
nating the possibility of the Pacific migration route.
Furthermore, the migration of Coriaria from South
America to Central America is not likely to be the case,
because C. ruscifolia and C. microphylla do not form a
sister group in Fig. 2C.
Based on the phylogenetic relationships of the extant
species and the estimated divergence time between
global branch swappings. Estimated divergence times between Groups 1 and 2 are indicated. Distribution, grouping of Good (1930), and
rbcL–matK grouping proposed in this study are shown on the right side of the tree. Areas are indicated with the following abbreviations: CAS,
continental Asia around the Himalayas; CAM, central America; IAS, insular Asia; MED, the Mediterranean; NZ, New Zealand; PAS, the
Pacific islands; PNG, Papua New Guinea; SAM, South America.
17
groups 1 and 2, the following scenario of the evolution
of Coriaria is proposed. The basal branching of the
Eurasian Coriaria suggests that Coriaria originated in
either Eurasia or North America. Coriaria was prob-
ably more widely distributed than at the present, at
least in Eurasia, because a fossil of C. longaeba was
found in southern France (Saporta, 1865) far from the
present distribution of its close relative, C. terminalis,
which is restricted to the Himalayas (Fig. 1). The
estimated divergence time between group 1 and group
2 is about 60 MYA, the early Tertiary, when Eurasia
and North America were closer to each other than at
the present (Smith and Briden, 1977; Briggs, 1987;
Hughes, 1994) and the Arctic circumpolar region was
warm enough to support temperate plant species
(Chaney, 1947; Woldstedt, 1954; Axelrod, 1960; Kobaya-
shi and Shikama, 1961). Based on this, it is possible
that the common ancestor of groups 1 and 2 expanded
its distribution through the circumpolar region.
Extant Coriaria in Eurasia are scattered in two
areas: the eastern and southeastern part of Asia,
including the Himalayas, and the Mediterranean. In
the phylogenetic tree (Fig. 2), C. myrtifolia of the
Mediterranean is located in a possibly more derived
position than the eastern species, suggesting that the
common ancestor of group 1 expanded the distribution
from east to west in Eurasia. The present disjunct
distribution should be considered to have been caused
by the climatic changes associated with the glaciation
and the drying in the Cenozoic (Good, 1930).
The common ancestor of group 2 distributed in North
America most likely diverged into two lineages. The
descendent of one lineage evolved into C. microphylla
and expanded its distribution into northern South
America after the land bridge between North and
South America was formed. Another descendent mi-
grated to the Pacific islands and diverged into several
species in Papua New Guinea (C. papuana), Fiji (C.
sp.), and New Zealand (C. arborea, C. sarmentosa, C.
lurida and other species which we did not sample in
this study). Central America and the Pacific islands
were not connected to each other in the Cenozoic, and
long-distance dispersal should be postulated to explain
the pattern of this disjunct distribution. The phyloge-
netic tree indicates that C. ruscifolia is not a direct
descendent of either the postulated North American
ancestor or C. microphylla but migrated from one of the
Pacific islands to Chile. Although we examined represen-
tative species covering the disjunct distribution of
Coriaria, other species are known from New Zealand,
Tahiti, and other Pacific islands. Nucleotide sequence
18 YOKOYAMA ET AL.
data of these taxa may help to infer more detailed
migration routes of Coriaria in the Southern Hemi-
sphere.
ACKNOWLEDGMENTS
We thank H. Ohba for his generosity in providing his unpublished
distribution map of Coriaria; K. Oginuma, H. Ohba, M. Takahashi,
H. Tobe, and K. Uemura for helpful discussions and critical reading of
the manuscript; H. Setoguchi and R. Watanabe for living materials;
K. Sato for cultivation; and T. Fukuda, K. Kobayashi, T. Omori, M.
Sato, and A. Yokoyama for assistance with sequencing. This work was
supported in part by Grants from the Ministry of Education, Science
and Culture, Japan (08554035 to M.H. and K.I.; 0845273 to M.S.).
REFERENCES
Airy Shaw, H. K. (1973). ‘‘J. C. Willis’ A Dictionary of the Flowering
Plants and Ferns,’’ 8th ed., Cambridge Univ. Press, Cambridge,
UK.
Albert, V. A., Chase, M. W., and Mishler, B. D. (1993). Character-state
weighting for cladistic analysis of protein-coding DNA sequences.
Ann. Missouri Bot. Gard. 80: 752–766.
Allan, H. H. (1961). Coriariaceae. In ‘‘Flora of New Zealand,’’ (L. B.
Moore, Ed.), Vol. 1, pp. 300–305. Government printer, Wellington,
New Zealand.
Axelrod, D. I. (1960). The evolution of flowering plants. In ‘‘The
Evolution of Life’’ (S. Tax, Ed.), pp. 227–305. Univ. of Chicago Press,
Chicago.
Bohm, B. A., and Ornduff, R. (1981). Leaf flavonoids and ordinal
affinities of Coriariaceae. Syst. Bot. 6: 15–26.
Bousquet, J., Strauss, S. H., Doerksen, A. H., and Price, R. A. (1992).
Extensive variation in evolutionary rate of rbcL gene sequences
among seed plants. Proc. Natl. Acad. Sci. USA 89: 7844–7848.
Bremer, K. (1988). The limits of amino acid sequence data in
angiosperm phylogenetic reconstruction. Evolution 42: 795–803.
Briggs, J. C. (1987). ‘‘Biogeography and Plate Tectonics,’’ Elsevier,
Amsterdam.
Chase, M. W., Soltis, D. E., Olmstead, R. G., Morgan, D., Les, D. H.,
Mishler, B. D., Duvall, M. R., Price, R. A., Hills, H. G., Qiu, Y.-L.,
Kron, K. A., Rettig, J. H., Conti, E., Palmer, J. D., Manhart, J. R.,
Sytsma, K. J., Michaels, H. J., Kress, W. J., Karol, K. G., Clark,
W. D., Hedroen, M., Gaut, B. S., Jansen, R. K., Kim, K.-J., Wimpee,
C. F., Smith, J. F., Furnier, G. R., Strauss, S. H., Xiang, Q.-Y.,
Plunkett, G. M., Soltis, P. S., Swensen, S. M., Williams, S. E.,
Gadek, P. A., Quinn, C. J., Eguiarte, L. E., Golenberg, E., Learn,
G. H., Jr., Graham, S. W., Barrett, S. C., Dayanandan, S., and
Albert, V. A. (1993). Phylogenetics of seed plants: An analysis of
nucleotide sequences from the plastid gene rbcL. Ann. Missouri
Bot. Gard. 80: 528–580.
Chaney, R. W. (1947). Tertiary centers and migration routes. Ecol.
Monogr. 17: 140–148.
Corner, E. J. H. (1976). ‘‘The Seeds of Dicotyledons,’’ 2 Vols., Cam-
bridge Univ. Press, Cambridge, UK.
Croizat, L. (1952). ‘‘Manual of Phytogeography: An Account of Plant
Dispersal throughout the World,’’ Junk, The Hague.
Cronquist, A. (1981). Coriariaceae. In ‘‘An Integrated System of
Classification of Flowering Plants,’’ pp. 136–139. Columbia Univ.
Press, New York.
Dahlgren, R. M. T. (1980). A revised system of classification of the
angiosperms. Bot. J. Linn. Soc. 80: 91–124.
Felsenstein, J. (1981). Evolutionary trees from DNA sequences: A
maximum likelihood approach. J. Mol. Evol. 17: 368–376.
Felsenstein, J. (1985). Confidence limits on phylogenies: An approach
using the bootstrap. Evolution 39: 783–791.
Felsenstein, J. (1995). PHYLIP: Phylogeny Inference Package, ver-
sion 3.572c, Univ. of Washington, Seattle.
Garg, M. (1981). Pollen morphology and systematic position of
Coriaria. Phytomorphology 30: 5–10.
Good, R. D. O. (1930). The geography of the genus Coriaria. New
Phytol. 29: 170–198.
Gregor, H.-J. (1980). Seeds of the genus Coriaria Linne (Coriari-
aceae) in the European Neogene. Tert. Res. 3: 61–69.
Hasebe, M., and Iwatsuki, K. (1990). Adiantum capillus-veneris
chloroplast DNA clone bank: Useful heterologous probes in the
systematics of the leptosporangiate ferns. Am. Fern J. 80: 20–25.
Hasebe, M., Omori, T., Nakazawa, M., Sano, T., Kato, M., and
Iwatsuki, K. (1994). RbcL gene sequences provide evidence for the
evolutionary lineages of leptosporangiate ferns. Proc. Natl. Acad.
Sci. USA 91: 5730–5734.
Hasebe, M., Ando, T., and Iwatsuki, K. (1998). Intrageneric relation-
ships of maple trees based on the chloroplast DNA restriction
fragment length polymorphisms. J. Plant Res. 111: 441–451.
Hotta, M. (1974). ‘‘History and Geography of Plants,’’ Sanseido,
Tokyo.
Hughes, N. F. (1994). ‘‘The Enigma of Angiosperm Origins,’’ Cam-
bridge Univ. Press, Cambridge, UK.
Ina, Y. (1995). New methods for estimating the numbers of synony-
mous and nonsynonymous substitutions. J. Mol. Evol. 40: 190–226.
Johnson, L. A., and Soltis, D. E. (1995). Phylogenetic inference in
Saxifragaceae sensu stricto and Gilia (Polemoniaceae) using matK
sequences. Ann. Missouri Bot. Gard. 82: 149–175.
Kimura, M. (1980). A simple method for estimating evolutionary rate
of base substitutions through comparative studies of nucleotide
sequences. J. Mol. Evol. 16: 110–120.
Kobayashi, T., and Shikama, T. (1961). The climatic history of the Far
East. In ‘‘Descriptive Palaeoclimatology’’ (A. E. M. Nairn, Ed.), pp.
292–306. Interscience, New York.
Li, W.-H. (1993). So, what about the molecular clock hypothesis?
Curr. Opin. Genet. Dev. 3: 896–901.
Maddison, D. R. (1991). Discovery and importance of multiple islands
of most-parsimonious trees. Syst. Zool. 40: 315–328.
Melville, R. (1966). Continental drift, Mesozoic continents and the
migrations of the angiosperms. Nature 211: 116–120.
Melville, R. (1981). Vicarious plant distributions and paleogeography
of the pacific region. In ‘‘Vicariance Biogeography’’ (G. Nelson and
D. E. Rosen, Eds), pp. 413–435. Columbia Univ. Press, New York.
Mickevich, M. F., and Farris, J. S. (1981). The implications of
congruence in Menidia. Syst. Zool. 30: 351–370.
Muller, J. (1981). Fossil pollen records of extant angiosperms. Bot.
Rev. 47: 1–142.
Olsen, G., Matsuda, H., Hagstrom, R., and Overbeek, R. (1994).
fastDNAml: A tool for construction of phylogenetic trees of DNA
sequences using maximum likelihood. CABIOS 10: 41–48.
Ooi, K., Endo, Y., Yokoyama, J., and Murakami, N. (1995). Useful
primer designs to amplify DNA fragments of the plastid gene matK
from angiosperm plants. J. Jap. Bot. 70: 328–331.
Ozaki, K. (1991). ‘‘Late Miocene and Pliocene Floras in Central
Honshu,’’ Kanagawa Pref. Mus., Yokohama.
Praglowski, J. (1970). Coriariaceae. In ‘‘World Pollen Flora’’ (G.
Erdtman, Ed.), pp. 15–22. Scandinavian Univ. Books, Copenhagen.
Saitou, N., and Nei, M. (1987). The neighbor-joining method: A new
method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:
406–425.
Saporta, G. de. (1865). Etudes sur la vegetation du Sud-Est de la
 MOLECULAR PHYLOGENY OF Coriaria
France a l’epoque tertiare, 2 (3). Ann. Sci. Nat. Bot., Ser. 5 3:
212–217.
Scholz, H. (1964). Sapindales. In ‘‘A. Engler’s Syllabus der Pflanzenfami-
lien’’ (H. Melchior, Ed.), Vol. 2, pp. 277–288. Borntraeger, Berlin.
Schuster, R. M. (1976). Plate tectonics and its bearing on the
geographical origin and dispersal of angiosperms. In ‘‘Origin and
Early Evolution of Angiosperms’’ (C. B. Beck, Ed.), pp. 48–138.
Columbia Univ. Press, New York.
Sharma, V. K. (1968). Floral morphology, anatomy, and embryology of
Coriaria nepalensis Wall. with a discussion on the interrelation-
ships of the family Coriariaceae. Phytomorphology 18: 143–153.
Shinozaki, K., Ohme, M., Tanaka, M., Wakasugi, T., Hayashida, N.,
Matsubayashi, T., Zaita, N., Chunwongse, J., Obokata, J., Yamagu-
chi-Shinozaki, K., Ohta, C., Torazawa, K., Meng, B. Y., Sugita, M.,
Deno, H., Kamagashira, T., Yamada, K., Kusuda, J., Takaiwa, F.,
Kato, A., Tohdoh, N., Shimada, H., and Sugiura, M. (1986). The
complete nucleotide sequence of the tobacco chloroplast genome:
Its gene organization and expression. EMBO J. 5: 2043–2049.
Skog, L. E. (1972). The genus Coriaria (Coriariaceae) in the Western
Hemisphere. Rhodora 74: 242–253.
Smith, A. G., and Briden, J. C. (1977). ‘‘Mesozoic and Cenozoic
Paleocontinental Maps,’’ Cambridge Univ. Press, Cambridge, UK.
Steele, K. P., and Vilgalys, R. (1994). Phylogenetic analyses of
Polemoniaceae using nucleotide sequences of the plastid gene
matK. Syst. Bot. 19: 126–142.
19
Sugita, M., Shinozaki, K., and Sugiura, M. (1985). Tobacco chloro-
plast tRNA Lys(UUU) gene contains a 2.5-kilo base-pair intron: An
open reading frame and a conserved boundary sequence in the
intron. Proc. Natl. Acad. Sci. USA 82: 3557–3561.
Swensen, S. M., Mullin, B. C., and Chase, M. W. (1994). Phylogenetic
affinities of Datiscaceae based on an analysis of nucleotide se-
quences from the plastid rbcL gene. Syst. Bot. 19: 157–168.
Swensen, S. M. (1996). The evolution of actinorhizal symbioses:
Evidence for multiple origins of the symbiotic association. Am. J.
Bot. 83: 1503–1512.
Swofford, D. L. (1993). ‘‘Phylogenetic analysis using parsimony,
version 3.1.1,’’ Illinois Natural History Survey, Champaign, IL.
Takhtajan, A. L. (1980). Outline of the classification of flowering
plants (Magnoliophyta). Bot. Rev. 46: 225–359.
Thorne, R. F. (1983). Proposed new realignments in the angiosperms.
Nord. J. Bot. 3: 85–117.
Woldstedt, P. (1954). ‘‘Die Allgemeinen Erscheinungen des Eiszeital-
ters.’’ Ferdinand Enke, Stuttgart.
Wolfe, K. H., Gouy, M., Yang, Y.-W., Sharp, P. M., and Li, W.-H. (1989).
Date of the monocot–dicot divergence estimated from chloroplast
DNA sequence data. Proc. Natl. Acad. Sci. USA 86: 6201–6205.
Wu, C.-I., and Li, W.-H. (1985). Evidence for higher rates of nucleotide
substitution in rodents than in man. Proc. Natl. Acad. Sci. USA 82:
1741–1745.