Short communication
Population genetic structure of the kuruma prawn (Penaeus
japonicus) in East Asia inferred from mitochondrial
DNA sequences
Tzong-Der Tzeng, Shean-Yeh Yeh, and Cho-Fat Hui
Tzeng, T-D., Yeh, S-Y., and Hui, C-F. 2004. Population genetic structure of the kuruma prawn (Penaeus japonicus) in East Asia inferred from mitochondrial DNA sequences. e ICES Journal of Marine Science, 61: 913e920.
Sequence analyses on the complete mitochondrial DNA (mtDNA) control region (992 bp) were conducted to elucidate the population structure of kuruma prawns (Penaeus japonicus) in East Asia. Five populations including 95 individuals were collected. They are separated into the Japan Sea (JS), the north and south of the East China Sea (NECS and SECS), the Taiwan Strait (TS), and the north of the South China Sea (NSCS) populations. There are 292 variable sites without any insertions and deletions. Nucleotide diversity in the total populations is 2.51 G 0.07%, and the variations within populations ranged from 2.61 G 0.93% (SECS) to 2.29 G 0.16% (JS). FSTvalues between the JS and the rest of the populations, between the NECS and NSCS populations, and between the SECS and NSCS populations show significant differences. The UPGMA tree of these five populations shows three distinct clusters; one includes the JS population; another includes the NECS population; the third includes populations from the rest of the areas. The analysis of molecular variance (AMOVA) shows clear genetic difference between the JS and the rest of the populations. Additional AMOVA analysis excluding the JS population indicates significant variation between the NECS population and the other three populations. We, therefore, conclude that three distinct populations exist in East Asia; one is in the JS; another is in the NECS; and the third is distributed in SECS, TS and NSCS.
Ó 2004 International Council for the Exploration of the Sea. Published by Elsevier Ltd. All rights reserved. Keywords: control region, East Asia, mtDNA, Penaeus japonicus, population structure. Received 10 October 2003; accepted 15 June 2004.
T-D. Tzeng: College of Liberal Education, Shu-Te University, No. 59, Hun Shan Rd., Hun Shan Village, Yen Chau, Kaohsiung County 824, Taiwan, R.O.C. S-Y. Yeh: Institute of Oceanography, National Taiwan University, PO Box 23-13, Taipei 106, Taiwan, R.O.C. C-F. Hui: Institute of Zoology, Academia Sinica, Taipei 115, Taiwan, R.O.C. Correspondence to T-D. Tzeng: tel: C886 7 6158000 ext. 4211; fax: C886 7 6158000 ext. 4299; e-mail:tdtzeng@mail.stu.edu.tw
Introduction
A concrete understanding of population genetic structure is of primary importance for the management and conserva-tion of genetic resources in exploited marine organisms (Hillis et al., 1996). Studies on population genetic structure of marine biota have frequently indicated that organisms with high dispersal capacity would have little genetic distinction over large geographic scales (Hellberg, 1996). These studies suggest that there are high levels of gene flow between marine populations. However, there is growing evidence that widespread marine organisms are more genetically structured than expected given their high dispersal potential and apparent lack of barriers to dispersal in the ocean (Palumbi, 1997; Benzie, 1999; Briggs, 1999).
Thus, there may be limits to the actual dispersal of marine organisms with high dispersal potential (Benzie and Williams, 1997). These limits vary widely with species, habitats, local ocean conditions, or historical events, and they may produce sufficient chances for genetic variation (Palumbi, 1994).
The kuruma prawn (Penaeus japonicus) is a widely distributed species throughout the Indo-west Pacific, ranging from eastern and southern Africa into the Red Sea through the entire Malay Archipelago to Taiwan, Korea, Japan, and Northern Australia, and they have moved through the Suez Canal into the Mediterranean (Hayashi, 1992). This species is one of the most important fishery animals in East Asia ( particularly the East China Sea and the Taiwan Strait). The life history of the kuruma prawn,
1054-3139/$30.00 Ó 2004 International Council for the Exploration of the Sea. Published by Elsevier Ltd. All rights reserved.
ICES Journal of Marine Science, 61: 913e920 (2004) doi:10.1016/j.icesjms.2004.06.015
comprising an offshore planktonic larval phase, an estuarine post-larval and juvenile phase, and an offshore adult and spawning phase (Dall et al., 1990), may allow moderate gene flow among populations.
Two morphologically distinguishable populations of kuruma prawn in the East China Sea and Taiwan Strait were discriminated byTzeng and Yeh (1999). However, the variation of morphological characters could be affected by genetic and environmental factors, so discrimination of populations based on morphological variation must be verified by genetic evidence to confirm that the variation reflects the true degree of reproductive isolation rather than environmental isolation (Pepin and Carr, 1992).
Mitochondrial DNA (mtDNA) has many attributes that make it particularly suitable for population genetic studies, including its rapid rate of evolution, lack of recombination, and maternal inheritance (Hoelzel et al., 1991). Since the control region of the mtDNA has been shown to be the most variable region in both vertebrates and invertebrates, this region is an ideal marker for characterizing geographical patterns of genetic variation within and between prawn populations (Simon, 1991). In this paper, we amplified and sequenced the complete control region to elucidate the population genetic structure of the kuruma prawn in marginal seas of East Asia.
Material and methods
Collection of samples
Five putative kuruma prawn populations were collected during 1995 and 1996 (Figure 1and Table 1). Specimens were iced or frozen immediately after capture and later kept
at75(C until extracted, but specimens from the Japan Sea were stored in 95% ethanol. Total genomic DNA was extracted from muscle tissue by using a standard DNA extraction technique with proteinase K digestion followed by phenol/chloroform purification.
Amplification and sequencing
A fragment of mtDNA was first amplified and sequenced for each specimen using primers P120-90 (50
-GATCTT-TAGGGGAATGGTGTAATTCCATTG-30) and P14586-609 (50-GTGTCTTCTTGAAGTCTG-30). Then the primers P15764-47 (50-GATAGCTTAAAGGTTTAACTAC-30),
P15481-60 (50-GAGTCTTTAACTTTTAATGACCCC-30), and P14857-80 (50
-GTGTAACAGGGTATCTAATCCT-GG-30) were designed according to the nucleotide sequen-ces of kuruma prawn obtained with primers P120-90 and P14586-609 and used for sequencing. Primer names indicate homologous positions on the P. monodon mito-chondrial genome (Wilson et al., 2000). The PCR protocol consisted of 39 cycles of denaturing at 95(C for 50 s, annealing at 50(C for 1 min, and extension at 72(C for 1.5 min.
Dideoxy chain-termination DNA sequencing was per-formed (Sanger et al., 1977) using Sequencing kit (version 2.0, United States Biochemical) with [a-35S]-dATP as label. The DNA sequencing reaction was carried out using the following cycling parameters: 29 cycles of denaturation at 95(C for 40 s, annealing at 50(C for 40 s, and extension at 72(C for 30 s. The sequencing reaction products were electrophoresed in a 6% polyacrylamide/7 M urea gel. The gel was fixed, dried, and visualized by autoradiography on a Kodak film for 24e72 h.
Sequence analysis
DNA sequences were aligned by using the PILEUP program in GCG (Genetics Computer Group, version 7.0;
Devereux et al., 1991). The beginning and end of the control region were confirmed by comparing with the published sequence of P. monodon (Wilson et al., 2000). Subsequent analyses were based on the complete control region sequence obtained from 95 individuals. The number of transitions (TS) and transversions (TV), and nucleotide diversities p (Nei, 1987) within populations were calcu-lated by using ARLEQUIN 2.000 (Schneider et al., 2000). To examine whether two of the populations are genetically different from each other, the FST statistic
(Wright, 1965) between five populations was estimated and tested using the program ProSeq (Filatov, 2002). The statistical significance of the estimate was tested through 1000 permutations. The dendrogram of five populations was constructed using the unweighted pair-group method with arithmetic means (UPGMA) based on the FSTvalues.
Gene flow (Nm), was estimated using the relationship
NmZ ((1/FST) 1)/2 (Hudson et al., 1992).
Figure 1. Five shaded areas showing the sampling locations of kuruma prawn in the Japan Sea (JS), the north (NECS) and south (SECS) of the East China Sea, the Taiwan Strait (TS), and the north of the South China Sea (NSCS).
Relationships between haplotypes were determined with the Kimura two-parameter distance model by using the Neighbour-Joining method in MEGA 2.1 (Kumar et al., 2001). Nucleotide diversities p (Nei, 1987) within populations were estimated by using ARLEQUIN.
Analyses of molecular variance (AMOVA) in ARLE-QUIN were performed to test the geographic divisions among populations. This approach is a hierarchical ap-proach analogous to analysis of variance (ANOVA) in which the correlations among haplotypes at various hierarchical levels are used as F-statistic analogs, designat-ed as F-statistic. AMOVA computes the proportion of variation among groups (FCT), the proportion of variation
among populations within groups (FSC), and the proportion
of variation within populations (FST). The significance of
these F-statistic analogs is evaluated by random permuta-tions of sequences among populapermuta-tions. We experimented with various groupings of populations suggested by population trees, FST values, and ocean division. The
groupings that maximize values of FCT and are
signifi-cantly different from random distributions of individuals are assumed to be the most probable geographic subdivisions. Isolation by distance was assessed by a Mantel test in NTSYS (Rohlf, 1997). We used the pair-wise FSTvalues
and the corresponding pair-wise geographical distances as the input data and 1000 permutations were performed to determine the level of significance. The approximate geographic distances between sampling localities were taken as the minimum distance map.
Results
The control region of kuruma prawn mtDNA contains 992 bp. Among 95 haplotypes identified from 95 individual mtDNA sequenced, there are 292 variable sites without any insertions or deletions (Figure 2). The TS:TV rate is 2.34:1. Nucleotide diversity (p) in the total populations is 2.51 G 0.07% but it varies within populations, ranging from 2.61 G 0.93% in the SECS population to 2.29 G 0.16% in the JS population (Table 1).
The FSTand Nmvalues are shown inTable 2. The FST
value across all populations shows a significant amount of genetic variation between five populations (FSTZ 0.0434, p ! 0.01). FST values between the JS and the other
populations, between the NECS and NSCS populations, and between the SECS and NSCS populations show significant genetic differences, but genetic variation be-tween the other populations is not significant. The Nm
values between all pairs of the five populations range from 0.0840 (JSeSECS populations) to 59.6404 (SECSeTS populations). Nmvalues between the JS and the rest of the
four populations are relatively lower ( from 0.0840 to 3.7203). The higher Nm values are found between the
NECS and SECS populations (12.2787), and between the NECS and TS populations (12.1297).
The neighbour-joining tree of the 95 haplotypes shows little genealogical structuring and is characterized by shallow branching and nodes not well supported by bootstrap values (not shown).
The UPGMA tree of five sampling areas is shown in
Figure 3. These five populations are clustered into two distinct groups; the first group includes the JS population; the second one includes the other four populations. The second group may be further divided into two subgroups; the first subgroup includes the NECS population; the second subgroup includes the SECS, TS, and NSCS populations.
The results of AMOVA for total populations are shown inTable 3. The AMOVA for five populations yields a small but significant FSTvalue of 0.0362, indicating that at least
one of the pair-wise populations reveals significant heterogeneity. Significant values of FCT are observed in
four of nine groupings. In four significant groupings, the division occurred in the JS population. The highest FCT
value (0.0542) is observed when the JS population and the rest are separately assigned into two groups (Table 3and grouping 2). Significant FCT values are also found in
groupings 3 and 4, indicating that an additional genetic discontinuity may also have occurred in the NECS population.
Table 1. Sampling locality, sampling date, sample code, sample size, and nucleotide diversity (p) with standard deviation (s.d.) in five kuruma prawn populations in East Asia.
Locality Sample code Sampling date Sampling size p Gs.d. (%)
Japan Sea (37(N, 137(E) JS May 1996 14 2.29 G 0.16
North of the East
China Sea (30(e31(N, 123(e124(E)
NECS Dec. 1995 12 2.54 G 0.20
South of the East
China Sea (26(e27(N, 122(e123(E)
SECS Dec. 1995 23 2.61 G 0.93
Taiwan Strait (24(e24(150N, 119(360
e120(E) TS Dec. 1995 24 2.32 G 0.84
North of the South
China Sea (21(e21(300N, 117(300 e118(E)
NSCS Dec. 1995 22 2.43 G 0.11
Total 95 2.51 G 0.07
915 Population genetic structure of the kuruma prawn
Figure 2. Variable sites in the 95 haplotypes found in the control region (992 bp) of 95 kuruma prawn individuals from five sampling localities. The numbers above the sequences correspond to the positions of the polymorphic sites. Dots indicate an identical nucleotide at the position relative to haplotype JS01.
T-D.
Tzeng
et
Figure 2. (co ntinued ) 917 Population genetic structure of the kuruma prawn
Table 2. FST (above diagonal) and Nm(below diagonal) values between five kuruma prawn populations in East Asia. Abbrevia-tions for populaAbbrevia-tions are defined in Table 1.
JS NECS SECS TS NSCS JS 0.0845** 0.0735** 0.0840** 0.0630** NECS 2.71 0.0200 n.s. 0.0202 n.s. 0.0457** SECS 0.0840 12.2787 0.0042 n.s. 0.0024* TS 2.7254 12.1297 59.6408 0.0093 n.s. NSCS 3.7203 4.9285 9.9741 26.5970 *p ! 0.05, **p ! 0.01, n.s. Z not significant (p O 0.05).
Figure 3. UPGMA tree showing relationships among five pop-ulations.
Table 3. The results of AMOVA for all five populations. Abbreviations for populations are defined in Table 1.
Groupings Variance component % Total variance F-statistics p
For all populations
1 Group 1 {JS, NECS, SECS, TS, NSCS} AG 3.62 FSTZ 0.0362 !0.0001
Based on UPGMA tree of five populations
2 Group 1 {JS} AG 5.43 FCTZ 0.0542 !0.0001
Group 2 {NECS, SECS, TS, NSCS}
3 Group 1 {JS} AG 4.38 FCTZ 0.0437 !0.0001 Group 2 {NECS} Group 3 {SECS, TS, NSCS} 4 Group 1 {JS} AG 3.94 FCTZ 0.0394 !0.0001 Group 2 {NECS} Group 3 {SECS, TS} Group 4 {NSCS} Based on the significance of FSTs
5 Group 1 {JS} AG 3.55 FCTZ 0.0354 Z0.0517
Group 2 {NECS, SECS, TS} Group 3 {NSCS}
6 Group 1 {JS} AG 2.71 FCTZ 0.0270 !0.0001
Group 2 {NECS, SECS} Group 3 {TS, NSCS} Based on ocean divisions and others
7 Group 1 {JS} AG 1.38 FCTZ 0.0138 !0.3994
Group 2 {NECS, SECS} Group 3 {TS}
Group 4 {NSCS}
8 Group 1{JS, NECS} AG 1.24 FCTZ 0.0124 Z0.0982
Group 2{SECS, TS, NSCS}
9 Group 1{JS, NECS, SECS} AG 0.36 FCTZ 0.0035 Z0.5958
Group 2{TS, NSCS}
AG is the among-groups component of variance. The best groupings have maximal values of AG. Note that the F-statistic estimators in the AMOVA are random variables and can take either positive or negative values, negative values indicating excess of heterozygotes. Such negative estimates should be interpreted as zero in the AMOVA (Schneider et al., 2000).
Additional AMOVA excluding the JS population was also performed. The results are shown inTable 4. The AMOVA for all four populations yields a significant FSTvalue of
0.0178, indicating significant genetic division among these four populations. Significant values of FCTare observed in
two of the five groupings. In two significant groupings, the division occurs in the NECS population, and indicates that the NECS population is an independent population. Genetic differentiation between populations is observed to be positively correlated with the distance of geographical separation between populations, which indicates that kuruma prawns conform to an isolation-by-distance model of maternal gene flow (r Z 0.83693, p Z 0.037).
Discussion
Although the neighbour-joining tree of the 95 haplotypes reveals few genealogical branches or geographic clusters, the results of cluster analysis, sequence statistic (FST), and
AMOVA indicate significant genetic division between these five populations. The cluster analysis indicates that these five populations can be clustered into three groups. One includes the JS population, the second includes the NECS population, and the third includes the rest of the populations (Figure 3). FST values between the JS
population and the rest, between the NECS and NSCS populations, and between the SECS and NSCS populations show significant genetic differences (Table 2), indicating that at least three isolated populations exist in these waters. Results of the AMOVA reveal three different populations in marginal seas of East Asia (Tables 3 and 4). Based on the above analyses, kuruma prawns in East Asia can be
discriminated into three distinct populations. The first population is in the Japan Sea, the second in the north of the East China Sea, and the third in the south of the East China Sea, the Taiwan Strait, and the north of the South China Sea. The genetic divisions between the populations in SECS, TS, and NECS are in agreement with the previous morphological study byTzeng and Yeh (1999).
Kuruma prawns migrate from inshore to offshore as they grow, but the migratory distance is limited (Dall et al., 1990). Thus, the dispersal of larvae is the primary source of gene flow, and ocean currents play a major role in the dispersal of this species. The spawning season lasts from April until October in the NECS population (Yamada et al., 1986). Along the eastern coast of China, kuruma prawn larvae from the north of the East China Sea may be transported to the Taiwan Strait by the China coastal current. However, during the spawning season of the NECS population, the China coastal current is not strong enough to flow through the Taiwan Strait, and only spreads to the north of Taiwan and to the middle of the Taiwan Strait (Wu, 1982). Higher levels of gene flow between the NECS and SECS populations (NmZ 12.2787), and between the
NECS and TS populations (NmZ 12.1297) have been observed, but lower Nm(4.9285) between the NECS and
NSCS populations has been found (Table 2). This occurrence of kuruma prawn larvae mixed results in homo-geneity among the NECS, SECS, and TS populations, but may not be large enough to eliminate the genetic difference between the NECS and NSCS populations (Table 2).
In the NSCS and TS populations, the main spawning season is from late spring to summer. During the main spawning season, the South China Sea warm water domi-nates the Taiwan Strait (Wang and Chern, 1989) and
Table 4. The results of AMOVA for the NECS, SECS, TS, and NSCS populations. Abbreviations for populations are defined in Table 1.
Groupings Variance component % Total variance F-statistics p
1 Group 1 {NECS, SECS, TS, NSCS} AG 1.79 FSTZ 0.0178 Z0.0047
2 Group 1 {NECS} AG 1.92 FCTZ 0.0191 !0.0001
Group 2 {SECS, TS, NSCS}
3 Group 1 {NECS} AG 2.01 FCTZ 0.0200 !0.0001
Group 2 {SECS, TS} Group 3 {NSCS}
4 Group 1 {NECS, SECS, TS} AG 1.23 FCTZ 0.0123 !0.2442
Group 2 {NSCS}
5 Group 1 {NECS, SECS} AG 0.74 FCTZ 0.0073 Z0.6727
Group 2 {TS, NSCS}
6 Group 1 {NECS, SECS} AG 0.58 FCTZ 0.0058 !0.4977
Group 2 {TS} Group 3 {NSCS}
AG is the among-groups component of variance. The best groupings have maximal values of AG. Note that the F-statistic estimators in the AMOVA are random variables and can take either positive or negative values, negative values indicating excess of heterozygotes. Such negative estimates should be interpreted as zero in the AMOVA (Schneider et al., 2000).
919 Population genetic structure of the kuruma prawn
provides a northbound gene flow from the north of South China Sea to the south of East China Sea. Very high gene flow between the TS and NSCS populations (NmZ 26.5970) is observed and prevents population
differentiation between these two populations (Table 2). In general, if the Nmvalue is greater than about 5, the gene
flow is considered sufficient to maintain a relatively homogeneous gene pool (Slatkin, 1987). There is a high gene flow (NmZ 9.9741) between the NSCS and SECS populations, but significant genetic difference is still observed (Table 2). This genetic difference between the NSCS and SECS populations may result from the re-cruitment of larvae from the NECS population.
During the last glacial maximum (LGM), the sea level was 130e150 m lower than the present level in the East China Sea and 100e120 m lower in the South China Sea. Consequently, the entire Bohai gulf, the Yellow Sea, and the Tsushima and Taiwan Straits were exposed, and the East China Sea was reduced into an elongated trough during LGM (Wang and Sun, 1994). As the Tsushima Strait was exposed during LGM, the Japan Sea and East China Sea were completely separated from each other. Therefore, gene flow between the two waters was completely suspended, explaining why FSTvalues between the JS and the rest of the
populations are at least twice the ones obtained between other populations. The disappearance of habitat had re-stricted marine species to the relatively limited areas and caused the mixing among populations and reduced the genetic variation between populations (Benzie and Wil-liams, 1997). After LGM, the distribution of kuruma prawns gradually extended corresponding to the rise of the sea level of the East China Sea. The similar magnitude of nucleotide diversity in each population provides part of the evidence that four populations may share common ancestry (Table 1).
Acknowledgements
We thank Dr Y. Uozumi, National Research Institute of Far Sea Fishery, Dr T. Minami, and H. Mastusato, Japan Sea National Research Institute in Niigata City, for kindly providing kuruma prawn samples from the Japan Sea.
References
Benzie, J. A. H. 1999. Genetic structure of coral reef organisms e ghosts of dispersal past. American Zoologist, 39: 131e145. Benzie, J. A. H., and Williams, S. T. 1997. Genetic structure of
giant clam (Tridacna maxima) populations in the West Pacific is not consistent with dispersal by present-day ocean currents. Evolution, 51: 768e783.
Briggs, J. S. 1999. Coincident biogeographic patterns: Indo-West Pacific Ocean. Evolution, 53: 326e335.
Dall, W., Hill, B. J., Rothlisberg, P. C., and Sharples, D. J. 1990. The Biology of the Penaeidae. Academic Press, London. Devereux, J., Haeberli, P., and Marquess, P. 1991. Genetic
Computer Group Manual, Version 7.0. University of Wisconsin, Madison, WI.
Filatov, D. A. 2002. ProSeq: a software for preparation and evolutionary analysis of DNA sequence data sets. Molecular Ecology Note, 2: 621e624.
Hayashi, K. 1992. Dendrobranchiata crustaceans from Japanese waters. Seibutsu Kenkyusha, Tokyo.
Hellberg, M. E. 1996. Dependence of gene flow on geographic distance in two solitary corals with different larval dispersal capabilities. Evolution, 50: 1167e1175.
Hillis, D. M., Mable, B. K., and Moritz, C. 1996. Molecular Systematics. 2nd edn. Sinauer, Sunderland, Massachusetts. Hoelzel, A. R., Hancock, J. M., and Dover, G. A. 1991. Evolution
of the cetacean mitochondrial D-loop region. Molecular Biology and Evolution, 8: 475e493.
Hudson, R. R., Slatkin, M., and Maddison, W. P. 1992. Estimation of levels of gene flow from DNA sequence data. Genetics, 132: 583e589.
Kumar, S., Tamura, K., Jakobsen, I. B., and Nei, M. 2001. MEGA2: Molecular Evolutionary Genetics Analysis Software. Arizona State University, Tempe, Arizona, USA.
Nei, M. 1987. Molecular Evolutionary Genetics. Columbia University Press, New York.
Palumbi, S. R. 1994. Genetic divergence, reproductive isolation, and marine speciation. Annual Review in Ecology and Systematics, 25: 547e572.
Palumbi, S. R. 1997. Molecular biogeography of the Pacific. Coral Reefs, 16: 47e52.
Pepin, P., and Carr, S. M. 1992. Morphological, meristic, and genetic analysis of stock structure in juvenile Atlantic cod (Gadus morhua) from the Newfoundland shelf. Canadian Journal of Fisheries and Aquatic Sciences, 52: 1924e1933. Rohlf, F. J. 1997. NTSYS-pc, Version 2.02. Software. Applied
Biostatistics Inc., Exeter.
Sanger, F., Nicklen, S., and Coulson, A. R. 1977. DNA sequencing with chain-terminating inhibitors. Proceedings of the National Academy of Sciences of the USA, 74: 5463e5467.
Schneider, S., Roessli, D., and Excoffier, L. 2000. ARLEQUIN, Version 2.000. A Software for Population Genetics Data Analysis. Genetics and Biometry Laboratory, University of Geneva, Switzerland.
Simon, C. 1991. Molecular systematics at the species boundary: exploiting conserved and variable regions of the mitochondrial genome of animals via direct sequencing from amplified DNA. In Molecular Techniques in Taxonomy. Ed. by G. M. Hewitt, A. W. B. Johnston, and J. P. W. Young. Springer-Verlag, New York. Slatkin, M. 1987. Gene flow and the geographic structure of natural
populations. Science, 236: 787e792.
Tzeng, T-D., and Yeh, S-Y. 1999. Analysis of morphometric characters of the kuruma shrimp (Penaeus japonicus) in the East China Sea and the Taiwan Strait. Journal of Fisheries Society of Taiwan, 26: 203e212.
Wang, J., and Chern, C-S. 1989. On the Kuroshio branch in the Taiwan Strait during wintertime. Progress in Oceanography, 21: 469e491.
Wang, P., and Sun, X. 1994. Last glacial maximum in China: comparison between land and sea. Catena, 23: 341e353. Wilson, K., Cahill, V., Ballment, E., and Benzie, J. 2000. The
complete sequence of the mitochondrial genome of the crustacean Penaeus monodon: are malacostracan crustaceans more closely related to insects than to branchiopods? Molecular Biology and Evolution, 17: 863e874.
Wright, S. 1965. The interpretation of population structure by F-statistics with special regard to systems of mating. Evolution, 19: 395e420.
Wu, B. 1982. Some problems on circulation study in Taiwan Strait. Taiwan Strait, 1: 1e7.
Yamada, U., Tagawa, M., Kishida, S., and Honjo, K. 1986. Fishes of the East China Sea and the Yellow Sea. Contributions from the Seikai Regional Fisheries Research Laboratory, No. 422.
