Kusaka et al.: Genetic population structure of Dipturus chinensis around Japan 105 
Table 3 
Pairwise values of molecular genetic diversity among populations (gp and Rg_p) estimated from mitochondrial cyto- 
chrome b gene (931 base pairs) (above the diagonal) and 4 microsatellite loci (below the diagonal), respectively, among 6 
populations of the polkadot skate (Dipturus chinensis) sampled in the East China Sea, Sea of Japan, and Pacific Ocean 
during 2010-2017. An asterisk (*) indicates that a value is significant after a sequential Bonferroni correction (P<0.001). 
Aomori 
Prefecture 
Kyoto Niigata Kochi 
Population location Danjo Islands Goto Islands Prefecture Prefecture Prefecture 
0.0269 
—0.0201 
0.0114 
0.8847* 
0.9342* 
0.9147* 
0.9235* 
0.8924* 
0.9234* 
0.9156* 
0.9237* 
0.8507* 
—0.0030 0.0357 
0.0037 
Danjo Islands 
Goto Islands 
Kyoto Prefecture 
Niigata Prefecture 
Kochi Prefecture 
Aomori Prefecture 
0.0002 
0.0029 
0.0093 
0.0806* 
0.1604* 
0.0194 
0.0277 
0.1060* 
0.1644* 
0.0031 
0.0800* 
0.1908* 
0.1325* 
0.2106* 0.1949* 
(in the Danjo and Goto Islands) and in the Sea of Japan 
(in the Kyoto and Niigata Prefectures) were high (pairwise 
Rg: 0.0800—-0.2106) and significant after a Bonferroni 
correction (P<0.001). The Rgp value between the popula- 
tions in the Kochi and Aomori Prefectures was also high 
(pairwise Rg: 0.1949) and significant after a Bonferroni 
correction (P<0.001). 
Discussion 
Genetic variation 
The high A and low x (Table 2) in mt cyt b observed in 
each geographic population, except for the one in Kochi 
Prefecture, correspond to the results from previous stud- 
ies in which the same genetic region as the one used for 
polkadot skate in our study was used for thornback skate 
(Raja clavata) in Portugal and the southwestern part of 
the North Sea (h=0.510-—0.720, n=0.006—0.007; Chevolot 
et al., 2006b) and for thorny skate in Newfoundland, Ice- 
land, the northeastern part of the North Sea, and Kattegat 
(h=0.719-0.916, t=0.0030—0.0071; Chevolot et al., 2007). 
In contrast, low h and low zx observed in the population in 
the Kochi Prefecture are similar to the results for thorn- 
back skate in part of the Azores Islands and in the Med- 
iterranean and Black Seas (h=0.000, z=0.000; Chevolot 
et al., 2006b). The mean Ap and Hy, for each population 
from the SSR analysis in our study are higher than val- 
ues reported by Griffiths et al. (2010) for the blue skate 
(D. batis) (Ap=3.65, Hp=0.349) and the flapper skate (D. 
intermedius), which was treated as the northern clade of 
blue skate (Ap=3.35, H,=0.366), around the British Isles; 
Griffiths et al. (2010) used almost the same SSR loci as 
those used in our study. Among the 6 geographic popula- 
tions in our study, the population in the Kochi Prefecture 
had the lowest Ap, and Hy was relatively lower for this 
population than for the others. Although factors affecting 
low genetic diversity of the population in Kochi Prefec- 
ture are unclear, the low diversity may be related to the 
demographic history of the population (see the “Evolution- 
ary history” section). 
Genetic differentiation among populations 
The maximum likelihood tree (Fig. 2) and the minimum 
spanning network (Fig. 3) based on the haplotypes of mt cyt 
b, reveal 3 lineages, clades A, B1, and B2. Clade A consists 
of populations in the East China Sea (populations in the 
Danjo and Goto Islands and 1 individual from the Koshiki- 
jima Islands) and in the Sea of Japan (populations in the 
Kyoto and Niigata Prefectures). The haplotypes observed 
for populations in Taiwan and off the Korean Peninsula 
were also included in clade A; however, there was only 1 
sample each from those locations. Addition of more sam- 
ples in analysis is needed to improve understanding of 
the genetic relationships of those geographic populations 
to the populations around Japan. Clade B1 is the popu- 
lation off the southern Pacific coast of Japan (in Kochi 
Prefecture), and clade B2 was made up of the population 
from the northern Pacific coast (in Aomori Prefecture). In 
both mt cyt 6b and SSR analyses, the 6 geographic popula- 
tions were separated into 3 groups, corresponding to the 
3 clades, by using hierarchical AMOVA analyses that 
maximized ®o7 and Rez values, respectively, although the 
values were barely not significant (Suppl. Table 1) (online 
only). Genetic differentiation among the 3 groups are also 
indicated by both the pairwise ®g7 and Rep values from mt 
cyt b and SSR analyses, respectively (Table 3). 
Our results indicate that the population structure of the 
polkadot skate is influenced by the complicated oceanic 
currents that surround Japan (Fig. 1). The clades A, B1, 
and B2 around Japan can be regarded as the Tsushima, 
Kuroshio, and Oyashio lineages, respectively. Similar 
genetic differentiations between the populations in the Sea 
of Japan and along the Pacific coasts have been reported 
for ocellate spot skate (Misawa et al., 2019b), in addition to 
reports for several coastal fish and invertebrate species 
(Kojima et al., 1997, 2004; Akihito et al., 2008; Katafuchi 
et al., 2011; Kokita and Nohara, 2011; Hirase et al., 2012; 
