Fig 1.
Map showing sampling sites for the sandy beach amphipod Haustorioides japonicus.
Empty circles and numerals show locations and ID numbers for sampling sites in the present study. Filled and colored circles indicate the population assignment of the sites in a previous study (Takada et al. 2018 [13]; see also Fig 2). Shaded areas indicate the coastlines during the last glacial maximum (–120 m). Map insert shows previously reported sites with H. japonicus (stars, Kamihira 2000 [16]; circles, Takada et al. 2015 [15]). The map was created with QGIS v2.18.0 (http://www.qgis.org) using layers freely available at Natural Earth (https://www.naturalearthdata.com/downloads/10m-physical-vectors/).
Table 1.
Sampling location, sample size (n), coordinates, collection date, and reference for sandy beach amphipod Haustorioides japonicus specimens used in this study.
Fig 2.
Minimum spanning tree of mitochondrial COI haplotypes from the northwestern Pacific (NWP) population of the sandy beach amphipod Haustorioides japonicus (A), and interrelationships among and geographic distributions of H. japonicus populations (B; this study and Takada et al. [13]). Branch length and circle size are proportional to the T92 + Γ distance and haplotype frequency, respectively. The haplotype ID is beside each node (see also S1 Table). ECS, East China Sea; SJS, southern Sea of Japan; SJC, central Sea of Japan; NWP, northwestern Pacific; SJN, Northern Sea of Japan.
Table 2.
Genetic diversity and neutrality indices for two group of populations of the sandy beach amphipod Haustorioides japonicus.
Fig 3.
Demographic timeline derived from the best two-epoch demographic model (model 3, Table 3).
An ancient epoch with a constant population size (θ1) is followed by the modern epoch of logistic population growth (θ2). The median estimate of the transition time is shown as a vertical red line. Horizontal and vertical blue lines and the blue shaded area show the median estimate for mutation-scaled female effective population size (θ), the time of the most recent common ancestor, and the 95% credible interval, respectively. The histogram shows the posterior density distribution for the two-epoch transition time.
Table 3.
Listed are the model ID, type of two-epoch model (TEM) and clock model, transition time (TT, 10−3 mutations/site), median marginal likelihood (MML), and 2×log Bayes factors (2×lnBF) for the models. Parameters for the best-fit model are shown in bold type. The result for the exponential TEM with strict clock is not shown because the run did not converged.
Fig 4.
Bayesian skyline plots based on the calibration of demographic transition (CDT) rate and a conventional mitochondrial evolutionary rate of 0.7%.
Median estimates of female effective population size are shown as lines and vertical lines show median estimates of the time of the most recent common ancestor. Global trends of historical temperature are superimposed on the plots (Jouzel et al. [31]).
Fig 5.
Posterior density distribution histograms of the split-time between Tohoku and other NWP individuals estimated by using CDT rate (A) and a conventional mitochondrial evolutionary rate of 0.7% (B), and global trends of historical temperature from Jouzel et al. [31] (C). Maximum clade credibility trees are also shown for each rate.