To reveal phylogeographical features of the sable Martes zibellina on Hokkaido Island, northern Japan, we analyzed the 5′-portion sequences of the mitochondrial DNA control region (535–537 base pairs), and compared their sequence variations with those in Russia. Genetic differences between the Hokkaido and Russian individuals, revealed by the present study, indicated that the Hokkaido population has not been well-differentiated after the immigration into Hokkaido from the Eurasian Continent. The intraspecific variations of M. zibellina showed interpopulation-differentiations due to ‘isolation by distance’ within Hokkaido, and they were smaller than those of the Japanese marten M. melampus, endemic to Honshu, Kyushu and Sikoku Islands of Japan. The genetic diversity in M. melampus populations introduced to Hokkaido was lower than that of M. melampus native on Honshu and Kyushu, and the two mitochondrial DNA haplotypes were distributed at restricted areas in southern Hokkaido. These findings suggest founder effects in introduced populations of M. melampus on Hokkaido. In addition, the clear phylogenetic separation between M. zibellina and M. melampus indicates that no hybridization between them have occurred on Hokkaido so far, although further studies using paternally and biparentally inherited genetic markers are necessary.

Taxonomic status of fossil Sorex shrew specimens from the late Middle Pleistocene of Honshu Island was addressed. Multivariate analyses of upper and lower jaws were conducted for the fossil specimens as well as for 8 species of Recent Sorex shrews from Japan and adjacent regions. It was confirmed that S. shinto and S. hosonoi already inhabited Honshu Island in the late Middle Pleistocene. Most fossil specimens previously referred to S. minutisslmus and S. gracillimus in Honshu should be regarded as S. hosonoi, although some specimens could not be completely distinguished from S. gracillimus. Thus, the appearance of S. hosonoi ascends, at least, to the late Middle Pleistocene during which time they were more broadly distributed, including northern Honshu, than at present (only in central Honshu). Most middle size fossils should be regarded as S. shinto, but at least one specimen was identified as S. caecutlens. It was also demonstrated that other species of Sorex shrews larger than S. shinto and S. hosonoi were present during the Middle Pleistocene of Japan. Most of the large fossils were considered to be S. isodon, which had not been recorded from and is now absent in the Japanese Islands. In addition, there was one specimen of another large species, S. unguiculatus, in Honshu during the Middle Pleistocene, which now is absent in Honshu.

We determined the environmental requirements for Lepus brachyurus lyoni on Sado Island, Japan, during the winter when food resources are low. Hare track survey sites were classified by vegetation type. The movement distances of hares, which indicated habitat preference, were evaluated as the relative length of tracks in a 1-ha area. In addition, we examined browse marks on plants and the protein content of plants to determine the preferred plant species. Movement distances of hares decreased significantly with increasing tree stand height, mean tree diameter, basal area of the tree stand, and forest canopy closure, but increased with branch-and-stem density. The percentage of browse marks on the branches of tree/shrub species decreased significantly with increasing branch-and-stem diameter. The rate of feeding increased significantly with increased protein content in the branches. These findings suggest that hares prefer shrub stands with high branch-and-stem density to mature forests, which have lower branch-and-stem densities. Hares also prefer branches and stems containing large amounts of protein. Young stands offering many shrub species as food resources and shelter constitute preferable environments for hares.

We compared the distribution patterns of five mammals in the Jomon period (12,000– 2,400 years before present time), middle Edo period (1730s), and the present (ca. 1978–2000) using three existing databases. Cervus nippon was widely distributed from Hokkaido to Kyushu during the Jomon and Edo, but absent from most of northern Honshu in the present age. Sus scrofa was widely distributed from southern Hokkaido to the Ryukyu Islands in the Jomon, but absent from Hokkaido in the 1730s and from northern Honshu in the present age. Macaca fuscata was distributed from the northern-most region of Honshu to Kyushu in the Jomon and Edo periods, but diminished in northern Honshu and exhibits a fragmented distribution in the present age. Ursus arctos and/or U. thibetanus were distributed from Hokkaido to Kyushu in the Jomon and 1730s, but greatly diminished in Kyushu in the present age. The distribution patterns of these five mammals changed little from the Jomon to the Edo periods, but were greatly diminished between the Edo and present periods. We suggest that the distribution of mammals changed due to direct (hunting) and indirect (habitat alteration) humanmammal interactions.

A new subspecies of the least weasel, Mustela nivalis (Mammalia, Carnivora) was described from the central mountains of Taiwan. The external and skull morphology, and conventional, G-banded and C-banded karyotypes of the present subspecies were compared with those of other subspecies. This new subspecies is distinct from the other subspecies by relatively longer tail, a broad and short interorbital region, short and narrow rostrum, anteriorly positioned infraorbital foramen, well developed protocone and parastyle of the fourth upper premolars, and anterior margin of right and left upper molars located in a straight line. The number of C-blocks in chromosomes of this new subspecies is less than those for other M. nivalis subspecies.

In order to assess the genetic usefulness of feces that were obtained from field, we conducted genetic identification by microsatellite analysis on fecal samples of the red fox (Vulpes vulpes) collected in the Shiretoko National Park, Hokkaido. Consequently, 59 fecal samples resulted in 22 as the minimum number of individuals. The cumulative P_(ID)sibs in the 22 samples was less than 0.01. Two pairs of fecal samples having the identical genotypes (P_(ID)sibs < 0.01) were considered to be dropped by the same foxes, and the distribution of these feces was almost overlapped with the home range of one fox, revealed by a previous study. Although few reports had revealed the absolute number of red foxes in field, the present study showed that the genetic analysis of their feces is useful for estimating a presumable number of individuals in the area. However, the lower genotyping success rates (23.3–69.8%) and genotyping reproducibility (53.5–88.4%) indicate the difficulty of genetic analysis by using these fecal samples. In order to improve the efficiency of the analysis, it may be effective to select and use a marker set which has smaller allele sizes, because the genotyping success rates and reproducibility increased when the average allele size decreased.