The problem of the maximum depth, down to which the sea level dropped during the last glaciation, is even more difficult and obscure than that of the high sealevels. Some evaluations have been proposed, being based on the estimations of depth of the Holocene sediments base in the coastal plain or on the consideration of submerged sediments or geomorphological features now found on the continental shelf. Many Japanese geologists estimate -100m to -140m for the last glacial low stand sealevel, but some doubt is cast on their data and opinions. Our estimate, -80±5m, was a conservative interpretation based on dates of shells and peat obtained from the shelf and coastal plain of Japanese Islands.
From the detailed survey of the submarine topography in the Tsugaru strait between Honshu and Hokkaido, six submarine terraces were found on the sill. It is noteworthy that the depth of those terraces on both areas of Honshu and Hokkaido is nearly the same. This fact suggests that there have been no distinct crustal movements after the building of those terraces on the sill of the strait. Among the submarine valleys developed on the continental shelf, one of which continued to a valley on land can be detected on the eastern part of the sill. From those circumstances of submerged terraces and valleys about 80m depression of sealevel at the last glaciation is deduced.
In most Pleistocene, the Japanese Islands was connected with the Korean Peninsula, and present major islands themselves were tied to each other. It is sure that large mammals such as elephants migrated into the Japanese islands through land bridges. In the early Shimosueyoshi transgression (Riss-Würm interglacial period), when the sealevel was about -100m, the Japan Sea connected with the Pacific through narrow passages located in the Korean and Tsugaru straits. At the time of maximum Würm when the sealevel was depressed to -80±5m, the land bridges between Honshu and other lands were never formed.
