This article is largely devoted to the historical review of researches concerning island arcs (Chapters I-III). In Chapter IV, such a writer's opinion as follows is presented briefly. 1. The Fossa Magna-Shichito-Bonin Arc, which is apparently younger than the Honshu Arc, is a manifestation of a certain deep seated zone of boundary dividing certain units also buried deep in the earth's mantle. The units and the boundary zone are thought to be of very ancient origin. Underlying the Palaeozoic geosyncline and the later geanticlines and geosynclines of the Honshu Arc. they had affected the Palaeozoic and Mesozoic history of the Honshu Arc, thus exhibiting various differences in rock facies, thickness of deposits, igneous activity and in time and magnitude of crustal movement between Northeast and Southwest Japan. At the beginning of Neogene epoch, the deep seated zone of boundary manifested itself resulting in the Fossa Magna. 2. A layered structure of the earth's mantle is probably the best solution of this problem. The low-velocity layer of GUTENBERG and the discontinuous distribution of deep earthquake foci in the vertical section suggest also the layered structure. Each layer is, further, divided into segments of varied horizontal extension, larger and polygonal below continents and oceans and smaller and elongated below the orogenic belts. 3. The significance of the Palaeogene-Neogene boundary and of the Neogene crustal development is very great so far as the western Pacific island arc province is concerned. The Neogene history in these areas is essentially independent of the preceding Palaeogene and Mesozoic history, though some tendencies to conform to the older elements are apparent.