414 FISHER AND HESS JCHAI-. 1/ 



to the southwest of a line drawn from New Zealand through Tonga, the Solomon 

 Islands and New Guinea, and that this line closely ])arallels a series of oceanic 

 trenches. Many geologists have taken this petrographic distinction, the "an- 

 desite line"*, extended northward along the deeps east of the Palau- Yap- 

 Marianas-Japanese -Kuril islands as the chief criterion for distinguishing 

 "continental" from Pacific basin areas. This raised the question of why, if 

 trenches are continental border structures, two such topographic features, the 

 Philii)pine and Ryuku trenches, lie so far outside the Pacific basin proper and 

 two others, the New Biitain and New^ Hebrides trenches, occur on the wrong 

 side of the andesite line. Recent seismic work in the Caribbean, the deep Bering 

 Sea and the Philippine Basin indicates that an oceanic crust lies between the 

 island arcs and the continents. Hess (1955) suggested that the occurrence of 

 andesite, rather than indicating the margin of the ocean basin proper, may 

 provide an index of deformation ; andesitic magmas might be produced where 

 deformation is strong enough to force the basaltic or lower crustal layer down- 

 w-ard sufficiently to cause partial fusion. It now seems more likely that andesite 

 of island arcs as well as basalt of the ocean basins comes from deep within the 

 mantle, the former occurring only in areas of strong shearing stress. Partial 

 melting of a small percentage (5% to 10%) of a given volume of mantle will 

 ])roduce a less mafic liquid than basalt, but this magma would be retained in 

 the rigid crystalline framework of the rock unless released by shearing. Partial 

 melting of a large enough percentage of the same volume to destroy the rigidity 

 of the framework could produce a separatable basaltic liquid under relatively 

 passive conditions. 



Another striking petrologic observation is the common occurrence of ser- 

 pentinized peridotites in island groups bordering the East Indian, w^estern 

 Pacific and Antillean trenches (Hess, 1937, et seq.). Hess had proposed that 

 these peridotites were intruded during the first great deformation of the 

 present mountain belts, presumably during a down-buckle of the crust, and 

 that their position ai)proximated the axis of the down-fold. From various 

 studies of the relations of these intrusions to datable rocks, Hess assigns a 

 probable mid-Mesozoic age to the main deformation in the Japan-Formosa - 

 Luzon-Borneo regions and a lesser age, perha])s late Cretaceous-Eocene, to 

 those intrusions in the Bonin-Marianas-Palau and Mindanao-Celebes-New^ 

 Guinea belts (Hess, 194S). 



Various workers have noted that in both island arc-trench and continental 

 margin-trench systems a row or double row of volcanoes, active at present or in 

 the fairly recent geologic past, lie sub-parallel to the trench trend. The vol- 

 canoes lie 150-350 km landward from the trench axes, with the most common 

 separation about 200 km. Such volcanoes, for example, comprise the islands of 

 western Tonga, much of the Marianas, Bonin, Kuril and Aleutian groups, the 

 ])eaks on the east shores of the southern Philipj)ines and the Kamchatka 

 peninsula, and the mountains bordering the west coast of Central America. The 

 island volcanoes do not ordinarily rise from the deep-sea floor but from the 

 crests or back slopes of geanticlinal ridges bordering the trenches. Gunn (1947) 



