SECT. 2] TRENCHES 413 



deep-focus earthquakes. From his analysis of specific examples he finds that 

 the plane of movement is nearly vertical and that the displacement is of strike- 

 slip character. His resolution, however, gives in each case two planes at right 

 angles, one parallel, or nearly so, to the trench and one at right angles. Mclntyre 

 and Christie (1957) suggest that the set which is parallel is the preferred 

 solution after analyzing Hodgson's earthquake data for the Tonga-Kermadec 

 region. Hess and Maxwell (1953) had previously suggested strike-slip faults in 

 this area at right angles to the trench and still hold this view. 



Seismic refraction investigations at sea have provided information on 

 crustal-layer structure down to the Mohorovicic discontinuity. Scripps Institu- 

 tion scientists have carried out such studies in the Tonga Trench (Raitt et al., 

 1955), Middle America Trench (Fisher and Shor, 1956; Shor and Fisher, 1961), 

 Peru-Chile Trench (Raitt and Shor, 1958) and Aleutian Trench (Shor, 1962). 

 Lamont and Woods Hole personnel have investigated the Puerto Rico Trench 

 (Officer et al., 1957, 1959, for example) and the South Sandwich Trench (Ewing 

 and Ewing, 1959). Russian scientists report seismic and gravimetric work 

 across the Kuril-Kamchatka Trench (Galperin et, al., 1958), and Gaskell et al. 

 (1958) report one short profile near the Challenger Deep in the Marianas 

 Trench. 



Measurements of the flow of heat through trench floors have been reported 

 for the Middle America Trench (Bullard et al., 1956) and the Peru-Chile Trench 

 (von Herzen, 1959). Heat flow was considerably lower near the trench axes off 

 Guatemala and Peru than was general in the deep-sea floor outside. No measure- 

 ment was obtained in the deepest part of the Chilean segment, off Antofagasta, 

 but a series of nearby measurements, normal to the trench axis, showed 

 progressively lower values as the deep was approached. Though the low values 

 observed (0.17-0.47 ^cal/sec/cm'-) might be due to rapid sedimentation in 

 trenches, these authors conclude that the abnormal values reflect real differences 

 in processes occurring far beneath the sea floor. They suggest that convection 

 currents in the mantle may bring heat to the surface in areas such as the East 

 Paciflc Rise, where high values of heat flow (up to 8.1 (jLcal/sec/cm^) have been 

 measured, and that the trenches may be regions where such currents are 

 sinking (Bullard et al., 1956). 



Magnetic profiles have been published for the Tonga (Raitt et al., 1955), 

 Kuril-Kamchatka (Galperin et al., 1958) and Aleutian (Keller et al., 1954) 

 trenches. From the absence of any consistent anomaly correlatable to the 

 trench on their eight crossings, the latter authors state that the anomalies 

 observed are produced by susceptibility contrast within the rocks of the sea 

 floor rather than by the topography of their surface. They conclude that the 

 structure which produced the susceptibility contrasts bears no relation to the 

 trench. Mason (in Raitt et al., 1955) reports an excess of magnetism on the island 

 side of the Tonga Trench and a deficiency to the east that cannot be explained 

 by reasonable basement configuration ; he also considers this distribution may 

 be due to susceptibility contrast rather than to structure. 



Marshall (1911) pointed out that, in Melanesia, andesitic lavas occur on and 



