412 FISHEK AND HESS [CHAP. 17 



scientists of the Laniont Geological Observatory, and others, have actively 

 pursued marine gravity exploration in trench areas ; notalile are their surveys 

 of the Puerto Rico (Tahvani et «/., 1959), Peru-Chile (Ewing et al., 1957), 

 Tonga (Talwani et al.. 1901) and Middle America (Worzel and Ewing, 1952) 

 trenches. CJajnanov (1955) reviews gravity observations in the Kuril- 

 Kamchatka trench. All of these surveys confirm the existence of a strip of large 

 negative isostatic anomalies centered slightly shoreward of the trench axis and, 

 commonly, smaller positive anomalies closer to the island group or mainland 

 and also seaward of the trench. In some localities, Timor and Barbados for 

 example, the negative strip extends beyond the trench, continuing across the 

 island or along a ridge that perhaps marks the ])osition of a former deep (Hess, 

 1938). The shape and dimensions of the anomaly curve led Vening Meinesz 

 (193(>) to propose that in the region of the negative belt there is an elastic down- 

 buckle of light crustal material into the substratum. Kuenen (1936) published 

 the results of model experiments in which he duplicated this hypothetical 

 structure that he called the "tectogene". Consideration of the probable stresses 

 and elastic limits of crustal materials and the studies of Bijlaard (1936) as well 

 as recent field studies of island arc structure caused Vening Meinesz to revise 

 his hy])othesis and to suggest that the anomalies reflect a plastic deformation 

 of crustal material in the mobile belts, with thickening and down-buckling 

 under compression (Vening Meinesz, 1954, 1955). He derived formulas (Vening 

 Meinesz, 1955) that indicate an asymmetrical development of the two sides of a 

 down-buckle at the ocean-continent boundary ; such asymmetry is confirmed 

 by the data of station and explosion seismology. 



Compilations of earthquake epicenter data, such as those of Gutenberg and 

 Richter (1954), establish that highest seismicity is associated with the circum- 

 Pacific trenches and those of the Caribbean and South Antilles. They indicate 

 that shallow shocks commonly occur near and slightly shoreward of the trench 

 axis and intermediate depth-shocks (70-300 km) below and inside the island arc 

 or mountain chain. When deep shocks (300-700 km) occur, their foci lie well 

 inside the arcuate structure, below an interior basin or inland from the momitain 

 range. From studies of the circum-Pacific earthquake sequences, BenioflF (1949, 

 1954) has suggested that great reverse faults or fracture zones dip steeply 

 shoreward along the island arcs of continental margins. He distinguishes two 

 types of structure, the dual oceanic and the triple marginal fault. The former, 

 typified by the Tonga and Kermadec earthquake sequences, has a lower com- 

 ponent (70 to 550-700 km) that dips steeply away from the ocean basin. The 

 marginal faults, perhaps most simply indicated by the Kuril-Kamchatka and 

 Peru-Chile sequences, have an intermediate component (70-300 km) of moderate 

 dip and a steep lower component extending to a depth of 550-700 km. For 

 neither fault type is the observational precision great enough to indicate 

 the dip of the shallow (0-70 km) component. Benioff believes the trench and 

 uplifted shoreward block are surface manifestations of reverse motion on the 

 faults or fracture zones. 



Hodgson (1957) has questioned this postulated thrust-fault character of 



