SECT. 2] TRENCHES 427 



that the anomahes are the result of deeper structure, not the topography itself. 

 A very similar situation is presented by the Java Trench and its central ridge. 

 The Barbados Ridge was explained by Hess (1938) as the squeezing up of in- 

 competent material from the axis of a down-buckle "like tooth paste out of a 

 tube". Deep drilling for oil on Barbados in recent years gives a picture con- 

 sistent with this interpretation so far as the sedimentary section is concerned. 

 But stronger evidence is supplied by the seismic profile given by J. Ewing et al. 

 (1957) showing the Mohorovicic discontinuity coming in from the Atlantic 

 and bending downward approaching the Barbados Ridge (Fig. 8b). Similarly, 

 Raitt, Fisher and Mason (1955) show the crust bending downward as one 

 approaches the Tonga Trench from the Pacific (Fig. 8c). On the island side one 

 might interpret the crust as behaving in much the same way except that the 

 mantle velocity for P waves below it has dropped from 8.2 km/sec to 7.0- 

 7.6 km/sec. This could be a result of serpentinization as discussed below. The 

 Tonga Trench has large negative gravity anomalies (Talwani, Worzel and 

 Ewing, 1961) but little sedimentary fill on its floor. On its eastern flank a guyot, 

 Capricorn Seamount, with its upper surface tilted toward the trench axis has 

 Ijeen described by Raitt et al. (1955). The bending downward of this upper 

 topographic surface of the guyot may be assumed to parallel what the crustal 

 layers have done beneath it and thus supports the seismic evidence mentioned 

 above. Menard and Dietz (1951) report a guyot on the axis of the Aleutian 

 Trench with its flat top about 900 fm deeper than guyots to the south of the 

 trench. 



The Puerto Rico Trench has been investigated much more intensively by 

 geophysical methods than any other trench area. The geology of the islands to 

 the south of it is now fairly well known (Mattson, 1960; Berryhill et al., 1960; 

 DonneUy, 1959 ; Helsley, 1960). More deductions may be drawn with regard to 

 its structure than that of other trenches. One may hazard a guess that it first 

 formed at about the beginning of Tertiary time, some 70 million years ago. 

 Until late Eocene, active vulcanism and the intrusion of dioritic plutons were 

 taking place but ceased at this time. Vulcanism, however, continued up to the 

 present along the Lesser Antilles. Paralleling this, the sediments of the Bar- 

 bados Ridge have been squeezed up from a trench axis but those in the Puerto 

 Rico Trench apparently have not been. From this one may deduce that the 

 Puerto Rico Trench is not actively going down at the present time, but that 

 forces must be still acting upon it to keep it far out of isostatic equilibrium. 

 Unfortunately, it is not a typical trench as would be the Tonga-Kermadec, 

 Aleutian or Kuril Trenches. Large east-west strike-slip faults with a sinistral 

 movement are known on the island of Puerto Rico, and N 60°E sinistral strike- 

 slip faults paraUeling the Anegada Trough have been mapped by Donnelly in 

 the Virgin Islands. Such faults may have modified to some extent the Puerto 

 Rico Trench, but probably as second-order effects not changing the gross 

 structure drastically. 



Talwani, Sutton and Worzel (1959) have computed very carefully mass 

 distributions under the Puerto Rico Trench section and to the north and south 



