Deep-sea trenches are the result of extension. This is not merely a matter of flexing a slab as it bends and begins to descend (sea-floor spreading hypothesis): the available seismic data show that the primary stress field results from more or less horizontal tension--at right angles to the axis of the trench--at most depths. Many persons reporting these results also have concluded that the same data "are consonant with" the underthrusting required by the sea-floor spreading hypothesis. Close study of these reports shows, however, that the underthrusting must be assumed first; even then it is difficult or impossible to reconcile the data with the assumption. This is so obvious that several recent authors have commented on the conflict between observation and hypothesis. >

Much published information is available about the structure, seismology, gravity, magnetics, volcanicity, heat flow, bathymetry, and sedimentation in and near deep-sea trenches. This information is contrary specifically to the assumption of compression. One of the most telling arguments against the compression assumption arises from seismic first-arrival patterns for earthquakes taking place approximately under trenches and island arcs; these produce largely tensional, or strike-slip, or ambiguous patterns (approximately in that order); the few compressional examples commonly have the compression axis oriented parallel with the alignment of the major structure, and ambiguous examples which might support underthrusting (if one solution is rejected) commonly require a horizontal thrust lane. Yet the conclusion of compression continues to be drawn, specifically for the reason of preserving the hypothesis in the face of mounting contradictions.

Efforts have been made to salvage the "plate consumption" idea by suggesting that the ocean crustal layer is "falling" through the mantle under or near trenches. This ad hoc assumption contradicts known gravity and density data, dip data, Q data, and the layering of the shallow mantle near trenches. Actually, deep-sea trenches are tensional because this is the primary stress field; there is no "down-going slab," and there is no primary horizontal compression.