The hypotheses of sea-floor spreading and plate tectonics require the removal of sediment from oceanic trenches either by crustal underthrusting or by folding against the base of a continental or insular margin. Accordingly, over a period of time the volume of sediment removed by way of spreading must be equal to the difference between the observable volume of undeformed terrigenous deposits in a trench and the volume contributed to it by continental erosion. To assess possible sediment loss from the central Chilean segment (23°–44° S.) of the Peru-Chile Trench, we have compared the volume of terrigenous deposits overlying the land, the continental margin, and filling the trench with that expected from continental denudation.

Our data indicate that an episode of sediment removal occurred at the base of the margin and adjacent deep-sea floor in Late Cretaceous and perhaps earlymost Tertiary time and may imply spreading. Nearly 100 × 10³ km³ of deposits of Tertiary age, chiefly Eocene to Pliocene, have accumulated on the margin, and perhaps an additional 5 × 10³ km³ in the trench. This amount of offshore sediment could be supplied by fairly low rates (3 cm/10³ yrs) of Tertiary erosion. However, many uncertainties in our denudation-sedimentation budget make it impossible to determine whether or not sediment reaching the base of the margin was removed tectonically in Tertiary time.

Between 27° and 44° S., the trench contains nearly 70 × 10^s km³ of turbidite deposits that we believe accumulated during late Cenozoic periods of glacially lowered sea level. The volume of turbidites in the trench is virtually equal to that expected from continental erosion, which is estimated to have probably been no greater than 5 cm/10³yr for the arid region between 27° and 31°, and 50 cm/10³yr for the humid and partially glaciated region from 36° to 42°. During this time of rapid erosion and trench filling, magnetic data indicate that convergence of lithospheric plates was taking place below the trench at a rate between 5 and 10 cm/yr. If turbidite deposits were swept from the trench at these rates, then continental denudation must have been exceedingly rapid: 20–40 cm/10³ yr for the arid zone, and 110–165 cm/10³ yr for the partially glaciated region. If more conventional estimates of erosion are valid, then either (1) late Cenozoic underthrusting has not taken place (or at a rate much slower than that implied by geophysical data), or (2) underthrusting at the prescribed rates has not involved the removal of a significant volume of sediment from the trench.