Cell biology: DNA versus membrane.

Density-driven sinking velocity Density-driven sinking velocity Slab-dip angle Plate-advance velocity Subduction-retreat velocity Subduction-retreat velocity Overriding slab Subducting slab transformations at 400–700 km down. The authors identify four central points. First, subducted material sinks at a velocity that is controlled solely by its negative buoyancy and the viscosity of the surrounding mantle. Second, subducting material cannot bend in the time it takes to sink, owing to its internal strength. Third, there is a small 'subduction retreat' component that lessens the total sinking velocity: the subducting ma terial sags under its own weight so that it rolls back somewhat as it subducts (Fig. 1). Fourth, because this subduction-retreat contribution is small, the velo city at which the subducting plate advances at the surface is close to the density-driven sinking velocity of the subducted material in the mantle. This 'free-subduction model' assumes that the contribution to the subduction-retreat velocity of the second, overlying tectonic plate is so small as to be negligible. If the plate-advance velocity is greater than the density driven sinking velocity, the subducting material will thicken, increasing its negative buoyancy and thus its sinking velocity. It will do this until the sinking velocity is consistent with the plate-advance velocity. If, by contrast, the plate-advance velocity is less than the sinking velocity, the subducting material will thin out and stretch. Eventually, the bottommost part will drop off, forming a 'drip'. For many subduction zones, the plate vel ocity is in good agreement with the sinking velocity predicted by this model. What is interesting is to look at the cases where the model fails. An example is the subduction zone that runs down the western coast of Central America. Here, an upper mantle of locally low viscosity, or a locally high density of the sub-ducting material, must be invoked to match the model velocity to observations — although there is no evidence that the region has such exceptional properties. This is also a region where seismic tomographic imaging shows subducted material extending down beyond 700 km, and Goes et al. suggest that this extra material, not accounted for in their model, is the source of extra density driving the higher velocity. They go on to postulate that instances where the model fails are indicators of material sinking into the lower mantle, and they look for other times in the plate record where plate motions are outside the estimates for density-driven sinking velocities …