On the existence of a second scale of convection in the upper mantle

Summary. This paper is concerned with an examination of the possibility that there might exist a small scale of convective circulation beneath the oceanic lithosphere. Recent suggestions that this might be the case have been made in an effort to understand why the bathymetry of the sea‐floor deviates from the prediction of boundary layer theory for ages in excess of about 100 Ma. The energy source which sustains the secondary motion is supposed to be found in the steep temperature gradient near the planetary surface which is itself presumably maintained by the large‐scale convective circulation associated with plate creation and destruction. Here we investigate the extent to which the temperature dependence of viscosity may act so as to stabilize the upper boundary layer against disruption by such secondary instability. If the viscosity profile is monotonie and the asymptotic upper mantle viscosity is about 10²²poise, as suggested by post‐glacial rebound data, then the existence of the second scale is extremely unlikely. On the other hand, if a sufficiently pronounced low viscosity zone does exist under old sea‐floor then the development of such a second scale cannot be ruled out completely. Some recently obtained geophysical evidence is reviewed which suggests that this is unlikely to be the case.