The matter of the viscosity of the mantle has started to become serious. In 1935, Norm Haskell estimated the viscosity to be about 1020 poise and there the matter stood for about half a century. For a little while, people worried about excess ellipticity of the Earth and attributed this to a “fossil bulge” that lagged the rotation rate. For this same little while, 1025 poise was thought to be the viscosity of the lower mantle, but then it was discovered that the equator was also out of shape by about the same amount, ruling out the “fossil bulge” idea. To cover their embarrassment, geodynamicists upped the viscosity of the mantle to 1021 by adopting S.I. (Satan's Invention) units. No one noticed for some time since it didn't really matter whether viscosity was given in stokes, poise, or pascal seconds. It was just a large number with a large uncertainty and no one had a feel for it anyway.