The relationship of shear stability in laboratory tests to shear stability in service was investigated in a diesel injector test, in the L-38 single-cylinder engine, in a multicylinder engine equipped for bearing film thickness measurement, and in road testing in an automobile and in trucks. All of the laboratory and road tests differed in severity of shearing and also showed that major VI improver types differed in the time needed for their kinematic viscosity to approach a fully-sheared value. Shear stability comparisons based on kinematic viscosities measured during the transition to the fully sheared condition did not accurately represent the relative shear stability of VI improvers in the fully-sheared condition. In the road tests, most of the oils were at their fully-sheared viscosity for the major part of their service life. The decrease of HTHS viscosity due to shearing in the road tests and laboratory engines was usually less than half the kinematic viscosity decrease. Bearing oil film thicknesses showed little or no change with test duration in the multi-cylinder lab engine and correlated well with HTHS viscosity.