The persistent, second order, anomalous signal found in the Brillet and Hall experiment is derived by applying 4D differential geometry in the rotating earth frame. By incorporating the off diagonal time-space components of the rotating frame metric directly into the analysis, rather than arbitrarily transforming them away, one finds a signal dependence on the surface speed of the earth due to rotation about its axis. This leads to a Brillet-Hall signal prediction in remarkably close agreement with experiment. No signal is predicted from the speed of the earth in solar or galactic orbit, as the associated metric for gravitational orbit has no off diagonal component. To corroborate this result, a repetition by other experimentalists of the Brillet-Hall experiment, in which the test apparatus turns with respect to the earth surface, is urged.