For spatial regressions with sinusoidal surfaces, the ordinary least squares estimator (OLSE) is shown to be asymptotically as efficient as the geeralized least squares estimator (GLSE) in that the covariance matrices of the two estimators have the same nontrivial limit under the same...