The purpose of this paper is to look into several common idempotent matrices of interest in the context of any mixed model least-squares adjustment. The cofactor (scaled covariance) matrices, based on the law of error propagation, are incorporated into th
