Abstract
This paper considers the concepts of leverage and influence in the linear regression model with correlated errors when the error covariance structure is completely specified. Generalizations of the usual measures are given. Extensions of residuals also naturally arise. The theory is illustrated using two examples