A simple problem of a two-body system perturbed by the disturbing function epsilon/r-squared is considered to show that the time-averaged equations of the true and mean anomaly are not necessarily equal. Secular effects due to perturbations in the mean and true anomalies are different not only in value but also in sign. A more general case where the disturbing function is periodic with respect to the mean anomaly is also considered. Here a discrepancy in the time-averaged solutions is due to the fact that the mean anomaly is an action-angle variable conjugate to the action variable L, while the true anomaly is not. When the action-angle variables are chosen as dependent variables, the Hamiltonian is a function only of action variables. One must first derive first-order solutions, substitute them into the right-hand side of the equations of motion, and then take time averages.