We discuss two papers [Spong, M. W. (1989) Syst. Central. Lett., 13, 15-21; Chang, Y.-Z. and R. W. Daniel (1992) Automatica, 28, 969-974] on the adaptive control of flexible joint robot manipulators. Our goal is to clarify the contributions of the two papers by correcting some technical errors and to provide some additional insight into the application of singular perturbation methods for the control of flexible joint robots.