In a recent paper Dhingra, J. S., R. L. Moose, H. Vanlandingham and T. A. Lauzon (1992). Automatica, 28, 395-399, have presented an estimation technique for a class of nonlinear systems. They claim to have shown by means of numerical examples that the proposed algorithm, called the Jump Matrix Technique (JMT), gave better results than the Extended Kalman Filter (EKF). The purpose of this note is to point out that a better implementation of the EKF would outperform the JMT.