A method is presented for determining the differential equations of the output moments of a linear random system in which the coefficients are components of vector Markov processes. The method is based upon well‐known results in applied stochastics. It appears considerably simpler than methods recently suggested for studying the behavior of linear random systems.A stability criterion previously published is obtained with little effort. Reasons are presented for disagreeing with published results that mean motion of linear random systems with Gaussian parameter variations is not influenced by these variations.