Methods for estimating the truncation error in processes of the Runge-Kutta type are divided into two categories; those which make use of the data of several consecutive steps, and those which use the data of a single stop only. Existing methods of both forms are discussed, and some new methods are suggested.