An explicit one-stage method is proposed for the numerical integration of systems of ordinary differential equations. A class of structural features of such systems is singled out. Economical numerical schemes up to the fourth order of accuracy are constructed using these features at the algorithmic level.