Four step methods for y″=f(x, y)

A systematic investigation is undertaken here on the possible practical consequences of the fact that the linear four step methods for y″=f(x, y) form a family with a certain structure. The mentioned property is found to be surprisingly rich in consequences, and three of them are of particular importance: (i) drastic reduction in the number of iterations at each step, (ii) increased flexibility with respect to the start of integration and to the modification of the step size during the integration process, (iii) possibility of direct and simple appraisal of the local truncation error.