Three earthquakes, two with previously determined fault-plane solutions, are selected in order to study the relation between the S waves and the source mechanism. The S waves are observed at favorable epicentral distances at stations distributed in all quadrants about the epicenter. The earthquakes are of a focal depth of 40 to 60 kilometers and belong to the aftershock sequence of the great earthquake of November 4, 1952. The direction of first motion and the plane of polarization of S are determined by the construction of particle-motion diagrams. In the case of the two earthquakes for which the fault-plane solutions have been published, no correspondence is found between the observed S wave data and the character of the S motion expected on the basis of the given nodal planes of P, whether the source be considered as a single couple or as a double couple. For the third earthquake it is found that the first motion of P is compressional along all rays leaving the focus downward and that the S waves are strongly SV polarized. No faulting mechanism can explain this distribution of the motion in the initial P and S phases. The motion is explained as corresponding to that generated by a simple force acting almost vertically downward. Graphical and analytical techniques of analysis determine the trend of the force at the source to be N 12° W, with a plunge of 85°. A reconsideration of the other two shocks shows that these, too, are better explained by a simple force source than by a faulting mechanism.