The vast and growing literature on the search for stratigraphic traps for petroleum ignores gravity gradients, for which the theory has been available since the heyday of the Eotvos torsion balance decades ago. These are discussed in this paper.The horizontal and vertical gradients can be measured with available gravimeters and (to a limited extent) with the Eotvos torsion balance. A major advantage of the gradient method is that surveying to determine position and elevation of the station is not required. Both theory and practice have been reported in the geophysical literature, but the important application to stratigraphic traps has not been mentioned. We evaluate here the method for locating 'pinchouts', a term which embraces 'stratigraphic' and 'unconformity' traps in Halbouty's (1972) classification. Both position and depth of the assumed pinchout are determined by the gradient anomaly. The magnitudes of anomalies of horizontal and vertical gradients are about equal. However, pending new instrumentation, only the horizontal gradient is practically useful for field surveys. The gradient method is quantitatively promising and, used in conjunction with other methods, should significantly advance the search for stratigraphic traps for petroleum.