Reflection seismic data can be imaged by migrating common midpoint slant stacks. The basic method is to assemble slant stack sections from the slant stack of each common midpoint gather at the same ray parameter. Earlier investigators have described migration methods for slant stacked shot profiles or common receiver gathers instead of common midpoint gathers. However, common midpoint slant stacks enjoy the practical advantages of midpoint coordinates. In addition, the migration equation makes no approximation for steep dips, wide offsets, or vertical velocity variations. A theoretical disadvantage is that there is no exact treatment of lateral velocity variations. Slant stack migration is a method of "migration before stack." It solves the dip selectivity problem of conventional stacking, particularly when horizontal reflectors intersect steep dipping reflectors. The correct handling of all dips also improves lateral resolution in the image. Slant stack migration provides a straightforward method of measuring interval velocity after migration has improved the seismic data. The kinematics (traveltime treatment) of slant stack migration is also accurate for postcritical reflections and refractions. These events transform into a p-tau surface with the additional dimension of midpoint. The slant stack migration equation converts the p-tau surface into a depth-midpoint velocity surface. As with migration in general, the effects of dip are automatically accounted for during velocity inversion.--Modified journal abstract.