Abstract
Training improves the ability of human observers to discriminate between two similar directions of motion. This gradual improvement is specific to the direction on which an observer is trained, and it endures for several months. Improvement does not affect motion perception generally, nor does it depend on recognition of details of the movement.