Persons who witnessed an automobile accident involving a green car were exposed to information that the car was blue. On a subsequent color recognition test, most subjects shifted their color selection in the direction of the misleading information and away from the actual perceived color. Shifting was greater for subjects who did not initially commit themselves to a color selection. Control subjects, who were not exposed to misleading information, distributed their chokes around the true color, and did not show a systematic color shift over time.