In three experiments the production and interpretation of perspective drawings by blind and sighted subjects were examined. Blindfolded sighted, early-blind, and late-blind subjects first attempted to produce raised-line drawings of a surface at a number of angles-0 degree (panel horizontal), -22.5 degrees, -45 degrees, -67.5 degrees, and -90 degrees (vertical). Congenitally blind subjects did not show foreshortening in their naive raised-line drawings. However, the congenitally blind subjects were able to understand aspects of perspective, and performed as well as the blindfolded sighted and late-blind subjects in a subsequent multiple-choice task. Subjects in the multiple-choice task were required to match tangible perspective drawings to a slanted board. Although the three groups performed alike, both groups of blind subjects performed better than blindfolded sighted controls on judgments involving drawings of the vertical panel in the second experiment. In a final experiment, in which vision and touch were compared, sighted subjects were required to adjust the angle of the panel to match foreshortened, perspective drawings. This experiment yielded significantly better performance with vision than with touch, but only for drawings of the board at the vertical orientation. The results suggested that congenitally blind people may benefit from haptic exposure to raised-line configurations representing geometric perspective.