A simple technique is described for assessing the sensitivity of the human visual system to gratings at threshold contrast. The technique has advantages for clinical use in that it is (1) inexpensive, (2) quick to administer, (3) portable, and (4) relatively free from bias. Forty-two diabetic patients and 84 normal controls have been tested. Fifteen diabetic patients (6/20 with retinopathy and 9/22 without) had test scores more than two standard deviations below the norm for age-matched controls.