The effect of cataract surgery was assessed in 103 patients whose visual acuity, as measured in dim light and with high contrast Snellen letters, was 20/50 or better. The patients desired cataract surgery. They complained of glare symptoms, had given up driving at night, or both. Contrast sensitivity was measured under simulated night conditions with a calibrated Vistech MCT 8000 device. No patient who had a contrast sensitivity of less than 27 at 6 cycles per degree (20/70 "equivalent acuity") had cataract surgery. The mean preoperative contrast sensitivity function showed abnormally low values for all frequencies (cycles/degree) tested. Postoperatively the values returned to normal. Of 99 patients who answered a postoperative questionnaire, 41 (41%) had been driving at night preoperatively and 80 (80%) were driving at night postoperatively. A control group of 24 noncataractous patients (48 eyes) was in the normal range of the test device used. These results support the view that the benefits of cataract surgery are not well predicted by Snellen acuity, even though functional vision may be significantly reduced.