Twenty-six mentally retarded children and young adults, most of them with severe neurological deficits, in the age range 5-24 years, were examined ophthalmologically, and their binocular visual acuities were tested with 'operant preferential looking' according to the technique of Mayer & Dobson (1980). Five of the subjects did not co-operate sufficiently for visual testing, although 3 had no ocular abnormalities that would affect binocular acuity. The visual acuity levels determined with 'preferential looking' ranged from 1.29 to 0.03 in the 21 subjects that could be tested. Two children with high myopia showed substantial increase in acuity with appropriate correction (from 0.1 to 1.06 in one child and from 0.05 to 0.31 in the other). Inability to fixate light or objects was found in subjects, who at acuity testing showed values of 0.15 or less. In patients with suspected or definite optic atrophy, visual acuities 0.28 or less were obtained. In general there was a good correlation between ophthalmological abnormalities (in fixation, pupil reactions to light, ocular motility, fundus appearance and refraction) and poor visual acuity. It is therefore concluded that visual acuity testing with 'preferential looking' techniques can give valid information on central visual function in mentally retarded children and young adults.