Optokinetic nystagmus (OKN) has been used to measure the visual resolution of a variety of species, but little is known about the development of OKN resolution in normal children. We used OKN to estimate the monocular visual resolution of 181 children, each tested at one of seven ages between 2 and 36 months. The child watched black-and-white vertical stripes moving horizontally across an 84 x 84 deg screen at 13 deg/s. The estimate of visual resolution was the width of the narrowest stripes which observers judged to elicit OKN. A test typically took 5-10 min and 95% of the children completed two tests of one eye. Estimates of visual resolution from the two tests agreed closely with each other:97% of the estimates from the second test were within 1 octave of those from the first test. Results for 165 children with normal eye alignment and minimal refractive errors indicated that, under our testing conditions, OKN resolution improves from 40 min at 2 months to at least 6 min at 36 months, with a plateau in development between 18 and 24 months of age. Children retested at a later age (n = 96) yielded results very similar to those of naive children of the same age. To evaluate the concurrent validity of OKN resolution, we tested one eye of each of 84 older amblyopic patients with OKN and, on the same day, with a linear letter chart at far. Results from the two tests were moderately but significantly correlated (r = 0.63, p<0.001). Although measurements based on optokinetic nystagmus, preferential looking, or visually evoked responses all indicate that visual resolution improves with age, they disagree on the shape of the developmental function, probably in part because they involve different neural centers. Therefore results from more than one measure provide a fuller picture of visual development during infancy. The ease of testing, good reliability, and significant concurrent validity suggest that our OKN procedure would be useful for testing the visual resolution of preverbal patients with eye problems.