Congenital nystagmus (CN) has been described as a 'fixation' nystagmus implying an inability to fixate a target. However, each cycle of CN contains a target-foveation period during which the eye velocity is at, or near, zero. Prolongation of foveation time, reduction of retinal image velocity and cycle-to-cycle foveation repeatability all contribute to increased visual acuity. We developed several methods to accurately measure the dynamics of foveation in CN; their use is illustrated on an individual with typical idiopathic CN and no afferent defects. During eight 5-second intervals of fixation on a stationary target, the horizontal standard deviation (SD) of the mean foveation position (FPOS) was 12.82 minarc and the SD of foveation velocity was 118.36 minarc/sec. The SD of the means of total eye position and of the non-foveating peak of the CN were 43.17 and 25.32 minarc respectively. The mean foveation-time interval (eye velocity less than or equal to 4 degrees/sec) was 57.27 msec. The SD FPOS for the best 1-second interval (4 successive CN cycles), in a typical 5-second record, was 0.71 minarc. Histograms revealed peaks of eye position at 0 +/- 10 minarc and of eye velocity at 0 +/- 240 minarc/sec. The small vertical component of the CN (16 minarc peak-to-peak) had a SD of 6.56 minarc. A nystagmus foveation function related to visual acuity was derived that was more sensitive than CN intensity. The increased visual acuity resulting from the use of convergence or base-out prisms was due to increased foveation time. Although it might appear that CN is a defect of fixation, this individual with CN had strong fixation reflexes in the sense that he was able to accurately (within 1 minarc) achieve (interbeat) and maintain (intrabeat) target foveation for appreciable periods of time. Our data support the hypothesis that individuals with idiopathic CN do not have a primary disturbance of fixation.