Scatter plots showing the amplitude versus velocity (maximum and average) relationship of horizontal saccades in 25 normal subjects and four groups of patients were statistically compared. Three patients with “subclinical” medial longitudinal fasciculus syndromes had significant slowing of adducting saccades, and two of these patients had unsuspected slowing of abducting saccades (although to a lesser degree). Five patients with olivopontocerebellar degeneration and three patients with myotonic dystrophy had significant slowing of saccades in both directions. Five patients with surgically documented acoustic neuromas did not have significant slowing despite brain-stem compression in three. It is concluded that the saccade velocity test can be a useful clinical tool in addition to its potential in clinical research.