PURPOSE: We sought to characterize the dyschromatopsia of optic neuritis, to determine the type and severity of color defect present and its relation to central vision and spatial acuity, to examine changes in this dyschromatopsia over time, and to determine the applicability of Köllner's rule to patients with optic neuritis. METHODS: We analyzed the raw data on color vision performance as assembled within the Optic Neuritis Treatment Trial (ONTT). The ONTT was designed to evaluate corticosteroids as a treatment for acute demyelinating optic neuritis and to allow long-term outcome and natural history analyses. Between July 1, 1988 and June 30, 1991, 488 patients were enrolled in this trial. All patients underwent extensive neurologic and ophthalmologic examinations including standardized testing of visual function that included testing of color vision. The ONTT population thus afforded a unique opportunity to characterize acquired dyschromatopsias in a large, homogenous, well-characterized cohort of patients with optic neuritis. We used quantitative analysis of FM-100 scores from this patient cohort to determine the severity of the dyschromatopsia, the selectivity of the dyschromatopsia (polarity of errors) and the type of dyschromatopsia (axis of confusion) by employing quadrant analysis of FM-100 scores. RESULTS: The results of high-and low-selectivity analyses of the FM-100 data showed that during the acute phase of optic neuritis, blue/yellow, red/ green, and non-selective color defects occurred; among patients with pure defects, blue/yellow defects were more frequent than red/green defects. At 6 months after the acute event, however, analyses showed that red/green defects were more common than blue/yellow defects. Among patients with selective color defects both acutely and at 6 months, the defect was as likely to change over time as remain the same. The likelihood of persistent dyschromatopsia at 6 months was related to the severity of initial central acuity loss, but the type of dyschromatopsia present (red/green versus blue/yellow) was not. CONCLUSIONS: Our results suggest that at the time of the acute attack of optic neuritis, the majority of selective color defects were blue/yellow defects, whereas at 6 months, more of the selective defects were red/green defects, though both types of defects (as well as nonselective defects) were seen acutely and at 6 months. Despite the rigorous inclusion criteria of the ONTT, the large number of patients we studied, correlation of color vision with visual acuity, and longitudinal follow up, this study showed that no single type of color defect was consistently associated with optic neuritis. Demyelinating optic neuritis does not obey Köllner's rule. Moreover, the type of defect present changed in some patients over the course of recovery. Thus, the type of defect may not even be consistent in individual patients as they recover. The type of defect appeared to be related to spatial vision at the time of the test, but the type of defect present at 6 months was not related to the severity of the initial visual loss. Therefore, in evaluating color defects associated with optic neuritis, the level of central visual function must be considered.