We performed a retrospective study on 100 eyes of 55 consecutive patients to compare the stereoscopic quality of simultaneous stereoscopic disk photographs and the reproducibility of optic nerve assessments using the Nidek 3Dx system and the Donaldson fundus camera. Photographs were taken on the same eyes on the same dates and reviewed independently by one clinician three times in a randomized, masked fashion. Mean cup/disk ratios using the Nidek 3Dx system were similar to mean cup/disk ratios using the Donaldson camera. There was significantly less variability of disk analysis using the Nidek 3Dx system than the Donaldson camera (F = 1.6, P = .01, df = 99). Overall mean stereoscopic quality was significantly better with the Nidek system (3.2 +/- 1.15 S.D.) than with the Donaldson camera (2.5 +/- 1.27 S.D., P < .01). Age was the only variable that correlated with stereoscopic quality for both cameras. The new Nidek camera provided significantly better stereoscopic quality images than the Donaldson camera and with significantly more reproducible optic nerve head assessments.