PURPOSE To compare the repeatability of three measures of corneal thickness: Orbscan Slitscan pachymetry, ultrasound pachymetry, and optical pachymetry. METHODS Twenty normal subjects were tested on three occasions. Two occurred on the same day and the third was on a different day at approximately the same time of day as one of the first two visits. Central corneal thickness of the right eye was measured with a Haag-Streit optical pachymeter, a Humphrey Model 855 ultrasound pachymeter, and the Orbscan system. Day-to-day and same-day repeatability was assessed by calculating the difference between the values from two visits and determining the mean difference, the SD, and the 95% limits of agreement (LoA) (LoA = mean +/- 1.96 SD). RESULTS Mean (+/- SD) central corneal thickness as measured by each instrument was as follows: 539 +/- 33 microm for optical pachymetry, 542 +/- 33 microm for ultrasound pachymetry, and 596 +/- 40 microm for Orbscan pachymetry. For day-to-day comparisons, optical pachymetry showed the poorest repeatability with 95% LoA of -61 to +32 microm. Ultrasound pachymetry showed better repeatability with 95% LoA of -22 to +24 microm. The Orbscan showed the best repeatability centrally with 95% LoA of -10 to +17 microm. Peripheral Orbscan pachymetry was less repeatable than that measured centrally but still more repeatable than central optical pachymetry. Similar results were found with same-day comparisons. CONCLUSION The Orbscan system is the most repeatable technique for measuring corneal thickness but shows a significant bias toward greater corneal thickness measures than both ultrasound and optical pachymetry.