Purpose.To assess the repeatability of measurements of ocular aberrations using wavefront sensing in a small group of observers and to assess the potential effect of measurement error on custom corneal correction due to this variability.Method.A Shack-Hartmann wavefront sensor was used to measure the ocular wavefront in nine eyes. Head position was stabilized using a dental bite bar, and the pupil was centred using a cathode ray tube monitor and circular grating. Twenty Shack-Hartmann images were collected for each measurement. Each observer had three sets of measurements taken; the first and the second after careful alignment and the final after regrasping the bite bar in the same position as for the second measurement, but without pupil realignment. The modulation transfer functions for each set were calculated, and the effect of best-aligned custom treatments on the modulation transfer function was estimated.Results.There were highly statistically significant differences in a large number of Zernike modes between the three sets of measurements. The modulation transfer functions calculated for the residual wavefronts after aligned custom treatment were below the diffraction limit. The root mean square wavefront errors were consistently better for the residual wavefronts obtained using the realigned data than using data taken without pupil realignment.Conclusions.Sequential measurement of ocular aberrations shows statistically significant differences in a large number of Zernike modes. If aberrations determined by a single measurement are to be used in a custom correction, the resulting modulation transfer function is likely to remain below the diffraction limit. Pupil realignment is critical in reduction of the residual root mean square wavefront values to a minimum.