PURPOSE Videokeratoscopes provide a wealth of information about the topography of the ocular surface. Although there have been numerous studies of the accuracy and precision of videokeratoscopes with inanimate test objects, little information exists on their precision (repeatability) for real eyes. METHODS To investigate the stability of the ocular surface in the inter-blink period, 10 patients were recruited for videokeratoscopy. Tear break-up time was measured and videokeratographs were acquired immediately post-blink and again at 4, 8, and 12 seconds post-blink. To permit statistical inferences to be drawn from the data, we acquired 24 videokeratographs for each of the four post-blink intervals. The videokeratograph data were interpolated (bilinear) to a common grid, and average and standard deviation (SD) maps were derived for each post-blink condition. t Tests were used to test the significance of changes observed in the topography. RESULTS The instantaneous power SD maps showed increasing variation toward the periphery, with most maps showing less than +/-0.5 diopters (D) of SD in the central 4 to 5 mm and variation in the periphery often reaching more than +/-1 D SD at the edge of an 8-mm diameter. When the 4-, 8-, and 12-second average maps were subtracted from the average map acquired immediately after blinking, regions of statistically significant ( p < 0.001) change were apparent in the upper and lower regions of the maps. The upper and lower bands of change were found to correlate with the natural position of the patients' lid margins. CONCLUSIONS For normal eyes, the central regions of videokeratographs show high stability in the inter-blink period. However, the upper and lower edges of 8-mm diameter maps show statistically significant variability, which appears to be related to the effects of eyelid pressure.