The reproducibility of intraocular pressure (IOP) measurement with the Goldmann applanation tonometer was investigated as part of a population-based epidemiological study. Sixty-two subjects were examined in a first measurement session. The IOP was measured three times consecutively in both eyes according to a fixed protocol. The mean standard deviation (SD) of these measurements was 0.8 mmHg. The mean intraobserver variation for the first measurement was 1.64 (SD 2.07) mmHg. For the median of the three measurements the intra-observer variation was 1.50 (SD 1.96) mmHg. The mean inter-observer values were 1.79 (SD 2.41) mmHg for the first measurement and 1.60 (SD 2.15) mmHg for the median measurement. The correlation coefficient between the median values of the three measurements of both observers was 0.81. No systematic differences were found between the two observers. Using the median value of three consecutive measurements reduced the inter-observer variation by 11% and the intra-observer variation by 9% compared with a single measurement.