To investigate how quickly Goldmann applanation tonometers used in clinical practice develop calibration errors, and to determine the frequency of checks required to detect these errors.Prospective check of the calibration error of all Haag-Streit Goldmann applanation tonometers in the department at month zero, month one, and month four. The tonometers were checked according to the Haag-Streit method using a standard calibration check weight bar by two independent observers. Calibration errors were classed as +/-0.5 to 2.5 mm Hg, +/-3 to 4 mm Hg, or >+/-4 mm Hg. Tonometers with a calibration error greater than +/-2.5 mm Hg were returned to the manufacturer for re-calibration.At month zero 2 of 34 (5.9%), at month one 3 of 29 (10.3%), and at month four 0 of 33 (0.0%) tonometers fell within the manufacturer's recommended calibration range of +/-0.5 mm Hg. A total of 14 of 34 (41.2%) tonometers at month zero, 10 of 29 (34.5%) tonometers at month one, and 17 of 33 (51.5%) tonometers at month four were identified to have calibration errors greater than +/-2.5 mm Hg.Goldmann applanation tonometers are not as accurate as the manufacturer's recommended calibration error tolerance of +/-0.5 mm Hg would suggest. Calibration error of less than +/-2.5 mm Hg is clinically acceptable. Calibration error checks should be carried out once monthly and tonometers with calibration error greater than +/-2.5 mm Hg returned to the manufacturer for re-calibration. Additional checks should be made if tonometers suffer specific damage. Ideally individual ophthalmologists should check calibration before each session.