To estimate variations in intra- and interindividual measurements of the corrected QT (QTc) interval, duplicates of 50 twelve lead electrocardiograms (100 photocopies, paper speed 50 mm/s) were given to each of nine investigators in random order. The electrocardiograms were recorded from patients with acute myocardial infarction consecutively admitted to a coronary care unit. Patients receiving drug therapy and those manifesting various arrhythmias were included. Two-way analysis of variance was used to evaluate the results from all 900 QTc measurements. Significant differences in these measurements were registered among investigators and were of major importance (p less than 0.001). This finding illustrates the difficulty in comparing mean values from different studies and emphasizes the difficulties in applying limits for a normal QTc interval to data obtained by different observers. Of less but still significant importance was the interaction between the investigator and electrocardiogram (p less than 0.001). Finally, the random error was calculated and proven to be of no importance (less than 0.5 mm) when more than 11 measurements were performed.