An electrocardiographic recording method with an associated reading guide, designed for epidemiological studies on Chagas' disease, was tested to assess its diagnostic reproducibility. Six cardiologists from five countries each read 100 electrocardiographic (ECG) tracings, icluding 30 from chronic chagasic patients, then reread them after an interval of 6 months. The readings were blind, with the tracings numbered randomly for the first reading and renumbered randomly for the second reading. The physicians, all experienced in interpreting ECGs from chagasic patients, followed printed instructions for reading the tracings. Reproducibility of the readins was evaluated using the Kappa (k) index for concordance. The results showed a high degree of interobserver concordance with respect to the diagnosis of normal vs. abnormal tracings (k= 0.66; SE 0.02). While the interpretations of some categories of ECG abnormalities were highly reproducible, others, especially those having a low prevalence, showed lower levels of concordance. Intraobserver concordance was uniformly higher than interobserver concordance. The findings of this study justify the use by specialists of the recording of readings method proposed for epidemiological studios on Chagas' disease, but warrant caution in the interpretation of some categories of electrocardiographic alterations