The diagnostic yield from ambulatory continuous ECG recording (Holter ECG) has been evaluated retrospectively in 200 consecutive patients referred from an urban population of 250,000 for arrhythmia diagnosis or evaluation. About half of the patients were in hospital, the other half were ambulatory. In 20/200 (10%) the examination was obviously useful: indication for therapy was documented or arrhythmia and symptoms did not coincide. 53/200 (26.5%) possibly benefited, with qualitative or quantitative observations of clinical interest, and 127 (63.5%) probably did not benefit (no previously unknown findings of importance). Of the 85 patients referred for syncope, six obviously benefited and ten possibly benefited from the examination.