An asymptomatic adult population of 196 men and women was studied with the echocardiogram to derive age- and sex-specific "normal" values for a number of clinically used echocardiographic variables. The results are in general agreement with previously published normal values. Body position during the examination, age and sex influence the echocardiographic results; body surface area correction normalizes most of these effects. The prevalence of occult abnormalities determined by the echocardiogram is 7%; the most common finding was mitral valve prolapse. Inter- and intraobserver variability was assessed. The interobserver differences found on analysis are statistically, but not clinically, significant. The echocardiogram appears to be a suitable tool to use in epidemiologic studies to detect selected cardiac abnormalities, but is limited for this purpose because some subjects in such a population cannot be adequately examined.