The indicator-dilution method for measuring left ventricular volume is reviewed as to theory, historical development, indicators and sensing devices employed, results obtained and criticisms of the method. This method is a further development of the indicator-dilution technic originally described by Stewart in 1894 for studying the circulation. In measuring ventricular volume, indicator is injected into the ventricle, and its concentration in the ejected blood is recorded immediately down-stream to the aortic valve by means of a system having a short dynamic response-time. The record obtained is a logarithmically decaying, step-function, indicator concentration-beat curve. Since the quantity of injected indicator is known, the end-diastolic, end-systolic and stroke volumes can be calculated if it is assumed that the aortic blood concentrations measured from beat to beat represent the concentrations in the ventricle before each ejection. Studies employing numerous indicators (concentrated sodium chloride solution, T-1824, indocyanine green, cold solutions and radio-active I-131-labeled Diodrast) on a variety of mammals ranging in body size from the horse to the rabbit, and including man, have shown that the left ventricle normally ejects slightly less than half of its end-diastolic volume with each beat. This ejected fraction is frequently reduced in cardiac disease. Although the volumes and residual fractions obtained by the wide variety of indicator-dilution technics are in general agreement, the end-systolic volumes and residual fractions estimated by means of the angiocardiographic technic are smaller. Possible reasons for these discrepancies are discussed.