It has long been known that the administration of digitalis to patients with heart disease produces striking clinical improvement in many cases. Students of cardiology have long been occupied with two closely related problems: (a) a differentiation of those cases that are improved, from those that are not benefited by the drug; (b) an explanation of its action, and an understanding of the manner in which it produces improvement. Little progress was made toward a solution of these problems until recent years, and even now, although advancement has been rapid, no more than a fair beginning has been made. A few facts have been established. It is well known that digitalis causes considerable delay in auriculoventricular conduction in many patients, and careful studies1have indicated that prolongation of a-v conduction of greater or less degree is a constant effect of the drug in all cases. The exact mechanism by which