In a recent article by one of us1 attention is called to the errors arising from theoretic deductions on the effect of digitalis in disorders of cardiac rhythm. Without a more exact understanding of the precise manner in which digitalis in varying doses affects the different cardiac functions, and also without a clearer understanding of the precise mechanism producing the different disorders of rhythm, it is necessary, in order to ascertain the response to digitalis, to investigate each arrhythmia directly. It is a matter of common experience to observe premature contractions, particularly of the ventricular type, called forth by digitalis. The ventricular premature contraction is especially common in patients with auricular fibrillation. In some of the latter, large doses are necessary; in others, rather small doses are sufficient to produce them. Our knowledge of the effect of digitalis on premature contractions occurring spontaneously is not quite so definite. In 1911