Analysis of 71 patients with atrial flutter delineated two categories of patients. One group (40 patients) had severe decompensated heart disease. Another group (31 patients) was free of structural heart disease or had disease not severe enough to result in congestive failure. Of the latter group 23 had an acute noncardiac illness. In 14 it was a disease of the bronchopulmonary system. Those patients with severe heart disease seldom reverted without electrical countershock during the first five days after hospital admission. The majority of the latter group reverted within five days without such therapy. Those observations are useful to the clinician who must select therapy for his patient with atrial flutter. Failure to appreciate this variability may lead to conflicting or misleading reports regarding the efficacy of newer forms of therapy.