The echocardiographic anatomy of double-inlet single left ventricle was studied in 57 patients, aged 1 day to 27 years (mean 6 years); the variables examined included morphology, size and function of the atrioventricular (AV) valves and ventricular septal defect and their relation to pulmonary stenosis, aortic stenosis and aortic arch obstruction. The visceroatrial situs was solitus and the heart was in the left side of the chest in all 57 patients. A d-loop ventricle was present in 21 patients and an l-loop ventricle in 36. The great arteries were normally related (Holmes heart) in 8 patients and transposed in 49. In all hearts, the right AV valve was anterior to the left AV valve. In 53 patients, the tricuspid valve (right valve in d-loop and left valve in l-loop) was closer to and had attachments on the septum. The tricuspid valve straddled the outflow chamber in eight patients. No significant difference was noted in the mean AV valve diameter when comparing mitral and tricuspid valves within the same group or between the groups with a d- or l-loop ventricle. The right AV valve diameter had a significant direct correlation with the aortic valve diameter and the size of the ventricular septal defect regardless of ventricular loop. Both AV valves were functionally normal in 34 patients. Among patients with AV valve dysfunction, the tricuspid valve tended to be stenotic in patients with an l-loop ventricle and regurgitant in patients with a d-loop ventricle. Mitral valve dysfunction was uncommon. The ventricular septal defect (46 patients) was separated from the semilunar valves in 24 patients (muscular defect) and adjacent to the anterior semilunar valve as a result of hypoplasia or malalignment, or both, of the infundibular septum (subaortic defect) in 19 patients. Multiple defects were present in three patients. The defect was unrestrictive in 26 patients, restrictive in 23 and could not be evaluated in 8. Pulmonary artery banding had been performed in 8 of the 26 patients with an unrestrictive defect and in 10 of the 23 patients with a restrictive defect. Only 4 of 19 subaortic defects compared with 16 of 24 muscular defects were restrictive. The size of the defect was significantly correlated with the measured pressure gradient. Among patients with transposition, only 2 of 13 with pulmonary stenosis had a restrictive ventricular septal defect compared with 15 of 30 without pulmonary stenosis. In patients with transposition, the defect size was significantly smaller when coarctation was present.