The geometric or anatomic diagnosis of mitral valve prolapse, as opposed to the pathologic diagnosis of myxomatous valve disease, is based on the relationship of the mitral leaflets to the surrounding anulus. Current echocardiographic criteria for this diagnosis include leaflet displacement above the annular hinge points in any two-dimensional view; implicit in this equivalent use of intersecting views is the assumption that the mitral anulus is a euclidean plane. Prolapse by these criteria is found in a surprisingly large proportion of the general population. In most of these individuals, however, prolapse is present in the apical four-chamber view and absent in roughly orthogonal long-axis views of the left ventricle. This frequently observed discrepancy between leaflet-annular relationships in intersecting views suggests an underlying geometric property of the mitral apparatus that would produce the appearance of prolapse in one view without actual leaflet distortion. To address this possibility, a model of the mitral valve and anulus was constructed. When the model anulus was given a nonplanar, saddle-shaped configuration, the clinical observations were reproduced: the leaflets appeared to lie above the low points of the anulus in one plane, and below its high points in a perpendicular plane. Therefore, the appearance of mitral valve prolapse can occur without actual leaflet displacement above the most superior points of the mitral anulus if the anulus is nonplanar. To determine whether this pattern is reflected in the human mitral anulus, two-dimensional echocardiographic views of the mitral apparatus were obtained by rotation about the cardiac apex in 20 patients without evident annular or rheumatic valvular disease. In all cases the mitral anulus, as reconstructed from these views, had a nonplanar systolic configuration, with high points located anteriorly and posteriorly. This is consistent with the findings of other groups in animals, and would favor the appearance of prolapse in the four-chamber view and its absence in long-axis views that are oriented anteroposteriorly. This model can therefore explain the frequently observed discrepancy between leaflet-annular relationships in roughly orthogonal views. It challenges the assumption that the mitral anulus is planar as well as the diagnosis of prolapse in many otherwise normal individuals based on that assumption.