The right ventricular ejection fraction (RVEF) can be shown theoretically as a mathematical function of the percent shortening in the 3 axial dimensions of the right ventricular cavity (the septum-free wall dimension (SF), the anterior-posterior dimension (AP), and the tricuspid valve-apex dimension (TA) or the long axis dimension (LA)). There is a need to decide which mechanism is the most important for the RVEF in cases with neither obvious regional wall motion abnormalities of the left ventricle nor right ventricular overload. Forty-four consecutive subjects (34 males/10 females) were enrolled: 16 had normal hemodynamic parameters without significant coronary artery stenosis, 15 had hypertrophic cardiomyopathy and 13 had dilated cardiomyopathy. Biplane right ventricular cineangiography was performed and the percent shortening of the SF, AP, and TA or LA were measured. The percent shortening in the SF (34.8+/-14.7%) was larger than that of the AP, TA, and LA (23.2+/-8.5, 21.0+/-8.3 and 18.3+/-7.0, respectively; all p<0.001). There was a linear correlation between the percent shortening of each dimension and the RVEF. The 95% confidence interval of the regression equation from the percent shortening of the SF and RVEF was located above those from the other percent shortenings, except for a lower RVEF. These results indicate that systolic shortening of the SF (ie, bellows action) plays an important role in the RVEF except for a lower ejection fraction.