Radiopaque markers implanted in the left ventricle allow one to follow the motion of specific elements of myocardium and compute a measure of relative left ventricular volume, called "eigenvolume." Workers in our laboratory partially validated this measure in excised dog hearts by comparing eigenvolumes with volumes of balloons in the left ventricle. Perhaps because the balloon did not fully fill the left ventricle, there was a significant nonzero intercept in the relation between eigenvolume and balloon volume. Therefore, the results only partially confirmed the eigenvolume theory because eigenvolume should be proportional to cavity volume. This study compares eigenvolume with left ventricular volume measured by biplane angiography in intact beating hearts following hemodynamic interventions (atropine, phenylephrine, nitroprusside, volume loading, and propranolol). We performed 32 angiograms in five dogs. A high correlation (r = 0.958 +/- 0.029 [SD]) was found between eigenvolumes and left ventricular volumes. In contrast to earlier in vitro validations and consistent with theory, the intercept of the relation between eigenvolume and left ventricular cavity volume (angiovolume) was not significantly different from zero. These data strengthen the hypothesis that eigenvolume accurately reflects true left ventricular volume in vivo. These same data were used to assess the immediate (i.e., within a few beats) effect of angiography. Left ventricular volume loading associated with the angiogram modestly (6%) increases left ventricular end-diastolic and end-systolic eigenvolumes without changing stroke volume or pressures, consistent with a mild negative inotropic effect.