We evaluated the use of a least-squares radionuclide separation technique to allow an increased number of myocardial blood flow measurements with radionuclide-labeled microspheres in dogs. Two sets of labeled microspheres were studied: a set of eight labeled with 125I, 153Gd, 57Co, 51Cr, 113Sn, 85Sr, 95Nb, and 46Sc; and a set of nine in which 125I and 46Sc were replaced with 114In, 54Mn, and 65Zn. For each microsphere label the nuclide activities determined by least-squares separation compared favorably with those actually added to in vitro samples containing a fixed amount of the other nuclides in the set. For the set of eight radionuclide-labeled microspheres, myocardial flow measurements made with the least-squares separation technique and the reference sample method were usually within 15% and almost all within 20% of direct measurements of coronary venous outflow in a right heart bypass preparation. Serial left atrial injections of 15-micron microspheres totaling 48 X 10(6) caused no significant changes in systemic hemodynamics, regional myocardial flows, or coronary pressure-flow relations, whether the coronary bed was autorelating or vasodilated with chromonar. We conclude that at least nine myocardial blood flow measurements can be made in dogs with acceptable accuracy and without evidence of dysfunction due to embolization of the coronary vascular bed. With appropriate validation, this method should be applicable to other organs and animal models as well.