After 24 hours of preservation under 15 mmHg perfusion pressure the recovery rates of isolated canine hearts were determined. Preservation was performed in a cold room maintained at 4*C with 4 different types of perfusates bubbled with a mixture of 95% 0y and 5% CO~ using a modified perfusion unit designed in our institute. The perfusates used were as follows; Group 1: Krebs-Henseleit solution, Group 2: Krebs solution added by albumin and PGE1. Group 3: Modified Wicomb*s solution, Group 4: Modified Collin*s solution. The extent of myocardial recovery was evaluated using a modified isolated carmine perfusion model by measuring heart rate, systolic arterial pressure, left atrial pressure[LAP] and cardiac output. In addition to the above hemodynamic parameters, biochemical and enzymatic assays from perfusates and electron microscopic changes of the myocardium were also studied. The results were as follows; 1] The heart recovery rates were 41.6%, 53.4% and 108.9% in groups 1, 2 and 3, respectively, and group 3 elicited the best result[p< 0.001]. The heart beat was never recovered in group 4. 2] Recovered systolic arterial pressures[mmHg] were 63.3% in group 1, 94.9% in group 2 and 94.3% in group 3. 3] LAPs[mmHg] were 20 in group 1, 13.5 in group 2 and 11.2 in group 3, which suggested that the best myocardial preservation was elicited in group 3[p< 0.05]. 4] Cardiac output, the sum of aortic stroke volume and coronary leakage, were 69.1% in group 2, and 90.7% in group 3, but these were not statistically significant[p=0.24]. No aortic stroke output was measured in group 1 and 4. 5] The degree of myocardial edema increase was 17.5` in group 1, 24.6% in group 2, 20.9% in group 3 and 55.3% in group 4. But there were no statistical differences in each group[p= 0.08]. 6] CPK-MB[U/L] levels were increased 750% and 332%[p< 0.05], glucose levels[mg/dl] 60.5% and 78.2% and SGOT[U/L] levels 523% and 333%, in groups 2 and 3, respectively. Biochemical and enzymatic assays could not be performed in group 1 and group 4, because of poor recovery of heart beat. 7] Electron microscopic findings in the myocardium of most groups revealed slight to moderate muscle cell and mitochondrial edema. But all these findings were within the limits of reversible change. From these above results, it is suggested that modified Wicomb*s solution seems to be the most useful physiologic salt solution for preservation of the heart. We propose that after further study and improvement, our portable continuous hypothermic perfusion system will contribute to the development of a better preservation method for donor hearts for human heart transplantation.