The goal of this investigation was to assess the accuracy of 18F-fluorodihydrorotenone (18F-FDHR) as a new deposited myocardial flow tracer and compare the results to those for 201Tl. Methods. The kinetics of these flow tracers were evaluated in 22 isolated, erythrocyte- and albumin-perfused rabbit hearts over a flow range encountered in patients. The two flow tracers plus a vascular reference tracer (131I-albumin) were introduced as a bolus through a port just above the aortic cannula. Myocardial extraction, retention, washout, and uptake parameters were computed from the venous outflow curves using the multiple indicator dilution technique and spectral analysis. Results. The mean initial extraction fractions of 18F-FDHR (0.85 +- 0.07) and 201Tl (0.87 +- 0.05) were not significantly different, although the initial extraction fraction for 18F-FDHR declined with flow (P < 0.0001), whereas the initial extraction fraction of 201Tl did not. Washout of 201Tl was faster (P < 0.001) and more affected by flow (P < 0.05) than 18F-FDHR washout. Except for initial extraction fraction, 18F-FDHR retention was greater (P < 0.001) and less affected by flow (P < 0.05) than 201Tl retention. Reflecting its superior retention, net uptake of 18F-FDHR was better correlated with flow than 201Tl uptake at both one and fifteen minutes after tracer introduction (P < 0.0001 for both comparisons). Conclusion. The superior correlation of 18F-FDHR uptake with flow indicates that it is a better flow tracer than 201Tl in the isolated rabbit heart. Compared to the other currently available positron-emitting flow tracers (82Rb, 13N-ammonia, and 15O-water), 18F-FDHR has the potential of providing excellent image resolution without the need for an on-site cyclotron.