We measured the characteristics of the decline in tension during isometric relaxation of canine papillary muscle. In the intact heart, relaxation begins with the isovolumic phase, but in experiments on papillary muscle previously reported the isotonic phase preceded the isometric phase during the course of relaxation. In our experiments, however, the isotonic bar was locked at the instant the muscle reached the end-systolic fiber length in order to hold the fiber at the length during the succeeding relaxation process. Therefore, we obtained a relaxation process similar to that occurring in the intact heart. The major results of these experiments are: (1) Maximum rate of the decline in tension (-dT/dtmax) is linearly related to the magnitude of total load. (2) -dT/dtmax is augmented by positive inotropic interventions and diminished by negative inotropic interventions. (3) An increase in preload results in only a slight increase in -dT/dtmax. (4) End-systolic fiber length itself is not a principal determinant of -dT/dtmax. (5) -dT/dtmax divided by total load is independent of the amount of muscle shortening. We, therefore, suggest that -dT/dtmax divided by total load cand be a useful index of the relaxation characteristics of cardiac muscle.