Measurements have been made of the changes in adenosine triphosphate and inorganic phosphate associated with activation and shortening in isolated sartorius muscles of Rana pipiens at 0 degrees C. The muscles were pretreated with 2,4-dinitrofluorobenzene to inhibit adenosine triphosphate: creatine phosphotransferase. When unloaded muscles were stimulated electrically, allowed to shorten fully and stay shortened during further stimulation, it was found that the usage of adenosine triphosphate (and production of inorganic phosphate) per pulse became less and less at greater rates of stimulation. The adenosine triphosphate usage per second reached a plateau and continued for a short while after the last pulse. This adenosine triphosphate is presumably used almost entirely for pumping calcium. Very little adenosine triphosphate was used per pulse by muscles in hypertonic solutions. These muscles passed normal action potentials but did not shorten or develop tension. Muscles which were re-extended after each of a series of almost completely unloaded shortenings had been completed used only the small amount of extra adenosine triphosphate needed for the little extra work done. Even the total breakdown of adenosine triphosphate was much too small to account for the heat of shortening, if this heat is assumed to be degraded free energy from adenosine triphosphate. Experiments were carried out according to Professor A. V. Hill's 'Further challenge to biochemists'. They showed that less adenosine triphosphate was used in the first 200 ms by a lightly loaded isotonic muscle than by an isometric muscle. Thus the heat of shortening, which is observed only while the muscle is shortening but not in the whole contraction-relaxation cycle, cannot be degraded free energy from adenosine triphosphate.