When a frog’s or toad’s sartorius is rapidly released during a maintained isometric contraction its tension drops immediately and is redeveloped later. The extent of release required to reduce the tension to zero is 3 to 4 % of the length of the muscle. This is much less than the 10 to 15 % originally stated by Gasser & Hill:​ the difference is explained. The amount of work done during release by the passive elastic element in series with the contractile element is affected only very slightly by speed of release:​ the damping of this element is small. The redevelopment of tension after release has been compared with the original development of tension when the stimulus began. Minor and variable differences only have been observed, and these are probably due to redistribution of length, during isometric contraction, between different regions of the muscle. At greater initial extensions the rise of tension during an isometric tetanus is much slower than at smaller initial extensions. This also is attributed to redistribution of length, within the muscle. At an initial extension not greater than that at which the developed tension is a maximum the system is ‘stable’ and the tension reaches its full value sharply:​ at extensions on the outer side of the maximum the system is ‘unstable’ and a long slow creep of rising tension occurs. The apparent complexity of the time-course of the heat production in an isometric twitch is discussed.