SUMMARY1. The length dependence of the mechanical latent period and early tension development, uncomplicated by latency relaxation, was determined from responses of single muscle fibres to restimulation at the peak of the isometric twitch at 15 'C.2. The onset of the reactivation response showed little (< 1x X 1O-7 N) or no latency relaxation.3. Reactivation latency was minimal (2-8 msec) and constant at 1P9-21 ,tm sarcomere length and it increased by about 3 msec with sarcomere extension to 3-2,sm.4. Reactivation responses showed two stages of early tension development, an initial phase in which tension acceleration increased, and a phase of maximum responsiveness in which tension acceleration was constant; the transition between the two phases occurred about 4-5 msec after the start of the stimulus at 2-2 jum sarcomere length and was delayed about 4 msec with increase in sarcomere length to 3-2 ,um.5. The square root of the maximum tension acceleration was directly proportional to the degree of overlap of thick and thin filaments in the sarcomere length range 2-3-3-2 ,um.6. It is proposed that the onset of the phase of constant tension acceleration marks the end of the period during which the activator, calcium, is distributed throughout the sarcomere.7. Analysis of early tension transients in relation to myofibril structure showed that length-dependent changes in reactivation latency and time of onset of constant tension acceleration were probably brought about mainly by alteration of the kinetics of distribution of activator within the myofibril and by changes in the diffusion distance between activator-release sites near the end of the sarcomere and the tension-generating sites.8. There was a 2 msec myofibril priming delay in the rise of tension in twitch responses that was not seen in reactivation responses; the possible origin of that delay is discussed in relation to structural changes accompanying activation and to competition between calcium-binding structures.9. The onset of twitch latency relaxation occurred within about 250 ,gsec after the time corresponding to latency of the earliest reactivation responses and appeared to signal the start of a process that took place after the arrival of calcium among the myofilaments. The origin of latency relaxation is discussed.