Isometric responses of individual gastrocnemius muscle units of both fast (F) and slow (S) twitch type were studied during repetitive intracellular stimulation of the innervating motoneurons in anesthetized cats. Paired stimuli usually produced maximum responses (measured in terms of tension-time area) with interpulse intervals between 5 and 10 msec. However, longer pulse trains produced maximum tension-time area per pulse with relatively long interpulse intervals (referred to as the optimum interval). For each unit, stimulus intervals around the optimum produced maximum modulation of tension during unfused tetani. Post-activation potentiation produced shifts in the optimum toward longer intervals. The optimum intervals for type F units were shorter (mean 49.4 msec) than those for type S units (mean 86.9 msec), but there was no difference between the groups when the mean intervals were expressed in ratio with the twitch contraction time, Tc (F units: 1.42 X Tc; S units: 1.31 X Tc). The range of optimum intervals for all gastrocnemius muscle units corresponded to motoneuron firing frequencies of 7-40 pulses/sec, well within the frequency ranges observed for gastrocnemius motoneurons firing under natural synaptic drive. Insertion of a single short interval (5--10 msec) at the onset of a relatively low frequency stimulus train produced a prolonged 'catch-like' enhancement of tension output, particularly in type S units. The duration of the catch-like effect depended on the interval between pulses in the underlying basic train, and was maximum for intervals between 1 and 2 X Tc. The catch-like effect was also seen in type F units but its duration was curtailed by another process called the 'sag' property. The 'sag' property occurred in type F units over a wide range of stimulus intervals (usually 0.5-3 X Tc) but was present in type S units only with relatively long intervals (greater than 2 X Tc). The 'sag' property appears to represent a disenhancement process intrinsic to muscle unit fibers activated in unfused tetani.