1. The firing thresholds of single myelinated fibres of frog sciatic nerves were monitored as a function of impulse activity in the fibre. The threshold was given by the number of coulombs in current pulses that excited a particular fibre half the time when delivered to the whole nerve. Threshold was tracked by a device that incrementally decreased the number of coulombs in the current pulse whenever the fibre responded and increased the pulse if it did not respond. 2. There was a pattern to the after-oscillations of threshold following activity. The fibres were briefly refractory, transiently superexcitable for about 1-1.5 sec and then entered a phase of raised threshold or 'depression' that lasted for many minutes. 3. Activity produced little change in the threshold curve during the refractory period. Strong depressions following prolonged activity prevented the threshold from returning to the base-line level within the time associated with the refractory period for the same fibre at rest. 4. After an impulse, superexcitability reached a maximum within 7-20 msec. This peak was larger as the number of impulses in a preceding burst increased and as the intervals between the impulses became briefer. Each successive impulse of a burst contributed less to the growth of superexcitability, and after the burst had 6-10 impulses additional impulses contributed nothing. 5. The depression phase was marked by the interaction between build-up, which depended on the activity rate, and recovery, which required as long as an hour or more for the threshold to be completely restored to resting level. These two mechanisms, one causing build-up and the other recovery, led to formation of dynamic equilibria. The threshold level at equilibrium increased monotonically with the activity rate. 6. The processes associated with superexcitability interact with those producing depression. In active fibres showing raised thresholds, impulses are followed by a relative superexcitability that persists for at least as long as an absolute superexcitability (with threshold below the resting level) can be measured in the same fibre at rest. 7. The duration of the superexcitable phase interpreted as a relative change in excitability was roughly the same regardless of the level of depression. 8. The magnitude of the oscillation in threshold was give to ten times larger than the grey region (the range of stimuli for which response is probabilistic). It is concluded that at regions of low conduction safety such as axonal branches, where weak forces can influence whether an impulse will pass, such pronounced and long-lasting after-effects of firing can be expected to modulate conduction of nerve impulses.