A train of regularly spaced rectangular pulses of voltage, amplitude modulated at a frequencyf,,, and applied to the control grid of a commercial valve working I conventionally, caused space current changes of up to 0.01 amp/cm2 from the oxide-coated cathode. An unmodulated pulse train, interlacing the first with a delay of td, sampled the anode current; the samples contained a component at the frequency fn, of low level, The component decayed as td increased, with a time constant of about 1 msec. I t was sometimes in phase with the component obtained when td=O, and sometimes in antiphase, The magnitude and phase were often dependent on the history of the valve. Thus an operation known to poison the cathode was usually followed, temporarily, by a large antiphase component.An in-phase component denotes an enhancement of emission following a positive pulse; an antiphase component denotes fatigue. There is evidence that the impedance at the interface between cathode core and coating is not responsible for the effects. The fatigue may well be largely a surface, rather than a bulk or matrix, effect. I t bears some relationship to effects reported elsewhere with pulses involving current densities of 10 amp/cm2.