After conditioning and extinction of keypecking with 25-sec intertrial intervals between key illuminations, an immediate change to 5-sec intertrial intervals reinstated pecking during trials. Brief illuminations of the chamber during intertrial intervals also temporarily restored extinguished keypecking. The same manipulations of trial tempo and chamber illumination usually weakened well established, still reinforced keypecking. No recovery of extinguished behavior occurred when the intertrial interval was shifted upward from 5 sec to 25 sec. These and other behavioral findings were examined in relation to (a) Pavlov's and Skinner's research and views on "disinhibition" and "external inhibition" and to (b) analogous phenomena in physiological studies of habituation and dishabituation. The reliable evidence of disinhibition obtained in the present experiments suggests the involvement of an inhibitory process in extinction.Disinhibition refers to the temporary reinstatement of an extinguished CR when an extraneous stimulus is presented (Pavlov, 1927). Demonstrations of this phenomenon, as well as the phenomenon of spontaneous recovery, helped convince Pavlov that an active inhibitory process is involved in the loss of a CR during extinction; extinction does not erase the excitation established by prior reinforcement, but builds up an inhibitory force that counteracts the excitation present in the situation. Furthermore, he viewed disinhibition as the "inhibition of an inhibition" rather than as the addition of some new excitatory effect to the situation. In support of the former alternative, he demonstrated that the same extraneous stimuli which produce disinhibition of an extinguished CR would usually decrease a well established and still reinforced CR ("external inhibition").Most experimental examples of disinhibition have been obtained with Pavlovian conditioning procedures. Skinner (1938) tried but was unable to demonstrate disinhibition of operant behavior (for example, by tossing rat Ss in the air during extinction of leverpressing). Similarly, Boakes (1973) was unsuccessful in his attempts to use a flashing light to disinhibit extinguished keypecking in pigeons. Skinner believed that evidence of disinhibition in operant conditioning would bolster the argument of other workers that a concept of inhibition is needed in analyzing behavior. His failure to obtain disinhibition was one of the reasons he concluded that inhibitory processes need not be postulated.