Relatively few attempts have been made to systemically examine the processes responsible for the ameliorative effects of response-contingent stimulation (RCS) on stuttering. It was hypothesized that the reductions in stuttering that frequently accompany RCS are the result of the stutterer being encouraged to access extant fluent speech that may not be fully evident during "contingency-free" (CF) conditions. A preliminary analysis of the hypothesis was conducted my monitoring RCS and CF stuttering frequency and speaking rate in 20 adult stutterers before, during and after a program of fluency training. Subjects were divided into "high" and "low" responders on the basis of their baseline response to the RCS procedure of time-out from speaking, after which they participated in a 32-hour program of fluency and training aimed at minimizing stuttering. After showing a degree of relapse during a subsequent 6-mon follow-up, high and low responders were found to be equally affected by time-out. This result contrasted the differential response shown by the two groups during the baseline phase, and is consistent with the hypothesis that improvements in fluency during RCS may occur when stutterers access extant fluent speech that is not otherwise being fully utilized.