Twenty-nine smokers were assigned to one of three groups: contingent punishment, non-contingent punishment and attention-placebo. The contingent punishment group had hot, smoky air blown in their faces as they smoked. The non-contingent group had fresh air blown in their faces while they were smoking and received the hot, smoky air while they were not smoking. The control group went through the same procedures as the contingent group except that they received no smoke. All Ss were seen for 7 treatment sessions spaced over three weeks and were followed up one month later. There were 10 trials (cigarettes) per session. A three-way analysis of variance using the number of cigarettes smoked before, during, and after treatment showed a significant trials effect but no treatment effect or trials X treatment interaction. A rise in smoking rates was noted at the follow-up period but smoking rates were still significantly lower than baseline. Trial duration and ratings of trial unpleasantness were not related to treatment success. The results of this study considered with previous use of aversive procedures indicates that contingent punishment is of limited value in the control of smoking. © 1969.