Behavioral and event-related brain potential (ERP) measures were used to assess the effects of tobacco smoking on selective attention. Two groups of abstinent smokers performed a Stroop color-naming task. The display color of a stimulus word determined the correct response, whereas word meaning was irrelevant. Meaning was congruent, neutral, or incongruent with respect to color. After completing two blocks of trials under abstinent conditions, subjects received a 15-min break before performing two more blocks. Subjects in the Smoking group (N=12) smoked two cigarettes during the break. Matched Control subjects (N=12) did not smoke during the break. Typical Stroop effects were found, as reaction time (RT) was shortest to congruent words, intermediate to neutral words, and longest to incongruent words. Overall RT decreased after the break equally for the Smoking and Control groups, whereas the magnitude of the Stroop effect was unchanged for either group. P300 amplitude decreased after the break for the Smoking group but not for the Control group, which implied that smoking rather than practice produced component decline. Error rate and P300 latency did not change after the break for either group. The results suggest that tobacco smoking may decrease the availability of general attentional resources required to evaluate colored word stimuli, whereas the specific stimulus processing mechanisms responsible for the Stroop effect are relatively unaffected.