Three studies examined the impact of legibility on compliance with smoking regulations. Waiting areas at a large airport were viewed as behavior settings (R. G. Barker, 1968) in which smoking rules could be communicated more clearly by applying K. Lynch's (1960) notions of legibility. Study 1 identified 3 types of areas: clearly no smoking, clearly smoking permitted, and ambiguous. As predicted, ambiguous areas had significantly more smokers than clearly defined no-smoking areas. In Study 2, legibility was varied systematically: Both legibility and user location in the areas contributed to users' knowledge of the rules, and there were more smoking violations in illegible areas. Study 3 indicated that when smoking occurred in a no-smoking area, people were more likely to act as maintenance mechanisms when in a legible area and when in the center of the area. Furthermore, over time, in legible no-smoking areas they became more assertive, whereas in illegible no-smoking areas they tended to leave the area. Evidence suggests that visual cues can successfully communicate the setting program in the absence of setting leaders, and behavior setting research can benefit from an understanding of psychological processes underlying legibility.