The results of a retrospective study of 3,937 cancer patients are presented, with controls matched by age and interviewer. A list of sites associated with smoking is established, including the lip, tongue, oral cavity, oropharynx, hypopharynx, larynx, lung, esophagus, and bladder. By considering the different criteria of consumption, we were able to separate these sites into two groups: (a) In cancer of the lip, oral cavity (excluding the tongue), oropharynx, hypopharynx, and esophagus, neither inhalation nor preference for cigarettes seems to be significant. (b) Inhalation and preference for cigarettes are significant for cancer of lung, larynx, and probably bladder. In the sites listed in (b), inhalation was an important factor, independent of the amount smoked or the method of smoking. The particular significance of the cigarette is less definite; it depends, in part, but in part only, on inhalation, which is associated with this type of smoking. To give a complete explanation of the significance of cigarettes, we have included other considerations: Smokers of the cigarette-only, pipe-only, and pipe-and-cigarette categories differ in certain respects and perhaps in their manner of smoking in general. The effect of inhalation appears, in this study, to be specific. Thus it is significant in some sites of cancer but not in others, and the study of a group of noncancerous patients brings out the relationship between inhalation and certain diseases.