In a recent paper we reported that there are considerable differences in the body weights of men categorized by their smoking habits (Khosla and Lowe, 1971). In a population of 10,482 steel workers we found that the men who had never smoked were much heavier than the smokers and that the men who had given up smoking for eight or more years were almost as heavy as those who had never smoked. The difference between the weights of non-smokers and smokers increased with age to the extent that over 40 years of age it was about 13 lb (5 9 kg).Our findings are in general agreement with two earlier surveys reported in the literature. In a largescale study of coal workers at three Scottish collieries, Ashford and his co-workers (1961) found that non-smokers were consistently heavier than smokers at each of six age groups, and at 40 years of age or over the difference was about 10 lb (4 5 kg). A survey of 8,548 Norwegian seamen and 9,114 men ashore (Natvig and Vellar, 1965) showed similar body weight differences in relation to smoking habits in the middle-age groups.However, Waller and Brooks (1972) found little difference between the body weights of non-smokers and smokers, surveyed while attending a public health exhibition in the city of London, and Pincherle (1971) has published a table showing four broad weight groupings (expressed as a percentage of average weight for age and height, calculated from the Metropolitan Insurance Company data) for the members of the Institute of Directors with the comment that the weight distributions could not be discriminated according to smoking habits. Because the Institute members are all from social classes I and II, it was suggested that the disagreement with our own findings might be attributable to differences in the trends of obesity and smoking habits according to social class. This is a point of some importance, for 78 % of men in our study were drawn from social classes III and IV. In this paper, therefore, we examine our data on the weights of the steel workers in relation to their social class as well as to their age and smoking habits.