Previous studies have shown that smoking i s associated with a high incidence of certain malignancies and a high incidence of metastatic spread of melanoma. The purpose of the present study was t o examine whether this high incidence of malignancy could be associated with certain aspects of immune function believed to be important in restricting tumour growth. Age-and sex-matched smoking and non-smoking normal subjects and male, smoking and non-smoking melanoma patients, were studied for the natural killing (NK) activity of their blood leukocytes against cultured melanoma and Chang cells. The levels of the various immunoglobulin classes in their sera and the E rosette levels of the normal subjects were also assessed. The results indicate that the N K activity of blood leukocytes from both normal subjects and melanoma patients who smoked was significantly lower against cultured melanoma cells than that of non-smokers. Smokers were also shown t o have lower IgG and IgA immunoglobulin levels i n their sera compared t o nonsmokers but no differences i n the percentage of E-rosetting (T) cells was detected. Recent studies provide some basis for the belief that the low N K activity and immunoglobulin levels i n smokers may be related. These results further suggest that a closer examination of the effects of this environmental hazard on the immune system and i t s relation t o malignancy i s needed.It is well known that smoking is associated with a raised incidence of a variety of malignant diseases such as carcinoma of the lung, oesophagus and bladder (Doll and Peto, 1976). It has also been shown that there is a higher incidence of metastatic disease, after 5 years, in male patients with malignant melanomas who smoke, than in those who are nonsmokers (Shaw et al., 1979). The underlying cause for this association between smoking and the increased incidence of malignancy is unknown and it is probable that multiple factors may be involved, including direct carcinogenic effects of products formed from tobacco during smoking.One of the possible associations, however, that has received little attention, is that related to the effects of smoking on the immune system. There is much experimental evidence from studies on both man and animals to show that smoking causes changes in the immune system (Holt and Keast, 1977) and surveillance by the immune system is considered important in control of tumour emergence and growth.In particular, recent studies have suggested that natural killer (NK) activity may play an important role in tumour surveillance (Kiessling and Haller, 1978; Baldwin, 1977; Hersey, 1979). Research originating in this laboratory has provided additional support for a surveillance role for NK cells in that melanoma patients with low post-operative NK activity against melanoma target cells had a significantly higher incidence of recurrence of melanoma than did patients with higher post-operative NK activity (Hersey et al., 1978).The present study was undertaken to determine whether an association between smoking a...