SHENNAN (1973) drew attention to a strong correlation between coffee consumption and national rates of mortality from renal cancer (simple correlation coefficient r =079). We obtained similar results using incidence as well as mortality rates, and found also a strong geographical correlation between renal cancer incidence and the consumption of animal protein (r 0 82; Armstrong and Doll, 1975). The present study was therefore undertaken to investigate the association between these variables and renal cancer in individuals. We sought also information regarding exposure to other agents which have been reported to be associated with renal cancer in either man or animals, such as tobacco (Bennington and Laubscher, 1968;Schmauz and Cole, 1974;Wynder, Mabuchi and Whitmore, 1974), lead (Van Esch, Van Genderen and Vink, 1962), aromatic amines (Poole-Wilson, 1969), leather (Schmauz and Cole, 1974), compound analgesics (Johansson etal., 1974) and 9 sulphonamides (Hansen and Bichel, 1952).The number of cases studied is small (139) and is effectively made even smaller by the need to consider separately the 2 main types of renal cancer in adults (adenocarcinoma of the parenchyma and transitional cell carcinoma of the pelvis). Statistically significant results could not be expected, therefore, unless we had isolated a factor that greatly increased the risk of the disease. In fact, few of the factors examined appear to have any effect but, with the numbers involved, the possibility of small effects, of the order of two-or three-fold increases in risk, cannot be excluded. Renal cancer is, however, a sufficiently rare disease that it will be difficult for any one group of investigators to accumulate enough experience to provide decisive answers. We are therefore reporting our results with a minimum of description but have tabulated them in some detail so that they can be combined with other observations in the future.