deaths were attributed to lung cancer in the United States among 1,452,454 from all causes i n a population of nearly 160,000,000 pe0p1e.l~ A survey in 1947 and 1948 by the National Cancer Institute in ten American cities reported (crude) pulmonary cancer-mortality rates per 100,000 white population of 15.4 in Birmingham, 15.5 in Pittsburgh, 15.9 in Dallas, 18.3 in Detroit, 18.6 in Chicago, 21.8 i n Philadelphia, and 22.5 in New Orleans; in three cities in which the figures for whites and nonwhites werc combined, the rates were 8.9 i n Atlanta, 14.8 in Denver, and 20.8 in San Francisco.15 It is not known to what extent these large differences are real but, in any event, the mortality is sizeable.
The risk can be expressed in another simple form. Of every 100 babies born in this country, nearly two will ultimately develop lung cancer, mostly between 50 and 80 years of age. I n any disease having such a small attack rate and only after so long a time, recognition of the etiological factors is extremely difficult, especially if they are multiple, if only intermittently present, and if the long interval from birth to disease represents the induction time from cumulative effects. T h e problem is quite different from that i n a n acute disease, such as measles, i n which nearly 100 per cent of persons are attacked, usually within the first ten years of life, shortly alter exposure to the etiological agent, antl by an agent that has multiplied and that is present when thc disease is manifest.
T h e problem of the etiology of lung cancer is greatly complicated by the probability that there are niultiple factors. T h e occasional tumors that occur in childhood (see reference 16 for references) are hardly attributable to the same agents as are those found i n uranium and cobalt miners o r in asbestos-, chromate-, nickel-, tar-, and other special-worker groups.10
T h e great majority of individual cases are