The only specific cancer for which there was evidence for a positive association with external radiation was lung cancer (see Tables VII and VIII and Fig. 2B). There were two lung cancers in the highest dose group and exclusion of high dose data results in a LRT statistic of 0.43 for the multiplicative dose-response model. The estimate of the lung cancer ERR per Sv is 1.68 with a likelihood ratio based 95% CI of (0.03, 4.94). Results in Table VII show that there was a strong SES effect, that baseline rates were higher for Y-12 and multiple facility workers, and that the relative risk for workers monitored for internal exposure was slightly elevated. Information on cigarette smoking for this cohort is not available for analysis and residual confounding by cigarette smoking cannot be ruled out---see Wing al for further discussion and references. Such confounding could bias dose response estimates in either direction. There was no evidence for an association between diseases of the circulatory system or nonmalignant respiratory disease and external radiation (see Tables VII and VIII and Figs. 3B and 3C. Note, however, that the largest LRT statistic in Table VIII occurred for emphysema for the low dose data (there were two deaths in the 320 to 640 mSv dose group with 0.16 expected based on U. S. rates).