and only one of these did not exhibit Jung cancer at death. A relationship observed between inittel Tung burden end time to death with cancer has been often used to infer a threshold burden below which no life shortening of dogs would be expected. This is shown in Figure 4.G.10 on 4-6 118. Note that the fibrotic deaths there have no bearing on cancer incidence wid inclusion of thes2 points in the constructing extrapolated curves is a senseless exercise. Tote also that the results are exhibited on a log-log graph which virtually obscures aT] differential detail. Most important, recognize tie nature of the experiment, i.e., the lung burdens were large, the results were saturated, and the number of animals was small. The _ crude relationship observed between initial lung burden and time to death with lung cancer does not necessarily imply that a threshold burden exists for beagles. Quite to the contrary, the range of exposures above the ‘inferred thresnold burden maybe interpreted as 4 region ofsaturated carcinogenic response, that is a burden regime in which lung cancer induction in a beagle populaticn approaches 100% during a normal life span. The point - is that the observed time to death is more likely related to the burden, through a population depletion effect, rather than through a burden dependent latent period. In the former interpretation appreciable cancer | would be anticipated zt lower burdens. This is again consistent with extensive observations of radioisotope-induced bone tumors in mice, which ‘support the interpretation that "latent period is constant and that the apparent relationship between increasing dose and decreasing time ts death with tumor is dee to the effects of dose-level on survival and on tumor expectancy." (See Toxicity of Ra-226 in Nice," M. Finkel et at, in Radiation-Induced Carcer, IAEA, Vienna, 1969.) The domain of this ccminent is broadened here in order to susmarize a8