=5— aoe? extrapolation to low dose rates is not only not ccenservative for alpha radiation induced tumors, but rather that there is a marked inverse dose-~ rate vs risk relationship. There is an increasing body of published experimental evidence that reflects this trend. Speiss and Mays‘??) observed that for 72“Ra alpha radiation induced bone sarcoma in man, the t:mcor incidence per rad approximately doubled for a four- fold increase in the spacing of ?7"Ra injections and that the observed incidence of bone tumors per rad in children was nearly twice that for adults. et al. (23) Upton show a significantly higher incidence of tumors in mice for a Biven neutron dose at more protracted periods of exposure. Moskalev and :. . Buldakov 24) showed that fractionation of the administered **°Py dose over larger periods of time increased bone tumor induction. The higher tumor incidence per rad for the smalier. lung burdens of crushed ?3°Pu0, micro _ spheres observed by Sanders (11) seems best explained by the limited alpha -drradiation of large numbers of cells by numerous very small, mobile particles of low activity per particle (see below). Hamsters subjected to low alpha doses from *!°Po distributed quite homogeneously in the bronchiolaralveolar region show a marked increase in the lung tumor incidence per rad at very low doses and dose rates (29) | And the incidence of bronchial ceuncer in uranium minere ‘reflects a higher tumor risk per rad at the lower doses ‘2®) for this low dose rate exposure group. The tobacco radioactivity results 1) indicate a significant tumor risk for the cumulative alpha radiation dose ‘from 7}°po in insoluble particles in the bronchi of smokers, involving much lower dose rates. Based on the above considerations it is evident that thetumor ris« is optimized when a very large number of cells and their descendants are subjected to only a few widely spaced alpha interactions with the snall