CONELDENTHAT or most desirable course of action is about 25 percent for a single weapon but about 50 percent when four weapons are employed (for a 4000-m cep, Fig. 19). When the cepis 12,000 m (Fig. 20) the differences are 15 percent for one bomb but 45 percent for four bombs. ) (f) Perhaps thé most outstanding finding from these comparisons is that cities differ in the optimum civil defense action they can take. The relative effectiveness of these actions varies with their size, population distribution, road net, and probable warning time. Thus for Boston undergroundshelter results in fewer casualties underall conditions of cEP and attack size (Figs. 19 and 20), whereas for Dayton minimum or maximum evacuation is preferable for small cep, and 4-hr (but not 1.5-hr) evacuation for large cep (Figs. 21 and 22). For the remaining targets, evacuation seems preferable when the cxp is small, and underground shelter best when the cep is large; however, the magnitudeof differences due to taking one or anothercivil defense action is by no means uniform. 100 -— | —T 80}- 7 z eee Se ee ee a @ wi a. ao Best available shelter = vi 40 |- " wi a 20+ /. 1-hr evacuation . Q-hr evacuction > oo ee 0 Underground shelter | | | 1 2 3 4 10-MT WEAPONS rs Fig. 28 — Percentage Washington Resident Population Killed by One to Four 10Mt Weapons, cer 12,000 m DEATHS FROM FALLOUT Method of Study It should be emphasized that none of the preceding comparisons considerthe probabilities of additional deaths from indirect radiation (fallout). To afford some insight into the dimensions of this threat a model attack was developed that delivered from two to four 10-Mt weapons on the Washington-Baltimore targets. A sample of 137 upper wind readings was drawn at random from 1955 us Weather Bureau teletype information. The sample composed about three-fifths of the available readings. These were separated into summer and winter readings and into low-, medium-, 40 ORO-—R~17 (App B)