CONFIDENTIAL iD Effectiveness Attacks with one to four 10-Mt ground-burst weapons aimed at the population centers of the target cities with circular probable errors (ceP) ranging from 4000 to 12,000 m were made, and the proportions of the target population killed were computed when the civil defense tactic was: seeking the best shelter now available, evacuating radially outward for a period of time equal to 1956 and 1959 expected warning times, and seeking underground Results indicated that mass evacuation is not as effective in reducing casualties as underground shelter when multiweapon attacks, large inaccuracies in delivery, or fallout from hearby targets are taken into account. Mass radial evacuation, if the population is unshielded, precludes the use of atomic warheads in antiaircraft guided missiles. The use of best existing shelter, attenuating radiation by 0.9, is the least desirable course of action, resulting in higher proportions of target population killed than either mass radial evacuation or underground shelter. Deaths in this case are due principally to blast and thermal effects and to the fact that lethal radiation doses are received by shelter occupants before rescue workers — impeded by debris and high radiation — can reach them. -m private or public shelter. Public and private shelters appear to provide the best protection from all effects and to tactics). If the us and ussr have equal capabilities in air offense and defense, a decided advantage will go to the nation with the best passive defense system. wo give military forces the greatest flexibility to meet the attack with any warhead at any altitude. The effectiveness of shelter close to the population need not be dependent on the successful functioning of the distant-early-warning network (as is the case with evacuation In this regard it appears that the nation that can place its population in shelter possesses a basic advantage — an ag-' gressor may be greatly deterred if he cannot be sure of striking a truly crippling blow. A shelter program would be particularly effective in this connection if it were accompanied by a gradual reduction in urban vulnerability. Compared to an estimated $50 million in direct costs for an evacuation program for 170 major cities, the range for shelter systems may be from $6 billion to $30 billion, but shelter construction costs need be sustained only once and maintenance costs are negligible. Also, shelter systems can contribute to the area’s economy and welfare by serving dual purposes, e.g., for subways and below-grade parking areas, without losing appreciable value as shelter. CONCLUSIONS Feasibility 1. The feasibility of mass evacuation has not been proven by actual test in any large city. There are serious difficulties facing realistic practice on the necessary scale. Current survival plan projects in a numberof large cities may find ways of overcoming the difficulties. 2. Mass evacuation of large cities to smaller villages and towns is not feasible within expected warning times. 9 ORO-R-17 (App B) “> Costs