ing 20 million shelters that would be identified out of moneys funded in this fiscal year, but the supplies for which are not funded in this fiscal year.

Beyond that, he would expect that the results of the survey, to the extent they were available during the year, would give us a basis to project longer-range programs. And at this time his feeling is that we don't know enough, and the reason we are having the survey is partly that we don't know enough in order to project such programs. (See p. 131.)

CASUALTY ESTIMATES BY WAR GAMING

Mr. ROBACK. Will you explain to the committee how these casualty estimates were arrived at?

Mr. YARMOLINSKY. On that question, I think I will defer to Mr. Cannell, if I may.

Mr. CANNELL. I think this is probably one of the fundamental questions that we ought to explore a little bit, because it is this part that you have to understand in order to understand that the program we are proposing works. I notice from the testimony over the last several days that we each in the course of considering this program have visualized the situation in which the program did not work. And we visualized that the situation where it did not work was representive of the total Nation. So I think that if I were to go into a little bit of detail about the range of enemy intentions, how they target, how we estimate the prevalence of various postattack environments, that this might help resolve why the program saves the number of people we say it will.

Now, the approach that we have used in doing this is to consider a range of enemy intentions. We would say that the least attack we would expect, would be an enemy intention to strike at our retaliatory bases only. So around that intention, taking estimates of enemy capabilities, say, for the present and for future time periods, we would target as an enemy targeter might, to strike out or destroy our retaliatory bases.

Having done that, we would estimate what are the effects to population and resources from an attack of that nature.

Now, that establishes the lower limit of destruction.

Then we would say what is the upper end of the range of destruction the United States could receive?

And so we would say that the enemy might have the intention of not only destroying retaliatory bases, but maximizing the destruction to the people or economic capability or industrial resources. It happens that those three categories are roughly coincident. So you can use people as a unit of measure, if you wish.

And so we take the estimated enemy capability and target, maximizing the effect of his weapons on that target system.

The weapons are distributed quite differently, depending on what his intentions are.

So in using this approach, what we are attempting to do is find out what is the situation, if his intention is at the lower limits; what does the country look like after the attacks, and what does it look like at

the upper limit, what is the proportion of surviving people relative to surviving resources in each of these conditions?

Then we have to review this problem, not merely, say, a year from now when the program is partially implemented, but we have to review it some years out in the future to check its obsolescence and its effectiveness over time.

So we would set these upper limits not only for 1962, but we would set those limits out for 1967 or 1968 or 1970.

Now, between those limits there are any number of specific attacks. For example, one could design an attack to strike at our retaliatory bases, plus, say, our petroleum refineries, or some other specific resource. Any of those intermediate attacks or special attacks on retaliatory bases plus a specific resource lie between the limits that we have set between population at the top end of this scale of destruction and military retaliatory bases at the lower end.

Now, once we have targeted in this fashion, we use computer techniques to estimate the number of people who would be blast casualties, and who would be within various levels of radiation.

So that when we say that the estimates of those who would be saved by the program we propose lies between something on the order of 15 to 35 million, the reason there is that range is that there is that range of kinds of attack you could receive. If he struck at one target complex, you would only save the lower number. If he struck at another combination, you would save the upper.

Now, if we tie some numbers to these intentions just to give a rough feel for the proportion-and I would like the privilege afterwards of verifying the numbers I quote, to be sure I have remembered them right--if we were to receive a military attack in the next several years, total casualties might be on the order of 40 million people, that is, if it struck only at our retaliatory bases.

Now, because those bases in general aren't located within blast distance in cities, and it is only when an enemy missile misses a target like Offutt by some considerable distance and it lands in Omaha, that you have an expected kill by blast. Thus a very large portion of those 40 million casualties, something on the order of 5 million might be blast casualties, and would not be subject to being saved by this program we propose.

The other 30 million or so, a big majority of those in this light attack could be saved. So that when one says that something on the order of 25 or 30 million, which is about the number we would expect would be killed in that kind of an attack, would be saved, that means that you have made a major improvement over our present posture.

At the other end of the scale, then, when the targeting might include population, the number of casualties prevented are of the same order of magnitude. But the problem is that there are a lot of people this program does not save. It saved quantitatively almost as much, but there are many others it did not save.

EXISTING SHELTER AND POPULATION DO NOT COINCIDE

Now, I would like to go from this general picture of how we make the estimates to allude to some very penetrating questions you asked, Mrs. Griffiths, about the city of Detroit and the distribution of the

shelter relative to people, because to the extent that the shelter and the people are not located together, the program we propose does not work.

Now, the preliminary estimates we have made from the surveys that have been made indicate that the problem you were concerned about is not quite as serious as it should seem on the surface. We think that shelter in existing buildings will be in proportion to the distribution of people much closer than it appears at first sight. I would like to explain why.

If we used a city that would be similar to Detroit or Indianapolis, there would be somewhere around 15 or 20 cities in the Nation in that general geographic area or size. If we were to use that as an example, the central city, roughly a mile in diameter, has about 5 to 10 percent of the population in it. And you are right that the people in that area may not be the highest element of society, and at night you are saving out of proportion people in that central core and there is too much shelter there compared to the number of people. So the point you make is true.

But I would say just as an aside that I understand that these central core people are very good voters, they sometimes vote more than once. But aside from that, I would say that in the central core-it is not a very large area-this proportion of shelter is far higher than the proportion of people that are there in the off hours. But in the next concentric ring around it, if you went another mile out and included a 3-mile circle, in that ring is the apartment house area in cities of that category, and somewhere around 20 to 25 percent of the people live in that area.

Now, the preliminary review of the statistics that we have from sample surveys indicates that there is a fair balance of people to shelter in that ring.

The next ring out, which carries you to about 5 miles, is in about that same proportion. Thus there is a tendency for the density to be high toward the center of the city.

THE SUBURBS-MORE PEOPLE THAN SHELTER

Now, if you go out to a 10-mile ring—and Indianapolis city limits approximate that very closely-that includes about 85 percent of the population. It is in that area as you go from somewhere around from the 5- to the 10-mile limit that we begin to run out of shelter.

It is not that we want to visualize, as it almost came out in the testimony, the people in the periphery moving from the outskirts of town and evacuating toward the center, because that is the wrong visualization of the problem. There isn't enough shelter in the central city to take account of that mass movement.

ASSIGNING PEOPLE TO SHELTER A FEW BLOCKS TOWARD THE CENTER

There is, however, a tendency for the shelter to be concentrated somewhat more toward the center of the city than the people, which means that in an operational sense when people are assigned the shelter they should move to, it is possible for everyone to move a few blocks at the minimum, say, or the maximum, say, of a mile or so in the general direction of the center, so that in anyone's individual mind

he is not moving from the outskirts to the center, rather he is moving a very slight distance toward the center.

It is almost like a train with a lot of cars and the locomotive moved toward the center of the city and the caboose in to the next closer ring. The idea was to maximize the use of the shelter, but the ones on the outskirts didn't move downtown, they moved very little inward. It is not a mass movement of people all rushing toward the center of the town, it is a very slight movement for any one individual.

The problem we are trying to solve in surveying and marking is, where does the shelter run out relative to the people who need this shelter space, so that we can build or provide by some other means the shelter in the periphery area.

Now, I don't know whether this tends to answer some of the questions. But I think that there was one more aspect of the problem that didn't come across clearly to me in the testimony that I could elaborate on very slightly.

AIMING ERRORS WILL SAVE PARTS OF CITIES

We tend, each of us, when we think of the attack to visualize the bomb aimed directly at us and ourselves in the fireball.

Now, if we were to think of these two sheets of paper as representing one of these cities, if it were Detroit, it would be a little bit elongated, if it were Indianapolis it would be almost round, and if an enemy were targeting such a city, he would tend to pick the geographic center for his designated ground zero, and he would allow the aiming error to distribute the weapon. A five megaton weapon in that area happens to coincide, in its blast destruction with a city about the size of Indianapolis. To illustrate, if the weapon landed on the designated ground zero when the weapon misses we save people on some side of the city. Because of aiming error, the area of blast and fire and primary effects, leaves uncovered a number of people who can be saved by fallout shelters. And so our program is effective. We are not saying we saved those that are under the blast, we are saving those that were not under the blast, because the enemy cannot put the weapon down on designated ground zero.

Mr. HOLLIFIELD. And this very point, that we cannot predetermine where ground zero will be, suggests we must prepare a system of shelters on the premise that the majority of the people will not be in a ground zero locality, and, therefore, they will get the benefits of even a modest type of shelter.

COMPETITION FOR FUNDS

Mr. CANNELL. I concur with you completely, I am only saying that we are faced always with the problem of maximizing the effectiveness of our dollars and cents. And it happens that the first money spent can buy more lives this first year by using structures that are already available.

It is not that we are trying to say that we won't do these other things, but we don't want to have Von Braun beat us to the moon, we want you to give us the money so we can do it and not give it all to him; you have to give us a fair racing chance in this.

And our fair racing chance comes in being funded along with him. If you do fund it, we will sequence this program in such a manner as to buy the most lives with each incremental expenditure of money.

Mrs. GRIFFITHS. I will agree with you; I think it is remarkable that we are putting $40 billion on the theory of Von Braun reaching the moon while the Secretary of Defense has already stated that $10 billion for shelter is not a proper amount, is ridiculous.

Mr. HOLIFIELD. Is absurd.

Mrs. GRIFFITHS. Let me point out to you that the Secretary of Defense, handling this program surrounded by admirals and generals who are devoting their time and attention to thinking of a weapon that will knock down the enemy's weapons, is in direct opposition to an enemy that is thinking of a weapon that will overcome any defense that they can make.

Therefore, it is in my opinion absurd to spend the kind of money that we are spending on defense that will ultimately fail and refuse to spend any money or even set up any plans for creating a civil defense that would ultimately succeed for some people.

PROBLEMS OF SUBURBS AND RURAL AREAS

Now, you have spent a lot of time telling me about what would happen in Detroit, with which I agree, with this exception.

My district begins 8 miles from the downtown. It extends for some 15 or 20 miles from the downtown area. It is the second highest owner-occupied district in the United States, which means that they are all individual dwellings, 86.8 percent live in the individual dwellings. There are few buildings in the entire area that are as much as two stories. I think even some of the main new high schools do not have basements.

So that any kind of shelter program that you are going to decide upon must, in my opinion, consider this type of an area. And I am sure it is not the only such area in the United States.

Mr. CANNELL. I was trying to explain that it is quite a common problem. We have to designate and find all these areas so that the program can be planned on a very logical overall basis.

We can't at random pick these areas without accidentally skipping some. So the survey has to be in a very systematic manner to each time find where is our best option for saving lives with the next increment of money.

Mrs. GRIFFITHS. I would like to ask you a question on this.

I asked the Secretary the other day, and he said that he did not believe that people in rural areas had any greater chance of avoiding destruction than people in city areas.

Is this true, or is it not? I asked him if the people in rural areas wouldn't have a better chance of evading destruction than people in urban areas. And he just said no.

Mr. CANNELL. Of course, it depends entirely on how it is targeted. But from a fallout standpoint, it is not a great deal different to provide the shielding one place than another.

Mrs. GRIFFITHS. From a blast, then.

Mr. CANNELL. You can't target rural areas with a blast.