Ill 
unfortunate, but not catastrophic. The British have had an extensive 
experience with that situation. Sir John might be able to refer to that. 
DR. SCHWARTZ: I don't think Windscale came close to a melt-down, 
though; whereas if they had been unable to cool the reactor, they would 
have been faced with the strong possiblity of melt-down, and the possible 
breach of containment which could have very well taken many lives. 
DR. NEEL: Well, you see, we are back into possibilities and prob- 
abil it ies . 
DR. SCHWARTZ: That is true. 
DR. FREDRICKSON: Mr. Helms. 
MR. HELMS: I am not enough of a mathematician to ask you the question 
properly, so bear with me here. We have in Appendix P of the documents 
attached to the Environmental Impact Statement that was published an excerpt 
from an article by Robin Holliday, in which he endeavors to set out the 
probabilities here, and I gather that that in part was the reason why, on 
page 99 of the thin Yellow Book, efforts were made to go through the A 
through M list of things that would have to happen before something escaped 
and harmed somebody. Now, are you attacking Holliday's approach to this? 
DR. SCHWARTZ: Specifically I have attacked Holliday's article on 
previous occasions. The Holliday article involves a lot of conjectural 
probabilities, to begin with. 
MR. HELMS: Well, I am trying to say now, if we could link his approach 
to the number of conditions, do you agree that the conditions on 99 and 
100 would have to occur before something — 
DR. SCHWARTZ: I am afraid I have not yet been sent a copy of that — 
MR. HELMS: Well, these are escape of foreign DNA, escape from the 
experimental situation, survival after escape, and so on, taking you through. 
DR. SCHWARTZ: Well, there are two problems here. One is that you 
have to identify all such sequences that can lead to escape. Now, the point 
of one of my quotations was that there are many malfunction sequences that 
have occurred in rocket testing and nuclear power plant operations that were 
not calculated. I mean, not only wasn't the probability assigned, they 
didn't even think of those possible sequences. So that is a very big problem, 
that when you are working without the benefit of actuarial data and you try 
to imagine all possible things that can go wrong and then compute the probabil- 
ities of all possible things that can go wrong, there is always a very great 
danger that you will forget about some possibilities. Who would think of the 
possibilities of workmen going around looking for air leaks with a candle in a 
nuclear plant? 
MR. HELMS: Let me ask you one more-- 
[ 315 ] 
