115 
We are now dealing with the opposite problem, and I might say, perhaps 
the reciprocal, Dr. Schwartz, of the mathematical problem you are approach- 
ing. We are dealing with an organism that we know has accumulated numerous 
deficiencies in 50 years of passage, and in EK2, half a dozen additional 
deficiencies, and I submit that no accident could restore all of those 
deficiencies together. You can wreck a thing by a single monkey wrench in 
the works. You can't make a complicated machine by throwing a monkey wrench 
in the works. 
DR. FREDRICKSON: Thank you. 
DR. SCHWARTZ: May I — 
DR. FREDRICKSON: No, Dr. Schwartz, please. 
DR. SINSHEIMER: May I respond to Dr. Davis? 
DR. FREDRICKSON: Dr. Sinsheimer, you may; you are a member of 
Commi t tee . 
DR. SINSHEIMER: It seems to me, Bernie, that the point is, of 
you can't make a complicated thing work by throwing a monkey wrench 
at least I hope not. But it seems to me that the question that Dr. 
was addressing can be put the following way. We say that an EK2 strain has 
10 8 probability, and we feel happy that that is a hundred million-fold less 
hazardous on its own. But in reality, it seems to me that is probably not 
going to be the true improvement. The true improvement is not going to be 
the survival of something like 1776. It is going to be more what is the 
probability that when- you are growing up a culture of 1776, it gets con- 
taminated with a more robust strain, the plasmid gets passed to it, and 
like Don said, some experiments had been done that that didn't happen, but 
how many experiments? What is the probability? Is it 1 0 ~ 2 ? Or 10~ 3 ? So 
in other words, some other route around it would be more probable than 10 -8 . 
That is all. 
DR. FREDRICKSON: Professor Rosenblith. 
PROFESSOR ROSENBLITH: Without associating myself with either the model 
of throwing a monkey wrench in or out, I would like to hope that NIH was 
supporting in this area studies by people who work in the area of risk assess- 
ment, and who make mathematical models of them, and would give us a variety of 
ways of looking at these processes. I think that seems to me one of the ways 
in which one can responsibly further this argument in addition to the experi- 
ments that are being carried on. 
DR. FREDRICKSON: I think now, if we may, I would like to turn to the 
subject of experimental Guidelines. We'll have the opportunity this evening 
to revisit some of these issues, provided we now speedily move through another 
section. I think we can do this without compromising the opportunity of 
the 
course 
in , or 
Schwartz 
[ 319 ] 
