UNIVERSITY OF CALIFORNIA, BERKELEY 
r 
' BERKELEY • DAVIS • IRVINE • LOS ANGELES • RIVERSIDE • SAN DIEGO • SAN FRANCISCO 
DEPARTMENT OF BACTERIOLOGY BERKELEY, CALIFORNIA 94720 
AND IMMUNOLOGY 
February 23, 1976 
Dr. Donald S. Fredrickson 
Director 
National Institutes of Health 
Bethesda, Maryland 20014 
Dear Dr. Frederickson: 
I want to congratulate you on the quality of the DNA Recombinant 
meeting. The background science and the rationale for the proposed 
guidelines were clearly explained, and the insiders and outsiders in 
the field were given every opportunity to express their opinions. The 
result was an amazingly good-willed and serious concentration on the 
real issues. 
In my opinion these issues cannot be resolved in a permanent way 
at present because the available data do not allow adequate evaluation 
of the relative potential for benefit and harm. (Phil Handler expressed 
it well when he commented that he had listened to all speakers and agreed 
with each one). On the other hand, controls are needed at present for 
the pursuit of the critical data and for the prevention of underground 
uncontrolled experimentation. On this basis it seems to me paramount 
that : 
1. A set of rules representing the best bets of the moment be adopted 
as temporary working orders. These should certainly not be pro- 
cessed as formal regulations and they could include a self-destruct 
clause by mandating review and revision every year. Nor should 
the first set of regulations be too rigid since they could result 
in oversell to the nonscientif ic community and prevent the possi- 
bility of future relaxing modifications. 
2. The NIH consider making on a crash basis facilities for pioneering 
experiments which will clarify the dangers and difficulties. These 
might even be experiments not allowed under the present rules, but 
their aims must clearly be to provide the information for better 
rules or provide the safe "organisms" and procedures which can be 
used in the field. These facilities might properly be at Fort 
De trick. 
[ 485 ] 
