CALIFORNIA INSTITUTE OF 
TECHNOLOGY 
PASADENA, CALIFORNIA 91100 
DIVISION OF BIOLOGY 
February 12, 1976 
Dr. Donald S. Fredrickson 
Office of the Director 
National Institutes of Health 
Bethesda, Maryland 20014 
Dear Don: 
I have been reviewing my letter to you of February 5 in the light of our recent 
meeting to see what statements I might wish to change or what I might add. 
I still am highly inclined to caution. The various potential hazards cited, 
for instance, by Wallace Rowe - while troubling enough - took into account only 
the potential for mischief that might be produced by extant gene combinations, 
when transported into or expressed within unfamiliar habitats. Little attention 
seems as yet to have been given to the potential hazards that might derive from 
the wholly new combinations of genes we shall create (or afford new opportunity 
for creation) by recombinatory events. 
I am not persuaded by the argument that the opportunities for the creation of 
these potentially dangerous gene combinations already exist in Nature - at some 
unknown frequency. Such events may indeed occur sporadically - and with what 
consequence? (New viruses do_ appear from time to time - whence?). Clearly we 
will be increasing the opportunities for such gene combinations by many, many 
orders of magnitude. I believe my analogy with the peril from radiation 
(despite our adaptation to background radiation) is apt. 
Nor am I persuaded that "biological containment" is adequate. I do not rest 
easy on Roy Curtiss' 10”-*- 9 style numbers because I expect other events with 
which he has not reckoned will come into play at much higher probabilities 
(i.e. the original cloning culture will be contaminated; cells will lyse in a 
pharynx and the DNA be taken up by another organism, etc. etc.). 
The irreversibility of the process - which can only increase over time as more 
and more laboratories and persons become involved - continues to disturb me. 
Let me try to make a few suggestions as to directions I would think might allow 
research to proceed yet provide a broader safety belt. I think it easiest to 
work backward from the conceivable applications (benefits). 
[ 443 ] 
