Dr. Donald Fredrickson 
December 27, 1977 
page -3- 
Such an indirect argument might be very specific, like the possibility of cloning 
a latent virus during a shotgun experiment; or very general, like the notion 
that among the array of recombinants cloned there will be some that disrupt the 
balance of nature. Either type of argument might justify special precautions 
for recombinant DNA research, provided it is spelled out in a form sufficiently 
explicit so that one can analyze it step by step and see whether one agrees 
with the assumptions and the logic. 
Jim Watson summarized the situation with respect to his own specific concern at 
the time of the moratorium - that of disseminating latent tumor viruses - and 
pointed out that the whole idea really didn't make sense when subjected to 
further analysis. Note that this is something quite different from the numerous 
allegations that recent data have diminished some people's fears about recombinant 
DNA. As I understood him, Jim was saying that his initial idea was basically a 
dumb one; that what was mainly needed was more thought, not more data. At any 
rate, this is a good example of a specific indirect argument that failed to pan 
out . 
On a much more abstract level, there is the general notion that, even though the 
typical DNA recombinant that turns up in a shotgun experiment is probably unable 
to survive in competition with natural organisms, exceptional recombinants may be 
even more fit than wild bacteria; and that if molecular biologists do enough 
cloning, they will eventually encounter such exceptions. The argument can be 
extended, as it was at Falmouth, to say that, even if all K-12 recombinants are 
unfit, there is a small chance that the cloned DNA will be transferred out of 
K-12 into some more robust bacterial strain, which the cloned DNA would render 
even more fit than it had been. Among these rare recombinants that are positively 
selected, some might have undesirable properties. Thus one can argue that the risk, 
however small, is not zero; especially as the magnitude of the "undesirable pro- 
perties" is undefined. 
The problem with this reasoning, if evaluated as serious scientific theory, is 
that it does not address the control situation at all; i.e., it never asks 
whether the expected rate at which disruptive gene combinations might arise in 
recombinant DNA research is greater than their rate of origin by natural genetic 
variation in bacterial populations untouched by human hands. I don't see this 
question as in any sense trivial; and if anyone else does, I hope I can one day be 
enlightened as to why. It's clearly true that, if I grow a bacterium harboring 
cloned DNA from a shotgun experiment, that bacterium might harbor gene combina- 
tions which, when transferred to some other bacterium, might convert the second 
bacterium into a pest or pathogen. But it's also true that, if I pick a random 
bacterium from the nearest sewer, that bacterium might harbor a natural plasmid 
which, when transferred to some other bacterium, might have the same result. 
In that second (natural) case, one can argue that most such transfers generate 
bacterial types that have already appeared, so that only a few plasmids arising 
as fresh mutants or recombinants would generate new problems; but that limitation 
is counterbalanced by the fact that a bacterial plasmid is much more likely 
to function in harmony with the recipient bacterium than a gene from some more 
distant source. It ends up as a draw, at least at present. (The fact that 
combinations from very distant sources are on the one hand more novel and on 
the other hand less likely to succeed than combinations from more closely 
related species has in fact led to some inconsistencies in the present Guidelines; 
[Appendix A — 158] 
