Dr. Donald Fredrickson 
December 27, 1977 
page -2- 
While the statement that everything non-novel is safe might seem supportive of our 
proposed revisions, it is in fact not the rationale for them and cannot be 
logically defended. We are all aware that the world is full of natural biological 
objects, untouched by molecular biologists, that constitute actual or potential 
hazards to human health or ecological stability. Natural as well as artificial 
genetic variation may either enhance or decrease the hazards in any particular 
case. The proper questions are therefore not whether DNA recombinants are safe 
but rather (1) whether bacteria carrying recombinant DNA are, as a class, less 
safe than bacteria picked at random from soil or sewage and (2) where specific 
DNA recombinants might pose any hazard, whether the risks involved differ signi- 
ficantly in degree or in kind from natural biohazards that are successfully handled by 
conventional methods. Where there is no novelty, the answer at least to the second 
question must certainly be negative; because whatever hazards might exist are with 
us already, and we can hardly protect the environment by trying to contain com- 
pletely things that are already occurring naturally all the time. 
3 . Are DNA recombinants expected to be dangerous ? 
The reasons for worrying about hazards can fall into two categories: (1) direct 
evidence (e.g., thalidomide causes birth defects in humans; DDT is bad for peli- 
cans) and (2) indirect arguments , that constitute reasonable extrapolations of 
existing data. A good current example of an indirect argument is the relation 
of fluorocarbon aerosols to the ozone layer. There is no direct evidence that 
the ozone layer has in fact been affected. But a great deal of hard chemistry 
backs up the prediction that damage to the ozone layer is a reasonable expec- 
tation. Another example is the effect on mankind of increasing the level of 
background irradiation. Below a certain level, no harm has been observed. We 
cannot point to any specific corpses as identified victims of low level radiation. 
But if we consider available data on humans at high doses, and animals and bacteria 
at low doses, in conjunction with physical and biochemical information on the 
mechanism of interaction of radiation with living cells we have reasonable cause 
for concern. 
The absence of direct evidence that recombinant DNA is harmful has been frequently 
mentioned, by Jim Watson among others. One can of course try to extend this data 
base still further, as is being done. However, many other human activities resemble 
recombinant DNA research in their apparent freedom from significant danger. Society 
does not routinely respond to every apparently harmless activity by demanding 
more extensive risk assessment data. Why is artificial recombinant DNA treated 
differently? 
I believe that one reason is a lingering misimpression dating back to Asilomar 
that someone did have a valid, perhaps profound, indirect argument, based on 
preexisting data or preestablished concepts from genetics, biochemistry, virology, 
epidemiology or evolutionary biology, that in some way predicted that artificial 
recombinants were more likely to be hazardous than natural organisms; i.e. , that 
the potential hazards of organisms with recombinant DNA were somehow more "potential" 
than those that are already inherent in living on a planet inhabited by other 
species of living organisms capable of natural variation. 
[Appendix A — 157] 
