take DNA from birds or plants. But it is difficult to see how 
philogenic order or any other single factor can accurately pre- 
dict risk when so many different factors (thirteen, according 
to the draft EIS) determine the likelihood that a particular organism 
will cause harm. It is possible to imagine an experiment with 
primate DNA which is less hazardous than one with prokaryote 
DNA. For example, an experiment which put primate DNA into 
E. Coli might produce an organism with such a low probability 
of survival and such a low probability that the primate DNA would 
be fully expressed and create a primate protein, that it poses 
virtually no risk to humans. Another experiment in which DNA 
from a prokaryote was inserted in E. Coli* might improve the sur- 
vival ability of E. Coli , be transmitted to a pathogenic strain 
of E. Coli , and make it more virulent, thereby substantially 
increasing the risk of disease in humans, whatever single factor 
is picked to rank hazards, it will always be possible to make 
a credible argument that some experiment which is rated low “risk 
is in fact more risky than something which was rated higher-risk. 
Some of the assumptions underlying the classification 
system are contradictory. In some situations higher containment 
levels are required because we know almost nothing about the 
hazards; in other situations, stricter containment is required 
because we have concrete information about the hazards. For 
23 
Appendix K — 1A9 
