- 2 - 
aitning directly at the construction of highly virulent primary hosts (e.g. the 
botulinum toxin experiment) or other operations of high potential risk, and 
(ii) assigning containment levels for all permissible experiments in such a 
manner that the probability of events a, b, c, or d above are lowered to an 
acceptable, reasonably safe level. 
The following paragraphs contain comments pertaining to general as well 
as specific aspects of the proposed revisions. Some of these comments 
pertain specifically to experiments with plant pathogenic bacteria, in which 
I have an active interest, and with which special problems have existed and 
new ones have arisen in view of the proposed revisions. 
1) Redefinition of recombinant DNA molecules. The proposed new definition is 
a meaningful one, indeed, and should be retained. 
2) Proposal to regulate only those experiments involving "novel" recombinant 
molecules . This is also a welcome change. It recognizes a basic conceptual 
dichotomy which existed in the original guidelines: the intent to regulate 
strictly one way of constructing a particular bacterial strain ( in vitro ) but 
not another ( in vivo ). Unfortunately, elements of this dichotomy have per- 
sisted in the proposed revisions (see under 3 below). 
Given the present state of affairs, in which neither the Committee nor 
the currently pending bills in Congress propose to regulate in vivo genetic 
manipulations, it is quite obvious to me that many kinds of in vitro cloning 
experiments which involve procaryotic DNA and host-vectors, (e.g. the cloning of 
a particular type of DNA in a host in which the DNA is normally present, or 
by which it can be acquired naturally) involve biohazards of a similar nature 
and magnitude to those inherent in many "standard" in vivo manipulations. 
Accordingly, I quite agree with the basic idea of distinguishing between 
experiments involving "novel" and "non-novel" molecules. 
[Appendix A — 111 ] 
