- 6 - 
such hosts can only be judged in very hypothetical and general terms. 
The hypothetical bacterium A which I have used in my analysis could 
well be, for example, Rhi zobium sp. or Agrobacterium tumefaciens , both 
of which are transformable, although not as efficiently and reproducibly 
as JE. col i . 
I would like, therefore, to suggest that a provision is made in the 
Guidelines to allow for the circumstances under which conjugative plasmids 
may be introduced into HV-1 procaryotic hosts carrying "novel" recombinant 
molecules of procaryotic origin to perform, for example, triparental crosses, 
but to require that the physical containment level for such experiments 
be raised one step. Perhaps experiments involving DNA from certain classes 
of pathogenic organisms could be excluded from such a provisions. 
I feel that such a provision can be justified on several grounds: (i) 
one could use additional safeguards to avoid co-transfer of the mobilizing 
plasmid to the recipient which receives the recombinant plasmid (e.g. 
suppressor-dependent or temperature-sensitive Tra" mutants, or plasmids which 
establish mating pairs but do not stably replicate in the recipient host); 
and (ii) although rarely pointed out in discussions of biohazards, the effec- 
tive prohibition of, or undue restriction on, in vitro genetic experiments 
will have adverse effects from the standpoint of biosafety in genetic exper- 
iments with bacteria in general: by effectively preventing the use of the 
cloning method through regulations such as those I have just described, re- 
searchers will naturally resort to classical in vivo methods such as conjuga- 
tional crosses in their work. Needless to say that such experiments, although 
not covered by the Guidelines and presumably occur naturally, are not biohazard 
free. Many researchers, including those of us working with plant pathogenic 
[Appendix A — 115] 
