11 
A. Experiments that should not be performed - We recognize that it 
can be argued that certain of the recombinants placed in this category 
could be adequately contained at this time. Nonetheless, our estimates 
of the possible dangers that may ensue if that containment fails are of 
such a magnitude that we consider it the wisest policy to at least defer 
experiments on these recombinant DNAs until there is more information to 
accurately assess that danger and to allow the construction of more effective 
biological barriers. In this respect, these guidelines are more stringent 
than those initially recommended (1). 
We therefore strongly advise that the following experiments not be 
initiated at the present time: (i) Cloning of recombinant DNAs derived 
from the pathogenic organisms in Classes 3,4, and 5 of "Classification 
of Etiologic Agents on the Basis of Hazard" (5), regardless of the vector- 
host system used, (ii) Deliberate formation of recombinant DNAs containing 
genes for the biosynthesis of toxins of very high toxicity (e.g., botulinum 
or diphtheria toxins), (iii) Deliberate creation from plant pathogens of 
recombinant DNAs that are likely to increase virulence and host range. 
(iv) Widespread or uncontrollable release into the environment of any 
organism containing a recombinant DNA molecule unless a series of con- 
trolled tests leave no reasonable doubt of safety, (v) Transfer of drug 
resistance traits to microorganisms that are not known to acquire them 
naturally should be deferred if they could compromise the use of a drug 
to control disease agents in medicine or agriculture. 
In addition, we recommend that at this time large-scale experiments 
(e.g., more than 10 liters of culture) with recombinant DNAs known to 
[ 82 ] 
