2 
At present, the hazards may be guessed at, speculated about, or voted 
upon, but they cannot be known absolutely in the absence of firm exoeri- 
mental data--and, unfortunately, the needed data were, more o^ten thar 
not, unavailable. Our problem then has been to construct guidelines that 
allow the promise of the methodology to be realized while advocating the 
considerable caution that is demanded by what we and others view as 
potential hazards. 
In designing these guidelines we have adopted the following principles, 
which are consistent with the general conclusions that .-/ere r ormu' ! ated at 
the International Conference on Recombinant DNA v olecjles r e’d at Asi'onar 
Conference Center, Pacific Grove, California, in February '975 '3): 
(i) There are certain experiments for which the assessed ootentia' nazard 
is so serious that they should not be attempted at the present time. 
(ii) The remainder can be undertaken at the present time provided that 
the experiment is justifiable on the basis that new knowledge or ze r e~'ts 
to humankind will accrue that cannot readily be obtained by use of cc r - 
ventional methodology and that appropriate safeguards are i r corporated 
into the design and execution of the experiment. Ir addition to a r 
insistence on the practice of good microbiological techniques, t^ese 
safeguards consist of providing both physical and biological barriers to 
the dissemination of the potentially hazardous agents, (iii) ~' n e level 
of containment provided by these barriers should match the estimated 
potential hazard for each of the different classes of recombinants . r or 
projects in a given class, this level should be highest at initiation and 
modified subsequently only if there is a substantiated change in the 
[ 73 ] 
