61 
One of these trends has consisted of the general tendency to increase 
the levels of containment applied to experiments with recombinant DNAs, and 
this can be illustrated by considering the fate of certain shotgun experi- 
ments as they have passed through the successive guidelines, starting with 
the letter that appeared in Science about one and a half years ago. 
The experiments I should like to consider are those in which random 
segments of DNA obtained from eukaryotic chromosomes are cloned in IS. coli . 
In the Science letter, these experiments were not included within the group 
for which we asked voluntary deferment. Rather, we limited our comment to 
the general admonition that they should be carried out with caution as they 
might engender certain hazards, some of which we noted. the general intent 
was to stimulate awareness in the scientific community of the potential 
dangers. 
In the next step, these experiments were divided into certain classes, 
and levels of containment were recommended for each class, first by a work- 
ing committee whose job it was to prepare a report for consideration at the 
Asilomar conference, and then by the organizing committee of that conference. 
Certain increases in containment levels occurred during that process. For 
example, the working committee recommended that experiments with the mam- 
malian DNAs, including those from primates, be carried out under conditions 
equivalent to P3 plus EK1. You can remember these terminologies. These 
were increased to a containment comparable or slightly higher than P3 plus 
EK2 by the organizing committee — a classification that was ratified by the 
conference participants. 
In subsequent consideration by the NIH Advisory Committee, the con- 
tainment conditions for these mammalian experiments were left at P3 plus 
EK2 in the Woods Hole guidelines, but were again increased last December 
at La Jolla. Thus, experiments with primate DNAs were separated from the 
general mammalian class and assigned P4 plus EK2, or P3 plus EK3 contain- 
ment levels, while the remaining experiments with mammals were retained at 
the P3 plus EK2 level. 
In a similar move, experiments with cold-blooded vertebrates, which had 
remained constant at P2 plus EK1 from Asilomar through Woods Hole were in- 
creased one level, that is to P2 plus EK2 at La Jolla. 
This trend to increase containment levels would clearly not be of con- 
cern to me were it based on data indicating that these experiments were 
potentially more dangerous than we had previously imagined. However, this 
was not the case. Indeed, the data that has accumulated since the Science 
letter was pubished argue in the opposite direction. These data include the 
finding that _E. coli K12, the strain of EL coli used in these experiments, 
is highly attenuated. It does not colonize in normal bowels of humans or 
[ 202 ] 
