121 
1 EK 1 and EK 2 strains of E. coli cannot escape from the 
2 laboratory by usual routes. Recently, unexpected results 
3 have shown that the EK 2 strain can survive spillage much 
4 better than had been anticipated. This indicates that 
5 caution is still in order, at least until the full results 
6 of the study become available in a year or so. 
7 It seems to us appropriate that the scientific 
3 data on this safety question be evaluated before rather than 
9 after giving hundreds of laboratories around the country 
10 the go-ahead to lower their precautions. 
n Another closed meeting, held at Ascot this year, 
12 served as an ostensible basis for the drastic downgrading 
^3 of containment requirements for recombinants involving animal 
14 virus DNA , including tumor viruses. Actually, the conference 
15 reportedly recommended P 2 containment, but an American 
16 subgroup subsequently met and opted for Pi and EKl . Let 
1 7 no one suppose that Pi actually represents containment. 
16 PI and even P 2 levels of precaution may prevent massive 
19 accidents, but they simply cannot prevent ultimate escape 
2 q from the laboratory. 
21 An experiment bearing on the safety of an escaped 
22 viral recombinant DNA has already been carried out, the 
23 Rowe-Martin risk-assessment experiment with polyoma DNA. 
2 4 The results have never been published, but they were 
25 reported to show that a recombinant plasmid carrying viral 
[ 213 ] 
