24 
engage in recombinant DNA research using this strain as he it, how fast 
would that strain evolve in the laboratory? Were the strain sampled at 
regular intervals after dispensation, how soon would ther arise genotypic 
or phenotypic differences from the parent? It is important to point out 
that markers are frequently lost during the normal handling of a complex 
auxotroph, and that some markers are so detrimental to a c ills growth 
that they are routinely lost during continuous handling (e.g., gln E, 
Greg Pahel, personal communication). How dispensation and use of any 
EK-2 candidate system will alter the genotypic and/or phenotypic properties 
of the system should be assessed prior to certification. 
§25. The strain X1876 produces large numbers of pSC] 1-containing 
minicel'ls which are capable of surviving under conditions :£ thymineless 
death. The capacity of these minicells to act as genetic honors in 
triparental matings should be discerned. Can the minicell "survive" 
to disseminate plasmid-carried information under conditions under which 
the host cells cannot (see this paper, §llc)? 
§26. The possibilities that the DAP-less and thymincless death 
phenotypes could be rescued by ini vitro crossfeeding (see this paper, 
§§8 and 9c) are neith er ruled out completely nor examined extensively. 
§27. The introduction of drug-resistance markers to ‘‘acilitate 
monitoring may confer additional survival phenotypes upon the host strains 
in environments not considered in the testing of y K 177G. 
§28. The possibilities that yl7 7 G / y^.&7 G DN'A could t: .nsform bacteria 
otljer than E. coli in the gut was not examined. There may exist in the 
normal bacterial flora of the human gut cells capable of retrieving 
[Appendix C — 212] 
