(Curtiss cj^ aj^. , Figure 28). 
§14. Survival during passage through rats: 
a) The design of the feeding experiment, in particular, the sample 
sizes of the rats used to test a particular bacterial strain (3, 6, and 9 
rats strain does not allow for fluctuations from a normal distribution 
in the variety of responses of rats to E. coli. E. coli is known to have 
widely different passage properties through individual humar.s, and there 
is no evidence to the contrary of the existence of a wide range of passage 
responses to E. coli in younger rat9 (to our knowledge, there is no set 
of experiments reported that speak directly to this point). Furthermore, 
there are no indications that the rat gut environment is similar enough to 
the human gut environment with respect to response of the gut to E. coli 
to allow for a direct comparison. Although we question the statistical 
significance of these results a. priori with respect to design, we offer 
two additional criticisms to assist in the design of similar experiments 
in the future. 
b) Examining the results, Curtiss _et_ _al. conclude, "It is therefore 
evident that the six hours it takes a fed strain to appear in feces is an 
insufficient length of time to permit enough growth for 100', DAP-less death 
to occur. In reality, we do not even know whether DAP is or is not present 
in the intestinal contents to spare dap strains from DAP-ltss death, 
-although the reduced survival of DAP strains can be used to argue that 
free DAP is probably not plentiful in the intestine." (Curtiss et al. , 
p.19) This is circular reasoning; it foregoes the conclusion that the 
death of the cell in the GI tract is due primarily to the DAP-less death 
phenotype. 
[Appendix C — 207] 
