Although the present study is conce

Although the present study is concerned with assessing growth differences between a set of E. coli strains the. term growth differences is, strictly speaking, somewhat misleading. We described growth as measured using optical density (absorbance), but it has been shown that the OD measure is not completely congruent to viable counts (CFU/ml) [19]. We do not, however, believe that this is a problem here since we are only interested in assessing growth difference, up to statistical certainty, between the strains. In addition, due to the conservativeness of the procedure here described, only large differences in growth are detected. This is an intended property of the procedure so that it is more likely that differences found are due to biological differences and not to natural random variations in growth. A limitation of the proposed model is therefore that strains exhibiting small growth differences, instigated by biological differences, may not be detected. A weakness of the OD based method is that antimicrobials can affect the maximum density reached by the cells meaning that microbial growth can occur without being detected using OD. This is nevertheless a weakness that has no bearing on the statistical method proposed for comparison, which can also be used with viable counts data.