Published survival curves of Escherichia coli in two growth media, with and without the presence of salt, at various temperatures and in a
Greek eggplant salad having various levels of essential oil, all had a characteristic downward concavity when plotted on semi logarithmic
coordinates. Some also exhibited what appeared as a ‘shoulder’ of considerable length. Regardless of whether a shoulder was noticed, the survival
pattern could be considered as a manifestation of an underlying unimodal distribution of the cells' death times. Mathematically, the data could be
described equally well by the Weibull and log normal distribution functions, which had similar modes, means, standard deviations and coefficients
of skewness. When plotted in their probability density function (PDF) form, the curves also appeared very similar visually. This enabled us to
quantify and compare the effect of temperature or essential oil concentration on the organism's survival in terms of these temporal distributions'
characteristics. Increased lethality was generally expressed in a shorter mean and mode, a smaller standard deviation and increased overall
symmetry as judged by the distributions' degree of skewness. The ‘shoulder’, as expected, simply indicated that the distribution's standard
deviation was much smaller than its mode. Rate models based on the two distribution functions could be used to predict non isothermal survival
patterns. They were derived on the assumption that the momentary inactivation rate is the isothermal rate at the momentary temperature at a time
that corresponds to the momentary survival ratio. In this application, however, the Weibullian model with a fixed power was not only simpler and
more convenient mathematically than the one based on the log normal distribution, but it also provided more accurate estimates of the dynamic
inactivation patterns.