Occasionally, experimental survival curves of micro-organisms exposed to a
lethal agent have a flat region and traditionally it has been interpreted a
s evidence of the existence of a "shoulder". However, if the survival curve
is considered the cumulative distribution of lethal events, which reflects
a spectrum of resistances, or sensitivities, then when the distribution's
mean, or mode, is large relative to its spread, a region resembling a "shou
lder" will be observed irrespective of whether the distribution is symmetri
c or skewed. Computer simulated survival curves generated with the Fermi an
d Weibull distributions as models demonstrate that the shape of the surviva
l curve alone is, therefore, insufficient to confirm any specific inactivat
ion mechanisms at the cellular and molecular level, although it can refute
the existence of some. Microbial mortality has also been assumed to be a pr
ocess following an exponential decay and hence that a certain degree of sur
vival is inevitable. It is not inconceivable, however, that there can be an
absolute thermal death time if the survivors are being progressively weake
ned by a prolonged exposure to the lethal agent. This testable possibility
is demonstrated with simulated survival curves generated with two mathemati
cal models. (C) 2000 Elsevier Science Ltd. All rights reserved.