In order to comply with the consumer demand for ready-to-eat and look 'fresh' products, mild heat treatment will be used more and more in the agrofood industry. Nonetheless there is no tool to define the most appropriate mild heat treatment. In order to build this tool, it is necessary to study and describe the response of a bacterial population to a mild increase in temperature, from the dynamic point of view. The response to a mild increase in temperature, defined by stress duration and temperature, consisted in a mortality phase followed by the lag time of the survivors and their exponential growth. The effect of the mild increase in temperature on the mortality phase was described in a previous paper (Bréand et al., Int. J. Food Microbiol., in press). The effect of the stress duration on the lag was presented in a previous paper (Bréand et al., Int. J. Food Microbiol. 38 (1997) 157-167). In particular, the biphasic relationship between the lag and the stress duration was observed and modelled with a four parameter nonlinear model: the primary model (Bréand et al., Int. J. Food Microbiol. 38 (1997) 157-167). The study presented in this paper deals with the effect of the stress temperature on the biphasic relationship between the lag time and the stress duration. The secondary models describing the effect of the stress temperature on this biphasic relationship, were empirically built from our experimental data concerning Listeria monocytogenes. This work pointed out that the higher the stress temperature, the narrower the range of stress duration for which the lag time increased. Since the primary and the secondary models of the lag time were available, the global model describing the effect of the mild increase duration and temperature directly on the lag was fitted. This model allowed an improvement of the parameter estimator precision. The potential contribution in mild heat treatment optimization of this global model and the one built for the mortality phase (Bréand et al., Int. J. Food Microbiol., in press) is discussed.