The purpose of this study was to evaluate maturation and storage conditions as a way to increase the safety of non-heat treated fermented sausages. The specific objectives were to investigate the effects of storage time and temperature on the levels of Escherichia coli, Listeria monocytogenes and Yersinia enterocolitica in fermented sausages and in broth, and to validate how well the broth experiments and some published models can predict inactivation in sausage. One strain each of E. coli, L. monocytogenes and Y. enterocolitica with induced acid tolerance was inoculated into sausage batters representing a typical Swedish recipe for cold-smoked sausages. The sausages were fermented at 27 degrees C for 39 or 48 h and then stored at different temperatures (8, 15, or 20-22 degrees C) for up to 44 days. The levels of the experimental strains, lactic acid bacteria, and pH, a(w), and lactic acid was measured during the maturation/storage period. Inactivation in BHI broths adjusted to pH 4.4 or 4.6, water activity of 0.93, and with 1, 1.3 or 2% lactic acid added was also studied. For all strains inactivation rates increased with temperature in both broths and sausages. At 8 degrees C the storage time required for a one-log reduction in sausage ranged from 21 days for E. coli, >16 days for L. monocytogenes, to 18 days for Y. enterocolitica. At temperatures of 20 degrees C or more, the storage time needed for a one log reduction was shorter: between 7 to 11 days for E. coli, 4 to 7 days for L. monocytogenes, and 1 to 4 days for Y. enterocolitica. A published model based on temperature only yielded a good prediction of E. coli inactivation in sausage. A linear model based on the rate estimated in broth yielded a fair prediction of L. monocytogenes inactivation. The performance of other inactivation models validated was unsatisfactory. Significant E. coli growth which occurred in batters without salt during initial phases of fermentation resulted in a subsequent increased inactivation rate, possibly due to increased susceptibility to stress of exponential phase bacteria. The results indicate that the practice of utilising a short maturation period and storage at refrigeration temperatures may result in unsatisfactory reductions of pathogens if present. Thus, inclusion of a maturation period above refrigeration temperatures before distribution may increase the safety of these products.