The destruction of Escherichia coli M23 OR.H- using lethal water activity levels and nonlethal temperatures was investigated. Death rates were measured for a combination of four growth-permissible temperatures (15°C, 25°C, 40°C and 45°C) and four distinctly lethal water activities (0.75, 0.83, 0.88 and 0.90). In addition, death rates were measured at two temperatures (4°C and 50°C) outside the growth range of E. coli. Death rate increased both at higher temperature or lower water activity. Inactivation curves resulting from exposure of E. coli to osmotic stress were biphasic. The initial rate of death was followed by a slower second phase decline, or 'tailing' effect. Addition of chloramphenicol to the suspending medium reduced the tailing effect and suggested that tailing was caused by de novo protein synthesis. Copyright (C) 1999 Elsevier Science B.V.