In published data the thermal destruction of Salmonella species in peanut butter deviates from pseudo-first-order kinetics. The reasons for such deviation are unknown. This study examined both the method used to measure the thermal destruction rate and the method of growth of the microorganisms to explain variations in destruction kinetics. Growth on a solid matrix results in a different physiological state that may provide greater resistance to adverse environments. In this study, Salmonella Tennessee and Oranienburg were grown for 24 h at 37°C under aerobic conditions in broth and agar media to represent planktonic and sessile cell growth, respectively. Peanut butter was held at 25°C and tested for Salmonella levels immediately after inoculation and at various time intervals up to 2 weeks. Thermal resistance was measured at 85°C by use of a newly developed thin-layer metal sample holder. Although thermal heat transfer through the metal device resulted in longer tau values than those obtained with plastic bags (32.5 ± 0.9 versus 12.4 ± 1.9 s), the bags have a relative variability of about 15 % compared with about 3 % in the plates, allowing improved uniformity of sample treatment. The two serovars tested in the thin-layer device showed similar overall thermal resistance levels in peanut butter regardless of growth in sessile or planktonic states. However, thermal destruction curves from sessile cultures exhibited greater linearity than those obtained from planktonic cells (P = 0.0198 and 0.0047 for Salmonella Oranienburg and Salmonella Tennessee, respectively). In addition, both Salmonella serovars showed significantly higher survival in peanut butter at 25°C when originally grown on solid media (P = 0.001) with a <1-log loss over 2 weeks as opposed to a 1- to 2-log loss when grown in liquid culture. Consequently, the use of cells grown on solid media may more accurately assess the survival of Salmonella at different temperatures in a low-water-activity environment such as peanut butter.