A cleaning simulator was used to determine the changes in soil and bacterial numbers on stainless steel surfaces over 36 soiling and washing cycles (each of 12 h) with four cleaning systems. The soiling milk contained 106 Streptococcus faecalis colony forming units/ml. One system included a pre-milk iodophor rinse (20 C), milk soil (30 C), post-milk water rinse (20 C), alkaline detergent (50 C) and a final water rinse (20 C). The second system was similar, but without the final rinse. In the third system, the post-milk rinse was omitted. The fourth system was similar to the second, except that water at 20 C was substituted for the detergent. The simulator technique proved effective for determining changes in soil and bacterial numbers on surfaces over time. The detergent was the most important system component for controlling bacterial numbers, with the sanitizer contributing some control and the rinses very little. The numbers of S. faecalis on the surfaces were related to the amount of surface soil (detergent efficiency), and also to the inhibitory effect of solutions used before the intercycle rest period of 9.5 h between washing and the next soiling. S. faecalis did not grow during the intercycle period of 9.5 h at 30 C and a relative humidity of 80%, even on surfaces where a significant milk soil was present.