Numerous investigations have been conducted to fill the need for simple, reliable, bacteriological tests to determine quantitatively the sanitary quality of food contact surfaces. They have resulted in the development of various swabbing (3, 4, 5, 7, l o ) , rinsing (2, l l ) , and agar contact methods (6,8,9,12). Some methods are better suited for examining flat surfaces than for examining food containers, eating utensils, or processing equipment, etc. ; in fact, there is no one method, presently available, that is applicable to the diversified surfaces encountered in processing plants, restaurants, and related food handling establishments. The plant sanitarian is therefore confronted with the problem of deciding which of the available methods will meet his requirements most satisfactorily. Moreover, information is not available on two important factors to be considered in making such a decision ; namely, the average proportion of the total contamination that a particular method recovers and the reliability of its results.Recently a direct surface agar plate laboratory method (DSAP) (1) has been developed which, based on triplicate determinations, is capable of detecting 88.5% to 9 . 3 % of the bacterial spore contamination on nonporous surfaces in 95 trials out of 100. This 95% confidence interval is based on experiments in which triplicate plate counts of a contaminating suspension were used as the bases for these percentage recoveries. In addition, the coefficient of variation of individual DSAP counts was shown to be 7.1% which is equivalent to the precision of the conventional agar pour plate method of enumerating bacteria. Although this method is probably too complicated for practical application routinely, it has provided an excellent reference procedure for quantitative comparison of surface recovery techniques in terms of the proportion of contamination detected and the precision of the results.Employing the DSAP method as the reference procedure a comparison was made of representative recovery methods and the results evaluated statistically.
EXPERIMENTAL PROCEDUREInoculation of test surfaces 1. iCTicrocorciis pyogciies var. aiireiis 209. Maintenance of stock cultures and preparation of milk suspensious were identical to the procedures previously described (1). The fii:al su\pensions contained approximately 10,000 to 13,000 organisms per ml. as determined by triplicate plate counts on tryptone glucose extract agar incubated at 35" C. for 48 hours.Vitreous china salad dishes were selected as test surfaces, because they are readily availablc, can be easily manipulated, and their smooth nonporous surfaces are typical of
