Characteristics of microbial colony counts on agar plates for food and microbial culture samples

Abstract

Microbial colony counts of concern of food products are one of the most important items in microbiological examinations. The distributions of colony counts per agar plate of food samples are considered to be reflected with microbial cell distributions in food homogenates. However, (i) the probabilistic distributions of the colony counts per agar plate at the dilution of counting and (ii) the relationship between the colony counts per plate and the number of agar plates for food samples have not been intensively studied so far. In this study, therefore, these two points were studied with raw food samples of raw minced beef and chicken and raw milk and microbial culture samples of Escherichia coli, Staphylococcus aureus, and Saccharomyces cerevisiae. Among four major probabilistic distributions, it was found that aerobic plate counts per plate of the foods were well described with negative binomial, Poisson, and normal distributions and that the colony counts per plate of microbial cultures were described well with binomial, Poisson, and normal distributions. The effect of the number of agar plates on the estimation of the mean of colony counts per plate of a sample was then studied with the data randomly resampled from the experimental data. The resampled data showed that with more number of plates the mean of counts fluctuated less and the coefficients of variation of colony counts per plate decreased further, which were coincident to the estimated by the central limit theory. Our study would provide useful information on the characteristics of colony counts per plate of food samples which are routinely examined.