Use of the bioluminometric ATP assay was evaluated for estimating total bacterial counts in ground beef. Minimum sensitivity was found to be 10⁶ cfu/g using a double filtration procedure for sample preparation. Although ATP content per cfu decreased approximately 10 fold during storage, correlation of total aerobic plate count (APC) with microbial ATP content was 0.96. Selective non-microbial ATP extraction with ATPase treatment was evaluated for use in conjunction with the double filtration procedure to increase assay sensitivity. The new method was effective for removing additional non-microbial ATP without reducing ATP in bacteria. Estimated APC values were generally accurate to within ±0.50 log for ground beef samples above the detection limit of 5 x 10⁴ cfu/g. ATPase treatment increased sensitivity of the ATP assay and APC estimation by about 1 log while increasing assay time by 40 minutes, for a total of 60 minutes for 4 samples assayed in triplicate. The ATP assay was evaluated for use with ground beef patties inoculated with mixed ground beef spoilage flora, <i>Pseudomonas</i>, or <i>Lactobacillus</i> and stored at 2°C or 10°C using oxygen permeable or impermeable (vacuum) packaging. Excellent correlation (r²=0.95) was obtained for each inoculum and storage condition over the range of 5 x 10⁴ to 1 x 10⁹ cfu/g, when estimated APC values were compared with experimentally observed APC values. Usefulness of the ATP assay for estimating APC values of frozen ground beef was evaluated. Retail ground beef and <i>Lactobacillus</i>- and <i>Pseudomonas</i>-inoculated beef were frozen and thawed at different rates and examined for APC and microbial ATP content. Results indicated that, although freezing and thawing lowered numbers of <i>Pseudomonas</i>, APC values and microbial ATP content closely correlated. APC estimates were generally accurate to within 1/2 log. The importance of using an ATP assay standard to correct for variable enzyme activity and presence of quenching factors was demonstrated, and improved formulae were developed for optimum assay standard use. Alternate regression methods were evaluated for estimation of APC values but did not yield enhanced accuracy. Only one regression equation was needed for estimating APC values of ground beef containing different types of bacteria stored in various ways. Therefore, little knowledge of ground beef history is needed in order to rapidly and accurately estimate microbial numbers in ground beef using the bioluminometric ATP assay. / Ph. D.