Document Type

Thesis - Open Access

Award Date


Degree Name

Master of Science (MS)

Department / School

Dairy and Food Science

First Advisor

Sanjeev Anand


Sampling and analysis are integral parts of dairy processing in quality control, process monitoring and control, and meeting regulatory standards. Accurate sampling is vital to avoid erroneous analysis results and interpretation. The sampling process depends on the appropriateness of sample collection, transportation, storage, and analysis conditions. The bacterial heterogeneity and cluster formation, structural heterogeneity of food, sample unit and size, frequency, and sampling methods, are also crucial in sampling. We aimed to evaluate the influence of sample volume, sampling frequency, and sampling methods in experimental processing lines of whole milk, skim milk, and whey. To standardize the experimental protocols, a study was conducted under lab conditions. Bacillus licheniformis vegetative cells were inoculated in reconstituted Non-Fat Dry Milk (NFDM) at 2.0 and 4.0 log cfu/mL levels under sterile conditions to represent low and high-count milk. The milk samples were held at 6⁰C in a refrigerated shaking incubator (rpm 115). At 0, 4, 8, and 12 hours, 10 mL and 100 mL milk samples were drawn, and composite samples were prepared by mixing the individual samples. The standardization experiments provided an indication of the influence of sampling factors on the Standard Plate Counts (SPC). This showed that sample volume, sampling methods, and frequency are important factors that influence the SPC of low and high-count NFDM samples. We studied the influence of sampling factors in an experimental pasteurization run of milk (whole and skim) and whey. An experimental High-Temperature-Short-Time (HTST) pasteurization unit was assembled in the Davis Dairy Plant of South Dakota State University. Whole milk, skim milk, and whey was inoculated with Bacillus licheniformis at 4.0 logs and pasteurized uninterruptedly for 12 hours at 72⁰C/16secs. Raw and pasteurized milk represented high and low-count milk, respectively. At 0, 4, 8, and 12-hours individual samples of 10mL and 100 mL were drawn from the low and high-count sides of the pasteurizer. Two continuous composite samples of the 12 hour run were also drawn separately from the low and high-count sides. SPC was enumerated using tryptic soy agar incubating the plates at 32°C for 48 hours, followed by ANOVA analysis. Results indicated that the SPC of 10 and 100mL samples were not statistically different. This demonstrates that sample volume did not influence the SPC; thus, 10 and 100 mL could be equally effective for microbiological analysis of low and highcount samples of milk and whey. We observed that SPC varied with the sampling intervals for low and high-count milk and whey. Therefore, in an extended process run of 12 hours, sampling at multiple intervals could be more representative of the entire process run. Composite samples collected during the 12-hour process showed higher counts than individual samples at 0, 4, and 8-hour counts, and it was closer to 12-hour counts. Composite sampling thus could provide an indication of inclusive microbial count in extended process runs which might help draw an inference on the entire process. However, further studies may be needed to evaluate the influence of any other factors on the composite sampling over a period. In case sampling at different time intervals is not feasible, composite sampling could provide better representativeness of the counts over the entire process. All the studies were conducted under controlled experimental conditions forming the basis for more extensive studies under specific commercial dairy processing conditions.

Number of Pages



South Dakota State University

Available for download on Monday, December 15, 2025



Rights Statement

In Copyright