Automated technique for sampling milk from farm bulk tanks: collaborative study.

An automated, in-line, mechanical technique for sampling milk from farm bulk tanks was evaluated in a collaborative study. The automated sampling device, which is mounted on the milk intake line, contains an electronically controlled peristaltic pump. The device takes a representative sample of the entire volume pumped through the system. Samples taken can be analyzed for both composition and microbiological quality. The study was performed in 3 phases. In the first 2 phases, samples taken by manual and automated methods were compared in analyses for somatic cell count, antibiotics, fat, protein, lactose, and solids-not-fat. The third phase, using a modified procedure, was designed to compare sampling methods in analyses for total bacteria count (standard plate count), psychrotrophic bacteria count, and coliform count. Evaluation of the data by a nested ANOVA indicated no difference between results for samples taken by the automated and manual methods (P = 0.05) in Phases 1 and 2, irrespective of whether the bulk milk was agitated before sampling. By introducing a sanitizing step between farms in Phase 3, the automated method also provided samples comparable with those taken manually for microbial analyses. The automated method has been adopted first action by AOAC International.

Enumeration of total bacteria and coliforms in milk by dry rehydratable film methods: collaborative study.

Eleven laboratories participated in a collaborative study to compare the dry rehydratable film (Petrifilm SM and Petrifilm VRB) methods, respectively, to the standard plate count (SPC) and violet red bile agar (VRBA) standard methods for estimation of total bacteria and coliform counts in raw and homogenized pasteurized milk. Each laboratory analyzed 16 samples (8 different samples in blind duplicate) for total count by both the SPC and Petrifilm SM methods. A second set of 16 samples was analyzed by the VRBA and Petrifilm VRB methods. The repeatability standard deviations (the square root of the between-replicates variance) of the SPC, Petrifilm SM, VRBA, and Petrifilm VRB methods were 0.05104, 0.0444, 0.14606, and 0.13806, respectively; the reproducibility standard deviations were 0.7197, 0.06380, 0.15326, and 0.13806, respectively. The difference between the mean log10 SPC and the mean log10 Petrifilm SM results was 0.027. For the VRBA and Petrifilm VRB methods, the mean log10 difference was 0.013. These results generally indicate the suitability of the dry rehydratable film methods as alternatives to the SPC and VRBA methods for milk samples. The methods have been adopted official first action.

Evaluation of the 3M Dry Medium Culture Plate (Petrifilm™ SM) Method for Determining Numbers of Bacteria in Raw Milk 1.

The 3M Company has developed a sample-ready system (Petrifilm ™ SM) for enumerating bacteria in milk and other food products. The testing unit consists of Standard Methods culture medium coated onto a base film and overlaid with a second film coated with a cold-water-soluble gelling agent and tetrazolium indicator dye. As such, the system is ready to accept samples of product. A pipette or 0.001-ml plate loop continuous pipetting syringe can be used for applying samples. In this study, both methods of sample addition were used and results compared with those of the Standard Plate Count (SPC) and standard Plate Loop (PL) methods for determining bacteria numbers in raw milk. In total, 108 samples were analyzed in duplicate by each of the four methods. The correlation coefficients (r) between the 3M-SPC and SPC, 3M-PL and PL, 3M-PL and SPC and PL and SPC were 0.946, 0.935, 0.941, and 0.974, respectively. Repeatability, as measured by mean log10 variance for duplicate determinations, was essentially the same for the four methods, and in all instances less than 0.005. The mean log10 differences between the SPC and 3M-SPC, and SPC and 3M-PL were, respectively, -0.177 and -0.168. The preceding statistical criteria suggest the Petrifilm™ SM method to be a suitable alternative to the SPC or the PL procedure.

Quantitative estimates of beta-lactam residues in raw milk around a reference standard: collaborative study.

A collaborative study was conducted to determine the reliability of a Bacillus stearothermophilus disc assay method for differentiating various concentrations of penicillin in raw milk. Participating laboratories tested 10 different samples (including one negative) in blind duplicate. Triplicate standards were alternated with triplicate unknowns around the periphery of each of 5 different plates. Zone diameters were measured and the difference in zone size of pairs of adjacent standard and unknown samples were analyzed by a paired t-test. Penicillin concentrations 0.003 IU/mL different from the reference concentrations were consistently distinguishable at a 95% confidence level. Such discriminatory power was determined to be possible with as few as 3 plates (9 replicates) per unknown. The method has been adopted official first action.

Collaborative evaluation of rolling ball viscometer for measuring somatic cells in abnormal milk.

The Ruakura rolling ball viscometer was evaluated in 3 laboratories along with currently approved instrumental methods for measuring somatic cells in milk and the Wisconsin mastitis test. Replacement of the Teepol reagent with Wisconsin mastitis test reagent in the rolling ball viscometer was also evaluated. Both repeatability and reproducibility were satisfactory for all methods evaluated. The instrumental methods each gave higher readings than the other 3 methods. Use of Wisconsin mastitis test reagent in the rolling ball viscometer improved both repeatability and reproducibility. Additional work on standardization is suggested to match rolling ball viscometer readings with those of the instrumental methods.

Quantitative Assay of Beta-Lactam Residues in Raw Milk Using a Disc Assay Method.

Numerous methods have been developed to determine presence of antibiotics in raw milk. Until recently, major effort had been placed on qualitative considerations, and primarily for detecting presence of penicillin (beta-lactam) residues. Only one method, the Sarcina lutea Cylinder Plate (CP) procedure, has been modified to provide for quantitative estimates. The CP method is a rather long, tedious test, requiring considerable technical skill. Need for a simpler, faster quantitative method was apparent. This paper describes a method for making quantitative estimates of beta-lactam residues around a fixed reference standard. The method uses Bacillus stearothermophilus in a disc assay test. Quantitative estimates above or below the reference level of antibiotic are computed through a paired-t statistical analysis. The test can be completed within 3 h.

Variability and Interrelationship of Various Herd Milk Components 1.

A study was made of the variation in herd milk composition and the interrelationship of various milk components as factors to be considered in milk pricing schemes, involving milkfat and/or other milk components. As measured by the coefficient of variability, fat was found to be the most variable component, followed by, in descending order of variability, protein, total solids, lactose, and solids-not-fat (SNF). Non-protein-nitrogen (NPN) level averaged 4.36 % of total nitrogen in 99 herd samples collected and analyzed in November, 1978. Coefficient of variation of NPN was 15.18. Thus NPN represents a significant factor in protein analysis where the Kjeldahl procedure is used as a reference standard. This fact lends weight to the desirability of basing milk purchase, if protein is involved, on "true" protein (total nitrogen minus non-protein-nitrogen) rather than total nitrogen.

Repeatability and Accuracy of Dye-Binding and Infra-Red Methods for Analyzing Protein and Other Milk Components 1.

Udy and Pro-Milk Mark II dye-binding methods and the Milko-Scan 104 infra-red device were evaluated for accuracy and repeatability in the analysis of protein in raw milk supplies. The infra-red device was also evaluated for accuracy in determination of milkfat, lactose and solids-not-fat, and was compared with the Milko-Tester for fat analysis. Repeatability of the three methods was, in all instances, less than 0.05% for protein. Standard deviation of accuracy (σy,x) for protein (total nitrogen × 6.38) for the Udy Analyzer, Pro-Milk, and Milko-Scan units was 0.063, 0.062, and 0.067, respectively. Standard deviation of accuracy of the Milko-Scan rated against Mojonnier and Milko-Tester methods for milkfat was, respectively, 0.054 and 0.050. Compared with liquid chromatography lactose determination and Mojonnier solids-not-fat determination, the Milko-Scan showed a standard deviation of accuracy of 0.083 and 0.073, respectively.

Precision of the Pro-Milk Method in Routine Determination of Protein in Dairy Testing Laboratories 1.

In a collaborative study of the Pro-Milk Mark II dye-binding method for determining protein in milk, repeatability (r) was found to be 0.053% protein, and reproducibility (R) 0.215% protein at the 95% confidence level. Standard deviations of these two measures of performance were 0.0188 and 0.0761, respectively. Repeatability was influenced by sampling error, mis-readings, equipment performance and routine control. Reproducibility was influenced by the same factors and, in addition, accuracy of calibration, with Kjeldahl as the reference method. Results of this study indicate the need to centralize laboratory calibration and to calibrate equipment over a wide range of protein levels.

Comparison of the Automated with the Semi-Automatic Coulter Counter Method and the Direct Microscopic Somatic Cell Count (DMSCC) on Raw Milk Samples 1.

The automatic Milk Cell Counter (MCC) and semi-automatic electronic cell counter (ESCC) of Coulter Electronics were compared with each other and with the direct microscopic cell count (DMSCC) on raw milk samples with various cell counts. The average DMSCC count on 241 samples of milk with Wisconsin Mastitis Test (WMT) results of 22 mm and higher was 55,000 cells/ml above the average MCC count when calibrated to a 4.4-µm minimum particle diameter. This difference is statistically significant at the 1% level. On 24 different raw milk samples of widely varying somatic cell count analyzed in replicate six times per sample, the standard deviations for replicate samples were 34,300, 34,900 and 136,000 for the MCC, ESCC and DMSCC, respectively. For these tests, the MCC had been calibrated to a 4.3-µm minimum particle diameter. The average difference between counts by the MCC and ESCC methods was only 6080/ml, but this was statistically significant at the 5% level. The average MCC count with the equipment set at 4.3-µm minimum particle diameter was 58,000 above the average DMSCC count.

Need for a Working-Factor in Evaluating Freezing Point Results on Raw Milk Samples 1.

The legal upper base for taking action in cases of water adulteration of milk is generally accepted as - 0.525°H. Henningson (2), in a study of 660 samples of milk known to contain no added water and representing 22 states in the U.S. and four Canadian provinces, found (at the 95% confidence limit) only 1% probability of a naturally-occurring milk sample falling above that level. Yet at a mean value of - 0.540°H, as noted in this same study, such base allows for 3% added water for the "average" dairy farm. The present study, made on samples without knowledge of their purity (lack of added water), tends to confirm this potential. At the same time it suggests the likelihood that, on a practical basis, a "working factor" set at some lower value could be useful to the dairy industry in coping with the problem of added water in milk.

Comparative Evaluation of the Difco Disc Assay Method for Detecting Penicillin in Milk and Milk Products.

The Difco disc assay method to detect antibiotic residues was evaluated and compared with the Bacillus subtilis disc assay and the Sarcinia lutea cylinder plate methods. The Difco method, using both color change and zone of inhibition as indicator, was able to detect penicillin to levels as low as 0.002 units per ml. Of 5200 raw milk samples analyzed for presence of penicillin in a commercial laboratory, the B. subtilis disc assay method identified and confirmed as positive 12 samples (0.239%). The Difco disc assay method identified 61 samples as positive for inhibitor, of which 48 (0.92%) were confirmed as positive for penicillin. Of 37 of these latter samples, analyzed also by the S. lutea cylinder plate method, 29 (78.4%) yielded positive results. The Difco procedure was found effective as a test method for detecting penicillin in finished fluid milk products.

Collaborative Study of the Coulter Counter-Chemical Method for Counting Somatic Cells in Raw Milk 1.

Variation between laboratories for Electronic Somatic Cell Counting by the chemical method (ESCC) was evaluated by a collaborative study. Eight laboratories counted somatic cells in 12 milk samples (six replicated samples) by the ESCC method. The somatic cell count for the same milk samples was also evaluated by the Direct Microscopic Somatic Cell Counting procedure (DMSCC) as a comparison for the level of error. The standard deviation of the variation of logarithms of ESCC counts between laboratories was 0.04368. The standard deviation for the variation of logarithms of DMSCC counts between technicians was 0.08617. The corresponding value for the DMSCC analysis of the last set of federal split milk samples was 0.141. An earlier study of electronic counting by the centrifuge method showed a standard deviation of 0.0711.