ANALYST, JANUARY 1986, VOL. 111 37 Automated Determination of Fat, Crude Protein and Lactose in Ewe Milk by Infrared Spectrometry Wendy M. Harris National Institute for Research in Dairying, Shin field, Reading RG2 9AT, UK An instrumental method is described for the automated determination of fat, crude protein and lactose in ewe milk by infrared spectrometry. When ewe milks are analysed for fat, crude protein and lactose on an instrument that has been calibrated for cow milk, differences occur between the instrument readouts and absolute values as determined by chemical methods. This work identifies these errors and also looks at the different storage conditions that may be used for samples that have to be analysed over a prolonged experimental period. Keywords: Ewe milk; infrared analysis; fat; protein; lactose The use of the infrared milk analyser is well established in the dairy industry for the analysis of bovine milk. The automatic determination of fat, crude protein and lactose in 1500 individual cow milks is carried out each week at the National Institute for Research in Dairying (NIRD).The instrument is calibrated according to the manufacturer's instructions' and its performance is monitored daily by analysing sub-samples of a bulk milk. The concentrations of fat, crude protein and lactose in this bulk milk are determined by standard chemical methods.2 The infrared milk analyser measures fat at a wavelength corresponding to a C-H vibration, protein by absorption of the peptide linkages between amino acids of protein molecules and lactose by absorption of the hydroxy groups in the lactose molecule.Wide variations in the composition of individual fatty acids and amino acids are found in milks taken from different mammalian species. Consequently, an instrument calibrated for cow milk will register incorrect results if used for the analysis of ewe milk. A review of the literature revealed little information on infrared analysis on mammalian milk other than cow. Barbosa and Miranda3 analysed goat milk for fat, crude protein and lactose, both by standard chemical methods and by an infrared analyser, which was calibrated for cow milk. Marked differ- ences were found, where the instrument overestimated for fat and lactose and underestimated for crude protein concentra- tion. A research project at NIRD required the automated analysis of fat, crude protein and lactose in ewe milk.This was carried out using an infrared analyser that had been calibrated for cow milk. Ng-Kwai Hang and Hayes4 used potassium dichromate at varying concentrations to preserve cow milk and noted the effects of subsequent storage on fat, crude protein and lactose levels, using infrared spectrometry. Samples with a concentra- tion of potassium dichromate of 8 mg ml-1 of milk were stored at room temperature for up to 10 d. Fat concentrations remained constant, but marginal changes in crude protein levels were observed. This concentration of potassium dichro- mate can cause allergic reactions in laboratory staff. Sjaunjas reported that bronopol (2-bromo-2-nitropropane-1,3-diol) is used in Sweden in preference to potassium dichromate, because of the health hazard that may stem from the use of the latter.This work identifies the errors involved in analysing ewe milk on an infrared instrument calibrated for cow milk. It also determines the storage conditions that can be used for samples that have to be taken over a long experimental period. Such storage permits large batches to be analysed at one time with the minimum of work time accorded to calibration. Experimental Apparatus A Milkoscan Model 203B Infrared Milk Analyser [Foss Electric (UK) Ltd., York], fitted with a double-beam infra- red spectrophotometer and a two-stage mechanical homo- geniser, was used. This model measures fat at a wavelength of 3.48 ym, protein at 6.50 pm and lactose at 9.55 pm.Dilu-vials disposable plastic vials of 20 ml capacity (supplied by Elkay Laboratories, Basingstoke, Hampshire) and a Model 243 polarimeter (Thorn Automation Ltd., New Basford, Notting- ham), fitted with a 20 mm sample cell, were also used. Reagent Potassium dichromate ( K2Cr207). Supplied as Lactabs Mark I11 tablets from Thompson and Capper, Runcorn, Cheshire. One tablet contained 30 mg of potassium dichro- mate and 20 mg of sodium chloride. Dilution of Milk It was necessary to dilute the ewe milks with de-ionised water, so that the concentrations of fat and protein, in particular, were brought within the parameters of instrument calibration. All analyses were then carried out on diluted samples and corrections made for the differences between fat, crude protein and lactose concentrations as determined by the instrument and against those measured by chemical methods.Test Samples Morning and evening milks from Mule sheep were individu- ally composited and the pooled sample diluted (25 ml of sample + 25 ml of de-ionised water). Sub-samples of 20 ml of the diluted test milks were prepared to give two separate batches of 59 individual samples, which were stored over the experimental period of 9 weeks. Further aliquots of the diluted test milks were prepared, to give three batches of 12 individual samples. These were used to monitor any changes in composition due to the effects of storage. Storage The first batch of 59 milks was stored frozen at - 18 "C and the other was preserved with potassium dichromate (1 Lactab Mark 111 per 20 ml of milk) and stored in a refrigerator at 4 "C.The other 36 sub-samples were split into three separate batches of 12 milks each. One was analysed fresh, another38 ANALYST, JANUARY 1986, VOL. 111 Table 1. Composition of fat crude protein and lactose in diluted and undiluted ewe milks. Results are in grams of constituent per 100 g of milk Fat Crude protein Lactose Sample Undiluted Diluted Corrected Undiluted Diluted Corrected Undiluted Diluted Corrected 1 4.70 2.38 4.76 6.12 3.07 6.14 2.53 1.26 2.52 2 5.43 2.72 5.44 6.37 3.21 6.42 4.25 2.10 4.20 3 4.90 2.49 4.98 6.35 3.20 6.40 4.73 2.38 4.76 4 6.62 3.31 6.62 6.19 3.11 6.22 4.03 2.01 4.02 5 4.76 2.40 4.80 6.23 3.14 6.28 4.92 2.48 4.96 6 5.02 2.55 5.10 5.87 2.96 5.92 4.88 2.46 4.92 Over-all mean 5.24 Range .. 4.70-6.62 5.28 6.19 5.87-6.37 6.23 4.22 2.53-4.92 4.23 Table 2. Comparison of methods for the determination of lactose in diluted ewe milk. Results are in grams of constituent per 100 g of milk Polarimeter Chloramine-T Sample 1 2 4 5 6 Over-allmean . . Standard deviation . . Range . . . . Result 4.78 4.66 4.38 4.34 4.38 4.40 4.48 4.46 4.58 4.52 4.60 4.64 Mean 4.72 4.36 4.39 4.47 4.55 4.62 4.52 0.14 4.36-4.72 Result 4.66 4.66 4.36 4.22 4.40 4.44 4.36 4.36 4.50 4.46 4.60 4.58 Mean 4.66 4.29 4.42 4.36 4.48 4.59 4.47 0.14 4.29-4.66 after frozen storage and the other after storage in the refrigerator, with added preservative. These periods of storage were in line with those applied to the batches of 59 milks. Reference Samples These samples were used to compare the composition of fat, crude protein and lactose, as determined by standard chemical methods and by infrared spectrometry.Eight fresh samples were selected to give reasonable coverage of the range of compositional values as expected in the test samples. These were prepared and diluted following the same procedure used for the test milks and then analysed immediately. Analysis All the stored samples were analysed at the end of the 9-week period by the infrared analyser, which had been calibrated for cow milk. The frozen milks were rapidly thawed by immersion in a water-bath, set at 40 L- 2 "C. Occasional mixing helped to accelerate thawing and when the samples were conditioned to 40 "C, they were well mixed by inversion and presented for analysis.The preserved samples that had been stored at 4 "C were conditioned in the same manner and analysed imme- diately. British Standard methods were used for the chemical analysis.2 The Rose - Gottlieb method for the determination of fat was used in preference to the Gerber technique, The Kjeldahl procedure, employing a mercury catalyst, was accepted for the determination of the total nitrogen. The total nitrogen includes non-protein nitrogen, which was not deter- mined in this work. The nitrogen results were converted into crude protein (TN x 6.38). Anhydrous lactose was deter- mined by the chloramine-T method. Results and Discussion Composition of Fat, Crude Protein and Lactose in Diluted and Undiluted Ewe Milks It was considered necessary to ascertain that the dilution of ewe milk did not alter the accuracy of the chemical methods used.Duplicate analyses were carried out on both dilute and undiluted fresh samples for fat, crude protein and lactose. The results are given in Table 1. The over-all means for fat, crude protein and lactose on the undiluted and corrected diluted milks show only marginal differences and it was concluded that dilution of ewe milk did not alter compositional values as determined by the chemical methods used in this study. Determination of Lactose in Diluted Ewe Milk The chloramine-T method is non-specific for lactose and consequently it was decided that the method be evaluated against the polarimetric procedure given by Grimbleby.6 Duplicate analyses were carried out on diluted samples for each method and the results are given in Table 2.Results for both methods are in good agreement and it was concluded that the chloramine-T method was acceptable for the determina- tion of lactose in diluted ewe milk. Differences in Instrument Readout and Results of Chemical Analyses on Ewe Milk The results reported in Table 3 show the differences in instrument readouts and results of chemical analysis on the eight reference samples. Instrument minus chemical results for fat, crude protein and lactose all showed differences with the greatest margin being found for fat, giving a mean difference of +0.39 (standard deviation, s.d. 0.06). Crude protein showed a difference of +0.34 (s.d. 0.08) and lactose -0.32 (s.d. 0.07). The standard deviation figures for fat and crude protein compare well with the expected levels for cow milks (s.d.0.05), although it was noted that for the lactose it was higher than normally found in routine laboratory analysis. Composition of Fresh and Stored Ewe Milk To test the effect of storage on the composition of ewe milk, fresh, preserved and frozen samples were analysed by an infrared milk analyser for fat, crude protein and lactose. The results of analyses on 12 samples for each treatment are given in Table 4. The fat results for the frozen milks are slightly lower than those reported for fresh and preserved milks. Fat emulsion is sensitive to freezing and the oil can separate out on thawing, thus accounting for decreased infrared absorption. As fat destruction appears to be linked with the fat level, the dilution of the ewe milks may have reduced this effect.ANALYST, JANUARY 1986, VOL.111 39 Table 3. Differences in the composition of ewe milk as determined by chemical and instrumental* methods. Results are in grams of constituent per 100 g of milk Fat Crude protein Lactose Sample 1 2 3 4 5 6 7 8 Mean . . Standard deviation Range . . Infrared 6.10 5.42 5.46 5.28 5.66 6.20 5.24 3.90 5.41 Chemical - Chemical infrared 6.44 +0.34 5.90 +0.48 5.86 +0.40 5.56 +0.28 6.10 +0.44 6.60 +0.40 5.66 +0.42 4.26 +0.36 5.80 +0.39 0.06 4.26-6.60 * Instrument calibrated for cow milk. Infrared 4.78 4.94 4.66 4.50 4.60 4.46 4.98 4.44 4.67 Chemical 5.02 5.30 4.94 4.90 4.96 4.82 5.24 4.92 5.01 4.82-5.30 Chemical - infrared +0.24 +0.36 +0.28 +0.40 +0.36 +0.36 +0.26 +0.48 +0.34 0.08 Infrared 5.06 4.66 4.66 4.78 4.82 4.88 5.32 5.26 4.93 Chemical 4.66 4.29 4.42 4.36 4.48 4.59 5.00 5.04 4.61 4.29-5.04 Chemical - infrared -0.40 -0.37 -0.24 -0.42 -0.34 -0.29 -0.32 -0.22 -0.32 0.07 Table 4.Composition of fresh versus stored ewe milk. Results are in grams of constituent per 100 g of milk Fat Crude protein Lactose Fresh Preserved Frozen Fresh Preserved Frozen Fresh Preserved Frozen Over-all Range . . 7.20-9.52 5.34-6.54 2.96-5.58 mean* . . 8.12 8.10 8.01 5.98 6.00 6.05 4.34 4.39 4.37 * Mean values of the analysis of 12 samples for each treatment. Table 5. Composition of frozen and preserved ewe milks. Results are in grams of constituent per 100 g of milk Fat Crude Protein Lactose Preserved Frozen Preserved Frozen Preserved Frozed Over-all mean* .. 6.04 6.04 4.34 4.29 5.22 5.27 Correlation Range . . . . . . 4.82-7.38 3.62-5.28 4.52-5.70 coefficient . . . . 0.964 0.998 0.993 * Mean of analysis of 59 samples. Composition of Frozen and Preserved Ewe Milk Potassium dichromate was used at a reduced level of 1.5 mg ml-1 of milk, compared with 8 mg ml-1 as used by previous workers.4 This was carried out in order to minimise the health hazard that may stem from the use of higher concentrations of potassium dichromate. The fat , crude protein and lactose concentrations in frozen and preserved ewe milks are given in Table 5. The results show that there is no difference in the over-all mean for fat and only marginal differences between those for crude protein and lactose. Correlation coefficients give a slightly lower estimate for fat (0.964), compared with crude protein and lactose (0.998 and 0.993).This may be due to the effects of freezing on the fat emulsion. Conclusion Diluted ewe milk may be routinely analysed for fat, crude protein and lactose, using an infrared milk analyser, provided adjustments are made for the differences in composition that occur when using an instrument calibrated for cow milk. The automatic infrared milk analyser at NIRD is used almost exclusively for bovine milk and therefore it was considered impractical to alter the calibration to meet an infrequent demand for ewe milk. Project leaders were given results for diluted milks and advised of the adjustments (plus their relative standard deviations) to make to the fat, crude protein and lactose figures to give more precise data. For continuous analysis of ewe milk, it may be more accurate to re-align the instrument calibration, to give the absolute levels of fat, crude protein and lactose. Infrared analysis on ewe milk can be carried out on diluted samples that have been stored over a period of weeks, either by preserving with potassium dichro- mate or by frozen storage. The author thanks Mr. E. Florence, NIRD, for helpful discussions during the course of this work. References 1. 2. “Milkoscan 203B, Manual of Operation,” A / S N. Foss Electric, HillerBd, Denmark, 1984. “Methods for the Chemical Analysis of Liquid Milk and Cream,” BS 1741 : 1963, British Standards Institution, London. 3. Barbosa, M., and Miranda, R., “Addendum to Special Publication No. 49. Challenges to Contemporary Dairy Analy- tical Techniques,” Royal Society of Chemistry, London, 1984. Ng-Kwai Hang, K. F., and Hayes, J. F., J. Dairy Sci., 1982,65, 1895. Sjaunja, L. A., Actu Agric. Scand., 1984, 34, 273. Grimbleby, F H., J . Dairy Res., 1956, 23, 229. 4. 5. 6. Paper A51122 Received April lst, 1985 Accepted August 14th, 1985