The use of Brix refractometer as a simple and economic device to estimate the protein content of sheep milk

Abstract

In this study, 737 individual sheep milk samples were collected 2 times from morning milkings to evaluate the relationships between Brix refractometer measurements and milk constituents— protein and fat percentages—to verify its ability to predict milk constituents. The Pearson's simple (rSP) and partial (rPP) correlations between milk constituents were calculated, and several first- and second-order regressions were tested to predict the protein and fat percentages. The rSP coefficients between the Brix refractometer measurement and fat (rSP = 0.46) and protein (rSP = 0.87) contents were different from the rPP coefficients, particularly for fat percentage (rPP = 0.04 and rPP = 0.82, for fat and protein percentages, respectively). The results of the forecasts can be considered satisfactory only for the first order regression that predicted the percentage of milk protein, while the regression that predicted the percentage of fat + milk protein presented a weak forecasting capacity.