Reliability of Five Methods for Protein Determination in Barley and Malt

Protein was determined in 44 barley and 49 malt samples by the Kjeldahl method. The data were used to establish a regression line of Kjeldahl protein and protein contents as estimated by biuret, dye-binding, alkaline distillation, and infrared reflectance (on two commercial instruments) methods. Protein contents of additional samples (76 of barley, 72 of malt) were determined by the Kjeldahl method (actual Kjeldahl protein). Protein contents of those were estimated by each of the five methods and were converted to the Kjeldahl nitrogen by use of the regression lines (predicted Kjeldahl nitrogen). Reliability of the five methods was evaluated by comparison of predicted with actual Kjeldahl protein of those additional samples. Six-rowed barleys contained, on the average, more protein than two-rowed barleys; varietal ranges in protein contents were smaller than location ranges; differences between barley and malt protein were small. Protein contents affected particle size of ground barley more than that of ground malt; ground high-protein barleys contained more large particles than the low-protein barleys. Malting reduced the particle size difference between high- and low-protein samples and increased amounts of fine particles. Correlation coefficients for linear regression between protein determined by the Kjeldahl and any of the other methods for calibration were significant at the 99% level and were generally above 0.97. The Kjeldahl method was the most precise and the biuret method the least precise. The biuret and dye-binding methods gave best agreement (for averages of two determinations) with the Kjeldahl method, infrared methods were intermediate, and alkaline distillation was the poorest.