Variability among sources and laboratories in nutrient analyses of corn and soybean meal. NCR-42 Committee on Swine Nutrition. North Central Regional-42.

Abstract

A cooperative research study involving members of a regional committee (North Central Regional Committee on Swine Nutrition [NCR-42]) was conducted to assess the variability in nutrient composition (DM, CP, Ca, P, Se, and amino acids) of corn and soybean meal from 16 sources (15 states, mostly in the Midwest) and to assess the analytical variability in nutrient assays among 22 laboratories (labs; 16 experiment station labs and six commercial labs). Corn samples were obtained from each participating station's feed mill during a 3-yr period (1989, 1990, and 1992), as were soybean meal samples during a 2-yr period (1989 and 1990). Both regular soybean meal (with hulls) and dehulled soybean meal were represented in the study. Samples were analyzed for DM and CP by all 22 labs, for Ca and P by 15 labs, for amino acids by 10 labs, and for Se by 6 labs. Each lab used its own analytical procedures. Samples of corn and soybean meal varied in their nutrient composition depending on the area of origin. The greatest variation among sources was in Se concentration, ranging from .02 to .29 mg/kg in corn and .08 to .95 mg/kg in soybean meal. Crude protein and lysine in corn were positively correlated, but the regression coefficient was low (r2 = .49). The relationship between CP and lysine for the two soybean meals combined was considerably stronger (r2 = .81). Lysine in corn increased by .018 percentage point and lysine in soybean meal (regular and dehulled combined) increased by .063 percentage point for each 1 percentage point increase in CP. Except for CP and Se, the analytical variability among labs was as great as, and in some cases greater than, the variability in nutrient composition among sources of corn and soybean meal. Within-lab analytical variability tended to be less than among-lab variability. Some labs performed certain analyses with considerably less variability and more accuracy than others.