Routine methods for the determination of solids-not-fat, moisture and fat (by difference) in butter - robustness, bias and precision

Routine methods for the determination of solids-not-fat (SNF) by the Kohman method and moisture by IDF Provisional Standard 137 in butter were tested for robustness, compared for bias with methods described in European Commis sion Regulation No. 880/98 (EC 880 method) and collaboratively studied. The SNF results were affected by the choice of solvent as well as the cooling technique (marble slab or desiccator) used for the beakers containing the d ried SNF, but not by the heating technique (oven or hot plate) used for the empty beakers and the beakers containing the SNF. IDF Provisional Standard 137 was found to be robust with regard to the heating technique (oven or h ot plate) and cooling technique (marble slab or desiccator) used for the em pty beakers as well as for the cooling technique (marble slab or desiccator ) used for the beakers containing the dried butter. IDF Provisional Standar d 137 and the Kohman SNF method gave results that were, on average, 0.07g/1 00g higher and 0.08g/100g lower, respectively, compared with the EC 880 met hod. No significant bias was found for fat-by-difference. Estimates for rep eatability (2.8s(r)) and reproducibility (2.8s(R)), respectively, were: SNF : 0.21g/100g and 0.44g/100g; moisture: 0.34g/100g and 0.40g/100g; fat-by-di fference: 0.40g/100g and 0.57g/100g. The precision estimates for moisture w ere larger than those stated in IDF Provisional Standard 137. This standard should be updated as its precision estimates were not established in accor dance with ISO Standard 5725. (C) 2001 Elsevier Science Ltd. All rights res erved.