Variations of moisture measurements in cheese

Data were accumulated during interlaboratory trials for cheese moisture det ermination from laboratories using officially recognized methods: AOAC; Int ernational Dairy Federation, and Standard Methods for the Examination of Da iry Products (SM), In one trial, ranges of means of 5 cheeses were 0.67, 0. 56, and 0.19% for 5, 9, and 8 laboratories, respectively. The lower ranges for the SM method were typical of 3 other interlaboratory trials, with rang es of 0.27, 0.34, and 0.34% for 6, 7, and 5 laboratories, respectively, Wit hin one laboratory, there were no significant differences among the 3 metho ds, but they all gave about 0.2% lower results than 2 other methods, one us ing freeze-drying, followed by drying in a vacuum, the other using cheese t hat was spread on sand and dried in a vacuum oven for 24 h, This finding in dicated that none of the officially recognized methods removed all the mois ture. Data showed that many laboratories tended to give either higher or lo wer results than the mean of all of them in a series of 7 interlaboratory t rials. Constant results, free of biases or systematic errors, are important in application of formulas for prediction of yield of cheese for purposes of yield control, but are difficult to obtain. It is proposed that results by a laboratory in interlaboratory trials be compared with those obtained b y one or more reference laboratories using a method that removes all the mo isture from cheese, The difference would be applied as a constant in the pr edictive yield formula. That difference would likely be best as a running m ean of differences in an ongoing series of trials, The reference laboratori es would use frozen samples for quality control to ensure uniformity of res ults among trials, Mean moistures of 36.10 and 36.11% were obtained on subs amples before and after freezing for 7 months.