Kjeldahl Data. As expected, because of the way the milk concentrates were produced, there was a difference (P < 0.01) in both CP and true protein across the 4 milks (Table 4). The UF process used to produce a milk protein concentrate with about 9% protein and the dilution of that concentrate to formulate milks with progressively decreasing protein content should have maintained a constant NPN concentration in all of the milks of different protein content. A constant NPN concentration in all the milks was expected (Table 4) because the primary compound in the NPN fraction of milk is urea (Walstra and Jenness, 1984; Webb et al. 1974), and urea should easily pass through a 10-kDa molecular weight cut-off UF membrane. However, a progressive increase (P < 0.05) in NPN concentration from P1 to P4 was observed (Table 4). Why was NPN higher with higher protein concentration? Two possible explanations exist. (1) Urea is only 50 to 70% of the NPN in milk (Walstra and Jenness, 1984; Webb et al., 1974) and one or more minor NPN compounds other than urea that are soluble in 12% TCA may not have
passed through the UF membrane or attached to some- thing larger (e.g., milk fat globule membrane) during the UF process. Or, (2) an error was present in the NPN method.