When milk was preheat-treated after either evaporation or homogenization with fat present the degree of heat stability induced in the resulting evaporated milk (EM) was considerably less than with the normal process in which the preheat treatment was carried out first. When the preheat treatment was carried out after both evaporation and homogenization, no heat stability was induced by the preheat treatment. Removal of the whey protein fraction from the milk increased the heat stability of both the EM homogenized before preheat treatment and the EM made by the normal process.

Formation of covalent bonds at milk sterilization temperatures was studied using caseins and casein peptides. At 120°C lysinoalanyl residues produced even at pH 7·0 were derived from intra-molecular interactions between phospho-serine and lysine; the conditions of formation were determined. It was also found that the formation of isopeptidic cross-links was significant with conditions more severe than those used for milk sterilization.

Potentiometric acid-base titration is a simple method that can be used successfully in the investigation of the protein system of milk and other protein-containing products. Its applications include the examination of conventional heating procedures, cheese ripening, the investigation of processed cheese and the effect of melting salts. The principle depends on the performance of pH-dependent phase transitions which are recorded by a single titration cycle and evaluated quantitatively by means of hysteresis loops.

Buffalo skim-milk is less heat stable than cow skim-milk. Interchanging ultracentrifugal whey (UCW) and milk diffusate with micellar casein caused significant changes in the heat stability of buffalo casein micelles (BCM) and cow casein micelles (CCM). Buffalo UCW dramatically destabilized COM, whereas buffalo diffu-sate with CCM exhibited the highest heat stability.

Cow κ-casein stabilizes αs-casein against precipitation by Ca better than buffalo º-casein. About 90% of αs-casein could be stabilized by κ: αs ratios of 0·20 and 0·231 for cow and buffalo, respectively.

Sialic acid release from micellar κ-casein by rennet was higher than from acid κ-casein in both buffalo and cow caseins, the release being slower in buffalo. The released macropeptide from buffalo κ-casein was smaller than that from cow κ-casein as revealed by Sephadex gel filtration.

Sub-units of BCM have less sialic acid (1·57mg/g) than whole micelles (2·70mg/g). On rennet action, 47% of bound sialic acid was released from sub-units as against 85% from whole micelles. The sub-micelles are less heat stable than whole micelles. Among ions tested, added Ca reduced heat stability more dramatically in whole micelles, whereas added phosphate improved the stability of micelles and, more strikingly, of sub-micelles. Citrate also improved the heat stability of sub-micelles but not of whole micelles.