Glutaraldehyde was used to fix the size distribution of casein micelles in skim milk before their fractionation by permeation chromatography on controlled pore glass. The effect of fixation was assessed by comparing the size and absorption profile for column fractions obtained from samples which were fixed before fractionation with those of unfixed samples and of samples that were fixed only after completion of chromatography. Micelle size was determined by photon correlation spectroscopy. Column profiles were obtained from absorption measurements at 340 nm and after pronase digestion at 280 nm to determine relative protein concentration. For comparative purposes the elution profile was divided into 4 peak areas, of which I-III contained most of the casein micelles, and IV consisted of the smallest micelles, soluble casein and whey protein. Average micelle size was unaltered by fixation but was larger in fractions from prefixed skim milk than those from unfixed samples. Fixation increased the absorbance readings but the elution profile at 340 nm for peak areas I-III was essentially the same for unfixed fractions and those obtained when fixation was applied after chromatography. However, areas I-III accounted for a much larger proportion of the total profile area when the latter procedure was followed. Total profile area increased with fixation time and temperature but this did not affect the elution profile. The results indicate that fixation with glutaraldehyde does not induce artefactual changes in casein micelle size.

Different concentrates from sheep's milk cheese whey, with up to 26·5% solids, were prepared in the laboratory by reverse osmosis, and their viscosities and heat capacities determined over the temperature range 5–80 °C. These products behaved like other liquid milk concentrates previously studied, so that similar analytical expressions could be used to predict these physical properties in cheese whey as a function of the temperature and solid content. A nomogram is also given for an easier computation of each property. The water in liquid dairy products and their concentrates up to 40% solid content appears to behave as pure, free water with respect to its partial contribution to the heat capacity of these products, its apparent heat capacity being 4·208±0·016 kJ. kg-1 °C-1.

A general theory of gel formation by cross-linking of polymer molecules, orginally developed for the study of rubber vulcanization, is proposed as a model for milk gel formation. The model divides gel formation and development into 4 stages, an enzymic and a non-enzymic stage preceding gelation, then a stage of material transition from the sol to the gelled state and a stage of increase in gel strength due to internal cross-link formations. The 2 post-gelation stages overlap to a variable degree depending upon particular circumstances. Gel formation and properties are interpreted in terms of the number of initial units which combine to form the gel and the number of cross-links eventually formed by each unit. When combined with a kinetic expression for the rate of cross-link formation the time course of gel shear modulus development or of the incorporation of material into the gel may be predicted. The classical asymmetric sigmoid shape of gel shear modulus development curves results from the third stage of material transition into the gel, which in turn depends upon the number of cross-links eventually formed by each initial unit. The modulus of mature gels is found to depend upon the square of material concentration, and gel breaking strength is found to be proportional to the modulus, in agreement with the predictions of the theory.

Milk fat globule membrane (MFGM) material was isolated from commercial pasteurized milks and pasteurized creams using a procedure specifically adapted for these products (McPherson et al. 1984a). Pasteurized cream membranes contained higher lipid levels while pasteurized milk membrane material had lower lipid contents than fresh raw MFGM. Electrophoretic analysis showed that membrane material from both commercial products contained, in addition to native MFGM polypeptides, significant amounts of β-lactoglobulin with lower levels of caseins and other skim-milk components. The incorporation of β-lactoglobulin was more pronounced in pasteurized cream membranes. Examination of the lipoprotein complexes present in these membrane preparations by linear sucrose density gradient centrifugation showed that pasteurized milk membrane material had a similar profile to fresh raw MFGM except that preferential binding of skim-milk components was found in some of the lipoprotein fractions. Pasteurized cream membranes showed considerably different density gradient patterns with only 2 lipoprotein fractions present, of which the major one was a low density complex.

The material surrounding the fat globules in commercial pasteurized homogenized milk was isolated by centrifugation of the freeze – thaw extract of the fat globules through a 52·5% (w/v) sucrose solution in simulated milk ultrafiltrate. The purified membrane material contained higher neutral lipid levels than fresh raw milk fat globule membrane (MFGM) material. Caseins and whey proteins, mainly β-lactoglobulin (β-lg) were found to be the major protein components of homogenized milk membrane material. Native MFGM polypeptides comprised only a small proportion of the total protein in these membranes. The total amount of membrane material isolated from homogenized milk was substantially higher than that from fresh raw milk. Analysis of homogenized milk membrane material by linear sucrose density gradient centrifugation showed a different pattern to that found with other milks, indicating the presence of new lipoprotein complexes. Commercial ultra heat treated (UHT) milks were also examined, but it was difficult to obtain isolated MFGM material free from core fat. However, fat globule fractions were isolated from these products by high speed centrifugation and these were analysed intact. Membrane material in these milks contained caseins and whey proteins (mainly β-1g). Natural MFGM polypeptides were present at extremely low levels. Phospholipid levels in UHT milk membranes were greatly reduced compared with membrane material isolated from other milks.

Hydration of Ca and Na caseinates which were prepared from whole casein at different pH levels (range 4·7–8·0) was determined by means of water vapour sorption measurements in the water activity (aw) range 0·58–0·95. Water uptake of Ca caseinates was systematically lower than that of Na caseinates prepared at equal pH, the differences increasing with increasing pH and aw. Plots of the water content ν. the amount of added hydroxide at constant aw revealed a linear relationship between water uptake and cation content, suggesting that the increase in water uptake is mainly determined by the amount and type of cations associated with the caseins. In the aw range tested, Ca2+ adsorbed about 2–7 and Na+ 3–18 mol water/mol of cation. This implies that replacement of one mol of casein-bound Na+ by Ca2+ is accompanied by a displacement of 1–11 mol water at aw 0·58–0·95. The loss of water is a consequence of conformational changes induced by the chelating and cross-linking effects of Ca2+, which also lead to micelle formation.

A method which was developed for assaying the extracellular lipases of psychrotrophic bacteria in milk (Stead, 1983, 1984) and which uses the fluorogenic substrate 4-methylumbelliferyl oleate has been adapted for use with skim milk powder (SMP), whey powder (WP) and whey protein concentrate (WPC). A five-fold increase in the concentration of sodium taurocholate (NaTC), in the mixture of NaTC and cetyltrimethylammonium bromide needed to dissociate lipase from milk proteins, removed the excessive sensitivity of the assay to variations in the concentrations of SMP, WP or WPC incorporated. Commercially available pancreatic lipase provided a suitable standard of lipolytic activity and as little as 1–2 μ could be detected in each assay system.

The ability of 42 microorganisms isolated from cheese, including strains of yeasts, Geotrichum candidum, Arthrobacter, Moraxella spp., Brevibacterium linens and Staphylococcus saprophyticus spp. to produce phenethyl alcohol (PEA) from L-phenylalanine was studied. All the yeast strains produced labelled PEA from L-[U-14C]phenylalanine. Phenylpyruvic acid was detected as an intermediate of PEA production and CO2 was produced by decarboxylation of this acid. For five strains tested, the level of PEA which had accumulated in the culture at the end of exponential growth phase represented 39–52% of the L-Phe added. None of G. candidum strains nor bacterial isolates produced PEA from L-Phe. Strains of Moraxella spp. and four strains of the S. saprophyticus group produced phenylacetaldehyde. For three strains of Arthrobacter spp., five of B. linens and all the Moraxella, 14CO2 produced from uniformly labelled L-Phe represented more than one carbon atom of the L-Phe molecule, suggesting that Phe was catabolized beyond the stage of phenylacetic acid. Production and disappearance of PEA during Camembert cheese ripening is probably related to the metabolic activity of microorganisms present at the cheese surface.

The viscosity of heat treated creams increased with time (t). The rate (k) was dependent on homogenization pressure, fat content and heat treatment temperatures. Viscosity and its rate of increase could be related by the expressions:

log η = log η0 + kt

(where η0 is the viscosity at t = 0),

log η0 = a1 + b1φ/d

(where φ is the fat volume fraction, d is the globule diameter) and

k = a2 + b2φ/d.

Different values of a1b1a2 and b2 were found for homogenized and unhomogenized creams.

Bovine milk fat globule membrane (MFGM) material was isolated from cream by a new technique which did not involve removal of caseins and whey proteins before destabilization of the fat globules. These components were removed by centrifugation of the membrane material extract through a concentrated sucrose solution (52·5% (w/v) sucrose in 10 mM-Tris HCl buffer, pH 7·5). Membranes collected at the sample–sucrose interface, while whey proteins remained in the supernatant and caseins migrated into the concentrated sucrose solution. The yield of membrane material using this procedure was 25% less than that from conventional methods. This reduced yield was due mainly to lower levels of lipid material, in particular triglyceride. Electrophoretic analysis showed that the polypeptide composition of ‘unwashed’ membrane material was similar to that of conventionally prepared MFGM. This method is particularly suitable for the isolation of membrane material from milks and dairy products in which the fat globule stability is reduced or unknown.

Zabadi was manufactured using reconstituted skim milk powder prepared with different ratios of water and sweet whey and the organoleptic properties and chemical composition of the milks were measured. Sweet whey could be used successfully to replace up to 50% of the water added for reconstitution, whilst with 75 or 100% whey a sweet off flavour appeared. However, these higher levels could be used for fruit flavoured zabadi, and about 40–60% less sucrose could be used because of the additional glucose + galactose contributed by the whey.

Teleme cheese made from milk stored at 4–5 °C for 1–5 d and then pasteurized was compared with that made from unstored control milks. In the raw milks titratable acidity, tyrosine value and acid degree value increased more rapidly in stored milks than in the controls. With pasteurized milks, variations in coagulation times were attributable to levels of psychrotrophs in the corresponding raw milks. Differences were minor for milk fat and N content between cheeses. Curds made from milks stored > 4 d differed markedly from the controls in pH, moisture, N soluble at pH 4·6 and trichloroacetic acid (TCA)-soluble N, but only small differences in levels of casein fragments were detected. Proteolysis in mature cheese (60 d old) made from milk stored > 3 d differed significantly from the controls, and casein fragments were detectable. Off flavour related to excessive protein breakdown developed only in cheese made from milk stored > 4 d. In 60 d old cheeses there was significantly more rancidity in those made from milk stored for > 3 d than in those made from control milk. Unacceptable rancidity became evident at 60 d in cheeses produced from milk stored > 4 d. For Teleme cheesemaking psychrotrophic bacteria multiplying in the milk to levels likely to be encountered in commercial practice are of far greater importance for their lipolytic than for their proteolytic enzymes.

Bovine, caprine, ovine and human milks have been clarified by addition of Triton X-100 and EDTA followed by a short incubation. Generally, addition of 4·2% Triton X–100 and 5·4% EDTA was sufficient and with skim-milk samples no dilution was necessary. With whole milks some dilution and greater amounts of Triton X-100 or longer incubation times or higher temperatures were required. These factors were interrelated to some extent and the optimum conditions will be determined by convenience and any particular requirements of the experiment. The attainment of optically transparent milk samples greatly facilitates colorimetric measurements; as examples we have shown that total protein may be estimated by the biuret reaction, NH2 groups measured with 2,4,6-trinitrobenzene sulphonic acid and SH groups by reaction with 5, 5′-dithiobis(2-nitrobenzoic acid). Acid phosphatase activity in bovine milk was followed colorimetrically by reaction with p-nitrophenyl phosphate to demonstrate the suitability of the clarification procedure in the measurement of enzymic reactions.

A fluorimetric assay, developed for the assay in milk of the lipase of the raw milk psychrotroph Pseudomonas fluorescens AR11 (Stead, 1983), was evaluated in both buffer and in milk using cell-free lipase preparations from 11 strains of lipolytic psychrotrophic bacteria isolated from milk. Lipases of all strains were more active against 4-methylumbelliferyl nonanoate (4-MUN) than against 4-methylumbelliferyl oleate (4-MUO) in the buffer system but most were less active against 4-MUN than against 4-MUO in the milk assay system. In both systems, 4-MUO had a much lower rate of non-enzymic hydrolysis than did 4-MUN. Lipase activities measured by radial diffusion in tributyrin agar and trioctanoin agar reflected activities against 4-MUO and 4-MUN, except that lipase from a strain of coliform was relatively more reactive against the 4-MU esters. When compared with assay in the buffer system the most effectively activated lipase in the milk system was that from AR11; of those from other strains, 6 were 50–69% as effectively activated, 2 were about 17% and 2 were 7–8% activated. This difference in behaviour was related to the direct inhibition of lipases of some of the strains by the mixture of sodium taurocholate (NaTC) and cetyltrimethylammonium bromide (CTAB) needed in the milk assay system to dissociate lipase from milk protein. Increasing the proportion of NaTC to CTAB increased the activities of the more weakly activated lipases but decreased those of others. The assay in milk may therefore underestimate the lipase from certain strains of psychrotrophic bacteria.

High-performance liquid chromatography was used to determine the amounts of all-trans retinol, 13-cis retinol and β-carotene in butters on sale in Britain. Samples originating in England and Ireland, New Zealand, Denmark and The Netherlands, and in France were analysed and each type of butter was measured in each season of the year. Fat, water, N and Na contents were also determined. The influence of season and country of origin on vitamin A content and on the level of salting are discussed.

Of the various factors influencing the volume transferred by the loop in the automated plate loop technique for enumerating bacteria in milk, the speed of the loop as it emerged from the milk sample was the most important. Increasing this speed from 3 to 6 cm/s resulted in a 23% increase in volume transferred by a standard loop of internal diam. 1·45 mm. The depth of dip was also important, but changes in temperature of the sample from 5 to 20 °C caused no significant change in volume transferred. The results indicate that if the transfer of 1 µl milk is required at speeds commonly used in the automated plate loop technique, the internal diam. of the loop should be reduced from the present standard 1·45 to 1·3 mm.

A micro scale syneresis assay, using the natural optical density of the whey released as a trace, is reported. This assay is suitable for use with raw, reconstituted skim-milk (RSM) or artificial milk. Using RSM the reproducibility and accuracy of the method compared favourably with larger scale laboratory assays. When the method was used to investigate the syneresis properties of artificial micelle milk (AMM), the rate and extent of syneresis and response to a major variable such as pH were very similar to those of RSM. This suggests that the assay used in conjunction with AMM is suitable for investigating the mechanism of syneresis at a molecular level.