Factors contributing to the initiation of lipid oxidation in cow's and mare's milk containing high levels of polyunsaturated fatty acids were studied. Addition of H2O2 just after milking, in slight excess of the stoichiometric amounts required to destroy ascorbic acid, delayed the development of oxidized flavours in cow's milk high in linoleic acid. Hydrogen peroxide treatment followed by the addition of α- or γ-tocopherols prevented lipid oxidation in cow's milk even when 0·1 mg Cu/l milk was added. When used separately in the presence of Cu these treatments were ineffective as was butylated hydroxyanisole treatment.

The lipid and ascorbic acid in mare's milk were remarkably stable to oxidation. Addition of 0·05 or 0·1 mg Cu/l, ethylenediamine tetraacetic acid, neocuproine, or H2O2 had very little effect on the loss of ascorbic acid and lipid oxidation in mare's milk.

Group N streptococci formed acetaldehyde and ethanol from glucose. As the enzymes aldehyde dehydrogenase, phosphotransacetylase and acetate kinase were present this would enable these organisms to reduce acetyl-CoA to acetaldehyde and convert acetyl-CoA to acetyl phosphate and acetate. A pentose phosphate pathway which converted ribose-5-phosphate to glyceraldehyde-3-phosphate was also present. Acetaldehyde could not be formed via the hexose monophosphate shunt or by direct decarboxylation of pyruvate, as the enzymes phosphoketolase and α-carboxylase were absent. Phosphoketolase activity was induced in Streptococcus lactis subsp. diacetylactis after growth on D-xylose. Group N streptococci also contained an NAD-dependent alcohol dehydrogenase which reduced acetaldehyde to ethanol while both NAD- and NADP-dependent alcohol dehydrogenase activities were found in Leuconostoc cremoris.

Assay procedures were developed for a number of enzymes in milk which apparently originate from leucocytes. The enzymes studied were acid phosphatase, N-acetyl-β-D-glucosaminidase, β-glucuronidase, arylsulphatase, α-mannosidase, and catalase. Quarter-milk samples were analysed for enzyme activity and results compared with the electronic cell count and the Wisconsin Mastitis Test. All enzymes measured except acid phosphatase and α-mannosidase showed good correlation with the electronic cell count. Of the other 4 enzymes tested, β-glucuronidase and arylsulphatase were unsuitable as diagnostic aids owing to the lengthy incubation periods required in their assay procedures. The assay of catalase, which involved the measurement of the initial rate of release of O2 using an O2 analyser apparatus, was rapid, sensitive and reasonably reliable, if fresh milk samples were used. The assay procedure for N-acetyl-β-D-glucosaminidase was considered to be the most reliable, simple and rapid enzymic method for estimating the number of somatic cells in milk.

Ovine and caprine milks showed a marked stability maximum at ~ pH7 in their heat-stability/pH curves, but became very unstable at higher pH values. A low level of κ-casein appears to be responsible for this low stability of ovine milk; β-lactoglobulin offsets the stabilizing influence of κ-casein at lower pH values. Removal of colloidal calcium phosphate from ovine or caprine milks had very little influence on their heat stabilities and it was necessary to reduce the concentration of soluble salts to a very low level before an effect was observed. At the pH of maximum stability ovine and caprine milks, although quite variable, had stabilities in the same range as bovine milks. Equine milk was very unstable to heat and the shape of the heat-stability/pH curve of most individual samples was similar to ovine and caprine milks.

β-Caseins isolated from buffalo's and cow's milk were hydrolysed either with rennet or with microbial proteases from Mucor miehei, M. pusillus Lindt or Endothia parasitica. The degradation products were separated by polyacrylamide-gel electrophoresis and the residual β-casein was determined quantitatively after various times. The electrophoretic patterns of the degradation products of buffalo and bovine β-casein produced by the different enzymes were not identical. β-Casein of buffalo's milk was hydrolysed by rennet and M. miehei protease at a slower rate than that of cow's milk. The reverse was found with E. parasitica and M. pusillus Lindt proteases. Carbamylation of buffalo β-casein was found to retard its proteolysis by all the enzymes but particularly by rennet and M. miehei protease.

A method is described for the quantitative estimation of a mixture of whey proteins by spectrophotometric scanning of the stained protein bands following polyacrylamide-gel electrophoresis. The incorporation of β-lactoglobulin A as an internal standard improves the accuracy of the technique. The method has been used to measure residual native whey proteins in milk after heating. Results are presented for whey protein denaturation in skim-milk after heat treatment at 130 and 140 °C for various periods of time.

Group N streptococci were found to cleave threonine to form acetaldehyde and glycine. Threonine aldolase, the enzyme catalysing this reaction, was found in all strains except Streptococcus cremoris Z8, an organism which had been shown previously to have a nutritional requirement for glycine. The enzyme was strongly inhibited by glycine and cysteine. The inhibition showed characteristics of allosteric inhibition and was pH-dependent. Inhibition by glycine, but not by cysteine, was highly specific. Analogues and derivatives of cysteine which contained a thiol group and a free amino group inhibited the activity of threonine aldolase. The presence of a carboxyl group was not necessary for inhibition. The cleavage of threonine by wholecell suspensions was stimulated by either an energy source to aid transport, or by rendering the cells permeable to substrate with oleate. Threonine did not appear to be degraded by enzymes other than threonine aldolase, as threonine dehydratase activity was low and NAD- and NADP-dependent threonine dehydrogenases were absent.

Addition of isolated κ-casein to milk reduced the destabilizing influence of forewarming, while the addition of isolated β-lactoglobulin (β-lg) to the κ-casein-enriched systems resulted in a loss in stability on forewarming. This effect was ascribed mainly to a β-lg:κ-casein interaction. Some individual milks were unaffected by forewarming while others were markedly destabilized by the same treatment. Addition of Ca2+ + Mg2+ and interchanging of milk sera by dialysis influenced stability.

It is concluded that milks which are destabilized by forewarming generally give a type A response (minimum in the pH–heat stability curves) and that preheat-stable milks generally give a type B curve (no minimum).

Clones of Streptococcus lactis and of Str. cremoris resistant to some bacteriophages that could virtually annihilate the parent strains were selected by plating very large numbers of the bacteria in the presence of the phage. Some of the resistant strains were shown to form acid rapidly in milk and to be satisfactory for cheese-making. They were not lysogenic. Adsorption of the phage to which the parent culture had been sensitive still occurred with 3 strains but not with 4 others. Rates of mutation from phage sensitivity to phage resistance were determined for 2 original strains of Str. lactis and were found to be of the same order as the rates quoted in the literature for other mutations. The mutation rates of phages to forms which could attack the resistant strains were too low to be measured.

Monospecific rabbit antibodies were produced against bovine chymosin (rennin) and bovine pepsin A and used for quantitative measurements of these enzymes by rocket immunoelectrophoresis. Bovine pepsin B was identified and measured quantitatively in tandem crossed immunoelectrophoresis, using a polyspecific antibody preparation. Chromatographically purified chymosin B and bovine pepsins A and B were used as standards. The contents of chymosin and bovine pepsin A were also determined by 2 forms of DEAE chromatography, and by an assay utilizing the ratio of milk-clotting activity at pH 6·0 and 6·5. Results obtained by measurements on 3 bovine rennets (calf rennet, adult bovine rennet, and a mixture of such types) by the different methods of analysis were in good agreement. The advantages of the immunochemical methods are direct and simultaneous identification and quantitation of enzymes.

The autolysin of Streptococcus cremoris had the specificity of an endo-N-acetylmuramidase as it hydrolysed the linkage between N-acetylmuramic acid and N-acetylglucosamine. The enzyme had no amidase or endopeptidase action. It reached highest activity in the exponential phase of growth and in the electron microscope seemed to fragment the coccal wall at the equatorial ring.

The highest incidence of lipolytic activity among the psychrotrophic Gram-negative flora of commercial raw milks was found in strains of Pseudomonas fluorescens and Ps. fragi. The lipases of all of the lipolytic strains remained wholly or partly active after heat treatment at 63°C for 30 min. Two of the strains tested further had lipases which retained 20–25% of their activity even when heated at 100 °C for 10 min. Cheeses made from milks in which strains or a single strain of lipolytic Gram-negative rods (GNR) had been allowed to multiply to > 107 colony forming units/ml became rancid after 4 months even though the GNR had been killed by pasteurization. The rancidity was characterized by a soapy off-flavour in cheeses containing free fatty acid concentrations from 3 to 10 times higher than those in control cheeses made from stored milks with low counts of GNR. Strong rancidity could be reproduced by adding the culture supernatant of a pre-grown lipolytic strain, but not the washed cells, to milk and pasteurizing it immediately before cheese-making, demonstrating the extracellular nature of the rancidity-inducing lipases.

Proteolysis in aseptic, chemically acidified (GDL) cheese and in starter cheese made under controlled bacteriological conditions (i.e. free of non-starter micro-organisms) was measured by gel electrophoresis, the formation of pH 4·6- and 12% TCA-soluble N, gel filtration and the liberation of free amino acids. The results show that rennet was mainly responsible for the level of proteolysis detected by gel electrophoresis, pH 4·6-soluble N and gel filtration i.e. large, medium and small peptides. However, rennet alone was capable of producing only a limited range of free amino acids; only methionine, histidine, glycine, serine and glutamic acid were produced at quantifiable levels (> 0·2 μmoles/g) in GDL cheese; it is suggested that free amino acids in Cheddar cheese are mainly the result of microbial peptidase activity. The levels of free amino acids in the starter cheese were considerably lower than values reported for commercial Cheddar.

Three techniques were used for preparing Cheddar cheese specimens for examination by scanning electron microscopy. Resulting micrographs prepared from fresh cheese made with calf rennet revealed that while all the techniques were satisfactory, different structural features were observed depending upon the method used. A modified critical point drying technique and a freeze-drying method displayed surface features whereas trypsin hydrolysis showed internal microstructure. The surface microstructure of the cheese was found to be formed of protein aggregated and fused into structural units of 1·5 μm. The internal microstructure appeared to be formed of thin compact walls. Cross sections prepared by freeze-drying displayed arrays in the protein matrix and locations where fat globules had been embedded.

Free fatty acids (FFA) and free δ-lactones in butter oil were quantitatively estimated by cold-finger molecular distillation at 35 °C followed by silicic-acid column chromatography of the cold-finger distillate and gas chromatography of the fatty acid and lactone fraction. Typical values for FFA and free δ-lactones in butter oil prepared from export standard sweet-cream salted Australian butter were in the range 0·07–0·24, 0·75–2·4, 4·5–8·6, 34–40 and 44–89 ppm for the C4,6,8,10,12 fatty acids and 0·09–0·18, 1·2–4·9 and 3·0–5·8 ppm for the C8,10,12 δ-lactones. These values are in the range of organoleptically desirable levels.

Electron microscopy showed that a varying proportion of cells of Streptococcus cremoris NCDO 924 grown in autoclaved skim-milk possessed a layer of extracellular material attached to the cell wall. Occasional filamentous extensions of this layer made contact with neighbouring casein micelles. The same surfacestaining material persisted during the production of cheese-curd, but after maximum scald it was predominantly filamentous in appearance. These filaments made frequent contact with the adjacent curd matrix and with milk-fat globules. Str. cremoris NCDO 1986 produced similar surface material when in curd, but not when grown in skim-milk. In all situations, the extracellular material stained with colloidal iron hydroxide, ruthenium red and periodic acid–thiosemicarbazide–silver proteinate, indicating that it was largely composed of an acidic carbohydrate. It is suggested that this carbohydrate facilitates the adhesion of starter bacteria to the cheese-curd matrix and that during the initial stages of syneresis this serves to prevent their expulsion from the curd with the whey.

Phenol, o-methoxyphenol, m- and p-cresol, indole and skatole have been isolated from good quality butter oil by cold-finger molecular distillation, and separated into phenolic and indolic fractions by solvent extraction and silicic-acid column chromatography. The individual compounds were quantitatively estimated by gas chromatography. When phenol, o-methoxyphenol, m- and; p-cresol, indole and skatole were added to volatile-free butter oil, their recoveries were 94, 47, 90, 75, 71 and 61% respectively. The technique, when applied to fresh butter oil, gave the following ranges of values for the natural levels of these compounds: phenol 0·005–0·022, o-methoxyphenol 0·002–0·10, m-cresol 0·002–0·010, p-cresol 0·002–0·004, indole 0·07–0·13, skatole 0·16–0·22 ppm of butter oil. The results, when compared with flavour threshold studies, showed that indole and skatole are important contributors to the natural flavour of butter oil, but that phenolic compounds are of only borderline significance.

Calf rennet, bovine pepsin, and porcine pepsin were used to produce cheese curd, using the same milk and lactic culture for each. Specimens were prepared for scanning electron microscope examination by a modified critical-point drying technique.

From examination of the micrographs, the curd made with bovine and porcine pepsin were similar in structure and in orientation of the coagulated protein, whereas the curd produced with rennet was different, having a more compact and organized structure.

Some strains of commercial cheese starters were obtained in higher yield after batch culture in whey-based media supplemented with soy peptones than in those supplemented with papain digested skim-milk. The soy peptones contained a greater proportion of small peptides than did the skim-milk digest. Also the average size of peptides was lower (3·4–3·9 residues) in soy preparations than in the milk digest (6·8). Small peptides isolated from the soy peptones were utilized by growing starters more completely than the equivalent peptides isolated from the papain digested milk. These observations may account for the difference in cell yields with soy and milk-digest peptones.

Differences in flavour scores and in the concentrations of free fatty acids, methyl ketones and H2S were measured in Cheddar cheeses containing various groups of non-starter bacteria or starter streptococci alone, made under controlled bacteriological conditions by the aseptic vat technique. The non-starter bacteria were made up of lactobacilli, leuconostocs, pediococci, micrococci and Gram-negative rods isolated in commercial creameries from raw milk or fresh cheese curd. These were added to the experimental cheese as single groups or as complete floras (reference floras). Several bacterial groups influenced the measured concentrations of the flavour compounds, but flavour differences were not correlated with these chemical differences. Only cheese containing a curd-derived whole reference flora or cheese made in open vats in the N.I.R.D. Experimental Dairy had significantly better flavour than starter-only cheese, but this improvement was not attributable to any particular group of bacteria.