Hydrogen peroxide accumulated in aerated cultures of Streptococcus lactis C10, the amount depending on the sugar being utilized. When the energy source was galactose or (to a lesser extent) lactose or maltose, H2O2 reached autoinhibitory levels. With glucose, H2O2 reached levels which did not cause obvious inhibition, and then declined. Washed cell suspensions utilized several media ingredients including glucose, galactose and yeast extract as substrates for O2 uptake, but cell-free extracts utilized none of these. The only substrate which supported O2 uptake by cell-free extracts was NADH. The activity of NADH oxidase was not affected by several intermediates of carbohydrate metabolism. Evidence was obtained to show that its synthesis was regulated via induction by O2 with an overriding catabolite repression type control, when cells were grown at the expense of glucose.

Whole milk was concentrated by ultrafiltration in a plant causing some homogenization of the fat. Comparisons were made with milk concentrated in a plant causing little homogenization and with milk homogenized conventionally. None of the processes appreciably affected the casein micelle size distribution. On rennet treatment of homogenized milk, casein micelle aggregation occurred more slowly, the protein network in the curd was less coarse and the rate of whey loss was reduced, compared with non-homogenized milk at the same concentration. In using concentrated milks for cheesemaking homogenization improved the composition of Cheddar cheese, because of increased fat and moisture retention, but curd fusion was poorer. Some aspects of the texture of the mature cheeses were improved, but the free fatty acid levels were higher. Values for the firmness of curds, formed from milks processed in different ways, did not relate to the extent of aggregation of the casein micelles. It is suggested that the complete cheesemaking process is driven by the tendency of the casein to aggregate.

The detailed chemical composition of 25 milks from different breeds of cow, sheep and goat were related to their properties of rennet clotting, coagulum development and syneresis at pH 6·4. Experiments in which concentrations of fat and whey proteins in milk were manipulated, and in which milks were homogenized at different pressures, were also carried out and the effects observed were related to the above processes.

The composition of milks varied widely and many relations between concentrations of components could be related to their known modes of secretion from the alveolus or to their structural functions in the colloidal phase of milk. Rennet clotting was related to total Ca concentration and also to the proportions of αs- and β-caseins present. Coagulum strength was strongly related to total casein concentration in milk but was unaffected by total fat content or by casein/fat ratios. Syneresis was inversely related to fat content of milk and little affected by coagulum strength. Homogenization retarded syneresis, probably by altering coagulum structure and modifying entrapment of fat globules within the coagulum.

A study has been made of the interdependence of the volume fraction and the degree of dispersion of the fat in creams subjected to single-pass homogenization. The effects of fat volume fraction φ could be incorporated into a semi-empirical correlation of the homogenizing variables established previously. In general, as φ increased from a low value the mean fat globule diameter at first decreased, reached a minimum and subsequently increased. The total area of the fat interface/m.l. volume when the globule diameter was a minimum proved to be independent of the homogenizing pressure and had a value of approximately 1·1 m2; the interface of maximum area was greater than this and, in contrast, its magnitude was pressure-dependent.

The overall results have been interpreted in relation to the variability of the disruptive energy within the homogenizing valve, competing effects of globule coalescence and hindered break-up, and the surface-active properties of the milk proteins.

Variations in the biochemical performance of 2 DL-type mixed strain starters CH 01 and FDn grown in a sample of skimmed goat's milk I (obtained 2–3 weeks after parturition) and of goat's milk II (8 months after parturition) were studied, and results compared with those of the cultures grown in skimmed cow's milk. The initial levels of citrate differed in the 3 milks, with goat's milk II containing 0·7 mg/ml, goat's milk I 1·4 mg/ml and cow's milk 1·8 mg/ml. The production of diacetyl, α-acetolactic acid and acetoin was very low in goat's milk II. In goat's milk I the production of these compounds was higher, but still considerably lower than in cow's milk. Production of ethanol (EtOH) was similar in the 3 milks incubated with starter CH 01, but with FDn the production of EtOH was distinctly higher in goat's milk II. The total amount of CO2 produced corresponded with the initial level of citrate; consequently the highest production was found in cow's milk and the lowest in goat's milk II. The use of different starters for fermented products made from goat's milk in early and late lactation is suggested.

A method has been developed, using the fluorogenic substrate 4-methylumbelliferyl oleate, to measure the activity of the extracellular lipase of the psychrotroph Pseudomonas fluorescens AR11 in milk. Inhibition of lipolysis by constituents of skim-milk was prevented by including both sodium taurocholate and cetyltrimethylammonium bromide in the reaction mixture. Cream also interfered with the reaction and skim-milk was used for high sensitivity. Natural milk lipase was readily inactivated by heat treatment to which AR11 lipase was relatively resistant.

The effects of thermization (a sub-pasteurization heat treatment), followed by cold storage at 7 °C for up to 3 d on the post-gelation behaviour of rennet-induced gels from alternate-day collected milks, pasteurized before testing, were investigated. Observations of the time course of gel rigidity modulus development and of the syneresis of cut curds were fitted to a general exponential relationship to aid interpretation. Milk performance was assessed by measurements of rennet coagulation time at pH 6·5, various gel rigidity and syneresis parameters and losses of protein and Ca to the whey. A small but significant reduction in protein losses to the whey (mean 0·05%) was the only effect resulting from thermization. Cold storage gave rise to significantly weaker gels than those from unstored controls. The gels from cold-stored milks attained 50% of their maximum rigidity modulus in significantly less time and lost significantly more protein (mean 0·11%) to the whey. Samples cold stored for 1, 2 or 3 d were not significantly different from each other and no other parameters were affected. No significant interactions between thermization and cold storage were found.

A new method was developed for making Domiati cheese, based on blending ultrafiltered reconstituted milk and lipolysed recombined cream. The chemical composition of the new product was similar to the traditional cheese. The flavour intensity increased with the amount of lipase used in the cream hydrolysis. During storage the acidity and tyrosine content of cheese from different treatments increased at nearly similar rates. The level of total volatile fatty acids was affected by the level of lipase used. Organoleptic scoring showed that cheese made with cream treated with the intermediate levels of lipase (2 or 5 g/500 g cream) ranked the highest, after 1 and 4 months of storage respectively.

Generation times and acid production after 6 and 24 h by 20 strains of Streptococcus lactis of dairy origin were determined in heat treated (80 °C for 30 min) and sterilized cow's and ewe's milk. Ewe's milk enhanced growth of the streptococci, with significantly (P < 0·001) shorter generation times and higher acid production after 6 h incubation than cow's milk, probably due to its higher vitamin content. The stronger buffer capacity of ewe's milk allowed a higher (P < 0·001) acid production after 24 h than cow's milk. A stimulatory effect of sterilization on generation time and acid production after 24 h was observed in cow's milk. However, the heat treated ewe's milk was shown to be a better substrate than sterilized ewe's milk for Str. lactis.

A new technique is described for the microfiltration of cheese brine. Soft cheese brine with viable counts/ml of up to 1·2 × 106 bacteria and 2·2 × 104 yeasts and moulds was filtered using a pleated tangential flow filtration cartridge of 0·2 εm pore size membrane. This filtration resulted in a 3–4 orders of magnitude reduction of bacteria and the complete rejection of yeast and mould from the permeate. Hard cheese brine with similar bacterial loads was treated with a 1·2 εm pore size membrane cartridge and the resulting permeate contained 2–3 orders of magnitude fewer bacteria and no yeasts or moulds. Flux decline during the filtration was a function of the number of microorganisms and the amount of N in the brine. A cleaning procedure using enzyme and acid washes was established in order to prolong membrane performance.

The apparent diffusion coefficients for NaCl in Romano-type cheese salted by brining or dry salting were ∼0·23 cm2/d. A plot of the natural logarithm of % salt-in moisture against distance was almost linear, the slope of which is the distance over which the salt concentration decreases by a factor of 1/e (0·368). Such plots may be useful in predicting the concentration of NaCl at any point in the cheese after any period of brining. After 9 d brining, the water to salt flux (ratio of outflow of water to inflow of NaCl) decreased from 2·42 at the surface to ∼0·5 in the middle zone of the cheese and increased again to > 1 at greater distances from the cheese/brine interface.

An extracellular proteinase-and lipase-deficient mutant of a psychrotroph, Pseudomonas fluorescens strain 32A, has been isolated and the absence of the proteinase enzyme confirmed by growth on differential media, enzyme assay and polyacrylamide gel electrophoresis. Competition between the parent and the mutant was observed when equal numbers of the 2 strains were inoculated together into raw skim-milk at 6 °C. Bitterness was detected at 6 °C in pasteurized skim-milk inoculated with the parent cells concurrent with the detection of proteolytic activity. In the case of the mutant, slight bitterness which did not increase with increasing cell numbers was detected in the absence of proteolysis. Mutant cells failed to grow on Na caseinate as the sole source of carbon. It was concluded that the extracellular proteinase, while not essential for growth in milk, does provide a selective advantage to the producer organism. This enzyme is, however, essential for growth on milk proteins and contributes to the development of bitterness in pasteurized milk.

Intracellular cell-free enzyme extracts (CFE) of cheese starter bacteria significantly enhanced the flavour-accelerating effect of commercial neutral proteinase without increasing the gross proteolysis in Cheddar cheese. Peptide and amino acid N were released more rapidly in CFE-treated cheeses, compared with those treated with neutral proteinase or untreated controls.

Cell-free extracts of both Lactobacillus bulgaricus and L. acidophilus demonstrated threonine aldolase activity, the end product of which was acetaldehyde, the major flavour compound of yoghurt. L. acidophilus also possessed an alcohol dehydrogenase activity capable of reducing acetaldehyde so that little yoghurt flavour was present in milks fermentation with this organism. Addition of threonine to fortified milk before fermentation with L. acidophilus increased acetaldehyde production and resulted in a well flavoured product similar to that of yoghurt made with L. bulgaricus. The contribution of these 2 enzymes to flavour production is discussed.