When examined under the electron microscope, bovine casein micelles were seen as aggregates of spheroidal granules arranged in spherical symmetry. The granules were of 2 kinds—one transparent in the electron beam and the other relatively opaque. With increasing acidity of the milk the regular arrangements of granules tended to break down, and bridges composed of granules, formed between neighbouring micelles. The average diameter of the granules was about 8 mμ, from which a molecular weight of about 225000 was calculated. Evidence was adduced for the identity of these granules with the macromolecules of native casein in equilibrium with whey.

Human casein micelles showed the same structural features. The average diameter of the granules was about 6 mμ, from which a molecular weight of about 100000 was calculated.

A high anti-O chaining titre against Salmonella pullorum-H can be obtained in colostral serum when the killed antigen is introduced into the mammary gland during the dry period. A significant titre, several times the normal level, is maintained in the milk serum for 3 weeks postpartum. There was a definite residual titre in the milk serum of non-infused cows against the O and H antigens of Salmonella typhosa and the O antigen of S. pullorum-H.

Coagulation of rennin-treated casein was inhibited by treatment of the casein with dimethylaminonaphthalene sulphonyl chloride. The effect was caused by substitution on lysine side chains of the κ-casein fraction, and inhibition was complete when 2–3 lysine residues/molecule were blocked. This level of substitution did not affect other properties of the κ-casein, such as the release from it of non-protein nitrogen (NPN) by rennin and its ability to stabilize αs-and β-caseins in the presence of Ca++. The evidence suggests that lysine side chains on κ-casein take part in the coagulation of rennin-treated casein.

While using an acid degree value test during work on lipolysis in milk, difficulty was encountered in extracting the fat from the milk of certain cows, especially milk of high butterfat content and milk from newly calved cows. This difficulty was overcome by increasing the quantity of reagent used. A study was made of the quantities of reagent required for milks of different butterfat content and for colostrum. No effect on the acid degree value occurred as a result of increasing the quantity of reagent. A modification of the test is recommended, using an increased quantity of reagent.

The viscosity index and adhesive strength of casein from the milk of 6 individual cows were determined on samples collected regularly throughout the 9-month dairying season. The viscosity index of the casein from each cow showed a seasonal decline while the adhesive strength showed a general increase during this time. Casein from one cow had a consistently low viscosity index attributed to the presence in it of a rare genetic variant of α-casein. It is suggested that the variations in viscosity index and adhesive strength are lactational rather than seasonal.

The physical effects of various cations in caseinate dispersions of high concentrations were investigated over a range of temperature and pH.

With calcium and strontium the temperature-viscosity relationships of the caseinates were abnormal in that the viscosity decreased rapidly from 30 to about 40 °C and a gel formed at temperatures in the region of 50–60 °C. On cooling, the gel reliquefied. No gel formed with barium, aluminium or magnesium. On cooling, magnesium preparations separated into 2 phases.

The supernatant phase from the magnesium caseinate and a corresponding phase prepared by centrifuging the calcium caseinate showed depletion of α-casein and enrichment of κ-casein and β-casein. The supernatant phase from the calcium caseinate showed the reversible gel formation on heating. The magnesium supernatant phase did not. κ-Casein and a mixture of κ- and β-caseins gave reversible gels at similar levels of calcium and pH.

For reversible gel formation to occur, calcium caseinate was required to be in fairly high concentration, to have a calcium content of about 1·0% of the protein and to be within the pH limits 5·2–6·0. The temperature at which gelation occurred was affected by the concentration of calcium and protein and by pH.

The behaviour of the material was compared with that of methyl cellulose with and without addition of urea.

Some potential commercial applications of the findings on viscosity relationships are outlined.

Some milk lipid materials representing the phospholipid protein complex—fat globule membrane from buttermilk (FGMI) and from butter serum (FGMII), intact fat globules (dialysed cream) and triglycerides (butter oil)—were subjected to copper-induced oxidation in systems containing distilled water, phosphate buffer of pH 6·60 or milk dialysate. These model systems usually contained 2% lipid material and were allowed to oxidize for 48 h at 3–5 °C. Oxidation intensity was measured by means of the 2-thiobarbituric acid test.

The oxidation intensity was dependent on the composition of the lipid material and of the aqueous phase. Ascorbic acid in absence of added copper catalysed the oxidation of the membrane materials and of globular fat. The oxidation intensity produced by FGMI (from buttermilk) differed significantly from that produced by FGMII (from butter serum) and the difference was well correlated with differences in their composition. Butter-oil preparations gave the lowest oxidation intensities, indicating that they are not the site for development of oxidative defects in aqueous media.

Total and differential cell counts were made on a total of 26272 quarter samples of foremilk taken at fortnightly intervals from cows in the Institute herd during the 12 years 1953–1965, and the samples were also examined bacteriologically.

The average cell count in each lactation up to the 7th was calculated.

The average total cell count of samples from which no staphylococci, streptococci or coliform organisms were isolated increased from the 1st to the 7th lactations from 0·19 to 0·60 million cells/ml, and from the 2nd lactation onwards the increase was due entirely to an increase in the number of polymorphs. With infected samples, however, changes in the average total cell count were caused by changes in the numbers both of polymorphs and of cells other than polymorphs.

The average total cell count of samples from which coagulase-negative staphylococci were isolated varied between 0·58 and 1·00 million/ml throughout the 7 lactations. For samples from which coagulase-positive staphylococci were isolated the cell count varied between 0·72 and 4·94 million/ml, the highest average total cell count occurring in the 5th lactation.

The average total cell count of samples from which streptococci were isolated varied between 1·48 and 4·62 million cells/ml, the highest value occurring in the 2nd lactation.

Lactic acid streptococci have an absolute requirement for magnesium and potassium and failed to grow in milk deionized with IRC 50 resin (sodium form) (Reiter unpublished; see also Fisons, 1958; Reiter, 1956). These findings were confirmed in preliminary experiments using the defined medium of Reiter & Oram (1962). Requirements for other metal ions have now been investigated. Two strains of Streptococcus lactis (ML3 and C10) and one of Streptococcus cremoris (HP) were used. They were maintained as milk cultures, and before use subcultured twice in the synthetic medium containing Ca, Mg, Fe, Zn, Mn, Co, Cu and Mo (cf. Reiter & Oram, 1962) and then finally in the synthetic medium containing only Mg and K (20 ppm.), and Fe (0·01 ppm.).

The rate of increase of the complex rigidity modulus of renneted skimmilk was measured, both in a modified form of the torsiometer previously described and in the thrombelastograph, an instrument designed for the study of blood coagulation. Milk diluted with water was also tested.

The effect of holding the milk is to slow down the process of coagulation without affecting the ‘coagulation time’ (i.e. its starting time). In undiluted milk a fall in pH from 6.66 to 6.18 does not affect maximum firmness but a later softening of the curd is accentuated in diluted and undiluted milk. Again, the coagulation time is not affected but the rate of firming up is reduced. Potassium oxalate, used to reduce the calcium content, shows similar effects.

N-Ethyl maleimide, a reagent specific for sulphydryl groups, had no effect on undiluted milk; but with dilute milk, at a pH of about 6·0, there was a decrease in maximum firmness and there were also differences in the later softening of the curd.

The significance of these effects on bonds formed during postenzymic coagulation is discussed.

Folic acid activity was determined for National Dried Milk and for 5 proprietary dried milk foods for babies, for a proprietary liquid milk baby food and for 8 brands of tinned evaporated milk. For comparison, values were determined for mature breast milk, for raw bulk cow's milk, for bottled pasteurized cow's milk and for goat's milk.

Human milk and raw and pasteurized cow's milk all had much the same folate activity, equivalent to about 54 µg folic acid/1. Values for goat's milk were much lower, around 6 µg/1.

Values for the reconstituted baby milks ranged from 9 to 65 µg/1, though only 2 of the products had markedly lower values than breast milk. The question is discussed whether the folic acid requirement in infants can be met by formulas based on cow's milk without having recourse to folic acid supplementation.

(1) A tentative scheme is proposed for the quantitative estimation of the effects of various factors on one major milk constituent, lactose, from analyses of the milk of individual cows within herds. The potential lactose content in the milk of individual cows is predicted from the observed potassium-to-lactose ratio in the milk, and the difference between the predicted potential and the actual lactose content is partitioned into fractions that are attributed to effects due to age, inter-quarter difference and changes with stage of lactation. The scheme was applied to 2 commercial herds, one (herd A) producing milk of low and the other (herd B) milk of normal SNF content.

(2) Lactation mean values for fat, SNF and lactose contents of the milk of individual cows showed a wide range of the same order in each of the 2 herds. Of the mean difference in SNF percentage of 0·30 between the herds, 0·14, or 47%, was due to a difference in lactose content.

(3) The predicted potential milk lactose content of the cows in the 2 herds ranged from 5·04 to 5·66 (g/100 g milk water). Herd mean values were 5·394 for herd A and 5·244 for herd B.

(4) Herd mean values for the effect on the lactose content (g/100 g milk water) of age, of changes with stage of lactation and of inter-quarter difference were −0·127, −0·080, and 0·073 respectively in herd A, and −0·197, −0·058 and −0·078, respectively, in herd B.

(5) The relative importance of the various factors in accounting for differences between the herds in the lactose content of their bulk milk were: predicted potential lactose content 61%, effect of age 28%, effect of changes with stage of lactation 9% and effect of inter-quarter difference 2%.

(6) The errors attached to the estimation of potential milk lactose content and the effects of age, of changes with stage of lactation and of inter-quarter difference are discussed.

(1) In a previous paper (Walsh, Rook & Dodd, 1968) a scheme was devised in which the potential milk lactose content for individual cows was predicted from the observed potassium-to-lactose ratio in the milk, and the difference between the predicted potential and the actual milk lactose contents was partitioned into effects due to age, changes with stage of lactation and inter-quarter difference. The scheme was based on repeated analyses for potassium and lactose of milk from the separate quarters of the udder of each animal at intervals throughout a single lactation, and was applied to 2 commercial herds. A simplified scheme, in which analyses are confined to milk secreted in mid-lactation and in which the measurement of the effect of changes with stage of lactation is omitted, has now been applied to a further 6 commercial herds. The results are reported here, together with those for the mid-lactation period for the 2 herds studied previously.

(2) The potassium and lactose contents of the milk of uninfected quarters of the heifers in 4 of the 6 herds and of the second-lactation animals in the 5th herd, which had no heifers, were found to conform with the relationship established previously (Walsh & Rook, 1964) on which the prediction of potential lactose content is based. The values for heifers in the 6th herd did not conform, and the results for this herd are therefore not reported.

(3) The range of herd mean values for SNF content was 8·17–8·62% and for lactose content 4·33–4·72%. The overall range for all herds for the predicted potential lactose content of the milk of individual cows was 5·06–5·71 (g/100g milk water). Herd mean values for predicted potential lactose content were, however, generally similar, ranging from 5·240 to 5·486 (g/100g milk water).

(4) The range of values for individual cows and the herd mean values for the effects of age and inter-quarter difference showed important differences from herd to herd. Herd mean values for the effect of age varied from −0·028 to −0·266 (g/100g milk water), and there was a similar range, of −0·074 to −0·251 (g/100g milk water), for the effect of inter-quarter difference. In herds where the effect of age was high, the effect of inter-quarter difference also was high.

(5) The effects of inter-quarter difference and of udder infections were similar. The loss in milk lactose content was about twice as high in quarters infected with streptococci as in quarters infected with staphylococci.

(6) Estimates of herd values for potential lactose content and for the combined effects of age and inter-quarter difference based on analyses of herd bulk milk agreed with corresponding estimates based on analyses of the milk of individual animals within the herd.

(7) The importance of variation in milk lactose content as a source of variation in SNF content and the relative importance of predicted potential lactose content and the effects of age and inter-quarter difference as a source of variation in milk lactose content are discussed.

The teats of milked and unmilked quarters of 6 cows were deliberately contaminated twice daily by immersion in a suspension of streptococci and staphylococci. Only the unmilked quarters became infected, 7 with Staphylococcus aureus and 3 with Streptococcus uberis. Neither previous infection with the same strain or a different strain of pathogen nor existing infection with Corynebacterium bovis appeared to influence the results.

Previous work had shown that when milk is boiled in a climbing film evaporator a reaction occurs between the casein micelles and the fat globule membrane, probably at the liquid–gas interface. It now appears that the micelle–globule complex is formed by combination of the κ-casein components of the micelles and the membrane of the fat globule. The reaction is not significantly affected by increases in concentration of individual constituents, or by changes in temperature or pH.

The history of the milk before its concentration is an important factor in its subsequent behaviour. Thus, cooling and stirring of the milk during storage, as normally occurs in a bulk collection tank, promotes the reaction in the evaporator and, for some milks, is essential for starting the reaction.

A high retention of whey by curds made from the concentrated milks, diluted with water to their original volume before renneting, indicates that curd properties other than fat retention are also altered during concentration.