Six hundred and seventeen isolates of Staphylococcus aureus from sub-clinical mastitis cases in 63 dairy herds in Northern Ireland were typed using a set of 25 phages. Ninety-four per cent of the isolates were typable, with nine phages, predominantly from groups I and III, being responsible for almost all of the lysis. Although 68 phage patterns were found, six of them typed 47·2 % of the isolates. One strain accounted for 14·7% of the isolates, but the largest number of strains (44) was restricted to individual farms. The epidemiological significance of these findings for on-farm mastitis control is discussed.

Levels of free fatty acids (FFA) were determined immediately after milking (initial FFA), and after 22 h storage at 4 °C, either without (FFA-22) or with a mechanical induction of lipolysis (FFA-22I), in milk from Friesian and Montbéliarde cows in the same herd, at identical stages of lactation and pregnancy and receiving the same feeding regimes. There was no difference between the two breeds in initial FFA and FFA-22 contents whatever the lactation period. This was also true during late lactation when lipolysis was further increased after feeding a poor quality grass silage. The observation that milk FFA-22I content from the Montbéliarde cows was higher than that from Friesian cows in summer, but not in winter, requires cautious interpretation as the activation treatments were not comparable, owing to the different ambient temperatures during activation in these two periods.

Levels of free fatty acids (FFA) were determined in milk from cows after 7 months of lactation and 4 months of pregnancy immediately after milking (initial FFA) and after 22 h storage at 4 °C (FFA-22). During the pre-experimental period, cows were at pasture. When housed indoors (experimental period) they were fed hay or grass silage for 3 weeks. Then all cows received grass silage for 3 weeks (post-experimental period). Feeding hay did not increase FFA-22 content in milk compared with pasture, but grass silage feeding enhanced FFA-22 content in milk compared with pasture (+130 %) or hay (+93 %). Increased lipolysis with grass silage was not due to underfeeding of the cows because grass silage and hay were both of high nutritive value. As pasture, hay and grass silage were of the same type (native mountain grassland), the high level of lipolysis occurring with grass silage probably resulted from the method of forage conservation.

Permeation chromatography on porous glass was carried out with skim milk at 20 or 30 °C using CPG-10 (50 nm) or a dual column system of CPG-10 (50 nm) followed by CPG-10 (300 nm). On columns of CPG-10 (50 nm) casein micelles were eluted at the void volume and were rapidly and efficiently resolved from non-micellar protein without micelle dissociation. The dual column system resulted in the additional resolution of the micelles into different size ranges. Examination of the resolved micelle fractions by electron microscopy showed a gradual decrease of weight average diameter (Dw) from 228·4 nm in the void volume fraction to 86·3 nm in the smallest micelle fraction. The translucent upper layer of a micelle sediment obtained by ultracentrifugation of skim milk at 30 °C consisted of casein aggregates intermediate in size between monomeric protein and the bulk micelle fraction as shown by its elution behaviour on CPG-10 (50 nm). These aggregates were enriched more than 2-fold with κ-casein relative to skim milk, were devoid of αs2-casein and had an estimated value of Dw of 33 nm. The ultracentrifugate serum contained ∼ 2·5% of total milk casein which had the elution characteristics of monomeric protein on CPG-10 (50 nm). It was concluded that the translucent sediment consisted of the smallest micelle fraction of skim milk and represented the minimum size range for casein polymerization in the natural milk environment. Overall, the results show that porous glass chromatography is an effective and convenient tool for the isolation and study of casein micelles.

Skim milk was heated rapidly to 130 °C in stainless steel tubing, and was then held at this temperature for periods of up to 1 h in a stainless steel holding vessel. Samples taken at intervals during the holding period were analysed for cations, inorganic and organic phosphate and protein in the total sample and in the supernatant after centrifugation at 60 500g. The cation and total phosphate (organic+inorganic) contents of the sedimentable material remained constant throughout the heating, although the caseins became extensively dephosphorylated. Dephosphorylated protein dissociated from the casein micelles during the first 20 min of heating, after which its concentration in the serum began to decrease, perhaps indicating the onset of the heat coagulation reaction.

The sizes of the casein particles in milk heated at 130°C were measured as functions of heating time and of initial pH. The effect of adding 10 mM-urea to the milks was also studied. Turbidities of the milks before and after dispersing the caseinate particles in 8 m-urea/EDTA/β-mercaptoethanol were measured, the latter measurement giving an indication of the extent of covalent cross-linking. Urea was found to modify the behaviour of the milk: diameters of the caseinate particles were considerably altered, and the extent of covalent cross-linking was diminished. These results agree with the observation that more protein is solubilized from the casein micelles when urea is present, indicating that a competitive reaction scheme involving cross-linking in opposition to micellar dissociation may be responsible for the ultimate precipitation of milk by heat.

Heat inactivation of a metalloproteinase, isolated from Pseudomonas fluorescens biotype I strain 112, was investigated in the temperature ranges 50–60 °C and 90–140 °C. At 90 °C the denaturation of the enzyme followed first-order kinetics with a decimal reduction time of 110 min and a velocity constant K of 3·5 × 10−4 s−1. Activation energy Ea was 100 kJ/mol for this temperature range. In the 50–60 °C region the proteinase was inactivated by autolysis, as shown by electrophoresis and gel filtration. At 55 °C the decimal reduction time was ∼ 22 s, at 57 °C it was 8 s. Rapid inactivation at 55 °C was only possible if the enzyme was heated from lower temperatures, but not if cooled down from 90 °C. This is due to a conformational change of the protein at this temperature. A model for the description of heat inactivation in the two temperature ranges is proposed.

The combined effect of ultrasonic waves (20 kHz, 150 W) and heat treatment applied simultaneously (thermoultrasonication) on the survival of a strain of Staphylococcus aureus was studied in 0·05 m-phosphate buffer pH 6·8 and ultra-heat-treated whole milk. This combined process decreased by 63% the decimal reduction times for the heat treatment when the organism was suspended in buffer and by 43% when suspended in milk. These effects were much greater than the additive effect of the two agents considered independently.

Changes in the composition of the microflora in Cheddar cheese ripened at elevated temperatures were examined. Cheeses were stored under the following ripening conditions: (i) control, 8°C for 32 weeks; (ii) 15°C, (iii) 17·5°C or (iv) 20°C, all for 8 weeks, then followed by 8°C for 24 weeks; (v) 15°C, (vi) 17·5°C or (vii) 20°C, all for 32 weeks. Generally, the total bacteria, lactic acid bacteria, lactobacilli and ‘undesirable lactobacilli’ (producing off flavours and CO2) counts in cheeses stored above 8°C were higher than in the control cheese. This effect seemed to be dependent on continued storage at the elevated temperature. Streptococcal counts rose slightly during ripening at elevated temperatures and fell slightly during ripening at the control temperature. Numbers of non-lactic acid bacteria were unaffected by elevated storage temperatures. In the only treatments where off flavours developed (17·5°C for 32 weeks, 20°C for 16 weeks or more) there was no apparent relationship between any bacterial group and the off flavours.

A mechanistic model of dairy cow metabolism is described. The model was constructed as part of a programme directed toward quantitative and dynamic analysis of concepts and data regarding factors which influence the partition of nutrients in lactating dairy cows. Sensitivity and behavioural analyses undertaken using the model indicate that concepts and data arising from tissue level experiments conducted in vitro can be used to structure and parameterize whole-animal models since the quantitative and dynamic behaviour of such a model is acceptable. These analyses indicate further that such models can be used to evaluate factors which influence patterns of nutrient utilization. By way of illustrating the model's utility regarding evaluation of concepts relating to the interpretation of energy balance experiments, results of a simulated energy balance experiment are presented. Apparent costs of milk synthesis are partitioned among biosynthetic costs, physiological costs, ion transport costs and the reductions in energy expenditures in synthesis of body components which accompany increasingly negative energy balances due to feed restriction.

The structure and characteristics of a model suitable for estimating digestion within the rumen and rates and patterns of nutrient entry in lactating cows are presented. The model consists of 12 state variables comprising a large particle pool, small particle pools representing insoluble dietary nutrients, soluble pools representing soluble dietary nutrients, and fermentation intermediates and end products. The model was constructed assuming continuous feeding, using Michaelis–Menten or mass action kinetics. The computer program was written in ACSL to run on a VAX computer. A fourth-order Runge–Kutta procedure was used for numerical integration.

Sensitivity and behavioural analysis demonstrated that overall stability and sensitivity of the model to individual parameters was generally satisfactory, but the need to improve the description and parameterization of aspects such as particle size in relation to availability, rate and affinity constants for amino acid degradation and rate constants for particle outflow from the rumen was established. Adjustments of the model to examine discontinuous feeding regimes were undertaken and initial results with respect to changes in fermentation rates, rumen acetate levels and microbial metabolism were considered realistic.

Comparisons with experimental data were considered satisfactory on forage-based and medium concentrate-containing diets, but with diets comprising 90% cereal, some inconsistencies, especially with respect to predictions of volatile fatty acid production rates, were observed. Reasons for this are put forward and suggestions for improvements in the model are discussed.

TWO models of lactating dairy cows were constructed by combining mechanistic models of digestion and metabolism reported previously and by adding elements to allow simulation of changes in tissue metabolic capacities over time. One model (day version) had an integration interval of 0·005 d and was suitable for simulation of within-day dynamics of nutrient supply and partition. The other (lactation version) had an integration interval of 1 d and was suitable for simulation of full lactation cycles. A number of simulation analyses were conducted to characterize and evaluate the models, to examine quantitative and dynamic properties of mechanisms which influence partition of nutrients, to identify aspects requiring further study, and to illustrate the potential usefulness of mechanistic, as compared to empirical, models in analyses of energy balance in lactating dairy cows.