1. Thirty-eight first-calf heifers in two equal groups were subjected for complete lactations to a rigid control of the duration of milking to either 4 or 8 min.

2. The treatment had no major effect on the milking rates of either group.

3. Whilst 4 min. was insufficient time to milk out completely the udders of about half the animals on this treatment, the machines on the 8 min. treatment were generally left on for a period after milk flow had stopped.

4. There was no significant difference between the 305-day lactation yields of the two groups or in the persistency. The 4 min. treatment did, however, decrease the yield of milk and fat in early lactation.

5. The 8 min. cows had more clinical and subclinical mastitis, though they did not have a higher proportion of eroded teats nor higher Whiteside test readings.

6. Teat erosion appeared to be correlated with slow milking.

1. A study was made of the milking of the two hind quarters of four cows by teat cup and by teat cannula at three levels of vacuum and by teat cannula at atmospheric pressure.

2. The eight quarters milked at significantly different maximum rates by teat cup but at nearly the same rate through a teat cannula, suggesting that the teat orifice is a very important factor controlling rate of milking.

3. The maximum rate of milking by both teat cup and teat cannula increased with increasing level of vacuum from 11 to 20 in. Hg. The rate of increase in the teat-cup milking was greater than in the cannula milking, suggesting that the teat orifice was stretched open at the higher levels of vacuum.

4. The average amount of strippings increased at the higher level of vacuum in teat-cup milking, due apparently to teat-cup crawl.

A study has been made of the effect of udder infections on the proportionate yield and composition of the milk of individual quarters of over 100 cows, and the data from eleven of these cows have been presented.

Typical examples have been given showing that subclinical Streptococcus and haemolytic Staphylococcus infections may persist for months and either cause a gradual reduction in the proportionate yield, and solids-not-fat percentage of a quarter or have no measurable effect, and that both types of reaction may occur simultaneously in the same udder.

Examples also show that following penicillin treatment of clinical cases caused by streptococci and staphylococci there is little recovery in the proportionate yield of the treated quarter until the following lactation, when recovery may be complete.

The experiments analysed in this series of papers were undertaken as the result of a discussion on methods of measuring the keeping quality of milk, at a Conference of Advisory Bacteriologists of the Ministry of Agriculture and Fisheries.

Owing to the disturbance and difficulty following the outbreak of war and the dislocation of handling and distribution, a serious decline in the keeping quality of raw-milk supplies was threatened.

The roll-tube method for estimating bacterial counts was devised by Esmarch(1), but was soon afterwards superseded by the introduction of Petri plates which have several advantages for cultural and counting purposes. The roll-tube method, however, has been revived by several workers (2,3,4), and where a simple, rapid and economical method is required it is very useful, especially if low counts are being dealt with. For this reason it is particularly suitable for yeast and mould counts in butter and has been so used in this laboratory for many years. In addition to its regular use for this purpose, several series of comparisons have shown that the roll-tube method is capable of giving results which compare very well with Petri plate counts upon milk, using the usual dilution technique in both cases.

1. A method has been developed for measuring the intensity of colour produced in the Aschaffenburg & Mullen phosphatase test in which use is made of permanent colour standards.

2. Using this method, the relationship between this phosphatase technique and the official Kay-Graham test has been investigated.

3. The limiting standards corresponding to 2.3 Lovibond Blue Units in the KayGraham test have been established for the 30 min. and 2 hr. procedures of Aschaffenburg & Mullen.

4. The advantages of the improved technique of the Aschaffenburg & Mullen test are discussed.

1. A new type of oxalate titration permitting determination of the nature of the colloidal phosphate of milk has been developed.

2. About 88% of the colloidal phosphate appears to be tricalcic, the remainder is thought to be dicalcic.

3. Estimates of tricalcium phosphate by the Ling titration are distinctly higher than the foregoing. This would seem to be due mainly to the divergent pH values to which milk and whey are titrated in this method. There is no evidence of any conversion of colloidal dicalcium to tricalcium phosphate during the Ling titration, as had been suggested.

In a paper by Mattick & Shattock (1) it is stated that: ‘The suspicion exists in the minds of those who investigated cases of gastro-enteritis that streptococci of the faecal type isolated from suspected cheese may have been responsible for the disease.’ They examined thirty cheeses and found that the faecal types of streptococci were common, and that the haemolytic varieties were rare. Since then the question has arisen as to the identity of the haemolytic streptococci found in cheese examined for the presence of the pathogenic streptococci. While we have never had occasion to examine cheese suspected of harbouring pathogenic organisms, the presence of haemolytic varieties of streptococci, their identity, and their role in cheese œcology is of general interest.

Some years ago discoloration was reported in mature New Zealand export cheeses. The defect occurred typically as bleached or pink areas surrounding cracks, though a less frequent ‘muddy’ discoloration was also noted. Discoloration was, on the whole, restricted to cheeses made in warm weather. Preliminary observations showed that the defect had certain characteristic features, viz.