In the foregoing parts it is indicated that great difficulty may be experienced in the control of the spoilage of commercial sterilised milk which results from the development of spore-bearing bacteria.

A detailed study of the actual sources of contamination is now being made, but in the meantime it may be pointed out that many of these bacteria are typical members of the soil flora. They may therefore enter the milk from the dust either at the farm or the factory. The greater the care with which the milk is handled from the time of milking until the end of the heating process, and the greater the efficiency of the cleaning and sterilisation of the plant and the empty bottles, the less will be the danger of spoilage arising from the use of infected milk. During the course of the investigations recorded it has been necessary to carry out numerous control experiments. The milk used has been Grade A (T.T.) milk and the utensils with which it came into contact and the empty bottles into which it was poured were clean and sterile. There has never been the least difficulty in obtaining a sterile, fresh-smelling and sweet-tasting product after heating at 100°C. (212°F.) for 30 minutes and storing at 22°C. (72°F.).

Sources of error in the adjustment of media to pH 3·5 have been discussed.

Titration curves for whey and beer wort with citric, lactic and tartaric acids have been constructed. It is shown that citric acid is the most convenient acid to use for adjusting media for yeast and mould counts. The data obtainable from these curves cannot be relied upon to give a medium of pH 3·5 ± 0·1, owing to the variations in the composition of whey and wort, but the required accuracy may be obtained by adding to the medium a bufEer mixture of citric acid monohydrate (64 parts) and potassium citrate monohydrate (36 parts) in a concentration of 2 per cent. In this way reliable results can be obtained without standardising acids or determining pH. values. This method may be used for any pH between 3 and 6.