1. Bacteriological studies have been made of some 400 samples of dried milk products comprising roller-dried separated milk, roller-dried full-cream milk, roller-dried buttermilk, roller-dried whey and spray-dried whey. The samples were derived from a total of plants operating throughout Great Britain.

2. The mean values for the plate counts at 37 and 55°C. of the roller-dried products were all below 1000 per g· as against ½–2½ million at 37°C. and 10,000–30,000 at 55°C. for the spray-dried products. The spray-dried products also gave three times as many positive presumptive coliform tests at all dilutions as did the roller-dried products.

3. The roller-dried whey and the roller-dried buttermilk both gave more variable counts at 37°C. and higher counts at 55°C. than the roller-dried separated milk and the roller-dried full-cream milk. The spray-dried whey gave higher and more variable counts at both 37 and 55°. than the spray-dried milk.

4. Marked differences were observed in the plate counts of samples derived from different plants, some plants giving consistently low counts and others giving higher and more variable figures. Differences were also observed in the counts at different stages of handling.

5. Evidence was obtained which indicates that mould counts do not provide a reliable indication of recontamination subsequent to drying.