Abstract: | SUMMARY: A co-ordinated experiment was carried out at two centres, in which colony counts from raw milk samples determined by the roll-tube and Petri dish methods were compared. A bulk medium was standardized for both laboratories, as also were details of technique. Statistical analyses showed that roll-tube counts were generally lower than the corresponding Petri dish counts, but the difference varied considerably from milk to milk. The variation between replicate sub-samples was about the same for both methods. Experience and the results both indicate that roll-tubes were slightly more difficult to count than Petri dishes. |