Abstract
The study of about 24,000 samples of dairy products by 56 laboratories proves that modified methods produce a greater spread between counts of truly good quality and truly poor quality products than does the present standard method. The numerical increase in count is slight in the case of truly high quality products.