Abstract
Summary: In experiments to determine the effect of feeding changes on milk yield and composition, udder disease or injury may easily invalidate the results, particularly those for milk composition. It is suggested that regular examination of the milk from each quarter of the udder would allow a diseased or injured quarter to be recognized and a true value for the effect of the feeding treatment to be obtained from the composition of the milk from at least two healthy quarters. Examples to illustrate the suggestion are given.