The collection of samples of milk from individual quarters has usually been done either by hand milking or by sampling by hand at the commencement of milking, before the application of the teat-cups for milking by machine. The former method is laborious, and may not be entirely accurate, as it depends a great deal on constant care throughout the milking to ensure that each quarter is milked into the correct container. If the quarters are milked separately by hand, the yield of milk from the individual quarters may be affected, and if the cow is not accustomed to being milked by hand, hand milking for quarter sampling may disturb her, and affect the yield, and so the composition of the milk. Quarter sampling by hand before the commencement of machine milking does not provide a representative sample of the yield of milk from the quarter, and it does not permit a measurement of quantity of milk produced by each quarter. In connexion with some work on the composition of milk from quarters showing a positive brom-thymol blue test (Hume(i), McDowall(2, 3)) and on the influence of such milk on the quality of cheese and butter made from it (McDowall(4)); the collection of full quarter samples of milk on a considerable scale was required. For this purpose a special milking-machine claw, a milking-machine bucket, and a strip bucket were constructed. These are illustrated in Fig. 1, I–IV.