1. During four periods equally spaced throughout the year raw milks coming from surrounding farms into the Buckingham receiving station were examined by (a) the 10 min. resazurin test, (b) the 1 hr. resazurin test, (c) the resazurin test in which the time to reach disk 1 was recorded, (d) the modified methylene-blue test, and (e) a keeping quality test based on taste and alcohol precipitation. Altogether, 2588 samples were examined.2. Table 2, in which the correlation coefficients are worked out between the results of the dye and the K.Q. tests, shows that there is little to choose between the methylene-blue test and the time to reach disk 1 in the resazurin test; but since an average of only 15 min. was saved by carrying out the resazurin test, there seems little advantage to be gained by the use of this test. Both of these tests are superior to the 10 min. and the 1 hr. resazurin tests.3. A comparison of Tables 5–7, in which the average K.Q. of the milks corresponding to different disk numbers or reduction times is recorded, shows that the methylene-blue test affords the best indication of the K.Q.4. A study of Table 3, setting out the relation between the 10 min. resazurin and the methylene-blue reduction tests, shows that 99% of milks giving a 0–½ reading in the resazurin test reduce methylene blue within 30 min. and 80 % of them within 10 min.; but that, of milks reducing methylene blue within 30 min., only 53% give a disk reading of 0–½ in the 10 min. resazurin test. If it is agreed that no raw milk on arrival at the creamery which reduces methylene blue within 30 min. and which has, according to Table 7, an average K.Q. of only about 5 hr. at the time of testing, is fit for human consumption, then it is clear that the present rejection standard in the 10 min. resazurin test passes nearly 50% of unsatisfactory milks.5. The data recorded in this paper all go to show that the best indirect index of the K.Q. of raw milk is afforded by the methylene-blue test as carried out by the method officially laid down by the Ministry of Health.6. On the other hand, if time is an overruling consideration, then a slightly less accurate estimate of the K.Q. of raw milk is afforded by the 1 hr. resazurin test. It must be pointed out, however, that the figures given in Table 6 suggest that below disk 2 there is some irregularity in the relation between the disk numbers and the K.Q. If the 1 hr. resazurin test should come into use for the routine grading of milk, a revision of the present tints at the lower end of the scale might prove desirable.