532 EXPERIMENT STATION RECORD. 



27.3 cts. per pound. The amount to be paid each of tlie 49 patrons on 



the 2 bases is shown, together with the gain or loss from paying by test. 



"It appears that iu consequence of the change from the old basis to the new, or 

 'test/ basis, 31 persons gained, in the aggregate, $-18. 11; that 18 persons lost, in the 

 aggregate, !f!48.45, and that the creamery pays practically the same aggregate sum 

 for the milk received. The gains and losses fail to balance completely because of 

 small fractions not counted. This means that over $48, or more than 4 per cent of 

 the month's business, is taken from people to whom it is justly due and paid to 

 people to whom it is not due. Many samples of milk are so near to the average 

 'test,' in this case 3.787 [per cent], that they neither gain nor lose materially. In 

 the present instance one half of the persons gain or lose less than a dollar. Were 

 all samples thus near the average it would make but little diiference on which of the 

 2 bases the payment •juight be made. But with the better or poorer grades of milk 

 the difference becomes more considerable, in one case over 23 per cent of the true 

 value of the milk." 



Eeference is made to the confusion which has arisen from expressing 

 the result of the test in terms of butter in some cases instead of in fat. 

 For instance, a sample of milk containing 4 per cent of fat is reported 

 by one creamery as testing "400" and by another as testing "444." 

 While in reality the tests are identical, the one being fat and the other 

 butter, this is not understood by the producer, and it is often impossible 

 for the station to say, when appealed to, whether the "test" is too low 

 or too high because it is not defined and may mean either fat or butter. 

 "Again, where the creamery reporting the test in butter offers 20 cts. 

 per pound, the creamery reporting the test in fat can ofier 23 cts. per 

 pound, and still actually pay a little less. . . . The test should uniformly 

 designate the number of pounds of fat in 100 lbs. of milk." 



The author also points out 2 possible errors in operating the Babcock 

 test, namely, inaccuracy in the calibration of the test bottles, and the 

 temperature at which the fat column is measured. Out of one lot of 

 57 test bottles bought by the station, 28 bottles, or 49 per cent, were 

 found to have an error greater than 0.1 per cent; and out of a second 

 lot of 59 bottles, 25, or 42 per cent, had a similar error. 



"About 7 per cent of the entire lot had an error exceeding 0.2 per cent. These 

 errors do not indicate carelessness on the part of the maker, as it is a difficult thing 

 to attain greater accuracy without a considerable increase in the cost of manufac- 

 ture; but they do show conclusively that the same customer's milk ought not to be 

 tested repeatedly in the same bottle." 



As to the effect of the temperature at which tlie fat column is read 

 off, the variations in the reading of a 4 per cent milk at temperatures 

 from 15 to 100° C. are given, together with the excess or deficit at tem- 

 peratures above or below 55° C. (131° F.), assumed to be the tempera- 

 ture at which the fat is normally read ; and the results are i^resented of 

 a number of trials showing the rapidity with which the column of fat 

 in the narrow neck of the bottle cools off in the air. For instance, 



