11 Dec, 1916.] Milktiuj Experiment. 751 



8. The cows which were milked at even periods gave more milk in 

 the evening than in the morning. The evening's milk was also slightly 

 poorer in quality than the morning's. This is contrary to generally 

 accepted exjjeriences, as in ordinary farm practice unequal periods of 

 milking are the rule. 



NOTE ON THE VARIATIONS IN THE AMOUNT OF FAT IN 

 MILK DUE TO VARIATIONS IN THE TIMES OF MILKING. 



S. H. Collins. 



In 1911* I deduced a formula from over 600 tests of milk, which 

 showed that if cows were milked at 6 a.m. and 6 p.m. the evening's 

 milk would be .2 j>er cent, fat poorer than the morning's milk. If 

 the cows were milked at 6 a.m. and 5 p.m. then the evening's milk 

 would be the richer to the extent of .3 per cent. fat. If milked at 

 6 a.m. and 4 p.m. the evening's would show .8 per cent, more fat, and 

 if milked at 6 a.m. and 3 p.m. the evening milk would be 1.3 per cent, 

 richer. 



The best test of any sinnlar law, based on averages, is to find out 

 if it agrees with facts discovered after the rule has been established. 



In the results explained by Mr. Walker it may be seen that in 

 twelve weeks' trials the evening milk (6 p.m.) is poorer than the 

 morning's (6 a.m.) by .25 per cent, fat with equal intervals of milking 

 and tliat the evening milk (4 p.m.) is richer than the morning milk 

 (6 a.m.) by .80 per cent, fat with unequal intervals. These results are 

 almost exactly what the rule predicts, namely, .25 against .20 and .80 

 against .80. As regards the weeks taken one at a time, there is only 

 one week with equal intervals, during which the experiment differed 

 from the rule by more than .2 ])er cent., and there were only three 

 weeks with unequal intei-vals where the experiment differed from the 

 rul? by mere than .2 per cent. 



If one milking was twelve minutes too late and the next milkin,g 

 twe've minutes too early, there would be a discrepancy of .2 per cent, 

 fat, or the same discrepancy would result from a delay of twenty-four 

 minutes from one milking, the other milk being at correct time. Since 

 accurate timing is quite impossible, the few cases where the experiment 

 appears to depart from the rule may be no departure at all, but merely 

 be the result of the difficulty of accurate time-keeping. 



These results confirm many previous ones in various parts of the 

 country and prove conclusively that the |>ercentage of fat in milk 

 de]>ends very largely on the times of milking and that the relationship 

 between times of milking and percentage of fat is a constant one. 



Tahlk LXXXTIT. — AvERAfJK Effect of Times of Milkint.. 



I'.vi'iiiii',' Milk will 1)0^ 

 jM.Mnr liy 2 "f, fat than tho moniini; milk. 

 r . !irr .. •O.'i % 



:< % 



-^^ % 



•8 0/ 

 ,.1 (>5% 



• Proceedlng-i of thf (Tnlver^ify of liiirliain I'lilliwophlon! Snclrfy, Vfil Iv., I'lirf I. 



