49 



ment by itself, however, does not furnish sufficient evidence to 

 draw any conclusions on this point, and especially as it has 

 been shown already that the loss of weight was largely due 

 to one cow in both lots. Moreover, the figures in Experi- 

 ment A, as shown in Tables II. and III., do not indicate that 

 when the cows are gaining in live-weight there is any improve- 

 ment in the yield or quality of the milk. It stands to reason, 

 of course, that if there is any large decrease in live-weight 

 due to ill-health or improper feeding, there should be a falling 

 off in the milk yield. Such, however, was not the case in 

 this experiment. 



EXPEBIMENT D. 



As soon as Experiment C was completed, the rations 

 were reversed, lot I. now receiving the heavy and lot II. 

 the moderate ration. The cows were left for one week to settle 

 down to the changed conditions before any samples of milk 

 were taken. The fourth experiment, here described as 

 Experiment D, continued from February 1, to April 25, 1905, 



TABLE XV. -QUANTITY OF MILK IN PINTS PER LOT PER DAY 

 FOR 12 WEKKS. 



