80 Rrport of Department of Animal TTusBANDRt of the 



or composition of tlie milk. In the former test, during 59 days, 

 18.4 lbs. of fat was formed in the milk which could not have had 

 its source in food fat or food protein and could hardly have 

 been drawn from the cow's body fat as she increased in weight 

 33 lbs. in the same time. In this test Cow 12 in 74 days pro- 

 duced 39 lbs. of fat similarly unaccounted for, with a body gain 

 of 15 lbs.; and Cow 2, in 4 days, 1^ lbs. These amounts of fat 

 must have come from the carbohydrates in the food. 



A lessening of protein supply in the food did not produce a 

 corresponding decrease of protein in the milk solids, but caused 

 a marked lessening of protein decomposition in the body. 

 Calorimeter determinations sl>ow that the heat value of urine 

 bears no constant relation to its nitrogen content, and also 

 prove that the formula used in computing heat energy of urine, 

 Nx5.343 Cal., is greatly in error, actual results being from 3 to 

 4 times as large as calculated by this formula. The energy 

 values of nutrients as given by Rubner, — protein and carbohy- 

 drates each 4.1 Cal. and fats 9.3 Cal. appear to be fully high 

 enough for herbivora, even when the loss due to escape of 

 unoxidized gases, methane chietly, is not considered. 



Over 49 per ct. of the available energy value of the rations 

 was used for maintenance, over 30 per ct. reappeared in the 

 milk solids, leaving a balance of from one-fifth to one-fourth of 

 the ration. The logical conclusion is that this balance, in part 

 at least, sustains the work of milk secretion. 



INTRODUCTION. 



Bulletin No. 132 of the New York Agricultural Experiment 

 Station presented the results of an. experiment to determine the 

 sources of milk fa't as related to the food suppl3\ The main con- 

 clusion therein stated was that milk fat can be formed in part, 

 at least, from carbohydrates, the data of the experiment point- 

 ing to this conclusion in a most convincing way. It was felt, 

 however, that so import aut a generalization, if correct, should 

 be sup])orted by results secured with more than one animal. 

 An ()p]K»i-tuiiity was d-sircd also for enlarging the scope of the 

 observations. KxiteriiKculs have been conducted, therefore, 



