New York AGRICULTURAL EXPERIMENT STATION. 49 
The milk fat could not come wholly from protein, taking the 
urine nitrogen to be the measure of the protein broken down, as 
we necessarily must. Neither could the combined effect of both 
food fat and decomposed protein account for all the milk fat, 
the deficit in the case of three cows being large. 
We are now confined to the alternatives of concluding either 
that previously formed body fat was drawn upon to supply the 
milk fat or that the carbohydrates were utilized for this pur- 
pose. The increased weight of two of the cows, and their gen- 
erally improved condition during the experiments, preclude the 
conclusion that 18.4 Ibs. and 39 lbs., respectively, of pure fat 
could have been withdrawn from their bodies. Such a result 
would have produced a marked change in condition and an 
appreciable diminution of weight. The case of Cow 2 is 
especially noteworthy. This was an undersized cow, thin in 
flesh at the birth of her calf, whose butter fat yield before, 
during and after the experiment was above 12 Ibs. per week, and 
it did not fall below 10 lbs. for some time, so that her butter 
production in a year was not less than 500 Ibs. During her 
largest flow of milk at least one-sixth of her fat product could 
not have been supplied from the normal herd ration she was 
eating, without drawing upon the carbohydrates. To assume 
that her small store of previously acquired body fat was drawn 
upon to the extent of this deficit would be the height of 
absurdity. Cow 12, eating a fat-poor ration, produced 39 lbs. 
of milk fat more than the ration could supply independently of 
the carbohydrates, and to all appearances she could have kept 
on for a year secreting normal milk from the ration she received. 
May we not finally conclude, then, that carbohydrates may be a 
source of milk fat? 
RELATION OF PROTEIN EXCHANGE TO MILK PRODUCTION. 
The relation of protein metabolism to the production of milk 
is interesting and important. It is recognized, of course, that 
there is a positive demand for protein in the secretion of milk 
solids, and it is not difficult to calculate the probable extent of 
4 
