16 BULLETIN £45, U. S. DEPARTMENT OF AGRICULTURE. 
sodium phosphate (Na 2 HP0 4 ) was selected because it is the most 
nearly neutral of the various sodium salts. A very pure prepara- 
tion of this compound, containing 9 to 12 molecules of water of 
crystallization, 7 can be obtained commercially at about $100 per ton. 
Sodium phosphate has been fed in only a few of the numerous 
phosphate- feeding experiments which have been conducted in the 
past. Gouin and Andouard (6) worked with it to some extent, but 
they gave no information in regard to the particular sodium phos- 
phate used or in regard to its water content. If we understand them 
right, their doses were very small in comparison to those which we 
finally used. We began with doses of 4.5 grams of phosphorus as 
di-sodium phosphate daily, and finally gave doses of 24 grams with- 
out producing any noticeable digestive disturbances. 
In deciding on the basal ration to which the phosphate was to be 
added we were largely influenced by the fact that the dry cows at 
Beltsville had previously been fed, as a matter of routine, a ration 
consisting of 4 pounds grain, 8 4 pounds legume hay, and 30 pounds 
corn silage. 
This ration carries decidedly more than enough protein and total 
nutriment to provide for the maintenance of a 1,000-pound cow, 
according to the Haecker and Savage standards (9) ; and according 
to the results obtained by Eckles (3), the surplus should be sufficient 
to provide for the development of the unborn calf. Using a slightly 
less liberal feed for the basal ration made it possible to compare 
the performance of the cows fed phosphate in addition with that of 
the general herd. In several cases, however, a somewhat more liberal 
ration was used both for the control and for the experimental animals. 
Alfalfa was selected as the hay to be used, partly on account of 
its high calcium content and partly because it had been used at Belts- 
ville in the past as often as anything else. 
The experiments fall into two general classes : First, those on ani- 
mals which had been fed for some years previously according to the 
routine used for the general herd; and, second, those on animals 
which had been on test and which, for at least a year preceding our 
experiments, had been fed much more liberally than the general herd. 
In the first class there are two smaller groups. Group 1 consists of 
seven experiments where the same animal served in one year as 
control and in another as experiment. Group 2 consists of experi- 
ments in which the records of animals on the phosphate feeding were 
7 The salt crystallizes with 12 molecules of water, but loses a considerable part of its 
water of crystallization readily on exposure to air. As obtained commercially, there- 
fore, it generally contains less than 12 molecules of water. 
8 The grain mixture fed has been varied from time to time. Those most frequently 
usc-d were grain mixtures B and C. The protein, total nutriment, and mineral con- 
tent of these are fairly typical for all the other mixtures. 
