54 STORRS AGRICULTURAL EXPERIMENT STATION. 
first test, another twelve-days’ test was made of the same herd. 
A comparison was thus made of the yields of milk and butter- 
fat with the two different rations. : 
During the third winter (1894-95) four herds were visited, 
each herd being under observation for twelve days at two dif- 
ferent periods in the same manner as the three herds studied 
in 1893-94, except that there was only a two-weeks’ interval 
between the two tests on the same herd. 
HERD TESTS DURING 1895-96. 
During the fourth winter (1895-96) two herds were studied 
in a similar way, except that in the one case the ration with 
much larger quantity of protein, and a much narrower nutti- 
tive ratio than usual, was used. Samples of the different feed- 
ing stuffs used in the tests were taken early in each test and 
sent to the laboratory for analysis. 
In the earlier tests, as soon as it was possible to obtain the 
results of the analyses, the proportions of nutrients in the 
ration fed was calculated, and suggestions were made for 
changes in the ration. After changes had been made and the 
animals had been upon the new ration for two weeks or longer, 
the herd was again visited and a new twelve-days’ test was 
made. In the tests during 1895-96 the Station representative 
stayed at the farm and made the change of ration. In these 
cases only nine days intervened between the two tests on the 
same herd, and it became necessary to calculate the first ration 
from average tables of analyses, as a basis for formulating a 
new ration. This was done with the idea that it would be 
best to have the Station make the change of feed. ‘The time 
proved rather short, however, for making the change, and the 
present winter (1896-97) we have gone back to the plan of 
allowing two weeks between tests on the same herd. 
The chief points upon which information was obtained were: 
Number of animals in the herd.—In considering the number 
of animals, only those which came into the test were included. 
Usually these were all of the cows on the farm which were in 
milk at the time of the test, except those which were nearly dry. 
Breed, age, and approximate weight of each cow.—The breed 
and age were obtained as accurately as possible from the owner. 
Since it was not practicable to take to the farm scales large 
a a 
