4 BULLETIN" 736, U. S. DEPARTMENT OF AGRICULTURE. 
while the experiment was in progress, the doors were lowered. For 
the purpose of the experiment there was no difference in the two 
structures. The new shed was built only a short distance from the 
main milking barn, so it was convenient to drive the cows from the 
open shed into the main barn to be milked; consequently both 
groups of cows were milked in the same structure during the last 
two years of the experiment. 
THE CLOSED BARN. 
The closed barn was of concrete construction, 36 feet by 59^ feet, 
with stall room for 26 cows, and was equipped with concrete floors, 
mangers, and gutters. The cows faced the outside walls, and the 
alleyway behind them was 8 feet wide. The feed alleys in front 
of the cows were 4 feet wide, and there was a 5-foot alleyway at each 
end of the barn. The 17 windows, 7 on each side, 1 on the north end, 
and 2 on the south end, provided 176 square feet of lighting space. 
A modification of the King system of ventilation was used. 
The concrete floors on one side of the barn were covered with 
various kinds of insulators, such as cork brick, creosoted blocks, and 
planks. One-half of the cows used in the experiment stood on the 
floors and the other half on the concrete. 
THE COWS. 
The herd throughout the entire investigation consisted of 1 pure- 
bred Guernsey, 2 pure-bred Holsteins, 10 grade Jerseys, and 8 cows 
of miscellaneous breeding. The records of all the animals stabled 
under the two systems could not be used, on account of the irregu- 
larity of calving, etc. 
PRODUCTION RECORDS. 
The herd was divided into two groups. During the first year 
one group was kept in the open shed and the other in the closed 
barn. The second year the groups were reversed. The third year 
the groups were again reversed, which gave three years' records for 
comparison. Owing to the irregularity in calving, all the cows have 
not three years' records which are comparable. Four cows had two 
years' records in the open shed, an average of which was taken and 
compared with their one year in the closed barn. Seven cows had 
two years' records in the closed barn, an average of which was 
compared with their one year's record in the open shed. 
Since the results of the housing are determined quite largely, if 
not entirely, upon the stabling period — November to March, in- 
clusive — only the records obtained for the five months were studied. 
These records do not in any case cover the entire period of five 
months, owing to the irregularity of some of the cows in calving, 
