212 
DR. H. T. BARNES ON THE CAPACITY FOR HEAT OF WATER 
cell tanks and regulator are not shown in either view. They were at the other end 
of the laboratory. 
Water Supply for Calorimeter .—The water supplied to the calorimeter passed 
through 40 feet of pure block-tin tube, \ inch in diameter, coiled up in the constant 
temperature tank. After passing through this and taking the temperature of the 
jacket water, it was passed from the tank to the water-jacket and into the 
calorimeter through a rubber tube placed inside the water circulating tube. By 
this means the water, after entering the tank, was entirely in the circulating system 
until it flowed out of the calorimeter. The head of water to maintain a steady flow 
was supplied from a reservoir placed on the floor above, and was connected to the 
apparatus by a glass-tube passed through a hole in the floor. In order to vary the 
supply, the water was passed through a series of fine tubes acting as resistances, 
which could be short-circuited by larger tubes connected across them. These larger 
tubes, offering no resistance, were thick-walled rubber tubes and connected to the 
ends of the fine tubes by T-pieces. When the water was to be passed through the 
fine tubes, the rubber tube was simply closed with a pinch-cock. The resistance 
tubes were two principal ones, 1 metre long and 1 millim. in section, and three lesser 
ones for fine adjustments. These were all immersed in a water-bath to keep them 
from changing in temperature suddenly, and thereby producing a change in the flow 
by changing the viscosity of the water. 
No device was used to maintain a constant head, as a slight falling-off in the flow 
was rather an advantage than otherwise, as it tended to compensate for the slight 
falling-off in the electric current supplied to the calorimeter by the large accumulators. 
Two large bottles, holding about 4 gallons each, formed the head and were connected 
in parallel. A layer of heavy paraffin oil was put over the water to prevent 
absorption of air by the distilled water, which was always used for the experiments. 
This was supplied to the bottles in sufficient quantity for about two experiments, and 
was run in under the oil through a T-connection in the tube connecting the two 
bottles together and with the supply tube for the apparatus. The water was first 
boiled in a large copper tank, and while still boiling was siphoned over into the 
bottles. It was allowed to cool before being used for an experiment. This method 
of boiling the water was used for all the earlier experiments below 60° C.; but it was 
found impossible to obtain steady conditions of flow above this limit, owing to the 
liberation of air inside the calorimeter even from the boiled water. This was some¬ 
what surprising and delayed the attainment of the final measurements at the high 
temperatures. • It appears that to remove the last trace of air from boiled water, it is 
necessary to submit it to extreme agitation. Sufficient agitation was supplied to the 
water, as it ran through the fine-bore tube of the calorimeter, to set free some of the 
air retained by the boiling water when it was run into the bottles under the oil. 
It was found necessary to devise some method of preparing absolutely air-free 
water before readings at the higher points could be obtained, To do this I found. 
