BETWEEN THE FREEZING AND BOILING-POINTS. 
215 
generally required for a complete ice determination. It may be said, however, that 
these experiments were among the easiest to obtain. 
At temperatures intermediate between the ice-point and room temperature, the 
regulator was used, and a stream of cold water run through a coil of metal tube 
placed in the tank. Water direct from the water mains usually sufficed for the 
cooler, but for lower temperatures the tap water was run first through a coil of tube 
immersed in a tank of ice. The desired temperature was attained through the 
equilibrium between the heat supplied by the lamp in the tank and the heat absorbed 
by the water flowing through the circulating tube. The tap Avater in the high- 
pressure mains varied, from about 5° in winter to about 18° in summer, in the 
laboratory where the apparatus was located. A temperature at least 2° lower than 
the desired temperature of regulation was required in the cooling water. The 
temperature of the laboratory varied from about 18° to 25° during the different 
seasons; but, in general, was steady to 1°, and often less during an experiment 
lasting 3 or 4 hours. When not using cooling water in the tank, the temperature of 
the apparatus had to be maintained a few degrees above the laboratory, so that the 
regulator could make up for the loss of heat to the surrounding air. Hence a 
temperature of about 26° or 27° in the tank was the most convenient point at which 
to work, which for an 8° rise in the calorimeter gave a mean temperature of about 
30° C. 
As the temperature of the jacket was increased more and more above the room 
temperature, more and more gas was required to aid the electric-heating lamp in 
maintaining a constant temperature. At a temperature of the jacket of about 
90° C. two gas flames were required, and, in addition, the 50 c.-p. lamp. Even for 
the highest points the regulation proved to be most satisfactory, so long as the gas 
pressure did not vary. It was found that the high temperature experiments had to 
be taken at times when there was no other gas being used in the building, as even in 
using the large constant-pressure gas governor fluctuations in the gas pressure in the 
building caused perceptible alterations to the regulator. Where the fluctuations 
were small and regular, the jacket water seldom varied as much as '01° during the 
time for obtaining the observations for each flow, even at the highest point 
attained. The wonderful efficiency of the regulator and circulating system, together 
with the preparation of air-free water, made it possible to obtain the observations at 
the higher temperatures with almost the same degree of accuracy as at the lower 
points. 
At whatever temperature the jacket was maintained, and before the electric-heating 
current was turned on, the calorimeter water was allowed to flow through the 
apparatus for some time after the steady conditions were attained by the regulator. 
The balance-point so obtained for the differential thermometers we have already 
termed the “ cold” readings, in contra-distinction to the readings obtained when the 
electric-heating current is turned on. The second balance-point, together with the 
