ii 
eight centimeters apart on all sides except in front of the glass door. T: 
was a copper vessel or tank connected by lead tubes In and L, to the tank 
T,. Both tanks were filled with oil. The oil in the tank T:, heated by one 
or more bunsen burners, passed through the tube L, into the tank T, and 
finally back through In into Ts. A stirrer, driven by an electric motor, 
aided in producing a rapid circulation of the oil. Tank T: and tubes In 
and L, were wrapped with several layers of asbestos paper. 
From a flask not shown in the figure water was siphoned to and 
through the tube G: into the evaporating dish D:. An overflow G, served 
to keep constant the depth of the water in the dish. The excess of water 
dropped on sponges S in an evaporating dish D,, itself drained by the tube 
T,. The sponges served to keep the space inside the box saturated with 
watery vapor, or nearly so. An inverted evaporating dish D, served to 
enclose almost completely the frame and liquid and thus insure the satura- 
tion of the space abcut the film on which the measurements were made. 
The dish D: rested om a wooden platform P supported at one end by 
a hinge and at the other ead by a cord C passing over a cylindrical metal 
rod which extended to the outside of the box. The height of the water 
surface was slowly raised or lowered by twisting the rod. 
A thermometer ti gave the temperature of the oil, t, the temperature 
of the water, t, the temperature of the space immediately above the water, 
and t, the temperature of the space outside the inverted evaporating dish. 
No measurements were made when the thermometers t., t,, and t, differed 
by more than a few tenths of a degree. This necessitated a wait of from 
one to five hours between readings at different temperatures. Three series 
of readings were taken, each requiring a continuous run of from ten to 
thirty-six hours—depending upon the number of observations made. 
Owing to the condensation on the wire W where it passed through the 
opening in the tank T, it was not practicable to carry the observations 
higher than 80°. An effort was made to prevent this condensation by driv- 
ing gently through the opening a stream of warm air from the outside. 
But this interfered somewhat with the action of the balance and the satu- 
ration of the space inside. It did not occur to the writer at the time to 
try heating the wire by means of an electric coil. 
For temperatures below room temperatures the asbestos was removed 
from the tank 'T: and the tank was surrounded by a large vessel containing 
[12—26988 ] 
