NOV. 22 , I915 
Improved Respiration Calorimeter 
323 
observations were recorded. This method afforded opportunity for 
errors in fundamental data, some of which might be obvious, but most 
of which would not be detected. 
measurement by electric-resistance THERMOMETERS 
To relieve the observer of the tedium of these observations, and 
especially to eliminate as much as possible of the personal element 
from the measurement, the mercury thermometers were replaced by a 
device for measuring the increase in the temperature of the water by 
the difference in electrical resistance of two coils of wire in the water 
circuit. Atwater and Rosa (4, p. 25; 5, p. 151) employed a device of 
this character in their original calorimeter, but did not develop it to 
measure temperature differences with the same degree of accuracy as 
the one here described. The latter device comprised two special resist¬ 
ance coils, a special Kohlrausch bridge, a sensitive galvanometer, and a 
lamp and scale for reading the deflections of the galvanometer. The 
specially mounted resistance coils, called the “ bulbs,” were inserted in 
the water line where the bulbs of the mercury thermometers had been 
and were connected with the special Kohlrausch bridge on the observer’s 
table, the two coils being in opposite branches of the bridge circuit, with 
the slide wire between them. The reflecting D’Arsonval galvanometer 
by which the bridge was shown to be balanced was suspended in such a 
position that the scale on which the deflections of the galvanometer 
were read was on a level with and directly in front of the eyes of the 
observer sitting at the table. The movement of the galvanometer was 
indicated by the movement of a vertical line of light along the scale, 
the light from a straight-filament electric lamp being reflected by the 
mirror of the galvanometer. To determine the difference in the tem¬ 
perature of the two coils, it was merely necessary to move the battery 
circuit contact along the slide wire of the bridge until the line of light 
was at the center of the scale, showing that the bridge wis balanced. 
The reading of the bridge scale was then recorded. To balance the 
bridge and read its scale was much more convenient than to read the 
mercury thermometers, and only one record was involved. 
Several types of resistance-thermometer bulb were tried in connection 
with this device. In one, insulated resistance wire was incased in a coil 
of small-bore lead tube, which was immersed in an enlargement in the 
water channel. This proved unsatisfactory for several reasons. One 
was that it did not respond quickly enough to changes in the temperature 
of the water, owing probably to poor thermal contact between the wire 
and the tube; and the mass of metal in the tube also tended to increase 
the lag. Another was that the space in the lead tube was not deprived 
of water vapor, and this eventually moistened the insulation of the resist¬ 
ance wire, so that a short circuit was established between the wire and the 
