90) 
for the measurement of the resistance is also represented. The galvano- 
meter was of the DrmsserHORST type with a period (undamped) 
of 4 sec. Before a calorimetric determination the resistance was 
measured, and the deflection of the galvanometer for a definite 
change in the ratio of the branches in the THomson bridge, the 
“sensitivity”, determined. Then with a definite ratio of the branches 
the movement of the galvanometer (Fig. 3 and 4) was followed 
during the experiment. After this the measurements of the resistance 
and of the sensitivity were repeated. The current for measuring the 
resistance was 5 milliamperes. The heating current was for the lead 
30 tot 50, for the copper 12 to 27 milliamperes. Considering the ratio 
between the resistances of the thermometer and the heating wires 
the development of heat in the temperature measurement does not 
come into account. 
We gladly record our cordial thanks to Dr. G. Housr and Mr. P. 
G. Carr for the aid they afforded us in undertaking the temperature 
measurements. 
Account had to be taken cf the circumstance, that the reduction 
factor, which gives the ratio between the galvanometer current in 
the Thomson bridge and the difference between the measured resist- 
ance and the resistance which would give no deflection, depends 
on the resistance in the variable branches of the bridge. A correction 
for this was always applied. 
§ 3. The resistance thermometer. For the resistance thermometer 
a gold wire was chosen, as in the range of the measurements 
(14—90° K.) gold is more suitable for interpolation than platinum *), 
and as on the other hand the indications of the gold thermometer are 
constant, if it is suitably treated before use (glowing before winding, 
then repeatedly cooling in liquid hydrogen and returning to room 
temperature’. Before each series of measurements the constancy of 
the resistance of the wire was tested by a measurement at the boiling 
point of hydrogen. 
In the temperature range of liquid hydrogen the gold thermometer 
was calibrated with the aid of a hydrogen vapour pressure appa- 
ratus as used in Comm. N°. 137d. The temperatures were deduced from 
formula (1) of that paper. In this way the scale of the gold thermo- 
meter is reduced to that of Ptr, which has been once more accu- 
rately fixed by the research of KAMERLINGH ONNes and Horst (Comm. 
N°. 1414). In the range of liquid oxygen this was done with the aid 
1) H. KAMERLINGH ONNES and G. Hoist. Comm. N°. 142a@ (June 1914). 
