( 876 ) 
§ 4. The thermoelement was Platinum-Constantin, the wires of which 
had been carefully annealed *). The Platinum-Copper junctions were 
“protected” by copper caps, according to the method generally adopted 
in this laboratory’). The electromotive forces were measured by 
compensation against a battery of five standard Weston-cells in 
parallel by means of the potentiometer arrangement already published.*) 
The thermoelement was calibrated in steam, oil bath at ordinary 
temperature, liquid nitrous oxide, liquid air, and liquid hydrogen; in 
N,O and liquid air by comparison with a gold resistance thermometer, 
and in liquid hydrogen by comparison with a platinum resistance 
thermometer. Both these resistance thermometers had been previously 
calibrated with the standard hydrogen thermometer. *) The sensitivity of 
this thermoelement is greatly diminished in the neighbourhood of 
liquid hydrogen temperatures; but at higher temperatures it proved 
sufficient for the present purpose. 
§ 5. Fig. 2 shows the arrangement of the apparatus used for 
producing the gas-stream and maintaining it at constant pressure. A 
BROTHERHOOD-cOMpressor, capable of compressing 10,000 litres per 
hour kept the cylinder C,, (capacity 25 1.) filled with dry air at a 
pressure of 70—80 atm., the pressure being indicated by a mano- 
meter M. The pressure in the expansion apparatus was regulated 
by means of the valve K, and its value was indicated by a 
second manometer attached immediately in front of the first copper 
spiral.’) To intercept any moisture that might have found its way 
inside the connections, a drying-tube, B, filled with solid AOH 
was inserted after the valve A. A small pressure cylinder, C,, of 
2—3 litres capacity, surrounded with ice, proved of considerable 
advantage in keeping the expansion pressure constant and in mini- 
mising the disturbing influence of possible irregularities in the flow 
of gas. After C, was coupled the expansion apparatus of fig. 1 with 
a manometer. The cooling spirals were immersed in water contained 
in a glass vessel, and this was completely surrounded by ice in a 
large earthenware pot. Before making a series of experiments, the 
apparatus was allowed to stand in ice for 3— + hours, so as to take 
1) For experiments with helium at tne temperature of liquid hydrogen, the 
thermocouple gold-silver would be preferable. (See Comm. NO. 107). 
2) Comm. Phys. Lab. Leiden N°. 27. 
3) Comm. Phys. Lab. Leiden. NO. 89 (These Proc. Feb. 1904). 
4) Comm. N'. 95e en 99¢ (These Proc. Sept. 1906 and Sept. 1907). 
5) It will be seen later that the pressure indicated by this manometer was not 
the actual pressure at which expansion took place. 
