( 877 ) 
up the same temperature throughout; the absence of current in the 
galvanometer showed when the temperature within the expansion 
chamber was the same as that of the surrounding bath. 
When an expansion was commenced, the deflection caused by the 
thermoelectric current in a HARTMANN and Braun dead beat galvano- 
meter (Comm. N°. 89, Proc. Nov. 1903) was watched; when the 
deflection became constant — usually in 3—5 minutes — showing 
that a steady state was reached, the electromotive force was measured 
by the potentiometer, accurate compensation being obtained by switching 
in an astatic Dusois and Rusens protected galvanometer (cf. Comm. 
N°. 89). When one measurement was completed, the valve A was 
altered until the next pressure in the series was reached, and the 
series Of Operations was repeated. 
$ 6. Expansions made with this apparatus were, however, some- 
thing unusual, as is at once evident from 
PAB EDE 
Series I. Brass valve A. Air at 0°C. 12/12/07. 
p=) Pom 
9.7 atm. warming effect 
AOE ie. do. - 
242 ,, + 0.86°C. 
290% + 1.49 „ 
Onl on + 2.76 ,, 
48.4, a LO, 
The fact that a warming effect was observed at lower pressures 
at once led to the suspicion that heat was being conducted along 
the valve to the expanding gas. During expansion, the gas-jet attains 
a high kinetic energy, and, correspondingly, an abnormally low 
temperature, and this kinetic energy is re-converted into heat only 
when the gas comes to rest after expansion; hence the expanding 
gas can take up heat from the valve, which, when added to that 
given by the reconversion of the kinetic energy, causes the final 
temperature to become too high. 
To determine the influence upon the heat-effect of the quantity of 
gas issuing from the valve, this was opened wider and another series 
of measurements made. 
