502 
PHYSICS: M. C. SHIELDS 
A DETERMINATION OF THE RATIO OF THE SPECIFIC HEATS OF 
HYDROGEN AT 18° AND -190°C. 
By Margaret C. Shields 
RYERSON PHYSICAL LABORATORY. UNIVERSITY OF CHICAGO 
Communicated by R. A. Millikan, June 26, 1917 
For the purpose of this study of the ratio of the specific heats of hydro- 
gen the Lummer and Pringsheim method has been adapted to a one 
liter flask with a precision comparable with the extreme precision it has 
given in large carboys. The method consists essentially in obtaining 
7 from observation of the cooling consequent upon small adiabatic 
expansions by way of the ideal gas equation 
(pi/P2)^-' = (siM)^. 
The cooling was in this case measured by a thermal element of 0 . 001 inches 
copper and constantan wires, lightly brazed at the junction, which were 
introduced through glass tubes drawn out inside to fine capillaries and 
spread nearly to the diameter of the flask, the junction being carefully 
placed at the center. The thermal E.M.F. developed by the expansion 
was measured by a null method with a WoM potentiometer, the gal- 
vanometer sensibility being such that an equilibrium temperature could 
be read to 0.0002? Because of the finite heat capacity of the junction 
and the surgings incident to the expansion, it was necessary to observe 
the cooling at a constant interval of time after the expansion, the interval 
varying from 0.6 to 1.0 seconds in different series of observations. With 
a small container there is during this time a considerable inflow of 
heat to the thermo junction by conduction and convection and by radia- 
tion, the total inflow being a function of the time and the mounting of 
the individual junction. Since this inflow is also proportional to the 
cooling, the error due to it is eliminated by taking a series of observa- 
tions for different excess pressures and plotting 7 as a function of Ap 
atmosphere, (this must be Hnear for the expansions used, which did not 
exceed 0.04 atmosphere) and taking the limiting value of 7 for = 0 
as the true 7. 
The method was first tested for air with three different junctions which 
gave lines of quite different slopes, but intercepts 
1.4014, 1.4019, and 1.4017. 
To this value computed from the ideal gas equation must be added a 
correction for internal work, 0.0012, as obtained from either the Berthelot 
