( 865 ) 
sion into heat of the energy generated, and also to ensure that the 
expanded gas should issue in a “quiet tranquil stream without jets 
or rapids’. (See $ 6, note 1). Natanson and Kester also used a 
porous plug. 
More recently O.szewski has determined ‘‘inversion-points”’ for 
various gases, and has returned to the original reduction-valve form of 
apparatus; but while JouLe and KerviN’s highest initial pressure was 
not greater than 6 atmospheres, OnszEwskI expanded from considerably 
higher pressures. Since the apparatus to be used in my research 
resembled that used by OtszewskKr, a careful examination of his 
experimental results and of the criticisms concerning them which 
have been published was necessary. 
§ 4. O1szewski’s first determination’) was of an “inversion-point”’ 
for hydrogen. This result was subsequently used by PorrerR*) as a 
means of discriminating between the validity of certain equations of 
state. Later Omszmwski investigated *) the dependence of inversion 
temperature upon initial pressure in the case of air and nitrogen, 
and found that the inversion temperature decreased with falling initial 
pressure. ‘The results of this research have been criticised in a 
theoretical investigation recently published by Haminton Dickson. *) 
Dickson reached the conclusion that OrszewskKr’s experiment differs 
fundamentally from that of Jourr and Kervin. This difference he 
attributed to the different kinetic energies possessed by the gas in 
the OLszEwskI experiment before and after expansion. It will be shown 
in a subsequent paper that conduction of heat and loss of pressure 
in OLSZEWSKI’sS experiments are probably much more important factors 
in the result than the change of kinetic energy. 
$ 5. Dickson’s criticism was based upon calculations made from 
VAN DER Waars’s equation of state, but, as this equation is not 
quantitatively correct, the results obtained are of doubtful value. 
But if we calculate from the real isothermals of the experi- 
mental gas, we ought to obtain close correspondence between the 
calculated and experimental values of the Joure-Kervin effect, in 
so far as no uncertainty is introduced by errors in the experimental 
1) K. OuszewskI: Phil. Mag. [6]. 3. p. 535. (1902). 
Ann. Phys. 7. p. 818. (1902). 
2) A. W. Porter: Phil. Mag. [6]. 11. p. 554 (1906). 
3) K. OLSZEWSKI: Phil. Mag. [6]. 13. p. 723 (1907). 
+) J, D. Hamitron Dickson: Phil. Mag. [6]. 15. p. 126. (1908). 
