G. JounstonE Stoney On Polarization Stress in Gases AQ 
from various considerations that the flow of heat 
Ga pV23V 
or, by a simple transformation, 
Glcg AIBN? 6 > 0 bo & (9) 
T being the temperature measured from absolute zero, Hence from the approxi- 
mate equations (8) and (9) we obtain the equation— 
koe —— . . . (A) 
/ 
which contains only quantities of which we already know enough to make use of 
them. qu. (A), may be thrown into a still more convenient form by writing P 
for the tension of the surrounding atmosphere of gas, which is nearly the same as 
the stress which the gas at the station we are considering would exert if de 
polarized. P will therefore vary nearly as pT, whence— 
2 
Kaeo) 
16. As an example of the application of these approximate formule, let us plot 
down on adiagram the value of «, the polarization stress, for various tensions of gas 
between a heater and cooler at constant temperatures and at a fixed distance 
asunder. 
Y Let the abscissas of the figure represent the tensions of the 
gas. Then the curve Obed, the ordinates of which represent the 
flow of heat, is known. The part representing conduction is the 
parallel line cd, and Obc represents the outflow of heat by that 
modified conduction which may be called penetration, which 
occurs when the exhaustion has proceeded so far that the num- 
ber of molecules in a unit tube is less than N (see above, section 
. 8). The curve Obed is therefore known, and if by equation (B) 
we plot down from it the values of «, the polarization stress, we 
find them approximately represented by the ordinates of a curve 
\ of the form Oace, the portion to the right of ¢ being coincident 
a with an equilateral hyperbola, while to the left of ¢ the ordinates 
(nn) fall short of the hyperbola, rising to a maximum and then falling 
ST CT , low tensions. Bearing 
this in mind, the accordance of the theoretic values with those determined experi- 
mentally by Mr. Crovkes and Mr. Moss, is satisfactory. 
! 
t 
{ 
! 
! 
1 
i} 
1 
\ 
1 
\ 
\ 
i} 
‘ 
\ 
1 
\ 
\ 
off'to zero. The position 
of this maximum can- 
not be obtained with 
certainty, because equa- 
tion (B) is less to be 
depended on at very 
12 
