AQ G. Jonunstonse Stoney On Polarization Stress in Gases. 
in, in such a sense that the state of the gas would not be disturbed by 
making the sides of the tube impervious to molecules, provided they were made at the 
same time perfect reflectors of molecules. By a reflector of molecules is to be under- 
stood a surface endowed with the property of throwing off any molecules that 
impinge upon it, with unabated speed and at an angle of reflection which lies in the 
same plane as the angle of incidence and is equal to it. The reflected molecules 
will affect the state of the gas within the tube exactly in the same way as the 
molecules passing in from outside had done before. We have now a portion of 
gas completely shut up inside a tube with sides that are perfect reflectors of mole- 
cules, and closed at the ends by pistons that are patches of the heater and cooler 
and which therefore scatter such molecules as reach them; and we know that 
the behaviour of this gas will be the same as that of the corresponding portion of 
the Crookes’s layer. We may call such a tube, a unit reflecting tube. 
7. Let the pistons of such a tube be kept at the temperatures T, & T,, and let 
gas be introduced into it. After a brief period of adjustment the gas will become 
stationary, 7.e., if a plane forming a cross section of the tube be considered, the 
molecular motions are such that the same number of molecules pass forwards as 
backwards through this plane per second. But they will pass it with unequal 
average velocities, because the vis viva of those crossing it towards the cooler must 
exceed the vis viva of those crossing it towards the heater, by an amount bearing 
a known ratio to the quantity of heat advancing. Hence the gas is polarized, the 
molecular motions being swifter when they are directed forward or towards the 
cooler and slower when directed backwards. 3 
8. Suppose that we begin with dense gas and gradually exhaust, and let us 
consider the succession of events that will arise as the exhaustion proceeds, 2.e., 
when n, the number of molecules in the unit tube, is progressively diminished. It 
is known that the flow of heat cannot conform to the laws of “conduction” unless 
the number of molecules exceeds a certain limit which we may call N, N depending 
upon the description of gas that is present, and upon the temperatures T, and T, 
of the pistons which close the unit tube. We must, therefore, divide the exhaustion 
into two periods, one lasting while the number of molecules in the tube exceeds N, 
and the other during the rest of the exhaustion. Throughout the first period the 
flow of heat follows the known laws of conduction, and therefore remains constant. 
Hence, during this part of the exhaustion the polarization of the gas (which may be 
measured by = v being the average velocity at any point of the layer, and 6v the 
average difference of the velocities forwards and backwards at that station), is so 
rapidly on the increase as quite to compensate in Kpv*sv (the expression for the flow 
of heat, p being the density at the station, and K a constant), for the diminishing 
density. During the second period, 7.c., when the molecules have become fewer 
than N, the polarization is still on the increase, but not so rapidly as before, and at 
