M. E. Pringsheim on the Radiometer, 123 



rated velocity than if it possessed the same temperature as the 

 surrounding gas. In consequence of this the warmer side of 

 the radiometer-vane is exposed to a greater repulsion or pres- 

 sure than the cooler side, and therefore moves backwards. 

 Just so these warmed gas-molecules repel a surface on which 

 they impinge, as they give up to it their surplus velocity. 



Moreover, this explanation requires our proposition on 

 radiometer -motion given above (p. 121) to be somewhat modi- 

 fied, since according to it a surface-element always undergoes 

 repulsion, even when it has the same temperature as the gas. 

 But as in reality a surface is never exposed to motion by the 

 gas without its back side being likewise in the rarefied gas, 

 the modification is in practice superfluous. 



We have not yet explained how it is that the motion does 

 not commence until the gas has reached a certain degree of 

 rarefaction; and this is the point in the explanation of which 

 the different theories essentially differ. Of all these theories, 

 that framed by Osborne Reynolds, communicated and adopted 

 by Schuster*, seems to me to come nearest the truth. It 

 rests essentially on the assumption that the predominant 

 motion in a determined direction communicated to the gas- 

 molecules by the passage of heat from a warmer body to the 

 gas cannot be again withdrawn from the gas by the collisions 

 of its molecules, but is only withdrawn when the motion 

 passes again from the gas to a solid body. Now, at the places 

 where the unilateral motion enters and leaves the gas certain 

 forces become operative. 



This theory agrees perfectly with the above-given general 

 law of radiometer-motion ; and the motion which appears in 

 the solid body is quite simply accounted for by the impact of 

 the gas-particles and the reaction of the case. That at the 

 same time the force acting upon an element of the surface is 

 proportional to the intensity of the heat-current follows im- 

 mediately from our assumption if it be presupposed that the 

 rarefaction is so great that, on both sides of the vane, each 

 surface- element is struck by an equal number of molecules 

 with ihe same mean velocity. Let the number be n, and the 

 component perpendicular to the surface-element, of the mean 

 velocity of the impinging molecules, v; then, assuming that 

 one side of the vane has the same temperature as the gas, and 

 putting the mass of a molecule = 1, the quantity of motion 

 communicated to the surface-element on this side in unit time 

 is equal to 2nv. If on the warmer side of the vane the mole- 

 cules rebound with the mean normal velocity-component 



* Nature, xvii. p 143 (1877). 

 K2 



