Existence of the Luminiferous ^Etliev. 479 



1877, and also in the pages of the ' Chemical News ' for 

 1879. 



Stripped of its mathematical accompaniments, this theory 

 may be briefly stated • thus : — If two surfaces at different 

 temperatures be opposite to each other with a gas between 

 them, there exists a force or stress tending to separate them. 

 This is almost self-evident in the case of a volatile liquid rest- 

 ing on a hot surface; for the stress then has to support the 

 weight of the drop in addition to the atmospheric pressure. 

 If one of these surfaces be movable, we should be able to 

 produce motion, and, under certain conditions, rotation, by the 

 action of this stress. This is just what Mr. Crookes has done. 

 In the radiometer the blackened surface of the vane becomes 

 slightly heated, owing to its absorption of the solar rays. 

 Excess of pressure between it and the sides of the containing 

 vessel is thus produced, resulting in the rotation of the fly. 

 The blackened surface of the vane becomes the heated surface, 

 the sides of the containing vessel the cooled. Now, if the 

 motion is produced by the mutual action of the two surfaces, 

 and in the radiometer the heater is movable while the cooler 

 is fixed, if we make the cooler movable and fix the heater, 

 then we ought equally to get rotation. This is the difference 

 between the otheoscope and the radiometer: in the first the 

 heater is fixed and the cooler movable; in the second the cooler 

 is fixed and the heater movable. Mr. Crookes has brought 

 forward abundant experimental proof of the existence of this 

 stress; and we have now to consider how it is brought about. 

 Imagine two surfaces of unit area marked out on the heater 

 and cooler, then the gas between them will be in active mole- 

 cular motion, but stationary as far as regards convection-cur- 

 rents and wind. Suppose perpendiculars drawn to the two 

 surfaces from the boundaries of these surfaces. Then we shall 

 form a tube whose two ends are at temperatures T Y and T 2 , 

 and whose sides (let us imagine) are perfect reflectors of mo- 

 rales. Since the gas is stationary, we must have that across 

 any section of such a tube equal numbers of molecules are 

 passing in opposite directions at the same time. But the 

 average velocity of those passing towards the cooler is greater 

 than the average velocity of those passing towards the heater, 

 because of the difference in temperature between the two sur- 

 faces, and, moreover, bears a known ratio to this difference. 

 Thus in such a gas a state of stress would be set up which 

 would produce a mutual repulsion between the two surfaces. 

 These are the aspects of the subject which are really important, 

 the further investigation of the motion of the molecules, i. e. 

 their impact, does not. concern us at present. It is sufficient 



2M2 



