102 
MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
Fig. 14. 
Exhaustion 117 M. Exhaustion '18 M. 
I have already shown (312) that when thin aluminium vanes are exposed to candle 
light the metal becomes equally warm throughout, and'a layer of molecular pressure 
is generated at its surface. The thickness of this layer of pressure, or the length of 
the lines of force of repulsion, varies with the degree of exhaustion, being longer as 
the exhaustion increases. The lines of force radiate from eveiy part of the warm 
metal, and being strongest in a direction normal to the surface it will be convenient 
only to recognise these in a discussion of their action. The force of repulsion is also 
greater the closer the repelled body is to the generating or driving surface, and the 
force diminishes rapidly as the distance increases, according to a law which has not 
yet been accurately determined,* but which does not appear to be tire law of “ inverse 
squares.” It will also simplify matters if the interfering action caused by the pressure 
generated by the warm glass bulb is neglected. 
In fig. 14, A, the exhaustion is low (117 M), and the thickness of the layer of 
pressure I assume to extend to a distance of 10 millims. from the surface of the 
aluminium vane. The lines in the diagram are represented as extending for that 
length from the hemi-cylincler in directions normal to the surface. Where the lines 
of force extend from the fly to the glass case, as at a to b and c to cl, pressure is 
exerted along the line, and repulsion ensues ; these are represented by black radial 
lines. But where the fines of force do not reach to the glass, as shown by dotted 
fines, no pressure is exerted. In the position and under the conditions shown in 
fig. 14, A, the rotation must be in the direction of the arrow or positive, as the 
pressure in the positive direction exerted by the fines of force between a and b more 
than counterbalances that in the negative direction exerted by the fines between 
c and cl. Inasmuch as some active fines of force are cut off by the screen, increase of 
the distance between the screen and the convex surface of the vane will allow more 
rays of force to become active, and will increase the speed in the positive direction. 
416. In fig. 14, B, I have represented the action when the screens and vanes are in 
the same position as in 14, A, but at an exhaustion of T8 M. Here the thickness of 
the layer of molecular pressure is supposed to extend to a distance limited by the 
* See the Author’s “ Experimental Contributions to the Theory of the Radiometer.” Proc. Roy. 
Soc.,, November 16, 1876, No. 175, vol. xxv. p. 310. 
