330 
ME. W. CROOKES ON EEPULSION RESULTING EEOM RADIATION. 
which luminous rays appear to exert on a blackened surface. In the first case, even 
when the gauge and the barometer were appreciably level, and the pump had been 
working for some hours, the superior repulsion of the black over the white was not so 
strong as it was in the second case when the gauge was several millims. below the 
barometer. 
131. These two experiments, the one showing a marked difference of action on a black 
surface between heat of low intensity and luminous rays, and the other showing that 
this difference may be neutralized by aqueous vapour, explain most of the anomalies 
I have met with ; and especially they prove how it was that my earlier experiments with 
black and white surfaces failed to show much difference. They also prove that still 
further improvement in the vacuum-producing apparatus would be advisable. I accord- 
ingly adopted Dr. Argus Smith’s and Professor Dewae’s plan of absorbing the residual 
gas by means of cocoanut-shell charcoal ; and I found that after a little experience this, 
although somewhat tedious, left little to be desired in the perfection of the vacuum. A 
glass tube about 6 inches long is tightly packed with small pieces of freshly ignited cocoa- 
nut-shell charcoal ; it is then drawn narrow at each end and sealed on to the apparatus, 
between it and the spiral glass tube*. The exhaustion proceeds as usual till the gauge and 
barometer are appreciably level ; the charcoal-tube is then heated to a temperature well 
within the softening point of the glass, when the occluded gases are given off from the 
charcoal and depress the mercurial gauge 30 or 40 millims. The pump is now worked 
rapidly until the gauge is brought up again ; the heating of the charcoal is repeated, 
when more gas is given off and the gauge is again depressed, although not so much as 
before. The pump is again set going, and these operations are repeated until heating 
the charcoal ceases to depress the gauge. The effect of this has been to repeatedly 
wash out the residual atmospheric air and aqueous vapour from the interior of the 
apparatus, and replace it by gas or vapour which has been occluded by the charcoal, and 
which we are justified in supposing will be again occluded by it even when very 
highly exhausted. When these operations are finished, and no more gas is carried 
down by the mercury, the apparatus is removed from the pump by sealing off the tube 
at the narrow part between the charcoal and the spiral, so as to leave the charcoal still 
in connexion with the apparatus. The two together are now set aside for some weeks, 
when the charcoal will gradually absorb the whole of the residual gas and leave the 
vacuum so nearly perfect that it will not conduct an induction-current of electricity. 
In most of my experiments this refinement is not necessary ; but in some, especially when 
working with the apparatus subsequently described, I prefer to adopt it. When it is 
considered that the charcoal has exerted its full action the tube containing it may be 
drawn off before the blowpipe, and the apparatus left ready for use. 
132. The repulsion being due to the action of radiation on the surface of bodies, it 
became of interest to ascertain whether doubling the amount of incident radiation 
would produce double the movement. In my earlier apparatus I could not detect any 
* In some of the subsequent woodcuts of apparatus (135, 145) this charcoal-tube is shown in its place. 
