ME. W. CEOOKES ON REPULSION RESULTING EEOM RADIATION. 
329 
The fusible metal bath was gradually increased in temperature up to dull redness, 
and the action of the copper ball heated in it was tested from time to time ; the tem- 
peratures were not ascertained, as they were above the boiling-point of mercury. The 
repulsion of the black disk increased until at dull redness the copper ball caused the 
rod joining the two disks to make an angle of about 40 degrees. At a full red heat the 
ball repelled the black disk very strongly, causing the rod to oscillate violently, and 
sometimes even to pass the axial position. 
129. A candle brought near the apparatus acted on the disks even more energetically 
than the red-hot copper. At a little distance off the movable rod set at an angle of 
45 degrees ; and by causing the candle to approach or recede, the angle formed by the 
rod varied in a corresponding manner, the torsion of the suspending fibre balancing the 
varying force of radiation. 
130. During the exhaustion of one of these pieces of apparatus, an action of aqueous 
vapour was observed which explained some of the anomalies I had met with in the 
course of this investigation. The apparatus had a little water in it ; and although the 
mercury-pump brought the gauge to within about 8 millims. of the barometric height 
in the course of ten minutes, the tension of the aqueous vapour prevented it from rising 
higher. After working the pump for several hours, and gently warming the different 
parts of the apparatus, the liquid water was evaporated, and only aqueous vapour 
remained. The gauge now rapidly rose to the height of the barometer, the apparatus 
necessarily being filled with the residual aqueous vapour. On bringing a lighted candle 
near the disks I expected to see the black one violently repelled ; but instead of that 
the connecting-arm set equatorially, showing that the radiation from the candle within 
a few inches of the disks repelled the white one as strongly as it did the black. The 
pump was kept in action, and oil of vitriol was passed through it once or twice (44). 
This was continued for about four hours ; and on testing the apparatus from time to 
time with a candle the repulsion of the black disk gradually increased, the arm setting 
at a greater and greater angle from the equatorial position, but at no time getting very 
strongly deflected. 
An accident happening to one of the tubes of the pump, it was necessary to let air 
into the apparatus ; it was passed in slowly over oil of vitriol. As soon as the pump 
was mended exhaustion of the apparatus was recommenced. As soon as the gauge rose 
within 6 millims. of the barometric height the candle was seen to repel the black disk. 
At 3 millims. the superior repulsion of the black over the white disk was sufficient to 
cause the arm to set 45° ; and as the exhaustion got better the repulsion of the black 
disk increased, until at the point when the gauge and barometer were level the candle 
exerted a strong action many feet off, and when brought close to the instrument set the 
bar and disks in most violent agitation, the black disk being driven violently away, and 
the connecting-arm swinging rapidly on each side of the axial position. 
This experiment shows that the presence of even a small quantity of aqueous vapour 
in the exhausted apparatus almost, if not quite, neutralizes the more energetic action 
