270 
ME. W. CEOOKES ON REPULSION RESULTING EROM RADIATION. 
No Screen.' Watef screen* 
1217 9-2 
89-5 8-8 
33'2 0-4 
also acted to the naked flame and behind water in accordance with theory, giving 
good rotation in one case and no movement in the other. 
Covered with a hot glass shade the platinnm compound was repelled, producing 
rotation. On removing the source of heat, and allowing the bulb to cool, reversion 
of movement took place. 
252. In addition to these couples, experiments were also tried in the apparatus 
with other substances likely to give valuable results. Metallic iron and gold, 
rolled together to thin foil and well polished on each side, were punched into disks 
and mounted on the fly, the gold facing one way and the iron the other. After 
exhaustion the iron was most repelled by a candle flame, rotation being produced. 
When the bulb was grasped in the warm hand or covered with a hot shade the gold 
was repelled, producing rotation, which reversed on cooling, 
253. Platinum and gold in thin foil, heated till the gold just fused to the surface of 
the platinum and then rolled flat, were tried in the same way. Exposed to the light 
of a candle the gold was slightly repelled, and with three candles continuous rotation 
was produced. Covered with a hot shade the platinum was slightly repelled. 
254. Zinc and copper were rolled together and tried in the same manner in the 
apparatus. One candle gave good rotation, the zinc being repelled. Heated with a 
hot glass shade the zinc was still repelled, rotation being produced, and the motion 
reversing on cooling. Behind water no movement was produced by the light of three 
candles. 
251. The following couple — 
Thallie oxide ....... 
Green platinum salt of Magnus . 
ANOMALIES EXHIBITED BY SELENIUM. 
255. The most sensitive couple experimented on is the chromic oxide and scarlet 
selenium (245), and accordingly I put these powders together in the form of a 
radiometer both for the sake of further experiments and for the convenience of 
retaining the actual apparatus in a permanent form for future illustration or research. 
Thin pith was used for the disks, and the powders were thickly painted on alternate 
sides, so that the same powder should always follow or lead, according to the direction 
of rotation. The radiometer was well exhausted and then sealed off. The experi- 
ments are sufficiently instructive to warrant my giving them somewhat in detail, as 
they illustrate in a striking manner the extremely complicated actions involved in 
these repulsions, and also serve as a warning against hasty generalisations. 
It should be remembered that the experiments collected together in this paper 
have extended over more than twelve months ; each experiment was fully described 
