348 
ME. W. CEOOKES ON EEPULSION EESULTING EEOM EADIATION. 
obedience of the instrument to the opposing forces, according as one or the other was in 
excess, was very striking. I may mention that only some of my radiometers act in this 
manner. It seems to require extreme lightness and great perfection of vacuum. 
The movable parts of the radiometer which shows this action best only weigh 
055 grain. 
163. These experiments had all been tried with surfaces made of pith, a very bad 
conductor of heat. It became of interest to ascertain what would be the action of 
a radiometer the fly of which was made of a good conductor, such as a metal. 
Experiments already recorded show that metals behave generally like pith. This has 
been proved in the case of magnesium (99, 100), aluminium (122), silver and bismuth 
(63), copper (64), brass (37-40, 61), and platinum (55, 62, 113, 114, 115); but none of 
these experiments have been tried under the different conditions to which I have lately 
submitted the radiometers. 
164. I selected thin rolled brass as the material wherewith to make the fly of 
a radiometer. The parts were all fastened together with hard solder, and no 
cement or organic matter was used, so that if necessary the radiometer could be sub- 
mitted to a high temperature without injury. In general appearance when finished 
it resembles the instrument shown in fig. 6. The moving portion weighed 13T grains. 
One side of the disks was silvered and polished, the other side being coated with lamp- 
black. The apparatus was exhausted with a charcoal reservoir attached. When 
exhausted it proved to be very sensitive, considering its weight, a candle 1^ inch from 
the bulb causing it to revolve about once a second, the black surface being repelled in 
the normal manner. 
165. The apparatus, standing motionless in a rather dark cold room, was covered 
with a warm glass shade. It immediately commenced to revolve the negative way, viz. 
silver side repelled, but very slowly. 
A few drops of ether poured on the bulb caused the arms to move rather rapidly the 
normal way. A hot shade put over whilst it was thus moving 
caused it to stop, and then begin moving the negative way. 
A small non-luminous gas-flame was held vertically beneath 
the apparatus, so that hot air should ascend and wrap round the 
bulb on all sides. The arms now revolved the negative way. 
166. The brass radiometer being somewhat heavy, one was 
made of aluminium, of the shape represented in fig. 9. The sur- 
faces were made large, and the whole moving parts were hard- 
soldered together. A siphon-gauge was attached, and the apparatus 
was connected direct on to the pump by a spiral, no charcoal-tube 
being used. One side of the wings was bright aluminium and the 
other was lampblacked. When exhausted the fly revolved 
very quickly to a candle a few inches off, the black being 
repelled. 
Pig. 9. 
