340 
ME. W. CEOOKES ON EEPULSION RESULTING EEOM RADIATION. 
coal-tube still attached to it. Ultimately, when all the residual gas has been absorbed 
by the charcoal, a flame is applied to the contracted part of the tube at h , and the 
charcoal-tube is disconnected. 
146. Before adopting the above method of making these instruments many experi- 
ments were tried, both to secure ease of manipulation and greater delicacy of action. 
The cup supporting the needle-point was made of ruby, sapphire, chrysolite, aquamarine, 
and agate ; it was, however, found that these offered no advantage over glass, as the 
friction was not sufficient to produce any abrasion of the glass by the steel point. The 
disks at the end of the arms were made of every imaginable substance which was likely 
to answer. Among these I may mention wood, paper, flies’ and butterflies’ wings, talc, 
mica, selenite, thin glass, metals of various kinds, ivory, cork, and pith (86). For general 
purposes I prefer pith, as it is easily cut into slices, is extremely light, dries readily in 
a vacuum, and does not evolve vapour subsequently ; besides which its natural white 
surface is almost as insensitive to radiation as any substance I have yet examined. 
The number of disks has been varied from ten, the maximum which can follow one 
another without available surface being uselessly obscured, to two, or even one, which 
latter form has been experimented with, and possesses some advantages. Six disks are 
a useful number ; but as the difficulty of making these instruments increases with the 
number of arms and disks to be got into the bulb, I prefer four disks for ordinary 
purposes. 
The material of which the arms are made has also been the subject of experiment. 
My earlier instruments (exhibited at the Soiree of the Royal Society on the 7th of April, 
1875) had straw arms. These are, however, too heavy, and are liable to evolve vapour 
after being kept in the vacuum for some time. An inorganic body is preferable ; and 
I have finally adopted either thin rolled brass or fine glass thread drawn from thermo- 
meter tubing. 
The colour of the disks has also been experimented on. During this part of the 
inquiry many curious results have been obtained, which will be described farther on. 
At present, however, I have found nothing better than lampblack for the black surface 
and the freshly cut pith for the white surface. In working with cork, metals, &c., 
where the natural surface is not white enough, oxide of zinc may be used as a coating 
for the white surface. 
147. The lampblack- is best applied to the pith surface in the following way : — 
Camphor is burnt, and a sheet of glass is held close over the flame. An abundant 
deposit of lampblack takes place. A brush dipped in alcohol is then rubbed over 
the deposit, and the surface is painted over with the mixture. The lampblack 
adheres very w 7 ell to pith, and in a few hours the alcohol and moisture have dried off, 
and a dead black, very even surface is the result. In some cases I smoke this again 
over burning camphor ; but this is not of much use, unless the first coating shows glis- 
tening patches or is not laid on evenly. 
148. I have proposed for this instrument the name of the Radiometer , as it serves to 
