423 
The unknown effect to be explained has the following pro- 
perties : 
1. It is a surface-effect more deeply seated than the gas- 
effect. 
2. It varies in the same proportion along the radius as the 
gas-effect (and is therefore probably proportional to the distance 
from the centre). 
3. The quantity of heat produced in disks of different 
thicknesses appears the same, as the radiation was found to be 
proportional inversely to the thickness. 
4. The effect is produced without perceptible diminution | 
when the disk is covered with a chamois leather blind, 
5. It is independent of the residual gas. 
6. It is different for different disks. 
Now the disk when rapidly rotating expands slightly, and 
consequently becomes cooled below the surrounding space. 
Hence during the time of rotation an equilibrium of tempera- 
ture takes place and it becomes heated up towards its former 
temperature, when the rotation is stopped the disk shrinks to 
its former size and gives out the heat it had taken in whilst it 
was rotating. Now let us see how this explanation would 
satisfy the required conditions. 
1. It is clearly a surface-effect when the rotation has not 
been continued too long, for the heat enters from outside and 
gets conducted inwards, and therefore it will be hotter outside 
than inside. It will be also more deeply seated than the gas- 
effect, for it is heated by radiation which affects a perceptible 
depth, while the gas-effect takes place on the surface only. 
2. It can be shown that the displacement at any point of 
the disk is, when small, very nearly proportional to the distance 
from the centre, or £=pr. The strains are therefore oan 
and é = » constant along and perpendicular to the radius. The 
33—2 
