248 MR. W. CROOKES ON REPULSIOtf RESULTING FROM RADIATION 
The standard candle being lighted and placed 500 millims. from No. 1 disk 
(standard lampblack), exhaustion by the pump is carried on till such a point of 
sensitiveness is reached that on removing the screen the standard disk is repelled 
so that the luminous index moves about 150 divisions on the scale. Under ordinary 
circumstances the standard candle is kept 500 millims. from the disk, but, when its 
force is enfeebled by interposed screens, the distance is diminished so as to increase 
the amplitude of swing of the luminous index. 
223. It is necessary to carefully exclude aqueous vapour from the interior of 
the torsion apparatus, otherwise the sensitiveness will be diminished (105, 130). 
In all the series of experiments with different disks, the exhaustion, too, must be 
as nearly as possible uniform, otherwise the proportion between the standard disk 
and the others will not be identical. With a very good vacuum, when, the apparatus 
is most sensitive, the amount of movement impressed on the experimental disks 
diminishes in greater proportion than that shown by the black disk ; in other 
words, as the vacuum becomes very good, the sensitiveness of the black surface 
increases at a greater rate than does that of most of the other surfaces. In 
illustration I may take an experiment in which a pith disk coated with precipitated 
oxide of zinc was compared with the standard black disk. The candle being 
900 millims. from the disks,- the experiment was tried as soon as the exhaustion 
was good enough to cause a fair movement of the index ray. The ratio between the 
black and the white disk was as 100 to 5 5 ’5. On continuing the exhaustion for some 
time and then repeating the experiment, the ratio between the black and the white 
became as 100 to 42 ’5, and when the exhaustion was at the usual height at which 
the experiments were tried the ratio was as 100 to 35. In all cases, the actual 
amount of repulsion on the disks was greater at the higher than at the lower 
exhaustions. When sufficient air is present in the apparatus to visibly depress the 
gauge, the repulsion on the black and white surfaces tends to get still more nearly 
equal. These results are curiously analogous to those described in my third paper on 
this subject,* pars. 128, 155, 170, 171, where it is shown that the less refrangible rays 
of the spectrum cause black and white surfaces to be repelled to almost the same 
extent ; the ratio between the black and the white increasing as the incident rays 
increase in refrangibility. In these instances, therefore, low refrangibility and low 
exhaustion produce similar results. The experiment described in par. 130 proves 
that at a high exhaustion the presence of residual aqueous vapour has the same effect 
in equalising the repulsive force of luminous radiation on black and white surfaces as 
is produced in dry air at a somewhat lower degree of exhaustion. 
224. In the following Tables will be found the mean results of experiments with 
various powders laid on the surface of mica or pith disks, f The deflection of the 
* Phil. Trans., Yol. 166, pt. 2, January 5, 1876. 
f The action was about 50 per cent, stronger with mica disks than with pith disks (see Table XII., 
par. 237). But when reduced to the usual standard of lampblack =100, the differences ceased to be 
greater than might be due to experimental errors. 
