276 
MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
apparatus was not very sensitive, and with so small an amplitude as 3 '5° much 
accuracy could not be expected. It was therefore put aside, and the experiments 
were not resumed for nearly four years. 
Apparatus No. 2 is the same in construction as No. 1 (fig. 5), made, however, more 
delicately, suspended by a finer torsion fibre, and sealed on to a pump with which I 
can easily bring the exhaustion to the most sensitive point. Instead of a hot platinum 
spiral, a standard candle is used. The polariser is the same large plate of real tour- 
maline used in the previous apparatus. With this, a large series of observations were 
made ; the means are given below, arranged in groups : — 
Artificial and natural 
tourmalines. 
Parallel. Crossed. 
Group 1 
24-8 
25-2 
„ 2 
13-5 
19-0 
„ 3 
14'0 
16-0 
„ 4 
18-5 
18-5 
„ 5 
17-0 
21-5 
„ 6 .... . 
18-0 
18-0 
„ 7 
16-0 
16-0 
„ 8 
16-0 
18-0 
„ 9 
17-0 
17-0 
Mean . 
172 
18-8 
260. So far the results show a very slight difference in 
favour of the crossed tour- 
malines. The artificial tourmaline being 
very small, and 
. exposing some bare glass 
round its edge, it was removed from the 
apparatus, and 
a thin slice from a natural 
crystal of tourmaline was suspended in 
its place. To avoid the inconvenience of 
electrification, I adopted a device suggested by Mr. C. F. Varley, which answered 
perfectly — viz., binding the crystal very loosely with the finest platinum wire, so as to 
form a gridiron-shaped grating about half a 
, millimetre from its surface, the wires being 
about a millimetre apart. As polariser, a 
large Nicol’s prism was used, for the loan 
of which I am indebted to Mr. Spottiswoode. 
With this arrangement a large number of observations were taken, with the following 
results : — 
Nicol’s prism and natural 
tourmalii 
ie. 
Parallel. 
— — \ 
Crossed. 
Group 1 
18°0 
16 - 5 
2 
20-5 
16-0 
„ 3 
20-5 
17- 
„ 4 
21- 
17- 
„ 5 .... , 
2D 
16-5 
„ 6 
21-5 
17- 
Mean . 
20-4 
16-7 
