MR, W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
311 
when subjected to high rarefaction is still in progress, and owing to the great 
expenditure of time required in getting accurate results it may be some time before 
the research is finished. I may, however, refer to the preliminary notice on this sub- 
ject, given in the ‘Proceedings of the Royal Society,’ November 16, 1876, for an 
account of some of the viscosity results obtained at that time, with a diagram of the 
curves of diminution of the viscosity of ah-, oxygen, and hydrogen, as the rarefaction 
proceeded. 
357. I have said that one-half of the mica plate swinging in the bulb of the 
viscosity apparatus is lampblacked. This lampblacked half is opposite a tube, p, 
furnished with a shutter, by means of which the light of a candle, q, can be thrown 
on the blacked plate. The repulsion thus produced is measured by the successive 
swings and final deflection of the index ray on the scale. 
358. When other gases than air are experimented on they are introduced into the 
apparatus through the tube, r, which is connected with gas reservoirs and appropriate 
apparatus for purifying and drying the gas. 
359. The apparatus, s, more especially under examination, is sealed to the tube, j. 
It is shown on a larger scale at C, D. The platinum ring in s is ignited by the 
battery t. As the power of this battery varies considerably during a series of 
experiments, to render it practically constant I adopted the following plan. One 
pole of the battery is connected with the wire, u, running direct to s ; the other pole 
is connected with a box of resistance coils, v, through which the current passes ; it then 
goes to the contact key, w, and thence to the other wire, x, up to s. By depressing 
the contact key, w, the current passes through the wire ring in s, and ignites it. 
The strength of the current passing through the ring can be regulated to a nicety by 
adding or subtracting resistance by means of the coils at v. At y and z wires shunt 
off a portion of the battery current and conduct it to the galvanometer A and the 
resistance coils B. The galvanometer is a very delicate one, and the resistance in B 
is so adjusted that the current flowing through the galvanometer shall deflect the 
needle 18°. This being adjusted once for all the resistance B is never altered. Any 
variation of battery power will now show itself on the galvanometer A, and the 
resistance of the coils, v, is immediately altered until the galvanometer needle is 
again brought to 18°. This method I have found sufficiently accurate for all 
purposes, and since its adoption I have been able to experiment day after day with 
the certainty of always having, practically, the same amount of battery power igniting 
the platinum ring. 
360. The apparatus, s, containing the rotating disk and vanes is shown enlarged at 
C, D. The construction of glass bulb, the rotating mica plate, E, the sloping vanes, F, 
and the platinum ring, G, connected with the outer wires, u, x, do not differ from 
those in the apparatus shown in fig. 31, paragraph 344. On the top of the platinum 
ring rests a disk of mica, H, lampblacked on the upper surface ; this cuts off direct 
radiation from the hot ring, and diffuses the heat somewhat over the surface of the 
