MR. W. CROOKES ON REPULSION RESULTING- FROM RADIATION. 
317 
ratus from the data in paragraphs 369 to 377. They are connected together by lines 
forming curves ; in the curve representing the “ candle repulsion ” I have interpolated 
a few observations from other experiments to fill up a gap between 5 9 -millionths and 
14-millionths, and to give a better idea of the direction the true curve would take. 
The “ candle repulsion ” and the “ viscosity ” curves are similar to those already 
published in the ‘ Proceeedings of the Royal Society.’* In describing these curves 
in November, 1876, I said that the viscosity of the residual gas in an air vacuum 
was practically constant up to an exhaustion of 250 millionths of an atmosphere, 
having only diminished from 0*126 at the normal pressure of the atmosphere to 
0*112. It now begins to fall off, and at 0*1 of a millionth of an atmosphere it 
has fallen to about 0*01. Simultaneously with this decrease in the viscosity, the 
force of repulsion exerted on a black surface by a standard light varies. It increases 
very slowly till the exhaustion has risen to about 70 millionths of an atmosphere ; 
at about 40 millionths the force is at its maximum, and it then sinks very rapidly, 
till at 0*1 millionth of an atmosphere it is less than one-tenth of its maximum. 
384. The diagram (fig. 34) entirely confirms the observations described last year. 
In order to keep the diagram within reasonable limits, I have omitted the obser- 
vations at lower exhaustions than 118 millionths of an atmosphere, but from that 
point upwards the parallelism is close. The candle repulsion rises to a maximum 
somewhere between 59 and 14 millionths of an atmosphere, and then rapidly sinks up 
to the highest exhaustion obtained. Simultaneously the viscosity drops rapidly at 
the high exhaustions. From these observations I might be justified in assuming that 
it would be a general law that above 40 millionths of an atmosphere the repulsion 
resulting from radiation would fall off as the exhaustion got nearer absolute, and this 
idea would be greatly confirmed by the experiment with the radiometer described in 
paragraph 334, and graphically illustrated in the curve on fig. 28. 
It now seems that this generalisation would be too hasty, and the utmost that can 
be said is that the statement is true for the particular instances described. When, 
instead of the feeble intensity of radiation which can penetrate glass from a candle 
some inches off, I substitute the intense energy of a red hot platinum wire a few 
millimetres off, the answer given to my interrogations is very different. There is 
now no maximum at 40 millionths, and subsequent rapid falling off, but a steady 
increase of speed from 67 revolutions a minute at 59 millionths, 150 revolutions 
at 14 millionths, 600 revolutions at 11 millionths, up to over 1000 revolutions at 
6 millionths, and still increasing speeds at 2 millionths and at 0*4 millionth. At an 
exhaustion where the repulsion set up by the candle is least, that caused by the hot 
wire is greatest. 
In air, at still higher exhaustions, I could detect no falling off of speed, but when 
the residual gas was hydrogen I thought that there was a diminution of velocity after 
1 millionth of an atmosphere had been reached. 
* No. 175, 1876, page 305, 
