MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
.315 
375. Pressure 6 millionths of an atmosphere. — The speed of the vanes is far too 
great to count, and their shape is scarcely distinguishable owing to the velocity. The 
disk rotates steadily at a rate of 22 turns a minute Viscosity = '029. Candle 
repulsion = 16. 
3 76. Pressure 2 millionths of an atmosphere. — By reason of their great speed, the 
vanes are now invisible, except as a nebulous ring. The disk makes 30 revolutions a 
minute. Viscosity = '020. Candle repulsion = 10. 
377. Pressure 0'4 millionth of an atmosphere. — The vanes and disk rotate as at 
the last pressure. There is no apparent diminution in the speed of the vanes, and the 
disk is going at the rate of 31 revolutions a minute. Viscosity = '016. Candle 
repulsion = 6. 
378. I could not get a higher exhaustion in this experiment, so I filled the 
apparatus with pure hydrogen. This was effected, previous to the first experiment, 
by sealing on to the pump a tube, shown at I, containing palladium foil saturated 
electrolytically with hydrogen. This alloy is perfectly stable in the cold at the 
highest exhaustions (233), but when gently heated the hydrogen is given off in quan- 
tities which can be easily regulated by careful manipulation. The gas was allowed 
to depress the gauge about 50 millims. The pump was then worked till a high 
exhaustion was reached, and more hydrogen evolved. After two or three heatings 
and pumpings, the residual air was assumed to have been washed out, and the obser- 
vations in hydrogen were commenced. 
379. At a pressure of 5 millims., when the platinum ring is ignited, the vanes 
are still, but the disk rotates in the negative direction. At 1'25 millims. both vanes 
and disk are still. At '8 millim. the vanes keep stationary as before, but the disk 
assumes a positive rotation. 
At *6 millim. positive rotation of the vanes commences, and that of the disk con- 
tinues at a little greater speed than before CyA 
At 47 millionths of an atmosphere the positive rotation of the vanes is too rapid to 
count. That of the disk is 10 revolutions a minute. 
380. Pressure 8 millionths of an atmosphere. — When the current is first turned on, 
the vanes commence to rotate rapidly in the positive direction, and the disk revolves 
more slowly in the positive direction. The speed of the vanes gradually increases up 
to a maximum of at least 1000 turns a minute ; and during this increase of speed the 
disk revolves slower and slower, until, when the vanes rotate at the highest velocity, 
the disk is quite still. The reason of this peculiar behaviour is probably this : the 
enormous speed of the vanes, acting on the disk through the viscosity of the residual 
gas, causes a tendency to negative rotation — i.e., in the same direction as the vanes ; 
this tendency to negative rotation balances the rotation due to the tangential action 
of the molecular pressure, which, deflected from the sloping vanes, would turn the disk 
positively. 
