312 
MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
black mica. Instead, therefore, of the molecular pressure starting from the wire, 
as in previous experiments, the blacked mica now becomes the driving surface. 
361. The whole of this complicated arrangement of apparatus is connected together 
by actual fusion of the glass tubes one to another ; no joint whatever occurs in any 
part, and a certain point of exhaustion being once attained, I can leave the apparatus 
to itself with the certainty that no leakage from without can occur. The apparatus 
by which the gas is introduced has, it is true, a tap, but when it has effectually done 
its work this tap is disconnected from the rest of the apparatus by fusion. 
362. An observation with this apparatus is taken in the following way. Having 
arrived at a point when a depression of the contact key tells me, by the behaviour of 
the rotating disks, that a useful observation can be taken, the pressure is first 
measured in the McLeod apparatus. The viscosity of the gas is then observed, and 
next the repulsion exerted on the viscosity-plate by the candle. At a very high 
exhaustion the appearance of the induction spark in the tube, h, is also noted, 
together with the spectrum given by it. The strength of the current being first 
regulated by the resistances, v, the key, w, is pressed down, and the direction and 
speed of the vanes and disk in s are taken by a chronograph recording to tenths of a 
second. Frequently duplicate or triplicate observations are taken at each pressure, 
time being allowed to elapse between the observations for the apparatus to become cool. 
363. In describing the direction of rotation of the vanes and disk, I shall call the 
direction they take at high exhaustions the positive direction, and the contrary the 
negative ; thus, the positive rotation of the disk and vanes will be as follows : — 
364. At a pressure of 761 millims., when the wire is heated, positive rotation takes 
place both of the disk and vanes. As the exhaustion increases the positive rotation 
of the disk diminishes, then stops, and at 100 millims. pressure the disk commences to 
rotate negatively, or the same way as the vanes go, thus : . The viscosity of the 
air is 0T24. 
The probable explanation of these actions is as follows : — In air at the normal 
density the red hot ring causes air currents to rise ; these strike the vanes and cause 
them to rotate positively. The sloping vanes also cause a deflection in the ascending 
current of hot air, and the disk, therefore, is struck at an angle by the air currents, 
causing it to rotate the opposite way to the vanes. The diagram (fig. 33) represents 
Fig. 33. 
this action, a, a, is the hot wire ring, the small arrows show the direction of the 
rising currents of hot air, which, striking against the vanes, are deflected in such a 
way as to cause the disk to rotate. The direction of rotation is shown by arrows. 
