306 The Motion of Radium in the Electric Freld. 
charge. Breaking the connexion restores the repulsive effect. 
If the spheres are connected with each other and the vanes 
are also connected one with another, then when a charge is. 
given to the spheres there is repulsion. 
Now evidently this behaviour is in favour of assuming that 
the repulsive effect is due to the preponderating influence of 
a charge given to the vane by induction, for such a charge 
would, when the vanes are connected, be repelled to the 
remote end of the beam. When the charged body is pre- 
sented at both ends of the beam and the vanes are connected, 
the repulsive charge must remain in the vanes, and ac- 
cordingly there is repulsion. 
Rather unintentionally a further experiment was made 
supporting these results. Vanes of very thin mica had been 
mounted radiometer-fashion at theextremities of an aluminium 
wire beam. The wire was stitched through the mica, and 
was therefore in good contact with both sides of the vane. 
When those vanes were coated on alternate faces with radium 
bromide the mill refused to rotate decisively in any one 
direction, but showed, if anything, a tendency to rotate in 
the “‘ attractive” direction. I at first ascribed this to the use 
of vanes too thin in substance, but subsequently, in accord- 
ance with the view that electrical connexicn between the 
vanes must destroy the rotation, I substituted a glass fibre for 
the wire. The mill now rotated freely and in the “ repulsive ” 
direction in the electric field. The experiment suggests that 
the mica itself must have acted asa conductor under the 
influence of the radium. 
On the question as to how far it is an induced or a con- 
ducted charge which leads to the repulsion of the radium- 
coated vane when the radium is presented towards the charged 
body, it is to be observed that not only some of the preceding 
experiments but also the rapidity with which the repulsion 
succeeds the charging of the sphere are in favour of ascribing 
the effect to induction. Quantitative measurements would be 
required to fully elucidate the matter. Possibly there is some 
direct ionic conduction of the charge as well as the inductive 
effect, and, of course, the experiments do not exclude the 
existence also of some mechanical effect from the transport of 
electrified particles, but they serve to show that the latter 
effect, must, at best, constitute but a small part of the force 
acting upon the vanes. 
Another source of rotation is to be sought for in the 
reactionary force attending the discharge of @-radiation. It 
is improbable that any intensification of this discharge in the 
electric field, sufficient to give rise to the effects observed, 

