184 ROYAL SOCIETY OF CANADA 



the focus point to the nearer wheel was 50 cm., and tli<> diani. of the 

 focus point was about 1-6 mm., being as nearly as could be judged the 

 same as the width of one of the slots in the M^heels. Under these con- 

 ditions, it was ])lain that the tippearance presented on the photographic 

 plate, if the wheels wei'e slowl}^ rotated, should be that of an umbra 

 l-(j mm. in width, fringed on either side by a penumbra of the same 

 extent. More exactly, if b is the width of a slot, and c the diam. of the 

 focus point, the width of the umbra should be 2b — c, and the total width 

 of the band including penumbra 26 -fc. This, in fact, proved to be the 

 ease. The etfect of aberration should have been to shift the bands in the 

 direction of rotation by a displacement equal to the distance turned 

 through by the wheels during the time taken by the rays to traverse the 

 interval between them. By means of a horizontal slot a httle less than 

 a quarter of an inch wnde in a brass plate, it was possible to expose one 

 half of the plate while the wheels were turning very slowly, and the 

 other half at the highest speed. Exposures were made for one minute 

 intervals alternately on the two halves in order to eliminate the etfect of 

 any possible change in the discharge, or in the relative })Ositions of the 

 focus tube and box. In general, each half was thus exposed for five 

 minutes. No displacement could in any case be detected, using a circum- 

 ferential velocity of 25 metres per second. 



(2) JJethod of Total Eclipse. — For the application of this method, 

 a brass tube was fitted along the axis of observation between the two 

 wheels. The ends of this tube were closed by discs having slots cut in 

 them of the same width and size as those in the edges of the wheels. 

 The slots in the ends of the tube were set very close to tho.se on the 

 wheels, and accurately parallel to them. The end of the tube nearest to 

 the photoi^raphic plate, was provided with a screw adjustment, by which 

 it could be shifted in a direction at right angles to the slots, while at the 

 aame time the parallelism of the slots was maintained as accurately as 

 po.ssible. 



If tlic position of the tube was adjusted so that any ])art of the slot 

 in the end of the tube nearest the X-ray focus, was open at the moment 

 when a slot in the wheel coincided with the slot in the tube at the other 

 end, the image obtained on the i)hotographic plate was an exact outline 

 of the whole width of the slot in the end of the brass tube nearest to the 

 plate. By the aid of ordinary light it was very easy to make the adjust- 

 ment so that one slot just began to o])en at the moment when the other 

 closed. Uuder these conditions, the two slots were never open togethei-, 

 and the light was just totally eclipsed. A movement of a thousandth of 

 an inch in the screw adjustment, was sufiicient to restore a very a])preci- 

 uble amount of light. It was therefore very necessary that the slots in 

 the wheels should be cut to correspond as accurately as possible. For- 

 tunately this had been foreseen, and the cutting of the slots was found 

 to be sufficiently exact, when tested in this manner. 



