[cox .4 callendar] experiments on the x-rays 183 



appréciable vibration. A.ssuining tliat it would be ]»ossible to observe an 

 aberration dis])lacement of one-tit'th of a millini. on the circumference, 

 we might expect to obtain some etiect provided that the velocity did not 

 exceed 100 kilometres per second. 



Construction of the Apparatus. — The wheels were made of brass one- 

 sixteenth of an inch thick. The discs were flattenetl and turned true on 

 a suitable hub, and were then soldered together at the edges so that the 

 radial slots to be cut in the edges might exactly correspond. The slots 

 were each a sixteenth of an inch wide, and half an inch deep, and num- 

 bered one hundred. The metal left between the slots was nearly four 

 times the width of a slot. This proportion of slot to space was necessary 

 in order to secure a total eclipse, because the rays necessarily formed a 

 conical pencil. When the slots had been cut, the discs were separated, 

 and fixed on a brass tube axis, at a distance of a metre apart, with the 

 corresponding slots in each on a line pai-allel to the axis of. rotation. 

 This precaution was essential for the method which'we proposed to adopt, 

 because although the slots wei-e cut on a very good milling machine, it 

 is doubtful whether the accuracy of the division Avould have been suffi- 

 cient to make each pair of slots give exactly similar etfects unless they 

 had been simultaneously cut. 



The brass tube carrying the discs, wastitted with steel pivots turning 

 in suitable bearings in the end of a long Avooden box, which was covered 

 with tin plate and at the ends with thick sheet lead. The X-rays were 

 admitted at one end of the box through a small tube fitted with a lead 

 cap. After passing thi-ough a pair of corresponding slots in the two 

 wheels, they were observed by means of a small tluoroscope, or by means 

 of a small camera, each protected b}' a double thickness of aluminum 

 foil, at the other end of the box. 



The adjustment and setting of the apparatus in each case could be 

 very easily and exactly performed by the aid of common light. In this 

 manner we tested the exact correspondence of the slots, which was found 

 to be very satisfactory, and also the steadiness of the apparatus when 

 driven at a high speed. 



With this apparatus it was possible to use three ditferent but closely 

 related methods, for the attempt to measure the velocity. These methods 

 may be called (1) the juethod of xVberration, (2) the method of Total 

 Eclipse, and (8) the method of Partial Eclipse. The methods all gave the 

 same result, but of the three the third method appeared to be the most 

 satisfactory. 



(1) The Method of Aberration. — For the apjilication of this method, 

 the axis of observation w^as aligned by optical observation of the small 

 red-hot focus point on the platinum plate, in such a manner that the 

 focus point was just visible through a pair of corresponding slots when 

 the latter were in the centre of the field of view. The distance from 



