1891.] Luminous Discharge of Electricity through a Gas. 87 



Fro. 2. 



Plan of Revolving Mirror. Upper halves of Bearings, A, carrying Oil Cup 



removed. 



touch two set screws, S, S, whose ends are also rounded, hardened, and 

 polished. Directly beneath the end of each axle is a cavity, C, which 

 serves to hold the oil running down from oil cups placed immediately 

 above. This arrangement was found to lubricate so well that there 

 was no appreciable heating even after long runs. The spindle is 

 made from square steel, the ends being turned down, and against two 

 opposite sides of the square centre portion, clutches E, E, for holding 

 the mirror M, are fastened. The whole is accurately balanced. The 

 bearings are attached to a heavy iron casting, which is firmly bolted 

 down to a heavy piece of masonry. 



In order to get a sufficiently rapid rotation of the revolving mirror 

 the Gramme had to be geared up. This was done by means of pulleys 

 mounted on ball bearings. 



A great many arrangements were tried in order to break the 

 primary circuit of the coil, when the mirror was in such a position 

 that the images of the luminous part of the tube would be reflected by 

 it into the telescope; after a great deal of time had been spent over 

 these they were all given up. The reason why they will not work is 

 pretty clear. The coil will not work when the primary circuit is 

 broken anything like so often as 500 times a second, so that if the 

 primary is to be broken by the mirror there must be very considerable 

 gearing down between the mirror and the break, in other words, the 

 mirror can be moved through a very considerable angle without 

 moving the break through more than a very small distance, but almost 

 the smallest possible movement of the mirror is sufficient to send the 

 images out of the field of view ; and it was found impossible to 

 diminish the play between the mirror and the break to such an extent 

 as to ensure that at a high rate of rotation the break took place syn- 

 chronously with the requisite position of the mirror. The method 

 finally adopted was the primitive one of using an independent mercury 

 bivak driven by a small Thirlmere water motor, and patiently looking 

 through the telescope until the break happened to occur just at the 



^ht moment. This, though a somewhat lengthy proceeding, was 

 not found in practice to be any longer than when synchronism be- 



