308 Sir Howard Grubb [May 25, 



necessary, and the telescope will follow any star through the whole 

 of its path in the heavens so long as it is above our horizon. 



But no matter how perfect the instrument may be in all the 

 details above spoken of, it is not possible to attain the necessary 

 perfection of motion without a good clock, and I thought it would be 

 interesting for you to see the working of such a clock, and have here 

 one which is identical with those used in the standard instruments 

 of the International Photographic Survey. This clock is the com- 

 bination of a good frictional governor, supplemented by a system of 

 control from an independent pendulum. Perhaps you will allow me 

 to explain why this control is necessary. A clock such as this will 

 go well and smoothly and keep good time from second to second, but 

 no uniform motion clock that I have ever met with can be depended 

 on for long periods. This one, I find, can be depended on to about 

 1 second in 600, but as it is necessary, or at least desirable, to be 

 able to depend on the clock for longer periods than this, while no 

 error of more than ^th part of a second can be permitted, it is evi- 

 dently necessary to supplement this by control from an inde- 

 pendent pendulum which can bo relied upon to the required amount 

 of accuracy. 



There is another very important reason why an independent 

 control is necessary. When an error occurs in the clock driving, 

 owing perhaps to some morsel of adventitious matter in the bearings 

 of the polar axis, or some little extra stiffness due to want of perfect 

 balance, &c, the tendency of all these governors is to bring the rate 

 of the clock which has been disturbed back again to the normal. 

 This answers perfectly well for visual work because, if such an 

 accidental error does take place occasionally, it merely means that 

 the star slightly shifts in the field and the wires can be again brought 

 up to the star and a satisfactory measurement taken, provided that 

 the image does not again shift during the few seconds required for 

 taking the observation. 



But in the case of the photographic telescope such an error would 

 be fatal, because the star has already impressed the photo plate at 

 one certain point. AVhen the error occurs the image shifts, and even 

 if the rate of the clock continues perfectly right for the whole of the 

 remainder of the exposure, the result will be of course a double or 

 distorted image. 



In photographic work w T e require some arrangement by which 

 any error which is accidentally introduced will be effectually and as 

 quickly as possible wiped out, the star image brought back again to 

 its original position on the plate and then the clock to resume its 

 normal rate, and this is a condition which no uncontrolled clock can 

 fulfil. The only solution which has yet been suggested to fulfil these 

 conditions is to have some means by which the clock of the equatorial 

 (which, as I said, goes well and smoothly for short intervals) is checked 

 and controlled every second from an independent pendulum. 



