Oct. 31, 1889] 



NATURE 



047 



B or F>', and transmitting a current which brings into 

 action the accelerator or retarder as may be required. 

 The period during which the accelerator or retarder re- 

 mains in action will depend upon the amount of the error 

 to be corrected, and the proportions of the pairs of wheels, 

 I, 2, and 3, 4. With the proportions described above, the 

 correction introduced is one-thirtieth of the rate, so that, 

 to correct an error of one-fifth of a second, the accelerator 

 or retarder, as the case may be, would have to remain in 

 operation V = 6 seconds. As soon as the correction has 

 been made, the lever. A, will resume its normal position, 

 and the bridge, B, coming then between the two platinum 

 plates, B b', a current will cease to be transmitted, and 

 the accelerator or retarder thrown out of action. 



It is to be noted that the apparatus above described 

 not only corrects any temporary disturbance of the equa- 

 torial clock rate, but cancels errors which have already 

 occurred. 



It will be seen that the third form of control is free 

 from the objections of the first and second. The detector 

 part of the apparatus is close to the screw spindle, only 

 removed from it by one pair of wheels, and the correction 



is not applied in the same manner by checking the speed 

 of the clock, but by introducing a differential gear, which 

 acts until the error be cured, and then drops out of gear 

 automatically. 



The fourth and last form of control, however, is that to 

 which I would invite your special attention, for I believe 

 it to be capable of results beyond all the others. 



I have endeavoured in it to select all the good points 

 of the other forms, and to combat the weak points. I 

 may not have as yet produced it in as perfect a form as 

 is possible, but I am satisfied it is capable of development 

 into a very perfect control, and even at present it is the 

 most perfect I have constructed. 



As long as the control applied its correction by altering 

 the speed of the governor, it was necessary to keep down 

 the vis inertia of the governors, but now, as the correction 

 is not applied in this way, I have made the governor very 

 heavy, and running at a very high speed. 



The vis inertia of the governor is represented by some 

 10,000 foot-pounds per minute ; consequently it is little 

 affected by any small or short differences in friction or 

 driving powers. 



Fig. 6. 



Again, at the suggestion of Dr. Gill, I make the governor 

 spindle gear directly into the counter spindle of the screw, 

 in order to have as few wheels and pinions to deal with 

 as possible. Errors of wheels behind the governors have 

 nothing whatever to do with the accuracy of its rate. 



As to the nature of the control, I use Dr. Gill's form of 

 detector and my own form of corrector, viz. accelerator 

 and retarder. I use Dr. Gill's detector because it seems 

 to be capable of being made on a larger scale than mine, 

 and consequently ought to be more delicate. 



I now propose to say a very few words on the optical 

 part of the instrument. 



The first question that naturally occurs is whether a 

 refractor or reflector should be used. I own that when I 

 first considered the subject I incUned to the belief that 

 reflectors would be found to be the most suitable ; and in 

 a paper of mine, which was read before the Royal Astro- 

 nomical Society last spring, I urged that comparative 

 trials should be made before a final decision was arrived 

 at. I have found reason, however, to modify my views 

 on this point. 



My reason for thinking that the reflector might possibly 



prove the best was founded on the consideration that in 

 reflecting instruments rays of all refrangibilities are brought 

 to a focus at one and the same point, whereas in the re- 

 fractor rays of various refrangibihty have different foci, and 

 the best we can do is to so arrange the curves that those 

 rays most active in impressing the photographic plate 

 may be brought as nearly as possible to the one focus. 



If we draw a curve which represents the position of the 

 focal point for various rays of the spectrum in an object- 

 glass corrected for photo work, it will be something like 

 this figure (Fig. 8). The same for a reflector will be repre- 

 sented by a straight line. Looking at these curves it is 

 certainly a natural conclusion that the reflector ought to 

 be best, and therefore it was that I urged that a fair com- 

 parative trial should be made between the reflector and 

 refractor as to their suitability for this work. 



The Congress, however, decided upon the use of re- 

 fractors, from the simple fact (as Dr. Gill says) that the 

 best work done (to that time) had been done by refractors, 

 not taking into consideration the very much more favour- 

 able conditions under which the refractor photographs 

 were taken. Since that time further experiments have 



