2 26 SMITHSONIAN MISCELLANEOUS COLLECTIONS [vol. 45 



he has no illumination of that field sufficient to discern the wires 

 of the tally. Suddenly a flash illuminates the field, and we may 

 suppose that by accident this occurs just at the instant when the 

 star image is bisected by the central wire. If now the flash had 

 recorded itself automatically on the chronograph, an observation 

 independent of the ordinary personal equation would have been 

 obtained, since we know where the star was at that recorded time 

 of observation. 



The device I have already described consisted in making this 

 unlikely accident occur for any star, and for numerous wires of 

 the tally. 



At the time when it was first devised, such a method consisting 

 not in correcting time observations for personal error committed, 

 but in preventing the committal of such errors, was unique, and it 

 had for the writer the good fortune of attracting the interest and 

 commendation of Professor Clerk Maxwell. But the writer was 

 prevented by distinct duties from developing it, and more recently 

 Repsold and others have attacked the same problem in another way, 

 with much success. Still it is possible that this older method may 

 prove of value, and I have thought best to give this brief account 

 of some recent trials of it. 



The instrument whose purpose is to illuminate the field at the 

 required instants, was constructed from my design by A. Hilger, 

 as follows : A conical pendulum consisting of a graduated rod with 

 a heavy ball which may be set at any height on the rod, is suspended 

 by two pairs of thin flat springs acting as a universal joint. At 

 the lower end of the rod is a needle which, as the pendulum revolves, 

 governs the rotation of a grooved arm carried by a clock-work 

 which also drives a drum carrying fine platinum brushes which make 

 instantaneous electrical contacts at each revolution. (Plate lvii.) 

 Thus the intervals between these contacts are governed by the rate of 

 the conical pendulum, which itself goes fast or slow according as the 

 ball is raised or lowered on the rod, this rod being graduated so 

 as to correspond with the declination of any star between 0° and 

 60°. Knowing the declination of the star, it can be arranged 

 beforehand, therefore, that the interval between contacts is equal 

 to the average interval between transits of the observed star across 

 the successive wires of the tally. By appropriate electrical connec- 

 tions this system of contacts is caused to produce a system of 

 instantaneous discharges in a vacuum tube, which is used in place 

 of a lamp to illuminate the cross wires of the telescope, and if the 

 star be first seen, for instance, half way between the first and second 



