THE STRUCTUKE OF THE NUCLEUS. 1G9 



make the next step in the direction of greater sensitiveness, by decreasing the dis- 

 tance apart of the upper and lower sectors of the quadrants. In the new instrument 

 the same light needle was used as above, and the new quadrants were about 10 

 cm. in diameter and 1 cm. high. Though the same capsule was used the damping 

 coefficient was increased by the less roomy quadi'ants. The needle was quite 

 aperiodic, but its period about 60 seconds. 



Again omitting the tabulated results for brevity, let me say that the experi- 

 ments were made in three sections. In the first the quadrants were chai'ged by a 

 water battery (nominally 48 volts) ; a diversion, however, was introduced by earth- 

 ing the middle of the battery. Supposing the two halves equal (^V^ = — V^), 

 = « ( Fg - Fi)( Fa - ( F, - Fi/2) = 2« Fg F, where Fg is 1/2 of the electro- 

 motive force of the batteiy. In the first part of the work I found a = .39. 



In the second part of the work the reason for this low result was investigated. 

 It appears that as the voltage of the quadrants increases from 10, 20 to 40 volts, 

 the deflections pass through a maximum, being actually greater at 20 volts than at 

 40. The explanation is suggested assuming that the two halves of the storage 

 battery are not equal, so that Fg + F, is not zero. As the balance was most 

 nearly maintained for 10 cells the constant becomes a = .84, an increase of nearly 

 24 times. . 



In the third part of the woi'k one quadrant was earthed, and the othei' chai'ged 

 with the full number of storage cells. The results were in general corroborative; 

 but here also the sensitiveness failed to increase with the potential difference of the 

 quadrants. 



A few idiostatic results were finally tried, showing that on the average a = .74. 



PRELIMINARY RESULTS WITH A HIGHLY CHARGED NEEDLE. 



11. Charge ly indaction. — I now proceeded to take advantage of the ob- 

 servation just made. Accordingly the quadrants were used for measuring the 

 electromotive foi-ce of two Clark cells while the needle was permanently 

 chai'ged by a storage battei'y. My first experiments consisted in seeing in how far 

 an inducing system would suffice. The capsule was charged directly and the 

 needle by induction. The results showed that the sensitiveness soon passes 

 through a maximum, more than ten cells connected with the capsule are scarcely 

 useful. The sensitiveness was low and variable. 



12. Charge by cotuluctio7i.— The needle therefore is to be charged by contact. 

 The only loss of charge can then take place through the silk threads or quartz 

 fiber. This method was very successful so far as small equipment with great 

 sensitiveness is concerned. Naturally the needle loses charge in the lapse of time, 

 which, however, can be more than supplied by ionijcing the air in the capsule. 



The results showed enormously increased sensitiveness. With but 10 storage 

 cells to the needle and capsule the deflection was 50 cm. per volt for a scale distance 

 of 4 meters. Material advantage is secured by charging the capsule. In makuig 

 these long excursions the period of the needle is increased to almost 2 minutes, 

 which is to be improved by cutting down the damping. 



