the Mtn/netic Effect of Electric Convection. 4G1 



conductor, beinii" independent of the value of the electrostatic 

 field produced l)y this voltage. For mica this critical value 

 is in the neighbourhood of J:300 volts ; for sulphur between 

 ISOOO and 12,000 volts; for ebonite, glass, and paraffin the 

 critical voltage is above 1200() volts. 



In the negative convection experiments the action of the 

 charge on the metallic surfaces was rendered nil by the 

 screening effect of the dielectric, or possibly by the neutrali- 

 zino- effect of the verv considerable charo-es absorbed bv the 

 dielectric. But this is only a supposition, — only a detailed 

 svstematic study of these dielectric effects can clear u[) this 

 very delicate point. 



Study of the Amount of Cliarge carried round under 

 carious Conditions in the '' Open Current "" Experiments. — 

 To complete our researches we undertook a study of the 

 apparent amount of charge carried round in the open- 

 current experiments, with the object, first, to make sure that 

 the third fundamental condition laid down at the beginning 

 of this article was realized, and, secondly, to clear up a little 

 more the role of the dielectric. 



The apparatus employed (fig. 5, p. 458) was the disk with an 

 ebonite core, OE, carrying 18 peripheral sectors, M. These 

 passed between the fixed sectors S and at the same time 

 touched the brush A : then, according to the direction of the 

 rotation, they either became directly discharged hy the 

 brush 0, or first passed between the supplementary sectors Si 

 and then were discharged at C. The sectors Si formed with 

 each sector M a capacity equal to that formed by the sectors S 

 with M. These sectors Si, connected to an electrometer, 

 permitted us to see what takes place, of an electrostatic 

 nature, in the air around the moving charged sectors. An 

 experiment consisted in measuring gaivanometrically the 

 quantity of charge taken up by the sectors as they came 

 under A, and the quantit}- of charge given up at C ; also the 

 reading of the electrometer connected to Si was noted. The 

 results were as follows : — 



1. Mocing Sectors M hare; Fij-ed Sectors S hare. — 

 (a) Whatever was the sign of the charge, the discharge- 

 current was almost invariably superior to the charging- 

 current^ the excess being variable, but in some cases as high 

 as 15 per cent. This dissymmetry was the greater when the 

 moving sectors were positive. 



[h) On the sectors Si the following peculiarities were 

 observed : — Suppose the disk in rotation in the direction of 

 the arrow 2, the sectors Si connected to the electrometer, 

 and all the rest of the apparatus to earth. No matter what 

 was the velocity of rotation, no deflexion of the electrometer 



