186 Mr. H. Pender on the Magnetic 



formed by tinfoil on ebonite were each divided in half. The 

 lower half of each was put up close to the disk (how close 

 Cremieu does not state) ; the upper halves were 10 mm. distant. 

 The foil covering a sector of 60° on each of the upper halves was 

 removed. The needle, inclosed in a copper tube with walls 

 *4 mm. thick connected to the earth, was placed directly over 

 the edge of the disk opposite this break in the condensing- 

 plates. Each sector of the disk was charged when immedi- 

 ately between the lower halves of the eondensing-plates. 

 When it came under the needle it was completely insulated 

 from the other two and from the charging apparatus. With 

 this apparatus Cremieu again found that no deflexion of the 

 needle was produced on charging or discharging the rotating 

 disks, so long as the needle was properly screened against 

 direct electrostatic action. If, however, the copper tube 

 above mentioned was replaced by a tube of some less perfect 

 conductor, e, g. graphite-paper, a deflexion could be obtained, 

 which had all the characteristics expected of an effect due 

 to electric convection. But this effect could be obtained 

 equally well when the magnets forming the astatic needle 

 were removed, and only the mica vaues on which they were 

 mounted were left. Here are Cremieu's final conclusions : — 



" Enfin, toutes les fois que le systeme suspendu, quel qu'il 

 soit, est protege par un ecran electrique tres conducteur, on 

 ne peut obtenir aucune deviation. 



" De plus, il arrive souvent qu'au moment ou Von change 

 le signe de la charge du di^que, une etincelle, ou meme une 

 sorte de pinceau de decharge a peine lumineux, se produit 

 entre le disque et les platenx fixes. 



" Ceci cause, sur Faiguille aimantee 5 des impulsions rever- 

 sibles a la fois avee le signe de la charge et le sens de la 

 rotation ; le peu de stabilite du zero de systemes aussi sensibles 

 fait qu'il peut resulter de ces impulsions des deviations 

 permanentes faibles/' 



According to Cremieu's explanations of these slight de- 

 flexions there is no reason why they should reverse with the 

 direction of rotation of the disk, In all probability, however, 

 these deflexions w r ere really caused by electric convection, 

 and should have reversed as observed. Though in the short 

 note which Cremieu has published concerning this experiment 

 there is no mention made of a caoutchouc coating over the 

 disk, the disk was undoubtedly thus covered, as in all the 

 previous experiments. The sparks observed were then due 

 to the passage of the charge from the eondensing-plates to 

 the caoutchouc surface. A slight charge might still have 



