444 Messrs. V. Cremieu and H. Pender on 



many modifications suggested by objections from numerous 

 sources. 



At the suggestion of Rowland, Pender took up the method 

 of Cremieu, improving it in some details. His experiments 

 were carried on first in the Physical Laboratory of the 

 Johns Hopkins University, and later in the country, far 

 from the disturbances of any industrial centre. The details 

 of these two series of experiments have been published 

 elsewhere *. The results were invariably positive ; and in 

 the second series the observed and calculated values 

 accorded within less than 5 per cent. 



Pender brought with him to Paris all his apparatus, 

 which we at once set up under conditions as nearly similar 

 as possible to the original disposition. The only condition 

 we were unable to realize was the same degree of mechanical 

 and magnetic stability. Nevertheless, we were able to find 

 again the qualitative effects of magnetic induction already 

 observed by Pender. We verified with a sufficient degree 

 of approximation the proportionality of these effects with 

 the speed of rotation of the disks and with the surface- 

 density of the charge carried round. We also found that 

 the effiicts of the two disks were superimposed or annulled 

 each the other in accordance with the theory of convection, 

 and that the effect obeyed the law of the distance between 

 the disks and induced coil. As to the order of magnitude 

 of the effects observed, the deflexions varied from 38 to 

 800 mm. on a scale placed 4 metres from the galvanometer. 



Repetition of Cremieu' s Experiment. — Cremieu at the same 

 time mounted his induction experiment, the conditions 

 being almost identical wdth those of his original experiment. 

 However, the galvanometer previously employed had been 

 destroyed ; so we were obliged to use another. i\lso the 

 conditions of stability were not so good as in the original 

 installation. 



Although the effects due to charging and discharging 

 the disk at rest were not entirely eliminated, we obtained 

 with a sufficient degree of certainty the same negative 

 results as in the original experiment. 



Verifications in Cremieu' s Experiment. — In these two 

 series of experiments we employed continuous metallic 

 disks turning between continuous parallel condensing plates. 

 In the course of the repetition of Pender's experiment, we 

 observed the following effect. When a voltao;e was em- 

 ployed sufficiently high to cause sparks to spring across 



* Phil. Mag. ser. 6, vol. ii. p. 179 (1901), et he. cit. toI. v. p. U (1903). 



