448 Messrs. V. Cremieu cmd H. Pender on 



and that taken bv the condenser after contact was much 

 greater, all else being- the same^ than the residual potential 

 after the discharge. 



These facts explain sufficiently the diminution of the effects 

 of induction of the disks in motion when a capacity is put 

 in parallel with the disks. The presence of this capacity 

 decreases the variation of the charge at each reversal of the 

 commutator, and, in consequence, the effects of induction, 

 which are proportional to this variation ; further, the indi- 

 cations of the electrometer, which gives the maximum potential 

 assumed by the disks, cannot be affected to the same degree. 



Consequently, it appears that the effects observed in Pender's 

 experiment are due to the charged disks in movement, as 

 demanded by the theory of electric convection. 



Verifications icithout fixed Condensing-jDlates. — To eliminate 

 entirely the questionable role of the fixed condensing-plates, 

 we made the following modification : — 



Pender's two disks, diameter 31 cm., were placed opposite 

 each other, 1 cm. apart_, and arranged to turn in opposite 

 directions. Concentric with the two disks was placed a coil 

 34 cm. internal diameter and 3 cm. thick. The coil, which was 

 entirely inclosed in a brass sheath, consisted of 1300 turns 

 of copper wire, with a total resistance of 60 ohms. The 

 commutator was arranged so as to connect this coil to the 

 galvanometer synchronously with the charging and dis- 

 charging of the turning disks. The deflexion o£ the galvano- 

 meter under these conditions was much less than in the 

 former experiments, on account o£ the considerable decrease 

 in the capacity of the two disks thus arranged. In spite of 

 that, however, the deflexion was sufficiently large to be 

 measured^ and the agreement between the observed and 

 calculated deflexion was quite satisfactory. 



Here we stopped our experiments with the induction 

 method. These experiments can, in fact, give no information 

 in regard to the nature of the magnetic field produced by the 

 movement of the two disks, whereas it is essential to know 

 whether this field is permanent wdth the rotation, or is merely 

 an instantaneous field produced at the moment of charging or 

 reversing the charge, resulting from magnetic perturbations 

 thereby set up. 



Experiments on the Direct Magnetic E feet. — First_, we utilized 

 the same arrangement as in the preceding induction experi- 

 ment. The coil was replaced by a delicate astatic system 

 inclosed in a metallic tube so arranged that the lower needle 

 of the system was 1 cm. above the upper edge of the disks, in 

 a plane equally distant from the two. The tube was protected 

 from the air-currents caused by the rotating disks by a large 



