the Magnetic Etfect of Eleetrio Convection. 449 



sheet of niiea properly })lace(l. The sensibility of the system 

 was cleterminetl by a test-coil. The reversal of a current of 

 10"* ampere in this coil gave a deflexion of 20 mm. on 

 a scale 4 metres distant. The convection-current possible 

 to realize in this experiment was only 2*5x10"'^ ampere, 

 and in consequence the deflexion expected was less than 

 5 mm. Hence the experiment was purely qualitative. 

 However, there was no doubt as to the effect. At each 

 reversal of the sign of charge upon the turning disks the 

 system was deflected in the expected direction, but it was 

 impossible to say, on account of the unsteadiness of the needle, 

 whether the deflexion was permanent or only an impulse. 



It is to be noted, that in spite of the smallness of this 

 result, it is nevertheless of great importance. Any perturbing 

 (effects due to the fixed condensing-plates or to currents which 

 might circulate in the moving disks are entirely eliminated ; 

 because the first are suppressed, and the second would destroy 

 each other, as the two disks turn in opposite directions. 



In the above experiment only a small part of the disk is 

 effective. The needle is in a very unfavourable position, 

 because the equipotential surfaces of the magnetic field are 

 very sharp over the edge of the disk, and further, it is poorly 

 protected against the air-currents, which produce a con- 

 siderable unsteadiness. These considerations led us to return 

 to Rowdand^s original method. 



A single disk turning in a horizontal plane was inclosed 

 in a box of ebonite, the inside of which, covered with tinfoil, 

 formed the condensing-plates, thus increasing considerably 

 the capacity of the moving disk. This arrangement allowed 

 us to put the astatic system very near the disk, in a region 

 where the magnetic field is horizontal and quite constant ; 

 and, finally, the effects of the air-currents were entirely 

 eliminated. We were therefore enabled to give the needle a 

 much longer period, without having to fear as much the 

 mechanical perturbations which might be produced during 

 the interval of time necessary for the system to take its 

 complete swing. 



This time the deflexions obtained were considerable, 20 to 

 30 mm., of the order and in the direction expected. But it 

 was still impossible to decide with certainty whether the 

 deflexions w^ere permanent or simply impulses. 



Pend.'r observed in his former experiments "^j in which a 

 perfect stability was realized, deflexions which were un- 

 doubtedly permanent, but he did not know at that time of 

 the magnetic effects which are produced in the neighbourhood 



* Phil. Ma.j^. vol. v. .Jan. 1903, p. 84. 

 FhiL Mag. S. 6. Vol. 6. No. 34. Oct. 1903. 2 G 



