Electromagnetic Efect of Convection-Currents. 447 



He uses a difference of potential corresponding to a spark- 

 length of 0"3 centim. in air between balls of 2 centim. dia- 

 meter, i. e. about 33 electrostatic units, equal to the sparking 

 potential between plane surfaces at 0*26 centim. The disk is 

 an imperfect conductor, and altogether it does not seem clear, 

 in spite of the precautions taken, that this is not due to 

 convection-currents. 



In the Berlin apparatus^ as stated above, the needle is 

 near the edge of the disk ; the magnetic effect produced is 

 assumed to be proportional to the surface-density multiplied 

 bj the linear velocity ; hence the force will be much greater 

 at the edge of the disk than near the centre ; but the field will 

 be more irregular, and so make accurate measurements more 

 difficult. 



In the present apparatus a uniform field is secured by using 

 two vertical disks rotating about horizontal axes in the same 

 line ; the needle system is placed between the disks, opposite 

 their centres. The disks are in the meridian ; they are gilded 

 on the faces turned towards the needle. Between the disks 

 are placed two glass condenser-plates gilded on the surfaces 

 near the disk ; and between these glasses is the needle. The 

 whole apparatus is symmetrical about the lower needle of the 

 astatic system. 



Each disk is surrounded by a gilded hard rubber guard- 

 plate in order to keep the density of the charge uniform at 

 the edges. The guard-plates are provided with adjusting- 

 screws to enable them to be put accurately in the plane of the 

 disks ; and the glass plates in turn have adjusting-screws for 

 securing parallelism with the guard-plates. The glass was 

 carefully chosen as being nearly plane. Disks, glass plates, 

 and guard-plates all have radial scratches, to prevent con- 

 duction-currents from circulating around the coatings. 



In the periphery of the disk are set eight brass studs 

 which penetrate radially for about 5 centim., then turning off 

 at a right angle run parallel to the axis until they come out 

 on the surface of the disks. They there make contact with 

 the gold foil. Metal brushes set in the guard-plate bear on 

 these studs, and in this way the disks are electrified. 



The figure (PI. IX. fig. 1) gives a vertical projection of 

 the entire disk-apparatus : — D D are the disks ; Gr G G G the 

 guard-rings; YYYY the condenser-plates; HE HE, hard 

 rubber rings fitting on the shoulders A A; XXXX bearing- 

 boxes for the axle ; P P P P supporting-standards ; E E inetal 

 bases sliding in the bed B B, and held in any position by 

 screws Z ; PF the bases carrying the glass plates, slidmg in 

 the same way as the others. S S S S are the adjusting-screws 



2H2 



