Electromagnetic Effect of Convection-Currents. 449 



grooves cut in their peripheries ; in these grooves wires are 

 wound. These serve as a galvanometer for determining the 

 needle-constant. When not in use they are held in the 

 position shown in the figure, but when it is desired to deter- 

 mine the needle-constant they are slipped on the shoulders 

 (AAA A) and pushed up in contact with the back of the disks. 

 Each has two turns : this arrangement will be referred to as 

 the disk-galvanometer. 



If a known current is sent through the disk-galvanometer, 

 and the geometrical constant be known, the part of the con- 

 stant depending on the field and needle is determined. 



The current is measured by a sine-galvanometer, placed in 

 another part of the room. To determine H at the sine-gal- 

 vanometer a metre brass circle is put around the sine-galva- 

 nometer, and the needle of the latter used as the needle of the 

 tangent-galvanometer thus made. Using this tangent-gal- 

 vanometer in connexion with a Weber electrodynamometer, 

 H at the sine-galvanometer is measured. 



The charging was by a Holtz machine connected to a 

 battery of six gallon Leyden jars. These latter are in circuit 

 with a revevsing-key, an electrostiatic gauge, and the disks. 



The potential w^as measured by a large absolute electro- 

 meter ; all previous observers have used spark- length between 

 balls, with Thomson's formula. Greater accuracy is claimed 

 for this work, largely on this account. 



In this instrument the movable plate is at one end of a 

 balance-arm, from the other end of which hangs, on knife- 

 edges, a balance-pan. This movable plate is surrounded by 

 a £uard-rinor. 



The lower plate is fixed by an insulating rod to a metal 

 stem, which slides up and down in guides. The distances are 

 read off on a scale on the metal stem. The zero reading is 

 got by inserting a piece of plane parallel glass whose thick- 

 ness has been measured. The lower plate and guard -ring 

 haA-e a diameter of 35 centim., and the movable disk a dia- 

 meter of 10 centim. 



The routine of the observations was as follows : — A deter- 

 nn'nation of H and the needle -constant (/S) was first made. 

 The electrostatic gauge was then set at a certain point, and 

 readings of difference of potential were taken. The disks 

 were now started, electrified, and a series of three elongations 

 of the needle taken ; the electrification reversed and three 

 more elongations taken, &c. 



About every five minutes speed-readings had to be noted, 

 and at each reversal it was necessary to replenish the charge 

 in order to keep the gauge-arm just at the mark. In this way 



