of Substances upon Voltaic Action. 451 



outer ends of the insulated zinc and silk-covered copper 

 wires were fixed by means of melted sealing-wax upon a 

 narrow strip of very thin wood for the purpose of keeping 

 them in their proper positions. The tubes and wires were 

 carefully adjusted so that none of them came into contact 

 with the lead. The frame moved in parallel wooden guides, 

 not shown in the sketch. The sliding contact-parts were 

 coated with blacklead, so that the to-and-fro movement was 

 very smooth. 



The sketch represents the glass tubes as being midway 

 between the bars of lead. 



The inner surfaces of the two pairs of bars were 5 inches 

 apart. The electrodes were amalgamated. The galvano- 

 meter employed in all the subsequent experiments with this 

 apparatus was a Thomson tripod one of 6117 ohms resistance, 

 made by Elliott. Two other forms of apparatus were con- 

 structed and employed, but they were unsuitable and need 

 not be described. 



Method of Manipulation, 



Immediately the circuit was closed the amount of deflexion 

 caused by voltaic disturbance was recorded ; the image of 

 light was then brought to zero ; and this was repeated until 

 the voltaic current either ceased or became quite steady. 

 When the liquid had only been recently made the disturbance 

 was considerable, and required an hour or more to subside ; 

 but after it had been in use a week or more the irregularity 

 subsided in a few minutes. In every case, however, the 

 image was ultimately brought to zero by means of the con- 

 trolling magnet, and the experiments were then made. 



Whilst sliding the movable part to and fro, great care was 

 taken not to suddenly stop or jerk it, and to observe that 

 neither the glass tubes nor insulated zinc wires came into 

 contact with the lead bars ; the insulated copper wires were 

 so thin and flexible that their occasional slight contact pro- 

 duced no disturbance. In consequence of the extreme 

 sensitiveness of the arrangement, considerable experience 

 was necessary in using the apparatus so as to obtain reliable 

 results. This great sensitiveness to movement was found 

 to be due to bubbles of air or gas upon the electrodes and 

 was subsequently obviated. 



With this apparatus the maximum deflexion caused b} r 

 motion of the tubes, either in the presence or absence of the 

 lead bars, commenced almost immediately and was usually 

 obtained in about fifteen minutes; whilst with Apparatus 

 " No. 2 " it required about 2| hours, probably in con- 



