88 Lord Kelvin and Mr. M. Maclean. [May 31, 



one for the water-dropper C, and one for the charging wire D. 

 Both the water-dropper, and the charging wire, ending with a pin- 

 point as sharp as possible, are insulated by solid paraffin, which is 

 surrounded by a metal tube, as shown in half size in Fig. 2. To 

 start with they were supported by pieces of vulcanite embedded in 

 paraffin. But it was found that after the lapse of some days, 

 (possibly on account of ozone generated by the incessant brush dis- 

 charges), the insulation had utterly failed in both of them. The 

 vulcanite pieces were then taken out, and solid paraffin, with the 

 metal guard-tube round it to screen it from electrically influencing 

 the water-dropper, was substituted. This has proved quite satis- 

 factory : the water- dropper, with the flow of water stopped, holds a 

 positive or a negative charge for hours. 



6. A quadrant electrometer E (described in " Electrostatics and 

 Magnetism," 346 353) was set up on the top of the vat near the 

 water-dropper, as shown in Fig. 1. It was used with lamp and semi- 

 transparent scale to indicate the difference of potential between the 

 water-dropper and the vat. The sensibility of the electrometer was 

 21 scale divisions (half -millimetres) per volt, and as the scale was 

 90 centimetres long, difference of potentials up to 43 volts positive or 

 negative, could be read by adjusting the metallic zero to the middle 

 of the scale. A fractional plate-electric machine was used, and by 

 means of it, in connection with the pin-point, the air inside the vat 

 could be electrified either positively or negatively. 



7. The vat was fixed in position in the Apparatus Boom of the 

 Natural Philosophy Department of the University of Glasgow on 

 the 13th of December, 1893, and for more than three months the air 

 inside was left undisturbed except by discharges from the pin-point 

 through the electrifying wire, and by the spray from the water- 

 dropper. Thus the air was becoming more and more freed of dust 

 day by day. Yet at the end of the four months we found that the 

 air was as easily electrified, either positively or negatively, as it was 

 at the beginning ; and that if we electrify it strongly by turning the 

 machine for half-an-hour, it retains a considerable portion of this 

 electrification for several hours. 



8. Observations were taken almost daily since the 13th Decem- 

 ber; but the following, taken on the 8th of February, the 12th of 

 March, and the 23rd of April, will serve as specimens, the results 

 being shown in each case by a curve. At all these dates the air 

 must have been very free from dust. Both daring the charging and 

 during the observations the case of the electrometer and one pair of 

 quadrants are kept metallically connected to the vat. During the 

 charging the water-dropper and the other pair of quadrants were also 

 kept in connection with the vat. Immediately after the charging 

 was stopped the charging- wire was connected metallically to the out- 



