A NEW CURRENT WEIGHER, ETC. 497 



in parallel, so that the current was about 2 amperes. After the washing with alcohol 

 had l>een continued for 20 minutes or half-an-hour there was a crackling noise, and 

 hundreds of tiny sparks appeared over the surface of the cylinder ; simultaneously the 

 lamps ceased to glow. A measurement of the insulation resistance between the coils 

 showed it to be of the order of 300 megohms ; the shorting pieces had been burnt out 

 with a most satisfactory result. To prevent the recurrence of the low insulation resist- 

 ance the washing was continued ; occasionally the lamps glowed, but with continued 

 washing the shorts were burnt out as before. When the insulation resistance was of 

 the order of 1000 megohms, with an applied pressure of 20 volts, the cylinder was 

 lifted from its bearings and placed in others secured to a framework resting on the 

 top of a bath of melted paraffin wax. About one-third of the circumference of the 

 cylinder dipped into the hot liquid. The cylinder was rotated until the marble was 

 sufficiently warm to keep the wax on its surface in a liquid condition ; it was then 

 removed for the wax to solidify, and afterwards dipped once more, in order to obtain 

 a thicker coat. The insulation resistance was measured while the cylinder was hot, 

 and also when the wax had solidified ; the latter value was always the greater. 

 After the lapse of a week or ten days, the ends and interior of the cylinder were 

 cleaned and preparations made for further measurements of diameters. The wax was 

 carefully removed from several parts of the cylinder and the strands cleaned by 

 rubbing with a small pad of silk ; the measurement of six or eight diameters was 

 then carefully made, the steel gauges being set up as before. A summary of these 

 measurements follows (see Table VII.), from which it is inferred that there was no 

 appreciable change in the diametral dimensions. 



In. one of the large cylinders the insulation resistance between the two upper and 

 the two lower helices was at first comparatively low, viz., 2000 megohms. The cause 

 of this .was found to lie in the internal ivory plugs through which the copper leads 

 passed. As it was impossible to remove these without stripping the cylinder, they 

 were slotted in such a way as to reduce theTsection of the conducting material ; the 

 insulation resistance was thus increased to 10,000 megohms. Insulation tests on 

 fresh ivory pieces were invariably satisfactory, but two such pieces inserted in the 

 ampere balance appeared to deteriorate with time, and eventually had to be replaced 

 by ebonite. 



To prevent damage to the surface of the wax with which the coils were coated, it 

 was thought desirable to cover it with a harder insulating material. Shellac varnish 

 was tried and used for the larger coils, but the suspended ones were untouched owing 

 to the results of experiments on equal surfaces of paraffin wax and shellac varnish. 

 The latter was found to be much more hygroscopic than the -former. From the 

 measurements made it is estimated that each suspended cylinder coated with paraffin 

 wax would change in mass by 6*8 milligrammes if removed from a dry atmosphere 

 to one saturated with moisture ; had the outer coating been shellac varnish the 

 corresponding change would be 146 milligrammes, 



VOL. covil. A. 3 8 



