574 



NA TURE 



[April 1 1, 1895 



represented in Fig. 2, on a scale of twice the size of the filter 

 which we have actually used, or of true size for a filter on a 

 tube of 2 cm. diameter, which for some purposes may be 

 belter. One of eleven similar discs, of size adapted to a 

 tube of 2 cm. diameter, and an outermost disc with projecting 

 lugs, are shown, true size, and with the gauge of the wire-cloth 

 which we have actually used, shown true size, in Fig. 3. The 

 eleven little circular discs of wire cloth are held in position by 

 bending over them the four lugs belonging to the outermost 

 disc, and all are kept compactly together by a short piece of 



Fio. 2, 



india-rubber tube stretched over them outside, as shown in 

 Fig. 2. 



§4. We commenced with a few experiments to test the effi- 

 ciency of the testing filler, R, with no tested filter at F, and 

 merely continuous block-tin pipe, FCG, from the can to the 

 paraffin tunnel. First, working the bellows with no electrifi- 

 cation of the needle point, wc found no sensible electric effect 

 on the electrometer, which proved that, whether from natural 

 electrification of the air of the laboratory, or by the action of 

 the bellows, or by the passage of the air through the long metal 

 pipe, no electrification sensible to our test was produced. After 



I 



Fig. 3. — Twenty-five wires 10 the rentimetre. Diameter of each wire is 

 "*i6 mm. Hence each aperture is 0*24 mm. square. 



thai we kept the needle point, «, electrified, either positively or 



vcly, (or five or six minutes at a time by turning the little 



machine, and we found large effects rising to about 3J 



v.ilu in five minutes, positive or negative, according as n was 



positive or negative. 



§ 5. The apparatus is now ready to test the efficacy of filters 

 or other appliances of different kinds placed at I fur the pur- 

 pose of diseleclrifying air which has been electrified, whether 

 positively or negatively, by the electrified needle point n. We 



NO. 1328, VOL. 51] 



began with a filter of 12 wire-gauze discs, placed at F and kept 

 in metallic connection with the tin pipe outside. This nearly 

 halved the electricity shown by the electrometer. We then 

 tried 24, 4S, 72, 96, 120 wire-gauze discs, successively, placed 

 in groups of 24, and separated from one another by short 

 lengths of 2 cm. of lead tube, in the line of the flow of the air 

 between F and G (Fig. i), all kept in met.illic connection with 

 the block-tin pipe and the outer case of the electrometer. We 

 were surprised with the smallness of the additions to the dis- 

 eleclrifying efficiency of the 12 strainers first tried : for ex- 

 ample, the filtrr of 120 wire g.iuzes only reduced the electrical 

 indication to a little less than one-half of what it was with ttie 

 12 which we first tried. 



We found that cotton-wool between the spaces in the groups 

 of 24 wire gauze's largely increased the diselectrifying effect. 

 Thus, with 72 wire gauzes and cotton-wool we succeeded in 

 reducing the electrical effect to ahout one-twelfth of what it 

 was with only a filler of 12 wire gauzes ; but hitherto we have 

 not succeeded in rendering imperceptibly small the electricity 

 yielded liy the outflowing air to the testing filler R in our 

 method of ob-ervation. 



§ 6. We intend trying various methods of obtaining more 

 and more nearly complete diselectrificatlon of the electrified air 

 flowing out of the can at F ; and this for air electrified other- 

 wise than by the needle point, as shown in the diagram : for 

 instance, by an electrified flame in place of the needle point ; 

 or again by bubbling through water or other liquids. Mean- 

 time, the mere fact that the electricity, whether positive or 

 negative, given to air by an electrified needle point, can be 

 conveyed through 3 or 4 metres of small metal tube (I cm. 

 diameter), and shown on a quadrant electrometer by a receiving 

 filter, is not without interest. We may add now that, with the 

 receiving filter removed and merely a fine platinum wire put in 

 the mouth of the paraffin tunnel, we have found that enough of 

 electricity is taken from the outflowing air to be amply shown 

 by the quadrant electrometer ; which renders even more sur- 

 prising the fact that the diselectrifying power of 120 strainers 

 of fine wire-gauze should be so small as we have found it. 



"On the Question of Dielectric Hysteresis." By Alfred W. 

 Porter and David K. Morris. 



The experiment described was intended to test whether the 

 dissipation of energy that occurs in the dielectric of a condenser 

 is due to true hysteresis (as claimed by Riccardo Arno), or simply 

 to viscosity. To discrimina'e between them it is essential that 

 the condenser be put through a cyclic series of states at such a 

 slow rate that all viscous effects shall have had time to subside 

 before a measurement of the charge corresponding lo a certain 

 potential difference is made. It is essential also to arrange 



PotcntiaL differenct. 



Fic. I. 



that any lest of the charge which involves a change of stale 

 should ilsclf form part of the cycle. This was accomplis.ied 

 by making the cycles as shown in the figure, the cyclic order 

 being RrURPSQltR; the portions PU and QU repre- 

 senting discharges through a galvanometer. If hysl'-resi-^ 

 exist, the change of charge QT will be greater than the change 

 I'T. As the result of twelve cycles the ratio qt/pf comes 



I I 



out I -r azrr ; a second series of twelve gives I — 73773- 



Hence the cha'gc is scnsilily the same for a given value of 

 the potential difference, whether that value has been arrived 

 at from lower or from higher values than itself. 



