DTSiOHARrxE OF ELECTRK'TTY. 247 



nceted togetlier so that it tunned a. (closed circuit, the other tube con- 

 tained nothiiiji' but air at a low pressure. When this double tube wsis 

 placed outside the primary the dischari^c^ went, at the passage of each 

 sjiar-k, througli the tube without a wire, while the tube containing the 

 wire remained <iuite dark. A similar experiment was tried by taking a 

 cylindrical tube and siis])ending in it a metal ling co-axial with the 

 tube; in this case it was easy to adjust the spark-length so that no dis- 

 charge ])assed through the tul)e when tlie primary Avas ]daced round it 

 at the level of the ring, while a discharge passed as soon as the primary 

 was moved above or below the ring. 



Another very convenient tube for showing this etiect is the one with 

 the hollow down the midd]<>, Fig. 11; when this is pumped so that dis- 

 charges can pass througli the outer tube the spark-length can be 

 adjusted so that the discharge stops immediately when a metal tube, a 

 test-tube containing a strong solution of an electrolyte, or a tube con- 

 taining air at a pressure at which it is electrically very weak, is placed 

 in the central opening. The discharge is renewed again as soon as the 

 tubes are removed. On one occasion, when the large tube was in a 

 peculiarly sensitive state, 1 was able to see distinctly the diminution 

 produced by a dielectric in the electro-motive intensity parallel to its 

 surface. The discharge stopp<'d as soon as a stick of sulphur or a 

 glass rod sufliciently large almost to till the opening was inserted, and 

 was renewed again as soon as these were withdrawn. It requires how- 

 ever the large tube to be in an extremely sensitive state for the efitect 

 produced by a dielectric to be apparent, and I have only on one occa- 

 sion succeeded in getting the tube into this condition. The effect on 

 that occasion however was so d<^tinite an<l regular that 1 ha^e no 

 doubt as to the existence of the screening eft'ect due to the dielectric. 



When the tube is of average sensitiveness <lielectrics do not pro- 

 duce any appreciable ettect, nor can the iniiuence of even comparatively 

 so good a conductor as distilled water be detected, and it is not until 

 after the addition of a. considerable quantity, 10 to 20 i>er cent of sul- 

 phuric; acid or anunonium cliloride that the insertion or withdrawal of 

 the tube stops or starts the discharge. 



A tube containing air at a low i)ressure is very eflicacious in stopi)iug 

 the discharge, and the result of the comparison of the relative effects 

 of an exhausted tube and a tube of the same size and shape contain- 

 ing a solution of an electrolyte are very remarkable. I fimnd that an 

 exhausted tube which contained air at a very low pressure (less than 

 yL of a millimetre) produced as nuich effect on the discharge in the 

 outer tube as a tube containing at least r>0,(K)() times as many mole- 

 cnles of ammonium chloiide. This would be expressed in the language 

 of electrolysis by saying that under the electro-moti\e intensity to 

 which it was ex])osed in Ihis e.\peiiu)ent the mohnmlar conducti\ity of 

 the gas is 50,000 times that of the electrolyte. The pro^jortion between 

 the number of air molecules aiul the luimber of molecules of an elec- 



