446 Prof. J. J. Thomson on the Discharge of Electricity 



presence in such a chain of any second body, even though it may 

 be a good conductor of electricity. Thus, when 

 a tube such as that in fig. 7 is used, which has Fig. 7. 

 a barometer-tube attached to it, so that by raising 

 or lowering the vessel into which the tube dips a 

 mercury pellet may be introduced into the dis- 

 charge-circuit, the spark-length in the primary 

 circuit may be so adjusted that a discharge passes 

 when there is a clear way round the tube, but 

 stops when a pellet of mercury is forced up so as 

 to close the gangway. I noticed a similar effect 

 in my experiments with a long vacuum-tube 

 described in the Proceedings of the Royal Society 

 for Jan. 1891. 



I had another discharge-tube prepared, of which 

 a section is shown in fig. 8, a, in which a dia- 

 phragm (ABj of thin copper plate was placed 

 across the tube; the diaphragm happened to catch at the bottom 

 of the tube, so that it divided the latter rather unequally, and 

 left a narrow passage round its edge. As much of the discharge 

 as there was room for went round the edge of the plate ; the 



remainder was not able to get through the copper, but formed 

 a closed circuit by itself in the larger segment of the tube. 

 In another tube/which is represented in section in fig. 8, /3, 

 the copper diaphragm was attached to the walls of the tube 

 by sealing-wax, so that there was no free way ; in this case the 

 discharge again refused to go through the copper, and split 

 up into two separate discharges, as in the figure. When the 

 tube was divided by copper diaphragms into six segments, as 

 in fig. 8, 7, no discharge at all would pass through. When the 

 primary was slipped up the tube above the diaphragm, a 

 brilliant discharge was obtained. These four experiments all 

 illustrate the difficulty which the electricity has in getting 

 transferred from a gas to another conductor. 



