178 



Scientific Proceedings, Roijal Dublin Society. 



terminal being connected to the plate and grid-circuits. The curves (fig. 1) 

 represent the characteristics xmder these conditions. , 



A few preliminary observations were made in the manner indicated in 

 fig 2 ; but these were merely made with a view to testing the possibility of 

 the method, and, as this mode of working is very insensitive, it is not 

 necessary to quote any of the I'esults obtained. 'J'he first modification tried 

 was to balance the current in the milammeter (fig. 2) by an opposed current 

 from a single cell in series with an adjustable high resistance. Some 

 experiments were made thus, a Paul Unipivot galvanometer with an 



farad 



Fig. 3. 



universal shunt being employed in place of the milammeter. As it was 

 convenient to utilize the laboratory supply for the 50-volt plate-circuit 

 battery, it was found impossible to connect the town supply of 200 volts a.c. 

 directly to A A in consequence of the insufficient insulation between the 

 a.c. and d.c. circuits. An insulated 200-volt a.c. supply was therefore 

 obtained, using a small step-up transformer from the 12-volt a.c. galvano- 

 meter lamp circuit. Beyond this no special care was taken as regards 

 insulation or shielding of the apparatus. A few observations were thus 

 made, using a graduated air-condenser of the rotating-sector type. This 



