68 



Scientific Proceedings^ Royal Dublin Society. 



the distance between the grid and the plate was about 4 mm. As the grid 

 is about 5'2 cm. in diameter, the capacity of the condenser so formed 



TT X 2'62 



would be about 



or about 4 cm. A decrease of the distance 



47r X 0-4' 



between the grid and the plate by O'l mm. would increase the capacity by 

 O'l cm. As the grid was charged to a potential of about 230 volts, and the 

 plate remained nearly at zero, the increase of negative charge induced on the 

 plate, i.e., the positive charge driven on to the quadrant by the approach of 

 230 X 01 



the grid, would be 



300 



or about 0'075 E. s. u. As will be seen 



later, one division of the electrometer scale corresponded to 1'5 x 10'^ E.s.u., 

 so this would cause a deflexion of fifty scale divisions. Obviously then a 

 variation of the distance between the grid and the plate of O'OOl mm. would 

 just cause an appreciable error. It is evident that a wire gauze grid 

 supported at its rim (which was the type used in the experiments referred 

 to) will sag considerably if the central part is even slightly warmer than the 

 rim, as it is certain to be if it is exposed to an intense beam of light. 

 Accordingly, in constructing the testing vessel shown in section in fig. 2, 

 special attention was paid to this point. 



The vessel, which is circular in plan, is made in two parts of solid brass 

 sufficiently massive to ensure that its temperature changes only compara- 

 tively slowly, especially as its walls are outside the light beam. A thin 

 paper washer is used at the joint, which is sealed with a little warmed tallow. 

 The two parts of the vessel are electrically connected and earthed, thus 

 preventing tlie possibility of any conduction leakage from the grid to the 

 plate. The vessel is carried on levelling screws not shown in the figure. 



The insulated plate has a small rim, as shown, to hold powders, or, is 



