70 Scientific Proceedings, Royal Dublin Society. 



The electrometer is of the Dolezalek type, with a coarse quartz suspen- 

 sion. The needle was always charged to about + 300 volts by means of a 

 hygroscopic battery before commencing a set of readings, and the sensitivity 

 tested. This varied somewhat from day to day, ranging betv.'een about 340 

 and about 370 scale divisions per volt, but only fell a few per cent, during a 

 set of readings, as the insulation of the needle was very good. The wires 

 connecting the testing vessel and the electrometer to the key were each 

 about 25 or 30 ems. long. They were everywhere within a few centimetres 

 of the slate bench, which, being a conductor, acted to a certain extent as a 

 protection against electrostatic induction effects ; but, as it was not necessary 

 to approach the apparatus during a test, no other electrostatic screen was" 

 used. A wooden screen was interposed between the testing vessel and the 

 electrometer to protect the latter from possible heating effects. The 

 connecting wires passed round the end of this screen. 



A double-barrel Geryk pump was used to exhaust the testing vessel, the 

 air-pressure being measured on a mercury gauge by means of a reading 

 microscope. By reducing the pressure the photo-electric current can be 

 magnified by collision and the sensitiveness greatly increased. 



It has been shown (see Allen, " Photo-Electricity," p. 64) that the 

 pressure p which gives the maximum magnification is given by the relation 



E 



f = -z — , where E is the voltage applied to the grid, d the distance 



from grid to plate, N the number of collisions made by one electron per 

 centimetre at a pressure of 1 mm. of mercury, and v the potential difference 

 through which an electron must fall in order to produce fresh ions at impact. 

 In this case E was about 230 volts, and d about 0'4 cm. For air N = 14'6, and 

 V = 25 volts [Townsend (8)], which gives ^j = 1-6 mm. of merciny. Most of 

 the experiments were carried out at a pressure of 2 mm. of mercury, which 

 is sufficiently close to the calculated value to give nearly the maximum 



magnification. The maximvmi current is equal to le^^, where e is the 

 base of the natural system of logarithms, and i is the current due to the 

 photo-electrons emitted. Putting in the above figures, we find that the 

 maximum current is 29'3i, so we may assume that working at 2 mm. the 

 magnification is nearly thirty-fold. This is borne out by a test on a zinc 

 plate, in which the current at 2 mm. pressure was about forty-five times that 

 at atmospheric pressure. This latter would be considerably smaller than 

 the full value corresponding to the photo-electrons emitted, as a certain 

 number of them would diffuse back to the plate in spite of the applied 

 electric field, so the agreement is quite satisfactory. 



To test the capacity of the plate-quadrant system, the deflexion produced 



