17 



being of '0065 inches diameter. The curves {A + ), {B+), 

 (C+), show the relation between the potential difference of 

 the electrodes, and the corresponding electrometer deflections 

 per 0*5 minute for distances of 7, 14, and 28 mm. between 

 the gauze and the flat end of the central cylindrical electrode, 

 the gauze being positive to the central electrode. The curves 

 (J — ), ^B — )y (C — ), are corresponding curves obtained when 

 the gauze was negative. Fig. 3 shows similar curves using 

 a brass guaze with 20 strands to the inch, each wire being 

 '010 in. in diameter. 



Table A. 



Diameter of Deflection for 400 volts. 



Material. Strands perin. -iv.v^o ,%, -r^ n — Z- 1 ^ TZ 



^ U ires, in. Deflection for 50 volts. 



Brass X ... 20 -010 1-4 



Brass X ... H '0185 1-25 



Brass X •. 14 015 1-33 



Brass X ... 40 -007 1-7 



Ironx ... 14 '013 1-25 



Ironx ... 21 -016 1-34 



Galv. Iron x 14 -0136 1-26 



Copper II ... 16 -013 1-14 



Copper II ... 32 -013 1*5 



X Represents crossed wire-mesh gauzes. 



II Represents parallel wire gauzes. 



A large number of other gauzes were similarly tested, in- 

 cluding some which were built in the form of grids, with one 

 set only of straight wires arranged parallel to each other. 

 As figs. 2 and 3 are, however, typical, the results obtained 

 with some of the other gauzes are given in Table A in terms 

 of the deflections per 0*5 minute at potential differences of 

 400 and 50 volts respectively between the cylinders with the 

 central electrode in all cases 28 mm. from the gauze. The 

 state of the surface of the gauze was found to have little 

 effect, even, when the gauze was wetted, and in the case of 

 iron allowed to rust slightly. 



At intervals during each set of experiments the draught 

 was stopped and a reading taken of the amount of ionization 

 which entered the chamber independently of the draught. 

 No special precautions were taken tO' dry the air, and on one 

 or two occasions the amount of ionization produced by the 

 uranium showed changes which were apparently due to slight 

 dampness of the uranium surface. This account'Od for the 

 reduction in the amount of ionization which was observed 

 on such occasions, as the path of the a particles through 

 the gas V70uld be slightly reduced. 



It will be observed in both figs. 2 and 3 that there is no 

 very close approach to saturation, although the approach is 



