Radium, Thorium, and Actinium in Electric Fields. 361 



could be reproduced. However, when this apparatus was 

 slightly modified, the increase of the experimental values 

 with the voltage practically disappeared. Fig. 2 shows 

 the modified form of Henderson's apparatus used by the 

 writer. The electrodes were two brass plates 10*4 cm. in 

 diameter, held, as indicated, at a distance of 2 cm. apart 

 in a bell-jar of 11*5 cm. internal diameter. The central 

 portions, whose diameter was 5*1, were removable, and the 

 activities on these were measured. Henderson used flat 

 brass plates separated by three vertical glass rods, the rods 

 and the edges of the plates being covered with paraffin wax 

 to stop brush discharge. He used cotton wadding to 

 prevent diffusion of active deposit from above or below 

 the plates into the region between them. The following 

 modifications were made by the writer to eliminate to a 

 much higher degree than was done by Henderson brush 

 discharge near the edges of the plates : — (1) The edges 

 were rounded off as shown in the figure, (2) a roll of silk 

 was used instead of cotton-wool, and (3) the upper plate 

 was supported from above. The results obtained with this 

 apparatus are given in column B. The continual increase 

 found by Henderson at high voltage has practically dis- 

 appeared. The difference between the values of column B 

 and the mean of column A is to be ascribed possibly to 

 some brush discharge still being present, and in the case of 

 the last three readings to the fact that for these readings 

 the plates were Yerj highly polished. The importance of 

 keeping the state of polish of the surfaces always the same 

 had not been recognized at this stage of the experiments. 



The high values given by the apparatus used by Henderson 

 were due to the action of brush discharge from the edges of 

 the plates or from cotton-wool fibres. If cotton- wool is held 

 against a pole of a Wimshurst machine in the dark, it glows 

 quite brightly, whereas good silk cloth does not. From any 

 point near the edges of the disks where a brush discharge 

 occurs, the gas is blown on to the opposite electrode, or more 

 probably on to the wall of the bell-jar, to which any neutral 

 deposit atoms carried by the gas will adhere. In a short 

 time all the gas between the electrodes will be entrained in 

 these so-formed jets, and only a fraction of the neutral 

 deposit will settle on the central portions of the plates. As 

 the voltage is increased so will the intensity of the brush 

 discharge, and with it the apparent percentage cathode 

 activity. This action of the brush discharge was illustrated 

 by filling a vessel similar to Henderson's with smoke. At 

 places on the glass wall opposite points on the electrodes 



