APPENDIX III. 



ELECTRIFICATION OF RADIOMETER VANES. 



One of the few defects of a radiometer is the electrification of its 

 vanes, which is caused by air currents in exhausting the apparatus or 

 by standing with the stop-cocks open to the exhaust pump. Since this 

 effect has not been recorded except by Stewart (loc, cit.) and myself, 

 and since no efficient remedies to prevent it have been given heretofore, 

 it seems desirable to indicate the progress made in this direction. 



The electrification of the vanes by air currents occurs in exhausting 

 the radiometer, especially when it is first assembled and the vanes are 

 new. This does not occur until a critical pressure of about o.i mm. 

 is attained, when the deflection is generally thrown entirely off the 

 scale, and the vanes may not return for many hours. In the present 

 work, for quite a while the joints of the apparatus leaked a slight 

 amount, so that the stop-cocks were left open to the exhaust pump, in 

 order that the change of pressure would not affect the sensibility of the 

 radiometer for a day's work. It was then found that the vanes became 

 electrified, without any apparent provocation, while making readings, 

 and that it was more aggravated in the hotter summer months. As 

 far as could be ascertained, the vane that was nearest to the rock-salt 

 window would be attracted to it. The mercury vapor was at once sus- 

 pected as the source of electrification. The remedy consisted in insert- 

 ing a glass tube containing gold foil spread on absorbent cotton, to 

 present a large surface. After that there was no further electrification 

 either from air curents or the mercury vapor, until the gold had become 

 amalgamated. As an example, the apparatus stood from October 20, 

 1902, to February 20, 1903 (four months) , without becoming electrified. 

 The pump was then cleaned and a little more gold foil added, when the 

 apparatus stood from March 10 to June 12, 1903, before it became elec- 

 trified. The gold was then found to be entirely amalgamated. Of 

 course, part of this could have been obviated by closing the stop-cocks, 

 but this in time starts them to leaking, hence the present method was the 

 more preferable. This was especially desirable after the joints had 

 become so tight that there was practically no leaking when the radio- 

 meter stood from June to September, 1903 (stop-cock closed) and main- 

 tained a sufficient sensitiveness that a candle several meters from the 

 slit threw the deflection clear off the scale. 



It is a peculiar fact that with gold foil in series old vanes do not 

 become electrified in exhausting the apparatus. Just the opposite is 

 true for newly blackened values, which on exhausting adhere to the 



124 



