Electrical Properties of Selenium. 235 



the selenium to gather up into drops, and it becomes as 

 impossible to spread as if it were mercury. The Bunsen 

 flame should be adjusted so that the temperature is only just 

 above the melting-point, 217°; the selenium then assumes 

 a plastic semi-fluid condition and can be easily manipulated. 



When a satisfactory surface has been secured, the cell is 

 immediately removed by means of forceps and placed upon a 

 thick copper plate to cool quickly. The Se surface should at 

 this stage be black and lustrous, and the resistance between 

 the electrodes sensibly infinite. 



The Bunsen flame having been turned down sufficiently to 

 lower the temperature to about 120°, the cell is replaced upon 

 the hot plate. In a few minutes parts of its bright surface 

 become dimmed, and shortly afterwards the whole turns a 

 dull grey colour. The temperature is then cautiously raised 

 until signs of melting just begin to appear, generally near one 

 of the edges. When this occurs the burner is instantly with- 

 drawn and the flame slightly lowered. The darkened spot 

 recrystallizes in the course of a few seconds, and the burner 

 is then replaced and left for four or five hours, during which 

 time the temperature of the Se should be only a few degrees 

 below the melting-point. Another hour is occupied in slowly 

 cooling the cell, the flame being gradually lowered and 

 finally extinguished. This process of long heating and slow 

 cooling is generally spoken of as u annealing." 



The resistance of such a cell as that described is generally 

 from 50,000 to 100,000 ohms in the dark, and 50 or 60 per 

 cent, less in diffused daylight. If alternately screened and 

 exposed, it will with an electromotive force of 50 volts or 

 upwards easily actuate an ordinary telegraph relay. 



§ 3. Hypothesis as to the Action of Light upon Selenium. 



Selenium as supplied commercially is never free from 

 impurities, and indeed I have lately been informed by a dis- 

 tinguished chemist, who has given some attention to the 

 subject, that perfectly pure selenium has not yet been obtained. 

 Graham Bell states that the samples collected by him in 

 different parts of the world were found to contain, among 

 other things, the metals lead, iron, and arsenic. 



Ten years ago I suggested * that metallic selenides, whether 

 originally existing as impurities in the Se or formed at the 

 expense of the electrodes during the process of " annealing," 

 might play a very important part, not only in relation to the 

 conductivity of crystalline Se, but also in bringing about the 



* Proc. Phys. Soc. vol. vii. p. 129; Phil. Mag. August 1885, p. 178. 



R 2 



