Films of Sodium and Potassium. Ill 



°be produced with selenium. Deposited at room temperature 

 in an exhausted bulb", in a manner similar to that employed 

 with sodium, the deposit is homogeneous and yellow or orange 

 in colour. If, however, the wall on which the deposition takes 

 place is heated above 60° 0. (the critical temperature for 

 selenium deposition) the colour by transmitted light is sky- 

 blue, and the film scatters very powerfully light of a ver- 

 milion-red colour. There is very little specular reflexion, 

 in which respect the selenium deposit differs from the 

 sodium films. The selenium deposits are permanent in air, 

 and the bulbs can be broken up for further examination. 



I have made casts in celluloid of the blue deposit, and 

 find that they scatter blue light, indicating a microscopic 

 structure. 



Doubtless much light could be thrown upon the whole 

 subject if some method could be devised of making an 

 accurate cast or replica of the sodium surfaces in some other 

 metal. Experiments along this line are in progress. Ex- 

 tremely complicated colour effects are produced if a trace of 

 some volatile hydrocarbon is present in the bulb. Some of 

 these effects were described in my earlier papers. In a 

 recent experiment I introduced a trace of a heavy lubricating 

 oil into the bulb, and distilled it along with the sodium 

 against a surface cooled to liquid air temperature. The film 

 was of a deep blood-red colour which disappeared at room 

 temperature, owing to its breaking up into granules. It is 

 unnecessary to go into these details at present, for until we 

 have some plausible hypothesis to «pply to the c:ises where 

 pure sodium alone is concerned, speculations as to the action 

 of liquid films in contact with the crystalline deposits are 

 hopeless. 



It is, perhaps, of interest to draw attention to a cnse 

 mentioned once by Lord Rayleigh in connexion with the 

 hypothesis that the cavities between the metallic crystals 

 act as selective absorbing resonators for the light- waves. 



Consider a surface shown in section in fig 5 (a grating for 



Fig. 5. 



example) in which the depth of the groove is one quarter 

 of a certain wave-length X. As is well known, \ will be 

 absent by interference in the central image, and appear in 



