1085 



couipomid ga\e rise to a decrease of the light al)sorption of the 

 vaporized tungsten. In this respect we have obtained good results 

 with Na,PO„ Na,WO„ KCN, iVnJ), jVaF, CaF, and with many 

 other substances. 



In tiie following way we still made it prol)able, that the vaporized 

 tungsten was present unaltered in the N^aC/ layer. When H,0 was 

 introduced into the lamp a quantity of hydrogen was developed. 

 As we had found already before that the NaCl did react neutrally, 

 so that there was no free /V(/, this development of hydrogen must 

 be due to the presence of free IF. 



While therefore the action of the salt layer depends little on its 

 chemical composition, the state of the salt on the wall often proved 

 to have influence. When much more salt was introduced into the 

 lamp, a white crystallized precipitate was formed just as on the 

 introduction of moist air. In this form the salt was much less active. 

 Lamps, the bulb wall of which is covered with salt crystals, are 

 not much better than those that do not contain any salt. A similar 

 phenomenon is observed with qldfis. When this is brought in small 

 quantities on the wire in pulverous state, and when this is distilled 

 on the bulb wall, this fine distillate shows a similar action as iV^iC/, 

 while the smooth bulb wall evidently is inactive. Still we cannot 

 generalize and say, that crystallized substances are inactive. With 

 crystallized NaF and KCl e.g. good results were obtained though 

 not so good as when these substances were in the amorphous state. 



In general however it was proved to be of importance for the 

 decolorating action of (he salt that it was present on the wall in a 

 very finely divided state. Now it is probable, that the W particles 

 coming with a relatively high velocity from the wire, will penetrate over 

 some distance into the salt and stick then in this layer in an extre- 

 mely disperse state. That in a similar state the light absorption can 

 be very small, had already been found by v. Weimarn ^) for colourless 

 gold-glass and gold-quartz and by Zsigmondy ') for colourless colloidal 

 gold-solutions, while v. Weimahn ') and Mc. Intosh and Edson ^) 

 have pointed to the phenomenon that colourless colloidal solutions can 

 be obtained by rapid cooling of true solutions. The particles are then 

 prevented from forming greater conglomerates. By means of the 



1) P. V. Weimarn, Zeitschr. fur Chem. und Ind. der KoUoïde, (11), 287, 1912. 

 ') R. Zsigmondy, Zur Erkenntniss der Kolloide, 1905 and Zeitschr. fiir phys. 

 Chemie, (56), 65, 1906. 



') P. V. Weimarn, Grundziige der Dispersoidchemie, p. 70, 1911. 



♦) D. Mc. Intosch and R. Edson, Journal Am. Chem. Soc. (38), 613, 1916. 



