087 



williont salt. We tlionglit now lliat tlie action of the salt might be 

 exphiineii in tliis way, that the electrically still separated molecnles 

 behave optically more or less as separated molecnles and give like 

 gas molecnles a line absorption spectrnm, while only when their 

 mntnal distances have become so small that the electrical condnction 

 sets in, also metallic absorption is observed. This proved however 

 not to be the case. Long before an electrically condncting layer is 

 formed, a grey colour and metallic absorption are found. 



In tig. 2 the intensities of a lamp withont salt (A) and of one 

 with salt (B) have been represented as fnnctions of the time. The 

 temperatnre of the wire was abont the same in both lamps'). The 

 diameter of the wire had been thns chosen that its decrease by 

 vaporization had no appreciable inflnence on the candle power. The 



lanij) that was nsed is represented in 

 fig. 3. In order to age the lamp it 

 was first bnrned dnring some time 

 with the iniiei" glass cylinder high np, 

 so that no tungsten conid i-each the 

 P cylinder wall. Then the cylinder was 

 lowered again and the intensity measnred 

 with definite intervals with a thermo-pile 

 through a CnCl, filter. In this way much 

 more accurate results could be obtained 

 than by photometric measurements. 

 E'ig. 3. Now it has been proved, that also in 



the lamps treated with salt the formed tungsten precipitate absorbs 

 from the very beginning and that the absorption is nearly proportional 

 to the number of the va|)orized 11 -molecnles. In the lamp without 

 salt the curve shows in the beginning a typical change of inclina- 

 tion, which could be rather well reproduced. 



In these thin layers we have evidently to do with a phenomenon analo- 

 gous to that mentioned by Stakk*), when he speaks of the cai'riers of the 

 spectral lines and assumes electrons that are optically free but elec- 

 trically bound. 



The phenomenon has however to be investigated much more 

 thoroughly before an insight will l)e obtained of what takes place 

 in these thin layers. We still want to mention the following experi- 

 ment. In a lamp immersed in liiiiiid air- a thin layer of chlorine was 



1) In the figure the numbers above the abscissae refer to (B), those below it 

 to (A). 



») Stark, Jahrbuch der Rad. und El. (Uj, 139, 1917. 



