J 074 



physical character. When the tungsten molecules are kept apart 

 from each other by the molecules of another substance, the strong- 

 absorption of the light by a continuous tungsten layer can be 

 prevented. 



This action is also shown by the "glass" on the wall. While the 

 smooth bulb-wall itself is evidently inactive, the "porous glass" 

 coming from the incandescent body proves to be able to envelop 

 the tungsten particles that are shot into it. When however the lamp 

 has been opened, so that the porous glass can crystallize, it has 

 lost much of its activity. 



^ 4. Aluuiinates, phosphates, oxides: 



It has been remarked already several times (comp. e.g. Chem. 

 Weekblad 13. 536 and seq. 19 J 6), that part of the troublesome 

 gases in a glow-lamp are only liberated when at the first use of 

 the lamp the wire has reached a high temperature. For this reason 

 it is desirable to use substances which at that high temperature 

 are vaporised only relatively slowly from the wire. In our experi- 

 ments we have found such materials in, silicates, several kinds of 

 glass, feld-spar esc. But besides by silicates this condition is also 

 fidpUed hi/ a/timinafcs, as e. g. magnesium-aluminate, further also 

 by Cu,(POjj (when not hydrolyzed) and other similar compounds. 



Our first experiments on magnesium-sulfate were made with spinel, 

 which we obtained from a well-known firm. The abo\e mentioned 

 property was found again. The analysis showed however, that the 

 composition was: 



SiO, 39,37,, Al,(), 22,5V„. b^O, 6,3Vo, CaO 34,37., MgO 0,27.. 



It could therefoi-e not be called sjiinel. Still this experiment was 

 very instructive. It showed namely that the presence of alcaline- 

 oxide was not absolutely necessary to obtain the effect. Later on we 

 will return to this. 



The natural spinel that was at our disposition being thus no true 

 spinel, we made artificially a compound by heating MgO and Al^O, 

 to 1800°. The product really showed the action discussed above. We 

 also tried other substances in this respect. When the temperature is 

 only sufficiently high and the layer on the incandescent body suffi- 

 ciently thin it is evident that all foreign substance on the wire 

 can be vaporized. Thus the CaO — when such a silicate is used — 

 from the glass will be sublimated on the wall under these circum- 

 stances. And then it is of course probable, that also some free Ca- 



