1071 



some WOs will have been left behind in the glass on the wire, probably also a 

 trace of Ca. The alkali-metal expressed as "KOH" corresponds to 2,0 mg KOH. 

 In fact after opening of the lamp 3,5 mg KOH was found on the wall. This 

 surplus is due to the fractional distillation of the glass. 



Experiment 2. In this experiment the vaporization of the glass was executed 

 more strongly, so that the wall b^'came very hot, while a great part of the 

 formed alcalimetal was solved again colloidally in the condensate on the wall 

 (which was then coloured brown). At the opening of the water-bulb momenla- 

 neous decoloration was found, while relativeh'^ much hydrogen was proved to 

 have been developed Results of the analysis: 38,2 mg *KOH", 5,3 mg "W". 

 To this corresponds 1.9 cm^ H^,. 



E. Reaction betiDcen tungsten and glass at 1000°. 



From the following experiment we see that no very higli tempe- 

 rature is necessary to show a reaction between tungsten and glass. 

 When tungsten powder and glass j)owder (e.g. Thüringian glass) 

 are mixed and then heated in an atmosphere of an indifferent gas 

 to 1050° during one hour, we find, that the i-eaction has taken place. 



In this way 0,65 mg WO3 (determined by the mercuro-tungstenate-methode) 

 was formed from 1 1 gr. tungsten and a few grams of glass. 



From another point of view still this action between tungsten and 

 glass is remarkable. Let us namely consider the case, that we wish 

 to seal electric conducting wires air-tight into an apparatus of hard- 

 glass. F'or this purpose we u)ust take a metal with a small dilatation 

 coelTicient, such a metal is the tungsten. During the sealing- in this is 

 coxered however with a layer of oxide and when this remained in 

 this state it would prevent, as has been found experimentally, the 

 connection between glass and wire from becoming air-tight. A( this 

 temperature however tungsten and the lower tungsten-oxides act on 

 the glass; the oxides do not remain therefore oxides and without any 

 further precautions a very sufficient shntling-off is obtained (Comp. 

 also E. Weintraub). ') 



F. Can the action, of the silicates be replaced hi/ alcali-oxides? 

 The reaction between tungslen and the glass capillary in the lamp 



is principally based upon the transformation : 



fv + z /'•" = iro.+a f . 



A a, O Na 



When a lamp is made, however, in which the filament is covered 

 not with glass, but with Na,!) ^) we have practically found nothing 

 of such a reaction. 



1) E. Weintbaub, U. S. patent 1. 154. 081. 



2) Comp. L. Hamburger, Gliem. Weekbl. 13, 516 (1916). 



