1088 



condensed on the bulb wall, u|)on this the tungsten was precipitated. 

 At the vaporization of the chlorine, when the lamp had been taken 

 out of the liquid air, the layer was torn in two over a length of 

 some centimeters, while the metal was left on the wall in the form 

 of continuous sheets proving that the distances between the tungsten 

 atoms are sufficiently small that their mutual attraction hold them 

 together. 



Finally we shall give the results that can be reached in the above 

 mentioned way by the application of a salt layer. 



From many experiments we concluded that the mean duration of 

 burning after which the iuleusity was sunk to 80 7c of 't^ original 

 value is for lamps with NaCl about 2,6 times, for lamps with 

 CaFj 3,3 times as long as for lamps without salt. 



Eindhoven. Laboratonj of Philips' 

 Glowlampioorks Ltd. 



