1068 

 TABEL 2. 



Precipitate mgr. KOH 



on the 1 

 glass wall. 



Exp. nO. 1. Exp. n». 2. Exp. no.3. 

 13.7 



15.7 



26.2 



Exp. nO. 4. 



Not deter- 

 mined 



Exp. nO. 5. 



19.4 



mgr. W I Only qual. 0.3 

 determina- 

 tion with ] 

 pos. result. 



10.3 



10.0 



4.0 



Spherules 



on the 

 filament. 



Only qual. determined, 

 deal of Si02; some CaO. 



Hardly any K.O and Na.O. A great 



Remarks referring to the table. When the lamp is opened (viz. when air is lelt 

 in) the principitate on the wall is found to contain K.O and Na20 and WO3. 

 For the sake of simplicity we have made all our calculations referring to 

 KOH and W. The percentage of "KOH" was determined acidimetricaliy, that 

 of W colorimetrically (Defacqz) and gravimetrically (mercuro-tungstenate). 



The quantity of tungsten found on the wall in these experiments was very 

 different. It depended to a high degree on (increases with) the temperature, 

 to which the wire is raised finally. Especially experiment 2 shows however, 

 that besides the alcaline percentage on the wall, caused by the decompo- 

 sition of the glass by the tungsten, a great part of the alcaline oxide is 

 due to a fractional distillation of the glass on the wall. 



is solved. This experiment, made under very rigid conditions (in 

 vacuo) is even a very useful confirmation of this view. 



A single observation on a similar phenomenon at high temperature has been 

 published by P. Lenard ^). He finds, namely that when potassium-borate is glowed 

 in a lamp, nearly all the potassium is vaporized from the pearl, while the boric 

 acid is left behind. 



Further may be mentioned that our experiments with glass had been finished, before 

 we became acquainted with the paper of E. Anderson and B. J Nkspell ^), in 

 which experiments are described on the vaporization of alcaiineoxides from cement- 

 materials by heating in a Fletcher-stove to + 1300°. 



These experiments are of course very rough and influences of the gases of 

 combustion can play a part, which is also evident to a certain degree from their 

 observations. But for the rest their results can be regarded as a confirmation of 

 what we found. It is remarkable, that they have also made experiments with 

 feldspar, with a substance therefore which compared with the cement materials 



J) P. Lenard, Ann. d. Phys. (4) 17, 204 (1905). 



2) E. Anderson and R. J. Nestell, Journ. Ind. a. Engin. Chem. 9, 253 (1917). 

 See also W. H. Rus:i and A. R. Merx Ibid. 9, 1035 ^1917). 



