1082 



opening sucli a lamp. ') When namely tlie moist air comes into the 

 bnlb the salt is seen to crystallize on the wall. When therefore the 

 decolorating action is due to a development ol' a halogen this thin 

 salt layer must, during the burning of the lamp continually develop 

 small quantities of chlorine. Neither by heat, nor by the action of 

 light sodium chloride develops chlorine. From the investigations of 

 Goldstein'), and WiedeiMann and Schmidt') it was known however, 

 that NaCl is intluenced by cathode rays. We have therefore considered 

 the possibility that the electrons, emitted by the incandescent wire, 

 miffht have a similar action. This was proved not to be the case. 



When electrons of high velocity impinge upon a NaCl layer, 

 a coloration of the salt is seen; brown or violet according to the 

 velocity. The brown salt is decolorated again by the action of the 

 day-light, the violet again is light-proof. Goldstein assumes now 

 that this coloration is due to a physical action, while Wiede- 

 mann and Schmidt came to the conclusion that the salt is decom- 

 posed and that then the formed sodium or sodium sub-chloride is 

 solved colloidally in the surplus of salt. Only in the latter case there 

 can be question of the formation of a halogen atmosphere. A closer 

 investigation taught us that in the case of the brown salt no chlorine 

 had been developed from the solid phase; it acted neutrally; in the 

 case of the violet salt on the contrary a decomposition has taken 

 place. 



As it has now been proved that jVaC/can be decomposed by an impact 

 of electrons we have put the question what velocity of the elec- 

 trons would be necessary for this. Hypothetically we have assumed 

 that the electrons would possess this energy, when their kinetic 

 energy is greater than the heat of formation of the salt. For NaCl 

 this heat amounts to 97,6 kg. cal. per Gram. mol. and thus to 

 0,676 X 10—1^ erg for a molecule. In this case electrons with a 

 velocity above 1,23X10^ cm/sec. would be able to bring about a 

 decomposition. Electrons with such high velocities really occur in 

 the lamp. We were able to prove this by means of a lamp with a 

 side tube which contained a Faraday cage. At the normal burning 

 temperature the charge was always — 5.7 Volt with respect to the 

 negative pole of the filament. This proved that electron velocities of 

 1.40 X 10^ cm 'sec. really occur. 



') In these experiments the thickness of the layer was about 1 — 2 X 10~^ cm., 

 viz. about 50 molecules. 



«) E. Goldstein, Wied. Ann. (54), 371, 1895 and (60), 491, 1897. 



8) E. Wiedemann and G. G. Schmidt, Wied. Ann. (54), 604, 1895; (56), 201, 

 1895 and (64), 78, 1898. 



