374 Prof. J. Trowbridge on the Spectra of 



point to the phenomena already described. According to 

 this hypothesis the rarefied water-vapour is dissociated at the 

 surface of the anticathode, which is thus greatly heated ; the 

 occluded hydrogen plays a part in this phenomenon. 



The behaviour of large aluminium electrodes in glass vessels 

 filled with ammonia-gas is also an interesting example of the 

 dissociation of water-vapour. The gas was obtained by 

 heating ammonia chloride, passing it over freshly slaked lime 

 and through drying-tubes filled with phosphoric pentoxide. 

 A sufficient amount of ammonia-gas was thus obtained for the 

 purposes of spectrum analysis. 



When a large condenser charged to a difference of potential 

 of twenty thousand volts was discharged through the rarefied 

 ammonia-gas, there being practically no self-induction in the 

 circuit, and the main effect therefore was due to the pilot 

 discharge, the light of the tube changed from a brilliant white 

 to a rosy red, and eye-inspection with a straight-vision spec- 

 troscope showed only the line-spectrum of hydrogen. One 

 would conclude from this inspection alone that there was pure 

 hydrogen in the tube. One might also surmise that the 

 oxygen of the water- vapour always present on the walls of the 

 glass vessel had combined with the aluminium terminals, 

 setting free the hydrogen which then carried the current. 

 The pressure, however, in the tube increased : and therefore 

 gas must have come from the aluminium. In the exhaustion 

 of j?-ray tubes provided with aluminium cathodes much time 

 and long treatment with condenser discharges is necessary to 

 drive out the gases from this cathode. The principal gas 

 seems, from the experiment with ammonia-gas, to be oxygen. 

 The same phenomenon is seenin tubes supplied with magnesium 

 terminals, but to a much less extent. It is not seen when the 

 terminals are of copper, iron, silver, platinum, or carbon. This 

 behaviour of aluminium toward oxygen is very suggestive 

 in regard to the ready passage of the <2?-rays through this 

 metal. 



I have been unable, with the conditions under which I have 

 worked, namely the use of very powerful discharges, to obtain 

 the spectra of hydrogen apart from water-vapour and hydro- 

 carbons. The study, therefore, of the spectra of hydrogen 

 compels one to study carefully the spectra of the hydrocarbons 

 and that of cyanogen ; for I am forced to the conclusion that 

 the combination of hydrogen with oxygen is a controlling 

 factor in all discharges through rarefied gases. The following 

 is a preliminary study of some of these compounds, which is 

 added at this stage of my inquiry to illustrate this theory. 

 When various gases are put in tubes provided with carbon 



