Hydrogen and some of its Compounds. 377 



Apparently these are the same lines found in the tubes 

 provided with carbon electrodes ; and also in tubes with 

 metallic electrodes which are filled with carbon compounds. 

 Observed visually, with a straight-vision spectroscope, all the 

 above cases appear identical. When, for instance, hydrogen 

 was put into the tube with carbon terminals and submitted 

 to discharges from an induction-coil, at first the line-spectrum 

 of hydrogen appeared. After the discharge had passed for 

 some time, this gradually changed into the characteristic 

 band -spectrum of carbon. To the eye alone the change was 

 equally noticeable ; the light being, at first, reddish, and 

 then chano-ino- to a white. Similar changes were noticed 

 when nitrogen and oxygen were used in the carbon tube. 



It seems to me that the following conclusions can be drawn 

 provisionally from the above : when various elementary gases 

 are introduced into wide tubes with carbon electrodes, and 

 exhausted to a pressure of 1-2 mm., and submitted to con- 

 denser-discharges, compounds of carbon with the various 

 gases are formed. With nitrogen this compound is probably 

 cyanogen ; with hydrogen, acetylene ; but when a photo- 

 graph of the spectrum in each case is taken, we get not the 

 spectrum of the compound nor that of the elementary gas, 

 but a carbon spectrum. This, however, does not mean that 

 we get the line-spectrum of elementary carbon ; for it is 

 certain that there is water-vapour present in the tubes, not- 

 withstanding the temperature to which it has been subjected. 

 The carbon may then unite with the oxygen of the water- 

 vapour, forming either carbon monoxide or dioxide; the 

 hydrogen being occluded by the terminals or the glass walls. 



Just as the spectra of gaseous carbon compounds in wide 

 tubes with metallic terminals appear identical with the spectra 

 of elementary gases in tubes with carbon electrodes submitted 

 to condenser-discharges, so we should expect that the spectra 

 produced would be the same in the two cases. This is found 

 to be true. The general appearance of the photographs 

 obtained with continuous currents is very different from those 

 obtained with condenser-discharges. In the former case 

 there is a marked band-appearance in addition to a line- 

 spectrum. The most prominent of these bands, in the region 

 studied, is the one beginning at 3884, fig. 2 (PI. VI.). With the 

 dispersion used this band consists of five prominent lines 

 crowding together toward the ultra-violet A somewhat 

 similar band, apparently of six lines, begins at 4216. Another 

 band, rather faint, consisting of a large number of fine lines, 

 shading off toward the ultra-violet, begins at 4126. The 

 bands beginning at 4216 and 3884 appear to be same as the 



