Iron Vapour in Air- Coal Gas Flame. 241 



With reference to the cone spectrum of iron, with which 

 this research is more specially concerned, no definite know- 

 ledge has been gained as to whether this spectrum is caused 

 by the formation or by the subsequent decomposition of the 

 nitride. Thus, it may be that the emission of the cone lines 

 is the result of a disruptive break-up of the molecule of the 

 iron compound which is fed into the flame (chloride, oxide, 

 &c). Such a break-up could be conceived to be brought 

 about, or at least facilitated, by the existence of a strong 

 chemical affinity between the nitrogen and the metal 

 atom. 



Further, all experimental evidence tends to show that in 

 the Bunsen cone the iron molecules are dissociated and caused 

 to radiate vigorously at a lower temperature than they would 

 do if subjected to thermal actions alone. In this latter 

 case, as has been established, the resulting spectrum is com- 

 posed of temperature lines only (mantles of air-coal gas, 

 oxy-coal gas, &c, flames, § 2). Hence, if it could be shown 

 that the decomposition of the nitride of iron takes place only 

 after it has entered the mantle, then the cone lines would be 

 caused by the process involved in the formation of the nitride. 

 But, as already suggested (§ 5), the nitride formed may be 

 quite unstable and its decomposition may take place whilst 

 it is still within the explosion region, and in that case it 

 would, of course, be extremely difficult to decide to which 

 stage in the transition processes concerned the cause of the 

 cone lines is to be attributed. 



Similar difficulties are encountered when we approach the 

 question as to the origin of band spectra. Are they the result 

 of molecular vibrations of a compound, as is generally as- 

 sumed, or are they due to some peculiar atomic vibration 

 caused by some special mode of chemical combination or 

 decomposition? Further, it may be asked with reference 

 to the appearance of the so-called cyanogen bands in flames 

 (§ 7) : How is it that we do not always observe these bands 

 in the air-coal gas flame? Is it necessary that the nitrogen 

 atoms which are responsible for the emission of these bands 

 in the flame should first pass through some special kind of 

 process, such as might be involved in their liberation from 

 the hydrogen atoms ? 



With regard to the origin of Class I. lines of iron, which 

 have so far been considered as due to thermal actions, it is, 

 of course, quite conceivable that some or even all of these 

 lines are in reality caused by chemical actions of quite a 

 different nature. The proposition that chemical actions, 



