12 Mr. G. A. Hemsalech on the Relative Behaviour 



are likewise emitted by the flame is the abruptness in the 

 decrease of intensity on passing from the region of chemical 

 to that of thermal actions, as is so well illustrated by the 

 triplets at 4046 and 4384. 



As the most prominent example of the lines of our supple- 

 mentary spectrum, may be cited the group at 4957. These 

 lines show absolutely no trace in the flame. 



Relative behaviour of flame and cone lines in high- 

 temperature flames. 



As M. de Watteville and myself have shown, all the flame 

 lines increase in intensity on passing to the oxy-coal gas,, 

 oxy-hydrogen, and oxy-acetylene flames, whereas the lines 

 of the supplementary spectrum appear only as traces in the 

 last-named flame. Those of the cone lines which show 

 feebly in the mantle of the air-coal gas flame are likewise 

 enhanced in the high-temperature flames. 



The sensitiveness to temperature changes is, however, not 

 the same for all the lines, some lines being more enhanced 

 than others. Thus the lines 3720, 3737, and 3746, which 

 are relatively feeble in the air-coal gas flame, appear rela- 

 tively bright in the oxy-acetylene flame. The characteristic 

 features of a high-temperature flame-spectrum are showm. in 

 the reproduction of a portion of the oxy-acetylene spectrum 

 on Plate I. fig. 3 (no. 1) which should be compared w T ith the 

 air-coal gas flame and cone spectra on Plate I. fig. 1. 



Comparison of flame- spectrum ivith that of self-induction 

 spark. 



On Plate I. fig 3 (no. 2), the spectrum of the self-induction 

 spark is placed in juxtaposition with that of the oxy-acetylene 

 flame. It will be seen that the flame lines, which are so 

 intense in No. 1, are relatively feeble in No. 2 ; examples 

 are the lines at 3720, 3860, and 3920. On the other hand, 

 the cone lines which are relatively feeble in .No. 1, such as 

 3816 and 4046, are relatively strong in the spark. As 

 M. de Watteville had already observed, the spectrum given 

 by the self-induction spark, indeed, approaches that of the 

 air-coal gas cone. There are, however, certain lines which 

 are enhanced in the self-induction spark, and these may be 

 assumed to be specially sensitive to electric actions. 



It thus seems that the chemical actions, to which the iron 

 vapour is subjected in the explosion region of the air-coal 



