2 JVIr. G. A. Hemsalech on the Relative Behaviour 



is desired to obtain, are enclosed in a glass bulb, through 

 which is led the air or oxygen circuit feeding a coal-gas, 

 hydrogen, or acetylene flame. The particles produced by the 

 sparking are generally chemical compounds, mostly oxides 

 or nitrates, and, as de Broglie has shown *, of ultra-micro- 

 scopic dimensions. They therefore remain suspended in the 

 gas for a considerable time and are carried by it into the 

 flame, where they are vaporized. 



Applying this and other methods to a study of the line- 

 spectra of iron and calcium we obtained the following general 

 and fundamental results t : — 



1. The character of the flame-spectrum of a substance 

 depends solely on the nature of the flame, and is not in the 

 least influenced by the electrical or other processes used for 

 preparing the material which is mixed with the gas. 



2. In the case of solutions it is immaterial which salt of 

 the element is used. 



3. The light radiations- from vapours in the outer mantles 

 of the various flames behave as if governed by the tem- 

 peratures of these flames. 



4. The spectrum emitted by metal vapours in the region of 

 the blue cone of the air-coal gas flame cannot be accounted 

 for by simple thermal actions, but is probably due to some 

 special chemical process prevailing in the explosion region. 



In the case of iron we established the existence of two 

 distinct groups of lines. One group, some 200 lines, is 

 emitted by the outer mantle of the air-coal gas flame, and 

 the lines merely increase in intensity on passing to flames of 

 higher temperature. We considered them as the fundamental 

 lines of the spectrum, which are easily excited by thermal 

 actions. The other group, of over 500 lines, is emitted by the 

 region of the blue cone only, and even in the oxy-acetylene 

 flame we could detect only traces of some of the lines. 



The object of the present research is 

 a. To investigate the relative behaviour of the various types 



of iron lines by means of new observational and 



experimental methods. 

 h. To gain some information with regard to the origin of the 



cone lines. 

 c. To extend the observations to the spectra of the closely 



related elements cobalt and nickel. 

 The present paper deals only with the subject-matter 

 mentioned under a. 



* de Broglie, Comptes Rendus, vol. clxvi. p. 624 (1908). 



t Hemsalech and de YVatteville. Comptes Rendm de VAcademie des 

 Sciences, vol. cxlv. p. 1266 (1907) ; vol. clxvi. p. 748 (1908); ibid. p. 859; 

 ibid. p. 962 ; ibid. p. 1889; vol. cxlvii. p. 188 (1908) ; vol. cxlix. p. 1112 

 (1909) ; ibid. p. 1369: vol. cl. p. 329 Q910). 



