of Magnesium, Copper, and Manganese. 313 



The following facts may be derived from these results : — 

 All lines which in the air-coal gas flame are equally strong 

 in cone and mantle, and therefore mark their character as 

 temperature lines, are of the " long " type in the furnace — 

 that is to say, their emission in the latter is, just as in the 

 flame, controlled by temperature. Again, all the lines of 

 the " tadpole " type which in the furnace are particularly 

 affected by thermo-electrical excitation are in the coal-gas 

 flame emitted solely by chemical excitation in the explosion 

 region. The three apparent exceptions — namely 4041, 

 4049, and 4070 — which have been classed as long lines in 

 the furnace emission, are probably also of the tadpole type, 

 as may be expected from their relative intensities in fringe 

 and luminous vapour ; they are perhaps a little more 

 sensitive to temperature than most other lines of this type, 

 and would therefore occupy a position intermediate between 

 long lines and tadpole lines. Lastly, nearly all the tadpole 

 lines are enhanced in the spark as compared with the arc. 

 Special interest is attached to the line 4235, which is a 

 prominent spark line. Thermo-electrical excitation in the 

 fringe brings it out with intensity 2, whereas under almost 

 similar conditions of temperature just outside the fringe, 

 thermo-chemical excitation hardly stimulates its emission. 

 Naturallv it is absent from the mantle of the air-coal gas 

 flame, but it is well brought out in the cone. Thus we find 

 also in the case of manganese a very close agreement 

 between the emission caused by chemical excitation in the 

 cone of the air-coal gas flame and that due to thermo- 

 electrical excitation in the red fringe of a single plate 

 furnace. 



PL IX. c shows the enhancement in the red fringe of the 

 three blue lines of manganese. 



§ 9. Summary. 



1. With an electrically heated plate of graphite, which 

 is not covered with a layer of carborundum, the red fringe 

 appears on both side?. The upper fringe will often extend 

 to a greater distance from the plate than the lower one. 

 This is probably caused by the upward rush of conducting 

 vapours which are expelled from the under surface and pass 

 round the sides of the plate into the space above. §2. 



2. In spite of the rapid upward motion of conducting 

 vapours above the plate, the border line of the upper fringe 

 is sharply defined and its spectrum emission stops abruptly 

 in marked contrast to that of the luminous vapour. §2. 



Phil. Mag. S. 6. Vol. 40. No. 237. Sept. 1920. Y - 



