of Magnesium, Copper, and Manganese. 311 



line 5700, the presence of continuous ground and bands 

 which in the mantle are stronger than in the cone has not 

 permitted to decide whether or not this line is emitted by 

 the mantle; for even if the intensity of this line in the 

 mantle were the same as in the cone, it would hardly be 

 possible to detect it for the reason just stated. 



No other line of this element has been observed in the 

 fringe or luminous vapour down to about X 3800. 



When comparing the results given by the single plate 

 furnace with those obtained for the mantle of the air-coal 

 gas flame, allowance must be made for the considerable 

 difference in temperature between the two cases ; for 

 whereas the temperature of the mantle is only about 

 1850° C, that of the protected space near the fringe is 

 probably in the neighbourhood of 2800° C. This great 

 difference may well account for the fact that the visibility of 

 the line 5700 in the protected space is not impeded by con- 

 tinuous ground to the same extent as in the air-coal gas 

 flame. According to my results with the plate furnace, this 

 line belongs most probably to type I., and I do not doubt 

 that if a little oxygen were added to the air-coal gas mixture 

 rhe intensity of the continuous ground would diminish and 

 the line show up plainly. With regard to the remaining 

 lines, the furnace results corroborate entirely Dr. de Watte- 

 ville's observations respecting their character. 



If now we compare the furnace results with those given 

 by a capacity spark in which the emission of lines is pre- 

 dominantly caused by electrical excitation, the striking fact 

 is brought out that the two lines which are most strongly 

 emitted in the latter — namely 5153 and 5218 — are precisely 

 those which in the furnace are excited solely by the therm- 

 electronic current. They are likewise emitted under 

 chemical excitation in the cone, but are absent from the 

 mantle : and the fact that no trace of them has been 

 observed in the luminous vapour beneath the fringe «"oes to 

 indicate that their emission is not stimulated by thermo- 

 chemical excitation up to a high temperature. Thus, also in 

 the case of copper, the spectroscopic effects produced by 

 chemical excitation in the air-coal gas cone and by thermo- 

 electrical excitation in the red fringe are the same. 



PL IX. b represents the spectrum given by copper vapour 

 in the protected space of the single plate furnace. Special 

 attention is invited to the unmistakable contrast between the 

 short lines in the fringe and the long lines in the luminous 

 vapour, which fact so manifestly bears witness to the 

 difference in origin between these two types of lines. 



