1880.] an Induction Coil with a De Meritens Machine. 175 



The spectrum of the secondary spark was then examined with 

 terminals of various metals. 



Aluminium. — The spectrum showed a faint continuous background 

 with the yellow sodium lines, and faint oxide of alnminium lines. 

 This was with a spark of half an inch. But although the spark was 

 subsequently lengthened, no difference in the spectrum was perceived 

 excepting that the continuous background was rendered more 

 bright. 



It would seem that these appearances are due to some such process 

 as the following : — The heat due to oxidation, added to that of the 

 discharge, is sufficient to volatilise the oxide of aluminium ; but that in 

 its passage across the interval between the terminals, the oxide becomes 

 so cooled that it gives a continuous spectrum. When the spark was 

 lengthened, the oxide, although perhaps at first more heated than 

 with the shorter spark, had more time to cool. 



Magnesium.. — In this, as in the former case, we have a faint con- 

 tinuous spectrum as a background, on which were seen the b group of 

 magnesium lines. One other line in the blue occasionally flashed out, 

 but was not permanently present. There was also a faint trace of the 

 oxide spectrum. The contrast between the cases of aluminium and 

 magnesium, in respect of the prominence of the oxide, or of the true 

 metallic spectrum, is doubtless due to the fact that in the former case 

 the oxide, and in the latter the metal, is the one which is more easily 

 vaporised. On sending a blast of air on the discharge, the blue line 

 .always disappeared ; the current of air having lowered the temperature 

 so far as to prevent the vaporisation necessary for its production. 



When the spark between magnesium terminals was made to pass 

 through hydrogen, the characteristic lines of hydrogen were seen, 

 apparently owing to a rise in temperature. This, as mentioned below, 

 does not occur with carbon poles. 



Platinum. — With terminals of this metal the spectrum was mainly 

 continuous, with the addition of the ubiquitous yellow sodium lines. 

 When the spark was short a few bands were faintly visible, some 

 apparently those of nitrogen, and others in the blue and violet belong- 

 ing to the oxide of platinum. When the spark was lengthened, these 

 bands disappeared, and nothing but the continuous spectrum (with the 

 D lines) was visible. 



It appears from these experiments that the application of the 

 de Meritens machine to the induction coil furnishes us with the 

 means of isolating certain lines of the metallic spectrum from the rest. 

 It has, in fact, enabled us to reduce at pleasure the spectra of aluminium, 

 of magnesium, and of platinum, to their most persistent lines ; precisely 

 ;as had already been noticed as occurring by natural processes in the 

 cases of sodium and of calcium. As a general rule, when the spark is 

 shortened, the metallic, or the oxide, lines come out, according as the 



