the Spectra of Metallic Vapours. 339 



It consists simply in directing the spectroscope upon a frag- 

 ment of sodium burning in oxygen gas. The incandescent metal 

 gives, of course, a continuous spectrum ; but the rays, in passing 

 through the cooler atmosphere of sodium-vapour which surrounds 

 the metallic nucleus, are selectively absorbed, and the dark 

 double line D, or Na «, appears with great distinctness on the 

 bright spectrum. As, however, the fragment of sodium is soon 

 consumed, I have used an apparatus resembling that employed 

 for making phosphoric anhydride, by which pellets of sodium 

 may be added as often as required. 



A moderate-sized deflagrating-jar is placed in a dish of sand. 

 In it is suspended a shallow iron cup, and through the same cap 

 which carries the latter is passed a short wide glass tube, so as to 

 be directly over the centre of the cup. Through the sand and 

 under the lower edge of the jar passes a bent glass tube, con- 

 nected with a caoutchouc bag of oxygen, and serving to intro- 

 duce a continuous slow stream of the gas to replace that con- 

 sumed by the sodium. The spectroscope should first be adjusted 

 as to position and focus by bringing it to bear on a candle placed 

 on the opposite side of the jar, so that its iiame may just be seen 

 over the edge of the iron cup. Then, while the jar is filling 

 with gas, the cup may be withdrawn, a pellet of sodium placed 

 in it and heated over a spirit-lamp until it begins to burn, and 

 lastly immersed in the jar. Fragments of sodium may be added 

 as required through the glass tube, and will readily burn if the 

 mass of soda in the cup be not allowed to cool below dark red- 

 ness. The experiment may thus be carried on as long as de- 

 sired; and, of course, two or more spectroscopes may be arranged 

 round a single deflagrating jar. 



I have not yet tried whether the spectrum thus produced can 

 be thrown on a screen, the deflagrating-jar being enclosed in a 

 Duboscq's lantern ; but, from its brilliancy in the spectroscope, I 

 have little doubt that it could be thus shown. The idea of this 

 method occurred to me about a year ago, and I have shown it to 

 many in Oxford ; but it seemed so obvious an expedient that I 

 thought it must have been already described. As, however, I 

 have been unable to find any account of it, I venture now at any 

 rate to bear my testimony to its efficacy. 



The same method is of course available in the case of other 

 volatile andoxidizable metals, as lithium, zinc, magnesium. 



I remain, 



Queen's College, Oxford. Yours, &c, 



H. G. Madan. 



