OK OUK KNOAVLEDGE OF SPECTKUM ANALYSIS. 265 



Hnggins: 'Phil. Trans.' dviii. p. 558 (18G8). 



Lielegg: 'Wien. Ber.' Ivii. (2) p. 593 (1868). 



Watts: 'Phil. Mag.' xxxviii. p. 249 (1869); xlviii. pp. 309 and 456 



(1874) ; xlix. p. 104 (1875). 

 Wiillner: 'Pog-g. Ann.' cxliv. p. 481 (1872). 

 Salet: 'Ann. Chim. Phys.' xxviii. p. 60 (1873). 

 Angstrom and Thalen : ' Nov. Act. TJps.' ix. (1875). 

 Lockyer: ' Proc. Roy. Soc' xxvii. p. 308 (1878) ; xxx. p. 335 (1880). 

 Liveing and Dewar : 'Proc. Roy. Soc' xxx. pp. 152, 494 (1880). 

 Piazzi Smyth: ' Ast. Obs. Ed.' xiii. (R.) p. 58 (1871) ; 'Phil. Mag.' 



xlix. p. 24 (1875) ; viii. p. 107 (1879). 



Few spectra have given rise to so much controversy as the spectrum 

 of carbon. We shall give an account of the most important experiments 

 which have been made on the subject. 



The Line-spectrum. — This spectrum appears when a very strong spark 

 is sent through carbonic oxide or carbonic acid. It has been observed and 

 described by Watts, Wiillner, Angstrom and Thalen. The best measure- 

 ments seem to be given by the two Swedish observers. Watts gives many 

 lines which are not found in Angstrom and Thalen's map, but it seems 

 possible that the separation of carbon and oxygen lines has only been 

 imperfectly effected by Watts. All observers seem agreed in ascribing 

 this spectrum to carbon. Though Huggins and Watts wei'e only able to 

 obtain this spectrum from carbonic oxide and carbonic acid, Angstrom and 

 Thalen seem to have seen it also in hydrocarbons when they used a large 

 condenser. 



2. The Band-spectrum {Candle-spectrum) . — This is the spectrum which 

 is observed at the base of every candle and gas flame. The controversy 

 on the carbon spectrum chiefly relates to this spectrum, there being a 

 disagreement of opinion whether it is due to the element carbon or to a 

 hj-drocarbon. The spectrum which has first been described by Swan 

 consists of a series of bands apparently fading away towards the blue, but 

 in reality easily resolvable into a series of lines. A good idea of the 

 appearance of this spectrum as it appears in spectroscopes of one prism, 

 is obtained from the drawing given in Lecoq de Boisbaudi'an's Atlas. 

 Angstrom and Thalen and Watts give more detailed drawings and 

 measurements. The spectrum was considered by Swan to be due to a 

 hydrocarbon, but Swan's experiments were only made with gases 

 containing hydrogen. Attfield discussed the question at great length in 

 1862, and came to the conclusion, that the spectrum was really due to 

 carbon. The experiments which were considered crucial by Attfield and 

 the great majority of observers were as follows : — 



1. A flame of cyanogen in oxygen shows, amongst other bands, this 

 spectrum most hrillianthj, after both gases have been carefully dried. 

 Cyanogen burning in air also gives the spectrum, but more faintly. 



2. Sparks taken in the following gases, at atmospheric pressure, 

 carefally prepared and dried, show the spectrum. 



Cyanogen. 



Carbonic oxide. 



Carbon bisulphide. 



These gases have only carbon in common, and, unless the experiments 

 are vitiated by impurities, they prove undoubtedly that the spectrum is due 

 to the element carbon. 



