552 Mr. C. W. Kaffety on some Investigations of 



Fig. 6, PL XIII., is reproduced from a spectrogram enlarged 

 with a cylindrical lens. 



Examination reveals a close agreement with the spectra 

 of the arc and the spark in the coal-gas flame, and with the 

 spark in an atmosphere of coal-gas. 



In addition, certain conspicuous new bands occur, some of 

 which appear to be identical with members of Deslandres' 

 negative groups *. 



Their chemical origin seems as yet to be unknown. The 

 approximate wave-lengths of the heads of the bands shown 

 in the spectrogram are as follows : — ■ 



XX 4102-5 



4068-0 

 3854-0 



3827-5 (?) 



The bands of cyanogen are completely absent. 



[In connexion with the spectrum of the arc-in-flame, there is a 

 very interesting point to be noticed with regard to the behaviour 

 of lines due to impurities in the carbon rods. Various elements 

 occur as impurities : calcium, iron, sodium, aluminium, silicon, 

 and magnesium are common, and often their spectra are strongly 

 represented. The action of the flame surrounding the arc is to 

 quench out these lines more or less completely, and a good 

 illustration of this is given in fig. 7, PL XIII. 



The two spectra here shown were obtained from the same 

 carbon rods — the upper one being that of the normal open arc, 

 and the lower one that of the arc in a coal-gas flame. In the 

 first, the lines of calcium— XX 4226-9, 3968-6, 3933-8— are very 

 strong; lines of silicon, aluminium, &c. also appear. In the 

 spectrum of the arc-in-flame these lines are either absent or very 

 faint, the most striking case being the practically complete 

 disappearance of the strongest calcium lines, H and K.] 



It was realized that to study in detail the modifications of 

 the hydrocarbon bands, and the other features of these 

 spectra, greater instrumental power was desirable ; for 

 although the optical performance of the smaller grating 

 spectrograph was excellent, an instrument of greater focal 

 length would have given better resolution on the photo- 

 graphic film. 



The spectra were accordingly photographed with a grating 

 spectrograph of 6-f'eet focus. The mounting of this instru- 

 ment leaves much to be desired when long exposures are 

 concerned, as its situation is subject to very considerable 



* See ' Index of Spectra,' W. Marshall Watts, Appendix V., p. 77; 

 Deslandres and D'Azambuja, C. R. 1905, cxl. p. 917. 



