1894.] in the Spectra of Carbon Electrodes, $c. 345 



" It is worthy of remark that Hartley and Adeney, among the 

 carbon lines measured by them, included those with wave-lengths 

 3881-9, 38707, 3589'9, and 3584'8, which, according to the foregoing 

 investigation, are described as the edges of cyanogen bands, and must 

 be struck out of the list of lines in the spectrum of carbon. Liveing 

 and Dewar, in their previous work on the graphite spark spectrum, 

 had not recorded them, but these were first presented as new 

 carbon lines by Hartley and Adeney." 



At the time the observations of Mr. Adeney and myself were made 

 there was no published description of the ultra-violet spectrum of 

 carbon, but before our work received publication, that of Messrs. 

 Liveing and Dewar appeared, and to this we referred (' Roy. Soc. 

 Proc.,' vol. 34, p. 429). The lines we described we still believe to be 

 none other than carbon lines, because of the absence of satisfactory 

 evidence to the contrary. 



The carbon spectrum, as modified by moistening graphite electrodes 

 and surrounding them with gases other than atmospheric air, was 

 closely studied by me and the different spectra were mapped (' Phil. 

 Trans.,' Part I, 1884, p. 49). Not only was the carbon spectrum 

 studied, but an endeavour was made to obtain the " cyanogen bands " 

 from saturated solutions of cyanides. No success attended these 

 experiments, even when on the one hand a stable cyanide such as the 

 potassium salt was used, nor when the most easily decomposed salt, 

 mercuric cyanide, was taken. On this ground there appeared to be 

 no sufficient reason for attributing the lines 3881'9, 38707, 3589'9, 

 and 3584'8 to cyanogen rather than to carbon. On the other hand, it 

 was shown that some of the lines were much strengthened and the 

 general appearance of the first group was altered when concentrated 

 solutions of zinc and calcium chlorides surrounded the electrodes, 

 but it was evident that the modification of the spectrum did not 

 arise from the zinc, calcium, or chlorine, and, therefore, more 

 probably the carbon lines were modified by the presence ox the saline 

 solutions. 



The stronger the solutions the more pronounced were the modifica- 

 tions in the spectra. Eder and Valenta do not appear to have re- 

 ferred to these facts, and it is quite possible they may have 

 overlooked pp. 58 and 59 of my paper, where they arc recorded and 

 discussed. 



I have recently again examined a large number of photographs of 

 spectra of saline solutions and several spectra of various flames 

 taken in the years 1880-81. The graphite electrodes immersed in 

 solutions show beautiful groups of lines which coincide with the 

 edges of certain bands in spectra of the flame of burning cyanogen. 

 These bands can be recognised in the groups III and IV on the 

 spectra photographed by Kayser and Runge. 



