the Spectra of Carbon and Hydrocarbon. 551 



intensities can be changed considerably by an alteration o£ 

 the length of the spark, and by other variations of the con- 

 ditions of discharge. 



When sufficient self-induction is introduced into the 

 oscillatory circuit, the line spectra of nitrogen and carbon 

 disappear, leaving a bright " Swan " spectrum with " the 

 three," the hydrocarbon bands, and those of cyanogen. 



The spectrum then becomes practically identical with that 

 of the arc-in-flame. 



A spectrum very closely similar is given also by the 

 spark with self-induction in an atmosphere of coal-gas, 

 the survival of the cyanogen bands being attributable, 

 presumably, to an admixture of some nitrogen with the 

 gas. 



The spectrum was obtained from the transformer spark 

 between carbon electrodes in a glass cylinder through which 

 a gentle stream of coal-gas was maintained, the gas being 

 brought into the sparking cylinder through the electrodes 

 themselves, which were drilled so as to form tubes. 



Here again, with the introduction of self-induction into 

 the discharge circuit, the line spectrum of carbon gives 

 place to the " Swan " bands. The spectrum from about 

 \ 4550 to X 3800 is reproduced in fig. 5, PL XIII. 



The modification of these spectra from that given by the 

 Meeker flame suggested the investigation of the spark under 

 alcohol, in order to ascertain whether or not the same new 

 lines were present in this case. The transformer discharge 

 was accordingly taken between electrodes of platinum under 

 absolute alcohol. With condenser in circuit the discharge 

 was intermittent and violently disruptive, giving a bright 

 continuous spectrum ; but when the capacity was discon- 

 nected, the spark became much easier to maintain and 

 appeared of a pronounced greenish colour. 



An examination of the spectrum revealed the " Swan " 

 bands brightly. Owing to the rapid liberation of solid 

 carbon in the liquid, photographic exposures were attended 

 with considerable difficulty, as the alcohol very soon became 

 turbid and semi-opaque. By using a fairly large quantity, 

 however, and by bringing the electrodes close up against the 

 side of the glass containing vessel in order to reduce the 

 absorption by the liquid, it became possible to give the long 

 exposures necessary. 



The spark needed constant adjustment which added very 

 considerably to the difficulty of obtaining good spectrograms. 

 The image on the slit of the spectrograph was apt to wander, 

 and required to be brought back on the same spot repeatedly. 



