the Spectra of Carbon and Hydrocarbon. 555 



The spectrograms of the spark in alcohol obtained with 

 the larger spectrograph were insufficiently exposed for 

 successful reproduction; but the negatives show clearly 

 a close agreement of the main features with those of the 

 arc-in-fiame, although the fainter lines are not recorded, 



The normal carbon arc shows only faint lines in the region 

 between about A, 4380 and the head of the cyanogen band 

 X4216, and a comparatively long exposure is required to 

 photograph them. 



For this reason lines due to impurities in the carbon rods 

 are very strongly recorded on the plate, even though their 

 actual intensities are not great. The region in question is 

 shown in fig. 11, PL XIII. 



An Investigation of the Fainter Regions of the 

 Hydrocarbon Band. 



In some spectrograms of the hydrocarbon bands obtained 

 with the smaller grating instrument, faint traces of a new 

 series of lines were detected in the region between \ 4107 

 and X4025. The discovery of radiations in this apparently 

 dark interval at the centre, between the main head and tail 

 series, was considered to be of sufficient interest to justify an 

 attempt to explore this region as thoroughly as possible in 

 the hope of connecting the two main parts of the hydro- 

 carbon spectrum. 



The first negatives with the grating spectrograph showed 

 that the main difficulty was the presence of a faint con- 

 tinuous spectrum extending throughout the band, which 

 tended to mask the lines. The latter were so very little brighter 

 than the continuous background on which they appeared 

 that increasing the length of the exposure beyond a certain 

 point did not serve to increase their visibility, but merely made 

 the whole image darker. Fortunately, however, the intensity 

 of the continuous spectrum seemed to be variable with the 

 precise state of the gas-flame, becoming fainter as the supply 

 of gas was reduced and the flame became more silent. 



In order to obtain a spectrum of greater luminosity, and 

 decrease the length of the exposures, it was decided to 

 employ a prismatic spectrograph. The instrument used 

 was provided with a compound prism of high dispersion 

 and a camera of about 9-inches focus. The effective working 

 aperture was fairly large, and the angular dispersion of the 

 prism for the region in question was very considerable. 



Several successful spectrograms were obtained, which 

 showed the new series more distinctly. 



The flame of the Meeker burner was adjusted and carefully 



