the Spectra of Carbon and Hydrocarbon. 553 



vibration with corresponding detriment to the best definition, 

 and it is also liable to temperature changes to some extent. 

 Under these disadvantageous conditions the recording o£ 

 the Meeker flame spectrum, necessitating exposures of many 

 hours' duration, was attended with considerable difficulty ; 

 and even greater difficulties were experienced in the case of 

 the spark under alcohol. 



The results obtained, however, were fairly satisfactory, espe- 

 cially in the case of the arc-in-flame, where the intensity of the 

 light rendered short exposures possible, and the detrimental 

 effects of intermittent vibration and temperature changes 

 could be avoided. 



The spectrum of iron in the carbon arc was employed as a 

 reference spectrum for wave-length determinations. 



The Meeker burner was placed in a horizontal position as 

 before, the luminous region immediately above the grid being 

 viewed from one side. The image thus formed on the slit by 

 the condensing lens was a narrow line of light, and fairly 

 uniform illumination was obtained over the entire width of 

 the spectrum. 



Fig. 8, PI. XIII., is reproduced from a spectrogram taken 

 with the larger instrument, and shows the main portion of 

 the hydrocarbon spectrum as given by the Meeker coal-gas 

 flame, with Fe : comparison. The very line and complicated 

 structure of the head of the main band (about X 4314) is 

 here revealed, although the densest region is still beyond the 

 resolving-power of the instrument. 



It is to be noted that in this case " the three " (\\ 4382, 

 4371, 4365) are inconspicuous. 



The same region of the spectrum of the arc-in-flame is 

 shown in fig. 9, PI. XIII. Seen on the scale given by 

 the increased instrumental power, the wealth of detail is 

 remarkable and the differences from the flame spectrum 

 become the more apparent. 



The greatly increased intensity of " the three" is seen to 

 be due in part to the appearance of new r components ; while 

 the flame-line of X 4324 has now become a group with 

 a new isolated line on the more refrangible side. Photo- 

 graphs of this region were obtained also in the 2nd order 

 spectrum of the grating ; one enlargement is reproduced in 

 fig. 10, PI. XIII. 



Examination reveals that, in addition to the appearance of 



new lines in the arc-in-flame spectrum, there are differences 



in relative intensities as compared with the flame spectrum. 



The flame shows two strong lines— X\ 4292*2, 4291*3; 



but in the spectrum of the arc-in-flame, the less refrangible 



