ON OUR KNOWLEDGE OF SPECTRUM ANALYSIS. 



125 



shows that the resemblance is only a general one, and that the relative 

 distances vary considerably in each set. 



Table V. 



Tt -will be noticed that the groups come nearer and nearer together as 

 they approach the violet, and that also the lines in each group are the 

 closer together the more refrangible the set. Roughly speaking, the 

 distance between the first and second line of each set is proportional to 

 the square of the wave-length ; in order that this relation ought to hold 

 rigidly, these distances for the ultra-violet sets ought to be respectively 

 7'5, 4-5, 3-9. Such a relation ought to hold if the lines are successive 

 harmonics of one fundamental vibration, according to Stoney's supposition. 



The fact that successive lines which belong to one vibrating system 

 come nearer and nearer together as they approach the violet or ultra- 

 violet end of the spectrum seems to be a pretty general one, and is well 

 exemplified by the system of hydrogen lines which Huggins found in the 

 star spectra. Table VI. gives the wave-lengths of the lines and their 

 differences. 



Huggins calls the line Hf , a and continues with tlie alphabet towards tbe ultra-violet. This designation 

 was chosen independently of the fact that the lines probably belonged to hydrogen. As the red hydrogen 

 line usually is called Ha, we have continued the same nomenclature towards the ultra-violet. Hence the 

 discrepancy with Huggins' designation. 



It has been pointed out by Johnstone Stoney that the second dif- 

 ferences sho".v greater irregularities than can be accounted for by errors 

 of observation, and that therefore the system of lines does not altogether 

 lie on a smooth curve when plotted down with wave-lengths as ordinates, 

 the ordinates lying at equal intervals of each other. Nevertheless the 

 fact that these lines approach each other rapidly, making up a fluting 

 on a large scale, and that generally characteristic groups when repeated 



