Chase.l 276 TApril 21, 



227. Identity of Spectral Lines in Different Elements. 

 Young {Am. Jour. Sci., xx, 355) and Liveing and Dewar {Proc. Roy. 

 Sac, xxxii, 225-31) have shown that many of the lines in different element- 

 ary spectra, which have been supposed to be identical, really differ slight- 

 ly in refrangibillty and can be separated by a sufficient increase of dis- 

 persive power in the trains of prisms. The number of separations which 

 has already been effected makes it very doubtful whether any case of ab- 

 solute coincidence can be found, where two elements are present in the 

 spectral incandescence. This has been thought, by some, fatal to Lock- 

 yer's and Thalen's hypothesis that all the lines are modifications of a few 

 basic lines. That such a generalization is too hasty, may be shown by 

 the following considerations : 1. Atoms are continually subject to in- 

 commensurable, as well as to commensurable tendencies. 2. There are 

 often various harmonic tendencies, which are simultaneously operative, the 

 final harmonic adjustment being determined "by the relative magnitude 

 of the individual tendencies. 3. The well-known experiment of oscillat- 

 ing balls, suspended from a horizontal cord, shows that the cyclical vibra- 

 tions are modified by each member of a harmonic group. 4. The slight 

 fluctuations in the lines of the solar spectrum make it probable that there 

 are similar fluctuations in chemical and cometary spectra. 5. This proba- 

 bility is increased by the differences of measurement which are made by 

 different observers at different times. 6. Propositions 2 and 5 are both 

 illustrated by the two harmonies which represent Tacchini's and ThoUon's 

 measurements (Note 226). 



228. Lithium Harmonies. 

 Liveing and Dewar {Proc. Boy. Soc.^ xxx, 93-9) have observed three 

 lines in the spectrum of lithium (3913, 3984 and 4273), besides Boisbau- 

 dran's line, 4131.7. The harmonies are shown below. 

 Harmonic Divisors. Harmonic Quotients. Observed. 



1 4273.02 4273 



1 + 7 ^. 4132.78 4131.7 



1 4- 15 ^ 3983.37 3984 



1 + 19 rt 3912.65 3913 



The coefficient of the first addition to the harmonic divisor is the same 

 jis the perissad divisor and as Front's coefficient of Li. The second and 

 third additions are respectively the artiad divisor and h the artiad divisor. 

 The harmony'is nearly as satisfactory, if we combine these lines with those 

 which are given by Huggins (see Proc. Am. Ph. Soc, xvii, 297). 

 Harmonic Divisors. Harmonic Quotients. Observed. 



1 6107.37 6107.3 



l + 40« 4796.64 4794.8 



1 -f 48 ffi 4599.23 4599.3 



1 + 63 a 4269.74 4273. 



l-f70a 4131.49 4131.7 



l-+-78a 3984.31 3984. 



] -f 82 a .^914.50 391 S. 



