214 W. Gibbs on the wave lengths 
omer Royal, lends to whatever comes from his pen a more than F 
ordinary interest and value. Those therefore who like myself 
have felt convinced that an explanation of the physical cause 
of the spectral lines of the elements could only be sought in 
lengths and the numbers upon Kirchhoff’s scale can be ex 
pressed for the whole scale from C to G by a single function 
% 
of the form 
a =a-+-bkh-+-ch?+-dk3+ek 
he forms five equations, the solution of which by the method | 
of least squares gives the numerical values of the constants 4, 
b, c, dande, By means of these constants Mr. Airy constructs > 
safe in practice. After the completion of this part of his | 
work, Mr. Airy became acquainted with Angstrém’s memol, — 
drawn and the corrections for eve 0-01 of k were obtained 
Fephically. Finally, the corrections for all the individual : 
ines were interpolated numerically and applied separately © 
each ee wave length. 4 
= 1s It appears that out of the 107 absolute measure 
. gheate of wave lengths made by Ditscheiner, and ‘by him refer 
ee. “k= ryouKl and f, =100000000%FrL a 
