DR. W. M. HICKS: A CRITICAL STUDY OF SPECTRAL SERIES. 



399 



between the calculated values and those found by sounding is remarkably close. The 

 only doubtful case may be that for m = 1. If 20559 be taken for this the formula is 

 only slightly changed and the agreement almost as good, the calculated value for 

 m = 3 giving C d\ = '05 instead of "00. It is excluded because it cannot belong to 

 a diffuse series in which the limit is S ( oo). 



TABLE of Parallel XD Series. 



After m = 2 the wave-numbers at the head of each set are those calculated from 

 the formula and are not, as usual, enclosed in [ ]. The links are entered as attached 

 to the respective observed lines. On the right of each is entered the C(dA) 

 calculated as if the error were on the observed line. They would be less often 

 considerably less if calculated on the series line itself. 



This is also 3e.P 3 (1) (8) the 90 A 2 linked F. It is probably not (280 D (2) but hides it. 

 VOL. COXX. A. 3 I 



