Series System in the Spectrum of Gold. 1& 



In m = 6, the displaced D 12 , D 22 would seem justified by 

 their order of intensities and same displacement — • 8. 



In m = 7, 3604*96 is exactly the calculated X for (SJ D l (7) . 

 But this line is also, as will be seen later, the combination 

 Pi(l) — pi(7). The values coincide. Whether the observed 

 is either or both there is nothing to decide. 



m = 8. The calculated D n gives a better v with D 22 . and 

 a closer value to the mean limit. 



The foregoing discussion supports conclusively : 



(1) The value of the limit Sf<*> ) = D(oo ) is 29469*85* 



±1-5. 



(2) The existence of summation D series. 



(3j The existence of extreme ultra red D lines, de- 

 pending very closely on 9 A as the mantissa of 

 the d 2 sequent. 



(4) A satellite separation about 249 or 258 due to 

 displacements of 27 Sj and 28 o\. 



With this limit and its uncertainty 1*5, and v = '3815'56-\-dv,. 

 dv <•!, 



A = 113935 + 27<fr- 5'46 f - 113935 + x, a?<10, 



8 = 1406-604 + '123..r, 



/(8A) = 30017-06--251#, 



/(9A) = 26734-86--237.i% 



It should be remembered, however, that the value of the 

 oun as determined from the S(oo ) and the doublet or triplet 

 separation appears often to show an indication of being- 

 very slightly smaller than when determined from D or F 

 data, although the latter are very consistent amongst them- 

 selves. It is possibly due, if it is real, to a difference in the 

 value of N between P and D sequences. In any case it is 

 always extremely small. Here the value from D and F data 

 would depend on the D (1) lines, and we have no direct 

 observations to give their values with sufficient exactness to 

 test. So far as evidence is at disposal it must come from the 

 sounding with <?, u,v links given above, and these point to a 

 slightly higher value of the D lines than those calculated. 

 In fact, the following arrangement would fit all the sounding 

 indications, within error limits. The observed lines are 

 supposed to be corrected by amounts less than the possible 



