Series System in the Spectrum of Gold. 9 



Paschen, who has measured lines in the far ultra red for 

 Cu and Ag, has unfortunately not dealt with Au, so that 

 direct comparison with observation is not possible. There 

 are, however, indirect methods which may be applied, viz., 

 the possible existence of summation lines (far in the violet), 

 the method of sounding as exemplified in [IV., V.], and the 

 presence of Ritz combination lines. In addition the d(l) 

 sequents should be the limits of a doublet F series. The 

 corresponding J) lines should be at 



D ]2 = 56206-00 + J--237.1-, 

 D 22 = 60021-56, 



•of which the D 12 should be considerably weaker than D 22 , and 

 a strong line intermediate to these should be expected as a 

 representative of the D n line. The separation of the D 12 and 

 D n gives the doublet separation of the F series and should 

 show a high occurrency. Fortunately we possess measures so 

 far down in the ultra violet by Handke *, and amongst them 

 is 60024-01 + 3*6 of intensity 1, but no 56206. There are 

 several strong lines in a suitable position for D u , but 55953 

 gives a satellite separation of 250. This separation occurs 

 so frequently throughout the whole spectrum as to fit in well 

 with the condition for a D n line. The whole set would then 

 be arranged as follows : — 



2733-70 + $ + -237.r 29471'W +lt+-llx±l-8 [56208-45+3-6] 



25255 25500 



[2986-25 + 2I+3-1] 29469-85+% (4) 55953-45 ±31 



6549-06+$+ -237 33286-53 +^±\-8 (1)60024-01+3 6 



A displacement of 28 & or 7 8 in the sequent produces a 

 satellite separation of 258'03, or 27 6\ of 248*88. The possible 

 errors in the observed do not enable us to settle which is the 

 more probable. We shall see later, when dealing with the 

 F series, that these separations are reproduced. The appear- 

 ance of the summation lines in the exact positions to be 

 expected, their relative intensities, and with a satellite 

 separation depending on the same oun-multiple displacement 

 as for orders m = 2, 3, together form weighty evidence in 

 favour of the existence of the D(l) lines as calculated. 



But we can obtain further evidence by sounding for the 

 actual D linos themselves. This method has been developed 

 in [V.] and depends on the linkage laws developed in [IV.]. 

 The e, u, v links in Au are so large that they can stretch from 



* Dissert. Berlin. The measures used are quoted from Lyman. 



