250 Mr. W. Sutherland on the 



Kydberg (Astro-physical Journal, v. & vi., 1897) I saw that 

 their work removed all doubt in a very interesting manner. 

 In the spectrum of the star f Puppis Pickering discovered a 

 regular series of lines, and then found that, if the Balmer 

 formula for the hydrogen lines is written in the form 



his series for the star could be written 



X=X (2m-f-l) 2 /{(2m + l) 2 -16^, 



with the same value of \ as that for hydrogen. But this- 

 last form can be written 



X=A (m + l/2) 2 /{(m + l/2) 2 -4} 3 



and therefore represents the chief series of harmonics which 

 we have just been studying. We will give side by side the 

 values of X in air as measured by Ames, by which we found 

 our values 4*518 and so on for m, and as discovered as a 

 series in the spectrum of ? Puppis : — 



m= 4*5 5*5 6*5 

 4534-8 4195-9 ... 

 ... 420.1-6 4026-5 



Thus Pickering's series for the hydrogen in £ Puppis gives 

 us the missing values 6*5, 9\5, and 10'5 in our series obtained 

 from Ames's measurements on terrestrial hydrogen in a 

 vacuum-tube. It is not to be understood that the f Puppis 

 lines give the fraction *5 any more exactly than it occurs in 

 our table, but the fact that Pickering found his series quite 

 independently of such considerations as we are using, and 

 brought it under the Balmer formula for hydrogen by the 

 introduction of the *5, with three additional instances of the 

 occurrence of this fraction beyond those obtained on earth, 

 goes a long way to prove that the departures from *5 in our 

 fractions should not cause us any serious misgiving, but 

 should be ascribed to subsidiary causes affecting the nodal 

 subdivision. 



Kydberg's procedure with Pickering's series is in accord- 

 ance with the wonderful laws of spectral structure discovered 

 by him and to be discussed in Section 4. Denoting l/X by n 

 so that n is the number of wave-lengths in a centimetre, he 

 developed the formula 



rc=K -B/(m + / a) 2 , (2) 



in which B is supposed to be the same for all substances, 

 while ?i and //, vary, and m has integral values. For the 



7-5 



8-5 



9-5 



10-5 



3924-5 



3858-85 







3924-9 



3858-6 



3817-2 



3783-4 



