109 



to the series v = (^{2) — f2(n)(n = 4, 5, . . .) in the orthohelium^) and lilliiimi") spec- 

 trum. For the same field tlie new lilliium lines were more intense Ihan the new 

 orthohelium lines in conformity with [he fact that the values of the functions f/'2 (") 

 differ less from unity for the lithium spectrum than for the orthohelium spectrum. 

 As it should be expected from the comparison with the elTect of an electric Held on 

 a system of one electron rotating in a central field it was further observed, that the 

 new lines showed a characteristic polarisation relative to the direction of the electric 

 force and were displaced more and more for increasing intensity of the field. In con- 

 formity with the theory the displacements observed were proportional to the square of 

 the electric force for small fields, while for larger fields it became graduallj' propor- 

 tional to the first power of the force. As regards the parhelium spectrum (f^in) differs 

 still less from unitj' than in lithium and we should therefore in this spectrum expect 

 the appearance of new lines of the above mentioned series of still larger intensities than 

 in the other spectra under consideration. These lines are not recorded by Stark as 

 separate lines, but due to the close coincidence in this case of the lines of the new 

 series and those of the ordinary diffuse series, the former enter as components in 

 the complicated effects described by Stark as the electric "resolution" of the lines 

 of the diffuse series. This is shown very clearly by some measurements of the effect 

 of electric fields on the helium spectrum published quite recently by H. Nyquist^) 

 who has used the method of Lo Surdo which allows one to obtain on the same 

 photograph the effect on the lines of electric fields of continuously varying intens- 

 ities. Besides components which are formed by continuous displacement from the 

 original lines, Nvquist's photographs show that the mentioned resolutions of the 

 two diffuse helium series contain components the intensity of which vanishes for 

 vanishing field and which for decreasing field tend to positions at definite distances 

 apart from the original line. In case of the parhelium spectrum these positions 

 correspond in the first place to the lines v = f^ (2) — ^ in), but besides these lines 

 there appear distinctly in the lower members of the diffuse series of both helium 

 spectra components the positions of which for decreasing field converge to the lines 

 given by J' = /'2(2) — fiin){n =4, 5). As regards Stark's observations of the "resolu- 

 tions" in the electric field of the higher members (72 = 5, 6, . . ) of the diffuse series 

 of the lithium spectrum and both helium spectra, we shall expect that for very 

 small fields there will appear, besides the new lines mentioned above, a number of 

 other new lines corresponding to v = [2 (2) — fj (n) where t> i. While according to 

 the considerations on page 35 the new lines for which t" — t' is either 2 or for 

 small fields should show intensities proportional to the square of the electric force, 

 we shall expect that the intensities of the latter new lines which correspond to larger 

 values of t" — r' will increase proportionally to higher powers of this force. Due lo 

 the exceedingly small deviations from unity of the values of ff, (/i) for sucli values 



') J. Koch, Ann. d. Phys. XLVIII, p. 98,(1915). 

 2) J. Stark, Ann. d. Phys. XLVIII, p. 210,(1915). 

 ä) H. Nyquist, Phys. Rev. X, p. 22G, 1917. 



D. K. D. Vidensli. Selsk. Skr., naturvidensk. og m.alhcm, AW., 8. Række, IV, I. 15 



Ö 



