Structure in the secondary hydrogen spectrum—IV

Abstract

Fulcher’s bands, as extended by himself and Allen, consisted of six sets of five lines in the red whose wave numbers obey a parabolic law very closely, both when the lines within each set are compared and when corresponding lines in each set are compared from set to set. There were also four sets of four lines in the green with similar properties. The numerical coherence of these lines has been ably discussed by Curtis, and I have nothing to add to that discussion except to say that on plates taken by Mr. Wilfred Hall with a 21-foot grating the line λ = 6127⋅40 (16315⋅64 v ) is shown resolved with one component which preliminary measures place at λ = 6127⋅21, the value suggested by Fulcher. It thus appears that the anomaly in this line is due to blending and not to a perturbation. In addition to the foregoing lines Fulcher included four other lines in the first four sets in the red, which he denoted by the letters S 4 , S 5 , S 6 , S 7 . These lines follow the parabolic law from set to set, but their interrelation within each set is far from obvious. An attempt to include them in a consistent scheme has been made by Dieke, who arranges the vertical series of Allen (S 3 , S 2 , S 1 , S 0 , S -1 ) as the first five lines of Q branches of which S 6 is the seventh line, whilst S 4 and S 7 are the first and second line respectively of associated It branches. There are serious objections to the arrangement proposed for the lines S 4 —S 7 , although I shall show that there is strong evidence for the correctness of the arrangement of Allen's vertical series as Q branches. In the first place, Dieke’s scheme has no place for any of the S 5 lines nor for the S 6 lines 5989⋅22, 6093⋅83 and 6399⋅45, whereas the credentials of S 5 seem as good as those of S 4 or S 6 , and one S 6 line seems as good as another; or, at least, the physical evidence for a connection between the S 6 lines and the S 3 —S -1 lines is no better for those which are included than for those which are excluded. In the second place, an examination of the numerical coherence of the wave-numbers of the lines in the R series proposed by Dieke shows that it is too irregular for this arrangement to be likely to be correct.