^{page 245 note *} Richardson, O. W. and Das, K., Roy. Soc. Proc., A, vol. cxxii, p. 688 (1929).CrossRefGoogle Scholar

^{page 245 note †} Sandeman, I., Proc. Boy. Soc. Edin., vol. xlix, p. 48 (1929).Google Scholar

^{page 245 note ‡} Gale, H. G., Monk, G. S., and Lee, K. O., Astrophys. Journ., vol. lxvii, p. 89 (1928).CrossRefGoogle Scholar

^{page 245 note §} The question of whether the multiplicity 3 of the state ³S has been correctly assigned is of small moment, since it is desirable to have a convenient label, such as ³S, to distinguish these states from the ¹S systems, some of which have recently been analysed accurately, in the ultra-violet by Hyman and Birge and in the visible by Richardson and Davidson. Mecke and Finkelnburg (Die Naturwissenschaften, April 1929) claim to have detected triplet structure in the initial ³P states of the Fulcher Bands, but not in the final ³S state.

^{page 246 note *} Allibone, T. E., Roy. Soc. Proc., A, vol. cxii, p. 196 (1926).CrossRefGoogle Scholar

^{page 246 note †} Poettker, A. H., Phys. Rev., vol. xxx, p. 418 (1927).CrossRefGoogle Scholar

^{page 247 note *} Believed to be unresolved.

^{page 249 note *} I. Sandeman, loc. cit.

^{page 250 note *} Sandeman, I., Nature, vol. cxxiii, p. 410 (1929).CrossRefGoogle Scholar

^{page 250 note †} Mulliken, R. S., Phys. Rev., vol. xxxii, p. 388 (1928).CrossRefGoogle Scholar Mulliken's method, given in a footnote to Table I of his paper, is primarily intended for bands of multiplicity 2, but can be extended by making the necessary changes to bands of odd multiplicity.

^{page 252 note *} National Research Council Report on Molecular Spectra in Gases, p. 233 (1926).Google Scholar

^{page 253 note *} See Loomis, and Wood, , Phys. Rev., vol. xxxii, p. 234 (1928).Google Scholar This rule, like that of Mecke, must be described in the words of Loomis and Wood as “semi-empirical.” It should also be noted that these tests refer to the old mechanics. In the present paper the notation refers throughout to the old mechanics, except where the application of the new mechanics to the term form F(j) is considered. The writer's analysis shows the differences between vibrational levels and the B constants to the first two decimal places to be to all intents numerically identical when calculated on either mechanics.

^{page 253 note †} E. Condon, ibid., vol. xxviii, p. 1182 (1926).

^{page 255 note *} Birge, R. T., Proc. Nat. Acad. Sci., vol. xiv, p. 12 (1928).CrossRefGoogle Scholar Mecke and Finkelnburg (loc. cit.) give the value of E_e for 2³S as 90083, based on their identification of the Lyman Bands with the transition 2³S→1¹S. Their values of ω₀ and ω₀x are also consequently smaller. As their work is not yet published, the value of Birge is retained pending the establishment of their hypothesis.

For the purpose of reckoning the quantity E_e for the 3³S state the null line of the band 0→0 has been taken as 11607·4, since the value 11605·7 given in Table IV has to be increased by ½(B₀″ — B₀′) = 1·7 to bring it into true accord with the old mechanics.