I HAVE recently been studying the vibrational energy functions of molecules, in connexion with a recalculation of heats of dissociation, and have noticed that the equation given by Dieke and Babcock (Proc. Nat. Acad. Sci., 13, 670; 1927) for the upper level of the atmospheric bands of oxygen has an unusually large coefficient for the n3 term. This, I now find, is due to an arithmetical error of 2 cm.-1 in their location of the origin of the O – O band. The correct figure should be 13,120.97 cm.-1, instead of their 13,122.97. The correction not only leads to the expected small coefficient for n3, but also brings their data into good agreement with the older constants, as given on p. 232 of the “Report on Molecular Spectra in Gases”. The resulting corrected equation is