On the structure of the Balmer series lines of hydrogen

It is well known that the Balmer series of the hydrogen spectrum, under moderate resolution, consists of doublets. Numerous measurements have been made of the doublet separation for Hα , but only a few of that for Hβ . These are as follows:— Hα . △λ (Å. U.) △ν (cm.-1 ) Michelson and Morley. . . 0·11 0·253 Ebert. . . . . 0·132 0·306 Michelson. . . . . 0·14 0·323 Houstoun. . . . . 0·065 0·153 Fabry and Buisson. . . . . 0·132 0·306 Meissner. . . . . 0·124 0·288 Merton and Nicholson. . . . 0·132 0·306 Merton. . . . . 0·145 0·34 Gehrcke and Lau. . . . . 0·126 0·293 Hβ . △λ (Å. U.) △ν (cm.-1 ). Michelson. . . . . 0·08 0·33 Merton and Nicholson. . . . 0·033 0·14 Merton. . . . . 0·093 0·39 Gehrcke and Lau. . . . . 0·0695 0·294 The only determinations which appear to have been made of the doublet separation for Hγ and Hδ are those recently obtained by Gehrcke and Lau. Their result for Hγ is △λ = 0·058 Å. U. (△ν = 0·307 cm.-1 ), and for Hδ is △λ = 0·0432 Å. U. (△ν = 0·257 cm.-1 ). The outstanding feature of the results given above, it will be seen, is the lack of agreement exhibited by the numbers for Hα and Hβ . While some spectroscopists have taken the Balmer series to be a principal one, others have considered it to be a subordinate series. From an inspection of the values recorded above, it is difficult to reach a definite conclusion respecting the type to which the Balmer series should be assigned. The general trend of the numbers representing the frequency differences is in the direction of a decrease, with the progress of the series towards the violet. This might lead one to conclude the type was that of a principal doublet series. Such a conclusion, however, is scarcely warranted, for Gehrcke and Lau, who found for the doublet separation of the first four members, Hα = 0·29, Hβ = 0·29, Hγ = 0·3, Hδ = 0·26 cm.-1 —values which show a tendency gradually to decrease—have stated their results are to be taken as supporting the view that the frequency difference of the doublet components is constant for all members of the series. Departures from this law exhibited by their numbers are to be considered as coming within the limits of error of the observations.