Stoney — Cmise of Double Lines in Spectra. 567 



the oscillation-frequencies of the successive lines is given by the formula 



in which A: is a constant for the whole series. By putting, successively, into this 

 formula the whole numbers 3, 4, 5, 6, &c., for ??^, it furnishes values for N, which 

 are the oscillation-frequencies of the successive lines. This still further establishes 

 the fact that these rays are caused by one event, or by one body of inter-depen- 

 dent events, occurring in the molecules of the gas. It can easily be seen that 

 the 1st, 2nd, and 4th of this series are lc\^, ]c^\, and /fff, in accordance with the 

 law which I had announced in 1870. 



There are many series of lines known to spectroscopists which form patterns 

 somewhat like that of the hydrogen series, and which we may presume are to be 

 referred to some one event or group of associated events occurring in the molecules 

 of the gas. The discovery of Balmer's law has stimulated other inquirers to 

 search for similar simple laws connecting the oscillation-frequencies in cases of 

 this kind ; and these attempts have at all events elicited useful approximate laws, 

 which have done science the service of making it possible for the investigator 

 in many important cases to pick out the members of an associated series of lines, 

 where the individual lines are too far separated, or too much mixed up with lines 

 not belonging to the series, for his eye to detect the association upon mere 

 inspection.* Most useful work of this kind has been carried on by Professors 

 Kayser and Runge in Germany and by Professor Rydberg in Sweden. It must 

 suffice here to give an outline of those results of Professor Rydberg's analysis 

 of the spectra of the monads lithium, sodium, potassium, ruthenium, and csesium, 

 to which a new and special meaning is imparted by the investigation in the 

 present Memoir. 



Both Professor Rydberg and Professors Kayser and Runge findf that the 

 spectrum of each of these elements contains and almost altogether consists of 

 three series of double lines. The distribution of the pairs constituting each 

 series over the spectrum is such as to form a pattern somewhat like that of 

 the great hydi'ogen series to which Balmer's law applies, although no equally 

 simple law has been detected connecting their positions. We shall presently see 



* The first ■work of this kind with which I am acquainted was the successful separation of one of 

 the bands of the spectrum of CO into two distinct series by Professor Alexander Herschel in 1883. See 

 Transactions R. 8. JSdinh., vol. xxxii., p. 454. It was carried out before the announcement of Balmer's 

 law, by the help of a harmonic law. 



f Professor Eydberg, " Eecherches sur la Constitution des Spectres d' Emission des Elements Chimiques " 

 {Transactions of Royal Academy of Sciences of Sweden, 1889, Baudot 23, No. 11). Professors Kayser and 

 Eunge, "Tiber die Linienspeotren der Alkalien" {Transactions of the Berlin Academy of Sciences, 1890, 

 St. xxviii., s. 555). 



