THE 



LONDON, EDINBURGH, and DUBLIN 



PHILOSOPHICAL MAGAZINE 



AND 



JOURNAL OF SCIENCE. 



[FIFTH SERIES.] 



JANUARY 1888. 



I. On a Method of Discriminating Real from Accidental 

 Coincidences between the Lines of different Spectra ; with some 

 Applications. By E. F. J. Love, M.A., Demonstrator of 

 Physics in the Mason College, Birmingham.* 

 [Plate I.] 



IN investigating problems the solution of which depends 

 on the coincidence of the lines belonging to different 

 spectra, e. g., in determining the presence of an j particular 

 substance in the Sun's reversing layer, it is usually considered 

 sufficient to demonstrate a close agreement between the 

 wave-lengths of a number of lines in the spectra. That 

 this method, however, taken by itself, does not suffice to give 

 us a reliable result was shown hy Schuster f, who demon- 

 strated that, in accordance with the Theory of Probability, a 

 certain number of coincidences between the lines of two 

 spectra might be expected to occur, even if the spectra be 

 quite unrelated; and showed how to calculate the maximum 

 number of coincidences possible on the assumption that no 

 relation exists between them. If no greater number than 

 this is found, the coincidences must be looked upon as 

 probably accidental. The method as given by Schuster is 

 employed to determine whether the lines of a spectrum are 

 harmonically related; but it could obviously be quite as well 

 applied to examine the relations between two different observed 

 spectra. It demands, however, a considerable amount of 

 rather troublesome computation. Further, cases may occur 



* Communicated by the Physical Society ; read November 26, 1887. 

 t Proc. Eoy. Soc. xxxi. p. 337, 1881. 

 Phil. Mag. S. 5. Yol. 25. No. 152. Jan. 1888. B 



