[ 698 ] 



LXXXVII. Belationship heticeen Spectra and Atomic Weights, 



To the Editors of the Philosophical Magazine. 

 Gentlemen, — 



IX his papers on a relationship between spectra and 

 atomic weiohts. and on the atomic weight of radium, 

 Dr. ]\Jarshall Watts'^ has tried to show how it is possible to 

 calculate the atomic weioht of an element from the atomic 

 weights of chemically related elements by means of their 

 spectra. He first makes use of a remark due to Ramage. 

 Eamage draws a diagram of the spectra of chemically- 

 related elements with oscillation-frequencies as abscissie and 

 the squares of the atomic weights as ordinates. In this way 

 each spectrum is represented by certain points on a parallel 

 to the axis of absciss?e, the distance from this axis being pro- 

 portional to the square of the atomic weight. Ramage now 

 observes that in many cases the points representing liomologous 

 spectral lines approximatehj lie in a straight line. This law 

 if true may of course be used to determine the atomic sveight 

 of an element from the atomic weight of two other elements 

 as soon as three homologous lines are known in their spectra. 

 Ramage is very careful to define whtit is meant by homolo- 

 gous f. Two spectral lines belonging to different spectra are 

 called homologous when they show the same characteristic 

 features. Thus we may say that the calcium line -^^'l^ is 

 homologous to the strontium line -4607 and to the barium 

 line 55o6. When a salt of one of these substances is brought 

 into the flame of a Bunsen burner this is the principal line. 

 In the arc-spectrum the three lines behave in a similar 

 manner, they are broadened and eiisily reversed. In the 

 magnetic field they are split into three rather broad com- 

 ponents and the spacing measured in the scale of oscillation- 

 frequencies is the same in all three cases. 



It is to homologous lines like these that we must look in 

 order to show the connexion between the spectra and the 

 atomic weights. 



It appears that Dr. AVatts in making use of Ramage's 

 remark for the calculation of atomic weights has not paid 

 attention to the definition of homologous lines. He assumes 

 lines as homologous which by their components in the mag- 

 netic field we know to be not homologous (for instance 

 Ba 4554 and Sr 3706, or Ba 4131 and Sr 3400). It is 

 evident that when the lines are densely distributed over some 

 part of the spectrum any straight line drawn through this 

 part of the spectrum in Ramage's diagram is sure to pass 



* Phil. Mag. [61 vol. V. p. 203 (1903) ; vol. vi. p. G4 (1903). 

 t Proc. of the Roy. Soc. vol. Ixx. p. 3 (1901 ). 



