Separation of the Spectral Lines of Thorium 17 



than the other. This action suggests that one has here a pair of 

 close double lines differing but slightly in separation. 



Table XX gives a few (38) scattering unseparated lines. 

 There seem to be some easily recognized pairs and groups here. 

 In the table, the pairs are shown in small brackets, and the groups 

 in large brackets. The lines not included in the brackets are more 

 scattered. For the shorter wave-lengths there is increasing prob- 

 ability of small separations escaping one's notice. 



General Remarks. — It is among the lines which have several 

 components that duplications are most easily recognized. It is 

 also easy to recognize their repetition from substance to sub- 

 stance. But this might be misleading. For if the magnitudes of 

 the separations for a particular number of components change 

 with increased number of lines they ultimately differ by only 

 small amounts, then one certainly would have separations found 

 in other substances, and a type, therefore, could have no par- 

 ticular meaning. This is too well known among triplets to need 

 comment. It is also true in thorium quadruplets. There are 

 in this substance a number of lines whose separations are reason- 

 ably similar to separations which I found in yttrium and zirco- 

 nium. With such a list of quadruplets and such diversity of sep- 

 aration as found in thorium, it is rather surprising, however, that 

 no quadruplet was found which approached reasonably near to 

 the magnitude of the separation in the quadruplet principal series. 



With the lines having five or more components in thorium there 

 are few duplications, and as the number of components increases, 

 the number of representatives of the different types correspond- 

 ingly decreases. So that there is practically no chance to apply 

 Preston's law in these lists. 



As found by myself and others, and in particular by Runge, 

 the successive separations in these several component's lines are, 

 for each particular line, multiples of small values called 'intervals,' 

 which Runge 1 first showed were multiples of aliquot parts of a 

 normal separation. In thorium these small intervals are not as 

 near to aliquot parts of the normal a as one should desire for 



'Phys. Zcit. 8, p. 15, 1907. 



105 



