Separation of the Spectral Lines of Thorium 13 



Table XVII contains a list of symmetrical triplets. In the 

 column of intensities the red component appears first, the unsep- 

 arated /^-component next, and the blue ^-component last. These 

 lines are further illustrations of the fact that there are no appre- 

 ciable steps in the value of triplets. When they approach each 

 other so closely in magnitude, it is scarcely possible to isolate 

 lines of a certain magnitude of separation and intensity and group 

 them into series or related lines. Under such circumstances it 

 would be much less meaningless to say that lines of similar types 

 repeat themselves from substance to substance, and therefore in- 

 fer a similarity of the substances. The same is true to a less 

 degree with the quadruplets. A few like separations of any type 1 

 are useless for comparing similar chemical substances or for in- 

 ferring a similarity of chemical behavior. However, these like 

 separations have an important bearing upon that very subject, 

 because they are natural starting points from which to begin the 

 search for a connected relationship in the separations. Only 

 when such relationships repeat themselves from substance to sub- 

 stance can one confidently assert that a similarity in the sub- 

 stances exists. 



In such a diversity of separations, it might seem that small 

 errors in the measurements for these triplets would, if corrected, 

 throw them into groups differing by a small but appreciable 

 amount. So that the actual number of separations would be 

 small in number and possibly related to one another in some sim- 

 ple way, or even related to a normal value. There are some 

 well-defined lines which have a separation of 1.07 to 1.08. Do 

 these belong to the normal triplet whose separation is 1.105 or 

 are they a class by themselves? I selected some of the sharpest 

 and easiest measures of these lines and subjected them to re- 

 newed measurements upon plates of both field strengths, and 

 was not able to change their value. I similarly treated a few 

 lines with separations about 1.04, 1.14, and 1.18 and with a sim- 

 ilar result. The whole could be answered satisfactorily with five 

 times the accuracy in measurement. 



1 Purvis, Trans. Camb. Phil. Soc. XX, no. VIII, p. 193; and Proc. Cainb. 

 Phil. Soc. XIII, pt. VI, p. 325; XIV, pt. I, p. 41; XIV, pt. III. p. 217. 



IOI 



