12 B. E. Moore 



Table XV contains seven 5-component lines. At least six of 

 these are unsymmetrical with respect to the ^-component only. 

 Line 4115.85 A has its components in the ratio of (o, 1, 3, 4) times 

 .417. The first three values represent the unsymmetricai /'-com- 

 ponents and the latter the symmetrical pair of ^-components. In 

 the /'-components of 3938.86 A one finds (o, 2, 3) times .41, which 

 is practically the same interval as in the previous line. The 

 ^-components of this line are not related to this interval. How- 

 ever, we may combine both p- and ^-components together as 

 multiples of the interval a/11. We then have (o, 8, 12, 14) times 

 a/ 11. Applying this interval to the previous lines gives larger 

 discrepancies in the readings. In line 3709.82 A the ^-components 

 are nearly in the ratio of one to two in separation and in the ratio 

 of three to one in intensity. 



Table XVI contains thirty-three quadruplets. Twenty-four of 

 these have unsymmetrical p-, and fifteen unsymmetrical ^-com- 

 ponents. There are six of these lines "lop-sided," i. e ; there are 

 more components on one side of the middle than on the other. 

 For three of these lines, 4619.67 A, 4318.65 A, and 4105.55 A, the 

 mean of the two separations upon one side is symmetrical with 

 the position of the third component upon the other side of the 

 middle. This is probably also true for 4050.02 A, but does not 

 hold for 4036.22 A. The distance between these two components 

 is very different for these lines. For line 4619.67 the two ^-com- 

 ponents on the one side do not appear separated on the weaker 

 field plate, but present the appearance of a broadened line. There 

 are one to two, one to three, two to three, and three to four ratios 

 represented in the separations of components, which are usually 

 expected to have equal values. Some of these ratios are repre- 

 sented more than once, but then the lines are unlike because the 

 ratio factors must be multiplied by different magnitudes to pro- 

 duce the observed separations. Two lines which have the same 

 magnitude of separation in />- or .?- usually have quite different 

 magnitudes in their respective s- or /'-components, as was noted 

 in the symmetrical quadruplets. This greatly reduces the possi- 

 bility of duplicates. 



100 



