The Magnetic Separation of Lines 15 
that the distance of the components of four lines are multiples of 
the value of +.20, and the components of two lines multiples of 
the value +.30. Hence the components of the six lines are mul- 
tiples of +.10. 
Although such a division is striking, it may be remarked that 
if one chooses the multiple value small enough all lines would 
naturally fall under the multiple proportions within the limits of 
error. Allowing an error of only .05, and all lines are at once 
multiples of .1. The same may be said of the before-mentioned 
multiple value .065, and no importance, within the limits of the 
present readings, can be attached to such small factors.. The 
value .065, or more accurately .06, appeared as a difference of 
two components, which difference was actually present when one 
directly compared the two plates, and the striking feature was 
that all components of this line were a multiple of this difference. 
So that, although errors of larger magnitude may arise, the differ- 
ence is still significant of the fact that the components of lines 
may stand in the very simple relation of multiples of small sep- 
arations. In the remaining quadruplets there can be no question 
of the presence of multiples of .20 and .30. There are no dupli- 
cates of the series character present. 
The following seventy-four lines are triplets. The intensities 
of the undisturbed p-components are not given. The ratio of the 
intensity of the s- to the p-components is usually one to two, 
though propably in many cases the ratio is nearer two to five. 
5 
