The Magnetic Separation of Lines 2 
position. Photographs of vibrations parallel to the lines of forces 
were taken upon one set of plates, and perpendicular to the lines 
of force upon another set of plates. This was effected in the 
usual way by means of a calcite prism. Photographs taken with- 
out the calcite, i. e. with both the parallel and perpendicular com- 
ponents on one set of plates, were used only to obtain the relative 
intensity of the parallel and perpendicular components. In these 
photographs the parallei and perpendicular components are often 
so close to each other as to prohibit their separation and hence a 
knowledge of their respective intensities. In triplets of large sep- 
aration, however, it was a matter of no difficulty. By a double 
comparison of such lines with corresponding components on 
parallel and perpendicular plates the ratio of the intensities of the 
parallel and perpendicular plates is established. The lines may 
then be compared on perpendicular and parallel plates, although 
the conditions of exposure and development may have been quite 
different. But withal there is a chance for a very large error in 
intensities, since there is a wide range in width and depth of the 
shadows, and no experimental method employed to compare the 
intensity of the shadow, as may be done in photometry or spectro- 
photometry. 
The variation in actinic sensibility of the plates throughout the 
spectrum is enormous, so that one must resort to differently pre- 
pared plates. Furthermore, the red sensitive plates, self-prepared 
with diacyamin, were far from uniform. Hence, a comparison of 
intensities of different lines, which is very important in investiga- 
tion for series, can be only roughly accurate within a short spec- 
tral range. This is generally sufficient to show that two or more 
near lines of like magnetic separation, but of great inequality in 
intensity, do not belong to the same series. This is all that can be 
expected from the intensities given in the following experiments. 
How great the variations in intensities may be under the different 
conditions of experiment may be seen by comparing the intensities 
here recorded with those given in Exner’s and Hascek’s tables 
which have been freely used. There is enough similarity in in- 
tensities to make their table serviceable for the identification of 
lines. 
