266 
DR. T. M. LOWRY: NATURAL AND MAGNETIC ROTATORY DISPERSION 
It can also be nsed very satisfactorily to read :— 
The green mercury line . 
The yellow sodium doublet . 
The yellow mercury doublet 
5460-97 
SSDO-IQO 
5896-16 J 
r 5769-160 
15790-49 J 
The two doublets are only available for reading small rotations and must be resolved 
into separate lines if large rotations are to be observed. 
The flame spectra of lithium, sodium and thallium, and the arc spectrum of mercury 
(Landolt, p. 433 ; Disch, ‘Ann. Phys.,’ 1903 (IV.), vol. 12, 1155 ; Schoneock, ‘ Zeit. 
YVreins Deutsch. Zuck.-Ind.,’ Tech. Part, 1903, vol. 53, 652), giving the six colours 
shown above, were the only line spectra that had been effectively made use of in 
polarimetry* at the time when the present investigation was commenced. In order 
to increase the number of available light-sources, recourse was had to the brilliant 
line-spectra of the metallic arcs, as described in Section 4 of the paper. With these 
new light-sources Peekin’s method can no longer be used, the glare of adjacent lines 
being so great that it is impossible to take accurate readings of the line that is being 
brought to extinction. Under these conditions it is necessary to resolve the light 
spectroscopically before it enters the polarinieter. This was done by means of a 
constant-deviation spectroscope, the instrument being arranged so that an image of 
the slit fell upon the triple-field of a polarinieter in the position normally assigned to 
the eye-piece of the spectroscope ; a wide slit opening symmetrically was inserted 
immediately in front of the triple-field and could be used to cut off the light from lines 
immediately adjacent to the one under observation. 
At an early stage in the work it was discovered that, whilst this arrangement was 
fairly satisfactory for lines in the middle of the spectrum, it was seriously defective 
for lines in the red and blue, on account of stray light from the green and yellow lines. 
Thus readings of magnetic rotation in carbon disulphide were depressed from 70°-90 
to 69°-33 in the case of the violet line Hg 4359 and from 58°-95 to 58°-28 in the case 
of the blue line Cd 4678 by stray light of longer wave-length. The readings, 40°-19, 
of the green line Hg 5461 were not affected by stray light; but the value for the yellow 
doublet Hg 5780 was raised from 35°-14 to 35°-19 and that for the red line Cd 6438 
from 27°-52 to 27°-68 by stray light of shorter wave-length ; the errors are here much 
less than in the case of the blue and violet lines but are sufficiently serious to destroy 
the value of the measurements as exact observations. 
This fault, which does not appear to have been recognised previously, is inherent 
in all devices of the Lippich type, in which the light is resolved spectroscopically 
* Disch had also taken readings with the red hydrogen line 6563 • 04 and with the mercury line 4916-41 
but had been obliged to use half-shadow angles up to 30°. 
