280 
DR. T. M. LOWRY: NATURAL AND MAGNETIC ROTATORY DISPERSION 
more different positions and reading with both positive and negative half-shadow 
angles. The results were remarkable :— 
rH.S.+ 4481-78 
^ LH.S.- 4481-88 
1-37 
1-58 
1-53 
1-50 
1-41 
1-54 
1-57 
2-34 
2-07 
1-89 
1-76 
2-01 
1-98 
1-99 
Mean . . . 4481 - 83 
1-85 
1-83 
1-71 
1-63 
1-71 
1-76 
1-78 
Difference. 0-10 
0-97 
0-49 
0-36 
0-26 
0-60 
0-44 
0-42 
Mean of all readings = 
4481-76. 
rH.S. + 3136-07 6-27 6-24 
^ LH.S. - 3137-00 6-85 6-55 
6-57 
6-32 
6-21 
6-37 
6-55 
6-60 
6-67 
6-66 
6-49 
6-49 
6-72 
6-74 
6-53 
6-53 
6-53 
6-51 
6-50 
6-71 
Mean... 3136-53 6-56 6-40 
6-53 
6-52 
6-47 
6-45 
6-54 
6-56 
6-59 
6-57 
6-60 
Difference. 4-0-93 -fO-58 -t-0-31 - 
0-08 
+ 0-40 +0-53 
+ 0-16 - 
0-02 
-0-07 - 
0-16 - 
0-16 
+ 0-22 
Mean of all readings = 3136'.53. 
It was evident that the difference between the values for the two half-shadow angles 
might be very serious, amounting to as much as 0°‘97 in the leevo- and 0°'93 in the 
dextro-cylinders, but that positions might be picked out in which the discrepancy was 
very small; further, the average value for each position was seen to be almost inde¬ 
pendent of the variations produced by reversing the half-shadow angle. It was hoped 
that by placing the quartz columns in suitable positions satisfactory results might still 
be obtained, though the differences between the average values for the cemented 
cylinders and for the cylinders in optical contact could not lightly be set aside. 
In seeking for the origin of the anomaly, the conclusion was reached that it might 
be accounted for by the presence of patches of material which failed to give a proper 
extinction, e.g., because they produced elliptical or circular polarisation. Such a bright 
patch, if occurring in the centre of the field, would require to be compensated by 
bringing the analysing Nicol more nearly perpendicular to the main polarising Nicol; 
this would mean moving the analyser to the left or to the right according as the 
polarising Nicol had been turned to the left or to the right relatively to the smaller 
fixed Nicols used to produce the “ half-shadow ” effect ; the presence of a non¬ 
extinguishing patch would therefore raise or lower the observed rotation and by 
approximately equal amounts according as the half-shadow angle was positive or 
negative, whilst the average rotation would be substantially the same as if the patch 
were absent. 
As it seemed possible that a non-extinguishing film or patch might have been 
produced by the optical contact, the cylinders were separated and examined in groups 
and finally in units but still showed the same phenomena. Finally, the existence of 
actual flaws was detected in at least 10 of the 12 cylinders by viewing the quartz 
with monochromatic green mercury light, with a zero half-shadow angle, the eye-piece 
being drawn right out of the telescope of the polarimeter, so that it was possible for 
the first time to focus upon the interior of the quartz specimens, instead of looking 
through them to the Nicols of the triple field. Black and green bands, resembling 
