274 
DR. T. M. LOWRY: NATURAL AND MAGNETIC ROTATORY DISPERSION 
tlie violet mercury line occupies the same dominant position for polarimetric work in 
the violet that the green line holds for work in the brighter portions of the spectrum. 
No other line in the blue or violet can be read with the same ease and accuracy, and 
jio other line in this region is available for general laboratory use. For all ordinary 
purposes it is sulScient to separate the line from the light blue (4916’41) and violet 
(4078’03 and 404678) lines without attempting to resolve the satellites (4348'0 and 
4341 'O) ; this can be done with an ordinary direct-vision prism on the eye-piece, and no 
further elaboration is necessary or desirable for readings of less than 1000°. With 
this simple arrangement and a vertical silica mercury lamp, readings can be taken with 
a half-shadow angle of 4° and with an accuracy but little inferior to that obtained in 
reading the green. These readings are of very great importance in measuring 
dispersion, as it is only on this line that the smaller differences of dispersive power 
can be perceived ; the variations at the red end of the spectrum are much smaller and 
much more difficult to detect with certainty. In a later paper it will be shown that 
the measurement of the ratio affords the best method at present available 
for characterising the rotatory dispersion of a substance. 
In order to ascertain the effect on the readings of including or excluding the two 
satellites, the following experiments were made. After having established the ratio 
of the violet and green rotations to be 1‘62700 for long columns of quartz, a further 
series of measurements was taken with a short rod 25‘902 mm. in length, the slit of 
the polarimeter being opened widely so that there was no separation of the satellites 
from the chief line. The readings were as follows :— 
Rotation. 
Ratio. 
Mercury green. 
661°‘47 
rooooo 
Mercury violet (current 3 amperes) . 
. 1076°‘33 
1‘62717 
4 
3 5 53 3 3 ^ 33 * • * 
. 1076°‘45 
1‘62729 
5 
33 33 33 ^ 33 • • ' 
, . 1076°‘52 
1‘62747 
These observations show that with a current of 3 amperes, the error produced by 
including the satellites amounts only to 1 part in 10,000 and would be inappreciable 
in readings of less than two right angles. In all ordinary polarimetric work involving 
readings of a few degrees only, the larger currents, giving an error of 1 part in 5,000 
at 4 amperes and an error of 1 part in 3,000 at 5 amperes, may be used without 
hesitation, if desired. 
(5) Silver. 547172, 5465‘66, 5209‘25. 
Of all the open metallic arcs, the silver arc is that which is most readily adapted to 
polarimetric work. The chief features of the spectrum are the dark green line 5209‘25 
and the bright green doublet, 5471‘72 and 5465"66, the more refrangible component of 
which approaches very closely to the green mercnry line 5460‘97. The dark and 
bright green lines could easily have been separated by a single direct-vision prism, but 
in practice all tlie open-arc spectra were resolved by means of the constant deviation 
