IN QUARTZ OF LIGHT IN THE VISIBLE REGION OF THE SPECTRUM. 
265. 
which formed the source of light for the polarimeter. The difficulty of having to move the light-source, 
collimator and prism for every fresh colour is effectively overcome by using a constant-deviation prism 
(Christie, ‘Roy. Soc. Proc.,’ 1878, vol. 26, 8) mounted on a spectroscope in which both slits are fixed 
and the colour is varied by merely rotating the prism (see F. Tw^MAN, ‘Phil. Mag.,’ 1907, vol. 1.3, 481). 
Lippich’s method of working has the great advantage that it can be applied to his half-shadow instruments 
with a double or triple field. 
Landolt’s Method. —Another method in which a continuous spectrum is used as a source of light is 
that of Landolt (‘Sitz. Akad.,’ Berlin, 1894, 323; ‘ Ber. Deut. Chem. Ges.,’ 1894, vol. 27, 2872), in 
which five filters are used to separate light in the red, yellow, green, light blue and dark blue portions of 
the spectrum, the optical mass-centres of the transmitted light at wave-lengths 6659, 5919, 5330, 4885, 
and 4482 being near to those of the Fraunhofer lines C, D, E, F, G at 6563, 5893, 5270, 4861, and 4308 
respectively. This method, which avoids the use of a spectroscope, is the only one that has been adopted 
widely in chemical laboratories (Winther, ‘ Zeit. Physikal. Chem.,’ 1907, vol. 60, 563; Walden, ‘ Zeit. 
Physikal. Chem.,’ 1906, vol. 55, 1; Tschugaefe, ‘Zeit. Physikal. Chem.,’ 1911, vol. 76, 469; Grossmann, 
‘ Zeit. Physikal. Chem.,’ 1910, vol. 75, 129). 
Perkin’s Method. —The use of a small direct-vision spectroscope in front of the eye-piece of the 
polarimeter was introduced by Sir William Perkin as a method of purifying the flame-spectrum of 
sodium. But it is capable of a much wider application as affording the simplest and one of the most 
efficient means of measuring rotatory dispersion in ordinary laboratory practice. Perkin used it in 
reading the magnetic rotation of red lithium and of green thallium light, salts of these metals being added 
to a platinum boat, already containing a sodium salt, heated by a jet of oxygen in the centre of the flame 
of a Bunsen burner. Used in conjunction with the lithium and sodium flames and the enclosed mercury 
arc, Perkin’s method satisfies all the requirements of a standard laboratory method for modern use. 
In reviewing the methods described above, it is necessary to emphasise the fundamental requirement 
for accurate polarimetric work, namely, that the field of the polarimeter must he uniformly lighted with mono¬ 
chromatic light. Of the methods in which a continuous source of light is used, Broch’s method cannot be 
used to measure optical rotations of small magnitude and (as will be shown in a later paper) it can be 
improved considerably even in the case of large rotations by using a polarimeter with a triple field and 
replacing the continuous spectrum by a crowded line spectrum, such as that of the iron arc ; Landolt’s 
method, on the other hand, breaks down completely in the case of rotations exceeding a very few degrees, 
the large blocks of spectrum employed giving no extinction at all with rotations of larger magnitude; the 
Lippich method is also unsatisfactory, especially in the case of large rotations, on account of stray light; 
this point is discussed in a later paragraph (p. 266). Of all the methods described above, the only one 
that is available for the accurate measurement of rotations, both of large and of small magnitude, is that 
of Perkin, in which a light-source containing two or three strong spectrum lines is resolved by means of 
a direct-vision prism placed in front of the eye-piece of the polarimeter. 
3. New Methods and Devices for Measuring Rotatory Disjyersion. 
Perkin’s method gives excellent results when applied to light comprising only a 
few spectrum lines. It is very economical of light and, after repeated trials of other 
methods, has been found to be the best method now available for reading :—• 
The red lithium line. 6708’2 
The green thallium line. 5350'65 
The violet mercury line. 4358'58. 
2 M 
VOL. CCXII.—A. 
