DR. T. M. LOWRY: NATURAL AND MAGNETIC ROTATORY DISPERSION 
•J86 
6. The water-jackets (fig. 3) were designed originally for use with organic liquids. 
But, unlike these liquids, quartz is an excellent conductor of heat and takes up a 
steady temperature within about 3 minutes, whereas liquids require at least this 
number of hours. But in spite of this excellent conduction, it is necessary to use 
great care to protect the ends of the quartz from atmospheric heating or cooling. This 
The jacket is double, a spiral of compo-tubing being used to direct the water in the outer section. The 
quartz is held between a cylinder of cork on the left and a strong spiral spring on the right. Four 
cover-slips are shown, one at each end of the two brass extension pieces, protecting the quartz from 
cooling by the outer air. A slot at the top of the jacket was closed by cotton wool. 
was discovered when cylinders which gave rotations of 1476°T4 and 1479°‘99 before 
measuring their lengths, were found to give 1476°’21 and 1483°'03 after measurement; 
these differences were traced to the somewhat large changes of atmospheric tempera¬ 
ture which had taken place in the interval. In these preliminary tests the cylinders 
were merely laid in a jacketed tube of 15 mm. diameter freely open at the ends and 
the variation was exceptionally large. In the final experiments all risk of temperature 
errors was eliminated {a) by using long columns of quartz so that the cooling of the 
ends was relatively much less important (Z>) by adding to each jacket a pair of heavy 
metal cylinders in good thermal contact with the jacket and having an aperture of 
only 10 instead of 15 mm. (c) by keeping the temperature of the room within about 
2° of that of the bath. Arrangements were also provided whereby four cover-slips 
could be interposed between the ends of the quartz and the outside air ; but these were 
not considered necessary and were discarded after the values for mercury green had 
been read. 
7. The lengths of the quartz cylinders were measured at the National Physical 
Laboratory to O'OOOl mm., with an estimated error of about 0’0003 mm. The 
conditions of working were sucli that greater reliance could be placed on measurements 
made with the separate cylinders and at atmospheric temperatures rather than on 
long columns jacketed to some fixed temperature. The accuracy of the length-measure¬ 
ments was of the same order as that of the measurements of optical rotation. The 
extreme variation between the Individual cylinders of Isevo-quartz corresponded with 
a difference of 0'0012 mm. in length or 0°‘03 in rotatory power. The difference 
between the two groups of four cylinders could be represented by an error of O'0016 mm. 
or 0°'04 and the difference before and after regrinding by the same figures; these 
differences are of the order of 0°'0002 per mm. and correspond with a possible error in 
the mean of aljout 0°‘0001 per mm. 
