by transmission through Fluids. 425 



Again, if the proportions of the combination be constant, but 

 the depth varied, the degrees of rotation will, as before stated, 

 bear a direct proportion to the depth; thus, since 10 inches of 

 a solution of cane-sugar, containing one-fourth its weight of 

 solid sugar, produces, according to my experiments, a rotation 

 of 50°, if the depth be 1 inch it would produce a rotation of 

 5°, and 36 inches would produce a rotation of 180°, and conse- 

 quently with such a depth there would be no apparent I'otation. 



Having thus shown that neither the direction in which the 

 analyser is rotated, nor the appearances presented at any gii^en 

 point, can alone enable us to determine whether the rotation 

 is right or left-handed, we proceed to explain that such in- 

 formation can only be obtained by observing the order of suc- 

 cession in which the different tints or shades of obscuration 

 present themselves. 



Removing the plane of polarization, by revolving the polar- 

 izer, any number of degrees not exceeding 90° either to the 

 right or to the left, it will be readily observed that the direc- 

 tion in which it has been moved might be ascertained by ob- 

 serving whether on moving the analyser the image X became 

 darker or brighter; the image X becoming darker when the 

 analyser is moved in the same direction as the other, but 

 lighter if moved in an opposite direction. This law, however, 

 it will be also seen is reversed if the rotation exceeds 90°. 

 This method, then, affords the means of ascertaining the di- 

 rection of rotation, provided it does not exceed 90° ; the rota- 

 tion being 77^/i/-handed, if, when the analyser is turned to the 

 right-hand^ the image which is the subject of observation be- 

 comes darker and darker, and /^-handed if it becomes lighter. 



It matters not whether X or O be the subject of observa- 

 tion, or whether X be light and O dark when set to 0°, pro- 

 vided the amount of rotation be estimated from the position 

 necessary to restore either ray to precisely the same condition 

 in which it appeared at 0°, -previously to the introduction of 

 the circularly polarizing medium. When, however, a double 

 refracting analyser is used, I would in all cases recommend 

 the experiment to be commenced by placing the analyser and 

 polarizer in such a position as that X is at its greatest point 

 of obscuration when the vernier or index points to 0°. When 

 a Nicol's prism or tourmaline is used as an analyser, pre- 

 senting only one image, then, previous to the introduction of 

 the circularly polarizing medium, it must be so placed with 

 respect to the polarizer as that the index of the vernier points 

 to 0° when the light is most completely polarized or obscured. 

 Uniformity in the mode of conducting these experiments will 

 greatly promote accurate results. 



