344 Prof. D. B. Brace on Light Propagated in 



back by a mirror, adjusted so as to be parallel to the lines of 

 force. The analyser showed the presence of depolarization 

 in the glass cylinders due to accidental double refraction, 

 although the same had been carefully selected and annealed. 

 No change in the two halves of the field of view could be 

 detected when the magnet was excited. The difficulty of 

 obtaining the highest sensibility with glass on account of 

 depolarization led to the following experiment with a liquid. 



In Plate IV., DE is the tube 48 cm. long, containing the 

 liquid to be examined, surrounded with a water-jacket 

 mounted between the poles of the electromagnet. The faces 

 of the latter were 50*7 cm. by 8*7 cm., accurately planed, and 

 7*3 cm. apart and placed as near parallel as possible, so as to 

 insure a uniform field of parallel lines of force. The optical 

 system and tube was mounted on an arm A, 70 in. long, 

 pivoted at P to a frame support resting on the bed-plate of 

 the magnet. The other end of the arm carrying the analyser 

 and telescope rested on the carrier S, which could be moved 

 laterally by the micrometer screw M. Sunlight passing 

 through the lens R was brought to a focus at L after being 

 reflected by the right-angle prism P. The ray was then passed 

 through the tube four times by reflexion at the mirrors E and D, 

 making a total distance of 198 cm. ; and finally through the 

 telescope T, which formed an image of the half-shade a little 

 in front of the analyser N. This image was observed with 

 the small telescope t, which could be moved in the direction 

 of the ray by the micrometer screw m. The " Lippich " L 

 could be rotated so that the ray rr could be polarized in any 

 plane with respect to the magnetic field. The divided circle 

 of the analyser could be read by the microscopes cc to '003°. 

 On account of diffused light and internal reflexion, the ray 

 was cut down by successive diaphragms to about 1 cm. 

 diameter. 



In some of the experiments the optical arrangements shown 

 at Hi and R 2 were used. In the former the ray, after passing 

 through the Lippich L, was reflected by a double right-angle 

 prism in the direction r. This prism was found, however, to 

 depolarize slightly, and did not give as great a sensibility as 

 R. In R ? the return ray passed within the mounting of the 

 Lippich, close to the nicols. This was used with small tubes 

 of liquids w T ith only two passages of the ray, and with less 

 than half the air-gap between the poles. Better definition 

 was obtained with this system, but the arrangement R, on 

 account of the greater length of the ray, gave the highest 

 sensibility. 



The liquid used was a saturated solution of iodide of 



