392 Prof. D. B. Brace on the 



mirror 1. This reflected it back, through a convergence 

 point in the axis of rotation below the brass L, to the second 

 concave mirror 2, and so on; the last mirror reflecting it, through 

 its point of convergence, successively into the sensitive-strip 

 cell S, the analysing nicol A, the condensing lens I, the total 

 reflecting prism T' and the focus / of the lens I, and, finally, 

 either into the eye or the observing telescope. The tube 

 carrying the thin cover-glass C could be adjusted by the 

 arm c so that the latter was exactly in the plane of the front 

 face of S, thus giving complete rotary compensation for the 

 two opposite paths of the rays within the liquid. The pola- 

 rizing nicol P could be rotated by means of an arm p, 

 attached to its retaining tube, so that it might be crossed 

 with the analyser A or varied at will. An extension arm, 

 some 73 cm. long and carrying an index, could be fastened 

 to this, and its rotation measured on a scale fixed upon the 

 trough. The strip S, upon which my sensibility depended, 

 I have already described*. This particular strip, which was 

 mounted so as to cover half the field of view, was 15 mm. 

 wide and 45 mm. long and O'l mm. thick, and was re- 

 markable in the sharpness of its edge. It was cut so that its 

 plane of greatest length made an angle of 70° with the optic 

 axis, and placed in the cell in such a way that the ray 

 traversed the strip at right angles to its optic axis. Instead 

 of using a-monobromonaphthaline, whose index is exactly that 

 of the ordinary ray, I used carbon disulphide, which had 

 been carefully clarified. The index of the latter is only 

 slightly below the ordinary ray, but not enough to cause 

 total reflexion for any of the ordinary rays. This liquid is 

 much clearer than the former, and gives superior definition 

 and a better vanishing line. The end plates, which were of 

 the thinnest cover-glass, were carefully tested for double 

 refraction, after being cemented on, and the cell filled with 

 liquid. To this cell was fastened an arm, a (fig. 3) with 

 which it could be rotated so as to vary the azimuth made by 

 the principal plane of the strip with that of the analysing 

 nicol, and thus the intensity of the field of view, settings for 

 a match being made with the arm p of the polarizer. The 

 submerged joints were cemented with a mixture of glycerine 

 and gelatine to prevent ingress of the oil to the nicols and 

 prisms. Preliminary adjustments of the beam of light were 

 usually made before filling the trough with liquid, after which 

 the ray had to be readjusted in order to bring it back into 

 the field of view again. Much care had to be exercised in 



* Phil. Mag. Jan. 1903, p. 161. 



