BY ROUGH AND WHITE SURFACES. 



207 



FiR. 2. 



an incidence of 90° up to an incidence of 0°, as far as it could be observed, owing 

 to the impossibility of examining the bands where they were reflected at or near 

 a perpendicular incidence. In order to meet this difficulty, I adopted the follow- 

 ing mode of observation : — 



The rough or white suface MN E 



being placed vertically, was illuminated ^ , .■■' , „ 



with the flame of a gas burner, or a 

 moderator lamp, placed at F. The 

 observer at E, or in any direction be- 

 tween F and N, observes the condition 

 of the bands of the polariscope when 

 they are placed parallel to MN. 



If we now take a single rough surface 

 of plate-glass blackened on its polished 

 side, and place it at MN, we shall observe 

 a neutral point about H, the polariscope 

 bands being negative or horizontal, with a white centre on the M side of H, and 

 positive or vertical, with a black centre on the N side of H. This neutral point is 

 obviously produced by the equal and opposite action of light polarised by re- 

 flexion and refraction. As these two lights, proceeding from light incident at H, 

 proceed from F, they cannot reach the eye at E by the ordinary law of reflexion. 

 They are portions, therefore, of oppositely polarised rays scattered in every direc- 

 tion by the rough surface of the glass. 



If we take the same plate of glass with its surface not blackened, and place it 

 at MN, we shall find the neutral point at G. 



If we invert this plate, so that its polished side is uppermost, the neutral 

 point is advanced to A, all the bands on the M side of A being, as before, negative, 

 and those on the N side of K positive,— a, result proving that the intensity of the 

 negative bands had been increased and the positive ones diminished. 



Conceiving that this efifect was produced by the light scattered by the rough 

 surface being polarised negatively by the refraction of the polished surface, I took 

 the plate with the blackened side, and laid upon its surface a plate of transparent 

 glass, in order to imitate the action of the unpolished surface in the preceding 

 experiment. The neutral point which, by the action of the single rough surface, 

 was at H, was now advanced to A by the refractive polarisation of the two 

 surfaces of the transparent plate. When this plate was placed at 7nn parallel to 

 MN, the very same effect was produced. By inclining' mw, the neutral point was 

 advanced from A to a, still farther, by increasing the inclination, and still farther, 

 by using several plates, the plane of refraction or incidence being parallel to a 

 plane passing through FGA. 



If we turn the plate or plates at mn round 90°, keeping the same inclination 



