OF ARTS AND SCIENCES. 295 



tiveness of the Arago polarimeter seems to be due to the fact, that, in- 

 stead of comparing the brightness of the field on two sides of the line 

 separating them, the comparison may be made on both sides of each 

 band, or throughout nearly the whole field. Accordingly, the rectan- 

 gular aperture was replaced by a series of metallic bars separated by 

 intervals exactly equal to their width. The double-image prism is 

 then placed at such a distance that the separation of the images shall 

 equal the width of the bars, so that the two images of the intervals 

 shall be exactly in contact. If this condition is fulfilled, the bands 

 will disappear, and the field will be perfectly uniform when the Nicol 

 is turned so that the two images have equal intensity. A slight mo- 

 tion of the Nicol will render the bands alternately light and dark, and 

 the exact point of disappearance may be determined with much pre- 

 cision. The sensitiveness appears to be even greater than that of the 

 Savart plate, since the change in brightness is abrupt at the edges of 

 the band, instead of varying continuously from the centre of the band 

 to the centre of the interspace. The bands were made by cementing 

 a piece of tin-foil to a plate of glass, ruling a series of parallel lines 

 upon it, and removing the foil from the space between the second and 

 third, and the fourth and fifth lines, &c. It was necessary that these 

 intervals should be made less than the interval between the first and 

 second, and the third and fourth lines, &c., by the width of the line 

 cut in the foil ; otherwise the bands would be narrower than the inter- 

 spaces by this amount. The bands might also be sawed out of a plate 

 of brass, and finished with a file. Since the separation of the images 

 produced by Iceland spar varies with the angular direction of the ray 

 traversing it, the bands will not disappear in all parts of the field at 

 once. This, however, is rather an advantage, since the eye is extremely 

 sensitive to a certain phase, in which the bands reappear equally in the 

 two sides of the field. A pointer serves to guide the eye to the centre 

 of the field. Should it be desirable to employ a large field, the rays 

 may be rendered parallel by a convex lens at a distance from the eye 

 equal to its focal distance. There are some advantages in substituting 

 a simple' plate of Iceland spar for the double-image prism. The rays 

 compared then come in the same direction, instead of being inclined 

 by an angle equal to that of the images of the prism. The two images 

 do not lie in the same plane. This may be remedied by inclining the 

 bars so that the right-hand edges shall always be in front or behind 

 the left-hand edges. If, however, the bands are narrow, and are 

 placed at the distance of distinct vision, this error is unimportant, and 

 the bands may be made to disappear almost entirely. 



