l»\v«vV \ in the Plane of Polarisation. 127 



The intensities of these images will therefore stand in the relation 



(tana eos^S)'- : 1. 



Thus if we take a = 45°, that is, place the plane of polarization 

 of the incident light at 45° to the vertical, the intensities will be 

 related as 



cos/32:l, 

 which will be for /S= 30°, 



3:4, 



and therefore already appears very perceptibly. The difference 

 will be still greater with larger angles of diffraction. The feebler 

 image is that produced by the horizontal vibrations. 



3. The first of these methods of observation is that chosen by 

 Stokes ; I used the second, which gives in a shorter time the 

 wished-for instruction. 



I used at first gratings which were produced by divisions in 

 glass, but without getting a distinct result ; these give very 

 irregular results with respect to the polarization of diffracted 

 light, which is easily explained. From the diffraction apparatus I 

 now took a Schwerd's lampblack grating, and this showed very 

 distinctly the difference in the brightness of two superimposed 

 pictui-es for somewhat large angles of diffraction. For this I 

 allowed sunlight to pass thi'ough a vertical slit into a dark cham- 

 ber, in which this light was first polarized by a Nicol's prism, 

 whose principal section was inclined at an angle of 45° with the 

 vertical. After this the light met the black grating by entering 

 through the glass, and passing out from the grating which was 

 uncovered behind. The diffraction spectra were observed as 

 usual by a telescope placed behind the lampblack grating, by 

 which the black grating remained undisturbed, and the telescope 

 alone was rotated horizontally about the middle line of the lamp- 

 black grating. 



AVhen, before the ocular of the telescope, a double-refracting 

 prism was placed, and this was turned so that the vertical thread 

 of the cross in both the visible spectra was seen in one vertical 

 line, both the images in the middle of the diffraction, for /3=0, 

 where the slit is seen white, were found to be equally bright. 

 And only towards the outside, when /3 was about 20', was one 

 of these perceptibly darker than the other. This difference in 

 the brightness of both the images of the parts of the diffraction 

 spectrum, seen at one time, increases from this point very per- 

 ceptibly. Here the feebler image is that produced by the hori- 

 zontal vibrations. 



If, now, we investigate, with this double-refracting prism in the 

 same position, the light that comes from a Nicol whose principal 



