122 On the Oplicul Properties 



nebulous light. The two middle ones, which were the jsrincipal 

 images, were equally luminous, and were polarized in an oppo- 

 site manner, like all other double images ; but each of the two 

 outer images was polarized in the same manner as the bright 

 image furthest from it. The most refracted of the two prin- 

 cii)al images was in th.is case the more perfect of tiie tAvo, and 

 exhibited a degree of prismatic colour so much greater than the 

 other, that it obviously belonged to a higher dispersive power. 

 When the light enters the prism, and emerges from it at equal 

 angles, the four images are not distinctly sej)arated, and are ex- 

 tremely imperfect. AV'lieu the angle of incidence at the first 

 surface of tlie prism is increased, the images become more and 

 more distinct, and better separated ; but, by diminishing the 

 angle of incidence, all the images approach one another, and 

 are confounded into one mass of nebulous liglit. 



With a plate of carbonate of barytes, which was about two- 

 tenths of an inch thick, and which had its surfaces at right an- 

 gles to the direction of the longitudinal joints, the image of a 

 candle was a large circular mass of light, when the incidence was 

 perpendicular. By inclining the plate, this mass v.as changed 

 into an annular image : by increasing the inclination, it assumed 

 the form of a crescent; and at a considerable angle of incidence, 

 it vvas separated into three imperfect images, or circular arches 

 of nebulous light, similar io those which were seen with the 

 fifth juism. The middle image, which was the brightest, con- 

 sisted of the ordinary and extraordinary image, which were not 

 separated, in consequence of the parallelism of the refracting 

 surfaces. In one position of the plate, these arches were crossed 

 bv other three sin)ilar arches, inclined to the first at an angle of 



10' or n\ 



The phaenomena which have now been described, differ in se- 

 veral respects from those which are presented by the agate. In 

 the carbonate of tarijles, the two images are distinctly sepa- 

 rated, and are, therefore, formed by two separate refractive 

 powers ; whereas in the agate, the bright image is placed in the 

 centre of the nebulous n:iass. In the carlonate of harytes, the 

 imperfect image occupies a small space ; but in the agate, it is 

 an elongated mass of light, extending about 71° in length, and 

 about !=■ T in breadth, on each side of the bright image. These 

 differences, however, arc probably owing to the different ways 

 in which the two minerals have been cut ; but it is not easy to 

 submit this point to direct experiment, on account of the diffi- 

 culty of procuring a mass of agate, from which a variety of 

 transparent prisms could be obtained. It follows, however, from 

 the theory of the depolarization of light, which I have explained 

 iu another pkicej and which is supported by all the evidence 



which 



