192 New Properties of Light. [March, 



paper on the subject ; but some idea of it may be formed from the 

 following brief statement. 



When the polarised image of a candle is doubled, by viewing it 

 through a prism of Iceland spar, the images vanish alternately in 

 every quadrant of the circular motion of the priim. If the vanished 

 image is depolarised or restored, by the interposition of a plate of 

 topaz, none of the images will vanish during the rotatory motion of 

 the prism. Now the topaz being a doubly refracting crystal 

 actually produces two images of the polarised candle, and tlierefore 

 upon turning the topaz round one of these images must vanish in 

 every quadrant, though from the images not being separated this 

 vanishing is invisible. Let the topaz be fixed in that position in 

 which none of the two images should vanish, which is its depo- 

 larising position, and let the apparently single, though really double 

 image, be viewed through a prism of Iceland spar. This prism will 

 now form two images of the double images, each of which consists 

 of two single images, and on turning the prism round one single 

 image of each of the double ones will vanish in eveiy quadrant; 

 but two images will always remain, as if no change had been going 

 on. The depolarisation of light is, thel-efore, the necessary conse- 

 quence of the polarising faculty of the interposed topaz, and could 

 not have existed imless the topaz had been a doubly retracting 

 crystal. This theory is capable of the most satisfactory proof by 

 substituting a rhomb of Iceland spar; a plate of agate, or a bundle 

 of glass plates, in room of tlie topaz ; or by transmitting a polarised 

 ray along the axis which joins the obtuse solid angles of a prism of 

 Iceland spar. 



All bodies, therefore, that depolarise light must necessarily give ■ 

 double images polarised in an opposite manner, like those formed 

 by calcareous sjjar. A great number of these bodies, such as rock 

 crystal, topaz, &e. exhil)It two images ; but others, such as mica, 

 amber, ice, gum aral)ic, caoutchouc, &c. present no direct indica- 

 tions of double refraction. In this latter class, therefore, the one 

 image lies above the other, and they may, or may not, be produced 

 by ditferent refractive powers. 



In regularly crystallized bodies, like topaz, &c. the neutral and 

 depolarising axes of one plate uniformly coincide with the neutral 

 and depolarising axes of all tlie rest, so that the crystal itself, 

 though composed of numerous plates, has still two neutral and two 

 depolarising axes. In substances, however, formed by successive 

 layers, like caoutchouc and gum arable, which liave no neutral 

 axes, but wliicli depolarise light in every position, the first layer is 

 deposited and crystallized so as to have two neutral and two depo- 

 larising axes; but there is no cause to determine that the second 

 layer shall have its axes coincident with those of t!ie first ; so that 

 after a number of layers are formed there will be a depolarising axis 

 in every direction. This reasoning is capable of direct confirmation 

 from an experiment, which 1 have described in my Treatise on New 

 i'liilosuphical lustrumcnts. If we take several films of mica, and 



