OF RADIAN'P HEAT THROUGH CRYSTALS. 101 



does not however possess the same demonstrative value as that 

 obtained in the case of rock-crystal, inasmuch as the mass of 

 the beryl was in some measure unhomogeneous. 



A cube of tourmaline cut, as regards its axis, exactly as the 

 former, and subjected to the same conditions of experiment, 

 showed an opposite deportment. In this case the quantity of 

 heat transmitted perpendicular to the axis was greater than the 

 quantity transmitted parallel to it. 



The deflections of the galvanometer by which these quantities 

 were measured are to each other in the proportion of 100: 158, 

 the former number denoting the quantity transmitted along the 

 axis, and the latter the quantity transmitted in a line perpendi- 

 cular to the same. The homogeneous character of the specimen 

 experimented with, does not permit of the idea that the observed 

 difference is due to any accidental irregularity of the mass of 

 the crystal. 



The foregoing experiments, therefore, prove that, in certain 

 crystals, radiant heat is transmitted in unequal quantities in dif- 

 ferent directions. 



It seemed to me of interest to pursue this inquiry further, by 

 applying polarized calorific rays* instead of the natural ones. 

 For this purpose the rays were caused to pass through a Nichol's 

 prism, previous to falling upon the cube of rock-crystal. The 

 prism was first so placed that the principal section, passing 

 through its obtuse angles, was vertical, and hence the plane of 

 polarization of the issuing rays horizontal. 



When the calorific rays thus polarized were permitted to 

 traverse the cube in the line of the crystallographic axis, a de- 

 flection of 8°*1 was produced. When the line of transmission 

 was perpendicular to the axis, the deflection was 8°* 3 ; the dif- 

 ference between the«e two values lies within the limits of the 

 error of observation, and hence they must be regarded as equal. 



Hence when heat, polarized in the manner indicated, is ap- 



• Strictly speaking, the heat as applied in the foregoing cases was not in a 

 natural state, but polarized elliptically by reflexion from the steel mirror of the 

 heliostat. 



But the peculiarity communicated in this way (the angle of incidence being 

 so small), compared with the linear polarization to which our experiments refer, 

 is of so vanishing a character, that, as far as our results are concerned, the light 

 may be regarded as being in its natural state. 



SCIEN. MEM.^Nat, Phil. Vol. I. Part II. I 



