DOUBLE REFRACTION IN ISOMORPHOUS SUBSTANCES. 261 



axis A of the maximum optical elasticity ; if, on the contrary, 

 a is more than 45°, the bisecting Hne is the axis C of minimum 

 optical elasticity. 



When one of the two quantities a and c are equal to h, the 

 direction of both the axes of elasticity, corresponding to equal 

 constants, becomes indeterminate in a plane perpendicular to the 

 third axis; the doubly-refracting medium has then only one 

 optical axis, which moreover coincides with that of maximum or 

 minimum elasticity, according as its proper numerical constant 

 is greater or less than the equal constants corresponding to the 

 indeterminate axes which are perpendicular to it. 



I have determined the direction of the three axes A, B, C, the 

 actual or relative values of the coefficients «, 6, c, by the ordinary 

 methods based upon the well-known laws of double refraction 

 and chromatic polarization. This determination is more or less 

 complicated for the various systems of crystals. I will briefly 

 state the course which I have followed in my experiments. 



In crystals which may be referred to a right quadratic prism, 

 to a rhombohedron or a hexagonal prism, the direction of the 

 single axis is known beforehand ; it therefore remains only to 

 determine the two principal indices, or at least their order of 

 magnitude. 



Wherever the refraction was measurable, I have always deter- 

 mined the indices directly. In the first place, I cut prisms in the 

 proper direction ; and on account of the striae, always left after 

 polishing a salt, cemented upon their refracting surfaces either 

 small fragments of thin sheet glass or laminae of gypsum. Per- 

 fect juxtaposition may be effected by means of soft cement. I 

 have, moreover, in all instances operated with at least two sepa- 

 rate prisms, and have not considered any results sufficiently 

 accurate unless they agreed with the calculated indices in the 

 second decimal place. The goniometer employed for measuring 

 the minimum deviation was that of M. Babinet, an instrument 

 well adapted for this purpose; but as the prisms were very 

 small and the light feeble, the opake thread at the focus of the 

 collimator was replaced by a luminous slit ; the readings were 

 made in the yellow part of the spectrum. 



When the measurement of the indices was impossible, 1 was 

 contented with determining the character of the optical axis by 



