DOUBLE REFRACTION IN ISOMORPHOUS SUBSTANCES. 263 



fore in most instances been satisfied with determining the angle 

 of the optical axes and the character of the bisecting line, and 

 shall describe further on the details of the processes which I 

 have adopted. 



In crystals which are referable to an oblique rhombic prism, 

 one of the axes of optical elasticity is always parallel to the hori- 

 zontal diagonal of the base ; it therefore remains to determine 

 the direction of the other two axes. For this purpose I adopted 

 two different methods. 



The method which I shall call the first, consists in cementing 

 two crystals together by a plane parallel to the horizontal diago- 

 nal of the rhombic base, taking care moreover to place them in 

 an inverse position. An accurate adjustment may be obtained 

 by allowing reflected light to fall upon the surfaces, which, not- 

 withstanding the torsion, must remain parallel, and thus ascer- 

 taining with certainty that the surfaces parallel to these diago- 

 nals belong to one and the same zone in both parts of the arti- 

 ficial hemitropic group. 



When the surfaces placed in juxtaposition are smooth, and 

 the cement by which they are joined sufficiently soft, it is not 

 difficult to effect an almost perfect coincidence. The cement 

 which I employed for this purpose is Canada balsam, which 

 hardens slowly on exposure to the air, or a thick solution of 

 gum-arabic. This latter substance cements certain crystals very 

 firmly, for example efflorescent salts, for which Canada balsam 

 is less suitable. 



From the hemitropic group I cut a plate with parallel sur- 

 faces perpendicular to the common horizontal diagonal of the 

 rhombic base ; and when it was to be cemented with Canada 

 balsam, ground it with water, or with alcohol when it was to be 

 cemented with gum. The halved plate included within its plane 

 two axes of optical elasticity belonging to each crystal ; and it is 

 obvious that from one crystal to the other the homologous axes 

 form an angle with each other, which is twice as great as their 

 inclination to the plane of the artificial hemitrope. 



The divided plate was then placed in an apparatus very similar 

 to that employed by M. Biot* in studying the colours developed 



• Precis elementaire de Physique experimentaley 3rd edit., vol. ii. p. 475, 

 and pi. 8. fig. 2. 



