37'2 PHILOSOPHICAL TRANSACTIONS. [ANNO 17QS. 



prism disappears, (fig. 8;) and this cut or section forms an angle of 122° 34', with 

 the plane on the extremity. It is unnecessary to observe, that the regularity of 

 the hexaedral prism, depends on that of the rhomboidal parallelopiped on which it 

 is formed. 



When, by detaching the laminae from the alternate solid angles of the regular 

 hexaedral prism, the planes resulting from this operation begin to run into each 

 other, and the crystal begins to assume the form of the rhomboidal parallelo- 

 piped to which it owes its origin, we frequently see the surface of these new planes 

 divided into an immense number of small rhombs, formed on them by the inter- 

 section of lines that are parallel to the sides, which belong to the rhomboidal form 

 of the new faces: (fig. 9.) These lines are owing to the extremities of the laminae 

 which have been deposited on the inferior faces, corresponding with those on which 

 we observe them ; and they serve to corroborate still further, the demonstration we 

 have given of the formation of the regular hexaedral prism in this substance. We 

 frequently see small rhombs traced on the surface of the planes, on the ends of the 

 hexaedral prism; (fig. 10.) This, no doubt, is occasioned also by the intersection 

 of the laminae, on the planes of the primitive rhomboidal parallelopiped. But 

 these rhombs, formed by the re-union of lines that join in angles of 6o° and 120°, 

 instead of 86° and Q4°, (like those we have seen traced on the faces which corres- 

 pond with those of the rhomboidal parallelopiped,) form angles of 6o° and 120°. 

 It would therefore be an error to consider them as indications of the form of the 

 elements of crystallization, as we are tempted to do from a simple inspection of 

 the crystal. These same lines form equilateral triangles with each other, as may 

 be seen in fig. 10. 



The cause of these small equilateral triangles, which sometimes project a little 

 over the planes on the ends of the prism, must now be obvious. If, during the 

 superposition of the crystalline laminae on all the planes of the rhomboidal paral- 

 lelopiped, it has happened, from any cause whatever, that the laminae deposited on 

 the 3 faces of the same summit, have not fallen exactly on those which preceded 

 them, or that they have experienced some deviation, or have not had the same 

 decrease as all the others, at the angle of 86°, these triangles must necessarily 

 occur; in the same manner it must be obvious why these small equilateral tri- 

 angular projections are frequently placed on one of the sides of the crystal. The 

 primitive form of the corundum crystal is therefore a rhomboidal parallelopiped, 

 whose solid angle at the summit is 84° 31', and that formed by the re-union of the 

 bases is 95 29'. The crystalline laminae are rhombs of 86° and 94 : these, in my 

 opinion, are double crystalline molecules; the single molecules I apprehend to be 

 isosceles triangles, of 86°, at the angle of the summit, and of 47° at those of 

 the base.* Though the rhomboidal parallelopiped of 86°' and 94° is the primitive 



* I am at present preparing a work, in which I shall, if circumstances permit me to finish it, give 

 the result of my observations, and my own opinion on this interesting part of mineralogy. I shall only 

 observe here, that though double molecules, square and rhomboidal, are frequently formed in the 



