I On the ey^aUine Forms of Corundum, |4t 



jteflion of the difierent particles of the rock which enveloped them : this naturally render* 

 the furface rough and dull. Cryftals of feld-fpar, formed in the granitic porphyroid rocks, 

 exhibit the fame kind of appearance, from the fame caufe. 



The corundum cryftals are, in general, opaque, or at leaft they have only an imperfe£l 

 tranfparency at the edges; when broken into thin fragments, the pieces are femi-tranf- 

 parent ; when held between the eye and the light, and examined with a powerful lens, it will 

 be perceived that their interior texture is rendered dull, by an infinite number of fmall flaw%r 

 "Crofling each other, much refembling the medullary part of wood, when it is viewed in the 

 fame manner. The degree of tranfparency of the fmall interftices, which are between thefe 

 flaws, is further evidence that this texture of fmall flaws occafions opacity, which augments 

 in proportion to the thicknefs of the fragments. 



This kind of internal ftrufture has alfo a very ftrong analogy with that of feld-fpar in 

 granite and porphyry. The endeavour to fplit thefe cryftals in a direction, either perpen- 

 dicular, or parallel to their axes, meets with a very confiderable refiftance: they may, in- 

 deed, be broken in thefe diretlions ; but the rugged and irregular furface of the broken parts* 

 clearly proves that the dire£tion in which the cryftalline laminae have been depofited one upon 

 another has not been followed. 



The regular hexaedral prifm of thefe cryftals cannot therefore be confidered as the form 

 of the nucleus of the cryftal ; and, confequently, is not the primitive form of the cryftals of 

 this fubftance. 



If, in order to difcover the direftion of the cryftalline laminae, a variety of cryftals be 

 examined, fome will hardly fail to be met with, which, on their folid angles,, formed by 

 the junction of the fides of the prifm, with the planes of the extremities, prefent fmall 

 ifofceles triangles. Thefe are fometimes greater, and fometimes fmaller, and form folid 

 angles, of 1 22° 34', with the extreme planes of the cryftal. They are, in fome inftances, 

 real faces of the cryftal, but moft frequently they evidently are the efFeil of fome violence 

 on that part. The fmoothnefs and brilliancy of thefe fmall faces, in the latter cafe, fhew 

 that a piece has been detached in the natural direction of cryftalline laminae. It is, indeed, 

 much lefs difHcult to feparate a portion of the cryftal at thofc angles, than at any other 

 part } and in following the natural direftion of the faces, with a little patience and dexterity, 

 all the cryftalline laminse may be detached, and progrefhvely increafe the fize of the 

 triangular face. 



This operation, however, cannot be done jndifcrimlnately on all the folid angles of 

 the cryftals, but only on the alternate ones at the fame extremity, and in a contrary direc- 

 tion to each other. As to the other angles, they may be broken, but it is impoffible to de- 

 tach them. When, inftead of the folid angles of an hexaedral prifm, fmall triangular planes 

 are met with (which frequently happens, whether caufed by violence or otherwife), they 

 are always placed in the direction above mentioned. If by following this indication of na- 

 ture, we continue to detach the cryftalline laminie, we fhall at laft caufe the form of the 

 hexaedral prifm to difappear totally, and in place of it, a rhomboidal parallelopiped will be 

 obtained (fig. 2.), of which the plane angles at the rhombs will be 86° and 94°; the folid 

 Vol. II. — March i 799. 4 A angles 



