6 On the eryftatttne Forms of 



rhombs, formed by the reunion of lines that join in angles of 60' and liO", inftcad of 86° and 

 94" (like thofe we have feen traced on the faces which correfpond with thofe of the rhomboidal 

 parailelopiped), form angles of 60° and 120°. It would, therefore, be an error to confider thetn 

 as indications of the form of the elements of cryftallization, as we are tempted to do from a 

 fimple infpeiflion of the cryftal. Thefe fame lines form equilateral triangle^ with one another, 

 as may be feen in fig. 10. 



The caufe of thefe fmall equilateral triangles, which fometimes projefi: a little over the 

 planes on the end of the prifm, muft now be obvious. If, during the fuperpofition of the cry- 

 italline laminae on all the planes of the rhomboidal parallelopiped, it has happened from any 

 caufe whatever, that the laminx depofited on the three faces of the fame fummit, have not 

 fallen exadlly on thofe which preceded them, or that they have experienced fome deviation, or 

 have not had the fame decreafe, as all the others, at the angle of 86°, thefe triangles muft ne- 

 ceflarily occur; in the fame manner it muft be obvious, why thefe fmall equilateral triangular 

 projedions are frequently placed on one of the fides of the cryftal. 



The primitive form of the corundum cryftal is therefore a rhomboidal parallelopiped, whofc 

 folid angle at the fummit is 84° 31', and that formed by the reunion of the bafes is 95° 29'. 



The cryftallinc laminas are rhombs of 86° and 94.°: thefe, in my opinion, are double cry- 

 ftalline molecules; the Angle molecules I apprehend to be ifofceles triangles, of 86" at the 

 angle of the fummit, and of 47° at thofe of the bafe*. 



Although the rhomboidal parallelopiped of 86° and 94° is the primitive form of the corun- 

 dum cryftal, yet it is rare to meet with that fubftance under this perfe£l: and determined form ; 

 and in moft mineral fubftances, it is more rare to meet with their primitive cryftals, than their 

 different modifications. Amongft Mr. Greville's numerous fpecimens of corundum, I have 

 met with only one which has this primitive form, and it is doubtful whether even this may 

 not be a fragment. 



The corundum cryftal prefents another modification, under which the regular hexaedral 

 prifm, inftead of having three alternate folid angles, at each of its ends (on which folid angles 

 are placed ifofceles triangular planes, forming a folid angle of 122° 34' with the planes at the 

 extremities upon which they are inclined), has alfo its angles fupplied by ifofceles triangular 

 planes; but thefe planes, inftead of 122? 34', form folid angles of 160" 42,' with the faid 

 planes on the extremities. (See fig. 11 and 12.) Thefe new planes, which conftitute a new 

 modification of the primitive form of corundum, are the refult of a different order, in the 

 decreafe of the laminae ; which, in the primitive form, are depofited on the planes of its pri- 

 mitive rhomboid by finglc rows of cryftalline molecules, and increafe the planes which ter- 



* I am at prefent preparing a work, in which 1 ftall, if circumftances permit me to (inifh it, give the refult 

 of my obfervations, and my own opinion on this interefting part of mineralogy. I fliall only obferve here, 

 that although double molecules, fquare and rhomboidal, are frequently formed in the procefs of cryftalliza- 

 tion, yet the real form of the cryftalline molecules feems to be triangular. By obferving the progrefs of the 

 rhomboidal parallelopiped, in its paffage to the form of an hexaedral prifm (fig 4 and 5), and by confidering 

 the prifm terminated, it feems evident, that the laft laminae which had been depofited, after the progreffive 

 decreafe in the rows of cryftalline molecules to one finglc molecule, muft neccffarily have been triangular. B. 



sninate 



