VOL. LXXXVIII.] PHILOSOPHICAL TRANSACTIONS. 373 



form of the corundum crystal, yet it is rare to meet with that substance under this 

 perfect and determined form ; and, in most mineral substances, it is more rare to 

 meet with their primitive crystals than their different modifications. Among 

 Mr. Greville's numerous specimens 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 crystal presents another modification, under which the regular 

 hexaedral prism, instead of having 3 alternate solid angles at each of its ends, (on 

 which solid angles are placed isosceles triangular planes, forming a solid angle of 

 122° 34', with the planes at the extremities on which they are inclined), has also 

 its angles supplied by isosceles triangular planes; but these planes, instead of 

 122° 34', form solid angles of l6o°42 / , with the said planes on the extremities. 

 (See fig. 11 and 12). These new planes, which constitute a new modification of 

 the primitive form of corundum, are the result of a different order in the decrease 

 of the laminae; which, in the primitive form, are deposited on the planes of its 

 primitive rhomboid by single rows of crystalline molecules, and increase the planes 

 which terminate the hexaedron : whereas, in this 2d modification, the decrease of 

 molecules is by 2 rows, which gives a more obtuse inclination, and forms new 

 planes. The surface is usually striated, parallel to the sides of the planes which 

 terminate this crystal; an appearance always announcing imperfection in the crystal- 

 lization, arising either from a change in the order of decrease or increase, or from 

 a less perfect union of the crystalline laminae. A section would shew gradual 

 risings or steps, as appears in fig. 14, which is a section of fig. 13, in the line adb. 

 These striae are not to be confounded with those in numberless substances, as in 

 tourmalines, schorls, &c. which arise from the longitudinal union of numberless 

 distinct crystals. The crystal resulting from this new mode of decrease in the 

 crystalline laminae, will represent one or other of the varieties shown in fig. 11, 

 12, and 13, according to the period when such decrease has begun in the process 

 of the crystallization; and, if it has begun very late, the new faces will only be 

 small, nay almost imperceptible isosceles triangles, forming solid angles of 

 l6o° 42', with the planes of the extremities of the prism, as in fig. 5; the mea- 

 sure of the angles however must be excepted. 



If this regular mode of decrease had begun with the first crystalline laminae 

 which were deposited on the primitive rhomboidal parallelopiped, the hexaedral 

 prism thence resulting would have been terminated by 2 very obtuse triedral pyra- 

 mids, whose planes would have been rhombs; and they would have been placed in 

 a contrary direction to each other, as may be seen in fig. 12, by the dotted lines. 

 I have not met with this variety, but its existence may be supposed. It happens 



process of crystallization, yet the real form of the crystalline molecules seems to be triangular. By 

 observing the progress of the rhomboidal parallelopiped, in its passage to the form of an hexaedral 

 prism, (fig. 4 and 5), and by considering the prism terminated, it seems evident, that the last lamina 

 which had been deposited, after the progressive decrease in the rows of crystalline molecules to one 

 single molecule, must necessarily have been triangular. — Orig. 



