the Corundum-Jlom from ^fta, 7 



ininate the hcxaedron : whereas, in this fecond modification, the decrcafe of molecules is by twc» 

 rows, which gives a more obtufe inclination, and forms new planes. The furface is ufually 

 ftriatcd, parallel to the fides of the planes which terminate this cryftal ; an appearance always 

 announcing imperfection in the cryftallization, arifing either from a change in the order of 

 dccreafe or increafe, or from a lefs perfed union of the cryftalline laminae. A fection would 

 Ihew gradual rifings or fteps, as appears in fig.' 14, which is a feflion of fig. 13, in the line 

 A D B. Thefc ftriae are not to be confounded with thofe in numberlefs fubftances, as in 

 tourmalines, fchorl?, &c. which arife from the longitudinal union of numberlefs diftindl cry- 

 ftals. The cryftal refulting from this new mode of decreafe in the cryftalline lamins, will 

 reprefent one or other of the varieties fhewn in f\g. 11, 12, and 13, according to the period 

 when fuch decreafe has begun in the procefs of the cryftallization ; and, i£ it has begun very 

 late, the new faces will only be fmall, nay, almoft imperceptible, ifofceles triangles, forming 

 folid angles of 160° 42' with the planes of the extremities of theprifm, as in fig. 5 j themea- 

 fure of the angles however muft be excepted. 



If this irregular mode of decreafe had begun with the firft cryftalline lamina?, which were 

 depofited on the primitive rhomboida! parallelepiped, the hexaedral prifm refulting therefrom 

 would have been terminated by two very obtufe triedral pyramids, wiiofe planes would have 

 been rhombs ; and they would have been placed in a contrary diredion to each other, as may 

 be feen in fig. I2, by the dotted lines. I have not met with this variety, but its exiftence may 

 be fuppofed. 



It happens fometimes, that the cryftallization has not been fo perfect as to defi:roy every 

 appearance of the faces of the primitive rhomboidal parallelopiped ; in this cafe there remains, 

 on the folid angle of ii2°, formed by the junction of the new faces with the edges of the 

 prifm, a fmall ifofceles triangle, as in fig. 13, which correfponds to thofe in fig. 5, of the pre- 

 ceding modification. 



The cryftals which explained the fecond modification, form alfo a part of Mr. Greville's 

 colledlion : one in particular, is highly worthy of notice; it is the moft perfe(St cryftal I have 

 ever feen of this fubftance. The furface of the faces of the prifm, although rough, is in-, 

 finitely lefs fo than that of the others, and much more brilliant. The planes on the ends 

 have the ufual polilh of cryftals ; its colour is a pale red, and its tranfparency may be com- 

 pared to that of wax. 



This fubftance prefents a third modification, in which the hexaedral prifm diminiflies in 

 diameter, as is apparent by comparing the diameter of its two ends ; in fome, it appears like a 

 regular hexaedral pyramid truncated (fig. 15). The cryftals of this modification are ufually 

 irregular, and feldom admit of a certain meafure of their angles ; but among the numerous 

 fpecimens, in Mr. Greville's colledtion, I have been able to afcertain, in the greater part, 

 that the hexagonal plane at the top, forms angles of about 1 2o<>, with the planes of the py- 

 ramid; and the hexagonal plane at the bafe, forms angles of about 78° with the planes of the 

 pyramid. In other inftances, the form of the pyramid varies greatly ; in fome the angle at 

 the upper plane was no*', and the angle at the bafe about 70°; in others, the angle at the 

 upper plane was about loo", and the one at the lower plane about 8of . 



