perfect kind, will emit a light of a very deep fire-colour, similar to that of a red hot iron.

5. The specific gravity of this stone varies in all its different kinds. The means deduced from a great number of observations afford the following numbers.

Imperfect corundum 3931. Perfect corundum, in the instance of Oriental ruby, 3977; and of sapphire, 4158. The difference seems to be proportionate to the degree of perfection of the crystallization, and consequently of the transparency of the stone.

6. We come now to the most extensive and most elaborate section of the paper which treats of the crystalline forms of the different kinds of corundum. The primitive form of all the kinds, whatever be their degree of perfection, we are here told is a rhomboid slightly acute, the obtuse angles of the planes measuring 94°, and the acute ones 86°; and it is asserted, that whatever the form of an individual crystal may be, it may always, by dividing it according to the lately established rules of crystallography, be ultimately reduced to this rhomboidal form. The manner in which crystals deviate from their primitive form, by the substitution of planes for the angles, effected by the retreat of rows of molecules, which constitute the crystalline laminæ, is amply discussed in a note; and nine modifications are described, forming a great variety of prismatic, pyramidal, and other crystals, of which some idea can only be formed by an inspection of the figures that accompany the paper.

7. The next section treats of the fracture and texture of this stone. In general we are told that all the kinds have a lamellated texture, the layers being in a direction parallel to the faces of the rhomboid, and that they break in a direction parallel to those faces. The ease, however, with which these laminæ may be divided, differs greatly in the different varieties; and this is ascribed to the degree of force existing in the attraction of the molecules which compose these crystals, as well as to the perfect adhesion of the crystalline laminæ composed of those molecules at all points of their surface. This attraction and adhesion, it is thought, varies with the colour of the stone, the blue or sapphire possessing those qualities in the highest degree, which accounts for the fracture of this stone being often in a direction oblique, and even at right angles to the planes of the laminæ.

8. The 8th section contains some observations on the phænomena of light exhibited by this stone. The prismatic, as well as the pyramidal crystals of corundum, when their extremities are terminated by planes which are perpendicular to their axes, very frequently exhibit on these planes a changeable variety of colours, known by the name of chatoyant. This property is ascribed to the reflection of light in the small intervals which remain between the crystalline laminæ in those parts where these laminæ are not in perfect contact. It follows hence that the most compact sorts of corundum will not exhibit this appearance. To the same property is also ascribed that beautiful reflection of light in the form of a star of six rays, frequently pro

duced in rubies and sapphires when cut in a particular direction. As to the manner of cutting these asteries, or star-stones, as they are usually called, it appears to be rather the effect of chance than of any determined theory. Some hints, however, are here given for the purpose, which chiefly recommend a proper attention to the primitive rhomboidal form of the crystal.

In the 9th section the author treats of the characters of the corundum afforded by chemical analysis; and here he anticipates the results of the investigation we are to be favoured with by Mr. Chenevix. These confirm what he has all along maintained in this paper of the identity of the several kinds of stones which he has classed under the name of Corundum.

The ingredients are uniformly found to be the same in all the species, differing only, and but in a small degree, in their proportions. The principal of these ingredients is argill or alumina, which, in the imperfect corundum from the Carnatic, Malabar, China, and Ava, consist of between 86 and 91 in 100 parts. In the sapphire it amounts to 92, and in the ruby to 90 hundredths. The other constituent parts are silica and iron; the former in the greatest proportion. Of the identity of stones which bear very different appearances, another example is here given in the instance of the felspar, reasons being assigned why some species of schorl, the adularia, and some other substances, ought to be classed with it.

The author proceeds next to consider the sort of corundum which does not exhibit the smallest rudiments of crystallization, and which mineralogists have agreed to distinguish by the name of Compact Corundum. It resembles, in many respects, a coarse jasper; but its much greater degree of hardness, and its much higher specific gravity, render its true nature easily distinguishable. It has a lamellated appearance. The red sort, in particular, gives pretty strong sparks when struck with steel. It is phosphorescent, like crystallized corundum. Its specific gravity is 3902.

Three sections treat next of the matrices of different sorts of corundum; which lead the author to dwell largely on a variety of substances which hitherto were thought to have no kind of affinity with this stone.

The first section treats of the matrix of imperfect corundum from the peninsula of India, chiefly from the Carnatic, and of the substances with which it is accompanied. This matrix, as far as our present knowledge extends, appears to be a stone of a particular nature; sometimes of a loose granulated texture, not unlike a coarse sandstone; and at other times of a closer grain, similar to the kind of marble known by the name of Coarse-grained Saline Marble: both kinds are of a pearly gray colour, sometimes slightly tinged with green, and have a degree of semi-transparency not unlike chalcedony. Their specific gravities are inferior to that of felspar. In this substance the crystals of corundum are imbedded, nearly in the same manner as those of felspar are dispersed in porphyry or certain granites. The accompanying substances are, 1. Lamellated fragments, not un

like felspar or adularia, and partaking of many of the properties of corundum itself, and even of its crystallization. 2. Fibrolites, which are described as small masses frequently crystallized, but different in some respects from all other mineral substances hitherto known. 3. Thallite, or the Epidote of Abbé Hauy. This is minutely described in three different states. 4. Hornblende, which is most constantly and most abundantly contained in the matrix here treated of. 5. Quartz, Talc, Mica, Steatite, Garnets, Zircon; all which, though manifestly dispersed in this matrix, are yet less frequent. And lastly, the presence of black Oxide of Iron is likewise evident, though not in such large proportions as in the matrix of the imperfect corundum from China.

Section 2. Of the matrix of imperfect corundum from China, and the substances with which it is accompanied.-This matrix is totally different from the preceding one, being a granite rock, composed of an aggregate mixture of felspar, fibrolite, mica, and attractable black oxide of iron: but none of that particular substance which has been mentioned as forming the basis of the preceding matrix is here observed. The four substances above mentioned are unequally distributed throughout the mass; some pieces being composed almost entirely of only one of them, while in other pieces the substances are mixed together in different proportions, and sometimes in nearly equal

ones.

Section 3. Of the matrix of perfect corundum from the Island of Ceylon, and the substances of which it appears to be composed.The author, on this head, speaks with some diffidence, as the precious stones comprised under the denomination of this kind of corundum are selected by the inhabitants from the sands washed down by the rivers or rivulets of the island, and have seldom been brought to Europe in any kind of matrix. He gives us, however, a list of the substances which compose the sands that are sent to us from Ceylon; although he will not venture to assert positively that these substances really accompany the corundum when in its matrix. They are, 1. Spinelle Ruby, which generally composes nine parts in ten of the whole mass of this sand, but in such small crystals or fragments as to render them of little or no consequence in trade; owing, no doubt, to the selection made in India before it is sent over. The few of a tolerable size that have been obtained, are here described as to their crystalline form, their colours, their peculiar matrix, in some of which was found an iron ore hitherto unknown, while some of them appeared evidently to be masses of adularia. 2. Tourmalin, in the form of a very obtuse rhomboid, with several varieties deducible from this form, and of different colours. 3. Ceylonite, so called by M. la Matherie, but distinguished by the name of Pleonaste in the Mineralogy of the Abbé Hauy. This is likewise of a variety of crystalline forms and colours. 4. Zircon, which, next to the Spinelle, is the substance most frequently found in the sand of Ceylon, the crystals of which, though very small, are yet in general very perfect. And lastly, though their numbers be very insignificant, some small scattered

fragments of quartz, felspar, calcareous spar, a brownish-yellow mica, and particles of attractable oxide of iron.

The paper closes with some account of corundum, which, contrary to the received opinion, that this stone was only found in the East Indies, has been thought to exist in other parts of the world. The author dwells mostly upon the appearances of a stone he himself discovered in the mountainous parts of the Forez in France, and which the Abbé Hauy considers only as a species of felspar. The Count alleges his reasons for classing it with the perfect blue corundum, known by the name of Sapphire. As to various stones found in Germany, in the Isle of Tirree on the western coast of Scotland, on Chesnut-hill near Philadelphia, and elsewhere, which have by some been considered as corundum, the author cautions us against acquiescing in those assertions till more conclusive arguments shall appear in their favour.

Analysis of Corundum, and of some of the Substances which accompany it; with Observations on the Affinities which the Earths have been supposed to have for each other, in the humid Way. By Richard Chenevix, Esq. F.R.S. and M.R.I.A. Read May 20, 1802. [Phil. Trans. 1802, p. 327.]

After a detail of several unsuccessful attempts to analyse this stone, which on account of its great hardness is both difficult to pulverize and to be reduced by saline agents, we find an ample description of the process, which was attended with the desired success. A piece of corundum, weighing 100 grains, was made several times red hot, and plunged into cold water; it was then pounded, first in a steel, and next in an agate mortar, and thus reduced into an impalpable powder. This powder was by means of dilute muriatic acid cleared from the ferruginous particles which adhered to it from the steel mortar. It was then put into a platina crucible with 200 grains of sub-borate of soda, and the mixture was exposed for an hour or two to a violent heat: the glass produced by this fusion was in about twelve hours dissolved, by boiling it in a proper quantity of muriatic acid.

The silica might now have been separated by evaporating the whole to dryness, but it was thought preferable to get rid of all the salts contained in the liquor by a precipitation effected by means of an alkaline carbonate. The precipitate thus obtained was then redissolved in muriatic acid, and the silica was hence cleared by evaporation. The remaining liquor was afterwards boiled with potash, by which means the alumina was precipitated. It was then redissolved by the excess of potash, from which the earth was finally obtained by muriate of ammonia. A small proportion of iron was separated by muriatic acid. Both these earths being now washed and dried, were ignited, and thus the exact weight of each was accurately ascertained. The author paid particular attention to the silica produced in this process; as Mr. Klaproth, who had formerly analysed this stone, declares that he never found any of this ingredient.

Next follow the tables of the contents of six kinds of corundum, viz. the sapphire, the ruby, and the corundum from the Carnatic, from Malabar, from China, and from Ava. The proportions of the first species are 54 silica, 92 alumina, 1 iron, and 1 loss. The proportions of the other kinds do not differ very considerably from these. The matrices of these stones being more easily fused than the six kinds above mentioned, the usual and well known mode of treatment by potash was found sufficient to render them soluble in acids. Although this mode be now very familiar to chemists, the author, however, in order to leave no cause for suspicion, describes the process he used with the matrix of the corundum from the peninsula of India. The results gave 42 silica, 371⁄2 alumina, 15 lime, 3 iron, and 2 loss, with a trace of manganese. By similar treatment the various substances contained in this and some other matrices, viz. felspar, fibrolite, and three sorts of thallite, were analysed, and the results are given in tables. It is remarkable that while all the other substances yield in different proportions the same ingredients as the matrices themselves, the fibrolite was found to consist only of silica and alumina, the quantity of iron it contained being so small as hardly to deserve notice.

In the prosecution of this inquiry Mr. Chenevix observed, that if a quantity of potash be for some time kept in fusion in a platina crucible, the latter will be found to lose some grains of its weight. The quantity of the metal thus lost he actually found in the potash; and hence he infers the affinity between these two substances, which affinity, it seems, is made use of by the Spaniards for detecting the platina contained in the ingots of gold sent from their American possessions. He also has occasion to show that potash which has usually been denominated a fixed alkali is not so, strictly speaking, since there is a degree of heat by which it may be totally volatilized. In a second part of the paper the author treats of the affinities which the earths are supposed to exercise towards each other when held in solution by acid or alkaline menstrua. There being a difference of opinion on this subject among some of the most eminent chemists, Mr. Chenevix has repeated many of their experiments, especially those of Guyton de Morveau. After descanting largely on the probable causes of error in this eminent chemist, as well as in Mr. Kirwan and others, he derives from his results the following general conclusions.

1st. That there exists an affinity between silica and alumina. 2ndly. That there exists a very powerful affinity between alumina and magnesia.

3rdly. That alumina shows an affinity for lime.

4thly. That Mr. Guyton was mistaken in every instance of affinity between the earths, except in one which had been observed before his experiments; and that he has attributed to a cause which does not exist, phænomena that must have resulted from the impurity of his re-agents.

And lastly, That neither the experiments of Mr. Guyton, nor an