154 Henry Clifton Sorby, Esq., on the 



We will begin our calculations with the mean specific gravity of 

 the square prismatic form, which is 1*773. Considering the rela- 

 tive frequency of the angles found in anthracite, I should say that 

 its best cleavage is parallel to the faces of the obtuse octahedron, 

 having its axes as 6, 5, and 3. Diamond also cleaves parallel to the 

 planes of the regular octahedron, and therefore it is most probable 

 that the atoms in both cases are arranged in a similar manner, viz., 

 in rows parallel to the faces of the octahedrons, so that their mutual 

 arrangement on those faces would be as shewn by figs. 5 and 8, 

 only in the diamond being spheres, and in the square prismatic form 

 spheriods with their axes in the ratio of 3 to 5 ; the former being 

 the axis of rotation of the ellipse generating it, which would be placed 

 in the crystal parallel to the axis of the prism ; and, therefore, the 

 section parallel to the faces of the octahedron would not exhibit 

 ellipses with axes of those relative values, but as shewn by fig. 8. 



Now, I have found by experiment, that the specific gravity of the 

 square prismatic form is 1*773, and this multiplied by 2 is 3-546. 

 The mean specific gravity of diamond is 3*521, which differs from 

 that deduced above by only jioth, which is much less than what 

 occurs in the statements of difi^erent experimenters. Whence I 

 think we may conclude that the atoms of the square prismatic form 

 and diamond are, to one another in their relative volume, as |^ to |. 



Again, supposing the atoms of coke to be spheres arranged cubi- 

 cally, so that on each face they would be related to one another in 

 the manner shewn by fig. 7, and those of diamond to be as above 

 described, and also that their relative volumes are as \ to \, it is 

 easily shewn mathematically, that their relative specific gravities 

 would be as 3 to 4^2. Calculating on these suppositions, and taking 

 the specific gravity of diamond at that above calculated from that of 

 the square prismatic form, the specific gravity of coke would be 

 1*880, which diff'ers from 1*891, found by experiment, by only y|oth. 

 Whence the volume of the atoms of coke is to that of those of diamond 

 as ^ to I, and to that of those of the square prismatic form, as ^ 

 to I. 



Again, supposing the atoms of graphite to be either spheres or 

 ellipsoids, having the same volume as the spheres in coke, but arranged 

 as a regular six-sided prism, if ellipsoids with their axes parallel 

 to the prism, that is to say, hexagonally parallel to the terminal 

 planes, as shewn by fig. 6, and rectangularly in the direction perpen- 

 dicular to them, which structure agrees perfectly with the fact of its 

 cleaving only parallel to the terminal planes, it is easily shewn 

 mathematically, that their relative specific gravity would be as 2 to 

 V3. Calculating on these suppositions, and taking the specific 

 gravity of coke at that calculated from that of the square prismatic 

 form, the specific gravity of graphite would be 2*172. The mean 

 found by experiment, as given above, is 2-177, which differs from 

 that calculated by only j^^th. 



