163 
cific gravity by only ^th. The volume of these two forms 
is, therefore, as J to J. 
Suppose the spherical atoms of diamond to be arranged 
as an octahedron, Fig. 1, in that case they would be placed 
with relation to one another on each equilateral triangular 
face, in the manner shown in Fig. 5 ; and such an arrange- 
ment agrees perfectly with the well known cleavage of this 
mineral. Also let us suppose the atoms of coke to be 
spheres, but arranged as a cube, Fig. 3, in which case they 
would be related to one another on each of its square faces, 
as shown in Fig. 7. On these suppositions it is easily shown 
mathematically that the relative volumes would be as 1 ; V2 ; 
and calculating on this supposition, and also that the volumes 
of the ultimate spheres in coke are to those in diamond as 
^ : we should have for the specific gravity of coke, by 
calculation, 1.880. Four very careful experiments, which 
only differed by ^th from one another, gave me a mean 
specific gravity of 1.891, which only differs from that calcu- 
lated by T |oth. 
Again, supposing the spheres of coke and graphite to be 
of the same magnitude, but in the latter arranged as an 
hexagonal prism, Fig. 2, in that case they would be related 
to one another on the terminal planes, in the manner shown 
by Fig. 6, and rectangularly in the direction perpendicularly 
to them, a structure indicated by the crystalline form and 
cleavage. If such be supposed, it is easily shown that their 
relative specific gravities should be to one another as </3 : 2; 
and this gives for the specific gravity of graphite, by calcu- 
lation, 2.172. The mean of six careful experiments gave me 
2.177, which only differs from that calculated by $Joth» 
I think, therefore, that these calculated results agree so 
very well with those found by experiment, that I am entitled 
to conclude, as I before said, that the four species of carbon 
are caused by its existing in atoms of the relative volumes 
of J, and J, arranged in such a manner as the form and 
