Dr Brewster on the Structure of Apo^liyllite. 7 
perpendicular to the axis, I was surprised to observe the tessel- 
lated structure shewn in Fig. 8, and to find that each of the six 
equilateral triangles had two axes of double refraction, of the 
same character and intensity as those which occurred in the 
rhomboidal prism. The general tints in the triangles AOB, 
DOE, descended to zero by a series of fringes at their junctions 
with the adjacent triangles, and the plane of the resultant axis 
bisected each of the angles at the centre of the hexagonal sec- 
tion. These facts put it beyond a doubt, that the bipyrami- 
dal dodecahedron was, in this case, neither derived from the 
rhomboid, which has always one axis of double refraction^', 
nor was the primitive form of the crystal ; but v/as form- 
ed by some new process, which had not been recognised by mi- 
neralogists. The perfect similarity between the properties of 
the rhomboidal prism and the bipyramidal crystal, suggested 
the idea that the latter might be composed of the former, in 
such a manner that each of the opposite faces of the pyramid 
were the two sides of the prism. Upon measuring the angle 
formed by these opposite faces, I found it to be 114®, the very 
same as the angle of the rhomboidal prism ; and, as I had ob- 
served that these two forms were the product of the same cry- 
stallization, and that the one was sometimes urrited to the other, 
I had no longer any doubt that this was the true origin of the 
bipyranridal dodecahedron of sulphate of potash. 
A structirre somewhat analogous to the one we have now de- 
scribed appears in the cruciform Harmotome, which Rome de Lisle 
supposes to consist of two compressed dodecahedrons, crossing 
one another at right angles, so that their axes are coincident. 
Hairy *[• considers this opinion as highly probable ; but he at 
the same tirue remarks, that the cruciform variety may still be re- 
garded as a simple crystal. In order to determine this point, I 
formed two planes perpendicular to tire axis, and having exposed 
the crystal to polarised light, I was enabled not only to confirm 
the conjecture of Rome de Lisle, but also to ascertain the man- 
ner in which the mutual penetration of the crystals takes place. 
* Since the publication of my paper on the laws of Polarisation and Double Re- 
fraction, in the F/itl. Trans, for 1818, I have ascertained that crystals whose pri- 
mitive form is the acute rhomboid, and the bipyramidal dodecahedron, have only 
one axis of double refraction. 
Tr4dt& d& Mineralogies tom. iii. p, IPT. 
