0 Dr Brewster on the Structure of ApophylUte. 
in some cases from the union of smaller crystals, whose homolo- 
gous sides are not parallel to each other. The Abbe Hauy 
very ingeniously conjectured that the hexaedral Arragonite 
was a compound crystal of this kind, consisting of four 
rhomboidal prisms ; and Mai us has shewn that such a com- 
bination is quite compatible with its doubly refracting struc- 
ture. If Arragonite had only one axis, as Malus believed, the 
mineralogical hypothesis would not have been at variance with 
the optical results ; but as I have elsewhere shewn * that Arra.- 
gonite has two axes, and that their united action is not symme- 
trical round the axis of the prism, it follows that the hypothesis 
of Hatiy is radically incompatible with the optical phenomena of 
Arragonite. 
These remarks are equally applicable to every hypothesis re- 
specting the formation of Arragonite, in which a given line in one 
crystal, lying in the plane of its resultant axes, is not parallel to a 
line similarly situated in all the other crystals. In many Arra- 
gonites which I have examined, the hexaedral prism is composed 
of irregular portions, one set of which has the plane of their re- 
sultant axes in a direction different from the other set ; but 
there was no appearance of a tessellated structure, as will be seen 
from Figures 9 and 10, where the shaded parts represent thosQ 
portions which have the plane of the axes in the direction AB^ 
while the white parts have the plane of their axes in the direc- 
tion AC. 
In examining the optical phenomena of Sulphate of potash, I 
have discovered a singular tessellated structure, which may 
throw considerable light upon that complicated crystalline form, 
the Bipyramidal Dodecahedron. My first experiments on this 
salt were made with rhomboidal prisms, whose angles were 114^ 
and 66 which I found to have two axes of double refraction. 
Among other crystals of sulphate of potash, I found hexaedral 
prisms and bipyramidal dodecahedrons f. The former had only 
one axis, and I therefore expected that the bipyramidal crystal 
would have the same property. Upon cutting a plate, however, 
« See Phil. Trans. 1818, p. 201, 202; — ax\di the Journal of the Royal Institu» 
tion, vol. iv. p. 112. 
*}' Count Bournon makes the primitive form of sulphate of potash the bipyrami» 
dodecahedron. See his Catalogue-) &c. p. 181. 
