Mr. keir on the 
53 % 
tinued long, and the tranfition of the glafs from a fluid 
to a folid ftate was very ilowly accomplilhed. The upper 
part of this glafs was found to be changed into a white, 
opaque, or rather femi-opaque fubftance, refembling in 
colour and texture fome of the white fpars. Under this 
cruft, which, in fome places, was a quarter of an inch 
thick, and in others more, the glafs was traniparent, but 
confiderably oblcured, and its colour was changed from a 
dark green to a dull blue. Ifi this femi-pellucid glafs were 
difperfed many white, opaque, regular cryftals, the form 
of which was generally that of a folid, whole fide-view 
is reprefented by fig. 1. and whole balls by fig. 2. The 
lurface of thefe cryftals feems to be bounded by lines 
rather elliptical than circular, which are fo dilpofed, that 
a tranfverfe lection of a cryftal, that is, a fe<ftion perpen- 
dicular to its axis, is an hexagon, as is lliewn in fig. 3. 
and 4. the former of which reprefents a view, and the 
latter a plan, of that fecftion. In the middle of each balls 
of the cryftal, a conical cavity appears, as is lliewn in fig. 
1 . and 2. The elliptical lines which bound the furface of 
the cryftals feem to be occafioned by the edges of many 
thin plates, fo arranged round the axis of each cryftal, 
that their longitudinal diameters are parallel to that axis. 
Of thefe plates, twelve are larger, more confpicuous, and 
better defined, than the reft. They are placed in pairs, 
at an equal diftance from each other, forming the fix 
angles of the hexagonal fe<ftion and bafis, as appears in 
fig. 1. 2. 3. and 4. The intervals between the pairs of 
plates, that is, the areas of the triangles into which the 
hexagonal 
