Dr. Eqfon on Cryjlallization. 39 
by a like number of pillars, which differ from 
each other both in fhape and in fize, as not any 
two of them have been found alike in every 
refpeCt. Thefe pillars are moreover divided into 
joints, at unequal diftances, which, by the afiift- 
ance of a crow of iron, may be forced afunderj 
and, what is very remarkable, a pillar, feparated 
at thofe joints, always (hews one end convex, and 
the other concave; the convexity being fome- 
times turned from, and fometimes towards the 
horizon •, but, in any Angle pillar, the direClioc 
is always the fame. 
Have we not every reafon to conclude, that 
the ifland of Staffa, and the Giant’s caufeways 
have been productions of volcanic matter, made 
liquid by fome fubterraneous fire, and, as foon 
as it cooled by flow degrees, cryftallized into that 
form which they now exhibit ? Had the fepara- 
tion of thefe pillars been owing to accident, they 
would have appeared like cracks in the earth, 
formed during a very dry feafon, without order, 
beauty or regularity j whereas the reverfe is very 
confpicuous. 
By attending to the fails and obfervations 
already related, it will, I am perfuaded, appear, 
that the fetting of metals is, in reality, a fpecies 
of cryliallization. This idea is ftrongly con,, 
firmed, by the laminated' texture which all the 
femimetals difcover, when broken in any direc¬ 
tion. The needle-like cryftals, fp conspicuous 
E 4 it) 
