4 Dr. Brewster's additional observations on the 
In order to observe the manner in which the cavities dis- 
appeared, I suspended one of the drops by a wire, and viewed 
it with a telescopic microscope when exposed to a strong heat. 
Soon after the drop became red hot, the cavities gradually 
contracted, and at last vanished, the centre of the cavity being 
the part that was last filled up. The drop had begun to melt 
at its smaller extremity, but the lines represented in fig. 2 
were still visible, the heat probably not having been suffi- 
ciently intense to affect its superficial structure. 
As the specific gravity of the crystallized drop is nearly 
the same as that of the annealed drop, the cavities must be 
produced by the contraction which the internal part experi- 
ences in cooling, for the sudden induration of the outer layer 
prevents the contraction from taking place in any other way. 
The manner, too, in which the cavities disappear, is a complete 
proof that they contain no air, and hence we may consider 
their magnitude, which increases with the size of the drop, as 
a measure of the contraction which the glass undergoes in its 
transition from the temperature at which it melts, to the ordi- 
nary temperature of the atmosphere.* 
I am informed by Dr. Hope, that he has frequently ob- 
tained unannealed drops of crown glass, in which there were 
no vacuities, and that they all burst spontaneously in the 
course of a few months. As there is at present no crown 
glass manufactory in this part of Scotland, I have not been 
able to make any experiments with drops of this kind ; but 
there is every reason to believe that they would exhibit the 
same optical properties, as those which are formed of flint 
* Upon this supposition, the contraction of glass in bulk, in passing from the first 
of these states to the second, will be or 0.02205. 
