110 Proceedings of Royal Society of Edinhurgh. [jan. 6, 
for an instant, allowing the glass to cool very slightly j and imme- 
diately thrusts it hack for another charge which covers the first. 
This is repeated several times, till a sufficient quantity of glass has 
been gathered, when he puts the whole mass hack into the mouth 
of the pot, and turning it round and round heats it up preparatory 
to blowing. 
Now, on considering the matter, it seemed probable that these 
little cavities would form at the surfaces thus produced and exposed 
successively to the air. Any dust which might fall on the surface 
would give a starting place for a cavity ; or a place where the con- 
tracting glass would part under the negative pressure produced in 
the way I have described above. 
Accordingly, I asked Mr Griffin if he could manage to gather a 
ball of glass of considerable size, without bringing it out from 
beneath the cover of the glass-pot, and he very kindly made the 
attempt and soon succeeded, and produced the ball marked C, 
which, when cooled in the manner already described, exhibited 
three or four beautiful large cavities, but none of the small cavities 
which are possessed by the others. 
I do not propose to enter at all into a discussion of what I may 
be allowed to call the lessons to be learned from a study of these 
phenomena, though there are several points which are well worthy 
of consideration. 
I wish only to refer to two matters which may be thought of in 
this connection. The first of these is one which was pointed out to 
me by Mr W. H. Barlow, F.E.S., the engineer of the Tay Bridge, and 
a Fellow of this Society. Mr Barlow is a member of the Ordnance 
Committee, and he was greatly interested in examining these 
globes, and in thinking of the possibilities of similar flaws being 
produced during the cooling of large castings, such as those used 
in the construction of big guns. 
The second question to which I would direct attention is, 
perhaps, one which is already in the minds of all who have looked 
at these globes. It is the case of the cooling of a body like our 
earth. It seems certain that if the interior of the earth is of 
material which shrinks in cooling, cavities such as these would of 
necessity be formed by the shrinkage of the interior parts after an 
outside shell has become rigid. 
