556 
Proceedings of the Royal Society 
any such, action in their experience. Movements of the liquid 
metal, after pouring, have been occasionally noticed, but these are 
accounted for by movements of cores or air; the necessity for 
immediately freeing the metal from the cores on setting, is pointed 
out as against any such supposed expansion, as the tendency is 
rather to shrink and tighten up against the inner part or core. 
One experiment which the author has recently made may he 
noted as pointing out the great difficulty in dealing with this 
subject, so as to get rid of disturbing causes such as air or gas 
entangled in the metal. 
The metal was run into a mould 8 in. x 5 in x 4 in., and a 
wrought iron bar about 18 in. x 1 in. x 1 in. was laid across thus — 
and the metal brought to about touching the bar ; after about a 
minute a thin moulder’s trowel could be almost passed between the 
bar and the still molten metal. Ai4 er setting, the trowel could 
hardly be put in as before, and the metal thereafter began to rise 
up on each side of the bar thus — 
The metal had now assumed an orange colour, the part 
below the bar being brightest. On examining the metal when 
cold it was found that the space or difference between the high 
and low parts was J of an inch. A few air holes being observed 
on the top of the casting, the metal was cut through at the 
raised and flat parts, when it was found that some honeycombing 
existed, due apparently to air bubbles. 
In some other experiments made, this uprise was not observed, 
and it appears probable that the uprise of the surface of the metal, 
as above described, was due to the pressure of air or gas, which by 
compression due to shrinkage was forced upwards, no other means 
