In former times when furnace shells were built not much more than a quarter as strong as they are to-day they were frequently burst by the expansion of the brickwork within them and it was therefore universal practice at one time to provide a space of 3 in. to 6 in. between the outside of the brickwork and the inside of the shell which was filled with some yielding material to allow for the expansion of the brickwork, and yet continually resist it. In modern practice, shells are so strong that they can resist the expansion of the brickwork and the latter is therefore laid up close to the shell and the small space necessary as a construction allowance is quite frequently grouted with concrete which, of course, is entirely unyielding. This means that the shell must be strong enough to resist the expansion of the brick or else it must stretch enough to accommodate the expansion of the brick within its elastic limit. Probably both these actions take place to some degree, but in any event this plan is not safe except with a shell of the strongest and best construction. If the brickwork were laid up tight inside a shell of older type the latter would be burst within a short time, as unfortunately has happened only too often even with the greatest precautions.

To prevent this is the reason for the custom, with such shells, of leaving a space between shell and lining and filling it with some material which is yielding and yet to some extent resistant, mentioned above. The necessity of resistance arises because if an open space were left between the brickwork and the shell the internal pressure would push the brick back, break the bond, and ruin the keying, so that some bricks would be likely to fall out, which would start a hole in the lining, and once started this would rapidly enlarge itself so as to destroy the lining. The bricks therefore must be held up to their places by some sort of filling between them and the shell.

Some furnacemen prefer for this purpose clay, but clay is only slightly compressible, and I prefer pumice cinder, made by running a light slag on to wet ground, or into a tank of water. The steam formed by the heat of the slag puffs the latter up into a huge sponge whose total volume is many times the volume of the slag from which it is made. As it is crushed under pressure its volume becomes smaller and smaller by the progressive destruction of the cells, first of the larger and then of the smaller sizes, but the material maintains its resistance to compression almost constant throughout a very great diminution of volume.

These are the conditions which we have to meet, as brickwork, unless rigidly bound, does not expand elastically like steel, but when expanded retains its new size, never contracting, or never to anything like the degree to which it had previously expanded, so that as its temperature varies it undergoes a slow but constant growth. Thus a furnace shell may be amply large at the beginning of a campaign and for several months or a year, but gradually, as the lining increases in size, when all the compressibility of the backing is gone, the direct force of expansion of the bricks is transmitted to the shell, and while the steel is stronger than the brickwork inch for inch, the latter is so much greater in area in any given section of the furnace that it generally bursts a thin shell when the "give" of the backing is gone.

In order to make the backing gas-tight the pumice cinder is sometimes mixed with clay, but this has the effect of reducing its compressibility especially as the clay is apt to burn hard after the furnace has been in operation for a while, and this therefore limits the amount of compression which a given portion of the backing can undergo without exerting undue stress upon the shell.