This is a subject deserving the most careful consideration. Stoves in modern practice are universally arranged in one straight row, where considerstions of room permit, but this exposes the maximum amount of surface to radiation, takes up the most room, and requires the longest hot-blast main.

About thirty years ago the late Samuel Thomas built the Pioneer Blast-Furnaces now owned by the Republic Iron and Steel Company in Birmingham, with the four stoves at the four corners of a square, with the hot-blast main running between them. This takes up much less room than the linear arrangement and shortens the length of the hot-blast main by the diameter plus clearance of two stoves, and where room is a matter of importance this plan is worthy of serious consideration.

The location of the stoves in relation to the furnace is a matter of much importance from another point of view. The hot-blast main becomes heated to quite a high temperature in spite of a heavy lining of firebrick, and as a result expands in ordinary cases from one to three inches. No expansion joint has as yet been designed that is satisfactory for this service on account of the high temperature to which it is necessarily more or less exposed, and in consequence one end or the other of the hot-blast main must move. Owing to the construction of the hot-blast valve seats shown in an earlier article these act as expansion joints to the extent of about half an inch or an inch. That is, the main dan slide on the flat surface of the joint to this extent without breaking anything. The bustle pipe is suspended from the furnace itself on swinging hanger bolts, and these are free to move more or less, but movements of the bustle pipe in relation to the center of the furnace are extremely undesirable because if it was put up truly concentric with the furnace originally every change must move it from the center, which means that the distances from the base of the tuyeres to the bottom of the penstocks is not constant, as it should be, but longer on one side of the furnace than on the other. This necessitates having special blow pipes for some or all of the tuyeres and greatly increases the spares necessary to be kept on hand, and, what is even more important, increases very greatly the difficulty of changing tuyeres or the like on account of the difficulty of securing the right length of blow pipe.

For these reasons the design of the hot-blast main, particularly in relation to expansion, should be very carefully considered before a final plan is adopted.