Page:Our big guns.djvu/35

( 29 ) gone on, stool manufacture has improved, and improved tests are complied with.

Having tested the two ends of the tube, and found that its breech end is not too "low" and its muzzle end not too "high," one is entitled to hope that the steel situated between these two limits will not be "lower" than the one, nor "higher" than the other but will bear such relation to the end tests, as its position between breech and muzzle would lead one to anticipate. And I believe that if the steel has been made by the open hearth process, this hope will not be a vain one, and I also believe that it may be justified in many cases where the crucible process is employed.

With respect to this latter process, I should say, there are certain foreign Governments who still insist on its being followed; indeed the French Government, for the special case of certain small parts, make its use, or made its use in 1882, when their tests were drawn up, obligatory.

But I have known a quite recent case, of the failure of the tube of a largo cannon—one of those I shall have to allude to hereafter—where within a very short distance, the carbon varied very nearly in the proportion of from one to two. This tube was made by the crucible process.

Assume, however, that the tube has been made by the open hearth process, and that this does secure regularity of composition, within the limits shown by the end test pieces, there still remains, and must always, I fear, remain, the risk of internal flaws in the casting.

You may test the quality, you may follow a process which may, and probably will, give uniformity of quality within certain limits, but you never can be sure there may not be a latent defect; and a defect that no subsequent process can completely cure, for remember, that steel must be forged at a temperature considerably below the welding point, and that thus, although the act of forging may press the walls of this hidden cavity together, that act cm never make sound metallic connection between them.