The requirement that they shall harden safely and do good work afterwards involves necessarily, high steel of suitable quality.
CHEMICAL CONSTITUENTS.
No engineer should, unless he be an expert steel-maker, attempt to specify an exact chemical formula and a corresponding physical requirement; in doing so he would probably make two requirements which could not be obtained in one piece of steel, and so subject himself to a back down or to ridicule, or both.
On the other hand, he may properly, and he should fix, a limit beyond which the hurtful elements would not be tolerated. Notwithstanding satisfactory machine tests, successful shop-work, and a liberal margin of safety, no steel can be relied upon that is overloaded with phosphorus, sulphur, manganese, oxygen, antimony, arsenic, or nitrogen.
In regard to silicon, it is common to have as much as 20 to 25 points in tires, with 55 to 80 carbon; such tires are made by the best manufacturers, and they endure well. But it is certain that good, sound steel can be made for any purpose with silicon not exceeding 10.
Structural steel can be made cheaply within the following limits:
| Silicon | < .10 | |
| Phosphorus | < .05 | |
| Sulphur | < .02 | |
| Manganese | < .50 | or even < .30 |
| Copper | < .03 | |
| Carbon to meet the physical requirements. | ||
Steel made within these limits and not overblown or overmelted must be better in every way than steel of
| Silicon | > .20 |
| Phosphorus | > .08 |
| Sulphur | > .05 |
| Manganese | > .60 |
| Carbon to meet the same requirements | |
A steel of the latter composition, or with no fixed limits, may be made cheaper than the first by a dollar or two a ton; but for any large lot it is believed that the first specification would be bid to at as low a price as if there were no specification; competition among manufacturers would fix that. At any rate there is no reason why an engineer should refuse to demand fairly pure material when he can do so at little or no extra cost.