It is fortunate for Americans that the introduction of lighting by coal gas into this country found the manufacture of wrought-iron pipes so well developed that they were immediately adopted for distributing the gas in houses, instead of the wretched pipes of thin lead which were for so many years employed for the purpose in England, and even now have not gone out of use. With us, gas-pipes are always of iron, put together with screw joints, and are tested, when the piping of a house is completed, by pumping air into them until the pressure in them is sufficient to raise the mercury in a manometer tube. Unless the pipes are so free from leakage that the mercury in the manometer remains at the same height for some hours, the gas company will not furnish gas to the house; so that the fact that the house is supplied with gas is evidence that the pipes have been tested by the company's inspector, and found tight; and, if once proved tight, iron pipes are not very liable to subsequent leakage, unless they have been tampered with or altered after testing. Where such leaks occur, they are most likely to be near the fixtures. Inferior gas-fitters, instead of making the outlets, or "nipples," of their pipes just long enough for the fittings to be screwed on, use waste pieces of random lengths; and it frequently happens, in consequence, that the contractor for the fixtures is obliged, before he can put on his brackets or centre lights, to unscrew the nipple provided for him, and put on one of proper length. As the pipes have already been tested, and he has not, therefore, the fear of the inspector before him, the fixture man is not always careful to make the new nipple as tight as the old one, and a concealed leak in the wall, where it is not easily discovered or remedied, may be the consequence. The inferior gas-fitters also frequently find it too much trouble to arrange their pipes so that the nipple projects exactly at right angles with the wall or ceiling, and, as a result, the fixtures, when attached to them, make various angles with the plastered surface. As this is ungraceful, the fixture man endeavors to bring the nipples to the proper position by screwing a long piece of pipe on them, to serve as a lever, and violently bending them to the angle required. The fittings for small gas-pipes are usually made of malleable iron, so that they can accommodate themselves to a certain amount of bending; but if they, or the pipes to which they are attached, should happen to be intractable, a bad leak may be caused in a partition or floor, where it can be reached only with difficulty.

Occasionally, gas-pipes may leak in other places. The inferior gas-fitters sometimes smuggle in split pipes into their work, filling up the seam with red lead putty, which stays in place long enough to enable the piping to pass the inspector's test, but falls out later, opening a dangerous leak; or a fitting may be broken by an accidental blow; or an old and leaky meter may have been furnished by the gas company, or it may not have been properly connected; and, in all these cases, leakages may be caused after the piping has been passed by the inspector; while a leak in a street main may fill the cellars of the neighboring houses with gas.

From whatever source it may proceed, a leakage of gas should never be investigated with the aid of a lamp, or candle, or lighted match. Innumerable explosions, many of them very destructive, have been caused in this way. The proper course, when any serious leak is suspected, is to feel the way, if daylight is not available, to the meter, and shut off the gas from the house. Such gas as has already escaped will very soon dissipate itself, and the pipes can then be examined with safety. If the defect is not found, a fixture, or a nipple, may be unscrewed, and some ether poured into the pipe, closing the opening immediately. The ether vapor will diffuse itself through the pipes, issuing at the leak, which can thus be traced by the smell, and corrected as circumstances may require.

Leaks in or about fixtures are much more common than those in pipes, and are, fortunately, more easily detected. Besides the leaks due to the substitution of a new nipple for one which has been tested, but is of improper length for attaching the fixture, other leaks are often due to a defective joint between the nipple and the fixture. Where the nipple for a ceiling light has been carelessly put in, so that the chandelier will not hang perpendicularly, it is often necessary to attach it to the nipple by a ball-and-socket joint, and such joints always leak. If the connection of the fixture with the nipple is perfect, there may be a leak at the key; and, if the keys are tight, the burners may not be well put on. In the latter case the gas will leak only when the burner is lighted, as the key is the limit of the portion of the fixture which contains gas at other times.

Probably nine-tenths of the leakages of gas which, almost undetected, poison the air of a house, are at the keys of the fixtures; and, in order to guard against or cure such leakages, the construction of a gas-key should be understood (Fig. 30). In substance, it consists of a tapering plug, carefully fitted to a conical hole through the body of the fixture. The plug has a hole through it, which can be turned so as to coincide with the bore of the fixture, allowing the gas to pass, or at right angles with it, shutting the gas off; the key itself, which is simply an extension of the plug, serving to turn it as desired. By use, the plug and its seat wear, the plug growing smaller, and the conical hole in which it moves larger, until there would be room enough between them to allow so subtle a fluid as illuminating gas to pass, if it were not for the short spiral spring, S, which is held by a screw to the end of the plug, but presses against the body of the fixture, so as to draw the plug with considerable force into its conical seat. If the plug were cylindrical, it would be drawn far in; but, being tapered, it can only be drawn in far enough to bring it into close contact with its conical seat; and, as the surfaces of the plug and its seat are worn away, the spring draws the plug farther in, still keeping the joint tight. In practice, however, the spiral spring is so short that its action does not last long; and, after a certain amount of wear has taken place, the plug is no longer drawn tightly down to its seat. The key may then be tightened by screwing in farther the screw which holds the spiral spring, so as to bring the latter once more into play; but continued wear again loosens the plug, and a time comes when the tightening screw can be turned no farther. The tension on the key could, even then, be restored by stretching out the spring, or by deepening the screw hole in the end of the plug; but either of these operations would probably be beyond the resources of an ordinary householder, and it is generally quite sufficient to remember, in opening or closing any gas-key, if worn and loose, to push it slightly upward, so as to leave it firmly sticking in its seat, independent of the tension of the spring. This simple precaution will keep the key from leaking for an indefinite period; and fixtures which would, from wear, leak quite seriously at the key if this were turned, as is common, with a downward pull, will be perfectly safe if the key is always pushed upward in turning.