and one can well imagine what would be the result in the case of riveted iron members in the skeleton structure of a building where such iron work is entirely concealed from view, periodical inspections being impossible. Rust is especially liable in the cellars and basements of buildings. The wrought-iron friction brakes of freight elevators in the cellars of stores, for example, are fequently found so consumed with rust as to be easily rubbed to pieces in the hand. Steel rivets are dangerous and they should never be used, unless of a very superior quality, so soft that hammering will not crystallize the material, and yet with sufficient tensile strength to insure perfect holding qualities. This is difficult to secure. Their use in columns for buildings is objectionable, as they rust badly under certain conditions; columns, therefore, should be without rivets, and the beam-bearing bracket shelf on cast-iron columns. should be cast in one piece with the column. EXPANSION OF IRON. It is generally supposed and frequently stated that there is a great difference between the expansion of iron and masonry by heat. This is not the case. For example, the length of a bar which at 32 degs. is represented by 1, at 212 degs. would be represented as follows: In the fire-proof building of the Western Union Telegraph Company, in New York, some years ago, a heavy brick pier, 7 or 8 ft. in diameter, adjoined the wall of the boiler furnaces. The difference in expansion in the brickwork next to this furnace wall as compared with that of the remaining brickwork of the pier was so great as to produce a crushing of the material from top to bottom of the pier for a depth of several inches, and it was found necessary to change the furnace wall and leave an air space between it and the pier. EXPANSION. While the difference in expansion between masonry and iron. incorporated with it is less per running foot than is generally supposed, and while the difference in expansion between a cubic foot of iron and that of a cubic foot of masonry would hardly be noticeable, especially if the iron were covered on all four sides, yet in stretches of 50 ft. or more, as in the case of iron I-beams and girders, the cumulative effect of expansion in uncovered iron might be a serious matter-quite sufficient with the rises of temperature due to a burning building to push out the bearing walls and wreck the building. Especially is this true of temperatures higher than 500 degs. It is unnecessary to suggest that metal differs from masonry in the important respect that heat does not travel throughout the entire length of the latter, while it does in the case of metal. In other words, while the difference between the expansion of a lineal foot of iron as compared with a lineal foot of masonry, marble, brick, etc., is very slight, the difference in conductivity is very great. The conducting power of silver, for example, being represented by 1, copper would be .845. cast iron .359, gold .981, marble .024, and brick .01-an important fact to be considerd in the construction of buildings. Brickwork raised to a white heat would not raise the temperature of other masonry in the same wall a few feet away but one end of an iron I-beam could not be raised to a white heat without raising the temperature of the beam for its entire length. It is a well-known fact that iron responds so readily to temperature that, in surveying land, a surveyor's 100 ft. iron chain will, in measuring the distance of a mile, result in a variation of 5 ft. between winter temperature and summer temperature, resulting in an error of one acre in every 533. Where iron beams and girders are inserted in walls without sufficient space left for their expansion under heat they are almost certain contem to overthrow the bearing walls by their expansion thrust. A large warehouse in Vienna in which such provision had been Plated by the architect was totally destroyed, with its contents, by reason of the fact that an officious subordinate, discovering the space in the wall purposely left at the end of each beam, deliberately poured liquid cement therein, which, having set, effectually thwarted the well-meant intention of the architect, and resulted in the destruction of the building. The expansion thrust of iron beams may be computed upon the following factor of expansion: Rolled iron of a length of 1,562 ft. willexpand one eighth of an inch for every degree of temperature. The heat of a burning building as already stated is enormoussufficient to fuse most known materials; it may safely be estimated to be at least 1,000 degs.; therefore a length of rolled iron of 1,562 ft. at 1.000 degs. of temperature would expand about 125 ins., and a 50 ft. length of iron girder would expand between 4 and 5 ins.,showing that there should be a play at each end at least 2 ins. if the iron is not fire-proofed. Inasmuch as in iron construction the iron beams and girders are usually anchored to the walls to steady them, the space should be left and the tie to the anchor should be by a movable hinge joint which would be of the same strength with an inflexible anchor for all tying purposes but would yield under the thrust pressure like an elbow and allow play of the beam, or stiff anchors should have elongated holes to allow expansion when beams are of great length. Girders are seldom over 25 feet long, but if bolted together, as is frequently the case, they may be 120 feet or more long, and a line of columns from cellar to roof of a building may easily have one continuous iron structure of 200 or more feet. It should be remembered, however, that this danger from the expansion of iron may be almost wholly counteracted by protecting it from exposure to fire through the use of nonconducting material. It is more important to protect girders than beams. The mistaken pride with which the owners of some buildings point to exposed iron beams in ceilings as evidence that the floors are "fire-proof," actually justifying the supposition that they are left exposed for such display, would be ludicrous if it were not serious. In buildings occupied for offices or dwellings, where there is not sufficient combustible material to endanger the beams, it is not so objectionable; but in warehouses and stores, filled with merchandise, such construction is dangerous, and if one of the upper floors should give way it would come hammering down to carry all below and thoroughly wreck the structure. In this connection it is well to say that combustible merchandise should never be stored 100 feet above the street grade, even in a fire-proof building, since the average fire department cannot reach it at that height. STONE STAIRCASE TREADS. Marble, slate, and other stones are certain to disintegrate or crumble when subjected to the joint action of heat and water. For this reason 90 per cent. of the staircases in modern fire-proof buildings would be found utterly unreliable in the event of fire, either for the escape of the inmates or for the use of firemen-a serious consideration. Stone treads are usually let into iron rabbet frames, and as these stone treads would give way in case of fire, it would be impossible for a person to find a footing on the stairways; 2-in. oak treads might actually last longer; but a safer staircase would be one the framework of which is of iron, the tread having an iron web or gridiron pattern, the interstices or openings of which should be small enough to prevent the passage of a foot, underlying the stone or slate, so that if the stone tread should disintegrate the staircase would still remain passable. It is possible to have the supporting tread of open work cast iron in an ornamental pattern, which in relief against the white marble tread resting on it would present a tasteful appearance from the underside or soffit of the staircase, with this great advantage that, in the event the action of fire and water should pulverize the marble or slate tread, it would still afford a safe support for the foot. In the case of the burning of the two fire-proof buildings, Temple Court and the Manhattan Savings Bank in New York, the slate treads yielded early in the fire, leaving staircases with openings the full size of the tread, which, within a few minutes after the fire started, were impassable for either firemen or inmates. It is astounding that this vital fault should be so generally overlooked in fire-proof buildings. I may here state that the Manhattan Savings Bank building did not deserve to be called "fire-proof" for the reason that it had hollow spaces under the wooden floor boards, and that the iron beams and girders were not protected. Some of them were large riveted box girders, which yielded quickly to the heat of burning goods and pushed out the side walls. It is generally supposed that it is not necessary to be careful as to stone treads in buildings occupied solely for offices separated in fire-proof hallways in which, it is claimed, there is nothing to burn; but in the case of one large fire-proof building of this kind in New York I found the space under the staircase in the basement story was used to store the waste paper and rubbish of the building material particularly likely to cause a fire by concealed matches, oily waste, cigar or cigarette stumps, etc., and to make a lively and quick fire quite sufficient to destroy stone staircase treads. Even where there is no combustible material in the hallway, if the staircase is near windows stone treads may be destroyed by exposure to burning buildings and by the combustion of window frames, dadoes, and other wooden trim. Enclosing walls. These should be of brick, the brickwork of the lower stories especially, if not of all, being laid in cement mortar. In fact, the specifications for a building in the compact part of the mercantile section of a city ought to be drawn in contemplation of the possible cremation of its contents. and the generation of heat considerably greater than 2,000 degs. Fahr. The heat of a wood fire is from 800 to 1,140 degs.; charcoal, about 2,200 degs.; coal, about 2,400 degs. Cast iron will melt at between 1,900 and 2,800 degs.; wrought iron, 3,000 to 3,500 degs; steel, 2,400 to 2,600 degs.; and if an architect should be required to draw specifications for a building adjoining others with knowledge beforehand that its entire contents, from cellar to roof, were to be totally consumed, and he were under a bond to pay damages to surrounding property, he would not be more severe in his exactions than should a building law protecting neighborhood rights in the enjoyment of property; for a mercantile or manufacturing building sometimes generates a greater heat in combustion than a smelting furnace. THE REFRACTORY QUALITIES OF CLAY. BY PROF. EDWARD ORTON. Jr., of Columbus, Ohio. Without attempting to go into the matter with any thorough ness, it can be said that a clay is a mixture consisting of the mineral, kaolinite or clay substance, with a very wide variety of other minerals, having no definite aggregate composition or proportion. The kaolinite may constitute nearly the whole of the mixture or it may be reduced in amount till it becomes insignificant. The accompanying materials vary similarly in character and amount, so that it is not possible to expect to find two clays which are exactly alike in ingredients and proportions. The kaolinite has fixed chemical and physical characteristics which are well known and sharply defined. Among the latter plasticity is the essential one, and this quantity is so important that it alone forms the means of deciding whether a mineral mixture can be called a clay or not. If there is kaolinite enough, present to make the whole material plastic, we call it clay. If there is not enough, we call it sandstone, or limestone, or an iron ore, or a coal, according to which one of the mineral ingredients is the most abundant and important in its makeup. These fundamental facts being understood, we can readily see why clays give a great range of refractoriness, for it is only reasonable to believe that each and everyone of the minerals entering into this complex mineral mass is certain to have not only its own individual pyrometric behavior, but also that it is certain to affect the pyrometric behavior of the other minerals with which it is associated. We have before seen that some materials are bases and some acids, or some positive and some negative. And when we know that the commonest of all the associated minerals found in clays comprise the commonest of all negative materials, called silicia or sand, and the commonest of the positive elements, such as lime, magnesia, iron and the alkalies, we can readily see that we have no simple problem on hand when we try to forecast the refractoriness which any given clay may exhibit. Before anything at all can be done in this direction, however we must, above all things, obtain a clear idea of the pyrometric qualities of the kaolinite itself, which is the soul or essence of the clay. As is well known to you all, clay substance itself, or kaolinite, as it has been christened by scientists, is a simple silicate of alumina, the simplest combination which these two compounds unite to form. Silica is the most abundant material in the earth's crust, and alumina is the next to it in the order of abundance. But while they are both present in innumerable other mineral forms, when mixed with other bases, they form together, few if any other compounds than kaolinite. Silica and alumina are both almost infusible by themselves. We cannot melt pure silica, unless we deliberately go at it to make a furnace which will do so. It can be melted with difficulty by the oxyhydrogen blow-pipe, or it can be melted easily and even vaporized by the electric arc, but these generate temperatures which are beyond all common attainment and above all use |