The wire lath partitions cannot be considered fireproof until they are plastered. Here, accordingly, the plaster forms an essential feature of the partition; and in case of any considerable portion of this being destroyed and exposing the metal frame, the partition could be repaired by replastering, provided the metal frame had not been injured.

The concrete blocks and the terra cotta blocks in the tests cited above were not injured by the fire and water test; and so, if the results under actual conditions were always as favorable as in these artificial instances, the expense of repairing this form of partition would appear to be less than in the case of the other forms. It should be noted, however, that the partitions tested were without openings, and that openings in a partition weaken its lateral stability. While the block partitions were uninjured, they might not show so favorable results where openings occur, because of the attendant loss of lateral strength. In this respect it is probable that the plaster and wire lath partitions, and those plaster board partitions having metal stiffening, would not be any more liable to failure with openings than without, because, as constructed, the metal frame is secured at floor and ceiling, and, where openings occur, the frame is also tied longitudinally.

Column Coverings. The particular form of covering to be used is affected by the section of the column. In general, however, this consists of terra cotta blocks, wire lath, and plaster, or a solid block of concrete or plaster. As before stated, the principal source of failure in all forms of covering is their liability to crack off or be knocked off. The more nearly, therefore, the covering can approach a monolith of substantial thickness, the better it will be. If it consists of blocks, these should be bonded or anchored so as to tie the whole together, and should be made with one and preferably two air spaces. If of plaster on wire lath, it should be cement of sufficient thickness; and if of concrete, cast in place, it should form a solid casing without joints and with an air space between it and the steel. In many cases, pipes are run in the column enclosure, so that in such instances the solid monolith is not practicable.

Corrosion of Steel. An important feature in all concrete-steel systems is the effect of the concrete on the steel. Some authorities have held that, on account of its alkaline nature, the presence of Portland cement in concrete is sufficient to prevent any corrosion of the steel. Observations of actual structures, and tests specially conducted, have shown, however, that under certain conditions steel will rust when imbedded in Portland cement concrete, while under certain other conditions it will not rust in such an environment. It has been held by some, for example, that this rusting will not occur unless sulphur is present in the concrete.

Professor Norton of the Massachusetts Institute of Technology has conducted a series of tests to observe the conditions under which steel in concrete will corrode. A number of mixtures of concrete were used, consisting of standard brands of cement and of both cinders and stone. The cinders showed very little sulphur present, and the concretes were distinctly alkaline. The metal imbedded was in the form of steel rods, sheet steel, and expanded metal. The results showed that when neat cement was used no corrosion occurred. It was also demonstrated that when corrosion occurred in either the cinder or stone concrete, it was coincident with cracks or voids in the concrete which allowed the moisture and carbon dioxide to penetrate. If the concrete was mixed wet, so as to form a watery cement coating over all the steel, this coating protected the metal even when cracks and voids were present.

Professor Norton announced the further conclusion that when rusting occurred in cinder concrete it was due to the iron oxide or rust in the cinders, which acted as a carrier of the moisture and carbon dioxide, and it was not due to the presence of sulphur. Also, that if cinder concrete was well rammed when wet, and was free from voids, it was about as effective as stone concrete in preventing rust.

His conclusion as regards the part played by rust in itself aiding the further corroding action by assuming the role of carrier for the active agents, shows the importance of having the steel free from rust when it is imbedded in the concrete.

The above observations and conclusions are of the utmost importance as establishing the conditions under which, in both stone and cinder concretes, steel may reasonably be expected not to corrode, and as showing clearly the precautions and methods that should be observed in such construction.

Paints. Paints used for the protection of steel, consist, like all other paints, of a pigment and a vehicle. The pigments used are generally red lead, iron oxide, carbon, and graphite. The vehicle commonly used is linseed oil; and generally this is boiled oil, although raw oil is sometimes used.

Observations covering a period of about four years were made by Mr. Henry B. Seaman, Member of the American Society of Civil Engineers, on various kinds of paint exposed to the locomotive smoke and gases on viaducts over the Manhattan Elevated Railroad in New York City. His report, published in the New York Evening Post, concludes that carbon and graphite paints stand such exposure rather better than others, and the carbon paints somewhat better than the graphite. None was entirely efficient. A detailed paper on paints for steel was prepared by Mr. G. M. Lilley, Associate Member of the American Society of Civil Engineers, and was published in Engineering News, April 24, 1902.

The value of paints as agents in the prevention of rusting of steel depends much upon the conditions under which the painting is done, the quality of the paint, and the treatment of the metal after painting.

The experiments of Professor Norton, already mentioned, have established that the essential thing is a coating of the steel which will not crack or peel off and is non-porous, and that the steel must be clean. The fact that in many cases paint has been applied over a coating of rust, does not, of course, afford any reason for condemning the use of paint because of its failure in such cases to prevent further corrosion.

If the paint can be applied in such a way as to form for the steel a continuous coating that will not crack, or blister, or peel off, it will probably be a very effective preventative of rust. All paints, however, are more or less porous, and to this extent inefficient.

It is, however, the opinion of the authorities who have given this subject most study, that, while more expensive, a thin coating of Portland cement applied continuously to a clean surface of steel is more effective than paint.

The alkaline character of the cement neutralizes the carbon dioxide which may be present, or which may tend to filter through to the steel. In this regard, therefore, it is probable that a small degree of rust in the steel before it is coated with cement would not be likely to cause further rust, as would be the case if the coating were of ordinary paint, since the carbon dioxide present in the rust would be neutralized by the cement.

FIRE-RESISTING WOODS.

There are several companies who have processes of treating wood to render it fire-resisting. These processes differ materially. None of them renders the wood absolutely fireproof, and tests have conclusively established that all such treated woods will burn if subjected to sufficient heat for a considerable time. Some authorities place this temperature limit at which ignition will occur, as low as 100° above the temperature required to burn untreated wood. Other authorities claim that the period during which wood will glow after it has been ignited and the flame removed, is as 1 to 10 for the treated and the untreated woods respectively.

The process of treating woods is to impregnate them with certain chemicals which serve to retard the giving off of combustible gases by the wood under heat, and which also, under the action of heat, themselves give off certain other gases that serve to extinguish combustion when started.

It has undoubtedly been demonstrated that treated wood will burn, and that the gases from it are combustible. It is, however, equally well established that treated wood will not ignite as readily as untreated wood; that it requires a higher temperature to maintain its combustion; and that when the source of heat is removed the wood will cease to glow more quickly than untreated wood.

A material has recently been put on the market in England under the name of "Uralite," which, it is claimed, can be worked like wood; and can be used largely in the same way, that is, either solid or as a veneer to form a fireproof covering. The basis of the material is asbestos mixed with whiting. The finished material is made of several thin layers felted together. For a description of this material, see Engineering, August 15, 1902.

OFFICE BUILDING FOR CHICAGO & NORTHWESTERN RAILWAY COMPANY, CHICAGO

OFFICE BUILDING FOR CHICAGO & NORTHWESTERN RAILWAY COMPANY, CHICAGO.

Trusses in court. Note columns extending to bottom chords of trusses. Boilers are hung from these columns. This truss is an example of the two-panel truss referred to on pages 133 and 293. For plan of building, see illustration on page 138