One of the great sources of trouble to any car-track, whether operated by steam or electricity, is at the joints of the rails. A great many devices have been employed for the purpose of making these joints as nearly perfect and as much like the remainder of the rail as possible. How important this is can be understood by another statement by Mr. Bowen in the paper previously referred to, in which he says that when the question came before him of renewing the State Street track in Chicago, he had a car weighing over four tons run over it, attached to a grip-car by means of a dynamometer. The same car was run over a track newly laid and at the same speed as over the old line. The dynamometer showed that it took 13.75 pounds more pull per ton to haul the car over the old line than over the new. That he attributed a great deal of this extra power required to the condition of the track at the joints can be seen from his conclusion that a new track with cast joints would last twelve years, and as there will be no low joints, the draw-bar pull will not increase much until the rail is worn down sufficiently to allow the wheel to run on the flange.

When it is remembered that some engineers figure that the force required to haul a ton on a well-constructed track should not exceed 8 pounds, the effect of the track being in bad condition can be plainly understood.

This trouble to joints has been obviated to a great extent by the recent practice of increasing the length of the rails from 30 to 60 feet. This reduces the number of joints one-half at once, and the average cost per rail is increased only about $2 per ton by extra length. Since electricity has been so generally introduced upon street railways as a motive power, and the rails have been used as a return conductor for the electricity, a great deal of attention has been paid to the joints. What is known as the castiron joint has been used with good success. This joint has been described in Mr. Bowen's paper as follows:

"The rails at the joint are scraped and brightened. A castiron mould is placed around the joint, making a tight fit. Into this molten iron 25 per cent scrap, 25 per cent soft and 50 per cent hard silicious pig iron is poured. The metal in contact with the mould begins to cool and form a crust, while the interior remains molten. This crust continues to cool and at the same time

contracts, forcing the molten metal strongly towards the centre, which makes a solid and rigid joint. The top and bottom part of the ball of the rail is afterwards filed off perfectly level, so that it is very difficult to detect the joint by riding over it or looking for it. Upon breaking the joint which has been cast welded, three spots will usually be found where amalgamation has taken place between the rail and the cast portion, one on each side of the web of the rail, and the other on the bottom."

He says that 17,000 joints were made in Chicago during the year 1895, and of these only 154 were lost, and that the joint in comparative tests has been shown to be far stronger than the rail itself, and that breakages that have occurred were due to flaws in the metal. The metal cast around the joint has eight times the cross-section area of the rail. Therefore, if the cast iron has onefourth the strength of steel, the joint will be twice as strong as the rail.

In Brooklyn, N. Y., 1600 of these cast joints were made in 1896, and only one failed during the first six months. In 2000 joints made by another company forty had broken.

Another method of welding rails has been described under "The Method of Track Construction in Buffalo." It was considered doubtful by many engineers whether such construction would be successful on account of its expansion and contraction due to changes in temperature, but it would seem from the few failures that not as great changes in temperature occurred as was expected. This is due doubtless to the rail being almost entirely surrounded by the pavement, preventing any direct action of the sun and keeping the greater part of the rail at the temperature of the pavement, so that no buckling has occurred on account of the heat, and that the elasticity of the rail or joint has been sufficient to take up the contraction due to cold weather. At all events, there seems to have been no trouble on that account from changes of temperature.

Coming now to the discussion of what is the best construction for a railway track in different improved pavements, Fig. 47 will show the method which the author would recommend for a street to be paved with granite. The rail is that shown in either Fig. 28 or Fig. 29, as may be deemed best by local authorities. The ties

should be of steel, spaced 10 feet apart and of the kind shown in either the Sioux City or Yonkers construction, and resting on top of a concrete beam, there being no connection whatever between the web of the rails. The objection urged to this method, that the rails cannot be kept in gauge, does not seem

valid when it has been used successfully by many companies, and when with the ordinary wooden-tie construction the tie is simply held to gauge by spikes on the bottom flange. With the method proposed, and a solid granite pavement built tight against the rail, it would seem that no difficulty would be encountered in keeping the rails to line and gauge.

Fig. 48 shows the plan proposed for the construction on a street

paved with asphalt. The rail is the same shape as that recommended for granite, except that it is but 6 inches deep. The rail

can be made of any required area to give the necessary strength, and the metal can be used to better advantage on a shallow rail, and with asphalt as great depth is not required as when granite block is laid. The arrangement of ties would be the same as in the other case, except that if the street-railway authorities have any preference for a tie from web to web of the rails, there would be no objection, as the concrete base for the asphalt can be laid around the ties without any difficulty.

There is considerable difference in the practice of different engineers as to whether asphalt should be laid up to the rails or whether blocks of some kind should be used. This will depend to a great extent upon the conditions of the street traffic. If the traffic be light, and the above construction is used, there can be no objection to finishing with asphalt up to the rails. Great care, however, must be exercised in doing this, and the asphalt should be tamped solidly under the lip and head of the rail, so that if wheels run along next to the rail, the asphalt will be sufficiently strong to resist the tendency to rut. In some cities, however, the entire space between the tracks and rails is paved with stone blocks, as many engineers think that this is a better construction. That it is more economical is probably true; but if the street be comparatively narrow, only a small portion of the street will be paved with asphalt if all of the track-space is paved with stone. In some streets also it is considered necessary to lay blocks of stone or brick on the outside of the rail as well as inside, and where the street-traffic is heavy this may be advisable. It should be remembered, however, that the theory of this construction of stone or brick is to prevent the tendency of the wheels to rut the pavement alongside the rails; but if, in the construction of a track, a rail is used that will present practically the same surface to traffic as the remainder of the street, neither inviting nor repelling wheels, this tendency is materially reduced.

The recommendation, however, would be, on streets of moderately heavy traffic, to place a row of toothing-stones arranged in pairs and set as headers on the inside of the rail, and on the outside lay the asphalt next to the rail. If the distance between the curbs is so wide that there is plenty of room outside of the track for the street-travel, and the street-railway authorities, for economi

cal reasons, wish to lay stone or brick pavement between the rails, there would be no particular objection if it be done in a thorough and substantial manner.

Fig. 49 represents a recommendation for a brick pavement. This is substantially the same as that shown in Fig. 48, except that no tie-rods should be used between the rails but at the base upon the concrete beam as recommended for granite. It is very difficult in using tie-rods between the webs to so place the holes that the rods will be exactly perpendicular to the rails, and

trouble always occurs in laying the blocks, whether of stone or, brick, between these bars on that account. It also makes an extrawide joint wherever these rods occur, and satisfactory results can never be obtained in that way.

The space between the upper and the lower flange of the rail, on the outside and on the inside, must be filled when a block pavement of any kind is to be used. Untreated and creosoted wood, sand, cement mortar, and specially burned tiling have been used for this purpose. Wood is probably the cheapest, and if the track is to sustain heavy traffic, so that it will require renewal every five or six years, untreated wood will probably be satisfactory; but if it is to remain ten or twelve years, it should be creosoted, so as to prevent decay before the rails will require renewing. Cement mortar gives good results, but is considered expensive and can be used but once. Specially burned bricks have been used with good results, although some engineers object to them on account of their being easily crushed.