The welding is done automatically and does not require an expert welder, but may easily be carried out by the permanent staff of any company.

The Paris Metropolitan Railway made some experiments on a short track, with the result that after one year's trial 1,200 joints were welded in 1904; the rails are 104 lbs. per yard. In 1906 a further 1,800 joints were welded, part of which were on exposed track, in which case lengths of 250 yards were welded together, having an expansion joint between. The results obtained having proved satisfactory, the directors decided to weld the whole system, some 10,000 joints, in a similar manner.

On the suburban railway from Berlin to Grosslichterfelde, the Allgem. Elektr. Ges. of Berlin, have welded 13 1/2 miles of track. The rails on this track are exposed ; three lengths of 45 ft. rails are welded together and connected to the next by an expansion joint.

The question of a continuous rail on exposed railways is not yet solved, for want of sufficient tests. That it is practicable within certain limits, and that it is desirable to have greater lengths of exposed track welded together, is admitted by permanent way engineers. In any case exposed rails can undoubtedly be welded without any risk in tunnels and subways where differences of temperature are very slight, and contraction and expansion therefore only minimal.

The matter of securing a proper joint for fastening together the ends of rails so as to make a smooth riding track without appreciable jar or jolt when the wheels pass a joint deserves, therefore, great consideration, and many forms of such joints have been suggested. All of these welded joints fasten the ends of the rails together, so that the rail is practically continuous, just as if there were no joints, so far as the running surface of the rail is concerned.

It was thought at one time that a continuous rail would be an impossibility because of the contraction and expansion of the rail under heat and cold, which, it was thought, would tend to pull the rails apart in cold weather and to cause them to bend and buckle out of line in hot weather. Experience has conclusively shown, however, that contraction and expansion need not be taken into account when the track is paved, provided it is well constructed and long lengths of rail are not left exposed to the rays of a hot sun for a considerable period. The paving tends to hold the track in line, and to protect it from extremes of heat and cold. The reason that contraction and expansion do not work havoc on track with welded joints is probably that the rails have enough elasticity to provide for the contraction and expansion without breaking.

It is found that the best results are secured by welding rail joints during cool weather, so that the effect of contraction in the coldest weather will be minimum. In this case, of course, there might be considerable expansion of the track in the hottest weather, but this does not cause bending of the rails, whereas occasionally, if the track is welded in very hot weather, the contraction in winter will cause the rail to break.

The following tests, carried out by independent experts and well-known engineers, show the relative strength of the "Thermit" rail joint compared with the rail itself, and proves that the running surface of the rail is not chemically altered during the process of welding, and that the electrical conductivity is higher than that of a joint with fish plates and bonds.

A. Tests made for Leeds Corporation Tramways.

Rails supported 5 ft. centre to centre, test made with 10 in. ram, 2 in. bearing on head of rail.

Up to 28 tons, no deflection, and then 1/64 in.

Test was then stopped, and it was found that there was no permanent set whatever.

Pressure brought on again to 65 tons, still only 1/8 in. deflection.

At 68 tons rail still sound. ,,70 ,, ,, broke, not through the weld.

B. Second Test made for Leeds Corporation Tramways.

Hydraulic test; dead load; 5 ft. bearings, 10 in. ram, 2 in. bar on head of rail.

RAIL WITH "THERMIT" WELDED JOINT.

Up to 60 tons no permanent deflection. Safe dead load at 68 tons.

70 tons fractured at side of the weld, the welded portion remaining intact.

FISH AND SOLE PLATE JOINT.

Fish plates 62 lbs. per pair, 2 ft. long, six 1 in. bolts; sole plate 46 lbs., 2 ft. x 8 in. x § in., twelve 7/8 in. bolts.

Permanent set at 85 tons § in.

Fractured at 102 tons.

Report of the Tensile Test carried out at Phoenix Works, 11th May, 1903, on "Thermit" welded Rail.

Test pieces of equal size (diameter § in., area 0.48 sq. in.) were taken from the head of the rail.

Result shows that the welded part has 87.75 per cent. of the strength of the original material, which is very satisfactory indeed. The test was carried out in the presence of Engineer Kurz, of the above works.

(Signed) H. Tiemann,

Engineer. Essen Ruhr, 12th May, 1903. P.S. - It should be noted that the above test refers to the rail without the ring of "Thermit" iron, by which it would be materially strengthened.

Tests showing the Relative Resistance of "Thermit" Joints compared with Fish Plate Joints and Bonds.

(A.) Test made on a Leeds section of rail, 100 lbs. per yard. The readings below are the mean of several:

(B.) Test made by Mr. J. Lord, Borough Engineer, Halifax (1904) :

(C.) Test made by Mr. I. E. Winslow, Consulting Engineer to Coventry Electric Tramways Co. (1905).

Resistance of consecutive "Thermit" joints, 4 ft. rail 80 lbs. per yard, with joint, also 4 ft. of rail without joint.

Conductivity Test.