=Silt.=--When silt was encountered, the shield was shoved into the ground without any excavation being done by hand ahead of the diaphragm. As the shield advanced the silt was forced through the doors into the tunnel. Forcing the shield through the silt resulted in raising the bed of the river, the amount that the bed was raised depending on the quantity of material brought into the shield. When the whole volume of the excavation was brought in, the surface of the bed was not affected; when about 50% was taken in, the surface was raised about 3 ft.; if the shield was driven blind, the bed was raised about 7 ft. When the shield was driven blind, the tunnel began to rise for about 2 ins., and the iron lining was distorted, the vertical diameter increasing and the horizontal one decreasing by about 1¹⁄₄ ins. It was found, however, that the tunnel was not affected when part of the excavation was taken, but if all of it was taken in or the shield was shoved with open doors, the tunnel was lowered. A powerful aid was thus found for the guidance of the shield; for, if high, the shield could be brought down by increasing the quantity of muck taken in, if low, by decreasing it. The junction of the shields under the river was made as follows: When the two shields of one tunnel, which had been driven from opposite sides of the river, approached within 10 ft. of each other, they were stopped; a 10-in. pipe was driven between them, and a final check of lines and levels was made through the pipe. One shield was then started up with all doors closed, while the doors of the stationary shield were opened for the muck driven ahead by the moving shield. This was continued until the cutting-edges came together. All doors in both shields were then opened and the shield mucked out. The cutting-edges were taken off and the shields moved together again, edge of skin to edge of skin. As the sections of the cutting-edges were taken off, the space between the skin edges was poled with 3-in. stuff. When everything except the skin had been removed, iron lining was built up inside the skins; the gap at the junction was filled with concrete and long bolts were used from ring to ring on the circumferential joint.