Under the present methods of bonding, the angle bar carries the greater part of the current, and bond wires frequently carry as little as 20 per cent. of it and sometimes even less. The rail resistance is lowest with new rails, but it gradually gets higher, due to rust and dirt formations between the angle bar and the rail. But even with new rail, the rail resistance varies greatly at different periods and even at different times during any twenty-four hours. This variation is entirely due to the fact that the angle bars carry more of the current than the bond wires, and that the bond wires under any condition are only large enough to carry the smaller part of the current from the battery. The lower the resistance of the bonds the less variable will be the rail resistance. The resistance of the angle bars increases greatly as rail resistance increases, as a result of which the angle bars rapidly carry less of the current. It is not infrequent to find the rail resistance to be as high as 0.20 ohms per 1,000 ft. of track, and we have known it to run as high as 0.264 in new rail, where especial attention had been given to obtaining as good bonding as possible. Under such conditions the angle bar carries very little of the current, the capacity of the bond is not sufficient to carry the current, and the net result is a failing track circuit, which is probably attributed to bad ballast, zinc treated ties or other causes. The principal defect in the track circuit is that of improper bonding. The only explanation as to why No. 8 iron wires became standard for bonding appears to be that the bond wires were cut from this size iron telegraph wire which was in general use at the time rails began to be bonded. It is important to obtain better bonding to obtain a minimum constant rail resistance. It has been recommended that: First--The use of galvanized wire bonds should be eliminated. Second--Forty per cent. copper clad bond wires should be used as a temporary expedient to replace galvanized bond wires. Third--Except for theft and crystallization, copper bond wires would be much more advisable. Fourth--Larger bond wires should be used, these bonds to be at least equal in carrying capacity to two 46-in. No. 6 solid copper or to two No. 2, 40 per cent. copper clad wires. Until recently it has been the general opinion of all experts on the track circuit that the rail resistance was rather an unimportant factor and that, as a general rule, the change in rail resistance could be disregarded in making track circuit investigations and calculations. Many bad track circuit conditions have been laid to bad ballast conditions, zinc treated ties, wet track, etc., which, if carefully analyzed, would have shown the trouble to be due to extremely high rail resistance. These faulty conclusions are being drawn nearly every day.