To use the device, open the cock and let a small amount of oil flow; apply fire to the pipe under the wheel, and the oil in the upper pipe is converted into gas, which flows out of the small holes in the lower pipe, takes fire, and heats not only the tire, but the upper pipe, thus converting more oil into gas. We had here a lot of blue flame jets and the same result as with gas, but at less cost. We had also a machine that was inexpensive and easily handled anywhere. Boxes were placed over the upper parts of the wheels, that the heat might pass closely to the tire. This device was extensively used by our people, and with great satisfaction. In one way care had to be taken, viz.: That in starting the fire it did not smoke and cover the tire with carbon or "lampblack," which is a non-conductor of heat. Experiments were made with air forced through gasoline, and with oil heated in a can to form gas. There was more danger in either of these than with our blowpipe device, and no better results were obtained, though the cost was greater. With the change of the wheels, the brakes had to be changed the same amount, that is, each one set in 11/2 in. This it was thought would either require new hangers or a change in the head or shoe in some way. We found that the hangers could easily be bent without removal. Fig. 34 shows three hangers after passing through the bending process. A short lever arranged to clasp the hanger just below the point, A, was the instrument; a forked "shore" is now placed, with the fork, against the point, A, and the other end against the car sill; press down on the lever and you bend the hanger at A; lower the lever to a point just below B, reverse the process, and you have the bend at B; the whole thing taking less than two minutes per hanger. A new bolt hole, of course, has been bored in the brake beam 11/2 in. inside the old hole. It takes but a short time after this to change the position of the head and shoe. [Illustration: FIG. 34] Before the day of change, a portion of the spikes were drawn from the inside of the rail to be moved, and spike set 3 in. inside of the rail. As a rule two spikes were drawn and the third left. At least every third spike was set for the new gauge, and in some cases every other one. There were several devices with which to set the spike. A small piece of iron 3 in. wide was common, and answered the purpose well. This had a handle, sometimes small, just large enough for the hand to clasp, while others had a handle long enough for a man to use it without stooping down. (See Figs. 35 and 36.) Another device is shown in Fig. 37, so arranged that the measurements were made from the head of the other rail. This was liked best, and, it is thought, gave the best results, as the moved rail was more likely to be in good line than when the measurements were taken from the flange. [Illustration: FIG. 35, FIG. 36 and FIG. 37]