A surprisingly large percentage of the fuel consumed is not utilized in drawing a train along the road. A portion of this percentage is used in firing-up. A portion is wasted when the engine is standing still, which is a considerable proportion of the whole time.

The policy of banking fires instead of drawing them reduces the injury resulting from great fluctuations in temperature, but in a general way we may say that there is but little, if any, saving in fuel by banking the fires, and therefore we may consider that almost a fire-box full of coal is wasted whether the fires are banked or drawn. Some tests were made on the Santa Fe, in which some large locomotives consumed from 1200 to 1660 pounds of coal merely in firing-up. But even the amount of coal required to produce the required steam-pressure in the boiler from cold water does not represent the total loss. The train-dispatcher, in his anxiety that engines shall be ready when needed, will sometimes order out the locomotives which remain somewhere in the yard, perhaps exposed to cold weather, and blow off steam for several hours before they make an actual start. Of course the amount thus attributable to firing-up is a very variable one, depending on the management, and therefore no precise figures are obtainable. But it has been estimated that it amounts to from 5 to 10% of the total consumption. A freight-train, especially on a single-track road, will usually spend several hours during the day on sidings, and when a single-track road is being run to the limit of its capacity, or when the management is not good, the time will be still greater. It has been found that the amount of coal required by an engine merely to keep up steam will amount to from 25 to 50 pounds per hour. If, in addition to this, steam escapes through the safety-valve, the loss is much larger. It is estimated that the amount lost through a 2 1/2-inch safety-valve in one minute would represent the consumption of 15 pounds of coal, which would be sufficient to haul 100 tons on a mile of track with easy grades. Again we see that the amount thus lost is exceedingly variable and almost non-computable, although as a rough estimate the amount has been placed at from 3 to 6% of the total. Another very large sub-item of loss of useful energy is that occasioned by stopping and starting. A train running 30 miles per hour has enough kinetic energy to move it on a straight level track for more than two miles. Therefore, every time a train running at 30 miles per hour is stopped, enough energy is consumed by the brakes to run it about two miles. There is a double loss, not only due to the fact of the loss of energy, but also because the power of the locomotive has been consumed in operating the brakes. When the train is again started, this kinetic energy must be restored to the train in addition to the ordinary resistances which are even greater, on account of the greater resistance at very low velocities. Of course, the proportion of fuel thus consumed depends on the frequency of the stops. It was demonstrated by some tests on the Manhattan Elevated Road in New York City, where the stops average one in every three-eighths of a mile, that this cause alone would account for the consumption of nearly three-fourths of the fuel. On ordinary railroads the proportion, of course, will not be nearly so great, but there is reason to believe that 10 to 20% is not excessive as an average figure. The amount of fuel which is consumed on account of curvature is, of course, a function of the curvature and will vary with each case. The only possible basis for a calculation of this amount must be somewhat as follows: Since all of the above calculations consider the average cost of a train-mile throughout the country, we must consider what is the average amount of curvature per mile of track throughout the country. By estimating, as will be developed in the next chapter, the proportion of the fuel expenditure which is due to this average amount of curvature and subtracting this, as well as the other subitems enumerated above, from the total average cost of a train-mile, we would then have the desired quantity, the cost of fuel per mile of straight level track. Although it is not easy to obtain reliable statistics showing the average curvature per mile of road throughout the United States, there is reason to believe that it is not far from 35° per mile. According to the method of calculations given in the next chapter, to determine the additional fuel consumed by the added resistance due to 35° of curvature per mile, we obtain the approximatevalue that about 4% of the fuel consumption will be due to this cause. To obtain an average figure for the resistance due to grade is perhaps even harder, but on the approximate basis that the average amount of rise and fall per mile of track is about 11 feet per mile, it would seem as if 25% of the consumption of fuel were due to grades. Summarizing the above items we would have:

Firing................

5 to 10%

Loss by radiation, etc...

3 to 6%

Stopping and starting ...

10 to 20%

Average curvature.....

4 to 4%

Average grade.........

25 to 25%

47 to 65%

Direct handling .................

53 to 35%

Average 44%

100 100

This gives, as an average figure for the increased consumption of fuel due to one additional mile of straight level track, 44% of the average consumption per mile.