This section is from the "The Economics Of Railroad Construction" book, by Walter Loring Webb, C.E.. Also see Amazon: The Economics Of Railroad Construction.

The effect on maintenance of equipment will be practically confined to the repairs, renewals and depreciation of steam locomotives (Items 25 to 27) and the same items for freight cars (Items 34 to 36). Very few roads have a passenger traffic which is affected by the rate of the ruling grade, since, for the encouragement of traffic, passenger-trains are usually added to the schedule in advance of the physical capacity of a locomotive to haul one or more additional cars. Therefore, in general, no allowance need be made for any effect on passenger-trains, or on the maintenance of the passenger-cars. But the cost of maintaining the freight-cars is actually reduced by having more trains and less cars per train. This means that, although the maximum drawbar pull will be the same in both cases, and will equal the maximum capacity of the locomotive, while on the ruling grades, the draw-bar pull while on the light grades, level track and down grades, which may mean 90% of the length of the road, will average very much less. It is impossible to make an accurate estimate of the amount of this saving which would be generally applicable. Wellington estimated it at 10%. Considering the very large proportion of freight-car maintenance charges which are evidently independent of draw-bar pull, the estimate is probably large enough and the error in adopting that figure is not very great.

Although, for either system of grades, the locomotive is supposed to work to its full capacity while on the ruling grades, there is also some saving for each locomotive when hauling a lighter train over the light grades, level sections and down grades. If, on account of the reduction in average draw-bar pull, the repairs of each of four locomotives were reduced 5% below the repairs of the three locomotives which could haul the same number of cars over lower ruling grades, then the repairs of the four locomotives would cost 4 x 95, or 380 compared with 3 x 100, or 300 for the three locomotives. This would mean that the additional locomotive should be assigned an added expenditure of 80% of the average cost for one. If the saving by the reduction of grade was only half as much, or one train in eight, so that seven trains were required to do the work of eight on the steeper grade, then the saving per engine would be correspondingly less. If it were 2.5% instead of 5, we would have, for the cost of eight engines, 8 x 97.5, or 780 compared with 7 x 100, or 700 for the seven engines. Again, we would have 80% as the additional net cost of the repairs on the additional engine. Of course the above estimates of 5% and 2.5%, as the saving on one engine, are merely guesswork, but the above demonstration shows that if the saving in repairs is proportional to the reduction in number of trains which is made possible by the reduction of grade, as is quite probable, then the cost of the repairs of the extra engine is the same, whether it is one engine out of four, or one engine out of eight, or one engine out of any other number. Therefore, although the estimate of 5% per engine as made above is a guess, it is probably near enough to the truth, so that there is comparatively little error in using the figure of 80% for the additional cost of repairs of the additional engine.

Continue to: