The economic considerations underlying the building of railroads are now fundamentally different from those existing fifty or sixty years ago. In 1840 the number of miles of railroad in the United States was 2818. In seventy years that mileage was multiplied by nearly 87. At that time the number of miles of line per 10,000 square miles (100 miles square) of territory was only 9.5. Now it is 808. At that time nearly the whole country was virgin territory, and it was not a question of the ultimate success of a road, provided it was constructed,with a decent regard for sound engineering principles, but a mere question of time before the country would develop sufficiently to support the road.
In a broad way we may now say that the railroads of the country are built. The great trunk lines have developed practically all of the available east and west routes, at least between the Atlantic and the Middle West, and all of the necessary lines from north to south. More tracks will doubtless be built, but they will be the expansion of one-and two-track lines into four, or even six-track roads, or the construction of short stretches of expensive reconstruction to obtain low-grade lines. If irrigation succeeds in transforming parts of the great West from deserts into fertile farms, there will probably be an enormous increase in railroad building in the West, but even this will be under different circumstances and conditions from those under which our railroads were built during the period from 1860 to 1890. If we examine a railroad map of the United States, it will not be easy to find a spot in the New England States (except Maine), in the Middle Atlantic States, or in the States around the Great Lakes, which is 20 miles from any railroad. If we consider that the whole country was gridironed with railroads running north and south and east and west at a distance apart of 25 miles, only one point in each square would be as far as 12.5 miles from any road. Such squares would have 50 miles of road for 625 square miles of territory or 8 miles per 100 square miles of territory. The average for the whole United States (8.08) is even now greater than this. Maine is the only State east of the 95th meridian in which the number is less than 8.
The practical meaning of the above is that the railroad building of the future in the eastern part of the United States will probably be confined to the construction of comparatively short cross-country lines whose chief purpose is to give additional facilities to sections of territory which already have railroad lines within ten to twenty miles. This again means that a railroad will not usually be able to monopolize all traffic in the territory through which it passes and for as many miles back from the railroad as railroad influence may be considered to extend, but that it must directly compete with other roads for its traffic. Of course it will have a virtual monopoly on sources of traffic which are so near to its line that no other road could obtain such traffic except at a prohibitive sacrifice, but it will mean that for some distance out of large cities, which are entered by several railroads from approximately the same direction, the traffic will be largely competitive. It will frequently be found that a very large proportion of the traffic of a railroad is competitive traffic and that the strictly non-competitive traffic, the local traffic which it picks up from its own immediate territory, is compara-tively unimportant. On this account we must largely modify the methods of estimation which would have been proper many years ago and also estimates based on the history of roads which have been long established, since in general those roads began under traffic considerations which are different from those of a new road of the present day.
We may first begin with the most rapid, easy, and approximate methods which have their value, because their rapidity and simplicity renders them easy to apply and they at least form a valuable check on other and more elaborate methods. In Table III are shown the gross "earnings from operation" for the whole United States for each year, from 1894 to 1910 inclusive, and in the next column the actual or estimated population. Dividing one by the other we have the average earnings per capita for the whole United States. It is interesting to note from this the comparatively low value of these receipts in 1894, which immediately followed the panic year of 1893, and how, by an almost uniform rise, the receipts increased with the boom times until 1907. Then, notwithstanding the growth in population, the gross earnings dropped nearly 200 millions and the average per capita dropped nearly three dollars. Since then they have recovered even more than that loss per capita, and now the per capita earnings are nearly double those in 1894. A study of the last column shows that, barring fluctuations, a normal increase in per capita earnings may always be expected. These values, after all, are but average values, and are the average of the whole ten sections into which the area of the United States is divided by the Interstate Commerce Commission. By reference to the map of the United States shown in Fig. 4, it will be seen that the whole territory of the United States is divided into groups. These groups are found to vary considerably in the character of their population, its density, the number of miles of railroad per hundred square miles of territory, and in the amount contributed per capita. In Table IV are shown the gross earnings for the year 1900 as divided among the different groups. Unfortunately the report of the Interstate Commerce Commission for 1910 does not contain similar data for that year, but the figures given, although now somewhat out of date, will still illustrate the principle.
The figures in the last column are not based on data which are so accurate that the figures may be considered to be precise, but the errors involved are certainly small and the figures are amply accurate for our purpose. The figures are seen to vary considerably from the average receipts per head of population for the whole United States. As a general statement it is true that the estimated earnings for a road to be constructed in the territory of any one of these ten sections would be given more accurately by the average figure for that section than by the figure given for the whole United States. Such a figure has its value as a first trial and preliminary estimate of the probable traffic of a road.
Gross earnings from operation
Earnings per capita.
Number of miles of line per hundred square miles of territory.
Number of miles of line per 10,000 inhabitants.
Gross earnings per mile operated.
Gross earnings per capita.
Group I. .. .
The earnings of a road, its operating expenses, and therefore its net revenue, vary from month to month throughout the year, and the variations are remarkably uniform during successive years. This variation is well shown in Fig. 5a, in which the lines show the actual variations in dollars per mile of line for the year 1911. The shaded areas show the maximum variation of each line during the preceding three years. This shows that the maximum revenue invariably occurs during October. The expenses are also a maximum, which practically means that expenses which may be somewhat controlled, such as extraordinary maintenance expenses, are purposely timed when the revenues to meet them are maximum. The revenues drop rather steadily until February and controllable expenses are likewise regulated accordingly. There is always a small increase in March and a drop in April, after which the rise begins for the big Fall business. The monthly variation in the operating ratio is also shown. The maximum is of course in January or February, when revenues are the lowest.