In the previous discussion it has been assumed that the cost of transportation will vary about in proportion to the density of the traffic and the length of haul. While there are no definite cost data available by which to bear out this assumption, the fact that train-loading is, in general, aligned with density of traffic (see graph "All Commodities") is a strong point in support of such a contention. The character of the fluctuations in density, length of haul and train-load in the graphs seems to indicate that the principle of decreasing costs does not operate uniformly, but tends to approach a limit. It is obvious that the economies due to increasing length of haul would not be very effective after, say, 100 miles had been passed. It also seems evident that there are physical limits (under any given set of conditions) to the efficient increase in train-loading. Moreover, in considering the trend of operating costs there is an important factor besides length of haul and train-loading. This is the handling of the trains over the tracks. How many trains can be economically passed over a given mile of track in a given length of time? Theoretically, there would be almost no limit, for freight trains might be imagined as passing in a solid procession with mere safety intervals between. Practically, the matter is not so simple. Freight and passenger trains must in many cases use the same tracks, and even freight trains have a variety of schedules, depending upon whether they are "fast freight lines" carrying perishable goods, ordinary service trains, or merely strings of empties. Moreover, all freight trains must be "made up"; that is, the cars must be sorted and put together in the switching and terminal yards. Hence the size and efficiency of the yards, the location of the passing or side tracks, the adequacy of the signal system, etc., all have a direct bearing on the operating schedules. It seems evident, then, that the principle of increasing returns must approach a practical limit in the matter of train-operation, just as in the matter of train-loading. But cannot the railroad expand its facilities, eliminate curves and gradients, build new yards and extra trackage, and thus again permit an increase in the number of trains? This is exactly what has been done in the past, millions of dollars having sometimes been spent in reconstruction of roadway to save a few minutes in train schedules. But today the problem is somewhat different. Terminal facilities have become cramped and crowded owing to the great increase in traffic density. In most cases the terminal yard forms the "neck of the bottle," and traffic on the main lines cannot flow freely while the yards are clogged with cars. For example, the New Haven Railroad, which has become a sort of vast terminal yard, was badly congested during 1919 and 1920, its average mileage per car per day being only about half the average for the United States. As a general conclusion it may be said that the principle of increasing returns has nearly ceased to operate for some of the largest roads. An important factor in terminal congestion is the greatly increased cost of city land and the consequent difficulty of expanding yard facilities. Nevertheless, if sufficient attention is paid to the terminal problem in the next decade or two it is quite possible that the working of the principle may again be extended with respect to train movement. This possibility calls attention to the necessity for distinguishing between the application of the law of increasing returns in a static system (that is, that existing at any given time) and its application to transportation considered in the historical or dynamic sense. In this chapter the statistical evidence has been analyzed both with respect to the various kinds of transportation existing at a given time (static) and the changes in the transportation system as a whole for the period 1867 to 1921 (dynamic). In considering the principle in its dynamic application, the importance of fundamental changes in the art of transportation must be noted. Thus the general electrification of railroads may |