"It is quite probable that the actual cost of transporting cotton piece goods from New York to Wichita _via_ Galveston does not exceed that of carrying them from New York to Kansas City via the cheapest route. The all-rail haul is to the latter point 1,300 miles and over. The ocean and rail movement involves a rail carriage of from 1,100 to 1,300 miles, depending upon the route selected. If the goods move through some Gulf port, there is a rail carriage of not less than 850 miles. If, therefore, the rate were to be measured by the expense of the service, it is probable that Wichita would today enjoy as low a rate as the Missouri river." The Wichita complication, moreover, works both ways. Wichita and Kansas City form two angles of a narrow triangle with its apex at the Gulf ports; but the distance from Kansas City is longer than from Wichita. Railroad competition brought it about, however, that the rate on export grain from Wichita _via_ Kansas City to the Gulf came to equal the rate from Kansas City to the Gulf _via_ Wichita. But the former was a much longer and more roundabout haul; and, moreover, was less than the shorter haul rate from Wichita to the Gulf direct. The analogy to the Hillsdale case, above-described, will appear clearly on inspection of the map. Summarizing the results so far reached, in all that concerns the two sorts of cases considered in the preceding paragraphs, our conclusion is that, when competition by rail at the distant point is alone present, and when the nearer point is on a roundabout route, a railway "is entitled to carry the traffic past X to Y (Philadelphia) for considerably less than nothing"; but, when the nearer point is on a direct line, the case is debatable. Proof that normal competition compels the lower rate at the remoter station must be uncommonly clear and conclusive. In other words, the facts that the rate at Y is not unduly low and also that the rate at X is not unreasonably high must both be firmly established. A distinct class of cases of local discrimination is suggested by the recent case of Montgomery, Ala., in the United States Commerce Court.[217] These like other cross line cases, akin to that of Wichita, Kansas, above-mentioned, arise in connection with practices as to the division of joint rates. They will be discussed in connection with pro-rating in our second volume.