A steady increase in the freight performance of American equipment is shown by official data of the Interstate Commerce Commission. Whereas in 1894 it required on an average 1,888 freight cars for every 1,000,000 tons of freight transported their capacity has so increased that the same amount of traffic in 1906 was carried by only 1,127 cars. In other words, an enormous increase in the freight service had been attained. On the other hand, the actual mileage performance of much of this equipment is extraordinarily low. It averages only about 9,000 miles annually or an equivalent of thirty miles a day. At a speed of fifteen miles an hour, this means that actual movement under a paying load, allowing for one-third of its journeyings empty, occupies but little over an hour and a quarter a day. The actual performance is, however, not quite as poor as appears. For, of course, this average includes the non-movement of all cars in bad order (sometimes one-tenth of the total); and also all idle equipment. This latter consideration is of great moment. Special cars, suitable only for seasonal business; and especially demurrage delays, often forty-eight hours or more, adversely affect the result. Where separate mileage records of "foreign" cars are kept, as on the Wabash system, it appears that their mileage is twice as high as for "home" cars. The difference is due to the fact that cars off their own rails, mainly are in actual demand and are kept moving. Probably the daily performance of loaded cars is not less than 150 miles. But a journey of this length, with two days' delay at each end at terminals, would bring the average down to about thirty miles. The public does not always appreciate these facts; and is often querulous. It is certain that the problem how to secure greater efficiency in the use of this equipment is as yet imperfectly solved.[61] [Illustration: Revenue and Cost Per Train Mile.]