is any difference between this principle and an old practice of forty years' standing. The principle is that "for any crossing the greatest economy will be attained when the cost per lineal foot of the substructure is equal to the cost per lineal foot of the trusses and lateral systems." The old practice was to make for economy the cost of a pier equal to the cost of the span that it supports, or, more properly, equal to one half of the cost of the two spans that it helps to support.

Is not the difference between these two methods perfectly plain? In one the cost of the pier is made equal to the cost of the trusses and laterals, and in the other it is made equal to the cost of the trusses, laterals, and the floor system. When one considers that the cost of the floor system is sometimes almost as great as one half of the total cost of the superstructure, he will recognize how faulty the old method was.

The following is the demonstration of the principle, simplified to the greatest practicable extent. Let us assume a crossing of indefinite length, for which the depth of bed-rock is constant, and let