has never yet been given any attention—at least none in print. It indicates within rather narrow limits a slight economy for the cantilever type, but the amount thereof and the location of the said limits are dependent upon several considerations, among which the most important are the following:

   A. Average length of spans considered.

   B. Ratio of live load plus impact to total load.

   C. Method adopted for combining reversing stresses, when pro-

portioning sections of members.

The best conception of this matter of economics can be obtained from a dissertation based upon an assumed layout of simple-truss spans, all of equal length,—for instance, a long succession of like spans of two hundred feet each, the panel-lengths being twenty feet. If now we extend the trusses of every other span one panel beyond each of its piers and suspend from the cantilevered ends thus formed the shortened intermediate spans, we shall have a cantilever bridge that will effect a saving in weight of metal in every span. It is evident without any figuring at all that the spans which contain the suspended trusses will weigh less than the simple-truss spans, because the suspended portion is decidedly lighter and the cantilever arms, being so short, cannot be very heavy. Again, the dead load stresses in the chords of the other spans are somewhat reduced, but probably not enough to permit of any reversion of stress when the span is empty with the adjoining spans loaded. The total stresses in chords are, therefore, materially smaller than those for the simple-truss spans, resulting in an economy of metal, notwithstanding the fact that there are two extra panel-lengths of top chord involved by the change from simple-span type to cantilever type, and that the vertical posts over the piers combined with the end main diagonals are somewhat heavier than the inclined end posts of the simple-truss span. On the other hand, though, the cantilever structure, having only one pair of pedestals per pier, involves a slight economy of "metal on piers" and permits the width of pier-top to be reduced a little below that required for the two pedestals of consecutive trusses in the simple-truss layout, which saving in some cases will extend from coping to bottom of caisson.

Next, let us assume that there are cantilever extensions of two panel lengths instead of one. There may or may not be a material saving of metal in those spans containing the suspended trusses in comparison with simple-truss spans, although there will be some reduction; for, while there is a decided lessening in the weight of metal per lineal foot in the suspended portion, the cantilever arms tend to become heavy. In the chords of the other spans, the dead-load stresses are made very small, permitting some reversion therein from the live loads on the adjoining spans; and the result will probably involve an increase in weight of truss metal. The net effect of the change in layout upon the structure as a whole is uncertain; but it is probably slightly uneconomic of metal.