If the assumption be made that three panel-lengths be cantilevered, it is almost certain that the total weight of metal in the structure will be augmented.

It is evident that the proportionate effect of the cantilevering under consideration is dependent upon whether the bridge is a railroad structure or a highway one with a paved roadway supported on a reinforced-concrete base, because the relative effect of reversion is far greater in the former case than in the latter; hence an amount of cantilevering of this kind that would be uneconomic in a railroad bridge might be truly economic in the corresponding highway structure.

As the span-length increases, the ratio of live load to total load decreases, and hence the proportionate effect on weight of metal due to reversing stresses diminishes. For this reason one can anticipate that the longer the average span the greater will be the relative importance of the economic feature of design under consideration.

Finally, the saving in metal (or the reverse) by this method of cantilevering is fundamentally dependent upon the manner in which the designing specifications take care of reversing stresses. If these be entirely ignored, as some engineers advocate doing, the cantilevering will effect a large economy of metal, even when the cantilever arms are comparatively long; whereas, if these stresses of opposite sign are cared for by adding to the larger three-quarters of the smaller and proportioning for the sum, the saving will be but little, if any. The most approved and up-to-date practice is to add to the larger stress only one-half of the smaller; and in that event some economy may be anticipated, provided that the length of the cantilever arms be not too great.

If, with a layout such as is being considered, there be found for openings of equal size an economy in a certain amount of cantilevering, the question arises "would there not be a further saving of metal, if the lengths of the continuous spans were to remain fixed and those of the other spans were to be moderately. increased?" The answer to this question, in all probability, is affirmative, although the economy involved would not be important.

In view of the preceding dissertation, it is evident that it is entirely impracticable to give any quantitative solution to this economic question, but that it must be solved for each case as it arises. In railroad bridge designing the matter is not important; because it is highly improbable that the value of the metal saved would offset the disadvantage of the reduction in rigidity that is unavoidable when changing from the fixed-span type to that of the cantilever. But in the case of a highway bridge with a heavy floor, it is an altogether different matter, because, as previously pointed out, rigidity is not so fundamentally important in highway structures as it is in railway bridges; and, moreover, the stiff concrete slab itself increases so greatly the rigidity of the steel construction that the detrimental looseness caused by the hinged attachments of the suspended span loses most of its importance.