About the time that cantilevers came into vogue, some twenty-five or thirty years ago, certain bridge designers entertained a wild idea to the effect that the new type involved some special virtue or feature of excellence or else that it was economic in first cost; because many cantilever bridges were built in places where simple-span structures would have been far better and cheaper. Possibly the thought of establishing an innovation induced some of the designers of those bridges to prefer the cantilever type to that of the simple truss. What a pity it is that such designers did not devote their time and energy to an attempt to introduce the steel-arch bridge into American practice! Had they done so, probably they would have been successful; because there is often true economy in the arch—besides it is far more aesthetic than either the cantilever or the simple truss. A long-span, cantilever bridge can be made agreeable to the eye by using artistic outlines and a well-studied web-system; and, again, its simple vastness produces a pleasing impression upon the beholder; but a small-span cantilever is ugly and causes a trained intelligence to propound to itself the question "why and wherefore?" without receiving a satisfying answer.

It is true that for certain fairly-narrow crossings the water is so deep, or the current is so swift, that the use of falsework is out of the question, and that the adoption of cantilevering during erection is necessitated. Such conditions, however, do not require cantilever bridges, but semi-cantilevers, i.e., structures that are cantilevers during erection and either simple spans or arches afterwards. This method of erection for simple-truss spans was first evolved by the author some twenty-five or thirty years ago, but was not actually used by him in construction until some time later. A description of it will be found in Chapter XXV of "Bridge Engineering."

It is evident to any engineer who gives the subject due consideration that a cantilever bridge is less rigid than the corresponding simple-truss structure, because its vertical deflections under live load are necessarily larger, thus permitting more vibration as well as greater irregularity in the track grade; hence, for steam-railway bridges, other things being equal, the simple-truss layout should be chosen—or even if it should cost somewhat more, because rigidity is an important consideration in the operation of steam-railway trains. For highway and electric-railway bridges, though, it is not of such great importance; consequently, if in these structures the

83