important point as the comparative economics of cantilever and suspension bridges. Some time in the not very distant future there are going to be built in this country many long-span bridges; and it behooves engineers to know in advance the economics of the different types of structures applicable thereto. Dr. Steinman deserves great credit for his energy and courage in attacking such a stupendous problem at such an early date in his professional career, without any records of weights at his disposal, and before he had had any actual experience in bridgework. In undertaking such an immense task he set a splendid example to other young engineers; and the incorrectness of his conclusion is no blot whatsoever upon his professional record. It would be well for engineering if there were in its ranks many more young men possessing the attributes of energy, ambition, and love for hard work to the same extent that he does. Such men will be badly needed in every branch of technics, if our profession is to take the high position in the community to which it is entitled by its importance to mankind.

Dr. Steinman can console himself with the reflection that he is not the only engineer who has devoted an entire treatise to the production of a wrong conclusion, for several decades ago an eminent French professor of engineering published a large book dealing with the economics of truss bridges, basing his calculations upon such incorrect premises that the result of his work was of no real value to the profession.

Comparing the results of the preceding calculations, as shown in Fig. 13c and Fig. 13i, it will be noticed that the span length for equal cost is much less for the combined railway and highway type of structure than for the strictly railway type. The reason for this is that in a modern highway bridge the proportion of dead load to live load is much greater than it is in a railway bridge, because of the large weight of the pavement, the supporting slabs, and the concrete footwalks. In the stiffening trusses of a suspension bridge it is generally the live load only which causes stresses that influence the sectional areas of the members, the dead load having no effect thereon whatsoever, but in a cantilever bridge it is the total live load plus the dead load which does so, with sometimes a little assistance from the wind load; hence it is evident that the smaller the proportion of live load to total load the more favorable it is for the suspension bridge. On this account, in strictly highway structures, the span length for equal cost will be much shorter than those thus far determined. Not knowing what the length would probably be, the author figured the costs of the two types for 1,500-foot, 1,200-foot, and 1,000-foot main openings, using carbon steel only; and from the results of the computations he plotted the curves in Fig. 13j. From this diagram it will be seen that the span length for equal cost is about 1,000 feet.

Recognizing that this investigation would not be complete without preparing a set of computations for strictly-highway structures of nickel steel, the necessary calculations were made for a 1,000-foot span of each