computations showed that, for long-span, double-track, steam-railway bridges of the Petit-truss type, there is no economy of metal whatsoever in making adjacent spans continuous over the piers, and that the matter of caring for or ignoring the effects of reversing stresses does not in any way influence the economics.

These results were so decided and the coincidence of weights was so exact that at first the author thought the entire question was settled; but it was suggested by Mr. Shortridge Hardesty, his principal designing engineer, that, if the divided-triangular truss adopted for the Sciotoville Bridge were investigated, a different result might be found; and it was decided to make the test. The layouts of trusses for the new comparison are shown in Figs. 11c and 11d. The computations were all prepared exactly as before, and it was found necessary to make two sets of figures for the continuous-truss layout before the correct dead load was determined. The findings of this second set of computations were as follows:

In the simple-truss spans the weights of metal for the divided-triangular and the Petit types were nearly alike, the slight difference which there was being in favor of the former; and the continuous-truss type showed a gain of twelve per cent over the non-continuous type when reversion was ignored and eleven per cent when it was properly provided for.

The computations made up to this stage of the investigation settled the economics for long-span, steam-railway bridges; but it was seen by a priori reasoning that the results would probably be somewhat different for standard highway bridges, consequently it was decided to repeat the calculations for the latter structures. To this end the total loadings (i.e., live load plus dead load) were retained, but a different division thereof was made by diminishing the live loads and increasing the dead loads, so as to correspond, as nearly as may be, with the ratios of those loads in modern highway bridges having a paved roadway, reinforced-concrete base-slab, and reinforced-granitoid sidewalks for the span-length under consideration.

The result of this third set of computations showed that for the divided- triangular-truss layout there was an economy of twenty-two per cent in favor of the continuous trusses, and that reversing stresses affected very few members and those so slightly that the influence of reversion could be completely ignored. It was not thought worth while to repeat these highway-bridge calculations for the Petit-truss layouts.

The computations made thus far settled the comparative economics of continuous and non-continuous trusses for long-span bridges, both railway and highway, also those of the Petit and the divided-triangular trusses for such structures; but it could not properly be assumed that what was found to be true for long spans would apply also to short ones, consequently it was decided to make a new set of computations for comparatively-short-span layouts for steam-railway bridges of both the divided-triangular truss and the Pratt-truss types.