being required mainly for stiffening certain truss members to resist erection stresses. Of course, if false work be used for the suspended span, the eight per cent excess will not be added.

The curves of percentages are based on two assumptions, viz.: first, the panels throughout the entire structure are of equal length, and, second, the lengths of the cantilever-arms and anchor-arms are the same. The first assumption is nearly always correct, for there is no advantage to be gained by varying the panel lengths in the various portions of the bridge. If the lengths of cantilever and anchor arms are unequal, the average weight of metal obtained for the latter by use of the curve will have to be corrected by the formula

where T' is the correct, final weight of truss and lateral metal in the anchor-arm, T is the weight of same found by the percentage curve, and r is the ratio of length of cantilever-arm to that of anchor-arm.

It should be observed that, in applying the percentage curves to structures having subdivided panels like those of the Petit truss, the main or double panel is to be used as the basis of calculation.

The method of applying the percentage curves is as follows: Let us take any opening and assume that there are six panels in each cantilever-arm, and that the weight per foot of truss and lateral metal in the suspended span is w, the panel length being p, and pw = W. It is to be observed that this method is applicable for any proportionate length of suspended span.

The weight of metal in the floor system, being independent of the span length and simply a function of the panel length and of the distance between trusses, is not considered in the investigation, but is, of course, to be added when figuring the total weight of metal in the structure.

The weight of truss and lateral metal in the cantilever-arm will be

1.2W + 1.4W + 1.65W + 2.0W + 2.4W + 3.OW = 11.65W.