the main opening. The explanation of this is that the weight per foot of the chords, though independent of the upward bending moment, increases proportionately to the downward bending moment with the length of span; while the weight per foot of the web, in so far as it is affected by the shears from exterior loading, the ruling factor in determining the sections of web members, varies inversely as the span length.

If the length of anchor-span be very short, say materially less than one half of the main opening, the weight per foot for trusses and laterals will have to be increased to 3.5 W, notwithstanding the fact that the entire top chords may then be built of eye-bars; but such short spans would probably be barred out by consideration for navigation interests.

The percentage curves of Plate X will not bear a rigid criticism, in that they make the weight of metal depend upon the number of panels. It is presupposed, however, that the panel length adopted is the most appropriate one for the bridge; and the curves will be found quite accurate whenever the proper panel length is used. With long panels the weight of metal per lineal foot found by the curves for cantilever and anchor arms is less than that found thereby for short panels. This is as it should be, but to a limited extent only; for it can be found by trial that an abnormally short or abnormally long panel length will give results too great or too small when checked by computations of weights made from actual designs.

These percentage curves enabled Mr. Hedrick to solve readily the next problem, viz., given the total distance between centres of anchorages and carte blanche as to the location of the main piers, to determine the length of each anchor-arm which will make the total weight of metal in the structure a minimum. He found this length to be two tenths of the total distance between the anchorages.

It must not be forgotten that for every dollar saved by reducing the total weight of metal through the shortening of the anchor-arms, it  will be necessary to  spend  about  twenty  cents  for  extra  concrete  in  the  anchorages.  On  this account,     for   the   conditions   assumed,   the   truly   economic   length   of   each