floor-beams should be employed and be riveted to the end posts of the trusses so as to make the lower lateral system a harmonious whole; and it matters not if the end posts be inclined.

Manufacturers are willing to use single angles in tension, but this is objectionable because of the violation of the rules of symmetry and the consequent causing of secondary stresses.

A mooted point in designing is the exact location of top-chord pins. The author believes they should always be placed either on or a very little below the gravity lines of the sections, for it takes an exceedingly small eccentricity to produce a high intensity of bending stress on the chord. It is not right to assume that the reverse bending moment due to the weight of the member between panel-points will entirely counteract the bending moment due to the eccentricity; because the form taken under loading by the center line of the long strut will be a waved line passing through the centers of the chord pins, being concave upward in one panel and convex upward in the adjoining one. The amount to lower the chord pins below the centers of gravity of the sections is to be determined by making the compressive intensity due to eccentricity equal to the tensional intensity due to bending from weight of member. In any case this adjustment is a matter of compromise on account of the shifting of centers of gravity from centers of figure by reason of the variation in make-up of section from panel to panel, the amount ordinarily varying from zero to a quarter or three-eighths of an inch. In large bridges, of course, the variation will be greater than this.

Manufacturers like to use cast iron in bridges on account of its comparative cheapness per pound; but on general principles the author tries to bar out all cast iron from his bridges, fearing that, if it be permitted in one place, the contractor will insist upon putting it into another. Cast iron is nearly always inferior to cast steel for any purpose.

There is an uneconomic detail which is too often employed in both trestles and elevated railroads, viz., the insertion of a heavy casting between the lower end of a column and the masonry. The author has never been able to perceive the philosophy of this detail; for it involves the planing of a large amount of extra surface as well as a considerable increase in the weight of metal. Moreover, the additional surfaces in contact do not militate towards rigidity, for perfect contact is not always attained. The object is evidently the spreading of the load over the masonry; but this could be accomplished just as effectively and at less cost by using a rolled base-plate of the proper size and carrying the load to it from the column section by means of vertical plates, horizontal connecting angles, and vertical stiffening angles. The necessity for giving the latter a tight fit at their lower ends requires some troublesome shopwork, but the additional cost thereof will not offset the expense of the extra amount of planing involved by the casting detail.

In respect to the general and detail principles of the economics of shop-