The author obtained a fine check on the correctness of this formula when testing the draw-span of his Jefferson City highway bridge. This span of 440' weighs 660,000 pounds, and was opened by four men in four minutes and fifty seconds. The power applied by the men was measured by dynamometers, and from the length of their path and from their pull the horse power was computed. It proved to be just a little less than unity, so near in fact that it was called unity. The velocity v was, on the average, 0.066 feet per second. Substituting in the formula gives

H. P. = 0.0125 X 660,000 X 0.066 ÷ 550 = 0.99.

It is possible that, if the experiments were to be made again, a greater divergence from the formula would be found, for the reason that the bridge is liable to work more easily after it has been operated a while.

The computation of stresses in ordinary draw-spans involves more or less ambiguity. The assumptions upon which the calculations are based are the following:

1.   The truss-rods in the tower are so light that they cannot transfer any shear past the pivot-pier; consequently, with a live load on one arm only, the said arm acts entirely independently of the other, thus making the draw for this loading consist of two simple spans.

2.   For live load on both arms, the reactions are to be found on the assumption that the draw is a continuous girder on four points of support, and by a formula based upon the Theorem of the Three Moments with a constant moment of inertia.

Plate IX gives a diagram from which can be read at a glance the percentage values of the reactions for any balanced load placed anywhere on any span. It is perhaps theoretically not quite perfect, because the values of the reactions depend slightly upon the ratio of dis-

tance between the two middle points of support to length of one arm; but any error made by assuming this ratio as constant for all drawbridges is a