It would also seem possible to tell in what direction, if any, the bend in the inclined stirrups should be made.  It is to be assumed, although not expressly stated, that the bends should curve from the center up toward the end of the beam, but an inspection of the stress triangles, Fig. 13, will show that the intensity of stress is just as great on the opposite side, and it is probable that, if any bends were required to reduce the maximum stress in the concrete, they should as likely be made on the side nearest the abutment.

From the stress triangles it may also be shown that, if the stirrups were vertical instead of inclined, the stress in the concrete on both sides would be practically equal, and that, in consequence, vertical stirrups are preferable.

The next issue raised by the author is covered in his seventh point, and relates to bending moments.  He states:  “* * * bending moments in so-called continuous beams are juggled to reduce them to what the designer would like to have them.  This has come to be almost a matter of taste, * * *.”

The author seems to imply that such juggling is wrong.  As a matter of fact, it is perfectly allowable and legitimate in every instance of beam or truss design, that is, from the standpoint of stress distribution, although this “juggling” is limited in practice by economical considerations.

In a series of beams supported at the ends, bending moments range from (w l^{2})/8 at the center of each span to zero at the supports, and, in a series of cantilevers, from zero at the center of the span to (w l^{2})/8 at the supports.  Between these two extremes, the designer can divide, adjust, or juggle them to his heart’s content, provided that in his design he makes the proper provision for the corresponding shifting of the points of contra-flexure.  If that were not the case, how could ordinary bridge trusses, which have their maximum bending at the center, compare with those which, like arches, are assumed to have no bending at that point?

In his tenth point, the author proposes a method of simple designing by doing away with the complicated formulas which take account of the actual co-operation of the two materials.  He states that an ideal design can be obtained in the same manner, that is, with the same formulas, as for ordinary rectangular beams; but, when he does so, he evidently fails to remember that the neutral axis is not near the center of a reinforced concrete beam under stress; in fact, with the percentage of reinforcement ordinarily used in designing—­varying between three-fourths of 1% to 1-1/2%—­the neutral axis, when the beam is loaded, is shifted from 26 to 10% of the beam depth above the center.  Hence, a low percentage of steel reinforcement will produce a great shifting of the neutral axis, so that a design based on the formulas advocated by the author would contain either a waste of materials, an overstress of the concrete, or an understress of the steel; in fact, an error in the design of from 10 to 26 per cent.  Such errors, indeed, are not to be recommended by good engineers.