From the tests conducted by the American Railway Engineering Association, it is found that the maximum impact which will be obtained at reduced speed is:

Less than 30% for speed of 10 miles per hour.

Less than 40% for speed of 15 miles per hour.

Less than 50% for speed of 20 miles per hour.

Less than 55% for speed of 25 miles per hour.

An inspection of the diagrams indicates that the effective span of the bridges and the characteristics of the engine loadings determine to a great extent whether a given loading can be run over the bridge, and shows that it is unsafe to attempt to decide whether any engine loading can be handled over a structure simply by knowing its total weight.

There is, unfortunately, a misunderstanding, among some railroad operating officials, as to the effect on bridge structures of such complex loadings as locomotives and cars. In these cases it is assumed that the effect is the same for all locomotives of the same total weight; and bridges are classified as being safe, or otherwise, for all locomotives of given total weights. If this practice must be resorted to, the limits set should be on a very conservative basis; for otherwise there would be danger of certain types of locomotives having a serious effect on some bridges, producing unsafe conditions. The practice would not be economical, because it would either lead to the premature renewal of some bridges or to an unnecessary ruling off of certain types of engines.

Where Low Classification Usually Occurs in Bridges

In older bridges there are certain parts where low classifications can usually be expected. These have been found to occur most often in the lightest members of the structure and in members which carry the smallest dead-load stresses. In proportioning a member, a part of the sectional area thereof can be taken as carrying dead-load stress and the remainder live-load stress. As the dead-load stress is constant, a smaller area would be required where a higher unit stress is used. This, therefore, leaves a portion of the area originally provided for dead-load stress available to carry live-load stress.

It is found that the floor systems of bridges have generally a lower classification than the girders or chords of the trusses. The low classification of stringers is generally in the section of the flanges near the center, in the riveting of flanges near the ends (particularly if they are shallow), and in the riveting connecting the stringers to the floor beams.

Floor beams, if of shallow depth, frequently show a low classification in flanges near the stringer connections, also in the riveting of flanges near the ends, and in splices connecting the webs of the floor beams to the gusset plates, particularly in types where the lower part of the floor beam is cut out to fit around the ends of the trusses or over pins.

In plate girders, the flanges frequently show low classification at points where the web is not fully spliced near the center and at points near the ends of cover plates. The flange riveting near the ends of girders frequently has a low classification, particularly where the girders are shallower at the ends.

Webs of plate girders show low classification near the ends of the girders where there is a relatively large expanse of web, unsupported by stiffeners. The web splices near the ends of the span have a low classification where only one line of rivets is used on each side of the splice.

In trusses, the posts and diagonals near the center of the span usually show a low classification. This is particularly true of the diagonals and counter-diagonals of light eye-bars or loop rods.

Suspenders, or hip-vertical members, frequently have a low classification. The