cuts down the quantity of metal, because, in the members where they are used, the weight of steel for both main sections and details is an absolute minimum; and, in general, pins weigh less than connecting plates. Whether secondary stresses receive proper consideration or not, a pin-connected bridge is somewhat lighter than the corresponding riveted one, and therefore ought to be less expensive. It is true that the fine shopwork requisite for the proper manufacture of pins and for the drilling of pin-holes makes the pound price for fabrication greater in the pin-connected structure; but this is offset more or less by its lower pound cost for erection, owing to the smaller number of field rivets to be driven, the shorter time required to make safe against loss of span by washout of falsework, and the reduction in overhead expense effected by minimizing the time for field operations.

From a study of the diagrams of weights of metal per lineal foot of span for pin-connected and riveted-truss bridges given in Chapter LV of "Bridge Engineering," it is found that the weights of the latter exceed those of the former by the following percentages:

Simple Spans

It is seen from these tables that the percentage of saving by using pin-connections increases gradually with the span-length. This is because of the effect of the reduced dead-loads.

Notwithstanding the fact that pin-connected-truss bridges are somewhat lighter, and possibly somewhat cheaper, than the corresponding riveted-truss bridges, the latter are decidedly preferable for all short or medium-length spans of railway structures (both steam and electric); because in short, pin-connected spans the vibration from rapidly-passing loads is so great that the motion of the eyes on the pins causes such a grinding of both that eventually the structure has to be replaced on that account. This was shown long ago to be the case in pin-connected elevated railroads; and lately it has been found necessary to replace pins in a number of rail-