Figure 8.—The Baltimore and Ohio Railroad’s Potomac River crossing at Harpers Ferry, about 1860. Bollman’s iron “Winchester span” of 1851 is seen at the right end of Latrobe’s timber structure of 1836, which forms the body of the bridge. (Photo courtesy of Harpers Ferry National Historical Park.)
The B. & O. was in every way a truly pioneer enterprise. It was the first practical railroad in America; the first to use an American locomotive; the first to cross the Alleghenies. The spirit of innovation had been encouraged by the railroad’s directors from the outset. It could hardly have been otherwise in light of the project’s elemental daring.
The first few major bridges beyond the line’s starting point on Pratt Street, in Baltimore, were of rather elaborate masonry, but this may be explained by the projectors’ consciousness of the railroad’s significance and their desire for permanence. However, the aforementioned economic factors shortly made obvious the necessity of departure from this system, and wood was thereafter employed for most long spans on the line as far as Harpers Ferry and beyond. Only the most minor culverts and short spans, and those only in locations near suitable quarries, were built of stone.
In addition to the economic considerations which prompted the company to revert to timber for the major bridges, there were several situations in which masonry construction was unsuitable for practical reasons. If stone arches were used in locations where the grade of the line was a relatively short distance above the surface of the stream to be crossed, a number of short arches would have been necessary to avoid a very flat single arch. In arch construction, the smaller the segment of a circle represented by the arch (that is, the flatter the arch), the greater the stress in the arch ring and the resulting horizontal thrust on the abutments.
Figure 9.—Bollman skew bridge at Elysville (now Daniels), Maryland, built in 1853-1854. (Photo courtesy of Maryland Historical Society.)
The piers for the numerous arches necessary to permit an optimum amount of rise relative to the span would have presented a dangerous restriction to stream flow in time of flood. By the use of timber trusses such crossings could be made in one or two spans with, at the most, one pier in the stream, thus avoiding the problem.
The principal timber bridges as far west as Cumberland were of Latrobe’s design. These were good, solid structures of composite construction, in which a certain amount of cast iron was used in joints and wrought iron for certain tension members. They were, however, more empirical than efficient and, for the most part, not only grossly overdesigned but of decidedly difficult fabrication and construction.
What is interesting about the Latrobian timber trusses, however, is the effect they appear to have had upon Bollman’s subsequent work in the design of his own truss. This effect is evidenced by the marked analogy between the primary structural elements of the two types. The Latrobe truss at Elysville ([fig. 2]) was only partially a truss, inasmuch as the greater part of the load was not carried from panel to panel, finally to appear at the abutments as a pure vertical reaction, but was carried from each panel (except the four at the center) directly to the bearing points at the piers by heavy diagonal struts, after the fashion of the famous 18th-century Swiss trusses of the Grubenmanns. It was a legitimate structural device, and the simplest means of extending the capacity of a spanning system. However, it was defective in that the struts applied considerable horizontal thrust to the abutments, requiring heavier masonry than would otherwise have been necessary.