[ Figure 6.]—Plan of Harpers Ferry bridge as built by Latrobe. The second Winchester track was later removed.

Of these pioneer civil and mechanical engineers, some were formally trained but most were self-taught. Bollman’s career on the B. & O. is of particular interest not only because he was perhaps the most successful of the latter class but because he was probably also the last. He may be said to be a true representative of the transitional period between intuitive and exact engineering. Actually, his designing was a composite of the two methods. While making consistent use of mathematical analysis, he was at the same time more or less dependent upon empirical methods. For years, B. & O. employees told stories of his sessions in the tin shop of the railroad’s main repair facility at Mount Clair in Baltimore, where he built models of bridges from scraps of metal and then tested them to destruction to locate weaknesses. It seems most likely, however, that the empirical studies were used solely as checks against the mathematical.

Figure 7.—Recent model of Bollman’s Winchester span. Only two of the three lines of trussing are shown. The model is based on Bollman’s published description and drawings of the structure. (USNM 318171; Smithsonian photo 46941.)

In the period when Bollman began designing—about 1840—there were fewer than ten men in the country designing bridges by scientifically correct analytical methods, Whipple and Roebling the most notable of this group. By 1884, the year of Bollman’s death, the age of intuitive design had been dead for a decade or longer.