Smith Trusses: Bringing Covered Bridges into the Industrial Age


Matthew Reckard, P.E.

J.A. Barker Engineering, Inc., Bloomington, Indiana




Figure 1. Cataract Covered Bridge, Owen County, Indiana. A 136 foot (41 m) span Smith truss.





       Robert Smith (1833-1898) is thought to have been the first builder to prefabricate bridges. The Smith Bridge Company, begun in 1867, operated out of a factory in Toledo, Ohio. In the company's early years they built hundreds of covered wooden bridges, the vast majority of them using Smith's patented truss system.

       Smith trusses were produced in a manner very much at odds with the romantic conception of covered bridges as a product of pre-industrial times. Instead they show many features characteristic of the late 19th Century industrial age we are more likely to associate with iron bridges, which Smith also designed and built.

       The author illustrates these points using as examples the 1876 Cataract Bridge in Owen County, Indiana, for which he designed repairs, and eight other Smith trusses in Indiana and Ohio.



1. Introduction


       Covered wooden bridges were built in great numbers in the United States and Canada during the 19th and early 20th Centuries. They have been popular with tourists for a century. Much of their allure is due to their romantic association with times before the advent of the motor vehicle. Covered bridges, designed for the horse-drawn buggy rather than the automobile, are a powerful symbol of an era seen as slower paced and simpler than today.

       More specifically, covered bridges are seen as a late preindustrial technological development, with roots in the medieval craft of timber framing. And in many covered bridges, especially older ones, this is true. The structures are hand crafted with simple tools. The timbers, cut from the nearby forests, display hew and blade marks from the broad axe and the pit saw. Joints are mortises and tenons fastened with trunnels (wood pegs).

       But in other covered bridges this vision is not correct. This is especially true of later Smith truss bridges. Although these are all-wood trusses (most haven't so much as a lateral tie rod of metal), there is not a hew mark or a mortise to be seen. They were sometimes erected hundreds of miles from where they were fabricated, and were fabricated using heavy machinery far from where the lumber was cut. Instead of late preindustrial designs, they are significant and fascinating examples of early industrial construction technology.



2. Robert Smith and the Smith Bridge Company



Text Box:  
Figure 2. Robert Smith
(photo: Miriam Wood Collection)

Robert Smith was born in 1833 in Miami County, Ohio, the son of a cabinetmaker who had emigrated from Maryland. As a young man he worked as a carpenter and is said to have invented a roof truss used in barns. He received his first bridge patent in 1867, and "in 1867 built 5 bridges; in 1868, 22; and in 1869, 75."[1] Smith moved his nascent Smith Bridge Company to Toledo in 1869, reorganized it as a joint stock company in 1870, and was the company's President until 1890. During these years the company produced hundreds of bridges annually.  Biographical information about Smith can be found in Waggoner (ed) and Neff.[2] A previous report by J.A. Barker Engineering contains a more lengthy bibliography.[3]

       The company prefabricated bridges in Toledo and then shipped them to the surrounding mid-Western states for erection by their own crews or others. Smith also received royalties for bridges designed and built to his patents by others. The total number of Smith trusses built over the years may have been in the thousands. At most a few dozen Smith truss bridges remain. One recent count listed only 22 (14 in Ohio, 6 in Indiana, one in Pennsylvania, and one in California), although it included no "single truss" Smiths (discussed later in this paper).[4]

       The company not only built Smith trusses, but also Howe and other truss types when their customers wanted them. They developed a composite truss, much like a Smith truss but with iron rods for the tension diagonals, for use on railroad bridges.  Furthermore, the company began building all metal bridges as early as 1870, and became known for swing and draw bridges. By 1890, when Smith sold the company, they had stopped making wooden bridges altogether.[5] Renamed the Toledo Bridge Company, the company was purchased in 1901 by J.P. Morgan who combined it with 24 other bridge companies to create the American Bridge Company, which survives today.[6]



LEAD Technologies Inc. V1.01

Figure 3. Smith Bridge Company factory, Toledo, Ohio (photo: Miriam Wood Collection)



      According to Waggoner, much of Robert Smith's success was due to his inventiveness in industrial machinery. Among his developments was "a gaining-machine, which does the work of 15 men, ... a process for making a steel eye-bar, ... a rotary saw, for making the joints of bridge-chords; and a multiple punch, by which six pieces of iron can be punched at one operation."  The "gaining-machine" probably cut notches in truss timbers.[7]


3. Smith Trusses


      Robert Smith received two patents for wooden bridge construction. The first (US # 66,900, dated July 1867) covered the truss arrangement; the second (US # 97,914, dated December 1869) covered a portal bracing system and castings used to anchor diagonal lateral braces to the chords. Truss drawings from these patents are shown as Figures 4 and 5. The second patent did not claim a new truss arrangement as part of the invention, but nevertheless shows a truss arrangement at midspan different than the one in the first patent.



Figure 4. "Type 1" Smith Truss, from patent drawings of 1867.



Figure 5. " Type 2" Smith Truss, from patent drawings of 1869.


Figure 6. "Type 3" Smith truss in the 1876 bridge at Cataract, Indiana.

Drawn showing original camber (slight arch). Today the trusses sag

slightly due to failure of some lower chord timber tension splices.


      A third truss arrangement at midspan was used beginning in the mid-1870s; like the second, this was never separately patented.  The bridge at Cataract, Indiana is an example of this "Type 3" Smith truss; a drawing of the trusses there is shown as Figure 6.

      As can be seen in Figures 4 through 6, all Smith trusses have vertical end posts flanked by unique bracing posts that slant sharply upwards from the end posts at the lower chord to meet the first regular diagonal web member. Most or all other Smith truss web members are diagonals. In the trusses shown in Figures 4 through 6 these are arranged in two overlapping sets. Each set forms a string of 'W's, like a Warren truss. The two sets are offset, however, so that diagonals cross in the middle to form a pattern of 'X's. For this reason Smith trusses have sometimes been called "double intersection Warren" trusses. The two sets of diagonals are sandwiched between triple chord timbers.

            Smith's first patent, however, notes other possible arrangements. The one shown in the drawing he calls "a double-truss bridge, as there are two sets of posts and braces. Bridges may be made with a single truss, or with three or more, depending upon the strength required." In practice Smith built "single truss" spans (with double-timber chords) only for very small spans. On longer spans he used three sets of diagonals between four chord timbers. In these cases the outer two sets of diagonals parallel each other with the middle one running counter to them.



4. Previous Literature about Smith Trusses


       There is limited discussion of Smith trusses in the literature. Modern observers of Smith trusses have been puzzled by several of their features. Among these are:

€       Some diagonals pass between and extend beyond the chord timbers while others don't.

€       Pairs of diagonals generally meet at a point on the chords, but near mid-span they sometimes intercept the chords a short distance apart.

€       The bridges were built using traditional English units (feet and inches). Yet some basic dimensions of the trusses don't seem to be sized in whole numbers of these units. At Cataract, for example, measurement between bolts where diagonals cross (the centers of the 'X's) yields an apparent stasndard panel length of 10' 11 5/8".

€       Dimensions one would expect to be repeated sometimes aren't. For example most bolts mentioned above are almost (but not exactly) 7' 4" above the lower chord, but the two nearest mid-span are almost (but not exactly) a foot lower than this.

€       Splices in the chords are located nearly midway between panel points, but never exactly midway. The difference varies from almost nothing to about ten inches.


        What little analysis there has been of Smith truss construction has, unfortunately, sometimes been wrong. Richard Allen, for example, says that Robert Smith "devised a new type of bridge truss, using a double set of trussed timbers without any tension members."[8] Not so: about half of the diagonals (and the entire lower chord) are tension members in a Smith truss. Miriam Wood, following Wilson's example, claims that Smith trusses "featured braces set at 45 degrees and counterbraces set at 60 degrees".[9] Not so: while the drawing for Smith's second patent shows diagonals at about these angles, I have found no diagonals near 45° in actual bridges, and only a few near 60°.

       Raymond Wilson, in a paper published in 1967, classified Smith trusses into four types.[10] This taxonomy is generally in use today. His "Type 1" is based on the 1867 patent drawing (Figure 4), distinguished by a pair of vertical posts at mid-span. Type 2, Wilson says, is "the 1869 Patent Truss" (Figure 5), distinguished by a pair of diagonals forming a 'V' at mid-span (although, as noted, the 1869 patent wasn't for a truss arrangement). Type 3 has diagonals crossed to form 'X's at midspan as well as elsewhere throughout its length, as seen in Figures 6 and 7. Wilson's Type 4 Smith trusses are like Type 3, except with a third set of diagonals.

       Wilson's taxonomy is problematic. He describes his four types as a chronological development. Yet as previously noted Smith's first patent (for "Type 1") describes using extra sets of diagonals (like a Type 4) for longer spans and heavier loads. Size, not age, correlates with the number of sets of diagonals in the eight Smith bridges I have inspected (the five largest bridges have three sets of diagonals; the three next smaller bridges have two sets of diagonals, and the smallest has only one). Further, Wilson's taxonomy doesn't include "single-truss" Smiths at all. My observations indicate that Wilson's Types 1, 2, and 3 do indeed represent a chronological development and are useful categories. I suggest Wilson's Type 4 should not be used, however; rather the number of sets of diagonals (i.e. single, double, or triple truss) should be noted.

       Wilson's classifications have other difficulties. He lists two 1867 bridges in Michigan (White's and Bradfield) as Type 4, but then notes them as "Variant. Two center posts." Given the two center posts and their early date they were, perhaps, what I would call " Type 1 triple-truss" bridges. However other sources list them as Brown trusses (a rare type patented by Josiah Brown in 1857).[11] It would be instructive to examine the bridges themselves but both have now been lost.

       Wilson classifies the Engle Mill Road Bridge (Greene County, Ohio) as a "variant" of a Type 3; I found it not a variant but a classic Type 3 with later alterations. Wilson lists many other bridges as "variants"; perhaps many were instead altered. Indeed, I have seen only one Smith truss bridge without major alterations, the Type 3 double truss bridge at Cataract, Indiana.



5. Deciphering Smith Trusses at Cataract


       No original plans for Smith truss bridges survive, as far as I know. The American Bridge Company no longer has any of the Smith Bridge Company's records.[12] Thus one has only the bridges themselves to learn from.

       While there is much popular interest in covered bridges, few enthusiasts are technically trained. Antiquarians, if they note oddities like those mentioned above, are likely to regard them as romantic mysteries. Engineers tasked with fixing the bridges may have little time for or curiosity about the details of their origins. My opportunity to carefully examine the Cataract Bridge, however, has revealed details that may assist both groups.

Figure 7. Interior of Cataract Bridge, a Type 3 double truss.


5.1 Structural Aspects


       The diagonal timbers in a Smith truss that extend beyond the chords are designed to be in tension at least some of the time. In Smith trusses the tension diagonals and the chords are both notched and bolted together to form a joint. The diagonal must extend past the chord so that there is a shear surface beyond the notch to resist the pull on the timber. At Cataract (and other Type 3 trusses) this includes all of the diagonals that extend outwards (away from mid-span) as they rise from the lower chord. These are marked 'C' in Figure 4. Smith's patent calls them the "posts".  As previously noted, Smith apparently mechanized the process of cutting notches in truss timbers.

       Even if mechanized, cutting notches takes time, and it also weaken the timbers. Since notches are not needed to transfer compressive forces, Smith truss diagonals that are always in compression are simply butt-jointed in the angle between a chord and a tension diagonal and spiked in place. On a Type 3 truss this includes most (not all) of the diagonals that extend inwards (towards mid-span) as they rise from the lower chord. Smith's patent calls them the "braces". They are marked 'D' in Figure 4. At the ends of these timbers three members - a chord, a tension diagonal, and a compression diagonal - meet at a single point.

       The pair of inward-leaning diagonals nearest to mid-span may be under tension under some loading conditions. For this reason their ends, like the outward-leaning diagonals, are notched and pass through the chord.

       If two tension members met a chord at the same point using notched joints, the timbers would have to be notched so deeply that they would all be badly weakened. Instead on Smith trusses the geometry is adjusted so that adjacent tension diagonals intercept the chords a short distance apart. In this way they can use the same notch type as all the other tension diagonals. On Type 2 trusses this adjustment in the geometry occurs once on the lower chord at mid-span (see Figure 5). On Type 3 trusses like Cataract it occurs three times on the lower chord and once on the upper  (see Figures 6; the upper chord is also visible in Figure 7).


5.2  Cataract Bridge Timbers


       One can tell the bridge at Cataract was an industrial product just by carefully examining the timbers. Many early covered bridges are made of mixed species of local timber, commonly oak and poplar in Indiana. Not so at Cataract, where the timbers are all white pine, a lightweight, easily worked wood, but one not native to the region. So the timbers were brought from hundreds of miles away  (probably Michigan or Wisconsin). This would have been exceedingly difficult without railroad transport (waterborne transport does not extend to Cataract). Early covered bridge timbers are usually hand-hewn or the product of a nearby rural sawmill. Not so at Cataract. All surfaces of the Cataract timbers, both in the trusses and in the upper and lower lateral cross-bracing between them, are smooth. The timbers' dimensions are precise. Every truss diagonal, for example, is exactly 6 ¾ inches thick, planed down from rough lumber 7 inches thick. In short, the Cataract timbers were finished by an industrial planing mill capable of handling pieces over 30 feet long. In 1876 these existed only in cities, far from Cataract.



Figure 8. Cataract Bridge, upper chord at endpost.

Truss timbers are white pine planed to precise dimensions.



Painted marks, often faint and obscured by dirt and animal droppings, were found on top of the truss chords at Cataract. Carefully examined, they turned out to be numbers match-marking all adjacent pieces, not only at the splices in the three principal chord beams but the dozens of spacer and shear transfer blocks between them on each chord. These are evidence that the bridge was prefabricated, dismantled, and reassembled on site. I must note, however, that I have observed similar marks on timbers of covered bridges made locally (presumably fabricated on shore, dismantled, and reassembled on falsework over the water).


Figure 9. Cataract Bridge upper chord showing marks left by the builders.



5.3 Cataract Bridge's Dimensions


Given that Cataract Bridge's timbers had been industrially processed, it seemed likely that it had an underlying geometry that would lend itself to industrial-type assembly. Careful examination of the timbers and their measured dimensions revealed such a simple plan.

       There are two scored marks, one foot apart, on the top of the upper chord timbers at mid-span. Outwards from these marks, in both directions, are six more marks at exact 11' intervals, with a seventh mark 2' 4" beyond the sixth, at the outside of the truss' vertical end-post. On the bottom of the lower chord there are also two scored marks one foot apart at mid-span. Outwards from these, like on the upper chord, there are more marks at regular intervals in both directions, but these are 10' 11" apart, one inch less than on the upper chord. Like on the top chord, the outside of the end-post is at a distance of six intervals plus 2' 4". Finally, one foot inwards from the first of the 10' 11" interval marks, on both sides of mid-span, is another mark. The arrangement is shown in Figure 9.

Figure 9. Basic dimensions of Smith trusses at Cataract Bridge.


These, and the 14' 8" height on the end posts, are all the measurements needed to lay out all of the timbers for the trusses, including the notches. The resulting structure will be slightly cambered (arched) due to the slightly smaller intervals on the lower chord.

       The following manufacturing sequence, or a similar one, seems likely:

1.   Assemble the chords (more about this below). Measure and mark them and the end posts as shown in Figure 9.

2.   Place end posts with their outer edges on the outermost marks on the chords, square them, then score them and the chords along their lines of intersection. Cut notches to those lines (or parallel ones transferred to other faces of the timbers with the framing square).

3.   Assemble end posts and chords. Square the joints, which requires bending the chords to the desired arc. A stringline or a straight line on the factory floor could be used to ensure the arc curvature is even.

4.   Place tension diagonals on the chords, aligned with the marks. Score them and the chords along their line of intersection, and cut notches to these lines as was done for the posts. The ends of the diagonals, deliberately too long to begin with, could be trimmed off 6 inches beyond (and parallel to) the notches now or later.

5.   Reassemble chords, end posts, and tension diagonals. Put compression diagonal timbers in position on the assembly, score at lines of intersection, cut to length, and place in the now-complete truss.

6.   Drill bolt holes, paint match markings, and disassemble for shipment.


Note that the only measurements required in this sequence are in the first step. From then on pieces are marked in place, assuring precise notches and member lengths. This simple layout system would allow relatively unskilled carpenters to produce well-fitted timber trusses rapidly and with little chance for errors.

Layout of the chords is also simple. Each has three parallel beams, each beam made from several timbers spliced end-to-end. The beams are notched frequently, with oblong blocks fitted between them to maintain the spacing and transfer and equalize stress between beams.

It might seem logical to place chord splices exactly midway between panel points on the truss. This might also give the best appearance. However this would require chord timbers to be many different lengths, since top and bottom panel lengths differ and there is an "extra" foot between panels at mid-span. This would require extra labor in construction and introduce chances for errors in cutting and fitting.

Instead at Cataract the staggered splices in the three beams are evenly spaced 11' 2" apart, 2" more than the interval of top chord marks described above. The result is that no splice is exactly midway between panel points, but all are close, and chord layout is very simple. Each of the 60 chord timbers in the bridge (except those lopped off at the ends) is 33' 6" long. All 24 "fish plates" used on the bridge's lower chord (tension) splices, all 100 lower chord spacer blocks, and all 124 upper chord spacer blocks, are identically sized and regularly spaced.

The upper chords are exactly 13 splice intervals long (145' 2") and could have been made with exactly 13 timbers and no waste. The lower chord was probably built similarly, but trimmed to exactly 140'. Building upper chords longer than lower chords was normal practice for covered bridges; the overhanging roof built on the former helps protect the latter from weather.

Upper and lower lateral bracing is simpler yet. The bracing is made of crossed timbers, all the same size and with identical notches at their ends, that form two strings of overlapping "W"s like the truss diagonals. The crossed, lap-jointed ends attach to the chords with bolts and the cast iron brackets covered by Smith's 1869 patent.



6. Comparing Cataract with Other Smith Trusses


       I have inspected eight Smith truss bridges besides the one at Cataract. These are listed in Table 1. All but one are, like Cataract Bridge, Type 3 single span structures built between 1875 and 1878. The exception, an 1870 Type 2 bridge, is discussed in the next section.



Bridge Name




Width (ft)

No. of Panels

Truss type

Yr. Built

Old Red

Gibson County, IN


14' 2"


Type 3 triple



Gibson County, IN


14' 6"


Type 3 triple


George Miller

Brown County, OH


16' 4"


Type 3 triple


North Pole

Brown County, OH


16' 4"


Type 3 triple



Brown County, OH


18' 6"


Type 3 triple


W. Engel Mill Road

Greene County, OH


16' 8"


Type 3 double



Owen County, IN


13' 8"


Type 3 double


Stevenson Road

Greene County, OH


16' 4"


Type 3 double


Carillon Park

Dayton, OH




Type 2 single


Table 1. Smith truss bridges examined by the author.


Truss height in the other eight varies from 14'-2" to 15'-4" clear between upper and lower chords. Their truss panel length shows little variation (10'-8" to 11'-5" measured on the top chord) except at Stevenson Road, where it is only 8'-11".

Many characteristics are the same in all the Type 3 bridges, among them:

€      All truss timbers are of planed white pine.

€      Notched tension joints are similar in all the bridges.

€      Chord splice intervals are always 2" greater than upper chord panel width, and lower chord panel widths are 1" or 1 ½" less.

€      Adjacent tension diagonals are always spaced exactly one foot apart.

€      Fish plates and chord spacer blocks are identical in all.


In short, the basic layout and construction system used at Cataract is common to all the Smith Company trusses I have seen that were built during the mid and late 1870s.


7. Carillon Park Bridge


       The Smith Bridge Company built this short (55' span) bridge in Greene County Ohio in 1870. Some sources list it as a Warren truss, although Wood notes that a "wooden Warren truss is something of an oddity."[13]  Burr truss-style arches were added to it in 1948 when it was moved to Dayton, and it has had other major repairs made to it. My brief recent examination of this bridge leads me to make the following observations:

€      This is a "Smith Type 2 single truss" bridge, a type described in Smith's patents but otherwise unrecognized in the literature. Some might consider this as a type of Warren truss, but the diagonals at midspan do not meet the lower chord at a single point as they would on a true Warren truss. Instead they intersect the chord a short distance apart, as on both Type 2 and 3 Smith trusses (see Figures 5 and 6). Note also that classic Warren trusses do not have vertical end posts as this bridge does.

€      Early Smith trusses have similarities to multiple kingpost (and Burr arch) trusses that are absent in later ones. This leads me to speculate that Robert Smith originally developed his truss as an "improved" multiple kingpost truss, with diagonal rather than vertical tension posts. On the Carillon Park bridge, for example, the tension posts are notched to receive the compression braces, just as on most kingpost trusses. Later Smith bridges (Type 3) have simplified joints without these notches. Furthermore, the earliest (Type 1) Smith trusses have braced vertical posts at midspan identical to those on multiple kingpost trusses (see Figure 4). Such posts are omitted on later Smith trusses. This hypothesis also provides an explanation for Smith's otherwise rather odd choice of the word "posts" for his tension diagonals: they are evolutionary descendants of kingposts.

€      Diagonal timbers of the Carillon Park bridge appear to be planed, but some (not all) of the chord timbers show saw blade marks. Perhaps in 1870 Smith's company had not yet acquired machinery capable of planing timbers as big as the chords. Perhaps those with saw marks are not original (the chords have clearly had major repairs). I did not have time to determine which is the case; more careful examination probably would.



7. Conclusions


       Detailed examination of nine Smith truss bridges has revealed much about their geometry, the ways they were built, and the intent of their inventor. They illustrate that Smith trusses were produced in a manner very much at odds with the romantic conception of covered bridges as a remnant of an era of highly skilled timber framers creating hand-crafted structures from local materials.

       Robert Smith designed bridges that could be produced rapidly in a central urban location using imported materials processed by industrial machinery and then shipped long distances by railroad for re-erection at their final destinations.

       The similarity of the trusses from one Smith bridge to the next, their simple layout, the uniformity of the planed timbers, the repetitive use of identical components, and the minimizing of required measurements indicate Smith was attempting to speed production, reduce workers' skill requirements, and minimize the chances for fabrication errors.

       All this is characteristic of the late 19th Century industrial age we are more likely to associate with the iron bridges that Smith also designed and built. Indeed, one may describe the Smith Truss as the inventor's attempt to bring modern industrial design and production techniques to an old technology - the covered bridge - so that it could remain economically competitive with the new bridge types, based on iron and steel, which were appearing on the market. In this he succeeded for more than two decades.

       This work lends insight into Smith trusses specifically. In a larger sense it illustrates the importance of considering original construction methods and constraints when trying to understand and conserve historic structures. It is the author's opinion that this is often overlooked. In doing so we miss a great opportunity to gain knowledge about the historic structures in our care.


[1] Clark Waggoner, ed., entries on Robert W. Smith and Smith Bridge Company, History of Toledo and Lucas County, Munsell & Co., New York, 1888, pp. 786-787

[2] Ibid; also Eldon M. Neff, ³Highlights in the Life of Robert W. Smith,² Connecticut River Valley Covered Bridge Society XI, No. 4 (Spring 1963).

[3] "Cataract Bridge: Historical Background", by Mark Brown 2001, rev. Matt Reckard 2002, J.A. Barker Engineering for Indiana Department of Natural Resources Division of Engineering, unpublished.

[4] Gasparini, Dario, Prof., Dept. Civil Engineering., Case Western Reserve Univ., Cleveland, OH, maintains a list of covered bridges by type. Personal communication, 2002.

[5] Waggoner, ed., and Neff, op cit.

[6] Victor C. Darnell, A Directory of American Bridge-Building Companies: 1840-1900 (Washington, D.C.: Society for Industrial Archeology, 1984), pp. 55-56, 85-86.

[7] Smith's second patent describes "the intersections of (lateral) braces, where they are gained together." The drawings show (and Smith bridges have) half-lap notched joints at these intersections.

[8] Allen, Richard S. "Covered Bridges of the Middle West", Stephen Greene Press, Brattleboro, VT, 1970, pp 20-23

[9] Wood, Miriam "The Covered Bridges of Ohio", Old Trail Printing Co., Columbus Oh, 1993, pp 32-34

[10] Wilson, Raymond E. "The Story of the Smith Truss", in Covered Bridge Topics, April 1967, publ. by Nat'l Society for the Preservation of Covered Bridges

[11] Allen, op cit, pp. 85, 135; Hyde, Charles K., Historic Highway Bridges of Michigan, Wayne State U. Press, Detroit, 1993, p. 57.

[12] Personal communication with American Bridge Company, 2002.

[13] Wood, op cit, p. 37; Allen, op cit, p. 134.