In the autumn of 1900, a carpenter who specialized in timber framing stood on a platform high above the Saint John River in Carleton County, New Brunswick, and drove a wooden peg into a hole he had bored with an auger through two crossing planks of white pine and into the beam beneath them. The peg was a tree nail, about an inch in diameter and eight inches long, cut from hardwood so that it would swell when wet and bind the joint tight without a piece of iron. He had been driving tree nails since before the sun cleared the hills east of the river. The river, thirty feet below him, ran fast with its October current ahead of the ice season that came down from the upper watershed.

The planks he was fastening formed the lattice of a truss bridge that would one day carry the weight of horses and wagons, and later machines neither he nor the engineer who drew the plans could imagine. He could see the structure on either side of him stretching across the river toward the far bank. The truss openings stood open to the sky, with scaffold planks running the length of every span. Other men performed the same motion at intervals across the growing bridge.

He would do it again before the light failed. He had been doing it for weeks. He would keep doing it for weeks more. The tree nail he drove that morning is still in place 124 years later.

The Hartland Covered Bridge over the Saint John River at Hartland, New Brunswick, is the longest covered wooden bridge in the world. It has held that title since it was completed in 1901. It has carried two-way traffic across the Saint John River for more than a century. It carried horses and wagons and the first automobiles to reach Carleton County in the years after 1910.

It carried the agricultural traffic of the Saint John River valley through two world wars, the Great Depression, the flood seasons of 1923 and 1936, and every spring ice melt since. It is 1,282 feet long. It remains in service today. This is the story of how the farming communities on both sides of the Saint John River at Hartland built the longest wooden bridge in the world.

It is the story of the carpenters who drove tree nails into a truss lattice above a river that floods. It is the story of the men who sank the pier foundations into the bed of a river that the Saint John reshapes every April. It is the story of why a bridge built without a roof in 1901 was given one in 1921, and how that decision is the reason the bridge still stands today. It is the story of a county whose economy was built entirely on the skill of working in timber, when that economy was dying in 1900, and where the men driving tree nails were the last generation trained in a craft built by the lumber trade and the shipyards before the iron age ended them.

To understand why the people of Hartland needed a permanent crossing over the Saint John River at the end of the nineteenth century, one must understand what the river is and what it does to the communities on its banks for about four weeks every spring. The Saint John River drains a basin of roughly 22,000 square miles across northern Maine, western New Brunswick, and southern Quebec. The river runs about 418 miles from its headwaters near the international border to the Bay of Fundy at the city of Saint John, dropping in elevation through the agricultural heart of New Brunswick before it reaches the coastal plain and the reversing tidal falls that give the city at its mouth its name. At Hartland, about midway along its course, the river is a mature waterway about 400 feet wide between its banks at low water, flowing over a bed of ledge rock and alluvial gravel on a gradient that produces a steady current free of the rapids found upstream.

In summer, when water levels drop, it is navigable by boat and flat-bottomed canoe. It is a manageable river in August. In April it becomes a different one. Snow accumulates in the Saint John basin from November to March.

It piles up in the upland forests, on the frozen surfaces of the lakes that feed the tributaries, and on the ice of the river itself, which freezes over completely in most winters from December to late March. In April, when temperatures climb above freezing, Atlantic storms bring rain, and the snowmelt across the whole basin accelerates at once, all that water runs toward the river at the same time. The Saint John at Hartland rises in the first week of April in an average year by ten to fifteen feet above its low-water course and spreads across the lowlands on both banks. In a year of heavy winter snow followed by an early warm rain, it rises higher.

In 1923, the spring flood reached levels that closed county roads, submerged farmland that had been worked for four generations, and destroyed buildings erected on ground families believed beyond the river’s reach. In 1936 the flood returned at similar levels. It has come every April since. It comes every April now.

Before the flood comes the ice. The Saint John freezes in December most years, building from the edges toward the center until the river surface becomes a road in January and February. Farmers crossed on foot or by horse and wagon in the depths of winter. Children crossed to school on the opposite bank.

Mail and goods moved over the ice. The ice road was reliable in the coldest winter weeks and unreliable at the start and end of the season, dependent on temperatures and subject to sudden changes that could make the crossing safe on a Tuesday and dangerous by Thursday. The ice broke up in late March or early April. The breakup moved with the current from the headwaters, breaking first in the warmer reaches and sending ice downstream against the lower river that was still frozen, piling into plates and blocks that jammed and then gave way in sudden surges of water and ice.

The ice run of the Saint John in the late nineteenth century was no small event. The river carried ice blocks weighing tons. Some were large enough to capsize a flat-bottomed boat caught among them. All of them moved at the speed of the river running toward the bay.

Between the ice road of February and the navigable river of May lay a stretch of weeks when neither crossing was reliably safe. The river was too broken for the ice road and too full of moving ice and strong current for the ferry. Carleton County had a ferry crossing at Hartland at the end of the nineteenth century. It crossed when the river allowed.

When the river did not allow, the crossing waited. Farmers with goods to take to market on the other bank waited. Children who needed to reach school waited. The mail was irregular.

Produce piled up. The weekly newspapers ran late. The doctor on the other bank was out of reach. The waiting was not merely an inconvenience.

It was a recurring economic cost spread across every farm, household, and business in the valley that depended on easy access between the two banks for more than nine months of the year. The case for a permanent bridge at Hartland had been raised in county council chambers, in the New Brunswick legislature, and in the local papers longer than anyone then serving in those bodies in 1895 could precisely remember. A permanent bridge across the Saint John required money the county alone did not have. It required provincial support allocated according to the province’s priorities and an engineering solution that could withstand the ice run and spring flood without being destroyed in its first year.

The wooden pile bridges the county had built over the smaller rivers of inland New Brunswick suited small rivers. The Saint John at Hartland was not a small river. The provincial charter for the Hartland bridge was issued in the late 1890s. The contract was let.

Work began. The engineering design chosen for the Hartland crossing was the multi-span wooden lattice truss. The Town lattice was no new technology in 1900. Ithiel Town had patented it in 1820, and it had built a long record of reliable performance across the northeastern United States and the Canadian Maritimes over the eighty years that followed.

Town’s core insight was that the diagonal lattice of the truss could be built entirely from standard-dimension planks without the massive heavy timbers that earlier truss designs required. Where Howe or Long trusses concentrated load in a relatively few heavy wooden members, the Town lattice spread the load across a very large number of lighter planks. The lattice diagonals ran at roughly 45 degrees from the lower chord to the upper chord, crossing the opposite diagonal plank at every intersection, and were pinned at every intersection with a tree nail. The design was buildable by timber framers without specialized training in structural engineering.

It was repairable with the same materials and tools used to build it. It was economical in its use of wood. It was the natural choice for a county whose craftsmen were timber workers. The Hartland crossing required seven spans across the full width of the Saint John River between its flood banks.

The spans were not equal in length but were fitted to the pier sites that engineers fixed by sounding the riverbed. Each span averaged about 180 feet between pier faces. The overall structure measured 1,282 feet from portal to portal. A Town lattice spanning 180 feet carried its load through the accumulated stiffness of hundreds of diagonal plank intersections, each in tension or compression, each depending on its tree-nail joint to transfer that load to the adjacent member.

The mathematics of the lattice truss at this scale had been worked out by engineers throughout the region for decades. How the design would behave under the particular conditions of the Saint John River, the flood and ice loads, the live loads of horse-drawn freight wagons, and eventually motor vehicles, was a matter of engineering judgment for whoever drew the plans for the Hartland crossing. The engineer responsible for the Hartland bridge is documented in the provincial records of New Brunswick. Whether that name stands prominent in the heritage literature of the structure or falls into the category of engineering decisions traceable only through archival research is a question the Fredericton archives answer more fully than the tourist brochures at the bridge entrance.

The pier foundations were the engineering problem on which the bridge’s survival depended. The Saint John at Hartland at low water runs over a bed of hard rock and alluvial gravel that looks stable to a farmer crossing it in August. The river in flood is not the river of August. The spring flood scours gravel from around any obstacle placed in the channel and deposits it downstream at the point where the current slows below what is needed to carry it.

A pier foundation resting on gravel that a flood can move is a foundation that a flood will eventually undermine. A foundation built to survive the spring flood must rest on solid rock beneath the gravel or be driven deep enough into stable gravel that scour around its perimeter cannot reach the base of the footing before the flood recedes and the disturbed material settles again. At every pier site, engineers had to determine the depth and character of the bearing material. Where solid rock could be reached, the foundation could rest on bedrock.

Where the alluvial material was deep, the foundation required decisions about the depth and density of the gravel and the maximum depth of scour a major flood could produce in the material surrounding the pier. The cofferdam method used by the work crews at Hartland was the method river-bridge builders across North America had used for in-water pier work throughout the nineteenth century. Wooden planks were driven around the perimeter of the pier site, the water inside the enclosure was pumped or bailed down to the bearing surface, the excavation was dug to that surface, the footing was placed, and the pier was built up inside before the flood season returned. Timing was the governing constraint.

The Saint John fell to its lowest levels in late summer. The construction season for in-river pier work began in July or August and ended when autumn rains began raising the water in October. Frost came in November. The flood came in April.

The construction window for foundation work in the channel was no more than four months in a favorable year. The men who worked inside the cofferdams at Hartland in the autumns of 1899 and 1900 worked under conditions imposed by New Brunswick’s October climate on any work done inside or alongside the Saint John. The water seeping through the planks into the cofferdam was cold. October temperatures in New Brunswick fall to freezing before dawn.

The foundation work at the bottom of the excavated cofferdam was done by men who worked in standing water that the pumping crews drew down as fast as they could, knowing the river pressed against the timber from the other side. The gravel and sand they removed from the excavation was raised by bucket. The concrete or stone masonry of the footings was lowered and placed by hand. A cofferdam that failed against the river’s current, or whose planks shifted or opened a gap along the riverbed, flooded.

The men inside got out. Whether they got out in time is not recorded in the publicly available accounts. The pier work was completed. The piers still stand in the river today where they were placed in 1899 and 1900.

The floods of 1923 and 1936 tested them at levels the engineers of 1898 had never directly measured, and the foundations did not move. The men who built the timber superstructure of the Hartland bridge came from the working economy of Carleton County and the wider Saint John River valley in 1900. To understand who they were and what they knew, one must understand what the New Brunswick lumber economy had been for the eighty years before these men drove tree nails into a lattice above the river, and what it had become by the time they were doing it. New Brunswick in the first half of the nineteenth century was one of the largest timber-producing regions in the British Empire.

The British timber trade, redirected to British North America during the Napoleonic Wars when Baltic supplies were cut off by the naval situation in the North Sea, had built an economy in New Brunswick organized entirely around the production of square timber for the British market. The white pine, yellow birch, and spruce that grew in the highlands above the Saint John River valley were cut by lumberjacks in winter camps, driven down the tributaries on spring floods to the main river, squared to export dimensions by broadaxe men working on the riverbanks, loaded onto timber ships at Saint John, and sent to the lumber yards of Liverpool and Bristol. The trade was enormous. At its peak in the 1840s and 1850s, the port of Saint John was one of the largest timber-shipping ports in the world.

The men who worked the woods, the river drives, and the shipyards of the Saint John system developed over several generations a body of knowledge about working large timber that was specific to the forests, rivers, and crafts of the Canadian Maritimes. They knew how to fell a white pine without shattering its trunk on frozen ground out of season. They knew how to square timber precisely with a broadaxe on a chalked line. They knew how to join heavy timbers with mortise-and-tenon joints and fasten those joints with tree nails, because wooden shipbuilding required exactly that knowledge, and the shipbuilding industry of the Saint John River had for decades been one of the largest wooden shipbuilding industries in the world.

By 1900 that trade was over. The white pine of the upper watershed had been cut to the limits of commercial feasibility. The British timber market had shifted back toward Scandinavia and toward iron-and-steel construction methods that required no structural wood at all. The New Brunswick shipbuilding industry, which had launched hundreds of wooden sailing vessels from the yards of Saint John, Fredericton, and the smaller yards at Gagetown, Jemseg, and the upriver towns, had been killed by the transition to iron and steel hulls and steam engines that the wooden yards could not retool to build at competitive cost.

The last great wooden ships left the Saint John River yards in the 1880s. By 1900 the yards were silent or gone. The men whose fathers and grandfathers had built those ships were working at other trades, or in the remaining timber operations, or in the service economy of a county whose industrial base had contracted severely since Confederation. The carpenters who drove tree nails into the Hartland bridge in the autumn of 1900 were, in some cases, the direct descendants of the men who had used the broadaxe in the riverbank timber yards and the caulking iron in the shipyards.

They were applying to a permanent highway bridge the skills the New Brunswick lumber economy had spent eighty years developing, and which the iron age had made commercially obsolete in the decade before they built the bridge. The knowledge of how a tree nail behaved in a driven joint, how it swelled with moisture to bind tight, how a lattice plank transferred its diagonal load to the next member at each intersection, was in their hands through a craft tradition that had shaped working life in the Saint John River valley since before any of them was born. Whether those men came from the Irish communities that had settled the valley after the famine migrations of 1847, which brought tens of thousands of Irish immigrants to New Brunswick in the worst famine years and permanently changed the province’s working economy and social geography; or from the Acadian communities whose families had been in the river valley for generations before the deportation of 1755 and had returned in the following decades to rebuild their presence in Carleton County and the upriver parishes; or from the Protestant Loyalist families whose connection to the Saint John River dated to the arrival of the United Empire Loyalists from the American colonies after 1783; or from the Wolastoqiyik, whose relationship with the Saint John River predated every European arrival by thousands of years, is a question the building accounts of the Hartland bridge answer only to the extent that the records of 1900 preserved it. The name of the superintendent appears in the engineering correspondence that refers to him.

The timber framers and cofferdam workers below the level of superintendent appear, if they appear at all, in the aggregate form preserved by the wage books of 1900. A number of men, a number of days, those rates, a total cost. The men whose daily labor assembled the structure exist in those books as an amount of money. They do not regularly exist as named individuals.

The physical conditions of the Hartland bridge site in the autumn of 1900 were those of a New Brunswick river crossing in the last building season before the river froze. The Saint John at Hartland in October is very cold. The water temperature falls toward four degrees Celsius in October and toward freezing in November. A man falling into the river in October from the working platforms above the truss faced water that would disable him within minutes.

This was not a theoretical consideration for the men working on the scaffold planks extending from the growing truss above the water. The scaffolds were planks nailed across transverse supports extending from the truss structure. The planks were wet from the morning fog that lay over the Saint John valley until mid-morning in October. They were slick with early-morning frost and dried to become less slippery by midday.

They became slippery again if the afternoon cooling was sharp. A man who slipped on the scaffold at mid-span of the 180-foot truss fell thirty or forty feet to the river. Whether the river below was deep enough at that point to receive him without putting him on the gravel bottom changed with the September water level, which changed with the summer rainfall, which the construction schedule did not take into account. The tree-nail work followed a rhythm.

A lattice plank was placed to cross the opposite diagonal. The carpenter bored through both planks and into the timber beneath with a hand auger, following the angle of the pattern set by the geometry of the clamp. The tree nail was cut to length, its tip slightly tapered to start it into the hole. The carpenter drove it with a mallet, and with each stroke it sank until it was flush with the surface of the plank or slightly proud of it.

Then he moved to the next intersection. On a Town lattice span of 180 feet, the tree-nail intersections numbered in the hundreds. The tree nails across the seven spans numbered in the thousands. The work was not skilled in the sense that it required years of training.

It was skilled in the sense that it required judgment to know when a nail seated properly and when it did not, and the discipline to bore holes accurately and drive nails fully rather than rushing the work. The timber for the Hartland bridge came from the forests of Carleton County in the upper watershed. In 1900 New Brunswick still had commercial quantities of softwood in the uplands away from the river valleys that had been cut over earlier. White pine suitable for the main members and spruce suitable for the lattice planks had been cut the previous winter in Crown timber limits north of Hartland, hauled to the river on winter sleigh roads, driven down on the spring freshet to the sorting boom at the Hartland landing.

The sawmill cut the logs into the dimensions the bridge required. The lattice planks were cut to consistent width and thickness. The main timbers were cut to the dimensions the engineer specified. The quality of the raw material determined the quality of the structure.

Timber with large knots in the wrong places, or with checks indicating internal stress from improper drying, was returned to the mill or re-graded for uses that carried no structural load. The superintendent at the timber yard knew what the lattice planks carrying diagonal tension had to be free of to remain sound for twenty or thirty years above a flooding river. Timber that did not meet that standard was not used in the truss. The Hartland bridge opened to traffic in 1901.

The horses and wagons that crossed it for the first time in the spring or summer of that year crossed a structure open to the sky above the deck. There was no roof. There were no siding walls. The truss lattice rose on both sides from the lower chord to the upper chord, and the upper chord was exposed to rain, sun, and frost.

This was not unusual for the time and place. Covered bridges required extra materials, extra labor, extra cost, and added dead load on the truss. The decision whether to cover a wooden bridge was one that a county or state made by weighing the cost of the cover against the extended service life it provided. The argument for covering was evident to anyone who watched an uncovered wooden bridge deteriorate.

A bridge exposed to the weather directly would last a decade or fifteen years in a climate like New Brunswick’s before its structural wood needed replacement. A covered bridge lasted decades longer. The cost of the cover at construction was a fraction of the cost of rebuilding after ten years. In 1901, the county and the province spent what they spent to get a permanent crossing at Hartland.

The covering was deferred. The uncovered bridge deteriorated at the rate that uncovered wooden bridges deteriorate in the Saint John River valley. Rain wetted the main timbers and the lattice planks. Summer sun dried them.

The freeze-thaw cycles of a New Brunswick winter pushed moisture into and out of the wood, into and out of the tree-nail holes and around their pins, every season. Timber exposed to wetting and drying, freezing and thawing, season after season, cracks on its surface, allowing water to enter the checks and work toward the heart, which begins to soften from internal moisture. The tree-nail joints, properly driven in 1900, began to loosen in their holes as the surrounding wood moved. The diagonal planks, straight and sound in 1901, began to show surface checks and end-grain wear, meaning the wood was losing the internal integrity on which its structural role depended.

By 1918 or 1919, the provincial authorities responsible for the Hartland bridge faced a decision that had been foreseeable since the day the cover was deferred. The bridge needed substantial intervention. The question was what form that intervention should take. The argument for covering came from men in Hartland and Carleton County who understood, from watching the timber structures of the river valley, what a roof and siding did to extend the life of wood.

The covered bridges still standing in the county, some of them dating to the 1870s, demonstrated concretely what the covering argument claimed in financial terms. The uncovered bridge built in 1901, deteriorating for nearly twenty years, needed either a systematic rebuilding of its damaged parts or the protection that would keep the remaining sound wood from following the damaged wood toward replacement. The province approved the covering in 1921. The work was done in 1921 and 1922.

A roof was built over the full 1,282-foot length of the deck. Wooden siding was hung on both sides of the truss with openings cut through the boards to let in light and airflow. The men who came to Hartland in 1921 to add the cover were not the same men who had built the bridge in 1900. Twenty-one years had passed.

The builders who had driven tree nails in the October cold above the river had aged into their forties and fifties, moved on, or died in the intervening years. The workforce of 1921 was a different generation working on a bridge they had not built. The tree nails they added in 1921 sit beside the tree nails of 1900 in a structure that does not visually distinguish between what was built in the original construction and what was added in the rehabilitation. The bridge carries the work of both generations without marking which belongs to which period.

What the covered Hartland bridge has carried for over a century since its covering was completed in 1922 is the full history of valley traffic, the permanent crossing that has tied the valley together through every April flood and every winter ice season since. The spring flood of 1923 was the first major test of the covered bridge at a scale approaching or exceeding the design assumptions of the pier foundations. The Saint John rose in April 1923 to levels that closed county roads at both bridge approaches, submerged the low-lying farm fields on the west bank for several miles, and put water over installations along the river that had not been flooded in living memory. The floodwater rose against the pier masonry and pressed on the cutwaters and moved behind the piers at velocities that scoured riverbed gravel around the footing bases.

The foundations set by the cofferdam crews in the riverbed in 1899 and 1900, in concrete on bedrock or in compacted gravel below the scour depth, were tested by hydraulic forces beyond the ordinary range, and the foundations did not move. The piers showed no uneven settlement. The truss above did not shift. The bridge was in service when the water receded.

The 1936 flood came at similar levels. The bridge held. It has held through every spring flood since. The automobiles that reached Carleton County after 1910 were lighter at first than the heaviest horse-drawn freight wagons the bridge was designed to carry.

As motor vehicle numbers in the county grew through the 1920s and 1930s, the vehicles got heavier. Farm trucks moving grain and livestock across the river imposed deck loads the engineer who drew the lattice in the late 1890s had not specifically designed for, because motor trucks did not yet exist in commercially meaningful form when the design was made. The province monitored the bridge’s load capacity at intervals and imposed vehicle weight limits according to what the timber truss structure could bear. The bridge carried what the limits allowed.

The limits changed as the timber aged and as reinforcement work increased or restricted the estimated capacity. The covered two-lane bridge carrying two-way vehicular traffic across 1,282 feet of a major river is not what modern highway engineers design when they design a river crossing. The approaches at either end of the Hartland bridge today are managed by traffic signals regulating flow across the single-lane sections of the crossing. The signals are electronic.

The bridge is wooden. In the year the bridge was built, no one in Hartland or the provincial highway department could have specified the signal system because it had not yet been invented. The bridge was built for the crossing requirements of 1901. It has been adapted to the crossing requirements of every year since.

The adaptations have accumulated around the structure without changing the structure itself. The Hartland Covered Bridge was designated a National Historic Site of Canada in 1980, seventy-nine years after the carpenters finished driving tree nails above the October river. The designation records the structure as the longest covered wooden bridge in the world and as a heritage landmark of New Brunswick and the Saint John River valley. It records the construction date.

It records the contracting company. The designation has produced a tourist presence in Hartland that draws visitors to walk the bridge, photograph it from the riverbank, and read the interpretive panels at the portal approaches. If you drive to Hartland and stop near the east entrance on the Hartland side of the river and walk inside the bridge, you enter a covered timber structure 1,282 feet long. At walking speed it takes several minutes to cross.

The deck beneath your feet has been replaced in maintenance cycles over the decades, but it remains entirely wooden. The same material the original carpenters used lies on the same structure. The lattice rises on both sides from the lower chord to the upper chord under the roofline. The diagonal planks of the lattice cross at their fixed angle, pinned at every intersection.

In the longer straight sections of the bridge, the lines of the lattice converging toward the far portal create a visual perspective shared by the interiors of covered bridges of the region. A long wooden tunnel tapering toward the light. Light enters through the side openings in long horizontal bands across the deck. The wooden trunnels at the lattice intersections are visible where the surface of the wood has checked slightly around the head of the pin.

Some of those tree nails were driven in 1900. Some were driven in 1921. Some were replaced in later maintenance work. The bridge does not identify which parts are original and which are not.

At the midpoint of the bridge you are 641 feet from either bank above the main navigation channel of the Saint John River. The pier beneath your feet was placed in the riverbed by the cofferdam crews in the autumn of 1899 or 1900, men working in cold water inside a timber enclosure while the river pressed against the other side. The lattice truss above you was framed by timber workers who drove thousands of tree nails in the weeks before the October frost ended the building season. The cover you stand inside was built by a different crew in 1921 and 1922, twenty years after the bridge itself, in response to a deterioration process that anyone who understood timber structures in the New Brunswick climate could have predicted on the day the original builders chose to do without it.

The national historic site designation does not name the men who drove those tree nails. The plaque at the bridge entrance gives the name of the structure, its date, and its classification. It does not give the names of the carpenters who worked on the scaffolds above the October river. The provincial records of the Hartland bridge construction are held in the Provincial Archives of New Brunswick in Fredericton.

Whether those records identify the individual workers who framed the truss spans and worked in the cofferdams in the autumn of 1900, or whether they preserve only the contractor’s totals and the aggregated labor cost entries that 1900 accounting kept, is a question the archives answer according to what the accounting of that era was designed to record. The contracting company that built the Hartland bridge appears in the financial records. The engineer responsible appears in the correspondence. The foremen appear where the construction letters mention them.

The individual carpenters and cofferdam workers below the level of foreman appear, if they appear at all, as a daily wage total. The men who drove the tree nails in October 1900 exist in those records to the extent that provincial bridge-contract accounting of that era kept the names of individual workers rather than aggregate costs. They were from the communities along the Saint John River that the lumber trade had built and that the trade’s decline had left with skilled hands and diminishing opportunity. They were in some cases from the Irish parishes of Carleton County, whose founding families had arrived in the 1840s with nothing and built a generation of familiarity with hard work in the river valleys of New Brunswick.

They were in some cases Acadian families for whom the Saint John River was a landscape centuries deep. They were working on a bridge in a valley that was their valley, a bridge the valley needed as a permanent crossing, because the ferry was not enough, the ice road was not reliable, and the spring flood isolated the two banks from each other for weeks every April. Whether any of those men came back after the bridge was finished, crossed it on horseback in the summer of 1901, and understood that the structure they had built would still be carrying traffic more than a century later is a question the records cannot answer. The bridge held.

The river ran beneath it. The tree nails were in the joints. The province put a roof on it in 1921. The floods of 1923 and 1936 tested it, and every spring since.

The ice runs have tested the foundations every April. The motor vehicles that replaced the horses have tested the deck capacity over the decades. None of those tests have closed the bridge permanently. The bridge is still open.

Vehicles cross it today. No covered wooden bridge anywhere in the world is longer. None has been built longer since 1901. The 1,282-foot lattice truss over the Saint John River at Hartland, New Brunswick, holds the record that the timber carpenters who drove the tree nails in October 1900 deposited in it.

The men who built it exist in the records at the Provincial Archives in Fredericton. The bridge stands in the river. The river rises every spring at levels that test the foundations.

The foundations hold.