On May 20, 2005, the Secretary of the Interior stood on a platform beside a concrete arch bridge in Williamson County, Tennessee, to open the Natchez Trace Parkway after its completion. The final section of roadway had been paved. The route the government had claimed ownership of seventy-one years earlier was now a continuous ribbon of asphalt stretching from Natchez, Mississippi, to Nashville, Tennessee, a distance of 444 miles. The man opening the road had not been born when Congress authorized it.

That is the first thing to understand about the Natchez Trace Parkway. It was one of the longest construction projects in the history of American highways, and the reason it took seventy years had almost nothing to do with engineering. The engineers were ready. The terrain was mostly favorable.

What the road ran into, again and again, state after state, was the oldest problem in American public works: the question of who owned the land it crossed, and whether they were willing to sell it. But before any of that—before the congressional authorization, the survey crews, the Civilian Conservation Corps camps, the bridge over Birdsong Hollow—this road was something else entirely. It was a footpath, and people walked it because there was no other way to get home. In the winter of 1806, a flatboat had been sailing for weeks.

It had come down from a tributary of the Ohio River, loaded with tobacco packed in large barrels, dried venison, barrels of salted pork, and whiskey. The current of the Ohio carried it into the Mississippi, and the Mississippi carried it south to Natchez, which in 1806 was a market town of a few thousand people, perched above a river landing where almost anything that could be floated from the interior could be sold. The sale was now over. The money was in the owner’s coat, or divided between his boots and his hatband—which was where a sensible man kept his money on this road.

The road north was the Natchez Trace. The alternative to walking was paying for passage on a keelboat working its way up the Mississippi against the current, which was possible but expensive and very slow. A flatboat that could cover 1,200 miles with the current in six weeks might take four months to return. Few men waited for that.

They sold the flatboats for timber and started walking. They called themselves Kaintucks—the men who made this trip—because most of them were from Kentucky and the regions north of the Ohio River. The name did not require a person to be born in Kentucky, only to have a certain willingness to sail a boat down a continental river system and then walk back. At the peak years in the first decade of the 1800s, roughly 10,000 people a year made the trip south by boat and north on foot.

They moved in groups for safety, because the trace had a reputation. The trace itself was old. The Choctaw and Chickasaw peoples had cut the first paths through these forests over the centuries, following the high ground between the headwaters of the Pearl, the Tennessee, the Cumberland, and the Duck rivers. Walking the high ground meant dry feet, clear sightlines, and fewer stream crossings—which is almost a definition of a good path on a foot trail through a roadless forest.

The Choctaw called one stretch of this trail Nashoba Chitto, the Great Wolf Road. By the time European traders began following these paths in the eighteenth century, they had been packed and eroded into a clearly visible trace. Thomas Jefferson ordered the trace improved for use as a post road in 1801. The Army was sent out to widen its worst sections, clear brush, and build simple inns at intervals of roughly a day’s travel—twenty to twenty-five miles apart.

These inns were called stands, and at the height of the system there were about fifty of them. Accommodations ranged from simple log buildings with dirt floors to more substantial structures, and travelers’ accounts ranged from grudging praise to outright contempt. The food was often bad, and the beds were a blanket laid over a board, but the horses were fed and watered, and after twenty-five miles on foot, a board counted for something. The federal government also negotiated rights-of-way with the Choctaw and Chickasaw for passage across native lands.

This effort was not unified, and it was not always fair, but it acknowledged that the two peoples owned the land the trace crossed. The outlaws who worked the trace acknowledged no such thing. Samuel Mason ran a criminal operation along the southern stretches of the trace for several years, robbing travelers and, according to some accounts, passing information to river pirates working the Natchez landing. He was eventually killed by two of his own associates in 1803 in a dispute over a reward.

One of those companions, traveling under the name John Setton, turned out to be Wiley Harpe, one of two brothers whose crimes had already driven them into hiding in Kentucky and Tennessee. Setton Harpe was recognized, arrested, hanged, and his head placed on a post beside the trace as a warning. The Kaintucks coming north passed it in the following weeks and, presumably, took the point. What remained for them to worry about after leaving Natchez and reaching the main trace was the weather, the condition of the road, and their provisions.

The trace was mud in wet weather and dust in dry. In both seasons, the Mississippi sun was scorching in summer and the December north wind was bitter. A healthy man in good boots could cover twenty to twenty-five miles a day. A sick man, or one with worn-out shoes, might manage ten.

The trace was 500 miles long. Andrew Jackson walked it twice. The first time, in 1813, he was leading troops home after an unsuccessful campaign against the Creek. The second time, in early 1815, after the Battle of New Orleans, he returned north as a national hero with the victorious army.

The victory at New Orleans was an overwhelming American triumph, and the news traveled quickly. The towns along the trace prepared celebrations. Jackson accepted the acclaim and kept marching. He had walked this road before, and he knew better than to stop in the mud.

But the most consequential journey on the Natchez Trace did not end in Nashville. It ended at an inn called Grinder’s Stand, in what is now Lewis County in central Tennessee. The traveler was Meriwether Lewis. October 1809.

Lewis was thirty-five years old and in poor health, by the accounts of those who saw him that year. His crossing of the continent with William Clark four years earlier had made him famous. Since then he had served as governor of the Upper Louisiana Territory, a job that had not gone well. His expense accounts had been questioned in Washington, his vouchers rejected, and his judgment doubted in official correspondence.

He was heading east to settle the dispute in person, carrying papers and traveling with a small party. He stopped at Grinder’s Stand on the night of October 10. Mrs. Grinder, whose husband was away, heard a gunshot in the night.

In the morning, Lewis was dying. He died before noon on October 11, seventy-two miles short of Nashville. Whether he killed himself or was killed is a controversy that has lasted more than 200 years. Thomas Jefferson believed he committed suicide.

Some of his contemporaries agreed. But the circumstances at Grinder’s Stand have always carried enough strangeness to keep the question open: the missing papers, the behavior of the people at the stand, the account gathered by naturalist Alexander Wilson when he visited the site the following year, which differed from what Mrs. Grinder had originally said. No one has ever settled it.

What is not in dispute is that one of the most famous Americans of his generation died alone, on the Natchez Trace, in an October darkness. A stone monument marks the place today, erected by the state of Tennessee in 1848. The parkway passes within sight of it. The trace on which Lewis died was already emptying of people.

In 1811, the first steamboat sailed the Ohio and Mississippi rivers under its own power. Within a decade, steam river travel had so radically changed the economics of the western trade that the Kaintuck trade was already fading. If you could pay for passage upstream and make it back to Louisville or Cincinnati in two weeks instead of walking for five, you took the steamer. The trace closed in on itself.

The government stopped maintaining it. By 1830, the trace was abandoned as a commercial road. The forest returned. Trees grew thick in the cleared sections.

The stands fell to ruin. The trace itself, packed down by decades of feet and hooves, had sunk below the surrounding ground level in some sections, becoming a trench in the forest floor, sometimes a foot and a half deep, sometimes more. Rainwater pooled in the trench and sat there. Roots cracked the compacted soil.

Weeds and brush crept in from the sides. Yet the trace did not disappear. Compacted, eroded soil does not recover easily. The depression that formed the hollowed sections was permanent, and in many sections, the mark of the original trail was still visible in the ground a century after the last Kaintuck stopped using it.

Locals knew where it was. Children of families who had run stands along the trace knew the stories. The trace stayed present in memory and local history long after it vanished from any practical use. In some places it was visible from above.

Pilots who flew over Mississippi and Tennessee in the early days of aviation reported that they could make out long sections of the old trace from the air, by the different appearance of the vegetation over that compacted ground. An abandoned road of eighty years still left a visible mark from a thousand feet up. In 1905, the Daughters of the American Revolution began placing stone markers along the old trace. The organization had been engaged in commemorating early American history for years, and the Natchez Trace was of paramount significance: a road that had carried the mail, soldiers, and commerce of the young republic across three southern states for a crucial generation.

The women who organized the marking project walked portions of it themselves through the Mississippi and Tennessee woods, searching for the depressions in the ground, measuring distances between known points, and placing tablets where they could be certain they had found the original trace. The marking project was slow and incomplete, but it accomplished two things. It kept the trace on the cultural and historical map of the region, and it partially recorded the locations of the remaining original trace. Thirty years later, when a group of Mississippi businessmen decided to pursue a national parkway along the old trace, they built on the DAR’s work.

The Natchez Trace Association was founded in 1934, its membership made up mostly of Mississippi business and civic leaders, and its goal was to secure federal authorization and funding for a scenic road following the old trace from Natchez to Nashville. The timing was not accidental. The New Deal was underway. Franklin Roosevelt’s administration was looking for massive public works projects that would provide wide-scale employment in rural areas.

Building a highway stretching nearly 500 miles through the forests of Mississippi, Alabama, and Tennessee—requiring surveying, clearing, grading, bridge building, and paving—was exactly the kind of labor-intensive work the relief programs were designed to fund. Congress authorized the Natchez Trace Parkway in June 1934, placing it under the National Park Service. It was to run from Natchez to Nashville, approximately 450 miles, following the general route of the old trace. Commercial trucks were banned.

Advertising was prohibited. It would connect to the surrounding road network only through tunnels and interchanges, with no surface crossings or traffic lights. The speed limit was 50 miles per hour along its length. There was to be a wide right-of-way, enough to insulate it from commercial sprawl on either side.

The vision was easy to draft. Securing the land was the next challenge—and it turned out to be the problem that took seven decades to fully solve. Building that right-of-way required the federal government to acquire land from thousands of individual owners in three states, many of them families who had farmed the same land for generations and were not eager to negotiate with a federal agency. The acquisition process began in 1934 and, in some sections, did not conclude until after 2000.

The Civilian Conservation Corps began work on the first sections in 1938. CCC camps were established at regular intervals along the proposed route across Mississippi, and the enrollees did the road-building work that the late Depression years required: clearing timber, constructing culverts, grading the roadbed, laying the gravel base. The machines they used were standard road-building equipment—crawler tractors and graders—and in the low loess hills of southwestern Mississippi, the work ran into ground that most of the men had never encountered before. Loess is wind-deposited silt, extremely fine-grained, with unusual structural properties.

When dry, it can hold an almost vertical cut bank thirty or forty feet high without collapsing. That is why the bluffs above the Mississippi River at Natchez stand nearly straight in places. When wet, it becomes slippery and flowable. The CCC crews working in the loess hills north of Natchez cut deep into those banks, and the walls held, which made the grading faster than it would have been in ordinary clay ground.

But the drainage had to be right. If water got behind a loess cut and saturated it, the wall would come down. The first paved sections in Mississippi opened in the early 1940s, extending north from Natchez through the hill towns of the state’s southwestern corner. Then the war came.

The CCC was disbanded in 1942. The young men who had been grading roads joined the Army. The Park Service kept a reduced staff working on the road through the war years, but the pace of construction fell off sharply. What had been a major relief-employment project became a routine maintenance operation.

After the war, construction resumed under regular congressional appropriations, which were smaller and less reliable than the Depression-era emergency funding. The Mississippi sections progressed year by year through the late 1940s and 1950s. The Black Belt of central Mississippi—where cotton had been the principal crop for a century, and where the old plantation economy had left its mark on the landscape in the form of large holdings and complex chains of title—was crossed with particular care. Absentee owners and heirs scattered across multiple states had to be accounted for in appraisals.

Land titles that had not been examined in decades had to be researched. One detail from those Mississippi sections is worth pausing over, because it reveals more about what the Park Service was trying to build here than anything else in the construction records. At a number of sites along the proposed route, surveyors walking the corridor found sections of the original sunken trace still intact. The trail was still there in the ground—narrower than in the Kaintuck years, overgrown at the edges, but clearly visible.

The ground was still lower than the surrounding forest floor, still compacted, still running north. The Park Service did not pave over those sections. Instead, it built the modern parkway to one side, preserved the sunken traces beside it, and designated small segments of the original trace with pull-offs where visitors could stop and walk down into the old trail. Those preserved sections are still there.

At more than a dozen points along the Mississippi sections, a gravel pull-off leads to a short walking trail that descends from the modern road down into the original trace. The floor is polished, compacted clay, a foot or more below the surrounding forest level. The passage is narrow, barely shoulder-width, and shaded by trees. It smells of leaf mold and old earth.

You walk in it for five minutes, maybe 200 yards, then return to the parking lot. Standing in one of those sunken traces, you are standing on the same ground the Kaintucks used. The same clay that turned to mud in the winter of 1806. The modern parkway is fifty feet away through the trees—a divided highway at 50 miles per hour—and the Park Service built that, and then preserved 200-yard stretches of the original footpath beside it.

That tells you everything about the kind of road this was meant to be. It was not a utility road. It was a road built around a place, designed to connect people to the history lying beneath it, fast for most of it, but slow and on foot where the history was close enough to touch. The Mississippi sections were largely complete by the early 1960s.

The road ran from Natchez through Port Gibson, past the ruins of Rocky Springs at mile marker 54—where an entire town had been abandoned in the late nineteenth century, leaving only a standing church in the woods—through the central hill country to Tupelo at mile marker 266, where the Park Service built its administrative headquarters, then to the Alabama state line. 310 miles paved and open. The Alabama section of the parkway was the shortest and least complicated. The road ran only 33 miles through the far northeastern corner of the state, crossing Colbert and Tishomingo counties, on its way between Mississippi and Tennessee.

The terrain resembled the Mississippi hill country. The land was rural and thinly populated, and there were no structural obstacles to acquisition. The Alabama section was completed during the 1960s. Then came Tennessee.

The Tennessee section of the parkway ran roughly 112 miles from the state line south to the junction with Interstate 40 on the western edge of Nashville. The southern end passed through Wayne and Lewis counties, among the most rural counties in the state. Land acquisition there was slow because the parcels were often large, the chains of title old, and the owners scattered. But there was no structural barrier to eventually completing it.

The road crept north through the 1960s and 1970s. At mile marker 385. 9, the parkway passed beside the stone obelisk in the clearing. The obelisk, made of Tennessee marble and erected in 1848, had weathered until its inscription was unreadable from any distance.

Grinder’s Stand, rebuilt in logs, stood nearby—two rooms joined by an open breezeway, the kind of building a traveler of 1809 would have recognized. Lewis was buried beneath the stone. The Park Service maintains the clearing, trimming the brush each season, so the obelisk is always visible from the road. Stand beside it in October, when the leaves are turning, and you are standing at the site of one of the strangest deaths in American history, within the bounds of a national parkway, with the sound of cars from the road 200 feet away.

North of the Lewis monument, the road entered more densely settled country: Maury County, then Williamson County. By the 1970s, Williamson County was the fastest-growing suburban county in Tennessee. Nashville was expanding south, and the corridor the parkway needed ran directly through the path of that expansion. Here, precisely, the engineering challenges—the main obstacle elsewhere—became the easy part.

Acquiring land in a growing suburb is different from acquiring farmland in Wayne County. Land values were higher, the owners had development plans, some parcels were already subdivided, and others were in the process of obtaining residential building permits. The federal government had to appraise each parcel, make an offer, negotiate, and when negotiations failed, file a condemnation suit in federal court, leaving the price to a jury. That process, repeated across dozens of parcels in a five-mile corridor through Williamson County, took years.

Then it took more years. Environmental review under the National Environmental Policy Act, signed in 1969, produced environmental impact statements, which led to public comment periods, which produced lawsuits, which produced revised environmental impact statements. The route of the parkway through the county underwent several revisions as acquisition difficulties and environmental findings interacted. Each adjustment required a new round of assessments and negotiations on the modified corridor.

One line in the Park Service’s historical summary of the Natchez Trace Parkway’s construction is worth quoting exactly as written. It states that the acquisition of the final parcels in Williamson County took longer—through years of negotiation and litigation—than the construction of the original sections built by the Civilian Conservation Corps in Mississippi. The men who graded the first sections of the road in 1938 finished their sections faster than the federal government’s lawyers finished arguing over the last few miles. By 1990, the Tennessee sections were built north to about mile marker 407, where the parkway crosses the Duck River.

The bridge carrying the parkway over the Duck River is a steel arch bridge, built in the parkway’s characteristic engineering style, which sought to be as unobtrusive as possible in the landscape it crossed. The river is clear over gravel at that point, lined with sycamores. Then, Birdsong Hollow. At approximately mile marker 438, about six miles from the northern terminus, the parkway reached the edge of a dry valley in the limestone country of Tennessee.

Birdsong Hollow was 155 feet deep and 1,572 feet wide at the point where the road needed to cross it. Its sides were steep, while the bottom of the hollow was flat and dry, with no permanent water. A road designed to parkway standards could not descend into the hollow and climb out again. The slopes were too steep and the turns at the base too tight for a divided highway.

It had to be crossed at the top. No ordinary bridge would do. The floor of the hollow was fragile by highway standards. Placing foundation piers in the hollow would have required construction equipment inside the valley, significant disturbance of the ground, and a structure visible from the hollow floor that would scar the very landscape the Park Service was trying to preserve.

The Park Service wanted a solution that would span the hollow without touching it. Figg and Muller Engineers were contracted by the Park Service. The solution they developed was a double-arch bridge—two parallel concrete arches, one for each direction of travel—springing from bedrock on both sides of the hollow and meeting in the middle, with no intermediate piers. The arch ribs carry the bridge deck’s load in compression, the same principle Roman stone bridges used two thousand years ago, here executed in reinforced concrete on a scale never before attempted in the United States.

To keep the arch ribs from exceeding their weight limit in the construction method, they were built hollow. Concrete shells were cast instead of solid sections, reducing the dead load of the arch itself without affecting its compressive strength. The construction method was balanced cantilever, working outward from both sides of the hollow simultaneously, casting identical sections of concrete on each arm of the arch. The weight of each new section is carried by the tension in the growing arch below it.

The arms grew toward each other over weeks and months, reaching farther and farther out over the hollow floor until the two sides met and settled in the middle. When the construction crews were at the peak of their reach, the open ends of the cantilever arms extended more than 200 feet out over the hollow, 155 feet above the valley floor, with the opposite rim of the hollow slowly approaching. The workers placing forms and pouring concrete worked at the end of a growing concrete arch with nothing but air beneath them. The two arches met in 1994.

The bridge opened to traffic in 1996, when enough of the road on both sides was complete to connect it to the rest of the highway system. The arch spans 582 feet in its longest single span. From the floor of the hollow, the two parallel white concrete arches rise into the Tennessee sky, looking so deliberate, so ordered, that it is hard to regard them as an engineering solution to a practical problem. But that is exactly what they are.

The arch exists because the hollow floor could not hold piers, and the concrete is hollow because the span could not bear the extra weight of solid ribs. Every visible feature of that bridge is a direct result of a circumstance. That is how good bridges are usually built. Completing the bridge did not complete the parkway.

The sections on either side of the hollow still had to be finished. Land acquisition continued into the late 1990s and the early 2000s. Some parcels required condemnation proceedings that went to a verdict in federal court. Others were settled in late-stage negotiations after years of correspondence and bargaining, on terms that eventually became acceptable to both sides.

By 2004, the right-of-way was fully secured, and contractors began grading and paving the final section. On May 20, 2005, the dedication ceremony was held near the bridge. The road was connected. The project was complete, seventy-one years after authorization.

It is worth pausing to consider what those seventy-one years mean in practical terms. The CCC enrollees who graded the first sections through the loess hills in 1938 were in their late teens and early twenties. By 2005, if they were still alive, they were in their mid-eighties or older. The Park Service engineers who laid out the routes through central Mississippi in the 1950s had retired by the time the bridge opened.

No one who worked on the parkway in its first decade lived to see the dedication. The road was built the way the original trace was built, piece by piece, with each generation inheriting the decisions of those who came before and making decisions that later generations would have to live with. If you drive it from south to north, you enter at the visitor center atop the Mississippi River bluff in Natchez, where the river appears below, brown and wide and slow. Across the river is Louisiana, low and flat.

Natchez itself climbs the bluff above the landing. Natchez Under-the-Hill, the rowdy riverfront quarter where Kaintucks spent some of their Natchez money before heading north, still stands below the bluff, tamer now, with a casino and some bars on the site of the old flatboat landing. From Natchez, the parkway runs north through the loess hills. The road cuts through them where necessary, with pale brown walls rising on both sides, forty feet high and more.

Between those cuts, the road runs along ridge tops, trees on both sides. No billboards. No visible commercial development. Just the road and the trees, at the same speed it has been since mile marker zero.

At mile marker 41, there is a pull-off on the right leading to one of the preserved sections of the original trace. You park and walk fifty feet, then descend from the modern road into the sunken trace. The floor is compacted clay, worn smooth, lower than the surrounding forest floor. The trace is narrow enough that two people walking side by side would crowd each other.

Trees arch overhead. You walk it for five minutes, then return to the car. The ruins of Rocky Springs are at mile marker 54. A town that once had a church, a post office, and hundreds of residents, abandoned in the late nineteenth century after yellow fever epidemics, exhausted soil, and the departure of a railroad line destroyed its economy.

The church still stands. The town around it is gone. The parkway passes without entering, but a trail connects the parking lot to the ruins. Past Jackson, the land flattens into the central Mississippi plains.

The parkway runs along ridge lines north toward Kosciusko, then through the hill towns of the upper Mississippi. No trucks. No billboards. No traffic lights.

After a lifetime of driving American highways, the absence of those things is a small, persistent surprise. Alabama passes and disappears in an hour. Thirty-three miles of hills, the same road, the same vegetation, the same speed limit. The southern sections of Tennessee look like sections of Mississippi moved north.

Wayne County is rural and wooded, the same mix of pasture and forest, the same ridgeline, then at mile marker 385. 9, the clearing and the obelisk. Stand beside Lewis’s grave in October, when the air is cold and the leaves are falling, and a car passes on the parkway, 200 feet away. Stand there another minute and another car passes.

People going to Nashville. People going to Natchez. None of them know what happened here on the night of October 10, 1809, and the road does not tell them, unless they stop and walk to the stone. North of the Lewis monument, the Duck River bridge, then higher ground, then signs of development through the trees.

Houses visible through the right-of-way screen. The corridor is still there, but narrower here than in Mississippi. The acquisition battles of the 1980s left their mark on how close the built environment comes to the edge of the parkway. At mile marker 438, the road reaches the edge of Birdsong Hollow, where the bridge begins.

The pavement spans the hollow on two parallel arches, 155 feet above the hollow floor, and if you stop at the parking area on the south rim and walk to the railing and look down, the floor of the hollow is very far below. A trail descends from the parking area on the north rim to the hollow, so you can walk to the bottom and stand beneath the bridge and look up. From below, the two arches look wider than they do from the deck, and the span between the hollow walls looks wider still. You can hear the cars crossing as a steady, low rumble through the concrete.

A twelve-ton car at 50 miles per hour, carried on two arches of hollow concrete, 155 feet above where you stand. The northern terminus is at mile marker 444. 1, an interchange with Tennessee State Route 96 and Interstate 40, eight miles west of downtown Nashville. There is a parking area, a small display building, and the sound of the interstate is audible from where you stand.

You exit the parkway onto Route 96 in about thirty seconds. Gas stations appear. A traffic light. A semi truck passes.

The experience of the preceding 444 miles ends at an unmarked line on the road. One detail from the 1934 authorization has stayed with me: Congress specified the length of the road at approximately 450 miles. The road that was built is 444 miles. The surveyors who laid out the final route, working from 1934 into the early 2000s, found a road six miles shorter than the original estimate.

Those six miles accumulated over seven decades of routing decisions, acquisition outcomes, land conditions, and engineering choices. No single person decided that length. It was the sum of everything that went into building the road. The original trace was perhaps 500 miles.

No one ever measured it precisely. It went where the ridges went, where the crossing was shallow, where the next camp was. It was not straight because the land was not straight, and the men who walked it made small decisions about the easiest way forward. At dozens of sites along the parkway, where the sunken traces still lie beside the modern pavement, you can step down into the original trace and walk it for a few minutes.

The floor is still worn below the surrounding ground level. The trace still points north. It follows the same logic the Choctaw used, and the Kaintucks, and the mail riders. Driving at fifty miles an hour on the parkway above, you cross 200 yards of the sunken trace in about fifteen seconds.

Walking it takes five minutes. Both paths go the same direction.