It lies beneath the flat red [music] clay country of southwest Georgia inside a park of more than 500 acres [music] where visitors drive a quiet loop past flags, white marble, and a small [music] granite pavilion.
Walk it on a summer afternoon and you will hear cicas, car [music] doors, and the low voice of a tour guide.
In the summer of 1864, you would have heard something else entirely.
You would have heard thousands upon thousands of men packed onto a bare sloping field.
A constant murmur of voices broken by coughing and the scrape of makeshift shovels.
You would have smelled smoke, mud, and a creek that was no longer safe to touch.
And you would have felt the Georgia heat pressing down on a place with almost no shade.
Its builders called it Camp Sumpter.
History remembers it by the name of the little railroad stop beside it, Andersonville.
[music] Today, the walls are gone.
Not one original log of the great stockade still stands, but the ground still holds things that refuse to sit quietly inside the story you were taught.
Dark stains in the clay [music] running in straight lines, each one marking the spot where a pine log 20 ft long was once driven into the earth.
Round pits a yard or two across, dug by hand and abandoned long ago.
A narrow tunnel that prisoners once dug toward freedom.
And until just a few years ago, a hospital that nobody could find.
Then came the detail that made no sense.
Andersonville is one of the most thoroughly documented prisons in American history.
Roughly 45,000 Union soldiers passed through it.
Nearly 13,000 died there, more than one in four.
And thanks to a secret list, copied by a teenage prisoner and smuggled out of the camp, about 95% of the graves in its cemetery carry a name.
And yet, the place itself has almost vanished.
According to the National Park Service, the exact location of the camp’s last hospital was largely unknown as recently as 2018.
A radar search for the gallows where six prisoners were hanged came up empty.
A cemetery with almost no mystery.
A camp with almost no trace.
For a century and a half, that gap was filled by stories.
Survivors blamed the camp’s commander.
His defenders called him a scapegoat.
And over a spring that prisoners believed was sent from heaven, survivors raised a monument of granite and marble.
The more the story hardened, the less anyone asked what the ground itself might say.
Then, beginning in 1973, archaeologists led by Lewis Larson Jr.
And Morgan Crook followed the stains in the clay with shovels and tels in National Park Service team opened more ground in 1989 and 1990.
In 2005, radar crews rolled across the field.
In 2018, specialists swept tens of thousands of square meters with radar, [music] magnetometers, and resistivity meters.
They are still working, and what they are finding does not always match what the books say.
[music] A quick note on how I will tell this story.
At a place like this, the difference between what is documented, what has been measured, and what is only interpretation matters more than usual.
I will mark each one as we go.
So, what was Andersonville built to be?
How did a place meant for 10,000 men end up holding more than [music] 30,000?
And what can the soil tell us that the paperwork cannot?
To answer that, we have to go back to the winter of 1863 and 1864 to a war in which the two armies had stopped trading prisoners.
In the early years, captured soldiers were usually exchanged man for man and sent home.
By 1863, that system was breaking down.
Each side blamed the other, and historians still argue about where the blame belongs.
What nobody disputes is the result.
Prisoners piled up faster than anyone could house them or feed them.
Richmond, the Confederate capital, was crowded and short of food.
Officials began looking for somewhere else to put thousands of men.
They chose a remote stretch of pinewoods in southwest Georgia beside a small railroad stop far from the front lines.
It had a rail line to bring men in.
It had a forest to build with and it had something else, a creek running right through the middle.
Think about the decision itself.
Somebody had to look at a patch of Georgia pine and decide that it would hold human beings by the thousands.
Somebody had to pick a number.
The number they picked was 10,000.
Hold on to that creek and hold on to that number.
Together, they would matter more than anyone planned.
How do you build a prison for 10,000 men in only 6 weeks?
Construction began in January 1864.
Crews cut pine from the surrounding woods and hewed the trunks into logs about 20 ft long and 8 to 12 [music] in thick.
Each one was set upright, 5 ft into the ground, side by side, until they formed a wall roughly 15 ft high.
The rectangle ran about 1,600 ft long and nearly 800 ft wide, enclosing a little more than 16 acres.
The work took about 6 weeks.
The first prisoners arrived on February 27th, 1864.
The stockade had been designed for 10,000.
Through the middle of that enclosure ran the creek, known as Stockade [music] Branch, a small tributary of Sweetwater Creek.
On paper, it was [music] the camp’s water supply.
In practice, before it ever reached the prisoners, it flowed through two Confederate encampments and passed the prison [music] bake house.
Latrines sat upstream and when heavy rain came they overflowed.
The only stream in the camp served for drinking, for washing and for sanitation all at once.
Picture that stream in the August heat running through the lowest ground in the camp, carrying the waste of one group to the lips of another.
That is why I asked you to remember it.
19 ft inside the wall.
A line of small wooden posts topped with a rail marked a boundary the prisoners called the deadline.
Cross it and the guards were under orders to shoot.
The official ration was a/4 pound of cornmeal and either a third of a pound of bacon or a pound of beef.
The camp had almost nothing to cook it with.
[music] There were no barracks.
Prisoners built what they called shebangs.
Scraps of cloth and sticks stretched over a patch of ground, and many simply dug holes to crawl into [music] despite the risk of suffocation.
And the men kept coming.
In June, the stockade was enlarged by about 10 acres to roughly 26 1/2.
By early August, more than 33,000 men were inside it.
Do the arithmetic, 26 1/2 acres is about 1.15 million square ft.
Divide that by 33,000 men and you get roughly 35 square ft each, a space about 5 feet by 7.
And that is before you subtract the marshy ground along the creek where no one could build or lie down.
Think about that for [music] a moment.
Every man you are picturing was sharing a space that size with his shelter, his sickness, and everyone beside him.
Now, here is where the ground begins to talk.
When a wooden post rots in the earth, [music] it leaves a ghost.
The soil where it stood turns darker and softer than the clay around it.
And an archaeologist who scrapes the surface clean can see it.
A circle of stain as sharp as a stamp.
In 1973 and 1974, Lewis Larson Jr.
[music] and Morgan Crook opened trenches north of the prison.
Working with shovels and tels, crews of three to six people at a time, they recorded the line of the outer stockade and a middle stade line beyond it.
They found the corners of the inner stockade, the position of the north gate, a gully, and what may be the places where the deadline posts once stood.
Their report did not appear until 1998, a quarter of a century after the digging, but it gave the park something it had never had, a measured map of a place that had been built to disappear.
Then in 2005, the Georgia Department of Transportation and the National Park Service sent ground penetrating radar across the southern part of the camp.
The team led by James Pomrred located the south gate and its stockade wall and traced portions of the outer, middle, and inner stockade walls.
Three lines, not one.
This was a layered barrier, not a simple fence.
And they measured the posts.
They reached down roughly 120 [music] to 135 cm beneath the surface.
That is about 4 ft, give or take.
The written record says five.
Be careful with that difference.
And remember that these trenches and scans sampled only a small slice of the camp.
So every finding is a window, not a whole map.
It does not prove anyone lied.
Soil washes away.
Ground levels change.
Radar is an estimate, not a ruler.
What it shows is something more modest and more interesting.
The physical evidence and the written record agree on the big picture and sometimes quarrel over the details.
That is exactly what an archaeologist hopes to find because the quarrel is where you learn something new.
But the radar saw something else in the middle of the camp.
The interior of the prison was riddled with pits.
The 2005 survey report describes it plainly.
The prison interior was rich with [music] deep pit features.
Round anomalies between 1 and 2 m across, [music] reaching depths of more than a meter and a half beneath the surface.
The authors believed they most likely represented handdug wells or house [music] pits.
In early 2018, a team from the National Park Services Southeast Archaeological Center came back with a bigger toolkit.
Between January 29th and February 8th, they surveyed more than 17,000 square meters on two slopes of the old camp.
Using four different instruments, each one sees something the others miss.
Radar bounces signals off buried changes in the soil.
A magnetometer picks up iron and burned earth.
A conductivity meter and a resistivity meter read how well the ground carries electricity, which changes where soil has been dug up and packed back down.
Together, [music] they let you see through the ground without disturbing it.
They found the line of the stockade.
They found the line of the deadline, and they found, in the report’s words, numerous circular pit feature anomalies, which appear to represent temporary shelters or wells.
Stop and think about what that means.
In a field with no barracks, a man who wanted shelter dug down, a hole a yard wide, covered with whatever cloth he could find, was a home.
And a man who wanted clean water dug down, too.
The National Park Services’s own handbook says [music] prisoners dug wells about 3 ft wide and in some cases as much as 75 ft deep with whatever tools they could find.
Careful though, radar cannot tell you whether a pit was a house or a well.
It can only tell you that the earth was disturbed.
The 2005 report itself says that only ground truthing, which means digging, will reveal what these features really are.
So when you hear that thousands of holes lie beneath that slope, remember that this is an interpretation, a very reasonable one, but not a proof.
There is one piece of digging that did happen.
In 1989 and 1990, the Southeast Archaeological Center excavated at Andersonville and recorded, among other things, the stains of stockade posts and an escape tunnel.
Somewhere under that ground, a group of men with almost no tools tried to dig their way out.
The center noted that some of what the shovels found differed from what the historical accounts said.
The details are buried in technical reports, but the message is clear.
The written story and the physical evidence do not always overlap.
Now, consider the other side of the pit.
Not all the danger at Andersonville came from the guards or the hunger.
[music] By the summer of 1864, a gang of prisoners known as the Raiders had begun to rob and attack their fellow [music] captives, preying on men too weak to fight back, other prisoners organized to stop them.
After a hurried trial run by the prisoners themselves, six of the raiders leaders were hanged on July 11th, 1864 from a newly built scaffold near the south gate.
[music] In 2018, archaeologists searched for that scaffold.
They also searched for the enclosures used for sick call and the barracks buildings of the guards.
On the opposite slope, they hunted for the settler’s store, the shop where a trader sold goods to the garrison.
On the south slope, they found neither the gallows nor the sick call enclosures.
The report offers the most likely reason.
Those structures were temporary.
A scaffold built for a single day or a fence built for a morning line of the sick leaves almost nothing behind.
A pit leaves a pit.
A post leaves a stain.
But a plank of pine that stood for hours or even for [music] weeks can rot into the soil and vanish without a trace.
The report also points to changes made to the site over the years since.
That means something unsettling about the archaeology of Andersonville.
The ground remembers what was dug.
It forgets what was built to be quickly torn down.
The archaeology of this place is in a sense the archaeology of holes.
Let me sort out what we actually know.
The way I promised at the start.
What is [music] documented?
The stockade lines, the gate, the line of the deadline, and an escape tunnel.
What has been measured?
The pits, [music] their size, and their depth.
What is interpretation that those pits were wells and shelters and that the missing gallows and sick call enclosures simply rotted away?
Each step down that list is a little less certain than the one above it.
So the surviving evidence tilts the story in a certain direction.
It preserves the struggle to survive, the wells, [music] the shelters, the escape attempts far better than it preserves the machinery of order.
If you read the camp only through what the soil kept, you would see far more of the prisoners effort to stay alive than of the structures that held them.
Quick pause.
If you are enjoying this, take one second to subscribe to Hollow Karen and tap the [music] bell.
This channel digs into American battlefields and prison sites with real archaeology.
And the next story is already waiting.
Now stay with me because the next part has a summer storm.
A spring [music] that nobody can fully explain and a number that you will not forget.
By the first days of August 1864, [music] the camp’s population had passed 33,000 and the weather was about to change.
On August 4th, 1864, the stockade held about 33,000 men, including more than 2,000 in the hospital.
By the end of that month, 2,933 of them would be dead.
That is nearly a hundred a day, every day for a month.
About four men every hour around the clock.
In the time it has taken you to watch this video so far, roughly a dozen would have died.
A prisoner named John Urban recorded the worst of it.
On a single day in early August, he wrote 175 prisoners died.
And in 3 days, the toll was nearly 500.
The camp’s commander, Captain Henry Wartz, was not blind to what was happening.
He wrote that [music] as long as 30,000 men are confined in any one enclosure, the proper policing is altogether impossible.
Hold on to that sentence.
[music] We will come back to it.
The men were dying mostly of disease.
Scurvy, diarrhea, and [music] dysentery led the list with typhoid and gangrine close behind.
All of them accelerated by malnutrition, filthy water, and crowding.
The camp hospital, which started inside the stockade and later moved outside, tells the story in miniature.
In the 10 weeks from late February to early May, more than 4,500 patients were treated there [music] and more than a,000 died, more than one in five.
Outside the walls, the war was not helping.
In late July, a Union cavalry raid under General George Stoneman set out to free the prisoners.
[music] It was defeated on July 29th, and about 500 of its men were captured and sent to Andersonville.
[music] Then came the rain.
In the first half of August, a violent storm swept across the camp.
The flood damaged a section of the stockade and in the aftermath, prisoners reported that [music] a spring of clear water had broken out of the ground.
The men who saw it had no doubt what it was.
They called it Providence Spring, a gift from God.
In one telling, a prisoner fashioned a sapling into a trough to carry the water across the deadline.
In 1901, survivors in the Woman’s Relief Corps placed a granite and marble monument over the spot.
It remains the best known symbol of Andersonville.
Now, notice what the sources do and do not agree on.
The date varies.
One National Park Service publication places the rain on August 9th.
Another puts the spring’s appearance on August 14th.
And Park Rangers Today offer a plain explanation.
Their account is that the spring was already there.
It had been covered over when the stockade was built and the storm washed the earth away and uncovered it again.
Careful though, that explanation is a rers’s interpretation of the geology, not an excavated finding.
What we have are two ways of telling the same event.
[music] One in the language of faith and one in the language of soil.
Both of them are about the same water.
But here is the part that should give you pause.
If the spring was such a miracle, [music] why did the dying continue?
Because clean water was only one of the killers.
A spring could not cure scurvy.
It could not shrink 33,000 men back into the space built for 10,000.
It could not make a handful of cornmeal into a meal [music] or turn a hole in the ground into a hospital.
Across the life of the camp, deaths averaged more than 900 a month.
Today, a pavilion of granite and marble stands over the spot with a sign warning visitors that the water is not safe to drink.
Think about that.
The most hopeful landmark in the camp is also a place where no one is allowed to drink.
And while the dead piled up faster than anyone could handle them, a small act of resistance was taking place in the camp’s medical office.
A 19-year-old cavalry sergeant from Connecticut named Dorance Atwater had been captured and sent to Andersonville.
He was assigned to work in the medical office where he kept the official register of the dead.
And at some [music] point, he came to suspect that no one on the Union side would ever see that list.
So [music] starting in August, he began secretly copying the whole thing by hand.
Every name, every regiment, [music] every date of death, the list grew by the day, and so did the danger of being caught with it.
What he did with that copy would decide whether 12,000 families ever learned where their men were buried.
To understand why that list mattered, you have to understand what happened to the dead.
When a man died at Andersonville, a label was attached to his body and he was carried to a place called the dead house opposite the south gate.
From there, the bodies were taken to a burial ground beside the camp.
Prisoners recalled that the dead were buried without their clothes because the living needed them.
The graves were not graves in any ordinary sense.
They were trenches about 3 ft wide with the men laid side by side, shoulderto-shoulder.
Even the Confederate guards who died there were buried in trenches of the same kind.
Each grave was marked with a numbered wooden stake.
If the numbers were lost, so were the names.
But someone had kept the names, and someone else had copied them.
In February 1865, Atwater was parrolled.
He left Andersonville on February 2nd carrying his secret list.
[music] The National Park Service puts the number at 12,636 names.
He smuggled it out of the camp and held on to it for the rest of the war.
The following summer in July 1865, Clara Barton accompanied an army expedition to Andersonville to identify and mark the graves.
The crew numbered 34 men, including two clerks, a foreman, 12 carpenters, 12 letterers, and seven laborers.
They compared the numbers on the old wooden stakes with the names on Atwater’s list.
Then they made new headboards, each one carrying the soldier’s name, company, regiment, date of death, and his original grave number.
Imagine being a mother in the north [music] in 1865 who had not heard from a son in a year and then reading his name, his regiment, [music] and the date from a list that a 19-year-old had copied by hand in a prison camp on July 26th, [music] 1865.
The burial ground was formally established as Andersonville National Cemetery on a site of about 27 acres.
Two main [music] roads divided into four quadrants with 18 sections for burials.
Around 1877, [music] the wooden headboards were replaced with marble.
The result is almost unbelievable.
According to the New England Historical Society, Atwwater’s list allowed the identification of about 95% of the graves.
In a war where so many soldiers were buried unnamed, that was a triumph of paperwork.
Do not overlook the other side of that number.
95% identified means 5% not out of nearly 13,000 graves.
That is still hundreds of men whose names the list could not supply.
Men who lie beneath markers that say only that they are unknown.
And what happened to the clerk who made it possible?
He was arrested.
The same source says that after he refused to give up a copy of his list, Atwater was court marshaled and sentenced to prison.
Clara Barton went to President Andrew Johnson and secured a pardon.
Atwater’s list was published by the New York Tribune in 1866.
Careful about the numbers.
The sources do not agree on the totals.
Depending on whom you read, the number of men who died at Andersonville is given as 12,912.
12,919 or 12,920.
The National Park Services page on the cemetery [music] says 12,920.
Atwater’s list is described as 12,636 names in one National Park Service article [music] and 12,912 in another.
These are small differences, but they are real and they are a reminder that even the best record is a human product.
So, we have a cemetery that is well documented.
Does the ground [music] agree?
In 2005, the radar team also scanned a section of the cemetery.
They looked at an early part of it, and their finding was a surprise.
The radar read all the graves there as single burials at depths of roughly 150 to 180 cm.
That is 5 to 6 ft.
The report describes these as predating the later mass burial practices that does not match the picture of shallow shoulderto-shoulder trenches you may have read about in other accounts.
Be careful with that, too.
The radar reading is not an excavation, and no one has opened those graves.
The accounts of trench burial may describe the busiest weeks of summer 1864 [music] when the pace of death was at its worst.
The early section may simply be earlier from a time when the camp was quieter.
This is a puzzle, not a conclusion.
But it is a reminder that even the most documented place in the Civil War [music] can still surprise you.
The same survey went looking for the dead house.
The building through which so many men passed on their last journey.
The radar did not find it.
The graves have names.
The building where the dead were gathered has not been found.
And that leaves one more question, the one that Americans have argued about since 1865.
If the dead are so carefully counted, who is responsible for them?
Captain Henry Wartz was arrested on May 7th, 1865.
He was tried by a military tribunal on charges of war crimes.
Former prisoners testified about cruelty and a Confederate surgeon’s report proved damaging.
Vertz was found guilty and sentenced to death on November 10th, 1865.
He was hanged in the courtyard of the old Capitol prison.
He remains the only Confederate official executed for war crimes.
For many survivors, that was justice.
For his defenders, it was something else.
They argued that Wartztz was a scapegoat for a collapsing government that could not feed its own armies, let alone its prisoners.
Historians [music] still divide on the question.
Some argue the deaths were the result of Confederate [music] policy and amounted to war crimes.
Others point to disease driven by overcrowding and to a widespread food shortage across the South.
Both sides have evidence.
There is also the dispute over who was to blame for the end of prisoner exchanges.
In July [music] 1864, W parrolled five soldiers to carry a petition north signed by a majority of the camp’s prisoners [music] asking the Union to resume exchanges.
The request was denied.
Remember the sentence I asked you to hold on to.
As [music] long as 30,000 men are confined in any one enclosure, the proper policing is altogether impossible.
You can read it as a warning.
You can read [music] it as an excuse.
You can read it as both.
The sentence does not change.
What changes is what you believe about the man who wrote it.
And now [music] let’s be honest about what the ground can and cannot say.
The ground cannot tell you what anyone intended.
A pit in the clay does not carry a motive.
Neither does a line of stains.
Archaeology cannot settle the argument over who bore the blame.
And anyone who tells you that a magnetometer reading proves a verdict on Henry Wartz is selling something.
What the ground can do is show the scale.
It shows a stockade built for 10,000 and packed with three times that number.
It shows pits dug by thousands of people with nothing else to live in.
It shows a creek that ran through the middle of the camp and a boundary line 19 ft from a wall.
The soil does not decide the case, but it makes it very hard to imagine that the paperwork told the whole story and the work is not finished.
In April 2018, the Southeast Archaeological Center surveyed 17,000 m south of the prison, searching for the third hospital, also called the shed hospital.
It was in use from December 1864 to April 1865 in the final months of the camp, and its exact location, according to the National Park Service, had been largely unknown.
The radar found what appears to be part of its western and northern stockade walls and part of a middle stockade.
After that survey, the total area scanned at Andersonville had reached more than 34,000 m.
[music] The team has said it intends to continue with the area around Market Street near the North Gate and possibly [music] the site of the second hospital still to come.
Think about what that means.
In 2018, archaeologists were still finding lost parts of the camp.
More than 150 years after it closed and put the scale in perspective, 34,600 m is about 8 1/2 acres.
And not all of that lies inside the stockade.
At its largest, the stockade enclosed 26 1/2 acres.
So even now after years of work, the survey has looked at only part of the ground.
Everything I have told you in this video comes from that part.
And plenty is still missing.
[music] The dead house has not been found.
The gallows has not been found.
The [music] sick call enclosures have not been found.
Every one of them was temporary and every one of them mattered.
[music] Today, a visitor can drive the loop around the old stockade ground.
A replica palisade, [music] a guard tower, and a gate stand near the center, and reconstructed shebangs sit near the far end.
The Providence Spring Pavilion stands on the grounds.
The National Cemetery holds the marble markers of about 12,900 men.
And the National Prisoner of War Museum tells the story of prisoners from all of America’s wars.
Please remember that Andersonville [music] is a protected national historic site.
Digging or removing anything from it is illegal and the real remains are still under the surface where the people who study them want to [clears throat] find them.
Stand on that slope at dusk when the cicas quiet down and the light turns long across the grass.
Somewhere under your feet are pits and posts and a tunnel.
And not far away, in rows, lie the men who were counted.
One by one, because a 19-year-old kept writing, “The dead were counted.
The place is still being found.”
If someone in your family served in the Civil War, tell me their name in the comments.
And if you want the next buried contradiction before it surfaces, subscribe to Hollow Karen.
