Three men talked by a pool, above what waited below

They had no backup lights. No guideline. No buoyancy compensator.

Between the four of them they had one small cylinder strapped to a leg, one regulator, and one torch tied to a forearm. None of the four held a cave diving certification. The pool they intended to dive was three metres deep.

At half past seven on the evening of 26 February 2001, the four students walked into the Ward entrance of Clarksville Cave.

Clarksville is a limestone cave in Albany County, in eastern New York State — the best known and most visited wild cave in the northeastern United States. The whole system runs about fifteen hundred metres and lies almost horizontally: no pits, no drops.

That is exactly what fools people.

Most of the passage in Clarksville is wide enough for visitors to walk upright. A few stretches have to be crawled. And a few stretches, all year round, are underwater.

The cave has three entrances. Gregory, Ward and Thacher — each opening into a different section. The four students went in through Ward.

Their destination was at the far northern end of the system: a chamber with a pool of water in it, known locally as the Lake Room.

At the bottom of that pool is an opening. About a metre wide and half a metre high.

Through that opening runs a flooded rock tube about fifty-seven metres long, called Pauley's Avenue.

People had swum through Pauley's Avenue before and come out the far side. Beyond it is cave that has never been mapped.

The four had heard the story.

Robert Svensson was twenty-three, a graduate student in computer science at Rensselaer Polytechnic Institute.

He had done a great deal of caving. In dry caving his experience was real.

Diving he had taught himself. No school, no certification, no paperwork of any kind.

The three friends with him: Joseph Baj, twenty-four, from Ozone Park, Queens. Jonathan R. Allison, twenty-three, of Troy. And Michael H. Chu, twenty-two, of Niskayuna.

Two of those three held open-water certifications — the kind that teaches you to dive in the sea and in lakes, places where if something goes wrong you go straight up and there is sky. Both had had theirs less than a year.

The third, Chu, had never dived at all.

None of this was their first trip to Clarksville. They had been in the cave many times. They were the ones who had found that the opening leading from the Lake Room into Pauley's Avenue had been blocked by fallen rubble.

The plan for 26 February was not to make the through-swim. The plan was just to clear the rock.

Clear it now, come back another day.

Because they were only going to shift some stones in a three-metre pool, they did not bring much gear.

The first stretch in from the Ward entrance is the widest part of the cave. A person walks upright.

Then the passage narrows. Past the area below the old north entrance the cave pinches down and you have to crawl. The four crawled a winding section, then reached the Lake Room.

The Lake Room is not large. The water is flat, black, about three metres deep. Four degrees.

Four degrees is the kind of cold that makes fingers slow within minutes.

The only light in the room was from their few hand-mounted torches sweeping back and forth over the wet walls. Their breath came out as thin smoke. The pool sat in the middle, so black it reflected nothing, and their voices came back off the stone ceiling sounding closer than they were.

Chu stayed on the bank — he could not dive.

The other three took it in turns to go down.

The work was almost comically simple: go to the bottom, pull rock, push silt, come up and breathe. When one man tired, they swapped.

But the bottom of the pool was covered in fine silt.

One light stroke of a fin and the whole body of water turned to milk. A torch shone into it became a white smear just in front of the mask. Beyond thirty centimetres it was meaningless.

And they had laid no line.

From that point on, the three men working on the bottom were operating almost entirely on memory and fingertips.

After a number of rotations, the cylinders began to run down.

One man was down to about a seventh. One was at about three fifths.

Svensson was at a third.

In cave diving there is a rule every textbook teaches on the first day: the rule of thirds. One third of your gas to go in, one third to come out, one third set aside for whatever nobody sees coming.

Which means that when a cylinder is down to a third, you should already be out of the water.

Svensson said he wanted to go down once more. Just to see whether the opening was clear yet.

His friends told him not to.

He did not listen.

Then they stopped arguing. Because it was a three-metre pool, directly beneath their feet, and what is dangerous about three metres.

Svensson went in.

He found his way to the opening on the bottom.

The water was still fouled. He could see nothing; he worked along the edge of the rock by hand, checking one section at a time.

In that kind of total darkness a person loses the sense of distance. Feel a little further, a little further.

By the time he stopped, half his body was inside Pauley's Avenue.

Then he realised his gas was nearly gone.

He decided to back out.

The place he had stopped in was a funnel-shaped restriction: going in, the rock guides you along; coming out, every edge of it turns around and braces against you. His hips and the cylinder on his back jammed against the wall.

Worse: it was not a straight section of tube. It was a bend.

Inside a tight bend, a diver cannot get the cylinder off his back. There is not enough room for the arms to fold behind him.

Robert Svensson was completely stuck.

Around him was black silt, rock, and the four-degree cold that had long since worked its way into him. The torch was still lit on his forearm, but it only showed a murky halo right in front of the mask and told him nothing more. At that range, his own breathing sounded very loud.

On the bank, three metres straight up, his three friends were still sitting and talking.

They heard nothing. Water does not carry a call for help.

Then Allison noticed that no more bubbles were breaking the surface.

He went in.

The water was as fouled as before. He could see nothing at all. He swept both hands around the opening in the floor — and touched a foot.

He felt more carefully. It was Svensson's foot.

Allison pulled.

At first it worked. He drew his friend out about thirty centimetres.

Then the rock closed like a pair of pliers and did not give another millimetre.

Allison's cylinder was the emptiest of the three. He was taking his last breaths from it. He had to let go and surface.

He told the other two.

Baj went down next — his cylinder had the most left, close to three fifths. He pulled. He could not move him either.

Baj surfaced, stripped off his gear on the bank, and started crawling back out toward the cave entrance to call for help. From the Lake Room to the mouth is a long crawl.

He had just left when Allison went back into the water.

This time he took the cylinder Baj had taken off. He dived down, wedged the cylinder in the place where he thought Svensson's hand might be able to reach it, and left it there.

Then he began doing the thing that people have repeated most often about this whole story.

He had no gas left. He took a breath at the surface, dived, pulled, and came up when his breath ran out. Took another breath. Went down again. Pulled again.

He did this over and over, in four-degree water. Each time he came up, his breath smoked in the cold air of the cave, and the surface broke and closed again behind him.

And each time he went down, more silt came up and the water got worse.

Allison later described the place in one sentence: like crawling into a storm drain packed with black mud.

More than ten minutes passed like that.

The foot at the bottom had stopped moving long before.

And the cylinder he had pushed down there was still exactly where he had put it. Nobody had touched it.

Only then did he let himself believe it.

At around nine at night, Baj reached the entrance and got word out.

But a cave rescue is not a thing that arrives once you have called. People who know the cave have to be assembled, diving equipment has to be gathered, and all of it has to be carried through a crawl that people find difficult with nothing in their hands.

Three hours later, the rescue team reached the Lake Room.

Close to a dozen divers went down in turns. They brought pry bars. They tried everything.

The rock would not let go.

At about half past six the following evening, they pulled every diver out of the water and changed approach: if you cannot get the man out of the water, take the water away from the man.

They reopened the north entrance — sealed off years earlier by the owner of the property above — ran hose in, and started pumping.

It took two days to drain the Lake Room.

The opening in the floor came into view under the lights. And a pair of feet came into view.

Another day of chipping rock.

At seven in the evening on 1 March 2001, Robert Svensson was brought out.

From the moment he entered the water to that point was three days.

For those three days, dozens of volunteers worked in shifts outdoors in a New York February, to bring a twenty-three-year-old home to his family.

The cause of death was not in question: he ran out of air.

As for the cause of the entrapment, the accident report came down to four points.

One, the passage was too narrow.

Two, the passage had a bend in it.

Three, it is possible the rock shifted while he was passing through — the same rubble the four of them had spent the evening clearing.

Four, the evidence indicated he had tried to turn around.

The fourth point is the one cave divers talk about most. In a funnel-shaped restriction with a bend in it, turning around cannot be done. But when a man is stuck in the dark and hears his own breathing getting harder, the body's first reflex is to turn back toward the way out.

And that reflex locked him in place.

When he was found, he was held so tightly that he could not get a hand down to the small reserve cylinder strapped to his own leg.

After the accident, the organisation that manages the cave sealed the north entrance again.

Then they did something more final: they filled part of the pool itself with rock.

The Lake Room is now about a metre deep.

Nobody can dig it out again. Nobody can dive it again.

Clarksville is still open. Visitors still come in through the Ward entrance, still walk the wide upright passage, still crawl the tight section, still reach the northern end of the system.

They stand beside a puddle that comes up to the ankle.

Most of them have no idea where they are standing.

WHY A THREE-METRE POOL KILLED SOMEBODY

What a "restriction" is in cave diving

The technical term is restriction — a section of passage narrow enough that the diver can no longer move freely.

There are grades. A mild one is a place where two divers cannot swim side by side. A severe one is a place where the cylinder has to come off, be pushed ahead, and the diver follows it through.

The most dangerous kind is not the narrowest. It is the UNEVEN kind: wide at the entrance, tightening as you go, the rock faces angled to one side like a funnel. Add a bend and the trap is complete.

That is exactly what Robert Svensson went into.

The list of rules on the wall of every cave diving school

Cave divers worldwide have a short list of rules, distilled from the analysis of hundreds of fatalities. The frightening part is that almost every fatality broke at least one line on that list.

1. Get cave training. An open-water certification is no use here. In the sea, if something goes wrong you go up. In a cave, above you is rock — there is no surface to go up to.

1. Run a continuous guideline to the exit. When the lights fail and the water fouls, the line is the only thing left that points out.

1. The rule of thirds. Never borrow from the reserve, not even once.

1. Do not exceed the depth you have been trained to.

1. At least three independent light sources per diver. One fails and you turn around immediately, no negotiation.

Against the night of 26 February: the four men broke rules 1, 2 and 5 the moment they walked into the cave. Rule 3 was broken on the last dive.

Decades of cave diving accident statistics in the United States point to the same place: most victims know how to dive, but have no cave training. The reason is not swimming skill. It is the things only a cave course teaches — fin techniques that do not stir silt (an open-water diver kicks downward, and in a cave one kick like that blinds the whole room), gas management as a compulsory calculation rather than a feeling, and above all the discipline of turning around.

Turn-around discipline runs against human instinct: at any sign that something is not right, you turn back NOW, even if the objective is only a few metres further on. A person who has not been taught this will always think, just a little further.

And the shallower the pool, the easier that thought comes. Three metres of water, what is there to be afraid of. The danger at Clarksville was never three metres of depth. It was fifty-seven metres of horizontal distance, with an unbroken rock ceiling above the whole of it.

A note against a common confusion

⚠ Two numbers in this account are easy to read into one another. "Three metres" is the DEPTH of the pool in the Lake Room — the vertical distance from surface to bottom. "Fifty-seven metres" is the HORIZONTAL length of Pauley's Avenue, the flooded rock tube underneath it, running sideways through the limestone. Robert Svensson did not die at a depth of fifty-seven metres; he was trapped at the restriction at the start of that horizontal tube, immediately below a pool three metres deep.

⚠ The four-degree water temperature comes only from retellings of the incident. It is plausible for a New York limestone cave in February, but contemporaneous news reports do not give the figure.

Sources

• https://storymaps.arcgis.com/stories/4b712d946d044f3287a97ab8948bb49b
• https://en.wikipedia.org/wiki/Clarksville_Cave
• https://webapps.bethlehempubliclibrary.org/spotlight/years/2001/2001-03-07.pdf
• https://www.timesunion.com/news/article/Cave-rescue-experts-relive-scary-moments-in-the-13094779.php
• https://caves.org/wp-content/uploads/2021/10/June_03_ACA99-01.pdf

AI-generated illustration. Not a photograph of the actual incident.