The cave divers who went back for their friends

In February 2014 two divers died at a depth of more than 100m in a huge cave system in Norway. The authorities said it was too dangerous to retrieve their bodies, but four friends of the men decided to take the risk - and seven weeks later they descended into the dark and glacial waters.

At the end of the Plurdalen valley in central Norway a 35m-wide river rises abruptly out of the ground.

If you dive into this strange pond, known as Plura, and swim underground for half a kilometre, you will emerge into a long, colourful cave.

Diving hobbyists can climb out of the water here to admire the grotto, before returning to Plura. But if you are highly trained and experienced - and an insatiably curious individual - you might continue on a course that quickly plunges much deeper, becoming narrow and difficult, through ice-cold, pitch-black water.

After negotiating this "sump" - an underground pocket of water - you will finally ascend to the cave of Steinugleflaget. And about 90m above the cave's vaulted ceiling lies your exit - a crack in the collapsed side of a hill.

On 6 February 2014, two divers cut a triangular hole in the ice at Plura, and, encased in waterproof dry suits and diving equipment, slipped into the water. Two hours later, after the first divers' dust had been allowed to settle, three of their friends followed behind.

The destination for all five men was Steinugleflaget. They were Finnish divers who knew one another from explorations they had made of the Ojamo mine, west of Helsinki. As was the custom of the group, no-one was in overall command, but the first diver to set off was Patrik Gronqvist, one of a trio of Finns who had discovered the passage between the caves the year before. He was diving with his good friend Jari Huotarinen, who was attempting the traverse for the first time.

The trip was at the extreme end of a dangerous sport. While most amateur divers might restrict themselves to dives of between 30 minutes and an hour, at a depth of 30m or so, the trip to Steinugleflaget would be a five-hour dive, with the aid of underwater scooters, to depths of more than 130m.

"The deeper part is very demanding, very cold water and narrow tunnels, and deep as well - it is the world's deepest sump that has been dived through," says Gronqvist.

At such depths and temperatures, a tear in a dry suit on the sharp cave floor could result in death.

There is also the possibility of equipment failure, and hypercapnia - carbon dioxide poisoning. "Carbon dioxide absorbs into the bloodstream much faster and easier at depth," says Gronqvist. Cave divers use "rebreathers" which artificially absorb the carbon dioxide they exhale, but these can become overloaded if the divers start breathing quickly, and at depth it is more difficult for them to control their breathing. "If you have to do anything physical - swim harder or faster or anything - that's very dangerous," says Gronqvist.

Hypercapnia can be deadly, but even a mild case may cause confusion and disorientation, which in a deep cave is liable to have serious consequences.

About an hour into the dive, shortly after the pair had swum through the deepest section and were about 110m lower than the cave entrance at Plura, Gronqvist realised that Huotarinen was not behind him. He went back and found his friend had become stuck in a narrow section of the cave, entangled in a cord connected to a piece of his equipment. He was using his torch to signal distress.
Huotarinen seemed to be starting to panic, which meant he risked breathing too fast. Gronqvist gave him a cylinder of gas to reduce the amount of carbon dioxide in his system, but while Huotarinen was switching mouthpieces, he started helplessly swallowing water.

To Gronqvist's horror, his friend died in front of his eyes - but getting agitated would put Gronqvist himself at risk of hypercapnia. After a brief effort to free the body, he forced himself to calm down.

There was nothing for it but to continue to Steinugleflaget - very slowly. Divers who have spent time in deep water cannot go straight to the surface because of the risk of decompression sickness, another potentially fatal hazard. The deeper the dive, the longer the decompression. Because he had stopped to help his friend, spending about 20 minutes at a depth of 110m, Gronqvist knew he would have to spend hours making additional decompression stops before surfacing.

He also knew that at some point the second group of divers would find Huotarinen's body blocking their way.