Some of the most important clues in the mystery of the SS Edmund Fitzgerald may be the smallest ones. And they could still be fastened to its deck.

When the doomed freighter launched in 1958, it was the longest ship moving through the Great Lakes, a 729-foot ore carrier designed to haul incredible loads through some of the most dangerous inland waters in North America. On November 9, 1975, the ship left Superior, Wisconsin, with more than 26,000 tons of taconite pellets aboard, bound for the steel mills of Detroit.

By the next evening, Lake Superior had swallowed the Edmund Fitzgerald whole.

The Fitzgerald wasn’t sailing alone. As it headed toward Whitefish Bay, Captain Ernest McSorley stayed in radio contact with Captain Bernie Cooper on the Arthur M. Anderson, which followed behind.

What happened during those chilling final hours, as Emily Sohn reports in this Popular Mechanics feature, shows how quickly an ordinary Great Lakes crossing turned into a deadly fight to stay afloat.

“The two captains—McSorley on the Fitzgerald and Bernie Cooper on the Anderson—stayed in contact by radio as the wind picked up early the next morning. By the afternoon of November 10, gusts of 90 mph were whipping snow into squalls. Waves swelled to the height of a two-story building and crashed over the ships with such magnitude that, Cooper would later say, up to 12 feet of water accumulated on his deck at times.”

Around 3:30 p.m., McSorley told Cooper that the Fitzgerald was damaged—snapped railings, broken vents on ballast tanks, and a list—and the ship’s radar failed. McSorley slowed down so the Anderson could catch up and help guide him toward Whitefish Bay.

When Cooper checked in just before 7 p.m., after two huge waves rolled over the Anderson and toward the Fitzgerald, McSorley said he was “holding his own.”

Fifteen minutes later, the Fitzgerald vanished from the Anderson’s radar and never came back. All 29 men were lost. Whitefish Bay was just 17 miles away.

How did a ship that large disappear so quickly without a distress call? After half a century, we’re still asking that question.

One answer may be the ship’s cargo hatches. On an ore carrier, the hatch covers seal the openings above the holds. In a vicious Lake Superior storm, the clamps that hold those covers down—normally regular working hardware—are part of the ship’s skin.

If the covers weren’t fully secured, or if waves damaged them, water could pour into the cargo holds. As that water accumulated, the Fitzgerald would sit lower in the water, list more severely, and become more vulnerable to waves.

While the Coast Guard and National Transportation Safety Board both pointed to cargo-hold flooding in their official explanations, investigators have never settled on one clean answer. Other theories point to structural weakness, grounding, overloading, or a catastrophic break on the surface that happened too quickly for the crew to send a distress call.

But why are we still so interested in those clamps? Because they’re still down there.

Ric Mixter, who visited the Fitzgerald wreck in a submersible in the 1990s, told Pop Mech that if he could return, he would take photos of every hatch clamp and investigate where the taconite ended up. And because nothing has been moved since the Fitzgerald sank, modern sonar, ROV cameras, and photogrammetry could document the hatch covers, clamps, coamings, cargo position, and damage patterns without disturbing the site.

For the full story of the SS Edmund Fitzgerald’s final voyage, the theories that persist, and the tech that could change what investigators know, read Pop Mech’s deep dive now.