The Edmund Fitzgerald sank on the evening of November 10, 1975, sometime between roughly 7:10 and 7:30 p.m. Eastern Standard Time, in eastern Lake Superior about 17 miles north-northwest of Whitefish Point, Michigan. All 29 crew members died. The ship broke apart and came to rest at a depth of about 530 feet. The exact second and the precise sequence of failure remain unresolved, but the timing window and the sequence of events are well constrained by radio traffic, weather records, and wreck-site surveys.

The documented sequence at a glance

Radio and weather records fix most of the evening tightly. The most important detail is the gap between the last two transmissions: a routine check-in that drew no concern, then nothing at all.

Time (EST, Nov 10, 1975) Event
2:00 p.m. Captain Ernest McSorley reports the Fitzgerald "rolling some" and slowing in heavy seas
About 3:30 p.m. The ship passes 20 miles southeast of the Arthur M. Anderson, which is also bound for Whitefish Bay
About 5:30 p.m. McSorley radios the Anderson: "We have a bad list, lost both radars"
About 6:30 p.m. McSorley asks the Anderson to keep it on radar; the Anderson agrees
About 7:10 p.m. Final transmission: "We are holding our own"
After 7:10 p.m. The Fitzgerald vanishes from the Anderson's radar, with no distress call
About 7:20–7:30 p.m. The Anderson begins searching and reporting to the Coast Guard
November 14, 1975 A Navy aircraft detects the wreck with a magnetic anomaly detector

The sink itself was almost certainly fast. The absence of an SOS, the ship's disappearance from radar, and the condition of the wreck all point to a breakup and plunge lasting seconds to a few minutes, not a gradual flooding. The Coast Guard's formal Marine Board of Investigation, completed in 1977, concluded the vessel sank in deep water without a distress signal. That is the authoritative statement of what happened; it is not an explanation of why.

How the ship likely broke and went down

The Fitzgerald's bow was high and buoyant when a ship is loaded; the stern rode lower. In extreme following seas, the bow can plunge off a wave crest and bury itself. Water then drives aft along the deck at high speed. The deck cargo hatches sit in that path. The conventional explanation, central to the 1977 Coast Guard report findings, is that hatches were damaged or loosened and seawater poured into the cargo holds. Once water mass loaded the forward holds, the bow lost buoyancy and the ship nosed under.

There is a competing mechanism with serious support. The Coast Guard's 1977 board favored flooding through the hatches; a 1978 National Transportation Safety Board report concluded the sinking probably resulted from flooding through ineffective hatch closures, possibly combined with hull damage on the shoals near Caribou Island the previous afternoon. Neither conclusion is settled beyond dispute, and the difference matters: hatch failure implies an access problem, hull damage implies a structural one.

The wreck itself supplies constraints. Surveys found the ship broken in two, with the bow section upright and the stern section overturned, resting close together on the bottom. A ship that flooded slowly and settled evenly would not typically separate. A hull that broke at or near the surface, or broke during a plunge, would. That geometry is the strongest physical argument against a long, gradual flooding scenario.

The storm that made it possible

The weather on November 10, 1975, was extreme but not unprecedented for Lake Superior. A low-pressure system intensified rapidly and moved northeast over the lake. Sustained winds reached hurricane-force territory, with documented sustained winds around 50–60 mph and gusts well above that; waves in eastern Superior were reported in the 20–25 foot range, with individual seas possibly higher. A November gale is normal for the lake; what mattered was the combination.

One underappreciated factor is sea state direction. The Fitzgerald was running before following seas, meaning the waves pushed from astern. Following seas are more dangerous for a laden bulk carrier than head seas for the simple reason that the wave crest and trough move relative to the hull, and the bow can lose support as the ship rides down the back of a wave. The captain's 2:00 p.m. report that the ship was "rolling some" is consistent with a vessel being worked hard, not just weathered in.

The deeper point is that the lake's short, steep waves interact with a long, rigid hull in ways an ocean forecast does not capture. Lake waves have shorter periods and steeper faces than ocean swell of similar height because the fetch is limited. A 25-foot lake wave is not equivalent to a 25-foot ocean wave. Forecasts that treated them as analogous would have understated the stress.

Where the evidence gets thin

This is where professional caution matters. The 1977 Coast Guard board's core findings have been challenged on procedural grounds. The board's access to the wreck and its interpretation of the hatch evidence were limited. The 1978 NTSB conclusion was a competing reading of the same data, not an independent physical test. Neither body had a complete picture of the vessel's final minutes. Multiple later hypotheses, including structural failure at the hull girder from accumulated fatigue, were offered but not established.

The wreck-site surveys added physical evidence rather than a verdict. The bow and stern sections rest near each other, and the bow section shows damage consistent with a high-energy plunge. But bottom-impact damage does not cleanly separate what happened on the surface from what happened on contact with the lakebed. A hatch that failed, a hull that cracked, and a wave that flooded the deck can all produce similar wreckage.

The practical consequence is that any confident single-cause claim is overreach. The best-supported reading is a rapid, catastrophic loss of buoyancy and structural integrity driven by a severe following sea and a flooding mechanism that no one can now observe directly. The exact mechanism remains an inference.

Several claims circulate widely and do not survive scrutiny. A frequently repeated line about the ship being "two feet from being a foot shorter" or similar quips is folklore, not record. Other common errors are worth flagging:

The last radio message was not a distress call. "We are holding our own" is a routine report and reading it as a cry for help projects drama onto a normal exchange. The Fitzgerald did not vanish without warning signs. The list and radar loss reported at 5:30 p.m. were real symptoms, though the captain did not treat them as an emergency. There was no confirmed SOS or flare. Any account of a distress signal is unsupported. The ship did not sink "in the middle of the lake." It went down near Whitefish Point, close to the eastern shore and the safety of Whitefish Bay it never reached. The wreck was not found by accident. An aircraft magnetic survey located it four days later; submersible dives the following year documented the condition.

A counterintuitive detail is worth holding onto: the safest-looking stretch of the voyage may have been the most dangerous. The Fitzgerald was past the open-water stress and heading for shelter when it was lost. The final approach to Whitefish Point can concentrate wave energy, and a ship already compromised by flooding or damage would be most vulnerable when it could not turn into the seas or slow without risking broaching. Proximity to safety is not the same as safety.

What the case teaches about large-lake shipping

The loss changed how Great Lakes vessels are handled in heavy weather. The practical lessons are concrete and still relevant:

A "bad list" and lost radar in a following sea are emergency signals, not routine weather reports. Treat structural symptoms as distress until disproven. Hatch integrity is a safety-critical system, not a cargo convenience. Verify closures and seals before heavy-weather transits. Following seas demand speed and heading management. A laden bulk carrier cannot maneuver like a small vessel. Keep a nearby ship on radar in heavy weather. The Anderson's role as a witness is why the timeline exists at all. Do not treat lake-wave height forecasts as ocean-equivalent. Period and steepness change the stress on a long hull.

The Coast Guard's 1977 findings remain the closest thing to an official verdict, but they were framed as probable causes, not proof. That distinction is the professional caveat the legend tends to erase. A board of investigation in 1977 could not have observed the final ten minutes any more than we can now. The conclusion was the best inference available, and it is still an inference.

The human cost is the part that does not require interpretation. Twenty-nine people died within sight of a safe harbor, in a storm that was forecast and survivable, on a ship that had been reporting problems for hours. The tragedy is not a mystery in the sense of being unknown; it is a case where known warning signs did not produce a known response in time.