On December 5, 1872, the British brigantine Dei Gratia, sailing from New York to Gibraltar, sighted another vessel acting strangely in the mid-Atlantic. She was making erratic course corrections, her sails partly set, partly furled. When the Dei Gratia closed and her boarding party climbed aboard, they found the ship deserted.

The vessel was the Mary Celeste — not Marie, though the misspelling has stuck in popular accounts for 150 years. She had left New York Harbor on November 7 bound for Genoa with a cargo of 1,701 barrels of denatured alcohol. Her captain was Benjamin Briggs, an experienced master with a reputation for sobriety and good seamanship. Aboard were his wife Sarah, their two-year-old daughter Sophia, and a crew of seven.

The boarding party found the Mary Celeste in seaworthy condition. Her cargo was intact. There was food in the galley and personal effects in the cabins. The ship's single lifeboat was missing, along with the chronometer, the sextant, and the ship's papers, except for the captain's logbook. The last log entry was dated November 25, ten days before she was found, when she was near the Azores.

What had happened in those ten days has been argued over ever since.

The pirate theory does not survive scrutiny. Pirates would have taken the cargo. The mutiny theory does not survive either. There were no signs of violence aboard, and the crew were not known to be discontented.

The early theories were lurid. Pirates. Mutiny. Sea monsters. A salvage hearing in Gibraltar in 1873, presided over by the British Attorney General Frederick Solly-Flood, treated the case as possible foul play and suggested the Dei Gratia's crew might have murdered the Mary Celeste's people for salvage money. The investigation found no evidence and awarded the salvors a payment well below the usual rate, suggesting the court harbored suspicions it could not prove.

The pirate theory does not survive scrutiny. Pirates would have taken the cargo. The mutiny theory does not survive either. There were no signs of violence aboard, and the crew were not known to be discontented.

The most persistent serious hypothesis involves the cargo. Denatured alcohol in wooden barrels expands and contracts with temperature. Nine of the 1,701 barrels were later found to be empty when the cargo was unloaded in Genoa. These nine, unlike the others, were made of red oak, which is more porous than the white oak used for the rest.

Maritime historian Stanley Spicer, whose 1993 book on the case remains a standard reference, proposed that alcohol vapor leaking from the red oak barrels could have built up in the hold. A spark, or even a sudden release of pressure when a hatch was opened, could have produced a flash explosion — loud and frightening, but not necessarily damaging to the ship's structure. A panicked captain might have ordered the crew into the lifeboat, intending to remain at a safe distance until the danger passed, attached by a tow line.

If that line then parted in the rough seas around the Azores, the lifeboat and ten souls would have been lost, and the Mary Celeste would have continued on under partial sail until found.

Recent work has strengthened this hypothesis. In 2006, the chemist Andrea Sella at University College London demonstrated that vaporized alcohol can produce a flash burn that is loud and bright but leaves no lasting scorch marks — exactly what the boarding party reported finding. The forward hatch on the Mary Celeste was found open and partly displaced, consistent with a pressure release from below.

Other researchers have proposed alternatives. The historian Charles Edey Fay, writing in 1942, suggested a malfunctioning chronometer led Briggs to believe he was nearer to dangerous shoals than he actually was, prompting an unnecessary evacuation. Marine geologist Jean Bruneau has suggested a waterspout or seaquake.

None of these explanations is perfectly satisfying, and the gaps in the evidence may never be closed. The Mary Celeste continued in service after 1872, passing through 17 owners over 13 years, gaining a reputation among sailors as cursed. In 1885 her final owner, Captain Gilman Parker, deliberately ran her aground on a reef off Haiti in an insurance fraud scheme. He was caught. The ship broke up.

Her wreck was relocated in 2001 by an expedition led by the writer Clive Cussler off the Rochelois Bank. The remains were modest — copper sheathing, a few timbers — and added little to what was already known.

What the case demonstrates, more than anything, is how a small set of strange facts can sustain enormous interpretive weight. Ten people left a sound ship in good weather and never came back. The likeliest explanation is mundane and tragic. The mystery endures partly because the mundane explanation cannot be proved, and partly because the alternative — that we will never know — is, for many readers, harder to accept than the strangest theory.