The hotel was very charming, with neatly cut lawns, a long sandy beach, and white buildings with a restaurant and a bar under a beautifully crafted thatched roof of palm fronds. The only drawback is that the coral comes up to within a few feet of the beach, and though it makes for great snorkeling, you’d scrape your chest off if you tried to swim in it. The food was gourmet quality. Seven different lobster dishes cooked in different ways and tinged in exotic sauces were only a small part of the menu. Then on to the bar for after-dinner drinks and hours of telling tales of shipwrecks and the people who search for them.

The boat showed up the next day. Owned by Allan Gardner of Highland Beach, Florida, Ella Warley II is a fifty-four-foot steel-hulled vessel specially designed by Allan for underwater research. She carries the latest dive and detection gear with state-of-the-art electronics. Allan is a very successful businessman who owns a large computer technology company. When not directing his corporate empire, he spends his time searching for shipwrecks in the Caribbean. My kind of guy.

Allan is a truly nice guy with the patience of Job who smiles constantly and, after a few shots of scotch, laughs continuously. Joining him on the voyage from Fort Lauderdale was the team from ECO-NOVA, including Mike Fletcher, master diver, and underwater photographers Robert Guertin and Lawrence Taylor — all friendly and good-hearted guys with enough esprit de corps to turn a grueling expedition into a proud moment of achievement and success.

John and I boarded a boat from a dock in a lagoon used by Carnival Cruise Lines ships to send their passengers to a tropical cove to enjoy a day of sun and surf. Like Allan, Jean Claude generously gave his time to come along, because he knew Haiti and could converse with the natives in Creole — a handy arrangement, as it turned out.

The voyage to Rochelais Reef began early the next morning. The seas were very rough, but I fortified myself with a bottle of Porfido tequila I had brought along for just such an occasion. Though a fairly stable boat, Ella Warley II rolled with the punches. With her flat bottom, she makes the perfect dive platform for underwater survey, but she’s not exactly what you’d call a luxury yacht. She was built for a purpose without the niceties of plush furnishings, a deep keel, or stabilizers. Plus she suffered from heads that always clogged when you flushed them.

Sleeping accommodations were austere. Allan, as owner, had the only stateroom. Two of the team slept in bunks in the wheelhouse. Two slept on the deck outside. Jean Claude and I shared the main cabin, me on a small foldout couch, he on the bench of the dining table.

We were both outcasts from the others because we snored. Jean Claude began from ten to two, while I took up the trumpet calls from two to six. Now I know what my poor wife goes through.

But our seven-man crew was tough. No one ever suffered seasickness or complained, except me.

We anchored for the night on the northwest tip of Haiti. The next morning, we sailed around Gonâve Island and reached Rochelais Reef by midmorning. As we approached, I was peering through binoculars into the distance where the reef was supposed to rise. An image materialized, and I adjusted the focus.

I turned to Allan and John and said, “If I didn’t know better, I’d say there is a village with huts sitting on the reef.”

Forty-five minutes later, we reached Rochelais Reef and anchored a hundred yards offshore. This place was an anthropologist’s dream. The story is that about eighty years ago, two brothers decided to take up residence on the reef to hunt conch. Over the decades, more native Haitians moved onto the reef, until now it is an island four feet above the water, built from more than a million conch shells. There are about fifty shacks erected from every scrap of flotsam you can imagine. We estimated the population at about two hundred. There wasn’t a tree or a bush to be seen. The sun beat down unmercifully on the conch-shell landscape. The nearest land was twelve miles away, and all the food and water had to be brought in by dugout canoes. We could not believe human beings could survive in such harsh conditions, much less spend their entire lives there.

John and his film crew, along with Jean Claude, took the thirteen-foot Boston Whaler over to the man-made island sitting atop the shallow reef. Naturally, the natives were curious about our presence. Jean Claude did not tell them we were looking for a shipwreck. They might have mistaken our intent and thought we were after treasure, which could have caused problems. He simply told them we were making a movie and pacified them about our intrusion by giving them ten gallons of gas for their outboard motors and a case of Coca-Cola.

We had studied the oldest accurate chart of the reef dating to 1910 and laid it over a modem chart. The reef had not changed. According to both charts, there was a pinnacle labeled “Vandalia Rock” on the southern end of the reef, but the natives assured us that no such rock existed. This would cause extra time in research later, exploring the possibility that a ship named Vandalia had also grounded on the reef.

The wind picked up, and Allan took the boat into a small bay on Gonâve Island, where we settled in for the night. With an early start the next morning, Allan dropped his cesium marine magnetometer over the side and began circling the reef for any magnetic anomalies. From fifty yards offshore, only one ten-gamma reading showed on his computer monitor. It came from the dead center of Rochelais Reef, where Mary Celeste was reported to have crushed her hull on the coral. The site also perfectly matched the direction a ship sailing from the southwest would have met the reef.

My demon must have taken a break.

Mike Fletcher suited up and dropped over the side, followed by Robert Guertin with his underwater video camera. The rest of us sat on the stem of the boat, soaking up the tropical breeze, wondering if Mike had found anything. Half an hour later, he returned to the boat and threw some copper sheathing, ballast rocks, and old wood with brass spikes driven through them onto the deck.

We had a shipwreck right where Mary Celeste was supposed to lie. But without finding the bell, which was no doubt salvaged, with the ship’s name in raised letters in bronze or inscribed ceramics, or some other artifact to identify her, we could only speculate. Everyone dove and retrieved what pitifully few artifacts we could find. The coral was some of the most beautiful I’ve seen in fifty years of diving, but I wished I weren’t there. What remained of the ship’s timbers was deeply buried in the calcareous growth that had rapidly buried the ship. After 116 years, the ship was entombed in an impenetrable burial shroud.

We found part of the anchor chain and an anchor. I tried to remove a bar from the coral, but it was stuck fast. Jean Claude and Mike brought up enough wood to fill a fair-sized bucket. And we removed some loose artifacts that were embedded in the sand. Every item was videotaped in position, tagged, and catalogued. Once the team returned home, the wood would be sent to laboratories to determine a date and source. It’s incredible how science and technology can tell you how old the wood is within years, as well as what part of the world it originally came from.

The ballast stones would also show characteristic mineralogy and texture that can identify the location from which they were extracted. They had to come from either the Palisades above the Hudson River, where Mary Celeste was rebuilt during the summer of 1872 in New York, or the mountains or shores of Nova Scotia, where she was originally constructed and then launched under the name Amazon. The brass spikes might give only an approximate age, but a clue might come from the copper sheathing. Whatever the case, it would take time to find the answers we sought.