His research is financed by a grant from the South African lottery board, which sometimes uses profits to back heritage projects.

He describes himself affably as "a one-man show." Major potential donors, like the Dutch government, want to see proof of the Meermin's remains before signing on to his project, he said. Eventually he hopes to generate enough interest to search for other slave wrecks in the vicinity, including the French ship Jardinière, which sank 28 years after the Meermin in the same bay.

 "This is a long-term thing," he said. "Twenty years, plus."

Mr. Boshoff's field reports dutifully note both his struggles and his strides. His first efforts to find the Meermin involved lashing a magnetometer to a foam surfboard and towing it down the beach behind a 4-by-4 vehicle. The next day, he hammered a six-and-a-half-foot pole into the sand at the points where the device had recorded the presence of metal, hoping to hit something.

"This was a singularly unsuccessful exercise," reads his report of November 2002. Later he determined the device was too rudimentary to register the presence of any metal deeper than five feet or provide accurate readings. 

Subsequent surveys from the air and on the ground went better. Last week Anglo-American, the mining and natural resources conglomerate, lent Mr. Boshoff a $40,000 magnetometer and the services of Albert Mandobe, a field assistant. For days Mr. Mandobe painstakingly paced the beach with a nine-foot metal pole strapped to his shoulder while the device recorded each magnetic abnormality as precisely as an electrocardiogram records heartbeats.

 

The magnetometer's measurements will be used to design a contour map that Mr. Boshoff says will guide excavation. Three sites look promising. The fourth, Mr. Boshoff has concluded, is probably the buried fence.

Success, he predicted, would ultimately be a combination of patience, skill and resources. Plus, he added, a measure of "blind luck."