The emerald-colored waters off Long Branch, N.J., were "gloomy and spooky" as Dan Lieb swam toward the two hulking silhouettes, sitting upright and side by side about 90 feet down.
The objects were heavily encrusted with marine life, but Mr. Lieb recognized the unmistakable lines, the wheels and boilers of identical locomotives, 160 years after they fell or were cast overboard.
"It looked like they were steaming across the bottom in a race," said Mr. Lieb, 56, of Neptune, N.J. "You could imagine them on tracks at a station with steam coming out of the valves, and people and luggage on the platform."
Five miles off the Jersey Shore, their presence is a mystery perplexing researchers. How did two pre-Civil War locomotives wind up there? Did they slip off a sailing ship during a storm? Were they purposely dropped into the deep?
Mr. Lieb, a technical illustrator and diver, is director of the Sunken Locomotives Project for the New Jersey Museum of Transportation, a nonprofit educational organization that took legal possession of the engines -- through a federal proceeding -- about nine years ago.
Research into the submerged locomotives also is being conducted by the New Jersey Historical Divers Association, said Mr. Lieb, president of the group.
"We had the option of bringing them up, leaving them on the bottom, or bringing up parts of them," he said. "We've been recovering pieces, examining them and trying to answer lingering questions."
The 15-ton locomotives, he said, are rare because they are Planet Class 2-2-2 T models, similar to one called the Pioneer now on display at the Baltimore & Ohio Railroad Museum in Baltimore.
Like the Pioneer, each one has six wheels, a pair of larger wheels flanked by smaller ones in the front and back, but the sunken engines are heavier and more powerful, what Mr. Lieb called a possible "missing link in locomotive technology."
"It's like the designers were pushing a specific design as far as it could go," he said. "There is no record of them being built and no journal or newspaper discussing their loss."
One of the few clues to their origin came from Paul Hepler, captain of the charter boat Venture III, who discovered the engines while mapping the ocean bottom with a magnetometer in 1985.
"I got a big mag reading and dove on it," said Mr. Hepler, 71. "I didn't know what it was at first because the water was dirty and the visibility was so bad back then.
"Once I got a better look at it in later dives, I could see they were locomotives," said Mr. Hepler, who retrieved the bronze bells and whistles.
Based on the markings and other information, researchers believed the engines were likely being shipped from a factory in Boston to a port in the Mid-Atlantic states, said Mr. Lieb, who had been invited by Mr. Hepler to dive on the site and has been back at least 20 times.
"They may have been dropped there, probably in order to shed weight from a vessel that was stricken in foul weather, or they might have gone over accidentally," he said.
Other pairs of locomotives were lost on the Great Lakes, on the East and West coasts, in the Gulf of Mexico and on major rivers, according to 19th century newspaper accounts.
But few have been as old or exhibited such unusual features.
"These were each 15 tons, but you could get a 35-ton locomotive at that time," Mr. Lieb said. "That's what's intriguing to us, because [the designers] were pushing these little locomotives as far as they could go.
"They are solid and beefy," he said, carried their own water and firewood and had no need of tender cars.
The locomotives "are of a type that didn't get saved for posterity," said Nick Zmijewski, collection manager for archives at the Railroad Museum of Pennsylvania in Strasburg. "They represent a locomotive whose glory days were fading when they were built" in the 1850s.
"This was the last gasp of that design before four driving wheels and four guide wheels became the standard."
The first up-close glimpse of the engines mesmerized Mr. Lieb in 2002.
"I'm a trained technical illustrator and looked at the locomotives as a drawing on paper," he said. "That's how clear the components were. You see what the engineer and designer had in mind. The mechanical function came through the veneer of growth, incrustation, and rust."