That said, the company, chaired by Christopher Rowe and run by Mark Gleave, the managing director, and John Kingsford, the operations director, is keen to stress the prosaic aspects of its business.

``We are not going after Spanish treasure cargoes,'' says Gleave, before hinting that actually they might. ``What we have sought to do is reassure investors that this is a very real business.''

The focus is on the recovery of high-value non-ferrous metals and other valuable cargoes from deep-water shipwrecks.

Instead of Spanish galleons, SubSea is concentrating on valuable bulk cargoes of metals in modern, steel-hulled wrecks that have been lost in international waters over the past couple of centuries. The company likes to establish clear ownership of the cargo and agrees on a salvage arrangement with the relevant insurer, which typically gets 90 percent of the value of metals recovered.

SubSea is using techniques developed by the offshore oil and gas industry in which Kingsford spent many years as a deepwater diver. Modern methods allow recovery of ships at depths well below the 300 meters that has, until recently, been the cut-off point for salvage operations.

``You are still on the beach at 300 meters,'' says Kingsford, who points out that only about 1 percent of the world's oceans are less than 300 meters deep and that most wrecks are in much deeper waters - of up to 4,000m.

Several projects have already been identified and agreement has just been reached with the underwriters of a wreck, codenamed Ella. The ship lies in relatively shallow Atlantic waters and carries a small 19th-century cargo of gold and silver bullion owned by a London-based insurer. Salvaging Ella is expected to cost US$2 million (HK$15.6 million), but revenues should be US$6 million.

This project will be used by SubSea as a trial run to test its equipment, personnel and procedures in a relatively low-risk environment and to carry out sea trials on a new survey and salvage vessel, Northern Horizon, which the company acquired last month.

The recovery of Ella, due for July, is to prepare for another early scheme, codenamed Celia, which contains a cargo of copper with a net worth of US$12 million to SubSea. The company estimates it will cost US$6 million to recover, leaving it with a clear profit.

SubSea disguises the putative sunken treasure with codenames to outfox rivals. And following the Ella and Celia cargo recoveries, it is planning a further five salvage projects over the next three years, with a total estimated gross cargo value of about US$200 million.

The first step in this business is research, Gleave says. This can include looking at records of shipping lanes, for example voyages from the mines of South Africa to the United Kingdom or from the copper mines of Peru to the United States. SubSea also looks at records of U-boat activity in 1941.

``The essential point is that we've got to see the original document,'' Gleave says. ``We need to check that the document comes from a collection of contemporary archives; we need to be as certain as we possibly could be - so that we could, if necessary, take our auditors to the archive and show them the lading documents or the insurance payout.''

He also says he has to be on his guard against charlatans: ``There are people out there who falsify information and sell it and some people have been badly stung by that.''

Gleave also points out that it is important to ensure that a seemingly promising target actually fits the history. ``There was a P&O ship coming back from South Africa in World War I with 45 tonnes of gold on board. There was not a shadow of a doubt that the gold was loaded and the boat got torpedoed off the Scilly Isles,'' he says, by way of example.

However, when a salvage crew found the boat 20 years ago and spent a large sum of money opening up the hull, they did not find the expected US$700 million worth of gold. Gleave says they should have known that in that war, gold mined in South Africa was sent to the United States and Canada and that, in this case, it had been transferred to a warship off Sierra Leone.

``Somebody wasted millions of pounds mounting the salvage operation because they didn't do the research,'' he says.

Once research has identified targets - SubSea reckons to have some 70 potential wrecks on its database, whose assets are supposedly worth US$1 billion - the next phase is to carry out a scan of the likely location using specialist sonar, which builds up a picture of the seabed. When the wreck is found, remotely controlled submersibles are used to take a look.

``Once a target is identified, we use a remote-access vehicle to survey the target, which costs up to US$500,000, depending on a number of variables, but once we've spent that money we can be very certain of the potential success of the next stage.''

The survey is designed to produce enough information for a team of specialist engineers to prepare a competitive tender for a fully priced recovery operation. Depending on the scale of the operation, SubSea is planning to use contractors - mainly from the oil and gas industry - to carry out the actual recovery.

Gleave says the company won't totally ignore treasure ships.

``It is impossible to spend 25 years in the archives without becoming aware of some of the bigger, more exciting historic cargoes, but the scale of the returns is so big that investors think you are taking the p***,'' he says.

An example is Nuestra Senora de Atocha, a Spanish treasure ship that sank in a hurricane off the Florida Keys in 1622. It was carrying 47 troy tonnes of Peruvian gold. The wreck was found in 1985 and has yielded US$500 million worth of gold and jewelry.

``Those things do exist and we know of them - but you can't start putting them on business plans or it produces numbers that people don't believe,'' says Gleave. ``But while searching for more prosaic targets, we can also sweep nearby for a more exotic target and we will be doing that.''

He says SubSea has identified some such targets but they will be a bonus. ``We know they are there and if they can be looked at for nil cost, we will do it, but the plan is to make solid, not spectacular, returns to investors,'' he adds.

``We are targeting US$60 million of revenue every year. The boring bit is fine, we can retire very comfortably on the boring bit.''

THE SUNDAY TELEGRAPH