Most people think of Hong Kong waters as polluted, full of plastic bags and old cars, with dead, black mud at the bottom. They're wrong.
In fact, the eastern coast of the territory is dotted with communities of life and colour.
Sweeper fish, rabbit fish, orange clown fish and striped butterfly fish mingle with the romantically-named sweetlips fish, weaving among fronds of sea-weed and bright red corals with names like gorgonian sea fan and Dendronephthya.
Moray eels and lion fish, damsel fish and 'anchovies by the million' have been spotted in some parts.
Starfish with names such as pentaceraster orientalis cling to rocks between lawns of sea anemones.
Sometimes an eerie translucent cuttle fish swims silently by; even sting-rays appear occasionally.
'I show photos to people and they say 'is that the Philippines, or Indonesia?' They're amazed when I say Hong Kong, they have no idea,' said diver and chief environmental scientist of Binnie Consultants, Gregor Hodgson.
Mr Hodgson has carried out a four-year government-funded study to map all the corals and look at the seabed to study the effects of dredging.
The aim was to find out what was there to gather information for environmental impact studies for dredging the sand from old river-beds to form huge reclamation areas like the airport and West Kowloon.
The work was undertaken by the Geotechnical Engineering Office (GEO), better known for its work on landslides than involvement below the waterline.
'Binnies were faced with a bit of a black hole because very little was known,' said Peter Whiteside, secretary of GEO's fill management committee.
It began partly because of a marine disaster: dredging very close to the shoreline of one Ninepins island in 1991 resulted in layers of fine sediment that led to a grey graveyard of dead table coral typically a metre or two wide.
'Table corals are 60 to 100 years old,' said Mr Hodgson. 'The effect [of the sediment] is very obvious and dramatic, there was a lot of damage, and you are not going to be able to grow another for 50 years.' But soft corals were unaffected and hard corals were returning. Eastern Ninepins sites were still good, he said.
Another environmental study of Mirs Bay led to it being declared off-limits as a dredging site.
It was partly at the urging of Government geotechnical engineer and former Perth diver John Massey, after green group objections and initial Binnie findings of corals worth saving. 'That was an historic event as it was the first time a major Hong Kong development project was cancelled largely due to environmental concerns,' said Mr Hodgson.
Binnie surveyed more than 100 reef sites from Lamma to Sha Tau Kok, producing more than 10,000 slides and categorising them according to conservation value.
Now the divers are moving westwards to the Soko Islands, where some corals have been spotted, but the sediment-laden Pearl River means fewer corals are expected further west than that.
Breaker Reef in the north-east is 'by far the most valuable site in Hong Kong', including semi-precious black and soft corals, he said. 'There's no question that it should be a marine park.' Marine parks are planned so far at Hoi Ha Wan at the mouth of Tolo Harbour and Cape D'Aguilar, on the southern tip of Hong Kong Island. But they are not the best places, he says.
Other good sites in terms of diversity and abundance are Double Haven and Ping Chau, says Mr Hodgson.
He is loathe to say where the best fish are because he fears fishermen targeting them, but his favourites are the damsels, the lion fish and butterflies.
Even some two-metre garoupa had been spotted, but most had been fished out.
'By the end of last year we reached the stage of having collected ecological, biological and oceanographic information to get a reasonable idea of what was there,' said Mr Whiteside. As for dredge sites, little impact had been found on the environment.
And in some parts where the holes had been filled, more species were found than before, he said.
This month a final contract is being let to set up 10 monitoring stations for studies of how the corals change over a one-year cycle.
Now green groups are concerned that the work should develop into a full management and monitoring plan for sea and land, to co-ordinate development plans way into the future.
'They're doing that all over the world,' said Mr Hodgson.
For instance, the huge surge of cold, low-oxygen water into Mirs Bay in 1994 which killed many of the corals there might happen on a regular basis, but no one has any certainty.
'It might happen every year but we don't know because we're not watching,' he said.
'There's a lot worth saving . . . the importance of establishing a long-term monitoring programme cannot be over-emphasised.'