Kirkpatrick Marsh looks a picture of environmental health, its waters fringed by waving grass and swarming with fat minnows.
But seeping out of these Anne Arundel wetlands into the Chesapeake Bay is a pollutant - methylmercury - that causes brain damage in people.
"It's quite crazy," says Carl Mitchell, a scientist at the Smithsonian Environmental Research Center in Edgewater. "These areas are so beautiful and pristine looking, but they also produce a lot of methylmercury."
Mitchell is among a growing number of researchers documenting the little-known role that wetlands play in transforming air pollution from coal-burning power plants into a form of mercury that contaminates fish.
As Maryland and other states look to build thousands of acres of wetlands to fight global warming, the research has significant implications. More wetlands would absorb more carbon dioxide, but they also could make mercury health hazards worse.
"Building wetlands has many positive benefits for the ecosystem and society at large," said Mark Marvin-DiPasquale, a microbial ecologist at the U.S. Geological Survey. "But if you increase the amount of wetlands by fivefold, does that mean you get five times more mercury?"
Mitchell is trying to figure this out as he studies marshes outside Edgewater, south of Annapolis.
Mercury isn't created by marshes. The element occurs naturally in rocks, dirt and coal, and it floats into the air when coal is burned in power plants and factories. This airborne form of the metal, called inorganic mercury, doesn't usually get into people, Mitchell said.
The metal becomes a health hazard when it's changed by microorganisms that multiply in the oxygen-deprived muck at the bottoms of rivers, lakes and wetlands. In this mud, bacteria transform inorganic mercury into methylmercury, which is at least 10 times more toxic and collects in the bodies of animals.
Tainted worms are gobbled by small fish, and these are eaten by big fish - such as tuna and shark - that are eventually eaten by humans. The methylmercury accumulates as it moves up the food chain, with the toxicity multiplying 10-fold at each step.
It's ironic that marshes appear to be a vital link in this toxic chain. Wetlands are normally regarded as environmental cornucopias, fountains of life for fish and birds. They clean the Chesapeake Bay by filtering out silt and other pollutants. And marshes absorb the carbon dioxide that is warming the atmosphere.
The ability of wetlands to sequester carbon and nurture wildlife is the reason a construction project of about 12,000 acres is being considered at the Blackwater National Wildlife Refuge on Maryland's Eastern Shore. A similar wetlands construction proposal in the San Francisco Bay area has been delayed for study into whether the new marshes would release significant mercury.
The issue is whether the amount of mercury flowing out of wetlands is small enough to be a minor health risk compared with the marshes' many positive qualities.
"It's a double-edged sword," Brian Branfireun, a hydrologist at the University of Toronto, said of wetlands construction. "There is a benefit to creating more wetlands habitat. But then there is a question of if you're creating a more serious mercury problem."
Mitchell is not suggesting paving over wetlands to improve public health. Rather, he said, society needs to stop spewing mercury into the air through the burning of coal. "The best away to get mercury out of fish is to stop putting mercury into the environment," he said.
But if the country is going to keep burning coal - the source of half of U.S. electricity - steps could be taken in marshes to reduce the production of methylmercury, Mitchell said. Adding activated charcoal to absorb the mercury in newly constructed wetlands might reduce the release of the contaminant, he said.
Other scientists have suggested building gravel beds under artificial wetlands to increase the flow of oxygen through the sediment, which could discourage the bacteria that create methylmercury. Or reeds and cattails could be planted instead of leafy plants like lily pads, which produce more bacteria when they die.
Mitchell said natural wetlands - such as Kirkpatrick Marsh, owned by the Smithsonian research center - should be left alone so the wildlife is not disturbed.
But to the east, on the other side of the Bay Bridge, the federal and state governments are studying a vast wetlands-building proposal at the Blackwater refuge. Funding for this effort could come in part from power companies that face new state greenhouse gas limits and want to build wetlands as a way of "offsetting" their pollution.
The Maryland Department of Natural Resources will look carefully at the mercury-producing potential of the proposed Blackwater wetlands as they are designed, said John Sherwell, manager of DNR's power plant research program.
"It's another wrinkle that needs to be looked at," Sherwell said. "We are pretty sensitized to the fact that you can get [methylmercury production] in marshes."
Mitchell is not the first to document the role of swamp bacteria in producing methylmercury. The link was discovered about two decades ago, in part by Cindy Gilmour, his supervisor at the Smithsonian research center.
Gilmour said Mitchell's new work is important for calculating precisely how much mercury is oozing from the marshes. That's necessary so researchers can judge the health risk, she said.
In one four-month period, from August 2007 through November 2007, Mitchell measured 27 mg of methylmercury flowing out of one tiny creek in Kirkpatrick Marsh. That's enough to contaminate more than 100 fish above Environmental Protection Agency health advisory limits, he said.
"It's not a trivial matter. The EPA estimates that up to 10 percent of women of child-bearing age eat enough mercury in fish that they are at risk," said Gilmour, a biogeochemist. "Mercury has developmental impacts on babies' brains, especially declines in IQ."
Because of mercury contamination, the federal government advises women of childbearing age and children not to eat more than six ounces of albacore tuna per week, and no shark, swordfish, king mackerel or tilefish.
On a recent morning, Mitchell plunged into the marsh wearing tall rubber boots, a pair of laptop computers slung over his shoulders. He sloshed between curtains of tall reeds.
Then he teetered across a pond on an eight-inch-wide wooden bridge that's rotten and wobbly and lacks hand railings. "If you're going to fall, fall to your left - not your right. It's deeper there," he warned, balancing his computers.
From a machine stationed on a platform in the swamp, he retrieved 12 bottles of water. Each was a sample the device had drawn from the creek over the previous 12 hours. He would tote them back to his lab to analyze their mercury content.
As he linked his laptop to another device measuring water speed and volume, an otter paddled up, as if curious about his findings. Mitchell delivered this sobering report: "There is certainly more methylmercury coming out of this marsh than going in."