Standing in waist-deep, murky water, Laura Murray raked the bottom of the creek, hoping to reap the reward of several years of work.

At last she found part of what she was looking for, some sprigs of sago pondweed. But there was no trace whatsoever of the redhead grass that she and colleagues from the Maryland Center for Environmental Science's Horn Point Laboratory had planted in Irish Creek.

"Before we planted," Murray noted, "there weren't either of the species in that bed."

A bed of widgeon grass already existed in the creek, located near the mouth of the Choptank River, but had shrunk dramatically in recent years. By bolstering it with new species, the scientists hoped to stabilize, and eventually grow, the grass bed.

It was one of the final projects supported through a multi-year research program aimed at finding ways to accelerate large- scale underwater grass restoration efforts.

And the Irish Creek experience, in a way, summed up the results of the projects funded between 2003 and 2007-scattered signs of success but also many failures.

Over four years, the Army Corps of Engineers' Research and Development Center and the National Oceanic and Atmospheric Administration's Chesapeake Bay Office funded more than $5 million in projects to help the Bay Program meet a goal of planting 1,000 acres of underwater grasses by 2008.

Efforts to plant underwater grasses in the Bay date to 1983, but in the 21 years prior to the research project, just 189 acres were planted-a rate of nine acres a year.

Most of those projects involved transplanting grasses from healthy beds to barren areas by hand, a labor- intensive job which often pushed restoration costs into the tens of thousands of dollars per acre. And, more than 90 percent of those projects failed within three years.

"We realized pretty quickly that we needed to be going in another direction and we needed methods that were more appropriate for large-scale restoration," said Deborah Shafer, a research marine biologist who led the project for the Corps.

The hope was to find an "agricultural approach" to underwater grass restoration. Instead of hand planting individual plants, most of the research was focused on finding ways to jump-start beds of eelgrass, a critical species in high-salinity areas, by using seeds.

Boosted by the project, the restoration rate jumped to 33 acres a year. Nonetheless, a recent report from the Corps and NOAA concludes that the 1,000-acre goal will be missed by a wide mark. Just 133 acres of underwater grasses were planted from 2003 through 2006. And that figure overstates progress, as most of the restoration projects failed to sustain plants over multiple years.

Nonetheless, those projects may have planted the seeds for future restoration. "I think we have made major steps in the past five years," Shafer said. "It is pretty phenomenal."

The Bay today has only about a third of the 185,000 acres of grass beds that scientists say it should support. While water quality is the main impediment to recovery, improved restoration techniques are needed to establish beds in barren areas where natural expansion is unlikely.

Underwater grasses are considered a key component of the Chesapeake and many other coastal ecosystems. They provide food and shelter for a variety of waterfowl, fish and shellfish. They help to filter pollutants out of the water and buffer shorelines from the erosive effects of waves.

Despite their importance, scientists say seagrasses are rapidly disappearing not only in the Bay, but around the globe as coastal waters become more polluted. Yet seagrasses have received only a fraction of the attention given to other coastal ecosystems, such as coral reefs.

The $5 million Bay project was the largest multi-agency research effort ever directed toward underwater grasses in the nation.

But researchers quickly realized how little they knew. They recognized that, as in agriculture, any large-scale restoration effort would need to rely on seeds to produce crops of grass. Unlike agriculture, no one was exactly sure how to deal with seeds. Various harvesting techniques had to be tested, including collecting seeds by hand and harvesting the seed-bearing portion of plants with mechanized equipment.

Harvesting was done in the spring, but eelgrass seeds don't actually germinate until the fall. Scientists tested techniques that distributed seeds in the spring, and tried storing seeds for distribution in the fall.

Storing seeds required its own research: Early in the work, millions of stored seeds died. Various temperatures, oxygen levels, salinities and other variables were tested to determine which yielded the best odds of survival.

More than 16 million eelgrass seeds were ultimately distributed at various sites at densities that ranged from 40,000 to 1 million seeds per acre.

About 1 percent of the distributed seeds took root and produced a plant. "A lot of people look at that number and say that's not very good," Shafer said. "I don't agree at all."

Instead, she noted, it shows that when handled properly, scientists can anticipate that 750,000 seeds would produce 7,500 plants per acre. That's a higher success rate than some research suggests seeds have in the wild, and it could produce the start of a healthy bed.

Unfortunately, that generally wasn't the case. Most of the beds planted in the Bay failed to survive over time, though. Eelgrass beds planted near Cherryfield Point and St. George Island in the Potomac River survived and spread, and are now visible in the annual aerial survey of underwater grass beds.

"That shows you that the techniques work," said Peter Bergstrom, a fisheries biologist who coordinated the project for NOAA's Bay Office. "But how do you get it to work more often?"

The most stunning example of planting success is in the coastal bays off Virginia's coast, where Bob Orth, an SAV expert from the Virginia Institute of Marine Science, has produced a thriving bed with the same techniques used in the Chesapeake.

A barren area was planted with eelgrass seeds from 2001 through 2004. It has developed into a lush bed covering 1,000 acres, which is easily visible from the air. In recent years, new satellite eelgrass beds have began taking root in nearby areas.

"The coastal bays are being planted and done with the same methods we use, and they are very successful," Shafer said.

What's different in the Chesapeake?

Shafer and Bergstrom believe water clarity, critical for underwater plants to receive adequate sunlight, tends to be poorer in the Chesapeake. And recent trends have shown worsening conditions in much of the Bay. In addition, eelgrass is sensitive to warm water, and monitoring has shown an increasing trend in Bay temperatures.

"There was some feeling when we started this in 2003 that the Bay was getting better so we would have more success," Bergstrom said. "That was before we knew the clarity was getting worse, or we knew the Bay was getting warmer."

Also, planting sites were selected based on water quality conditions observed near healthy underwater grass beds. It appears likely that establishing a new bed requires better conditions than those needed to maintain an established bed.

"If you have the wrong site, it doesn't matter how you plant it." Bergstrom said.

Bergstrom and Shafer concluded in their report that new research may be needed to refine conditions required for establishing grass beds.

They also said more work needs to be focused on grasses in the mid portion of the Bay. Much of the project was aimed at eelgrass because it is the only species that thrives in high salinity portions of the Bay and is therefore considered especially important.

But the medium salinity water in the mid-Bay holds more than 60 percent of the Chesapeake's potential underwater grass habitat. Historically, it supported several varieties of underwater grass. Today, the region is dominated by widgeon grass, a species notorious for its wide year-to-year fluctuations as water conditions change.

The widgeon grass beds in the Lower Choptank, for instance, peaked at more than 6,000 acres in 1997 and again in 2002. But in more recent years, the beds have shrunk to less than a quarter of that size or less.

Murray and her colleagues wanted to try stabilizing the beds by mixing in sago pondweed and redhead grass, hoping that increased plant diversity would make the bed better able to withstand changing conditions.

As with the eelgrass projects, it was no easy task. So little was known about the species, it was unclear whether it was best to move entire plants, or spur growth with seeds or plant fragments, and researchers tested various approaches both in the wild, and in ponds.

Results are mixed. Redhead grass didn't take root at all; but scattered patches of sago pondweed have become established and begun expanding. In the summer, Murray said, the mixed species bed is lush and thick. "You couldn't get a boat through it," she said

Other techniques need to be tried in the mid-Bay, Shafer said. One idea is spreading seeds from multiple species at the same time to improve diversity, as in Irish Creek.

"Each of those species responds to a little different salinity and temperature, which is what you want," she said. "The conditions are so variable, you want to hedge your bet, and put out a variety of plants so one can grow if it has the right conditions."

Ultimately, she said, such work may someday help underwater grasses to be planted with the reliability of a corn field. "Literally," Shafer said, "what we are trying to do is farm an underwater crop."