The Marmes Rockshelter was one of the most significant archaeological sites in the Pacific Northwest, yielding thousands of Stone Age artifacts -- along with the oldest human remains yet to be found in Washington state -- before it was inundated by the reservoir behind Lower Monumental Dam on the Snake River in southeastern Washington. As the dam neared completion in the fall of 1968, Washington Senator Warren G. Magnuson (1905-1989) used his political clout to secure an emergency appropriation to build a horseshoe-shaped enclosure around the rockshelter and an adjacent floodplain, hoping it would keep the area dry enough for archaeologists to continue their work. Unfortunately, the enclosure, built on a gravel base, filled as quickly as the main reservoir. A team of scientists from Washington State University hurriedly covered what they could with plastic and sand and then watched helplessly as the site disappeared beneath 40 feet of water in February 1969.
Evidence collected before the site had to be abandoned showed that it had been used for shelter, storage, and burials for more than 11,000 years. Among the most compelling finds was a 10,000-year-old cremation hearth. The remains of at least 38 individuals were identified. Most had been cremated and then buried, with grave goods that included shell beads, bear teeth, and, in one case, the hoop from an infant’s cradleboard. The oldest of the human bones came from a man who died about 10,000 years ago -- 700 years before the famed Kennewick Man, whose ancient skeleton was found on the bank of the Columbia River near Kennewick in 1996. The excavators also found a wealth of tools, weapons, ornaments, and other detritus of life in what is now Washington state at the end of the last Ice Age.
The loss of this valuable window into Stone Age culture in the Northwest galvanized public opinion, leading to stronger measures to protect historic and archaeological resources. "Science, public interest, and politics converged at Marmes Rockshelter," Richard D. Daugherty (1922-2014), an anthropologist who directed the excavation for WSU, wrote in the preface to a recently published report on the project. The result was a renewed commitment to preserve "discoveries about our human selves." Magnuson predicted that the Marmes Rockshelter would serve as "a landmark precedent in our nation’s responsibility to its own heritage," one that "will be felt for decades to come" (Hicks, xiv, xv).
The Ranchers and the Professor
In one sense, the story of the Marmes Rockshelter began in 1952, when rancher John McGregor of Hooper, Washington, invited Daugherty -- a newly hired associate professor at what was then Washington State College -- to look at several caves and rockshelters in the cliffs above the Palouse River, near its confluence with the Snake. Daugherty directed the excavation of one of the caves the next year. Although he described the site "productive," he did not return to the area for another 10 years.
Construction of Lower Monumental Dam began in 1961. The next year, Daugherty and his colleagues at WSU received a modest federal grant to survey the archaeological resources in the Palouse River Canyon, one of the areas to be flooded by the dam. Daugherty thought the most promising site for excavation was at the mouth of the Palouse River. While tents were being set up for the excavation team, he took two students on a field trip to look at a rockshelter on a nearby ranch owned by Roland J. Marmes. "It was immediately apparent that this was a rich site," he wrote (Hicks, xiii). He abandoned plans to investigate the mouth of the river in order to concentrate on what became known as the Marmes Rockshelter.
The rockshelter was basically an alcove, 40 feet wide and 25 feet deep, beneath an overhanging ledge of basalt. The researchers theorized that humans first discovered it while on foraging expeditions, beginning not long after the last of the great Ice Age floods had finished carving out the scablands of Eastern Washington. Evidence of consistent human use over thousands of years suggested that the rockshelter was a "tethered" site: a regional base that people returned to regularly.
In the first season of excavation, the researchers found numerous storage pits in the shelter, some lined with fragments of mats made from grasses or reeds. Cool temperatures and consistent humidity made the shelter useful as a cache for food. Numerous traces of plant foods and animal bones suggested that the site offered access to abundant food sources. Relatively few fish bones were recovered, perhaps because of the limits of the methods used by the excavators, or perhaps because fish runs had not yet recovered from the turbidity of the Ice Age floods.
By the end of the first season, the researchers had also found the partial skeletons of 11 individuals, including three (two adults and an infant) that were located in a strata of earth that had been covered by a layer of ash produced by the eruption and collapse of prehistoric Mount Mazama in southern Oregon more than 6,500 years ago. Radiocarbon dating of shells and other organic material found next to these remains indicated that they were about 8,000 years old -- among the oldest found up to that point anywhere in the West.
Despite the significance of the discoveries, the WSU team gave only cursory attention to the Marmes site during the next two years. Funding for archaeological research was extremely limited at the time, when four dams on the Snake River were in various stages of planning or construction. Eighty other archaeological sites had been identified within the area to be flooded by Lower Monumental Dam alone. Daugherty and his colleagues couldn’t justify spending all the time that was left on operations at just one site.
Fewer excavators worked at the rockshelter in 1964 than in the previous season, and most were new to the project. Record keeping was inconsistent; some work was needlessly duplicated. By modern archaeological standards, the collection methods were haphazard, due in part to the inexperience of many of the workers. Some people kept careful field notes about where they found artifacts; others did not. Most of the teams used quarter-inch screens to sift the dirt for artifacts, instead of the eighth-inch screens that are more commonly used today. Smaller objects -- such as the bones of fish -- are easily lost when sifted through larger screens.
The archaeological excavations ended at the end of the 1964 field season. However, Roald Fryxell, a young geologist and member of Daugherty’s original team, returned the next year in order to conduct further studies on the geologic strata at the rockshelter. He asked landowner Roland Marmes to use his bulldozer to cut a trench through deposits at the mouth of the rockshelter, down the talus slope in front, and onto the floodplain below. Fryxell followed the blade as Marmes carefully carved through the deposits, four inches at a time. In the floodplain, at a point about 40 feet in front of the shelter and about 12 feet below the surface, the bulldozer uncovered what were later identified as human bones.
Although Fryxell had watched the bulldozer blade expose the bones, he could not prove that they had not fallen there as a result of previous excavations. He and others revisited the site often during the next two years, but it wasn’t until April of 1968 that they discovered more bones, clearly undisturbed, and subsequently dated to being more than 10,000 years old: the oldest found up to that point anywhere in the Western Hemisphere. (More recent discoveries, in Idaho and California, predate the so-called "Marmes Man.")
Science and Sensationalism
Emergency salvage operations began immediately, as did a campaign to bring public attention -- and funding -- to the work at the site. On April 29, Fryxell and Daugherty appeared at a press conference hosted by Senator Warren G. Magnuson -- one of the most powerful people in Washington, D.C., at the time. They announced that they had discovered bones that could be as old as 13,000 years (a couple of millennia older than their actual date); that some of them had been charred and split in a way that suggested cannibalism; and that untold wonders yet to be uncovered were about to be flooded by a dam.
The story caused a sensation. "Human Fossil Held Oldest in Americas," The New York Times reported, in a story that quoted Fryxell as saying the bones belonged to "a young pre-Indian nomad whose fellow tribesmen may have ‘literally had him for dinner.’ " He said the team had found one long bone, believed to be from a human, that had been split lengthwise. "Aborigines did this to get the bone marrow to eat," he said, concluding, "It is entirely possible Marmes Man was eaten by his buddies."
By June 1968, the Marmes Rockshelter was a hive of activity, crowded with scientists, students, and onlookers. There was a palpable feeling of excitement and expectation, fueled in part by the media attention directed at the project and reinforced by a continuous stream of visitors. Field notes written at the time reflected the workers’ sense that they were in the midst of "cutting-edge science" (Hicks, 15).
A number of important new discoveries were made, including a tiny bone needle, with an eye about the diameter of a modern straight pin, obviously used for very fine stitching. It was found on the floodplain, as were other needles, awls, and small bone tools, suggesting that work requiring some delicacy was done outdoors, where the light was better.
The major discovery inside the rockshelter was the cremation hearth, found at a deep level that hadn’t been excavated earlier. The hearth consisted of a series of small rings of rock and rock piles, peppered with shards of bone, rock chips, and debitage (waste flakes resulting from the manufacture of stone tools). Virtually all of the bone was extensively burned, making it difficult to determine if it was human or animal. Adding to the difficulty was the fact that all of the bone fragments found in the hearth area were small; the largest being a piece of skull that was just two and a half inches long.
Later analysis indicated that the hearth was used primarily for human cremation, in accordance with ritualistic burial practices that were common among Stone Age peoples. Most of the bones and bone fragments in the hearth area had been burned and then broken into smaller pieces and reburned. This practice could account for the marks on the one bone (a fragment from an apparent femur) that Fryxell and others pointed to as evidence of cannibalism in Marmes culture. Such an argument, one anthropologist concluded, "shows a profound lack of understanding of the typical cremation process, where several times the bones are gathered after the fire has died down and then burned again or broken into small pieces for further disposal" (Hicks, 148).
Race Against Time and Water
Lower Monumental Dam was originally scheduled to be completed in December 1968, just eight months after the press conference in Magnuson’s office. Under pressure, the Army Corps of Engineers, builders of the dam, agreed to finance new excavations, in an effort to salvage as much of the archaeological record at the Marmes site as possible. The digging began in May 1968 and continued through February, with most workers on 18-hour shifts, despite one of the coldest winters on record.
Daugherty, supervisor of the original excavations, had turned his attention to other projects, leaving day-to-day management of the salvage operation to Fryxell and Henry T. Irwin, an assistant professor of archaeology. Faced with the knowledge that the site would probably be under water in a matter of months, they decided to sacrifice precision to speed. Roland Marmes’ bulldozer was put to work again, along with a backhoe. The machines churned through the upper layers in order to get to older, deeper levels, probably destroying many small artifacts and possibly more human remains in the process.
Frantic for more time, the WSU officials appealed to Magnuson for help. The powerful chairman of the Senate Appropriations Committee convinced President Lyndon B. Johnson to approve an emergency appropriation of $750,000 to build a cofferdam around the site. (A cofferdam, often used in bridge building, is an enclosure with pumps that can be used to keep an area dry.) Construction of the 1,900-foot-long, 500-foot-high cofferdam began in November 1968. Meanwhile, the Corps grudgingly agreed to postpone for a year the date for filling the Lower Monumental reservoir.
The cofferdam was built on a base of glacial gravel that was 130 feet deep in places. Engineers expected that water would seep through the gravel and into the sheltered area. They had planned it pump it out. They did not expect that it would pour in at the rate of 45,000 gallons a minute. When the Lower Monumental Dam’s floodgates were finally closed, in February 1969, the cofferdam's pumps quickly failed.
In a last, desperate measure, Fryxell and other team members built heavy wooden cribs around two unexcavated squares that seemed particularly promising. They lined the existing trenches with plastic sheeting and then, for five around-the-clock days, covered the plastic with 8,000 cubic yards of sand, as a possible gift to the future. If the reservoir should ever be drained, it may be possible to dig once again into the past.
Half-drowned remnants of the leaky cofferdam still stand, visible from an overlook about a quarter-mile above Lyons Ferry State Park.