Skip to main content Skip to navigation
Washington State University
Museum of Anthropology The Marmes Rockshelter Site


In 1952, John McGregor, a rancher from Hooper, Washington, took Richard Daugherty of Washington State University to see caves and rockshelters along the lower Palouse River Canyon. Among that group of sites was one on property owned by Roland (aka Squirt) Marmes (pronounced Mar-muss).

In 1962 Daugherty returned to the area to excavate the Palus Village site (45FR36A), located at the mouth of the Palouse River, which would be inundated following the construction of the Lower Monumental Dam on the Snake River. When the area originally intended for excavation at Palus Village was found to be severely disturbed, the decision was made to switch excavation efforts to the nearby rockshelter on Roland Marmes’ property.

Marmes Rockshelter prior to excavations.
Marmes Rockshelter prior to excavations.

Excavations 1962-1965


Major excavations began in the summer of 1962. Daugherty and Fryxell defined stratigraphic units and an arbitrary coordinate system that was used to guide and control all later excavations. The coordinate system consisted of a grid of squares five feet on each side. An arbitrary one-hundred-foot datum point was set on the high point of the berm at the mouth of the shelter. The 100’ datum was situated at N80/W0. N60 was arbitrarily chosen as the division between the rockshelter and floodplain areas of the site (Fryxell and Keel 1969:28).

Seventeen units were excavated in six-inch levels to a depth of around five feet during the ten weeks of the 1962 field season. Eight stratigraphic layers were recorded, with layer I at the bottom.

The wall of an excavation unit showing deposition layers
Stratigraphy at Marmes Rockshelter.


Seven to eight hundred artifacts were found and eleven burials were uncovered. The burials were removed from the site as plaster coated blocks of sediment and excavated at the laboratory of Anthropology at Washington State University (Breschini 1979). Three of the burials were found below a layer of Mazama ash, which at that time was dated to ca. 6500 B.P., but is currently dated at 7627 cal BP.

Pit features found in upper layers intruded into earlier layers and demonstrated the site’s function as a storage location.

1963 and 1964

The same grid system, site datum, and excavation strategy were followed during subsequent field seasons. Upper areas were shoveled clear without screening the sediment, as they were considered disturbed. Few notes survive and no reports were written. In 1964, the site was designated a National Historic Landmark (Rice 1969:7).


In 1965, Fryxell returned to conduct detailed stratigraphic research. He engaged Roland Marmes to cut a bulldozer trench to a depth of 14 feet below the surface of the floodplain. A concentration of bone fragments was uncovered, but it could not be determined whether the bone had come from the sides or the bottom of the trench. Carl Gustafson, zooarchaeologist with the Department of Anthropology at Washington State University, identified the bones as elk, deer and human.

Richard Daugherty and Roald Fryxell supervising the excavation of a trench outside the rockshelter.
Bulldozing at Marmes. (Richard Daugherty with pipe)