Museum of Anthropology

The Marmes Rockshelter Site

Site Interpretations


Hicks’ research design (Hicks 2004:25-41) addresses a series of questions within four major research themes. His research was constrained by problems arising from the excavation and post-excavation procedures. Problems related to the original research design and excavation procedures include:

  • Missing field catalogs for 1962 and 1963 and only a partial catalog for 1964

  • Failure of many excavators to use level forms

  • Incomplete field notes

  • Inconsistent collection procedures during floodplain water-screening which prevented correlation of larger items found in the floodplain with items found during the water-screening

  • Emphasis on materials associated with certain aspects of the occupation (i.e., lithics and faunal remains) to the detriment of materials associated with other aspects of the occupation (i.e., botanicals).

Problems related to post-excavation procedures include:

  • Two collections management teams, one at Washington State University and one at the Colville Confederated Tribes, employing slightly different approaches to database construction.

  • Missing items and items without provenience information. These include half of all formed stone tools recovered, some human remains and associated funerary objects, and bulk samples from the floodplain.

Despite these constraints, by applying detailed and intense analysis of the available written records and cultural materials Hicks was able to address a number of important questions within each theme.


Age of the Cultural Deposits


Among the questions asked regarding cultural chronology are:

  • When was the Marmes Rockshelter site first used?

  • Does the early assemblage correlate with any others in the area?

  • What is the relationship between the cultural materials in the rockshelter and on the floodplain?

  • What is the continuing relationship over time between these areas of the site?

  • Is there occupational continuity at the Marmes Rockshelter site?

  • If yes, was the occupation by the same cultural group?

  • What explains discontinuities in the occupational sequence?

Daugherty (1962) proposed a theoretical framework for the inland northwest which he called the Northwest Riverine Areal Tradition. This tradition is one of three area traditions, along with the Desert Area Tradition (or Desert Culture) and the Southwest Agricultural Area Tradition, that make up the Intermontane Western Tradition. He saw these as developing partly in response to environmental changes.

The Marmes Rockshelter site is important in developing a cultural chronology for the Inland Northwest because it contains a good sequence of radiocarbon dates in addition to well-dated ash deposits. Leonhardy and Rice (1970) built upon Daugherty's (1962) theoretical framework in developing their cultural typology for the Lower Snake River Drainage

Hicks (2004) was able to add seven new radiocarbon dates to this sequence by dating items found below the earliest demonstrated use of the rockshelter. A new date of 11,230±50 B.P. from culturally modified bird bone located on top of glacial lake sediments at the bottom of the excavation unit overlaps with the 10,810±300 B.P date from rockshelter deposits. The new date pushes back the earliest opportunity for humans to use landforms in the Snake River drainage and the Palouse River Canyon following the draining of Glacial Lake Missoula and shortly after the Glacier Peak eruption. Glacier Peak tephra dates the Richey-Roberts Clovis Cache, which suggests that use of both Richey-Roberts and Marmers Rockshelter occurred at roughly the same time.

Dates from other early assemblages also overlap with dates from the Marmes Rockshelter site. These are Paulina Lake in central Oregon (35DS34) and Cooper’s Ferry in west-central Idaho (10IH73) (Hicks 2004:393-394).

New dates from the floodplain fit within the range for Stratum I of the Rockshelter, and also show that the floodplain continued in use “at least into the Cascade Phase” (Hicks 2004:395).

Cultural material is found in all deposits above Mazama ash. All phases of Leonhardy and Rice’s (1970) projectile point typology are represented and each type overlaps with earlier and later types. Use may have ceased right after the deposition of Mazama ash, until sedimentation and soil formation began again. Technology and material types for lithic tools reflect the regional trend of early use of cryptocrystalline raw material, more intensive subsequent use of basalt, especially during the Cascade phase, and then a return to use of cryptocrystalline material. There is also no hiatus between the Cascade and Tucannon phases at Marmes Rockshelter, which suggests that evidence of such a regional hiatus may have resulted from limited data.


Environment at the Time of Occupation and Climate Change over Time


Hicks (2004) addressed three broad questions about the nature of the past environment and changes in climate over time:

  • What was the sequence of climatic change?

  • What was the effect of Mazama ashfall?

  • What is the relationship between the geological and sedimentary history of the area and its use by Native Americans?

Fryxell’s original climatic sequence for the period before the Mazama ashfall has been refined in recent years. Fryxell used rate of roof fall within the rockshelter as a proxy for climate change. His general trends are correct, but the timing of the trends has been modified. The warming and drying trend begins around 9500 B.P. (Chatters 1998; Wigand in Hicks 2004). Warm and wet conditions prevail from 6000 to 4000 B.P, followed by a cooler and wetter period which lasts until ca. 3000 B.P. Beginning around 3000 B.P., there is gradual drying, which continues in the present.

One indication of climate is the type of shellfish found in the Snake and Palouse Rivers. Gonidea prefers warmer, slower-moving water, which is an indication of less vegetation. This species dominates the pre-Mazama strata. Margaritifera prefers cooler water, which is a sign of more vegetation. This species dominates from about 2700 B.P. Between 6700 B.P and 2700 B.P., both species co-exist.

Mazama ash probably contributed to the aeolian base of the area for about 3,000 years. The climate was the warmest and driest at the time of eruption, and the vegetation cover was minimal, therefore there would have been considerable redeposition of ash. Around 6000 B.P., the vegetation cover increased and there was less redeposition of ash. The ashfall’s effect on food resources might have included a decline in mammals and “clumping” of certain prey populations. Short-term changes in the seasonal round in response to variable ash deposition, however, would not be visible in the archaeological record (Hicks 2004:408).

There is no evidence of site abandonment. Hicks concludes (2004:408) that “effects on soil development…are variable and probably fairly limited in time.”


Conclusions about the Occupation of the Marmes Rockshelter


Hicks’ (2004) questions about the function of the site include:

  • What was the nature of the material culture, what changes are apparent over time, and what does this tell us about how people used the site?

  • What kinds of changes can be seen in resource use over time?

  • Are there indications of seasonal use?

  • What changes might be observed in the use of terrestrial species?

  • What is the contribution of fish to the diet, and which species were utilized?

  • What changes in the use of plant foods can be observed?

Following Binford’s (1980) model, Hicks looked for a transition from mobile foraging to a more semi-sedentary lifestyle and attempted to pinpoint the timing of this change. Could Marmes be identified as a seasonally “tethered” site, or a field camp within a collector strategy?

The original research design, which focused on vertical culture-historical investigation, constrained a more horizontal, spatially oriented settlement pattern approach. Hicks also wanted to avoid an oversimplified, ethnographic analogy with local post-contact cultural groups.

Hicks found that there was differential use of the floodplain and the rockshelter. The floodplain contained more debitage and only about 5 percent of all tools. All of the bone needles were found in the floodplain deposits. Ninety-five percent of all tools were found in the rockshelter, as were all the bone needles. Additionally, the storage features containing botanicals in the upper levels suggested a storage focus in the later part of the occupational sequence.

The site had been used for habitation on a continuous basis over a long period of time. The evidence for this includes tool manufacture and maintenance, food procurement, and processing of food as well as other materials. With the exception of Stratum IV and Stratum VIII, the site was a residential camp. It may have been occupied during the winter based on location, the density of artifacts, the density of fire-related features inside the rockshelter, and the heavy use-wear on tools.

The early use of the floodplain, as well as Stratum I and Stratum II in the rockshelter, probably consisted of generalized activities in the rockshelter and specialized manufacturing activities on the floodplain.

There is a small amount of evidence to show the relationship between use of the floodplain and rockshelter. The cremation hearth suggests use as a socio-religious structure, but this does not appear to be the main focus of the occupation. Many other hearth structures reflect day-to-day occupation. These features, plus intensive tool use, exhaustion of tools, and disposal of exhausted tools suggests use as a mobile forager base camp. Later cultural activities appear to be restricted to the rockshelter, probably with “significant fluctuations in the intensity of the occupations” (Hicks 2004:411). The most intensive occupations, as evidenced by tool manufacture, occur in Stratum III and Stratum IV.

Early subsistence patterns show use of large and medium-sized mammals, fish, and shellfish, consistent with function as a residential base camp in a mobile foraging system. There is no evidence at this time that large mammals were overexploited. Later, there is evidence of more use of small mammals, an increase in the use of fish, a decrease in the use of shellfish, and an increase in storage of plant foods. This is consistent with a function of field camp in a logistical collecting system. It is possible that the site was related to a pithouse village (45FR36C) at the site of what became Palus Village (45FR36A). In the late prehistoric period, there is little material that fits a logistical collector model.

Unfortunately, there is not enough salmon bone in the assemblage to determine whether the site was used as a salmon processing location. Plant food remains are also rare in the assemblage, but this is more a result of the original research emphasis on environmental change, rather than an indication that few plant foods were processed at the site.




Questions related to trade include:

  • What materials came from outside the Lower Snake River region?

  • Where did these materials originate?

  • What are the implications with regard to social contact?

  • Was the impetus for trade technological, subsistence, or other?

Non-local toolstone includes obsidian, which comprises less than one percent of formed tools. There are five distinct sources for the obsidian found in the assemblage. Obsidian from Indian Creek, in eastern Oregon, is found in all strata. Obsidian from Whitewater Ridge, in eastern Oregon, is found in Strata III, VI, and VII. Obsidian from Gregory Creek, also in eastern Oregon, is found only in Stratum I. Obsidian from Timber Butte, in western Idaho, is found in Stata IV/V and VII. A fifth source is unidentified (Hicks 2004:401). Based on the size of tools over time, it appears that these materials were procured by the inhabitants of the site, either by travel to these areas specifically for obsidian, or embedded in the annual round.

Other non-local materials are non local cryptocrystalline chert, petrified wood from the mid-Columbia area, a cryptocrytalline materials that resembles Knife River flint from the Plains (see Ozbun et al. in Hicks 2004:224), and marine shell. Marine shell was almost certainly obtained through trade. Olivella biplicata was used to decorate clothing and as jewelry. It is found at many sites in the Inland Northwest, and shells exhibit a similar pattern of preparation by grinding over a long period of time. Examples are found in all strata at the Marmes Rockshelter Site. A pendant made from abalone shell (Haliotis sp.) is found in Stratum III.


Human Remains


No additional analysis of human remains was conducted for Hicks’ (2004) report. Any associated funerary objects which were inadvertently included in material analyses are reported but were not photographed. Hicks (2004) addressed only those questions regarding the presence of human remains which could be answered using previously completed, publicly available research.

  • Were the human burials intentional?

  • When were the people buried?

  • Are there changes in the burial patterns?

The minimum number of individuals found has been calculated at thirty-six, plus many small skeletal elements and bone fragments. It is probable that other fragments were present in the sediments from the bulldozer cut made in 1968. Twenty-six individuals came from the rockshelter, six were located in the cremation hearth, and four were found in the floodplain. The human remains span the period from ca. 9000 B.P. to ca. 660 B.P. Some remains from the Marmes Rockshelter are older than Kennewick Man (Hicks 2004:378).

The burial patterns identified by Sprague (1959) for the late precontact period appear to be valid for the Marmes Rockshelter as well. Many individuals were oriented north-south, buried with olivella shell ornaments, stone tools, red ochre, and other cultural items.


Overview and Introduction


Discovery and Excavations 1962-1965


Excavations 1968 


Curation Projects 1994 through 2005




Hicks (2004) Appendices






Museum of Anthropology , PO Box 644910, Washington State University, Pullman WA 99164-4910, 509-335-3441, Contact Us