July, 1997

A Recently Recorded Late Woodland Campsite

on the Dempsey Divide: 34RM773, Thurmond Ranch #112

J. Peter Thurmond & Charles B. Batchelor

The second author has been collecting artifacts from a large campsite on the East Fork of Sergeant Major Creek in recent years which was not included in the 1990 synthesis of the archaeological sites on the Dempsey Divide in Roger Mills County (Thurmond 1990). The site has produced ten corner-notched arrow points (and no dart points; see artifact photo and table below), and thus appears to have been used during the time when the bow and arrow had entirely supplanted the atlatl and dart in this locale. Based on the radiocarbon dates from Dempsey Divide site 34RM208 (see "Thirty Thousand Years of Change on the Dempsey Divide", this website), we presume that 34RM773 dates about AD 900-1000.

The location of 34RM773 certainly follows the pattern of previously recorded major campsites in the area (Thurmond 1990), lying on the Ogallala boundary ecotone, on a ridge overlooking a perennial, spring-fed stream. Running water is available within 200 m, and the site offers a commanding view of the terrain to the north and northeast; you can see the Canadian/Washita divide by day, and the lights of the town of Leedey thirty miles away at night. One would assume, given the complete northern exposure, that people did not camp here in cold weather. It is a brutal spot in a January blizzard, but pleasantly breezy in July.  The visible site covers a large area of some 17,500 m², measuring 250 m  SW/NE x 70 m NW/SE, and probably represent multiple camping events. 34RM773 as recorded extends south of the ranch bound­ary fence, and is visible on two adjoining neighbors, where heavy grazing has produced good surface exposure and significant deflation by wind and water erosion. Dense grass cover complete­ly obscures the surface on our side of the fence, but it is likely that the archaeological materials continue 50-75 m north of the fence. Total site area is therefore probably greater than 20,000 m² (about 5 acres).

Where the surface has been denuded by grazing, a fairly dense concentration of burned rock (thermally fractured quartzit­ic cobbles, presumably used prehistorically to line hearths and/or boil food in skins or baskets; Wandsnider 1997) and lithic debitage lies immediately adjacent to a lag deposit of basal Ogallala Formation gravels. Cobbles in the gravel deposit range up to 25 cm in diameter. All stages of reduc­tion are represented in the chipped stone waste within the camp­site, from tested cobbles and cores to tiny biface thinning and retouch flakes. Lithic procurement, tool manufacture and tool refurbishment were all performed at the site. Most of the lithic procurement debitage is Ogallala quartzite. Some of the latter stage decorticate flakes of Ogallala quartzite appear to derive from heat-treated cores or bifaces. Most of the smallest decorti­cate thinning and retouch flakes are of Alibates agatized dolo­mite. Fragments of two ground sandstone milling stones (manos) were observed on the site. Milling slabs (metates) and manos are fairly common both on campsites and as isolated caches along the Ogallala boundary ecotone on the Dempsey Divide, and probably relate to Late Archaic/Woodland forager processing of acorns from the ubiquitous scrub oak trees and shinnery oak brush.

The site extends to within 100 m of where the Tertiary Ogallala Formation feathers out atop the Permian Elk City sandstone, and the ridgetop is occupied by a thin mantle of the very base of the Ogallala. The archaeological deposit is essentially surfi­cial, and has been severely disturbed by livestock traffic and agricultural terracing throughout its visible extent. Shallow burial by aeolian deposits is possible on our side of the fence, but has not been assessed.

The lithic assemblages from the Late Archaic and Woodland sites on the Dempsey Divide are pretty much the same, with the exception of the projectile points (Thurmond 1990: 118-123). For this reason, sites of the two time periods were lumped together in the 1990 OAS Bulletin article on this study area (ibid.). However, as we continue to collect diagnostics from the sites on the ranch, it is becoming possible to sort out which are Late Archaic, which are Late Woodland, and which are both (or early Woodland, with a mix of dart and arrow points and no pottery). The Late Archa­ic to Late Woodland sites in the Thurmond Ranch survey area can now be classified as follows: Late Archaic only: 34RM208B, 208C (also Protohistoric), 332D, 333D, 334C, 336B, 446A-C, 450, 461, 470A, 507, 508, 510, 513, 532, and 601; Late Archaic/Woodland: 34RM208A, 211, 332B, 454 and 514; Late Woodland only: 34RM331, 335, 336C (also Early Archaic), 447 (also Historic), 449, 602A, 611 (also Paleoindian), and 773. Pure Late Archaic components outnumber Late Woodland ones by roughly 2:1.

In recent years, Don Wyckoff and the first author have radiocarbon dated many samples from Copan soil (cf. Hall 1977, 1990) exposures in the vicinity. Formation of the melanized soil, indicative of mesic climatic conditions, begins as early as 50BC at the Finch Canyon (Section 5-12N-24W) and Higgins Creek #1 (Section 31-13N-23W) exposures. As we discuss elsewhere in this website, the Copan soil actually appears to be a welded conglomeration of three buried soils; see "Thirty Thousand Years of Change on the Dempsey Divide" and "Delaware Canyon Radiocarbon Dates". Our terminal dates at the top of the Copan soil are clustering around AD1000. The onset of more xeric conditions at or about that time appears to coincide with a habitation shift from the Ogallala boundary to the arable flood­plain soils of the Washita and its larger tributaries (Baugh 1984, Moore 1988), and presumably a subsistence change from hunting and gathering to farming and hunting.

Table 1. Chipped Stone Tools Collected from 34RM773.

Specimens A-I & K are corner-notched arrow points. Specimen L is a crude side-notched arrow point. Specimens J & M are unnotched triangular arrow points. Specimen N is an edge-damaged decorticate billet flake with a lipped platform and minute shatter along 22 mm of lateral edge, at an edge angle 15°.

TH = maximum thickness; TL = total length; BL = blade length; BW = blade width; HL = haft length; HW = haft width; NW = neck width; BD = basal depth; NA = neck angle.

References Cited

Baugh, Timothy G. (editor)

1984  Archaeology of the Mixed Grass Prairie, Phase I: Quartermas­ter Creek. Oklahoma Archaeological Survey, Archaeological Resource Survey Report 20.

Hall, Stephen A.

1977  Geology and Palynology of Archaeological Sites and Associat­ed Sediments, in The Prehistory of the Little Caney River, University of Tulsa Contributions in Archaeology 1:13-41.

1990  Channel Trenching and Climatic Change in the Southern U.S. Great Plains. Geology 18:342-345.

Moore, Michael C.

1988  Archaeology of the Mixed Grass Prairie, Phases II and III: Hay and Cyclone Creeks, and Predictive Modeling in the Quartermaster Watershed. Oklahoma Archaeological Survey, Archaeological Resource Survey Report 33.

Thurmond, J. Peter

1990  Archaeology of the Dempsey Divide, a Late Archaic/Woodland Hotspot on the Southern Plains. Bulletin of the Oklahoma Anthropological Society 39:103-157.

Wandsnider, LuAnn

1997  The Roasted and the Boiled: Food Composition and Heat Treat­ment with Special Emphasis on Pit-Hearth Cooking. Journal of Anthropological Archaeology 16:1-48.