Missouri State University

Big Eddy 2005


Special Publication No. 4

Center for Archaeological Research

edited by
Neal H. Lopinot, Jack H. Ray, and Michael D. Conner

With Contributions by
Richard G. Baker
E. Arthur Bettis III
Edwin R. Hajic
Jacob Letts
Neal H. Lopinot
Patrick Moss
Gina S. Powell
Jack H. Ray
Lori Wozniak


This report presents the results of geoarchaeological, paleoecological, and archaeological research undertaken in the lower Sac River valley and in Archaic deposits at the Big Eddy site.

A model of regional landform development is presented based on field investigations, 22 radiocarbon dates, and studies of detailed topographic maps, high-altitude color-infrared aerial photographs, digital orthophotoquadrangles, historical black-and-white aerial photographs, and soil survey maps. Thirteen landforms have been delineated for the lower Sac River valley. They consist of multiple relatively high, undifferentiated terraces; two low terraces and associated paleochannels; two floodplain levels and associated paleochannels; alluvial fans; undifferentiated channel belts of tributary creeks; tributary-creek channels; and the Sac River. Initial impressions of the associated sediment assemblages, depositional environments, geochronology, and known locations of prehistoric cultural deposits at multiple sites along the lower Sac River channel, combined with geomorphological analyses at the Big Eddy site, indicate that a large percentage of the prehistoric record occurs in paleolandscape positions that are now shallowly to deeply buried.

The pollen, macroplant, and stable carbon isotope ratios provide proxy records of Holocene vegetation and climate change in the lower Sac River valley. Although anthropogenic affects on &Mac182;13C values need to be considered when reconstructing paleoenvironmental conditions, the carbon isotope results from the Big Eddy site are consistent with an overall increase in regional aridity and/or an increase in the seasonality of precipitation during the Younger Dryas between about 13,210 and 11,940 B.P. Studies of pollen and macroplant deposits indicate increases in nonarboreal vegetation and the appearance of prairie species from at least 8090 B.P. to 7520 B.P., followed by increases in forest vegetation between at least 6300 B.P. and 4500 B.P. The pollen and macroplant data indicate that the lower Sac River valley may have been relatively buffered from the effects of the Hypsithermal during the middle Holocene, with more stable vegetation patterns than elsewhere to the north and west. Significantly higher sedimentation rates and increased &Mac182;13C values in the Sac River valley during the late Holocene suggest an increase in large floods and mean annual precipitation that resulted in the expansion of forests (C3 vegetation) on the floodplain.

A survey was undertaken for a 49-km (30.4 mi) section of the Sac River from Stockton Dam to the U.S. 54 bridge over the Sac River, through portions of Cedar and St. Clair counties. This cutbank survey resulted in revisits of 35 previously identified sites and the identification of 11 new sites. In addition to conducting a canoe survey of the entire corridor, many of the sites were revisited on more than one occasion. The study benefited greatly from a wealth of information provided by artifact collectors, some of whom kept records of in situ finds. The cutbank survey has demonstrated the richness of the Sac River valley, particularly with respect to sites having buried Paleoindian and Archaic components. Including the Montgomery site (23CE261), which was not within the survey area, 18 sites exhibit Paleoindian components. Thirty-three sites are multicomponent, and 28 have three to as many as 12 separate components. The survey results have required a re-evaluation of the NRHP status of all previously identified sites. The cutbank survey also indicates the necessity for a flexible management plan involving in part periodic monitoring of sites and/ or cutbanks in the lower Sac River valley. This should be undertaken at a considerably more consistent and frequent basis than has been performed heretofore. When undertaken in concert with additional coring and testing of those portions of sites within sloughing easements, but away from cutbanks, preemptive archaeology can be undertaken when sites are being threatened but before erosion causes the destruction of significant deposits.

The findings of excavations undertaken principally in 2001 and 2002 at the Big Eddy site are described in this volume. These pertain to excavations in six block areas (Blocks F-K) within: (1) Late Archaic deposits in the thick late Rodgers Shelter submember; (2) Late Archaic, Middle Archaic, and Early Archaic deposits in the thick middle Rodgers Shelter submember; and (3) Middle Archaic and Early Archaic deposits in the thin middle Rodgers Shelter submember. Within the thick late submember deposits, a stratified sequence of Smith-Etley (ca. 4500-3600 B.P., but primarily 4200-3800 B.P.), Williams (ca. 4100-3900 B.P.), Stealth (ca. 3600-3400 B.P.), Kings (ca. 3800-3000 B.P.), and Afton (ca. 3000-2700 B.P.) components were defined beneath a relatively thin veneer of mixed Woodland and Mississippian deposits at and near the surface. Two Late Archaic phases, the Smetley and Sac phases, have been defined.

Big Eddy apparently contains relatively little evidence for Middle Archaic occupation, but the site was used periodically by a host of Early Archaic groups. From earliest to latest, excavations within the thin Middle Rodgers Shelter submember resulted in the definition of eight Early Archaic components following Dalton/San Patrice (ca. 10,500- 9800 B.P.): Breckenridge (ca. 9800-9700 B.P.), Scottsbluff (ca. 9600-9500 B.P.), Indeterminate Early Archaic (ca. 9500-9000 B.P.), Cache River (ca. 9000- 8700 B.P.), Graham Cave (ca. 8600-8200 B.P.), Rice Lobed (ca. 8200-8000 B.P.), Hidden Valley (ca. 7900- 7200 B.P.), and Searcy (ca. 7800-7100 B.P.). For the most part, the evidence from Big Eddy suggests that individual styles of Early Archaic projectile points are representative of different cultural groups. The same conclusion generally applies to Late Archaic projectile points.

Patterns of lithic raw-material selection and use characterizing the early part of the Early Archaic period were first established during the Late Paleoindian period. Beginning with the Graham Cave occupations, however, there was a shift toward the use of Burlington chert and away from a preference for Jefferson City chert. This pattern continued with varying degrees of intensity through the end of the Late Archaic period. Intentional heat treatment of chert appears during late Early Archaic times, being well manifested first in the Hidden Valley and Searcy components at Big Eddy. Although varying in intensity, the practice of heat treating chert continued throughout the remainder of prehistory.

The Big Eddy site was used throughout most of the Archaic stage, but it was sometimes used for very specialized tasks on a very brief and intermittent basis and at other times as a residential locus occupied on a seasonal or multiseasonal (perhaps even year-round) basis. Although bone preservation is exceedingly poor, archaeobotanical evidence indicates that hickory nuts, walnuts, and perhaps acorns and fleshy fruits were important plant food resources throughout Archaic and Woodland times. The record at Big Eddy also points to an emphasis on chenopod harvesting beginning around 4500- 4300 B.P. This is clearly manifested during the Smetley and Sac phases, or from about 4200-3600 B.P. The evidence is suggestive of the cultivation of this starchy seed, or at least the manipulation of the habitats favoring chenopod. Given the presence of squash at Big Eddy during at least the latter part of the Smetley phase, some form of cultivation of chenopod would seem to be likely. Pericarp thicknesses lend support to this contention, suggesting that chenopod was then undergoing the process of domestication.