The Ritual Landscape of Late Precontact Eastern Oklahoma: Archaeology from the WPA Era until Today

By Amanda L. Regnier, Scott W. Hammerstedt and Sheila Bobalik Savage

Revisits and updates WPA-funded archaeological research on key Oklahoma mound sites

As part of Great Depression relief projects started in the 1930s, the Works Progress Administration (WPA) sponsored massive archaeological projects across Oklahoma. The WPA crews excavated eight mound sites and dozens of nonmound residential sites in the Arkansas River Valley that date between AD 1000 and 1450. These sites are considered the westernmost representations of Mississippian culture in the Southeast.
 
The results of these excavations were documented in field journals and photographs prepared by the field supervisors and submitted in a series of quarterly reports to WPA headquarters. These reports contain a wealth of unpublished information summarizing excavations at the mound sites and residential sites, including mound profiles, burial descriptions, house maps, artifact tables, and artifact sketches. Of the excavated mound sites, results from only one, Spiro, have been extensively studied and synthesized in academic literature. The seven additional WPA-excavated mound sites—Norman, Hughes, Brackett, Eufaula, Skidgel, Reed, and Lillie Creek—are known to archaeologists outside of Oklahoma only as unlabeled points on maps of mound sites in the Southeast.
 
The Ritual Landscape of Late Precontact Eastern Oklahoma curates and contextualizes the results of the WPA excavations, showing how they inform archaeological understanding of Mississippian occupation in the Arkansas Valley. Regnier, Hammerstedt, and Savage also relate the history and experiences of practicing archaeology in the 1930s, incorporating colorful excerpts from field journals of the young, inexperienced archaeologists. Finally, the authors update current knowledge of mound and nonmound sites in the region, providing an excellent example of historical archaeology.

• Language: English
• Publisher: University Alabama Press
• Release date: Aug 27, 2019
• ISBN: 9780817392390

Book preview

THE RITUAL LANDSCAPE OF LATE PRECONTACT EASTERN OKLAHOMA

THE RITUAL LANDSCAPE OF LATE PRECONTACT EASTERN OKLAHOMA

ARCHAEOLOGY FROM THE WPA UNTIL TODAY

Amanda L. Regnier, Scott W. Hammerstedt, and Sheila Bobalik Savage

With contributions from David A. Baerreis, Forrest E. Clements, Lynn E. Howard, Phil J. Newkumet, and Kenneth G. Orr

The University of Alabama Press

Tuscaloosa

The University of Alabama Press

Tuscaloosa, Alabama 35487-0380

uapress.ua.edu

Copyright © 2019 by the University of Alabama Press

All rights reserved.

Inquiries about reproducing material from this work should be addressed to the University of Alabama Press.

Typeface: Minion and Futura

Cover image: WPA foreman James Bud Hunter fills out paperwork; courtesy of the Sam Noble Museum of Natural History

Cover design: David Nees

Library of Congress Cataloging-in-Publication Data

Names: Regnier, Amanda Leigh, author. | Hammerstedt, Scott W. (Scott William), author. | Bobalik Savage, Sheila, author.

Title: The ritual landscape of late precontact Eastern Oklahoma : archaeology from the WPA until today / Amanda L. Regnier, Scott W. Hammerstedt, and Sheila Bobalik Savage ; with contributions from David A. Baerreis, Forrest E. Clements, Lynn E. Howard, Phil J. Newkumet, and Kenneth G. Orr.

Description: Tuscaloosa : The University of Alabama Press, [2019] | Includes bibliographical references and index.

Identifiers: LCCN 2019001162| ISBN 9780817320256 (cloth) | ISBN 9780817392390 (ebook)

Subjects: LCSH: Indians of North America—Oklahoma—Antiquities. | Oklahoma—Antiquities. | Excavations (Archaeology)—Oklahoma—History—20th century. | Archaeology—Oklahoma—History—20th century. | New Deal, 1933-1939—Oklahoma. | United States. Works Progress Administration of Oklahoma.

Classification: LCC E78.O45 R44 2019 | DDC 976.6/01—dc23

LC record available at https://lccn.loc.gov/2019001162

Contents

Acknowledgments

Abbreviations

Introduction: Oklahoma, the Great Depression, and Relief Archaeology

1. The Life and Times of an Oklahoma WPA Archaeologist

2. Field and Laboratory Methods of Oklahoma WPA Archaeology

3. The Norman Site, Wagoner County

4. The Hughes Site, Muskogee County

5. The Brackett Site, Cherokee County

6. The Eufaula Mound, McIntosh County

7. The Skidgel Mound, Le Flore County

8. The Reed and Huffaker Sites, Delaware County

9. The Lillie Creek Site, Delaware County

10. Synthesizing WPA Archaeology in the Arkansas Drainage

Notes

References

Index

Acknowledgments

A project like this one can only be completed with the help of many people. At the Sam Noble Museum of Natural History, we thank Marc Levine, Susie Fishman-Armstrong, Elizabeth Leith, and Elsbeth Dowd for facilitating access to collections and documents over the years. Copies of the WPA Quarterly Reports that make up much of this volume are housed at the Western History Collections of the University of Oklahoma Library and at the Sam Noble Museum of Natural History. James Brown, Patrick Livingood, and George Sabo have served as valuable sounding boards for many of our ideas. Ann Early, Richard Drass, and Susan Vehik graciously answered our questions about artifacts from their research areas. Robert Brooks, director emeritus of the Oklahoma Archeological Survey, was extremely supportive when this research was in its early stages. Bill Savage answered our questions about eastern Oklahoma during the 1930s. Funding for radiocarbon dates came from the University of Oklahoma’s College of Arts and Sciences and the Oklahoma Archeological Survey. Laurie Tinsley from the University of Oklahoma Academic Records Office generously went into university archives and provided information about the field supervisors’ coursework. Tracy Newkumet Burrows gave us permission to use the photograph of the four WPA field supervisors saved by her grandfather, Phil Newkumet. Shawn Lambert drew the copper artifacts from the Reed site. We would be remiss if we did not mention the WPA field supervisors and other workmen. The research conducted by Forrest Clements, David Baerreis, Joe Bauxar, Lynn Howard, Phil Newkumet, Kenneth Orr, Herbert Antle, Fred Carder, and Sarah White Clements has made it possible to produce this volume. Lynne Sullivan and an anonymous reviewer provided helpful suggestions on an earlier draft.

Finally, we would like to acknowledge the members of Caddo Nation of Oklahoma and the Wichita and Affiliated Tribes. Thank you for your continued support of our research.

Abbreviations

Introduction

Oklahoma, the Great Depression, and Relief Archaeology

The bleak landscape of rural Oklahoma during the Dust Bowl is an indelible part of American collective memory of the Great Depression and the 1930s. In his 1939 novel, The Grapes of Wrath, John Steinbeck created the saga of the fictional Joads, a family of rural, uneducated Okies forced to leave the land that could no longer support them after a series of crop failures. Thanks to the novel and photographs of the era, many Americans can easily call to mind images of impoverished Oklahoma tenant farmers headed westward in rickety, overloaded Ford Model T’s on a quest for a promised land that never materialized.

Steinbeck’s Joad family originated in Sallisaw, a town in eastern Oklahoma’s Sequoyah County, near one of the most archaeologically rich areas of Oklahoma. The Arkansas River forms the southern boundary of Sequoyah County and the northern boundary of Le Flore County. The river terraces of northern Le Flore County were particularly rich in archaeological sites for as many as 12,000 years. The Spiro Mounds were the largest and longest enduring of these sites in the Arkansas River drainage. The endemic poverty, precipitous economic decline, and environmental instability of eastern Oklahoma in the 1930s formed the backdrop for the destruction of the Craig Mound at Spiro, one of the worst incidents of looting ever documented at a North American archaeological site. In 1935, a group of hard up local men managed to destroy what was likely the largest single cache of precontact native art north of the Rio Grande (La Vere 2007). The ensuing archaeological cleanup, conducted by the University of Oklahoma and funded by the Great Depression relief funds funneled through the Works Progress Administration (WPA), was the largest in a series of relief archaeological projects undertaken across the state in the 1930s. Between 1934 and 1940, archaeological excavations were conducted at Spiro and at a series of related mound sites. The federally funded relief project came at a time when many locals were desperate for income, and it provided steady jobs for scores of impoverished residents of rural eastern Oklahoma.

Residents of Sequoyah County, and the rest of eastern Oklahoma, were in terrible financial shape in the 1930s, but Steinbeck mischaracterized the landscape when he described crops desiccating and topsoil blowing away due to prolonged droughts and high winds (Hall 2012). The exceptional drought and losses of topsoil of the Dust Bowl were centered in the Great Plains environment of western Oklahoma and the Texas panhandle, where the primary cash crop was wheat (Figure I.1). Eastern Oklahoma has more trees, rolling hills, and a substantially wetter climate; corn and cotton were the primary cash crops. The dire economic situation that gripped eastern Oklahoma in the 1930s arose directly out of economic and agricultural developments of the preceding decades. Eastern Oklahoma had been in serious economic trouble for most of the 1920s, a result of fluctuating cotton prices, the boll weevil epidemic, growing debt loads for farmers, and the collapse of the oil and coal industries (Bryant 1975). The economic situation had been so bad for so long that the stock market crash of 1929, which signaled the start of the Great Depression, did not even make the Sequoyah County newspaper (Hall 2012:42).

Demographically, Sequoyah County serves as a microcosm for the rural counties of eastern Oklahoma, where the men who made up the WPA excavation crews resided. Historian Ryan Hall (2012:38) uses US census data to characterize the county population during the 1930s. Despite being part of the historic Cherokee Nation in the nineteenth century, Sequoyah County was overwhelmingly white in the 1930s. Only about 20% of county residents were born in Oklahoma; most came from southern states, especially Arkansas. More than 60% of the farms in Sequoyah County were operated by tenants or sharecroppers. Less than 70% of children between the ages of 7 and 13 attended school, and at least 10% of the population was illiterate.

State and Federal Great Depression Relief Programs

Although federal relief programs benefited the vast majority of employable Oklahomans through the 1930s, statewide popular support for New Deal relief initiatives ebbed and flowed. The precipitous decline of Oklahoma’s economy during the 1920s left Oklahomans clamoring for political change. Populist Democratic candidate William Alfalfa Bill Murray was elected governor of Oklahoma in 1930. Murray had a long role in Oklahoma politics, serving as president of the 1906 Oklahoma statehood convention and writing many articles of the state constitution. He ascended to the governorship with an outsize personality in a rumpled white suit and as a champion of farmers and an enemy of the rich, whom he blamed for the state’s economic plight (Bryant 1968).¹ Most importantly, Murray promised relief, which had been short in coming, since President Herbert Hoover’s federal government had been largely ineffective and the Oklahoma state government had directed relief responsibilities downward to cash-strapped local governments (Goble and Scales 1983:5). In 1931, Murray began appropriating state funds for relief, but by 1934, the state had contributed only a paltry $1.2 million dollars to the cause (Bryant 1975:173).

The first incarnation of federal work relief programs in Oklahoma occurred under the auspices of the Federal Relief Administration (FERA), created by Congress in May 1933 after Franklin D. Roosevelt defeated wildly unpopular and ineffective incumbent president Hoover in the 1932 election. The FERA provided grants to state and local governments to create jobs for unskilled male and female workers and white-collar positions. The Civil Works Administration (CWA), which funded the first archaeological efforts in Oklahoma, was created to provide job opportunities during the winter of 1933–1934. Governor Murray and the federal government found themselves at odds over CWA funds² because the state could not afford to match federal support, violating a major condition of receiving relief monies. Further, Murray’s distribution of funds appeared to be based on patronage and political support. Owing to these perceived improprieties, federal officials seized control of the distribution of CWA relief funds from the state government in February 1934 (Goble and Scales 1983:9).³ Under federal control, the CWA built dams, college dormitories, airports, roads, and courthouses across Oklahoma (Bryant 1975:174). Despite this, relief efforts prior to 1935 are described as inadequate, haphazard, and ineffective, with the bulk of Oklahoma counties receiving less than $50 per capita of support (Bryant 1975:175).

In early 1933, Roosevelt created the Civilian Conservation Corps (CCC) to recruit young men to work on projects related to natural resources. Also referred to as the tree army, CCC crews built roads, trails, and campsites and constructed buildings at state and national park facilities to help citizens get closer to nature (Means 2013:4–5; Pasquill 2008). Even as support for other relief projects wore thin toward the end of the 1930s, the CCC remained popular. Many of their projects aided the Soil Conservation Service, where they were desperately needed in Oklahoma to combat soil loss due to erosion and the Dust Bowl, although their effectiveness at such efforts is dubious (Bryant 1975:187).

When his outsider populist agenda and eccentric character wore thin, Murray was voted out of office in 1934, defeated by oil baron E. W. Marland. Marland promised to bring the New Deal to Oklahoma (Goble and Scales 1983:11). His visionary plan for Oklahoma was to create five agencies, consisting of planning, housing, flood control, industry, and highway boards, at a staggering cost funded by tax increases (Goble and Scales 1983:12–14). Despite the failure of Marland’s ambitious plan, federal relief efforts in Oklahoma ramped up with the creation of the Works Progress Administration in 1935. Between just 1935 and 1937, the WPA employed 119,000 Oklahoma men and women (Mullins 2009).

In Oklahoma, more than half the WPA funds were spent on road construction, but the most visible legacy are the many civic buildings constructed as part of the relief effort. These public buildings, typically made of local stone, include school buildings, National Guard armories, fire stations, courthouses, and hospitals, as well as less visible architecture such as sidewalks and drainage features. They remain prominent features on the Oklahoma landscape, particularly in rural areas. In addition, the WPA funded sewing rooms and provided domestic training for women, geological and mineralogical studies, historic preservation and reconstruction, music projects, public art, and a folklore project to collect stories from former slaves, pioneers, and members of native communities (Mullins 2009). Oklahoma WPA headquarters were in Oklahoma City. A state administrator oversaw the nine districts, each of which had an area office. The Oklahoma WPA archaeologists mention interacting with personnel from both the central Oklahoma City office and several area offices, in McAlester, Muskogee, and Tulsa.

Public support for Oklahoma WPA projects remained strong through 1936. By 1937, the WPA had poured $43 million into the state, which was augmented by $10 million in matching funds from the state and other sources (Bryant 1975:186). About 70% of these funds went to wages (Mullins 2009). While Oklahoma had the third-greatest income decline in the country during the Depression, it ranked only twenty-second in per capita relief funds (Bryant 1975:187). Because of mismanagement and corruption, federal relief efforts never reached their full potential. The state’s mismanagement of Social Security benefits to the aged, disabled, and needy children (Goble and Scales 1983:16) is a good example of the ineptitude with which the state government handled relief. The legislature funneled benefit funds from the newly created federal Social Security Administration to corrupt county boards, which then distributed them. Roughly one-third of individuals who received benefits were political supporters of county commissioners and neither needy nor elderly. When federal authorities discovered this corruption, a national scandal ensued, and the Social Security Administration suspended federal payments to Oklahoma (Goble and Scales 1983:19). This left many farmers who were still dealing with the fallout of the Dust Bowl with no recourse. Others opposed the WPA because they saw it as a make work program equivalent to welfare (Bryant 1975:187). The WPA funds in Oklahoma went largely to rural areas, so the population centers of Oklahoma City and Tulsa saw little benefit from relief funds, and residents of those areas saw little reason to support the relief effort.

By 1938, sentiment had begun to turn and the New Deal became central to the race for Oklahoma governor. The WPA workers received a large raise just before the election in the hopes of getting New Deal supporters reelected (Bryant 1975:188). Roosevelt appeared in Oklahoma City, rallying behind the candidates who favored relief programs. Both efforts proved unsuccessful, and strident New Deal opponent Leon Phillips was elected governor. Phillips filed lawsuits attempting to stop the construction of reservoirs along the Red River, delaying those projects for upward of two years (Bryant 1975:190). After 1938, federal relief funds in Oklahoma steadily dwindled, but half the state was still on some form of relief in 1939 (Hall 2012:46). By 1940, as rainfall levels rose above normal and drought conditions abated, the hard-hit farming sector had begun to turn around, and the citizens of rural Oklahoma began to think that they had less need for relief funds (Lambert 1983:83). While this was certainly the perception and anti–New Deal sentiment was ever increasing, the WPA still employed almost one-third of Oklahoma’s heads of household in 1940 (Hall 2012:46). As the growing threat of the European and Pacific wars loomed, relief funds were increasingly diverted to projects that would further national defense. At the same time, Oklahoma’s population had changed. The departure of the most hard-hit and indebted farmers in the mid-1930s meant that only the most successful remained. Oklahoma became more urbanized throughout the 1930s, while relief efforts in the state were still overwhelmingly directed to rural areas. Perhaps most importantly, popular skepticism toward the federal government, an ever-present sentiment in Oklahoma, began to emerge again as the state economy strengthened (Bryant 1975:192).

WPA Archaeological Programs

For American archaeologists, the Great Depression calls up a mental image as indelible as the overloaded Okie Model T, one of large groups of men (and sometimes women) in shabby overalls holding shovels next to partially excavated earthen mounds. As Haag (1985:278) succinctly states, There has never been a greater revolution in American archaeology than that engendered by the New Deal period. Federally sponsored relief programs resulted in the expansion of anthropology and archaeology programs in American universities, greatly increased available archaeological knowledge with projects conducted on unprecedented scales, and initiated a generation of new archaeologists who introduced improved field techniques to the discipline (Haag 1985:278). Before the advent of relief archaeology, a typical field investigation included 10 to 15 crew members in the field for 12 to 16 weeks at a cost of roughly $2,500 (Setzler 1943:206). Federally sponsored project crews averaged 150 members and lasted 36 to 38 months in the field, generating an unprecedented wealth of archaeological data. The massive scale WPA excavations provided a crucial foundation for the modern archaeological understanding of mound construction, community organization, domestic architecture, and mortuary practices of the precontact Southeast. In 1943, Smithsonian Institution archaeologist Frank Setzler (1943:207) assessed that the WPA accomplished in 6 to 7 years what would have taken pre-1930 archaeologists 50 years! While the antiquated field methods and variable quality of notes can be extremely frustrating for modern archaeologists, the artifact collections and notes from these and other federally funded excavations remain fertile ground for contemporary archaeological research (e.g., Dye 2016; Means 2013; Pritchard and Ahlman 2009).

The first federally sponsored archaeological relief projects in the United States were the FERA-funded 1933 excavations at the Marksville site in Louisiana (Lyon 1996:28). Archaeological work expanded under the CWA during the winter of 1933–1934, with projects sponsored by the Smithsonian Institution in Georgia, Florida, Alabama, North Carolina, Tennessee, and California that employed 1,500 people (Lyon 1996:29–37; Setzler and Strong 1936). The CWA sites were concentrated in the southeastern United States because of both the mild winter conditions and the abundance of unemployed potential crew members (Setzler and Strong 1936:301). At the same time, the Tennessee Valley Authority (TVA), created by Roosevelt in 1933, initiated projects in northern Alabama and Tennessee in advance of the creation of a system of dams and hydroelectric power plants along the river (Olinger and Howard 2009). The CWA provided early labor crews for TVA archaeology, aimed at excavating sites slated to be lost to reservoir inundation (Lyon 1996:30). The CCC crews were drawn into archaeology across the country, particularly when archaeological sites were located on state and national park and forest lands. In some cases, agencies pushed the use of CCC crews because labor costs were substantially lower than those incurred by other relief programs (Lyon 1996:155).

Archaeology was particularly well suited to the goals of federal Depression relief efforts for several reasons. Large numbers of men could be put to work quickly on lengthy projects (Means 2013:8). Costs to establish a project were relatively cheap; 90% of funds could go to labor, with the other 10% covering field supplies (Setzler and Strong 1936:306). The projects resulted in scientific knowledge rather than a concrete product, so a great number of workers could be employed without any worry of oversaturating a market. Relief archaeology was not without its drawbacks, as Setzler and Strong (1936:307) point out. Spectacular archaeological sites were a limited resource, and early relief efforts lacked well-planned timetables and budgets. Field supervisors also frequently wound up supervising 100 or more unskilled workers, making it impossible to ensure that rigorous field methods were always followed (Fowler 1986:145).

The creation of the WPA in August 1935 changed the administration of relief archaeology. The WPA considered archaeology a white-collar, professional project, but unlike other projects in that category, no central supervisory authority existed. Proposed projects were submitted on a state-by-state basis; half the states in the nation requested WPA funds for archaeological projects in 1935 (Lyon 1996:64). The WPA recruited the National Park Service (NPS) to approve and provide technical supervision for projects in hopes of avoiding funding applications by unqualified entities. The NPS lacked staff with technical skills to supervise and evaluate proposed projects and turned to the Smithsonian for assistance establishing standard procedures for archaeological projects, including recording, curation, and reporting of finds, and for reviewing proposed archaeological work to ensure its quality (Lyon 1996:66).

By 1938, 200 archaeologists were directing projects and nearly every paper at the Society for American Archaeology meeting was based on WPA projects. Concerns continued to mount about the variable quality of archaeological work. The WPA appointed anthropologist Vincenzo Petrullo of the University of Pennsylvania Museum to serve as national consultant for archaeological projects. Petrullo admirably attempted to divert WPA funds from archaeological projects that lacked skilled supervisors, failed to plan for long-term archaeological collection storage, and did not produce archaeological reports with Operating Procedure No. W-18 (Lyon 1996:67). As of July 1938, project supervisors were required to prove their qualifications to lead fieldwork, to curate artifact collections in a public institution with free access, and to produce quarterly progress reports of artifacts recovered, sites surveyed and found, hypotheses tested, and conclusions reached. The Quarterly Reports, aimed at helping researchers prepare a final report when projects concluded, proved unpopular with WPA archaeologists due to the additional time commitment (Lyon 1996:69).

By 1939, the WPA was supporting 32 archaeological projects at a cost of more than $2 million. The new rules that came along with the 1939 WPA reorganization created huge headaches, as projects had to receive 25% of their funds from the state and workers had to rotate off the project after 18 months. As the threat of World War II loomed, the WPA continued to reorganize; after 1940, defense projects took priority over aid, and funds available for archaeology steadily declined (Means 2013:10). This trend continued in the following years, particularly after the United States was drawn into the Pacific theater in December 1941, and the archaeology program was shuttered for good in April 1942 (Lyon 1996:78).

Federal relief archaeology programs were essential for the development of the broader discipline of southeastern archaeology. Initial projects were sponsored by the CCC, and archaeological programs funded under the FERA, the CWA, the WPA, and the TVA generated massive amounts of archaeological collections from sites across the Southeast. Perhaps even more importantly, these projects led to the establishment of a number of university anthropology museums across the Southeast to house those collections, including those in Kentucky, Tennessee, and Louisiana (Sullivan et al. 2011:70–71). In the Southeast, the most extensive WPA excavations occurred in Alabama, Georgia, Kentucky, Louisiana, North Carolina, and Tennessee (Lyon 1996).

Federally Sponsored Relief Archaeology in Oklahoma

Federally sponsored relief archaeology in Oklahoma was directed from the start by the Department of Anthropology at the University of Oklahoma. Between 1934 and 1942, the WPA archaeological project recorded or excavated sites in 14 of Oklahoma’s 77 counties (Figure I.2) across the state. The WPA paperwork and artifacts curated at the Sam Noble Museum of Natural History (SNMNH) include notes, artifacts, photographs, and Quarterly Reports from 134 different archaeological sites. The bulk of the sites with curated materials were in the eastern portion of the state, with 47 in Le Flore County and 31 in Delaware County (Rogers 1978). Excavated contexts include villages, rock shelters, platform mounds, burial mounds, midden mounds, and cemeteries. Rogers (1978) speculated that the WPA identified many more than these 134 sites. He based this conclusion on Clements (1938a:2), who reported crews had found more than 100 sites in Delaware County by mid-1938. Work in Delaware County continued for nearly another two years after that report, so the count surely had risen by the time the project was shuttered.

Prior to the federal relief projects, precious little archaeology had been conducted in Oklahoma. The first scientific excavations in the state were directed by Joseph Thoburn (Wright 1947). Thoburn conducted excavations while he served as a professor in the history department at the University of Oklahoma from 1913 to 1917⁴ and continued fieldwork after he was appointed to direct the Oklahoma Historical Society (OHS) (La Vere 2007).⁵ Thoburn retired in 1931, just before Oklahoma relief archaeology programs began. He excavated most notably at Spiro, providing the only known photograph of the Craig Mound prior to its destruction by looters in the 1930s.⁶ Thoburn worked in Le Flore, Delaware, and Kay Counties, as well as in the Oklahoma panhandle (Wright 1947:404). Otto Spring served as Thoburn’s assistant, working out of Delaware County. Even after Thoburn retired, Spring continued to excavate at sites in eastern Oklahoma using Thoburn’s by then antiquated methods, much to the chagrin of the OU relief archaeologists. In the early 1930s, four precontact cultures drew the attention of Oklahoma archaeologists: the ‘slab-house’ culture in the Panhandle; the Ozark bluff-shelter culture in the northeastern section of the state; the Caddo house mound culture in the east central and central sections; and the moundbuilder culture on the Arkansas and its northern tributaries in eastern Oklahoma (Bauxar and Bell 1950:2). These categories of interest were drawn directly from Thoburn’s work with OU and the OHS.

The relief excavations summarized in this volume began in January 1934 at the Norman site in Wagoner County. The Norman excavations were funded by the CWA (Finkelstein 1940b). The Oklahoma relief efforts were not part of the broader Smithsonian Institution CWA initiative. Correspondence from the time indicates that the project funds were funneled from a state FERA office as part of the relief allocation for Wagoner County (Eddleman 1935). Norman was selected for the effort because at the time it was the only known site with large mounds of a manageable size to complete excavation in the time allotted.⁷ The CWA paid for a project director and a very limited crew of two laborers, locals from Wagoner (Finkelstein 1934b:3).

As the CWA ended operation in March 1934, funds were subsumed back under the broader authority of the FERA; under this relief package, work at Norman continued and excavations began at a slab-house site in Texas County. The FERA project employed approximately 30 people (Reed 1938). In 1936, the WPA funded additional excavations at Norman, at Red River drainage precontact Caddo sites in Choctaw County, and finally at Spiro. The Spiro excavations began in the summer of 1936 with the cleanup of the looter-ravaged Craig Mound, and crews continued to work on Spiro and outlying related sites in the surrounding Fort Coffee region for three more years. The scope of the WPA archaeological project grew with time, as excavation projects expanded into northeast and central Oklahoma. In mid-1938, the project employed 140 people (Reed 1938:3). Through February 1937, the federal government had provided $18,000 to the project. Between February 1937 and June 1938, that number increased by $75,000. The OU and the University of Tulsa contributed a total of $11,000, and oil baron Frank Phillips made a generous $6,000 donation to the project (Phillips Bequest 1938:4; Reed 1938:3).

The archaeological project employed more than 200 workers at its height in late 1938. At this time, crews in eastern Oklahoma were concentrated in three areas—northwest of Spiro near Muskogee, at outlying sites around Spiro, and in far northeastern Oklahoma. From this peak until its end in 1942, field crews and funds steadily dwindled. With Spiro exhausted, the field supervisors were forced to scramble to find new archaeological sites to keep crews working, chasing down leads across the eastern portion of the state. As the archaeological field crews shrank in size in 1939, laboratory and other operations in Norman were established and expanded. In Norman, the project employed both laboratory analysts and artists tasked with illustrating artifacts and building exhibits for public consumption. The exhibits and art showed the value of the federal funds directed at the project and educated the public about Oklahoma’s past.

The Oklahoma Quarterly Reports

This book focuses on the results of WPA excavations conducted at seven sites in the Arkansas drainage and incorporates published data from two additional mound sites, Spiro and Harlan (Figure I.3). All nine of these sites have mounds constructed between AD 1050 and 1400. Table I.1 lists the WPA sites, major features, dates of excavations, and field supervisors. Excavations at the WPA sites occurred from 1934 to 1940. The information presented is drawn largely from the Quarterly Reports required after Petrullo assumed control of federal WPA archaeology in 1938. The Oklahoma Quarterly Reports include descriptions of features excavated, site maps, mound profile and artifact sketches, and some results of field analysis of artifacts. Field supervisors were responsible for submitting these reports to project director Forrest Clements at OU, who then submitted them to Washington. The Quarterly Reports seem to have created a headache for the field supervisors, as evidenced by some of their correspondence. Supervisor Phil Newkumet writes to Kenneth Orr in a 1939 letter, How you doing? Finished quarterly report? Why ain’t you writ? Why ain’t I? (Newkumet 1939d).

The Quarterly Reports were well received in Washington. Clements brags,

Our report was received in Washington with considerable enthusiasm and I think we should keep up the detail of these reports. Naturally I know, and you know that they fall far short of perfection, but apparently our last report was superior to anything else that had been sent in. Let’s try to make the next one much better (Clements 1939d).

As time wore on, Clements had more and more trouble getting the Quarterly Reports from the field supervisors. He prodded supervisor Phil Newkumet twice over the December 1939 report, writing first, As you know your quarterly report is due and I wish you to complete this one even if you have to cut it short. If you have to omit certain items, these can be included in your next quarterly report. It is extremely important that we get these reports into Washington with the least possible delay (Clements 1939o). In February 1940, Clements (1940b) writes again, stating, We cannot find your quarterly report and suppose you did not have it finished when you were here. The report is past due and we must send it to Washington in the near future. Please send it to us as soon as possible.

While we use Spiro as a point of comparison for these sites and draw upon the experiences of the Spiro WPA field supervisors, we do not discuss the WPA work at Spiro. As they were not required at the time, there are no periodic progress reports summarizing the Spiro excavations. Esteemed archaeologist James A. Brown (1996) spent more than 50 years working through the notes and photographs from the WPA excavations and summarized this work in a two-volume monograph. We also use the Harlan site, a multiple mound center excavated in the late 1940s/early 1950s by Robert E. Bell (1972), as a basis for comparison. Harlan is second in size only to Spiro in the Arkansas drainage. Bell’s excavations were far better documented and employed more modern methods than the WPA projects, making them highly useful as a comparative tool.

Our aim with this narrative is to provide a history of the WPA excavations at the eastern Oklahoma mound sites, explain the methods employed by the WPA archaeologists, and place these seven mound sites in the Arkansas drainage in modern archaeological context. While a great deal of research attention has been rightly lavished upon the remarkable discoveries at Spiro, particularly the Craig Mound, the many related sites have been considered only sporadically; no modern synthesis of these sites has been attempted. Because of this, Spiro falsely exists in a vacuum, with little being understood about relationships between the largest and best-known site and contemporaneous mound centers in the surrounding region.

Before we delve into descriptions of specific sites, we provide additional project background. Chapter 1 introduces the project managers, field supervisors, and crews who participated in the excavations and uses correspondence, journals, and reports on file to reconstruct life as a WPA field supervisor in 1930s Oklahoma. The laboratory in Norman also contributed progress reports detailing the number of artifacts processed, restored, and illustrated and describing other special projects, including exhibit construction and the production of a film promoting the archaeology project. Chapter 2 examines field and laboratory methods and describes special projects financed by the WPA, including an archaeological film and various exhibits. Chapters 3 through 9 include the Quarterly Reports, written by the field supervisors, and our own background and interpretations of excavations at the seven mound sites. Our aim is to present the primary data on what was recovered by the WPA and to elevate these sites beyond their current status as dots on a map to archaeologists working outside of Oklahoma.

As mentioned, the earliest New Deal mound excavations occurred at Norman, in Wagoner County. A 1934 report from these excavations is on file at SNMNH. Those excavations are covered in Chapter 3. The first mound site excavated after Spiro, covered in Chapter 4, was the Hughes site in Muskogee County. After the Hughes excavations were complete, the project shifted east to Cherokee County and the Brackett site, described in Chapter 5. The next excavated site was Eufaula, located farthest away from Spiro, well to the west, and chronicled in Chapter 6. Excavations at the Skidgel Mound, just a few kilometers west of Spiro, occurred sporadically from mid-1938 through December 1939 and are described in Chapter 7. At the same time as the eastern Oklahoma mound sites were being worked, field crews in northeast Oklahoma worked on rock shelters, mound sites, and villages in advance of the creation of a reservoir on the Grand River. Two mound complexes were excavated in northeast Oklahoma: Reed/Huffaker, described in Chapter 8, and Lillie Creek, described along with an associated residential site in Chapter 9. Chapter 10 summarizes the WPA projects and provides a new synthesis and interpretation of the AD 1100–1700 Arkansas Valley.

The young field supervisors who oversaw excavations in various regions of the state were responsible for pulling the Quarterly Reports together from the field. Many were written while excavations were ongoing. For projects with multiple Quarterly Reports, we created a cohesive narrative of a given site by stitching together multiple reports. While we have edited the reports for grammar, refined the organization, and changed some of the archaic archaeological terms,⁸ the words of the Quarterly Reports are the authors’ own, and not ours. We provide the contextual introductions and the interpretations that place these sites in context. The archaeological data summarized in these reports are immensely useful for understanding the late precontact Arkansas drainage. Much of it has not seen the light of day since WPA operations shuttered, which is why we have synthesized this information in this book.

The Arkansas Drainage of Eastern Oklahoma

From its headwaters in the Colorado Rockies, the Arkansas River flows east more than 1,400 miles to its junction with the Mississippi River. It flows eastward into Kansas and then dips south into Oklahoma near Ponca City. It flows southeastward and enters Arkansas at Fort Smith. Several major tributaries join the Arkansas in eastern Oklahoma. Each of these drainages also had significant archaeological sites. The Grand River flows south/southwest from Kansas, where it is known as the Neosho River. Both it and the Verdigris River join the Arkansas near the present-day town of Muskogee. Norman, in Wagoner County, and Reed and Lillie Creek, in Delaware County, are all along the Grand River. Norman is now submerged under Fort Gibson Lake, and Reed and Lillie Creek are under Grand Lake o’ the Cherokees. Hughes, in Muskogee County, is just south of the junction of the Verdigris and Arkansas. It is not inundated, but it is in a residential area and has been heavily impacted by decades of modern construction. The smaller Illinois River flows southwest from the Arkansas Ozarks and joins the Arkansas River in Sequoyah County. Brackett is in Cherokee County on the Illinois River and is not submerged. Harlan, which is partially submerged, is also in Cherokee County, on a tributary that joins the Grand just downstream. Just downstream from the Illinois-Arkansas confluence, the Canadian River conflues with the Arkansas in McIntosh County. The Canadian flows primarily eastward across central Oklahoma before joining with the parallel North Canadian River and the Arkansas. Eufaula once was at the confluence of the Canadian and North Canadian and is now submerged by Eufaula Lake. Skidgel (and Spiro) are both along the Arkansas, near the Oklahoma-Arkansas state line.

The WPA mound sites are situated in three biotic districts, as identified by Blair