Scraping the Bowl: New Data on Prehistoric Tobacco Use in the Southeast

30 Days of Tennessee Archaeology 2016, Day 30

Stephen B. Carmody1, Ryan Hunt2, Jon Russ2, Jera R. Davis,3 Natalie Prodanovich,2 Benjamin McKenzie,1 and Christopher Van de Ven1
1 Department of Earth and Environmental Sciences, Sewanee: University of the South
2 Department of Chemistry, Rhodes College
University of Alabama, Office of Archaeological Research

In last year’s 30 Days of Tennessee Archaeology blog post, Secrets in the Smoke: Prehistoric Tobacco Use in Tennessee, we discussed the results of both pollen analysis and gas chromatography/mass spectrometry (GC/MS) analysis on archaeologically-recovered pipes and pipe residues from the state of Tennessee. Thanks to research funds provided by the Eastern Band of Cherokee Indians, we are happy to report new findings. Having nearly exhausted the collections at the McClung Museum of Natural History and Culture at the University of Tennessee, Knoxville, we were fortunate to gain additional access to pipes from sites in Alabama, Georgia, Mississippi, and North Carolina. We selected pipes based upon cultural and temporal affiliation and for the amount of residue. Once selected, residues were carefully removed, so as not damage the pipes. Afterwards, faculty and students from the Department of Chemistry at Rhodes College in Memphis, Tennessee extracted organic compounds from the residues and analyzed them using GC/MS. This year we sampled 55 pipes, of which 35 produced positive evidence of tobacco use.


Pipe recovered from the Seven Mile Island site (1LU21).



Pipe fragment recovered from the unnamed 1MS121 site in Marshall County, Alabama.

Thirty-one Alabama pipes comprised the majority of our sample. These are curated at the University of Alabama, Office of Archaeological Research. Twenty-seven derive from the Moundville site in west-central Alabama and four derive from three Tennessee River Valley sites in north Alabama: 1La14, 1Lu21 (see above), and 1Ms121 (see image to right). With the exception of the pipe from 1La14, a Copena mound site, all pipes are Mississippian in age. In total, sixteen of the Moundville pipes and all four Tennessee Valley pipes tested positive for nicotine.


We also tested two pipes currently housed at the Fernbank Museum of Natural History in Atlanta, Georgia and one other curated at the Antonio J. Waring Jr. Laboratory at the University of West Georgia.

The two Fernbank pipes come from the Glass site in Telfair County, Georgia, a late prehistoric Native American site (ca. AD 1450-1550) visited by Spanish conquistador Hernando de Soto. Dr. Dennis Blanton of James Madison University, Virginia has been excavating at Glass since 2006. One of the two pipes from Glass tested positive for nicotine.

One pipe fragment from the collections at the Waring Lab was also submitted for analysis. The first came from the Buzzards Roost site, a historic Creek town located in Taylor County, Georgia. The fragment produced a positive nicotine signature.

North Carolina


Pipe fragment recovered from the Biltmore Mound. Image courtesy of Dr. Larry Kimball.

Dr. Larry Kimball, Professor of Anthropology and Director of the ASU Laboratories of Archaeological Science at Appalachian State University in Boone, North Carolina, provided us with a sample from his excavations at the Biltmore site (31BN174) in Asheville, North Carolina. The pipe was recovered from a posthole in the center of a Middle Woodland Connestee phase platform mound, indicative of the Hopewellian culture in western North Carolina, eastern Tennessee, and northern Georgia (see profile below). The Middle Woodland date for the pipe makes it possibly the earliest evidence of tobacco use in North Carolina and the Appalachian region.


Drawing of Biltmore Mound showing Posthole 21, where the pipe fragment was recovered. Image courtesy of Dr. Larry Kimball.


Six pipes were also made available to our study from the Feltus site (22JE500) located in Jefferson County, Mississippi. The three-mound site dates to the late Baytown through middle Coles Creek periods (AD 700 – 1100). From 2006 to 2012, faculty and students from the Research Laboratories of Archaeology at the University of North Carolina, Chapel Hill excavated at Feltus and they are hoping to return in 2017. Each of the six pipes submitted for analysis tested positive for nicotine.



Pipe tested from the Fains Island site (1JE1).

Six pipes sampled from the collections at the McClung Museum produced evidence for the use of tobacco. Five of these pipes were recovered during archaeological excavations, and one was donated to the museum from a once privately-held collection, the Barnes Collection.

Three of the pipes were recovered from sites excavated during the Chickamauga Basin Project prior to the construction of the TVA Chickamauga Reservoir, between 1936 and 1939 (Lewis, Lewis, and Sullivan 1995). The pipes were recovered from the Dallas site (8HA1), the Hixon site (1HA3), and the Ledford Island site (16BY13).


Pipe analyzed from the Toqua site (40MR6).

The Fains Island site (1JE1 – see image above) was excavated in 1935 as part of the Douglas Reservoir project. The site is a Mississippian period open-air habitation site with a platform mound. Nicotine was also detected in the residue of a pipe recovered from this site.

Toqua site (40MR6) in Monroe County, Tennessee, is an open-air habitation site with Mississippian and Cherokee components. A pipe recovered from a proto-historic feature tested positive for nicotine.

A pipe from the Barnes Collection donated to the McClung museum (see image below) was also analyzed and produced evidence for tobacco use.

In total, we analyzed 55 pipes from 17 archaeological sites. Thirty-five showed evidence for tobacco use at 13 different sites. Previous evidence for tobacco use at these seeds consists of a single seed from a mound at the Moundville site (Knight 2010). We believe that both our success rate and the new cultural patterns emerging demonstrate the value of GC/MS analysis.


Pipe from the Barnes Collection housed at the McClung Museum, UTK.

Since the beginning of our project, we have tested 75 pipes from 21 sites for either tobacco pollen or for a nicotine signature. Forty pipes from 17 sites across the Southeast have tested positive for tobacco use. Moving forward, we hope to continue building our database. Eventually we hope to have enough spatial and temporal data to discuss differences in the use of tobacco over time and space, explore the role of other plants in the smoking complex, and better understand the timing and transmission of tobacco into the region.


Map showing sites where evidence of nicotine was discovered.