Maize Histories

The spread of maize from its initial area of domestication in the Balsas River Valley in south-central Mexico approximately 8,700 years ago (Piperno et al. Reference Piperno, Ranere, Holst, Iriarte and Dickau2009) has long interested researchers, given the social and ecological transformations that tended to accompany this crop. North of Mexico, maize entered the economies of indigenous foragers in the US Southwest by approximately 2500–2000 BC and was part of the package of cultural traits that moved with Basketmaker II phase farming groups early in the first millennium AD (Hanselka and Vierra Reference Hanselka, Vierra and Vierra2017). The pathway(s) that maize took into the Plains is currently unclear, but macrobotanical evidence in the eastern Plains points to its appearance between cal AD 688 and 977 (Adair Reference Adair2012). Recent microbotanical evidence suggests that maize was part of Indigenous systems in the eastern Plains as early as cal 361–197 BC (Adair et al. Reference Adair, Duncan, Young, Bozarth and Lusteck2022). This apparent lag signals a similar discrepancy between macrobotanical and microbotanical remains in the Eastern Woodlands.

Recent work by Mary Simon and colleagues (e.g., Emerson et al. Reference Emerson, Hedman, Simon, Fort and Witt2020; Simon, Hollenbach, and Redmond Reference Simon, Hollenbach and Redmond2021) and John Hart and coworkers (e.g., Hart Reference Hart2022; Hart and Lovis Reference Hart and Lovis2013; Hart et al. Reference Hart, Thompson and Brumbach2003, Reference Hart, Brumbach and Lusteck2007) has provided considerable insight into maize histories in the American Bottom and Great Lakes regions, respectively. Simon (Reference Simon, Raviele and Lovis2014, Reference Simon2017; Simon, Hedman, and Emerson Reference Simon, Hedman and Emerson2021) has demonstrated that maize is rare in the American Bottom before AD 900/1000, reshaping a common narrative in which gradually increasing amounts of maize were an integral part of Woodland period cultural evolution in this region. Even maize cited at Middle Woodland sites has been shown to be intrusive or originally misidentified as maize (Simon Reference Simon2017; Simon, Hedman, and Emerson Reference Simon, Hedman and Emerson2021). The sole Woodland maize date from western Illinois broadly that stands up to scrutiny was recovered from the Edgar Hoener site and dates between cal AD 657 and 775 (see Simon Reference Simon, Raviele and Lovis2014). This new perspective on early maize use fits well with what we are finding in the MOV.

Expanding our understanding of early maize use, Hart and colleagues (Reference Hart, Thompson and Brumbach2003, Reference Hart, Brumbach and Lusteck2007) identified maize phytoliths in residues from pots recovered from central New York dating to as early as 300 BC. Macrobotanical evidence is less common, but sites like Grand Banks in southern Ontario have produced directly dated maize cupules as early as cal AD 330–649 (Crawford et al. Reference Crawford, Smith and Bowyer1997:114).Footnote ¹ These early dates on the opposite side of the continent from the south-central Mexican heartland of maize domestication have led some scholars to question their validity or at least find these dates confusing (Emerson et al. Reference Emerson, Hedman, Simon, Fort and Witt2020). However, considered broadly, these data offer an intriguing case for an early introduction of maize into the Great Lakes region.

The different types of evidence (i.e., macro- versus microbotanicals) create a situation in which we may talk past each other; yet, taken at face value these studies demonstrate some of the complexities in understanding the varying pathways that maize took as it entered subsistence systems of Indigenous foragers in the Eastern Woodlands before its establishment as a focal aspect of agricultural systems around AD 1000. The MOV reflects one possible avenue of movement between these regions and provides an otherwise missing perspective. The Ohio River Valley has long been seen as a thoroughfare for the movement of people and ideas (e.g., Griffin Reference Griffin1952). At least in the early years of the Late Precontact period (around AD 1000–1300), people were moving between the Great Lakes, MOV, and the American Bottom regions (Cook Reference Cook2017; Cook and Price Reference Cook and Douglas Price2015), and there is also some intriguing evidence for movement during the centuries preceding this period (Cook and Price Reference Cook and Douglas Price2015).

Now that remains from the Edwin Harness Mound in Ohio have been shown to be plants other than maize (Simon, Hollenbach, and Redmond Reference Simon, Hollenbach and Redmond2021), there are no directly dated maize macrobotanical remains from solid contexts before AD 1000 in the MOV. More common are secondary lines of evidence, including maize contextually associated with charcoal dates and carbon isotope evidence from human bone. For example, features from the Woods site in West Virginia containing maize date between cal AD 569 and 775 (Shott Reference Shott1990; Wymer Reference Wymer and Seeman1992, Reference Wymer, Johannessen and Hastorf1994). Similarly, maize is reported from Woodland contexts at the Sand Ridge site in southwest Ohio (Riggs Reference Riggs1998), but stratigraphic mixing at the site likely means that these kernels were intrusive. Direct analysis is necessary to confirm these dates and ensure that they are not examples of contamination or of non-maize plants like those identified elsewhere. Additionally, isotopic evidence from northeast Ohio indicates the presence of maize in the diets of individuals around cal AD 775–1022 (Redmond Reference Redmond2012:124). More work is clearly needed to confirm site-specific and regional patterns regarding the possibility of pre–AD 900/1000 maize use in the region. The Turpin site in southwest Ohio provides a compelling place to begin that work (Figure 1).

Figure 1. Map of sites and areas mentioned in this study.

Maize from Turpin: Spanning Late Woodland and Early Fort Ancient Time Periods

As part of our ongoing work to better understand the emergence of maize agriculture in the MOV, we conducted field and collections-based research at the Turpin site over the last decade (Comstock Reference Comstock2017; Comstock and Cook Reference Comstock and Cook2021; Cook Reference Cook2017). This multicomponent site contains both preagricultural (Late Woodland; around AD 400–1000) and Late Precontact Fort Ancient culture (around AD 1000–1300) occupations and is the type-site for the post-Hopewell Newtown phase (around AD 400–700) in the region (Griffin Reference Griffin1952; see also Comstock Reference Comstock2017; Cook Reference Cook2017). Excavations by the then Cincinnati Museum of Natural History (CMNH) revealed intensive Newtown Phase occupations, termed the “Old Village” component, that also purportedly contained maize (Oehler Reference Oehler1973). To determine whether these maize remains were intrusive like much of the maize recovered from Woodland period contexts in the American Bottom (Simon Reference Simon, Raviele and Lovis2014), six maize cob fragments were selected from CMNH collections for dating at the NSF Arizona AMS Laboratory: three from “Old Village” contexts and three from Fort Ancient contexts (see Cook Reference Cook2017).

The results of this analysis are consistent with the contexts from which the samples were taken (Table 1; Figure 2). Cob fragments sampled from the three Newtown phase “Old Village” contexts all resulted in pre–AD 1000 distributions. One date from the “2nd Old Village,” a context that we interpret to reflect a separate occupation based on CMNH parlance, produced a 12-row maize cob fragment that dates between cal AD 552 and 649, making it one of the earliest directly dated maize cobs in the Eastern Woodlands. Two additional cobs date to the centuries before AD 1000, although their accuracy suffers from the nature of the calibration curve during this period. One eight-row cob fragment dates between cal AD 684 and 973 and a 10-row cob fragment dates between cal AD 774 and 974. Note that these pre–AD 1000 distributions do not overlap calibrated ranges associated with Fort Ancient contexts and were recovered from contexts with distinctly different material culture. We thus interpret these early maize dates as earlier than established villages of maize agriculturalists in our region. Maize cob fragments sampled from suspected Fort Ancient contexts date between cal AD 1040 and 1275 (Table 1) and are consistent with other dated contexts from the site and the early Fort Ancient period more broadly (see Comstock Reference Comstock2017; Comstock and Cook Reference Comstock and Cook2021; Cook Reference Cook2017).

Table 1. Maize Dates and Contexts from CMC Excavations at Turpin.

Notes: Calibrations performed using OxCal 4.4.4 with the r.5 atmospheric data from Reimer and colleagues (Reference Reimer, Austin, Bard, Bayliss, Blackwell, Ramsey and Butzin2020).

* Row numbers estimated by Kristen Gremillion.

Figure 2. Calibrated distributions of maize dated from the Turpin site.

The pre–AD 1000 maize dates presented here point to maize use at the Turpin site earlier than has been so far discovered in the MOV, and at least one is earlier than most maize macroremains found in the Midcontinent. Considered at a regional scale, these dates place the MOV in consideration for one of multiple early areas of maize use in the midcontinent. We should note that unlike some regions in which there is a lag between early maize evidence and bone isotope evidence for maize consumption, Turpin has produced evidence for early maize remains and early maize consumption. For example, Greenlee (Reference Greenlee2002:191) notes that bone collagen from half the sampled individuals interred in the stone mound, a mortuary facility at Turpin dated between cal AD 600 and 900, produced elevated δ¹³C values (between −9‰ and −13‰). Analysis of dental enamel from an individual in this context dated between AD 651 and 867 returned δ¹³C values of −14.7‰ (M1) and −14.5‰ (M3), indicating the presence of maize in their diet (Cook and Price Reference Cook and Douglas Price2015:125).

These findings fit well with subtle evidence that imply networks of connectivity during the AD 500–1000 period. For example, the discovery of Scallorn-like arrow points (Comstock and Cook Reference Comstock and Cook2021) could point to connections between the MOV and western parts of the Midcontinent. Additionally, similarities between some artistic motifs such as the Newtown gorgets (Seeman and Dancey Reference Seeman, Dancey, Emerson, McElrath and Fortier2000:598) and the Missouri “Jaguar” gorget (Wood 1999) could point to continued regional connections after the apparent dissipation of Hopewell networks. Taking this a step further, we could also see regionally diagnostic pottery forms like the well-known Newtown “angled shoulder” as a precursor to similar vessels in the American Bottom that Griffin (Reference Griffin1943) called “Cahokia shoulders”: indeed, Perino (Reference Perino1964) also suggested that some vessels, dating between about AD 500 and 900, were precursors to Mississippian angled-shouldered vessels. All these data can be used for testing a variety of hypotheses regarding Woodland and Late Precontact‒era connections throughout the Eastern Woodlands.

Toward Multiple Working Hypotheses of Maize in the Eastern Woodlands

Hart and colleagues (Reference Hart, Lovis and Anne Katzenberg2021) suggest that multiple lines of evidence are needed to refine our understanding of regional maize histories. Disparities in the lines of evidence used and possibly in the lines of evidence available in each region make it difficult to form interregional comparisons. Yet these issues also provide opportunities to work together to solve this problem. For example, pan-regional research standards that incorporate analyses of macrobotanical remains, microbotanical remains, and pottery residues would provide multiple lines of evidence while leveling the analytical playing field. Currently, however, the disparities in methods applied to the problem do not allow us to meaningfully parse regional differences.

We also suggest that, as maize histories continue to be considered, refined, and integrated, multiple working hypotheses (sensu Chamberlin Reference Chamberlin1890) are needed to incorporate complex and varied outcomes. This approach allows us to consider a variety of possibilities while working to foster interregional research. We initiate this process by considering four hypotheses (of many possibilities) that could account for the complex maize histories evident.