2help

The functional dimensions of earth oven cooking: An analysis of an accidently burned maize roast at the C. W. Cooper site in West-central Illinois

Gregory D. Wilson, Amber M. VanDerwarker

University of California, Santa Barbara

Despite the prevalence of earth ovens, a subclass of pit features used for cooking, there is little consensus regarding how these cooking features were used or about the foodstuffs that were prepared in them. To provide a more detailed understanding of earth oven cooking in the archaeological record, we analyze the archaeobotanical contents, stratigraphy, and morphology of a cooking pit recently excavated at C. W. Cooper, an early Mississippian (A.D. 1150–1200) site in west-central Illinois in order to contribute to research on late Prehistoric foodways. Filled with nearly 100 ears of maize, this earth oven presents the opportunity to document the process of undertaking a maize roast. The volume of maize and its presence within a dense concentration of cooking, processing, and storage facilities allows us to consider the communal nature of outdoor earth oven cooking in the 12th century Central Illinois Valley and the socioeconomic dimensions of commensal politics more broadly during the Precolumbian era.

Keywords: Mississippian, earth oven, maize, pit features, cooking techniques, Midwestern archaeology

Introduction Pits are common archaeological features that were

once used to store and process foodstuffs in the

ancient American Midwest and Midsouth. Archaeo-

logical analyses of pit features have contributed

substantially to understandings of late prehistoric

foodways in the North American Midcontinent

(Binford et al. 1970; Fortier 1983; Harris 1996; Holt

1996; Kelly et al. 1987; Koldehoff and Galloy 2006;

Stahl 1985). For example DeBoer’s (1988) famous

study of the Mississippian pit features revealed that

subterranean storage features could have been used

to conceal agricultural surpluses from central admin-

istrators seeking to extract tribute payments.

Moreover, Bardolph’s (2014) recent analysis of early

Mississippian (A.D. 1100–1150) pit features in the

southern Central Illinois River Valley demonstrated

that local groups maintained traditional Woodland

era conventions of food storage and preparation

while adopting an array of other Mississippian

practices and forms of material culture. This paper

contributes to archaeological research on pit features

and foodways through an investigation of the

functional dimensions of earth ovens, a subclass of

pit features used for cooking. Earth ovens are

generally identified based on the presence of oxidized

soils, burned fill zones, and concentrations of fire

cracked rocks. Despite their prevalence in the

archaeological record in the eastern United States,

there is little scholarly consensus regarding how these

cooking features were used or the kinds of foodstuffs

that were prepared in them. To highlight both the

antiquity and ubiquity of this form of cooking we

draw on abundant ethnographic, ethnohistoric, and

archaeological evidence of earth oven cooking in

other areas of the New World.

We add to this body of literature on earth ovens

through an analysis and discussion of the contents

and structure of a cooking pit recently excavated at

the early Mississippian C. W. Cooper site (A.D. 1150–

1200), a small bluff top village in west-central Illinois

(FIG. 1). Specifically, we consider the archaeobotani-

cal assemblage recovered from inside this pit, in

addition to evaluating the pit’s stratigraphy and

morphology. Filled to the brim with nearly 100 ears

of maize, this earth oven feature allows us to

document the process of undertaking a maize roast.

The volume of maize and the location of this feature

within a dense concentration of cooking, processing,

and storage facilities also enables a consideration of

Correspondence to: Amber M. VanDerwarker, Department of Anthropology, University of California, HSSB 2001, Santa Barbara, CA 93103-3210. Email: [email protected]

166 � Trustees of Boston University 2015 DOI 10.1179/0093469015Z.000000000118 Journal of Field Archaeology 2015 VOL. 40 NO. 2

the communal nature of outdoor earth oven cooking

in the Central Illinois Valley during the 12th century.

The Antiquity and Use of Earth Oven Features in the New World The earliest evidence of earth ovens in the New

World dates between 8500–8000 B.C. and can be

found in Alaska, the Pacific Northwest, and the

Great Plains (Thoms 2008, 2009). In addition to these

early cases, Thoms (2008, 2009) has documented

abundant cases of earth oven/hot rock cooking

throughout the western United States, combining

ethnographic and archaeological evidence to achieve

a fair amount of resolution, thus allowing for the

identification of these features on archaeological sites

(Fish et al. 1985). Earth ovens have also been

documented in the archaeological and ethnographic

records of Mesoamerica (Dering 1999; Leach and

Sobolik 2010; Salazar et al. 2012) and South America

(Van Den Bel 2010). Unfortunately, while there is

abundant archaeological evidence of earth oven use

in the eastern United States, ethnographic evidence of

this type of cooking technology is generally lacking

from the eastern United States.

Most ethnographic evidence of earth oven cooking

in the New World highlights the cooking of meat and

plant products high in fiber and carbohydrates. The

majority of documented plant species include root

and tuber plants, or those high in toxicity, such as

Camassia, Allium, and Agave (Fish et al. 1985; Leach

and Sobolik 2010; Wandsnider 1997). By cooking

these plants in the earth oven, they ‘‘render foods

more nutritious and digestible’’ (Thoms 2009: 573).

Moreover, the general size of the cooking feature is

Figure 1 Select sites from the Central Illinois River Valley.

Wilson and VanDerwarker An analysis of a maize roast at the C.W Cooper site, IL

Journal of Field Archaeology 2015 VOL. 40 NO. 2 167

often correlated with the amount of food being

prepared, thus their size can be used to infer the scale

of the event for which the food is meant (Salazar et al.

2012; Thoms 2008, 2009).

Recent ethnobotanical and ethnographic research

in the Maya lowlands by Salazar et al. (2012)

documents the roasting of whole maize ears in

traditional Maya earth ovens. The context of this

particular maize roast was a first fruits feast

celebrating the initial maize harvest of the season

(Salazar et al. 2012: 291). The earth oven was dug to a

depth of about 60 cm, the deepest of all the ovens

documented in the study, which is not surprising

given the communal nature of the event. A fire was

built inside the pit, and once the fire died out, the

cooks covered the coals with leaves, on top of which

were placed the maize ears. The maize ears were

slightly wet to retain moisture during the cooking

process. More leaves were then laid atop the row(s) of

maize, following by an earthen cap, after which the

oven was left unattended overnight, or for several

days, depending on the level of heat (Salazar et al.

2012: 291).

Ethnographic and ethnoarchaeological accounts

closer to the study area provide more specific insight

into how ancient people of the Midsouth would have

used earth ovens (Black and Creel 1997; Hough 1926;

Peacock 2008; Teit 1900; Thoms 2008). Our survey

of this literature highlights two main methods of

heating: indirect, placing heated rocks on the ground

surface or in a shallow pit for slow cooking, and

direct, building a fire inside a pit. The indirect method

was practiced by the Chopunnish (Nez Perce) in the

Pacific Northwest:

[T]he Indians prepared a large fire of dried wood, on which was thrown a number of smooth stones from the river. As soon as the fire went down and the stones were heated they were laid next to each other in a level position and covered with a quantity of branches of pine, on which were placed bear flesh, the boughs and the flesh alternated for several courses, leaving a thick layer of pine on the top. On this heap was then placed a small quantity of water, and the whole covered with earth to the depth of 4 inches. After remaining in this state about three hours the meat was taken off (Elliot Coues 1893 [in Hough 1926: 36]).

There are also ethnographically documented exam-

ples of indirect cooking techniques in which a fire is

built over the floor of a rock-lined pit. Foodstuffs

were then laid overtop the stones once the fire burned

down to coals (Hough 1926: 37).

An example of the direct method comes from an

observation of a group of California Indians. In 1877

Stephen Powers (in Hough 1926: 37) wrote that

‘‘…under stress of hunger [the Tlelding] ate soaproot

(Chlorogalum pomeridianum), the poisonous properties

of which they extracted by baking it in large quantities

in the ground, covering it over with green leaves and

building a fire over it, which was allowed to burn many

hours, when the root is said to be sweet and palatable.’’

Wandsnider’s (1997) ethnographic cross-cultural study

of pit cooking technology found that the use of these

two methods largely depended on the types of food

being cooked. Fires were often built inside pits for the

purpose of high temperature cooking. Alternatively,

heated rocks covered with layers of vegetation were

often used to prepare foods requiring a moist cooking

environment. However, sometimes the two methods

were used in conjunction for a long duration and high

temperature cooking.

Several scholars have attempted to archaeologi-

cally differentiate between these direct and indirect

cooking techniques (Binford et al. 1970; Fortier 1983:

120–124; Koldehoff 2002; Stahl 1985). Research by

Binford et al. (1970) at the Hatchery West site in

Clinton County, Illinois has provided one of the

earliest and most detailed investigations of Late

Woodland (A.D. 500–1000) and Mississippian (A.D.

1000–1400) earth ovens. Two subclasses of earth

ovens, deep and shallow, were identified at the site.

Deep earth ovens averaged 65 cm in depth and were

flat-bottomed pits with oxidized floors and side walls.

Shallow earth ovens averaged 43 cm in depth and

were generally basin-shaped in profile; these pits

differed from deep earth ovens in that they lacked in

situ evidence of soil oxidation but instead possessed

traces of redeposited, oxidized soils and other heat-

altered materials mixed into different fill layers.

Binford and colleagues (1970) interpreted the deep

earth ovens to be indirect cooking facilities and the

shallow earth ovens to be direct cooking facilities.

Recent experimental research, however, indicates

that these functional assignments may be incorrect;

Koldehoff (2002) observed experimentally that cook-

ing fires placed directly on pit floors resulted in a

pattern of oxidized subsoil that penetrated several

centimeters along their basal walls. Moreover, this

pattern of oxidation only occurred after a pit was

exposed to a high temperature fire for several hours

(Koldehoff 2002: 190). Thus, the presence of oxidized

soils in the deep earth ovens at Hatchery West

indicates they were probably direct cooking facilities.

It is also likely that the shallow earth ovens lacking

oxidized soils were used for either indirect cooking or

short-term direct cooking activities.

The excavation of a number of deep earth oven pits

with straight walls and flat bases at the Late

Woodland Mund site (A.D. 500–650) in the Ame-

rican Bottom region of western Illinois has provided

additional detailed information on the identification

of earth ovens (Fortier 1983: 120–124). These features

were morphologically similar to the deep earth ovens

Wilson and VanDerwarker An analysis of a maize roast at the C.W Cooper site, IL

168 Journal of Field Archaeology 2015 VOL. 40 NO. 2

at Hatchery West but lacked similar evidence for soil

oxidation along their bases. A number of earth ovens

from the Mund site also each contained a discrete

stratum of burned limestone near the base. In several

cases, this burned limestone layer was found to lie

directly beneath a black greasy layer, which has been

interpreted as decomposed thatch or matting (Fortier

1983). Considering the presence of burned limestone

and the absence of oxidation, Fortier argued that

these represented indirect cooking facilities (Fortier

1983: 124).

It is interesting that the vast majority of the earth

ovens excavated at Late Woodland sites, such

as Range and Woodland Hill (A.D. 500–800) and

terminal Late Woodland (A.D. 800–1050) sites in the

American Bottom region, lacked oxidized soils

altogether. It appears that the act of placing heated

rocks and foodstuffs into a relatively deep pit may

have been the predominant cooking technique during

the Woodland occupation of this region. As we

discuss below, this trend does not appear to have

been the case for the Central Illinois River Valley,

located approximately 170 km north of the American

Bottom.

The C. W. Cooper Site Earth Oven It is clear from the above discussion that there is

significant variation among features identified as

earth ovens at archaeological sites located in Illinois

and in North America. At present it is difficult to

interpret this variability in functional terms, espe-

cially considering that most of the known earth ovens

in archaeological contexts represent disposal (not

cooking) in their final context. The recent excavation

of a pit feature containing the remains of an

accidently-burned maize roast at the C. W. Cooper

site in Fulton County, Illinois provides the opportu-

nity to better understand the functional dimensions

of these ancient cooking facilities.

The C. W. Cooper site is located on the western

bluffcrest of the Illinois River Valley (FIG. 1). Pre-

vious excavations at the site in 1982 and 1983 by the

Upper Mississippi Valley Archaeological Research

Fund uncovered portions of a small Mississippian

village dating to the latter half of 12th century A.D., as

well as a later Oneota village dating to the early 14th

century. Surface artifact distributions cover an

estimated 3 ha although the densest concentration

of subsurface features at the site is restricted to a

3000–4000 sq m area that begins at the edge of the

bluff and extends northwards. Further excavations

funded by the National Science Foundation were

conducted at the site between 2011 and 2013. These

excavations encompassed 204 sq m of the site,

uncovering portions of seven early Mississippian

period (A.D. 1150–1200) buildings and 47 pit features.

A complex of five superimposed pit features,

including the earth oven in question (Feature 13),

was found along the southern edge of these excava-

tions (FIG. 2). Feature 13 was located roughly in the

middle of this pit complex and dates to the

Mississippian period of the site. It was superimposed

by one pit (Feature 12) that was superimposed by

another (Feature 1). This earth oven had a circular

opening that was 120 cm in diameter. It extended

52 cm in depth with out-curving walls toward the top

and a flat base. This feature was excavated by natural

stratigraphic layers which allowed for a detailed

reconstruction of its use and abandonment.

The base of the Feature 13 pit consisted of a layer

of orangeish-red oxidized clay that extended several

centimeters into the surrounding subsoil (FIGS. 3, 4).

This zone of oxidation also extended up the lower

sides of the pit. Above this oxidized stratum was a

thin layer of charred fibrous plant material, resem-

bling wild grass, which appears to have lined the base

Figure 2 Map of the University of California, Santa Barbara

excavations at the C. W. Cooper site.

Wilson and VanDerwarker An analysis of a maize roast at the C.W Cooper site, IL

Journal of Field Archaeology 2015 VOL. 40 NO. 2 169

of the pit to facilitate cooking (FIG. 5). Atop this

fibrous plant material was a dense layer of charred

maize. This layer was removed in its entirety and

processed as several flotation samples totaling 107 L.

Multiple, intact charred cobs with kernels still

attached were present in this layer, as were thousands

of maize kernels, cupules, and smaller cob fragments

(FIG. 6). Unfortunately, these whole ears of maize

(complete with cobs and attached kernels) did not

remain intact during removal. Typically, these maize

ears would have been removed from the pit once fully

cooked; thus, the sheer abundance of burned maize in

this oxidized pit feature indicates that the pit was

abandoned during or after the cooking process and

was never cleaned out.

Our reconstruction of the cooking methods, based

on the pit’s stratigraphy and contents, indicates that

the maize was covered with a layer of clay upon

which a fire was built—as evidenced by wood

charcoal overlaying the clay. In its final form, this

layer of clay was burned until it had hardened,

suggesting that it had been intentionally laid in

uniform thickness over the maize as a protective

buffer from the fire. The near incineration of the

maize appears to have been the result of the intensity

or duration of the fire built above this clay layer. This

fire was evidently too hot or left unattended for too

long, resulting in the burning of the maize until it

became an inedible mass. More than 300 carbonized

wild grass and weedy seeds were identified above the

clay layer, and their abundance above the maize layer

strongly suggests that they were intentionally laid on

top of the clay layer, perhaps as kindling for the fire.

After it was discovered that the maize had burned

beyond edibility, the feature was capped with refuse

and abandoned.

The plant data A total of 11 flotation samples (107 L of soil) were

taken from Feature 13 representing the entirety of the

feature fill (with the exception of the fibrous plant

Limits of excavation

Burned maize

Burned clay

Refuse deposit

Refuse deposit Feature 13Feature 12

SS

Rocks

Ground surface Feature

9 Feature 10 Feature 1 SS

0 20 cm

Figure 3 Profile map of the Feature 13 earth oven at the C. W. Cooper site.

Figure 4 The oxidized clay layer overlying the burned maize

in Feature 13, C. W. Cooper site.

Figure 5 Fibrous plant material at the base of the Feature 13

earth oven, C. W. Cooper site.

Figure 6 Burned maize at the bottom of the Feature 13 earth

oven, C. W. Cooper site.

Wilson and VanDerwarker An analysis of a maize roast at the C.W Cooper site, IL

170 Journal of Field Archaeology 2015 VOL. 40 NO. 2

layer below the maize ears) (TABLE 1). Samples were

floated using a Model A Flote-Tech machine during

the summer of 2011; light and heavy fractions were

bagged separately and shipped to the University

of California, Santa Barbara (UCSB) Integrative

Subsistence Laboratory (ISL) for identification and

analysis. Although the materials from the light and

heavy fractions were processed and sorted separately,

data from the two fractions were combined for

analysis. According to standard practice, the light

fractions were weighed and then sifted through

2.0 mm, 1.4 mm, and 0.7 mm standard geological

sieves. Carbonized plant remains from both fractions

were sorted in entirety down to the 2.0 mm sieve size

with the aid of a stereoscopic microscope 10–4x.

Residue less than 2.0 mm in size was scanned for

seeds, which were removed and counted; in addition,

taxa encountered in the 1.4 mm sieve that were not

identified from the 2.0 mm sieve were also removed,

counted, and weighed. Maize cupules and acorn

nutshells were also collected from the 1.4 mm sieve as

these materials tend to fragment into smaller pieces

and can be underrepresented in the 2.0 mm sieve.

Botanical materials were identified with reference

to the macrobotanical comparative collection at

the UCSB ISL, various seed identification manuals

(Martin and Barkley 1961; Delorit 1970), and the

USDA pictorial website (2011). All plant specimens

were identified to the lowest possible taxonomic level

(TABLE 2), although taxonomic identification was not

always possible as some plant specimens lacked

diagnostic features altogether or were too highly

fragmented. As a result, these specimens were

classified as ‘‘unidentified’’ or ‘‘unidentified seed.’’

In other cases, probable identifications were made;

for example, if a specimen closely resembled a maize

cupule, but a clear taxonomic distinction was not

possible (e.g., the specimen was highly fragmented),

then the specimen was identified as a probable maize

cupule and recorded as ‘‘cf. maize cupule.’’ Once

the plant specimens were sorted and identified, we

recorded counts, weights (g), portion of plant (e.g.,

maize kernels versus cupules), and provenience

information. Generally, most of the seeds identified

in the samples were too small to weigh, and thus only

counts were recorded. Hickory nutshell and maize

remains were both counted and weighed. Seven of the

11 flotation samples produced exceptionally large

quantities of plant remains, yielding sample weights

in excess of 1000 g (with two exceeding 3000 g). In

order to deal with these large samples in a timely

manner, we recorded the total weight for each of the

11 samples and then subsampled them using a riffle

splitter. After the samples were split, the subsamples

chosen for analysis were also weighed. We used this

weight information to extrapolate figures for plant

weight, wood weight, and taxa counts/weights from

the subsample to the total sample. By dividing the

total sample weight by the subsample weight, we

arrived at a figure that defined the proportional

relationship between the sample and its respective

subsample. This figure was used as a multiplier for

the actual counts/weights recorded from the sub-

sample to determine the proportionally equivalent

counts/weights for the total sample.

In total, 52,966 maize fragments weighing 760

grams were identified in this feature (this does not

include the possible ‘‘cf.’’ identifications); more than

44,000 of these fragments were kernels, with the

remainder consisting of cupules and cob fragments

(TABLE 2). Measurements were taken on complete

maize kernels and cupules as part of a related project

(Gracer et al. 2013) and reveal that maize kernels

from Feature 13 are wide and crescent shaped (FIG. 7).

Cob fragments from the 1983–1984 excavations at

C. W. Cooper reveal the presence of 8-, 10- and

12-row maize varieties. These cobs, graciously pro-

vided by Dr. Lawrence Conrad, derive from screened

and hand-excavated contexts and were analyzed by

VanDerwarker in the UCSB ISL. Angle measure-

ments from the Feature 13 maize cupules confirm the

presence of 8-, 10-, and 12-row cobs, in addition to

cupules from the occasional 14-row cob; the majority

of cupules measured, however, come from 10- and 12-

row cobs, suggesting that the maize in this feature

may represent the Midwestern Twelve Row variety

Table 1 Inventory of flotation samples taken from Feature 13 at C. W. Cooper.

Catalog # Sample volume Total sample weight* (g) Plant weight (g) Wood weight (g)

6 26 3161.76* 345.66 9.50 10 10 208.62 1.80 0.85 18 13 485.55 5.36 2.83 19 6 180.79 3.50 0.69 23 10 1711.43* 2.01 0.63 24 6 1038.82* 2.15 0.55 25 11 3052.51* 17.27 1.46 28 4 2545.00* 46.93 3.76 29 1 324.60 2.95 0.27 30 13 1481.15* 167.76 6.68 31 7 1130.20* 201.19 2.33

*denotes that a subsample was taken.

Wilson and VanDerwarker An analysis of a maize roast at the C.W Cooper site, IL

Journal of Field Archaeology 2015 VOL. 40 NO. 2 171

common to the broader Illinois region (Fritz 1992).

Based on average kernel counts from 10- and 12-row

cobs of North American heirloom varieties (from the

UCSB ISL comparative collection), we estimate that

these kernels come from approximately 95 ears of

maize. If we assume a consumption rate of three ears

of maize per person, then this maize roast could have

fed about 31–32 people. Some of the cupules, however,

may derive from 8-row cobs, which would translate

into more ears of maize represented by the kernels.

Regardless of row number, it is clear that this maize

roast was meant to feed a large group of people, on the

order of several households.

While the maize represents 97% by count (99% by

weight) of the plant materials recovered from the

Feature 13, the remaining taxa identified nevertheless

bear mentioning (TABLE 2). A fair amount of nut-

shell fragments was also encountered, dominated by

hickory (Carya spp.), which was found scattered

Table 2 Inventory of plants identified in Feature 13 at C. W. Cooper (cf.5probable but uncertain identification).

Common name Taxonomic name Counts Weights (g)

MAIZE Maize cob fragment Zea mays 68 24.79 Maize cob fragment cf. Zea mays cf. 30 0.14 Maize cupule Zea mays 8826 32.52 Maize cupule cf. Zea mays cf. 6 0.01 Maize kernel Zea mays 44,074 702.42 Maize kernel cf. Zea mays cf. 9 0.09 Total maize (cf. not included) 52,968 759.73 NUTS Acorn Quercus spp. 28 0.13 Acorn meat Quercus spp. 2 0.08 Acorn cf. Quercus spp. cf. 14 0.1 Hazelnut cf. Corylus sp. 1 Hickory Carya spp. 318 3.49 Hickory cf. Carya spp. 4 0.08 Black walnut Juglans nigra 1 0.05 Total nuts 368 3.93 WEEDY SPECIES Amaranth Amaranthus spp. 6 Bedstraw cf. Galium spp. 4 Bulrush Scirpus spp. 19 Bulrush cf. Scirpus spp. cf. 11 Carpetweed Mollugo spp. 1 Cheno/am Chenopodium/Amaranthus 1 Chenopod Chenopodium spp. 310 Goosegrass Eleusine indica 2 Grass family Poaceae 173 Knotweed/smartweed Polygonum spp. 9 Knotweed/smartweed cf. Polygonum spp. cf. 4 Maygrass Phalaris caroliniana 1 Purslane Portulaca spp. 51 Sedge family cf. Cyperaceae 4 Spurge family cf. Euphorbiaceae 1 Tick clover Desmodium spp. 7 Total weedy species 604 FRUITS Maypop cf. Passiflora incarnata cf. 4 Nightshade Solanum spp. 2 Total fruits 6 MISCELLANEOUS Bean cf. Phaseolus sp. cf. 4 Bearsfoot Polymnia uvedalia 1 Squash/gourd rind Cucurbita/Lagenaria 4 Sunflower Helianthus annuus 4 Sunflower cf. Helianthus annuus cf. 1 Total miscellaneous 14 UNIDENTIFIABLE 588 2.19 UNIDENTIFIABLE SEED 25 Total plants 54,618 766.09

Figure 7 Maize kernel image from the Feature 13 earth oven,

C. W. Cooper site.

Wilson and VanDerwarker An analysis of a maize roast at the C.W Cooper site, IL

172 Journal of Field Archaeology 2015 VOL. 40 NO. 2

throughout all the zones of the pit feature. Given that

nutshell was generally used secondarily as a fuel

source, it is not surprising that it was identified in the

upper layers of the pit. The same reasoning can be

applied to the presence of grass seeds and weedy seeds

(which include bedstraw [Galium spp.], chenopod

[Chenopodium spp.], bulrush [Scirpus spp.], purslane

[Portulaca spp.], smartweed/knotweed [Polygonum

spp.], tick clover [Desmodium spp.], and unidentified

members of the grass family [Poaceae]), as these

species were likely used as kindling for the fire set

atop this earth oven. The presence of these species in

the lower levels of the feature is best explained as

intrusive; the intrusion of small seeds and nutshell

fragments into the maize layers of the earth oven

likely originated in a small area of the pit where the

burnt clay layer did not entirely cover the feature.

Discussion and Conclusions A close consideration of the morphology, stratigra-

phy, and archaeobotanical contents of Feature 13 at

the C. W. Cooper site has provided important insight

into one method in which earth ovens were used to

cook food in the late prehistoric Midwest. In this case

a clay layer placed directly over a large number of

maize ears which served to shield the maize from direct

contact with a fire built above it. This particular

cooking event represents a direct method insofar as a

fire was built inside (but at the top of) the pit, but could

also be considered indirect in that a thick clay layer

buffered the maize from direct contact with the fire. It

is clear that the method of cooking used in this feature

differs from techniques used in the majority of the

ethnographic examples, in which the cooks built the

fire (or placed hot rocks) at the base of the earth oven

(Hough 1926: 37; Salazar et al. 2012; Thoms 2008,

2009; Wandsnider 1997). Indeed, the Maya account of

their maize roast, which is the case most analogous to

the C. W. Cooper earth oven, documents the use of hot

coals at the bottom of the pit (Salazar et al. 2012);

despite this singular difference, the layering of vegeta-

tion, maize, and earth was identical in both cases.

Given that no coals or hot rocks were placed below the

maize layer in the C. W. Cooper earth oven, the

oxidized base of this pit is testament to the high

intensity and long duration of the fire built atop it. Pits

with oxidized soils have also been identified at other

late prehistoric era sites in the Central Illinois River

Valley. Esarey (2000) reported the presence of a

number of Late Woodland pits exhibiting oxidized

bases at the Liverpool Lake site in Mason County

Illinois. Moreover, Bardolph (2014) recently docu-

mented a number of heavily oxidized pits at the early

Mississippian Lamb site (A.D. 1100–1150) in Schuyler

County, Illinois. Because most of these earth ovens

were cleaned out and later refilled with refuse, it is

unknown if they were used in exactly the same way as

the C. W. Cooper site feature. The presence of heavily

oxidized soils in these features, however, indicates

their likely use in direct cooking activities.

One of the most important insights gleaned from

the analysis of the C. W. Cooper earth oven is the

large amount of food (ca. 95 maize ears) it was used

to cook. The scale of this maize roast reinforces

previous assertions that such features were used in

the preparation of communal meals, possibly feasts.

Indeed, expansive excavations conducted at the Late

Woodland Range site in the American Bottom region

revealed that earth ovens were often located in areas

shared by multiple households, an arrangement that

led Kelly (1990) to interpret them as communal

cooking facilities. Based on the data presented here,

the nature and scale of the cooking that took place in

Feature 13 indicate that earth ovens at the C. W.

Cooper site were also used for the communal

preparation of foodstuffs. Whether or not the maize

roast that occurred in Feature 13 was conducted as

part of a first fruits celebration, as documented in the

Maya case, is unclear. Ethnohistorical studies from

the southeastern United States provide ample evi-

dence of annual events marking the first maize

harvest, known regionally as the Busk or Green

Corn Ceremony (Hudson 1976). While it is tempting

to interpret the C. W. Cooper earth oven as part of an

annual harvest feast, it is simply not possible to

achieve this level of specificity. We can say that the

C. W. Cooper maize roast was meant to involve a large

gathering of people, perhaps the entire community

that was residing that the site. This is an important

point as the late 12th century was an era when the

regional population was beginning to establish larger

and more nucleated settlements. Feasting events such

as the one represented in the Feature 13 earth oven

deposit may have been instrumental in integrating

these expanded communities.

This investigation of earth oven functionality is

part of a broader, ongoing effort to understand

diachronic changes in foodways in the late Prehistoric

Central Illinois River Valley. Our recent analysis

of changing plant subsistence during the Late

Woodland–Mississippian transition revealed a sig-

nificant increase in maize production and consump-

tion during the early Mississippian period, the latter

period represented by the C. W. Cooper plant

assemblage (excluding the Feature 13 maize remains)

(VanDerwarker et al. 2013). It is interesting that the

earth oven feature at C. W. Cooper represents the

persistence of a Late Woodland cooking technique

traditionally embedded within a context of commun-

alism but that simultaneously incorporates a rela-

tively new dietary element—in this case, maize. While

maize was certainly present in Late Woodland diets

Wilson and VanDerwarker An analysis of a maize roast at the C.W Cooper site, IL

Journal of Field Archaeology 2015 VOL. 40 NO. 2 173

of the Illinois Valley, it did not become a significant

staple food in the region until the early Mississippian

period (VanDerwarker et al. 2013). While we may

never know the reason this large outdoor roasting

event failed, the fact that it did fail provided us the

opportunity to see how earth ovens were constructed,

used, and the types of foods prepared within in the

Central Illinois Valley and perhaps the broader

Midwestern United States.

Acknowledgments This material is based upon work supported by the

National Science Foundation under grant 1062290.

We acknowledge Amy Gusick and Kristin Hoppa for

their careful excavation of this feature, as well as their

participation in preparing a preliminary conference

version of this manuscript that was presented at

the 2011 Southeastern Archaeological Conference in

Jacksonville, Florida. We also acknowledge the work

of Jennifer Alvarado in sorting and counting the

abundant maize remains from Feature 13. Finally,

we thank Dana Bardolph and Danielle Kurin for

providing comments on earlier drafts. Thanks also to

Larry Conrad for providing access to maize cobs

from earlier WIU excavations of the site.

Gregory D. Wilson (Ph.D. 2005, University of North

Carolina, Chapel Hill) is Associate Professor at the

University of California, Santa Barbara. His research

is concerned with issues of social inequality, identity

politics, and violence in Precolumbian North and South

America, through a lens of household and community-

centered archaeology.

Amber M. VanDerwarker (Ph.D. 2003, University of

North Carolina, Chapel Hill) is Associate Professor at

the University of California, Santa Barbara. She has

been involved in field and laboratory work in Mexico,

eastern North America, and Peru. Her research

encompasses a variety of methods, regions, and themes

that focus on the relationship between humans and food

in the New World, especially during the period of

transition to agriculture.

References Bardolph, D. 2014. ‘‘Evaluating Cahokian Contact and

Mississippian Identity Politics in the Late Prehistoric Central Illinois River Valley,’’ American Antiquity 79: 69–89.

Binford, L. A., S. R. Binford, R. Whallon, and M. A. Hardin. 1970. Archaeology at Hatchery West. Society for American Archaeology Memoir 24, Washington, D.C.: Society for American Archaeology.

Black, S. L., and D. G. Creel. 1997. ‘‘The Central Texas Burned Rock Midden Reconsidered,’’ in S. L. Black, L. W. Ellis, D. G. Creel, and G. T. Goode, eds., Hot Rock Cooking on the Greater Edwards Plateau: Four Burned Rock Midden Sites in West Central Texas. Studies in Archeology 22. Texas Archeological Research Laboratory, University of Texas at Austin and Archeological Studies Program 2. Austin: Texas Department of Transportation, Environmental Affairs Division, 269–305.

DeBoer, W. R. 1988. ‘‘Subterranean Storage and the Organization of Surplus: The View from Eastern North America,’’ Southeastern Archaeology 26: 1–20.

Delorit, R. 1970. Illustrated Taxonomy Manual of Weed Seeds. River Falls: Wisconsin State University.

Dering, P. 1999. ‘‘Earth Oven Plant Processing in Archaic Period Economies: An Example from a Semi-Arid Savannah in South- Central North America,’’ American Antiquity 64: 655–674.

Esarey, D. 1988. ‘‘The Liverpool Lake Site: Results of Archaeological Monitoring at 11Mn163, Mason County, Illinois,’’ manuscript on file, Dickson Mounds Museum, Lewistown, Illinois.

Fish, S. K., P. R. Fish, C. Miksicek, and J. Madsen. 1985. ‘‘Prehistoric Agave Cultivation in Southern Arizona,’’ Desert Plants 7: 100, 107–112.

Fortier, A. C. 1983. ‘‘Mund Phase Features,’’ in A. C. Fortier, F. A. Finney, and R. B. Lacampagne, eds., The Mund Site (11- S-435). American Bottom Archaeology, FAI-270 Site Reports 5. Urbana: University of Illinois Press, 111–154.

Fritz, G. J. 1992. ‘‘‘Newer,’ ‘Better’ Maize and the Mississippian Emergence: A Critique of Prime Mover Explanations,’’ in W. I. Woods, Late Prehistoric Agriculture: Observations from the Midwest. Studies in Illinois Archaeology 8. Springfield: Illinois Historic Preservation Agency, 19–43.

Gracer, A., A. M. VanDerwarker, and G. D. Wilson. 2013. ‘‘Farming and Food Insecurity in the Mississippian Central Illinois River Valley: Metric Data on Maize Kernels and Cupules as a Means to Establish the Number and Type(s) of Varieties.’’ Poster presented at the 78th Annual Meeting of the Society for American Archaeology, Honolulu, Hawaii.

Harris, W. G. 1996. ‘‘Form and Function of Crab Orchard Tradition Pit Features in the Big Muddy River Drainage of Southern Illinois,’’ unpublished Master’s thesis, Southern Illinois University, Carbondale.

Holt, J. Z. 1996. ‘‘AG Church Site Features and Community Organization,’’ Illinois Archaeology 8: 58–84.

Hough, W. 1926. ‘‘Fire as an Agent in Human Culture,’’ Smithsonian Institution, United States National Museum, Bulletin 139. Washington, D.C.: Government Printing Office.

Hudson, C. 1976. The Southeastern Indians. Knoxville: University of Tennessee Press.

Kelly, J. E. 1990. ‘‘The Emergence of Mississippian Culture in the American Bottom,’’ in B. D. Smith, ed., The Mississippian Emergence. Washington, D.C.: Smithsonian Institution Press, 113–152.

Kelly, J. E., A. C. Fortier, S. J. Ozuk, and J. A. Williams. 1987. The Range Site: Archaic Through Late Woodland Occupations. American Bottom Archaeology FAI-270 Site Reports 16. Urbana: University of Illinois Press.

Koldehoff, B. 2002. The Woodland Ridge Site and Late Woodland Land Use in the Southern American Bottom. Transportation Archaeological Research Reports 15. Illinois Transportation Archaeological Research Program, Urbana-Champaign: University of Illinois.

Koldehoff, B., and J. M. Galloy. 2006. ‘‘Late Woodland Frontiers in the American Bottom Region,’’ Southeastern Archaeology 25: 275–300.

Leach, J. D., and K. D. Sobolik. 2010. ‘‘High Dietary Intake of Prebiotic Inulin-Type Fructans in the Prehistoric Chihuahuan Desert,’’ British Journal of Nutrition 103: 1558–1561.

Martin, A. C., and W. D. Barkley. 1961. Seed Identification Manual. Berkeley: University of California Press.

Peacock, S. L. 2008. ‘‘From Complex to Simple: Balsamroot, Inulin, and the Chemistry of Traditional Interior Salish Pit- Cooking Technology,’’ Botany 86: 116–128.

Salazar, C., D. Zizumbo-Villarreal, S. B. Brush, and P. Colunga- Garcı́a Marı́n. 2012. ‘‘Earth Ovens (Pı́ib) in the Maya Lowlands: Ethnobotanical Data Supporting Early Use,’’ Economic Botany 66: 285–297.

Stahl, A. B. 1985. The Dohack Site (11-S-642). American Bottom Archaeology FAI-270 Site Reports 12. Urbana: University of Illinois Press.

Teit, J. A. 1900. The Thompson Indians of British Columbia. Memoir of the American Museum of Natural History 1. New York: G.E. Stechert and Co.

Thoms, A. V. 2008. ‘‘The Fire Stones Carry: Ethnographic Records and Archaeological Expectations for Hot-Rock Cookery in Western North America,’’ Journal of Anthropological Archaeology 27: 443–460.

Wilson and VanDerwarker An analysis of a maize roast at the C.W Cooper site, IL

174 Journal of Field Archaeology 2015 VOL. 40 NO. 2

Thoms, A. V. 2009. ‘‘Rock of Ages: Propagation of Hot-Rock Cookery in Western North American,’’ Journal of Archaeo- logical Science 36: 573–591.

United States Department of Agriculture (USDA). 2011. ‘‘National Plant Germplasm System,’’ (http://www.ars-grin. gov/npgs/images/sbml/).

VanDerwarker, A. M., G. D. Wilson, and D. N. Bardolph. 2013. ‘‘Maize Adoption and Intensification in the Central Illinois

River Valley: An Analysis of Archaeobotanical Data from the Late Woodland through Early Mississippian Periods (AD 400– 1200),’’ Southeastern Archaeology 32: 147–168.

Van den Bel, M. 2010. ‘‘A Description of Late Archaic Rock-Filled Pits in French Guiana,’’ Revista de Arqueologı́a 23: 60–71.

Wandsnider, L. 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.

Wilson and VanDerwarker An analysis of a maize roast at the C.W Cooper site, IL

Journal of Field Archaeology 2015 VOL. 40 NO. 2 175