LAKEHE AD UNIVERSITY, THUNDER BAY, ONTARIO 2014
Plant Microfossil Analysis of Middle Woodland Food Residues,
Northern Minnesota
A thesis submitted in partial fulfillment of the Master of Environmental
Studies: Northern Environments and Cultures
By: Alexandra Burchill
PERMISSION TO USE
Copies of this document will be available at the department of Northern Environments
and Cultures, as well as Lakehead University’s library. My supervisor Dr. Matthew
Boyd and graduate coordinator Martha Dowsley, or in their absence, the Chair of the
Department of Anthropology or the Dean of the Faculty of Science and Environmental
Studies may grant permission for the copying of this thesis for scholarly purposes. Any
copying of this thesis for the purpose of financial gain is not allowed without my written
permission.
Scholarly use of the document must have acknowledged recognition
towards Lakehead University and me.
I
ABSTRACT
Northern Minnesota lies within the southern edge of the Boreal Forest and, as a result,
archaeological sites in this region typically have poor organic preservation and thin,
disturbed, stratigraphy. For this reason, little is known about specific plant foods and
their importance at many sites. In order to fill this gap, my research focuses on the
extraction of plant microfossils (starch, phytoliths and pollen) from carbonized and noncarbonized food residues associated with Middle Woodland (100 BC – AD 500)
components. My results show that wild rice was widely consumed during this time
along with cultigens such as maize. No additional evidence suggested farming, so
there is a possibility of trade with periphery groups to acquire the cultigens recovered
from the microfossil analysis. These results demonstrate the importance of plant
microfossil studies as a tool for identifying subtle evidence of wild and domesticated
plants in regions characterized by poor organic preservation, small seasonally-occupied
sites and other fundamental limitations. The mixed economic strategy apparent in some
Northern Minnesota sites re-defines a diet of native and domesticated cultigens, which
can be applied to the wider archaeobotanical literature of northeastern North America.
II
ACKNOWLEDGEMENTS
Many individuals assisted with this research project and expressed interest in my
topic throughout the last two years. I would first like to thank my supervisor, Dr. Boyd
for his mentorship. He showed great patience while editing my thesis and guidance in
developing different ideas and lines of thought. I am grateful for the opportunity
provided by Dr. Boyd to study at Lakehead University in the MES: NECU program and
to have such an interesting and fulfilling research topic.
I would also like to thank my committee members, Dr. Hamilton, Dr. Varney and
Dr. Mulholland for their edits and suggestions. Dr. Hamilton provided literary resources
during my two years at Lakehead and was always available to answer any questions.
Dr. Mulholland, my external reviewer, provided helpful insights into Minnesota
archaeology and paleoethobotany.
Clarence Surette and Megan Wady patiently taught me the required procedures
for phytolith and starch extraction and processing. Clarence was always close by to
help in the laboratory, from loosening caps twisted too tight on centrifuge tubes to
extracting the last milligram of residue from lithic and ceramic samples.
The Minnesota archaeology community was extremely helpful throughout the last
two years and provided an endless amount of information and support. Lee Johnson
provided the samples used in this thesis project and all of the available site information.
Heather Hoffman located and mapped the nine sites used in this thesis. Specific
archaeologists that assisted with this project include Bill Yourd, LeRoy Gonsier, David
III
Radford, Jennifer Shafer, David Mather, and Jim Cummings. I would also like to thank
all the individuals that worked on the archaeological sites from which samples were
collected.
Additional support came from Manitoba archaeologist Dr. Leigh Syms, who shared
his vast knowledge and expertise on prehistoric plant use. Anthea Kyle thoroughly
edited my thesis document in the spring of 2014.
Jill Taylor-Hollings spent time
ensuring that all the ceramic samples were correctly identified, and she answered any
questions I had about ceramics in general.
My colleagues in the Northern
Environments and Cultures program, 2012, provided friendship and an endless amount
of support and enthusiasm for my thesis project. There are many others that provided
support for this thesis project such as family and friends and for that I am grateful.
Dr. Bubel and Dr. McGeough of the University of Lethbridge were my inspiration to
pursue a Masters degree in archaeology. Both of these amazing professors put up with
meetings, questions, and inquiries and taught me everything I knew about archaeology
prior to my master’s. Dr. Bubel is not only an inspirational professor, but a friend as
well, and her hard work and dedication in the field and laboratory are an example I will
always look up too.
IV
TABLE OF CONTENTS
1.0 INTRODUCTION-------------------------------------------------------------------------------- 1
1.1 CULTURE HISTORY-------------------------------------------------------------- 1
1.2 METHODS---------------------------------------------------------------------------- 6
2.0 STUDY AREA----------------------------------------------------------------------------------- 7
2.1 MODERN ENVIRONMENT------------------------------------------------------ 7
2.2 PALEOENVIRONMENT---------------------------------------------------------- 10
3.0 ARCHAEOLOIGICAL SITES---------------------------------------------------------------- 20
3.1 ARCHAEOLOGICAL SITES----------------------------------------------------- 20
3.2 THIRD RIVER BORROW PIT SITE-------------------------------------------- 22
3.3 BIG RICE SITE---------------------------------------------------------------------- 25
3.4 KYLELEEN’S TALL PINE-------------------------------------------------------- 28
3.5 KYLELEEN’S BENT PINE--------------------------------------------------------28
3.6 WINNIE COTTAGES---------------------------------------------------------------29
3.7 WINDY BEAD------------------------------------------------------------------------ 30
3.8 SAGA ISLAND---------------------------------------------------------------------- 31
3.9 LOST LAKE-------------------------------------------------------------------------- 32
3.10 NO BEARD------------------------------------------------------------------------- 32
3.11 SAMPLES--------------------------------------------------------------------------- 33
4.0 THE WOODLAND TRADITION IN THE UPPER GREAT LAKES----------------- 35
4.1 INTRODUCTION-------------------------------------------------------------------- 35
4.2 EARLY WOODLAND (1200 BC TO 250 AD)
ACHAEOLOGICAL COMPLEXES-------------------------------------------------- 35
4.3 THE MIDDLE WOODLAND (100 BC TO 500 AD)
ARCHAEOLOGICAL COMPLEXES------------------------------------------------39
4.3.1 MIDDLE WOODLAND CHRONOLOGICAL ISSUES-------------------45
4.4 LATE WOODLAND (500-1750 AD) ARCHAEOLOGICAL
COMPLEXES----------------------------------------------------------------------------- 49
4.5 PALEODIET OF WOODLAND PEOPLES----------------------------------- 51
5.0 PHYTOLITH AND STARCH ANALYSIS IN ARCHAEOLOGY--------------------- 82
5.1 INTRODUCTION-------------------------------------------------------------------- 82
5.2 PHYTOLITHS------------------------------------------------------------------------ 82
5.3 STARCH GRANULES------------------------------------------------------------- 88
5.3.1 STAINING-------------------------------------------------------------------------- 93
5.3.2 BIREFRIGENCE------------------------------------------------------------------ 93
5.3.3 GELATINIZATION--------------------------------------------------------------- 94
5.4 LIMITATIONS------------------------------------------------------------------------ 95
5.5 CASE STUDIES--------------------------------------------------------------------- 99
6.0 METHODS---------------------------------------------------------------------------------------- 101
6.1 OVERVIEW OF FOOD RESIDUE ANLAYSIS------------------------------ 101
6.2 SELECTION OF ARCHAEOLOGICAL SAMPLES------------------------ 102
6.3 LABORATORY PROTOCOLS-------------------------------------------------- 103
6.4 IDENTIFICATION OF SAMPLES-----------------------------------------------111
6.5 COMPARATIVE REFERENCE KEY------------------------------------------ 121
V
6.6 UNKNOWN PLANT SPECIES-------------------------------------------------- 124
6.7 CONTAMINATION----------------------------------------------------------------- 125
7.0 RESULTS----------------------------------------------------------------------------------------- 127
7.1 INTRODUCTION-------------------------------------------------------------------- 127
7.2 RESIDUE SAMPLE SIZES------------------------------------------------------- 127
7.3 THIRD RIVER BORROW PIT--------------------------------------------------- 138
7.4 BIG RICE------------------------------------------------------------------------------138
7.5 KYLELEEN’S TALL PINE-------------------------------------------------------- 140
7.6 WINNIE COTTAGES-------------------------------------------------------------- 140
7.7 WINDY BEAD------------------------------------------------------------------------ 141
7.8 SAGA ISLAND---------------------------------------------------------------------- 142
7.9 LOST LAKE-------------------------------------------------------------------------- 142
7.10 NO BEARD------------------------------------------------------------------------- 143
7.11 KYLELEEN’S BENT PINE------------------------------------------------------144
7.12 GRASS PHYTOLITHS----------------------------------------------------------- 149
8.0 INTERPRETATIONS AND DISCUSSION---------------------------------------------- 151
8.1 INTRODUCTION-------------------------------------------------------------------- 151
8.2 MAIZE--------------------------------------------------------------------------------- 152
8.3 WILD RICE--------------------------------------------------------------------------- 156
8.4 SQUASH AND COMMON BEAN--------------------------------------------- 158
8.5 CULTIGEN ACQUISITION------------------------------------------------------- 163
9.0 CONCLUSIONS--------------------------------------------------------------------------------- 173
REFERENCES--------------------------------------------------------------------------------------- 178
APPENDIX-------------------------------------------------------------------------------------------- 200
APPENDIX 1------------------------------------------------------------------------------ 201
APPENDIX 2------------------------------------------------------------------------------ 202
APPENDIX 3------------------------------------------------------------------------------ 202
APPENDIX 4------------------------------------------------------------------------------ 203
APPENDIX 5------------------------------------------------------------------------------ 204
APPENDIX 6------------------------------------------------------------------------------ 205
APPENDIX 7----------------------------------------------------------------------------- 206
APPENDIX 8----------------------------------------------------------------------------- 207
APPENDIX 9----------------------------------------------------------------------------- 207
VI
LIST OF FIGURES
Figure 1.1: Map of study area ---------------------------------------------------------------------------- 3
Figure 1.2: Map of Minnesota counties in the study area ----------------------------------------- 5
Figure 2.1: Map of glacial lobes in Minnesota ------------------------------------------------------- 11
Figure 2.2: Glacial movement during the Late Wisconsin in Minnesota----------------------- 12
Figure 2.3: Vegetation in relation to different glacial surges versus time --------------------- 15
Figure 2.4: Varve thickness, pollen levels for trees and shrubs, diatom flux and
elemental percentages versus time at Elk Lake --------------------------------------15
Figure 3.1: Location of the nine study sites in Northern Minnesota ---------------------------- 21
Figure 3.2: East planview of Feature 1, Unit 2S and 8S ------------------------------------------ 24
Figure 3.3: Big Rice site feature 25, bottom of level 12 ------------------------------------------- 26
Figure 3.4: Ceramic sherds from the Windy Bead site -------------------------------------------- 30
Figure 3.5: Windy Bead ceramic decorative attributes -------------------------------------------- 31
Figure 3.6: Laurel sherds from the Saga Island site ----------------------------------------------- 32
Figure 4.1: Ceramic typologies from the Sub-Arctic, Eastern Woodland and Northern
Plains versus time --------------------------------------------------------------------------- 39
Figure 4.2: Spatial extent of Brainerd assemblages in Minnesota ----------------------------- 42
Figure 4.3: Map of Minnesota displaying the Gull Lake Dam site, King Coulee,
Ogema-Geshik, Cass Lake --------------------------------------------------------------- 44
Figure 4.4: Brainerd net-impressed sherd from the Windy Bead site -------------------------- 45
Figure 4.5: Distribution of Laurel assemblages across Eastern North America -------------54
Figure 4.6: Laurel rim sherd from the Saga Island site -------------------------------------------- 56
Figure 4.7: Map of Late Archaic and Early Woodland sites with evidence of maize ------- 65
Figure 4.8: Location of sites with early evidence of maize. Maize was utilized sometime
within the available dates, and may not date to the beginning of site
occupation ------------------------------------------------------------------------------------- 69
Figure 4.9: Location of sites with domesticated bean --------------------------------------------- 75
Figure 5.1: (A) Cassava starch, an example of a compound starch type. (B)
Mouldy potato starch granules exhibiting staining. (C) Gelatinized azuki
bean cv. Erimo starch granules. (D and E) Damaged starch granule,
recovered on a grinding stone from the No Beard site, (D) XPL, (E) PPL -----92
Figure 6.1: Flowchart of methods for non-carbonized food residue ---------------------------- 104
Figure 6.2: Flowchart of methods for carbonized food residue ---------------------------------- 105
Figure 6.3: Carbonized food residue on ceramics from the Windy Bead site (A), and the
No Beard site (B). Non-carbonized food residue on lithics (ground stone
artifacts) from the No Beard (C) and Windy Bead sites (D) ----------------------- 107
Figure 6.4: Maize wavy-top rondel phytolith ----------------------------------------------------------115
Figure 6.5: Maize starch grain under XPL (A) and PPL (B) -------------------------------------- 116
Figure 6.6: Squash phytolith ----------------------------------------------------------------------------- 117
Figure 6.7: Starch granules identified as squash --------------------------------------------------- 118
Figure 6.8: Hook shaped phytoliths of Phaseolus vulgaris --------------------------------------- 119
Figure 6.9: Phaseolus starch grain --------------------------------------------------------------------- 119
Figure 6.10: Wild rice rondel phytolith from the MacGillivray site ------------------------------- 120
VIII
Figure 6.11: Comparative plant microfossils. A (PPL) + B (XPL) maize starch granules;
A (PPL) + B (CPL) bean starch granules; E (PPL) + F (CPL) squash starch
granules; G (PPL) wild rice rondel phytolith; H (PPL) maize rondel
phytolith---------------------------------------------------------------------------------------- 124
Figure 7.1: Brainerd and Laurel ceramics analyzed in this study.
(A), (B) Brainerd sherds from the Third River Borrow Pit site, (C) Brainerd
sherd from the Windy Bead site, (D) Brainerd sherd from the Saga Island
site, (E) Laurel rim from the Saga Island site, and (F) Laurel sherd from the
Lost Lake site --------------------------------------------------------------------------------- 129
Figure 7.2: Two grinding stones used for residue analysis. A is from the No Beard site
and B is from the Windy Bead site ------------------------------------------------------ 132
Figure 7.3: Locations of samples extracted from the two grinding stones. (A) No Beard
site, sample one is from a dry bush, sample two is a wet brush, sample
three is from sonication, sample four is sonication of the worked edge; (B)
Windy Bead site, sample one is from a wet brush, sample two is from
sonication--------------------------------------------------------------------------------------- 132
Figure 7.4: Diagnostic phytolith and starch data from sediment samples. Big Rice: BR,
Windy Bead: WB. Amounts are expressed as percentages of the total
microfossil count (n=250) ------------------------------------------------------------------ 134
Figure 7.5: Microfossil assemblage from sediment samples. Big Rice: BR,
Windy Bead: WB. Amounts are expressed as percentages of the total
microfossil count (n=250) ------------------------------------------------------------------ 135
Figure 7.6: Diagnostic phytolith and starch data from ceramic samples. Third River
Borrow Pit: TRBP, Big Rice: BR, Saga Island: SI, Windy Bead: WB, Winnie
Cottages: WC, Kyleleen’s Tall Pine: KTP, Kyleleen’s Bent Pine: KBP,
Lost Lake: LL, No Beard: NB. Amounts are expressed as percentages of
the total microfossil count (n=250) ------------------------------------------------------- 136
Figure 7.7: Microfossil assemblage recovered from ceramic samples. Third River
Borrow Pit: TRBP, Big Rice: BR, Saga Island: SI, Windy Bead: WB, Winnie
Cottages: WC, Kyleleen’s Tall Pine: KTP, Kyleleen’s Bent Pine: KBP,
Lost Lake: LL, No Beard: NB. Amounts are expressed as percentages
of the total microfossil count (n=250) ---------------------------------------------------- 137
Figure 7.8: Diagnostic phytolith and starch data from lithic samples. No Beard:
NB, Windy Bead: WB. Amounts are expressed as percentages of the
total microfossil count (n=250) ------------------------------------------------------------ 138
Figure 7.9: Microfossil assemblage recovered from two grinding stone samples.
No Beard: NB, Windy Bead: WB. Amounts are expressed as percentages
of the total microfossil count (n=250) --------------------------------------------------- 139
Figure 7.10: Damaged starch granule, recovered on a grinding stone from the
No Beard site, (A) XPL, (B) PPL -------------------------------------------------------- 143
Figure 7.11: Microfossil photos of Z.mays ssp. mays. observed in carbonized food
residue, soil and stone tools (A) Third River Borrow Pit site. (B) Big Rice
site. (C, D, H) Windy Bead site. (F) No Beard site -------------------------------- 145
Figure 7.12: Microfossil photos of Zizania sp. observed in carbonized food residue, soil
and stone tools (A, C) Big Rice site. (B) Windy Bead site ---------------------- 146
IX
Figure 7.13: Z.mays ssp. mays. starch granules observed in carbonized food residue,
soil and stone tools photographed under XPL and PL light. (A, B)
Windy Bead site ---------------------------------------------------------------------------- 147
Figure 7.14: Possible Phaseolus vulgaris starch granule from a Big Rice soil sample.
(A) XPL light, (B) PL light ---------------------------------------------------------------- 147
Figure 7.15: Possible Cucurbita sp. starch granules taken under XPL and PL light.
(A, B) XPL and PL light photos taken from a Big Rice site soil sample.
(C, D) Starch granules are from carbonized food residue originating from
the Saga Island site. (E, F) Starch identified in carbonized food residue
from the Windy Bead site ---------------------------------------------------------------- 148
Figure 7.16: (A) Squash seed collected during excavation at the Big Rice site. (B) Wild
rice seed portion collected during excavation of the Big Rice site -------------149
Figure 8.1: Map of nine sites in the study area and areas of positive plant identification--162
Figure 8.2: Avonlea Complex distribution ------------------------------------------------------------- 170
X
LIST OF TABLES
Table 1: Seeds from Area A ------------------------------------------------------------------------------ 28
Table 2: Major ceramic types dating to the Early, Middle and Late Woodland periods in
Minnesota ----------------------------------------------------------------------------------------- 39
Table 3: Brainerd ware dates across Northern Minnesota --------------------------------------- 46
Table 4: Plant taxon and diagnostic characteristics of phytoliths, starch and pollen -------113
Table 5: Sample weight of carbonized residue from ceramics,
non-carbonized residue from lithics and soil, along with provenience
Information ---------------------------------------------------------------------------------------- 129
Table 6: Ceramics from the ten chosen study sites ------------------------------------------------ 131
XI
DEFITION OF TERMS
CULTIVATION: Human involvement such as the weeding, pruning, and tending to plants to affect
the plant’s life cycle.
DOMESTICATION: Occurs when a species is genetically altered to benefit the interests of humans.
FOOD RESIDUE ANALYSIS: Food residue analysis can be conducted on a wide variety of
archaeological remains to determine the presence, and/or absence of different plant
and animal species.
PHYTOLITH: A rigid structure of solid silica that forms when plants absorb groundwater. The silica
is deposited in intracellular, and/or extracellular structures of the living plant.
POLLEN: Reproductive seed plants in the form of a fine to coarse powder.
STARCH: Carbohydrate produced by plants as a form of energy.
THREE SISTERS: Agricultural crops that were of great importance to Aboriginal groups in North
America. The Three Sisters agricultural system consists of maize, squash, and
beans and these crops were often planted together.
WOODLAND PERIOD: Chronological division used to differentiate prehistoric cultures in North
America. In Eastern North America, the Woodland Period dates from 1000 BC to
1000 AD.
XII
1.0 INTRODUCTION
1.1 Culture History
During the Paleoindian, Archaic and Woodland Periods in North America,
Aboriginal groups depended on a wide variety of floral species. The reasons for this
may be ideological, economical and/or environmental, but it is understood that diet
always consists of plants, whether they are native to the occupied area, or traded
with outside populations (Syms et al. 2013). Wild plants were widely consumed in
Eastern North America (Yarnell 1993), up until the introduction and dispersal of
domesticated plants such as maize (Zea mays ssp. mays), beans (Phaseolus sp.)
and squash (Cucurbita sp.).
The question remains of how these domesticates
reached certain areas and when they were adopted into subsistence strategies.
The focus of this thesis is on the Middle Woodland period in Northern
Minnesota. This period, which dates from 100 BC to 500 AD, was characterized by
distinctive lithic technologies, and new ceramic forms. Associated projectiles are
broadly, shallow, or corner notched, and/or have an expanding or straight stem.
Middle Woodland ceramic wares had decorative differences, were thinner, and
made with a finer temper; these attributes differentiated them from Early Woodland
ceramics.
It is generally assumed that Middle Woodland peoples occupying
Northern Minnesota were semi-sedentary hunter-gatherers relying on mammals,
and wild rice (Zizania spp.), along with starchy nuts, fruits, and seeds (Arzigian
2008; Mulholland et al. 1996; Peter and Motivans 1983; Shafer 2003).
This
subsistence base appears to have persisted until the arrival of the Three Sister
crops (maize, beans, and squash) between 800 AD and 1200 AD, which flourished
1
until Precontact (Adair and Drass 2011; Hart et al. 2002; Wilson 1917). One of the
research goals for this study is to readdress these currently held notions about
horticultural subsistence during the Middle Woodland in Northern Minnesota.
Other research goals include reconstructing the plant component of the
paleodiet and identifying whether trade or local agriculture allowed for the presence
of domesticates at these sites.
Cultivation, which is the weeding, pruning and
tending of plants can be identified through multiple lines of evidence such as pollen
and agricultural tools. Cultivation does not mean that something is fully
domesticated, but that the plant's life cycle has been impacted in some way by
human selection (Syms et al. 2013). Food production begins with the deliberate
care towards a certain species, which may be invisible to archaeologists, but
through evidence of agricultural tools, diagnostic phytoliths, starch granules and
pollen, it becomes visible.
Domestication occurs when humans, causing the
domesticated species to become dependent for survival (Staller 2010), genetically
alter plants or animals. This is visible in the archaeological record from seed sizes
and diagnostic features of phytolith and starch granules.
Archaeological sites chosen for this research project are located in Northern
Minnesota, within the Great Lakes- St. Lawrence Forest (Figure 1.2).
2
Figure 1.1: Map of study area (base map adapted from USGS 2013; Environment Canada;
Minnesota Department of Natural Resources).
This area was chosen because food residue analysis has been applied to only
a few sites in Northern Minnesota and recent results indicate the presence of
cultivated plants in adjacent regions of Northwestern Ontario, the Canadian prairies,
Southeastern Minnesota and the Dakotas (Boyd et al. 2006, 2008; Lints 2012;
Mulholland 1986; Perkl 1998).
For example, the Martin-Bird site from Northern
Ontario yielded maize, wild rice and squash (Boyd et al. 2014). Lints’ 2012 study of
ten sites spanning the Northern Plains showed the widespread dispersal of beans
and maize by at least AD 700. The King Coulee site in Southeastern Minnesota
contained Cucurbita pepo seeds and the Big Hidatsa site in North Dakota showed
evidence of maize cultivation (Mulholland 1993; Perkl 1998). Prehistoric sites are
common along the Mississippi Headwaters and beside rivers and lakes in Central
3
and Northern Minnesota, making this study area a rich source of information about
Woodland diet (Mulholland et al. 1996).
The sites chosen for this study are associated with Brainerd and Laurel pottery
(Hohman-Caine et al. 2012); these two complexes represent the earliest potteryproducing cultures in Northern Minnesota and are the most widespread expressions
of the Initial (Middle) Woodland period (Anfinson 1979; Arzigian 2008).
These
populations relied on a diet of bison, deer, elk and wild rice (Arzigian 2008;
Lugenbeal 1976; Hohman-Caine and Goltz 1995).
Nine sites were chosen to give a comprehensive representation of paleodiet
during the Middle Woodland across the region. The sites chosen include the (1)
Third River Borrow Pit; (2) Big Rice; (3) Winnie Cottages; (4) No Beard; (5) Windy
Bead; (6) Lost Lake; (7) Saga Island (8) Kyleleen’s Bent Pine; and (9) Kyleleen’s
Tall Pine. From these sites, 31 ceramic sherds, two grinding stones and five soil
samples were analyzed using the methods described below. The sample size was
selected to allow for multiple lines of evidence to interpret patterns of trade and/or
small-scale agriculture.
The counties included in the study area are Cook, Lake, St. Louis,
Koochiching, Itasca, Carlton, Cass and Aitkin (Figure 1.2). Both Brainerd and Laurel
ware are present in these eight counties, making this area of Northern Minnesota a
suitable location for the study of Middle Woodland plant consumption.
4
Figure 1.2: Map of Minnesota counties in the study area (source:
http://Geology.com ).
Current research from these periphery areas pushes back the timing for the
introduction and dispersal of domesticated cultigens, making it likely that maize,
beans and squash were present in Northern Minnesota (Boyd et al. 2006, 2008;
Lints 2012; Mulholland 1993; Perkl 1998). Acidic soils prevent the preservation of
organic materials such as seeds; therefore, not many macrobotanical remains have
been recovered, which has hindered the discourse on prehistoric subsistence
activities in Northern Minnesota. Food residue analysis from a number of different
contexts has proven to be an invaluable method for inferring past diet and floralbased subsistence in cases were there is poor survival of organic material.
5
1.2 Methods
Food residue analysis is the microscopic identification of microfossils to
determine paleodiet. The microfossils in this study include phytoliths, which are
products of silica absorption from ground water; starch granules, carbohydrates in
which plants store energy; and pollen grains, which are powders containing the
microgametophytes of reproductive seed plants. These microfossils are especially
useful in areas of poor organic preservation and they are influencing the dietary
interpretations arrived at with conventional techniques (Crowther 2012; Berman and
Pearsall 2008; Boyd and Surette 2010; Haslam 2004; Piperno and Dillehay 2008;
Piperno et al. 2009).
Phytoliths survive in environments of good preservation because they are
inorganic and very durable (Piperno 2006). Starch granules have been found in a
variety of well and poorly preserved environments with some associated deposits 2
million years old (Torrence and Barton 2006). Pollen is the male gametophyte, and
depending on species they can travel long distances and have been found in
deposits dating to 8,050 BC (Fearn and Lui 1995; Wright 1993). Morphological
characteristics led researchers to recognize key diagnostic features of phytoliths,
starch and pollen, allowing for accurate species identifications (Brown 1984; Fearn
and Lui 1995; Torrence 2006).
6
2.0 STUDY AREA
2.1 Modern Environment
Climate
Northern Minnesota has a continental climate, characterized by warm
summers and cool to cold winters with an average annual rainfall of 60.96 to 63.5
centimetres. Temperatures are wide ranging, with July maximums of 21.1° to 27.8°
Celsius and January minimums of -11.1° to -8.9° Celsius (Anderson et al. 1993;
Mulholland et al. 1996). Three dominant air masses control the climate: southerly
winds carry humid, warm air from the Gulf of Mexico, northerly winds bring dry and
cool air from the Arctic and westerly winds bring dry Pacific air (Anderson et al.
1993).
Soils
Lakela (1965) describes the soils in the central part of Northeastern Minnesota
as gray-brown loam, with sand and peat overlying fine sand. The western and
southwestern parts of St. Louis County contain a gray loam soil over limy subsoil.
The soil pH of a jack pine stand in the Superior National Forest measured 4.8 to 5.2
(Ahlgren and Ahlgren 1965).
Soils unique to the Third River Borrow Pit site
includes, sandy loam Typic Glossaqualfs, loamy fine sands Alfic Udipsamments,
mixed typic Fluvaquents and Aquic Udipsamments.
These soils reflect an
environment of sand plain origin in a floodplain (Mulholland et al, 1996). Peters and
Motivans’ (1983) evaluation of the Big Rice site identified soils of the BaudetteSpooner-Toivola area including sands and loamy sands.
7
Geology
The geology of the region is an important environmental aspect for
understanding resource availability. Rocky outcrops used as lithic resources date to
the Early Precambrian, Middle Precambrian, Late Precambrian and Cretaceous.
Geologic formations consist mostly of Precambrian bedrock; exceptions include
areas of Cretaceous sediments in the Mesabi Range and Quaternary sediments
(Lakela 1965). Northeastern Minnesota includes the Duluth Complex, a section of
the Middle Precambrian basin with the Biwabik Iron Formation, three Lower
Precambrian batholiths (Saganaga, Giant’s Range and Vermillion), Lower
Precambrian volcanic-sedimentary (greenstone) belt and Upper Precambrian
continental lava flows (Ojakangas and Matsch 1982). Glaciofluvial deposits overlie
these rock formations and from the north to the south and eastwards, geologic
formations become younger. Metasediments, conglomerates, Knife Lake slates and
Keewatin greenstones underlie the Northern area.
Iron formations, unfolded
quartzite and argillite overlie these older rock formations. To the south, intense
folding acts on Huronian rocks, which are intruded by granite. The North Shore
Highland is a younger rock formation within the Keweenawan system. Diabase,
gabbo and other similar rocks intrude lava flows and intercalated sediments dip east
to southwest towards Lake Superior (Lakela 1965).
Vegetation
Mulholland et al. (1996) and Lakela (1965) describe current vegetation in
Northern Minnesota as being mostly composed of jack pine (Pinus banksiana), with
spruce (Picea) thriving in lower areas of successional growth with white pine (Pinus
8
strobus), red pine (Pinus resinosa) and balsam fir (Abies balsamea). Low juniper
(Juniperus communis) only grows in rocky forests and exposed bluffs, while
creeping juniper (Juniperus horizontalis) can be found on the Fox Islands and Spilt
Rock. Yew (Taxus canadensis) and white cedar (Thuja occidentalis) inhabit islands,
rockier areas and moist lowlands. Tamarack (Larix laricina), black spruce (Picea
marinana) dominant upland bogs and crevice plant life such as mosses, ferns and
lichens thrive on ledges and cliffs. The mix of coniferous and deciduous forests in
Northern Minnesota is typical of the mixed forest transition zone in the Great LakesSt. Lawrence Forest.
Deciduous species present includes, oak (Quercus), elm
(Ulmus L.), ash (Fraxinus), and basswood (Tilia americana).
Wild rice (Zizania sp.) is a native plant that grows in shallow lakes and
streams, and for at least the last 2,000 years, wild rice has been an important form
of sustenance for Minnesotan populations (Valppu and Rapp 2000; Yost 2007; Yost
et al. 2013). Yost’s (2007) study on Zizania concluded the presence of wild rice
phytoliths and pollen in modern lake deposits in Minnesota, which correlated to plant
density and stand location. The presence of wild rice phytoliths and pollen is found
in paleodeposits as well, and suggests a long-term reliance on wild rice by past and
present populations (Yost 2007).
Lake sediment cores from central Minnesota
analyzed by Yost et al. (2012) yielded an abundance of Zizania phytolith and pollen
grains. One core reached a depth of 1,400 cm, which dates to around 8,050 BC at
the bottom depth. Zizania was identified in this core beginning at 460 cm, with an
increase in pollen microfossil abundance towards the top of the sample (McAndrews
1999; Yost et al. 2013).
9
Ethnographic accounts by Jenks (1903) show that wild rice was important
spiritually, and as a form of sustenance. The Ojibwa occupied Northern Minnesota
in the last 250 years, and collected wild rice every autumn from many of the streams
and lakes. The name “Mun-o-min-ik-a-sheenh-ug, or Rice-makers”, was given to
the Ojibwe residing on the rice lakes of the St. Croix River (Jenks 1903). Currently,
wild rice can be found in shallow, muddy lakes and streams in Northern Minnesota
and is being cultivated and harvested by retailers with wild rice farms on the
Mississippi shores (Mallard Club Wild Rice 2014; Thomas 1996).
The
archaeological significance of wild rice in Northern Minnesota will be addressed in
Chapter four.
2.2 Paleoenvironment
Glacial History
Glacial history of the area has had a tremendous impact on current landforms
and topography shaping, and to some extent, how Woodland peoples interacted
with their environment (Wright 1993).
Ojakangas and Matsch (1982) described
current landforms formed by glaciation in Northeastern Minnesota as a peneplain
surface of rugged terrain, where outcrops are quite common, because drift is
minimal except in areas of moraines. The Black Duck and Guthrie are two till plains
reflective of these glacial movements forming modern topography (Hill and Huber
1996). Wright (1993) noted that the Alexandria and Itasca moraine now cover large
extents of Northeastern Minnesota (Figure 2.1).
10
Figure 2.1: Map of glacial lobes in Minnesota (Maple Hill Park.org).
From 12,500 to 9,900 B.P glacial advances occurred in the region leaving behind
three major proglacial lakes (Anderson et al. 1993; Hill and Huber 1996; Wright
1993). The three major drainage basins are the Rainy River, headward-eroding
tributaries of the St. Croix and upper Mississippi, and lastly Lake Superior, in which
North Shore streams flow (Ojakangas and Matsch 1982).
Wright (1993) and Ojakangas and Matsch (1982) described three lobes of
glacial ice active in the area between 30,000 and 14,000 years ago and accountable
for many of the Quaternary landforms visible today. They are: the Superior Lobe,
Rainy Lobe and the St. Louis Sublobe (Figure 2.2).
11
Figure 2.2: Glacial movement during the Late Wisconsin in Minnesota
(Wright 1993).
The Wadena Lobe, which is thought to be contemporaneous with the Rainy
Lobe moved southwards from Canada through north-central Minnesota (Wright
1993). This formed the Itasca and Alexandria Moraines about 20,000 years ago
(Hill and Huber 1996; Wright 1993; Dean and Megard 1993; Ojakangas and Matsch
1982). The final glacial advance of the St. Louis Sublobe of the Des Moines Lobe,
between 14,000 and 12,000 years ago, formed most of the landforms visible today
(Hill and Huber 1996; Dean and Megard 1993).
12
The Laurentide ice sheet expanded into Minnesota from Northern regions in
Canada around 30,000 years ago, and the Rainy lobe covered the upland west of
the Lake Superior basin, while the lowland now occupied by the Red River
channelled the Des Moines lobe (Wright 1993; Ojakangas and Matsch 1982). The
Wadena Drift represents the oldest surficial drift in the area, evident from the
Wadena drumlin field, marking the movement of the Wadena ice lobe. Prior to
14,000 years ago, the eastern margin of the Wadena Drift was deposited as the St.
Croix moraine of the Rainy lobe and the Itasca moraine outwash buried the northern
margin.
The Des Moines lobe from the west covered the Alexandria moraine
complex; however, the Wadena lobe may have deposited the core of the moraine.
Climate fluctuations caused the Wadena lobe to retreat and re-advance south of
Itasca Park, forming the Itasca moraine. The Superior and Rainy lobes terminated
at the north-south St. Croix moraine. Due to subsequent climatic shifts around
14,000 years ago, the Wadena lobe at the Itasca moraine ceased, as did the
Superior and Rainy river lobes at the St. Croix moraine. Only the Des Moines
moraine was active after the Rainy, Superior and Wadena lobes retreated from the
terminal moraines. The Des Moines lobe finally reached the terminus area 14,000
years ago and upon retreat, it left behind the Anoka sand-plain. A few proglacial
lakes remained after the retreat of the St. Louis Sublobe; Glacial lake Upham,
Glacial Lake Aitkin and Glacial Lake Koochiching, which drained into the developing
Lake Agassiz (Wright 1993).
Around 11,700 years ago the Laurentide ice sheet retreated north, and Lake
13
Agassiz grew in size, spreading northwest (Wright 1993).
Glacial ice began to
retreat 18,000 to14,000 years ago, and ice blocks within moraines began to melt,
which is evident from basal lake sediments (Wright 1993; Dean and Megard 1993).
By 11,500 BP, glacial ice had retreated from Northern Minnesota leaving Glacial
Lake Koochiching and Agassiz to remain covering large extents of land (Mulholland
et al. 1996). The ice eventually retreated so far to the north to enable the northward
displacement of Lake Agassiz, and after about 8,000 years ago, it eventually
drained eastwards towards into the St. Lawrence basin and finally northwards into
Hudson and James Bay (Tyrell Sea) (Wright 1993). The Marquette re-advance of
Superior lobe glacial ice took place around 10,000 BP, and this had an effect on
Minnesota’s northshore and areas east of Duluth, Minnesota (Phillips and Hill 2004).
Paleovegetation
Following ice retreat, different plant and animal species began to populate
Northern Minnesota. Evident from Appendix 1, Woodland peoples from 50 BC to
950 AD relied on some of these species, forming various subsistence strategies.
Bishop (2007), Gibbon (2012a), Hoover (1963), Mulholland et al. (1996), Peter and
Motivans (1983) and Shafer (2003) determined that the main faunal elements found
at some sites derive from deer and caribou, which were specifically listed in site
reports. Though what is found versus what was actually consumed may be quite
different because of high soil acidity levels rapidly decomposing smaller faunal
elements and macrobotanicals.
Phytolith and pollen analysis have both greatly contributed to paleovegetation
literature from the Great Lakes and research has shown vegetation changes
14
throughout the Holocene (Figures 2.3 & 2.4). Regionally, widespread vegetation
shifts correspond to climate, however locally; pollen shows shifts in relation to other
variables such as, fire, environmental processes and succession (Bradbury et al.
1993; Boyd et al. 2003; Wright 1993).
Figure 2.3: Vegetation in relation to different glacial surges versus time (Wright
1993).
15
Figure 2.4: Varve thickness, pollen levels for trees and shrubs, diatom flux and
elemental percentages versus time at Elk Lake (Anderson et al. 1993).
Between 20,500 to 14,700 years ago, tundra-like vegetation such as, Arenaria
rubella, Vaccinium uliginosum var. alpinum, Dryas integrifolia and Silene acaulis,
thrived in central Minnesota, and when Northern Minnesota began to deglaciate
around 14,000 years ago this vegetation expanded northwards (Birks 1976; Dean
and Megard 1993; Wright 1993).
Shrubs such as, willows, alder, Shepherdia
canadensis and Empetrum dominated regional flora between 14,700 to 13,600
along with dwarf birch. This was prior to the establishment of spruce-dominated
forests 13,600 years ago. Between 13,600 and 10,000 years ago, little change
occurred, except with the migration of deciduous vegetation into the region (sprucebirch, spruce-pine forests), and local changes to wetland species due to lake level
fluctuations. Ten thousand years ago, Pinus banksiana expanded, forming the
mixed coniferous-deciduous forest that can still be seen today (Birks 1976;
16
Mulholland et al. 1996; Wright 1993).
Similar to Mulholland’s et al. (1996)
assessment, Wright’s (1993) pollen study yielded high percentages of spruce before
the introduction of pine around 10,000 B.P. Dry, warm conditions allowed for a
prairie presence and moist, warm conditions developed deciduous forests creating a
combination of mixed forest and boreal zones (Anderson et al. 1993; Perkl 2009).
During the Middle Holocene (7800 to 4500 BP) Northern Minnesota was
affected by the eastward spread of prairie vegetation, and considerable fluctuations
in lake levels and stream flow (Bartlein and Watts 1993; Hohman-Caine and Goltz
1995; Wright 1993).
This is evident from pollen stratigraphy showing eroded
sediment caused by lower water levels at Lake Itasca (Wright 1993). In parts of
North American and Europe, the Middle Holocene consisted of drier and warmer
conditions than modern day, which influenced the migration of prairie-savannah
vegetation into the Great Lakes-St. Lawrence Forest (Anderson et al. 1993). Prairie
vegetation migrated approximately 120 km northward and eastward, and because of
the drought prone sandy soils, woodland prairie transition plant species moved
northeast an extra 75 to 100 km (Anderson et al. 1993; Hohman-Caine and Goltz
1995). Pollen cores record complete or partial drying of lake basins around 4,000
years ago, allowing for weed like annual vegetation to intersperse in wetter
conditions. This would have caused a decreased reliance on aquatic resources that
prehistoric peoples were heavily dependent on (Anderson et al. 1993; HohmanCaine and Goltz 1995). This period of warming (4,000 ka) was caused by climatic
oscillations and transitions in the three airstreams affecting Minnesota. The drier
climatic conditions caused the coniferous forests to be replaced with sagebrush,
17
grass and oak savannah (Anderson et al. 1993; Wright 1993). Varve thickness at
Elk Lake suggests a 600-year period, between 5.2 and 4.6 ka, of rapid reversal in
the trend towards warm conditions (Wright 1993). This reversal was caused by a
climatic variability pattern, linked to solar flux, zonal winds, and magnetic storms
(Anderson et al. 1993; Bradbury et al. 1993). Between 6900 and 4000 BP the pine
and oak or oak savannah intermingled with pine forest gradually moved westwards
(Hohman-Caine and Goltz 1995). Elk and bison would have been successful in this
environment, as evident from archaeological sites in the region (Hohman-Caine and
Goltz 1995; Appendix 1).
The climate during the Early to Middle Woodland Period (1000BC to 900 AD)
was cooler and defined by movements of deciduous forests replacing the prairies in
Northern Minnesota (Arzigian 2008). This environment would have fostered the
growth of aquatic and woodland plant resources, many of which appear in floral and
faunal assemblages from Woodland archaeological sites in the region. The return of
moist conditions around 4,000 ka led to the re-establishment of mixed wood tree
taxa such as birch and pine (Anderson et al. 1993; Anfinson 1997; Hohman-Caine
and Goltz 1995). Climatic patterns stabilized with high precipitation with resultant
increased lake levels. Aquatic floral and faunal populations would have increased
with lacustrine stability. Preceding the movement of white pine westward into Lake
Itasca County by 2700 B.P, there is a short increase in mesic deciduous forest.
Deciduous forest would have introduced species such as basswood (Tilia
americana) and wood nettle (Laportea canadensis); Hohman-Caine and Goltz
(1995) propose that fibres from these species may have been used for the net and
18
cord marks on Brainerd ceramic surfaces. The vegetation, upon early European
settlement, was a combination of aspen-birch, conifer and bog swamps, pine flats,
and jack pine barrens (Marschner 1974).
The climate, soils, geology, glacial history, fauna and flora of Northern
Minnesota determined resource availability and the chosen subsistence strategies of
Woodland peoples. The climate over time influenced environmental fluctuations and
glacial movement during the Holocene and up until the Woodland period, affected
the movement of people and animals. Acidic soils unique to Northern Minnesota
and a Boreal ecozone determine which materials survive, and which degrade,
changing site interpretations of subsistence. Glacial history of Northern Minnesota
changed the landscape to shape resource distribution and access. Lastly, the fauna
and flora that diffused into Northern Minnesota after ice retreat adapted to various
climatic shifts during the Holocene. The paleovegetation reacting to these shifts
eventually influenced what floral subsistence was available for Woodland
populations.
19
3.0 ARCHAEOLOGICAL SITES
3.1 Introduction
Sites were selected based on age, accessibility to material, stratigraphic
context and access to published reports. Age is an important factor because of
recent results showing cultigens in Manitoba, Saskatchewan and Northwestern
Ontario during the Late Woodland (Boyd et al. 2006; 2008; Boyd and Surette 2010;
Lints 2012). Material accessibility allows for a suitable number of samples to be
processed from a chosen site.
A larger sample size ensures more complete
interpretations and analysis, since aspects of food production can be better
assessed. With a clear idea of the sample’s stratigraphic context, it is easier to
determine age and interpret how samples relate to each other. Site reports provide
the details described above, so with records of where samples originate from within
sites, and/or site background, interpretations become more reliable. Sites chosen
include: Third River Borrow Pit, Big Rice, Windy Bead, Third River, Winnie Cottages,
Lost Lake, Kyleleen’s Tall Pine, Saga Island, and No Beard (Figure 3.1).
20
Figure 3.1: Location of the nine study sites in Northern Minnesota (Heather Hoffman
2014).
21
3.2 Third River Borrow Pit Site (21-176)
Located in Itasca County, Minnesota in the Blackduck District of Chippewa
National Forest, the Third River Borrow Pit (TRBP site) was described by Mulholland
et al. (1996) as containing abundant Brainerd pottery as well as some Laurel
remains. A feature containing fish bone, charcoal and pottery had four radiocarbon
dates processed from the charcoal and organic residue (Appendix 3).
The TRBP site report outlines that the site was discovered in an eroding bank
where cultural material began to emerge. The site is on a terrace above a floodplain
where post-glacial landforms and processes, namely the Bemidji Sand Plain glacial
landform, dominate the landscape. Sediments are predominantly sandy loam Typic
Glossaqualfs, while the northern area of the site contains loamy fine sands Alfic
Udipsamments.
A third type of soil is mixed typic Fluvaquents and Aquic
Udipsammets. Site vegetation consists of spruce in low areas and jack pine with
white and red pine on higher ground. Surrounding water bodies such as Dixon Lake
contained wild rice in the past, suggesting wild rice cultivation by TRBP site
inhabitants.
Twelve units were excavated at the TRBP in 10 cm levels, or until sediment
layers could be distinguished. Unit 2S contained the largest amount of ceramics in
the southern area; Unit 5S and Unit 11N also contained large quantities of ceramics.
Sixty-five lithic pieces were recovered reflecting the various stages including three
tools and 62 pieces of debitage; there is a 1:30 ratio for lithics to ceramics, indicating
an area for ceramic production. The type of lithic material indicates a variety of local
stone, including Hudson Bay Lowland chert, Swan River chert, basalt and quartz,
22
among others. Exotic Tongue River silicified sediment and Knife River Flint were
also identified. Fire cracked rock is associated with ceramics, indicating a wide
range of functions, such as ceramic firing, dispersal of hearth stones, and boil pits.
In all, 2,104 ceramic body and rim sherds were recovered during all stages of survey
and excavation.
Brainerd ware was the predominant type of ceramic found at the site; 50.6%
was Net Impressed and 6.3% was Horizontally Corded. Laurel ware comprises
10.1% of the recoveries and these have circular punctuates, are undecorated,
and/or have vertical linear impressions.
In total, there are 18 Brainerd Net
Impressed vessels, four Brainerd Horizontally Corded vessels, and seven Laurel
vessels. Residue was found on the interior of Brainerd corded sherds from Feature
1 associated with wood charcoal.
A Horizontally Corded sherd from the same
feature also has carbonized residue on the inside located at an angle to the rim.
Two features of interest were excavated at TRBP site. Feature 1/1A is located
in the corner of Unit 2S. The contiguous Unit 8S contains the second feature, and is
labelled Feature 1 (Figure 3.2).
Both features contained abundant amounts of
charcoal and ceramics. Within Feature 1A, a Brainerd Horizontally Corded body
sherd with adjoining rim sherd was found, as well as a second one 44 cm below the
surface within Feature 1A. These sherds contained carbonized food residue, and
are associated with charcoal deposits suggesting a hearth.
23
Figure 3.2: East planview of Feature 1, Unit 2S and 8S (Mulholland et al. 1996).
Flotation analysis was conducted on both Feature 1 and Feature 1/1A, and the
results yielded uncharred and charred floral material.
Chenopods, specifically
Chenopodium hybridum var gigantosperm were identified with other weedy flora.
Both in prehistoric and historic times, chenopods have been an important food
source (Mulholland et al. 1996; Yarnell 1964). Other plant species that Mulholland
et al. (1996) identified are elderberry (Sambucus sp.), smartweed (Polygonum sp.),
blueberry (Vaccinium cf. angustifolium), wild strawberry (Fragaria virginiana) and
raspberry (Rubus idaeus) macroremains.
Radiocarbon dates were taken from samples such as organic residue, and
charcoal; results showed varying dates (Appendix 9). Mulholland et al. (1996) give
24
two reasons for differences in returned dates between samples found 0.5cm apart in
the feature. One reason is that the carbonized food residue may be providing offset
dates. The second reason is that the charcoal may be extraneous, or contaminated
thus giving a younger date. Dating issues are common in the region, as explained
more in Chapter 4, section 4.3.1.
TRBP site is a predominantly Brainerd site
providing an excellent opportunity to study the makers of Brainerd ware and the
introduction of early cultigens into the region.
3.3 Big Rice Site (09-034)
This multicomponent site discovered in 1981 is located in Superior National
Forest, St. Louis County, Minnesota in the Superior National Forest Laurentian
District (Shafer 2003). As opposed to the TRBP site, the Big Rice site has
predominantly Laurel ceramics.
The Big Rice site is a large site where some
inhabitants produced Laurel ware and relied on wild rice (Shafer 2003). Valppu and
Rapp (2000) note the importance of the Big Rice site when their results showed that
wild rice was first processed and consumed in relation to the production of Laurel
ceramics during the Early/Middle Woodland around 50 BC (Appendix 2). The site is
situated above the Big Rice Lake in the sandy sediments of the Big Rice outwash
plain.
The soils at the site are composed of loamy sands and sands, and the
substratum is composed of gravel and sand. Unlike the TRBP site, the Big Rice site
is multicomponent, containing archaeological material from the Paleoindian,
Woodland and Contact periods. Valppu and Rapp (2000) note that stratigraphic
mixing is common in the upper levels from the site, evident from the mixture of Late
25
Woodland ceramic wares with Laurel materials. Soil samples analyzed for phytolith
and starch remains in this study were extracted from the base of features
surrounded by unmixed glacial deposits to prevent contamination and ensure that
the sediment was collected form Middle Woodland deposits. The processing and
gathering of wild rice at the site is suggested by ricing jigs and parching pits (Valppu
and Rapp 2000). Parching pits are evident from black or dark brown circles with
associated fire hearths (Figure 3.3, Shafer 2003).
Figure 3.3: Big Rice site feature 25, bottom of level 12, note
trowel for scale (Peters and Motivans 1983).
Other ceramic wares present at the site include Selkirk, Blackduck and Sandy
Lake making Big Rice one of the longest occupied sites in Northern Minnesota. A
minimum of 159 different Laurel vessels were found at the Big Rice site and these
26
vessels are separated into different design categories such as: 15 Laurel
Boss/Punctuate, 11 Laurel Plain, 27 Laurel Incised, 17 Laurel Push-pull, 67 Laurel
Dentate, 11 Pseudo-Scallop shell and 11 Plain (Shafer 2003).
Other artifacts found at the site include lithics, red ochre, copper in the form of
pressure flakers and historic material. Lithics are comprised of Knife River flint,
jasper taconite, Lake of the Woods chert, Hudson Bay Lowland chert, Gunflint silica
and Knife Lake siltstone.
Types of tools include retouched and utilized flakes,
projectiles and scrapers representing a low ratio of lithics to ceramics.
Different faunal material associated with the Laurel occupation suggests
habitation between July and October, furthering the case for wild rice processing,
which involves crop gathering during these months. Floral materials that were not
associated with any cultural period in particular include Chenopodium sp. seeds,
Polygonum sp. seeds, pincherry (Prunus pensylvanica) seeds, red-berried elder
(Sambucus pubens), blueberry (Vaccinium sp.), wild rice (Zizania aquatica and Z.
palustris) grains and wild red raspberry (Rubus idaeus) (Table 1, Valppu and Rapp
2000).
27
Table 1: Seeds from Area A (Valppu and Rapp 2000).
Species
Galium cf. aparine
Zizania aquatic
Chenopodium sp.
Polygonum sp.
Unknowns
Abies balsamea (needles)
Rubus ideaus-type
Gramineae (undiff.)
Prunus virginiana
Sambucus pubens
Diervilla lonicera
Scirpus sp.
Solanum dulcamara
Amelanchier sp.
Picea sp. (needles)
Prunus Americana
Total
Total Number
390
305
299
115
68
60
42
42
37
32
31
29
16
4
4
3
1477
Total Percent
26.4
20.6
20.2
7.8
4.6
4.1
2.8
2.8
2.5
2.2
2.1
2.0
1.1
0.3
0.3
0.2
100.0
3.4 Kyleleen’s Tall Pine (07-505)
Kyleleen’s Tall Pine is a habitation site located on a primary terrace by
overlooking Isabella River and contains large quantities of various lithics and
ceramics. An unspecified number of jasper taconite, Gunflint silica and Knife Lake
siltstone lithic objects were recovered, in addition to one jasper taconite scraper.
This is a multicomponent site with Brainerd, Blackduck and possibly Sandy Lake
ceramics. It was investigated in 2012 by a University of Minnesota field school after
the Pagami Creek fire swept through the area revealing cultural material. Modern
vegetation recorded included grasses, shrubs, pine trees, reeds and shoreline sweet
gale.
3.5 Kyleleen’s Bent Pine
28
This is another fire affected site located close to Kyleleen’s Tall Pine. The fire
exposed Middle Woodland cultural material. Site forms were not available for
Kyleleen’s Bent Pine (personal communication, Lee Johnson).
3.6 Winnie Cottages (02-526)
The Winnie Cottages site is located in Cass County, Minnesota, near an oxbow
of the Mississippi River. It was first surveyed in 1996, and from the excavation in
2007 a variety of lithics and ceramics were recovered. Winnie Cottages is situated
in the Bemidji Sand Plain; therefore, soils are mostly sands and gravels carried in by
meltwater from the Koochiching Lobe (11-13,000 years BP). The lithic assemblage
is small in comparison to ceramics and is comprised of 756 flakes and a core.
Ceramics total 3,387 sherds, and consist of mostly cordwrapped (1,950 sherds),
smoothed (120 sherds), and net-impressed (32 sherds). Lithic materials are jasper
silica, Gunflint silica, Knife Lake siltstone, chert, quartz, Knife River flint and
quartzite. Tools recovered numbered 25 in addition to five Late Woodland projectile
points. Conclusions based on the lithic assemblage are that local sources were
relied on more than exotic sources, and that tool manufacture and modification was
occurring on site. Local raw material that comprises most of the lithics includes,
Tongue River silicified sediment, quartz and cherts. Unit 5N 10W at a depth of 25 to
30 centimeters contained Brainerd ceramics and flakes with associated charcoal.
Other ceramics found elsewhere in the site include Blackduck, Psinomani, St. Croix
and Onamia. Ceramics indicate mostly Late Woodland occupation with a smaller
Brainerd component pre-dating the Blackduck, and Sandy Lake ceramics.
One
quartzite groundstone was recovered and excavators noted that there was no
29
evidence of battering. Faunal remains are few and poorly preserved, with some
identifiable ungulate, deer and rodents.
3.7 Windy Bead (05-373)
Windy Bead is a multicomponent semi-permanent habitation site that stretches
from the Paleoindian to Historic Periods.
The site is located in Lake County,
Minnesota, near Basswood Lake, on a westward extending peninsula. Survey and
reconnaissance was conducted in 1984, 1985, and testing took place in 2001, with
excavations in 2002 and 2003.
Different material collected includes buttons,
calcined bone, lithic debitage and tools, trade beads, musket balls, ceramics,
triangular projectile points, copper beads, copper awl/punch, groundstones, and cut
kettle brass/copper. Most of the ceramics are Laurel and have dentates, punctates,
incised rims or Pseudo-Scalloped Shell decoration (Figures 3.4 & 3.5). Soils are
organic-silt grading to sorted aeolian tossed sands and loess. Vegetation on site is
predominantly birch, balsam, hazel and cedar.
Figure 3.4: Ceramic sherds from the Windy Bead site.
30
Ceramic Decoration at Windy Bead
Dentate
Psuedo-scalloped Schell
Push-pull
Linear Stamp
Incised
Blackduck Types
Sany Lake Types
Net Impressed
Varia
Figure 3.5: Windy Bead ceramic decorative attributes (Lee Johnson n.d).
3.8 Saga Island (02-162)
The Saga Island site is located on an island near the American-Canadian
border in Cook County, Minnesota.
Laurel and Historic Period material was
recovered from the habitation site during survey and reconnaissance in 1982,
assessment in 1984, and testing in 2000. Ceramics, trade beads, heavy gauge
birdshots, calcined bone, and lithic debris make up most of the cultural material
(Figure 3.6). The soil is black humus over a dark brown sandy loam interspersed
with bedrock.
Local vegetation includes red and white pine and the on-site
vegetation is described as grass.
31
Figure 3.6: Laurel sherds from the Saga Island site.
3.9 Lost Lake (2-00LL)
Lost Lake material was recovered in the 1950s-1960s by an avocational
archaeologist and Chippewa National Forest employee.
It is described as a
Prehistoric Indian village site located in Beltrami County, Minnesota.
Remains
include 110 lithic flakes, a small faunal assemblage, and large quantities of
ceramics, which are predominantly Laurel, Blackduck, and Brainerd.
Brainerd,
Laurel, and Late Woodland sherds were analyzed in this study for food residue
analysis.
3.10 No Beard (05-264)
The No Beard site is located in Lake County, Minnesota; 20 units were
excavated in 2000 and 2002.
Faunal remains derive mostly from beaver, elk,
molluscs, fish, reptiles, but many of the bones are unidentifiable calcined fragments.
A number of bifaces, copper tools, projectile points, scrapers, and ceramics were
also recovered. The majority of lithic material is Gunflint silica, Knife Lake siltstone,
jasper taconite, jasper siltstone, Hudson Bay Lowland chert, Knife River Flint,
quartzite, jasper, and granite.
Laurel ceramics were identified during the
32
excavations.
3.11 Samples
Food residue is found in a variety of contexts, including carbonized food
residue found on ceramics, and non-carbonized food residue found on lithics and in
soil samples (Horrocks 2005; Gott et al. 2006; Thompson et al. 1994). Food residue
can also be identified from dental calculus (Hardy et al. 2009); however; obtaining
dental calculus samples was beyond the scope of this study. Carbonized food
remains can be found adhering to the inside and outside of ceramic vessels, which
is a clear indicator of the foods consumed by the vessel producers (Boyd et al.
2006; 2008; Boyd and Surette 2010, Roper 2013; Thompson et al. 1994). Noncarbonized food remains are evidence of plant presence in soil samples, or
processing methods with grinding stones (Horrocks 2005; Gott et al. 2006; Zarrillo
and Kooyman 2006). Residue can vary in thickness, depending on preservation
and the extent of processing or cooking. Thicker residues are more likely to yield
samples of sufficient size (over 50 milligrams) to permit radiocarbon dating to help
interpret aspects of when certain plant species were available to site occupants
(Roper 2013).
Food residues extracted from ceramics are direct indicators of plant
consumption because the recovered phytoliths, starch granules and pollen
represent the plants cooked in meals (Boyd and Surette 2010). Ceramics that were
used more often for cooking will have a thicker layer of residue and preserve the
underlying residue (Figure 7.4). Lithics such as grinding stones or fire-cracked rock
33
are useful indicators of plant use because microfossils tend to become entrapped in
microfissures and preserved (Barton 2007; Zarrillo and Kooyman 2006).
The
analysis of two groundstones analyzed from the Windy Bead and No Beard site
revealed evidence of grinding (Figures 7.5 & 7.6). Soil samples are another context
used to analyze non-carbonized food residue in the form of phytoliths, starch and
pollen (Horrocks 2005). Over time, organic materials break down and decompose,
leaving behind identifiable traces of plant material including the microfossils in the
soil (Horrocks 2005).
Often, soil samples will be collected from archaeological
hearth features in search of micro-botanical traces of past meals.
Although samples were analyzed from nine sites in Northern Minnesota, only
eight sites had accessible literature. The ninth site, Kyleleen’s Bent Pine, is located
in a region decimated by fire, which led to the exposure of Middle Woodland cultural
material on the surface. This site was included in this study, because the Middle
Woodland period ceramics contained food residue. Only the Third River Borrow Pit
and Big Rice have dated material from residue and/or organic charcoal.
The
analyzed cultural material in the form of Brainerd and Laurel ceramics are
confidently dated to Early, Middle Woodland Period in Northern Minnesota. All nine
sites are located in the Superior National Forest and Chippewa National Forest, a
heavily forested region well known for the many lacustrine sources dotting the
landscape.
34
4.0 THE WOODLAND TRADITION IN THE UPPER GREAT LAKES
4.1 INTRODUCTION
The Woodland Period lasted from around 1,200 BC to 1,600 AD throughout
much of the Eastern Woodlands and adjacent biomes in North America, and is
chronologically situated between the Archaic and Historic Periods (Hamilton et al.
2011). Diagnostic artifacts, stratigraphic context and AMS dates differentiate these
sub-periods, but are not uniformly expressed throughout the Eastern Woodlands
(Thomas 1996). For example, the Early Woodland has not been identified in
Northern Minnesota, leading researchers in Northern Minnesota to generally use the
term “Initial Woodland” to describe the Middle Woodland, and “Terminal Woodland”
to describe the Late Woodland (Hohman-Caine and Goltz 1995; Shafer 2003;
Thomas 1996). This study will use the conventional tripartite division to separate
the Woodland Period in Minnesota.
4.2 EARLY WOODLAND (1200 BC TO 250 AD) ACHAEOLOGICAL COMPLEXES
Changes that took place during the Late Archaic to Early Woodland transition
are reflected in transforming site settlement patterns, technological innovations and
spiritual practices (Fitting 1969; Perkl 2009). The changes in settlement patterns
are thought to reflect gradual response to population increases (Arzigian 2008).
Ceramic production begins during the Early Woodland and projectile point
morphologies change from dart to triangular points. Burial mound ceremonialism
began with large burial mounds becoming a conspicuous feature of the Woodland
Period in some regions (Lugenbeal 1976; Shafer 2003; Thomas 1996).
35
Several models have developed to explain the transition from the Late Archaic
to Early Woodland in southeastern Minnesota. These include climate change and
adaptation to changing conditions, population expansion causing warfare, and the
replacement of Late Archaic populations by intrusive groups or immigrants with
different cultural backgrounds and technology (Arzigian 2008; Perkl 2009). In the
1950s generalizations were offered that Archaic populations settled in camps, while
subsequent Woodland populations lived in villages as their populations increased
(Fitting 1969), but this distinction might be over-stated. This over-statement reflects
older ways of thinking, when the nature and subtly of the Archiac to Woodland
transition was considered much simpler.
The study of settlement patterns is important because it can reveal humanenvironment
interactions,
migrations,
and
economic
development
patterns
(Silbernagel et al. 1997). To procure settlement data from the Archaic to Historic
Periods, Silbernagel et al. (1997) examined a number of different sources including
ethnographies, ethnohistories, resource availability, in addition to archaeological
investigations and geomorphic mapping. Based on the data, Archaic lacustrine sites
are located primarily in complex shoreline areas of shallow water and were probably
occupied when freshwater fish were spawning. These areas also provided access
to deer, wetland plants, beaver, moose and chert bedrock outcrops. Woodland sites
are generally focused on interior wetland resources and are sometimes
reoccupations of older, Paleoindian, sites (Silbernagel et al. 1997). Site settlement
reflects resource access because certain resources become available through trade
and seasonal exploitation at different times of year and in various locations.
36
Silbernagel et al. (1997) suggests that Early Woodland and Late Archaic sites
exhibit similar seasonal occupation cycles, with population aggregation during winter
months and dispersal in the summer.
Thomas and Mather (1996) suggest the
opposite; their data from the McKinstry site support the notion that populations
concentrated in the warm months and dispersed in the winter. In the Boreal Forest
to the north, broad spectrum foragers concentrated in warm months and dispersed
in the winter, this was reinforced by subsistence economies utilizing wild rice
gathering in the fall, fishing in the spring/summer, and land mammals in the winter
(Meyer and Hamilton 1994; Reid and Rajnovich 1991). Site variation between Late
Archaic and Early Woodland settlement patterns emphasizes differences in
resource access throughout Minnesota.
Technological innovations during the Early Woodland include changes in
projectile point morphology and the introduction of ceramics around 200 BC
(Anfinson 1997). Small projectile points with expanding bases, associated with the
Point Peninsula complex, appear to diffuse Northeast during the Early Woodland
(Fiedel 2001; Lee 1954).
Many scholars characterize the Early Woodland as
beginning with the introduction and production of ceramics for efficient food
processing (Fiedel 2001; Hamilton et al. 2011). The transition from bark, leather
and/or wood containers to fired clay can be explained as resulting from a greater
need for food storage and higher cooking temperatures. Other factors affecting the
need for ceramic vessels include the introduction of new plant species that may
have required higher cooking temperatures and increasing populations that
warranted more food storage (Fiedel 2001; Skibo 2013). Additional possibilities for
37
the transition to ceramics are the ability to recycle ceramic containers, create multifunctional vessels and use them in ritual functions (Skibo 2013). Although these are
possible models, why ceramic technology emerged is a complicated question and
one that is still unanswered. Ceramics play an important role in the consumption,
production, processing and storage of food, and may have affected changing
subsistence practices (Fiedel 2001; Skibo 2013).
Ceramic types are used to differentiate between the Early, Middle and Late
Woodland; these types are defined by differences in decoration, style, shape and
vessel size (Figure 4.1). Different types of ceramics are present in Minnesota, and
they have assisted in dating associated assemblages based on vessel
characteristics (Table 2). Early Woodland ceramics have thick tempered vessel
walls and are conoidal in shape with minimal decoration.
The following figure
illustrates ceramic development over time in different ecological areas of North
America.
38
Figure 4.1: Ceramic typologies from the Sub-Arctic, Eastern Woodland and
Northern Plains versus time (Hamilton et al. 2011).
Table 2: Major ceramic types dating to the Early, Middle and Late Woodland periods
in Minnesota (Anfinson 1979).
Early Woodland
Middle Woodland
Late Woodland
La Moille Thick (500-350 BC)
Brainerd (1600 BC to 400
AD)
Blackduck (800-1400 AD)
Malmo/Kern Series (800 BC
to 200 AD)
Laurel (100 BC- 1000 AD)
Blue
Earth/Correctionville
Phase (1000-1600 AD)
Fox Lake Phase (200BC-900
AD)
Cambria Phase (1000-1300
AD)
Howard Lake Phase (200 BC
to 300 AD)
Pokegama Smooth (0 to 500
AD)
St. Croix (500 to 800 AD)
Lake Benton Phase (900 to
1500 AD)
Onamia Series (800 to 1000
AD)
Orr Phase (1300 to 1800 AD)
Great Oasis Phase (900-1250
AD)
Sandy Lake Ware (1000 to
1750 AD)
39
4.3 THE MIDDLE WOODLAND (100 BC TO 500 AD) ARCHAEOLOGICAL COMPLEXES
New ceramic forms, earthworks, burial mounds, and the construction of food
storage pits define the Middle Woodland period in North America (Arzigian 2008;
Mason 1981). In Northern Minnesota, the best indicators of the Middle Woodland
are distinctive lithic technologies, new ceramic forms, and burial mound
ceremonialism (Shafer 2003). Associated projectiles can be broadly, shallow, or
corner notched, and/or have an expanding or straight stem.
The lithics are
manufactured from local and exotic materials including both Burlington chert and
Knife River Flint (Arzigian 2008; Hohman-Caine and Goltz 1995; Thomas and
Mather 1996; Shafer 2003). Wood working tools, small to medium sized scrapers,
wedges, harpoon points, beaver incisors used as cutting tools, and copper items
such as pressure flakers, beads, and awls, are also present in Middle Woodland
occupations (Arzigian 2008; Donaldson and Wortner 1995; Shafer 2003).
Gibbon and Hohman-Caine (1980) recognize a shift to focal subsistence
practices as opposed to more diffuse strategies, which affected settlement patterns
in some areas of the Upper Great Lakes. Instead of re-settling different areas to
take advantage of seasonal resources, populations settled in one area for longer
periods of time and exploited the local resources. The increased density and size of
crop production by some Middle Woodland populations may have resulted in larger
settlement sizes (Fritz 1993; Scarry and Yarnell 2011). These crops could include a
variety of plant species researched in this study such as maize (Zea mays ssp.
mays) and wild rice (Zizania sp.). Middle Woodland settlements were occupied
40
more heavily in the summer and populations relied on plant cultivation and fishing.
Winter settlements were smaller and more dispersed, with an increased reliance on
small and large mammals for sustenance. During the winter months, most groups
would have preferred wooded areas with wind protection; it is therefore not
surprising that in southwestern Minnesota, there are site clusters around lakes, on
islands, and on peninsulas
were fuel and shelter might be concentrated and
protected from periodic destruction by prairie fire (Anfinson 1997). Such locations
would have also granted groups access to the denser concentrations of aquatic
resources and provided protection from enemy groups (Anfinson 1997; HohmanCaine and Goltz 1995). The transition from the Middle to Late Woodland ended by
600 AD to 800 AD in Eastern Minnesota (Gibbon and Caine 1980).
Brainerd Ceramic Assemblages
Sites with Brainerd ceramics are located in north central Minnesota (Figure
4.2), Montana and Manitoba (Anfinson 1979; Hohman-Caine and Goltz 1995; Norris
2007). Current research suggests that Brainerd ware in Northern Minnesota is the
same as Rock Lake ware found in Manitoba, and extending through Saskatchewan
into Alberta (Arzigian 2008; Hohman-Caine and Goltz 1995; Hohman-Caine et al.
2012; Norris 2007). The literature is somewhat divided between research in Canada
and the United Sates, which in the past caused different labels to be applied to the
same ceramics. As of 2013, there are 195 multicomponent sites and 99 single
component Brainerd sites identified in Minnesota (Personal communication with
LeRoy Gonsior, Minnesota Historical Society Archaeology Department, 2013).
41
Figure 4.2: Spatial extent of Brainerd assemblages in Minnesota (Anfinson
1979; Hohman-Caine et al. 2012).
Hohman-Caine and Goltz (1995) place Brainerd ware in the artifact
assemblage of a cultural manifestation called Elk Lake culture. The Elk Lake culture
dates from the Late Archaic to the Early Woodland (Hohman-Caine and Goltz 1995).
This varies from previous assumptions that Brainerd ware production began during
the Middle Woodland. Brainerd ware dates are subject to debate, and there is a
wide variety of contrary data suggesting Late Archaic/Early Woodland production.
Projectile points associated with Brainerd ware range from 27 to 45 millimeters in
length.
Additional lithic tools associated with the Elk Lake culture include
rectangular/square
wedges
or
chisels,
and
scrapers
of
various
sizes.
Archaeological evidence shows that bison, deer, elk and wild rice were significant
42
dietary sources for Elk Lake populations in the cool, moist prairie/woodland ecotone
(Arzigian 2008; Hohman-Caine and Goltz 1995). The radiometric dates collected by
Hohman-Caine and Goltz (1995) begin at 1550 BC, and end around 750 BC, and
this may indicate a preference for the prairie-woodland ecotone present at the time
in northcentral Minnesota.
The number of Brainerd sites increased in the area
around 2,700 years ago coinciding with the return of white pine to the region’s
western edge (Wright 1968).
Forests had fully returned to the region after the
decline of Brainerd sites around 350 BC (Hohman-Caine and Goltz 1995; Wright
1993).
Brainerd ware was first identified at the Gull-Lake Dam site (Figure 4.3), a
multicomponent habitation and mound site located in central Minnesota. There are
two periods of occupation at the site; the first occupation is 800 BC to 200 AD, and
the second Brainerd occupation dates from 600 AD to 900 AD (Zellie 1988).
Notably the second set of dates from 600 to 900 AD varies considerably from the
dates proposed by Hohman-Caine and Goltz (1995). The site consists of twelve
round, elongated, and linear mounds close to the eastern side of the dam.
Stratigraphic contexts from Gull Lake Dam show that producers of Brainerd ware
may have used these earthen burial mounds. At this site, the Brainerd ware is also
associated with scrapers, brown chalcedony debitage and side-notched projectile
points (Anfinson 1979).
43
Figure 4.3: Map of Minnesota displaying the Gull Lake Dam site, King Coulee,
Ogema-Geshik, and Cass Lake (Hohman-Caine and Goltz 1995; Perkl 1998;
Thompson et al. 1994). Base map from http://www.geology.com.
In Northern Minnesota sites, the main morphological trait differentiating
Brainerd from other ceramic wares is horizontally corded or net-marked surfaces
(Arzigian 2008; Thompson et al. 1995).
Vessel interiors are plain and only
sometimes exhibit discontinuous net impressions (Hohman-Caine and Goltz 1995).
Horizontally corded Brainerd ware reveals external surfaces impressed with fine
cords in a z-twist pattern, applied obliquely or horizontally to the rim (Anfinson 1979;
Hohman-Caine and Goltz 1995).
Net-impressed Brainerd ware exhibits net
dragging, causing exterior net impressions and intentional smoothing impressions to
occur (Figure 4.4). Brainerd vessels, resemble Laurel and Malmo vessels, and tend
44
to have minimal neck constriction with conoidal body forms, and vertical rims
(Arzigian 2008; Thomas 1996).
Figure 4.4: Brainerd net-impressed sherd from the Windy Bead site.
4.3.1 MIDDLE WOODLAND CHRONOLOGICAL ISSUES
There is considerable debate regarding the age of Brainerd Ware pottery in
Northern Minnesota (Thompson et al. 1995). Much of this debate surrounds a set of
surprisingly early dates obtained on food residue from Brainerd ceramics, charcoal
and bone. These early dates suggest that while Brainerd producers existed during
the Middle Woodland Period (Personal Communication with Taylor-Hollings,
University of Alberta, 2014), the early dates suggest Brainerd also dates as early as
the Late Archaic to Early Woodland Periods (Hohman-Caine and Goltz 1995;
Hohman-Caine et al. 2012).
Most of the controversial AMS dates have been
collected from organic residue adhering to Brainerd ware, leading to the possibility
that they are too old due to contamination from old carbon (Hohman-Caine and
Goltz 1995). Table 3 provides an overview of Brainerd ware dating and the changes
45
that occurred from the 1950s when Brainerd ware was classified as Late Woodland
to the 1990s, when Brainerd ware was reclassified as Late Archaic.
Table 3: Brainerd ware dates across Northern Minnesota (Hohman-Caine et al.
2012).
Sources
Gull Lake Dam and
Headwaters Reservoir
Date
Late Woodland
Sources
Wilford (1955) and Evans
(1961)
Relationship between
Brainerd and Laurel
ware at McKinstry and
Lockport sites.
Brainerd ware also
underlies Blackduck
and Sandy Lake
pottery at several
locations.
“Stratigraphic
associations with late
Middle Woodland St.
Croix Stamped
ceramics at the Gull
Lake Dam site.”
18 dates, 16 of which
were radiocarbon
dates.
Middle Woodland
Lugenbeal (1978) and
Thompson et al. (1995)
AD 600 to 800
Anfinson (1979)
1600 BC to 400 AD
Hohman-Caine and Goltz
(1995)
One factor for these earlier dates may be the freshwater reservoir effect (FRE),
which is defined as the introduction of old carbon into archaeological materials
causing offset dates (Arzigian 2008; Hamilton et al. 2011; Roper 2013). The FRE is
similar to the marine reservoir effect, which occurs when dissolved minerals or hard
water affects
δ14
C values (Fischer and Heinemeier 2003). The marine reservoir
effect contributes to the offsetting of dates on marine materials, as well as on bone
collagen from populations that consumed large amounts of marine resources
46
(Yoneda et al. 2002).
Unlike the marine reservoir effect, the FRE has not
conclusively proven to contribute to offset AMS dates on organic residue at inland
sites (Fischer and Heinemeier 2003). Until solid evidence comes forth and unless
the dates are significantly offset, it is possible that these earlier dates are correct
(Hart et al. 2013; Hart and Lovis 2014).
For example, Hohman-Caine and Goltz (1995) analyzed 18 radiometric dates
gathered from organic residue on Brainerd ceramics or associated contexts at nine
sites. These dates expand the time range of Brainerd ware well outside of the
accepted Middle Woodland range of 0 BC to 500 AD. Most of the dates HohmanCaine and Goltz (1995) collected are from ceramics and charcoal from hearths.
Direct dates deriving from cooking residue on vessels are more beneficial since they
directly relate to the use of the pottery in food preparation, while the wood charcoal
dates rely upon indirect associations of wood burning in hearths with pottery that
may or may not be directly linked stratigraphically to the pottery production and use
(Hart et al. 2013). In addition, wood charcoal is less secure than organic residue
because of the risk of contamination from modern root penetration of the charcoal,
or from bioturbation processes that might result in mix carbon from unrelated
sources into the sample. This is a real possibility in light of the taphonomic
complexity of deposition, generally poor stratigraphy and chronic forest fires in the
Mississippi Headwaters region.
Another consideration when indirect dating
archaeological deposits using charcoal from hearths is that the fuel wood might
derive from old trees that might predate the human occupation by several hundred
years (Hohman-Caine and Goltz 1995). For these reasons, Hohman-Caine and
47
Goltz (1995) accept the dates from ceramic residue and reject the wood charcoal
dates, which would mean that Brainerd ware is the earliest ceramic ware in the
upper Great Lakes dating from 1600 BC to 400 AD (Arzigian 2008; Hohman-Caine
and Goltz 1995; Thomas 1996).
Hohman-Caine et al. (2012) conducted a large study of Brainerd ware
chronology using AMS and OSL (Optically Stimulated Luminescence) dates. Of the
81 ceramic, bone and charcoal samples, only 51 dates are deemed acceptable
because of possible contamination in the thin, disturbed, sediments of Northcentral
Minnesota. Additionally, perhaps caused by the FRE, the dates from some of the
ceramic samples might be artificially too old for pottery production in the region.
Residue dates that are significantly early for the cultural assemblage also have more
negatively or less positive
13
C/12C isotope ratios, suggesting a connection between
offset dates and more negative isotope ratios. The issues with Brainerd ware are
that the early dates may be contaminated from a number of sources (Hohman-Caine
et al. 2012). As more studies are conducted on the FRE and ways to adjust for FRE
contamination are developed, researchers will be able to adjust the early Brainerd
dates and provide a more accurate start for Brainerd production in Minnesota
(Hohman-Caine et al. 2012; Roper 2014). From the eight acceptable OSL dates, 39
acceptable ceramic residue dates, and projectile point morphology, Hohman-Caine
et al. (2012) situate Brainerd ware within the Early Woodland, beginning at 2750 B.P
and ending at 1700 B.P.
A study of
14
C ages from ceramic residue by Roper (2013) finds age offsets
between AMS dates on carbonized food residue and charcoal. This recent study
48
attempts to address offset AMS dates, such as the ones encountered by HohmanCaine and Goltz (1995). Possible causes for these offset dates could include the
FRE, contamination from shell-tempered sherds, nixtamalizing of corn, organic
matter in vessel fabric, fish consumption, vessel temper, and the use of hardwater
for cooking (Roper 2013). Results from Roper’s study show that the introduction of
old carbon into residue may be a large factor in offset dates and a major cause of
dates that are too early for the cultural assemblage (2013), and could explain the
erroneously old dates collected from Brainerd ceramics by Hohman-Caine et al.
(2012).
In contrast, while Hart and Lovis (2014) do not dispute the presence of a FRE,
or the apparent inaccuracies of AMS residue dates, they do recognize the low
validity applied to statistical tests on offset AMS dates from organic residue. Low
statistical validity is a result of researchers not considering sample size and
conducting tests on low quantities of data. To exemplify studies with low statistical
validity, Hart and Lovis (2014) bring into question Roper’s (2013) study on offset
dates from organic residue and her possible causes, including shell temper in
vessels and nixtamalizing corn. Hart and Lovis (2014) contend that shell temper in
vessels causing offset AMS dates is an unlikely scenario because CaCO 3 is only 12
percent carbon and the shell quantities exposed to leaching are small, making the
ancient carbon concentrations in shell unable to constitute the resulting offset
carbon 14 ages of 115 to 442 years. Since Brainerd ware is not shell tempered, any
contribution shell temper may have to providing offset dates is nullified.
Hart and Lovis (2014) also reject nixtamalizing of corn to explain anomalously
49
old dates on residue because of low statistical validity. Nixtamalizing describes the
alkaline processing used to produce hominy with lime from shell lime, limestone or
wood ash. This is done by heating calcium carbonate (in the form of shell, or
limestone) between 600 to 900 degrees Celsius. In order to account for age offsets,
there has to be 80 percent carbon from at least a 50-year-old wood, or 40 percent
carbon from a 100-year-old wood. Only 0.1 percent to 5 percent lime by maize
weight is used in traditional nixtamalization, and only 5 percent to 30 percent carbon
is present in wood ash; therefore, the proportions of wood needed to result in
significant age offsets are unlikely.
As discussed by Hart and Lovis (2014), evidence of fish consumption as a
source of ancient carbon causing offset AMS dates on organic residue has also
been rejected because it is considered statistically weak. The carbon content in fish
varies by species type and fat content, so in statistically valid offsets there must be
large quantities of fish cooked in relation to other foods and/or a high percentage of
old carbon in the aquatic system (Hart and Lovis 2014). In effect, simply recovering
fish or mollusc remains at archaeological sites does not confirm the presence of the
FRE (Hart et al. 2013).
Carbonates from marl deposits, bedrock and calcareous materials are likely
the main sources of radiometrically dead carbon resulting in questionably early
dates (Arzigian 2008; Hohman-Caine et al. 2012; Hart et al. 2013). Ancient carbon
is sequestered in streams and lakes and can enter freshwater systems through
groundwater, contributing to the geologic substrate of an area. When calcareous
bedrock, soils with calcareous substrata or glacial till begin to weather, the
50
groundwater becomes contaminated with ancient carbon (Hart et al. 2013).
Hohman-Caine et al. (2012) analyzed the nature and extent of geologic formations
near Brainerd sites in order to identify if the carbonate content in geologic deposits
was causing offset dates. Although surficial glacial drift does cover the study area,
the later (i.e. more acceptable) Brainerd dates apply in areas where there are
significant surficial carbonate deposits. If carbonate deposits were the cause of
offset Brainerd dates, marl deposits, bedrock and calcareous material would have
been found near areas of older dates, not younger ones (Hohman-Caine et al.
2012). Although, it is possible that ceramic vessels were not produced near clay
sources and the material used to make vessels was located near areas of carbonate
deposits and carried into areas of ceramic production.
The paste used to make vessels, as well as boiling hard water in the finished
vessel are two additional factors that may introduce old carbon into samples causing
offset OSL and AMS dates (Hohman-Caine et al. 2012). Clay sources used to make
Brainerd vessels were generally collected near surface deposits that were depleted
in carbonates; therefore, the ceramic paste could not include old carbon.
Full
alkali/acid/alkali pretreatments on Brainerd sherds prior to AMS dating would
potentially remove any possibility of old carbon contamination from boiling hard
water in the finished vessel. The paste used to make Brainerd ware and boiling
water in the finished vessel are unlikely sources of offset OSL and AMS dates
(Hohman-Caine et al. 2012).
Hohman-Caine et al. (2012) and Fisher and Heinenmeier (2003) identify that
significant indicators of offset dates are found in carbon and nitrogen stable isotope
51
values present in chemical analysis. Put simply, the isotopic composition of key
elements (notably the relative abundance of
13
C relative to
C (δ13) or
12
15
N to
14
N
(δ15) reflect different photosynthetic pathways, the difference between terrestrial
versus marine foodwebs, or the relative importance of some cultigens. This has
implications for radiocarbon dating since elevated values of
13
C/12C and 15N/14N may
indicate food sources that could contribute to offset dates. In contrast, when Hart et
al. (2013) reviewed δ13C and δ15N values from the Finger Lakes region in New York
and from sites in Greenland, they found that increased consumption of maize at
sites in the Finger Lakes region directly correlates to an increase in isotropic ratio or
elevated
13
C/12C values (maize is a C4 plant that is more likely to integrate
13
C into
its structure). Enriched δ15N values from modern plants have been found in some
Greenland archaeological middens containing marine organisms ( 15N is found in
enriched abundances in marine contexts, and decomposition of marine food wastes
allows a higher abundance in terrestrial plants growing on such middens that would
normally be expected of terrestrial plants). Elevated N isotopic ratios have also
been linked to the enrichment of N15 in soils from organic fertilizers (Hart et al.
2013). As a result,
13
C/12C and
15
N/14N values may be poor indicators of offset
dates in some regions because external factors cause increased or decreased
values.
Despite numerous studies, no conclusions about the exact cause of offset
AMS dates could be made; furthermore, old carbon does not seem to have
implicated Brainerd ware dates in Northern Minnesota (Hart and Lovis 2007a and
2007b; Hart et al. 2013; Hohman-Caine et al. 2012; Roper 2013). It is likely that
52
sample contamination is the result of a number of factors, or that some of the early
dates are correct (Hart and Lovis 2014; Roper 2013). Archaeological chronologies
tend to be rigid and “law-like,” making it difficult for researchers to challenge
established chronologies. There is no dispute that the FRE may be the cause of
offset dates, but at issue is how dramatic that offset might be. Until solid evidence
comes forth, researchers should entertain the thought of some early dates being
correct (Hart et al. 2013; Hart and Lovis 2014).
Laurel Phase (100 BC – AD 1000)
The Laurel phase is associated with some of the earliest and most widespread
pottery in the region (Mason 1981; Thomas 1996). Notably, given assumptions
about aquatic diets in the Boreal Forest, residue dates on Laurel material should be
older than expected, due to the FRE. Wilford was the first scholar to place Laurel
material within the broader scope of Minnesota archaeological material (BrandzinLow 1997). Laurel assemblages extend all the way from Northern Minnesota, to
east-central Saskatchewan, Northern Ontario and the southern half of Manitoba
(Anfinson 1979; Brandzin-Low 1997; Thomas 1996; Figure 4.5).
53
Figure 4.5: Distribution of Laurel assemblages across Eastern North America (Meyer and
Hamilton 1994).
Morphological similarities between ceramics and stratigraphic associations
show that Laurel ware may have been influenced by Howard Lake, and/or Malmo
producing cultures (Anfinson 1979). Laurel ceramics have a similar distribution to
Blackduck assemblages, which has some researchers speculating the replacement
of Laurel ware with Blackduck ware at the end of the Middle Woodland (Thomas and
Mather 1996).
Dr. Andrew C. Lawson and George Bryce first recorded Laurel ware in 1884 at
the Minnesota Mounds. The Mckinstry Mound 1 was crudely excavated in 1896 by
L.H. Kempton and G.G. Hulber (Brandzin-Low 1997; Thomas and Mather 1996).
This site was possibly a fishing station, in addition to a locus for mortuary
ceremonialism, based on the location of the site and the presence of burial mounds.
Lithic production does not seem to have been a major activity at this site, and the
54
large amount of ceramics suggests a firing area (Thomas and Mather 1996).
Laurel sites are commonly found near water, in the vicinity of rivers, lakes and
rapids (Dawson 1981; Janzen 1968).
At many of these sites, there is aquatic
resource exploitation of wild rice, fish and other wetland plant species (Meyer and
Hamilton 1994). Laurel populations utilized seasonal resources, such as fish in the
summer and land mammals in the winter (Meyer and Hamilton 1994).
Laurel ware is composed of many different traits and/or decoration styles
(Anfinson 1979; Meyer and Hamilton 1994). Vessels are typically coiled when there
is an absence of decoration. The rim is straight, everted or inverted slightly, while
the lip is flattened and unthickened.
There is a slight neck constriction, the
shoulders are gently rounded, the base is conoidal and the body is elongate globular
(Anfinson 1979; Thomas and Mather 1996). Decoration is usually on the upper lip
and neck, and consists of mostly punctates, incising, dentate stamping, bosses,
push-pull bands, cordwrapped stick impressions and pseudo-scallop shell stamps
(Figure 4.6, Meyer and Hamilton 1994).
55
Figure 4.6: Laurel rim sherd from the Saga Island site.
The Laurel culture is connected with Hopewellian groups from Ohio and Illinois,
as suggested by the presence of burial mound ceremonialism sometimes with exotic
grave inclusions in some contexts, most notably northern Minnesota and the
boundary waters (Hamilton et al. 2011; Mason 1981; Thomas and Mather 1996).
Interactions between Avonlea groups on the Canadian Plains and Laurel groups
from the Boreal Forest are also evident from chronological and archaeological
evidence (Meyer and Hamilton 1994). Avonlea groundstone celts that are similar to
ones found in Laurel sites were recovered at the Miniota site in Manitoba (Landals
1995). The Gravel Pit site in Northern Saskatchewan contained Avonlea and Laurel
material in close stratigraphic association, signifying trade and cultural contact
(Klimko 1985; Meyer and Walde 2009). The appearance of Avonlea ceramics and
artifacts in Laurel components, and vice-versa, supports cultural interactions
between these groups, particularly along the forest and plains ecological boundaries
56
(Meyer and Hamilton 1994; Morgan 1979).
Associated with Laurel ceramics are various lithics made from Yellowstone
obsidian, and Knife River Flint and a preponderance of local materials suggesting
extra-regional exchange networks (Lugenbeal 1976; Thomas and Mather 1996).
The rest of the Laurel artifact assemblage includes: a high ratio of scrapers, antler
harpoons, copper tools, ornamental artifacts, notched and stemmed projectile
points, and hafted beaver incisors (Stoltman 1973). Burial mounds accompanied by
red-ochre and secondary burials are found at some Laurel sites (Brandzin-Low
1997; Janzen 1968; Thomas and Mather 1996).
Evidence of wild rice has been recovered from a number of Laurel sites in
Northern Minnesota and Northwestern Ontario, and this local wetland plant was an
important component of Laurel diet (Boyd and Surette 2010; Surette 2008; Valppu
and Rapp 2000). Charcoal from the Big Rice site in Northern Minnesota was AMS
dated to 50 BC and the charcoal was stratigraphically associated with wild rice
kernels (Zizania aquatica). In the upper levels of the Big Rice site, many ricing
features containing mixed ceramics were also recovered (Valppu and Rapp 2000).
Surette (2008) identified wild rice phytoliths from organic residue in several Laurel
vessels from Northwestern Ontario. Hamilton and Nicholson (2006) suggests that
organic remains recovered at Laurel sites indicate a broad hunting and gathering
subsistence strategy that relied on seasonal rounds.
These seasonal rounds
included wild rice gathering in the fall and sturgeon fishing during the spring
spawning runs (Thomas and Mather 1996).
Laurel settlement patterns reflect the seasonal rounds made to obtain a wide
57
array of faunal and floral resources. Bands of 50 to 100 individuals joined larger
groups during the spring, and in the fall, families split off and made for winter camps
(Hamilton and Meyer 1994; Reid and Rajnovich 1991). Populations were more
sedentary than their Archaic ancestors, suggested by the identification of fishing
villages, an increase in artifact density at Laurel sites, the appearance of ceramics,
and the nature of domestic architecture (Thomas and Mather 1996).
At the
Ballynacree site north of Lake of the Woods in Northwestern Ontario, there are five
Laurel house structures with dates around 1240 AD, although this is rather late for
Laurel and the context and association of these dates may be off. One oval lodge
measuring three to six meters in width was excavated and this structure would have
been able to support 10 to 13 people (Reid and Rajnovich 1985).
The type-site for the Laurel phase is the Smith Mound or Grand Mound. The
Smith Mound is a multicomponent habitation and mound site located on the
southern bank of the Rainy River (Anfinson 1979). The Smith site contains remains
from both Laurel and Blackduck habitation, which are separated into four phases of
occupation by Lugenbeal (1976). The Smith site is only one mound site among
many in the Rainy River area. A habitation area located close to mound contains
faunal remains, lithic tools, and other remains indicating daily site use (Lugenbeal
1976). There is a high degree of seasonality at the site and the faunal assemblage
supports fishing during warmer times of the year. Additional artifacts that were
found include antler harpoons, cold hammered copper, and cut beaver incisors
(Anfinson 1979).
58
4.4 LATE WOODLAND (500-1750 AD) ARCHAEOLOGICAL COMPLEXES
The period from AD 500 to AD 800 represents the transformation to increased
settlements and burial mound ceremonialism. During this transitional shift, ceramics
also change shape reflecting Late Woodland varieties, and small triangular projectile
points (arrow tips) replace stemmed and notched dart forms (Gibbon 2012b).
Starting around AD 700, the Upper Mississippian Oneota tradition expanded and
Mississippian groups migrated into the Upper Great Lakes region (Egan-Bruhy
2013). Between AD 600 and AD 800, characteristics of the Late Woodland were
widespread across central and northern Minnesota and the Blackduck, Kathio, and
Clam River complexes swept across the region (Gibbon 2012b). Around AD 1050,
Middle Mississippian impacts were felt in predominantly eastern Minnesota, these
impacts included a dependence on maize agriculture (Egan-Bruhy 2013). From AD
1250 to AD 1750 Oneota and Plains village traits dispersed throughout Minnesota
and with these traits, the Psinomani Complex moved outwards from central
Minnesota, and the Blackduck Complex shifted northwards (Gibbon 2012b).
During the Late Woodland Period, house structures increase in number and
size and there appears to be an intensive trading network in the Eastern
Woodlands, as supported by exotic cherts (Ferris 1999; Fitting 1969). Some cultural
groups inhabiting the Upper Great Lakes relied on subsistence agriculture, which
had an impact on population sizes and settlement patterns (Gibbon 2012b; Hamilton
et al. 2011).
The majority of Late Woodland sites are sedentary villages, local
hunting and gathering areas, and fishing stations (Gibbon 2012b). In addition to the
cultivation of domesticates, wild plants and various mammal species formed the
59
subsistence base of Late Woodland populations (Hamilton et al. 2011). Wild rice
harvesting became more intensive in Northern Minnesota during this time and
towards the end of the Late Woodland, some sites began to represent semisedentary lifestyles supported by wild rice harvesting (Gibbon 2012b). Most of the
Late Woodland sites are predominantly habitation and mound sites (Arzigian 2008).
The number of burial mounds increase, which reflects the Arvilla burial complex
expanding across central Minnesota (Gibbon 2012b). The Arvilla burial complex
includes circular, or linear mounds, the use of yellow and red ochre, Broad Side
Notched and Prairie Side Notched projectiles, small mortuary vessels, disarticulated
and flexed secondary burials, and bone and shell ornamental grave goods in deep
pits (Gibbon 2012b). Late Woodland burials are generally intrusive into older burial
mounds and they are partially flexed in sitting or semi-sitting positions (Arzigian
2008; Hamilton et al. 2011).
A number of bone tools were recovered from Late Woodland sites, including
bear canine ornaments, spatulas, awls, fleshers, cut beaver incisors, and unilaterally
barbed harpoons.
Typical lithic artifacts include side-notched points, triangular
points, side scrapers, trapezoidal end scrapers, tubular-shaped drills, thumbnail
scrapers, oval and lunate knives (Arzigian 2008; Gibbon 2012a). Notched triangular
points correspond to the introduction of the bow and arrow around 600 AD (Blitz
1988). These triangular points probably appeared earlier; however, after 50 BC the
bow and arrow was the primary hunting technology (Hamilton et al. 2011). Copper
artifacts such as fishhooks, beads and gorges are also unique to the Late
Woodland. Ceramics appear to transition away from thick, conoidal shaped vessels
60
in the Middle Woodland to thin-walled globular and shouldered vessels with
constricted necks and outflaring rims at the start of the Late Woodland (Gibbon
2012b; Hamilton et al. 2011; Thomas 1996).
4.5 PALEODIET OF WOODLAND PEOPLES
Woodland subsistence practices changed during the Early, Middle and Late
Woodland, reflecting the adoption and cultivation of various floral species.
The
Three Sisters or maize, beans (Phaseolus sp.) and squash (Cucurbita sp.) formed
an agricultural system that was adopted at various times during the Woodland
Period (Fiedel 2001; Lovis et al. 2001; Mt. Pleasant 2006; Staller 2010).
The
introduction of the Three Sisters restructured the subsistence economies of various
groups with an increased reliance on agricultural practices. Native plants were still
utilized in addition to the Three Sisters; however, over time their significance
dwindled (Fritz 2011; Scarry and Scarry 2005; Syms et al. 2013). Answering when
and why certain domesticated crops became important to past cultural groups is
complex, because populations had differential access to a variety of plant species
through trade and seasonal exploitation over time (Boyd et al. 2014). Traditional
subsistence strategies added to the complexity because indigenous flora may have
been more important culturally and/or economically than non-native plant resources.
Plant selection may be culturally determined when certain plant resources are
chosen over others due to taste preferences and/or their importance as traditional
food sources. Economic factors also influence food selection since plants have
different nutritional values and harvest rates; therefore, the choices are based on
resource stability and security.
For example, although resources such as
61
chenopods were present during the Middle and early Late Woodland, populations
preferred to rely on wetland fauna and flora (Lovis et al. 2001). They may have
used traditional food sources and native plants because they were more accessible
and better tasting than weedy annuals.
Another example is maize, which was
present throughout the Woodland in some areas; however, it was not consumed
significantly until the Late Woodland, suggesting food selection and a continued
reliance on native species (Lovis et al. 2001).
Domesticated plant species have been adapted genetically from wild
progenitors and can become more beneficial over time with larger harvests (Staller
2010; Syms et al. 2013). Evidence of domesticate use at Middle Woodland sites
shows a shift from hunting and gathering economies to more horticulture-based
strategies, focused on supporting population increases (Arzigian 2008). During the
Late Woodland, agricultural practices intensified and some populations in the
Eastern Woodlands were fully sedentary (Arzigian 2008).
Lovis et al. (2001) conducted research at the Schultz site in Michigan to try to
understand more about the subsistence practices of Woodland peoples. This site
was occupied between 600 and 1100 AD, which Lovis et al. (2001) consider a
critical point in the transformation to agricultural subsistence. Data from the site
show intense harvesting of select plant foods, and a limited use of domesticates.
Mirroring the movement to more intensive agricultural practices is a population
increase during the Middle Woodland.
Once population growth began, groups
collected resources with larger harvests, leading to a greater reliance on
domesticated species (Lovis et al. 2001).
62
Agricultural techniques for the procurement of maize, beans, and squash are
the product of tradition and/or efficiency. Maize has been grown in the Americas at
different times and in various ways (Boyd et al. 2010; Scarry and Scarry 2005). For
example, Scarry and Scarry (2005) describe maize cultivation in Late Prehistoric
societies of the Southwest United States as varying in complexity because of
cultural differences. In some areas, farming was communal and gender specific (Mt.
Pleasant 2006). Men would assist the women with clearing the fields, planting and
harvesting. In the early spring or late winter, men burned stumps and underbrush to
clear fields. Women typically prepared fields, weeded, and did most of the planting
and harvesting. Hoes and digging sticks were used to create hills forming ranks,
and these ranks were spaced out in meter intervals. With maize, 4 to 10 grains
were placed in each hill, while beans and squash were planted between the maize
so that the vines could grow on the maize stalks (Fritz 2011; Mt. Pleasant 2006;
Scarry and Yarnell 2011; Staller 2010). Maize requires warm soil temperatures and
well-drained soils, but since northeastern soils are often wet and cold in May, mound
agriculture was used to improve soil drainage and create air flow (Mt. Pleasant
2006). When the crops sprouted, the plants could be thinned, weeded and hoed to
promote root development, and when plants ripened, they were harvested primarily
by women and prepared for storage. Plants given by each household were stored in
community granaries after a harvest.
The Cherokee populations followed this
general practice, as did the Tunica, Powhatan, Iroquois and Chickasaw, but with
slight variations (Mt. Pleasant 2006; Scarry and Scarry 2005).
63
Maize
Prior to 1492, while maize was unknown in Europe, it was cultivated
extensively in the western hemisphere for sustenance, religious, and artistic
purposes (Buckler and Stevens 2005; Staller 2010). The earliest evidence of maize
is from the Oaxaca Valley in Mexico at the Guila Naquitz Cave dating back 6250
years (Buckler and Stevens 2005; Piperno and Flannery 2001). Maize is a C4 grass
and the photosynthetic pathway used by C4 plants creates molecules with four
carbon atoms. This higher demand for Carbon results in them being less
discriminating against the heavier (and rarer)
13
C, therefore proportionately more 13C
end up in C4 plants than in C3 pants that are more common in temperate and
northern temperate climate zones. Maize is also a member of Gramineae, or the
grass family Poaceae (Buckler and Stevens 2005; Staller 2010; Vogel and Van der
Merwe 1977).
Varieties of maize can be associated with a specific culture and this is evident
when a group would consciously select phylogenetic varieties of maize (Staller
2010).
The phylogenetic varieties of maize are a cause of dispute among
researchers tracking domestication events in the genus Zea (Kato 1984). Teosinte,
the wild progenitor of maize, produces 6 to 12 kernels in two rows, while modern
maize produces 20 rows of kernels (Buckler and Stevens 2005). Currently, five
species of Zea are recognized: Zea nicaraguensis, Zea luxurians, Zea perennis, Zea
mays L., Zea diploperennis. Four subspecies divide Zea mays L.: Zea mays L. ssp.
mays, Zea mays L. ssp. mexicana, Zea mays L. ssp. huehuetenangensis, Zea mays
L. ssp. parviglumis (Buckler and Stevens 2005). Zea mays L. ssp. parviglumis is the
64
likely ancestor of maize based on simple sequence repeat markers and isozyme
data (Buckler and Stevens 2005; Doebley 1984; Matsuoka et al 2001). Current
archaeological discourse supports a single domestication event from teosinte (Z.
Mays ssp. parviglumis) with later diversification (Benz 2006; Buckler and Stevens
2005; Piperno et al. 2009; Matsuoka 2001; Staller et al. 2006). This diversification
may have occurred in North America as maize adapted quickly to the new
environment, allowing for greater productivity and climate tolerance (Figure 4.7,
Staller 2010).
Figure 4.7: Map of Late Archaic and Early Woodland sites with evidence of maize (Fern
and Lui 1995).
The single domestication event may have occurred in southern Mexico around
9,000 years ago (Staller et al. 2006; Staller 2010). A phylogenetic approach was
taken by Matsuoka (2001) to reconfirm microsatellite data, and results support a
single monophyletic lineage in the highlands of Mexico, originating from Zea mays
65
ssp. parviglumis. Afterwards, two discrete expansion events may have led to the
spread of maize in some regions of the Southwest United States and South America
(Matsuoka 2001). Alternatively, phytolith and starch data from the Central Balsas
River Valley in southwestern Mexico place the origins of domesticated maize in the
tropical rainforest 8,700 years ago, rather than in the Mexican highlands (Piperno et
al. 2009). Although evidence strongly points to a single domestication event, some
researchers still support the idea of multiple domestication occurrences (Kato Y
1984; Staller 2010).
Until future research can fully interpret the origins of
domesticated maize, it will be considered a Mexican cultigen dating to the early
7050 BC (Piperno 2009).
Several studies push back the dates of maize cultivation and consumption in
North America (Figures 4.7 & 4.8). Most researchers select either a set of sites, or
an individual site to identify patterns of early domesticated plant consumption; the
results are confined to the particular study area and do not extend throughout the
entire region. Hart et al. (2003) conducted a study on the introduction of certain
crops to the northeastern United States, with a focus on Zea mays. Three sites from
the northern Finger Lakes region (New York State) had archaeological contexts
analyzed with phytoliths, carbon isotope, along with AMS radiocarbon dating. The
results of Hart et al. (2003) from one site push back the entry of maize and squash
in the Finger Lakes region to 100 BC, in contrast to dates from the Middle Woodland
Period.
Maize cultivation began sooner than the Late Woodland Period in northwestern
Ontario, which is contrary to what the literature suggests about the timing and
66
emergence of maize in the Boreal Forest. An example is the Martin-Bird site, which
is located on Whitefish Lake in north-western Ontario, 100 kilometers southwest of
Thunder Bay. A Terminal (or Late) Woodland occupation (600-1400 AD) yielded a
copper tool-making cache and a Blackduck burial mound (Dawson 1987). In Boyd
et al. (2014), 68 carbonized food residue samples were analyzed from
archaeological sites around Whitefish Lake. Eighty-six percent of these samples
were obtained from the Martin-Bird site, including a mortuary vessel recovered from
the centre of a burial mound.
Carbonized food residue adhering to this vessel
revealed a wide range of microfossils: the site contained not only maize starch
granules and phytoliths, but also starch from domesticated bean and wild rice
phytoliths.
Starch granules with the characteristics of squash were identified,
although these can also be found in a variety of wild plant species. The mortuary
vessel and other residues analyzed at Martin-Bird reveal a broader history of maize
cultivation in the Boreal Forest than previously thought (Boyd et al. 2014).
Residues containing evidence of maize and beans have also been found in
Saskatchewan and Manitoba (Figure 4.8).
67
Figure 4.8: Location of sites with early evidence of maize. Maize was utilized sometime
within the available dates, and may not date to the beginning of site occupation (Boyd et al
2014; Hart et al. 2003; Lints 2012; Mulholland 1993)
Lints’ (2012) study examined eight sites in these two provinces and found that maize
and beans were dispersed among Plains inhabitants by 660-710 AD. Various native
plant taxa, including wild rice, were also found indicating a wide-ranging plant diet
(Lints 2012). In North Dakota, at the Plains Village Big Hidatsa site, Mulholland
(1993) identified maize phytoliths from sediment collected in test pits. Big Hidatsa
was occupied from the mid-1700’s to 1845, and both explorers and ethnographers
record corn as a staple food for site occupants (Wilson 1917). The distribution of
phytoliths from different parts of maize plants at the Hidatsa site may be the result of
husking the cobs in the fields and carrying the ears into the habitation area for
drying, threshing and winnowing. Cross-shaped phytoliths are only found in maize
husks and the leaves, which is why only small quantities were recovered from
68
village contexts. The husk and leaves of corn would not have been brought into the
village area in large amounts (Mulholland 1993).
Vogel and Van der Merwe (1977) investigated human skeletal evidence from
New York State dating to the 11th century AD. These remains revealed a period of
change in North America, in which maize was diffusing from Central American
sources (Vogel and Van der Merwe 1977). C4 plants, such as maize have more
positive C13/C12 ratios than other terrestrial plants.
Isotopic studies involve
measuring the carbon isotope ratios, which determine whether C3 or C4 plants
make up most of the diet (Katzenberg et al. 1995). Four sites are analyzed in Vogel
and Van der Merwe’s (1977) study; two are pre-horticultural and two are
horticultural. Results show a noticeable difference in the isotopic ratios between the
two groups of sites, confirming the presence and diffusion of maize into North
America. Katzenberg et al. (1995) conducted a similar study, looking at the stable
nitrogen and carbon isotope content of human skeletons in Southern Ontario from
400 to 1500 AD. Their results revealed a steady increase in δ13C values from 400
to 1500 AD and a slight decrease in δ15N values during the early Historic Period.
The decrease in δ15N could also indicate a decline in aquatic foods, perhaps with
plant protein from beans. From 600 to 1250 AD, maize consumption gradually
increased in Southern Ontario and during this time, the importance of meat protein
decreased slightly in the early Historic Period (Katzenberg et al. 1995).
Squash
Domesticated squash is one of the Three Sister crops associated with early
agriculture in North America (Staller 2010).
Pepo squash (Cucurbita pepo ssp.
69
ovifera) was present in Eastern North America 2490 BC, and is the oldest plant
manipulated for human consumption in the Eastern Woodlands (Simon 2009; Smith
2011). Cucurbita pepo, along with other squash species, may have been used for
protein rich seeds, shamanistic tools, fishing net floats, and/or as containers (Scarry
and Yarnell 2011; Simon 2009; Smith 2011). One non-edible type, Lagenaria (bottle
gourd squash) was used as a container because it is lightweight and strong.
Lagenaria was either brought to North America around 10,000 B.P by Paleoindian
peoples, or it was carried by ocean currents from Asia to America (Smith 2011).
Lagenaria did not reach Eastern North America until roughly 7,300 B.P (Scarry and
Yarnell 2011; Smith 2011). Of the edible varieties, wild Cucurbita are bitter and
small with protein rich seeds that can be cut and dried, eaten green, roasted, and
boiled (Perkl 1998). Domestication processes led to the development of larger,
thick-fleshed fruits and sweeter squashes (Perkl 1998). Three different varieties of
Cucurbita were cultivated and domesticated in various regions of North America.
The first variety, Cucurbita pepo ssp. pepo, includes marrow, zucchini and pumpkin.
The next variety is Cucurbita argyrosperma, known as cushaw squash. Cucurbita
moschata includes butternut squash and Kentucky field pumpkin. Pumpkin and
cushaw were later additions to the Eastern Woodlands, arriving shortly after 1000
AD (Scarry and Yarnell 2011). Most commonly recovered from archaeological sites
in North America, Cucurbita pepo var. ovifera includes crookneck and acorn squash,
and scallop and ornamental gourds (Smith 2011).
Cucurbita pepo cultivation began in Eastern North America throughout the
Mississippi corridor and subsequently moved to the north and west (Perkl 1998;
70
Simon 2009). It was previously thought that squash in the Upper Midwest was the
result of trade from Mesoamerican sources; however, current research shows that
the eastern United States was a center of independent domestication (Perkl 1998;
Simon 2009). Protein allozyme analysis yielded results showing that native squash
is most similar to Cucurbita pepo subspecies ovifera, either texana, or ozarkana.
These two species originate from central Texas and the Ozark Mountains; therefore,
trade must have occurred as Cucurbita pepo is outside the given range of the wild
ancestor, which raises the question of how widely distributed wild squash was and
the impact trade had on extending the range (Simon 2009).
The manipulation of squashes and gourds is more apparent during the Late
Archaic than the Woodland Period, with Cucurbita pepo rinds and seeds recovered
at sites across the Eastern Woodlands (Yarnell 1993). Wild cucurbits were present
in Maine at 5700 B.P and Central Pennsylvania at 5400 B.P (Fiedel 2001; Simon
2009; 2011). Phylogenetic characteristics are visible from the archaeological record,
suggesting plant domestication and selection for favoured traits. In particular, rinds
become thickened, seeds are enlarged, rind surfaces are lobed and there is a 2 mm
baseline thickness in comparison to the wild progenitors (Yarnell 1993).
Domesticated squash (Cucurbita pepo) was found in the form of 21 fragmented
and whole uncharred seeds at a multicomponent habitation site in King Coulee,
Southeastern Minnesota (Figure 4.3). Accelerator mass spectrometry radiocarbon
dates indicate that Cucurbita pepo was present during the Late Archaic, 580 BC,
and during the Late Woodland (780 AD). This established the earliest evidence of
domesticated plants in the Upper Midwest, and pushed Cucurbit cultivation north of
71
the known Midcontinental range (Perkl 1998). It is likely that the Cucurbita pepo
found at the King Coulee site in southeastern Minnesota was traded along the
Mississippi corridor because evidence for cultivation at the site has not been found
as of yet (Perkl 1998).
Domesticated Bean
There were five species of domesticated bean grown by North American
Indigenous populations: jack bean (Canavalia ensiformis), tepary bean (Phaseolus
acutifolius var. latifolius), common bean (P. vulgaris), small lima, or sieva (P.
lunatus), and scarlet runner bean (P. coccineus) (Fritz 2011).
Four of the five
species originated in Mexico; the fifth species (tepary bean) originated from
Mesoamerica or the Sonoran Desert of the Southwest United States (Fritz 2011).
Phaseolus spp. (a wild species of bean) was dated between 10,600 and 8,500 B.P
at Guilá Naquitz Cave in the highlands of Mexico (Piperno 2011).
Beans are
typically boiled without parching, so they are not charred as much as other cultigens;
this makes them difficult to identify at sites because they do not preserve as well
(Fritz 2011), a characteristic that affects our understanding of the antiquity of
domesticated beans (Boyd et al. 2014).
Of the Three Sister crops in Eastern North America, common beans appear to
be the latest arrival, with dates ranging between 1250 and 1300 AD in the central
Mississippi Valley. In the Northeastern Woodlands, beans were dated no earlier
than 1200 AD (Fritz 2011; Hart et al. 2002). Beans are complimentary to maize,
because they are an excellent source of protein and contain amino acids missing in
maize (Bernal et al. 2004; Boyd et al. 2014; Hart et al. 2002).
Beans also fix
72
nitrogen in the soil, which increases agricultural soil fertility (Bernal et al. 2004; Fritz
2011; Smith 1992).
Archaeological evidence also suggests that beans were the last of the Three
Sister crops to arrive in the Boreal Forest, while maize and squash agriculture took
place much earlier. One reason for this may be that traditional trade networks and
cultural subsistence practices caused the exclusion of domesticated beans until the
Late Woodland (Boyd et al. 2014; Hart and Scarry 1999; Fritz 2011). Additionally,
beans are susceptible to spring frost and difficult to cultivate in shorter growing
seasons (Boyd et al. 2014; Mt. Pleasant 2006).
Common bean (P. vulgaris) remains were AMS dated to 1070 +/- 60 AD in the
Northeastern Woodlands, and eight other sites contained pre-1300 AD dates. The
date of 1070 +/- 60 AD came from wood charcoal in a feature nearby a storage pit
full of charred maize, beans, and squash at the Roundtop site in New York (Hart and
Scarry 1999). The eight sites containing pre-AD 1300 dated beans are located in
the Susquehanna River Valley, New York, with the exception of one site in
Pennsylvania. Hart and Scarry (1999) suggest that the early date at the Roundtop
site of 1070 +/-60 AD pushes back the date for intense cultivation of the Three Sister
crops in Northeastern North America (Figure 4.9).
73
Figure 4.9: Location of sites with domesticated bean (Hart and Scarry 1999; Lints 2012).
Wild Rice
Wild rice (Zizania sp.) belongs to the family Poaceae, and the tribe Oryzoidae
and is an aquatic semi-annual grass. There are four different species of wild rice: Z.
aquatica L., Z. palustris L. (Northern USA and Canada), Z. latifolia Turcz. (Asia), and
Z. texana A. S. Hitchc. (Texas) (Archibold et al. 1985; Bunzel et al. 2002). Wild rice
is gynaecandrous, meaning that the anthers mature later than the female flowers,
ensuring cross-fertilization. Germination begins with temperatures between 4º and
6º degrees Celsius, and growth begins in the spring with a life-cycle of 110 days
(Archibold et al. 1985). Z. aquatica was originally thought to be the only wild rice
consumed by Northern populations; however, Z. palustris is also a northerly species
so both species may have been utilized at some northerly sites. The chaff portion of
wild rice plants produce phytoliths and after the rice is processed, enough chaff may
remain with the seeds to account for wild rice phytoliths in pottery residue
74
(Mulholland 1993).
Paleovegetation changes throughout the Holocene include the introduction and
dispersal of wild rice (Z. sp.) into Northern Minnesota. Pollen cores from this area
suggest that wild rice populations were increasing around 1000 BC in some lakes
(Yost et al. 2013). Human populations for over 9,000 years have occupied the lakes
and surrounding area of Mille Lacs State Park, North-central Minnesota. Yost et al.
(2013) identified the past presence of wild rice (Zizania spp L.) in this area by
analyzing phytolith content of lacustrine sediment samples. Birks (1976) confirmed
the presence of wild rice in central Minnesota with Zizania seeds found in detrital
copropel. Birks’ (1976) study showed that Z.aquatica spread in Rice Lake because
the lake became more shallow through sediment accumulation, resulting in the
replacement of Najas flexilis by wild rice. This trend may be seen across other
central Minnesota lakes around 10,500 years ago (Birks 1976).
Wild rice (Z. palustris L.) phytolith evidence was also found on Brainerd
ceramics at the Call Lake I site, dating between 700 and 800 BC, and at the OgemaGeshik site, dating to 170 AD (Hohman-Caine and Goltz 1995; Thompson et al.
1994). Only the rondel bases were used to determine the presence of wild rice
phytoliths in Thompson et al. (1994), so the residues should be re-analyzed for
diagnostic indicators of wild rice. Charred wild rice grains were also found in St.
Louis County, Minnesota, associated with Laurel pottery (Thompson et al. 1994).
Archaeological data from Middle and Late Woodland material supports the
interpretation that wild rice was processed and consumed throughout central
Canada (Boyd et al. 2014). Whitefish Lake in Northwestern Ontario was cored by
75
Boyd et al. (2013), yielding evidence of wild rice (Zizania sp.) phytoliths. Wild rice
colonized Whitefish Lake by 6,100 cal BP, likely in response to rising lake levels
(Boyd et al. 2014). The analysis of 176 sites across the Boreal Forest and northeastern Plains revealed that some Middle to Late Woodland populations consumed
a diet consisting of wild rice, maize, beans, and squash in addition to native plants.
Microfossils recovered from sites on the Canadian Plains contained more bean
evidence than wild rice, while microfossils extracted from the Boreal Forest samples
had less evidence of bean. In the Boreal Forest, wild rice is known to have played
an extensive dietary role in the subsistence economies of prehistoric and historic
populations. The wild rice microfossils recovered from the Canadian Plains suggests
trade with Boreal populations, and/or a wider distribution of the plant (Boyd et al.
2014; Lints 2012). The lower quantities of bean recovered from sites in the Boreal
Forest (north-east of the Plains) can be explained by a complementary reliance on
maize and wild rice. In the north-eastern Plains, wild rice may have been of lesser
importance than bean because wild rice is not easily available, and beans and wild
rice contain similar nutrients (Boyd et al 2014). Wild rice becomes less
complementary to maize, as opposed to bean, after it is cooked because the protein
degrades; therefore, there may have been a greater emphasis on bean throughout
time (Boyd et al. 2014; Hart et al. 2007)
Lake Ogechie in Central Minnesota was important to Ojibwe peoples who
harvested wild rice from the shallow lake before European contact (Yost et al. 2013).
Wild rice is typically harvested with two people per canoe, traversing shallow wild
rice stands in lakes (Yost et al. 2013). One individual stands at the stern while the
76
other stands amidship and, using a cedar knocking stick to bend the wild rice stalks
over the canoe bow, with a second stick to thrash the rip kernals from the stalk and
into the canoe. The rice is later dried over fires and trampled in a thrashing pit to
remove the husks (Wilcox 2007).
Native Plant Species
Prior to the arrival of domesticates in the Great Lakes, many different native
plant species were used for medicinal, technological, ritual and subsistence
purposes (Marles et al. 2000; Simon 2009; Crawford 2011). Marles et al. (2000)
interviewed elders with ethnobotanical knowledge from around the Boreal Forest in
Northwestern Ontario.
Each elder supplied different accounts of plant use,
suggesting that there are local variations in how individuals and bands select,
prepare and process plants (Crawford 2011; Marles et al. 2000).
A wide variety of native plants would have been available to Archaic and
Woodland Period populations in Northern Minnesota, and approximately 2,000
native plant species were available across the Northeast United States (Crawford
2011). Some of the different plant species available would have included wild fruits
and vegetables.
Different types of common wild fruit species include various
cranberries Viburnum trilobum (high-bush cranberries), Vaccinium vitis-idaea
(mountain cranberries) and Vaccinium oxycoccos (bog cranberries).
Additional
types of berries were consumed: Fragaria * ananassa (strawberry), Prunus
pensylvanica (pin cherry), Amelanchier alnifolia (saskatoons), Rubus chamaemorus
(cloudberry), Ribes uva-crispa (goose-berry), Shepherdia canadensis (buffalo
berry), Empetrum nigrum (crowberry), and Cyanococcus (blueberry).
Other
77
varieties of wild fruit species are Ribes (currant), Corylus (hazelnut), Rosa rugosa
(rosehip). Some traditional vegetables include tree species such as the inner bark
of Pinaceae (Pine), Betula (Birch) and Populus (Aspen).
Additional traditional
vegetables include Typha latfolia (cattail), Poales (reeds), Chamerion angustifolium
(fireweed) and Allium (wild onion) (Crawford 2011; Smith 2011).
The wide variety of plant foods added nutrients, vitamins, carbohydrates and
phytochemicals to traditional diets (Scarry 2003; Marles et al. 2000; Smith 2011),
but throughout the Archaic and Woodland Periods, different wild plant species
became available through trade and/or access to new resources. The following
describes the wide variety of native plants recovered from Early, Middle and Late
Archaic contexts. Recovering evidence of plant use at Archaic sites is difficult in
Northern Minnesota because of site age and poor preservation. Often, Archaic plant
assemblages yield only a few wood fragments or pieces of nutshell. Early Archaic
sites from the Eastern Woodlands had better preservation and contained walnut
shell, grape seed, hickory shell, chenopod testa, acorn shell, and oak, maple, elm
and pine wood assemblage (Johannessen 1993; Simon 2009). Middle Archaic sites
in the Eastern Woodlands contained macrobotanical remains similar to those found
in the Early Archaic sites, including nutshell, hickory, and acorn (Scarry 2003). A
wide variety of fleshy-fruit seeds were recovered, such as blackberry, raspberry,
sumac, grape, and persimmon.
Tubers, bulbs and roots were also important,
although less evident in the archaeological record. Wild bean (Strophostyles) and
giant ragweed (Ambrosia trifida) are commonly found in Middle Archaic sites from
more southerly areas of the Eastern Woodlands (Simon 2009). Late Archaic Period
78
sites in the Northeast United States have better preservation, yielding more
macrobotanical remains and information on Archaic plant use (Crawford 2011).
Seeds became more common, along with weedy plant species and thick-shelled
hickory, although nutshell (black walnut and acorn) and wood still dominate plant
assemblages (Crawford 2011). Hogpeanut tubers, American lotus, wetland sedge,
rushes and wetland plants such as wild rice were also consumed (Simon 2009).
Weedy annuals were exploited, and human movement dispersed the range of plants
(Yarnell 1993).
Squash (Cucurbita pepo), sunflower (Helianthus annuus), chenopodium
(Chenopodium
berlandieri),
giant
ragweed
(Ambrosia
trifida),
knotweed
(polygonum), maygrass (Phalaris caroliniana), little barley (Hordeum pusillum) and
sumpweed (Iva annua) are weedy annual species included in the Eastern
Agricultural Complex, which was spread from the southeastern United States to the
Eastern Woodlands (Scarry and Yarnell 2011; Simon 2009; Syms et al. 2013).
These native plants made up a large part of Archaic and Woodland diets up until the
introduction of the Three Sister crops, when they became incorporated into
subsistence economies over the span of the Woodland Period (Scarry and Yarnell
2011). The native chenopod, sunflower, sumpweed or marshelder and maygrass
were the most widely consumed weedy annuals (Scarry and Yarnell 2011; Yarnell
1993).
Two other members of the Eastern Agricultural Complex, which arrived later
than previously mentioned plant species, are little barley (Hordeum pusillum) and
erect knotweed (Polygonum erectum) (Crawford 2011; Scarry and Yarnell 2011;
79
Simon 2009). Giant ragweed (Ambrosia trifida) and barnyard grass (Echinochloa
muricata) are also common weedy annuals, although the ranges of these plants
were restricted. Chenopodium has the largest range out of the weedy annuals and
is commonly found along major rivers. Chenopod seeds have convex margins and
thick coats. Sumpweed is the second most common weedy annual, and a fully
domesticated subspecies. Sunflower seeds have a larger range than sumpweed
and, based on AMS dates from the Hayes site in Texas, were domesticated by 2200
BC (Simon 2009).
Macrobotanicals found in Brainerd occupation levels at the
Roosevelt Lake site yielded possible chenopod seeds, seeds from strawberry,
raspberry, and nut shells including hazelnut and acorn (Arzigian 2008). Thompson
et al. (1994) also detected chenopod and amaranth starch granules.
Bone, shell and stone hoes have been recovered at a number of sites across
the Eastern Woodlands and Canadian Plains (Boyd et al. 2006; Scarry and Yarnell
2011). Hoes would have been used to harvest agricultural crops efficiently, and
although chert sickle blades can be used to cut larger bundles of stems, there is not
much archaeological evidence to support the use of sickle blades in North America.
Squash was most likely handpicked and plants with small heads may have been
either beaten or hand stripped allowing the seeds to fall onto cloth or in a basket.
Attached pericarps or glumes on oily seeds or smaller grains were removed by
parching, threshing and winnowing the plants (Scarry and Yarnell 2011).
The Woodland Period in Northern Minnesota is characterized by changes in
settlement patterns, technology and subsistence practices.
Settlements tend to
become larger towards to the Late Woodland and reflect growing population sizes.
80
Ceramic technology emerges during the Early Woodland and over the course of the
Woodland Period there are a number of stylistic shifts, including the transition from
conoidal to globular vessels. Lithic technology changes to reflect hunting strategies
and triangular points appear around 600 AD corresponding with the use of the bow
and arrow. Subsistence practices changed over the Woodland Period, when
populations incorporated non-native cultigens such as the Three Sisters, into diets
dependent on wild plants and various mammals.
During the Early to Middle
Woodland Period, populations maintained a broad-based subsistence strategy that
focused on wild rice gathering in the fall and the fishing during the spawning run in
the spring. Wild plants and mammals supplied additional nutrients and
carbohydrates to Woodland groups inhabiting Northern Minnesota. The introduction
of the Three Sisters at different times during the Woodland Period marked an
increased reliance on subsistence agriculture, which would have resulted in a
localized subsistence strategy.
Although native plants and mammals were still
important dietary components, the significance of these species dwindled with an
increased reliance on the Three Sisters agricultural system.
The timing and extent of the Three Sisters during the Woodland Period leave
difficult questions to answer in Northern Minnesota because of poor site
preservation and thin, disturbed, stratigraphy.
Macrobotanical remains such as
plant seeds do not survive well in the acidic soils; therefore, alternative methods are
needed for inquiry into the floral subsistence of the Woodland Period. Food residue
analysis and the identification of phytoliths, starch granules, and pollen grains can
guide interpretations and answer questions about when and where specific plant
81
species may have been utilized.
5.0 PHYTOLITH AND STARCH ANALYSIS IN ARCHAEOLOGY
5.1 INTRODUCTION
Food residue analysis is a collection of methods that can be used to interpret
the paleodiet and subsistence practices of ancient peoples.
Plant microfossil
analysis is one such method that involves the extraction and microscopic analysis of
phytoliths, starch granules, and other remains, from archaeological samples. For
the past 30 years, food residue analysis has taken archaeologists from general
assumptions about food consumption to facts based on evidence found on
microscopic slides (Boyd et al. 2014; Pearsall 1982; Powers 1992). This chapter
describes the structure of phytoliths and starch granules, corresponding literature,
and reviews current archaeological analyses and applications.
5.2 PHYTOLITHS
Phytoliths develop when plants deposit solid silica within an extracellular or
intracellular location after silica is absorbed from groundwater in a soluble state
(Piperno 1988; 2006). When the plant dies, the silica microfossils produced by this
process are incorporated into sediments and/or soils. A number of factors such as
soil composition, water content, taxonomic affinity of the plant, and climatic
environment of growth determine how phytoliths develop in plants. Soluble silica
(H4SiO4) is absorbed by plant roots, and is carried upwards to aerial organs by
water-conducting tissue in the transpiration stream. Some silica is manifested in the
growing plant as silicon dioxide (SiO2), filling in the cell walls, lumina, and
intercellular spaces (Piperno 2006). Phytoliths are composed of mostly amorphous
82
(noncrystalline) silicon dioxide (SiO2) with 4 to 9% water composition. They contain
tiny amounts of impurities such as Fe, Mn, P, Cu, organic C, Al, N and Mg. These
elements minerals reside within the cytoplasm of living cells and when the cell fills
with silica, they remain enclosed within the phytolith. Biogenic silica from plants has
a specific gravity of 1.5 to 2.3 and is optically isotropic, ranging in refractive index
from 1.41 to 1.47. Colours range from light brown to opaque to colourless under
transmitted light. Carbon is sometimes visible inside phytoliths from trapped cellular
contents and dark spherical cavities, and when no carbon is visible, phytoliths may
contain cellular cytoplasmic matter.
Organic and burnt phytoliths have a lower
specific gravity than other phytoliths and an altered refractive index (Piperno 2006).
Phytoliths were first analyzed and recognized in Europe, where early accounts
originated from Germany and England (Piperno 2006; Powers 1992). Generally,
European applications of phytolith analysis changed from simple plant identification
to paleoeconomic and paleoenvironmental reconstruction (Lu et al. 2006; Piperno
1988). One of the first accounts of phytolith observation was made by Christian
Gottfried Ehrenberg, who wrote about microscopic flora, which he termed
“phytolitharia” or silicified plant material. Even earlier than this, Struve submitted the
first publication of phytolith research to the University of Berlin in 1835. From the
1960s to the 1980s, publications such as those in the Annals of Botany dealt
specifically with cell silification and drew more attention to the topic. Microfossil
research during the late nineties focused on various grass species, phytoliths
originating from different plant parts, and morphological differences (Powers 1992).
These examples of botanical research laid the groundwork for future archaeologists
83
interested in subsistence strategies and food economies.
Intense interest in phytolith research began around 1960, with one of the first
projects taking place in Peru (Pearsall 1982). Since the 1960s, the identification,
classification, and processing of phytoliths has developed allowing for more
reliability and validity in plant identification (Rover 1971; Twiss et al.1969).
Currently, researchers are focused on plant identification, the migration/diffusion of
plant species, reconstructions of past environments, and limitations.
Limitations
associated with phytolith analysis include assemblage bias, different counts used by
various labs, species identification, classification schemes and the understanding of
taphonomic processes (Pearsall 1982; Rovner 1971; Shillito 2011).
There are two debated mechanisms explaining how plants absorb soluble
silica, and Piperno (2006) maintained that both mechanisms are in fact correct.
These mechanisms involve the active transport of monosilic acid by metabolic
processes, and the nonselective flow of monosilic acid from groundwater with other
elements in the transpiration stream. With the first described pathway the plant
expels energy metabolically during silica absorption, and with the second
mechanism, plants do not expel energy. Once the monosilic acid enters the plants’
vascular system, a chemical/biological process begins the transformation into silicon
dioxide (SiO2) (Piperno 2006).
Over the past three decades, research has established that phytoliths are
identifiable taxonomically and are present in pollen and spore producing plants
(Rover 1971). Similar shapes can also be found from plants in the same genera,
species and family. Two factors determine how phytoliths are shaped: the first is the
84
kind of cell accumulating silica and where it is located within the plant; the second
refers to incomplete silification of a part of the lumen resulting in nonconforming cell
shapes. Classification systems and phytolith terminology were not well standardized
in the past; however, current research is addressing these limitations. Phytoliths
can be identified by specific characteristics including dimensions, size, shape,
orientation of cells, their relation to surrounding structures, ornamentation of walls
and thickness of cell walls (Brown 1984; Rover 1971). The Society of Phytolith
Researchers established a standard code for phytolith identification, which includes
using anatomical origins, descriptors for surface and shape attributes (Piperno
2006). To summarize Piperno (2006), the only way researchers will be able to
confidently identify phytoliths is to develop rigid keys and have a grasp on the
regional flora.
Different types of phytoliths have been identified based on certain
morphological characteristics. C3 and C4 grass phytoliths comprise a number of
grass species, which inhabit different climatic regions and which take two forms of
phytoliths. C3 species are concentrated in regions of high latitudes and elevations.
Pooid phytoliths, the most common type of C3 grass phytoliths, are crescent,
elliptical, circular or oblong. C4 species include Chloridoid and panicoid phytoliths,
Chloridoid phytoliths are saddle-shaped and present in arid to semi-arid warm
regions with low soil moisture, while panicoid phytoliths comprise dumbbells and
crosses that are found in warm, tropical to sub-tropical environments (Twiss 1992).
Some of the major phytolith types found in the grass family (Poaceae) are short cell
phytoliths or silica bodies, which can be divided into three subfamilies native to the
85
Great
Plains:
namely,
bilobates/crosses
(subfamily:
Panicoideae),
saddles
(Chloridoideae), and circular/oval/rectangular forms (Pooideae) (Bozarth 1992;
Piperno 2006; Twiss et al. 1969). Ehrhartoideae forms are located in forested and
aquatic regions, and Panicoideae forms can be recovered from tall tropic grasses
and prairie grasses in the United States. Bambusoideae grasses reside in forested
areas
containing
Centothecoideae,
Danthonioideae,
Arundinoideae (including Phragmites and Arundo).
Aristidoideae
and
Three subfamilies found in
forest understories include Pharoideae, Anomochlooideae and Puelioideae (Piperno
2006).
Although it is possible to discern phytolith subfamilies, some overlap exists
between taxonomic levels, caused when plants produce morphologically different
phytoliths (Brown 1984; Piperno 2006).
Glumes, lemmas, and palaeas contain
phytoliths different from ones found in the sheaths, leaves, floral bracts and culms of
plants. Asteriform and dendriform are two shapes found specifically in the bracts of
plants. Gymnosperms contain less silica and phytoliths from needles such as Pinus
spp., Picea, Tsuga and Pseudotsuga. Basal Angiosperms produce phytoliths of
various sizes, and include species such as, Amborellaceae and water lilies.
Monocotyledons diverged from basal angiosperms and are mostly aquatic (Piperno
2006).
Seeds and fruits from nongrass monocotyledons, eudicots and basal
angiosperms produce many phytoliths identifiable at the genus level.
Certain
families produce distinctive seed and fruit phytoliths such as Cucurbitaceae,
Musaceae, Ulmaceae, Euphorbiaceae, Cyperaceae, Burseraceae, Moraceae and
86
Marantaceae (Bozarth 1987; 1992; Piperno 2006). Phytoliths can also be found in
wood, specifically in xylem vessels and/or parenchyma cells. These form three
different phytoliths, including irregularly shaped particles covered with tiny
protuberances, aggregate grains, and spherical grains with small projections
(Protium panamense) (Bozarth 1992; Piperno 2006).
Piperno (2006) describes three less common phytoliths as Mesophyll
phytoliths, vascular tissue and epidermal phytoliths. Mesophyll phytoliths are found
in the ground tissue of leaves under the epidermis; the two types of tissue silicified
include spongy Mesophyll and palisade Mesophyll. Vascular tissue is composed of
the phloem and xylem, the area of the plant dealing with food storage and water
conduction. Epidermal phytoliths form a continuous layer of plant surfaces, and are
commonly silicified and are either sinuate or wavy.
Epidermal seed and fruit
phytoliths are planar with sinuous outlines of 4 to 8 sides. Uncommon types of
phytoliths are less useful because they are not taxonomically relevant (Bozarth
1987; Piperno 2006).
Multiplicity and redundancy in the context of phytolith production in plants can
cause misinterpretations of climatic data. Twiss (1992) provides an example of how
Aristida and Zea produce both Panicoid and Pooid phytoliths. Although one type of
phytolith may be abundant in one subfamily, it can occur in taxa throughout different
subfamilies. Twiss (1992) emphasizes that phytoliths are an indicator of the grass
kind and in order to identify specific taxa, different parts of preserved plants are
needed. C3 and C4 plants can grow in the same environment and this is common
in areas where temperatures fluctuate with a cool, moist season and a warm, dry
87
season (Twiss 1992).
5.3 STARCH GRANULES
Starch granules have been recovered from a variety of well to poorly preserved
contexts with some associated deposits two million years old.
Starch can be
recovered from archaeological food residues, dental calculus, charred tubers,
preserved bread, stone tools, coprolites and soils, among other contexts. In Roman
times, starch was first recorded; however, it is only within the past 20 years that it
has been studied as a form of dietary evidence (Haslam 2006; Loy 1994; Torrence
and Barton 2006). Plant domestication, mobility patterns, diet, artifact function, and
vegetation histories are some of the different areas of starch research (Torrence and
Barton 2006).
Different parts of plants produce starch as a form of energy storage. Starch
production begins when photosynthesis converts sunlight into energy within the
chloroplasts of plants, prompting a few reactions that split water into hydrogen and
oxygen. The initial reaction recombines free hydrogen with absorbed carbon dioxide
forming glucose. Glucose is the base on which complex carbohydrates (cellulose
and starch), proteins and fats form. Some glucose is transported from chloroplasts
to specialized organs of plants, where it is converted within amyloplasts to storage
or transitory starch. Starch begins to form at the hilum where layers are laid down
on top of eachother each day with normal growing conditions (Gott et al. 2006;
Haslam 2004). When the plant requires energy, the starch is converted to sugar
and is transferred to the necessary areas. Temporary, transient or transitory starch
is formed in the chloroplast when photosynthesis rates are high (Haslam 2004).
88
Transient starch is converted to sugar during the night and moved to necessary
areas of the plant or transformed into storage starch in the amyloplasts.
Donald Ugent pioneered systematic starch analysis in the 1980s when he
recognized various starch types with diagnostic characteristics.
He also
experimented with chromatography, staining, spectrophotometry and microscopy.
Unlike phytolith analysis, starch has been thoroughly studied by botanists (Torrence
and Barton 2006). Critical studies by Loy et al. (1992), Liu et al. (2013), Piperno et
al. (2004) and Pearsall et al. (2004) focused on individual granules derived from
stone tool residues.
Starch consists of alternating amorphous and crystalline shells of hard and soft
material about 120 to 400 nm thick. Towards the granule surface, some starches
have higher concentrations of amylase. Amylose chain and amylopectin clusters
protrude from the center of a granule as the foundation for the next growth layer
(Gott et al. 2006). Lipids, proteins, and phosphorous are sometimes found in trace
amounts either in the interior or on the surface of starch and these components
impact starch characteristics (Gott et al. 2006).
Glucose units of amylose and
amylopectin molecules are essentially a ring of six carbon atoms numbered one to
six. With amylose, the first and fourth carbon atoms of each glucose unit are linked
together. In amylopectin, while most of the glucose units are joined by the first and
fourth carbon atoms, four to six percent of the glucose units have a link with the
sixth carbon atom and form a separate branch point. Amylose is composed of 1,500
glucose units, while Amylopectin is composed of 600,000 glucose units.
Gelatinization and the starch granules’ reaction to staining are dependent on the
89
ratio of amylopectin to amylose, which is controlled by environmental and genetic
factors. Amylose content of 20 to 30 percent can be found in most economic plants
(Gott et al. 2006).
Starch grain morphology is largely dependent on the plant’s genetic
composition, while modifications can be made to grain shape and size by external
and internal factors (Gott et al. 2006). Similar to phytoliths, different parts of a plant
will produce morphologically distinct starch granules.
Starch granules have a
central hilum, which may contain non-starch amyloplast material and fissures
radiating from the hilum (Henry et al. 2009). Large starch granules have visible
lamellae or growth layers, and open hila or vacuoles are evident with some plant
species. Equatorial grooves are present in barley, wheat and rye, while pores cover
starch surfaces in the subfamily Panicoideae (e.g., millet, corn, and sorghum) (Gott
et al 2006; Haslam 2004, 2006).
The shape classes that assist with identifying starch granules are spherical,
oval, rounded, kidney-shaped, irregular, disc, elongated and polyhedral.
multiple shapes are produced within a single plant.
Often,
For example, the Triticeae
(wheat, barley and rye) exhibit two forms of granules: small spherical ones or large
disc-shaped lenticular ones.
Starch sizes can be anywhere from one to 100
microns, with some rare exceptions. High water concentrations in the granules can
account for an increase in size and the location from which the starch originates in
the plant. Starch originating from the center of the plant will be elongated, while
starch found near the periphery of the stem will be small and round. Granule age
also impacts size: older granules are typically larger than younger granules of the
90
same species from the same storage site. Depending on the environment, a more
stressed organism with lower nutritional status will yield fewer and smaller starch
granules because of limited carbohydrate intake (Gott et al. 2006).
Starch granules have been classified into three categories based on the way
they form in the amyloplast: simple, compound and semi-compound (Gott et al.
2006).
Simple granules have only one component in the amyloplast, while
compound granules contain subgranules or granula that form within a single
amyloplast and have a polarizing cross.
Sweet potato, rice, oats, other edible
grasses, cassava and quinoa are all examples of plants with compound granules
(Figure 5.1).
Semi-compound granules start as compound, but the separate
subgranules fuse together forming a layer of amorphous starch. Semi-compound
granules from Amaranthus retroflexus (American pigweed) and Scilla ovatfolia
(Hyacinthaceae) bulbs exhibit two or more hila, and one exterior surface (Gott et al.
2006).
91
A
B
C
Figure 5.1: (A) Cassava starch, an example of a compound starch type (Rocha et al.
2010). (B) Mouldy potato starch granules exhibiting staining. (C) Gelatinized azuki
bean cv. Erimo starch granules (Hsieh et al. 2000).
The shoots, stems, roots, leaves, fruit and even pollen grains of plants contain
starch. Reserve starch is concentrated in storage organs such as seeds, fruits,
tubers and roots; therefore, the identification of reserve starch provides evidence of
what plant parts were exploited in the past. Nonedible plant parts also contain minor
amounts of starch and these may appear on tools for example, in fibres used to
make rope. Storage starch is found in leaf tissues, and this in addition to transient
starch could isolate places where leaves are used for bedding and/or roofs at
archaeological sites (Gott et al. 2006).
92
5.3.1 STAINING
Depending on the type of stain, staining is used to change starch color and
help identify damaged starch granules (Figure 5.1, Gott et al. 2006; Haslam 2004;
Lamb and Loy 2005).
The amylopectin-amylase ratio causes a colour change
reaction when iodine staining is applied to transform starch to red-purple colour.
Gott et al. (2006) found that iodine-staining starch is a useful identification tool, with
plants such as potato with a high amount of amylose turn a purple color, while
starch from maize with high amount of amylopectin will stain red.
Lints (2012)
tested both Trypan Blue and Congo Red stains, discovering that both agents were
equally effective identifiers of damaged starch grains.
5.3.2 BIREFRINGENCE
Birefringence refers to measured differences in the refractive index of optically
isotropic solids, which is apparent when viewed under cross-polarized light
(Olympus Microscopy Resource Center, 2012).
Starch appears white and
illuminated contrasting the surrounding darker microfossils, and this contrast is
reduced by the starch extraction process (Barton and Fullagar 2006). Birefringence
occurs because the semicrystalline nature of starch and a highly ordered molecular
structure causes polarised light to travel at different velocities through the granules.
Chemical processes can cause the lamellae to also appear less marked and the
hilum may have centrally developed cracks. These cracks have to be differentiated
from fissures radiating from the hilum, because this can lead to misidentification
(Gott et al. 2006). Birefringence also allows for the detection of an extinction cross
or Maltese cross in the starch granule centre. The arms of the cross rotate when
93
the polariser is turned, and arm length and angle vary between plant species (Gott
et al. 2006). Birefringence allows for precise identification due to the contrasting
pattern and illuminated aspects of the starch (Barton and Fullagar 2006).
5.3.3 GELATINIZATION
Damage to starch granules during cooking can result in gelatinization. Starch
is insoluble, but permeable to water and at low temperatures with water contact,
starch begins to swell in a reversible process. With the application of heat, swelling
remains reversible until the point of gelatinization when the hydrogen bonds holding
together the linear molecular chains of amylopectin and amylose are disrupted.
Salts, alkali and acids can also cause gelatinization at room temperature. Starches
with greater amounts of amylose generally have a raised temperature of
gelatinization, meaning that more heat is needed to gelatinize the granules.
Gelatinization causes the loss of native structure and morphology, and once
granules are gelatinized fully they become extremely difficult to identify
microscopically and classify taxonomically (Figure 5.1, Haslam 2004; Zarrillo et al.
2008). Baked starch granules will have sharp crosses, strong birefringence, and
intact forms compared to boiled and/or steamed starch granules (Crowther 2012).
How starch granules gelatinize and change shape depends on the species of
interest; for example, cereal granules will swell but not burst while those of potatoes
will swell to a large size and then burst (Gott et al. 2006). Although it is difficult to
identify gelatinized starch, finding cooked starch is evidence of plant preparation
processes.
94
5.4 LIMITATIONS
Phytolith and starch analyses are recent applications to the archaeological
discipline and although many studies have shown the value of applying these
methods, some limitations have surfaced over the years.
This can include:
deterioration of grains due to mechanical wear, taphonomic processes, poor
preservation, contamination of archaeological samples by modern starch, and
unidentified species.
Mechanical Wear
Microfossil damage can result in misidentification because changes can alter
the diagnostic features of microfossils, either rendering them completely
unidentifiable, or altering their appearance to resemble starch from another plant
species (Zarrillo et al. 2008).
The interaction with enzymatic saliva caused by
chewing has a significant effect on starch breakdown. Some forms of starch, which
are found in coprolites, are able to avoid being broken down by amylase in stomach
acid and the mouth (Torrence 2006). Milling, pounding or the grinding of plant
material with groundstones can also cause mechanical wear. The harder the plant
material is processed, the more likely microfossil damage will occur (Babot 2003).
Under cross-polarized light, this wear takes the shape of tears, a warped extinction
cross and/or a loss of birefringence. Under both plain-polarized light and crosspolarized light, the starch granules may appear collapsed, incomplete, truncated,
burst, and/or joined to surrounding plant material (Babot 2003). When maize starch
exhibits signs of grinding, the extinction cross of 90º will appear to have shifted, and
the microfossil may be covered in fissures or holes (Figure 7.10). Damage can be
95
identified by applying Typan Blue to samples, which penetrates small cracks and
darkens areas of damage (Barton 2007; Haslam 2004; Lints 2012).
Taphonomic Processes
Taphonomy, in this context, refers to the chemical and physical processes
influencing starch and phytolith preservation, degradation and postdepositional
movement (Haslam 2006). The different aspects of taphonomy impacting phytolith
and starch analysis include effects of bacteria, fungi, enzymes, soil movement and
mechanical wear.
Starch is able to survive in a variety of environments and has been recovered
from ceramic sherds, stone tools, dental calculus, sediment and soil (Boyd et al.
2014; Hardy et al. 2009). Although this may be so, many factors can implement the
breakdown of starch. Fungal and microbial enzymes are the main sources of starch
degradation (Haslam 2004; Torrence 2006).
Fungi and bacteria affect starch because like other soil microorganisms they
breakdown soil components over time. Fungi can grow hyphae, which extends
microhabitats where enzymes are secreted to decompose organic matter. Bacteria
growing in clusters on surfaces can occupy a few cubic millimeters of soil and are
dependent on root growth, tillage for movement and rainfall (Haslam 2004). Haslam
(2004) notes that starch rapidly degrades in soil, due to fungal, microbial, and
enzyme activity close to the surface of soils. Haslam (2004) describes enzymes as
necessary components to the fungal and bacterial decomposition of starch, and the
biological catalysts responsible for lowering the activation energy required for
chemical reactions. Amylases are the enzymes related to starch breakdown, and
96
these enzymes are derived from microbial or animal cells, bacteria, fungal spores,
enzymes enclosed in dead cells and active plant matter. Depending on the type of
soil, different enzymes will be present acting to breakdown starch (Haslam 2004).
The soil pH, oxygen content, temperature, and physical dispersal are the main
physical and chemical processes behind the degradation and movement of starch in
soils (Haslam 2004; Torrence 2006). Preferential decay of starch in sediments, as
opposed to protected surfaces and niches on archaeological material may also be
another factor in the rapid decay of starch in soil (Haslam 2004). Only recently have
studies attempted to discern the degree of starch movement in soil over time. The
recent popularity of microfossil analysis has its detractors with researchers such as
Haslam (2004), claiming issues of possible contamination due to starch movement
in soil.
Similar to starch, phytoliths can be deposited into soils and sediments in a
number of ways.
Piperno (2006) first addresses plant decomposition, which
deposits phytoliths in the upper soil horizons (A horizon).
Fire, wind, human
transport, animal droppings and various other scenarios also deposit phytoliths in
soil. Conducting residue analysis on the surrounding sediment within the matrix is
one way to avoid contamination and test for movement in soil (Piperno 2006).
Jenkins (2009) describes aeolian and alluvial forces, which act to pit, and/or
corrode single-celled forms and breakdown multi-celled forms. Phytoliths post-burial
are susceptible to soil conditions, and do not survive as well in alkaline conditions or
when pH is above nine. It has also been shown that illuviation or the downward
movement of phytoliths in soil is possible from the A to B horizon depending on soil
97
characteristics (Jenkins 2009).
Preservation
The analysis of microfossils is valuable because phytoliths and starch granules
can survive in areas of poor preservation such as the neotropics, or boreal forest
(Boyd and Surette 2010; Zarrillo et al. 2008). Despite the exceptional preservation
of these microfossils, the phytolith and starch granules are preserved in different
ways that are not fully understood as of yet. Starch granules can collect on artifacts
and microorganisms, while the build-up of sediment or plaque on residue preserves
and protects the microfossils from decay (Torrence 2006). Plaque, for example,
traps starch before it can be dissolved by amylase in dental calculus (Hardy et al.
2009). In addition, starch has a microcrystalline structure that serves to protect it
from degradative elements (Torrence 2006).
Phytoliths survive in environments of poor preseveration because they are
inorganic and very durable. Phytoliths have a tendency to preserve better in dry,
terrestrial, arid soils, whereas pollen grains and other organic microfossils do not
(Boyd 2005). Piperno (1988) describes phytoliths as robust and capable of being
preserved in adverse conditions, whereas other microremains may degrade from
dissolving or burning. Phytoliths are composed of calcium oxalate or silica so they
can survive conditions that would destroy organics (Piperno 2006). Despite the
excellent preservation in some locations, taphonomic processes described above
affect starch and phytolith survival and identification (Jenkins 2009). Preservation of
phytoliths and starch granules factors in how well the remains can be identified.
Another factor affecting identification is contamination, which can occur in the field
98
or lab.
5.5 CASE STUDIES
Despite the limitations, current research applying phytolith and starch granule
analyses to archaeological contexts demonstrates the value of using this
methodology. This is especially the case in areas of poor preservation like the
boreal forest (Boyd et al 2014). Various studies use alternate methods for extracting
plant microfossils; however, they achieve the same objective of identifying new
aspects of paleodiet.
Using a variety of proxies including phytolith and starch granules, Boyd et al.
(2008) were able to interpret the northern geographic limits of maize consumption in
North America. This particular study looked at maize phytoliths and starch granules
from ceramic and soil residue.
Recovered maize phytoliths includes diagnostic
wavy-top rondels with entire bases, which is a unique form found in Zea mays and
produced in the cob portion of maize. Starch granules are positively identified as
maize if they are 20 µm in size, have a polygonal to rounded shape, a 90° extinction
cross under cross-polarized light, and a linear X or Y shaped central fissure (Boyd et
al. 2008). In Boyd et al. (2008), the outlined processing methods calls for 5 to 40
milligrams of carbonized food residue. The residue is digested in 50% HNO 3 for 12
to 24 hours, and repeated washing and centrifuging rid the sample of any remaining
acid. After processing the sample and isolating the phytolith and starch granules on
microscopic slides, the samples are analyzed using high-powered microscopy. The
samples yield evidence of both maize and beans, and results show that maize was
widespread in the study area by at least AD 700 (Boyd et al. 2008).
99
Berman and Pearsall (2008) illustrated the effectiveness of starch analysis to
reconstruct a diet consisting of Zea mays, Capsicum and Manihot esculenta in the
central Bahamas.
Maize, sweet potato, beans, and yautía were identified from
starch granules extracted from two ground stones associated with Archaic
populations in the Caribbean. In the same study, Berman and Pearsall (2008) also
analyzed 28 microliths from the Three Dog site and 15 had evidence of residue.
The starch and phytoliths were extracted using the piggyback method described in
section 6.3 of chapter six. Results interpreted the paleodiet of Archaic populations
and showed how and why these Caribbean inhabitants used domesticates (Berman
and Pearsall 2008).
Food residue analysis applied in a variety of regions has served to expand
knowledge of early plant use and consumption (Boyd et al. 2014). Methodological
applications include the use of food residue analysis on ceramics, lithics and matrix
samples. Recent studies have shown that these methods are successful in pointing
to when and where domesticates were consumed in the past (Boyd et al. 2014;
Piperno 2009; Piperno et al. 2009).
Starch and phytolith analyses are
advantageous methods that allow for accurate plant identification. The questions of
what people were consuming and when are difficult to answer in situations of poor
site preservation, bad context, a lack of historical texts, and/or ethnographies.
Rigorous analyses of starch and phytolith microfossils allow archaeologists to
compare samples spatially, and provide information on plant exploitation and
subsistence strategies. Overall, both the developing theoretical frameworks and
methodological applications are changing past notions of subsistence.
100
6.0 METHODS
6.1 OVERVIEW OF FOOD RESIDUE ANALYSIS
The botanical components of past subsistence strategies have become an
important aspect of archaeological inquiry.
Recent studies show the utility of
applying food residue methods to archaeological samples, because results allow for
new interpretations about site inhabitants and subsistence strategies (Haas et al.
2013; Henry et al. 2014; Liu et al. 2010; Madella et al. 2013). Current research in
this field focuses on strategies of identifying when and where cultigens have been
adopted (Blake 2006; Boyd and Surette 2010; Staller et al. 2006).
Plant microfossils such as starch granules and phytoliths are examined to
identify flora presence and use at archaeological sites.
These microfossils are
present in archaeological contexts as carbonized or non-carbonized food residues,
soils, and dental calculus (Henry and Piperno 2008; Horrocks 2005; Piperno et al.
2009; Rosenswig et al. 2014). Ceramics and fire-cracked rock may yield carbonized
residue, while fire-cracked rock, lithics, and sediment samples may yield noncarbonized plant remains.
This analysis will be applied to specifically chosen
samples from Middle Woodland sites within Northern Minnesota.
101
6.2 SELECTION OF ARCHAEOLOGICAL SAMPLES
Several few factors dictated which archaeological samples were suitable for
food residue analysis in this thesis. The first factor is whether the samples come
from a dated component, and whether they fall within a reasonable range for the
Middle Woodland period. AMS dates on residue, charcoal and bone, in addition to
samples directly associated with Middle Woodland ceramics and lithics, place
samples in an appropriate age bracket for this study.
A second factor pertains to the nature of the samples.
Specifically, the
ceramics must contain evidence of residue and the lithics should have a visibly
worked surface. Suitable samples include ceramics, lithics, fire-cracked rock and
matrix from archaeological features. Published material describes these contexts as
strong indictors of plant processing, preparation, and consumption (Boyd et al. 2014;
Yost et al. 2013; Zarrillo and Kooyman 2006).
The third and last major factor is the quality and amount of archaeological
information on the sites chosen for this study. Site forms and reports have been
accessed for eight out the nine sites in the study area. The ninth site was included
because of residue weight on the obtained ceramic samples.
Some alternate
material has been published on the sites; however, it is not a significant amount and
relates to larger studies of Brainerd and Laurel ware in the Upper Midwest
(Hohman-Caine et al. 2012; Mulholland 1997).
102
6.3 LABORATORY PROTOCOLS
Figure 6.1: Flowchart of methods for non-carbonized food residue.
103
Figure 6.2: Flowchart of methods for carbonized food residue.
A specific set of laboratory procedures was used to process the microfossils
adhering to the archaeological samples (Figures 6.1 and 62). Before the methods
are described in detail, it is necessary to review preliminary procedures. Two types
of tests were completed prior to processing, to rule out any risks of contamination in
the lab. The first test considers airborne contaminants in the processing laboratory.
Five slides were placed in strategic locations around the lab in weigh dishes and two
drops of silicone oil are added to trap any airborne particles. After 24 hours a cover
slip was placed on the silicone oil and the slides were analyzed with a high powered
104
microscope under a magnification of 20 times. The results were some fibres, and
dust particles, nothing significant such as microfossils to warrant re-testing the lab
for airborne contaminants. The second test consisted of, or involved, blank samples
that were run through the normal laboratory procedures for processing sediment and
residue samples (Figures 6.1 and 6.2). Most of the steps were included to utilize
laboratory equipment, thereby testing the cleanliness of the equipment and
laboratory environment. Similar to the test for airborne contaminants, the results
yielded some fibres and dust particles, but microfossils were not recovered. After
finding the test results contained no contaminants of concern, archaeological
material could be processed.
Methods used at Lakehead University were adapted by Surette (n.d.), and
based on Chandler-Ezell and Pearsall (2003) and Horrocks (2005). The extraction
procedures vary depending on the type of collected sample (Figures 6.1, 6.2, and
6.3).
105
A
B
D
C
Figure 6.3: Carbonized food residue on ceramics from the Windy Bead site (A), and the No
Beard site (B). Non-carbonized food residue on lithics (grinding stones) from the No Beard
(C) and Windy Bead sites (D).
Non-Carbonized Food Residue
Non-carbonized food residue develops from food remains that are not burnt,
although cooking could still produce this form of residue. This form of residue can
be found within archaeological contexts such as matrix samples, fire-cracked rock
and lithics.
Fire-cracked rock can contain carbonized and/or non- carbonized
residue, and the methods used to extract the microfossils are similar to lithic
extraction so fire-cracked rock is included under the non-carbonized section.
The analysis of soil samples begins with clay removal. If the sample contains
a large amount of clays, deflocculation and gravity settling are necessary. If not, a
graduated cylinder is filled with pure water up to the 100 millilitre mark and then the
sample is added until the water rises to the 150 millilitre mark, yielding 50 ml of soil
106
sample to work with. The sample is rinsed out with pure water to remove it from the
graduated cylinder into a new beaker. This is covered with calgon, which is a watersoftening agent and pure water is filled up to a measured 8 cm line drawn on the
beaker. The beaker is placed on a hot plate with a magnetic stirrer to mix the soil
sample. A pipette is used to remove and discard the supernatant; this process is
repeated until the supernatant is clear. Using a 118 µm disposable Nitex™ cloth
sieve, the contents of the beaker are filtered into a new beaker, and the sample is
rinsed from the original beaker and through the sieve. In some cases the remaining
larger particles are kept for macrofossil analysis. The rest of the sample is poured
from the beaker into centrifuge tubes.
The tubes are filled with the sample and pure water is filled to the 50ml line.
The tubes are centrifuged, the supernatant is removed with a pipette and the
sediment is concentrated in a single tube. After these steps, the starch granules are
extracted from the sample using density separation. Sodium metatungstate at a
specific gravity of 1.7 g/L is added to the 50 millilitre centrifuge tube, and filled up to
the five millilitre line. After centrifuging and pipetting the sample, the supernatant is
added into a new centrifuge tube and the remaining sample in the first tube can be
used for phytolith analysis.
Sodium metatungstate at a specific gravity of 2.3 g/L is added to the phytolith
samples; both phytolith and starch granule centrifuge tubes are filled up to the 50
millilitre line with pure water and centrifuged three times. After the last wash both
sets of microfossil samples are added to 1.5 ml microcentrifuge tubes with enough
ethanol to cover the samples. This is later deposited on microscopic slides.
107
Before groundstone residue is processed, four different samples are taken
from the tool. The four different samples are collected to extract as much residue as
possible from around and within the tool; thereby increasing the potential microfossil
counts and information on floral material (Surette n.d). Sample one is the sediment
in contact with the stone tool, sample two is the sediment deriving from dry brushing
of the stone tool surface, sample three is sediment from wet brushing and rinsing of
the stone tool surface, and lastly sample four is the sediment recovered from
sonication of the stone tool. Sample one is from a separate sediment bag collected
in the field or directly from the bag containing the stone tool.
Sample two is
extracted by dry brushing the stone tool with a toothbrush in a tray large enough to
accommodate the artifact. Sample three is extracted by washing the stone tool with
water and gently brushing it with a toothbrush in a tray. Sample four is extracted by
immersing the stone in a dish filled with water inside the sonicator, and sonicating
for 30 minutes. Once the materials from each tube are dried, the sample weights
are measured.
The phytolith and starch granules from stone tools are processed using density
separation. A metal spatula is used to place 60 mg of each sample into a new
centrifuge tube. Five millilitres of sodium metatungstate solution at a specific gravity
of 1.7 g/L is placed in each tube, which is centrifuged. A pipette is used to remove
the supernatant into a new centrifuge tube. This is re-centrifuged and re-pipetted to
increase starch extraction, which yields zero to ten millilitres of extracted sample.
The remaining material in the first centrifuge tube can later be used for phytolith
extraction.
Each tube is filled up with distilled water, and centrifuged; this is
108
repeated two more times to remove all of the sodium metatungstate. Ethanol is
added to the tubes with a squeeze bottle and the samples are pipetted into
microcentrifuge tubes to be later applied to microscopic slides.
After the starch extraction is finished the phytoliths can be extracted; sodium
metatungstate at a specific gravity of 2.3 g/L is poured in each tube with the
remaining materials from starch extraction.
A pipette is used to move the
supernatant into a new centrifuge tube, then centrifuged and pipetted to increase the
amount of phytoliths extracted. Each tube is filled with distilled water, centrifuged
and the supernatant removed with a pipette. This is repeated two or three more
times until all of the sodium metatungstate is removed. Ethanol is added to the
centrifuge tube, and the remaining samples are pipetted from the bottom of each
tube and placed in microcentrifuge tubes.
Carbonized Food Residue
Carbonized remains are formed from the release of carbon, which is later
preserved as residue on ceramics surfaces, and/or fire-cracked rock.
The
carbonized residue adhering to ceramics is gently scraped off with a small dental
pick or scalpel under a stereoscopic microscope (Figure 6.3). The residue is then
placed in a Petrie dish and from there into a microcentrifuge tube where the starch
granules can be extracted first followed by the phytoliths.
All the residue is used for starch extraction to reach a microfossil count of 250.
The samples are broken apart to make the extraction easier; afterwards, 6%
hydrogen peroxide is added to each centrifuge tube. After they are placed in the
orbital shaker, the samples are washed and centrifuged. If the sample still reacts
109
after this step, it is split into two samples and washed, centrifuged again. The
sample is then filtered with a 118 µm Nitex™ cloth and placed into microcentrifuge
tubes; at this point, ethanol is added to each tube and it is ready for mounting on
microscopic slides. The sample that did not go through the 118 µm Nitex™ cloth the
first time can be used for phytolith extraction; the larger particles are rinsed into the
centrifuge tube. Both the phytolith and starch extracts are centrifuged and washed,
the starch is placed into microcentrifuge tubes with ethanol for mounting, and the
phytolith sample is allowed 24 hours to dry.
After drying, 50% nitric acid is added to cover the samples and placed in a
water bath at 55 degrees Celsius for 12 to 24 hours. Durafilm is placed over the
glass openings and the rack is then added to the water bath under the fumehood.
Twelve to 24 hours are needed to sufficiently digest the residues; afterwards, the
samples are washed and centrifuged to remove any excess acid. The samples are
pipetted into microcentrifuge tubes where ethanol is added.
At this point, the
samples are ready to be mounted (Surette n.d).
After the extraction process the extracted material is mounted. Using a new
pipette, one to five drops from each sample are placed on a clean, labelled slide.
Starch samples can be stained with Trypan blue using only 10 ml of the agent. The
samples dry under the fumehood, four to six drops of Entellan are added to the
slides and a cover slip is placed on top. The phytolith and starch samples are
mounted with 30% thiodiethanol, because it has refractive properties that make the
microfossils more detailed. Toothpicks are used to manoeuvre the cover slips and
help clean the excess Entellan around the slide. Slides dry under the fumehood for
110
three days to ensure that samples do not get contaminated by airborne particles.
After the three days, samples can be analyzed using high powered microscopy.
6.4 IDENTIFICATION OF SAMPLES
High-powered microscopy was conducted with an Olympus Differential
Interference Contrast (DIC) microscope.
This type of microscope was chosen
because it allows for precise identification of diagnostic microfossil features (Shillito
2011).
Phytolith identification is based on the Brown (1984) key for phytolith
identification with the exception of rondels. This particular key uses three classes to
divide phytoliths into eight shape categories, which are further sub-divided into minor
shape categories. The three main classes are trapezoids, bilobates and saddles,
and they differ considerably when looking at tribes and subfamilies.
The key
represents 52 genera, 112 species of grass, and 16 tribes; there is also identification
information for 12 non-grass species. For rondel phytoliths, a key developed by
Surette (2008) is used for identification. Phytolith types are variable within different
plants and different plant parts. A range of phytolith shapes is present depending on
whether one is looking at the culms, leaves, roots and/or inflorescence (Shillito 2011;
Surette 2008). Phytoliths are identifiable at higher taxonomic levels, and although
many families have not been identified, cross-analyzing reference collections, and
comparative collections assist in identifying species to low taxonomic levels.
Starch granules are identified by examining grain size and thickness, shape
and position of hilum, presence or appearance of lamellae, and polarization cross
type (Boyd et al. 2008; Torrence 2006; Zarrillo et al. 2008).
For successful
identification, both cross-polarized light and plain light are used.
With starch
111
granules, taxonomic identification often relies on comparisons with individual
granules and reference granules (Wilson et al. 2010). Lints (2012) developed a key
for starch identification based on 45 plant species native to the boreal forest and
northeastern prairies, separated into five different classes: bell-shaped, circular,
elongated, angular, compound, basal and lastly irregular. Along with the phytolith
and starch identification, diatoms, and pollen grains are counted and identified if
possible. Phytoliths and starch granules have been shown to be more useful for
identifying plant remains than pollen in some cases, because they can be found in
larger quantities on archaeological material. Phytolith and starch assemblages from
domesticated plants such as maize, squash and beans will be described below, in
addition to native flora. Some starch granules from genera such as Cucurbita sp.,
Phaseolus vulgaris sp. and Zizania sp. are insufficiently distinctive from wild plant
confusers so in these cases, they are considered possible identifications. Table 4 is
composed of different plant species studied in this thesis and the diagnostic
characteristics that allow for confident identifications.
112
Table 4: Plant taxon and diagnostic characteristics of phytoliths, starch and pollen.
Taxon
Phytolith
Criteria
Zea
mays Wavy-top
ssp. Mays
Ruffle-Top
Concave
side
Entire base
Zizania spp.
Four spikes
or more
Indented
base
Honeycomb
Texture
Circular
Cucurbita sp.
Phaseolus
vulgaris sp.
Starch
Criteria
90 degree
extinction
cross
X or Y
shaped
fissure
18-25 mm
Six
compacted
sides
Hookshaped
3-8mm
Circular
10-20 mm
Circular
Rounded
cross-arms
25-40mm
Elongated
to Oval
Touching
Extinction
Cross
Pollen
Source(s)
55-100
mm
Exine
pattern
Single
pore
Boyd and
Surette
2010
Fearn and
Liu 1995
Holst et al.
2007
Pearsall et
al. 2003
Zarrillo et
al. 2008
Surette
2008
Boyd et al.
2014
Yost et al.
2013
Bozarth
1987
Duncan et
al. 2009
Bozarth
1990
Piperno
and
Dillehay
2008
Boyd et al.
2006
Maize (Zea mays ssp. Mays)
Maize (Zea mays ssp. mays) is identified based on a set of diagnostic features
found in phytoliths and starch granules. Over 40 different species of domesticated
maize have been identified, and 500 wild species may contribute maize-like
113
phytoliths, and because of this, a number of morphological characteristics are used
to validate the presence of maize in archaeological samples (Piperno 2009). Maize
rondel phytoliths will have either a wavy-top, or ruffle-top with concave sides and an
entire base (Figure 6.4, Bozarth 1993; Pearsall et al. 2003; Boyd and Surette 2010).
Figure 6.4: Maize wavy-top rondel phytolith (source: Pearsall 2004).
Wavy top and ruffle top are two types of rondels used to differentiate maize
from wild grasses, and they occur in the cupules and glumes of cobs (Pearsall et al.
2003; Piperno 2009).
Piperno (2009) has defined the three major indicators of
maize rondel phytoliths as a ruffle-top, wavy-top, and half decorated, though many
other phytoliths can be found in maize, some of which are also present in other
plants. The wavy-top is the only phytolith found solely in maize and the major
indicator used in this thesis (Boyd and Surette 2010; Haas et al. 2013; Pearsall et al.
2003; Zarrillo et al. 2008).
114
Grass seeds, especially maize kernels, produce large quantities of starch
granules, with morphological characteristics unique to Poaceae (Piperno et al.
2004).
In areas where teosinte is not present, maize starch granules can be
distinguished from native grasses in these regions. Maize starch granules commonly
range in size from 18 to 25 micrometers (µM) in length. In many of the native
grasses, granule size ranges from a maximum length of two to 18 mm, and an
average of three to 11 mm (Haas et al. 2013; Holst et al. 2007). Maize starch
exhibits a y, or x fissure, contains up to six compacted sides and a 90° extinction
cross (Figure 6.5).
Based on the reference collection developed at Lakehead
University, these features are unique to maize and are not present in any native
flora (Lints 2012).
A
B
Figure 6.5: Maize starch grain under Cross-Polarized Light XPL (A) and PlainPolarized Light PPL (B).
Squash (Cucurbita)
The main identifying feature of squash is large circular shaped phytoliths with a
honeycomb texture (Figure 6.6). Bozarth (1987) describes the diagnostic features
115
of squash as hemispheroidal and spheroidal phytoliths that have scalloped surfaces
with concave depressions. The phytoliths extracted from squash are produced in
the fruit and can be found in the seeds, peduncles and rinds of the plant (Bozarth
1987). The lack of these phytoliths in the flesh of the fruit means that squash will
tend to be under-represented or invisible in archaeological cooking residues
(Bozarth 1987).
Figure 6.6: Squash phytoliths, A-D illustrate different samples (source: Piperno et al. 2007)
116
Figure 6.7: Starch granules identified as squash, A to F illustrate different samples of
squash (source: Duncan et al. 2009)
Squash starch grains are easily identified from the cross-arms viewed under
XPL (Duncan et al. 2009). However, the grains appear identical so it is impossible
at this time to identify different varieties of squash from starch granules (Figure 6.7).
As well, there are some wild plants that produce similar looking granules.
Common Bean (Phaseolus vulgaris)
Common bean or Phaseolus vulgaris yields phytoliths that are recognizable by
large hook shaped microfossils, as illustrated in figure 6.8 (Bozarth 1990). Bozarth
(1990) described a few varieties of beans that all take on this hook shape; these
varieties include P. vulgaris and P. lunatus.
Phaseolus lunatus and P. vulgaris
starch range from 25 to 40µm in length, and are typically elongated in shape. A
diagnostic feature of bean grains is the touching extinction cross (Figure 6.9).
117
Figure 6.8: Hook shaped phytoliths of Phaseolus vulgaris, A to D illustrate different
samples (Bozarth 1990)
Figure 6.9: Phaseolus starch grain (Lints 2012).
118
Wild Rice (Zizania sPp.)
Wild rice (Zizania ssp.) phytoliths are identified by the presence of a variety of
morphological features. The criteria for identification are that the rondels must have
an indented base, with multiple spikes on the top (Surette 2008).
Figure 6.10: Wild rice rondel phytolith from the MacGillivray site, Ontario.
Various forms of wild rice can be identified down to the species level, these wild rice
rondels exhibit four spikes, and three indentations (Figure 6.10). Wild rice will also
produce rondels with only three spikes and an indented base, although this is not
considered diagnostic because native flora can also produce similar rondels.
Zizania starch was described in the comparative starch key produced by Lints
(2012), however; this example has shown the difficulties in applying the key to
archaeological samples. Zizania starch is only 3 to 8µm in size and a wide variety of
native plants contain morphologically similar traits.
119
Native Flora
A wide variety of native plants, other than wild rice, cannot yet be identified
because comparative studies of native species have not yet been developed, and
such work is beyond the scope of this thesis. Although the comparative reference
collection at Lakehead contains over 300 plant species from the Boreal Forest and
north-eastern Plains, many native plants from the Minnesota area could not be
identified with the current reference collection. Native plants are able to be identified
from rondel phytoliths with the key developed by Surette (2008); however, starch
granules are more difficult to identify because many native plants produce starch
that is similar in shape and size.
As summarized in Appendix 1, some of the native flora identified from Middle
Woodland sites in Northern Minnesota includes Chenopodium gigantospermum,
raspberry (Rubus idaeus), blueberry (Vaccinium angustifolium), wild strawberry
(Fragaria virginiana), elderberry (Sambucus) and smartweed (Polygonum sp.). The
distribution
of
Chenopodium
gigantospermum
(Chenopodiaceae)
(mapleleaf
goosefoot) is found over most of North America (University of Minnesota Herbarium
n.d; Smith 2006; Personal Communication with Surette, Lakehead University, 2012).
This common plant spreads easily and macroremains were recovered at the Big
Rice site (Shafer 2003). Wild raspberry (Rubus idaeus) is a common plant found at
some Woodland sites in Northern Minnesota (Shafer 2003). Raspberries have red
fruit that grows from hollow shells when stalks are separated and they are found in
thickets (Peterson 1977). Vaccinium angustifolium (Ericaceae) or blueberry is also
120
found in the Northern Minnesota area (Shafer 2003).
This perennial shrub is
common in low and high thickets and can be eaten fresh, dried or cooked. The
berries form a star pattern with five calyx lobes (Peterson 1977; University of
Minnesota Herbarium n.d). Wild strawberry or Fragaria virginiana (Rosaceae) is
also native to Northern Minnesota (University of Minnesota Herbarium n.d; Shafer
2003).
Strawberry starch granules have a platy structure within a larger oval
enveloping shape (Knee, n.d). These low lying plants can be dried, eaten fresh or
cooked; the leaves can also be eaten raw or boiled (Peterson 1977). Sambucus
canadensis or elderberry macroremains have been found at archaeological sites
(Shafer 2003) in the study area. Juicy, purple or black berries top flowers with five
petals, and they are mostly eaten raw (Peterson 1977).
Some varieties of
smartweed or Polygonum sp. (Polygonaceae) are considered invasive weeds and
wetland species (University of Minnesota Herbarium n.d; Shafer 2003). Polygonum
sp starch granules are shaped like teardrops and roughly 0.15 mm (Hart 2011).
6.5 COMPARATIVE REFERENCE KEY
Reference collections derive from modern flora, sometimes sampled near
archaeological sites, and from these plants of known species, phytoliths and
starches are extracted to compare to archaeological samples. These collections
provide background information on modern plants and morphological characteristics
of plant microfossils. Detailed descriptions and comparisons of phytolith and starch
morphology and size allows for the recognition of squash and gourds (Curcurbita
ssp.), wild rice (Zizania spp.), beans (Phaseolus ssp.), maize (Zea mays ssp. mays),
and various native flora. Comparative and experimental studies can help extend
121
reference collections by adding in modern plant species that may have been present
and used in the past (Ollendorf et al. 1988). Different regions produce variations in
native vegetation based on climate and topography so it is important to become
familiar with regional vegetation.
This study includes two different reference
collections at Lakehead University to cover both Parkland and Plains ecological
zones.
The first reference collection is an assortment of Plains flora and was
compiled by Lints (2012), while the second collection focuses on Boreal Flora and
this is supplied by Lakehead University (Figure 6.11).
122
A
B
A
C
D
E
F
G
H
Figure 6.11: Comparative plant microfossils. A (PPL) + B (XPL) maize starch
granules; A (PPL) + B (XPL) bean starch granules; E (PPL) + F (XPL) squash
starch granules; G (PPL) wild rice rondel phytolith; H (PPL) maize rondel
phytolith.
123
6.6 UNKNOWN SPECIES
The inability to identify certain species simply because they have not been
added to reference collections is another limitation that over time will be lessened.
In order to identify unknown species for this study, different reference collections
were cross-referenced and modern collections were assessed.
The cross-
referenced databases include several online sources such as Plants for a Future,
United States Department of Agriculture Plants Database (http://plants.usda.gov;
http://www.pfaf.org), and the University of Minnesota Herbarium. The comparative
reference collection at Lakehead University was also used to compare unknown
microfossils in archaeological samples to comparative samples from the Canadian
Plains and Boreal Forest.
The application of phytolith and starch analyses with the methods described
above are an effective means of identifying non-native and local plants in
archaeological contexts.
Archaeological contexts include ceramics, lithics or
sediment samples containing carbonized and/or non-carbonized food residue. The
procedures for carbonized and non-carbonized food residue vary primarily because
of the chemical agents used for density separation and the initial steps used to
extract the microfossils. The analysis of microscopic slides with an Olympus (DIC)
microscope allows for the precise identification of plant microfossils based on
diagnostic traits.
124
6.7 CONTAMINATION
Contamination is the result of a number of factors attributed to human error in
the field or lab (Crowther et al. 2014). In the field, contamination can result from not
keeping excavated archaeological materials separate, resulting in the mixing of
surface soils and sediments from different layers. A way to test for sediment mixing
is to use control samples from within archaeological layers and the surrounding area
to determine the presence of contaminants in the archaeological samples. Another
form of contamination occurs when excavators do not take the necessary
precautions to wash their hands after breaks, which can cause artifacts to be
exposed to food remains (Cummings 2007; Personal Communication with Megan
Wady, Lakehead University, 2012).
In the lab, contamination can result from airborne starch and poor control over
people entering the laboratory (Crowther et al 2014). Monitoring for the presence of
airborne starch can be done by taking samples of areas where airborne starch may
come to rest and testing for their presence/absence. Although Crowther et al. (2014)
describes the low statistical validity and unreliable results with passive starch traps,
passive traps coupled with blank contamination tests are routinely used at
Lakehead. By using multiple tests for contamination there is a decreased risk of
undetected cross-contamination by airborne microfossils, or microfossils entrapped
in laboratory equipment. The laboratory should be kept as clean as possible to
eliminate risk factors, and this includes daily wipe downs. As well, all used materials
should be cleaned in a sonicator or by hand with industrial soap. When experiments
are in progress, a sign should be posted on the doorway asking people not to enter
125
the laboratory, thereby controlling in and out traffic. Results from Crowther et al.
(2014) on starch contamination showed that items such as non-powdered gloves
and Calgon could cause starch contamination. In addition, the uses of bleach or
weak acids to decontaminate the laboratory are ineffective (Crowther et al. 2014).
At Lakehead University, the deflocculent that we use for soil processing tested
negative for starch contamination; therefore, the results obtained in this study are
not affected by lab contamination.
With contamination in mind and the proper
procedures followed to maintain a clean environment, it is unlikely samples have
become compromised in the lab (Crowther et al. 2014).
126
7.0 RESULTS
7.1 INTRODUCTION
The following chapter presents the results from carbonized and noncarbonized food remains found adhering to ceramics, residue on lithics and soil
samples. These remains are from the nine sites described in Chapter one, which
date primarily to the Middle Woodland and are associated with Brainerd and/or
Laurel pottery contexts.
Images collected from the comparative collection at
Lakehead University in Chapter six can be cross-referenced with the archaeological
microfossils described in the following chapter.
7.2 RESIDUE SAMPLE SIZES
Sample sizes varied between carbonized and non-carbonized residue at all
nine sites, with some samples weighing less than five milligrams, while others
yielding over 70 milligrams (Figure 7.1, Table 5).
127
A
B
C
D
E
F
Figure 7.1: Examples of Brainerd and Laurel ceramics analyzed in this study. (A),
(B) Brainerd sherds from the Third River Borrow Pit site, (C) Brainerd sherd from the
Windy Bead site, (D) Brainerd sherd from the Saga Island site, (E) Laurel rim from
the Saga Island site, and (F) Laurel sherd from the Lost Lake site.
128
Table 5: Sample weight of carbonized residue from ceramics, non-carbonized residue
from lithics and soil, along with provenience information.
Site
Sample Code
Third
River
Borrow Pit
1 or 1-355-229.1-4
2 or 1-355-276.1-4
3 or 1-355.274.2
4 or 1-355-266.2-13
5 or 1-355.300.1
6 or 1-355-298.1-4
7 or 1-355-295.1-2
8 or 1-355-216.1
9 or 1-355-296.3-5
10 or 1-355-285.1
Big Rice
1 or 09-034-IB15
3 or 09-034-385
2 or 09-034-335
Windy Bead
AB37 or 0537335AB37
AB36 or 78AB36
AB24 or 0537378AB24
AB41 IS AB41
No Beard
AB61
Kyleleen’s
Bent Pine
AB5
AB6
AB11
AB62
Kyleleen’s
Tall Pine
Saga Island
Lost Lake
AB45
AB90
814AB16 or AB16
814AB18 or AB18
814AB19 or AB19
AB43
Sample Size mg
(Ceramics)
1 (78.1) Unit 25/1E Level 4 SE
quad
2 (5.0) Unit 25/0.5 E Level 5 NW
quad
3 (50.2) Unit 25/0.5 E Level 5
SW quad
4 (2.1) Unit 25/0.5E Level 3 SW
quad
5 (6.5) Unit 35/0 Level 5 SE
quad
6 (13.6) Unit 35/0 Level 4 NE
quad
7 (22.5) Unit 35/0 W Level 4 NE
quad
8 (23.7) Unit ON/1W Level 6 SE
quad
9 (13.0) Unit 35/0 Level 4 SE
quad
10 (9.4) Unit 25/0.5 E Level 8 W
½ of NE quad
1 (5.9)
2 (1.2) Lot 225 ACC 112
3 (5.9)
Sample Size mg
(Lithics)
AB37 (24.5) Unit 1 Level 7 NE
quad
AB36 (6.6) Surface ACC #565
Lot 78
AB24 (18.6) Surface ACC #565
Lot 78
AB41 (34) Unit 1 Level 4 SW
quad Lot 20
AB61 (5.1) Unit 14 ACC #317
Lithic 1 Unit 3 Level 6
Lot 98 SE quad ACC
#565
L5: Wetbrush (59.0)
L6: Sonicate (19.6)
Sample Size mg
(Soil)
S1 (27.54) 98/94 LV.
4 LOT #201 S1 S2
(34.44)
85/42 Lv 3. Area 2 S2
S3 (50.0) Unit 5
S4 (50.0) Unit 4
S5 (50.0) 914-03
Lithic 2 Unit 11 Level 11
Lot 501 SE quad ACC
#317
L1:
Working
Edge
Sonicate (64)
L2: Wetbrush (166.8)
L3: Sonicate (56.7)
L4: Dry Brush (67.8)
AB5 (>7) Surface
AB6 (<7) Surface
AB11 (20) Surface
AB62 (11.5) Surface ACC #996
Lots #1-3
AB45 (45.6) Unit 4 Level 6 SE
quad
AB90 (44.8)
AB16 (21)
AB18 (7.4)
AB19 (5.2)
AB43 (49.5)
129
Winnie
Cottages
AB60
AB40
AB44
AB60 (45.6) 5N 11W 20-25 cms
SW quad
AB40 (51.5) 10N 75W 10-15cms
NW quad
AB44 (9.3) 0N 13W 10-15cms
SW squad
B
The sample weight from ceramic residue was not high enough to include any
additional studies such as AMS dating. The Windy Bead, Saga Island and Lost
Lake sites produced ceramics with the thickest residue, while the Kyleleen’s Bent
Pine and Big Rice site had thinner layers of ceramic residue. The ceramic sherds
analyzed from the other five sites had varying amounts of residue. Most ceramics
are either Horizontally Corded, Net Impressed Brainerd ware, and/or a type of
Laurel decorated ware (Table 6). Different native plant species were identified from
rondel phytoliths, in addition to non-native flora (Appendix 9).
These are only
possible identifications that were counted based on the reoccurrence of rare rondel
types in the plant assemblage. A comparative key is not available for Northern
Minnesota flora; therefore, the possibility that a native plant exhibits similar
characteristics as the identified rondel phytoliths cannot be ruled out.
130
Table 6: Ceramics from the ten chosen study sites.
Site
Third
River
(21IC176)
Borrow
Pit
Sample Code
Ceramic Attributes
1 or 1-355-229.1-4
2 or 1-355-276.1-4
3 or 1-355.274.2
4 or 1-355-266.2-13
5 or 1-355.300.1
6 or 1-355-298.1-4
7 or 1-355-295.1-2
8 or 1-355-216.1
9 or 1-355-296.3-5
10 or 1-355-285.1
1: Thick-walled ware. Brainerd Net or Fabric Impressed.
2: Brainerd Net Impressed with clear knotholes.
3: Unsmoothed surface finish, Fabric or Net Impressed.
4: Questionable. Middle Woodland.
5: Textile Impressed, probably not Brainerd ware. Late or
Brainerd Cord:
Impressed surface finish. Dates to the
Woodland.
6: Brained Net Impressed.
7: Parallel grooved Brainerd ware. Thick, grit tempered
Middle Woodland.
8: Might be Net or Fabric Impressed Brainerd Ware.
9: Middle Woodland, Laurel ware.
10: Cord Impressed surface finish, Brainerd.
1: Laurel pattern, but it might be Late Woodland (Selkirk?).
2: Probably Net Impressed could also be Fabric Impressed.
3: Weathered Laurel Push Pull decoration.
Big Rice
1 or 09-034-IB15
3 or 09-034-385
2 or 09-034-335
Windy Bead
AB37
0537335AB37
AB36 or 78AB36
AB24
0537378AB24
AB41 is AB41
AB61
No Beard
Kyleleen’s Bent Pine
Kyleleen’s Tall Pine
Saga Island
Lost Lake
Winnie Cottages
or
or
AB37: Laurel Smoothed.
AB36: Laurel, thick grit temper.
AB24: Brainerd with a tool impression on the lip corner.
AB41: Laurel Dentate Stamped, weathered.
AB61: Presumably Laurel. Dentate or verging on Dentate and
Pseudo Scallop Shell.
AB5: Late Woodland Textile Decoration.
AB6: Late Woodland Textile Decoration.
AB11: Late Woodland Textile Decoration.
AB62: Presumably Brainerd, Middle Woodland. There is a
strange temper.
AB45: Laurel Pseudo Scallop Shell, Middle Woodland.
AB90: Chevrons, Dentate Stamped Laurel, Middle Woodland.
814AB16: Net Impressions, Brainerd.
814AB18: Thin-walled Laurel Stamped.
814AB19: Textile Impressed Late Woodland.
AB43: Laurel or Sandy Lake Smoothed Plain.
AB60: Combed Blackduck Late Woodland.
AB40: Net Impressed knots, Brainerd ware.
AB44: Brainerd Net-Impressed.
AB5
AB6
AB11
AB62
AB45
AB90
814AB16 or AB16
814AB18 or AB18
814AB19 or AB19
AB43
AB60
AB40
AB44
Only two groundstone artifacts were analyzed from the Windy Bead and No
Beard sites and both produced substantial sample weights (Figures 7.2 and 7.3,
Table 5).
There were only two grinding stones tested in this study because
groundstones dating to the Middle Woodland with evidence of plant processing are
C
D
131
especially scarce. A working edge was visible on the No Beard groundstone, which
was sonicated separately to identify differences in microfossils from the worked
edge and the rest of the stone.
A
B
Figure 7.2: Two grinding stones used for residue analysis. A is from the No
Beard site and B is from the Windy Bead site.
A
B
B
Figure 7.3: Locations of samples extracted from the two grinding stones. (A) No
Beard site, sample one is from a dry bush, sample two is a wet brush, sample
three is from sonication, sample four is sonication of the worked edge; (B) Windy
Bead site, sample one is from a wet brush, sample two is from sonication.
The five soil samples from Big Rice and the Windy Bead site yielded larger
quantities of microfossils than the other samples and had substantial sample
weights (Table 5). The five soil samples contained a wide array of native plant
rondel phytoliths (Appendix 6). Soil samples were collected during the excavation
process from archaeological features including two hearths from the Big Rice site.
132
The following figures (7.4 to 7.9) include results from both carbonized and noncarbonized residue. Figure 7.4 to 7.9 are quantitative diagrams that depict the type
of microfossil recovered from ceramics, lithics and sediment samples.
Figure 7.4: Diagnostic phytolith and starch data from sediment samples. Big Rice:
BR, Windy Bead: WB. Amounts are expressed as percentages of the total
microfossil count (n=250).
133
BR (1)
BR (2)
WB (1)
WB (2)
WB (3)
E
lo
0
n
g
a
te
la
P
lo
n
g
a
lo
0
E
n
g
h
20 0
S
t
or
h
ric
20 0
T
e
a
e
th
in
o r
m
o e
ra
m th
G
S O
s
s s
te
te te
s
la la
P P
te
la
te te
P
a
E
20 0
om
e
s
D
20 0
ou
b
le
O
li
ut
n
e
a
s
S
20 0
dd
le
s
L
o
20 40 0
n
g
S
in
u
o
u
s
T
ra
p
e
zo
id
s
S
20 40 60 0
h
t
or
in
S
u
o
us
T
ra
p
L
e
o
20 40 0
g
on
s
N
id
zo
n
in
S
u
ra
p
in
s
u
b
s
ilo
uo
id
zo
S
e
n
B
20 0
o
N
T
t
or
us
h
o
S
20 0
pe
s
ra
T
e
at
zo
id
o
s
P
20 0
0
0
r
U
20 40 60 80 100 120 140 0
e
s
h
s Ot
te
ba es rm l
l o s f o de
ly ros l i on
u
C B R
0
20 0
0
s
h
lit
o
yt
h ch
P r
a
d t
e S
ifi n
s
ss w
o
m n
la n
to e
nc nk
ia ol
U
D P
0
Figure 7.5: Microfossil assemblage from sediment samples. Big Rice: BR, Windy
Bead: WB. Amounts are expressed as percentages of the total microfossil count
(n=250).
134
ar
ch
st
ar
is
ulg
rc
h
sv
sta
lu
eo
Ph
as
ur
e
Po
ss
ibl
e
ibl
ss
Po
Ze
a
Ro
nd
el
m
ay
s
(Z
i
Cu
c
za
sta
nia
rc
h
bit
a
)
sp
s)
ay
m
a
(Z
e
el
Ro
nd
TRBP (8)
TRBP (9)
TRBP (1)
TRBP (10)
TRBP (7)
TRBP (6)
TRBP (5)
TRBP (4)
TRBP (3)
TRBP (2)
BR (1)
BR (2)
BR (3)
SI (1)
SI (2)
SI (3)
WB (1)
WB (2)
WB (3)
WB (4)
WC (1)
WC (2)
WC (3)
KTP (1)
KBP (2)
KBP (3)
LL (1)
LL (2)
LL (3)
LL (4)
NB (1)
0.0 0.4 0.8 1.2 1.6 2.0 0.0
2.4
4.8
7.2
9.6 0.0 0.4 0.8 1.2 1.6 2.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Figure 7.6: Diagnostic phytolith and starch data from ceramic samples. Third River
Borrow Pit: TRBP, Big Rice: BR, Saga Island: SI, Windy Bead: WB, Winnie
Cottages: WC, Kyleleen’s Tall Pine: KTP, Kyleleen’s Bent Pine: KBP, Lost Lake: LL,
No Beard: NB. Amounts are expressed as percentages of the total microfossil count
(n=250).
135
Pollen
s
Diatom
Phytolit
hs Unk
nown
cf Amel
anchier
alnifolia
starch
Starch
Unknow
n
er)
Bullifor
ms
Rondel
(Oth
Polylob
ates
Crosses
Bilobate
s
Short N
on Sinuo
us Trape
zoids
Long N
on Sinuo
us Trape
zoids
Short S
inuous
Trapezo
ids
Long S
inuous
Trapezo
ids
Outlines
Saddles
Double
Elongat
e Plate
s (Gram
ineae)
Elongat
e Plate
s (Smoo
th)
Elongat
e Plate
s (other
)
Short P
lates
Trichom
es
TRBP (8)
TRBP (9)
TRBP (1)
TRBP (10)
TRBP (7)
TRBP (6)
TRBP (5)
TRBP (4)
TRBP (3)
TRBP (2)
BR (1)
BR (2)
BR (3)
SI (1)
SI (2)
SI (3)
WB (1)
WB (2)
WB (3)
WB (4)
WC (1)
WC (2)
WC (3)
KTP (1)
KBP (2)
KBP (3)
LL (1)
LL (2)
LL (3)
LL (4)
NB (1)
0 0 0 0 0 20 0 20 0 20 0 20 40 0 20 0 20 0 20 0 20 0 0 0 0 20 40 60 0 20 0 0 20 40 60 80 100 120 140 0 20 40 60 0
Figure 7.7: Microfossil assemblage recovered from ceramic samples. Third River
Borrow Pit: TRBP, Big Rice: BR, Saga Island: SI, Windy Bead: WB, Winnie
Cottages: WC, Kyleleen’s Tall Pine: KTP, Kyleleen’s Bent Pine: KBP, Lost Lake: LL,
No Beard: NB. Amounts are expressed as percentages of the total microfossil count
(n=250).
136
Figure 7.8: Diagnostic phytolith and starch data from lithic samples. No Beard: NB,
Windy Bead: WB. Amounts are expressed as percentages of the total microfossil
count (n=250).
137
ez
oid
s
sT
r ap
ez
oi d
s
wn
uo
u
Dia
20 0
to
Po m s
lle
n
ate
s
lylo
Cr bat
os
es
Ot ses
he
rR
on
de
l
Sta
rch
Un
kn
o
Po
ob
Bil
Lo
Sh
ng
or t
No
n
No
nS
i nu
ou
sT
or t
Sin
u
Sh
Si n
rap
ou
sT
ez
rap
oid
s
ids
zo
pe
Tra
us
uo
Sin
ng
Lo
ng
Elo ate
ng Pla
Elo ate tes
(G
ng Pl
Sh ate ates ram
ine
Pla ( S
or t
Tri Pla tes m oo ae)
tes
ch
( ot t h)
om
he
r)
es
Do
ub
le
Ou
tlin
es
Sa
dd
l es
Elo
0
NB(1)
WB (1)
WB (2)
NB (2)
NB (3)
NB (4)
0
0
0
0
0
20
0
20
0
20
0
20
40
60 0
20
40
0
20
0
20
40
0
0
20
40 0
20
40
60
80 100 0
0
Figure 7.9: Microfossil assemblage recovered from two grinding stone samples.
No Beard: NB, Windy Bead: WB. Amounts are expressed as percentages of the
total microfossil count (n=250).
7.3 THIRD RIVER BORROW PIT (21-176)
Phytolith and Starch Content of Carbonized Food Residue (Figures 7.6 and 7.7)
The Third River Borrow Pit site had ten ceramic sherds analyzed for phytolith
and starch granules (Tables 6 and 13). One sherd produced a diagnostic Z. mays
ssp. mays. rondel phytolith, while all the sherds had unknown phytoliths and
additional rondel phytoliths (Figure 7.11). One sample yielded 60 unknown starch
granules, the other nine samples had fewer than 12 unknown starch granules
(Figure 7.7).
138
7.4 BIG RICE (09-034)
Phytolith and Starch Content of Carbonized Food Residue (Figures 7.6 and 7.7)
Three ceramic sherds were analyzed from the Big Rice site (Tables 6 and 13).
Two sherds yielded three diagnostic Z.mays ssp. mays. rondel phytoliths (Figure
7.11). Nine Zizania sp. rondel phytoliths were identified on one sherd (Figure 7.12).
There were 67 other rondel phytoliths produced by one sherd, while the other
sherds had only one unidentified rondel phytolith (Figure 7.7). Z.mays ssp. mays.
starch granules were recovered from one sherd and fewer than three unknown
starch granules were located on all three sherds.
Phytolith and Starch Content of Non-Carbonized Food Residue (Figures 7.4 and 7.5)
The Big Rice site had two soil samples analyzed for microfossils; the first
sample contained charcoal and possible wild rice, while the other sample was
extracted from a hearth rock ring (Appendix 8). Charred wild rice was recovered in
substantial quantities from the site, some of which was radiocarbon dated (Figure
7.16). The sample from the feature containing charcoal and possible wild rice is in
level four, and the sample from the feature containing a hearth rock ring is in level
three (Peter and Motivans 1983). The highly disturbed sediments at depth suggest
cultural mixing between Middle and Late Woodland occupation layers.
samples contained Z.mays ssp. mays. rondel phytoliths.
Both
There were also four
Z.mays ssp. mays. starch granules recovered from one sample (Figure 7.4).
Between the two samples, 15 Zizania sp. were identified (Figures 7.4 and 7.12). A
single possible Phaseolus vulgaris or common bean starch was recognized,
although it is smaller in size than the grains that appear to be characteristic of
139
Phaseolus (Figure 7.14). A possible Cucurbita sp. starch granule was also identified
from the same soil sample, although starch evidence for squash is inconclusive as
discussed above (Figures 7.4 and 7.15). Both samples contained a large amount of
other rondels, with one sample having a count of 90 and the other a count of 50.
Diatoms of different algae species were recovered from both Big Rice site samples
(Figure 7.5).
7.5 KYLELEEN’S TALL PINE (07-505)
Phytolith and Starch Content of Carbonized Food Residue (Figures 7.6 and 7.7)
Only one ceramic sherd was analyzed from Kyleleen’s Tall Pine and three
unknown rondels, an unknown starch and a diatom were recovered, in addition to a
low number of phytoliths (Figure 7.7).
7.6 WINNIE COTTAGES (02-526)
Phytolith and Starch Content of Carbonized Food Residue (Figures 7.6 and 7.7)
Three sherds were analyzed from the Winnie Cottages site and all three had
variable sample weights compared to samples from other sites. One ceramic sherd
had 9.3 milligrams of residue and the other two sherds had around 48 milligrams of
residue. One maize rondel phytolith was identified and two of the samples contained
five to eight other rondel phytoliths (Figures 7.6 and 7.7). One unknown starch, a
diatom and two pollen granules were recovered from all three samples (Appendix 7).
7.7 WINDY BEAD (05-373)
Phytolith and Starch Content of Carbonized Food Residue (Figures 7.6 and 7.7)
Four ceramic samples were analyzed from the Windy Bead site. One Laurel
sample contained a diagnostic maize rondel phytolith (Figure 7.11, Appendix 7).
140
Possible Cucurbita starch granules were identified from one Laurel and one
Brainerd sherd (Figures 7.6). Fewer than 12 unidentifiable rondels were recovered,
as well as four unknown starch granules from one sample and 62 on another (Figure
7.7).
Phytolith and Starch Content of Non-Carbonized Food Residue (Figures 7.5, 7.5, 7.8,
and 7.9)
There are three soil samples associated with the Windy Bead site, all of which
produced over 300 microfossils (Figure 7.5, Appendix 8). The rondel count was
especially high for all three Windy Bead soil samples, with individual sample counts
ranging from 24 to 139 rondels (Table 5) (mean=89). The three samples contained
Z.mays ssp. mays. rondel phytoliths (Figure 7.4). Zizania sp. was identified on two
out of the three samples from rondel phytoliths (Figure 7.12). In addition, a possible
Cucurbita starch granule was identified in a single soil sample.
Diatoms and
unknown pollen were located in two out of the three soil samples (Figure 7.5).
Raphides, which are thin calcium oxalate rods or crystals commonly found in some
native wild plants such as plants of the families Amaranthaceae, Araceae,
Asteraceae, and Poaceae were recovered from one sample (Arnott and Webb
2000).
The stone tool analyzed from the Windy Bead site (Figures 7.2, 7.3, 7.8, and
7.9) was wet-brushed and sonicated, producing less microfossils than the No Beard
stone tool, but still containing a much higher microfossil count than most ceramic
samples. Very few rondel phytoliths were recovered, although one is diagnostic of
Zizania sp. Among the 121 starch granules was one maize phytolith and one maize
141
starch granule (Figures 7.8, 7.9, 7.11, and 7.13).
7.8 SAGA ISLAND (02-162)
Phytolith and Starch Content of Carbonized Food Residue (Figures 7.6 and 7.7)
Two sherds were analyzed from the Saga Island site, both with sample weights
around 45 milligrams (Table 5, Appendix 6).
A maize rondel phytolith was
recovered, along with a Zizania sp. phytolith from the same Laurel sherd. Both
ceramic samples yielded a possible Cucurbita starch granule (Figure 7.6). Between
the two sherds, four unknown starch granules, six rondel phytoliths and one
unknown pollen grain were recovered.
7.9 LOST LAKE (2-00LL)
Phytolith and Starch Content of Carbonized Food Residue (Figures 7.6 and 7.7)
The four Lost Lake ceramic sherds had one sample with very thick residue,
while less than 10 milligrams of residue was extracted from the other three sherds
(Table 5, Appendix 7). Only one Zizania sp. phytolith was recovered, and between
the four samples, seven other rondel phytoliths were identified.
Phaseolus vulgaris starch granule was identified on a Laurel sherd.
A possible
Unknown
starch was present in three samples, with the highest count at 16. One sample
contained 62 diatoms, while the other three samples had an average of two
recovered diatoms (Figure 7.7).
7.10 NO BEARD (05-264)
Phytolith and Starch Content of Carbonized Food Residue (Figures 7.6 and 7.7)
A single Laurel sherd was analyzed from the No Beard site (Appendix 7).
Although there was little residue, 152 microfossils were recovered (Figure 7.7). 150
142
starch granules were identified from clusters located on several areas of the
microscopic slide.
Phytolith and Starch Content of Non-Carbonized Food Residue (Figures 7.8 and 7.9)
Four different samples were extracted from the No Beard grinding stone, one
sample was sonicated from the working edge, while another was sonicated from the
rest of the stone tool (Figures 7.2 and 7.3). The last two samples were obtained
through wet-brushing and dry-brushing the tool.
Diagnostic Z.mays ssp. mays.
rondel phytoliths were identified from the sonicated and wet brushed samples
(Figures 7.8 and 7.9).
sonicated sample.
A Zizania sp. rondel phytolith was recovered from the
Some samples produced large quantities of other rondel
phytoliths, with counts as low as seven and as high as 33. Z.mays ssp. mays.
starch was identified on three of the four samples (Figures 7.10 and 7.13). There
was a significant amount of unknown starch ranging from 25 to 50 granules (Figure
7.9). Diatoms were located on two samples, in addition to one unknown pollen grain
(Appendix 5).
A
B
Figure 7.10: Damaged starch granule, recovered on a grinding stone from the No
Beard site, (A) XPL, (B) PPL.
143
7.11 KYLELEEN’S BENT PINE (THIRD RIVER)
Phytolith and Starch Content of Carbonized Food Residue (Figures 7.6 and 7.7)
Out of the three samples from the Kyleleen’s Bent Pine site, one yielded
significantly high quantities of starch with a count of 68 (Figure 7.6, Appendix 7). All
three ceramic sherds produced relatively low microfossil counts with no identifiable
types. The lack of microfossils produced by cultivated or other identifiable plants
may be due to low sample weight obtained from the ceramics or, alternatively, their
absence at the site.
The following images (Figures 7.11 to 7.16) illustrate examples of different
plant microfossils and macrofossils such as maize phytoliths and/or starch, wild rice
phytoliths, possible squash and bean phytoliths, and plant seeds recovered during
data analysis and excavation.
144
A
B
C
D
E
F
Figure 7.11: Microfossil photos of Z.mays ssp. mays. observed in carbonized food
residue, soil and stone tools (A) Third River Borrow Pit site. (B) Big Rice site. (C, D,
E) Windy Bead site. (F) No Beard site.
145
A
B
C
D
E
F
Figure 7.12: Microfossil photos of Zizania sp. observed in carbonized food
residue, soil and stone tools (A, C, D, E, and F) Big Rice site. (B) Windy
Bead site.
146
A
B
A
Figure 7.13: Z.mays ssp. mays. starch granules observed in carbonized
food residue, soil and stone tools photographed under XPL and PPL light.
(A, B) Windy Bead site.
A
B
Figure 7.14: Possible Phaseolus vulgaris starch granule from a Big Rice
soil sample. (A) XPL light, (B) PPL light.
147
A
B
C
D
E
F
Figure 7.15: Possible Cucurbita sp. starch granules taken under XPL and
PL light. (A, B) XPL and PL light photos taken from a Big Rice site soil
sample. (C, D) Starch granules are from carbonized food residue
originating from the Saga Island site. (E, F) Starch identified in carbonized
food residue from the Windy Bead site.
148
A
B
Figure 7.16: (A) Squash seed collected during excavation at the Big
Rice site (Peter and Motivans 1985). (B) Wild rice seed portion
collected during excavation of the Big Rice site.
7.12 GRASS PHYTOLITHS
A large percentage of recovered plant microfossils were comprised of C3 and
C4 grass species. It is difficult to interpret C3 and C4 plant presences from soil
samples because of possible contamination depending on how far below the surface
depth the samples were collected. The probability of cross-contamination is lower
with lithics because residue deposited during milling becomes encased in micro-
149
fissures on grinding stones (Torrence and Barton 2006). However, contamination
does occur with lithic residue, and any contamination can be assessed by
comparing the lithic dry brush, wet brush and sonication samples. The phytolith
distribution for the No Beard grinding stone is characterized by high amounts of
Pooid types, much smaller amounts of Chloridoid and low counts of Panicoid
phytoliths (Figure 7.9).
The sonicated samples contain lower amounts of grass
phytoliths than the dry and wet brush overall, although the sonicated sample from
the working edge has an even lower percentage of grass phytoliths than the other
sonicated sample. Both sonicated samples contain a higher percentage of diatoms
than the dry and wet brush. Similarly, the grinding stone from the Windy Bead site
had a lower count of recovered grass phytoliths, and large quantities of unknown
starch (Figure 7.9). Carbonized food residue provides an even less likely chance of
cross-contamination because exterior residue becomes encrusted on ceramics,
protecting underlying residue layers.
Overall, the carbonized ceramic samples
yielded higher quantities of Festucoid or Pooid phytoliths, although Chloridoid
phytoliths were present, as well as Panicoid phytoliths in very low amounts.
The extracted residue from all nine sites varied in size and resulting
microfossils. The ceramics had the lowest samples sizes, while the sediment had
the highest sample size.
Many unknown phytoliths and starch granules were
recovered, suggesting future work to develop extensive comparative keys.
Diagnostic rondel phytoliths, and starch granules of Zea mays ssp. mays were
identified, along with Zizania rondel phytoliths and possible Phaseolus vulgaris and
possible Cucurbita starch.
150
8.0 INTERPRETATIONS AND DISCUSSION
8.1 INTRODUCTION
This chapter is divided into two sections.
The first section discusses
domesticated plant evidence recovered from the archaeological samples, and this is
applied to the broader context of archaeobotanical literature. The second section
reviews lines of evidence indicating whether trade, and/or local horticultural
production of domesticated plants offer the most viable explanation for the
microfossils found in Brainerd and Laurel components analyzed for this thesis.
Interpretations and re-assessments of the floral component at archaeological sites in
Northern Minnesota and peripheral locations tie together these two sections.
Phytolith and starch content from all nine sites demonstrates a reliance on various
floral species from the Northern Minnesota area, specifically maize, wild rice, and
possibly beans, and squash. Extensive research on comparative material from the
region, which is beyond the scope of this project, would allow specific species of
wild plants to be better identified.
One of the objectives of this study is to reinterpret floral aspects of paleodiet at
the nine sites in Northern Minnesota. Based on just the microfossils, there is no way
to determine the relative importance of specific plants identified in my samples, as
this would stretch the data. The literature suggests that over the course of the
Middle Woodland and into the Late Woodland, the reliance on cultivated plants may
have increased (Fritz 2011; Scarry and Scarry 2005; Syms et al. 2013).
This
increase could have been caused by expanding populations and group networks,
which may have resulted in the trade of domesticated plants.
151
It is unlikely that the recovered microfossils are due to cross-contamination for
a number of reasons. With ceramic samples, only the outside residue layer is
exposed to surrounding sediment. The sediment is unlikely to have any modern
contamination because starch does not move through soil at a great distance and
some samples are located quite far below surface level (Haslam 2006). A ceramic
sherd from the Saga Island site is in Level four, 20-25 centimetres below surface
level. Groundstones contain microfissures that can trap plant remains and preserve
them (Barton 2007; Zarrillo and Kooyman 2006). With the three different types of
lithic extractions, wet-brush, dry-brush and sonication, it is unlikely that all three
would yield modern contamination.
Contamination can occur in the field from
handling the artifacts after being in contact with modern food residue. However, it is
improbable that cross-contamination from modern food residue affected all 33
ceramic and lithic samples used in this study. In addition, maize is the most likely
source of modern contamination (Crowther 2014).
In this study, two lines of
evidence maize starch granules and phytoliths were recovered, greatly decreasing
any possibility of modern contamination.
The two lines of evidence decrease
contamination risks, because in most cases modern contamination is recovered as
either an abundance of maize starch, or maize phytoliths, but not both.
As discussed in Chapter 4, section 4.3.1, Brainerd ware dates are contested
because of the early dates returned from samples obtained throughout Minnesota.
Determining whether or not some of these dates are correct is not for the purposes
of this thesis. For this study, the few dates obtained at the Northern Minnesota sites
will not be used for interpretation of the cultigens recovered from plant microfossils.
152
Instead, the conventional Middle Woodland period dates of 100 BC to 500 AD will
be used to interpret the timing of domesticated and native cultigen use.
8.2 MAIZE
Both carbonized and non-carbonized food residue from ceramics, lithics and
soil samples yielded maize rondel phytoliths and starch granules (Figure 7.4, 7.6,
7.8). The Third River Borrow Pit, Big Rice, Windy Bead, Winnie Cottages, Saga
Island and No Beard sites had samples that tested positive for Z. mays ssp. mays
rondel phytoliths, and/or starch granules. It is worth noting that the sites testing
negative for Z. mays ssp. mays (Kyleleen’s Bent Pine, Kyleleen’s Tall Pine, and Lost
Lake) had samples with very thin carbonized residue and low sample weights.
Therefore, it is possible that the absence of maize remains in these samples was an
effect of small sample size. With the exception of three sites, all sites tested positive
for maize indicating widespread consumption across the study area.
While
important, it is difficult to assess the importance of this plant to Laurel and Brainerd
diet based on residue data alone.
The archaeological literature suggests that by the Late Woodland Period,
maize began to play an increasingly large dietary role across the Upper Great Lakes
and elsewhere (Boyd and Surette 2010; Lovis et al. 2001; Mt. Pleasant 2006; Scarry
2005). Isotope studies of human skeletal remains in the Eastern Woodlands by
Vogel and Van der Merwe (1977) and Katzenberg et al. (1995) show an increase of
maize consumption during the Woodland Period.
Gibbon and Hohman-Caine
(1980) argue that during the Middle to Late Woodland transition in Eastern
Minnesota, populations switched from diffuse to local subsistence patterns. This
153
means that populations began to settle in areas for longer periods of time and
narrow their subsistence base to include plant and animal species close to the area
where they chose to settle. No previous micro-botanical evidence of cultigens has
been recovered from Middle Woodland sites in eastern Minnesota, indicating that at
this earlier time the subsistence reflected a “diffuse hunting-gathering strategy”
(Gibbon and Caine 1980).
This is the basis of the assumption that Aboriginal
populations in Minnesota did not farm maize until Mississippian settlements moved
into southern portions of the State around 800 AD to 1200 AD, well outside of the
Middle Woodland time range (Gibbon and Caine 1980).
However, a lack of
evidence for maize cultivation during the Early to Middle Woodland does not
necessary mean that cultigens were not consumed.
Indeed, as this study and
others (Boyd and Surette 2010; Hart et al. 2003; Katzenberg et al. 1995) show,
cultigens were widely consumed during the Middle Woodland period, from New York
State to Northwestern Ontario.
There is some uncertainty about relying on macrobotanical remains to confirm
the presence/absence of maize; however, both the macrobotanicals and
microfossils provide important data about the timing and dispersal of maize.
Occasional recoveries of maize and other plant macroremains in Early and Middle
Woodland components in the Midwest have been interpreted by some as the result
of postdepositional intrusion of later plant remains into these earlier contexts (Hart et
al. 2003). This has caused some uncertainty in the timing of maize consumption
based on macroremains. Macrobotanical remains deteriorate over time at a faster
rate than plant microfossils. This leads to further uncertainty in the timing and
154
emergence of maize cultivation because the older the site, the less likely plant
remains will be recovered (Hart et al. 2003). In addition, an excavation methodology
that doesn’t include the routine use of flotation will most likely not lead to the
recovery of macrobotanicals. Archaeological components dating before the Late
Woodland period generally do not have well preserved identifiable macroremains.
This has formed the idea that maize cultivation doesn’t appear until the Late
Woodland (Crawford 2001; Lovis et al. 2001), and several authors have suggested
that it may have been used non-intensively before this time (Adovasio and Johnson
1981; Crawford et al. 1997). The Meadowcroft Rockshelter in Pennsylvania
contained maize cob fragments associated with charcoal deposits dating to 522 BC
and 473 BC (Adovasio and Johnson 1981; Hart 2008). The Grand Banks site in
Southern Ontario contained maize fragments directly dated to 1861-1416 BP 363
AD and 362 AD (Crawford et al. 1997; Hart 2008). The earliest maize
macrobotanical evidence in New York State dates from AD 119 to AD 718–880
(Hart et al. 2007). Maize began to increase in importance in Southern Ontario
during the Late Woodland, which is evident with Princess Point archaeobotanical
remains (Crawford 2011). The dates obtained from macrobotanical remains
suggest that maize was widespread in the Upper Midwest and Northeastern North
America, and towards the Late Woodland, there is an increase of maize reliance
and dispersal (Adovasio and Johnson 1981; Crawford et al. 1997; Crawford 2011 ;
Hart et al. 2007; Hart 2008).
Microfossil analysis is changing past interpretations of geographic and
temporal ranges of maize cultivation and its dietary importance to Woodland
155
populations. Microfossil data is also affecting the known ranges of the Three Sisters
(maize, common bean and squash) when utilized together. Boyd and Surette (2010)
used microfossil analysis to date maize consumption in Northwestern Ontario to the
Middle Woodland (ca. 500 AD). A multi-proxy approach using AMS dating, phytolith
and starch analysis and stable carbon isotope analysis dated maize to the seventh
century AD in New York State (Hart et al. 2003). In addition, results from this study
confirm the presence, processing, and cooking of maize during the Middle
Woodland period in Northern Minnesota. This challenges the previous assumption
of a sole reliance on native flora in northern regions of Minnesota during the Middle
to early Late Woodland periods (Gibbon and Hohman-Caine 1980; Mason 2001).
8.3 WILD RICE
The combination of plant microfossils, macrobotanicals and ethnographic
accounts suggest that wild rice was an important resource for populations in the
Great Lakes-St. Lawrence area. Wild rice rondel phytoliths were recovered from the
Big Rice, Saga Island, Lost Lake, No Beard and Windy Bead sites (Figures 7.4, 7.6,
and 7.8). The sites where wild rice tested negative are the Third River Borrow Pit,
Kyleleen’s Bent Pine, Kyleleen’s Tall Pine and Winnie Cottages sites. The results
may be a product of microfossil analysis limitations, as wild rice rondel phytoliths
can be morphologically varied, however; for this study, only four-spiked rondels with
indented bases were used for positive identification of wild rice. The lack of support
for wild rice in samples from Northern Minnesota is surprising considering a wide
range of evidence suggesting intensive wild rice harvesting periods.
This wide
range
of
evidence
includes macrobotanical remains
and
156
ethnographic records. Wild rice macrobotanical remains were recovered from the
Big Rice site by other analysts (Valppu and Rapp 2000). A number of studies have
also demonstrated the antiquity of wild rice in Minnesota lakes (Drewes 2008;
McAndrews 1969: Vennum 1998; Yost and Blinnikov 2011).
Wild rice was an
important food source for Aboriginal people in this region as summarized in Chapter
four, section 4.5. Some scholars have observed, furthermore, that wild rice and
maize are complementary as both plants are high in carbohydrates, protein, iron,
and B vitamins (Hart et al. 2003; Keane 1997; Vennum 1998). Another link between
maize and wild rice consumption is the spiritual importance of both crops to
Aboriginal populations (Wilson 1917; Vennum 1998). Using microfossil evidence,
Thompson et al. (1994), and Valpuu and Rapp (2000) identified wild rice in
Minnesota samples dating to 2,000 years.
Vennum (1998) also provides
ethnographic accounts of wild rice being harvested and consumed by Ojibway
peoples in Northern Minnesota.
Wild rice, or manoomin, meaning ‘good berry’ or ‘good seed’ in Ojibway, was
used ceremonially, as well as non-ceremonially.
Historical and ethnographic
accounts supplied by Vennum (1998) place wild rice in Aitkin County Minnesota, Big
Rice Lake in the Mille Lacs area, Mud Lake in Northern Minnesota and Neds Lake in
Anoka County, although it was probably present in many of Minnesota’s shallow
lakes and rivers. Wild rice has also been documented in Wisconsin, Michigan,
Southern Ontario and Iowa. The Oneota in Wisconsin and the Huron in Southern
Ontario harvested wild rice. Over 12 provinces and states report carbonized wild
rice remains and prehistoric wild rice harvests are evident in South Dakota,
157
Tennessee, Alabama, Kentucky, Ohio, Pennsylvania, Ontario, and Illinois (Vennum
1998). In the Northeast United States, wild rice harvests may have begun around
600 BC and the productivity of wild rice stands would have fluctuated with climate
change (Crawford 2011). Yost and Blinnikov (2011) recovered wild rice (Zizania
palustris sp.) in Northern Minnesota bordering the Laurentian Mixed Forest and the
Eastern Broadleaf Forest ecosystem (Arrigada 2006).
In addition to microfossil
studies and ethnographies, Drewes’ (2008) summary of ecological reports states
that Minnesota has 907 lakes with wild rice, which are located all over the state but
tend to concentrate in Northern Minnesota around the northern portion of the
Eastern Broadleaf Forest. Laurel populations also consumed wild rice in the Boreal
Forests of Northwestern Ontario (Boyd et al. 2010; Surette 2008 ). The presence of
wild rice in Northern Minnesota and adjacent Northwestern Ontario, suggests that
wild rice was widespread and of great dietary importance to Northern Minnesotan
populations during the Middle Woodland.
8.4 SQUASH AND COMMON BEAN
Possible squash (Cucurbita sp.) and common bean (Phaseolus vulgaris) were
recovered in low amounts from four sites. Squash is difficult to identify due to the
native plant confusers, so for the purposes of this study squash is only a possibility.
Common bean is also only a possibility because, although the starch meets most of
the diagnostic characteristics, it is not of the right size proportions for bean. The
Saga Island, No Beard, Windy Bead and the Big Rice sites contain evidence of
possible squash starch granules, and the Big Rice site has one squash seed
originating from collected macrobotanical remains (Peters and Motivans 1983). The
158
Big Rice, No Beard and Lost Lake sites also have samples with recovered possible
Phaseolus vulgaris starch granules.
As discussed above, Cucurbita sp. starch is similar to starch from a variety of
wild plants making it difficult to discern; therefore, the starch granules found were
identified as ‘possible Cucurbita sp.’ Squash phytoliths were not present in any of
the ceramic, soil, or lithic samples. Squash phytoliths are only produced in the rind
and stem, which are inedible plant portions, so it is unlikely that they would have
been found in carbonized residue (Bozarth 1987). If the possible squash starch is
actually squash, this may signify trade as opposed to small-scale agriculture
because phytolith evidence is expected if squash was being cultivated. Although
the squash seed at the Big Rice site supports small-scale agriculture, the minimal
stratigraphic separation between occupations makes it impossible to place the
squash seed temporally in the Middle Woodland Period. Squash seeds were also
recovered from Late Archaic, 580 BC, deposits in Southeastern Minnesota and
trade has been suggested with groups along the Upper Mississippi River Valley to
obtain domesticated squash (Perkl 1998). In the central Mississippi River Valley,
some of the earliest squash macrobotanical remains date from AD 434–613 to 681–
889 (Hart 2007a). The earliest evidence for squash in the eastern United States is
from seeds in Illinois and Tennessee dating around 5050 BC (Hart 2008; Smith
1992). In Maine, macrobotanical remains of a squash rind dated around 4767-4345
BC (Hart 2008; Petersen and Asch Sidell 1996). Squash rinds from the Plains Nebo
Hill site were AMS dated to the Late Archaic (2169 BC and 2200 BC) (Adair and
Drass 2011). Cucurbita pepo rinds and seeds have been recovered from more Late
159
Archaic than Early Woodland contexts in the Eastern Woodlands (Adair and Drass
2011).
Archaeological evidence suggest a hiatus in which Archaic populations
intensely cultivated cucurbits. This activity may have lessened towards the Early
Woodland and resumed during the Middle Woodland (Pauketat 2004). Cucurbita
pepo may be one of the oldest plants cultivated in the Eastern Woodlands (Smith
2011), and towards the Late Woodland Period, the significance of this plant may
have increased with development of the Three Sisters cultivation system.
Possible evidence of common bean (Phaseolus vulgaris) was present at the
Big Rice site, where maize and possibly squash was also identified from various
forms of microfossil evidence. The lack of bean phytoliths in food residue is not
surprising because bean phytoliths are only produced in bean pods and only the
seeds may have been present at sites (Lints 2012).
Based on these lines of
evidence, it is possible that the Three Sisters crop system was in place by the
Middle Woodland Period; however, this view is speculative until further analysis
connects the three major economic crops at Northern Minnesota sites.
The archaeological literature suggests that beans did not arrive until the Late
Woodland in the Midwest (Fritz 2011; Hart and Scarry 1999). Reasons why it may
have been the last of the Three Sisters to arrive could have been the result of
traditional trade networks and/or a conscious choice.
Beans are also more
susceptible to spring frost and require a longer growing season, making them less
viable in more northerly settings (Boyd et al. 2014). Despite the limitations for bean
cultivation and consumption in more northerly locations, results in this thesis and
those collected by Lints (2012) show that beans predate the Late Woodland in some
160
locations. Beans are difficult to identify at sites because they are not charred as
much as other cultigens, affecting how they preserve compared to other plant
remains.
The poor preservation of beans at locations where organic remains
already have a low survival rate may be affecting the perceived antiquity of bean in
the Midwest. Beans have been dated to 1250 to 1300 AD in the central Mississippi
Valley (Fritz 2011) and 1200 AD in the Northeastern Woodlands (Hart et al. 2002).
In addition, macrobotanical evidence for bean consumption has been recorded at
the Round Top site in New York dating to 1273-1400 AD (Hart 2008). Over 24 sites
with 35 Phaseolus vulgaris remains have been dated to pre-1300 AD in the
Northeast United States. In the eastern Central Plains bean macroremains were
AMS dated to around 1100 AD (Adair n.d). Common bean appears to arrive last to
regions with the Three Sisters agricultural system. They were consumed earlier in
the Eastern Woodlands, and Central Plains, and are less utilized less in the
Mississippi Valley, where macroremains yield younger dates.
161
Figure 8.1: Areas of positive plant identification based on microfossil data collected
from Middle Woodland occupations in Northern Minnesota sites.
From the microfossil results obtained in this study, it is clear that Brainerd and
Laurel populations consumed a wide array of plant species, both cultivated and wild.
Greaves and Kramer (2014) outline mixed economies and the role of cultivated and
native plants in the hunter-gatherer diet. Their research has shown that cultivated
plants offer similar nutritional and foraging values to wild plants, and the reliance on
both could be a sign of diversified diets with fall back strategies. Greaves and
Kramer (2014) use the term fall back strategies to describe foods that may be used
162
as back-up in unpredicted times of food scarcity when the main sources of nutrition
are missing. The authors comment that cultivated foods with no nutritional, or taste
advantage, may be used as fall-back strategies. In addition, cultigens were around
thousands of years prior to the advent of agriculture in some areas.
The
replacement of wild foods by cultivated ones varies, and this is determined by
environments with the possibility for agriculture (Greaves and Kramer 2014). The
populations inhabiting Northern Minnesota may have used mixed economic
strategies such as the ones explained by Greaves and Kramer (2014). Wild rice
was a possible main food source that was supplemented by fall-back cultigens such
as maize, beans, and squash. The results from this study may be another classic
example of archaeological data indicating a mix of food production and foraging
economies, with low-level use of cultigens and foraging (Greaves and Kramer
2014).
8.5 CULTIGEN ACQUISITION
Determining cultigen acquisition is complex and involves looking at the
archaeological record in terms of trade versus farming. Based on my results there
are no conventional signs of farming at the Minnesota sites.
Due to a lack of
evidence for farming, this section will present a series of scenarios supporting trade
with surrounding horticultural groups. The first scenario describes possible trade
with groups in Southern Minnesota. Scenario two discusses trade during the Middle
Woodland period with the rise of Hopewellian exchange networks in the Central
Mississippi Valley (Crawford 1997; Fritz 2000).
The third scenario touches on
163
Oneota connections with groups in Southern Minnesota and evidence of
macrobotanical remains from Oneota sites. The last scenario covers possible trade
connections with groups in Northwestern Ontario and the Northeastern Plains.
Conventional signs of farming can include agricultural tools, macrobotanicals,
and microfossil evidence. One of the clear indicators of maize farming is pollen,
which was not recovered in any of the nine sites. Pollen from Z. mays ssp. mays is
extremely rare, preserves in lake and wetland environments and does not travel far
due to its relatively large size and high specific density (Fearn and Liu 1995). Since
maize pollen is extremely rare in cores and food residue contexts, the case for
maize agriculture is difficult to make; however, the rarity of maize pollen might
suggest that although no evidence was collected, that does not mean it was not
there. The relative infancy of micro-botanical studies means that further work may
build sample sizes and the recovery rate of maize pollen. Despite the rarity of maize
pollen, it does not travel far from the original source. Therefore, the absence of
maize pollen opposes any argument of small-scale maize agriculture during the
Middle Woodland period in Northern Minnesota.
Maize starch granules and wavy-top rondel phytoliths were identified from
residue adhering to two grinding stones. These stone tools were used specifically for
grinding, with one tool having a visibly worked edge (Figure 7.2; Figure 7.3). The
presence of maize starch granules on both tools from the Windy Bead and No Beard
sites, and a rondel phytolith from the No Beard site, may suggest maize processing
at both locations. Although maize processing is a possibility, it is difficult to interpret
whether the maize was obtained through trade or locally grown. In either situation,
164
the maize is processed in a similar way. Whether maize is obtained locally or
through trade, the kernels would still need to be milled so the presence of
microfossils on the stone tools does add any perspectives to the argument of trade,
or small-scale agriculture.
Wilson’s (1917) ethnographic account of Buffalo Bird Woman, a Hidatsa
agriculturalist, provides insight into how maize was produced and prepared for
storage and trade. When the corn ripened in the fields it was husked by the entire
community, a process that lasted about ten days. Corn was strung together and
dried on drying stages, and often they were braided to prevent the wind from
shelling out the dry kernels. After the corn had dried, the grain was removed from
the cob by threshing. Historically there were many goods flowing throughout the
Upper Great Lakes, one example is trade connections between the Sioux and
Hidatsa. The Sioux and others sought out the agricultural goods of the Hidatsa
villages and were known to trade one tanned buffalo robe for a string of braided corn
(Wilson 1917). Evidently, corn kernels were removed from the cob prior to trade
and storage, a practise that would have decreased the number of rondel phytoliths
adhering to the final product.
Transporting corn kernels is much easier than
transporting the entire cob so it is likely that Z. mays ssp. mays microfossil counts
would be much lower at locations where populations are receiving agricultural
goods, rather than producing them. This strengthens the case for trade with external
groups, although local agriculture should not be ruled out based on the plant
microfossils recovered in this thesis.
The Cucurbita sp. seed found at the Big Rice site is consistent with the
165
identification of the squash starch, although the seed dates are uncertain. Phytoliths
produced in the squash rind are typically found at locations where production is
taking place (Bozarth 1987). The lack of Cucurbita sp. phytoliths in the samples,
and other indicators of local cultivation may support trade up the Mississippi
corridor, as opposed to small-scale cultivation (Perkl 1998).
The Headwater Lakes Region, which is where a few of the nine study sites are
located, is heavily forested with numerous lacustrine sources (Gibbon 1994).
Gibbon (1994) argued that this environment would have been unsuitable for maize
horticulture, and instead locals relied on the extensive wild rice stands.
The
environment is unsuited for maize agriculture because of the relief and heavily
forested area. Wild rice parching pits were excavated at the Big Rice site and
numerous wild rice macrobotanical remains were recovered (Valppu and Rapp
2000). The topography is rocky at most sites and some of the modern vegetation
includes young balsam poplar, willow, spruce, jack pine and a variety of berry
species. The Big Rice, Third River Borrow Pit, Windy Bead and No Beard are some
of the larger sites considered in this study. Most of the sites are described as semipermanent occupations, where populations would settle to take advantage of
seasonal resources like fish, or wild rice. Although the nine sites in this study are
not in areas well suited for agriculture, small-scale seasonal cultivation should not
be ruled out completely. Instead, conventional archaeological evidence recovered
in the future will pin point when agricultural procurement of the Three Sisters began
during the Woodland Period in Northern Minnesota.
The described scenarios in this section outline connections from surrounding
166
cultural groups to Northern Minnesota, which may explain cultigen acquisition in
small, northerly sites. There are a number of cultural connections evident from
morphologically similar technologies and burial practises throughout the Midwest
United States.
Another aspect of these interactions includes the trade of
domesticated crops such as the Three Sisters. The first scenario outlined in this
section describes the trade of cultivated plant species from Southern to Northern
Minnesota using the river systems of the Upper Mississippi River Valley (Perkl 1998;
Crawford 1997). The Upper Mississippi Valley is located between the headwaters of
the Mississippi River in Lake Itasca and the junction of the Mississippi and Iowa
Rivers (Gibbon 1994).
An imaginary line separates the wild rice harvesters of
Northern Minnesota and the gardeners of Southern Minnesota and Iowa (Gibbon
1994).
Southern Minnesota has a milder climate than Northern Minnesota and
during the Middle Woodland gardening was fully implemented in Southern
Minnesota (Gibbon 2012). Towards the Late Woodland, agricultural procurement
began to increase, as seen in the recovery of carbonized maize from Plains Village
sites such as Oakwood Lakes, southwestern Minnesota (Anfinson 1997). The next
two scenarios describe the possible exchange of domesticated cultigens from the
Mississippi Valley and the Midwest United States to populations in southern
Minnesota.
Trade or cultural diffusion was thought to have occurred from Hopewellian
centers, several of which are located in the Mississippi Valley (Mason 1970). This
would have made trade with Northern groups in the Upper Mississippi Valley
feasible through the lake and river systems of Minnesota. By 550 BC, the Three
167
Sisters were established in the Mississippi Valley. Zea mays ssp. mays, Phaseolus
vulgaris, and Cucurbita sp. were brought in from the Southwest United States where
cutivation began by 2000 or 1500 BC. The Mississippi Valley domesticates later
diffused into the Northeastern Woodlands (Lockard 2008).
A Middle Woodland
period Hopewellian center was located in northwestern Missouri and this is
considered the most westerly example of Hopewellian influence in North America.
Large villages were situated along Missouri River tributaries and radiocarbon dates
suggest the establishment of Kansa City Hopewell sites around AD 1, lasting until
AD 500 (Johnson and Johnson 1998). Maize and squash were confirmed at these
Hopewellian sites around AD 250 (Adair 1988). Maize cob fragments, cupules and
kernels were recovered from Middle Woodland Kansa City Hopewell sites and Adair
and Drass (2011) comment that this may indicate the northward spread of maize
from the Central Plains.
Cucurbita pepo has been directly AMS dated from
Hopewellian occupations during the Middle Woodland (Adair and Drass 2011).
Interaction with eastern Hopewellian centers in the Illinois River Valley is suggested
by similarities in projectile-point and ceramic morphology (Johnson 1998). The large
amount of macrobotanical evidence from Hopewellian sites confirms how
widespread maize cultivation was in the Mississippi Valley, and how easily
exchanged it was with northern groups.
Oneota components are present at a number of Late Woodland sites in
Southwestern Minnesota, and horticulture was widespread in Oneota and related
Midwestern cultures (Anfinson 1997). Oneota is referred to as a culture that links
the East with the West and is located throughout the Prairie Peninsula. Oneota
168
peoples are defined by maize cultivation primarily, although there is also evidence of
squash, sunflower, beans and native plants at Oneota sites (Egan-Bruhy 2013;
Henning 1998).
During the Late Woodland period, the Blue Earth Region in
Southern Minnesota reveals evidence of Oneota burial practises (Henning 1998).
Again, this reflects the northwards movement of plant cultigens from southern
groups up the waterways of Minnesota.
Other speculation supporting how Brainerd and Laurel populations may have
obtained domesticated plants includes interactions with populations on the
northeastern Plains during the Middle Woodland period (Figure 8.2).
Figure 8.2: Avonlea Complex distribution (Hamilton et al. 2011; Department of
Economics and Geography, Hofstra University).
Laurel and Avonlea populations may have traded for food resources, evident
from cultural connections between the two groups (Meyer and Hamilton 1994).
169
Laurel and Avonlea groups were contemporaneous in some areas and may have
had
extensive
trading
networks
between
temperate
and
subarctic/boreal
environments (Meyer and Hamilton 1994; Meyer and Walde 2009).
Laurel
populations are known for wild rice cultivation (Valppu and Rapp 2000) and the
results from this study and Boyd and Surette (2010) demonstrate a subsistence
economy reliant upon maize, wild rice and possible squash and beans. Boyd and
Surette’s (2010) study of Laurel food residue in northwestern Ontario proposed
trade, as opposed to small-scale agriculture, as an explanation for the presence of
maize in the most northerly Woodland sites.
Lints (2012) study found large
quantities of bean starch in most Avonlea sites tested on the Northern Plains, and it
is speculated that trade might have occurred between contemporaneous Laurel and
Avonlea populations. Ceramics with a mix of Avonlea and Laurel traits were also
collected at the Gravel Pit site in Saskatchewan demonstrating close cultural
connections between the two groups (Meyer and Walde 2009).
An important aspect of understanding cultural connections and the possible
trade of cultivated plants is to acknowledge parallels between Brainerd, Laurel, and
Avonlea ceramic technologies.
When the production of Brainerd ware ceased,
Avonlea pottery emerged in the second or third century AD, appearing in the
parklands and prairie border northwest of Brainerd sites (Hohman-Caine and Goltz
1995). However, these wares may not be sequential at all and some unsurveyed
areas dating to the same time may contain both wares, suggesting that Avonlea and
Brainerd were contemporaneous (Hohman-Caine and Goltz 1995).
It is hypothesized that parallel-grooved and net-impressed ceramics associated
170
with Avonlea groups originated from southern localities, such as Manitoba and
Minnesota (Norris 2007). Avonlea sites on the north-eastern Plains have higher
counts of maize and beans, than Laurel sites in northwestern Ontario (Lints 2012).
Wild rice may have been traded with Avonlea populations for maize and/or beans
(Boyd et al. 2014) since both plants are lightweight, easily stored and transported
(Lints 2012). In addition, Avonlea populations may have intercepted wild rice stands
in the southern Boreal Forest at some times of the year (Boyd et al. 2014). Food
residue results collected in Northern Minnesota, northwestern Ontario, and the
northeastern Plains shows the widespread consumption, and in some locations
small-scale agriculture of the Three Sisters, in addition to wild rice, during the Middle
to Late Woodland in Northcentral North America (Boyd and Surette 2010; Lints
2012).
The results from Middle Woodland sites included in this thesis indicate a
dietary reliance on native plants, possibly including the Three Sisters. However,
there is lack of supporting evidence for agricultural procurement of the Three
Sisters. Connections with southern groups along the Mississippi Corridor and less
direct contact with Avonlea populations on the Northern Plains would make trade
feasible for obtaining crops that were not locally grown, such as maize, beans, and
squash. Wild rice was locally grown and may have been traded as a means of
obtaining non-local floral species and other goods. Botanical remains discussed in
site reports from the study area are limited, typically including only modern flora.
Poor preservation due to acidic soils constricts the available data on plant use, but
alternate food residue analysis methods such as phytolith and starch analyses have
171
redefined archaeobotanical knowledge of the Middle Woodland Period in the
Northern Minnesota area (Figure 8.1).
This redefines how we think about ancient
hunter-gatherers in the context of small northerly sites, which may be subject to a
misinterpretation of subsistence due to the lack of macrobotanical remains. This
study has demonstrated the importance of looking for invisible botanical information,
that could potentially re-define the hunter-gatherer label placed on many small,
northerly sites and re-interpret site data to reveal a mixed economic strategy reliant
on a wide range of domesticated and native cultigens.
172
9.0 CONCLUSIONS
Current research on plant-based subsistence in eastern North America is
continuously pushing back the entry dates of local plants and non-native cultigens
such as the Three Sisters. In addition, new data has reinvented the idea that past
populations of the region were mobile hunter-gatherers during the Archaic and
sedentary horticulturalists during the Woodland Period (Kuehn 1998). Recent
studies in the Eastern Woodlands have pushed back the introduction of native
cultigens to the Late Archaic (Crawford 2011; Yarnell 1993). These cultigens were
not uniformly introduced across the entire region, and may have appeared
sporadically until cultivation intensified during the Late Woodland. New dates for wild
squashes, sunflower, sumpweed and chenopod show that these species were
domesticated before 900 BC, and utilized during the Late Archaic (Fritz 1993;
Yarnell 1993).
The introduction of the Three Sisters (maize, beans and squash) occurred at
different times during the Woodland Period. It is theorized that beans arrived last,
while squash and maize arrived sometime during the Middle to Late Woodland from
the Southwest United States. Current research has also readdressed these models
and pushed back the entry of domesticated cultigens in the Boreal Forest, Northern
Great Plains, and Great Lakes (Boyd et al. 2014; Mulholland 1993; Perkl 1998).
Ethnographic accounts after European contact record intense agricultural activity in
relation to these crops and extensive trade networks (Adair and Drass 2011; Wilson
1917). Evidently, the presence of the Three Sisters drastically affected the diets and
subsistence base of Woodland populations, fully transforming societies from
173
foragers of the Early Woodland to agriculturalists of the Late Woodland (Hamilton et
al. 2011; Scarry 2005).
Not much research on early subsistence strategies had been completed in
Northern Minnesota, due to small site sizes, poor organic preservation, and other
factors. Soils in Northern Minnesota prevent the preservation of marcobotanicals
and other organic remains, causing a low recovery rate of plant remains and fauna.
Phytolith and starch analysis has confirmed the consumption of these cultigens in
some areas of Northern Minnesota. The microfossil counts and diagnostic types
found on 31 ceramic samples, two grinding stones and five soil samples from nine
sites dispersed across Northern Minnesota indicate widespread plant use.
The presence of maize phytoliths at half of the sites, along with starch granules
indicates that it was consumed in Northern Minnesota. The earliest evidence of
maize came from the Third River Borrow Pit site on a horizontally-corded Brainerd
sherd dating to 390 BC. Maize was also confirmed on Laurel sherds at the Big Rice
and Saga Island sites, pushing back the entry date for maize to 280 AD. This gives
an accurate representation of when maize cultigens may have been exchanged with
external populations, and predates maize consumption by groups in northwestern
Ontario and the northeastern Plains (Boyd et al. 2014).
The presence of possible squash (Cucurbita sp.) and domesticated bean
(Phaseolus vulgaris) on Laurel material may indicate trade. Squash was thought to
have been carried up the Mississippi Corridor to Southeastern Minnesota during the
Archaic (Perkl 1998). Microfossils identified in the form of phytoliths confirms that
squash was present in Northern Minnesota prior to 500 AD, so it was likely traded
174
with Southern Minnesotan groups who established it during the Late Archaic, or
even earlier in time (Perkl 1998).
Interpretations made with bean starch granules will be considered speculative
until more evidence is found; however, this puts the estimated date for the
introduction of bean to 500 AD from the prior date of 1300 AD (Hart et al. 2002).
Beans are still considered the last of the Three Sister crops to significantly influence
the diets of Woodland populations. It is clear from evidence presented in this study
and others, that beans could have been utilized across much of the subarctic,
Canadian Plains and Northeastern Woodlands prior to AD 500 (Hart et al. 2002;
Lints 2012).
A wide array of native plants and wild rice were present in Brainerd and Laurel
contexts, as shown with this study.
Wild rice has already been confirmed in
Northern Minnesota and around the Lake Superior Basin; evidence of wild rice from
this study readdresses the dietary importance of wild rice to Middle Woodland
populations (Boyd and Surette 2010; Valppu and Rapp 2000; Wilcox 2007; Yost et
al. 2013). The development of local comparative keys from Northern Minnesota
would greatly aid the interpretation and species identification of some of these wild
plants. Ethnographic accounts and macrobotanical remains suggest the reliance on
a wide variety of berries, nuts, wild fruits and vegetables (Marles et al. 2000; Simon
2009).
Previous research describes Brainerd and Laurel populations as huntergatherer/foragers generally subsisting on wild rice, local plants and a wide variety of
mammals. Seasonal rounds would have been made to exploit these resources, and
175
wild rice would often be stored for winter months (Shafer 2003; Wilson 1917). Food
residues analyzed in this study supports this subsistence model and in addition
shows a wider variety of plant consumption than previously assumed. Wild rice has
been present in Northern Minnesota for the last 10,000 years, although the
presence of wild rice would have fluctuated with climate. Recovered Zizania
phytoliths and charred remains indicate local harvesting.
Similar to maize,
domesticated squash was possibly traded; evidence includes starch granules and a
seed found at the Big Rice site. Bean evidence in the form of starch granules
suggests that they were likely traded as no other microfossils were recovered.
Based on the comparison of microfossils from all nine sites investigated in this
study, not all populations subsisted on the same plant species or engaged in their
trade. More research using microfossil analysis should be carried through to aid in
interpreting and defining the extent of trade versus local agriculture at other
Northern Minnesota sites. Wild rice was possibly a key asset for these Middle
Woodland populations to obtain non-native resources such as the Three Sisters.
Brainerd and Laurel populations inhabited Northern Minnesota from roughly
100 BC to 500 AD, representing a time when both domesticated cultigens and wild
rice were paired in a mixed economic approach. The presence of maize and
possibly squash and beans throughout the nine sites in Northern Minnesota has
altered the archaeological discourse in Northern Minnesota by identifying when and
where the Three Sisters began to emerge through trade. In addition, the resulting
interpretations have confirmed the widespread use of wild rice and native plants that
would have been of economic importance before the spread of agriculture during the
176
Late Woodland Period.
This study extends evidence of known plant use in
Northwestern Ontario, the Canadian Plains, north-eastern Plains, and the Eastern
Woodlands by identifying another northerly location of early use of the Three Sisters
paired with wild rice. These data are refined because in most locations, microfossil
analysis has never been completed and a limited knowledge of plant use before
European contact was common. Future research should aim to reconstruct notions
of paleodiet with a multi-proxy approach at sites pre-dating the Woodland Period.
177
REFERENCES
Adair, J. M.
n.d. Defining Central Plains Agriculture: The Arrival of the Common Bean
(Phaseolus vulgaris). Museum of Anthropology, University of Kansas.
Adair, J. M.
1988. Prehistoric Agriculture in the Central Plains. Publications in Anthropology
16, University of Kansas, Lawrence, Kan.
Adair, M. and Drass, R.
2011. Patterns of Plant Use in the Prehistoric Central Plains and Southern
Plains. In B. Smith. ed. The Subsistence Economies of Indigenous North
American Societies. Washington: Smithsonian Institute Scholarly Press, 330352.
Adovasio, J. M. and Johnson, W. C.
1981. The appearance of cultigens in the upper Ohio Valley: A view from
Meadowcroft Rockshelter. Pennsylvania Archaeologist 51(1-2), 63-80.
Ahlegren, F. I. and Ahlegren, E. C.
1965. Effects of Prescribed Burning on Soil Microorganisms in a Minnesota
Jack Pine Forest. Ecology 46(3), 304-310.
Anderson, Y. R., Bradbury, P. J. and Dean, E. W.
1993. Elk Lake in Perspective. In Elk Lake, Minnesota: Evidence for Rapid
Climate Change in the North-Central United States, ed. Bradbury, Platt. J and
Dean, E. Walter,1-6. The Geology Society of America, Inc, Boulder, Colorado.
Anfinson, F. S.
1979. A Handbook of Minnesota Prehistoric Ceramics.
Occasional
Publications in Minnesota Anthropology No. 5. Fort Snelling: Minnesota
Archaeological Society.
1997. Southwestern Minnesota Archaeology: 12,000 Years in the Prairie
Lake Region. Minnesota Historical Society Press.
Archibold, O. W., Weichel, J. B. and Fitzgibbon, E. J.
1985. Factors in the Ecology of Wild Rice in Northern Saskatchewan. The
Canadian Geographer 29(2), 100-112.
Arnott, J. H. and Webb, A. M.
2000. Twinned Raphides of Calcium Oxalate in Grape (Vitis): Implications for
Crystal Stability and Function. International Journal of Plant Sciences 131(1),
133-142.
178
Arrigada, J. E.
2006.
Plant Communities of Lake Ogechie, Mille Lacs Co., MN.
Unpublished Report. Department of Biology. Saint Cloud State University,
St. Cloud, MN. 56301.
Arzigian, C.
2008. Minnesota Statewide Multiple Property Documentation Form for the
Woodland Tradition.
University of Wisconsin-La Crosse: Minnesota
Department of Transportation, Mississippi Valley Archaeology Center.
Babot M. del. P.
2003. Starch Grain Damage as an Indicator of Food Processing. In Phytolith
and Starch Research in the Australian-Pacific-Asian Regions: The State of the
Art. Papers from a conference held at The Australian National University,
August 2001, Cranberry, Australia.
Bartlein, J. P., and Watts A. W.
1993. Paleoclimatic Interpretation of the Elk Lake Pollen Record. In
Bradbury, Platt. J. and Dean, E. Walter. eds. Elk Lake, Minnesota: Evidence
for Rapid Climate Change in the North-Central United States. Boulder,
Colorado: The Geology Society of America Inc, 275-294.
Barton, H.
2007. Starch residues on museum artefacts: implications for determining tool
use. Journal of Archaeological Science 34, 1752-1762.
Barton, H. and Fullagar, R.
2006. Microscopy. In Torrence, R. and Barton, H. eds.
Ancient Starch Research. Walnut Creek: Left Coast Press Inc, 47-52.
Benz F. Bruce.
2006. Maize in the Americas. In John E. Staller, Robert H. Tykot, Bruce F.
Benz. eds. Histories of Maize: Multidisciplinary Approaches to the Prehistory,
Biogeography, Domestication and Evolution of Maize. San Diego and New
York: Elsevier Academic Press, 9-18.
Berman, J. M. and Pearsall, M. D.
2008. A the Crossroads: Starch Grain and Phytolith Analyses in Lucayan
Prehistory. Latin American Antiquity 19(2),181-203.
Bernal, R. G., Tlusty, B. Estevez de Jensen, C. van Berkum, P. Graham, P. H.
2004. Charactersitics of Rhizobia Nodulating Beans in the Central Region of
Minnesota. Canadian Journal of Microbiology 50(12), 1023-1031.
179
Birks, B. J. H.
1976. Late-Wisconsinan Vegetational History at Wolf Creek, Central
Minnesota. Ecological Monographs 46, 395-429.
Bishop, E. H.
2007. Lake Superior: The Ultimate Guide to the Region. Duluth, Minnesota:
Lake Superior Port Cities, Inc.
Blake, M.
2006. Dating the Initial Spread of Zea mays. In John E. Staller, Robert H.
Tykot, Bruce F. Benz. eds. Histories of Maize: Multidisciplinary Approaches to
the Prehistory, Biogeography, Domestication and Evolution of Maize. San
Diego and New York: Elsevier Academic Press, 55-72.
Blitz, J. H.
1988. Adoption of the Bow in Prehistoric North America. North American
Archaeologist 9(2), 123-145.
Boyd, M., Running, IV. G. and Havlom, K.
2003. Paleoecology and Geochronology of Glacial Lake Hind During the
Pleistocene- Holocene Transition: A Context for Folsom Surface Finds on the
Canadian Prairies. Geoarchaeology: An International Journal 18(6), 583-607.
Boyd, M.
2005. Phytoliths as Paleoenvironmental Indicators in a Dune Field on the
Northern Great Plains. Journal of Arid Environments 61, 357-375.
Boyd, M., Surette, C. and Nicholson, B. A.
2006.
Archaeobotanical Evidence of Prehistoric Maize (Zea mays)
Consumption at the Northern Edge of the Great Plains.
Journal of
Archaeological Science 33, 1129-1140.
Boyd, M., Varney, T. Surette, C. and Surette, J.
2008. Reassessing the northern limit of maize consumption in North America:
stable isotope, plant microfossil, and trace element content of carbonized food
residue. Journal of Archaeological Science 35, 2545-2556.
Boyd, M. and Surette, C.
2010. Northernmost Precontact Maize in North America. American Antiquity
75(1), 117-133.
180
Boyd, M., Lints, A. Surette, C. and Hamilton, S.
2014. Wild rice (Zizania spp.), the Three Sisters, and the Woodland Tradition
in Western and Central Canada. In Maria E. Raviele and William A. Lovis.
eds. Reassessing the Timing, Rate and Adoption Trajectories of Domesticate
Use in the Midwest and Great Lakes. Special Paper No. 8. Midwest
Archaeological Conference, Inc.
Bozarth, R. S.
1987. Diagnostic Opal Phytoliths from Rinds of Selected Cucurbita Species.
American Antiquity 52(3), 607-615.
1990. Diagnostic Opal Phytoliths from Pods of Selected Varieties of Common
Beans (Phaseolus vulgaris). American Antiquity 55(1), 98-104.
1992. Classification of opal phytoliths formed in selected dicotyledons native
to the Great Plains. Phytolith Systematics Advances in Archaeological and
Museum Science 1, 193-214.
1993. Maize (Zea mays) Cob Phytoliths from a Central Kansas Great Bend
Aspect Archaeological Site. Plains Anthropologist 38(146), 279-286.
Brad C.
2013. www.Geology.com. “Minnesota Physical Map.” Geoscience News and
Information Geogeology.com N.p., Web. June 17, 2014
Bradury, P. J., Dean, E. W. and Anderson, R. Y.
1993. Holocene Climatic and Limnologic History of the North-Central United
States as Recorded in the Varved Sediments of Elk Lake, Minnesota: A
Synthesis. In Bradbury Platt, J. and Dean E. Walter (eds) Elk Lake Minnesota,
Evidence for Rapid Climate Change in the North-Central States. Special
Paper 276. The Geological Society of America, Inc, 309-329.
Brandzin-Low, V.
1997. The distribution and characteristics of Laurel ceramics from northern
Manitoba: a comparative study. MA Thesis, University of Saskatchewan
Department of Anthropology and Archaeology.
Brown, A. Dwight.
1984. Prospects and Limits of a Phytolith Key for Grasses in the Central
United States. Journal of Archaeological Science 11, 345-368.
Buckler, E. S. and Stevens, N. M.
2005. Maize origins, domestication, and selection. In Motley, T. J., Zerega, N.
and Cross, H. eds. Darwin's Harvest. New York: Columbia University Press,
67-90.
181
Bunzel, M., Allerdings, E. Sinwell, V. Ralph, J. and Steinhart, H.
2002. Cell Wall Hydroxycinnamates in Wild Rice (Zizania aquatica L.)
Insoluble dietary fibre. European Food Research and Technolgy 214, 484488.
Chandler-Ezell, K. and Pearsall, M. D.
2003. Piggyback microfossil processing: joint starch and phytolith sampling
from stone tools. The Phytolitharien 15, 2-8.
Crawford, W. G., Smith, G. D. and Bowyer, E. V.
1997. Dating the Entry of Corn (Zea mays) into the Lower Great Lakes
Region. American Antiquity 62(1), 112-119.
Crawford, W. G.
2011. People and Plant Interactions in the Northeast. In Bruce D. Smith. ed.
The Subsistence Economies of Indigenous North American Societies.
Washington, D.C: Smithsonian Institution Scholarly Press, 503-516.
Crowther, A.
2012. The Differential Survival of Native Starch during Cooking and
Implications for Archaeological Analyses: A Review.
Archaeological
Anthropological Science 4, 221-235.
Crowther, A., Haslam, M. Oakden, N. Walde, D. and Mercader, J.
2014. Documenting Contamination in Ancient Starch Laboratories. Journal of
Archaeological Science 49, 90-104.
Cummings, S. L.
2007. Manual for Pollen, Phytolith, Starch, FTIR, AMS Radiocarbon, and
Macrofloral Sampling. Golden, Colorado: Paleo Research Institute.
Dawson, K. C. A.
1981. The Wabinosh River Site and the Laurel Tradition in
Northwestern Ontario. Ontario Archaeology 36, 3-46.
1987. Northern Ontario and the Early Contact Period: The Northern Ojibwa
1615 - 1715. Canadian Journal of Archaeology 11, 143-187.
Dean, E. W. and Megard, O. R.
1993. The Varve Chronometer in Elk Late: Record of Climatic Variability and
Evidence for Solar Geomagnetic 14C Climatic Connection. In: Bradbury Platt,
J. and Dean E. Walter (eds) Elk Lake Minnesota, Evidence for Rapid Climate
Change in the North-Central States. Special Paper 276. The Geological
Society of America, Inc, 45-68.
182
Doebley, J.
1984. Maize introgression into teosinte: a reappraisal. Annals of the Missouri
Botanical Garden 71, 1100–1113.
Donaldson, S. W. and Wortner, S.
1995. The Hind Site and the Glacial Kame Burial Complex in Ontario. Ontario
Archaeology 59, 5-95.
Drewes, A. L. D.
2008. Sustaining a Ricing Culture: An Integrated Landscape Approach to
Understanding Harvest, Distribution and Management of Wild Rice (Zizania
Palustris) Across the Upper Great Lakes Region. ProQuest.
Duncan, N. A., Pearsall, D. M. and Benfer, R. A.
2009. Gourd and squash artifacts yield starch grains of feasting foods from
preceramic Peru. Proceedings of the National Academy of Sciences 106(32),
13202-13206.
Egan-Bruhy, C. K.
2013. Ethnicity as Evidenced in Subsistence Patterns of Late Prehistoric,
Upper Great Lakes Populations. Commonwealth Cultural Resources Group,
Inc.
Fearn, M. and Lui, K.
1995. Maize Pollen of 3500 B.P. from Southern Alabama. American Antiquity
60(1), 109-117.
Ferris, N.
1999. Telling tales: Interpretive trends in southern Ontario Late Woodland
archaeology. Ontario Archaeology 68,1-62.
Fiedel, J. S.
2001. What Happened in the Early Woodland? Archaeology of Eastern North
America 29, 101-142.
Fischer, A. and Heinenmeier, J.
2003. Freshwater Reservoir Effect in
Pottery. Radiocarbon 45(3), 449-466.
14
C Dates of Food Residue on
Fitting, E. J.
1969. Settlement Analysis in the Great Lakes Region. Southwestern Journal
of Anthropology. 25(4): 360-377.
183
Fritz, G.
1993. Early and Middle Woodland Paleoethnobotany. In C. Margaret Scarry.
ed. Foraging and Farming in the Eastern Woodland. University Press of
Florida, 39-56.
2011. The Role of “Tropical” Crops in Early North America. In Bruce D. Smith.
ed. The Subsistence Economies of Indigenous North American Societies.
Washington, D.C: Smithsonian Institution Scholarly Press, 503-516.
Gibbon, E. G. and Hohman-Caine, A. C.
1980. The Middle to Late Woodland Transition in Eastern Minnesota.
Midcontinental Journal of Archaeology. 5(1), 57-72.
Gibbon, E. G.
1994. Cultures of the Upper Mississippi River Valley and Adjacent Prairies in
Iowa and Minnesota. In Karl H. Schlesier. ed. Plains Indians, A.D. 500-1500:
The Archaeology of Past Historic Groups. University of Oklahoma Press:
Norman and London, 128-148.
2012a. Upper Rice Lake (21CE04): A Multiple Component Woodland Site in
Clear-water County, Minnesota. The Minnesota Archaeologist 71: 74-88.
2012b. Archaeology of Minnesota: The Prehistory of the Upper Mississippi
River Region. University of Minnesota Press.
Gott, B., Barton, H. Samuel, D. and Torrence, R.
2006. Biology of Starch. In R. Torrence. and H. Barton. eds. Ancient Starch
Research. Walnut Creek: Left Coast Press, 35-47.
Greaves, D. R. and Kramer, L. Karen.
2014. Hunter-gatherer use of Wild Plants and Domesticates: Archaeological
Implications for Mixed Economies before Agricultural Intensification. Journal
of Archaeological Science 41, 263-271.
Haas, J., Creamer, W. Mesía, H. L. Goldstein, D. Reinhard, K. and Rodriguez, V. C.
2013. Evidence for maize (Zea mays) in the Late Archaic (3000–1800 BC) in
the Norte Chico region of Peru. Proceedings of the National Academy of
Sciences 110(13), 4945-4949.
Hamilton, S. and Nicholson, B. A.
2006. Aboriginal Seasonal Subsistence and Land Use on the Northeastern
Plains: Insight from Ethnohistoric Sources. Plains Anthropologist 51(199),
253-280.
184
Hamilton, S., Taylor-Hollings, J. and Norris, D.
2011. Human Ecology of the Canadian Prairie Ecozone ca. 1500 BP:
Diffusion, Migration and Technological Innovation. In B. A. Nicholson. ed.
Human Ecology of the Canadian Prairie Ecozone. Regina: Canada Plains
Research Center, 99-152.
Hardy, K., Blakeney, T. Copeland, L. Kirkham, J. Wrangham, R. Collins, M.
2009. Starch granules, dental calculus and new perspectives on ancient diet.
Journal of Archaeological Science 36(2), 248-255.
Hart, C. T.
2011. Evaluating the Usefulness of Phytoliths and Starch Granules found on
Surface Artifacts. Journal of Archaeological Science 38, 3244-3253.
Hart, P. J. and Scarry, C. M.
1999. The Age of Common Beans (Phaseolus vulgaris) in the Northeastern
United States. American Antiquity 64(4), 653-658.
Hart, J., Asch, D. Scarry, C. and Crawford, G.
2002. The Age of Common Bean in the North Eastern Woodlands. American
Antiquity 76, 377-385.
Hart, J., Thompson, R. and Brumbach, H.
2003. Phytolith Evidence for Early Maize (Zea Mays) in 2003 the Northern
Finger Lakes Region of New York. American Antiquity 68(4), 619-640.
Hart, J. P., Brumbach, H. J., & Lusteck, R.
2007. Extending the phytolith evidence for early maize (Zea mays ssp. mays)
and squash (Cucurbita sp.) in central New York. American Antiquity, 72(3),
563.
Hart, J. P., & Lovis, W. A.
2007a. A multi-regional analysis of AMS and radiometric dates from
carbonized food residues. Midcontinental Journal of Archaeology 32(2), 201261.
2007b. The freshwater reservoir and radiocarbon dates on cooking residues:
Old apparent ages or a single outlier? Comments on Fischer and Heinemeier
(2003). Radiocarbon 49(3), 1403-1410.
2014. A Re-Evaluation of the Reliability of AMS Dates on Pottery Food
Residues from the Late Prehistoric Central Plains of North America: Comment
on Roper (2013). Radiocarbon 56(1), 341-353.
185
Hart, J. P.
2008. Evolving the three sisters: The changing histories of maize, bean, and
squash in New York and the greater Northeast. Current Northeast
Paleoethnobotany II 512, 87-99.
Hart, J. P., Lovis, W. A. Urquhart, G. R. and Reber, E. A.
2013. Modeling Freshwater Reservoir Offsets on Radiocarbon-Dated Charred
Cooking Residues. American Antiquity 78(3), 536-552.
Haslam, M.
2004.
The decomposition of starch grains in soils: implications for
archaeological residue analyses. Journal of Archaeological Science 31(12),
1715-1734.
2006. Potential misidentification of in situ archaeological tool-residues: starch
and conidia. Journal of Archaeological Science 33(1), 114-121.
Henning, R. D.
1998. The Oneota Tradition. In W. Raymond Wood. ed. Archaeology on the
Great Plains. University Press of Kansas, 345-415.
Henry, A. G., Hudson, H. F. and Piperno, D. R.
2009. Changes in starch grain morphologies from cooking. Journal of
Archaeological Science 36(3), 915-922.
Henry, A. G., & Piperno, D. R.
2008. Using plant microfossils from dental calculus to recover human diet: a
case study from Tell al-Raqā'i, Syria. Journal of Archaeological Science, 35(7),
1943-1950.
Henry, A. G., Brooks, A. S. and Piperno, D. R.
2014. Plant foods and the dietary ecology of Neanderthals and early modern
humans. Journal of human evolution 69, 44-54.
Hill, L. C. and James, K. H.
1996. Geo-ecological Dynamics in the Glacial Lakes Duluth-Agassiz Region
(Northern Minnesota) During the Late Pleistocene and Early Holocene.
Current Research in the Pleistocene 13, 119-121.
Hohman-Caine, A. C. and Goltz, E. G.
1995. Brainerd Ware and the Early Woodland Dilemma.
Archaeologist 59, 109-129.
The Minnesota
Hohman-Caine, A. C., Syms, E. Lee. and Goltz, E. Grant.
2012. The Age of Brainerd Ceramics. Minnesota State Historical Society
Contract No. 4107232, Hackensack, Minnesota: Soils Consulting.
186
Holst, Irene., Moreno, J. E. and Piperno, R. D.
2007. Identification of teosinte, maize, and Tripsacum in Mesoamerica by
using pollen, starch grains, and phytoliths. Proceedings of the National
Academy of Sciences 104(45), 17608-17613.
Hoover, H.
1963. The Long-Shadowed Forest. London: Souvenir Press.
Horrocks, M.
2005. A combined procedure for recovering phytoliths and starch residues
from soils, sedimentary deposits and similar materials.
Journal of
archaeological science. 32(8), 1169-1175.
Hsieh, H. M., Swanson, B. G. and Lumpkin, T. A.
2000. Abrasion, Grinding, Cooking and the Composition and Physical
Characteristics of Azuki koshi an. Journal of food processing and
preservation 24(2), 87-106.
Janzen, D. E.
1968. The Naomikong Point Site and the Dimensions of Laurel in the Lake
Superior Region. Anthropological Papers: Museum of Anthropology, University
of Michigan No. 36.
Jenks, A.
1903. The Wild Rice Gatherers of the Upper Great Lakes: A Study in
American Primitive Economics. Man 3, 124-126.
Jenkins, E.
2009. Phytolith taphonomy: a comparison of dry ashing and acid extraction on
the breakdown of conjoined phytoliths formed in Triticum durum. Journal of
Archaeological Science 39(10), 2402-2407.
Johannessen, S.
1993. Food, Dishes, and Society in the Mississippi Valley. In Scarry, M. ed.
Foraging and Farming in the Eastern Woodlands. University Press of Florida,
182-205.
Johnson, M. A.and Johnson, E. A.
1998. The Plains Woodland. In W. Raymond Wood. ed. Archaeology on the
Great Plains. University Press of Kansas, 201-235.
Johnson M. C.
1998. The Coalescent Tradition. In W. Raymond Wood. ed. Archaeology on
the Great Plains. University Press of Kansas, 308-345.
187
Kato, T. A.
1984. Chromosome morphology and the origin of maize and its races.
Evolutionary Biology 17, 219–253.
Katzenberg, M. A., Schwarcz, H. P, Knyf, M. and Melbye, F. J.
1995. Stable Isotope Evidence for Maize Horticulture and Paleodiet in
Southern Ontario, Canada. American Antiquity 60, 335-350.
Keane, K.
1997. Wild Rice (Zizania L. spp.). Root Woman and Dave.
Klimko, 0.
1985. Nipawin Reservoir Study, Volume 7 The Gravel Pit and Eastcott Flat Sites.
Saskatchewan Power Corporation, Saskatchewan.
Knee, M.
N.d. Plants Cells. Ohio State University.
http://www.hcs.ohiostate.edu/hcs300/cell1.htm.
Kuehn, S. R.
1998. Evidence for Late Paleoindian-Early Archaic Subsistence Behaviour in
the Western Great Lakes. American Antiquity 63(3), 457-476.
Lakela, O.
1965. A Flora of Northeastern Minnesota. Minneapolis: University of
Minnesota Press.
Lamb, J. and Loy, T.
2005. Seeing red: the use of Congo Red dye to identify cooked and damaged
starch grains in archaeological residues. Journal of Archaeological
Science 32(10), 1433-1440.
Landals, A.
1995.
The Miniota and Broadview Sites: Two Avonlea Loci on the
Northeastern Plains. Manitoba Archaeological Journal 5(1), 1-27.
Lee E. T.
1954. The First Sheguiandah Expedition, Manitoulin Island, Ontario.
American Antiquity 20(2): 101-111.
Lints, A.
2012. Evidence of Domesticates From Early Ceramics of the Northern Plains:
An Analysis of Avonlea Cultural Materials (AD 300-1000). MA Thesis,
Lakehead University, Department of Anthropology.
188
Liu, L., Field, J. Fullagar, R. Zhao, C. Chen, X. and Yu, J.
2010. A Functional Analysis of Grinding Stones from an Early Holocene Site
at Donghulin, North China. Journal of Archaeological Science 37(10), 26302639.
Liu, L., Bestel, S. Shi, J. Song, Y. and Chen, X.
2013. Paleolithic human exploitation of plant foods during the last glacial
maximum in North China. Proceedings of the National Academy of
Sciences, 110(14), 5380-5385.
Lockard, A. C.
2008. Societies, Networks, and Transitions: A Global History. Houghton
Mifflin Company.
Lovis, A. W., Egan-Bruhy, C. K. Beverly, A. S. and Monaghan, G. W.
2001. Wetlands and Emergent Horticultural Economies on the Upper Great
Lakes: A New Perspective from the Schultz Site. American Antiquity 66(4),
615-632.
Loy, T., Spriggs, M. and Wickler, S.
1992. Direct evidence for human use of plants 28,000 years ago: starch
residues on stone artefacts from the northern Solomon Islands.
Antiquity 66(253), 898-912.
Loy, T.
1994. Methods in the analysis of starch residues on prehistoric stone tools.
Tropical archaeobotany: Applications and new developments, 86-114.
Lu, H. Y., Wu, N. Q. Yang, X. D. Jiang, H. Liu, K. B. and Liu, T. S.
2006. Phytoliths as quantitative indicators for the reconstruction of past
environmental
conditions
in
China
I:
phytolith-based
transfer
functions. Quaternary Science Reviews 25(9), 945-959.
Lugenbeal, E. N.
1976. The Archaeology of the Smith Site: A Study of the Ceramics and
Culture History of Minnesota Laurel and Blackduck. PhD Thesis, University of
Wisconsin, Department of Anthropology.
Madella, M., Lancelotti, C. and García-Granero, J.
2013. Millet microremains—an alternative approach to understand cultivation
and use of critical crops in Prehistory. Archaeological and Anthropological
Sciences, 1-12.
Mallard Club Wild Rice
2014. “The Mallard Club Wild Rice.” The Mallard Club Wild Rice, Minnesota.
www.mallardclubwildrice.com. N.p. Web June 17, 2014.
189
Mason, R.
1970. Hopewell, Middle Woodland, and the Laurel Culture: A Problem in
Archaeological Classification. American Anthropologist 72(4), 802-815.
1981. Great Lakes Archaeology. The Blackburn Press.
Maple Hill Park.
2013. “Glacial Geology and Soils of Maple Hill.” Friends of Maple Hill Park.
www.maplehillpark.org. N.p. Web June 17, 2014.
Marschner, J. Francis.
1974. The Original Vegation of Minnesota. Compiled from U.S. General Land
Office Survey Office Notes. Office of Agricultural Economics. U.S. Department
of Agriculture.
Marles J. R., Clavelle, C. Monteleone, L. Tays, N. and Burns, D.
2000. Aboriginal Plant Use in Canada's Northwest Boreal Forest.
Resources Canada, Canada Forest Services.
Natural
Matsuoka, Y., Vigouroux, Y. Goodman, M. M. Sanchez, J. Buckler, E. and Doebley,
J.
2001. A single domestication for maize shown by multilocus microsatellite
genotyping. Proceedings of the National Academy of Sciences 99(9), 60806084.
McAndrews, J. H.
1969. Paleobotany of a wild rice lake in Minnesota. Canadian Journal of
Botany 47(11), 1671-1679.
1999. Wild Rice and Upland Vegetation History near Lake Mille Lacs,
Minnesota. In Mather, D. and Abel, E. eds. The Lake Onamia- Trunk Highway
169 Data Recovery Project, Mille Lacs County, Minnesota. Maple Groove:
Loucks and Associates, Inc.
Meyer, D. and Hamilton, S.
1994. Neighbours to the North: Peoples of the Boreal Forest.
In K. H. Schlesier. ed. Plains Indians A.D. 500 – 1500.
Oklahoma Press, 96–127.
University of
Meyer, D. and Walde, D.
2009. Rethinking Avonlea: Pottery Wares and Cultural Phase. Plains
Anthropologist 54(209), 49-73.
190
Morgan, Grace R.
1979. An Ecological Study of the Northern Plains as Seen Through the Garrett
Site. University of Regina, Occasional Papers in Anthropology, No.1.
Mt. Pleasant, J.
2006. The Science behind the Three Sisters Mound System: An Agronomic
Assessment of an Indigenous Agricultural System in the Northeast. In John E.
Staller., Robert H. Tykot, and Bruce F. Benz. ed. Histories of Maize:
Multidisciplinary Approaches to the Prehistory, Linguistics, Biogeography,
Domestication, and Evolution of Maize. New York: Academic Press, 529-535.
Mulholland, C. S.
1986. Identification of Plants in a Sediment. In Rovner, Irwin. ed. Plant Opal
Phytolith Analysis in Archaeology and Paleoecology. Occasional Papers 1 of
The Phytolitharien N.C. State University, 123-128.
1993. A Test of Phytolith Analyses at Big Hidatsa, North Dakota. In Pearsall,
D. M. and Piperno, D. R. eds. Current Approaches in Phytolith Analysis:
Applications in Archaeology. Journal of Archaeological Research 21, 175-216.
Mulholland, C. S., Mulholland, L. S, Shafer, R. J. Valppu, H. S. and Rapp, Jr. G.
1996. Evaluation of Two Sites on the Dixon Lake Road/Third River,
Blackduck District, Chippewa National Forest, Itasca County, Minnesota.
University of Minnesota, Duluth: Archaeometry Laboratory.
1997. Mitigation of the Third River Borrow Pit Site (21-IC-0176, 09-03-010355), Blackduck District, Chippewa National Forest, Itasca County,
Minnesota. University of Minnesota, Duluth: Archaeometry Laboratory.
Norris, D.
2007. Presence of Net-impressed and Horizontally Corded Ware in Southern
Manitoba: The Relationship between Rock Lake and Brainerd Ware. MA
Thesis, University of Saskatchewan, Department of Archaeology.
Ollendorf, A.L., Mulholland, C. S., and Rapp, G. Jr.
1988. Phytolith Analysis as a Means of Plant Identification: Arundo donax
and Phragmites communis. Annals of Botany 61, 209-214.
Olympus Microscopy Resource Center.
2012. http://www.olympusmicro.com/ Olympus America Inc.
Ojakangas. J. R. and Charles L. M.
1982. Minnesota’s Geology. Minneapolis: University of Minnesota Press.
Pauketat R. T.
2004. Ancient Cahokia and the Mississippians. Cambridge University Press.
191
Pearsall, M. D.
1982. Phytoliths: Application of a New Phytolith Technique in Archaeology.
American Anthropologist 84(4), 860-871.
Pearsall, D. M., Chandler-Ezell, K., & Chandler-Ezell, A.
2003. Identifying maize in neotropical sediments and soils using cob
phytoliths. Journal of Archaeological Science, 30(5), 611-627.
Pearsall, D. M., Chandler-Ezell, K. and Zeidler, J. A.
2004a. Maize in Ancient Ecuador: Results of Residue Analysis of Stone
Tools from the Real Alto site. Journal of Archaeological Science 31(4), 423442.
Pearsall, D. M., Chandler-Ezell, K. and Chandler-Ezell, A.
2004b. Maize can still be Identified Using Phytoliths: Response to Rovner.
Journal of Archaeological Science 3: 1029: 1038.
Perkl, E. B.
1998. Cucurbita pepo from King Coulee, Southeastern Minnesota. American
Antiquity 63(2), 279-288.
2009.
The Late-Archaic-Early Woodland Transition in Southeastern
Minnesota.
PhD Thesis, University of Minnesota, Department of
Archaeology,
Peters, R. G. and Motivans, A. K.
1983. Evaluation: The Big Rice Site (F.S. #09-09-034). Cultural Resource
Management on the Superior National Forest 1983 Annual Report: Duluth,
MN, USDA Forest Service, edited by G.R. Peters, 280-317
Peterson, J. B. and Asch Sidell, N.
1996. Mid-Holocene Evidence of Cucurbita Sp. from Central Maine. American
Antiquity 61, 685-698.
Peterson, A. L.
1977. A Field Guide to Edible Wild Plants Eastern and Central North
America. New York, New York: Houghton Mifflin Company.
Piperno, D. R.
1988. Phytolith analysis: an archaeological and geological perspective. San
Diego: Academic Press.
2006. Phytoliths: A Comprehensive Guide for Archaeologists
Paleoecologists. Lanham, MD: Rowman AltaMira Press.
and
192
2009. Identifying Crop Plants with Phytoliths (and Starch Grains) in Central
and South America: A Review and an Update of the Evidence. Quaternary
International 193, 146-159.
Piperno, D. R., Weiss, E. Holst, I. and Nadel, D.
2004. Processing of wild cereal grains in the Upper Palaeolithic revealed by
starch grain analysis. Nature 430(7000), 670-673.
Piperno, R. D. and Tom, D. D.
2008. Starch Grains on Human Teeth Reveal Early Broad Crop Diet in
Northern Peru. Proceedings of the National Academy of Sciences of the
United States 105(50), 19622-19627.
Piperno, R. D., Ranere, J. A. Holst, I. Iriarte, Jose. and Dickau, R.
2009. Starch Grain and Phytolith Analysis from Early Ninth Millennium B.P.
Maize form Central Balsas River Valley, Mexico. Proceedings of the National
Academy of Sciences of the United States of America 106(13), 5019-5024.
Piperno, R. D.
2011. The Origins of Plant Cultivation and Domestication in the New World
Tropics. Current Anthropology 52(S4), S453-S470.
Piperno, R. D. and Flannery, V. K.
2011. The Earliest Archaeological Maize (Zea mays L.) from Highland
Mexico: New Accelerator Mass Spectrometry Dates and their Implications.
Proceedings of the National Academy of Sciences of the United States of
America 98(4), 2101-2103.
Phillips, B. and Hill, C.
2004. Deglaciation History and Geomorphological Character of the Region
between the Agassiz and Superior Basins, Associated with the ‘Interlakes
Composite’ of Minnesota and Ontario. In The Late Paleo-Indian Great Lakes:
Geological and Archaeological Investigations of Late Pleistocene and Early
Holocene Environments, eds. L. J. Jackson and A. Hinshelwood, 275-301.
Mercury Series Archaeology Paper 165. Quebec: Canadian Museum of
Civilization,
Powers, H. Alix.
1992. Great Expectations: A Short Historical Review of European Phytolith
Systematics. In Rapp, Jr. George and Susan C. Mulholland. eds. Phytolith
Systematics: Emerging Issues,15-36.
Rapp, Jr. G. and Mulholland, C. S.
1982. Phytolith Systematics:
Emerging Issues, 1-15.
An
Introduction.
Phytolith
Systematics:
193
Reid, C. S. and Rajnovich, G.
1985. Laurel Architecture: Five Case Studies. The Minnesota Archaeologist
44(2), 5–30.
1991. Laurel: A Re- evaluation of the Spatial, Social and Temporal Paradigms.
Canadian Journal of Archaeology 15:193–234.
Rocha, T. D. S., Carneiro, A. P. D. A. and Franco, C. M. L.
2010. Effect of enzymatic hydrolysis on some physicochemical properties of
root and tuber granular starches. Food Science and Technology
(Campinas) 30(2), 544-551.
Roper, C. D.
2013. Evaluating the Reliability of AMS Dates on Food Residue on Pottery
from Late Prehistoric Central Plains of North America. Radiocarbon 55(1),
151-162.
2014. A Response to Hart and Lovis. Radiocarbon 56 (1), 355-359.
Rosenswig, R. M., Pearsall, D. M. Masson, M. A. Culleton, B. J. and Kennett, D. J.
2014. Archaic period settlement and subsistence in the Maya lowlands: new
starch grain and lithic data from Freshwater Creek, Belize. Journal of
Archaeological Science 41, 308-321.
Rovner, I.
1971. Potential of opal phytoliths for use
reconstruction. Quaternary Research 1(3), 343-359.
in
paleoecological
Scarry, M. C.
2003. Patterns of Wild Plant Utilization in the Prehistoric Eastern Woodlands.
In Paul E. Minnis. ed. People and Plants: in Ancient Eastern North America.
Smithsonian Institute, 50-105.
Scarry, C. M. and Scarry, F. J.
2005. Native American ‘Garden Agriculture’ in Southeastern North America.
World Archaeology 37(2), 259-274.
Scarry, C. M. and Yarnell, A. R.
2011. Native American Domestication and Husbandry of Plants in Eastern
North America. In Smith D. Bruce. ed. The Subsistence Economies of
Ingenuous North American Societies. Washington D.C.: Smithsonian Institute
Scholarly Press, 483-503.
194
Shafer, R. J.
2003. A Seriation of Ceramics from the Big Rice Site (21SL163, FS 09-09034), St. Louis County, Minnesota. MA Thesis, University of Minnesota,
Faculty of Graduate School.
Shillito, L.
2011. Grains of Truth or Transparent Blindfolds?
Debates in Archaeological Phytolith Analysis.
Archaeobotany 22, 71-82.
A Review of Current
Vegetational History
Silbernagel, J., Martin, R. S. Gale, R. M. and Chen, J.
1997.
Prehistoric, historic, and present settlement patterns related to
ecological hierarchy in the Eastern Upper Peninsula of Michigan, USA.
Landscape Ecology 12(4), 223-240.
Simon, L. M.
2009. A Regional and Chronological Synthesis of Archaic Period Plant Use in
the Midcontinent. In Thomas E. Emerson., Dale L. MacElrath, and Andrew C.
eds. Archaic Societies: Diversity and Complexity across the Midcontinent.
Fortier: State University of New York, 81-114.
Simon, M. L.
2011. Evidence for variability among squash seeds from the Hoxie site
(11CK4), Illinois. Journal of Archaeological Science 38(9), 2079-2093.
Skibo, M. J.
2013. Understanding Pottery Function (Manuals in Archaeological Method,
Theory and Technique). New York: Springer Science and Business Media.
Smith, D. B.
1989. Origins of agriculture in eastern North America. American Association
for the Advancement of Science.
1992. Rivers of Change: Essays on Early Agriculture in Eastern North
America. Alabama: The University of Alabama Press.
2006. Eastern North America as an Independent Center of Plant
Domestication. Proceedings of the National Academy of Sciences 103(33),
12223-12228.
2011. ed. The Subsistence Economies of Indigenous North American
Societies: A Handbook. Washington D.C.: Smithsonian Institute Scholarly
Press.
195
Staller, E. J.
2010. Maize Cobs and Cultures: History of Zea mays L. Heidelberg Dordrecht
London New York: Springer.
Staller, J., Tykot, R. and Benz, B.
2006.
Histories of Maize: Multidisciplinary Approaches to the Prehistory,
Linguistics, Biogeography, Domestication, and Evolution of Maize. United
States: Elsevier Academic Press.
Stoltman, James B.
1973. The Laurel Culture in Minnesota. Minnesota Prehistoric Archaeology
Series, no. 8. Minnesota Historical Society. St. Paul.
Struever, S.
1965. Middle Woodland Culture History in the Great Lakes Riverine Area
American Antiquity 31(2), 211-223.
Surette, L. J. C.
2008. The Potential of Microfossil Use in Paleodiet and Paleoenvironmental
Analysis in Northwestern Ontario. MA Thesis, Lakehead University,
Department of Geology.
Syms, E. L., Speirs, K. and Durawa, H.
2013. Aboriginal Peoples as the First Cultivators Centuries of Plant Cultivation
before the Europeans (Rev). Draft
Thomas, M. M. and Mather, D. eds.
1996. The McKinstry Site (21KC2): Final Report of Phase III Investigations for
Mn/DOT S.P. 360-44, Replacement of T.H. 11 Bridge 5178 over the Little Fork
River, Koochiching County, Minnesota. Loucks Project Report 93512,
prepared for the Minnesota Department of Transportation. Maple Grove:
Loucks & Associates, Inc.
Thomas, Matthew.
1996. Ceramic Analysis. In M. M. Thomas and D. Mather (eds.), The McKinstry
Site (21KC2): Final Report of Phase III Investigations for Mn/DOT S.P. 360-44,
Replacement of T.H. 11 Bridge 5178 Over the Little Fork River, Koochiching
County, Minnesota. Loucks Project Report 93512, prepared for the Minnesota
Department of Transportation. Maple Grove: Loucks & Associates, Inc.
Thompson, G. R., Kluth. R. A. and Kluth. D. W.
1994. Tracing the Use of Brainerd Ware through Opal Phytolith Analysis of
Food Residue. The Minnesota Archaeologist 53: 86-95.
196
1995. Ceramic Artifacts. In Justin, Michael. ed. Archaeological Data
Recovery at the Roosevelt Lake Narrows Site, Cass County, Minnesota.
Woodward-Clyde Report 93e209, prepared for Minnesota Department of
Transportation. St. Paul, Minnesota: State Historic Preservation Office.
Torrence, R.
2006. Ancient Starch Research. Walnut Creek, CA: Left Coast Press Inc
Torrence, R. and Barton, H. eds.
2006. Ancient Starch Research. Walnut Creek, California: Left Coast Press.
Twiss, P. C., Suess, E. and Smith, R. M.
1969. Morphological classification of grass phytoliths. Soil Science Society of
America Journal 33(1), 109-115.
Twiss, P. C.
1992. Predicted world distribution of C3 and C4 grass phytoliths. Phytolith
systematics 113-128.
University of Minnesota
Bell
Museum
of
Natural
History
Minneapolis,
http://www.bellmuseum.umn.edu. N.p. Web June 17, 2014.
Minnesota.
Valppu, S. and Rapp, G.
2000. Paleoethnobotanical Context and Dating of the Laurel Use of Wild
Rice: The Big Rice Site. The Minnesota Archaeologist 59:81–87.
Vennum, T. Jr.
1988. Wild Rice and the Ojibway People.
Society Press.
St. Paul, Minnesota: Historical
Vogel, J. C. and Van der Merwe, N. J.
1977. Isotopic evidence for early maize cultivation in New York State.
American Antiquity 42, 238–242.
Wilcox, L.
2007. Going with the Grain: on Minnesota Lakes, Native Americans Satisfy a
Growing Hunger for ‘Slow Food’ by Harvesting Authentically Wild Rice the
Old-Fashioned Way. Smithsonian 38(6), 94.
Wilson, J., Hardy, K. A. R. Copeland, L. Wrangham, R. and Collins, M.
2010. Automated classification of starch granules using supervised pattern
recognition of morphological properties. Journal of Archaeological
Science 37(3), 594-604.
197
Wilson, G. L.
1917. Agriculture of the Hidatsa Indians: An Indian Interpretation. University
of Minnesota, Studies in the Social Sciences, No. 9.
Winham, P. R. and Calabrese, A. F.
1998.
The Middle Missouri Tradition.
In W. Raymond, Wood. ed.
Archaeology on the Great Plains. University Press of Kansas, 269-308.
Wright, Jr. E. H.
1993. History of the Landscape in the Itasca Region. In Bradbury, Platt. J.
and Dean, E. Walter. eds. Elk Lake, Minnesota: Evidence for Rapid Climate
Change in the North-Central United States. The Geology Society of America,
Inc, Boulder, Colorado, 7-18.
Yarnell, R. A.
1964. Aboriginal Relationships between Culture and Plant Life in the Upper
Great Lakes Region. Anthropological Papers 23. Museum of Anthropology,
University of Michigan.
1993. The Importance of Native Crops during the Late Archaic and Woodland
Periods. In C. Margaret, Scarry. ed. Foraging and Farming in the Eastern
Woodland. University Press of Florida, 39-56.
Yerkes, W. R.
1988. The Woodland and Mississippian Traditions in the Prehistory of
Midwestern North America. Journal of World Prehistory 2(3), 307-357.
Yoneda, M., Tanaka, A. Shibata, Y. Morita, M. Uzawa, K. Hirota. M. and Uchida, M.
2002. Radiocarbon marine reservoir effect in human remains from the
Kitakogane site, Hokkaido, Japan. Journal of archaeological science 29(5),
529-536.
Yost, L. C.
2007. Phytolith Evidence for the Presence and Abundance of Wild Rice
(Zizania spp.) from Central Minnesota Lake Sediments. MS Thesis, St. Cloud
State University, Minnesota. Department of Geography.
Yost, C. L., and Blinnikov, M. S.
2011. Locally diagnostic phytoliths of wild rice (Zizania palustris L.) from
Minnesota, USA: comparison to other wetland grasses and usefulness for
archaeobotany
and
paleoecological
reconstructions.
Journal
of
Archaeological Science 38(8), 1977-1991.
198
Yost, L. C., Blinnikov, S. M. and Julius L. M.
2013. Detecting Ancient Wild Rice (Zizania ssp. L.) Using Phytoliths: A
Taxonomic Study of Modern Wild Rice in Minnesota (USA) Sediments.
Journal of Paleolimnology 49(2), 221-236.
Zarrillo, S. and Kooyman B.
2006. Evidence of Berry and Maize Processing on the Canadian Plains from
Starch Grain Analysis. American Antiquity 71(3), 473-499.
Zarrillo, Sonia., Deborah M. Pearsall. Scott J. Raymond. Mary Ann Tisdale. and
Dugane J. Quon.
2008. Directly Dated Starch Residues Document Early Formative Maize (Zea
mays L.) in Tropical Ecuador. Proceedings of the National Academy of
Sciences 105(13), 5006-5011.
Zellie, C.
1988. Upper Mississippi River Headwaters Reservoirs Damsites Cultural
Resources Investigation. Lake Winnibigoshish, Leech Lake, Pokegama Lake,
Sandy Lake, Pine River and Gull Lake. St. Paul, Minnesota: Landscape
Research.
199
APPENDIX
Archaeological Site
Flora
Fauna
Big Rice
Zizania spp. (Wild Rice)
Odocoileus virginiana (Deer)
Tubers
C. Canadensis (Beaver)
Hog-nut
Bison bison (Bison)
Fruits
Testudines (Turtle)
Nuts
Chordate (Fish)
Z. mays (Corn)
Gastropoda (Land Snail)
Cucurbita (Squash)
Bivalvia (Mussels)
Cyanoccus (Blueberry)
Rubus idaeus (Wild Red Raspberry)
A. alces (Moose)
O. zibethicus
(Muskrat)
Sambucus (Red-berried Elder)
Erethizontidae(Porcupine)
Prunus pensylvanica (Pincherry)
Lepus (Hare)
Abies (Fir)
Aves (Variety of Birds)
Populus tremuloides (Aspen)
Acer (Maple)
Crataegus (Thornapple)
Betula (Birch)
Tilia Americana (Basswood)
Schultz
Zizania ssp. (Wild Rice)
Nuts
Upper Rice Lake
C. Canadensis (Beaver)
O. zibethicus
(Muskrat)
C. Lupus (Dog)
U. americanus (Black Bear)
Aves (Bird)
Chordate (Fish)
Anura (Frog)
Bivalvia (Mussels)
Bison bison (Bison)
A. alces (Moose)
Cervus canadensis (Elk)
O. virginianus (White-tailed Deer)
Mephitidae (Skunk)
Mustelinae (Mink)
Lynx lynx (Lynx)
200
Spermophilus (Gopher)
Third River Borrow Pit
Picea (Spruce)
Canis lupus (Wolf)
Pinus strobus (White Pine)
Ursidae (Bear)
Pinus resinosa (Red Pine)
O. zibethicus
(Muskrat)
Pinus banksiana (Jack Pine)
C. Canadensis (Beaver)
Mustela (Mink)
Lutrinae (Otter)
M. pennanti (Fisher)
Leporidae (Rabbit)
Anseriformes (Waterfowl)
Chordate (Fish)
A. alces (Moose)
Rangifer
Caribou)
tarandus
caribou
(Woodland
Bison bison (Bison)
Cervus canadensis (Elk)
Cervidae (Deer)
McKinstry (21KC2)
Chenopodium sp.
Acipenser fulvescens (Lake Sturgeon)
Viola canadensis
Alces alces (Moose)
Rubus idaeus
Marmota monax (Woodchuck)
Sambucus spp.
Castor canadensis (Beaver)
Rumex sp.
Moxostoma (sucker)
Pinus sp.
Picea sp.
Abies balsamea
Zizania Aquatica
Appendix 1: Floral and faunal remains recovered from a few northern Minnesota
Woodland sites, representing some species in the Great Lakes-St. Lawrence Forest
(Bishop 2007; Gibbon 2012a; Hoover 1963; Mulholland et al. 1996; Peter and
Motivans 1983; Shafer 2003).
201
Sample and
Laboratory
Number
Association
Material
Date with
Standard
Deviation
N/a
Feature with
pit and wild
rice
Charcoal
1670 + 45
BP
Appendix 2: Radiocarbon date from a sample at the Big Rice site (Shafer
2003).
Radiocarbon Age (BP)
Sample
Association Material
and
Laboratory
Number
C1-A Beta Top
of Charcoal
94420
Feature 1 Birch
(cf Betula
Brainerd
papyrifera)
C2-A
Base
of Charcoal
Beta
Feature 1
101863
Brainerd
Calibrated Age
Date with Conventional Intercept 1
Standard Age
Deviation
Sigma
2
Sigma
1700+70
1700+70
AD380
AD250- AD
210425
535
1850+50
1860+50
AD145
AD100- AD
65260
235
C3-A Beta ThroughCharcoal 1590+50
101864
out Feature
1 Brainerd
1620+50
AD430
AD405- AD
350530
560
Crust 1
2320+60
BC390
BC405- BC
320
505330
BC
330205
Brainerd
Organic
HorizonResidue
tally
Corded
Sherd
in
Feature 1
2460+60
Appendix 3: Radiocarbon dates from charcoal and organic residue samples at the
Third River Borrow Pit Site (Mulholland et al. 1996).
202
Feature
11
22
33
36
36
Sample
Conventional
1 Max Cal Age
14C Age (B.P.) (Intercepts) Min
Cal Age
(BA 94091)
1910 +/-100
B.C.
35
(A.D.81) A.D.
229
(BA 94092)
2040 +/- 100
B.C.
172
(44.5.5) A.D.
66
(BA 94093)
2020 +/- 90
B.C.
160
(39.29.22.10.1)
A.D. 72
(BA 94094)
1060 +/- 80
A.D 894 (989)
1024
(Beta-75859) 600 +/- 60
A.D.
1299
(1327, 1346,
1393) 1410
2 Max Cal Age
(Intercepts)
Min Cal Age
B.C. 156 (A.D.
81) A.D. 339
B.C.
358
(44.5.5) A.D.
133
B.C. 349 (39.
22. 10.1) A.D.
133
A.D. 780(989)
1159
A.D.
1283
(1327. 1346.
1393. 1435)
Appendix 4: AMS dates from Area A of Zizania aquatica seeds (Valppu and Rapp
2000).
203
Site Name
Microfossils
No Beard(1)
Windy Bead (1)
Windy Bead (2)
No Beard (2)
No Beard (3)
No Beard (4)
Elongate Plates
(Gramineae)
6
4
2
9
5
3
Elongate Plates
(Smooth)
1
1
0
0
0
0
Saddles
2
1
22
7
22
1
0
18
14
20
0
0
7
1
7
4
0
15
22
30
3
0
20
10
18
0
0
3
2
3
Long Sinuous
Trapezoids
33
51
7
29
24
8
44
37
18
28
18
2
30
29
8
11
21
9
43
4
0
0
45
7
1
0
9
0
1
1
14
12
3
0
38
9
0
3
3
0
0
0
0
0
8
0
0
1
0
1
3
1
0
33
1
1
31
0
0
7
0
1
1
10
4
1
0
0
0
0
1
0
0
0
0
0
0
1
30
0
0
20
0
0
95
0
0
47
0
0
50
1
1
25
1
0
Elongate Plates
(other)
Short Plates
Trichomes
Double Outlines
Short Sinuous
Trapezoids
Long Non
Sinuous
Trapezoids
Short Non
Sinuous
Trapezoids
Bilobates
Polylobates
Crosses
Rondel (Zea
mays)
Rondel (Zizania)
Other Rondel
Zea mays
starch
Possible
Cucurbita
starch
Possible
Phaseolus
vulgaris starch
Starch
Unknown
Diatoms
Pollen
Appendix 5: Microfossils recovered from lithic samples.
204
Appendix 6: Microfossils recovered from ceramic samples (Table one of
two).
205
Appendix 7: Microfossils recovered from ceramic samples (Table two of
two).
206
Site Name
Microfossils
Big Rice (1)
Big Rice
(2)
Windy Bead
(1)
Windy Bead
(2)
Windy Bead
(3)
Elongate Plates Gramineae
7
18
15
11
22
Elongate Plates Smooth
0
2
6
9
8
Elongate Plates Other
3
6
12
11
18
Short Plates
5
0
0
13
0
Trichomes
15
13
24
15
18
Double Outlines
10
16
7
11
23
Saddles
34
17
18
31
34
Long Sinuous Trapezoids
41
38
62
34
26
Short Sinuous Trapezoids
Long Non Sinuous
Trapezoids
Short Non Sinuous
Trapezoids
17
31
18
23
17
10
17
23
29
22
18
16
29
20
11
Bilobates
3
1
3
9
25
Polylobates
1
0
2
6
4
Crosses
0
0
1
0
1
Bulliforms
0
2
0
2
3
Rondel Zea mays
1
2
2
1
1
Rondel cf Zea mays
0
0
0
0
0
Rondel Zizania sp
10
5
0
5
2
Rondel cf Zizania sp
0
0
0
0
0
Rondel Other
90
50
24
139
104
Unclassified Phytoliths
3
0
4
2
3
Zea mays starch
4
0
0
0
0
Possible Cucurbita starch
Possible Phaseolus
vulgaris starch
1
0
0
0
1
1
0
0
0
0
Unknown Starch
19
1
6
5
15
Diatoms
4
6
1
1
0
Pollen
0
0
0
1
1
Appendix 8: Microfossils recovered from sediment samples.
Rondel Type
Site Name
Windy Bead Unit 4
Big Rice 09-034 2754g
Big Rice #3
Saga Island AB45
Zea mays
Zizania
Zizania
Zizania
Zizania
Hesperostipa comata
Danthonia spicata
Hesperostipa comata
Appendix 9: Possible plants identified from rare rondel phytoliths.
207
Alexandra Burchill