Archaeobotanical Analysis: Principles
and Methods
by Karen R. Adams
Introduction
Sample Types
General Principles and Conventions
Charred vs. Noncharred Remains
Wood Charcoal vs. Charred Nonwood
Specimens
Naming Conventions
Methods
Macrofossils
Wood Charcoal
Charred Nonwood
Specimens
Modified Vegetal
Items
Microfossils: Flotation Samples
Processing
Subsampling for
Wood Charcoal
Subsampling for
Charred Nonwood Specimens
Examination and
Identification
Indices
Additional Resources
References Cited
To Cite This Publication
Introduction
1
Archaeologists collect ancient plant remains in an effort to understand
prehistoric plant use by the inhabitants of a site or locale and to reconstruct
the nature and composition of vegetation communities that existed in the
past. Plant remains collected from archaeological contexts provide information
about both wild and domesticated plant resources used for food, fuel,
construction, and other purposes. Because people generally used resources
that were available locally, plant assemblages can also provide important
clues about the characteristics of the natural vegetation that formerly
grew in the vicinity of sites. Under certain circumstances, data derived
from ancient plant assemblages can also provide insights into seasonality
of site use and abandonment, paleoenvironmental conditions, and human
nutrition and health; this is especially true when the plant data are
evaluated in conjunction with other kinds of evidence, including pollen,
tree-ring, faunal, and human osteological data.
2
The purpose of this publication is to describe for researchers
and other interested readers the types of archaeobotanical samples collected
by the Crow Canyon Archaeological Center and the methods used to process
and analyze them.1 This information should
allow others to evaluate not only the methods employed, but also the data
obtained as a result—data which are reported in detail in The
Crow Canyon Archaeological Center Research Database and in individual
chapters in Crow Canyon's series of on-line
site reports.2 Although this publication
is designed primarily as a companion piece to the site reports, researchers
and students might find the information contained herein useful in guiding
their own studies as well.
3
This publication deals with plant remains collected as individual
vegetal specimens and plant remains recovered in flotation samples. It
does not deal with pieces of wood collected as tree-ring samples, which
are processed and analyzed by the Laboratory
of Tree-Ring Research in Tucson, Arizona, or with pollen samples,
which are processed and analyzed by various independent contractors. For
a description of how tree-ring and pollen samples are collected in the
field, refer to Crow Canyon's on-line field
manual.
Sample Types
4
The plant remains collected from sites excavated by the Crow Canyon
Archaeological Center fall into one of two broad categories depending
on their size and the method used to collect them. Macrofossils are pieces of plants that are large enough to be seen with the unaided
eye and that are retrieved by hand during the excavation and screening
of sediment in the field. The majority of macrofossils are submitted to
the laboratory as vegetal specimens (data-entry code: VEG), and most often
they consist of pieces of wood charcoal, although corn cobs, squash seeds,
and other large, distinctive plant remains are occasionally collected
as well. Also included in this category are plant materials that were
intentionally modified into tools, containers, or other serviceable items;
these are submitted to the laboratory as basketry (BAS), textile (TEX),
or "other modified vegetal" (OMV) specimens. Macrofossil remains constitute
a subjective sample of the contents of a given deposit, because collection
depends on what an individual excavator decides to retrieve. For example,
all modified vegetal specimens recognized in the field are collected,
but it is impractical for excavators to retrieve every piece of wood charcoal
observed in the course of digging a burned structure.
5
Microfossils are tiny pieces of plant material recovered
specifically in flotation samples (data-entry code: FLO). Small seeds,
fruits, and other reproductive parts make up the majority of items in
this category, but tiny fragments of wood and other plant parts may be
recovered as well.3 Flotation samples consist
of a standard volume of sediment (usually 1 liter), and they are routinely
collected from contexts where plant remains are expected or perceived
to be plentiful. These contexts include primary refuse (for example, ash
in thermal features), secondary refuse (middens and trash fill in structures),
collapsed structure roofs ("roof fall"), and other cultural deposits.
Because they are collected and processed systematically, flotation samples
provide relatively unbiased data about prehistoric plant use at a site.
For a more detailed discussion of when and how flotation samples are collected
at sites excavated by Crow Canyon, refer to the field
manual.
General Principles and Conventions
Charred vs. Noncharred Remains
6
Both charred and noncharred plant remains are recovered from sites
excavated by the Crow Canyon Archaeological Center. Charring is usually
considered to be the result of prehistoric cultural activities; noncharred
remains are more likely to be modern or to have been introduced through
natural processes such as rodent activity and wind (Minnis
1981*1:147). Therefore, most often only charred (including partly
charred) materials are considered in analyses of Crow Canyon archaeobotanical
assemblages; the presence of noncharred specimens is noted, but these
items are not usually included in assessments of prehistoric plant use.
There are exceptions to this rule, however, and every instance is considered
on a case-by-case basis. For example, the recovery of a single charred
domesticated wheat (Triticum) grain from Troy's Tower (Site 5MT3951),
a Pueblo III site tested by Crow Canyon (Varien
1999*10), can be explained by the fact that in modern times wheat
stubble had been burned in the agricultural field in which the site is
located (Adams 1999*3). On
the other hand, the recovery of a noncharred gourd container from Sand
Canyon Pueblo (Site 5MT765), another Pueblo III site excavated by Crow
Canyon, clearly is a case of extraordinary preservation of a prehistoric
artifact (Adams 2000*1).
Wood Charcoal vs. Charred Nonwood Specimens
7
Charcoal is defined as any "dark or black porous carbon prepared
from vegetable or animal substances" (Merriam-Webster
1996*1:192). Most archaeobotanists, however, restrict the use of the
word "charcoal" to describe burned wood specifically—that is, charred
pieces of trunks, branches, and twigs from both trees and shrubs. The
charred remains of the nonwoody parts of trees and shrubs (for example,
fruits and seeds) and the charred remains of nonwoody plants (herbaceous
perennials and grasses) are generally not described as charcoal, even
though they, too, clearly fall under the strict definition. Instead, such
specimens are usually referred to in more-specific descriptive terms,
such as "reproductive" or "vegetative" parts, without continued reference
to their charred condition. Because of the potential for confusion, especially
among readers who are not archaeobotanists, I have begun referring to
the two categories as "wood charcoal" and "charred nonwood specimens."
This convention should make clear that plant remains in both categories
are burned, but it also preserves the custom with regard to the use of
the word "charcoal" that has established itself in the archaeobotanical
literature.
Naming Conventions
8
Because many plants are known by multiple common names, which vary
according to geographic region and local custom, the only consistent and
reliable means of reporting data is to use standard scientific nomenclature.
Therefore, in all Crow Canyon site reports, plants are referred to first
by their scientific names, then by their common names. The latter are
provided either parenthetically in text or in tables showing the correspondence
between scientific and common names. Once the correspondence has been
established in a given publication, both scientific and common names are
often used interchangeably, depending on the context. The taxonomic nomenclature
(both scientific and common names) used in Crow Canyon publications conforms
to A Utah Flora (Welsh
et al. 1987*1) whenever possible.
9
When a specimen is identified by a combination of two genus names—for
example, "Amelanchier/Peraphyllum" or "Prunus/Rosa"—it
means that the specimen could be a member of either of the named genera
and that a more precise determination was not possible. The order in which
the names appear does not indicate which genus identification is most
likely correct; the names are simply listed alphabetically. Some specimens
in a given sample are identified to the genus level; others in the same
sample may be identified to both the genus and species levels. This occurs
because some specimens are better preserved than others. For interpretive
purposes, these items most often would be combined at the broadest level
of identification (see discussion of taxon diversity, paragraph
28).
10
One special case involving the identification of plant remains
to two possible genera requires explanation. Archaeobotanists use the
common name "cheno-am" (lowercase) to refer to seeds that could be from
plants in either the genus Chenopodium or the genus Amaranthus. Palynologists, on the other hand, use "Cheno-am" (uppercase) to refer
to pollen grains that could be from either the family Chenopodiaceae or
the genus Amaranthus. Except for the difference in capitalization,
there is no obvious clue that the definition used by archaeobotanists
is somewhat narrower than that used by palynologists.
11
The taxonomic level to which any given specimen is identified depends
on a number of factors, including the "comfort level" of the individual
analyst, the condition of the specimen, and whether or not other similar-appearing
items are known from the region. (Only angiosperms, or flowering plants,
are identified to the subclass level—that is, identified as Monocotyledons
or Dicotyledons.) Following Bohrer
and Adams (1977*1:41), analysts indicate one of three levels of confidence
for each identification: "absolute," "-type," or "compares favorably."
An "absolute" identification to the level of species indicates that all
species of the relevant genus in the local environment have been examined,
and the specimen in question appears identical to the named species. The
label "-type" signifies that the specimen has morphological characteristics
that closely resemble those of the named species, but that other plants
in the area might also have similar-looking parts. "Compares favorably"
(abbreviated "cf.") is used when a specimen is difficult to identify because
of poor preservation or some other factor, but its characteristics suggest
that it is more likely to be of the named species than of some other species;
the use of this abbreviation indicates a "best guess" assessment. Although
species is the taxonomic category used in the foregoing explanations,
the three qualifiers can be used for identifications to any taxonomic
level. In Crow Canyon reports, analytic confidence is indicated in tables
by the inclusion of either "cf." or "-type" preceding or following the
scientific name (for example, cf. Pinus or Pinus-type).
Scientific names that do not include either of these qualifiers should
be assumed to indicate "absolute" identifications (for example, Zea
mays refers to a positive identification to this species). Including
qualifiers every time a scientific name is mentioned in a discussion,
however, would be cumbersome; for that reason, they are generally omitted
from text. Readers should refer to the tables that accompany the text
if they wish to know the level of confidence of a given taxonomic identification.
12
In cases in which a prehistoric specimen cannot be assigned to
a specific taxonomic category, the item is recorded as "unknown." If similar
or identical specimens are found with some frequency, additional efforts
at identification are made. If those efforts fail, the "unknown" is described
in terms of its key morphological and anatomical features. If recognized,
nonbotanical materials such as bone, coal, termite fecal pellets, insect
parts, gastropods, and "black spherical bodies" are recorded as present
in samples.
Methods
Macrofossils
Wood Charcoal
13
Charred wood preserves well in archaeological contexts because
it is more resistant than noncharred wood to decay and insect damage.
Some specimens even retain an outer layer of bark, although most do not.
Identification of wood charcoal in archaeological assemblages can provide
valuable information about the types of wood used for fuel and in construction.
Subsampling
14
Because wood charcoal is usually plentiful in archaeological contexts,
economy dictates that only a subsample of all the specimens submitted
be examined. The first step in analyzing wood charcoal is to pour the
entire contents of a given bag of plant material (which may contain both
wood and nonwood vegetal specimens) onto newspaper or a lab tray so that
each individual item is visible. The analyst then attempts to select an
unbiased subsample of 20 pieces of wood charcoal by choosing specimens
of different sizes and retrieving items from different sections of the
newspaper or lab tray. If fewer than 20 pieces of charred wood are present,
all are examined; if more than 20 pieces are present, additional specimens
that appear morphologically distinctive to the naked eye are examined
after the first set of 20 has been analyzed.
Examination
15
Each piece of wood charcoal is snapped in half so that a fresh
transverse (cross) section is exposed. This section is then examined under
a dissecting binocular microscope at magnifications of 10 to 45X. Often
the pieces are placed in a sand-filled petri dish to stabilize them for
ease of viewing.
Identification
16
Each piece of wood charcoal examined is identified to the finest
of four possible taxonomic levels, listed here from most general to most
specific: subclass, family, genus, and species. Wood charcoal identifications
are based on anatomical traits (for example, rings, vessels, rays, and
background patterning) viewed in cross section. Ancient wood charcoal
is identified using Crow Canyon's modern wood charcoal comparative collection
from the region, backed by plant voucher specimens in the University of
Arizona herbarium in Tucson. Because of similarities among woods of some
plant families or genera in the region, more wood types may be present
in Crow Canyon archaeological assemblages than are actually indicated
in the data summaries. For example, wood identified as Pinus may
be either Pinus edulis or Pinus ponderosa. The identification
criteria used in the analysis of wood charcoal collected from sites excavated
by Crow Canyon are described in detail in Plant
Identification Criteria, an on-line resource for researchers.
Charred Nonwood Specimens
17
Nonwood plant specimens include both agricultural and wild resources,
and they usually consist of reproductive parts such as seeds or fruit.
Excavators collect recognizable pieces of domesticates, such as maize
(Zea mays) cobs and kernels, squash (Cucurbita) seeds and
rinds, and bean (Phaseolus) seeds. They also collect pinyon (Pinus
edulis) nutshells, yucca (Yucca baccata) fruits and seeds,
and a wide variety of other wild plant parts, including such vegetative
parts as leaves, stalks, and roots.
Examination
18
All charred nonwood specimens are removed from bags of macrofossils
whose contents have been thinly spread out on newspaper or a lab tray
so that all specimens are visible. Each item is examined under a dissecting
binocular microscope (10 to 45X) and, if possible, identified. The nonwood
materials are bagged and labeled separately from the wood charcoal found
in the same sample.
Identification
19
The identification criteria for reproductive plant parts include
size, shape, surface texture, and points of attachment. Seeds are identified
using the author's modern comparative collection, backed by University
of Arizona herbarium voucher specimens. Reference texts, such as Delorit
(1970*1), Landers and Johnson
(1976*1), and Martin and
Barkley (1961*1), are also used. Vegetative parts such as leaves and
stems are identified on the basis of internal anatomy and morphology.
If an item cannot be identified by reference to the collections or texts,
it is measured and described as an "unknown."
Modified Vegetal Items
20
Plant materials were sometimes modified (cut, split, woven, or
otherwise shaped) in order to fashion them into various utilitarian or
decorative items, such as baskets or other containers, mats, plaques,
sandals, and tools. These items—whole and fragmentary, charred and
noncharred—occasionally preserve in archaeological deposits and,
when intact enough to be recognized in the field, are collected as macrofossil
specimens. Because of their generally fragile condition, special care
is used in handling them during excavation, processing, and analysis.
Examination
21
Modified vegetal artifacts are fully described both quantitatively
(measured, counted) and qualitatively (text descriptions). Microscope
examination is sometimes required to determine the anatomical details
of the constituent elements. Very rarely, small portions of the items
are broken off to expose fresh transverse (cross) sections.
Identification
22
The identification conventions cited above for wood charcoal and
charred nonwood specimens (see paragraph 16 and paragraph 19) are also applied to modified vegetal
items. The raw materials of construction are identified to the most specific
taxonomic level possible. Although many modified vegetal items are constructed
of a single plant or plant part, some are made up of more than one plant
or plant part. In such cases, each of the constituent elements is identified
and described, if possible. The terminology used to describe modified
items as artifacts conforms to conventions published by specialists familiar
with each type of artifact.
Microfossils: Flotation Samples
Processing
23
The steps used to process flotation samples are outlined in detail
in the on-line laboratory
manual, but the procedure basically involves placing an individual
sample in a bucket, adding water, then gently stirring the mixture to
free the organic materials. The inorganic material that settles to the
bottom of the bucket constitutes the "heavy fraction," which is collected,
allowed to dry, and stored. Charred and noncharred plant remains that
float to the surface are poured into a fine (0.355-mm) mesh to be captured
as the "light fraction." The light fraction is allowed to dry before being
sifted through a series of geologic sieves. This process separates the
light (organic) fraction into 4.75-mm, 2.80-mm, 1.40-mm, 0.71-mm, and
0.25-mm subsamples, which are then individually bagged and labeled. Although
the size of the original mesh used to capture the entire light fraction
(0.355 mm) is larger than the finest mesh used to create the subsamples
(0.25 mm), smaller particles that adhere to larger particles when they
are wet can detach as the residue dries and be caught in the 0.25-mm screen
used during dry screening. Plant remains also continue to break into smaller
pieces whenever a sample is handled.
Subsampling for Wood Charcoal
24
Analysts examine the flotation light fraction in two steps, using
a dissecting binocular light microscope at magnifications ranging from
10 to 45X. The first step involves subsampling the wood charcoal. The
analyst begins by choosing 20 pieces of wood (using the method described
in paragraph 14) from the 4.75-mm portion, because
the large size of the individual specimens allows for more-confident identification.
If the 4.75-mm portion does not contain 20 pieces of wood, some of the
larger specimens in the 2.80-mm portion are selected to achieve the goal
of examining 20 pieces total. After the initial subsample of 20 pieces
is analyzed, additional pieces that appear morphologically distinctive
to the unaided eye are also examined and identified.
Subsampling for Charred Nonwood Specimens
25
The second step in analyzing plant remains from flotation samples
involves looking for nonwood remains such as seeds, fruit, other reproductive
structures, and nonwood vegetative parts. Samples less than 50 ml in total
light-fraction volume are analyzed completely. For samples with light
fractions that measure 50 ml or more, the following protocol applies:
The 4.75- and 2.80-mm portions are completely examined under the microscope
for seeds and for other reproductive and vegetative parts. An approach
known as the "species area curve" (Mueller-Dombois
and Ellenberg 1974*1:52–53) is then used to subsample the 1.40-
and 0.71-mm portions. This approach maximizes the number of taxa recorded
while minimizing the volume of sample sorted (Adams
1993*1:196). The goal is to identify the maximum number of taxa represented
in a sample, rather than to record the total number of whole or fragmented
items of each taxon. The 1.40- and 0.71-mm portions are sorted in increments
of 0.90 ml. Each subsample of 0.90 ml is measured with a graduated cylinder.
No new taxa must be identified in three successive 0.90-ml subsamples
for the sample sieve size to be considered completely analyzed. If remains
of new taxa are observed, then an additional three 0.90-ml subsamples
are examined until no new taxa are identified. In a similar manner, the
portion captured in the 0.25-mm screen is sorted in increments of 0.30
ml, again using the species-area-curve approach. Materials that pass through
the 0.25-mm screen are not examined, because seeds of this size are usually
also preserved in screens with larger mesh sizes and because fragmented
items of this size are extremely difficult to identify.
Examination and Identification
26
The conventions cited above for examining and identifying macrofossils
(paragraph 16 and paragraph
19) also apply to items recovered in flotation samples.
Indices
Taxon Ubiquity
27
Taxon ubiquity, that is, the number of samples in which the remains
of plants of a given taxon occur within a total universe of samples analyzed,
provides insight into the frequency of use of a plant resource in prehistory.
Ubiquity is calculated for flotation samples only, and the measure is
expressed as a percentage. To calculate ubiquity for a given taxon, the
site or other sample set of interest must first be selected, then the
number of flotation samples in which plant parts of the given taxon occur
is counted. This number is divided by the total number of flotation samples
analyzed for the site or sample set, and the resulting figure is converted
to a percentage. For example, the presence of cheno-am seeds in six of
15 analyzed flotation samples would constitute a ubiquity of 40 percent.
From this, it would be inferred that cheno-ams were a commonly used resource.
Sterile and nonproductive samples—that is, flotation samples that
yield no botanical remains and samples that yield only unidentifiable
botanical remains—are included in the total number of samples analyzed.
Taxon Diversity
28
In any given sample set, and even within individual flotation samples,
there may be items identified to similar, but not identical, taxonomic
levels. When assessing taxonomic diversity within a given sample, analysts
at Crow Canyon take a conservative approach and combine certain records
for the purpose of counting the number of different taxa represented.
For example, if a flotation sample contains both Artemisia tridentata–type
wood charcoal and Artemisia-type wood charcoal, the two taxonomic
levels are combined and considered to represent only a single taxon (Artemisia)
for the purpose of determining the diversity of potential wood types in
the sample. A single taxon represented by more than one plant part may
be handled slightly differently, however, depending on the parts present
and the types of prehistoric uses they are believed to represent. For
example, Pinus-type bark scales, wood charcoal, and cone scales
identified in a single flotation sample would be counted as a single taxon
because all are believed to represent the use of pine for fuel. However,
if Pinus-type nutshell fragments were identified in the same sample,
they would be counted independently in discussions of food taxa because
they are believed to represent the use of pine for food.
Taxon Density
29
Archaeobotanists often calculate the number of plant parts per
unit of sediment volume examined—for example, the number of juniper
seeds per liter of flotation sample. Taxon density is not calculated for
Crow Canyon samples, because not all examples of each taxon/part combination
within each sample are identified or counted, and only rarely are samples
examined in their entirety. As stated earlier, the goal when analyzing
flotation samples from Crow Canyon sites is to identify the maximum number
of taxa represented in a sample, not to record the total number of whole
or fragmented specimens of each taxon/part combination that is present.
Additional Resources
30
A number of additional resources available on Crow Canyon's Web
site supplement the information presented in this document:
- Ethnographic
Uses of Plants: This database lists historic uses of all plant
parts identified in plant assemblages from sites excavated by the Crow
Canyon Archaeological Center. Full references to the primary ethnographic
literature are provided.
- Plant
Identification Criteria: This document details all metric and
nonmetric data relevant to the identification of all taxa and plant
parts in assemblages from sites excavated by Crow Canyon. Descriptions
of selected taxa are accompanied by photographs, most of them taken
through the microscope.
- The
Crow Canyon Archaeological Center Research Database: This large
database includes data for all analyzed archaeobotanical samples and
specimens collected at sites investigated by Crow Canyon.
1Archaeobotanical samples and specimens
are collected by Crow Canyon archaeologists and participants in the Center's
research and education programs. The materials are processed and analyzed
by both the author and numerous student interns working under the author's
supervision.
2Five on-line chapters (Adams
1999*3; Adams and Brown 2000*1; Bowyer and Adams 2004*1; Murray and Jackman-Craig 2003*1; Rainey and Jezik 2002*1),
as well as one chapter published in traditional print form (Adams
1993*1), also include descriptions of the methods used to process
samples and analyze plant remains. The methods described in these earlier
publications document the gradual refinement of the procedures used at
Crow Canyon, culminating in the protocol described in the current document.
3Pollen grains, which are not visible to
the naked eye, constitute a distinctive category of plant remains subject
to specialized sampling, processing, and analysis techniques. Although
pollen data are sometimes presented and discussed in Crow Canyon publications,
those data are provided by independent palynological consultants, whose
methods are described in their individual reports.
Karen R. Adams (Ph.D., University of Arizona,
1988) is an independent consultant in archaeobotany with more than 30
years of experience in the American Southwest and northern Mexico.
To cite this publication:
Adams, Karen R.
2004 Archaeobotanical Analysis: Principles and Methods [HTML Title].
Available: https://www.crowcanyon.org/plantmethods. Date of use: day month
year.*
*Example: Date of use: 26 November 2004. |
