6
Excavation Methodology and Field Procedures
CHRISTINE A.HASTORF AND MATTHEW S. BANDY
I N THE 1996 FIELD SEASON, we exposed broad,
previously unexcavated areas of the site in
order to better understand the amorphous and
complex soil deposits we had found in the 1992
field season. The areas we excavated were marked
out with the units designated by their southwest
comer, based on the same grid system used in the
surface collection of the site. Each excavation unit
begins with an arbitrary 2 by 2 m unit at 10 cm
deep but switched to culturally defined areas as
soon as soil matrix changes were visible. Some
units have odd numbers due to a lack of theodolite
access at certain times during the early excava-
tions. All excavated soil was passed through a
.635 (1/4" mesh) screen in measured buckets of 10
liters, except for the soil collected as a flotation or
the archival soil (pollen-phytolith-archive)
samples, both plotted on each locus excavation
plan. In 1996, every tenth bucket of soil was
passed through a .32 cm (1/8" mesh) screen to
gain a subsample of the smaller animal bone and
lithic debitage. All excavated soil is therefore
measured volumetrically, allowing us to calculate
the density of the artifacts within each excavated
locus across the site. All artifacts were bagged and
tagged (and washed) by artifact type. These
artifacts are currently stored in the Tiwanaku
regional museum under the auspices of the
Direcci6n Nacional de Arqueologia y
Antropologia (DINAAR), the official Bolivian
archaeological commission. Each artifact type is
processed differently post-excavation. The
ceramics are soaked to extract the salts and then
washed in the field. All of the lithics and the bone
are also washed in Chiripa. Ceramics that had any
evidence of organic encrustation are set aside and
scraped to collect the organic remains before
washing. All complete artifacts are photographed,
all diagnostic ceramics drawn. Large and prehis-
torically sealed ground stone was washed using
distilled water, with all of the sediment captured
for pollen analysis.
We conducted water flotation soil processing
to collect a systematic sub-sample of charred
plants, micro-fauna, and other small artifact types.
Our flotation methods included collecting at least
one standard-sized soil sample from every exca-
vated locus. Field excavators were instructed to
30      Early Settlement at Chiripa, Bolivia
sample each locus by collecting one 10-liter
"bulk" flotation sample (Popper and Hastorf
1988), and in certain contexts, like use-surfaces or
middens, we also collected a second 10-liter
"scatter" or average soil sample. In addition,
across certain surfaces, many bulk soil samples
were taken, usually one sample every 50 centime-
ters. A "bulk" soil sample is a single 10-liter block
of soil, with a recorded x, y, z provenience. A
"scatter" sample is a collection of soil distributed
from throughout the locus matrix to create an
average view of what was deposited within the
soil. These two strategies are implemented in
parallel to provide a fuller view of the artifactual
material from a specific locations (Lennstrom and
Hastorf 1992) (see Whitehead, chapter 12).
Micromorphology samples were collected
occasionally in vertical columns as intact solid
blocks of soil approximately 20 x 10 x 10 cm.
Such soil thin section analysis can present an in
situ and detailed view of the soil matrix found in
archaeological sites with more specific views of
what was laid down as well as the conditions of its
deposit (see Goodman below, chapter 10). A
series of micromorphological soil profiles were
collected from the excavations at the Monticulo,
Santiago, and Llusco areas. Additional off-site
"natural" samples were collected as controls.
STRATIGRAPHIC ORGANIZATION
OF THE EXCAVATIONS
To build a sequence of past events, and
therefore a Harris Matrix of every excavation area,
we use the term locus to denote the smallest
visibly defineable and excavated unit (Harris
1979). Each locus represents an action in the past.
These matrices allow us to stratigraphically
reconstruct the sequence of past events. Appendi-
ces 1-3 display the three main area matrices from
the 1996 excavations. On the form of each locus,
we always described the soil of each excavation
locus, often taking both color and black-and-white
photographs. The locus is to be distinguished
from the stratigraphic event, which is a unit of
stratigraphy. The stratigraphic event is a natural
property of the matrix, resulting from the pro-
cesses by which the site was formed. Events were
labeled by area of the site, such that all D-events
are from the mound area, B-events are from
Santiago, and A-events are from Llusco. The
locus, by contrast, is an archaeological unit of
provenience, formed by the manner and sequence
in which the site is excavated. Ideally, each locus
should belong to only one stratigraphic event,
although stratigraphic events can contain many (or
no) loci. An event is a unit of homogenous soil
linked to an activity or process; for example, an
ash lens, an intentional fill level, a floor, or the cut
of a pit or of a foundation trench. Stratigraphic
events occurred in the past in a particular and
determinate sequence. One of the primary goals
of our excavation was to reconstruct this sequence.
Reconstruction of a detailed stratigraphic
sequence allows for fine chronological control, as
well as a more detailed understanding of the
processes, both natural and cultural, that are
responsible for the formation of the site as it exists
today. Features were also assigned when identi-
fied. We define a level by culturally visible
changes in the deposits. We use the term architec-
tural sub-division (ASD) for an architectural
feature such as a structure. These identities are all
recorded on every appropriate locus form and on
our database. While analysis occurs at the locus
level, interpretation is at the event level.