You are on page 1of 16

Klarich_Craig_SAA2001.

doc

Page: 1

Geophysical Survey in the Lake Titicaca Basin: Uncovering


Elite Domestic Architecture at Pucar, Peru.1
Elizabeth Klarich and Nathan Craig
University of California, Santa Barbara
Please do no cite in any context without permission of the authors.

Introduction
It is the aim of PROYECTO de ARQUEOLOGA DOMSTICA de PUKARA to use data collected
from excavations of elite households to determine how aspiring political actors in the Lake
Titicaca Basin gained and maintained power during the Upper Formative Period (500 BC- AD
400). The site of Pucar (200 BC- AD 400), located in the northwestern Basin in Peru, was one of
two regional population centers during the Upper Formative and covered an area of at least 1km2.

Following local convention, the spelling Pucar refers to the archaeological site and the modern town and the
spelling Pukara refers to the prehistoric culture. [note: since the presentation of this poster in 2001, we have
shifted to using the spelling Pukara to also refer to the archaeological site, following the conventions of the
Peruvian National Institute of Culture]

Klarich_Craig_SAA2001.doc

Page: 2

Monumental architecture, exotic goods, and complex residential areas are evidence for an elite
presence at Pucar, but the foundations of elite power are unclear. Excavation data from
households, the foci of production, consumption, storage, and distribution activities within
preindustrial societies (Hirth 1989: 441), will be used to determine what type of economic,
political, and ritual strategies were involved in the development and maintenance of Pukara elites.
Excavations by Alfred Kidder II in 1939 exposed a large compound in the central area of the site
with characteristics typical of an elite household (e.g. large size, quality of construction, central
location, both domestic and non-domestic features). In order to develop an excavation strategy at
a site that is both large and without surface architecture, a geophysical survey was conducted
during September of 2000.

In this poster results are discussed, including descriptions of significant subsurface features
encountered, comparisons are made with the results of previous excavations, and future
excavation strategies are outlined.2

Further information on the project is available at http://titicaca.ucsb.edu/pukara

Klarich_Craig_SAA2001.doc

Page: 3

Background to Pucar
Pucar was one of two major population centers in the Titicaca Basin during the Upper Formative
Period, contemporaneous with early occupations at Tiwanaku in Bolivia (Stanish n.d.). The site is
located approximately 80 km northwest of the shores of Lake Titicaca at an elevation of 3950
masl. The reported size limits of the site vary from 150 ha (Chvez 1992) to 6km2 (Mujica 1991),
illustrating the lack of consensus on the size and scope of Pucar. The site has been characterized
as an early urban center (Mujica 1978; Rowe 1963), a ceremonial center (Chvez 1992: 42), the
center of a complex chiefdom (Mujica 1991; Stanish et al. 1997), and more generally as the
center of a ceramic style region (Steadman 1995). Previous archaeological research at Pucar has
clarified the layout of the monumental public architecture and surrounding mound complexes,
defined Classic Pukara decorated ceramics and monoliths, and determined two areas of distinct
residential architecture (Chvez 1992; Franquemont 1986; Kidder 1942; Mujica 1991).

Klarich_Craig_SAA2001.doc

Page: 4

Two major excavation projects and a limited testing project have been conducted at Pucar. In
1939, Alfred Kidder II directed the first and most extensive excavation project at the site. Kidder
excavated in six areas: I, II, and III near the riverbank; IV on the alluvial plain (or pampa) to the
east of the monumental architecture; and V and VI on the terraces and main platform of the
Qalasaya complex. In the 1950s, Kidder and Manuel Chvez Balln returned to the site to collect
samples for C-14 dating (Franquemont 1986; Lumbreras and Amat 1968; Paredes 1985;
Steadman 1995; Wheeler and Mujica 1981). The second large-scale excavation project was
sponsored by COPESCO/ UNESCO and conducted from 1975-1980 (Paredes 1985; Wheeler and
Mujica 1981). The project included additional excavations of the sunken court area, which
established the presence of a pre-Pukara construction under the Classic Pukara courts, and
reconstruction of the terraces and stairways of the Qalasaya complex. A report published by
Wheeler and Mujica (1981) provides a summary of the architectural features, pottery types, and
faunal remains from Formative, Late Intermediate, and Late Horizon occupations of the site.
Kidders 1939 excavations, specifically the exposure of two very distinct areas of domestic
architecture and refuse, serve as the empirical foundation for organizing the proposed research.
The first of these areas included an excavated structure from Area III, near the banks of the Ro
Pucar. Kidder described the structure as a non-elite domestic area dated to the Classic Pukara
Period (summarized in Chvez 1992: 59-68; Kidder ms). Excavations in Area IV, located in the
central area of the site, exposed a second area with a domestic component with the following
characteristics:
Layout
Scale
Construction Materials

Architectural Features
Other Features
Materials Collected

Compound subdivided into structures, open areas, a mound/ platform.


The two principal exterior walls measured ca. 35 m. in length.
Exterior compound walls were constructed of large, worked stones
with a probable adobe superstructure. Multiple walls subdividing the
compound were built with unworked stones.
Rooms/ structures, walls, stone-lined canals, & pebble flooring.
3m. baked clay hearth, human cyst burial, a deposit of ca. 100 human
mandibles/ skull fragments deposited near a large slab.
Pukara decorated and plainware sherds, charred animal bones,
domestic refuse, & a carved monolith (prior to excavation).

Klarich_Craig_SAA2001.doc

Page: 5

Area IV, with its relatively large structure size (compared to the structure by the river),
complexity of architecture, high value objects, and proximity to monumental architecture fit all
of the common criteria used for characterizing elite households (Smith 1987). The nature and
function of this area, and its relationship to both the area of monumental architecture and nonelite structures by the river, can only be clarified through further excavations.

Rationale for Geophysical Survey at Pucar


Due to the nature of building materials utilized in this region, there is little to no visible
architecture on the modern ground surface. In Area IV, Kidder noted that he would not have
excavated in the area if local informants had not reported finding a monolith there (Chvez
1992:59). At present, the central area is quite flat, with only a few slight mounds and linear
features that may represent subsurface structures or walls. Surface pottery is also scarce and
predominantly dates to the Late Intermediate Period occupation at the site (post- AD 1100, see

Klarich_Craig_SAA2001.doc

Page: 6

Carlevato 1988). In order to determine if additional compounds were located in this central
pampa area of the site we conducted a geophysical survey to the north of Kidders Area IV. With
geophysical techniques we hoped to locate areas of architecture, differentiate compounds, and
detect features within those areas.

Geophysical Methods3
Geophysical remote sensing detectors are used here as prospection and discovery tools to locate
and identify subsurface anomalies. Interpretation of these sub-surface anomalies is then used to
direct future excavations. In general, geophysical surveys are useful when sites are large and the
archaeological features a researcher wishes to investigate are very specific.
It is useful to utilize more than one method of geophysical survey when a site is stratigraphically
complex or has multiple periods of occupation (Greenfield 2000: 167), as is the case at Pucar.
Use of multiple methods of data collection is also important when trying to target specific
archaeological periods if sites are vertically stratified and if there has been horizontal
displacement by water movement or by plowing. At Pucar, there have been multiple periods of
occupation and the site has been heavily plowed during historic times.
Two kinds of geophysical remote sensing detectors were used for the survey. These were a
Geometrics 858 a cesium magnetometer and Geophysical Survey Systems International (GSSI)
SIR-2000 ground penetrating radar (GPR) with a 400 megahertz antenna. To complement the
geophysical research, areas of the site were also mapped using a Leica TCR1105 Total Station
and Trimble Geoexplorer II GPS receiver. Data from the magnetic survey were processed using
Geometrics MagMap2000 software, while GPR data were processed with the RadanNT software
package from GSSI. Surface maps generated by the Leica Total station were imported onto a
laptop computer using the Leica Survey Office software, and GPS data were processed and
exported using the Trimble Pathfinder Office software package. Output from each of these four
data sources was reassembled in a common map projection and coordinate system using ArcView
3.2 GIS software from Environmental Systems Research Institute (ESRI).
The Geophysical Survey: Field Methods
When conducting geophysical survey, space is generally partitioned into grids that can be
surveyed in regularly spaced transects (Breiner 1999: 12; Conyers and Goodman 1997: 25). Once
recorded by a detector, these field data can be imported to a computer for further processing and
visualization to delineate anomalous entities within the data. Increasingly, researchers are
importing the results of surveys into GIS systems so that the results of more than one type of
survey can be viewed simultaneously. This aids tremendously in the identification of patterns,
and frequently leads to the discovery of subtle patterning in multiple data sets that may not be
visible with a single method of remote sensing.
At Pucar, a series of four 50m x 50m grid blocks were laid out using the Leica Total Station.
The magnetometer survey was collected along North-South oriented lines and the ends of the
transect lines were marked every meter using wooden stakes. The GPR survey was collected
along East-West oriented transect lines and the eastern and western ends were marked at one3

Special thanks to Mark Aldenderfer of the University of California, Santa Barbara, for the use of his equipment and
crew for the geophysical survey at Pucar.

Klarich_Craig_SAA2001.doc

Page: 7

meter intervals. Additional magnetometer blocks were surveyed in order to determine the shape,
orientation, and nature of the magnetic anomalies detected in the first four blocks surveyed. The
area was expanded with two 20m x 50m blocks to the east and a 30m x 50m block to the south of
the original survey area. Time did not permit the survey of these additional blocks with the GPR.
In addition to the geophysical survey, the pampa area was mapped in detail using the Leica Total
Station and the GPS. Field walls, both prehistoric and modern, were mapped, as were the
boundaries of modern agricultural fields, surface features, and Kidder's backdirt piles and
excavation areas from 1939. Using a map of Kidder's excavation Area IV, we were able to
determine the orientation of the architecture, the size, and which walls were still exposed on the
surface. This guided our survey strategy and helped greatly in interpreting the results from the
geophysical surveys.
Magnetometry

Magnetometers have been used to identify numerous buried anthropological entities including:
walls, structures, pottery, bricks, roof tiles, fire pits, buried pathways, tombs, buried entrances,
and monuments (Breiner 1999: 45). At Pucar the G-858 was used in a pre-excavation context
(Greenfield 2000: 167) primarily as a discovery tool to aid in understanding the location and
arrangement of sub-surface deposits. Detection of anomalies in archaeology generally relies on
the fact that the object has stronger magnetic fields than the surrounding matrix.

Klarich_Craig_SAA2001.doc

Page: 8

Heating is known to increase magnetic susceptibility through the production of magnetite


(Marmet et al. 1999: 168). Before heating, small regions that are called domains located in each
magnetite crystal are randomly oriented. While hot, particularly at high temperatures, domains
reorient themselves. During cooling, domains tend to align themselves in the general direction of
the ambient magnetic field and parallel to each other. This realignment and parallel orientation of
domains creates a net magnetization that is fixed with respect to the object (Breiner 1999:10).
Some features like walls and tombs have been located because their construction resulted from
the displacement of uniformly magnetic soil and the features are represented in the magnetic
survey as voids (Breiner 1999: 45). Walls do not always have weak magnetic fields in relation to
the surrounding dirt. Depending on a walls construction material and the magnetic properties of
the surrounding soil, the wall may have higher or lower magnetic properties than the substrate.
Magnetograms, or rasterized output from a magnetometer survey (Neubauer and EderHinterleitner 1997: 179), are becoming an increasingly popular form of representation. A basic
understanding of magnetics is required to interpret magnetograms. Magnetic fields are polar,
having both a positive and negative component. Objects where both the positive and negative
poles can be observed are referred to as di-poles. Adjacent light and dark patches in the
magnetograms represent di-polar anomalies that are visible because of the greater intensity of the
local magnetic field emanating from some object buried beneath the surface.
After the survey was complete the magnetometer data were transferred to the laptop computer
where they were processed using Magmap2000. The data were plotted in Magmap2000s 2D
view and displayed with a hillshaded greyscale plot. The plot of each survey was captured to a
jpeg and imported into ArcView GIS where they could be aligned to real world coordinates for
comparison of anomaly patterning in the other data sources.
Ground Penetrating Radar

Klarich_Craig_SAA2001.doc

Page: 9

Ground penetrating radar has also been used extensively as a prospection and discovery tool in
archaeology (Arnold et al. 1997; Clarke et al. 1999; Conyers and Goodman 1997; Savvaidis et al.
1999; Tohge et al. 1998). The GPR method relies on the transmission of high frequency radio
waves (Conyers and Goodman 1997: 23; Arnold et al. 1997: 161). Electromagnetic energy is
propagated downward into the ground in a conical shape by an antenna. The two-way travel time
elapsed between transmission, reflection off of buried objects or discontinuities, and reception of
the wave at the surface by an antenna is measured by the detector. Since the GPR method
measures two-way travel time and intensity of energy reflected, it is possible to produce a depth
profile where the horizontal axis represents distance along a transect line and the vertical axis
represents depth in units of time (see table below).
The depth and resolution of a GPR survey are functions of the wavelength of the antenna used
and the local geologic contexts. Longer wavelength antennas are capable of detecting features
that are buried deeper, but with long wavelength antennas smaller features may not be detected.
Likewise, shorter wavelength antennas do not provide a picture of great depth, but permit the
detection of much smaller objects. The depth that an antenna can detect buried features is a
function of the soils relative dialetric permeability (RDP) (Conyers and Goodman 1997: 53).
Increases in soil moisture and salt contents lead to increasing RDP creating greater signal
attenuation. Lower RDP permits greater depth penetration.
Since GPR employs measurements of 1) distance along a transect line, 2) time since transmission,
and 3) intensity of wavelength reflection, there are many ways that radar data can be visualized.
Single transect lines viewed in cross section produce a linescan plot (see figures below).
Combining multiple transect lines into a single file and viewing all transect lines as a common
time since transmission produces a time-slice plot (see figures below). Linescan plots offer the
advantage that different depths, measured in time, can be observed. Linescan plots can be used to
examine site stratigraphy over large areas (Baker et al. 1997: 1098; Clarke et al. 1999: 107). On
the other hand, time-slice plots provide a view with horizontal extent permitting the resolution of
the shape and extent of subsurface targets (Savvaidis et al. 1999: 72). However, time-slice images
are limited to a single time interval. Linescan plots can be quite useful when trying to explore or
predict vertical relationships, while time-slice plots provide information on the extent of an
anomaly at a given depth.
Block Number
1
2
3
4

Slice Depth (ns)


4.57
8
3.16
4.31

Slice Width (ns)


1
0.1
0.8
0.8

Once the data were collected and transported to the laptop computer they were processed using
RadanNT. All of the radar surveys were assembled into a 3D project where the data were viewed
in a linescan plot (figures below). The four surveys were horizontal high pass filtered to remove
system ringing. The filter was set to 1023 scans. Each survey was then transformed with a
migration filter to correct for the hyperbolic refraction patterns produced by the conical shape of
the radar signal transmitted into the ground. Parameters for the migration filter were set using
Radans graphical interface. The results of these post-processing steps are shown below. Once

Klarich_Craig_SAA2001.doc

Page: 10

transformed, each block was plotted as a 3D cube where various time slices could be examined
for linear anomaly patterning. Screen shots of the time slices that showed linear anomalies were
captured to jpeg and imported into ArcView GIS where they could be aligned to their respective
real world coordinates.

Linescan Plot of GPR transect 1 in Block 1 unfiltered

Linescan Plot of GPR transect 1 in Block 1 after horizontal high pass filtering with the filter set to 1023 scans.

Linescan Plot of GPR transect 1 in Block 1 after migration filtering with filter parameters set using RadanNTs
graphical interface.

Klarich_Craig_SAA2001.doc

Page: 11

Results and Discussion


GPR and magnetic anomalies were found in each of the survey blocks, including both linear
anomalies and anomalies of limited lateral extent (Desvignes 1999: 86). A number of the linear
anomalies both extended across more than one survey block and some were perpendicular to the
orientation of the transect. This indicates that the anomalies seen do not represent instrument
error or some artifact in the data, but rather supports that the anomalies represent some subsurface
feature of anthropogenic origin.
Anomalies Visible in the Magnetic Survey
Linear anomalies can be seen in the magnetogram by linear arrangements of di-poles (Figure).
Patterns of small magnetic anomalies in a linear arrangement may likely be the product of wall
building. Kidder (m.s.) reports that sandstone and andesite were both used in the construction of
walls. Andesitic rocks should produce visible di-polar anomalies. Since they would have been
moved from their source deposit, we would expect the di-poles to have a random orientation.
Ferric sandstones, which are likely to be present since the local sandstones are reddish, may also
produce small or weaker anomalies. Like the andesite, these would have random orientations due
to removal from their source deposits.
Large globular anomalies were also seen in the magnetic survey. These globular di-pole
anomalies may either represent deep anomalies from a large source, or could possibly be
produced by smaller, shallower objects. These may represent thermal anomalies, such as a
cooking hearth or an area for firing ceramics. However, since this area is inhabited and plowed by
the local Quechua, there is a possibility that some anomalies are produced by historic or modern
pieces of metal. We made a dedicated effort to systematically examine the area of the
magnetometer surveys and remove fragments of metal on the surface.
Anomalies Visible in the Ground Penetrating Radar Survey
Linear anomalies are seen in the radar data by the patterned differences in reflection intensity at a
given time slice. Numerous linear anomalies were observed in the four GPR survey blocks.
Additionally, there are several places with clearly defined corners where roughly perpendicular
linear anomalies meet. Often where these corners meet, one can see either higher or lower radar
refraction patterns within the spaces bounded by linear anomalies.
Linear Anomalies found in both surveys
Once all the data were imported and properly aligned to their real world coordinates in ArcView,
it was possible to begin looking for consistent patterns existing in both forms of geophysical
survey. To do this we used the flicker method, which involves turning a layer off and on rapidly
to find anomaly patterns that were congruent in both of the surveys. If the anomaly was distinct in
the radar data and if there was some evident pattern in the magnetometer survey, then the
anomaly was digitized. In total, 59 linear anomalies were observable in both surveys. There are
many other anomalies that are observable with only one of the methods and these were not
digitized.

Klarich_Craig_SAA2001.doc

Page: 12

The patterns of lines interpreted from the geophysical survey are quite similar in their size and
layout to the walls that Kidder found in his Area IV excavations. Linear anomalies most likely
represent walls, partitioning space into different activity areas. Aldenderfer defines two broad
categories for interpretation of the use of interior space open and busy based on the
presence or absence of features in the floor (1991). Open spaces, lacking features, would be more
conducive to gatherings of people. In contrast, busy spaces include the presence of pits, hearths or
platforms that most likely reflect room function and would have made the aggregation of people
difficult (Aldenderfer 1991:244). There are indications from the magnetometer survey that
similar variation in the use of space may be present at Pucar. Busy areas have multiple
magnetic anomalies that may represent wall fall, but they may also be areas where small fires
were set repeatedly or ceramics or heat-treated lithics were deposited. Conversely, there are
spaces that appear open and undifferentiated. These areas may be more complex once
excavated, but at this point there does appear to be variation in the use of space in this area of the
site. Lastly, there do not appear to be consistent patterns either of open or busy spaces in the radar
survey. Some open spaces appear as voids while others show a certain amount of patterning
within the bounded space.
Generally, where linear anomalies are encountered in the radar survey the area inside the bounded
space seems to have lower radar refraction patterns. However, this pattern does not always hold
true, and the variation probably reflects variation in the nature of the construction methods and
room contents. One thing seems clear the area inside these spaces is different than the outside.

Klarich_Craig_SAA2001.doc

Page: 13

The linear anomalies that were interpreted in both geophysical methods appear to be of roughly
the same size as the walls that Kidder found in his Area IV excavations. However, the orientation
of the walls appears to be different than those previously excavated. The Area IV walls are
oriented along similar lines as the terraces of the Qalasaya monumental architecture. The linear
anomalies found in the 2000 geophysical survey are quite different. It is important to note,
however, that the orientation of the linear anomalies is consistent within the two different sets of
survey blocks and are congruent between the two different kinds of detectors.

From Survey to Excavation

The ultimate goal of the geophysical survey was to develop an excavation strategy for a large
area (at least 300 m2) within a very large site (at least 1km2). Using the magnetometer and GPR

Klarich_Craig_SAA2001.doc

Page: 14

surveys we have determined that the area had a dense occupation and that there were at least two
different uses of space within the area (open and busy). The time slices of the GPR survey
can be used to argue that the deposits are shallow (above 10 ns.), which is also in accordance with
Kidders excavations in Area IV (Chvez 1992; Kidder ms).
For future excavations, we have decided to focus on Quad 2 based on the following factors:
density of anomalies (both linear and non-linear), presence of both busy and open spaces,
proximity to Kidders Area IV, and congruity between subsurface data and the limited surface
remains. Because of the nature of geophysical survey results, we have chosen to excavate in large
blocks and begin with shallow excavations to locate architecture. Three 5m x 5m excavation
blocks (divided into 1m x 1m units) will be placed in areas with anomalies in both surveys: one
open area, one busy area, and one block at an intersection of an open and busy area. The
blocks will be excavated to clear the plow zone (at a depth of 10-15 cm) and define architectural
units. The excavations will be expanded using the decapage method in order to expose large areas
and determine the contemporaneity of architecture, features, and activity areas. The results of the
excavations will then be used to further interpret the geophysical data from unexcavated portions
of the survey.

References
Aldenderfer, M.S.
1991 Continuity and Change in Ceremonial Structures at Late Preceramic Asana,
Southern Peru. Latin American Antiquity 2(3):227- 258.
Arnold, J., E. Ambos and D. O. Larson
1997 Geophysical Surveys of Stratigraphically Complex Island California Sites: New
Implications for Household Archaeology. Antiquity 71:157-168.
Baker, J., N. Anderson and P. Pilles
1997 Ground-Penetrating Radar Surveying in Support of Archaeological Site
Investigations. Computers and Geosciences 23(10):1093-1099.
Breiner, S.
1999

Applications Manual for Portable Magnetometers. Geometrics, Sunnyvale.

Carlevato, D.
1988 Late Ceramics from Pucara, Peru: An Indicator of Changing Site Function.
Expeditions 30(3): 39-45.
Chvez, K. L. M.
1988 The Significance of Chiripa in Lake Titicaca Basin Developments. Expeditions
30(3): 17-26.
Chvez, S. L.
1992 The Conventionalized Rules in Pucara Pottery Technology and Iconography:

Klarich_Craig_SAA2001.doc

Page: 15

Implications of Socio- Political Development in the Northern Titicaca Basin.


Unpublished PhD dissertation, Michigan State University.
Clarke, C., E. Utsi and V. Utsi
1999 Ground Penetrating Radar Investigations at North Ballachulish Moss, Highland,
Scotland. Archaeological Prospection 6:107-121.
Conyers and Goodman, L. and D. Goodman
1997 Ground-Penetrating Radar: An Introduction for Archaeologists. AltaMira, Walnut
Creed.
Desvignes, G., A. Tabbagh and C. Benech
1999 The Determination of the Depth of Magnetic Anomaly Sources. Archaeological
Prospection 6(2):85-105.
Franquemont, E.
1986 The Ancient Pottery from Pucara, Peru. awpa Pacha 24:1-30
Greenfield, H. J.
2000 Integrating Surface and Subsurface Reconnaissance Data in the Study of
Stratigraphically Complex Sites: Blagotin, Serbia. Geoarchaeology 15(2):167201.
Hirth, K.
1989

Domestic Architecture and Social Rank in a Mesoamerican Urban Center. In


Households and Communities, edited by M. Eachearn, D. Archer and R. Gavin,
pp. 441-449. Calgary Archaeological Association, University of Calgary, Calgary.
1993 The Household as an Analytic Unit: Problems in Method and Theory. In
Prehispanic Domestic Units in Western Mesoamerica, edited by R. S. Santley and
K. G. Hirth, pp. 21-36. CRC Press, Boca Raton.
Kidder-II, A.
1942 Preliminary notes on the Archaeology of Pucara, Puno, Peru. Actas y Trabajos
Cientficos de XXVII Congreso Internacional de Americanistas (Lima 1939)
1:341-345.
n.d.
manuscript- 1939 Excavations at Pucara (care of Elias Mujica).
Lumbreras, L.G. and H. Amat
1968 Secuencia aqueologica del altiplano occidental de Titicaca. 37th International
Congress of Americanists, 1966 (2):75-106.
Marmet, E., M. Bina, N. Fedoroff and A. Tabbagh
1999 Relationships between Human Activity and the Magnetic Properties of Soils: A
Case Study in the Medieval Site of Roissy-en-France. Archaeological Prospection
6(3): 161-170.
Mujica, E.

Klarich_Craig_SAA2001.doc

Page: 16

1978

Nueva hiptesis sobre el desarrollo temprano del altiplano, del Titicaca y de sus
areas de interaccin. Arte y Arqueologa 5-6:285- 308.
1991 Pukara: Una sociedad compleja temprana en la cuenca norte de Titicaca. In Los
Incas y El Antiguo Peru: 3000 Aos de Historia, pp. 272-297. vol. 1. Sociedad
Estatal Quinto Centenario, Madrid.

Neubauer, W. and A. Eder-Hinterleitner


1997 Resistivity and Magnetics of the Roman Town Carnuntum, Austria: an Example of
Combined Interpretation of Prospection Data. Archaeological Prospection 4(4).
Paredes, R.
1985

Rowe, J.
1963

Excavaciones Arqueolgicas en Pukara, Puno. Tesis de Licenciatura, La


Universidad Nacional San Antonio Abad del Cuzco, Peru.

Urban Settlements in Ancient Peru. awpa Pacha 1:1-26.

Savvaidis, A., G. Tsokas, Y. Liritzis and M. Apostolou


1999 Location and Mapping of Ancient Ruins on the Castle of Lefkas (Greece) by
Resistivity and GPR Methods. Archaeological Prospection 6:63-73.
Smith, M.
1987

Stanish, C.
nd

Household Possessions and Wealth in Agrarian States: Implications for


Archaeology. Journal of Anthropological Archaeology 6:297-335.

Ancient Collasuyu: The Evolution of Social Power in the Lake Titicaca Basin of
Peru and Bolivia (manuscript in possession of author).

Stanish, C., E. de la Vega M., L. Steadman, C. C. Justo, K. L. Frye, L. O. Mamani, M. T. Seddon


and P. C. Chuquimia
1997 Archaeological Survey in the Juli-Desaguadero Region of Lake Titicaca Basin,
Southern Peru. Fieldiana 29 (Publication 1488).
Steadman, L.H.
1995 Excavations at Camata: an Early Ceramic Chronology for the Western Titicaca
Basin, Peru. Unpublished PhD dissertation, Department of Anthropology,
University of California, Berkeley
Tohge, M., F. Karube, M. Kobayashi, A. Tanaka and K. Ishii
1998 The Use of Ground Penetrating Radar to Map an Ancient Village Buried by
Volcanic Ash. Journal of Applied Geophysics 40:49-58.
Wheeler, J. and E. Mujica
1981 Prehistoric Pastoralism in the Lake Titicaca Basin, Peru (1979-1980 Field
Season). Report submitted to National Science Foundation.

You might also like