ANTHROPOLOGICAL PA-PER-S

-OF'AURLISOY
AUA ITR

'T
AMERICAN MUSEMO
THE.
I

SEUK

V'OLUME, XXXIV, PARTII

:EX-CAVATIONS`,AT, TIAHIUANACO.
ly~W. C. BENNE~TT'

B3i Ou-DN Q T
THE

AM(ERIAN

TRUTE

MUSEUM OF NATURAL:HISTORY.
NWm YORiK Cm
1934

THE AMERICAN MUSEUM OF NATURAL HISTORY

PUBLICATIQNS IN ANTHROPOLOGY
In 1906 the present series of Anthropological Pape was authorized by the
Trustees of the Museum to record the results of researchpconducted by the Depart-..
ment of Anthropology. The series comprises octavo volumes of about 350 pages
each, issued in parts at -irregular intervals. Previous t 1908 articles devoted to
anthropological subjects appeared as occasonal pap in the Bulletin and also n the.
Memoir series of the Museum. Of -the Anthropological Papers 33 volumes have
been completed and 2 volumehave-- been iued zin part. A complete list of these
publications 'with prices will, be furnished when requested. All comxunicatioza
should be addressed to the-Librarian of the Museum..'
The current volume is:-,
VOLUME-XXXIV:I. Navajo Texts. By- Plin-y Earle Goddard. Pp. 1-179. 1933. Price,4$1.75.,
II. Kinship in the Admiralty Islands. By Margaret-Mead. Pp. 181-358,
and 17 text figures. 1934.- Price, $1.75
III. Excavations at Tiahuanaco. By W. C. Bennett. Pp. -359-494, and 35
'text figures. 1934. Price, $1.50

ANTH ROPOLOGICAL PAPERS

OF

THE AMERICAN MUSEUM OF NATURAL HISTORY

VOLUME XXXIV, PART III

EXCAVATIONS AT TIAHUANACO
-

BY W. C. BENNETT

BY ORDER OF THE TRUSTEES

OF

THE AMERICAN MUSEUM OF NATURAL HISTORY

NEW YORK CITY
1934

EXCAVATIONS AT TIAHUANACO
BY W. C. BENNETT

359

PREFACE
During the six months from March to September, 1932, Mr. John
G. Phillips and I made a survey of the archaeological possibilities in the
southern highlands of the Andes, namely on the altiplano of Bolivia and
the southern region of Peru. This trip was made possible by the patronage of Mr. Frank Phillips. Our purpose was to continue the survey work
started by Doctor Ronald L. Olson on the coast of Peru, and to retrace
the steps of Adolph F. Bandelier to understand better his extensive notes
and collections now in the American Museum of Natural History.
The first trip took us from La Paz, around the southeastern corner
of Lake Titicaca, from Achacachi, Macalaya, Ancoraimes, Carabuco to
Escoma, and inland over the Andean divide to Timusi, Chuma, Mocomoco, and Italaqui. On this trip we located, described, photographed,
and prepared groundplans of over forty Inca type villages and fortifications. These were composed mainly of aggregations of houses and walls
built of split, but not squared, mountain stone, piled without any
cementing of adobe or mud. The houses had narrow doorways, niches
and shelves, but no windows, and corbeled arch stone roofs. Both
round and square houses were found. Forts were situated on prominent
points. A typical fort consisted of a square patio surrounded by small
houses with all the doorways facing inward. A series of three or more
thick walls surrounded this unit in concentric circles. Many of these
ruins contained crudely built towers, usually with several skeletons inside.
The second trip was to Cochabamba and the surrounding region.
Here we visited mound sites which promised fruitful excavations. An
Inca type village site at Illuri indicated stratification. We visited Incallacta, a famous late ruin which Nordenskiold described many years
before.
The Bolivian Government granted us permission to excavate ten
test pits of not more than ten square meters in surface area each, at
any spots about the ruins of Tiahuanaco which we chose, the depth of
the pit to be determined by the thickness of the cultural strata. This
excavation was carried on during the months of June and July of 1932.
Since the Tiahuanaco site is without doubt the most important archaeological center in Bolivia, and since our work revealed ceramic stratigraphy of importance to the whole Andean problem, it is this section of
our work which is the subject matter for the present paper.
After the Tiahuanaco work we visited the Islands of Titicaca and
Coati, and the Peninsula of Copacabana, where we were, in truth, follow361

362

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

ing the trail made by Bandelier thirty-five years before. The islands
proved interesting for their possible Tiahuanaco affiliations, as well as
for their excellent examples of Inca architecture.
Leaving Bolivia, we proceeded to Cuzco and its environs. Cuzco
is a famous archaeological center and contains, within the present
town, almost a complete survey of the Inca culture. We visited Sacsahuaman, the fortress above Cuzco, Ollantaytambo, and Macchu Pichu,
famed citadel of the Inca described by Hiram Bingham.
Again returning to the coast of Peru we spent some time at Lima,
examining the ruins in the vicinity and the museum collections.
While it is not possible to mention the names of all the multitude of
friends who assisted our work in Bolivia, we would like to extend our
appreciation, together with that of the American Museum, to the Right
Honorable Daniel Salamanca, President of the Republic of Bolivia,
and to the late Honorable Sr. Alfredo Otero, Minister of Instruction, for
their interest and co6peration in allowing us to excavate at Tiahuanaco;
to the American Minister Edward F. Feeley, and to other members of
the American diplomatic and consular service, Messrs. Robert Joyce,
Paul Daniels, and Robert Fernald, for their kind assistance; to Sr.
Juan Perou, for graciously providing us a home at his ranch while working at Tiahuanaco; to Professor Arthur Posnansky, Sr. Frederick Buch,
Doctor Alberto Villegas, Sr. Luiz Hertzog, and Mr. Kenneth Manning,
for their interest and actual assistance in our work.
The report which follows covers only the excavations at Tiahuanaco.
The survey work will be continued in 1934.
WENDELL C. BENNETT.
November 18, 1933.

CONTENTS
PAGE
.361
.367
.369

PREFACE
INTRODUCTION
EXCAVATIONS
PIT I
PIT II
PIT III
PIT IV
PIT V
PIT VI
PIT VII.
PIT VIII.
PIT IX
PIT X
THE COLLECTION
POTSHERDS
Plain Wares.
Extra Thick Ware
Thick Polished Ware
Painted Line on Plain Ware
Painted Ware
Polished Ware
One-Color Ware
Two-Color Ware.
Three-Color Ware
Four-Color Ware

Special Painting
Isolated Fragments
POTTERY DESIGN.
Classic and Decadent Style
POTTERY SHAPES.

Shape A, Kero-Shaped Cup

Shape B, Hollow Base Libation Bowl
Shape C, Flaring Rim Bowl
Shape D, Vase
Shape E, Small Flaring-Sided Bowl
Shape a, Horizontal Handled Dishes
Shape b, Shallow, Plain Rim Bowl
Shape c, Small, Flaring Rim Bowl
Shape d, Narrow Rim, Wide Open Bowl.
Shape e, Modeled Puma Bowl
Shape f, Goblet
Shape g, Modeled Heads
Shape h, Angular Rim Bowl
Shape i, Narrow Neck Globular Bowls
Shape j, Round Bottom Bowls .
Shape k, Small Open Wide Rim Bowls
363

.369
.372
.375
.375
.378
.385
.385
.387
.390
.391
.392
.392
.392
.394
.394
.394
.395
.396
.397
.397
.398
.399
.399
.399
.400
.403
.406
.406
.408
.412
.413
.413
.414
.414
.414
.414
.416
.416
.416
.416
.416

.417
.417

364

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

PAGE

. .417
Shape 1, Round Base Open Bowl
Shape m, Large, High Shouldered Jug . .417
..418
Shape n, High Shouldered Vase
..418
Shape o, One-Handled Pitcher
..418
Shape p, Angular Body Bowl
..418
Shape q, Spouted Bowl
..418
Shape r, Animal Effigy Handle
..418
Shape s, Small Modern Vases

COMPLETE AND PARTIALLY COMPLETE BOWLS.

.419

..424
*4 Clay
. .425
. .425
Stone
427
Copper .
. .427
Bone
.428
Shell
..428
STONE SCULPTURE
..429
The Large Monolith
439
The Small Monolithic Statue .
..443
A Slightly Carved Stone
..444
Stone Heads
..444
Circular Grindstone
. .445
CULTURAL STRATIGRAPHY
..448
EARLY TIAHUANACO
..453
CLASSIC TIAHUANACO
..456
DECADENT TIAHUANACO
..459
POST TIAHUANACO AND INCA
. .460
TIAHUANACO STONE SCULPTURE
464
STYLE 1, THE CLASSIC MONOLITHIC STATUE. .
ARTIFACTS

466
STYLE 2, VIRACOCHA, GATEWAY OF THE SUN FIGURE. .
.467
STYLE 3, SQUARE PILLAR TYPE STATUES .
468
STYLE 4, ANIMAL HEAD SEATED FIGURES IN SITTING POSITION
468
STYLE 5, KNEELING HUMAN STATUES
469
STYLE 6, STONE HEADS
470
Substyle 6a, Headband Type
470
Substyle 6b, Flat Stones with Faces.
471
Substyle 6c, Rounded Boulder Heads
471
Substyle 6d, Modeled Face Heads
471
Substyle 6e, Cornerstone Faces.
472
Possible Additions to Style 6
472
STYLE 7, CARVED LIZARD, SALAMANDER, OR TOAD
473
STYLE 8, GEOMETRIC PATTERNS.
. 473
MISCELLANEOUS PIECES
474
Group I, Realistic Stone Carving
474
Group II, Conventionalized Classic Figure and Head
474
Group III, Technically Decadent Pillar-Like Statues and Head
475
Group IV, Geometric Style Carving
. 476
CONCLUSIONS.
492
BIBLIOGRAPHY

LIST OF ILLUSTRATIONS
TEXT FIGURES

PAGE
1. Plan of the Ruins of Tiahuanaco (after Posnansky) showing the Location
of the Pits excavated by the Expedition .facing 370
2. Groundplan and Cross-section of Pit I
..371
3. Groundplan and Cross-section of Pit II
..373
4. Groundplan showing Position of Path and Pit, and Cross-section of Pit IV 377
5. Groundplan of Pit V
...379
6. Cross-section of South Wall of Pit V indicating the Stratification Lines of
Stone, Ash Beds, and Clay
..380
7. Cross-section of West Wall of Pit V
..381
8. Cross-section of East Wall of Pit V, confirming the Analysis of the
Previous Sections
..382
9. Cross-section of North Wall of Pit V, showing the Position of the Canal
384
10. Groundplan of Pit VIII
..388
11. Cross-section of North Wall of Pit VIII
..389
12. Key to the Pottery Shapes Described
..407
13. Shapes and Designs on Early Type Tiahuanaco Pottery
410
14. A Typical Early Tiahuanaco Bowl Shape with Four Examples of the
Type of Inner Rim Design
. .411
15. Shapes and Designs of Classic Tiahuanaco Types. .
.415
16. Classic Tiahuanaco Shapes and Designs
..420
17. Shapes and Designs of Decadent Tiahuanaco Style . .422
18. Pit VII showing Arrangement of Sculptured Stone . .429
19. Drawing of Large Monolith in Pit VII, indicating Principal Design Areas
which are described in the Text
. .430
20. Large Monolith lying on its Back in Pit VII, now mounted on the Prado
in La Paz
. .432
21. Design on the Side of the Headband of the Large Monolithic Statue
433
22. Design on Side of Head of Large Monolithic Statue . .433
23. Elaborate Design from Center of Chest of Large Monolithic Statue
434
24. Left Hand of Large Monolithic Statue, showing the Five Proportioned
Fingers ..
435
25. Right Hand of Large Monolithlic Statue, illustrating the Distorted
. .435
Position of the Thumb and Four Fingers
26. Design below the Forearm of the Large Monolithic Statue.
436
27. Elaborate Design on the Side Arm of the Monolithic Statue .
436
28. Waist Design of the Monolithic Statue showing the Running Human
. .437
Figure
29. Design on the Center of the Waist of the Monolithic Statue between the
Two Running Human Figures
.438
30. Waistband Design on the Monolithic Statue . .439
31. The Smaller Bearded Statue beside the Large Monolith in Pit VII
440
32. Front, Side, and Back View of Smaller Bearded Statue of Pit VII
441
33. Slightly Carved Stone of Pit VII
..442
34. Edge of Ceremonial Grindstone found in Pit VII . .443
35. Chart of Pits and Levels arranged according to Periods
446
365

I.

INTRODUCTION
The Tiahuanaco ruins are situated on the high plateau of Bolivia,
south of Lake Titicaca. The Tiahuanaco station on the Guaqui to La
Paz railway is 21 kilometers from Guaqui. It is 3,825 meters above sea
level. Posnanskyl gives the exact location as 160 33' 26" latitude south
and 680 48' 6" west of Greenwich. The principal ruins lie within an
area about 1,000 meters from east to west by 450 meters from north to
south.2 Geographically, Tiahuanaco lies in a valley about 11 kilometers
wide, between the ranges of Kimsa-chatta to the south, and Achuta to
the north. Although Lake Titicaca probably covered this territory in
past geological times, there is now a difference of 34.73 meters between the
level of Tiahuanaco and the lake.
The ruins of Tiahuanaco have long been famous. The artificially
fortified hill of Acapana, the megalithic enclosed structure of Calasasaya,
the elaborately jointed stone architecture of Puma Puncu, and the
artistically carved frieze of the Gateway of the Sun, are only a few of the
wonders that have attracted the attention of traveler and archaeologist.
From Cieza de Leon, who visited the site before 1550, to the present day,
numerous writers have devoted chapters to impressions and descriptions
of the ruins. Stiubel and Uhle in their prodigious account of Tiahuanaco
give an excellent survey of the principal bibliographic references to
Tiahuanaco, in which they include direct quotation of the most pertinent
passages.
In this century many writers have described Tiahuanaco with varying degrees of elaborateness, but comparatively little scientific work has
been done. Sttbel and Uhle include in their book the results of their
rather limited excavating, as well as detailed descriptions and measurements of most of the important structures and stones exposed at the
ruins in their time. For three months in 1903 Georges Courty carried
on excavations in which he unearthed many new structures (e.g., the
"monolithic " stairway; the small eastern temple; the "cloaca maxima"
of Acapana; the "palacio"). Unfortunately, only a brief summary of
his work was ever published.3 Arthur Posnansky has concentrated for
the past thirty years on the problems of Tiahuanaco. His work is
published in the volume, Eine Praehistorische Metropole in Sudamerika,
as well as in many short papers. Several other local workers have made
large collections at the ruins, but have not published accounts of their
labors.
1Posnansky, 1914, 78.
2As computed by Means, 1931, 123.
3Crequi-Montfort, 531-551.
367

368

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Posnansky's plan of the ruins is admirable. Between Posnansky and

Uhle and Stiibel accurate pictures and descriptions are given of the
existing structures. The principal lack in our knowledge of the ruins is
a stratigraphical pottery series. It is generally conceded that at least
two, if not three, periods or phases of culture are represented at Tiahuanaco. The distinction is based on stylistic rather than stratigraphic
proof. All indications at Tiahuanaco point to the probability of good
stratigraphy, not of buildings, but of potsherds. There are many evidences of a considerable population and an extended time period, both
of which are conducive to stratification in a limited area.

EXCAVATIONS
The permit granted by the Bolivian government was for excavating
pits with a surface area not over ten square meters. Sinking a pit directly
into the ground is not the most ideal manner of excavating. The practical difficulties of handling dirt become enormous when any depth is
attained. However, the method has some advantages in small scale
excavating at Tiahuanaco, where the surface of the land about the ruins
is comparatively level.
The pits were excavated one at a time in order that the work could
be personally supervised. The arbitrary unit of one half meter was
adopted for depth levels. While in some pits a smaller unit might have
proved more desirable, the half meter measure was generally satisfactory. The material from each half meter depth was separated into boxes
and classified as a unit. No selection was made at the time of excavating,
and the workers were instructed to save every artifact, even to the last
plain potsherd. Excavation was icontinued in each pit until undisturbed sand or clay was reached. In two pits (IV and VIII), water
seepage prevented the completion of'the work.
Obviously an arbitrary depth unit of one half meter does not often
coincide exactly with a cultural depth division. The pit excavation
method prevents the determination of any horizontal stratification until
the work is completed, thus making it imperative to establish arbitrary
units of some kind and coordinate them with the cross-section later.
The present chapter includes a field descrip5tion of the pits in the
order of their excavation. The plan included here -shows the location of
the pits (Fig. 1).
PIT I

This pit was located on the north side off the cut made by the

Guaqui to La Paz railway, 420.3 meters west of Tiahuanaco station, and

just south of the outer wall of a small church. ''-The railioad lin64ihakes a
cut about four. meters deep through this section and evidence of two
meter accumulation is everywhere visible. The soil from 'the railkbad cut
was not piled on the sides, but was used as a'fill a little further to the
west. The site is near the reported location of the' old Tiahuanaco
burial ground.
The pit was 5 by 2 meters, running parallel with the railroad and the
church wall.
i)
Level 1, 0-0.5 meters deep. The top soil was ext'ry6ha`rd and
contained very little material. Beneath this the ground was soft and the
369

370

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

appearance of the soil suggested relatively recent disturbance. A fragment of glass and some pieces of modern pottery confirmed this opinion.
Three shallow burials were encountered. They appeared modern. Very
little pottery was found.
Level 2, 0.5-1.0 meters deep. The soft soil continued and the whole
level was filled with modern burials, nine in total. Some were in extended position and others were scattered. Neither skull deformation
nor other peculiarities gave any suggestion of antiquity to these burials.
A few fragments of broken and cut animal bone and some pottery were
the only artifacts discovered.
Although an old cemetery is located in the yard of the church next to
this site, I could find no record of burial outside the church, although the
custom is common enough. In order to subdue the excitement caused
by the first unit of excavating, all of these bones were re-buried when the
pit was filled in.
Level 3, 1.0-1.5 meters deep. The soil of this level was mostly sand
and only slightly disturbed. In the center, however, was a bed of heavy
black ash. This was oval in shape, about 1 by .75 meters and it extended
from 1.0 to 1.4 meters deep. The collection for this level was separated into two divisions. Level 3 refers to all the material, exclusive of the ash bed,
while the contents of the bed are designated Level 3a. Level 3, then, furnished some animal bones, an ordinary grindstone, and some plain pottery.
Level Sa, 1.0-1.4 meters deep, ash bed. The ash bed contained a
considerable quantity of pottery as well as a partially complete kero and
a small flaring bowl. The pottery was mostly cooking rather than
mortuary ware, thus negating any possible connection with the burials
above. Stratigraphically Levels 3 and 3a seem distinct from Levels 1
and 2.
Level 4, 1.5-2.0 meters deep. This level consisted of pure sand,
apparently undisturbed, except for one section at the east end of the pit.
No artifacts of any kind were found.
Level 5, 2.0-2.5 meters deep. The disturbed section at the east end
of the pit at Level 4 continued to a depth of 2.3 meters where the powdery
remains of a skeleton were found. As well as could be determined the
burial was intrusive; it was extended from north to south, with the
head at the north end. At the head end were a small flaring bowl
(broken), a kero (broken), and a narrow neck vessel which had once had
a spout (Cf. pp. 417, 418).
The analysis of the pit (Fig. 2) shows two levels. Levels 1 and 2
represent a modern stratum, while Levels 3 and 3a, and probably the

rr -

|O AT o 5N
.w

LIZ
a: PALACE
.PLC
EASV
\
INNER
X

I~i~~1 P17111 COURT

___

a3DOL
PIT Vil
A

ACAPANA

\\TIANUANPACO
POE SLO

Pit TIX
a

PIT

01

PAZ

ALroa

LA
-iAQIy
-~~~~~~~/~ATER

GODS

SCALE 1:3000
PeT
IV

PIJNCU
~~~~~~~~~~~~PUM-A
AND PIT VI
TO

Fig. 1. Plan of the Ruins of Tiahuanaco (after Posnansky) showing the Location of the Pits excavated by the Expedition.

0.1

GROUND PLA N

a!_-

at

H/umus soil (some sherds)

2

3a-

3
2.3
5

Layer of modern burials

ash bed 5ome culture

mfirs

~Undisturbed
FBurial wifh three po

ts

NORTH WALL CR055 SECTION

SCALE
I

M.

Fig. 2. Groundplan and Cross-section of Pit I. The relation of the intrusive

burial to Level 3 is shown, while both levels are under the upper modern levels.

371

372

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

intrusive burial of Level 5, represent a cultural stratum of Tiahuanaco

type unmixed with Inca ware. The analysis of style will be given in the
subsequent chapters.
PIT II
This pit lies within the megalithic enclosure of Calasasaya, just a
few meters out from the southwest corner of the inner temple. Measuring
from the southern corner of the projecting west wall of Calasasaya, the
pit lies 38 meters within the temple in an east by north direction.
There is some discrepancy in the historic accounts of the nature
of the interior of Calasasaya. Although the inner temple is generally
conceded to have been lower than the surrounding level, there is no such
concordance on Squier's statement' that the outer temple was in the
form of a "terre-plein" faced with the stone walls of which megalithic
traces remain. Indeed most of the travelers compare Calasasaya with
Stonehenge. Stilbel and Uhle2 discard speculation and insist on the
present configuration which they consider as generally level as compared
with the surrounding territory. Posnansky3 confirms the partial terreplein idea which he computes at 2.30 meters high (measured by the
height of the monolithic stairway-a factor which Uhle and Stiubel did
not take into consideration, because the stairway had not been discovered
in their time). Along the north and east walls of the temple Posnansky
even describes an outer supporting terrace which gave the stepped
appearance typical of Tiahuanaco structures. None of the authors,
scientific or otherwise, have published any cross-sections of the temple
as it exists today. Posnansky's view seems the most rational. The
temple was not built on an absolutely flat plain, and the interior floor
was probably leveled at a height which reached its maximum at the
lower east front where the monolithic stairway was located. Even today
the temple is generally higher than the surrounding plains. Consiaerable dirt must have been washed away, as is evident in the fill of the
inner temple and the small, semi-subterranean temple (more to the
east). The wash was sufficient to cover the monolithic stairway and the
large stone statue, of which more will be said later. Furthermore,
agriculture which would assist the wearing down of the terre-plein, has
been carried on within the temple enclosure up to very recent times.
Pit II measured 3.3 by 3 meters, with the long side parallel to the
west wall of the temple.
iSquier, 276-278.
2Part I, Tafel 2 and Part 2, 16.

aPosnansky, 1914, 108-109.

OL-4--

~L

I
I
______
______

~z-a

GROUND PLAN

atCt
_
Layer of squared stones.
0.5 __

Ash bed__

2.0

~~~Some ash

3ome

ash

Soft sedion -with

.5_ part of human skull.

Cloy
2.0

Clay
ao~ ~ ~~~~~~~Ca

Undi.sfurbed cfay.

5CALE

CRo55

SECTION

6!-a'

Fig. 3. Groundplan and Cross-section of Pit II. The confusion of the cut stone
arrangement in the top two levels contrasts with the elaborate finish of a pavement
discovered nearby during recent excavations by the Bolivian Government.

373

374

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Level 1, 0-0.5 meters deep. The soil of this first level was very solid.
Starting at a depth of 0.2 meters cut stones were encountered. Sixteen
stones were found, ranging in size from 16 by 11 by 9 centimeters to 70
by 45 by 15 centimeters. These stones were rather evenly distributed
throughout this level, but showed no signs of arrangement, as a paving,
for example. Some were on edge, others flat, still others oblique. Most
were of sandstone, a few of lava, and two of a fragile mudstone. All were
cut square, but neither carved nor jointed. One had a groove around
each end. Pottery, mostly plain ware, and bone fragments were scarce
in this level.
Level 2, 0.5-1.0 meters deep. The dirt continued with about the
same texture and solidity, though with some trace of ash. In the center
of the south half was a small ash bed and in the southwest corner, a
large one. Nineteen more cut and squared stones were found in this
level, with the same scattered arrangement. One stone was grooved,
but the others were plain building blocks. All of them might have served
for wall fill (between megaliths) or pavement.
Level 3, 1.0-1.5 meters deep. A few scraps of bone and pottery fragments, as well as a little ash, indicated that this level was disturbed.
The ground was softer than before. In the southwest corner a soft,
slightly ashy bed, furnished part of the frontal bone of a human skull,
though no trace of a burial was evident.
Level 4, 1.5-2.0 meters deep. The texture of the earth changed to
clay. Pieces of broken bone and a little ash indicated slight disturbance.
For security, the southern half of this pit was excavated to three
meters and the southwest corner to 3.5 meters in depth, but without
further indication of disturbance in the very solid clay.
No evidence of cultural stratification is furnished by the crosssection of this pit (Fig. 3). Levels 1 and 2 seem contemporaneous with
the cut stone blocks and Level 3 is too skimpy for separate consideration.
Since the pit was excavated well within the temple it seems unlikely that
the cut stones were from the outer wall, but would more probably have
been part of the paving or of some inner wall construction. Since for
several centuries surface stone of tractable size has been consistently
removed from the ruins for local building, it can safely be assumed that
the stones of Levels 1 and 2 have long been buried. Furthermore, on the
basis of this assumption, all three levels, as a unit, might be considered
as contemporaneous with the ex-terre-plein and thus the temple itself.

1934.]

Bennett, Tiahuanaco Excavations.

375

PIT III
According to Posnansky's survey (Fig. 1) the principal ruins of
Tiahuanaco were surrounded by an artificial moat or canal. The northwest corner of the area thus surrounded is comparatively free from surface
temple remains and thus seemed a likely place for dwelling site material.
Accordingly, Pit III was excavated in this section, about 245 meters
northwest of Calasasaya. The pit! was 3 by 3.3 meters in size.
Level 1, 0-0.5 meters deep. The soil was soft and mixed with considerable ash. Fragments of bone and potsherds were abundant.
Level 2, 0.5-1.0 meters deep. In the southeast corner a large fire
bed was found. More bone and potsherds were encountered. Before
the meter was completed, solid clay which showed no trace of disturbance,
was encountered. Tests to a depth of 2 meters verified its undisturbed
nature.
The potsherds were overwhelmingly plain, the first level contained
only 0.34 per cent painted ware and the second level 100 per cent plain
ware. Analysis is difficult in! this situation.
PIT IV
The section on the south side of the railroad cut, diagonally opposite
(west) of Pit I, was said by all my Indian workers to be the richest site
around Tiahuanaco. The section along the railroad indicated much
greater richness of material than Pit. I had furnished, and} so I was not
averse to excavating again in this locality. Pit IV was 26 meters farther
west along therailroad track, and 15 meters to the south of it. The pit
was 3 by 3.3 meters, with the long side parallel to the track.
Level 1, 0-1 meter deep. The first half meter was badly mixed with
dirt piled up by the railroad workers. Unfortunately, this material got
mixed with the artifacts from the second half meter and so the whole
had to be treated as one level. In the second half meter, near one meter
depth, was a heavy ash bed in the east side of the pit. Here were found
innumerable broken bones and potsherds of both Inca and Tiahuanaco
type. In general the Inca sherds were stratigraphically higher. The
ash bed was about 30 centimeters thick.
Level 2, 1.0-1.5 meters deep. Just below one meter depth a row of
flat stones (two stones wide) ran diagonally across the pit. This was
definitely a path and so marked an old; ground surface. The ash bed,
mentioned in Level 1, confirms the idea of an old ground surface at this
depth. Beneath this path, near the center of the south side of the
excavation, the soft texture of the soil indicated a pit. This pit, a little

376

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

over one meter in diameter, continued to 3.5 meters depth. Since it

formed a distinct unit in the excavation it has been treated separately
in the analysis. Consequently Level 2 refers to the material from 1.0 to
1.5 meters deep, exclusive of the contents of the pit. Inca sherds, most
of them from the upper part, were found in this level. Since the indication of an old ground surface divided this level from Level 1, I am inclined
to class the Inca ware as intrusive, that is, belonging to Level 1 rather
than Level 2.
Level 3, 1.5-2.0 meters deep. Like Level 2 this section also excludes
the pit material. Before two meters depth was reached undisturbed
sand has been found in all parts, except the pit before mentioned. No
Inca ware was found in this level, nor was there any demarcation between
Levels 2 and 3.
Level 4, 1.0-2.0 meters deep (pit). The first meter of this pit was
stone-lined on the outside. This lining did not continue to the bottom
of the pit. No Inca ware was found.
Level 5, 2.0-3.0 meters deep (pit). Many potsherds mixed with dirt
and ash were found, but no trace of burial. The pit extended into the
undisturbed sand which lined it on all sides. It was clearly demarked
by the soft texture and dark color in contrast to the yellow undisturbed

sand.
Level 6, 3.0-3.5 meters deep (pit). Considerable water seeped into
the pit at this depth. Consequently, any burial in it would probably
have disintegrated. However, the conglomerate fill of the pit did not
suggest burial, but more refuse. A complete jar was found (broken)
at this level. Half of one potsherd was found in the upper part of the
pit (Level 4) and the other half in Level 6, indicating that the whole pit
is but a single cultural unit.
A summary of the cross-sectional stratigraphy represented by Pit
IV (Fig. 4) shows first a mixed layer, one meter thick, containing a good
number of Inca as well as Tiahuanaco ware fragments. This mixed
layer is definitely divided from the others by an old ground surface, as
indicated by the stone path and the ash bed. The pit, Levels 4-6,
starts just below this walk, which is actually over it, suggesting that the
pit is older than the walk and therefore older than Level 1. Since the
pit starts at this point and shows no inner stratigraphy, it must be
considered as one cultural unit. No Inca pottery occurs in the pit.
Levels 2 and 3 lie stratigraphically below Level 1. Since there is no
demarcation between them, they may be- considered as a single cultural
unit, older than Level 1. This conclusion neglects the presence of Inca

-I
~~--__

-ug e
a'2
Rso
fo,

---

PIT

't

w_\7

~~~~~
__

GROUND PLAN

Ash, sherds, Inca relics.

os

--

_7

_I

Mfixed ash and sherds.

wall or pafh

C---=-

- _---_

__
-

2
154~

1-_

Mixed
4

mixed

2.0

sand

Sand

3.0

CROSS 5ECTION OF NORTH m SOUTH CENTER CUT

SCALE

Fig. 4. Groundplan showing Position of Path and Pit, and

Cross-section of Pit IV. A surface line, indicated by the path,
separates the upper level fr om the lower two and the pit.
377

378

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

ware in the Level 2 collection. However, this may be accounted for by

the fact that the old surface lay slightly below one meter depth, while
Level 2 (in accordance with the arbitrary division of half meters) started
at one meter. That the Inca ware is intrusive is confirmed by the pit
which starts at the old surface level and contains no Inca ware. Since
the pit (Levels 4-6) cuts through Levels 2 and 3 and starts at the top of
Level 2 it is probably of later date than those levels. Thus in Pit IV
there are three cultural divisions: Level 1; Levels 4-6 (pit); Levels 2-3.
PIT V
To the east of the fortified hill, Acapana, are the remains of a temple,
now barely discernible. The site is marked by a few dressed stones and a
low hill, or rise. Between this ruin and Acapana is a flat stretch of
ground about 150 meters wide. Pit V was made in this section, 126
meters to the east of Acapana. The pit was 3.3 by 3 meters with the
long side running east to west.
Level 1, 0-0.5 meters deep. A layer of humus, clay, and some ash
formed the first half meter. Many potsherds and broken bone fragments
were found. In the southwest corner, at 40 centimeters depth was an
extended (east-west) burial with a llama skeleton beside it. To uncover
this it was necessary to extend the pit somewhat to the southwest. To
the north of this burial was another extending into the west bank.
Level 2, 0.5-1.0 meters deep. In this layer the pottery fragments and
llama bone continued, though no more burials were found. At 65 centimeters depth a line of ash mixed with bone fragments and other material
stretched across the pit.
At one meter depth, in the southwest corner, a solid clay bed formed
a floor, the limits of which could not be fully determined. However,
in the southeast corner, several unworked stones were roughly grouped
at this same level.
Level 3, 1.0-1.5 meters deep. Against the south wall of the pit, west
of the center, two cut stones formed a platform resting on two smaller
stones. Perhaps this was an accidental arrangement, but the tops of
the stone were on the same level, one meter deep, as the clay floor mentioned in Level 2. Another- ash line, varying in thickness, extended
across the pit at a depth of 1.10 meters, dipping to 1.25 in the southeast
corner.
Level.4, 1.5-2.0 meters deep. The most interesting find in this half
meter was a canal in the north wall of the pit, running parallel to the
wall and clear across the pit. It dipped slightly to the east, away from

EAST

I-

be

WEST

GROUND PLAN

.40

:t.

CANAL SECTION

IT Y.

CANAL OF ROUGH STONE

SCALE

as

;,.

Fig. 5. Groundplan of Pit V, indicating Platform cut out of

Corner to facilitate Deep Digging, also the Position of the Covered Stone Canal, 1.70-2.0 M. Deep in Relation to the Pit, and
a Cross-section of the Canal.
379

380

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Acapana (Fig. 5). It was made of rough unworked stones. The inside
measurements were 25 by 35 centimeters and the outside were 40 centimeters wide and 30 centimeters deep. The canal was from 1.70 to 2.0

Humus level
2

nworlked sfones

8rial
As

)Cat
>
tostones_________
S
-----

bones|
-

-...

3As
Thin alsh
2.0~~~~~2

Pq~~~~ixed

Z5

Mixed

7 -8
4.0 ___

(top of Canal line)

Spec

Barren

h bed 7L
4

clay layer

Disturbed 3lightly.
Clay
7
-Te--

Ash, bone, sherds.

CROSS SECTION

OF

50UTH COT

SCALE

i.om.
Fig. 6. Cross-section of South Wall of Pit V indicating the Stratification Lines
of Stone, Ash Beds, and Clay.
0

0.5

meters deep. The top and bottom, as well as the two sides were lined
with stone, thus indicating that it was intended for use underground,
quite likely,for drainage. A thin ash line at the level of the top of the
canal extended as a shallow layer throughout the pit.

1934.]

J~Unkfore-d

Bennett, Tiahuanaco Excavations.

381

While the canal seemed intended for use underground there was no
evidence of intrusion which would suggest its depth at the time of construction. As the cross-section indicates, two layers of ash extend in
unbroken beds above it at 0.65 and 1.25 meters depth, respectively.
Thin though it is, the ash bed extends across the pit at the level of the
o

SOUTH

6
floor
3

AsThin

6
4.0

NORTH

sh liner

sfones

alixed

layer

__--~----__________________._________

7 Barrsni layer of clay washeua

d
ltones.
2.5a

2~0

time,
thoghherooisbyo

0---5---------

Ash, bonev poitery

CROSS5 SECTION

Of

WEST WALL.

SCAL E
0

0w5

O~~1. M.

Fig. 7. Cross-section of West Wall of Pit V. In the first level is a burial with
llama bones. The ash beds, a clay floor, the canal, and unworked stone layer are the
significant stratificatioii indicators.

top of the canal, suggesting that there was an old surface level at
time, though the proof is by no means convincing.

one

Level 5, 2.0-2.5 meters deep. Below the canal at 2.20 meters depth
is an ash bed. In the southwest corner at 2.40 meters depth was a bed
of flat unworked stones. Above this was a bed of ash and bone which

0
U.

0.5
2A
1.0

1l.5

C.ncnal

Thin ash line

2.5

Mixed

3.0

8arren layer
3.5

_______

____

____

Mixed
___

40

__

4Mixed

q
45

CR055 SECTION OFEA5ST CUT

SCALE

Fig. 8. Cross-section of East Wall of Pit V, confirming the

Analysis of the Previous Sections.

382

1Bennett, Tiahuanaco Excavations.

1934.1

383

extended into the west wall. Pottery and bone fragments were found
throughout the pit. Worked stone, bone and copper fragments were
also found.
Level 6, 2.5-3.0 meters deep. A layer with considerable clay mixed
with dirt and ash. More bone and pottery fragments were found. Some
worked bone tools, part of a stone dish and a pottery disc (whorl?)
were also found.
Level 7, 3.0-3.5 meters deep. Except for an ash bed (cf. Level 7a)
in the southern part of this level the clay content, while slightly disturbed,
was practically barren. A few potsherds and fabricated pieces were
found in the upper part of the layer.
Level 7a, 2.70 to 3.30 meters deep (ash bed). This fire pit was 70
centimeters wide and penetrated one meter into the southern wall of
the excavation. It contained considerable ash, bone, potsherds, stone,
and some copper. The material was classified separately.
Level 8, 3.5-4.0 meters deep. Under a 15 centimeter bed of apparently undisturbed clay and stone a bed of ashes about 0.5 meters thick
was found along the east wall. Scant traces of disturbance were observable throughout the whole pit at this level. The ash was heavy and wet
through ground water seepage.
Level 9, 4.0-4.60 meters deep. At 4.60 meters depth undisturbed
clay was encountered. In the center of the southern part of the pit
was found a thick ash bed containing pottery, bone, stone, copper ore,
and other artifacts.
This was certainly the most significant pit in all the excavations.
It showed the best really stratified deposits. As the cross-sectional
drawings (Figs. 6-9) indicate, the depth was not attained by artificial
pits (as in Pit IV), but by an accumulation of dwelling refuse and wash.
While no definite remains of houses were found, the heavy ash beds
certainly suggest dwelling, and the pottery is mostly of the ordinary
kitchen variety.
Analysis of the stratification results in the following array. First,
Level 1 contained deformed skull burials and llama skeletons and was
separated from Level; 2 by an ash bed or layer. (This division is slightly
below the half meter line.) Then Level 2 was based by a clay floor in
the southwest corner, unworked stones in the southeast corner, and
dressed stones in the southern center. This line definitely separates
Level 2 from Level 3 in the southern half of the pit. In other parts of the
pit, Levels 2 and 3 are separated by an ash layer, not completely horizontal in its distribution. The cut stones already mentioned may be

384

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

associated with Level 3, though the lower part of the same level is more
closely associated with Level 4. In fact, nothing divides the lower part
of Level 3 from Level 4. Level 5 forms a distinct stratum, being
0

0.5

Ash,
2
t.0

Ash

3
'.5

2.5___

qMixed

Barren clay

8
40

,JMixed
f Clay

Gulf ure, ash, bone.

4.5

CRO5 SECTION aF NORTH CUT

SCALE
O 5
QS

0w.

Fig. 9. Cross-section of North Wall of Pit V, showing the Position of the Canal. The depth and slight slant of this canal suggest
that it was used as part of a drainage system.

separated from Level 6 by a line of unworked stones along the western

part of the pit. The demarcation between Level 4 and Level 5 is not so
clear, though Level 5 seems to be distinct from the canal level on the
northern part of the pit. Level 6 is separated from Level 5 in part of

1934.]

Bennett, Tiahuanaco Excavations.

385

the pit. There is no distinction between Level 6 and the scanty mixed
section of Level 7. The ash bed of Level 7a starts in Level 6 and continues into Level 7, the lower part of which is a barren layer of clay and
washed stones. Thus Levels 8 and 9 are distinctly set off from Level 7.
A thin layer of clay divides Levels 8 and 9, though it does not seem very
significant. In r6sume, then, the divisions are as follows
Level 1
Level 2, including the upper part of Level 3
Level 3, lower part, and part if not all of Level 4
Level 5
Level 6, Level 7, Level 7a
Level 8, Level 9
PIT VI
Puma Puncu forms a distinct unit southwest of the main ruins of
Tiahuanaco. Pit VI was excavated in the field in front (to the east) of
this ruin. The results of the excavation were disappointing. A handful
of plain sherds was found in the first meter. Then undisturbed sand was
struck. To the west of the hill of the Puma Puncu ruins is a flat
section which Posnansky has designated "docks." A test pit in this
section encountered undisturbed sand at one half meter depth. The
poverty of these pits contrasts markedly with the richness of the
others.
PIT VII
To the east of the monolithic stairway of Calasasaya is a small
temple 21 by 22 meters in size. This has been designated by Posnansky
as the "First Period Temple." It was a semi-subterranean structure
about 1.80 meters below the surface of the base of the monolithic stairway.' The temple was completely covered up to the first of this century.
Squier,2 writing in 1877, does not mention the temple although his map
shows an angular wall in front of the sun temple which might doubtfully
represent it. It was in the excavations of Georges Courty that the structure was first revealed in a three meter depth of d6bris and wash.3
Pit VII was made within the northern half of the temple. It was 4
by 2.5 meters, parallel to the northern wall of the structure. At a depth
of one half meter the head of a large monolithic statue was encountered
in the southern central portion of the pit. To uncover this it was necessary to extend the excavation 6 meters to the south at a 3.5 meter width
'Posnansky, 1914, 81-83.
2Squier, 276.
3Courty, 534-538.

386

Anthropological Papers A merican Museum of Natural History. [Vol. XXXIV,

(Fig. 18). Unfortunately, the change in excavation technique made it

impossible to preserve stratified levels. However, relatively few potsherds were found. Furthermore, excavation could not be continued to
undisturbed ground because of the interference of the statue.
The statue is described in detail in another section (p. 429). It
extended with the head to the north and the base to the south, with a
slant of 20 degrees to the west of north. The headband was encountered
at a depth of 0.5 meters and the base at 1.80 meters. The figure lay on
its back and the assumption is that it faced south when erected, though
there is the possibility of a quarter twist in the fall from a position facing east.
The statue was 7.30 meters long, including a base of 1.80 meters.
At the feet of the monolith (top of the base), on the west side at 2.10
meters depth, a circular blue stone was found (cf. p. 444), and under this
at 2.30 meters depth a typical stone head' was encountered. Near this
were some blue, lapis lazuli, beads. To the east of the feet of the monolith and parallel to it was a smaller statue, 2.55 meters long. Its head
was at a depth of 1.90 meters and the feet were at 2.60 meters. Beside
the small statue, still farther to the east were two stones, one carved
with a crude face, the other unworked. Above these at 1.30 meters
depth was another of the stone heads with a double face. Many squared
and dressed stones were'found in all parts of the pit, though in disturbed

positions.
An analysis of this temple is inserted here although the argument
anticipates the classifications of pottery and stone working which follow
in later sections. The temrple, as described by Posnansky and Courty,
was semi-subterranean, with the sides faced with stone applied in the
technique of upright slabs set at intervals and filled in between with
smaller stones. Set into these facing walls were blocks which had carved
stone heads projecting from them.
An attempt to determine the chronological significance of this
termple involves the consideration of several points. The facing wall fill
is no longer visible, but to judge from Posnansky's statement and photographs it was composed of a conglomeration of well-finished and crude
blocks. From the photographs, some of the blocks appear to have been
drilled, although this is'not certain. The general appearance is that of a
wall composed of stones collected from different places, and neither cut
nor placed in accordance with any preconceived plan. The stone heads
are stuck into the wall in irregular positions. A classification, of the
'Cf. Posnansky classification, 1914, 83-84.

1934.]

Bennett, Tiahuanaco Excavations.

387

stone heads, found within this temple, divides thenm into four subtypes
(see Stone Sculpture, Type 6): first, a group of heads, with headbands,
and faces of the general classic Tiahuanaco type, projected from
rectangular blocks; second, a group of flat stones with faces crudely
chipped on one side; third, a group of rounded boulders with faces on
one side; and finally, two fairly realistically carved heads with cylindrical
projections for wall mounting. The heads too give the idea of collected

material.
A further analysis of the stone sculptures found shows a large,
Classic monolithic statue, decorated in the "Gateway of the Sun" style,
side by side with an angular, unincised, bearded statue. Whether this
smaller statue is considered earlier or later than the large one, it is without question not contemporaneous in style. When the circular blue
stone, the crudely carved uprights, and the sculptured stone heads are
added to the variety it is obvious that the conglomeration is not stylistically contemporaneous. Here again the best explanation is that of a
late temple in which material was brought together from all parts of the
ruins.
Courty found some of the heads with paint still preserved on them.
Finally, it is possible that the large statue was one of the two described
by Cieza de Le6n as he stood and looked "mas adelante" (ahead) from
Acapana, in which case it was still standing in the early sixteenth

century.'
The pottery analysis shows mainly Decadent and Inca sherds with
none of the early Tiahuanaco types established by stratigraphy in the
other pits.
The total effect of this analysis indicates that this small, semisubterranean temple was late Tiahuanaco at best. Since the temple had
been filled in by wash, and since the excavations did not penetrate its
floor because of the interference of the large statue, the material from the
Pit VII is considered as representing a single, late period. Unfortunately,
this analysis sheds no light on the chronology of the sculpturing styles
represented in the temple.
PIT VIII
This pit was excavated on the flat plane to the west of the fortress,
Acapana. This location is within the canal-enclosed temple area, but not
particularly close to any of the known ruins. It lies 112.7 meters southwest of the southern corner of the projecting west wall of Calasasaya
'See Uhle and Stubel, part II, 33, for an argument that Cieza was looking eastward from Acapana.
The head of a large statue, now in:Pa Paz, was also found to the east of Acapana.

388

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

and due south of the "Palacio" temple. The pit measured 3 by 3.30
meters.
Level 1, 0-0.5 meters deep. Clay soil and a considerable quantity of
potsherds characterized the first half meter, but it was not notably
distinguished.
ILevel 2, 0.5-0.75 meters deep. The ordinary half meter unit was discontinued in this instance because of the quantity of sherds. At 50

CANAL

Cut

Sfone

II

__

___

GROUND PLAN
SCALE

Fig. 10. Groundplan of Pit VIII. A path of six dressed slabs laid end to end
runs north to south through the center of the pits and continues on both sides. Some
of these slabs are drilled for use in some previous construction. A loose stone slab
stands by itself at the same level. At the west edge of the pit is a canal, composed of
two walls, without covering.

centimeters below the surface a canal was found which was made of cut
stones (Fig. 10). It extended north and south and was constructed
with a parallel double row of dressed stones, but without stone base or
covering. It was evidently intended as a surface canal. Each side wall
was composed of a double layer of finished stones, carefully fitted.
The base was apparently a mixture of clay and small stone. The canal
measured 50 centimeters in outside height, 70 centimeters in outside
width, 50 centimeters in inside depth, 30 centimeters in inside width.

Bennett, Tiahuanaco Excavations.

1934.]

389

At 70 centimeters depth a stone path crossed the pit from north to

south. This was composed of a single row of flat dressed stones laid end
to end. The stones were taken from some previous building as some of
them were bored with construction drill holes. The stones measured as
follows:2 58 by 78 by 10 centimeters with two holes drilled in sides at the
center
Mixed loyer

I
0.50

2
3

0.75

3O

Canal

Heavy 5herd layer

CIQAs
Clay__ ;
4

1.0~~~~~~21S

PaMhway6e

W'aler w.orn stones$v 0--`,o

Ash sherds

,4sh~A mixed

IWafer level culture contfiues'

CROSS SECTION oF NORTH WALL

SCALE.
0

0.8
Ifrn.
Fig. 11. Cross-section of North Wall of Pit VIII. The relation of the pathway
to the canal is shown. Also the nature of the open canal of cut stone blocks is more
evident, without covering but with clay base. Above the path to the edge of the canal
is a heavy sherd layer. Below, the levels continue, without marked stratigraphy, to
the pit which was never completed because of underground water seepage.

3 57 by 87 by 15 centimeters with slanting holes drilled in the four

corners
4 58 by 83 by 14 centimeters with no holes
5 40 by 28 by 12 centimeters
On the same level with this row was a single stone in the southwest
corner, which ineasured 45 by 65 by 12 centimeters.
This whole quarter meter contained a high ash content.

390

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Level 3, 0.75-1.0 meters deep. The base of the canal was part of
this layer. To the east of the stone path was a bed of water-worn stonies
and next to this was a heavy ash bed.
Level 4, 1.0-1.5 meters deep. A half meter of ash, clay, and debris
with no distinguishing stratigraphic characteristics.
Level 5, 1.5-2.0 meters deep. The ash and disturbed clay continued
throughout the entire half meter, though the dirt was gradually getting
damper.
Level 6, 2.0-2.5 meters deep. At 2.20 meters depth a pit started
surrounded by undisturbed clay. The pit was about as big as the surface
area of the excavation, but the undisturbed clay was easily distinguishable. Water seepage continued, and the dirt and pottery had a greenish
tint from vegetable mold. At 2.50 meters depth the water became so
deep that work had to be discontinued, although the undisturbed ground
had not been reached.
Level 1 is not clearly separated from Level 2, except that Level 2
contains a concentration of potsherds in an almost solid mass (1878 in a
quarter meter) (Fig. 11). The division is roughly confirmed by the line
of the top of the canal, although it is not clear just how the canal functioned. Level 2 is, however, definitely separated from Level 3. The
stone path, the single stone, the concentration of potsherds, and the
water-worn stones all confirm an old surface level. The stone path is in
itself a secondary form of stratification as the stones are utilized from
some previous building. Below Level 3 there are no sharp divisions.
Levels 4, 5, and 6 follow in sequence and must be grouped or separated
on the basis of the collection analysis.
PIT IX
This pit was excavated in the yard behind the house of Sefior Rodriguez. It is only a short distance back of the railway station, to the
northwest. Its position corresponds to that of Pit III on the north.
Level 1, 0-0.5 meters deep. Considerable quantity of heavy crude
pottery. Two extended burials in fragmentary condition.
Levels 2, 3, 4, 0.5-2.0 meters deep. Nothing distinguished one layer
from another in the actual excavation, except perhaps the concentration
of the heavy pottery in the upper layer. At 1.90 meters deep undisturbed
sand was encountered in most of the pit.
Level 5, 2.0-2.70 meters deep. Two side pits continued below the
general level of the undisturbed ground, and the material from these pits
has been classed together as Level 5.

1934.1

Bennett, Tiahuanaco Excavations.

391

PIT X
This pit was dug in the old cemetery with the intention of finding
some good, high grade pottery samples. It proved satisfactory in this
respect, though disappointing stratigraphically. The material was divided into two levels. Since the pit was started on a slight slope, Level
1 represents a depth of 40 centimeters at the south and 70 centimeters at
the north, and Level 2 was 90 centimeters at the south, 1.20 meters at the
north. Except for a few side pits to 1.40 meters, no greater depth was
attained. Although nineteen complete or partially complete clay vessels
were found in this pit, there were many indications of previous disturbance. One beautiful kero-shaped vessel had pieces missing which the
most diligent search did not reveal.
In this section the analysis has been limited to the actual excavations.
Later, this analysis will be extended to include the classification of the
collections and the cultural implications revealed.

THE COLLECTION
The classification of the Tiahuanaco excavation materials is made
on the basis of pits and levels. Whole and almost complete pots are
treated separately from potsherds. While this division is not maintained,
in the final analysis of the collection into its possible cultural significance,
it was thought best to limit the percentage tables strictly to the sherds.
In this section the collection is described and classified in its totality.
Some notes are made as to the possible stratigraphic significance of some
of the types, but the bulk of this latter task is left for the next section.
POTSHERDS
The sherds from each level of each pit were first divided into two
groups: plain and painted. The plain sherds were further classified
according to color of ware, and the painted sherds according to the
number of colors used. Further subdivisions were based on designs and
forms. The basic tables summarize these classifications for each pit
and each level.
By far the greater number of sherds are of plain ware. The classification of over 14,500 sherds results in the following percentages.69.78
22.24
7.98

Plain sherds
Painted sherds
Miscellaneous

100 .00

When viewed as a characteristic of individual levels in a pit, the

relation of plain to painted ware takes on some significance (cf. Table 1).
For example, in Pit VIII the analysis reveals a change of proportion of
plain ware from about 55 per cent and painted about 35 per cent in the
first three levels, to about 93 per cent of plain and about 3 per cent of
painted ware in the second three levels. This seems indicative of cultural
change.
Plain Wares. Most of the plain ware comes from cooking vessels.
The commonest shape is the olla with flat bottom, wide open mouth,
flaring rim, and two flat loop rim or side handles. The larger part of the
ware is between .4 and 1.5 centimeters thick, seemingly welt fired, and
rather fine in texture. The temper is of sand, often mixed with mica.
The handles, although varying in detail, are essentially uniform. They
are flat loops, ranging in width from 1 to 3 plus centimeters, and are
invariably vertical, whether attached to body or rim. Some of the wider
handles are slightly concave rather than flat, though the difference is not
significant. A few rounded handles are exceptions to the general class.
392

TABLE I
PERETAGE DITRIBON OF WARES AccoaDING TO PIT AND LEVEL

"

0
Q~~~~~~~
~~~~~~~~~~~~~~~~~~~
-.I
1l
l
l
18
1= I
I
I
N

Pit I

0
~~~~

'~~~~

61.8521.63 4.12 8.24 6.18 3.09

67.34 30.60 20.40 10.20
3 29.16 70.82 16.16 33.33 20.83
3a 81.06 17.41 8.33 2.27 3.78 3.03

1
2

1 83.60 14.74 6.55 5.73 .82 .82 .82

2 59.70 35.81 11.94 14.92 1.49 1.49 5.97
3 76.0024.0016.00 4.00 4.00

Pit III

1 86.11
2 100.00

PitIV

1
2
3
4
5
6

PitV

1 62.69 30.84 7.96 15.42

2 53.62 31.72 9.83 15.86
3 51.51 28.28 14.14 13.13
4 68.98 18.05 5.09 9.72
5 80.50 13.31 5.28 5.28
6 89.93 2.88 2.16
7 93.44 6.56 4.92 1.64
8 91.46 6.10 6.10
9 95.69 .86 .86
7a 93.75

.34

13.19

63.55 29.84 7.8910.97 6.69 4.29

3.86
52.39 44.26 12.42 15.25 8.61 7.25 .73 .49
46.77 40.75 2.57 5.15 11.58 19.74 1.71 9.01
60.73 37.88 14.15 5.02 12.78 5.02 .91
38.74 53.44 11.71 9.01 16.81 14.41 1.50 4.20
45.57 46.79 8.85 3.79 17.7115.18 1.26
2.98
11.55
18.18
1.39 1.39 .46 11.11
2.52 .23
5.50
.72
5.03

5.86 .17
1.01

7.35 1.47

1 83.79 8.82

Pit VII

1 68.06 22.58110.32 7.74 3.55

Pit VIII

1 50.56 39.58 6.06 15.91 12.50 4.92 .19

2 57.88 27.84 7.61 7.88 7.45 4.53 .37
3 59.65 32.90 16.67 7.02 5.70 3.51
4 91.56 5.53 2.30 2.15 .77 .31
5 94.79 .74
.37
.37
6 95.81 2.79 .28 2.51

Pit IX

83.42 9.60
84.54 10.52
70.99 12.79
82.57 4.23
5 76.00 5.11

PitX

1
2
3
4

3.14
4.38
4.44
1.25

2.62
3.57
3.41
2.98
3.20 2.40

1.51

1.64
4.47

.82
1.49

.34

.34

99.97 97
99.98 49
99.98 24
99.98 132

.82 99.98 122

99.98 67
100.00 25
99.98 288
100.00 193

2.44
3.44
6.25

2.44

.17 99.98 1166

99.95 813
99.96 233
.43
.45 99.96 219
99.98 333
.60
1.26 99.93 79
.08

.12

2.49
.34 .34
.23
1.44
.86 1.72
2.50 3.75

.86
1.47
.32

99.99
99.98
99.99
99.99
100.01
100.00
100.00
100.00
99.99
100.00

201
580
99
216
436
139
61
64
116
80

99-96 68
.97 .32

.64

99.98 310

8.711.14
.19
.95
99.99 528
.53 12.30 1.44
.21 99.99 1878
1.12
.11
4.82 2.64 ?.44 .44
.44
1.32
100.01 228
2.91
.46 .31
.61 1.53
100.00 652
4.47
1.12
1.49
.37 1.49
99.98 269
1.40
1.12
.28
100.00 358
.87
.12 .06 .44

.12

13.19
15.20

52.89 42.38 9.85 5.32132913.14 .78

2.50 2.20

76.37

1.651

.55

23.081 9.071

1.49 1.49

.99
1.21 .17 1.03
1.01 1.01
1.39 .46
.23 .23
.72

7.09

.80

6.04 3.57 2.75

1.03
2.04

3.48
3.09
2.02
1.85
.69
2.16

2.79 1.05
5.93 1.04
1.69 .69 .19
4.19 .74
3.75 1.19
14.67 1.36 .17
2.40

t4*a

C)

2.06

1.51

5.88
.97

4
EH
2 1
0

2.93 1.54 .77 .17

2.81 1.35 1.10 .24
3.43
2.57 .43
1.35
.45 .45
3.60
2.40 .30 .30
7.57
3.79 1.26
1.26

5.97 1.49

Pit VI

3.09
2.04

13.40

Pit II

.34

~C

~~~

&H4~

.31 .63 .16

.16 .47 .47

.55

.17 99.99 573

99.99 1598
.17 99.98 586
99.99 637
100.00 125

99.97 639
100.00 364

1934.]

Bennett, Tiahuanaco Excavations.

393

Classified roughly according to the color of the ware, or in some

cases the slip, the following percentages prevail
Orange
Brown
Black
Yellow
Red-Brown
Brown-Black
Scraped
Miscellaneous

Per Cent
40.40
23.61
17.34
6.62
2.34
4.50
1.29
3.90
100.00

Orange ware, the commonest found in all pits, is generally covered

with a thin slip. Somne of this ware is over 1.5 centimeters thick, and
some is exceptionally thin. It is not notably distinct from the other
wares, and cannot be considered a type, other than a plain ware.
Brown plain ware is generally thin, though, like the orange, all
thicknesses occur. Both light and dark browns can be distinguished.
Most of the brown ware is dark in clay color as well as slip. The black
plain ware is easily confused with a burnt-black orange and a burnt dark
brown. However, black finished (though unburnished) ware is found.
This ware is generally thin. A thin, deep black, unpolished, micatempered variety of the plain black ware is found in Pits II, VIII, and V,
at the lowest levels. In general, the dark browns and black wares are
more frequent in the lower levels, while the orange dominates the surface
and upper levels, though this distinction is not sharp enough to direct
conclusions.
Yellow ware is thick and well fired. The slip is a dull yellow. Its
distribution throughout the pits is sporadic. Included under the miscellaneous total are red, cream, cream gray, and colorless wares which
occur infrequently. Some of the combined colors, such as reddish-brown,
brownish-black, although necessary in the preliminary arrangement, are
not of great significance in the final classification.
Another group of plain ware is called "scraped." This ware is
moderately thin, finished in orange, brown, and black. The whole surface
has been scraped with a flat instrument (perhaps a piece of gourd), and
the scraping lines, roughly parallel, are clearly discernible. This ware
at first appeared distinct, but later it was seen to be a common process,
ordinarily not left in such an obvious stage of incompleteness, but more
smoothly finished. This scraping treatment is even found on some of the
early painted ware, over color and all.

394

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Extra Thick Ware. Pieces of an extra thick ware, ranging from 1.5
to 2.5 centimeters, occur in various parts of the pits, though mostly in
the upper levels. The fragments are all from large vessels. Both grit
and mica-tempered sherds are found. Some are crude, unfinished segments, though always durable. Others are well finished on the outside,
almost polished, and painted red. One piece had a thin black painted
line. One crude thick piece, part of the base of a bowl, had a thick projecting ridge on which the bowl rested, much after the style of the "hollow-base" bowls. The diameter of this base was estimated from the arc
segment as 34 centimeters. Another fragment of a base of equal diameter
was well finished, quite flat, and had no suggestion of the projecting
ridge. The rim pieces in this ware are of straight-sided, plain-edged,
large bowls with the actual border slightly thicker than the bowl sides.
There is no flare of any kind, but the edge is squared off. One heavy loop
vertical handle started from the rim. Another of the same type started
from the body, near the base. These fragments suggest a large, flatbottomed, open, straight-sided, tub-shaped bowl with a plain rim.
Complete bowls of this type are actually found in the collection of the
Cochabamba Museum. Somewhat similarly shaped thick bowls are
found on the coast of Peru with Tiahuanaco derived designs painted on
them.
Thick Polished Ware. A distinctive plain ware is a thick, well-fired,
sand-tempered clay with a burnished finish, on the outside only, in red,
orange or brown. The thickness averages around .8 centimeters. Most
of the pieces are fragments from large vessels. One rim type, without
flare, has the cross-section of an inverted equilateral triangle with the
base up. Thus from a 1.5 centimeter rim edge width it narrows rapidly
to the regular .8 centimeters thickness. This rim type suggests that the
vessel was a large open bowl. This ware is not continuously distributed
throughout the excavations, but it appears and disappears with varying
degrees of frequency at some level in almost every pit. The significance
of this distribution will be discussed in the next section.
Painted Line on Plain Ware. Although logically this ware belongs
with the painted group, it is described with the plain wares because the
clay, tempering, shape, and finish of the vessels are much more like, the
plain than the painted pottery. It has the appearance of a plain unslipped ware with lines painted on it. This class includes several varieties. First, is an orange ware with rough engraving, that is, incised lines
are rudely sketched, without creating any definite design. A wide dull
black line is also applied on an orange ware without any definite design.

1934.]

395

Bennett, Tiahuanaco Excavations.

Some orange ware combines the black and the incised line. There is also
an orange ware with both black and white lines. The black line alone is
found on yellow and brown ware. Finally, black and white lines, and
black, white, and brown lines rarely occur on a plain red slip ware. This
latter ware has step designs, curves, and more definite patterns. In
summary, these variations occur as follows
Orange with incisions
Orange with black line
Orange with incision and line
Orange with black and white line
Yellow with black line
All others

Per Cent
9.20
68.96
12.87
3.45
4.14
1 .38
100.00 (435 pieces)

The ware is comparatively thick, varying around 1 centimeter. It is

well-fired, sand-tempered, and has the feel and ring of modern building
tile. The vessels are all large. One typical shape is a large urn-like
vessel with a straight collar, topped with a sharply flaring rim, the edge of
which is at right angles to the perpendicular collar. From the flat bases
and the collars which remain it appears that the shape of the whole vessel
must, in some examples at least, have been a somewhat truncated heartshaped bowl about 70 centimeters high and wide. On the basis of shape
and general texture much of the plain thick orange and yellow ware
might be included in this class.
Painted Ware. The proportion (22.24 per cent) of the total sherds
classified as "painted " obviously includes a variety of wares. However,
one of the impressive facts about Tiahuanaco painted pottery is its consistency in color, design, and form which can be observed throughout
the collections. Thus the references which are constantly made in the
literature to "Tiahuanaco style" are not at all meaningless, but are
directed to those qualities which distinguish the wares of the Tiahuanaco
site as a whole from other Andean types. It is my purpose in this discussion, first, to describe Tiahuanaco ware as a cultural unit, and secondly,
to attempt to distinguish phases within the Tiahuanaco unit. It is
from this first point of view that the painted ware is here described,
reserving the more subtle problem for the following section.
Painted ware includes, in this analysis, the following subdivisions:-

396

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Polished ware
One-color ware
Two-color ware
Three-color ware
Four-color ware

Per Cent
28.84
30.14
23.03
16.65
1.34
100.00 (3291 pieces)

Polished Ware. This group includes all painted ware which shows
polish, but no design. While most of these pieces are parts of vessels
which quite probably had painted designs, there are also some complete
vessels of one color which show no painting other than this basic slip
color. Divided on the basis of slip color, the polished ware may be
summarized thus:
Red polished
Orange polished
Brown polished
Black polished
Miscellaneous

Per Cent
51.74
25.92
11.06
8.01
3.27
100.00 (949 pieces)

The red polished ware dominates because a red slip is the basic
color for most of thei painted ware and consequently would occur with
more frequency when the pottery is broken into sherds. Black polished
includes both plain and incised ware. It is a distinct ware, as black is not
a base color in painted ware. Most of the black ware sherds are from
kero-shaped cups. The incisions are fine and sharp. The polish is
usually brilliant. Out of seventy-six black polished sherds only fourteen
are incised. To these must be added two partially complete kero-shaped
cups of black incised ware. Although the black polished ware is rather
generally scattered throughout the levels of the pits, the incised ware is
limited to Pit I, 2 and 3a; Pit IV, 1-5; and Pit V, 2. From the distribution, the incised black ware seems more recent than the plain polished
black.
Some parts of polished, plain rimmed, curved base, open bowls seem
form
a unit distinct from the rest of the polished sherds. One group
to
includes fragments of this type bowl in a poorly finished, though thin,
red colored ware, which has the appearance of being scraped both inside
and out. Six fragments occur in Pit VIII, 5, and seven fragments in
Level 6 of the same pit. Another group in this same class bowl is of a
dark polished brownish-red ware. This series is well finished, without

Bennett, Tiahuanaco Excavations.

1934.]

397

the scraped appearance. One piece occurs in Pit V, 9, one in Pit VIII, 6,
and five fragments (including four parts of one bowl as one piece) in
Pit VIII, 4. The isolation of these polished bowls, which are not typically
Tiahuanaco in shape, in the lowest levels of Pits V and VIII suggests
that they may form a special group.
One-Color Ware. The term "one color" is used here to refer to a
design of one color applied over a basic slip color. Since the base is
often a painted color, the pottery might well be referred to as having two
colors. However, for the purpose of simplification, the use of one-color
design over a basic color will be called "one color" ware. The following table shows the emphasis on two combinations:
Per Cent
41.94
36.90
6.75
3.02
3.83
1.71
5.85

Black-on-red
Black-on-orange
Black-on-brown
Black-on-yellow
White-on-red
Yellow-on-red
Miscellaneous

100. 00 (992 pieces)

Even these one-color pieces are sometimes nothing but parts of

more complex polychrome vessels. However, complete pots with only
one-color designs are common enough in the collections.
The miscellaneous group includes quite a variety of combinations
which from the fragments cannot be identified with whole pots. These
combinations are listed with the number of examples of each
Black-on-cream
Black-on-gray
Black-on-light brown
Yellow-on-brown
White-on-orange
White-on-brown
Brown-on-orange

6
2
3
6
3
10
3

Brown-on-brown
Brown-on-yellow
Orange-on-red
Orange-on-brown
Red-on-orange
Red-on-yellow
Red-on-glaze white

4
6
1
1
5
7
1

The black-on-orange and black-on-red wares are rather uniformly

distributed throughout the pits, with the black-on-red dominating
slightly in the lower levels. Black-on-brown, however, is practically
limited to the uppermost levels (Pit I, 2; Pit II, 1; Pit IV, 1, 2; Pit V,
1, 2; Pit VI and VII; Pit VIII, 1; Pit IX, 1 and 2).
Two-Color Ware. This classification includes all fragments which
have two design colors on a single color background. Classed according
to color combinations, the alignment is as follows:

398

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Black and white-on-red
Black and yellow-on-red
Black and white-on-orange
White and yellow-on-red
Black and white-on-brown
Black and yellow-on-brown
Brown and white-on-brown
Miscellaneous

Per Cent
48.42
22.03
12.53
4.09
2.64
2.64
2.11
5.54

100.00 (758 pieces)

Although the one-color grouping showed a fairly equal use of orange

or red as a background color, in the two-color group red is more frequently
the base. Likewise the brown is less frequent.
The less frequent color combinations, classed as miscellaneous,
include the following variety:Black, white-on-yellow
Black, white-on-gray-brown
Black, yellow-on-orange
Black, orange-on-red
Black, orange-on-orange
Black, red-on-orange
Black, gray-on-orange
Black, gray-on-red
Black, red brown-on-yellow

2
1
5
1
4
4
2
1

Brown, white-on-orange
Brown, yellow-on-red
Red-brown, white-on-brown
White, yellow-on-orange
White, yellow-on-brown
White, red-on-orange
Yellow, red-on-brown
Black, brown-on-orange

3
2
1
1
3
4
2
5

Three-Color Ware. Pottery with three design colors on a background

is here grouped together. Black, white, yellow-on-red is the dominating
combination, to the exclusion of most others, as the following table of
percentages shows:
Per Cent
Black, white, yellow-on-red
Black, white, yellow-on-orange
Black, white, yellow-on-brown
Black, white, gray-on-red
Black, white, red-on-orange
Brown, white, yellow-on-orange
Miscellaneous

80.61
4.24
3.43
2.83
2.42
2.83
3.64

100.00 (548 pieces, which include 53

Inca and specials not used
in above percentages)

The miscellaneous
nations.

group

includes the following variety of combi-

1934.]

399

Bennett, Tiahuanaco Excavations.

Black, white, gray-on-orange
Black, white, gray-on-brown
Black, white, red-on-red
Black, white, brown-on-orange
Black, yellow, brown-on-wbrown
Black, yellow, brown-on-red
Brown, white, yellow-on-light brown
Brown, white, light brown-on-red

1
5
2
1
5
1
1
2

Four-Color Ware. Four design colors on a background are the

criteria for this class. In the excavations relatively few of these pieces
were encountered, though four and even five-color ware is not infrequent in the collections of mortuary pottery. The majority of the ware
falls into one color combination as shown below.Black, white, yellow, gray-on-red
Miscellaneous

Per Cent
86.36
13.64

Black, white, yellow, gray-on-brown

Black, white, yellow, red-on-orange
Black, white, yellow, brown-on-red
Black, white, gray, red-on-orange
Black, yellow, brown, orange-on-orange

100.00 (44 pieces)

1
2
1
1
1

Special Painting. In the lower levels of two pits (V and VIII), a

special painted ware was found which differed from the typical Tiahuanaco ware. This was a thin ware, mica-tempered, and soft-fired,
characterized by buff-colored, slipped and plain clay. Two varieties
are found, red-on-buff and polychrome-on-buff. The first set consists of
plain rim pieces from small open curved base bowls, with a red band
design usually along the border, but in some pieces on the body too.
The second set consists of a variety of colors, black, white, red, brown,
and orange, on a buff slip, used in combinations of two or three. They
are burnished colors which produce a glossy effect not noticeable in other
Tiahuanaco pottery. The designs are more varied than in the red-onbuff class, though still simple stripes (vertical) and angles. The small,
open, plain rim bowl is not the only shape suggested by the sherds. Two
pieces are sectional parts of "hollow base" bowls; two others suggest
kero-shaped vessels; another piece is one of the rim-loops of a scalloped
edge incense bowl.
Isolated Fragments. Certain pieces present peculiarities which fail
to assimilate with any of the group classifications. Although these are
not definitely foreign wares, it is more convenient to describe them
separately, rather than force their classification.

400

Anthropological Papers American Museum of Natural Histor y. [Vol. XXXIV,

Two fragments of small, open, bevel rim bowls were found in Pit
VIII, 2. These are thin, well-fired pieces, with polished colors. The
rims are painted red, which is apparently the basic paint. Over this is a
black paint which almost covers the body solidly, although one piece
shows a small square of white paint.
One sherd from Pit V, 9 is a thin fragment, with red and deep brown
burnished colors. The step design is used and the outline between blocks
of color is deeply incised. This incision on a painted pattern is decidedly
non-Tiahuanaco in style.
Two pieces are painted in white, yellow, red, and brown, apparently
on a black background paint. One fragment is a part of the rim of an
incense bowl, painted on the inside. The design suggests a fish head,
and it is executed without outline, in non-characteristic fashion. These
pieces are both from Pit VIII, 4. Whole vessels in collections have
similar colors, designs, and technique. The shape is like a spittoon (See
discussion and drawing, pp. 411, 451).
In Pit VIII, 5 was a thick piece of mica-tempered pottery with part
of a complex design in black, white, yellow-on-red. The design, rather
than the colors, gives the impression of an exceptional piece.
POTTERY DESIGN
Decorative design cannot be perfectly determined on the basis of
potsherds. In this report most sherd designs are classified according
to their identification with those on complete pots. A great many sherds,
however, are so small that it is impossible to determine the exact nature
of the completed design, and these are classified by some arbitrarily
selected measure. On this basis the potsherds from the various pits have
been classified according to design type and pit level. The numerical
results are summarized in Table 2.
A complete classification of Tiahuanaco designs, based on museum
specimens and all branches of decorative art, is beyond the scope of this
publication. Something of this sort has been attempted by Posnansky,l
though not as completely as in the unpublished second volume of the
same work. In the present paper the description of design elements is
limited to the actual collections obtained by my excavation. Thus the
study is by no means exhaustive.
Again, the consistency of Tiahuanaco subject matter and style is
impressive, despite the quantity of variation in detail. 81.99 per cent
of the decoration of the potsherds is geometric, while only 18.01 per cent
'Posnansky, 1914.

TABLE 2
FRUQuzNCY DISTRIBUTION oF DESzGN Tys IN PTs AD Lvtis

Pit I

74

2
3
3a

7 2 31
4

1
Pit6
I
2
3

Pit III

1
2
3
4

PitV

6
7
8
9
7a

Pit VI

PitoVII

Pit VIII

Pit IX

Pit X
Total

15
2

1
2

1
2
3
Pit1
V
4
5
6

1
2
3
4
5
6

0~~~~

14
1

15
5
13
5

0
0
170 22 6
13
21
5221
117
10
2
39
44
50 1 21 2 12
19 1 5
1

7
7

3
41

1
4 4

5
2

3
12

22

32
7 5
1
15 3 10 16 22 2
66
2
1
7 1
1
1941
29
1
5?
1
1

8
2

1
6
3
3

3 234
6 239
82
3

3 137
30

1
2
1
1

52

1
2

46
127
14
28
35
1
1

0
0
0
1

21 3 2

112
228
22
17

23 1 17 10
3 1424 29
2 1
2 1
3
1

(5)

(4)

13
1

2
1 2
6 11 11 3
5 2

1
2

83
27

9
4

56 8

2
3 11
3 45 931 2
3
3
1

(2) (2)
(2)

1
28
43 97
32
2 1
1

1247 75 94

1
2

1
2
3
4
5

2
5

38

1
1

177
382
37
28
2
9

(1)

1
3 12 2
7
1 1

3
1

6
3

13

1
2

7
2

026716 6

2
1

22
3

38 119

34
8

11

1638 75i

37
98
49
3
5

203
51

292210

1934.1

Bennett, Tiahuanaco Excavations.

401

is zoomorphic. This difference in proportion is perhaps augmented by

the size of the sherds, which tends to emphasize the geometric nature of
the design; that is, complete pots contain both zoomorphic and geometric
designs in the composition, and a small sherd may not tell the whole story.
All sherds with designs in single straight lines were classed as
linear. This linear group dominates the whole design classification
including 56.42 per cent of the total decorated sherds. Since many
types of design can be broken up into linear segments, the classification is
not absolutely a true one. When the lines were wide, rather than thin,
the sherd was classed as a "band" design. Bands of color do occur on
most complete pots, sometimes as part of a composition, and sometimes
as the only design element.
Wavy designs refer to the perpendicular wavy lines, so commonly
found in decorating open flaring bowls. Generally the wavy lines alternate in two or three colors. Often the wavy line section is just a fill
between two areas of more complex design. A broken wavy line also
occurs. Here the rows of wavy segments generally encircle the vessel.
A curved line on a sherd is so classified. Circles also form an obvious
descriptive group. The circles are usually in rows and are composed of
an outline with a dot in the middle. This design element is seldom used
alone. A double circle is composed of two solid discs joined by a short
bar.
Lines are commonly combined with other elements. Thus the
subdivisions of the linear designs are: Line and circle, line and curve,
line and wavy line, line and square, line and band, line and diamond, and
line and angle. When the design is more elaborate than a simple line
pattern or a simple combination it is classed as a "complex linear."
Step designs are typical of Tiahuanaco ware. The step pattern,
although relatively obvious and simple, has considerable variation.
Posnansky' gives the types in his Plate XLVI. including the design
elements which are here called zigzag, angular, angular scroll, and cross.
An angular design consists of a V-shaped figure, often repeated, one within
the other. When two rows of such figures, one upright, one inverted,
are placed side by side, the effect is that of parallel rows of zigzags.
The scrolls are simple and occur in connected series around a vessel.
When the turns are formed by angles instead of curves this design is
called "angular scroll." The double S is somewhat like a scroll on end.
The two separate S's are placed one above the other, and in this arrangement used in parallel vertical rows for the design.
'Posnansky, 1914.

402

Arnthropological Papers Americtn lMuseum of Natural History. [Vol. XXXIV,

An interlocking design is difficult to describe, but is relatively simple

as the drawing (Fig. 17b) shows. It is really mis-named, but perhaps the
term will suffice. A continuous zigzag wide line has the intervening
V-shaped spaces filled with L-shaped lines which project alternately from
the ascending and descending line of the zigzag. Small circles almost
always accompany this design.
The miscellaneous geometric classification includes such obvious
figures as: squares, crosses, triangles, zigzags, and stars. Obviously any
of these design elements can be and often are combined; the classification is based on the dominating element. Some designs are frequently
associated with certain vessel shapes, as will be shown in the section
on shapes (p. 406).
The 18.01 per cent of zoomorphic designs fall into a few classes,
thus:Parts of complex designs
Condor designs
Puma designs
Human representations

Snake designs
Bird designs
Unidentified animal parts

Per Cent
29.90
31.66
9.55
18.84
2.01
.75
7.29

100.00 (398 pieces)

The first group, parts of complex designs, includes the sherd designs
which were obviously part of a complex figure, but which were too fragmentary to be identified with one of the known types. Condor, puma,
human, and snake designs are (with the possible exception of the snake)
the most typical Tiahuanaco design elements, and as such will be discussed at greater length. Bird designs are rare, and relatively late, as
the same type is found in the late Chullpa ware throughout Bolivia.
Finally, the group of unidentified animal parts includes a number of
sherds which, with slightly more evidence, might be included with the
puma group. As a matter of fact, the combination of this group with the
pumas would give a total percentage of 16.84 which is more nearly the
correct ratio of the prominence of puma designs.
The relation of design group to color division is shown in this table:Ware

One-color
Two-color
Three-color
Four-color

Zoomorphic
Per Cent
2.03
15.37
48.31
82.22

Geometric
Per Cent
97.97
84.63
51.69
17.78

Total Pieces
938
735
474
45

1934.1

Bennett, Tiahuanaco Excavations.

403

The high correlation between complexity of design and multiplicity of

color is an obvious characteristic of Tiahuanaco ware.
With the tabular and descriptive discussion of subject-matter in
design presented above, the more difficult subject of style of design is now
broached with some caution. If design is treated without regard for
stratigraphic position of sherds, two styles, which might be called " classic " and "decadent" are immediately evident. Like the colors, the
design elements are essentially uniform throughout Tiahuanaco ware.
However, in the colors, a division in brilliance is noted, and in the
designs a division in style is noted. The distinction in style and color is
readily observed, but the establishment of chronological distinction will
be left for the section on cultural stratification.
Classic and Decadent Style. It is not possible to present a thorough
analysis of Tiahuanaco style at this time. Moreover, such an analysis
has never been satisfactorily made, although Posnanskyl does list many
of the elements found. However, the detailed study must be reserved
for the future. A brief review of some of the salient features of Tiahuanaco style may serve to distinguish two periods, Classic and
Decadent. A division of color treatment into a rich, varied group and a
drab, restricted group has already indicated the Classic-Decadent distinction. Subject-matter, on the other hand, was shown to remain consistent
throughout, though a more detailed study would reveal the introduction
of new design elements and the elimination of old motives in the Decadent phase. It is the style of decoration which most clearly marks the
division.
Since this account is based on the collection actually excavated,
supplemented, to be sure, by notes on museum and private collections,
the style elements referred to are from pottery samples instead of stone.
This separation of ceramic and stone design is not completely artificial.
The winged and tailed king condor, the profile head, full front body warrior, and even the elaborate curled tail puma so commonly found on
ceramics are rarely, if ever, portrayed in stone. On the other hand,
the elaborately crowned, staff-holding, straight front view Viracocha
god, the running condor and human figures (of the Gateway of the Sun
and statues), the sun god heads (of the base frieze of the Gateway of the
Sun), are not, within my observation, found on Tiahuanaco ceramics.
Distinctions of this kind can be further elaborated. Thus fish heads,
caracoles, the "female sex" (?) sign, and other details common in stone
design are rarely found on ceramics. From our limited knowledge of
lPosnansky, 1914.

404

Anthro pologic4 Papers American Museum of Natural History. [Vol. XXXIV,

Tiahuanaco textiles it may be said that textile and stone designs form a
group as opposed to ceramic designs. Furthermore, this distinction
between stone carving designs and ceramic designs is applicable only to
the Tiahuanaco site, and certainly does not apply to the Epigonal or
derived Tiahuanaco style of the central and southern Peruvian coast.
In these last-mentioned regions the design elements are frequently taken
directly from the Gateway of the Sun and other stone carving.
Aside from the subject-matter which is a distinguishing characteristic
of Classic Tiahuanaco design, a group of traits can be given which
together typify this Classic Tiahuanaco style, although taken individually
a trait may not be unique.
With the exception of the wide, open, flaring rim bowls, where the
exterior is left unadorned and the design is confined to the flaring rim,
the designs are regularly restricted to the outside of the vessel. A scallop
or a wavy line may adorn the inner rim of a vessel, but the principal
design area is on the outside. All of the outside surface is treated equally,
that is, the design is not restricted to one half or one section of the vessel.
Drawings, particularly zoomorphic ones, are regularly outlined with
a heavy dark line, and then filled in with varying colors' and detail. The
figures are drawn in profile, both head and body, except in human figures
on which only the head is in profile. There are no attempts to vary this
rule in the typical Classic work. No pumas are found with four legs
or three-quarter front heads. However, the figures are generally right
side up and in natural positions. The proportions are fairly true. Spots
on the pumas, crowns and feathers on the condors are attempts at realism. Such a restricted and conventionalized style may be called pictorial only in its emphasis on the identification of the subject-matter as
opposed to an emphasis of decorative design. However, little attempt at
scenic composition is made. Figures or units are regularly repeated
around the vessel, two, three, or four times. Ordinarily no variation,
except for colors, occurs in the repetition of the units. Nor is more than
one zoomorphic figure often used. Combinations of man and puma or
condor are exceptional.
Restrictions and qualitative distinctions of color and shapes which
serve to distinguish the Classic from the Decadent periods are mentioned in other sections. The combination of all factors makes the mass
lIn the illustrations of pottery designs in this paper the accompanying color key has been employed.

YELLOW EnORANGE

n GRAY

RED

BROWN

1934.1

Bennett, Tiahuanaco Excavations.

405

distinction of truly Decadent wares and truly Classic a fairly obvious

one. However, detailed analysis proves embarrassing, because of the
absence of sharp distinctions at any given point. If the degenerate style
represents a true cultural decay, rather than a sharp overthrow, or
cultural change, the change of style would be gradual and at the same
time irregular in respect to its component elements.
The design in the Decadent style continues to be located principally
on the outside of all vessels, but there is often a restriction in the actual
area covered. A band around the top of a kero-shaped cup will compose
the only decorated area, for example, but on the whole, the Tiahuanaco
penchant for well-distributed design is unviolated.
Outline is still the most common figure treatment technique, but
solidly filled black figures are not infrequent. Also the fill and the detail
show much less variety in the Decadent than in the Classic style. A
general simplification of technique, color, shape, design, and subjectmatter (especially figure treatment) is perhaps one of the most characteristic distinctions for the Decadent epoch.
The attempt to identify the subject-matter of the drawing (which
was called pictorial in order to avoid the term realistic) has become subordinate to pure design. Details of Classic animal drawings are used as
design elements in the Decadent style. First heads are used without
bodies, then eyes without heads. No longer is it important to have the
figure in a semi-realistic position. The repetition, especially of heads,
quite often fills out the pattern by alternating one upright head with one
upside down. Front views are fairly common, especially for human
heads. There is a general confusion of elements and combinations which
would have been sacrilegious to the Classic designers. Puma heads are
attached to bodies with wings and tail feathers, and condor heads have
feline paws. A puma head, a snake body, and a condor tail are a common combination. Perhaps the easiest method of describing the
Decadent type is by specific figures.
The elaborate and dignified king condor figure of the Classic epoch
becomes depleted in the Decadent. The feet, tail, or wing are often
omitted. The body may be reduced to a curved solid band with a wing
attached. Heads are commonly used alone. Tail feathers are regularly
used alone as a design unit. Another common depiction of the condor
is that of an elongated band representing the body (without feet or legs),
but with an upturned head at one end, a tail at the other end, also upturned, and a wing in the middle.
The Classic human figure has rather realistic features, curved profile,
rounded or winged eyes, and an elaborated headdress. The body is

406

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

presented in front view and the hands clasp objects, such as axes, trophy
heads, copper pins, and so forth. In the Decadent style the design is
often reduced to the heads alone. Also its depiction may involve an
elongation, in which the profile and the back of the head are separated
from the eye by an abnormally long distance. The eye in turn may be
more conventionalized and often becomes the whole design element.
Furthermore, the heads may be turned upside down or they may face
each other so that one nose serves for two profiles.
There is little doubt that a detailed study of design would reveal
definite stages in the Decadent styles, leading eventually to the so-called
"Chullpa " style of the region. However, it is doubtful whether excavation will ever confirm such minute divisions stratigraphically, although
their existence cannot be denied.
POTTERY SHAPES
It is difficult to determine the shape of a vessel accuratelv from a
small fragment. Consequently, this discussion of Tiahuanaco shapes
draws upon the whole vessels, as well as sherds, and upon supplementary material from the Tiahuanaco collections in museums. However,
the discussion is by no means exhaustive.
Just as in other ceramic traits a few were found which were dominantly Tiahuanaco, so five shapes typify the collections. Thus the identifiable sherds and the complete vessels give the following frequencies:Shape A, kero-shaped cup
Shape B, hollow base incense
Shape C, wide flaring bowl
Shape D, vase
Shape E, flaring cup
All others, shapes a-t

196
129
37
13
88
93

Total

556

Table 3 shows the relation between shapes and depth of occurrence

in the various pits. This evidence will be used in describing the characteristics of each shape, as well as in the later section on cultural sequences. The relation of shape to design is summarized in Table 4.
First the various shapes will be discussed. Fig. 12 is a key chart to
these shapes.
Shape A, Kero-Shaped Cup. The kero-shaped cup is the most typical
shape of all Tiahuanaco ware (Fig. 12A). It is represented in the typical
stone-carved god figures, usually as held in the left hand. It occurs in all,
save perhaps the earliest, Tiahuanaco periods. Some form of this kero-

TABLE 3
FREQuENCY DISTRIBuTION OF VESSEL SHAPES

Shapes
Pit I

Pit I

1
2
3
3a
5

E
OC aD

IN

Prrs AND LEVELS

b c d e fg

jk

n o

rs

p q

2
3
1

1
2

1
1

1
1

31

PitIll

1
1
2
21

Pit IV

1
2
3
4
5
6

40
29
12
5
16

1
2
3
4
5
6
7
8
9
7a

2
18
1
1

Pit V

Pit VI

Pit VII

Pit VIII

Pit iX

Pit X
Totals

4 52
1
4 5210
1
5 1
1
1
1
1
3 5
1
3

33
1
5
1
2

1 2
1
2

1
1

1?

1
1
1
1
1

1
1

1
1
4
4
14
4

2 1
1

1
3

2
9

11

1 1

2
1

1
4

11
2
1

11

1
2
3
4
5
6

10 3
13
2
2 3
4
4

1
2
3
4
5

3 2 3
2
14
2
2

1
2

4
3

196

50

60

2
21
1

10
2
3

2
1

41 1
11 1
2

53 2
5 1
1 3

129

1
2
1

1
1

5
6

4
3

8
8

19

37

13

88

1
1
1

11I 1
2

17

2
2 1

51 3 7 11 4 4 8 6 6 3 1

TABLE 4
FREQUENCY RELATIONSHIP OF SHAPE TO DESIGN
0

~~~

ShapeA
Bolivia Museum
B

68
3

7
3

15

Bolivia Museum
9

D
Bolivia Museum

Bolivia Museum

1
1

2
1
2 1

14

26

25

7
10
8/6 3/5

2
8
1

1
2

4
5

/1

8
4/

5
1/

53

104
45

32
19

1+ 7

34
7
12

/4

1/3

/2

/2

/1

/3

/3

/3

11

1
1

1
2

/21I/

d
Bolivia Museum

C
Bolivia Museum

3
11

~ ~

Llama
64
3
Llama
121
1
1
2

13

Bolivia Museum
jlivia
M
Bolivia Museum

4Circle
4

1
1

/2

/1

1;olina
Museum2
Bolivia Museum
14

2
7

Bolivia Museum

11

p, q, r

B3olivia Museum

1/2 1

A~~~~~~~

(Jcz:

t'Ce
C

Fig. 12. Key to the Pottery Shapes Described. A-E, the major shapes encountered both in excavation and in collections; a-s, less frequent shapes contained in the
Expedition collection.
407

408

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

shaped vessel is represented in Epigonal and other cultures derived from

Tiahuanaco influences. It even extends into the Inca period: where the
true kero made of wood, supplants it, although a black ware cup is still
found in abundance. The typical plain kero is about 6 to 8 centimeters
high. The sides are perpendicular from the round base to about the
middle of the vessel from which point they flare out until the diameter
at the rim is over twice that of the base. Aside from this gradual flare
the rim is quite plain. The commonest variant of the plain form is the
kero with a horizontal ridge around the body. A groove on the inside
of the cup marks the position of the ridge on the outside. Most ridges
are narrow, but one variant has a raised band about 1.5 centimeters
wide. The ridge inay be located around the upper half of the cup, the
middle, the lower half, both around the upper and lower sections; and,
finally, it may be double around the middle. A kero with a head modeled
slightly in relief at its base is another variant. The head may represent
a bird, a puma, or a human being. A related shape, typical of Cochabamba, is a vessel with a very constricted base, from which the sides
spread out, in funnel style, to a rim diameter four times that of the base.
The straight-sided, non-flaring, kero-shaped vessel, rather typical of
Coastal Tiahuanaco style, was not found in my excavating, nor is it
common in the existing Highland collections.
Shape A-vessels are found in almost every pit and, with the exception of the lowest levels, at most depths. They are particularly abundant
in Pit IV where the percentage of painted sherds runs high. Many
varieties of design are found on vessels of this shape. However, the wavy
lines, scrolls, and complex angular designs so typical of some other
vessels (Shape E, for example) are not found, and there is considerable
evidence that these designs are comparatively late. Both Classic and
conventional or Decadent zoomorphic designs are found on the keros,
indicating that the shape persisted, though the design changed.
Shape B, Hollow Base Libation Bowls. This libation bowl is another
of the most typical shapes found at Tiahuanaco (Fig. 12B). Its distribution thtoughout the levels of each pit is even more complete than
for the kero-shaped cup. However, it is apparently limited to the Tiahuanaco site and immediately adjoining regions, such as the Island of
Titicaca. In general, these vessels may be described as squat, open bowls
with two handles, in which the base is inserted like a partition about 2
centimeters from the bottom of the vessel. This gives rise to the
descriptive term "hollow base." The sides are concave in profile, the
rim being half again as great as the base in diameter. Otherwise the rims

1934.1

Bennett, Tiahuanaco Excavations.

409

are plain or undulated. Careful examination of the collections, both

museum and excavation, allows the libation bowls to be subdivided as
follows
a. A flat bottom (not "hollow") perpendicular-sided, cup-shaped
bowl, with from four to six rim undulations. These bowls are decorated
either by dull-point engraved lines filled with red, yellow, and white
paint, or by plain painting in the shiny colors typical of the "special
paint" ware already mentioned. Designs consist of steps and angular
scrolls, and some un-outlined birds or animals. A small modeled animal
head is found on the rim of some of these as a variation. Because of the
undulations, animal heads, and general shape, these vessels have been
classed as a subtype of Shape B. Means' illustrates one of these cups as
probably representing the Tiahuanaco I period. This deduction is
borne out in general by my excavation, but, unfortunately, I did not
find enough of this type to check it absolutely.
b. Bowls of plain, unadorned, brown ware with the base indented
less than one centimeter (average 0.6 centimeter), with sides without
extreme flare and without handles, form a rather definite group which is
further identified by its distribution in Pits III, 2; V, 8; VIII, 4, 5, 6.
Its significance is emphasized when compared with the "special"
painted ware and the horizontal handle (Shape a), also limited to the
lower levels of Pits V and VIII.
c. A group of plain orange ware has the bottom partition indented
2 to 3 centimeters. The ware is well fired and of medium thickness.
The bowls are of average size, about 12 centimeters high, 12 centimeters
in base diameter, and 15 centimeters in rim diameter. Six fragments
have stubs of vertical handles. Most of this orange ware has no further
decoration, but on some a trace of white or red paint is still evident.
Vessels of this type occur in Pit IV, all levels, and others are found in
Pits II, 1; V, 2; IX, 1; X, 1.
d. This group differs from the preceding (c), only in being painted.
Design is confined to the outside, although the inside is usually painted
some plain color, commonly red. These are two-handled bowls, with the
rim plain and slightly wider in diameter than the base, and the base
raised from 1.5 to 4 centimeters. The handles start below the partition
and extend to a point below the rim. They are all set vertically, although one exceptional bowl with a horizontal handle is represented in
the Museo Nacional de Bolivia. The typical specimen has a design
composed of a row of circles around the base. The upper body designs
'Means, 1931, 107, Fig. 5.

L~~
a'

II1111411111111 Lil

IiilifilIl

nsr->

II
II 1 1I
1
IillilliilVlillilI

bp

II III II 11111

111Q!LOwIlilpittli

AVMlEi "l-WRfiUW?
C'

d'

7I
--

Cc- i

S_
-

Fig. 13. Shapes and Designs on Early Type Tiahuanaco Pottery. a-a', incised design filled with colors (in incisions) on a flat
bottomed, wavy rim libation cup (Museo Nacional de Bolivia); b-b',
the puma head type flat bottom cup with painted design (Museum
fur V6lkerkunde, Munich); c-c', long-necked decanter shape with
the typical angular design in several glaze colors (Museo Nacional
de Bolivia), the most typical design of all ware of this period;
d-d', a variation of design on the same type bowl as shown in
c-c' (Museo Nacional de Bolivia); e, f, g, plain ware shapes of this
period encountered in the excavations.
410

411

Bennett, Tiahuanaco Excavations.

1934.1

vary somewhat, although a trident feather design is very common.

Condor designs occur, but other zoomorphic figures are rare. As in the
kero-shaped cups, the "late designs" (wavy lines, double S, scroll,
complex angle) are not found with any frequency. Many colors are

III~~~~~i

Fig. 14. A Typical Early Tiahuanaco Bowl Shape with Four Examples of the
Type of Inner Rim Design. The spittoon shape showni in a-b illustrates the characteristic Early design w-hich occurs around the outside of most of these bowls. c,
zoomorphic figure on bowl (Musee d'Ethnographie, Paris); d, zoomorphic figure
on bowl (Museum fur Volkerkunde, Munich); e, zoomorphic figure on bowl (Mus6e
d'Ethnographie, Paris); f, zoomorphic figure on bowl (Museum fur Volkerkunde,
Berlin). c-f are painted in glaze colors on a black background section, a technique
which Means calls "rimming."

commonly employed in the designs, polychromes of four and five colors

typical.
e. Except in Subtype a all the bowls mentioned so far have had
plain rims. Another set of hollow base bowls has an undulating rim.
These undulations may be regular loops, angular lugs, or pierced proare

412

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

jections resembling one type of horizontal handle. Fragments of this

type are not frequent in my own collection, but complete, undulating
rim bowls are common enough in the Tiahuanaco collections.
f. Bowls with modeled animal heads as rim projections form a final
subgroup. The combination of modeled animal heads and rim undulations (Subtype e) is common. Modeled heads are also found on plain
rim vessels (like Subtypes c and d) in which case the bowls have only one

handle.
Modeled heads include puma, condor, one monkey (Pit V, 9), and a
llama form. My collection includes only the puma, condor, and monkey.
One form of solid puma head is modeled directly on the rim edge, with
or without small differentiated ears. The head faces out on an upward
incline. The modeling is simple and only slightly augmented by incised
grooves. Three of these were found (Pit II, 2; Pit V, 6 and 8). They
form a distinct class in contrast to the other types found. Another
modeled puma head is hollow and much larger. The modeling is in
greater prominence. The head is not directly attached to the rim,
but connected by means of a long, hollow neck. The head has either
two large flat ears or a thin, flat, square collar which separates it from the
neck. Rarely are both ears and collar employed. The condor heads are
mostly of this collared type. The monkey head has a neck, but no collar,
and is hollowed out with the mouth as a spout. One fragment of the
collar from Pit IV, 5 is notable for its coloring. One side is covered
with a thin red slip, while the other has an indistinct design in black,
white, green, and red chalky colors. Other heads are either plain or
colored with a single slip.
Seventy-five typical fragments of hollow base bowls of my collection are distributed throughout the types as follows:Subtype a
b
c
d
e

1
12
18
32
3

9
f
Incised design on libation bowls is found in several of the Bolivian
collections, although unfortunately I did not find any fragments. There
is again some indication that the incised type is older than the plain or

painted.
Shape C. The wide open, flaring rim bowl is characteristically Tiahuanaco (Fig. 12C). Rising from a slightly projected base the short sides

1934.1

Bennett, Tiahuanaco Excavatioms.

413

are convex and terminate in a broad, flat rim which angles outward and
upward. All decoration is on the inner side of this rim. The designs on
these vessels are typical of the Classic style, while truly Decadent designs
are not found. Condor and puma and, rarely, human designs are most
frequent. A typical bowl of this shape measures 12 centimeters high,
12 centimeters in base diameter, and 24 centimeters in rim diameter,
while the width of the rim is 7 centimeters.
Shape D, Vase. A lozenge-shaped vase with a high flaring rim is
fairly typical of Tiahuanaco, although not as frequently found as other
shapes (Fig. 12D). Zoomorphic designs are the most common type of
decoration on these vases, and although both Classic and Decadent styles
are included, the Classic is dominant. Even the geometric designs favor
the typical step and linear patterns rather than the late wavy line and
scroll types. These vases are handleless. The design is located around
the body of the bowl between a band around the neck and a band a few
centimeters above the base.
Shape E, Small Flaring-Sided Bowl. A small bowl (Fig. 12E) has a
circular flat base about 8.5 centimeters in diameter, from which the sides
flare out gradually to a diameter slightly more than one and one half
times the base, or about 14 centimeters. The height of these bowls is
about 8 centimeters. The rims are plain. There are no handles. In
some ways this shape resembles an unelongated kero-shaped cup (Fig.
12, Shape A), although there is no evidence for its being a derived forn.
These bowls are the typical medium for the Decadent style design.
Zoomorphic forms of design are in the minority and those found, with
few exceptions, are Decadent in style. The geometric designs are predominantly wavy lines, steps, complex angles, scrolls, and double S's
which practically never appear as independent designs on vessels of
Shapes A, B, C, and D. The colors of the bowls are not the rich, deep
colors of the Classic Tiahuanaco, but rather the flat, opaque tones of the
Decadent style. The shape is numerous in the collections, and although
generally confined to the upper levels of the pits, has a disturbing distribution. A straight-sided (instead of concave) variant has a wide
distribution in connection with the spread of Tiahuanaco influence.
It is interesting to note that Tiahuanaco style on the coast of Peru has
both this shape and the kero-shaped cups in straight-sided variants.
This shape is preserved in the Inca ware of the Bolivian Highlands.
It is also one of the most characteristic shapes of the so-called Chullpa
ware, widely distributed throughout the Bolivian plateau and Lake
Titicaca region.

414

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Shapes A-E, described above might be called the primary Tiahuanaco types, judged on a quantitative basis. Sparsely represented
in the excavations are numerous other shapes which deserve mention.
They are briefly described below as "secondary" types. It does not
follow that other excavations might not materially increase the importance of some of them.
Shape a, Horizontal Handled Dishes. Horizontal handles are rare in
Tiahuanaco pottery. They appear occasionally as variations of vertical
body handles. The group included here is composed of horizontal rim
handles. All but two are plain ware. All come from three pits: Pits
V, 7a and 9; VIII, 4, 5, 6; III, 1, 2. It has already been mentioned
that pierced rim undulations of Shape B libation bowls have the appearance, when broken off, of horizontal rim handles. However, only one
specimen seems to be of this type. The other sixteen are divided into
two groups. Nine are loop handles which project at only a slight angle
from the vertical extension of the rim. The fragments of the dishes are
too small to determine the shape accurately, but the suggestion is of a
shallow platter shape (Fig. 12a). Seven are loop handles which project
almost at right angles to the rim, and are apparently parts of deeper,
more bowl-shaped dishes. I have not seen any complete Tiahuanaco
specimen of this type either in the literature or in the existing collections,
but Bingham illustrates a similar type of shallow dish as rare at Machu
Picchu.'
Shape b, Shallow, Plain Rim Bowl. This shape has already been
mentioned in the discussion of the special group of red and brownishblack polished ware. It is a simple, convex-sided open bowl, with a
plain rim (Fig. 12b). The sides curve to a flat base. Fragments of this
type were found only in Pit V, 9, and Pit VIII, 4, 5, 6.
Shape c, Small, Flaring Rim Bowl. The simple bowl (Fig. 12c) is of
relatively little importance, except that in Pit VIII, 4 it occurs with the
special, shiny paint which distinguishes the lower levels of Pits V and
VIII from the general run of Tiahuanaco ware.
Shape d, Narrow Rim, Wide Open Bowl. This bowl is somewhat like
Shape C, except that the rim is narrower and less flaring, and the decoration is on the outside (Fig. 12d). Two vertical handles extend from just
below the rim edge to the curved body. A projecting ridge encircles the
flat base. The design, a scrolled snake body with human face and feet,
is in rich Classic colors (Fig. 17d) and, though rare, seems typically
Tiahuanacoan.
lBingham, 267, Fig.. 12a.

t/E

61?

W ! X ~ ~ ~ ~ ~ ~ ~ ~ ~l I
b

e
)F
Fig. 15. Shapes and Designs of Classic Tiahuanaco Types. a, the kero-shaped
goblet (A) is one of the most typical shapes. b, c, d, designs copied from goblets of
the same shape; b, head from a goblet (Museum fuir V6lkerkunde, Munich), more
elaborate than most Classic stvle human heads, but still represents the type; c,
masked warrior figure with the battle ax and trophy head was found by the expedition; d, Classic condor figure (Museum fur V6lkerkunde, Munich); e-f, twohandled and collar-puma head varieties of the "hollow base" libation bowls, f,
after Schmidt, p. 537, Fig. 1.
415

416

Anthropological Papers American Museurm of Natural History. [Vol. XXXIV,

Shape e, Modeled Puma Bowl. An effigy bowl in puma shape (Fig.

12e) is found in some of the collections of Tiahuanaco ware. The body is
formed by the elongated bowl. Four feet are attached; a large modeled
head projects in front, and an upright hollow tail in back. Only fragments of this type of bowl are in my collection. Two ears and part of the
body come from Pit IV, 5. The body is decorated with the double circle
connected with a bar, so typical of puma designs. Ears of modeled
pumas, perhaps of this type bowl, come from Pit V, 2 and 3. Pit IX, 2
furnished a hollow tail. In function it might, be classed with the incense
or libation bowls of Type B (Fig. 12B).
Shape f, Goblet. This goblet-shaped drinking vessel (Fig. 12f) may
be a special variant of the Shape A kero. The straight-sided circular
lower half bulges in the upper half in a prominent convex arc. The plain
rim is somewhat restricted and has none of the flare of the typical keros.
The design is typical of Classic Tiahuanaco style. However, there is a
variety in polished black ware which suggests Inca influence.
Shape g, Modeled Heads. If modeled heads on vessels are considered
only as part of the decoration, then there is no justification for this class
(Fig. 12g). However, the number of pieces with modeled human heads
is so small that it is easier to describe them as a group.
One group, similar to Shapef, is perhaps another variant of Shape A
keros. However, the modeled human head at the base of the goblet gives
it a distinctive appearance. The vessels with the modeled head bases are
not as high as the typical kero-shaped cup, and the sides curve inward
rather than outward. Heads of this type are found in plain, unfinished
ware, in polished and painted red ware, and in polished black ware.
Another group is composed of rin pieces, possibly of kero or Subtype f vessels. In the complete vessel the head would be upside down,
because the chin is modeled close to the rim, and the eyes modeled and
painted in below.
A final group, a large bowl with restricted mouth and with a high
collar, is modeled in the shape of a human head. A single handle extends
from collar to body opposite the modeled face. This type is probably
Inca, as almost the same style of vessels is found at Cuzco.
Shape h, Angular Rim Bowl. A few fragments of bowls with 2
centimeter wide rims which extend at right angles from the body were
found in the excavation, but no pieces were uncovered which would
correctly indicate the complete bowl shape
Shape i, Narrow Neck Globular Bowls. A medium-sized bowl with a
constricted neck and high straight collar (Fig. 12i) is sparsely represented

1934.1

Bennett, Tiahuanaco Excavations.

417

in the collection. One complete vessel with a projecting spout has a

snake body design with a puma head. Others of this shape are represented by collar fragments from which little of the design can be told.
Shape j, Round Bottom Bowls. An open, recurved-sided, plainrimmed bowl (Fig. 12j) is rather frequently found. It is somewhat the
same as the Shape E open bowl (Fig. 12E), but differs from it in its
round bottom. However, that one feature is sufficient to place this bowl
in a class by itself. The bottom is not completely round, but the junction of the sides and the base, instead of being characteristically angular,
is a bulging curve which extends well under the vessel.
Typical of this form is a zigzag line design called "interlocking."
This is combined with circles. Almost every example of this design is in
yellow lines and white circles on a red background. Other designs found
are in Decadent style, buit it is not possible to say that the shape too is
Decadent.
Shape k, Small Open Wide Rim Bowl. This is a small edition of Shape
C (Figs. 12C, k). It is not classified with the large ones, because of the
great discrepancy in size. Polished red vessels of this kind are most
typical, while painted ones are rare. These bowls are 10 centimeters in
rim diameter and 6 centimeters high, as contrasted to 24 and 12 centimeter measurements of the large size.
Shape 1, Round Base Open Bowvl. The round base open bowl again
exhibits a curving contact of base with convex sides (Fig. 121). The
bowl is wide, but the rim diameter is less than the body width. There is
an angular flaring rim of medium width. A scroll design on the body of
these vessels is typical, and is executed in black-on-red. In fact, it is
rare to find one of these bowls without a scroll design. Since other
occurrences of scroll designs have been noted on Decadent vessels it
might be assumed that this too is a late type.
Shape m, Large, High Shouldered Jug. This group consists of large
jugs with constricted necks and high collars (Fig. 12m). The rim may be
plain or flaring. The base is flat and the sides diverge in a straight line
up to the curved, high shoulders. Some have two loop handles set low
on the body.
The design, of step or condor, together with triangles, is in dark
brown-on-yellow. In design, general treatment, and form they resemble,
or at least suggest, the Atacamenio typical shape described by Uhle for
northern Chile.'
lUhle, 1919, Pis. XVII, XIX.

418

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Shape n, High Shouldered Vase. A shape which resembles m is a

high shouldered vessel which substitutes a wide mouth without collared
neck for the high collar of the jug (Fig. 12n).
Shape o, One-Handled Pitcher. One-handled pitchers (Fig. 12o) are
mostly in plain ware, or, if painted, are ornamented with simple geometric designs. They have a wide distribution in the Bolivian Highlands, in association with the "Chullpa" sites. They are undoubtedly
a late form in Tiahuanaco.
Shape p, Angular Body Bowl. A deep, wide-mouthed, flaring rim,
one-handled bowl has as its distinctive feature a body which is sharply
angular at about the center (Fig. 12p).
Shape q, Spouted Bowl. Bowls with spouts are rare in the typically
Tiahuanaco culture. One has already been mentioned as a variant of
Shape i. While my own excavation uncovered only the one mentioned,
and the spout of another, they are found in other collections, and distributed throughout the "Chullpa" culture of Bolivia. The usual style is
an extended body spout which is joined to the body by a bar, thus forming a spout handle (Fig. 12q). This spout handle is found on bowls of
several shapes, including the angular variety, Shape p.
Shape r, Animal Effigy Handle. Some rims have handles extending
to the body which are slightly shaped to resemble animals (Fig. 12r).
While some of these bowls are globular, others are angular bodied. The
Cochabamba collections of Tiahuanaco Decadent style abound in this
type of handle.
In general it may be said that Shapes p, q, r form a group of late
styles. Design and distribution confirm this statement.
Shape s, Small Modern Vases. Scarcely deserving mention are small,
crudely made vases of modern ware, found in the pits in association with
modern burials (Fig. 12s).
On the basis of color, design, and distribution the shapes may be
temporarily subdivided as follows (distribution refers to pit levels, and
also to non-Tiahuanaco sites like the Island of Titicaca and Cochabamba) :Specials
Early?
Ba, b
a
b
c

Classic

Bd, e, f
C
d
g
h

Both Classic
and Decadent
Bc
ADE

eij

Decadent

f kl
mno
p qr

A (Cochabamba
constricted base)

1934.]

Bennett, Tiahuanaco Excavations.

419

COMPLETE AND PARTIALLY COMPLETE BOWLS

Since most of the Tables, with the exception of those for the pottery
shapes, have been made up on the basis of sherds, exclusive of complete
vessels, it is necessary to give a brief descriptive account of the complete
and partially complete bowls. The more important of these bowls have
been illustrated in the drawings for the section on Cultural Stratigraphy.
The complete bowls can best be described in a catalogue list, arranged
according to pit, shape, style, colors, and design.
The ten pits yielded only forty-eight bowls sufficiently complete
to permit reconstruction. Most of the pits were made in the dwelling or
temple areas of the ruins and not in the cemetery. Pit X, the only one
really in the old cemetery section, yielded nineteen restorable vessels,
and it was a previously disturbed site.
1. Pit I, 3a; Shape A, ridge in center; Decadent style; black, white-on-red;
design of bands and irregular, rounded point cross.
2. Pit I, 5; Shape A, angular ridge in center; Decadent style; blackC, white,
yellow-on-red; design of front view faces, two upright, two inverted,
around base.
3. Pit IV, 1; Shape A, plain; Decadent style; black, white, brown-on-red brown;
design of two Decadent heads around top of vessel (Fig. 17f).
4. Pit IV, 5; Shape A, band around top; Decadent style; black ware; incised
lines around band.

5. Pit IX, 3; Shape A, plain; Decadent style; black, white, yellow-on-dull red;
design of band, bird, and large S-shape around top.
6. Pit X, 2; Shape A, plain; Classic style; black, white, yellow, gray-on-red;
design of four warrior figures with puma masks, carrying trophy head in
left hand and battle ax in right (Fig. 15a, c).
7. Pit X, 2; Shape A, plain; Classic style; black, white, gray-on-red; design of
two curled tail pumas with back legs bent.
8. Pit IX, 2; Shape C; Classic style; black, white, yellow, gray-on-red; design
in two rows around inside rim, at top a row of step, curve, and bird designs;
below a row of step and scroll.
9. Pit X, 2; ;ha4e'C; Classic style; black, white, yellow, gray-on-red; design of
three, curled tail pumas, separated by double S; around inside rim (Fig.
16 b, b').
10. Pit X, 2; Shape D; Classic style; black, white, yellow, gray-on-red; design of
two rows of steps (Fig. 16a, a').
11. Pit X, 2; Shap'e D; semi-Classic style; black, white-on-red-brown; design of
two condors with double band bodies, wings, and no feet.
12. Pit I, 3a; Shape E; Decadent style; black-on-orange; design of step and
perpendicular wavy lines.
13. Pit I, 5; Shape E; Decadent style; black, yellow-on-red; design of perpendicular wavy lines on outside, scallops around inner edge.
14. Pit IV, `5; Shape E; Decadent style; black, white, yellow-on-brown; design
dof t'wo Decadent heads around outside (Fig. 17e".)

'1'

4a

a,

b6

Ie

I\
I

_.'N'D",

..!

PI

z, "'

x
d

Fig. 16. Classic Tiahuanaco Shapes and Designs, all copied from Bowls found
by the Expedition. a-a', vase shape (D) with typical step design; b-b', open bowl
shape (C) with design on inner rim, of Classic pumas; c-c', constricted neck bowl
(with spout, i) with serpent design, in Classic technique, but not so typical as other
designs; d-d', this bowl and design are both unusual Classic style, but none the less
encorporate. The shape is classified as (d) in the list. The design is self explanatory.
420

Bennett, Tiahuanaco Excavations.

1934.]

421

15. Pit IV, 5; Shape E; Decadent style; black-on-orange; design of scallop around
inner edge.
16. Pit IV, 6; Shape E; Decadent style; black-on-red; top half of outside is black,
the rest of the bowl red.
17. Pit IX, 2; Shape E; Decadent style; black, white, yellow-on-red; design of
two llamas, circles and S around outside.
18. Pit IX, 2; Shape E; Classic style; black, white, yellow, gray-on-red; design
of three rows of steps around the outside.
19. Pit IX, 4; Shape E; Decadent style; black, white, yellow-on-red; design of
crude puma with condor wing and band body.
20. Pit X, 1; Shape E; Decadent style; black-on-red; design of line and scroll.
21. Pit X, 2; Shape E; Decadent style; black, white-on-red; design of line and
circle chain.
22. Pit X, 2; Shape E; semi-Classic style; black, white, orange-on-red; design of
two full, crowned condors, with winged animal body with two feet; design
in colors with black as a background, not an outline.
23. Pit X, 2; Shape E; semi-Classic style; black, white, yellow-on-red; design of
two pumas with tails that stick straight up.
24. Pit X, 2; Shape E; Decadent style; black, white-on-brown; design of a row of
steps and curves above a row of perpendicular wavy lines.
25. Pit X, 2; Shape E; Decadent style; black, white, yellow-on-red; design of two
upright and two inverted figures with condor tails and puma-like heads

(Fig. 17e, e').

26. Pit X, 2; Shape E; style uncertain; yellow, black-on-red; design of three llamas
with packs and leader, in yellow on black background; bowl small and red

inside; polished.
27. Pit X, 2; Shape E; style uncertain; yellow, white, black-on-red; design like
26, in yellow and white on black background; small bowl, red inside; not
as well polished as 26.
28. Pit VIII, 4; Shape b; Early style; plain, polished, reddish-brown.
29. Pit IV, 6; Shape like c; Decadent or Inca style; plain orange with horizontal
side handle.
30. Pit X, 2; Shape d; Classic style; black, white, yellow, gray-on-red; design of
scroll-like body with projecting legs and front view faces; circles around
inner rim; step design around outer rim (Fig. 17a, a').
31. Pit IX, 3; Shape g; style uncertain; plain, unslipped brown modeled head with
coca wad in cheek.
32. Pit I, 5; Shape i; with spout (broken); Classic style; black, white, yellow,
gray-on-red; design of snake body with puma head, step design around
base of collar, broken wavy lines around rim (Fig. 16c, c').
33. Pit IX, 2; Shape j; Decadent style; black, white, yellow-on-red; design of
Decadent heads around side.
34. Pit X, 2; Shape j; Decadent or semi-Classic style; yellow, white-on-red;
interlocking design with circles around outside (Fig. 16d, d').
35. Pit IV, 5; Shape k; Decadent style; plain red.
36. Pit V, 2; Shape k; Decadent style; plain red.
37. Pit IX, 3; Shape k; Decadent style; black-on-red; crude step on inner rim.
38. Pit X, 2; Shape k; Decadent style; plain red.
39. Pit X, 2; Shape k; Decadent style; plain red.

,. ":,-i 2,_:-r'
-

f-

'C,

Fig. 17. Shapes and Designs of Decadent Tiahuanaco Style. a-a', Cochabamba
style kero-shape with line and dot design (collection of Frederick Buch, La Paz);
b-b', rounded base cup (j) with interlocking design with circles, this design in yellow
with white circles is characteristically associated with this shape; c-c', two-handled
jug (m) with step design, design and shape are not unlike the typical Atacameno jug;
d-d-', open, rounded base bowl (1) with black.-onfred scroll design, a frequent combination; e-e'-e", shape E bowl, characteristic of Decadent phase with two samples of
design, one a depleted condor-puma combination, the other an elongated human face
in profile; f, kero-shape goblet (A) with Decadent human face design.
422

423

Bennett, Tiahuanaco Excavations.

.1934.1

40. Pit IV, 3; Shape 1; Decadent style; black-on-red; scroll design around outside
(Fig. 17d, d').
41. Pit IV, 6; Shape m; Decadent style; black, white-on-plain orange; irregular
step and horizontal wavy line design on outside (Fig. 17c, c').
42. Pit IX, 3; Shape o; Chullpa style; plain orange and with two side handles.
43. Pit IX, 3; Shape o; Chuilpa style; plain red and small.
44. Pit X, 2; Shape o; Chullpa style; plain dark brown and with one handle and
lug on front side.
45. Pit I, 1; Shape s; modern style; plain orange with one handle and modeled
face.
46. Pit V, 1; Shapes; Chullpa style; black-on-orange; small cross design.
47. Pit V, 1; Shape s; Chuilpa style; plain orange.
48. Pit X, 2; Shape s; Chullpa style; plain orange.

In the complete vessels the style is easier to determine than in the

sherds. To be sure there are intermediate types which cannot definitely
be assigned to one style, but this is only a confirrnation of the theory
that Tiahuanaco culture represents a continuous change. The distribution of complete bowls according to style and shape is as follows:Shapes

Style
Early
Classic
Semi-Classic
Decadent
Chullpa (Post)
Uncertain

2A, 2C, D, E, d, i
D, 2E, j
5A, liE, j, 5k, 1, m, c-like
3o, 4s

2E, 9

Totals
1
8
4
25
7
3

48

This is the same general alignment as that given at the end of the
section on shapes. Since the distribution of complete bowls according
to Pit and Level also confirms the sherd analysis it is summarized here
without further comment:Pit
I, 1
3a
5

IV, 1
3
5
6

Shape
s
A
E
A
E
i
A
1
A
2E
k
E
m

c-like

Style
Modern
Decadent
Decadent
Decadent
Decadent
Classic
Decadent
Decadent
Decadent
Decadent
Decadent
Decadent
Decadent
Post, or Decadent

Pit Totals

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

424

Pit

Shape

Style
Carried Forward

V, 1
2

2s
k

VIII, 4
IX, 2

b
C
E

~E
..

A
9

4
X, 1
2

k
2o
E
E
2A
C
D
D
2E
3E
2E
d

j
2k
0
S

Pit Totals
15

Chullpa
Decadent
Early

3
1

Classic
Classic
Decadent
Decadent
Decadent
Uncertain
Decadent
Chullpa
Decadent

10

Decadent
Classic
Classic
Classic
Semi-Classic
Semi-Classic
Decadent
Uncertain
Classic
Semi-Classic
Decadent
Chullpa
Chullpa

19

Total Bowls

48

The complete bowls are included with the sherd classifications in the
analysis of Cultural Stratigraphy. The collection of complete bowls is
not altogether typical of the Tiahuanaco site. The really fine mortuary
pottery is only sparsely represented. Some of the pits, like IV, contained
a fair percentage of Classic sherds, but the complete bowls are, without
exception, Decadent. Perhaps this is really the true picture, and it does,
indeed, confirm the analysis given of this pit in the section on
Excavations.
ARTIFACTS
Artifacts are not abundant in the collection. Potsherds outweigh
all others, and other fabricated pieces are disproportionately scarce.
Thus I describe the artifacts which occur in the collection for the sake
of completeness, and not with any implication that a typical or significant array for Tiahuanaco is portrayed. The excavation evidence,

1934.]

Bennett, Tiahuanaco Excavations.

425

when combined with notes on distribution throughout the area, suggests

that a careful study of the artifacts would produce significant results.
Clay. Pierced discs of clay, probably made from potsherds, are the
only suggestion of spindle whorls found. They are about 4 centimeters
in diameter and .5 centimeter thick. One in Pit V, 6 has an elongated
oval shape, 4 by 3 centimeters and .6 centimeters thick. A triangular
notch is cut from each of the four sides.
Unpierced clay discs from 2 to 6 centimeters in diameter are found
throughout the excavations. They are likewise abundant in Bandelier's
collections from " Chullpa " sites in the Bolivian Highlands. They were
possibly used as gaming pieces.
Fragments of pottery with holes, perhaps to facilitate mending, are
frequently encountered. This is another trait found in the Chullpa sites.
An earplug was found in Pit I, 3a. It is of brown finished clay and
made in modern spool shape. The diameter at each end is 3.2 centimeters and the total length, 2.2 centimeters. A hole 1 centimeter in
diameter pierces the spool. The groove around the center is 1.2 centimeters wide and .4 centimeter deep.
Most interesting of the clay artifacts is a set of hemispherical buttons. The flat surface has a hollowed-out depression in the center.
From this depression two small holes extend to the sides of the button,
running parallel to the flat surface. The average flat surface diameter
of the nine buttons found is 2.1 centimeters and the average thickness is
1.5 centimeters. There is very little deviation from these averages.
Only one button is decorated with parallel grooves across the rounded
surface. The buttons were found only in one pit. Four are from Pit
VIII, 4, and five from VIII, 5. In Pit X, 1 a stone of the same general
shape was found, but it was neither hollowed nor pierced. Bandelier
found similar buttons on the Island of Titicaca in sites which had no
Inca pottery.
Stone. Although rather well carved stone bowls are found in Tiahuanaco collections my excavations only uncovered four fragments.
In Pit III, 1, was a fragment of a deep granite bowl of fairly well finished
workmanship. The stone is finished on both sides. Across the outside
is a raised band, 3.8 centimeters wide and .4 centimeter high.- In Pit
IV, 2 was a section of a well-polished tripod bowl of black, fine-grained
stone. The bowl section is 5.5 centimeters high and 11 centimeters in
diameter. The thickness varies from 1.5 centimeters at the bAse to .8
centimeter at the rim edge. Only the base of one tripod leg remains,
and this measures 4 by 1.1 centimeters. Presumably there were two

426

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

more such flattened projections for legs. The bowl looks suspiciously
Incan. A small, red sandstone, rounded mortar) finished inside and out,
was found in Pit V, 9. It is 3 centimeters high, 7.5 centimeters in diameter, and .7 centimeter thick at the base. The base is rounded, the sides
convex, and the rim plain. The fourth fragment is really part of a
granite mortar, finished on the outside and rough on the inside. It was
found in Pit IX, 1. It is 7.5 centimeters high, 17 centimeters in diameter,
and 1.5 to 3 centimeters thick. It is a circular, flat base, straight-sided,
bowl-shaped mortar.
Hammerstones are of granite, cylindrical, with rounded edges, and
straight sides. A typical hammer measures 4.5 centimeters in length and
3.5 centimeters in diameter. One from Pit VIII, 1 is slightly grooved
in the center.
Plain stones are utilized as polishers, but are otherwise unfabricated. They are of miscellaneous shapes and sizes, and most are of
sandstone. There is one muller-shaped grinder or polisher from Pit
IV, 5 with a base 8 by 6 centimeters, a top 5 by 4 centimeters, and 8
centimeters in height.
In Pit X, 1 was found a granite ball 3.5 centimeters in diameter.
All the abrasive stone work is typical of both Chuilpa and Tiahuanaco
work in Bolivia.
Pieces of chalk which were used as polishing stones were found in
several places. When the chalk piece has been squared off on four sides
of a broken fragment it looks like part of a chalk bowl. However, the
inside is rough and has the appearance of some type of plaster, although
that it is cannot be substantiated fully.
In Pit V, 9 a T-shaped polished blue basalt ax was found. It
measures 5.8 centimeters across the top, or butt end, 7.5 centimeters in
length and 1.8 centimeters in maximum thickness. Posnansky found five
axes of this type, all at considerable depth. The only other ax shape is a
polished green celt butt found in Pit IV, 5. The tang measures 2.5
centimeters in width and length and 1.2 to 2 centimeters in thickness.
Where the tang meets the body of the celt the width is 4 centimeters.
The fragment is 5.3 centimeters long to the break, where it measures
4.5 centimeters in width and 2.5 in thickness.
Flakes of flint, slate, quartz, mica, lapis-lazuli, and obsidian are
found in most of the pits. Small lapis-lazuli beads are also found (Pit V,
9 and VII). Most of the flakes are unworked, but one piece of obsidian
from Pit V, 6, roughly triangular in shape, had one edge retouched.
In Pit VIII, 3 a piece of red flint was roughly shaped into a triangular
arrow point.

1934-1

Bennett, Tiahuanaco Excavations4427

Stone "tops" are rather numerous. Typical specimens measure

about 4.5 centimeters in length and 3 centimeters in butt-end diameter,
tapering from the flat end to a dull point. Most of them are made of red
sandstone, though a few are of granite. They vary from elongated to
squat shapes, but the general aspect is always the same. The shape is
pecked out, and never polished. Tops occur in most levels of the Tiahuanaco excavations, as well as in the Chullpa sites of Bolivia.
Although both modeled and carved stones are found in other collections, only two examples of modeling, one doubtful, are in my collection.
Statues are excluded from this analysis. In Pit VIII, 2 was found a small
red sandstone animal, roughly resembling a bear (?). The fore feet are
stubby, but distinct, while the hind feet are merged into one. The head
is roughly shaped, but no features are shown. A similar specimen was
seen in the Cochabamba collections from a non-Tiahuanaco site.
Another from Pit IV, 4 is of a fine-grained black stone which has a
projection in the shape of an ear. The fragment is too small to judge
its exact nature.
Miscellaneous stone fragments include the following
Pit V, 3 and V, 5
Pit V, 6
Pit X, 1
Pit III, 1
Pit IV, 6

Sections of polished slate

Small piece of granite, squared on four sides
Piece of red sandstone, polished on one side
Polished piece, possibly bowl rim
Triangular piece of sandstone

Copper. Small pieces of copper sulphate and other copper ores

were found throughout the excavations, but actual artifacts of copper
were exceedingly rare. Poor preservation accounts for the lack. A thin
copper piece with one edge straight and the others slightly curved was
found in Pit VIII, 4. It measures 4.8 by 3.5 centimeters. A small hole is
pierced in the center, which suggests that it may have been sewn on
cloth. Also in Pit VIII, 4 was a small bar of copper 4.2 centimeters long.
From one rounded end .6 centimeters in diameter it tapers to .3 centimeters in diameter. A plain headed topo or pin was found in Pit IX, 3.
It is 10 centimeters long and .4 centimeters in diameter at the rounded
head, from which it gradually tapers to a point.
Bone. Bone artifacts also are not abundant. Two plain needles
were found in Pit VIII, 2 and 4, respectively. The first is 13 centimeters
long and .4 in diameter, while the second is 8.5 centimeters long and .35 in
diameter. Both are round in cross-section, squared at one end, pointed
at the other. Other bones are pointed for some type of tool, although not
as completely finished as the needles. In Pit X, 1 was a polished and

428

Anthropological Papers American Museum of Natural History. IVol. XXXIV,

pointed bone 14 centimeters long. It is flat in cross-section, 1 by 3

centimeters. Broken pieces of the same type were found in Pit VIII, 2
and 5. Also in Pit VIII, 2 was a short pointed bone, 5.8 centimeters
long and 1.7 wide, unshaped except for a short, wedge-like point. In
Pit VIII, 3 was a long fragment of bone with a dull point at one end;
and a similar one was found in Pit X, 2. In Pit V, 2 was an antler tip,
apparently not artificially retouched.
A fragment of a llama metatarsal came from Pit VIII, 3. It is
notched with a series of cross grooves and is similar to the rasping sticks of
Mexico. It is 13 centimeters long.
The best of the bone artifacts is a well-finished spatula from Pit V,
6. It is 10.5 centimeters long, with the blade 4.5 centimeters in length
and 1.5 wide. The handle is .75 centimeter wide, narrow at the butt end,
and pierced. The blade is thin, longitudinally hollow on one side, well
polished, and sharpened on the front edge.
Two polished rib fragments from Pits VII and V, 9 may have served
as scrapers. In Pit V, 6 was a piece of bone cut in semicircular fashion
(3.5 by 3.2 centimeters). In Pit V, 5 was a section of a bisected ball
socket, 3 centimeters in diameter and 1 centimeter thick, with a hole
hollowed out of the center.
One piece of a long, hollow bone bead, 2.7 centimeters long and 1.1
centimeters in diameter, was found in Pit VIII, 2. In Pit V, 4 a tibia of a
bird was cut off close to the joint, perhaps in the manufacture of beads.
A pierced bone disc 3.8 centimeters in diameter and .1 centimeter
thick was found in Pit X, 1. A fragment of a warped piece of cut bone
still showed a hole, shaped like a cross (half broken away). It came from
Pit V, 5.
Shell. Only two pieces of shell artifacts were found, both in Pit X,
2. One is a thin pierced disc 1.5 centimeters in diameter. The other is a
thin elongated disc, pierced in the center. The edge is undulating. It
measures 1.8 by 1.2 centimeters.

STONE SCULPTURE
The excavation of Pit VII furnished the only sculptured stone
material. The array of statues and other objects has already been
discussed (cf. p. 386 and Fig. 18), but the description of the material
has been reserved for this section. The one significant conclusion gained
from the excavation of Pit VII is that all the heterogeneous styles of
stone work encountered are stratigraphically contemporaneous. Since
the styles represented in this pit are widely divergent, the stratigraphic

1934.1

429

Bennett, Tiahuanaco Excavations.

contemporaneity does not imply simultaneous manufacture. Indeed,

if the small temple (within which Pit VII was located) is considered as of
late origin it is quite possible that statues and stone work were brought
in from other parts of the ruins of Tiahuanaco and encorporated in the
building. This practice of utilizing statuary and building materialrin
constructing a new temple on the ruins of the old is not uncommon'in
archaeological America.

t:$
II ~

0~ ~0
~

2
~

'1ETERS4

Fig. 18. Pit VII showing Arrangement of Sculptured Stone. A, large monolithic
statue, the head is partly in the original pit, which was extended, on an angle, to
uncover the body; B, smaller bearded statue, with its plain base aligned with that
of the big statue; C, slightly sculptured stone; a, ceremonial, carved, grindstone
(now in Museo Nacional de Bolivia); b, rounded boulder stone head, found at 2.10
m. (now in Museo Nacional de Bolivia); c, Two-faced, crudely carved stone head
(now in Museo Nacional de Bolivia); d, long stone slab, uncarved.

Without further discussion of the theoretical significance of this"pit,

the sculptured objects will be described and compared with others from
Tiahuanaco and elsewhere.
fli
The Large Monolith. To facilitate the description of this statue a
drawing (Fig. 19) is included here which shows the principal design
areas. It is cut from a single block of red sandstone. For convenience,
the summary of the principal measurements follows:-

E-4
*

bCt
bi2

U)

o
-4

0
biJ
0

0
.,-

0
C;-

0a

bb

1934.1

431

Bennett, Tiahuanaco Excavations.

Total length of statue

Length of plain base
Length of figure
Length of head
Width at headband
Thickness at headband
Width at shoulders
Width at waistband
Width at top of base

Meters
7.30
1.80
5.50
1.90
1.05
1.05
1.27
1.15
1.15

The total effect of the statue is that of a monolithic pillar (Fig. 20).
In gross size it is the largest statue yet found at Tiahuanaco, but in style
and design it is quite typical. The features, the shape of the head, the
arms, legs, and feet are modeled in the round, but not with sufficient
realism or emphasis to destroy its columnar effect. The characteristic
Tiahuanaco angularity and straight line is preserved in the carving.
Delicate designs are incised on the headband, chest, waist, and other
parts. These designs are completely obscured in the mass effect of the
large statue. When upright it must have been very difficult to distinguish the designs around the headband, for example.
The top of the head is flat. Above the headband (Fig. 19A) are
parallel grooves, now worn but once sharp-edged, which probably represented a feather headdress. These are 60 centimeters in length. There
are five grooves along the front, and they continue on both sides. The
headband itself (Fig. 19B) is 35 centimeters wide and slightly raised.
In front it is too worn to distinguish any design, but on the left side it is
more clearly seen, and probably continued in the same style in the front.
Fig. 21 shows the headband design which consists of a running figure
with a scepter in his five-fingered hand. The scepter is double above and
single below, all ends terminating in fish heads. The feet have three toes
each, and the back foot has a condor head extended from it. The legs
are decorated with fish head designs. The features are unfortunately
obscured, but the peculiar, turned-up, ringed nose suggests that the
figure is either zoomorphic or masked. This idea is amplified by a curledtail design and a wing decorated with condor heads. The figure has a
headdress composed of three projected fish heads and an abbreviated
tail-feather (trident) design.
The face of the large monolithic statue is again typically Tiahuanaco.
A double ledge crosses the forehead below the headband, and continues
down both sides, leaving a flat face with rounded cheeks, approximately
95 centimeters long and wide. The nose is badly broken, but projected

Fig. 20. Large Monolith lying on its Back in Pit VII, now mounted on the Prado
in La Paz. The smaller bearded statue is set up against the side of the pit and the
serpent design on its side can be seen. The two other slabs, one slightly carved
(Fig. 18, C, d) are in the right foreground.

432

Bennett, Tiahuanaco Excavation4.433

1934.1

at one time. The mouth is also worn away, although it was probably
rectangular in shape with rounded corners. Whether teeth in two rows
were represented, as in other statues, is impossible to state. The eyes are
formed of two concentric ovals below which are two tears, and possibly
a third, although the lower one may be the eye of a condor head. On the
outer side of each eye is a wing ending in a condor head on the lower
curved cheek. On the upper part of this wing is incised another condor
head with an extended neck. On the side of the head, the two ledges
mentioned above form a series of three steps, and the lower ends of the
ledges end in condor heads. The ear starts at the lower edge of the head-

0~~~~

0
Fig. 21

0
Fig. 22

Fig. 21. Design on the Side of the Headband of the Large Monolithic Statue,
depicting a Running Figure, Masked (?), with Curled Tail.
Fig. 22. Design on Side of Head of Large Monolithic Statue.

band. It is rectangular in shape, with a square notch cut out of the edge
on the front side, and a triangular notch at the lower back corner. Behind
the ear is an incised face (Fig. 22) with a headband, T-shaped nose, two
tear eyes, and a small mouth.
The section (Fig. 19D) across the chest of the monolith is elaborately
designed. A collar composed of a chain of stylized condor heads is just
below the chin. On the right shoulder (Fig. 19D, 1) is a running human
figure which can be described simply by referring to the top and bottom
row of almost identical running figures on the Gateway of the Sun. In
the middle of the chest (Fig. 19, D3) is the design shown in Fig. 23. It is
an elaborate design. The top part is a front view face with a concave-

434

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

sided T-shaped nose, a small oval mouth, and two rounded eyes each
with a bar ending in an oval projected below (a variant of the tear
design). Above the head, perhaps as a band, are three double rectangles,
each with a double oval above it. On each side of the face is an irregular
band ending in condor heads. The central part of the design is composed
of three connected ovals, one below the other. These are outlined with
broken bands. On each side of this oval series is a design like a stylized
three-finger hand and thumb, below
which is a horizontal bar, and, still
& &)(ffi)(g
Th
00
farther below, two series of three conJ
,1
t[B[Et
W
Enected
R
41
circles. The lowest part of the
['ri
design is another face, completely outlined with a broad band, with two
yoval-rectangle eyes, a concave-sided
YJ
T-shaped nose, and a rectangular
[ +j3 t
Jj mouth. Figs. 19 D, 4 and D, 5 represent
tthe running condor and human figures
0[ O
of the Gateway of the Sun, repeated as
1
o
~~~~~~before.
00
The next design section (Fig. 19E)
2 p) [0cD T~
(o5
{
(
t~JJ B C' > g includes the two hands of the monolith.
<
)
L11 IC (2 (2)The left hand (Fig. 24) holds a kero@)Q)4
shaped cup. The hand is outspread
showing five fingers in correct proportion, each with the nails depicted. The
(
(U)l
cup has a narrow band around the
upper part (above the hand), decorated
with a four segment design which ends
in fish (?) heads. The base of the cup

1i

m~ID
vLJ}

l1

Fig. 23. Elaborate Design from is decorated with a step figure, each end
Center of Chest of Large Monolith- decorated with a condor head. Above
the cup are six fish with angular bodies.
ic Statue.
The right hand is in a closed (grasping) position, so distorted that the
thumb is on top pointing towards the center of the body, and the four
fingers, fully exposed, point outward (Cf. Fig. 25). The object held in
this closed right hand is hard to identify. Below the hand are five projections, the central one terminating in an oval, and the outer two on
each side ending in fish heads. Above the hand is a horseshoe-shaped
band with a fish head at each end. Above the band project two small
and one large fish head, and within the band is a "female sex" sign surmounted by two or four projecting fish heads.

Fig. 24. Left Hand of Large Monolithic Statue, showing the Five Proportioned
Finigers. This hand holds a kero-shaped cup with a band above and a decorated base.

Fig. 25. Right Hand of Large Monolithic Statue, illustrating the Distorted
Position of the Thumb and Four Fingers. The object grasped in the hand is hard to
distinguish as the incision is badly faded on the statue at this point.
435

436

Anthropologicat Papers A merican Museum of Natural History. [Vol. XXXIV,

Above both arms, on the body (Fig. 19E, 1 and E, 4) are running
condor figures of the Gateway of the Sun style. Below each forearm is
the design shown in Fig. 26. It is composed of a double "female sex"
design between two horizontal bars, above which are four condor heads,
and projecting from each end of which is another condor head.
On the arm on the side of the statue appears the elaborate figure
shown in Fig. 27. It is difficult to describe this. The upper part represents an animal head with a headdress and a disc suspended from the
neck in typical puma style. Posnanskyl also calls this a variant of the

Fig. 26. Design below the Forearm

of the Large Monolithic Statue.

Xgo
Fig. 27. Elaborate Design on
the Side Arm of the Monolithic Statue.

puma, but, in all, the teeth and nose suggest the llama. The headdress
is composed of a bar which curves up in front and terminates in a fish
head. In the center above this bar is a tail feather design and on each
side a three-leafed plant. According to Mrs. Bandelier this is the Inga
Pulgerima, "flor del Inca," one variety of which is known in the highlands, but the tufted plant represented in the drawing is found only in
the low country. On the back of the headdress is a clearer representation of the same plant. The central stalk has two sets of side branches,
each ending in the flower. The lower part of the design is confused.
The plant appears again on a single stalk. Amidst an array of bars and
lines appear three front view faces of animals, each of a different type.

1Posnansky, 1914, PI. XXXXVIII, 3, 6.

1934.1

Bennett, Tiahuanaco Excavations.

437

Around the waist (Fig. 19 F) of the monolith, on the front side, are
three designs. Two (Fig. 19 F, 1; F, 3) are the running human figure
from the Gateway of the Sun (Fig. 28). The center design (Fig. 19
F, 2) is shown in Fig. 29. The upper part is a simple face with a segmented band at the top and ending at the sides in condor heads. A
middle rectangular section has a condor head projected from each side.

Fig. 28. Waist Design of the Monolithic Statue showing the Running Human

Figure, almost Identical with those on the Gateway of the Sun.

The lowest part is a rectangle outlined with a broken band, with three
bars inside, roughly suggesting two eyes and a nose.
The waistband is elaborately decorated (Fig. 19 G). On each side
(Figs. 19 G, 1 and G, 3) is a design which Posnanskyl has temporarily
called "female sex" designation. This is surrounded by branches ending in condor heads, two from the base and four from the top. From the
center part of the top, between the two sets of condor heads, is another
projecting bar. The central figure of the waistband design (Fig. 19 G, 2)
is possibly a sun face (although this is obscure) surrounded by ten
lPosnansky, 1914, Fig. 63, 7, 8.

438

Anthropological Papers American Museum of Natural History.. [Vol. XXXIV,

branches ending in fish heads (six from the top and two from each side).
On each side of this design are four fish heads connected with each other
with a continuous bar. The waistband is 45 centimeters wide.
The legs (Fig. 19 H) are 1.40 meters long and each one is 50 centimeters wide at the top. A wide groove separates one leg from the other.
The feet are squared, but five toes are indicated. From the knees down
the decoration is worn off, but from the waistband to the knees it is
still clear. On each leg, in front, are horizontal rows of four large, high
relief circles with a smaller circle incised within each
o a\1 7ti one. This decoration also continues on the side of
01
O0l the legs. It quite possibly represents embroidered
medallions on a gauze base cloth.
The sides and back of the large statue are aloD (
>N

(:most solidly designed.

The side of the headband

has the human figure with curled tail already de[=L9 cj7g scribed. The side shoulder has the Gateway of the
Sun running human figure and the running condor
figure is on the side upper arm. Two running
U
G
U human figures are below the arm on the side. The
____
side waistband has the same figure as the front with
f
some
f
'c
slight variation.
The back side of the headband has two running
Fig. 29. Design on condor figures back to back, thus differing from the
the Center of the curled tail figure on the side. Below the headband
Waist of the Monolithic Statue between nine braids are designed down the back of the neck,
the Two Running separated with grooves, designed with a series of inverted V grooves, and ending on the shoulders in
Human Figures.
condor heads. On the back of each shoulder is a
sun face, as on the frieze of the Gateway of the Sun. Each face has two
winged eyes with two tears. The faces are surrounded by a crown of
projecting heads and designs. Thus across the top the projections run:
condor head, disc, disc, tri-feather, disc, disc, condor head. Down each
side the series is the same, except that a condor head is used in place of
the tri-feather. Below each of these faces is a stepped base with projecting discs and condor heads.
In the middle of the back, above the waistband, is the design of a
front view human figure, like the Viracocha of the Gateway of the Sun.
This figure has the same front view face with winged eyes. The head is
surrounded by a crown composed of seventeen projected faces. The
arms project on each side, with elbows bent. Each hand has three

1934.]

439

Bennett, Tiahuanaco Excavations.

fingers and thumb, which hold some object (not a scepter) which has two
projecting branches ending in puma heads. The body of the figure has a
waistband and some design. The spread out feet have projections on
each side which end in a flower design. The whole figure rests on a
stepped base as on the Gateway of the Sun. On each side of this figure
is a design in three parts. At the top, just under the shoulder sun face, is
a "female sex" design, with four projecting condor heads, two from the
base and two from the top. Below this is a simple face with three plain
projections from each side and three projections from the top which end

ll~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1i

s. t e bo i

"

e sx" dsig, te fr fs

Fig. 30. Waistband Design on the Monolithic Statue. The

may be seen below the waistband.

pol'ka-dot design

in tri-feathers. At the bottom iS a "male sex" design, the four fish heads
projected from above.
The waistband in the back has the same design as in the front. The
polka-dot pants also continue, but it is quite clear that they never extended to the ankles, but stop about the middle of the leg.
The base is plain and was inserted in the ground to maintain the
statue in its upright position. The statue has been moved to the east
end of the Prado in La Paz.
The Small Monolithic Statue. Immediately to the east of the large
monolithic statue, adjacent and parallel to it, a smaller monolith was

440

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

found. The base of this smaller statue was so closely aligned with that
of the large one that one assumed the two had stood side by side and
that they had fallen, or been pushed over, at approximately the"same

Fig. 31. The Smaller Bearded Statue beside the Large Monolith in Pit VII.

time. This assumption of contemporaneity emphasizes the contrast in

styles. The following summary of measurements shows the contrast
in size:Total length of monolith
Length from head to waist
Length of head
Width of head
Thickness of head
Width at waist
Thickness at waist

Meters
2.55
1.60
.87
.45
.32
.45
.32

1934.1

Bennett, Tiahuanaco Excavations.

441

This statue is also made of red sandstone. The columnar effect obscures the modeling, but the designs are cut in much higher relief (Figs.
31-32) comparatively speaking, than on the large statue. The technique
of workmanship is noticeably more crude. Two lightning rays meet in
triangular points on the forehead of the statue and continue down the
sides of the head, joining the bar of the T-shaped nose, and running
into the beard which surrounds the mouth. This beard, in high relief,

q/) ApS )X
Fig. 32. Front, Side, and Back View of Smaller Bearded Statue of Pit VII.
The front view shows the curled up beard and the lightning rays on the forehead,
both of which are connected by a raised band on the side. The shoulders do not
project as prominently as the front view drawing suggests. The puma and serpent
designs are typical. The back view is not certainly correct, but the erosion was bad,
and now (1934) nothing can be seen.

curls up on each side of the mouth and forms a point on the chin. The
mouth itself is a small oblong, as are the two eyes. In place of ears on
the side of the head are two pumas in relief. Although they are carved
in profile, four feet are depicted. They have ring noses, small round eyes,
rounded ears, and tails that curve downward (Fig. 32).
The'arms of the statue do not meet on the chest in the typical style,
but are arranged with the left hand on the stomach and the right hand
on the chest. Both hands have five fingers spread out (Fig. 32).

442

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Below the arms is a narrow, undecorated waistband. Under this

are two pumas in low relief, depicted with four-legged, side view bodies,
and front view heads. The hind legs and the base of the tail which curves
over the animals' backs are carved on the side, rather than on the front
of the statue (Fig. 32), although the figure is unbroken. The faces have
two loop ears, and regular round eyes, oval mouth, plain nose features.

Fig. 33. Slightly Carved Stone of Pit VII, with Nose, Eyes, Mouth, and Groove
separating Head. This is now (1934) completely weathered away.

On the sides of the statue is a serpent-like figure, with the head at

its shoulders and the body down the sides in five angular zigzags to the
curled-up tail at the base. The head has an oblong mouth, a blunt
nose, two round eyes, and two backward pointing ears.
On the back of the statue is the suggestion of some figure carving,
although it is badly worn away.
No delicate incised ornamentation is found on this statue. The
decoration is all in relief.> The positions, designs, type figures, and
technique all differ markedly from that of the large monolithic statue.

1934.1

443

Bennett, Tiahuanaco Excavations.

A Slightly Carv;ed Stone. Still further to the east of the large monolith were two stones of about the same size, one plain, and the other with

Fig. 34. Edge of Ceremonial Grindstone found in Pit VII. This design is continuous around the edge.

a face faintly depicted (Fig. 33). This latter stone measured as follows:Total length of stone
Length of head
Width of stone
Thickness of stone

Meters
1.50
.55
.40
.33

The head was delineated by a shallow groove encircling the stone.

A T-nose, rectangular mouth, and square eyes, all slightly carved, com-

444

Arnthropological Papers AmericanMuseum ofNatural History. [Vol. XXXIV,

pleted the front facial features, and the sides contained a small notched
rectangular ear.
Stone Heads. Two stone heads were found, one at 2.30 and the
other at 1.30 meters depth. The first one is a rounded boulder
with round eyes, rounded nose, and oval mouth all depicted in relief.
The other head is an irregular rough stone, with two sets of eyes and
mouths hollowed out, as if representing a double face. The work is very
crude.
Circular Grindstone (?). At 2.10 meters depth to the west of the
feet of the large statue was a circular, fine-grained, blue stone. It
measured 63 centimeters in diameter and 16 centimeters in thickness.
One side was flat, but the other was indented about 5 centimeters, with a
drain cut through one part of the resulting rim. Around the edge was a
continuous wavy band, with the edges sharply delineated. A circle was
cut in each loop of this wavy band. The only use suggested for this
odd piece was as a grinding stone. To say the least it is not typical of
other Tiahuanaco specimens (Fig. 34).

CULTURAL STRATIGRAPHY
Throughout the preceding description of the Tiahuanaco pottery
collection distinctions in shapes, colors, and designs have been made.
In the section on excavations actual distinctions between levels in the
pits were described. The next step is the correlation of these two sets
of evidence. This has been arranged according to the excavated pits in
the graphic form of a diagram (Fig. 35). A few words about the set up
of each pit may clarify the analysis of the four cultural phases which
follows. On the basis of the collection, four major divisions are distinguishable: A, Early Tiahuanaco, or perhaps pre-Tiahuanaco; B,
Classic Tiahuanaco; C, Decadent Tiahuanaco; D, Post Tiahuanaco and
Inca.
In my first classifications I tried to subdivide these four phases of
culture. Both deductively and inductively there is evidence to support
such further subdivision. However, at present it seems more practical
to establish definitely the four phases which are demonstrably present,
and leave the more subtle division for further research.
Referring to Fig. 35, I shall comment briefly on each pit. In Pit I
only two cultural phases are represented. Levels 1 and 2 were associated
with recent burials and modern pottery. The Tiahuanaco sherds present
are almost exclusively of the Decadent type. Pit I, 3, 3a and the intrusive burial, Pit, I, 5 represent a single cultural unit of Decadent type.
This section is stratigraphically distinct from the top levels, Pit I, 1
and Pit I, 2. Although one bowl in Pit I, 5 is of late Classic type, the
others associated with it are undeniably Decadent. The bowl mentioned
had the spout missing, broken off with an old break, which might indicate that it had been preserved from the previous cultural epoch.
Pit II, being definitely a part of the old terre-plein of the Sun
Temple, is probably contemporaneous with it, and therefore belongs to
the Classic phase. The artifact analysis confirms this. No distinctions
between layers were evident and so the whole is treated as a Classic
unit.

Pit III is not seriously considered in the analysis because of the

shallow depth (1 meter). The high percentage of plain over painted
ware, and the presence of three horizontal handle fragments, suggests
that it might be an Early Tiahuanaco site-. However, none of the special
painted wares were found.
Pit IV is an annoying puzzle. Inca pottery fragments in Levels 1
and 2 indicate either a late position, or disturbance. Levels 2 and 3
445

0
U)

._

0
n
O

E-r
bo
"e
*i =

*t
._

C4 o Y
e O

78*:

>

p;

C40

0"

0
H

*
ni*

sv

1:

0=
V0

0*}
;e

lc
0

0D=

.d

ni7C;

1934.]

Bennett, Tiahuanaco Excavations.

447

form a unit, separated stratigraphically from Level 1 by an old surface

level, and the collection indicates the Decadent chronological phase.
The greatest problem is presented by Levels 4, 5, and 6, successive depths
in a pit which starts at the old surface level between Levels 1 and 2,
and continues to a depth of 3.5 meters. While the stratigraphy would
indicate that Levels 4, 5, and 6 were more recent than Levels 2 and 3,
the actual collection analysis indicates a greater percentage of Classic
ware. The complete bowls from this section are, on the contrary, all
Decadent. This apparently mixed stratification can only' be accounwted
for by speculation: consequently, I will leave the decision for future
evidence, being satisfied at the moment with description alone. Because of this discrepancy I do not include Levels 4, 5, and 6 in the final
analysis.
Pit V presents the best stratigraphic set-up. Unfortunately, the
selection of arbitrary half-meter levels does not nmake the level divisions
coincide with the cultural divisions, but in spite of this defect the stratification is evident. Level 1 is dominantly Decadent in materials. The
surface stratum contains some "Chullpa" or Post-Tiahuanaco material
also. Level 2 is about half Decadent and half Classic, due to the failure
of my arbitrary level to coincide exactly, and also the actual existing
historical condition which did not break sharply from Classic to Decadent phases. However, a clay floor, an ash bed, and a row of bones all
indicate a stratigraphic division between Level 2 and Level 3. Thus, in
this pit Levels 1 and 2 which are predominantly Decadent in style are
distinct from Levels 3, 4, and 5 which are predominantly Classic in style.
Another alignment of water-worn stones, and semi-barren clay divides
Levels 3, 4, and 5 from Levels 6-9 below. Furthermore, Levels 6-9
represent the Early Tiahuanaco style. Thus in Pit V, Early, Classic,
and Decadent Tiahuanaco are stratigraphically distinct, with the
possible addition of Post Tiahuanaco as upper surface material.
Pit VI is not considered at all. Pit VII, not completed because of
the monolith uncovered, was shown in the excavation analysis to be composed of the fill of a sunken temple. A mixture of Classic, Decadent, and
Inca material is the logical result of a washed-in fill. By deduction, Pit
VII would be a rather later site, especially if my analysis of this temple
as a late rather than an early structure is correct (cf. p. 386).
Pit VIII is another well-stratified site, confirming the analysis of
Pit V. Level 1 is predominantly Decadent in style. Level 2 is not so
distinctly Decadent, but is a mixture of Classic and Decadent. Between
Level 2 Und 3 a stone pathway, a bed of water-worn stones, and an ash bed

448

Anthropological PapersAmericanMuseumofNatural History. [Vol. XXXIV,

indicate an old surface level division. Level 3 is predominantly Classic

in style. Below this level, Levels 4, 6, and 6 are segregated only by
greater depth. However, Levels 5 and 6 contain the same Early Tiahuanaco material already mentioned for Pit V, 6-9, while Level 4 contains
both Classic and Early material.
Pit IX is too disturbed to serve in any final analysis. As part of an
old burial ground it presents a conglomeration of Classic ware mixed
with Decadent in unstratified confusion. The surface material, and Level
1 in general, is perhaps most recent, as the presence of some Inca ware
indicates. On the basis of the distribution of thick orange ware with
crude decoration of black, or black and white lines, Levels 1 and 2 may
be distinguished from Levels 3, 4, and 5. The thick polished ware
dominates in the lower three levels. The stratification is, however, far
from satisfactory.
Pit X, while disturbed by excavation in recent years, and possibly
by intrusive burials at all times, is predominantly Classic in type. It is
located in the center of the old burial ground, from which the finest
Classic pottery is extracted. Out of nineteen restorable vessels found
in this pit, eleven are of general Classic type, six Decadent, and two
Chullpa. In the excavation no intact burials were found, and many of
the bowls were only partially complete, all of which indicates disturbance.
From this array, certain levels have been selected as most representative of the cultural phases. These are as follows:Early Tiahuanaco: Pit V, 6-9; VIII, 4-6
Classic Tiahuanaco: Pit II, 1-3; V, 3-5; VIII, 3; X, 1-2
Decadent Tiahuanaco: Pit 1,3-3a, 5; IV, 2-3; V, 1; VIII, 1
Post or Inca: Pit I, 1-2; IV, 1

The tables included in this section are arranged according to this grouping. In a grouping such as Pit V, 3, 4, 5, the percentage is calculated for
the group as a unit.
With this prelimninary discussion as a basis the four phases of
Tiahuanaco will now be considered.
EARLY TIAHUANACO
of
this
Early phase is established by its stratigraphic
The position
In Pit V it runs from 2.5 to 4.5 meters, and
and
VIII.
in
Pits
V
depth
in Pit VIII, from about 1.25 to 2.80 meters. In both pits two other cultural phases are superimposed above it, and in Pit V a definite line of
demarcation isolates it. No buildings can be associated with this Early
Tiahuanaco phase, although this is due mainly to lack of sufficient evidence. Ash beds and fire pits are associated.

10 "i4 0M
M
00
1 00 00 O Co N
Co 0 t 0 tOC

-> -.

es

CO

V-_

H4;00
c0

O U:O V- L
t- Co 10
_I O
0
0. .I
co. d4 L:O
0 Co d4
co
CD C

l-i

t.

Co

es

--

oq

10Co0

ko

ko
CYo

OU~

C14O

ON

kOo

00

c1
10

LO

rA1

c0
i

~~

P.4

oCo

0
Ca-

0r- to~

c~o

00

m10 o

Co

to0

10

0
10

OC)

'-4

0r.10

Co

oLO1

r4

CoD

co t o co
Co
O
0C4OCo)I
Q
CC

00
~~~
~~~~Co~~~~~-4

0
O

p.4

Om to
m co c

0~ ~ ~-

CoC0 o
104
tCtI

Co

m CYCoCo Cq
OC

O
O

--4

p.0

C3
p4D

(1

450

Anthropological PapersAmericanMuseumof Natural History. [Vol. XXXIV,

Although most of the pits show a predominance of plain sherds, the

levels of this Early phase are over 90 per cent plain (Table 5). More
distinctive is the corresponding small percentage of painted sherds, less
than 5 per cent. Plain sherds are mostly of brown or black, unslipped
clay, in contrast to the orange clay of the later phases.
Special wares, which are found almost exclusively in the lowest
levels of Pits V and VIII, have already been described (cf. p. 399). A
specially scraped and polished reddish-orange and reddish-brown ware,
a red-on-buff and a polychrome-on-buff are the characteristic painted
wares of this section. The paint has a glossy tone distinct from the
Classic colors. The designs are mostly linear or angular. Zoomorphic
designs represent unidentified animals in colors on a black background,
thus omitting the black outline which is characteristic of most
Tiahuanaco art.
Several vessel shapes occur exclusively in this Early phase. Thus
the horizontal, rim handle dishes and bowls (Shape a) described on p. 409,
are only found in the lower levels of Pits V and VIII, and in Pit III.
Likewise the shallow open bowl and the small, flaring rim olla (Shapes
b and c) were only found in this cultural stratum. Hollow base libation
bbwls (Shape B) were divided into six subtypes in the description on p.
408 et seq. Subtypes a and b are associated with this earliest phase (Fig.
13). Subtype a, a flat base bowl with undulating rim and small upturned
puuma head on the rim, is not as clearly associated as is Subtype b, a
plain brown open libation bowl with slightly (.6 centimeter) indented
base. These latter are found in Pits III; V, 8; VIII, 4, 5, 6. They are
undecorated. One sherd with deeply incised decoration may indicate
the antiquity of incised incense bowls at Tiahuanaco, although the
suggestion was not further substantiated. The modeled puma heads of
this Early phase are of the solid variety, with small ears, and slightly
modeled and incised features. They are attached almost directly to the
rims of the vessels, without long necks, and with the heads facing slightly
upward. None of the collared, hollow variety of modeled heads is found
in these levels.
Further elaboration of the Early period styles can be made from an
examination of the Tiahuanaco collections. The collection of Frederick
Buch and the Museo Nacional in La Paz, of Georges Courty in the Trocadero Musee d'Ethnographie of Paris, the Posnansky collection in the
Museum fur VoLkerkunde in Munich and the Berlin Museum fur Volkerkunde, all contain a small number of vessels which can be identified with
the Early Tiahuanaco period. The identification is based on the com-

1934.1

Bennett, Tiahuanaco Excavations.

451

parison of sherds found in the Tiahuanaco excavations with the specimens in these collections. The chief criteria are the colors, polychrome
black, white, red, orange and brown-on-buff colored clay, the shiny
quality of the paint, the linear and angular designs, the treatment of
zoomorphic design, and the shapes.
The collections present one new group of vessel shapes, all variations
on a long-necked decanter (Fig. 13c-d). The typical decoration of these
decanters is an angular design around the body, and perhaps a stripe or
two of color around the neck. There is some minor variation in both
shape and design. The collections also suggest that the wavy rim cup
(Fig. 13a-b) with a flat base and a deeply incised design, with the
lines filled with color, is closely akin to the wavy rim cup with puma
head (Shape B, Subtype a), and thus belong to the Early period.
Finally, a flaring rim bowl type is identified with the Early period by the
angular design around the outside, and the zoomorphic designs around
the inner rim (of which two fragments were found in Pit VIII), as well as
by the colors of the paint and the treatment. These zoomorphic designs
are quite distinct from the Classic Tiahuanaco (Fig. 14). The subjects
cannot be positively identified. The three-ring crown suggests the male
condor treatment, but other features do not confirm this. Even the
shape of the bowl is unique. It is a bowl with a wide flat base and evenly
convex sides that meet a wide flaring rim. The rim extends inward as
well as outward and leaves a very constricted mouth opening. This
shape is like the typical American spittoon.
The presence in the collections of complete specimens from this
Early period suggests that there is some site or cemetery more accessible
than my 1.5 to 4.5 meters depth. Unfortunately, none of the collections
examined have specific information about the precise location of these
pieces in the ruins. Further examination of the Tiahuanaco ruins should
reveal some locality in which this type of material can be more completely

studied.
Small clay buttons (cf. p. 425) are a distinctive trait of the Early
phase of Pit VIII. Nine of these were found in this one pit. The Early
levels had a considerable quantity of small stone artifacts. Hammerstones, polishers, chalk, slate, obsidian, quartz, flint flakes, all were of
ordinary types, but a T-shaped ax was more distinctive. Bone spatulas
and needles, as well as pointed and notched bones, were found. Copper
fragments and a piece of a pin, topo, indicated a knowledge of metals.
Thus far the Early phase has been characterized by positive traits.
Considering the typical Tiahuanaco elements which are absent in the

CO0CO

Cs

O~
0
0

C)
UO

0
00

COC

C)

CO

"

o-

r--

106
M

0
C.)

b~~~C11 18

0
z

C t-

|
N
1'd

r-

E0
Q

mC1

oc

ho
-4'

PA

Ep4
0

05

00

-'-

00c
t -

e~~~~~~~~~~~~~c
0
> ~ ~ ~~on8

00~

5S
e
Cc1tso

z4

)a

.4g00

4)

d4-

r.

C) p m

Ez1

k;

1934.1

Bennett, Tiahuanaco Excavations.

453

Early levels, the phase has even more definite isolation. In Table 6 the
high percentage of Classic style pottery in Pit VIII, 4-6 all comes from
Pit VIII, 4 which, as described, is half Classic, half Early.
The only evidence which I have found from other localities which
might check the identification of the Early Tiahuanaco ware is the
material from Bandelier's collections at Kea Kollu Chico on the Island of
Titicaca, where small bowls with deeply incised linear designs, small
puma head bowls, a fragment with angular design in glaze colors, and
small clay buttons are found in non-Inca association. Although
Bandelier called this site "Chullpa" the plain material is not precisely
typical of other Chullpa sites around Bolivia.
* CLASSIc TIAHUANACO
Throughout the description of the Tiahuanaco collections in this
report the distinction has been made between the Classic and Decadent
styles in colors, design, treatment and subject-matter, and shapes of
pottery. Thus in this section only a summary of these differences will
be given. Fortunately, the distinction is substantiated by stratigraphic
proof.
Classic levels are stratigraphically lower than Decadent, higher
than Early. Thus the Classic stratum of Pit II extends from .20 to 1.5
meters deep; Pit V, from .75 to 2.5 meters deep; Pit VIII from .75 to
1.25 plus meters deep. Classic is separated from Early by a definite
striated division of Pit V (between Levels 5 and 6). It is separated from
Decadent by definite divisions in Pit V (between Levels 2 and 3), and
in Pit VIII (between Levels 2 and 3). The Classic of Pit II is probably
associated with the ex-terre-plein of Calasasaya temple. At least Levels
1 and 2 of this pit were filled with squared and dressed stones. In Pit
V, two dressed stones were associated with Level 3. In Pit VIII the
well-finished stone-faced canal might be associated with the Classic
layer, number 3, althouigh it is impossible to establish the relationship
definitely.
Classic sherds have a higher percentage of painted wares than the
Early phase, but this does not distinguish them from the Decadent (cf.
Table 5). The standard color combinations are black and white-onred; black, white, yellow-on-red; black, white, yellow and gray-on-red
There is a higher percentage of four color wares than in any other group.
The colors are rich in tone and skilfully applied. Usually the whole
vessel is polished. The basic paint, or slip, is red, rather than the orange
of the Decadent. The designs are limited in variety, being principally a

10101010

aS
C)

10 1 00
C'q 10 CO C
0000
.

0 CO

r--

csq

CO

05
I

"

P-4

t'~~10t
-

A
00

t'.
Cy

es

N IR402

00 C.O

Ild

__

cli

LD

Nco O xo
-I CCDO~~~~~~~~10~
ao

02

Ez
Q

r-l

f.0Io

-~~~~~~~~

010o10 0
r,.
c'q

*O
a)

C9

0'
-4

,_O
00 C00

C)

>

C..)

e
co

_ C,i_

O0 X

OC5

OC

0
0

z4

V
U CO
02 O _
co

E--

C0

CO

CO-

1t.
C0
-q
C

t-CO CO2t
U

O1
4C) c
-4

C)
C,

CO

> -- ~CO-CO

10 _
oo '-4S
0 C

C)

-~~

OoC
0~ ~ ~ ~ ~ ~ r-

21~~'' 002o

04

a)

-0

05
02a)

Ca

I
oo

05
Q a)

S>
0)
00

o o

EH Ev

1934.1

Bennett, Tiahuanaco Excavations.

455

complete puma figure; an elaborate condor; and a human figure.

Geometric designs are predominantly step pattern. The designs of
figures are usually profile views (two legged puma; profile human head
with front view body). All figures are outlined in black, while the other
colors are used as design fill. Usually, a single unit design is repeated
around the vessel, with regular color changes.
The percentage of sherds of the Classic treatment far exceeds those
with Decadent treatment in the Classic levels, as Table 6 shows. The
ratio is 75 per cent Classic to 25 per cent Decadent. Also the Classic
levels contain a higher percentage of zoomorphic designs than do any of
the others (cf. Table 7).
The shapes have already been classified into Classic and Decadent
types. This division is not absolutely borne out by the stratigraphic
evidence. Unfortunately, the pits which showed the best stratification
produced no complete vessels, so that determination of shapes is based
on an often dubious sherd reconstruction, and on the existing collections. However some facts seem reasonably established. Thus the
kero-shape cup (A) appears first in Classic levels. It continues throughout the Decadent. I have no doubt that several modifications in style
of keros accompany this change, but I have insufficient evidence to
determine all of them. Hollow base libation bowls (B) are common,
especially the painted, well finished variety (Subtype d) and the undulating rim and puma head types (Subtypes e andf: cf. Fig. 15f). The puma
heads are hollow, with well modeled features, long necks, and square
collars. The wide rimmed open bowl (Shape C; cf. Fig. 16b) is almost
invariably associated with the typical Classic designs. The Shape D
vase is a constant Classic type. Shape E, the small concave-sided open
bowl, has its origin in the Classic, but the development and extensive
use is in the Decadent. The minor shapes are too sparsely represented
to permit of more speculation than has already been employed, although
two with Classic designs are illustrated in Fig. 16.
Stone work is still extensive in the Classic period. There are
hammers, plain and one grooved, polishers, bowls, and many fragments
of flint, slate, chalk, mica, quartz, and a green stone. One piece of copper
was found. Bone needles and discs, as well as many fragments of cut
bone, are found. Two shell discs, pierced, were found in Pit X.
1
Considering the culture of Tiahuanaco as a whole, including building and stone carving, it seems probable.that the Classic phase will
eventually be subdivided. The change from sandstone to lava building
materials without a shift out of the Classic design style is evidence of
this. However, more extensive excavation must first be made.

456

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

In Bolivia the Classic style is, so far as is now known, almost

limited to the actual ruin of Tiahuanaco. Even the relatively nearby
Island of Titicaca presents no material of true Classic style. The
Decadent style, on the other hand, has a wide geographical distribution.
DECADENT TIAHUANACO
Most surface and first and second level material is of the Decadent
Tiahuanaco type. The division between Classic and Decadent levels
has already been mentioned. Decadent is likewise distinguished from the
Post Tiahuanaco phase in Pit I (between Levels 2 and 3) and in Pit IV
(between Levels 1 and 2). However, I believe that the Decadent phase
represents the last actively functioning Tiahuanaco culture. In Pit
VIII, 2, a stone path is associated with the Decadent level. Significantly,
this path is composed of stones obviously extracted from some building
or ruin, as they are bored for copper clamps.
The Decadent phase contains a relatively high percentage of
painted pottery fragments. However, these are mostly of one and twocolor ware, and practically lacking in four-color ware (cf. Table 5).
The colors, as previously mentioned, are dull, and the pottery itself
lacks the fine polish of the Classic ware. Orange, rather than red, is the
,commonest basic color. Brown is also used as a base.
There is a high percentage of plain thick orange ware. An orange
ware with a poorly painted black or black and white design is common.
Extra thick ware (cf. p. 394) is found. Black polished ware is typical of all
phases of Tiahuanaco, but the addition of incised design is a Decadent trait.
Designs are composed of the same basic elements as in the Classic,
with a few additions, but there is a greater emphasis on the geometric
patterns and a lessening of the percentage of zoomorphic figures. Step
pattern, perpendicular wavy lines, double S, horizontal wavy, scroll,
-and angular patterns are used as complete design elements, whereas
their appearance in Classic decoration is as secondary fill. This tendency
'to elaborate parts of design figures is demonstrated in a breakdown of
Classic design. Thus the full puma and king condor and human figure
'drawings are reduced to puma heads, eye designs, etc. The treatment
'on the whole is much cruder in technique (cf. Fig. 17 for typical shapes
and designs). The outline method is still used. Front views of heads
are used. White, yellow, and orange colors are used as independent
'design elements, and not just as fills.
-Generally'the proportion of Decadent to Classic ware in the upper
is approximately 28 per cent Classic to 70 per cent Decadent
levels
pit(cf. Table 7).

l-

on
't 10

Cd

01

on
.

zv
pi

j-

0
p

a)
C

_/i

m
0

o4
w
DO

pn

c0

01

i-

C0

cOJ

VD

E-q

01

1n

>IT-

.=4
ce
10
m
C)
&O
m
w

~~~i
c~~~~~c~
Ic
C

458

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

A reiteration of the analysis of shapes emphasizes the following

major differences. Kero (Shape A) still continues, but with ridges more
common and a special constricted base variety (cf. Fig. 17a). The
hollow base bowl is found (Shape B), mostly in the plain orange, plain
rim, two handle variety with deep base indentation (Subtype c).
Shape E is most prominent. Many new shapes are introduced, although
these are better confirmed from other collections than from my own.
Stone work in artifacts is much rarer in the Decadent levels than in
the Classic. However, hammers and polishers are found. Stone tops
are most prominent in the Decadent, although they occur sporadically
in the two preceding phases. Worked bone is conspicuously absent. A
piece of a copper topo represents the metallurgy (cf. Table 9).
Logically, a distinction like this one between Classic and Decadent
styles, no matter how clearly separated the two types are in their extremes, is bound to include transitional types. This is certainly true in
the actual classification. Many pieces might be either Classic or Decadent, or neither one nor the other. These subtleties of change will eventually be worked out. Distributions confirm both the independence of
the Decadent phase and the necessity of determining further subdivisions
in it. Cochabamba, for example, is a site with purely Decadent style
pottery. The Island of Titicaca, Pelechuco, Charasani, and many
other sites in the Bolivian Highlands, present Decadent style decoration.
The contrast in the distribution of the Classic and the Decadent in the
Bolivian highlands is inexplicable.
POST TIAHuANAco AND INCA
Inca sherds are found on the surface around the Tiahuanaco ruins.
The sherds are of typical Cuzco pattern, mostly from aryballoid jars,
animal-handled shallow plates, and straight-sided open bowls. The
decoration, colors, and texture present a decided contrast to all Tiahuanaco ware.
The levels at the Tiahuanaco excavations which are classed as Post
Tiahuanaco, still contain a high percentage of Decadent Tiahuanaco
types. This is due partly to the fact that true segregation of the Post
Tiahuanaco from the Decadent is not possible in the frequently disturbed
upper pit level. It is also due to the lack of sharp distinction between
the two periods. Inca ware is distinct, but other Post Tiahuanaco
material is still influenced by Tiahuanaco design, colors, and shapes.
The so-called "Chullpa" ware is a name loosely applied to this Post
Tiahuanaco material. Unfortunately, "Chullpa" has been applied to

TABLE 9
DISTRIBuTIoN OF ARTIFACTS IN TIAHUANACO PERODS
(Also Miscellaneous Pottery Types)

Artifacts

V
6-9

Clay Buttons
Clay Whorls

VIII
4-6

V
3-5

II
1-3

VIII
3

1
2
1

3
1

V
1

VIII
1

IV

1-2

1
3

3
1
1

1
1
x

1
xx
x
x

Green-stone'

Copper Ore
Copper Artifact

IV
2-3

1
Beads1
2
Chalk
2
Flint
2
Slate

Quartz
Lapis-Lazuli

I
3-3a

Shell Discs

Mica
Obsidian

X
1-2

Post and Inca

x
Worked Bone
Bone Spoon
Bone Needles
Bone Points1
Bone Discs

Grindstone
Stone Tops
Hammers (cylindrical)
Hammers (grooved)
T-Shaped Ax
Polishers
Bowls

Decadent

Classic

Early

x
x

Pottery Types8
Extra Thick
Black (poULshed)
Black (incised)
Raised Collar

Incised Design

1
1

3
3

4
6

1
8
4

1
2

1934.]

Bennett, Tiahuanaco Excavations.

459

almost all plain wares of the Highland, and thus it has lost any specific
meaning. However, the " Chullpa " sites of Bolivia present an interesting problem in themselves. The coarse pottery shows decoration influence from both Tiahuanaco and Inca. It is not possible to say that
all "Chullpa" ware is chronologically post-Inca, or post-Spanish, but
undoubtedly most of it is. Indeed, the pottery made today by the
Aymara Indians is distinctly of the Chullpa type.
There is no evidence at Tiahuanaco that the Inca ware bears any
direct relationship to the previous cultures at the ruins. Stratigraphically it is superimposed on Tiahuanaco, but no pottery designs or shapes
suggest an amalgamation of styles which might further imply cultural
contemporaneity.

TIAHUANACO STONE SCULPTURE

The early descriptive accounts of Tiahuanaco contain many casual
references to stone statues or idols. Most noteworthy is the narrative
of Cieza de Le6n written in the middle of the sixteenth century. Standing on the fortified hill, Acapana, he saw two stone statues in the form
of human figures "farther on" (mas adelante deste cerro). They were so
large that they appeared like small giants. They wore long clothing
different from that of the natives and seemed to have some ornament on
their head. Father Diego de Alcobasal writes:There are also many other stones carved into the shape of men and women so
naturally that-they appear to be alive, some drinking with cups in their hands, others
sitting, others standing, and others walking in the stream which flows by the walls.
There are also statues of women with their infants in their laps, others with them on
their backs, and in a thousand other postures . . .

Although these accounts suggest considerable variety of stone work,

they are not sufficiently detailed to justify any classification. The later
workers, such as d'Orbigny and Squier give both descriptions and drawings, but despite these their omissions make acculrate classification
difficult. The work of Uhle and Stuibel,2 Courty, Posnansky, and others
among the latest workers give photographs and descriptions in adequate
detail. These materials, supplemented by my own notes and photographs, furnish the basis for the following descriptive classification of
Tiahuanaco stone carving.
Stone Sculpture at Tiahuanaco
1. Statue now called "El Fraile" (the Friar) (Uhle: No. 2)
Southwest corner of Calasasaya. Found by Uhle, 1877, in situ.
Uhle und Stilbel: Tafel 31, 1
Posnansky: 1914, frontispiece
Posnansky: 1912, Fig. 16
Means: 1931, Fig. 69
Schmidt: 570, 1
Bennett photograph
2. Large statue found in Pit VII of Bennett excavations, in situ.
Small temple to east of monolithic stairway (Calasasaya); now in La Paz
Photographs and drawings, cf. pp. 429-439.
3. Head of large statue. Found by d'Orbigny, 1833. (Uhle: No. 1)
Found east of Acapana, but is now in Museo Nacional in La Paz
Uhle und Stubel: Tafel 35 (collection of early drawings)
Lehmann und Doering: Pl. 19
Posnansky photographs
'As recorded by Garcilaso de la Vega, cf. Means, 1931, 121, and from Cieza de
in the original by Uhle und Stkbel, Part II, 1.
2Part II, 31-32.
460

Le6n, as quoted

1934.]

Bennett, Tiahuanaco Excavations.

461

4. Statue near Puma Puncu (Uhle: No. 3)

Once between Puma Puncu and Tiahuanaco village; now west of Puma
Puncu.
Uhle und Stubel: Tafel 31, 2 and 2a.

5.

6.

7.
8.

9.
10.
11.

12.
13.
14.
15.
16.

Posnansky photographs
Bennett photographs
Middle "Water God." Found by Courty, 1903, in situ.
Middle of three statues along railroad east of Tiahuanaco station
Courty: 533-534, Fig. 3
Posnansky, 1912, Fig. 25
Posnansky photographs
Bennett photographs
West "Water God." Found by Courty, 1903, in situ.
Courty: 533-534
Bennett photograph
West of three statues along railroad east of Tiahuanaco station
East "Water God"
East of three statues along railroad east of Tiahuanaco station
Bennett photographs
Central figure on "Gateway of the Sun"
Monolithic gateway now in northwest corner of Calasasaya
Illustrated in many places, with photographs and drawings, principally:Posnansky: 1914
Posnansky: 1912
Schmidt: 565, 1
Lehmann und Doering: P1. 2
Means, 1931, Figs. 70-71
Angular, headless statue
Museo Nacional de Bolivia, La Paz
Posnansky photograph
Parts of monolithic statue
Near Gateway of the Sun, Calasasaya
Bennett photograph
Statue in five pieces
Center of Calasasaya
Bennett photograph
Double statue
Posnansky: 1912, 23, Fig. 27
Double statue
Posnansky: 1912, Fig. 26
Double statue
Posnansky: 1912, Fig. 26
Double statue (Uhle: Nos. 12 and 13)
Schmidt: 450, illustrates one of these
Feet of monolithic statue
West center of Calasasaya
Bennett photograph

462

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

17. Head of statue (or perhaps bust) (Uhle: No. 19)

In court of church at Tiahuanaco Pueblo
Uhle und Stubel: Tafel 32, 5 and 5a
18. Well modeled torso of statue (Uhle: No. 6)
Found east of Acapana, near Nos. 3 and 36 by d'Orbigny, 1833
d'Orbigny: 343, and Atlas, P1. 7, bottom
19. Sculptured plaque. Found by Courty, 1903, in situ.
East part of Calasasaya, near stairway
Courty: Fig. 10
Posnansky photograph
20. Sculptured plaque
Musee d'Ethnographie, Trocadero (78-8-69)
Bennett drawing
21. A "typical" statue (Uhle: No. 4)
La Paz
Uhle und Stilbel: Part II, 31. Uhle says this is mentioned by Squier: 297,
and Mitre: 12 and 40
22. Pillar-like, blue stone statue (Uhle, No. 10?)
Calasasaya, southeast of Gateway of Sun
Bennett photograph
Uhle und Stiubel: Tafel 32, 6 and 6a
(This may be another statue, but the appearance is very similar. Seen in
Tiahuanaco pueblo on the west side of the plaza, as a house bench)
23. Simple statue with phallic emblem (Uhle: No. 9)
In a valley flat, south of Acapana, along railroad line (?)
Uhle und Stilbel: Tafel 32, 3
24. Small, bearded statue. Bennett excavation, Pit VII, in situ.
Small temple east of monolithic stairway (Calasasaya)
Photographs and drawings, cf. pp. 439-442
25. Unfinished, pillar-like statue. Found by d'Orbigny, 1833, in situ. (Uhle: No. 5)
East of Acapana, near Nos. 3 and 36
d'Orbigny: 342-343, Atlas, P1. 7, middle
26. Kneeling animal figure. (Uhle: No. 16)
Pueblo of Tiahuanaco, in small side street, northwest of Plaza
Uhle und Stilbel: Tafel 32, 2 and 2a
Posnansky photograph (?)
27. Kneeling zoomorphic figure (Uhle: No. 17)
Left of entrance of Tiahuanaco churchyard
Uhle und Stuibel: Tafel 32, 1 and la
28. Kneeling animal figure with head missing
Puma Puncu
Bennett photograph
29. Kneeling animal figure (Uhle: No. 18 ???)
Region of Copacabana (Now in Museum fur Volkerkunde, Berlin)
Schmidt: 451
30. Seated statue with turban (Uhle: No. 14)
South side of Tiahuanaco churchyard entrance
Uhle und Stilbel: Tafel 33, 1 and la

1934.1

31.

32.

32a.
32b.
32c.
32d.
33.
34.

35.

Bennett, Tiahuanaco Excavations.

463

Posnansky: 1914, Figs. 30-32

Bennett photograph
Seated statue with turban (Uhle: No. 15)
North side of Tiahuanaco churchyard entrance
Uhle und Stilbel: Tafel 33, 2 and 2a
Posnansky: 1914, Fig. 29
Posnansky photographs
Bennett photograph
Stone heads
Small temple east of monolithic stairway (Calasasaya)
Posnansky: 1914, Tafel XXXVI, Figs. 14-16
Posnansky: 1912, Figs. 2-4
Courty: Part 2, Fig. 5
Posnansky and Bennett photographs
Heads with headband
Flat face heads
Rounded boulder heads
Modeled face heads
Carved lizard or toad. Found by Courty, 1903, in situ.
Small temple east of monolithic stairway (Calasasaya)
Posnansky: 1914, Pl. XXXXI, Fig. 33
Carved toad (Uhle: No. 21)
Between Tiahuanaco pueblo and Calasasaya
T. von Tschudi: 294
Carved stone corner. Found by Courty, 1903, in situ.
Four of these as entrances to "altars" just outside the northwest
corner of Calasasaya
Courty: Part 2, 541, Fig. 12

Posnansky photograph
36. "Sphinx-like" statue (Uhle: No. 18 ???)
d'Orbigny: 341, P1. 7 top
37. Monolith with carved "paddle" (Uhle: No. 11)
East of Acapana, near No. 3
Uhle und Stilbel: Tafel 32, 7
38. Roughly carved stone with face. Bennett excavation, Pit VII, in situ.
Small temple east of monolithic stairway (Calasasaya)
Photograph cf. Fig. 33
39. Part of stone head carved in relief
Posnansky photograph
40. Flat stone carved in low relief, in zigzags and curves
Posnansky photograph. (There are two photographs which may represent
two such stones, or the two sides of this one)
41. Conventionalized condor head, angular carving
Buch collection, La Paz
Means: 1931, 132, Fig. 72. (Also Means: 1931, Figs. 73-75 of specimens
of similar type, reported as from Peru, in Trocadero Museum, Paris)
42. Non-Tiahuanaco style statue
Near the village
d'Orbigny: 347 (description)

464

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

STYLE 1, THE CLASSJC MONOILITHIC STATUE

One group of the largest statues forms a compact stylistic unit.
The individual pieces vary only slightly from the "type" pattern, as
is shown in Table 10. Statues 1 to 7, 10 and 21 all fall into this group.
A description of the type has already been given in some detail with
reference to the large statue in Pit V'II (No. 2), but a summary is given
here.
The statues are large. The height of the figures, exclusive of the
bases, ranges from 2.05 to 5.50 meters: the widths from .40 to 1.05
meters. The height of the head is from 28 to 40 per cent of the total
figure height. All are made of reddish sandstone, with the exception of
the large head, No. 3, which is of trachyte.
Except for Nos. 1 and 10 these statues have a vertically grooved
headdress, probably representing feathers. All in this group, without
exception, have wide, raised, decorated headbands. The bands are flat,
and the decoration incised, or in low relief. The faces are either square or
slightly rectangular; across the forehead and down both sides is a double
step ridge. The back of the head is flat (again excepting Nos. 1 and 10
with rounded heads), and tresses are indicated in some by vertical rows
of rectangles. The nose is straight-sided and projected. The mouth
is a raised, round cornered rectangle. with a groove inside, although two
statues, Nos. 3 and 4, have a double row of teeth indicated. The cheeks
are rounded. The eyes are round-cornered squares. All statues with
details still preserved have wings on the eyes, as well as a tear band with
three round (or rounded-square) tears flowing from the eyes. The chins
are projected from the chest and slightly modeled. Two statues, Nos.
1 and 3, have condor heads incised on the chins. The ears are variations
of the rectangle with a notch in the front side and another in the lower
back corner. The variations consist of other notches cut from the back
side and from the upper back corner.
The shoulders are rounded, and some slant forward somewhat.
They project back from the head slightly. The chest is flat and, almost
without exception, a decorated area. On statues Nos. 1, 2, 4, 5 the arms
are placed so that the hands meet on the chest; but on Nos. 6 and 7 the
arms and hands extend straight downward at the sides, the hands
covering the waistband. The left hand is depicted as open, with five
fingers which, in Statues 1 and 2 are correctly proportioned as to length.
When the left hand is at the chest position it holds a kero-shaped cup
with a raised, decorated band around the upper part. In the statues
with the hands on the chest the right hand is depicted as closed, and as

1934.]

Bennett, Tiahuanaco Excavations.

465

open in those statues with the hands at the sides. Again, five fingers
are indicated. The closed right hand is distorted as shown in Fig. 25,
and holds a scepter-like object.
All the statues have a wide, flat, waistband, decorated in low
relief or by incision. Nos. 1 and 2 have pants decorated with raised
decoration, and Nos. 4, 5, 7, 10 have a skirt that extends below the knees.
The legs, below the garments, are usually distinguished by a groove which
separates them. The feet are modeled and have five toes each, except for
Nos. 4 and 7 which have rounded feet. Only No. 1 has a jointed ankleband. All the statues have plain bases which project into the ground.
Statues 11, 16, and 18 might well be classed as a minor variation of
the preceding group. They are really not sufficiently distinct to warrant
a separate division. No. 11 is in five parts today. It is of a fine-grained
andesitic rock, and well carved, though more in the round than the
preceding group. It has the wide decorated headband, the squareshaped face with the double-step side ridges, the projecting nose. and the
rounded square eyes. The chest is rounded and decorated. The arms
hang down at the side; the hands placed over the flat waistband have
five, proportioned, squiare-tipped fingers. Medallion decorated pants
extend down to the ankles. The legs are rounded and completely separated. A raised band encircles the ankles. The feet are rounded. All
the statues stand on a plain base. No. 16 consists only of a pair of feet,
of the type just described. No. 18, as drawn by d'Orbigny, might well
be the same as No. 11, except that it is described as without head and
feet.
The headless statue in La Paz, No. 9, is difficult to classify. The
sides are straight, the angles sharp. The block effect too is more apparent
than in the others previously described. It is made of lava rock. The
arms are placed so that the left hand rests on the stomach and the right
on the chest. Both hands are open with five fingers, only the thumb being
proportionately shorter. No objects are held in the hand. The waistband is raised and decorated with the male and female sex (?) designs
with projecting condor heads, so typical of other statues and the Gateway of the Sun. A large fish head is designed on each leg, as on statue
No. 5. The legs are separated by a groove. While the designs connect
this statue with the first style, the general technique is more like that of
Style 3. The position of the hands certainly suggests the small monolithic
statue of Pit VII (No. 24, cf. p. 440).
All of the statues thus far described are monolithic. The total
effect is that of an upright pillar, modified into a semblance of the

466

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

human form. The carving is well done, the conventions skilfully

executed, the designs delicately incised.
STYLE 2, VIRACOCHA, GATEWAY OF THE SUN FIGURE
The type for this group is the central figure on the Gateway of the
Sun. Others of the same type are the "double" statues, which consist
of a single squared pillar with two figures projected from one side in high
relief, one right side up, the other upside down, and above it, the heads
close, but not actually connected. Posnansky illustrates three of these,
Uhle refers to two figures, and Schmidt illustrates one such double
statue (probably one of Uhle's). Following my numbering, this group
includes Statues 8, 12, 13, 14, 15.
These are not really statues, but rather figures carved in high relief
on one side of a block. The material is andesitic. The figures are short,
about .50 centimeters high. The double statues are essentially identical
with the central figure of the Gateway of the Sun, but are simpler. The
few differences are noted in the description.
The Viracocha figure has an elaborate crown extending around the
head and consisting of projected puma heads, discs, etc. The double
statues have the flat headband, embellished with a projected diadem,
and mounted above with the vertical groove feather decoration. All
have square faces, with a single stepped ridge across the forehead and
down the sides. One double statue has two tresses on each side of the
face as well. The noses have concave sides. The mouths are round
cornered rectangles. The cheeks are rounded and no ears are represented.
The eyes are indented and round, elaborated with plain wings, and three
round tears.' The chins are slightly projected. The shoulders are
straight. The chests are flat; that on the Viracocha figure is decorated;
those on the double statues are plain, except for a raised ridge which
extends horizontally across the chest. The arms are projected, one on
each side. Each hand has three fingers and an opposed thumb and holds
a scepter. The Viracocha figure has a narrow segmented waistband with
straps running over the shoulders. All figures have skirts which extend
to the ankles. The legs below are separated. The feet project forward,
without toe distinctions. The Viracocha figure stands on an elaborately
decorated, stepped base, and the double statues stand on a plain stepped
base.
One cannot fail to notice the strong resemblances in style between
this group and the first one, in spite of the differences in application.
'Not all the double statues preserve all the details. However, since they
the comparison is based on a summary of the traits of all of them.

are all

essentially similar,

1934.]

Bennett, Tiahuanaco Excavations.

467

The double statues have the headband with vertical grooves above.
The treatment of the face is practically identical in the two groups,
particularly in regard to the winged, three-teared eyes. Table 10 shows
many other detail comparisons. When the design elements are compared
the resemblance between the two groups is even more striking. The running human and condor figures of the Gateway of the Sun are also found
on the headband of No. 5 and on the chest and waist decorations of No.
2. The puma, condor, and fish head designs are common elements of
both the Gateway and the monolithic statues, except for the crowned
condor design which I have not seen on the statues, and the curled tail
puma which is not on the Gateway. The male and female sex (?) signs
with projecting condor (and fish) heads occur on the Gateway and as
common elements of decoration on Statues 2, 3, 5, 9. The puma head
on a fish body which forms the center of the pedestal of the Viracocha
figure on the Gateway is also used on the headband of Statue 3. The
rectangular segment bar is a design in common. Simplified suggestions
of the sun faces from the frieze of the Gateway are found on Statues 2,
11, and possibly 4. A Viracocha figure is found on the back of Statue 2.
The net result of this comparison leads to the conclusion that
Styles 1 and 2 are manifestations of the same cultural phase. On the
basis of comparative design and technique, the sculptured plaques Nos.
19 and 20 may be included in this double group. No. 19 has four curledtail pumas carved along one edge, very similar to those on the headband
of No. 3. No. 20 has the heads of two human figures holding scepters in
much the same style as the running human figures on the Gateway of
the Sun. There are several indications that the second style may be a
slightly later phase than the first. The stone is volcanic rather than
sandstone; the Gateway seems slightly more formalized; the Viracocha figure has only four digits instead of five. However, there is no
indication of any great difference in time between the two styles.
STYLE 3, SQUARED PILLAR TYPE STATUES
This group of statues (Nos. 22, 23, 25, and possibly 24) look like
decadent forms of Style 1. Simplicity, angularity, and mediocrity characterize them. The statues included in Style 3 do not form a compact,
easily distinguished stylistic unit, but are the odd monolithic statues
which do not fit into the other styles. They are from 1.5 to 2.5 meters
high and are all over .40 meters wide.
No. 22 of a bluish volcanic rock has a square face projected from a
square background, with a square ear on one side, simple facial features,

468

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

and roughly worn back head. The shoulders are square and hardly
distinguishable from the head block. The squared arms continue from
the sides so that the five-fingered, squared hands almost meet in a band
across the front. The waistband is narrow and decorated with triangular
grooves. A phallus projects above the waistband in the center. The legs
are separated by a groove, and the feet, slightly projected, have five toes.
No. 23 is quite plain. The head is distinguished from the body and
topped with a band or disc-like crown. No features are marked. Short,
plain arms hang at the sides. There is a narrow waistband. No legs or
feet are distinguished. At the base is a well-modeled phallus with
testicles. No. 25, as drawn by d'Orbigny, is of this same type. The
headband is more distinct and a nose is projected from the flat face.
Flat arms without details hang at the sides, covering the narrow, plain
waistband. The legs are divided by a groove. There is no phallus.
No. 24, the small monolith of Pit VII, has already been thoroughly
described (cf. p. 441). It is clearly different from the other statues in
this style, except for the pillar-like effect of the whole. The lightning
rays, highly projected features, the beard, the puma-design ears, the
snake design on the sides, the pumas on the skirt, and the position of the
five-fingered hands on the chest and stomach really place this statue in a
distinct class.
STYLE 4, ANIMAL HEAD FIGURES IN SITTING POSITION
This group forms a compact, definite stylistic unit. There are four
statues of this type, Nos. 26, 27, 28, 29. The figure is in sitting or
kneeling position, with both legs doubled under it. The head is clearly
animal. The jaw is wide open, the ears back on the rounded head are
small and rounded. The eyes, when represented, are projected and
circular. Short arms hang over the waistband on each side. The upper
part of the arm, including the shoulder, is shaped into semi-realistic
form. The left arm has a simple face instead of a hand at its extremity.
The right hand holds a hammer-shaped object like a stone-headed club.
A narrow waistband has a wedge-shaped piece in the back. The bent
knees are distinguished by a groove, and the edge of a skirt is delineated
on one of the figures. All are lava figures. No. 29 is from Copacabana,
but is clearly of the same style.
STYLE 5, KNEELING HUMAN STATUES
The two statues that flank the entrance of the church patio in Tiahuanaco village (Nos. 30 and 31) have frequently been described as a
distinct type. There is plenty of evidence for this distinction. They are

1934.]

Bennett, Tiahuanaco Excavations.

469

of red sandstone and far more naturalistic in general treatment than other
Tiahuanaco sculpturing. The position of the statues, apparently kneeling, is in itself distinctive. The use of curves, instead of angles, the
natural position of the head and arms, the realism of the features, all set
them in a separate class. The headbands are wide and decorated with a
scroll groove which gives the whole band the appearance of a twisted
turban. The headband of No. 31 has a coiled diadem in front which may
terminate in a human face, and in back there is a definite face with small
round eyes, modeled base nose, and open mouth with two rows of teeth.
Above the headbands the head is modeled slightly, perhaps to represent
a cap or head cloth. A ridge which crosses the forehead, under the headband, and down the sides of the head, represents either hair, or the
head cloth just mentioned. The hair along the back of the neck is indicated. A curved ear is projected on the side of the head. The profile
of the face is convex, in contrast to the straight faces of the other statues.
The forehead is wide. The oval eyes are inset. There are neither wings
nor tears. The nose is projected. Cheek bones are prominent and
enhance the "modeled" effect of the face. The mouth is projected and
curved upward at the ends.
In No. 30 the right knee is bent and the right arm clasps it, though
there is some suggestion that the right arm is holding some obiect. The
left hand is at the side. The hands have four and possibly five fingers.
No. 31 appears to be kneeling on both knees. The left hand is
clearly placed on the left knee, with five fingers and finger nails depicted.
The right hand is crossed on the chest. Accoi ding to Uhlel und Stiubel
the seamed edge of a cloth garment is depicted, as well as a collar piece.
These two figures certainly represent a style quite distinct from the
others at Tiahuanaco.
STYLE 6, STONE HEADS
Since almost all the stone heads from Tiahuanaco have been found
in the small temple east of the monolithic stairway of Calasasaya they
have formerly been classed together as contemporaneous, not, however,
without notations on the variations in style repiesented. The interpretation of this temple as chronologically recent (cf. p. 387) allows for
stone heads to have been collected from all parts of the ruins to be
placed in this temple. With this interpretation the various styles of
stone heads are not necessarily contemporaneous. There are five general styles represented in the carving of the stone heads. These are
described in the following pages.
'Text to Tafel 33.

470

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Substyle 6a, Headband Type. These stone heads are really faces
projected from squared stone blocks evidently intended as units in wall
construction. All have plain wide, flat, headbands around the forehead.
The headdress above the band is also plain. The faces are square,
although some are slightly rounded at the chin. A single stepped ledge
runs across the forehead and down the sides of the face. In some the
side ledge is accentuated, as if representing side whiskers. The ears are
rectangular, the nose wedge-shaped, and projected. The eyes are round,
some projected and some sunken. Details of wings and tears do not
appear. The mouth is a hollowed oblong surrounded by protruding
lips. Some possibly have files of teeth. In some the cheek bones are
prominent.
Courty,' who discovered many stone heads, says that a red ocher
(and sometimes an ultramarine blue) could be seen in the slits of the
eyes, ears, nose, and mouths when the heads were first unearthed. He
also says that the material is a metamorphosed trachyte with a white
fracture.
Posnansky2 says that the majority of the stone heads are of this
type. The general impression is certainly suggestive of the Style 1
monolithic statues. All the details are not exact parallels, but there are
numerous comparable points. The resemblance is more noticeable when
this type of head is contrasted with the others described below.
My present belief is that these stone heads are contemporaneous
with the monolithic statues of Style 1 and were used as wall decoration
in the buildings of that period. Later, they were collected from the ruins
and re-used in this late temple to the east of the monolithic stairway of
Calasasaya. It is the only way of accounting for the discrepancy between these well-squared and comparatively well-sculptured building
block units, and the crude, half cut, half rough stone walls of the semisubterranean temple.
Puma heads are also projected from stone blocks in this same style.
They are classic representations of the puma in stone. They tend to
confirm the connection between Style 1 and Style 6a.
Substyle 6b, Flat Stones with Faces. This group, also fairly numerous,
consists of flat, thin, irregular or roughly squared stones with simple
features, depicting a face on one surface. They are made of a chalklike stone. The nose is a vertical raised bar. There are some with slight
brow delineations, but no suggestion of the T-shaped nose. The eyes are
1Courty, Part 2, 536.
2Posnansky, 1914, 83.

1934.1

Bennett, Tiahuanaco Excavations.

471

round, small, and projected. One head in the Munich collection has a
circular protrusion on each cheek, below the eyes, perhaps representing
a single tear. The mouth is generally represented by a horizontal groove,
though some of the mouths are slightly elaborated with protruding lips.
The contrast between Substyles 6a and 6b is so obvious that it
hardly needs elaboration. The first set is fairly finished in technique
and stylization and the second group is crude, simple, and classed as
scuilpture only by the grace of an inclusive definition.
Substyle 6c, Rounded Boulder Heads. These heads are made from
slightly modified, round, sandstone boulders, with one side flattened and
carved with simple features. The nose is raised, with concave sides and
slight bulge at the base. (One has a plain straight nose). The eyes are
round and raised and one head in Munich has a single tear on the cheek.
The mouth is a hollow with a raised lip outline, in round cornered
rectangle shape, sometimes turned slightly up or down at the corners.
This substyle is very similar to the preceding one (6b), but the workmanship is slightly better. Substyles 6b and 6c form a group for which
there are no other parallels in Tiahuanaco sculpturing. The type is
certainly not characteristic of the ruins. There are some indications
that it may be a post-Tiahuanaco phase.
Substyle 6d, Modeled Face Heads. Two stone heads form a separate
group which Posnansky has called "portrait" types.1 They are flat,
but realistically shaped modeled faces. Behind the head is a cylindrical
projection for wall-mounting. The foreheads are round. The noses are
rounded on top and modeled to shape at the base. Hair is depicted on
one head. The lids are portrayed on the modeled eyes. The cheeks are
rounded. The mouths, one horizontal and one turned up at the edge,
have realistically shaped lips. Two rows of teeth are visible. The chins
are modeled.
The naturalism of these faces suggests the two statues of Style 5.
Unfortunately, there is no other evidence to associate the two styles.
The dowel-like projections on the back of these heads for wall attachment are not typically Tiahuanaco.
Substyle 6e, Cornerstone Faces. Four small square cornerstones
(No. 35) were found by Courty2 at the entrance to the "altars" just out'
side the northwest corner of Calasasaya. The four are identically alike.
The face is on two sides of- the block, the edge of the block neatly bisecting the nose and the mouth. The carving is in high relief, with sharp
1Posnansky, 1914, 83-84, also Tafel XXXVI, Figs. 15, 16.
2Courty, Part 2, 541.

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

472

edges. The nose has straight sides, but widens out in a flare at the base.
The mouth is oval. The eyes are round, and a large question-mark
shape surrounds each. The ears are rectangular.
Another piece in the same general style, although zoomorphic
instead of human, is No. 41. It is a conventionalized head, perhaps of a
condor, in high relief on a stone block. A raised disc serves as an eye and
similar discs ornament the space above and behind the head. Means
writes that this
represents the second phase of culture there, (but) does not resemble the carving of
the (Gateway) frieze at all save for the fact that they have the same square-edged
cutting. It seems to me on aesthetic grounds that it represents a later Period, one in
which, because of excessive conventionalizing tendencies, designs had lost much of
their coherence.1

Possible Additions to Style 6. The roughly carved stone of Pit VII

(No. 38) which is described on p. 443 is little more than a face roughly
and simply delineated at the top of a large rectangular stone. The simplicity is suggestive of Substyles 6b and 6c, and this upright might well
be a variation in the same style.
d'Orbigny2 illustrates a figure (No. 36) of which he says:. . .

un sphinx informe qui represente une tete humaine et derriere des esp6ces

d'ailes;

. .

"

From his drawing this "sphinx" looks like a variation of Substyle 6a,
with headband and all. The drawings in d'Orbigny's Atlas are not noted
for their accuracy, in any case.
STYLE 7, CARVED LIZARD, SALAMANDER, OR TOAD
According to Uhle und Stubel,3 T. von Tschudi found a great stone
with a crude animal figure which he called a toad, but with a wide, three
piece tail (No. 34). Courty found a similar, though much smaller stone
(No. 33), in the small temple east of the monolithic stairway of Calasasaya. Posnansky4 calls this animal a sapo con cola de pez (toad with a
fish-tail). It has four legs, a small flat head, and a large, spreading, flat
tail. It is carved in high relief on a rough stone.
The style of workmanship is crude and does not differ greatly from
the technique of Substyles 6b and 6c. I have designated it as a distinct
style only because of the animal represented. I think that the distribution of carved lizard-like animals will be different from that of flat or
boulder stone heads.
'Means, 1931, 132.

2d'Orbigny, 341; also Atlas, P1. 7 top.

sUhle und Stuibel, Part 2, 32.
4Posnansky, 1914, Tafel XXXXI, Fig. 33.

~1934.]

Bennett, Tiahuanaco Excavations.

473

STYLE 8, GEOMETRIC PATTERNS

No. 40 represents another distinct style found at Tiahuanaco. This
stone is part of a wide slab which probably stood upright as a monolith.
It is completely carved on the flat surfaces. The design is characterized
by its symmetry and by the employment of such geometric patterns as
squares, crosses, zigzags, triangles, circles, and scrolls. The center of the
slab has a toad-figure outlined by a diamond shape. Surrounding this
is a complex scroll pattern, symmetrical on all sides. The lower half of
the slab is decorated with zigzag and cross designs.

MISCELLANEOUS PIECES
Uhle und Stiibell describe and illustrate a monolithic, green schist
upright, 1.80 meters high (No. 37). This is decorated on one side by
raised designs. At the top are two triangles, then a kero-cup shape, a
horizontal bar, and, at the bottom, a paddle shape with two grooves
across the blade. This monolith is called "El Fraile" by the authors,
while the statue (No. 1) which is now known by that name locally, is
merely referred to as a large statue. This upright does not fit into any
of the classifications, though its closest parallel is with Style 3.
No. 39 is the broken top part of a monolithic statue. The top is
smooth. Then comes a wide raised band which surrounds the stone.
This headband is designed with a figure with a snake-like body with large
spots. The head is elongated, and two appendages branch out from the
chin like feelers (?). The mouth is small and oval. The nose has straight
sides and a widened base. The eyes are square. Two small ears project
back from the flat head. Below the band, discs on the end of stepped
bars are the only visible decoration. The style of the decoration and
carving is Classic Tiahuanaco, though the figure represented is not; that
is to say, I have not found it on other stone carved specimens. However,
a Tiahuanaco-style poncho in the Metropolitan Museum of Art has a
tapestry representation of this identical figure.
Further grouping and chronological implication of the eight stone
sculpturing styles listed here is, as yet, without foundation in actual
concrete data. However, a few tentative suggestions of possible groupings and possible chronology can be made on the basis of subjective
evidence. Certain similarities between styles have been mentioned
throughout this classification and a short summary grouping might
assist in clarifying the problem. Proceeding on this basis, the eight
styles and substyles might be reduced to four groups:lUhle und Stiubel, Tafel 32, 7.

474

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Group I, Realistic Stone Carving (Styles 5, 6d, and possibly the

Puma Heads of 6a). The two kneeling statues in front of the church and
the two modeled type stone heads certainly form a group which is more
realistic in style than the other Tiahuanaco stone sculpturing. These
may well be older than the other material, as Posnansky and others
believe. On stylistic grounds it would be difficult to derive these realistic
statues from the highly conventionalized Classic monoliths. They were
probably found near the pueblo church where they now stand and thus
are not part of the main group of ruins. Realistic carving, generally
considered as early, has a wide highland distribution, even up to Huaraz
in Northern Peru, as will be shown in the concluding chapter. The
sculptured puma heads, with projections for wall insets, might be included in this group, not so much for their realism as for their distribution which parallels that of the realistic statue. However, at Tiahuanaco
they are more closely related to the stone head group.
Group II, Conventionalized Classic Figure and Head (Styles 1, 2, 4,
6a). The large monolithic statues and the high relief figure treatment
of the Gateway of the Sun and the double statues form one fairly compact stylistic group, although the individual pieces may vary considerably. The stone heads with headbands are similar in general treatment. If the puma head variety of the stone heads is included there is a
logical connection with the animal head figures in sitting position (Style
4). This group includes sculpturing in both sandstone and lava. The
suggestion has already been made that the lava group may be somewhat
later than the sandstone group. However, the whole of Group II is
solidified when contrasted with Groups I or III. Furthermore, the
known distribution of Group II style is fairly limited to Tiahuanaco.
Group II represents Tiahuanaco stone sculpturing at the height of its
development.
Group III, Technically Decadent Pillar-like Statues and Heads
(Styles 3, 6b, 6c, 7). The residue of statues and sculpturing not definitely
assignable to any particular style is gathered here in a group which has,
as mentioned previously, the common characteristics of simplicity,
angularity, and mediocrity. Some of the statues still retain a few characteristics of the Classic style, but, on the whole, they appear to be
Decadent forms. However, some new stylistic elements appear which
keep the group from being purely a degeneration of the Classic.
The inclusion of the carved lizards or toads, the boulder heads, and
the flat stone heads in this group is purely arbitrary. Simplicity of
treatment is about the only common factor. However, chronologically,

1934.1

Bennett, Tiahuanaco Excavations.

475

these heads, like the statues, seem more recent than the Classic Tiahuanaco. Both styles are found in the small temple to the east of Calasasaya (Pit VII). The final identification of Group III as a contemporaneous stylistic unit definitely later than Group II awaits objective
confirmation.
Group IV, Geometric Style Carving (Styles 6e, 8). One slab is
elaborately carved with geometric designs. Four cornerstone heads
have a geometric stylization of treatment. Since the slab has a toad
figure in the center, the carved lizard-toad group (Style 7), might also
be included here. This is further confirmed by the slabs with lizard and
geometric patterns found by Squierl at Hatuncolla.
The chronological stylistic sequence of Groups I, II, III is postulated tentatively. Group IV is difficult to place. It is later than
Group II, but whether contemporaneous, earlier, or later than Group
III, cannot be determined. However, the whole chronological sequence
awaits further study for confirmation.

'Squier, 385-386.

CONCLUSIONS
At the present writing the Tiahuanaco problem can be conveniently
divided into three parts: the local phases and correlations at the Tiahuanaco site itself; the position of Tiahuanaco in Bolivian archaeology;
the distribution of Tiahuanaco materials and influence throughout the
Andean area.
This paper has dealt only with the local problem of Tiahuanaco,
and furthermore, principally with one section of that problem, namely,
ceramic stratigraphy. In summary, four ceramic phases are represented:
Early Tiahuanaco; Classic Tiahuanaco; Decadent Tiahuanaco; Post
Tiahuanaco and Inca.
The Early Tiahuanaco levels contain over 95 per cent plain sherds.
A straight rim, rounded bottom, open bowl; a flaring rim, globular
bowl; and a shallow, horizontal-handled dish are the typical, unpainted
shapes. The painted ware shapes are principally variations on a long
constricted neck, globular base decanter; a flat bottom, wide flat flaring
rim, spittoon-shaped bowl; and a flat bottom, fretted rim incense bowl,
with or without a modeled puma head on the rim. The designs are zigzag linear with alternating colors, or zoomorphic with peculiar animals,
not outlined in black, but painted in colors on a black background.
Black, white, red, orange, and brown shiny colors are applied directly
to an unslipped, buff-colored clay. Design areas are limited. Some of
the decoration made by filling incised lines with colors probably belongs
in this group. Small clay buttons and a T-shaped stone ax are distinctive artifacts of these levels.
The Classic Tiahuanaco levels contain over 50 per cent painted ware.
The most typical ceramic shapes are the flaring sided, kero-shape goblets;
the hollow base, two-handled and collared puma head types of incense
bowl; a flat base, wide rim, open bowl; and a flaring rim base. Designs
are painted in black, white, yellow, and sometimes gray and brown on a
red slip. Linear outline in black is typical. Design areas repeat around
the vessels. The most frequent designs are a curled-tail puma in profile,
a condor with wings and tail feathers, a complete human figure with
profile head, and geometric units based principally on the step-fret
pattern. The colors are rich, the designs well finished, and the vessels
are technically well made and polished.
The Decadent Tiahuanaco style is a degeneration of the Classic,
combined with some new elements. Variations of the kero-shape goblet
and.the hollow base incense bowl are still found, but the most typical
shape is a flat bottom, flaring sided cup. New shapes are angular bodied,
476

1934.1

Bennett, Tiahuanaco Excavations.

477

one-handled pitchers; two-handled, high shouldered jugs; and open,'

curved flaring rim, round bottom, two-handled bowls. Black and white
applied over an orange slip is the typical color combination. There is a
greater emphasis on geometric design, the scroll, douible S, zigzag, and
step becoming prominent elements. Zoomorphic design represents a
break-down of the Classic condor, puma, and human figure, into crude
imitations, or designs based on a condor wing, a puma head, a human
eye, etc. Together with this break-down in completeness of design
occurs a general decadence in finish, application, texture, and polish of
the ware.
The Inca and Post Tiahuanaco ware is stratigraphically in the top
half meter or on the surface. The Inca designs are typical of the geometric patterns of Cuzco pottery; the shapes are principally aryballoids
and animal-handled shallow plates. There is nothing that suggests
the derivation of Inca forms from Tiahuanaco types. In other words,
the Inca material is superimposed on Tiahuanaco, but with no implications of historic connection. The Post-Tiahuanaco, like "Chullpa" is
a common Bolivian type discussed later in this section.
Although this stratigraphic set-up is the result of my excavation
of ten pits at Tiahuanaco, it is not conclusively proved. Only two pits
showed the complete series. Since whole vessels of the Early Tiahuanaco
type are found in collections it is probable that it occurs in positions more
favorable for preservation than the two to four meter depths of my excavations. Undoubtedly some grave material is of this type, and future
excavations must determine whether the grave material is isolated or
associated with other wares. Pit X, of the present excavations, was a
fairly pure Classic site, suggesting tentatively that the Classic phase
can be isolated by burials. The Decadent phase is perhaps the most
definitely established, by grave isolation, stratigraphy, and distribution.
It is generally agreed that there were at least two phases of Tiahuanaco culture. Most writers either follow Posnansky's designation of
periods, based on material and building technique, or draw their conclusions from speculation. Means' does state that Tiahuanaco I pottery
is distinct and within that group "there is a contrast between extremely
coarse incised vessels and almost equally coarse, but painted, vessels."
Courty alone actually mentions stratified ceramic sites2:
Stratigraphically, under the Aymara cemeteries with common red pottery, I
have found the beautiful polished vases on which are painted the subjects which are
easily reencountered in the sculptures of the great monuments of Tiahuanaco, on the
idols and on the gateway named 'of the Sun'.
'Means, 1931, 112.
2In Crequi-Montfort, 66.

478

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

However, the major distinctions in Posnanksy's Tiahuanaco I and Tiahuanaco II have not been based on ceramics. The distinctions are based
primarily on building materials. Sandstone and lava rock are the principal materials used. Some buildings are constructed entirely of sandstone,
others entirely of lava, and still others with both stones. Furthermore,
there is a contrast in architectural style between buildings with walls
constructed with a series of upright megalithic pillars filled in between
with smaller stones, and buildings constructed with notched and jointed
blocks. In the all-sandstone buildings the megalithic upright technique
is employed and in the all-lava buildings the notched and jointed blocks
are used. In all probability, the sandstone megalithic style precedes the
lava jointed-block style, although absolute proof is lacking. Furthermore, there is no evidence of any great chronological or cultural discrepancy between the two styles. On the contrary, the fact that Calasasaya, Acapana, and Puma Puncu were started with one material and
finished, or continued, with another, without any radical changes in
building plan, indicates a rather close connection between the two styles.
It is not possible at this time to associate ceramic periods with the
building periods. The Classic levels had cut stones in association in three
pits, but this is not sufficient to identify a building style. Pit II was in
the terre-plein of Calasasaya temple and the Classic material is quite
possibly contemporaneous with the building. Unfortunately, however,
both sandstone and lava materials have been used in the construction
of Calasasaya. Neither the Early nor Decadent Tiahuanaco levels
could be associated with definite buildings, although dressed and drilled
stones, used secondarily as paths in the Decadent level, might indicate a
post-building phase. Pits excavated with this association problem in
mind might well reveal some connections.
An examination of the e.xisting and illustrated stone sculpture found
at Tiahuanaco resulted in a descriptive classification of eight styles. A
tentative, subjective analysis of these eight styles arranged them in four
stylistic groups with a possible chronological succession as follows:Group I. Realistic stone carving (kneeling figures; modeled heads). An Early
Tiahuanaco phase.
Group II. Conventionalized, classic figures and heads (the large monoliths; Gateway of the Sun type high relief; stone heads with headbands). The
Classic Tiahuanaco.
Group III. Technically Decadent pillar-like statues and heads (including a variety
of mediocre, simple, statues; boulder and flat carved stone heads).
A Decadent Tiahuanaco phase.
Group IV. Geometric style carving (slabs with geometric design; angular cornerstone carved heads). Uncertain position.

1934.]

Bennett, Tiahuanaco Excavations.

479

The analogy between the style groups for stone sculpturing and the
ceramic periods must not be forced. The zoomorphic designs of the
Early Tiahuanaco, while not in the same conventions as the later periods,
are not completely realistic. There is little doubt that the Classic Tiahuanaco ceramics are contemporaneous with at least part of the Group II
sculptures. Still, as mentioned in the design style discussion (p. 403),
there is a distinction between the design subject of stone carving and
pottery painting. Decadent Tiahuanaco has no demonstrable relationship with Group III sculptures. Until more concrete evidence is presented for actual connections between styles, the analogies must be
treated with caution.
Group II style sculpturing is again the only one that can be associated with building techniques. The same designs which occur on some
of the statues are found on the Gateway of the Sun, the Gateway of the
Pantheon, and on individual stone building slabs. The analysis of stone
sculpture showed a possible slight time variation between the sandstone
and lava statues in Group II. This compares favorably with the difference of sandstone and lava materials in building technique which likewise suggests some, but no great chronological discrepancy.
Future excavations at Tiahuanaco have then several concrete problems: a further checking of ceramic stratigraphy and the correlation of
stylistic differences in building, materials, and stone sculpture with the
ceramic periods, as well as with each other.
The position of Tiahuanaco in general Bolivian archaeology is the
second division of the problem. Reconnaissance archaeological work has
been rather extensively carried out in Bolivia, principally by Adolph F.
Bandelier' who worked for several years in the highlands and whose diary
and Bolivian manuscript are now at the American Museum. Others,
namely Nordenskiold, Posnansky, Uhle, Squier, have also explored
Bolivia. Considerable work has been done by local enthusiasts. Part
of my own work of last year was devoted to a survey around Titicaca
Lake. An examination of the present status of Bolivian archaeology
emphasizes the fact that there is only one Tiahuanaco ruin. Posnansky,2
who has traveled about the altiplano for years searching for Tiahuanaco
remains, writes:The only monuments around Lake Titicaca which we believe could belong to
the Tiahuanaco epoch would be, according to their style, the ruins of Sillustani . . .

Max Uhle considers the massive stone terrace at Llojepaya on the south
shore of Copacabana as of Tiahuanaco type, as well as the ruins of
'The I8sand8 of Titicaca and Koati.
2Posnansky, 1913, 11.

480

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Taaq'ani on the western end of Cumana Island, and Luquimata on the

peninsula of Taraca,1 Several cut stones on the Island of Titicaca might
be considered Tiahuanaco in style. Cut stone seats and benches are
common around the peninsula of Copacabana, although not definitely
assignable to Tiahuanaco. The ruins on the Islands of Titicaca and Coati,
on the Peninsula of Huata, around the northern shore of Lake Titicaca
from Achacachi to Escoma and back into the mountains along this same
stretch, around La Paz and down the plateau to Patacamaya, Sicasica,
and the whole region around Illimani glacier, are decidedly not of
typical Tiahuanaco style. Those around the lake are fortified Inca
type villages of piled, undressed stone with narrow doorways, corbelled
arch roofs, niches and shelves, but no windows. Those around Sicasica
are adobe towers used both for dwelling and burial, and commonly called
"Chullpas." To the southeast, in Santa Cruz, Bolivia, at Samaipata
cut stone has been found.2
In general, it can be said that up to the present time no ruin has
been found in Bolivia of typical Tiahuanaco cut stone, upright or jointed
technique, yielding equally typical Tiahuanaco ceramics and artifacts.3
However, Tiahuanaco is distinctly a ceremonial site, composed of an
aggregation of temples. It is not, therefore, particularly surprising that
the distribution of Tiahuanaco culture throughout Bolivia must be
sought by habitation rather than temple sites.
Tiahuanaco ceramics, as contrasted to buildings, have a wide, but by
no means regular, distribution in Bolivia. Early Tiahuanaco style is
found, according to my present knowledge, on the Islands of Titicaca
and Coati and at a site called Tciripa on the lake shore north of Guaqui.
Bandelier's collection from Kea Kollu Chico on the Island of Titicaca
contains Early Tiahuanaco fragments associated with a high proportion
of plain ware, which he designated "Chullpa," but, significantly, without any Inca sherds. This site was badly jumbled with skeletons and
refuse and was associated with no buildings. Of the pottery he writes4:The pottery was mostly of the coarser type, but we obtained several gaudily
painted specimens with plastic decoration recalling some previously secured at
Tiahuanaco.
'Cf. Means, 1931, 172, note 28.
2Means, 1917, 325.
3The 1934 Expedition has discovered two Tiahuanaco type ruins. One at Pajchiri on Cumana
Island with cut sandstone steps and upright blocks, and another at Lucurmata on the shore of Titicaca
between Guaqui and Chililaya. The latter promises to be a Classic Tiahuanaco lava megalithic enclosure with all ramifications. Also at Tciripa, on the lake shore, Decadent Tiahuanaco has been associated
with an upright facing slab building technique.
4Bandelier, 1910, 173.

1934.1

Bennett, Tiahuanaco Excavations.

481

This pottery has the shiny red, black, yellow, and orange colors on a plain,
unslipped background. The zigzag design is common. Zoomorphic
figures identical with the Tiahuanaco specimens are painted in colors on a
black background. The flat and hollow-base incense bowls with both
plain and fretted rims are typical shapes, and some have the modeled
puma heads on the rims. An incised incense bowl occurs, and one with
the frets pierced, as in one form of Early horizontal handle. Globular
shapes, one with a tripod knob base, are found, but not the long neck
decanter shape. The horizontal handled dish and the spittoon-shaped
bowl are likewise missing. The typical clay buttons are found, however.
The "Chullpa" ware associated with this Early Tiahuanaco style is
composed largely of one- and two-handled plain pitchers and shallow,
flat-bottomed dishes with straight sides and one or two flat grooved rim
lugs. This type is not the common Chullpa ware found all over highland
Bolivia. No further information as to the associations of the Early
Tiahuanaco ware found on Coati and at Tciripa is as yet available.
The Classic Tiahuanaco ware, according to present information,
is limited in distribution to the immediate vicinity of the ruins. Even
the nearby Islands of Titicaca and Coati do not have pure Classic Tiahuanaco pottery shapes and design. The collections of Bandelier from
many sites in highland Bolivia contain no Classic Tiahuanaco material.
It is possible that future work will reveal more sites.'
The Decadent Tiahuanaco ware is found in pure and derived forms
throughout much of Bolivia. Some of the sites around the lake are'the
Islands of Titicaca, Coati, and Cumana, the Peninsulas of Huata and
Copacabana, Tciripa, Chililaya, Pucarani, Pelechuco, Charasani, and
Sillustani. Around La Paz are several sites, such as Llogheta, HankoHanko, Poto-Poto. The Cochabamba region, Mojos, and northeast
Bolivia also furnish ware'of Decadent Tiahuanaco style. On the Island
of Titicaca the Decadent Tiahuanaco style was found in a site which
contained neither Inca nor Early Tiahuanaco styles. Likewise, at
Charasani certain cists contained Decadent Tiahuanaco ware mixed
with a plain ware, but no Inca ware. Other cists contained a plain ware
with neither Decadent Tiahuanaco nor Inca. In other words, the
Decadent Tiahuanaco is fairly well isolated as a distinct style in sites
other than Tiahuanaco itself, where it is a continuation of Classic. The
Cochabamba style, while containing Decadent Tiahuanaco elements,
has a freshness of treatment, and a variety of design style, which suggests
a secondary center or other influences. The Mojos and Northeast
'In 1934 Classic ceramics have been seen at Tciripa and on the Island of Pariti.

482

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

Bolivia style, as described by Nordenski6ld1 and Posnansky2 is a derived

rather than a Decadent Tiahuanaco style, and, while the kero-shaped
goblet and the small, flaring-sided bowl (Tiahuanaco Shapes A and E)
are found, a great many new shapes are encorporated, including such nonTiahuanaco forms as solid and hollow tripod leg bowls. Indeed the
Cochabamba style may turn out to be a link between Tiahuanaco and
Mojos, with influences from both sides. The possibility of subdividing
the Decadent Tiahuanaco into phases was previously mentioned. The
distributional evidence and the differences presented by Tiahuanacolike materials from Bolivia give still further bases for future subdivisions.
Unfortunately, the distribution of stone sculpture is very poorly
known. The style Group I has an interesting Andean distribution which
will be mentioned in the next section. A well modeled head with snake
headband in the Berlin Museum fur Volkerkunde came from Puno,
Peru, and is probably of Group I style. Group II, the Classic monolithic
statue, has only been found at the Tiahuanaco site. Bandelier found
crude carved uprights on the Peninsula of Huata, and a boulder-stone
head on the Island of Coati. From his description I judge these carvings
correspond to the crude simplicity of Group II1.3 The slabs found by
Squier at Hatuncolla correspond in part to the geometric designs of
Group IV. It is interesting to note that he reports a belief that these
were used in an Inca structure.4 Schmidt5 illustrated stone heads from
Ancoraimes, Titicachi, and Sampaya.
In spite of the limitation of knowledge of these distributions, it is
interesting to note that the Classic style of building, stone sculpture, and
ceramic shapes and designs are concentrated at the actual Tiahuanaco
site.
In Bolivia, in the highlands, both Inca and Chullpa sites are quite
distinct, both from each other and from Tiahuanaco. The term " Chullpa" has been loosely applied to all plain, unidentified pottery. It has
already been shown that the ware called "Chullpa" by Bandelier from
the Island of Titicaca, can be divided into that associated with Early
Tiahuanaco and that without association. While much of the "Chullpa" ware may be post-Inca, there is no logical reason why some of it
could not be much older. The ware found by Bandelier around Illimani, in such sites as Chichillani, Kokani, etc., is quite distinctive.
'Nordenskiold, 1924; 1917, 17-18, etc.

2Posnansky, 1922, 195.

3The Argentine Mission to Bolivia in 1933 found two statues at Mocachi, Copacabana, one of early
style (Group I) seated figures and another almost identical to the small bearded statue of Pit VII (No.
24), with beard, spread hands, and serpent figure on the sides.
4Squier, 385-386.
5Schmidt, 453, Figs. 1, 2, 3.

1934.1

Bennett, Tiahuanaco Excavations.

483

The commonest shape is a shallow open bowl with a constricted flat base
and diverging straight sides with plain rims. A duck-shaped vessel, a
large olla, and a globular, high collar bowl with two side handles and a
truncated base (like an imitation aryballoid) are other typical forms.
Bowls are either plain or simply decorated with black line design. The
decorated areas are commonly on the inside of the open bowls and consist
of simple elements repeated several times. This ware is much more
closely related to Inca than to Tiahuanaco types and probably is the
creation of local Indians under Inca political control. Still it might
possibly be the incipient Inca style. Future study will reveal the Chullpa divisions and their chronological relations.
The distribution of Tiahuanaco culture and influence throughout
the Andean Area is a complex problem, restricted on all sides by lack of
definite information. Means' gives a good summary of the present
status of the problem, and Uhle, Tello, and Kroeber have contributed
much detailed information. The purpose of this section is to sum up the
problem as seen from Tiahuanaco itself. A detailed and thorough study
of Tiahuanaco distributions is outside the scope of this paper.
At Tiahuanaco the archaeological evidence consists of building
techniques, stone sculpture, ceramics (with their shapes, designs, and
colors), textiles (by indirect evidence), metal working, and small artifacts. All of these sub-sections have not as yet been thoroughly integrated, but all are included in the general term "Tiahuanaco style."
Tiahuanaco style is distinct, particularly when contrasted with the
coastal styles of Nazca, Lima, and Chimu, and the late style of the Inca.
The distribution of the Tiahuanaco style can be traced, regardless of the
medium of expression. However, within the style classification, stone
working, including building and sculpturing, and perhaps metal working
has, in general, highland distributions, while ceramics and textiles have
mostly coast affiliations. This division can be partly explained by the
distribution and preservation of materials-stone is more available in
the highlands, pottery and textiles are preserved better on the coast.
Information is particularly lacking on the distribution of Tiahuanaco building technique. As Means2 has stated, the northern wall of
Sacsahuaman, and some walls of Ollantaytambo, near Cuzco, are probably pre-Inca and of Ti-huanaco type. The pyramid at Vilcashuamatn
is likewise probably related.3 Tello4 describes terraced, truncated
'Means, 1931, Chapter IV.
2Means, 1931, 137.
3Means, 1931, 109.
'Tello, 1928, 272.

484

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

pyramids in the Huaraz region of northern Peru, and also sacred corrals
of which he writes:.

they are formed of great stones, planted vertically and arranged in rows in

the same style as the enclosure of Kalasasaya in Tiahuanaco.'

If building technique were analyzed into such elements as stepped

pyramids, sunken courts, megalithic upright enclosures, subterranean
rooms, jointed blocks, association of statues with buildings, carved
stones used in building, etc., a great many parallels could be drawn between the Chavin-Huaraz section of northern Peru and Tiahuanaco.
Furthermore, some of the adobe structures on the coast may eventually
be associated with Tiahuanaco style, and Means2 says that the Tiahuanaco II period modified clay-ball architecture of the Coast into
various size bricks.
Stone sculpture was highly developed both in the northern Andean
highlands, around Chavin and Huaraz, and in the southern highlands at
Tiahuanaco. Numerous resemblances between the two regions have
long been noted. The work of Julio C. Tello in the northern section has
furnished material for even more detailed studies. A common type of
seated male and female statue found in Aija, Huaraz is described by
Tello.3 The fairly realistic heads with rounded cheeks, projected chins,
modeled lips, oval or round eyes, crescent ears, simple headbands with
decoration, and the designated hair down the back, all resemble the
statues of Group I (Style 5) at Tiahuanaco. The Huaraz group is more
ornate, with decorated shields, representation of textile patterns, trophy
heads, etc., and is, as a whole, more formalized than the Tiahuanaco

correspondents.
Sculptured puma heads with block attachments for insertion in
walls are also found in both regions. Concerninlg the distribution of these
Tello writes:
Heads of cats . . . are found in abundance in nearly all the North Andean
ruins; and in Tiawanako, Cuzco, Huanaco Viego and other archaeological centers of
the highlands.4

Of all of these the Tiahuanaco forms are the simplest, the Chavin the
most ornate.

Door lintels with relief designs are common in the Huaraz region.
As described by Tello5 the design is generally composed of a seated
human figure with a cat on each side. The cats have front-view faces
1Tello, 1928, 279.
2Means, 1917, 328.
3Tello, 1923,237-241; and 1928, 279-281.
4Tello, 1923, 257.
'Tello, 1923, 231.

1934.]

Bennett, Tiahuanaco Excavations.

485

with loop ears and profile bodies with curled tails. At Tiahuanaco the
two cats in relief on the base of the small statue found in Pit VII are
almost identical with the northern type, having front-view faces with
loop ears and profile bodies with curled tails. Other squared stones with
relief carving of profile pumas at Tiahuanaco might well have served as
lintels.
The elaborately carved Chavfn obelisks and monoliths are described
in detail by Tello.1 Markham2 made a comparison between the Tiahuanaco and Chavin monoliths and the idea that both types are elaborations of a basic, widespread culture has been confirmed many times since.
As Means3 expresses the relationship, both sets of sculpturing are " derived
from the same fundamental concept." The similarity lies in the style
of cutting and treatment rather than in design details. The problem then
arises as to whether Chavin is, historically speaking, influenced by Tiahuanaco culture or whether the reverse is true. Tello thinks that the
Chavin culture is more basic. Means believes that the Chavin style is
more mature, conventionalized rather than realistic, and composed of
elements which could only be assembled by a combination of traits from
Tiahuanaco culture with those from the coast cultures of Chimu and
Nazca. The time required for such a spread accounts for the greater
maturity of style at Chavin, as contrasted with the Tiahuanaco prototype.4
Not only in the monoliths, but also in the other sculpturing styles
mentioned above, the Chavin forms are more elaborate than the Tiahuanaco. On purely stylistic grounds Means' analysis seems sound. His
theory also accounts for the notable gap between the two regions, namely
in the central highlands around Cuzco. Furthermore, the spread of Tiahuanaco culture is well established by the ceramics and textiles of the
coast, while Chavin style, although it has influenced some of the northern
coast, has certainly not had the same wide distribution. Unfortunately,
until more work is done to establish direct or implied stratification of
materials, and until more is known about distributions, the stylistic
analysis must stand on its own merits.
Ceramics and textiles furnish little evidence of Tiahuanaco influence
in the highland regions. A vessel in the Berlin museum,5 said to come
from Cuzco, is in the Tiahuanaco style. Means6 also states that Tello
'Tello, 1923.
2Markham, 1910, 389.
3Means, 1931, 138.
4Means, 1931, 143-144.
5Cf. Means 1931, 137.
SMeans, 1931, 145.

486

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

found painted cloth with Tiahuanaco influence in Huancabamba, in

northern Peru. On the whole, however, the coast is a much more fertile
field for the tracing of Tiahuanaco style in ceramics and textiles.
The distribution of Tiahuanaco style and the Decadent or derived
Tiahuanaco style (called Epigonal) on the Peruvian coast can best be
summarized by Kroeber's' statement:
Tiahuanaco and Tiahuanacoid (Epigonal) era. Probably to be included are:
Tiahuanaco (at that site); Chullpa ware of the Titicaca region; sporadic occurrences
from the Nazea area; Ica Epigonal; Pachacamac Tiahuanaco and Epigonal; ProtoLima from Nieveria-Cajamarquilla; Middle Ancon I; and Epigonal ware at Chancay
and Supe; part or all of the site A ware at Moche.

Tiahuanaco style has had the strongest influence at Pachacamac and the
Central Coast sites, where it is the earliest culture found (with the
possible exceptions of the shell mound material from Ancon and Puerto
de Supe). Uhle, in his work at Pachacamac2 established the stratigraphic
position of the Tiahuanaco style and also divided the material into a
Tiahuanaco and derived, or Epigonal style. Unfortunately, both styles
were found in the same graves. This division of the Tiahuanaco style,
always in grave association, has been found at most of the sites of the
Central Coast (such as Supe, Ancon, Nieveria), at Moche in the Trujillo region, and as far south as Nazea. Olson in his field notes of the
Myron I. Granger Peruvian expedition of the Museum reports the finding
of a "pure" Tiahuanaco style in the valley of Nazea. The Tiahuanaco
style is not found in abundance and never isolated from the Epigonal.
The Epigonal style is, on the contrary, well represented in the collections. In its northern distribution it divides (or encorporates) into a redwhite-black geometric, a pressed relief ware, and a tripod, cursive style;
while in the southern distribution it includes a red-white-black textile
pattern ware. North of the TruLjillo region, according to Kroeber3
"indirect Tiahuanacoid and Central Peruvian influence was carried to
the northern-most area of Piura by Late Chimu."
The Tiahuanaco style textiles have the same distribution and the
same dual division into pure and Epigonal as the pottery. To quote
O'Neale and Kroeber:4. . .Tiahuanacoid Chimu and Early (Middle) Lima can be equated roughly
with Epigonal Nazea-Ica in time; and like it they show a preponderance of tapestries,
the decay of embroidery, a persistence of knitting, and flat braid.

The Tiahuanaco style ware on the coast is well finished, with five or
six color designs. The colors are black, white, yellow, gray-on-red as in
'Kroeber, 1925a, 232.
2Uhle, 1903, Chapter X.
3Kroeber, 1925a, 229.
41930, 35.

1934.]

Bennett, Tiahuanaco Excavations.

487

the Classic Tiahuanaco highland style. Condor, puma, and human

designs are used. Innumerable design details are exact parallels of the
highland prototype. In the textiles the designs are even closer to those
found on the Tiahuanaco highland stone carving. Textiles such as those
illustrated by Uhle1 and Reiss and Sttbel2 have designs closely similar to
the running figures on the Gateway of the Sun. On the coast textiles the
filling of squares with figures is also like the Gateway of the Sun treatment. The figure hands have the Viracocha-like three fingers and thumb.
Many other parallel details could be shown. The cloth utilizes tans,
browns, red, orange, and some blue. All these resemblances allowed
Uhle3 to refer to this material as definitely derived from the great Tiahuanaco, including recognizable designs, although already decadent.
However, particularly in reference to the ceramics, Uhle himself was not
quite convinced of the absolute identifications of the coast style with
highland Tiahuanaco. After listing a number of differences between the
two regions he writes:All these details taken together help to prove that, in addition to the district of
Lake Titicaca, there must have been another important locality to which this style
was peculiar. The specimens from Pachacamac seem to come mainly from the
latter, as they were not produced at Pachacamac and only a few of them might possibly have been brought from the shores of the Titicaca.4

Kroeber too cautions the hasty identification of coast style and highland
Tiahuanaco:
. . . it is well to remember that in spite of its resemblances to the non-Inca ware
from the Titicaca region, it differs from this. It has for instance, forms apparently
never reported from the Titicaca area the double spout, bird and spout, jar with
tapering face spout-besides numerous differences in designs.5

From the point of view of highland Tiahuanaco the divergence of the

coast ceramics is even more marked. None of the Classic Tiahuanaco
shapes are found on the coast except straight-sided varieties of the keroshaped goblet and the flat-bottomed, shallow open bowl, varieties
furthermore, which do not occur at Tiahuanaco itself where both shapes
have flaring sides. The curled tail puma, the Classic crowned condor, the
typical form of the step design, the human warrior figure, and other
typical Classic designs do not occur on coast pottery. Unfortunately,
textiles are not preserved at Tiahuanaco. It has been already suggested,
however, that the design work on the monolithic statues and perhaps
even on the Gateway of the Sun are derivations of textile designs.
'Uhle, 1903, Chapter X and plates.
3Uhle, 1912,323.
4Uhle, 1903, 22.
5Kroeber, 1926a, 273.

2VolI. II, PI. 49.

488

Anthropological Papers American Museum of Natural History. [Vol. XXXIV

Furthermore, it has been shown that a general division can be made

between the stone carving designs (including by implication the textiles)
and ceramic designs. Since the coast textiles show very close Tiahuanaco
affinities, it is possible to suggest that the Tiahuanaco style was largely
carried to the coast through the medium of textiles, the designs of which
were utilized for decorative elements on the existing coastal pottery
shapes. That textile designs were applied to ceramics is clearly demonstrated by the Kroeber and Strong' red-white-black textile pattern ware at
Ica. As an example of such a transfer I cite the following comparison.
The Viracocha figure of the Gateway of the Sun is found on a textile
from Pachacamac.2 The Viracocha figure does not occur as a typical
pottery design at Tiahuanaco. However, several large U-shaped bowls
from Pacheco on the coast of Peru have this same figure,3 and simplifications of this figure are typical decorations in the Epigonal period. Other
illustrations could be given, but at this time I merely wish to point out
this possible explanation of some of the ceramic differences between Tiahuanaco highland and coastal style, without suggesting that the explanation is the final or exclusive one.
The Epigonal style of the coast is generally inferior to the coast
Tiahuanaco. Fewer colors are used, the surface is soft and has little
polish, the designs become stiff, and tend towards the geometric and
curvilinear. Design subject becomes broken into elements, although the
human face and front-view figure with projecting trident feathers are
typical. At Supe, Kroeber4 lists the typical shapes as goblets, double
spouts, low-plate bowls, skull, and cat-head jars. Local styles, such as
the pressed relief decorated ware of the northern coast, and the redwhite-black textile pattern of the southern coast, originate in various
regions. In textiles, too, geometric designs and stylized heads are found
in elaborate tapestries. The colors are more or less the same.
Epigonal and Tiahuanaco style are stratigraphically contemporaneous on the coast. The Tiahuanaco has been considered the older on
stylistic grounds alone. That this might not be the true picture was
suggested by Kroeber and Strong at Ica5 who considered the Middle
Ica I style, chronologically determined as later than Epigonal, to present
more similarities to Tiahuanaco style. They suggest that at Ica:
. . .

Epigonal came first of the three, and that Tiahuanaco and Middle Ica re-

presented local developments-perhaps synchronous ones-out of the wide-spread and

'Kroeber and Strong, 1924, 113-115.
2See Uhle, 1903, P1. 4.
3See Yacovleff, 77.
4Kroeber, 1925b, 241.
5Kroeber and Strong, 1924, 118 and 120.

1934.1

Bennett, Tiahuanaco Excavations.

489

simple Epigonal style in an era of upward swing of culture. In this case Ica Epigonal
would still be truly epigonal; but to Proto-Nazca and other pre-Tiahuanaco cultures
instead of to Tiahuanaco.'

However, the textile sequence for the same region established by O'Neale
and Kroeber2 shows that Nazca-Ica Epigonal ties up with general
Epigonal, while Middle Ica is a more formalized trend towards Late
Ica.
Epigonal, while probably still influenced by highland Tiahuanaco
as Uhle suggests,3 represents a local coastal decadence as well. There are
many differences between Decadent Tiahuanaco of the highlands and the
Epigonal. It is quite possible that some of the new forms which appear
in the Decadent Tiahuanaco represent a back-wash from the coast, as
Kroeber has suggested in correspondence. In general analysis the
deterioration follows more or less parallel lines in both highland and
coast cultures. There is the same reduction in technical skill, number of
colors, polish of ware. Realistic figures are reduced to parts. Geometric tendencies prevail in design application. However, more detailed comparisons are necessary before the inter-relationships can be
determined.
In Chile two Tiahuanaco-influenced styles are also found. The
northern coast and the interior is more directly influenced by highland
Tiahuanaco. The earliest pottery found is in the Tiahuanaco style,
although this chronology is not based on any stratigraphy. The pottery
is decorated largely with geometrical designs, according to Latcham,4
with none of the anthropomorphic and animal figures of the Gateway of
the Sun. These latter are found on cloth and wood carving, however.
Latcham5 considers the Tiahuanaco influence as derived from the
Decadent phase. The Epigonal in the southern part, as described by
Uhle6 is characterized by the suppression of original figurative elements
and the reduction of these to geometric elements, among which the step
design dominates. The step or row of triangles as a design is characteristic of Atacamenio style, and Uhle7 thinks that it has influenced the later
Tiahuanaco style. At least the most typical shape of the Atacamefio
painted ware a high collar, bulging-sided, two-side handled jug, which
narrows to a flat base-as well as its step pattern decoration is one of the
new shapes found in the Decadent Tiahuanaco highland period. Future
'Kroeber and Strong, 1924, 118.
21930, 42.
'Uhle, 1912, 323.

4Latcham, 1928, 67.

5Latcham, 1928, 69.
SUhle,
1919, 25.
7Uhle, 1919, 24.

490

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

work may reveal some connection between the Epigonal of the Peruvian
and Chilean coasts. Uhlel found a Tiahuanaco and a Decadent style at
Arequipa, Peru, which suggests that the highland influence spread all
along the coast.
In Ecuador, to summarize Means2, Tiahuanaco style can be seen in
the carving on the edges of the stone seats and carved slabs of Manabi,
and on a puma-urn found at Caraques by Saville. Their carving is like
the style of Hatuncolla, as described by Squier, or Group IV of the Tiahuanaco stone sculpture classification. On the coast of Ecuador a
ceramic period (perhaps the Caras) which grades from archaic to Tiahuanaco II is recognized. This sequence is repeated in the highland.
Gold objects from Patecte, Sigsig, and Chordeleg (after Saville) show
traces of Tiahuanaco II style.
Uhle3 arranged the Argentine archaeology into four periods. The
first is a primitive or archaic period. The second is called " Draconian'
and is characterized by an engraved and painted ware. The incising on
black ware is much like Tiahuanaco in style and details. The third
period is designated "Pre-Inca Calchaqui" and has painted wares like
Decadent Tiahuanaco. Wooden tablets compare with stone tablets of
Tiahuanaco. The final period is Incaic. A more detailed study will
reveal many parallels between Tiahuanaco and Argentine archaeology.
The Tiahuanaco site seems to me to be the center, perhaps chiefly a
ceremonial center, of a higher culture which developed out of the general
Andean culture. At the Tiahuanaco site certain techniques of building,
stone carving, and pottery manufacture were developed locally. Some of
the techniques spread, but, more definitely, the Tiahuanaco style spread
throughout much of Bolivia, Peru, Chile, and Argentine. On this point
Kroeber writes:
In short, except perhaps in its presumptive immediate homeland on the Bolivian
plateau, the Tiahuanaco style nowhere appears alone but is regularly associated with
the supposedly derivative Epigonal or with local styles or with both. On the other
hand, it is the one style other than the Inca which is found over almost all Peru.4

Tiahuanaco should be considered as one manifestation of a general highland culture. The Chavin style might be considered another manifestation of that highland culture, perhaps counter-influenced by Tiahuanaco,
which may have reached its specialized development at an earlier stage.
But Chavin cannot be accounted for completely as an offshoot of Tiahuanaco. Kroeber, in reference to the middle periods of north Peru,
writes:'Uhle, 1912, 324.
2Means, 1931, 164-168.

3Uhie, 1910, 518.

4Kroeber, 1925a, 212.

1934.]

Bennett, Tiahuanaco Excavations.

491

There can be little doubt that the bulk of the stylistic elements is of highland
rather than coastal origin, but highland from Ecuador to Bolivia rather than of one
Peruvian district; the Classic Tiahuanaco manner is represented only in a small
minority of pieces.

Styles are often localized in the Andean region, but the problems of
Peru cannot be solved from one locality. The earliest periods on the
coast used wool in textiles, which means that highland coastal contacts
were already established, because the llamas live only in the highlands.
Studies must therefore allow for considerable interplay of cultures. One
hopes that the days of treasure hunting and wild subjective speculation
on history have passed.

'Kroeber, 1926b, 111.

BIBLIOGRAPHY
BANDELIER, ADOLPH F.
1910
The Islands of Titicaca and Koati. New York, 1910.
BINGHAM, HIRAM
1915
Types of Machu Picchu Pottery (American Anthropologist, n.s.
vol. 17, pp. 257-271, Lancaster, 1915).
COURTY, GEORGES

See Cr6qui-Montfort, Count G. de

CRP,QUI-MONTFORT, COUNT G. DE
Fouilles de la Mission Scientifique frangaise a Tiahuanaco (Ver1906
handlungen des XIV Internationalen Amerikanisten Kongresses, Stuttgart, part 2, pp. 531-551, Stuttgart, 1906).
D'ORBIGNY, ALCIDE
1844
Voyage dans l'Amerique M6ridionale, Tome III, Partie I, Partie
Historique. Paris, 1844.
KROEBER, A. L.
The Uhle Pottery Collections from Moche (University of Cali1925a
fornia Publications in American Archaeology and Ethnology,
vol. 21, no. 5, pp. 191-234, Berkeley, 1925).
The Uhle Pottery Collections from Supe (University of California
1925b
Publications in American Archaeology and Ethnology, vol.
21, no. 6, pp. 235-264, Berkeley, 1925).
The Uhle Pottery Collections from Chancay (University of Cali1926a
fornia Publications in American Archaeology and Ethnology,
vol. 21, no. 7, pp. 265-304, Berkeley, 1926).
1926b
Archaeological Explorations in Peru. Part I. Ancient Pottery
from Trujillo (Field Museum of Natural History, Anthropology, Memoirs. vol. 2, no. 1, Chicago, 1926).
KROEBER, A. L. AND STRONG, WILLIAM DUNCAN
The Uhle Pottery Collections from Ica (University of California
1924
Publications in American Archaeology and Ethnology, vol.
21, no. 3, pp. 95-133, Berkeley, 1924).
LATCHAM, RICARDO E.
La Alfareria Indigena Chilena (Publicase a Expensas de la Comisi6n
1928
Oficial Organizadora de la Concurrencia de Chile a la Exposici6n Ibero-Americana de Sevilla. Santiago de Chile, 1928).
LEHMANN, WALTER (assisted by DOERING, HEINRICH)
1924
The Art of Old Peru (Publication of the Ethnological Institute of
the Ethnographical Museum, Berlin). New York, 1924.
MARKHAM, CLEMENTS R.
A Comparison of the Ancient Peruvian Carvings and the Stones of
1910
Tiahuanacu and Chavin (Verhandlungen des XVI Internationalen Amerikanisten-Kongresses, Wien, Erste Halfte,
pp. 389-394, Wien und Leipzig, 1910).
MEANS, PHILIP AINSWORTH
A Survey of Ancient Peruvian Art (Transactions of the Connecticut
1917
Academy of Arts and Sciences, vol. 21, pp. 315-442, 1917).
1931
Ancient Civilizations of the Andes. New York and London, 1931.
492

1934.]

Bennett, Tiahuanaco Excavations.

493

NORDENSKI6LD, ERLAND
Die 6stliche Ausbreitung der Tiahuanaco-kultur in Bolivien und
1917
ihr Verhialtnis zur Aruak-kultur in Mojos (Zeitschrift fur
Ethnologie, vol. 49, pp. 10-20, Berlin, 1917).
1924
Forschungen und Abenteuer in Sudamerika. Stuttgart, 1924.
O'NEALE, LILA M. AND KROEBER, A. L.
1930
Textile Periods in Ancient Peru (University of California Publications in American Archaeology and Ethnology, vol. 28, no. 2,
pp. 23-56, Berkeley, 1930).
POSNANSKY, ARTHUR
1911
Tihuanacu y la Civilizaci6n Prehist6rica en el Altiplano Andino.
La Paz, 1911.
1912
Gufa General Ilustrada para la Investigaci6n de los Monumentos
Prehist6ricos de Tihuanacu 6 Islas del Sol y la Luna (Titicaca y Koaty) con breves apuntes sobre los Chullpas, Urus y
escritura antigua de los aborigenas del Altiplano andino. La
Paz, 1912.
Una Falsa Critica de Max Uhle. Berlin, 1913.
1913
1914
Eine Praehistorische Metropole in Sudamerika (Una Metr6poli
Prehist6rica en la America del Sud). Berlin, 1914.
1922
Breves Noticias de una Rama Cultural Tihuanacu al Noroeste de
Bolivia (Annaes do XX Congresso Internacional de Americanistas, Rio de Janeiro, vol. 2, segunda parte, pp. 195-197,
Rio de Janeiro, 1928).
REISS, J. AND STtBEL, A.
1880-1887 The Necropolis of Ancon in Peru. A Contribution to our Knowledge of the Culture and Industries of the Empire of the Incas,
being the Results of Excavations made on the Spot. Translated by Professor A. H. Keane. 3 vols. Berlin, 1880-1887.
RIVERO, MARIANO EDWARD AND VON TSCHUDI, JOHN JAMES
1853
Peruvian Antiquities. Translated into English, from the original
Spanish by Francis L. Hawks. New York, 1853.
SCHMIDT, MAX
1929
Kunst und Kultur von Peru. Berlin, 1929.
SQUIER, E. GEORGE
1877
Peru. Incidents of Travel and Exploration in the Land of the
Incas. New York, 1877.
TELLO, JULIO C.
1923
Wira-Kocha (Inca, vol. 1, pp. 94-110, Lima, 1923).
1930
Andean Civilization: Some Problems of Peruvian Archaeology
(Proceedings, Twenty-Third International Congress of Americanists, New York, pp. 259-290, New York, 1930).
VON TSCHUDI, JOHANN JAKOB
1869
Reisen durch Sudamerika. 5 vols., vol. 5, Leipzig, 1866-1869.
UHLE, MAX
1903
Pachacamac (University of Pennsylvania, Department of Archaeology). Philadelphia, 1903.

494

Anthropological Papers American Museum of Natural History. [Vol. XXXIV,

UHLE, MAX
1912a

Los Origines de los Incas (Actas del XVII Congreso Internacional

de Americanistas, Buenos Aires, vol. 1, pp. 302-354, Buenos
Aires, 1912).
La Relaciones Prehistoricas entre el Per6 y la Argentina (Actas del
1912b
XVII Congreso Internacional de Americanistas, Buenos Aires,
vol. 1, pp. 509-540, Buenos Aires, 1912).
La Arqueologia de Arica y Tacna (Boletin de la Sociedad Ecua1919
toriana de Estudios Historicos Americanos, vol. 3, no. 7,
pp. 1-48, July-Oct., 1919).
UHLE, MAX AND STtBEL, A.
1892
Die Ruinenstaette von Tiahuanaco im Hochlande des Alten Peru.
Breslau, 1892.
YACOVLEFF, EUGENIO
1932
Las Falconidas en el Arte y en las Creencias de los Antiguos
Peruanos (Revista del Museo Nacional, vol. 1, no. 1, pp. 35III, Lima; 1932).

I-. .,_'.T;l;--41,

.a,L-_

,,,1-,.
I- ,--,..,I-1 ", I;-1..,, -'_,. ,'_-`l&-k,"_.-,
I
l
z
.
_
I1,,,,
._ - .-_,,'y:1..j ,-.-,- `.,;_ :'# `,.-

z
W
,
.
"
_
-I-.,.
. I-,- "-j-,_,_l'-, - _.*x
-, ,.1.l -",_ .',-_ i. _,'
.---,'.-__. '_ -1-;.

_--.
',.- ,,,
-- -..- ,Z" ..,W-,'--.
!..
,,-,.- 11.--.6z
.-'._T.._L__
','-,.-,"
.-'._T.._L__
ri-"n
__ ,,,,
-4 .7..,".-W-,--.
.,
.,
._ y ., -,_. .,-,-,--,
1,1.,4.,,,-.*'7p
.,I.11N
_._, ,L
.,_,.-,-,.
,--,-.
.-..."
-`,.,,,--,,.r
.,
-'-- -,.,,II..,
..-..
__,,
.k' :,,._..',._l'A"',
I',,
I,'--..,. z" _,..- A-,
".'I,
I--- ll.
i-."
11 ,_,--,.,-,
.._-__---z
-.ll.
_. . lp.
.-,.".--7.',
.t,
,,,
.r,4. ._ -1,
-',..
17.
,,
-,
,., i..-_,.., "', 1,-_r_.___;,.._-;,
.-A,
,I-._L_..
,., , '_..,:,"-,,-,
-_.. ,"-.,i.,.
`. ,.-1).-.1,
,,
_!:. ,,.K
,;,,--"
't".. -_i-'l-.,_-,,
-c ,',.
-., -, -,
;;;.-,', j,,.,.,._,..t,,
.,,
I-----11I`Z.l,
-.,
-, ,,
-'--7.,-'_-.
",.
-,,
...., , -",
-.."
',.
..
__.,
-,
...
.l._._,
,.:
4.
I.-11
V,-,..
*,
.,I,.11
-_-,.
--,,.-,
41,
-V,-,.,
._e-17,.,.'
,,,.._,,....,.
._...-..
_--..-.;.--..--w;l,.;r. 2_..
..,.,
-,
-- ..-,_...i,
.-l
,.-..
I--i
.'A...-,-.-,-j,.i, ,:;.-...I--,:
,1,
,.
.4,r..-,
..t
.,_.__,:
_-..
--.n_---1.,
.. .._,
,,,.-.
-1 i.4-_
.. /,;.,.Q..l I., '11.i,,.' .. r,,.tP-,'-,-_
-.,.
,,".
_.,.._A,--zt
,.I"
.,
4j5 -:.,
-.
- I' .,...
.7.-'ti-.11
-,-,
-.-v--,,
" ., .- .--.-."-.,-.i
.l ,,,,--..,,--,
,,
1- -.,.,-.,
..,,-."'.---".,
-.,_....
-..,I-,..-,,,.
,,.',..
.;".,-,
It".,...,.
,.,,-.,-, ,- $'
,--.",
..".11.
j, '.zi 1-,I
'k,
1-I.----_',
,,.,
.L_,.
", .,i-_., .",..,..,.c,
_11, .- ,.,
_:.
..
,,,,,-...
-.-,!n.!I-t,
-,,
.,', I .'.-,,:n-!..,
.,...,.
-.,.e
--;,,--,".
,.-. x.
,".._-!..,_.-I
._,I,_
i, C,..,
---, -._-,.-'y
Ni6,
.,.
i.s. ,4',.7"
i.s.
...
.__
,;.
.--4
-,-".'.!!..,
.z
",-;.,.'.
.
7
-.'.',-. . . ..
,% 'r .,_I,,,-3c
Ii.._1-I1,-...,;..",-;
.,,,;
. ,%Ir-;... '. ..,.1 4z
.,K,':.,..-- t, .. ...,.:-,'-_-..,".,-._,j7,
P',.,
---,,.,
!,,,....,-,,'.-,.
"1.-,
-,_.,--I.
"_,..
--.'...
,.. ,4- ..,
,
.,-'l
.1 .11 "-1.1-1
., "., _; ,.:-,' -"-7,- ..I.,
; -,,4 _; ,.:-,' -'-,-7,
.., .*,
I.
,L., 4;..,-.- .'-,!_;. ';._,...,oic,.,.-_.
:..-, -_.,.,_.
.1 ,,_,,.
"_,-,--,I-,
..,
,3,
-1 -;.:r
.". --.._.&
I'-',.,""'."
..
" -.6-111-1- 11 I-,1_,_..
I,-(;i:-,,
.:,.A-,;;.,,
.,,
I,
--i"'
-_.
,.I-.,
--"-,
` 'L _. _._.
I -(;--,.
i: ---,.,:.Z_-i.,
,:.:_t7--.,,;
-7.,,;l
.:..-, !, -4.
,,.
.:.
.,.",-,.
-'-,,.-, -:I11,.,
I" ..'!:,, 1: "__..
": ',_,
.',-_..
-;.k,
1,
l',Z?-A
.!. ",
-,. I.,.
'' -, _,z;-.,
-_.
.,,.-.'_.V,"Z-'.--_`.7,
.__.-.,_fL.
',---.
-4
,._,_11, '. .-.
,r ll,.&.I
., ...11"I.,
7 lll. ,,,
I.C'7111'.
al_,_ ..--;'-il 11 ,'k,
,..
,I,.__-,,-1--:.._14,..ll
,-- 41
-."7.
--- ,.,--,..III---..
-..-..
.;,_
.,:,.-,-x,,_;t'_
.7
.f'll.,.V.....
i:tl.I.:I
--. ;0,.,.. .,,
,.',.;..,
-r,
.,:- ri...--I-,
t,
-1
_."
-.
-'4.*
.- ".',
*..-.,-..l
I'll.' ;.,_,-.-.:;,..<,-;t
,:4..
-;,."11
... "-1
1..-,I
__,!.._
L7- -1,,-,
,1.ozl .131. -,-f __'.:
--,
-,- .i.-.-A 1,;..--."
-,.lj.,-,-_- ",,.,--,.,..
I,....,;.,_,`,N
.; -,`)."-,t,,A..
" -.,,,
.,."._,1.:-1---'- ",-,-.,;.,,--;. -, , -" ","- ,., .,.--i
.-,
,.--.
i -, 1 --4,A.',-.,
- ---,:
,..,
I.,_-.---.
,..
...,-1,,,_': .,.,,-,
1, .1.t-,- ._,-..
._;',,"_.)
--_.,,_"_.;;:__'
-;,.,_
I-..,%.J.,
-.;,-..
.,I...
_%,,
..',
_-.kI
Z;
....,.1__.
I---_,,.,,.._"
--1_-,""
_7.
?.
I-,.-..t.
!,-,..
- -_.,. .11.
1, .-,,
-i", r-,-.
;. -..,. .-,.!.`-,.-,',.,.k. !,'.-,._,"_'1_11, '. _". .,
11.",
-,... .,,.,
,%"' ',,; llrl.,-_,;,, .--__I.:,
T,. ,.",;.,
.-,S,
1.... I-,_,-__-_.-_.
,..`wk.
--,
-. -.,
'..,_";
.--.
.z .1 .1 ..'..
,.;-. I.,
..
-.
-, ..";.1.1:
"-Ah
,.,
,.,.I'.
,.---,;..,
..X_ -I;:, j-u-,""I. I -;----A
. ., -,k'.,,`,--.,
.,.7' ..-`
--,,__
-",- -,_
",-INIRI
).:f
-,._..- " ',::.
,, 7__..'.... "._.,.--.-,;.
-.,.---.,-_"-,-.--'.I.".l
-L
-" -" -x
,1'1_-.
-,
.,.,,-,1', '.`, -,-,,..,j,
.!"-.,
-, .;_-rl.i
.,"'..
','7'
-1 , .. ,,.,,
---.,,,,..,...
--V,.,-,-.
,.
-, ,.,';,,..
!,. ...11. --1_..,-- _,., --R,.
__",-_....,_
-. --.-._
.,
,.;J.._.-`.,',-..A,--..,
'',
..,.
.,-'.
I:.,
--, -1..._.-,.-,.--, ,.,
!;"
,,-- ",,',
:.--. -%-.-,.,-.
.-_I,.%-,-x
,;i..,....
-.,:.,-.--,."
,,-,..,.
.---,,:,
,,-----t
-,,--I
-z.
-,-,:,.
._.,.lr.x,
-" -.--,.
- -1..l ...
,...
-- ---4
".., .,-...
'.-,,.-I
.,
'I.
_-._i_
,,I
-,,.
_;. _--,.
"'_..' " "",-..-,
..,`,-.
..,
,-?a.,!,".
.,'a
;417,o,_.41
--_
,,
-.-,..)
1, ;e,
--,3.1,:,,
I,
!.,. _,
,i.,
.'. ,-7_.,,.,, .. ,.
'T"
,_+
-,
;,
-. ---'I"',
.,, ".'..";.- ",
-. -f,
,..,
-:'-"".
.1.."4.-,r
-,,
.' 3 -, .'*i._-, -.,.'1_7-,-,
:,_ ,3 -, *i", -.I., 1.-,
...I,..,,-.___
', y... -,-,__,-,
..,..
_. .,.--,,--_
-----,:.:.,-,
'.-",
-.,__..,.,,
I__-,
n__.
.'..
ex
.-4w,-'r
--.,
".
,,,.,.
",,.
,-_'.
.-",.
-.--,",
..
,. _,.-,,,
-.,,_; .,...
,.,..-.lt..,- --, ..,. .,-,."
_- -- -]L, 4.-,,..
-,-"
,,.
.__ -,_I ,1.
.1!1--,
1- .'-...--I---.--.-.,-:1,--- ,,,._
.. --.,,--.-;_.
._
..-, : ..'.I,r.,
,--'
1,
..
"_-,.!,.-...,
- - "k
....,1. -- ., ,-.---'t..
I,t,A.I,
-.....
_-.'- -., '. -_ _-.f,'- -- ",.
-_.. 1. -1, i.f --.
;j;"-,,-w,_-"
1- .- ..--,
,;". ,,1".-.#.
1_ ,,.-4, .,,,,
,,,-,0(
-,;.
-,-_
,.(l
1,.I
-1,.-.";---. t",.,,-..,17.-,k-.,
10 , ,;- .,,
I.
t;l.- ',,-,.,
. _. .l `1o,. j..--Z.-,;,.
...'k--.
I"t`.,,I-le
-.;.,.- "--".I-,..
..,__
'- t. ..-,..-r.. .i,a,,li.- ,'.-, -o,..,-1, ,'';. ;,,., ..'_
-Z_-',
,U.,I,
. ._.,-, 11,
-.
..
1.-.r.t.
..,,'-,.'4__
"
:1
..,
":.,O__.Z.
":.`
..--,
,-,-.
"
`,1,
-,7.'.
_- .I
7,.
.l_.';
_.k,...- .-,,-C
-,
k_;
,-, ,,.t' -, .1, ,-., -.,-,
.__,.,".,_
._
-7 _--.'_
.- ,,,..-.t,,4..-,.,.
_.l
,-'.
i'..
-..l,I,-,-.
,-.'.
-..-,
-"Q
-6
"-I,-.,
,..
,,..
..Z'
,'. ..zl
1':5.
.I,.,,.
1-/,.I. -.5-1,
.,l,,I
- ._:;.-...,.
-,
,.-11.
1". --;.,
.-";111
-,,,
,__-..,
"o,11.""
.-,..,,,..',-,
.,;
.t,p-,.,
";.`,
4r.,_
L..,.."--.,
.j
.,,.
'.-",-_ .-.-,
,,?'.
..`,
..,---,-.--- -,.';-..,I-.".,..,._
..."
".,:;"-,.,.,,.i.t
,.
_1-n,,_.-,.
,,..-' -.i
,.
-", ,I
,- ,,.
._4..
'.. .,,.,
._.":n.!_y,-,
:,
I.,.,
-l'-.
_;
1..L.it.:!:,..
.'
-1 ."
-.- -.--'
.-:t ...I,..-!..:"
` ".l ,--,,-.-,.._1-- ,I-.L
.":n.!_,"..l,_:.
-"C.-'C...-,
.,.
-, C$
_-,...--,-:
-,..
-,
i.s.
.-"__---1,
;,
1.
,,,j.'s.
-,.
.z
..,.,I.:z
.1 -,
r.,l
..
,%,r--.
1, 4---.
I%lr-.,
-,
.,'.,.,..I-'y
_.,,;-; -- -,Ip.,-,%;i.
,I!,I,,I,:.
.
,.
.-,-,,
I'r
-,,.
- --.,1..,.'l -..-,'!.
.,I..-..
--,- .,-'-.,---,
,,'4._.,,I
..'
,..11
..;,.
._".
`..,
`.,.,.,,-.--_
I,
", .,--,..:-.....;.z .`7'_.___
... /I
,.1
I'__.et._;._-,.,
...",,,1,
..,.."
'I
.."
,.:
I.
.,
....-I.,
1.1L.,
_.. I,.lll ,"I
,7,,,_
Z_;'r 1_. .,I.,
;"
_:..'l -, ',.,f.,.
,,,.,-I,
,- I-,
1....,T-6.
I-11._ .1-.-1.1
".,.w.,
.1,,-1., .`._.14`-,;.,,"-. ,,., .:t----,,--;.,-,k-,;_;-_--ml`.
.',-.
".
.., I--(;.- .,,;--1
II
--..
-,
-.l-,tt.-e
"..-".Iw
_7-1,;-f-,.-,-4
-,-. ,".jp,
-;,.-,,..i..,..-,__
,. '!:,, 1:.i- -,_-__.-.,
_.:. 1.1.
"
.; -,,:.-,_-,
-_-4.,
...._
-1, .'. _...,
.,.'
". .1_
.., -,
,_ Z?.,,
"I
,Z..
.-;'
. 41
_1_.--,""
illl _. ,....'.
.
:.tI
.,,.-,-%,.,-,;
I-,4,
-,.,.-,
1;
-.
k.
7
,.
'k,._..,..:--Z;-----'
.---_:,_ --;_..
.:-,--,
:.._14,..ll,_4
-._
--":,
,,1..
".
,,,.
,-,-,
",
'-..-..---,
.O.f'll.,
,
..,.,.t,
...-.*.
.- -,-._-f"--;,.,...(',.,.,,,.'
.;!._".,,. .k
-',.
.._.
,..
,_.t
I'll.'
..
.,_
.J;,... .,,t:; 4,g...._
-..".,..`,.,;,
'N,...,
_,;.,.- _,,.-.
--_.-..
-!-.
,_..I--"'.
.,_.1__
i.-':
,-_- ",.w,-,-I.:
--..1v,
1,
I.,
,..I.,,.`-._
-.-1,;.,-..;-.,--.,,;.,,--,
-,:.
%
't
-....--4
,I- 11
,.
, 1.t
-,-"V,
,
.
,
.1_1., -'- ,,- -1 -A-'
--.
._
. ,,,-,.
;.-f
'A.-,.,,
._
-i._
,.,..
,...Ir
-,.,-'
.
..";.,
..-.,'1
'
;
.
,-,
.,-...-...,,._.',
,,.;
,. 4j,
_ -".-'-.'-- -.
-'. ;. .. .-I,
I'll-!,-__ ,- .1'_
,..-,,_ ",
,_-i,,,,_,.,
-.:,--1-.-1;
I'll,.k','!
._4k
,I, tzr ". .--,.
___-, " ,.,.;.
-Ill_..;-1-,.1..-.
-,.-,,.,.k.
!,.-,.",S,-.
-1.
.,-,.-,.,
_; .'Irl. ,-.,
-_,_,,.
-;,
--,-,
,""-,
"..
-",-..
',tl'. -ZI.,,",-'' ,_Ahj
j;.. ..
;:..
,,_.
..X_ .I;:,
_.--,-1.yw..'
-,.,
-,.%1.
:.- __,._-.",-,-.,--_",,
-.----T",
--.-,
-.'L ,',.L".,;.,-.%-.
-x
,--"
,-- .','7"""'
., )._..,1_,.'.,.I Lv. ,,-,._I-._L
.,.-,..",
,,--,
.1, ,-.
.;i
',,'.`,",1'1_-.
'i'.1
...,.._
.- .i,.-.I-.'!-.
.,."'..,q"44.kl.
- .),; 1,-I
."
iz
:.--,,
._ f ,--,Z' .!--,.'.,--"".,.,..I
_f, _. .I11.
q.., 1-,
1.,-'.;
--.,-..,
-..
,.,
.,
-,'._
,. ""7. _.-Z ,.,-,
-.-,4.
-.,--,:,,
.,.-,.
..:.
-,_I,,_.I,
".,
,,
,.
%..-..f,
-.-%.___
.I.
..l- -1 1. 11- .-..-,'t,
I -."_%,
-,Z,4,- ,,---I .".
.
--,-I 1.1 -"Ir".-,..
- ..,.
,., ,F Pl!--,_
__
.,.. -1,,,,
'k;
-,,_." ,,I..
-'.,I.
--;---.
-1-1,,_..,
,.;.
t.-.-,
I
.1_?a-l-,__
,,,:.
,i- ;.,--7-'
,+'
1,,.
_r;
..T.
I'.
,-,"I,,..,- ._",<X.
.,,_""I
1,47_.
'. -,,.4"
-,, -'.'.'l
-.,'.-- -- ". k i...,-- 1,..)
,.,-.,.
,"'t _;.,,,
.,.' ", ;,
..,.,,-.
',. -..l
-,,,---"
:,_
:'-.,-.-,
--., .3 -, -'*i.--,""
.,.. '-,..,,-',.---,^
.o -4, .,.-k
.---,.
.,.I.---_. .1.-"_1r
,,..._,
--,.., - ,
11 ,":.,
", ,.
''.r,-""'..
-..._-'
.__
1- -_ .l ,`.,
I ;, _." -:'.,
"-.,
4w-:-.
."Z-,
,x-,
,'f-,:.:.,,-:_..`_
-,"I....;'- "..
_,.,..
1;-,,.-. -,,-.-:,.,. .,.
,;._ .1. ....
- , -.,
.. .-".
-1 --`.rj'_-.
_.. _t.- .'-.,II
-___
,I;..
Z.-i-,
- -,, ,k4l.
--.!
;j"
,-,l
-- -.
-- --' _-....--._
--,-- ,".,..i_ ,.,.-._,
.,. .,,,
-1.f',,.
..
'_.'- -.,'.-_ --_n"
_4"
_- "
--...-..-'

.,- _

-,,_

".-.I-.
__

..,

.!,

.; .,,

...

"?

_1

-,

-1-1

-v '.

-,

"_.

.1

..

,_.

.,I.-,-".
-

..

.I-I,--.
,,,

""tr,,

"i,:;r

Ar.
f,.,ZZ-.".
ii.,.,.,;;t_,.."__"--.-.
--'.l
.-.1.m
..1

I-,,.

'.

..

_-

'k,

1,..i.

..,

--

".

.-_

-,

.-

-,

.,

iz

-,-

"

.-_

I,
.

..-

,;'

._,_

,.."-,.

.-

,r

,z.

..

'-,

"",

--

i,:,

.rj

,'.

..

V;

"__t_--,--..
,.,7,. ..-"--.
I-,"..,..l
;1..-i, -17..-*."V."
__.,.,_.I_..,,-_.,_,,
,.:k,
;.

.,

__

--

.,,.:
-I-F

.l

.,.1v,
-,
,.l

,;.

,I

,:..i.

"

.,

,-

.-.

,,

-,

--.
.,

,.-

,
.-',
"-,

-,,..

"

.- _.

.'.,-

-.

,.. -.
4w,-

._:

ll.;,,._ 1,1:. "-_ -,..,-

..-

".;;.

,,-

-,-,
,,- - -

-.,.,

.,.._,
.!

,..._,
'_

..

"I

..

"

_.

I,

,,

,,

".,
.',

I.c,

.,

-,

,-.
I-,
-- ,---

-i,

.-

--i

",

_-

-,

.1

-,

P'

".

_1

.,
,,.,
- -I11
11
"
I--

.,

-..

.,

.&

,_,-p..

'..--..:.,. .1-I
-, "_I
-11.
.-i4::-'-, 11,
.j.,,-1,--,__-_.,_.,_l
,1.0,__..,,,Ii
."',,-,_,.
"'.,IN
t_,.-"_.;;:,_
,.
--,..
.- ... .1'..,,.._..
,,k'.,,`-.,-1,_...",x,_.-.-I.
.-,-.-,,-----._1.,..,l,,.,-_.,
-- ,. .- ,_'.'", ..
,-,,
,.,'.l
.-,,
Z.,A,
"
'
,
I...--_
---,;.",...
-_.
.;.,3T
-'-'t,

_._,

__.

..

1.--

"

1-

V.--1---

,7,
_..

...-

;.,_

,_.

:;.

".,

4_

-:

I.,
!.
,, --..-

-,-

;-

-,

.,.-

,,

I,..I.__,_,;

,..

..

--

.l

..

",

:, -..4i",

I-1.

'),

-"' .m l

-fm

,,.,-

,-,.

-,t4

,.k
!:,','
..,,'-,-,,!`.,
,_-`,.
.,.
li7
....-,Z-,,
I.N,,-,,-I.
-,-,. ";i.
.,-.l-..,.,-41-,-,-,

.-

,,,-,

.-f"_ ,,.(l _,..,,.,- ,,,- ,,.(l ...... . -..,, .1--.

,,--,..,-"
..1.
-1...;!;",
-,.r,-;
.- ,,,-....,_-.-.-,t",
:r,__-_..
1_.
10 ')."-,,
I,"t,'. -vk--,
.i;.,.llv -11,.,
16
-.',. -,,,, ,,,4.-t.
-. "... ." --t.-_t,!,-ll,,F.,_-,
.-'.'-,-..
',.;;.__.:- -4 . "-I
.*... ,,::. ".-zF, ,c;.
,.,.. --_
-',
'. '- :.1
I -,.I1,
'';.
--,c
,--, " "!.-.,_..
-,,
-j-,
-fl
..,
I..
-.r" Z-,-,,_
",_
".
.. ;, -,.,..'.o"_.,.,
,.,,,,
ll.:.;.. .--,l
i,....7.
4
,. ,.'4 -41
.,,,.w
-.-,'
., .-,,
,.-.,.
.-.:.,-,':
....,.'
1.
--,:,
4i.`
-,t,-_
-:.l
A.
S
.I
.,._.
-x,'-- -,,.,.,
.l
;,
.,------,
-;.-,
,A,
._
,:.,
.-,--_.__'11_1'j__,.-,
.,
..
..'.
-.,
.,.I--,-77;.-.W'.P".,,
.,,..
II-.
.':4."._.,
__J?'I..__,-,"--.'_-_
-!..
',.
,,.,,
- "._Z_
I,"
1A.-,
,,I.l
--,
I-_.*
-lnl..
_1A.I,e
-,---1.
__.,;.i__,!.-_.. .-:,.
zI1.. I.- _,_..l
_,__ ,..
-.,*- .,,'
-,., - _.
"...'71%;.i
-- ",!Z.
-_41
.:._.,)
..,---;,- --_..,."_. -!---..,-..,Z;,,...
Z_..
_.. -4-1I1,.,.,,Z
-,.4..
_.._
'_ --.
.-',",...;'"..-_. --, .,,_,
-,.,
,.
..-,,
,_
---...1...
,-" ..,.,-_r-,:.,'
.1,..
--,.z-.:.'
,,_
-,;-,,
4_.,
.,..
.,-.._
,'4
z..t, .-;._,.
.,,I_,_
_7.' -,,`..
..,
-. ;:. ... ..
f.._
..
.--_. t..
..-,-,",I
.,.
;.
,,,..
-, .I.,
-.,,, r",-,_
-.
-.,-, ,,,.X-F
,,,- _;.._
ty.'..--.._'4-f,.-,--.-,,4.,,...l Irl, .... 1;. 1.:., -;L ,,,.1
.,C. ---"g..--....-.i4.
.,I,
Z..,,T-,..:I..
-I' ri ,, ".,..:_.. .-t_--1. V,- -,Z.,_T",k
-I, '.. ,.---t-.
,--1.:
'-,,,-..
,..-,
-, -, -.
_,.
.,
."-,-...
-,--,
I-.1
,.I---7, .,- -,.,,-.,
z.v"I
..--.-'%,
l,'i.
.,._..A.ij..,
0,-:---.
.,--,,i,-",'
,; --..
,,--,-,-,,-"
.l,-:-1
T-Ic
.:-,.
1, ,-. .1.
_11.1-_.....
--,,.--.
"`f
1.
,..r
,,i.
._
1.
.,
--.-.
- -.,.,L.
-, ,, ."",..-",
- ',- `.--,, ;--:-",
,..
,_._ ,,Zlj.
- ,-Z -- ".- --_..`.,.:-, "W-,
-.,,_ --.:, ."I
,-.
,.;..-,,
'.'.?. --,
_4,.-, ,._
....,
_..-;.I,',-`.. -1",.,,;, ,., !......;.,,,
. .-..
, ___.._
..v_f:,,,t44.:,.--t7..
---,..--.-1!1V .,.,
,.
, ..
.,.;.
"t,,,i:;:.
1 -- ,.. -"-)_v__,:".-_4'
..,.l
...I1:;l
-- .,.
!..A.
,.,.,
..-..
-.1..",ir
,,_.,,,.,4',
.,...,_.,-....-..
.l -:'7-.k,-,---`l;..,
""
., ,. ..,--C,;
,.,
__.,
__., ..
"..,
_-..
II,,
--1.__.1_.
II,l
-,1.I"
-,-.,.:o.
:..--,'-'._%.,
,..- '4---,-.-.-.
,Z.
-,
",,,W-,--,
--,-,
t_4..-_-t... -..
_z..-,
----.
,--,-__,,,
-Z-,--.,-.,
,I..
!..-,
-I-,
.,.
- "',
,.-,j
,--'.
.:_ -'-!.'- .4.*.IL
,.,.., .1
4'. ..,,
,11;,. _._,,--,-.".
-".,
"'--,,:.
,.44
,
.S.z
",,--- -I'.4- ,.,,i.,-, ,,,, ,;.-'Q!..
_,.,
-I;1.
,,., .'-.,q.
,,..,
`_ .. ,Q
-I..,-4
_.J._,
T-- 'z...
,.,-1.,,-.r.,.t.,
_,,-.
,--_.V,:-.
-- ,.--r.' , ,_0.,
,.:!..,._,'.-,.t.
.k,
:.
,.., ,;.-.,,
"4_.",-.-,
-,.,
.&.",
-. -,z...
:.,(-'-#.,. ,..
--.-___
.,_ ,,I_llz
.."1.1.4
.-<--t,;.-.
....
..,.,t
"..,
-..e.,.....
,!.'_._,-..
,Z.I,
lk
.--._,.
I-,
"`,-.,
..H
,1,It
._.,,,
-,I-.-,.
....:
..- -.-,-f-II't,
-,I.,
,,
i,.,",
t'__'_.
t"
",
-,-,.,
t,
I,-'4.;v
Ir-1.-:.;.
..-.
..
6, ..,"
,,. _,.: -, I, ..,.--.1
--<4,,,L---..,
- --,
.$t ..-_
,-..
r.-.
I.11.1
,,.-..-:k-.-,1,
1,-.-,.
;;:,-,,,.",
I,
!_.,.. II :. -"--'"-.,.':,'
-..-,Z,
,,..,,!.
,..,
-,--1_,j., _'l
_v_
-"
- " -;"--"

_.".

1.

"

-,

,;.

.-

_.

4_

-1

-"..
.,.4..-_',--,-;,I

--

.__,.

_I.,

o_

._

-.",

7,i..,.

..,

.-:

,,.;,

-1

_.,

,e-

.,

..

...

,:.,.

V"

-r ., .,

-';'

".

_-

%;

"

_'

_Ltjl.lrf,.', _
,%i-_-.r., :".;_-,-`-..
_,..--..
-:__..___..-,
,..-,--1,.
11

,#--n't--.,

1..

,.."

1,

---

,_..,

-,

-,

,,,

".,

_.
--.-

--

_...

,-.
-----,,,,.,q--71.,
-11
4.1i,
..-.f.,.--:.
-_.-"-1,.._....,
t..-%x
.IIz- ,..
I.._-'-.,..-..'.'..-.,..-_-,.:,
--..r
I-,.)._.,,".-,;.
171zm,_,;'l
ll._
,-.,,__

- I,

,-

.e

.,

,,,

--

r(

---_."-

-t

'I,

,.

.,

,.

,_;,.

,.r

7",.-',,
v,"._`-- -.11-

1'.,

...

..

.,

,.

_'..,

1:.

-_
", 1.,,

,--

_.

-,
.- .,.,

-,

.,

--.,-

Ow,

',-..-

ir,

_1

.-7---f

,.

JV

'47- J.,

t--7.,_-.."
,,-_'l-,
-1,--i......-

',--,,

_.

,
",.
_.I-1. .-Z-,

._

.-

.,

ftz

~~~~A
~~~~~~A