Andean Past

Volume 5 Article 18

1998

Textiles from the Lower Osmore Valley, Southern

Peru: A Cultural Interpretation
Ran Boytner
University of California Los Angeles, [email protected]

Follow this and additional works at: https://digitalcommons.library.umaine.edu/andean_past

Part of the Archaeological Anthropology Commons

Recommended Citation
Boytner, Ran (1998) "Textiles from the Lower Osmore Valley, Southern Peru: A Cultural Interpretation," Andean Past: Vol. 5 , Article
18.
Available at: https://digitalcommons.library.umaine.edu/andean_past/vol5/iss1/18

This Article is brought to you for free and open access by DigitalCommons@UMaine. It has been accepted for inclusion in Andean Past by an
authorized administrator of DigitalCommons@UMaine. For more information, please contact [email protected].
 TEXTILES FROM THE LOWER OSMORE VALLEY, SOUTHERNPERU:
A CULTURAL INTERPRETATION

Ran Boytner
University of California at Los Angeles

Introduction such project, the Proyecto Colonias Costera&

de Tiwanaku (PCCT) (Owen 1991a, 1991b,
During the tenninal stages of the Ti- 1993). Specifically, it analyzes the textiles
wanaku state, the peripheral regions were the from burials at the site of EI Algodonal, which
first to feel the impact of the state's collapse. is located in the lower third of the coastal val-
The middle Osmore Valley, around the mod- ley (Figure 1) and was excavated by PCCT in
em city of Moquegua in Southern Peru, was 1989-1990. The focus of the research is to
one of these places (Figure 1). The site of compare the textiles from 110-Tumilaca/
Omo, which had served as the Tiwanaku cen- Cabuza burials and Chiribaya burials, which
ter in the middle valley for several centuries, can be clearly distinguished by the distinct
was abandoned during the tenth century A.D. style of ceramic vessels associated with each.
at the beginning of the Tumilaca phase (Ber- Additional specimens from the sites of Chiri-
mann et al. 1989:282; Goldstein 1989). baya Alta and Yaral (Figure 1) were examined
Around this time, some of the local population in order to increase the size of the Chiribaya
moved into the coastal segment of the Osmore sample and to evaluate the site level variabil-
Valley, which had not previously.been inhab- ity in Chiribaya material. The samples from
ited by Tiwanaku-related people (Owen these two sites were selected randomly from
1991a:l). . burials with associated ceramics in the AI-
gorobal phase (early Chiribaya) style (Jessup
These new, cOIIlplex societies in the 1991). Only a small fraction of the textiles
coastal Osmore Valley were based on irriga- from these sites were analyzed, however, and
tion agriculture and constructed a long canal future analyses of the complete textile collec-
to incre~e agricultural production. They were tions from these sites will certainly increase
also camelid herders (Owen 1991b). Perhaps ourunders~ding of the coastal valley peo-
the most striking aspect of this intrusive soci- ples and cultures. .
ety is that it was clearly divided into two dis-
tinct, .contemporaneous groups, archaeologi- The initial problem that this research ad-
cally identified as the 110-Tumilaca/Cabuza dresses is the definition of the 110-Tumilaca/
culture and the Chiribaya culture.' Both cul- Cabuza and Chiribaya textile styles. Though
tures left extensive archaeological remains, the two cultures are clearly differentiated by
including habitation sites, cemeteries, canals, their ceramics (Jessup 1991; Owen 1991a;
and field systems, which can often be unambi- Goldstein 1989), this is the first attempt to
guously assigned to one culture or the other. correlate these ceramic differences with the
This dual occupation of the coastal Osmore textile products of these two cultures. How-
Valley lasted from around A.D. 950 to A.D. ever, the scope of this research is broader.
1250, while the Chiribaya culture persisted to
perhaps as late as A.D. 1375 (Owen 1992). Based on the definition of each culture's
textile repertoire, the textiles of each group are
The lower Osmore Valley has been the analyzed independently to address specific
focus of various projects. Some are com- problems. The first is the problem of the ori-
pleted, while others are still not published. gins of the two cultures. The second problem
This research analyzes the textiles from one is to assess the relationships of both cultures
with the Tiwanaku culture in tenns of the pro-
cesses that occurred during the collapse of the
1 For extensive discussion of the differences between Tiwanaku state, and the overall influence of
the two cultures see Owen (1993).

ANDEAN PAST 5 (1998):325-356.

ANDEAN PAST 5 (1998) -326

this collapse on the remote middle and coastal and identified by Owen and myself, using Jes-
valley. Third, the social and political structure sup's (1991) seriation ofChiribaya ceramics.
both between and within the two coastal val-
ley cultures will be addressed. Differences of Methodology
quality, decoration, and motifs in textiles were
recorded for each culture, and were analyzed Each piece of textile was laid flat on a
in order to suggest individual status differ- plastic supporting screen and mechanically
ences, as well as cultural identities. Finally, cleaned with a soft brush. Each was photo-,
the apparent lack of gender differences in the graphed and then recorded on forms adopted
textile record of both cultures will be dis- from an earlier version that was used by Niki
cussed. Clark for her research on the textiles from the
middle Osmore site of Estuquifia (Clark
The majority of the textiles analyzed, 395 ]988). Many pieces were illustrated with col-
fragments, came from the site of EI Algodonal ored draWings. These drawings are not to
(Tables 1 and 2). All of the textiles recovered scale, but a scale was included in the photos.
from the cemetery of EI Algodonal are con- Data recording emphasized types of patterns
sidered to be 110-Tumilaca/Cabuza textiles. and decorations, using both drawings and
Every diagnostic ceramic vessel and sherd written descriptions. Measurements of width
from this cemetery pertains to the 110- and length of the complete article were taken
Tumilaca/Cabuza tradition; not. one single when present or reconstructible, as were yam
Chiribaya sherd was found there (Owen thickness and color, spin and ply directions
1991b); The cultural affiliation of this mate- (where evident), yam count, weaving and
rial is quite secure. Textiles from burials in decorative techniques, and character of deco-
the EI Algodonal habitation area were .as- ration and form (tunic or shirt, bag, etc.). Af-
signed to one or the other culture according to ter being recorded, each textile was wrapped
intact ceramic ves~els found in situ in the in acid free paper and stored in a box con-
same burials. taining all the textiles from the same prove-
nience. To reduce curatorial problems, some
Many of the textiles came from mummy extra large textiles were stored separately in a
bundles excavated from intact burials. A large special large box.
number of textiles from EI Algodonal, how-
ever, were isolated pieces in various condi- The data were coded and entered into
tions of preservation collected from disturbed dBase 111+directly from the original recording
burials or severely looted surface areas at this forms, and data on the age and sex of associ-
site. The intact mummy bundles were un- ated human remains were added. Ages and
wrapped by Juana Lazo (Universidad Cat6lica sexes were provided by Shelly Burgess (also
"Santa Maria", Arequipa), Niki Clark (Univer- see Burgess n.d.). The computer database in-
sity of Chicago), and their associates prior to cludes variables selected for their relevance to
this project. The present research deals only specific research questions. These variables
with the textiles themselves, not the ways in and research questions include:
which they were used to prepare mummy
bundles.2 1. Presence or absence of mending was re-
corded as one possible measure of raw mate-
Other textiles came from the site ofChiri- rial availability, and as a possible indicator of
baya Alta (54 pieces) and Yaral (12 pieces) economic and social status differences within
(Tables 1 and 2). The cultural affiliations of and between different sites and cultures.
these textiles were based on the intact ceramic
vessels found in the same burials as the tex- 2. Textile area and density were measured as
tiles. All of these ceramics were examined possible indicators of status, assuming that
larger as well as denser pieces of cloth may
indicate higher status (Roach and Eicher 1965;
2 Research addressing the issue of mummy bundle con- Murra 1989; Weiner and Schneider 1989).
struction will be carried out by Clark and Lazo.
327- Boytner: Osmore Textiles

One density measure used is "crossing count", 7. For decorations that contained geometric or
calculated as the number of warp/weft cross- naturalistic motifs, as opposed to simple color
ings per square centimeter. In practice this stripes, individual types of designs and motifs
figure is the product of warps per centimeter were defined and recorded (Tables 12 and 13).
times wefts per centimeter. Taken together,
crossing and yarn thickness give a general 8. This category was primarily used for the
measure of the density of the cloth. In the 110-Tumilaca/Cabuza textiles, because the
majority of cases, the measurement of the total Chiribaya textiles from EI Algodonal were
original area was impossible to achieve be- usually decorated only with color stripes. In
cause not enough of the textile was preserved the Chiribaya sample from Chiribaya Alta, on
to allow the extrapolation of the original size the other hand, many of the motifs were much
of the item. more elaborate. Because the sample was
small, no typology was established for these
3. Direction of spinning and plying (where motifs, which were coded simply as "elabo-
evident) was recorded in order to detect po~si- rated forms".
ble cultural differences. It quickly became
clear that no spinning differences between the A Comparison between 110-Tumilacal
two culturescouldbe observed. . Cabuza and Chiribaya Textiles

4. Type offweaving technique was recorded The initial research problem was to iden-
for the main weaving surfac~ (as distinguished tify differences betw~en the 110-Tumilaca/
from weaving used for finishing or decora- Cabuza and Chiribaya textiles. This analysis
tion). In rare cases where a variety of weaving emphasizes .variables that might have been
techniques were used on one single textile, intended to indicate cultural differences, that
each individual technique was recorded. is, traits that are easily visible on the textiles'
These data were collected in order to indicate surface. Shape, main weaving technique, fin-
possible cultural differences. ishing, and decoration are the primary vari-
ables considered. These components are the
5. Finishing techniques were recorded by visible parts that are the end results of the in-
types and the specific location on the piece. tention and labor invested in the creation of
Finishing techniques included technical and each textile by the weavers of each specific
decorative aspects of salvages and seams. Lo- culture.
cations on the piece included seams on the
sides of tunics and bags, bottom edges of tu- Most of the analysis focuses on tunics and
nics and top edges of bags, edges of necks and bags because other forms are represented by
arm openings, and other edges of other types very few examples. Tunics were the most
of textiles such as panuelos (handkerchiefs), prevalent type of textile among those that
mantas (blankets), etc. These variables were were identifiable (Figures 2, 5, 6, and 12; Ta-
examined not only for cultural differences, but ble 1). The shape of tunics is consistent
also for differences by sex, age, and site (Ta- among the 110-Tumilaca/Cabuza and Chiri-
bles 7, 8, 9, and 10) within and between each baya tunics from EI Algodonal, Yaral, and
culture. Chiribaya Alta. The tunics are all sleeveless,
with a slit neck, and range from square to
6. Wherever decoration was present, its ap- trapezoidal in comparable frequencies in both
pearance was recorded. The techniques used, cultures (Figures 2 and 3). Tunic size varies
the location on the piece, and the design pat- from 54 by 57 cm to as large as 120 by 107
tern were all noted. A series of 17 categories cm. No specific tunic size could be associated
(Tables 12 and 13) was developed to code with each culture. All but one tunic were sin-
these data, and certain categories were further gle web products, folded and sewn along the
analyzed as described below. warps, leaving holes for the arms. A neck slit
was left unwoven and when the structure of
the tunic was trapezoidal, warps were added in
ANDEAN PAST 5 (1998) -328

the shoulder area. Warps run length-wise on Only seven of the textiles were woven
all tunics investigated. completely in techniques other than warp face.
These pieces are too few to indicate any intra-
Bags were the second most abundant type valley cultural patterns, and they might
of textiles identified (Figures 3 and 9, Table equally represent exotic objects or unique in-
1). Many individuals had a bag full of coca dividual preferences.
leaves buried with them (Owen 1991b), and
some had two such bags. Bags were rectan- Based on both yarn diameter and crossing'
gular or trapezoidal in shape and had similar count (density), bags are generally woven
average sizes in both cultures. The same finer than tunics, a characteristic found in both
weaving technique utilized for tunics was used cultures. 110-Tumilaca/Cabuza bags have
in bags. The pouch portions of bags are single wefts 0.8-1.0 mm in diameter, and warps from
web products; warps run vertically and are 0.6 to 1.0 mm in diameter, while Chiribaya
sewn on the sides leaving one end open as the bags have wefts between 0.8 and 2.0 mm in
mouth of the bag. When bags are trapezoidal, diameter, and warps 1.0 mm in diameter (only
warps have been added to the lower section, two Chiribaya bags were analyzed from El
making the bottom wider then the top (Figure Algodonal).
9).
Ilo-Tumilaca/Cabuza tunics at EI Algo-
No cultural or gender differences in bag donal have about the same weft thickness as
size or shape could be observed in the sample. bags (0.96-1 mm.), but have 20% smaller
It seems that the shape and size of bags were warps (0.66-0.74 mm.), with an accompanying
individual choices. Unfortupately, due to lo- reduction in yam crossil}g(108-156 c.p.s.c. in
gistical problems, bag samples were available bags, 91-101 c.p.s.c. in tunics) (Tables 3-4).
only from Algodonal and Chiribaya Alta. The same phenomenon is evident among
There is, however,. no reason to believe that Chiribaya tunics from El Algodonal. Al-
the bags from Yaral would differ significantly. though only a small sample is available, there
is little significant difference between the two
The most common technique used for the cultures (Tables 3-4).
main weaving of the textiles is a plain weave
1/1 warp-faced technique (Table 6). In this Finishing techniques were sometimes used
technique, the warp yarns of the textile are the for decorative purposes in both the 110-
ones that are seen on the surface (VanStan Tumilaca/Cabuza and Chiribaya traditions.
1958a; D'Harcourt 1962; Albers 1963; Coll- Many tunics have elaborated finishing at the
ingwood 1987). The warp-face technique pre- neck and the arm hole. These places were
dominates in both the 110-Tumilaca/Cabuza evidently the most stressed parts of the gar-
and the Chiribaya textiles. ment and tended to be the first to break down.
For this reason, tunics were often reinforced at
Some of the textiles were woven with a these points, and often this reinforcement also
combination of weaving techniques including served as decoration. Even in many simple
the warp-faced and simple plain weave tech- tunics, some extra weaving or embroidery was
niques, among others. These textiles have low done at these points in more sophisticated
crossing counts (13-15 crossings per square ways than the main weaving itself. Less frag-
centimeter, c.p.s.c.) and show a disorganized ile parts of the tunic, however, such as the
mixture of weaving techniques. These textiles bottom or the sides, were decorated only if
apparently represent pieces made for purposes other parts of the tunics were decorated. Fin-
that did not require standard quality cloth, or ishing was used occasionally for decoration at
by people who did not have access to material the sides of the tunics. In tunics of this type,
and/or labor resources available to other the side seam is sewn together below the arm
members of their society. hole with one or more colored yams, and often
with unusually fine yams. The bottom edges
of tunics vary little and were generally. only
329- Boytner: Osmore Textiles

reinforced for functional reasons. However, in ing each line vary in colo.-3to create geometric
a few cases the bottom edges of tunics were designs. The most dominant color is red, a
not only reinforced but were also embroidered color that during the Late Horizon was com-
to further increase this part's durability. The monly used among the Inca to indicate high
sides and the top part of bags were almost al~ status (VanStan 1958b:53-54; Murra
ways done with decorated finishing, which 1962:719; Uhle 1903). The lines also contain
formed an integral component of the overall blue, green, and yellow. The geometric de-
decoration of these objects (for range of fin- signs are always presented in a "double fea-
ishing techniques and frequency of use see ture" format, with each side decorated with
Tables 7-10). modified "M" motifs, solid triangles, elabo-
rated S shapes, or combinations of these (Fig-
Cultural differences in finishing types are ure 4). Usually this "double feature" format
not clear. Ilo-Tumilaca/Cabuza and Chiribaya was laid out in an almost, but not quite, sym-
people from the different sites used all the metrical pattern. Minor but discernable dif-
various finishing techniques, and weavers of ferences make each side distinct from the
each culture applied each technique to some other, creating a "pseudo-symmetrical" design.
location on their textiles (Tables 7-10). There The designs consist of both vertical reflection'
were preferences for using specific techniques and 1800 rotation or pmll and p1l2 respec-
on specific locations on textiles in each culture tively according to Ascher (1991:160). Add-
at each site (Tables 7-10). I attribute these ing the color scheme, a design both vertically
differences, how~ver, to the small sample of reflects itself in differe:gtcolors, or pm'II, of
Chiribaya textiles from El Algodonal, and.the the main body, and also has vertical rotation,
even smaller sample of Ilo~Tumilaca/Cabuza or p'll2 of the added triangles (ibid.:164-165).
textiles at Chiribay.aAlta and Yaral (Tables 1
and 2). A clearer idea of patterns in decora~ The other major location for decoration on
tive finishing techniques will be available only Ilo-Tumilaca/Cabuza tunics was done on the
when the full analysis of the Chiribaya Alta, sides, starting under the arm hole and ending
Chiribaya Baja, and Yaral textiles is com- at the bottom, covering the seam that joins the
pleted. The important point is that no specific front and back panels of the tunic. This stripe
finishing type or technique can be exclusively usually comprises 5 to 11 loop-stitch embroi-
identified with one or the other culture. dered lines done in the same range of colors as
the ones used for the under-the-neck stripe.
Cultural differences in the textiles are evi- These side stripes are typically 20 to 100 cen-
dent only in the most obvious and intentional timeters long, depending on the length of the
of their features: their decoration. It is diffi- tunic. The geometric designs include ones
cult to identify differences in any of the other similar to those used under the neck, plus ad-
categories examined. However, cultural dis- ditional motifs found only on side stripes
tinctions are very clear in the types, locations, (Figure 11). In this case, because the stripe is
and techniques of decoration on the textiles. much longer, the design pattern is allowed to
repeat. Unlike the two sides of the neck
110-Tumilaca/Cabuza tunics are typically stripe, the side stripes on a given tunic appear
decorated with an embroidery technique that to be identical to each other, or vertical reflec-
creates geometric designs in two different lo- tion pmll (Ascher 1991:160). This may be
cations on the tunic. The most visible and because the stripes are long, narrow, and
common is a narrow horizontal stripe directly placed widely apart from each other, so that
under the neck (Figure 4, Table 11). This even if they were different, the observer might
stripe is composed of five to seven adjacent not notice the asymmetry. All the tunics that
loop-stitch embroidered lines usually no more
than 10 centimeters long. The yams compos-
3 For analysisof dyestuff from this textile collection,
see Boytner and Wallert 1993; Wallert and Boytner
1996.
ANDEAN PAST 5 (1998) -330

have side decorations also have neck stripes, line (2-4 millimeters) composed of two to
but some tunics have neck stripes only. three colored warp yarns. The symmetry of
the design is completely different from those
Another way that the Ilo-Tumilaca/Cabuza on the Ilo-Tumilaca/Chiribaya. It is done in
people decorated their tunics was to place se- glide reflect, or pial (Ascher 1991:1960),
ries of divided stripes along the two vertical which presents a completely different concept
exterior sides of the tunic (Figure 5). Each of of spatial organization of the design.
these divided stripes is comprised of a wide
stripe with a narrow line of a contrasting color Bags were more elaborately decorated than
running down the center. A wide series of tunics. Unlike the tunics, there is little varia-
these divided stripes, in different color combi- tion in bag colors; they are either brown or
nations, is placed on the two sides of the tunic. red. Brown bags were always made of what
This decoration usually occupies about 2/7 of seems to be natural brown wool with a few
each side of the tunic, leaving the center area narrow vertical colored bands. Red bags are
plain. The stripes are not done in the embroi- usually red overall with the same type of ver-
dery technique, but rather by weaving dyed tical colored bands.
warp yarns to create the lines as part of the
primary structure of the fabric. At El Algo- Typically, each bag is decorated with three
donal, this kind of decoration was found only vertical, wide bands containing various de-
in Ilo-Tumilaca/Cabuza contexts.. However, signs (Figure 9). The designs are done with
this decoration was also found in Chiribaya supplementary warps in red, blue, green,
contexts at Chiribaya Alta and I suspect that it white, brown, and yellow. The motifs used on
was associated with high status and prestige the stripes are geometric and repeat along the
privileges. . vertical axis (Figure 8).' The stripes are sym-
metrical and each stripe is similar in its deco-
Chiribaya tunics were decorated differ- ration to the others. Also, the back and front
ently from Ilo-Tumilaca/Cabuza ones. Based of the bag are identical (although differences
on associated ceramics, four Chiribaya tunics in the front and back decoration of bags have
were recovered for El Algodonal (Table 1). been noted in many bags scattered on the
One of these tunics is undecorated and plain. looted surface of Chiribaya Alta).
The other three, however, are decorated in a
distinct and similar style, unlike the one seen The differences between Ilo-Tumilaca/
on the Ilo-Tumilaca/Cabuza textiles. The Cabuza and Chiribaya bags at El Algodonal
limited number of tunics from Chiribaya Alta are slight. All bags from Ilo-Tumilaca/Cabuza
investigated here (20 tunics) are similar to the contexts have linear patterning (Figure 8 left),
Chiribaya tunics found at El Algodonal, which and only one bag from a Chiribaya context
suggests that the. El Algodonal sample is a had different, triangular patterning (Figure 8
reasonable representation of Chiribaya mate- right). Both cultures used the same supple-
rial. mentary warp technique to create the band
motifs. Because only one example of the tri-
The only decorative technique commonly angular pattern was found, it is unclear if it
used on Chiribaya tunics at El Algodonal was reflects a real cultural difference. Many such
done with colored warp yarns that form stripes bags can be seen on the looted surface of
in the primary structure of the cloth. This Chiribaya Alta; only future research will be
decoration results in long colored stripes on able to assess these differences.
the sides of the tunic, starting at the shoulder
and ending at the bottom edge. On each side The differences in decoration on Ilo-
of the tunic, there is a narrow line in one color, Tumilaca/Cabuza and Chiribaya tunics might
usually 1-2 centimeters wide, and a wide line be attributed to one of two possibilities. The
of a different color, usually 3-7 centimeters in first possibility is that there was a chronologi-
width (Figure 6). On a few tunics, this pair of cal difference between the two cultures. In
stripes is delimited on each side by a narrow that case, the distinct decorations would be
331- Boytner: Osmore Textiles

because the two cultures never met. However, The large size of the Chiribaya stripe motifs
recent radiocarbon dates from sites in the val- and the location of the 110-Tumilaca/Cabuza
ley indicate a substantial length of time during neck stripe on the high central part of the tunic
which 110-Tumilaca/Cabuza and Chiribaya makes them both easy for an observer to dis-
people coexisted in the valley (Owen 1992). tinguish, and fit well with Wobst's observa-
Moreover, 110-Tumilaca/Cabuza and Chiri- tions concerning stylistic elements that are
baya burials were found in the same habitation used to communicate ethnic affiliations:
area at El Algodonal (Owen 1991b), in the These two cultures must have interacted con-
same cemeteries in Yaral with no stratigraphic tinually, and because no evidence of warfare
differences between them (Lozada 1991; Lo- was found in the valley, they evidently inter-
zada et ai. 1991), and probably also in the acted in a peaceful manner (Owen 1992).
same cemeteries in Chiribaya Alta, although Nevertheless, the two groups probably com-
these.data are still preliminary. Also, there are peted for .control over resources, mainly water,
5 textile pieces from Chiribaya Alta, 3 from which were severely limited in the narrow
looted surface contexts and 2 from tomb exca- valley.
vations, all of which have 110-Tumilaca/
Cabuza and Chiribaya style decorations on the Hodder (1979:450) suggests that in times
same piece. This mixing of decorative tech- of economic stress, cultural differences be-
niques might be .expected among contempo- tween groups are often highlighted. These
rary cultures. The technical similarity in tex- differences are manifested as a stronger em-
tiles, including weaving technique, yarn thick- phasis on characteristic material culture sym-
ness, yarn density, spin, and size and shape of bols, making the boundaries between cultures
tunics, bags, and other forms also suggests particularly distinctive. At the same time, one
that the two cultures coexisted, at least for a possible response to difficult times is to in-
limited time, in the'lower valley. crease cooperation between the .groups in or-
der to enhance the total efficiency of resource
Another possible explanation for the dif- use from a given area. While the level of co-
ferences in the tunics of the two cultures is operation is increasing, each participant will
that they were used to mark ethnic differences. expand his or her presentation of, and empha-
As the largest single piece of cloth worn by sis on, symbols of cultural identity. He or she
these people, and possibly the one mostly will try to exaggerate these symbols in order
commonly worn, tunics were well suited to to make them as distinct as possible from
use as cultural identity markers. Ethnohistoric those of the other culture.
material suggests that clothes were important
tools for marking cultural identity, status, and The situation in the Osmore Valley seems
roles in Andean societies. Textiles were so to have worked in a similar way. The two
important that a naked enemy was considered cultures coexisted in the narrow valley, where
to be powerless and not a threat (Murra 1962; resources were limited and the environment
Murra 1989:280; Clark 1988:1). Moreover, was not similar to the areas from which they
"no political, military, social, or religious had come (probably the middle Osmore Val-
event was completed without textile volun- ley). Coping with limited means for survival,
teered or bestowed, burned, exchanged, or sac- cooperation formed between the two cultures,
rificed" (Murra 1989:293). probably with each developing its own techni-
calor economical specialization. Parallel to,
Although the 110-Tumilaca/Cabuza and and because of, this situation, distinct cultural
Chiribaya tunics share many similar traits, it is identities were highly emphasized, and so the
easy to distinguish between them because the use of individual cultural markers was impor-
differences are found in obvious, visible traits tant.
that were easily manipulated to serve explicit
message functions. Wobst (1977:333) sug- The size of the designs that comprise the
gested the importance of size and location of textile differentiation, especially the 110-
symbols used to transmit cultural messages. Tumilaca/Cabuza markers, suggests another
ANDEAN PAST 5 (1998) -332

important aspect of the two cultures' interac- were the same in both cultures, which may
tion. An observer could identify the cultural indicate that they have a common origin.
affiliation of an individual only from a dis-
tance of a few meters or less. The implication Camelids and camelid wool were clearly
is that in general there was no need for early of great importance to both' the 110-
reconnaissance when individuals from the two Tumilaca/Cabuza and Chiribaya people of the
cultures approached one another. Individuals coastal Osmore Valley. Masses of raw'
could approach one another without threaten- camelid wool were found in some burials,
ing or being threatened. Wobst (1977:322) sometimes on the bottom surface of the tomb,
has suggested that ethnic markers visible only and sometimes inside the mummy bundle.
at such short range imply interaction, rather Many of the burials contained camelid bones,
than aggressive conflict, between the groups. most often the cranium, mandible, and/or the
In this. sense, the textile data support the con- foot bones (Owen 1991b; Lozada 1991; Lo-
clusion that the relationship between the 110- zada et al. 1991). Although all the textiles
Tumilaca/Cabuza and Chiribaya cultures in examined from burial contexts were made of
the valley was one of cooperation and peaceful wool, excavations in the habitation areas of El
coexistence. Algodonal and other coastal Osmore sites fre-
quently encountered bolls, seeds, yarns, and
Evidence for Highland Origins . textile fragments made of cotton in midden
contexts, so cotton was clearly available and
Andean textiles typically embody great used by the people of El Algodonal (Owen
ideological power, and in order to preserve 1992: personal communication). Their exclu-
this power, textile manufacturing methods and sive use of wool textiles in burial contexts is
decoration designs tend to be conservative unusual among Andean coastal cultures,
(Conklin 1983:20;' Frame 1986:52; Murra where cotton yarns were often incorporated
1962:713,721). Rene-e, textiles can be effi- with wool yarns for ceremonial, as well as
cient tools to assist in determining geographi- utilitarian textiles (Murra 1989:153; Rowe
cal origins. Comparing textiles of suspected 1986:153). The fact that the vast majority of
immigrants with textiles from areas where yarns used were single yarns S-spun is diffi-
they are thought to have resided for long peri- cult to explain. S-spun yarns typical of the
ods of time may indicate cultural relationships North Coast tradition are usually cotton warps
or place of origins. (Rowe 1984:89). Moreover, this type of yarn
construction is different from Tiwanaku tex-
As the 110-Tumilaca/Cabuza and Chiribaya tiles and other Late Intermediate textiles found
people were cooperating in the valley, the in the region, which are typically Z-spun 2S-
need for clear and immediate distinction was plied (Conklin 1983; Lothrop and Mahler
evident. Tunic decorations were highly em- 1957; Oakland 1986). Is it possible that the
phasized as the primary medium for exhibiting 110-Tumilaca/Cabuza and Chiribaya cultures
cultural identity. Once the goal of distinctive- both adopted North Coast yarn preparation
ness was achieved, further distinctions were technology? The answer for this question is
unnecessary, and the other textile traits re- still far from clear and this investigation sug-
mained the same. Clear status differences gests no explanation for the phenomenon.
within each culture probably existed, but can-
not be observed in the investigation of the This unusual emphasis on wool, and the
textiles from El Algodonal. The brief analysis similarities between the textiles of the two
of the textiles from Chiribaya Alta clearly in- cultures, suggest that both 110-Tumilaca/
dicates intra-cultural social differences (see Cabuza and the Chiribaya traditions derived
below), but only a complete analysis of this not from the coast, but rather from the high-
site's textile assemblage will identify the tex- lands, where camelids traditionally played a
tiles characteristic of each social class. Bag more central role in the economy and ideol-
decoration techniques and forms, main weav- ogy. The similarities in textiles suggest a
ing techniques, and other textile characteristics close relationship between the two cultures,
333- Boytner: Osmore Textiles

perhaps even a common origin. It is clear, with decorated bands identical both in tech-
however, that at the time when the two cul- nique and color to those used on bags. This
tures are first archaeologically recognizable in tunic, together with the general similarities in
the valley they were already distinct from each ceramic motifs, suggests that the people of the
other. Based on the textiles alone, it is impos- coastal Osmore Valley knew the Tiwanaku
sible to select a precise point of origin in the ideology and were familiar with the former
highlands, but the textiles are certainly con- Tiwanaku center in the middle Osmore valley,
sistent with other archaeological data that sug- at the site of Omo. Although the presence of
gest that at least the 110-Tumilaca/Cabuza the Tiwanaku culture in the middle valley is
culture derives directly from the terminal Ti- clear (Goldstein 1989), the relationship be-
wanaku tradition of the middle Osmore, near tween the ideology of Tiwanaku and those of
Moquegua. the 110-Tumilaca/Cabuza and Chiribaya cul-
tures is still unknown. Clarifying these rela-
110-TumiiacalCabuza, Chiribaya, and the tionships will increase our understanding not
Tiwanaku State Ideology only of political and social processes in the
valley's history, but also of the processes and
Our understanding of the nature of rela- consequences of the collapse of the Tiwanaku
tionships between the Chiribaya and Ti- state.
wanaku cultures is still vague. A distinctive
tunic that casts some light on the ideological There are interesting contrasts between
relationships of the 110-TumilacafCabuza peo- Tiwanaku textiles and coastal Osmore textiles
ple and the Tiwanaku state was recovered in their treatments of symmetry and perspec-
from a looted burial at El'Algodonal. This tive. In the 110-Tumilaca/Cabuza tunic deco-
tunic is unique in its decoration, and is the rations, as well as on the Chiribaya ones, there
only piece examined that has depictions of an is always a slight distortion of symmetry. Di-
anthropomorphic figure. The figure is done viding the tunic on a vertical axis from the
with supplementary warps on stripes vertically neck down, the two sides are never symmetri-
decorating the plain weave, 1/1, warp-faced cal mirror images. On the contrary, they are
tunic. The image of the figure alternates with distinctively different from side to side. In
an image. of a square within a ~quare done in 110-Tumilaca/Cabuza decoration, the horizon-
the same technique. Between two figu- tal decorative stripe below the neck slit is al-
rine/square stripes, there is another color stripe ways slightly asymmetrical along its central
similar to the one used for decorated bags vertical axis (Figure 4). In Chiribaya decora-
done with supplementary warps without any tion, the sequence and width of color stripes
apparent design (Figure 13). This figure is on the tunics is always different on each side.
different from any other 110-Tumilaca/Cabuza Each side has the opposite combination of
or Chiribaya motif encountered. Although color and stripe width of the other (Figure 6).
this figure lacks the usual staves and the lines This phenomenon might not be immediately
radiating from its head, its face is depicted in a obvious to the observer, but the overall im-
frontal view and not in profile. Its general ap- pression of the decoration is not balanced.
pearance and temporal context suggest that it This ~ame idea was executed with more so,.
represents a simplification of a Tiwanaku de- phistication in tunics from Chiribaya Alta,
ity figure found on earlier Tiwanaku textiles where changes of color sequence and width
from the region (see Conklin 1983, especially did not occur from side to side, but rather from
T7 on p.15). front to back, with the change in stripe pat-
terns occurring at the top of the shoulder (Fig-
The weaving technique used to depict the ure 12). This type of change required a more
deity on this tunic is similar to the supple- sophisticated technical mastery of weaving,
mentary yam technique used in decorated because the change at the shoulder was done
bands of 110-Tumilaca/Cabuza and Chiribaya along a single weft yam. Even if we put the
bags. In addition, pa,rallelto the stripes bear- tunic on its side and we look for the "Horizon
ing the depiction of this deity are other stripes phenomenon" as one possible way of weaving,
ANDEAN PAST 5 (1998) -334

and thus observing,4 the tunic (Conklin signs (Weiner and Schneider 1989:1). Both
1986:125), the symmetry of the design is cultures selected motifs widely used in the
clearly broken. Neither the characteristic of Andean region at the time (tenth and eleventh
symmetry nor that of the "Horizon effect", centuries A.D.), including S shapes, triangles,
both typical of Tiwanaku textiles (Bird and and squares (see Figures 4, 8, 10, and 11)
Skinner 1974; Conklin 1970, 1983, 1986), are (Frame 1986:55; Rowe 1986:156-157). These
evident in either the 110-TwnilacaJCabuza or motifs were probably part of a pan-Andean
the Chiribaya textiles. Both of the coastal ideology common to many cultures in the re-
Osmore groups replaced the most notable Ti- gion at that time. But why did the 110-'
wanaku organizational principles with the dis- Tumilaca/Cabuza and Chiribaya people
tinctly different concept of "pseudo- choose. not to adopt Tiwanaku insignia after
symmetry". the collapse of the Tiwanaku state? Were not
the symbols historically recognized as illus-
A few Chiribaya tunics found at Chiribaya trating power and prestige, representing wide-
Alta embody yet another variant of the spread and powerful ideology? Why did the
"pseudo-symmetry" organizational principle, people of the valley simply. adapt the symbols
or the glide reflect pial (Ascher 1991:160). of the collapsed entity to their own use, util-
These tunics have a horizontal band of zigzag izing an already-known set of symbols to indi-
decoration under the neck done in continuous cate social status and economic power?
zigzag embroidery (Figure 7). This decoration
probably originated from the functional need Goldstein (1989) found evidence that sug-
to reinforce this part of the tunic, as described gestS deliberate destruction of many of the
earlier. . It is almost the same size as the typi- structures and burials at the Tiwanaku center
cal 110-Tumilaca/Cabuza loop-stitch embroi- at Omo. I suspect, with Bermann et al.
dery stripe, and shows a comparable "pseudo- (1989: 162), that the absence of the standing
symmetrical" organization. The stripe is ef- deity and Tiwanaku condor was part of the
fectively cut along its vertical center line, and same process. Both the people in the middle
the sequence of colors inverted, so that the top valley and those in the coastal Osmore Valley
of one side is depicted on the bottom of the rejected the Tiwanaku ideology as such. It is
other side, and vice versa. Yet, this band is possible that the overall structure of the ideol-
not similar in its shape or size to the 110- ogy persisted, because many decorative motifs
Twnilaca/Cabuza band, and can be clearly typical of the Tiwanaku repertoire remained in
distinguished from it. It had no boundaries, the Ilo-Tumilaca/Cabuza and Chiribaya tex-
no complicated geometry, and simply utilized tiles. The symbols such as the standing god
the "pseudo-symmetry" rule in its most sim- and the eagle that were associated directly
plistic way. with Tiwanaku dominance, however, simply
vanished from Ilo-Tumilaca/Cabuza and
The motifs' selected for each decoration Chiribaya textiles and ceramics, and the tradi-
were not random choices of geometrical de- tional Tiwanaku forms of symmetry were de-
liberately broken.

4 Conklin (1986) suggested that the correct way to look It is currently not completely clear how the
at South Andean Middle Horizon textiles should be populace of the middle Osmore Valley related
sideways because this is the way they were woven. The to the Tiwanaku state around A.D. 950. Their
concept of looking into the horizon, where the wider apparent rejection of the explicitly Tiwanaku-
lines represent areas closer to the observer, and thinner related aspects of the ideology, however, may
lines represents areas far from the observer is useful indicate that there was some conflict between
here. When a textile was constructed to make a gar- them and the representatives of the Tiwanaku
ment, however, instead of presenting the lines in hori- state in the valley. Evidence from Omo
zontal mode, the garment was constructed to present (Goldstein 1989) suggests that the Tiwanaku
the lines in vertical mode, making it difficult for the temple at the site was deliberately destroyed
untrained eye to notice the view-to-the-horizon concept by the local 110-Tumilaca/Cabuza people.
depicted on it.
335- Boytner: Osmore Textiles

Rejection of Tiwanaku dominance was re- at Yaral versus 0.3-0.49 mm at Chiribaya

flected, then, not only by demolishing power Alta), and much higher yarn densities (15-101
centers, but also by abolishing the symbols of c.p.s.c. at Yaral and EI Algodonal versus 112-
their power. The Tiwanaku icons served not 167 c.p.s.c. at Chiribaya Alta). Because of
as a source for power among the valley's peo- their high density, tunics from Chiribaya Alta
ple, but as reminders of being subject to pow- required substantially longer periods of time
ers they did not welcome. It seems that the act to weave. Unidentified fragments of cloth
of rejection was so powerful that it was re- from El Algodonal have about the same yarn
membered by the valley people immigrating thickness and density as the tunics, suggesting
to lower elevations, who never readopted Ti- that the general quality differences can be ex-
wanaku design modes. tended to much of the El Algodonal textile
assemblage.
Intra-Valley Social Stratification
The textiles from Chiribaya Alta were also
Inter-site differences in the textile assem- more finely finished than those from £1 Algo-
blages strongly suggest hierarchical status re- donal and Yaral, and greater amounts of work
lationships between the sites, and, indirectly, were invested in finishing. The use of more
among individuals. For example, there are than one color, particularly what seem to be
dramatic differences in the labor and material dyed colors, was much more common at
expended on bags from El Algodonal as com- Chiribaya Alta than at El Algodonal or Yaral
pared with those from Chiribaya Alta. The (Tables 7-10).
yarns used for bags from Chiribaya Alta. are
40-60% thinner than those from El Algodonal Although the finishing of bags from Chiri-
(0.8-2 mm. at El .Algodonal versus 0.39-0.7 baya Alta is more elaborate than that of bags
nun at Chiribaya Alta), while the density of from El Algodonal and Yaral, the differences
the textiles was higher at Chiribaya Alta (108- are only slight. Bags of both cultures were
156 c.p.s.c. at El Algodonal versus 170-194 evidently important objects in burials. They
c.p.s.c. at Chiribaya Alta) (Tables 3-4). Con- contained coca for the dead, and they were '

sidering that the bags from both sites are always highly decorated.. Because bags are
similar in size, it appears that people buried at relatively small, they required comparatively
Chiribaya Aha had bags that required signifi- little wool and labor to make. Even poor peo-
cantly more time to weave. ple could evidently afford to have relatively
elaborate bags (sometimes more than one) to
Comparing the tunics from El Algodonal accompany them in death. In some cases,
to those from Yaral, it is evident that the tu- bags were mended in various places. Despite
nics from Yaral are coarser and less dense being tattered, they were very elaborate in
(Tables 3-4). While the shapes of the tunics their weaving and decoration. These bags
from Yaral are similar to those from El Algo- were probably used and repaired by their bear-
donal, their crudeness suggests that the people ers for a long time because they could not af-
of Yaral could not afford as much weaving ford to acquire new ones.
time and investment in raw materials. It
seems that the Yaral people were poorer than At £1 Algodonal, as well as at Yaral, only
the people of El Algodonal. a small fraction of the textiles are decorated,
and those that are have only minimal amounts
As was the case with bags, the tunics from of decoration. Some of the decorated textiles
Chiribaya Alta are finer and denser than those were used for a long time, and were mended
from El Algodonal and Yaral. Although there before the person was buried. This was not
is not much difference in weft thicknesses the case at Chiribaya Alta. In burials of both
(0.75-1 mm at El Algodonal, 0.87-2 mm at cultures at Chiribaya Alta, the decorative
Yaral versus 0.6-1.98 mm at Chiribaya Alta), techniques were the same as at El Algodonal
Chiribaya Alta tunics have much thinner and Yaral, but the density was higher (Tables
warps (0.6-0.74 mm at El Algodonal, 0.69-1.7 5 and 11). Ilo-Tumilaca/Cabuza tunics from
ANDEAN PAST 5 (1998) -336

looted surface contexts at Chiribaya Alta have lent size El Algodonal and Yaral textiles, be-
side stripes up to 2S loop-stitch rows wide, cause the decorated areas are larger and the
although the neck decoration is about the same decorative techniques more complex.
as at El Algodonal. The designs in these
stripes are more elaborate, and involve com- Finally, there is a greater variety of textile
plex geometrical forms (Figure 10). Although types at Chiribaya Alta than at El Algodonal.
the range of colors is the same and red is still Hats, "handkerchiefs" (panuelos), belts, bag-
the dominant color, the greater area of decora- belts ifaja-bolsas), and other types that are.
tion and the use of more colors in any given rare at El Algodonal are common at Chiribaya
stripe allowed the tunics to create a stronger Alta. All in all, the textile evidence seems to
impression. reflect the location of Chiribaya Alta above
the valley floor, on top of the steep slopes, far
The relative elaboration of decoration at from water resources, and surrounded by forti-
Chiribaya Alta is even greater in Chiribaya fication walls. This apparently was a central
textiles. Although the same colors are used, site and elite cemetery for the valley.
many Chiribaya Alta tunics have the addi-
tional complication of color sequences that The clear site-level status differences in
change at the top of the shoulder, and many the textiles suggest that both cultures were so-
have elaborate zoomorphic and abstract mo- cially stratified. The site of Chiribaya Alta
tifs,most commonly a lizard figure, but also was used by the elite of both cultures, and it is
including serpents, felines, and birds. These possible that the two cultures shared not only
motifs are executed in the same supplementary the same site hierarchy and geographical envi-
yarn technique that is used on bags. This ronment, but also the same social structure.
technique must have increased the cost of .the The textiles alone do not reveal the specific
tunic, because it requires not only more col- nature of the social hierarchy, nor the exact
ored yarns, but also a completely woven sur- relationships between the two cultures. Fur-
face below the decoration, and hence larger ther research integrating the textile record with
amounts of wool for the same size tunic. Such other elements of the material culture may
sophisticated decoration also implies access to clarify this issue.
more skilled labor.
Gender Differences
Colored yarns were evidently valuable,
because in 90% of both bags and tunics, only The tunics of both cultures come in rec-
the warp yarns are colored. These articles tangular and trapezoidal varieties. Clark
look completely colored, even though only (1988:14) suggested that rectangular tunics
some of their yarns were colored. This was an were worn by men and trapezoidal tunics were
efficient way to minimize costs without sacri- worn by women at the Late Intermediate pe-
ficing appearance. riod site of Estuquifia in the upper Osmore
drainage. At El Algodonal, however, there
Unfortunately, it is difficult to estimate the was no evidence of differentiation in tunic
time required to make these textiles. As shape by gender at any of the investigated
shown by Franquemont in his ethnographic cultures. This may be due to the poor quality
research (1984:322-323), weaving time is of preservation and the small number of com-
highly variable, depending on the loom and plete tunics with individuals of known sex
spinning technique, the age of the weaver, the from El Algodonal. Both the 110-Tumilaca/
time of year when the weaving was done, and Cabuza and the Chiribaya sample suffer from
the quality of wool. Nevertheless, Fran- these limitations.
quemont's work suggests that these textiles
would have taken hundreds of hours to pro- There appear to be no significant differ-
duce. Whatever the exact time was, the time ences between bags buried with males and
spent to make the Chiribaya Alta textiles females of the 110-Tumilaca/Cabuza culture at
would have been greater than that for equiva- El Algodonal. No differences in yarn thick-
337- Boytner: Osmore Textiles

ness of wefts or warps were observed (Tables textiles from the two cultures strongly sug-
3-4). Bags buried with males were slightly gests cooperation and coexistence in the
denser than those buried with females (Table coastal Osmore Valley. Nevertheless, the in-
5), but the 10% difference in density is proba- herent differences in the symmetrical organi-
bly not significant given the large standard zation of designs provide clear evidence of the
deviation of the sample. distinctiveness of each cultural identity.
Ascher (1991) claims that although there is no
No gender preferences were detected in necessary functional explanation for the rela.,
finishing types in either culture. Nor are there tionships between the spatial design organiza-
any evident gender differences in decorative tion and the cultural interpretation of the
motifs at EI Algodonal (Tables 12-13). The world, this relationship does exist. Moreover,
sample of Chiribaya decorative motifs is too it reflects distinctive and particular cultural
small to assess gender differences, but at least sets of ideas and beliefs. Thus, the differences
among the 110-Tumilaca/Cabuza people, there in the organization of design in the two cul-
were evidently no specific textile motifs that tures reflect the differences between them, and
were strongly associated with either gender. provide an indication that each is a distinct
and individual cultural system.
Overall, the textiles do not suggest any
gender preferences in weaving techniques, Two important issues remain unanswered.
finishing, or even decoration. It. is, however, First, what is the true nature of the relationship
important to note that the number of individu- between the 110-Tumilaca/Cabuza and Chiri-
als of known gender from El Algodonal is baya cultures? How is it that these two cul-
small, and it is possible that such preferences tures engaged in such close relations, while so
simply cannot be detected in:this sample. The strongly maintaining ahd emphasizing their
apparent absence of gender differences is par- individual identities? Second, there is the
ticularly surprisinR because Jessup (1991) re- question of how and why the 110-Tumilaca/
ported that in his excavations in Chiribaya Cabuza culture disappeared from the valley. It
cemeteries, he found different ceramics in is possible that the competition within the
burials of males and females. Societies in symbiotic relationship of the two cultures fi-
which males and females share the same status nally ended the 110-Tumilaca/Cabuza people's
are not unknown in the Andes, but they are not ability to control resource~, forcing them out-
common. Only future investigation of textiles side the valley. The absence of signs of war-
from other coastal Osmore Valley sites would fare and the lack of evidence of diffusion or
resolve this problem and will either support or blending suggests that the 110-Tumilaca/
contradict the above observation. Cabuza people probably left the valley peace-
fully, rather than integrating into the Chiribaya
Conclusions society. This is, of course, only one possibil-
ity, and should be addressed using the full
The textiles analyzed here comprised only range of available archaeological data.
a small segment of the cultural universe of the
110-Tumilaca/Cabuza and Chiribaya people. Although the main purpose of this research
The conclusions offered serve best to bolster was to define differences between the two
other lines of evid~nce, and to suggest hy- coastal valley cultures, it also helped define
potheses for future research. One such con- other questions. It addresses issues regarding
clusion is that textiles point to a common 110-Tumilaca/Cabuza and Chiribaya ideologi-
highland origin for both 110-Tumilaca/Cabuza cal and political structures. Furthermore, it
and Chiribaya cultures. The absence of Ti- examines a process of change in these struc-
wanaku deities in textile decoration is signifi- tures that occurred by the end of the Middle
cant, and combined with the evidence of in- Horizon. By looking at its margins, we may
tentional destruction of the Tiwanaku center at better understand the collapse of the Tiwanaku
Omo, can be interpreted as a rejection of Ti- state and the events that followed the destruc-
wanaku ideology. The great similarity of the tion of central authority. Once we understand
ANDEAN PAST 5 (1998) -338

these processes in the lower Osmore Valley, D.S. Rice, C. Stanish, and P.R. Scarr, pp. 269-286.
BAR International Series 545ii.
we may apply our model to other regions of Bird, Junius B. and M.D. Skinner
the Tiwanaku state and examine the results. 1974 The Technical Features of a Middle Horizon
Eventually, we may be able to achieve a better Tapestry Shirt from Peru. Textile Museum Journal
understanding of the general ideological and 4(1):5-11.
political processes of the Andean region. Boytner, Ran and Arie Wallert
1993 Dyes from the Tumilaca and Chiribaya Cul-
tures, South Coast of Peru: What Can We Learn?
Acknowledgements Paper presented at the 33rd Annual Meeting of the'
Institute of Andean Studies, Berkeley, California.
I wish especially to recognize Bruce Burgess, Shelly D.
Owen's assistance in conducting my research n.d. Human Remains: Preservation and Demogra-
and in writing this report. There was no limit phy. In Mortuary Investigations at Chiribaya Alta:
A Late Intermediate Period Site Near 110, Peru,
to his willingness to share data and experience edited by Sloan Williams and Jane E. Buikstra (iq
with me at any time and place. I do, however, press).
bear sole responsibility for the interpretation Clark, Niki R.
of these textiles. Thanks go to Timothy Earle . 1988 Archaeological Textiles in Their Socio-
and Christopher Donnan for their constructive Cultural Context: Late Prehistoric Period, Es-
tuquifta Site, Far South Peru. Paper presented at
criticism and helpful advice in the various the 16th Midwestern Conference on Andean and
stages of this research. Thanks are also due to Amazonian Archaeology and Ethnohistory. Ann
Jane Buikstra for granting access to material Arbor, Michigan.
from Chiribaya Alta and Yaral, as well as to Collingwood, Peter
Shelly Burgess, Marla Cecilia Lozada, Liliana 1987 Textile and Weaving Structures. London: B.T.
Batsford Ltd. .
Ulloa Torres and to the rest of the staff at the
Conklin, William J.
Azapa Museum of the University of Tarapaca 1970 Peruvian Textile Fragment from the Beginning
at Arica, Chile. Niki Clark, Cecilia Quequ- of the Middle Horizon. Textile Museum Journal
ezana, and Manuel Patcheco deserve thanks (I): 15-24.
for their enthusiasm i~ sharing data, and for 1983 Pucara and Tiahuana.so Tapestry: Time and
their interest and dedication to the under- Style in Sierra Tradition. Nawpa Pacha (21):1-44.
1986 The Mythic Geometry of the Ancient Southern
standing of Andean prehistory. I would espe- Sierra. In The Junius Bird Conference on Andean
cially like to acknowledge Patricia Anawalt Textiles, edited by Ann P. Rowe, pp. 123-135.
and Barbara Sloan of the Center for the Study Washington D.C.: The Textile Museum.
of Regional Dress at the Fowler Museum D'Harcourt, Raoul
1962 Textiles of Ancient Peru and their Technique.
(UCLA), and Katherine Spilker of the Los Seattle: University of Washington Press.
Angeles County Museum. Southern Peru Frame, Mary ,

Copper Company (SPCC) provided valuable 1986 The Visual Images in Fabric Structures in An-
support for this research. Lastly, I would like cient Peruvian Textiles. In The Junius Bird Con-
to offer my deepest appreciation to Karen ference on Andean Textiles, edited by Ann P.
Rowe, pp. 47-80. Washington D.C.: The Textile
Marshik who dedicated long, patient hours to Museum.
correcting and re-correcting many drafts of Franquemont, Edward M.
this paper. I could never thank her enough. 1984 Cloth Production Rates in Chinchero, Peru. In
The Junius Bird Conference on Andean Textiles,
References Cited edited by Ann P. Rowe, pp. 309-329. Washington
D.C.: The Textile Museum.
Albers, Anni Goldstein, Paul S.
1963 On Weaving. Middletown: Wesleyan Univer- 1989 Omo, A Tiwanaku Provincial Center in Mo-
sity Press. quegua, Peru. Ph.D. dissertation, Department of
Ascher, Marcia Anthropology, University of Chicago.
1991 Ethnomathematics: A Multicultural View of Hodder, Ian
1979 Economic and Social Stress and Material
Mathematical Ideas. Pacific Grove, California:
Brooks/Cole Publishing Co. Culture Patterning. American Antiquity 44(3):446-
454.
Bermann, M., P. Goldstein, C. Stanish, and L.M.
Watanabe Jessup, David A.
1989 The Collapse of the Tiwanaku State: A View 1991 General Trends in the Development of the
from the Osmore Drainage. In Ecology, Settlement ChiribayaCulture,South CoastalPeru. Paperpre-
and History in the Osmore Valley, Peru, edited by
339- Roytner: Osmore Textiles

sented at the 56th Annual Meeting of the Society 1958b A Peruvian Tasseled Fabric. Notes in Anthro-
for American Archaeology, New Orleans. pology 3. The Florida State University, Depart-
Lothrop, S.K. and Joy Mahler ment of Anthropology.
1957 A Chancay-Style Grave at Zapallan, Peru: An Wallert, Arie and Ran Boytner
Analysis of its Textiles, Pottery, and other Fur- 1996 Dyes from the Tumilaca and Chiribaya Cul-
nishings. Papers of the Peabody Museum of Ar- tures, South Coast of Peru. Journal of Archaeo-
chaeology and Ethnology 50( 1). logical Science 23:853-861.
Lozada, Maria C. Weiner, Annette B. and Jane Schneider
1991 Mortuary Excavations at La Yaral, Southern 1989 Introduction. In Cloth and Human Experi-
Peru. Unpublished Manuscript. ence, edited by A.B. Weiner and J. Schneider, PI>'.
Lozada, Maria C., E. TOlTes,and Jane E. Buikstra 1-29. Washington, D.C.: Smithsonian Institution
1991 New Chiribaya Funerary Patterns at La Yaral, Press.
Southern Peru. Paper presented at the 90th Annual Wobst, Martin H.
Meeting of the American Anthropological Asso- 1977 Stylistic Behavior and Information Exchange.
ciation, Chicago, Illinois. In For the Director: Research Essays in Honor of
Murra, John V. James B. Griffin, edited by C.E. Cleland, pp. 317-
1962 Cloth and Its Function in the Inca State. 342. University of Michigan, Museum of Anthro-
American Anthropologist (62):710-728. pology, Anthropological Papers 61.
1989 Cloth and Its Function in the Inca State. In
Cloth and Human Experience, edited by A.B.
Weiner and J. Schneider, pp. 275-302. Washing-
ton: Smithsonian Institution Press.
Oakland, Amy
1986 Tiahuanaco Tapestry Tunics and Mantles from
San Pedro de Atacama, Chile. In The Junius Bird
Conference on Andean Textiles, edited by Ann P.
Rowe, pp. 101-121. Washington D.C.: The Textile
Museum.
Owen, Bruce
1991a Colonization and Complexity in the Twilight
of Tiwanaku: The Coastal Osmore Valley, Peru.
Paper presented at the 56th Annual Meeting of the
Society for American -Archaeology,New Orleans.
1991b Proyecto Colonias Costeras de Tiwanaku,
1989-1990. Field report.
1992 Coastal Colonies and the Collapse of Ti-
wanaku: The Coastal Osmore Valley, Peru. Paper
presented at the 57th Annual Meeting of the Soci-
ety for American Archaeology, Pittsburgh.
1993 A Model of Multiethnicity: State Collapse,
Competition, and Social Complexity from Ti-
wanaku to Chiribaya in the Osmore Valley,.Peru.
Unpublished Ph.D. Dissertation, Department of
Anthropology, UCLA.
Roach, M.E. and lB. Eicher (eds.)
1965 Dress, Adornment, and the Social Order. New
York: John Wiley and Sons, Inc.
Rowe, Ann P.
1984 Costumes and Feather Work of the Lords of
Chimor. Washington D.C.: The Textile Museum.
1986 Textiles from the Nasca Valley at the Time of
the Fall of the Huari Empire. In The Junius Bird
Conference on Andean Textiles, edited by Ann P.
Rowe, pp. 151-182. Washington D.C.: The Textile
Museum.
Uhle, Max
1903 Pachacamac. Department of Archaeology of
the University of Pennsylvania, Philadelphia.
VanStan, Ina
1958a Problems in Pre-Colombian Textile Classifi-
cation. Tallahassee: Florida State University.
ANDEAN PASTS (1998) -340

1"';:':1Padftc ocean
_
- Arable floodplain
Osmore river

Figure 1. The Osmore valley (from Owen 1991a).

 341- Boytner: Osmore Textiles

Figure 2. Variation of shapes among tuni~s.

neckslit top
top
.1
~
, ,
,
under-the-neck
stripe band
(stri :)
-- --
... +- side
side~

I . I ,\
i
bottom
i
bottom

Figure 3. Key to specific area locations and terminology used in the text: left, tunic; right, bag.
ANDEAN PAST 5 (1998) -346

111111 III I II III III

III III III 11./

III III

III III III III III III

III II III III II III

II\ IIIIII II\ II)III

III III III III 1\/ll'

[] IJ · .
c

Figure 13. The tunic with the front face deity found on the looted surface at El Algodonal.
 w
~
j.J
EL ALGODONAL CHIRIBA YA ALTA YARAL

I-TIC CHIRIBAYA I-TIC CHIRIBAYA I-TIC CHIRIBAYA

SEX M F N/A TO M F N/A TO M F N/A TO M F. NIA TO M F N/A TO M F N/A TO

BAG 6 2 7 15 2 2 4 1 -I 8 3 9 20

TUNIC 17 18 41 76 4 4 2 18 20 1 1 2 7 9

UNKNOWN 16 19 90 125 2 1 3

2-PLY YARN 2 4 7 13 3 3

RE-PLmDYARN 10 9 27 46 2 3 1 6 3 4' 7

ROPE 4 4 4 12 3 3 1 1

HANDKERCHIEF 2 3 S 4 4
(PANUELO)

OTIIERS 12 1 13 1 1 1 1 2 2

EL ALGOOONAL CHIRIBAY A AL TA YARAL

EXCAVATION 99 54 24

LOOTED SURFACE 218 o o

~
~:'!

i~
(Top) Table 1. Textile types found on each site.
(Bottom)Table 2. Collectiontypes.
~
:::to
~
c.)
 EL ALGODONAL CIURIBAYAALTA YARAL

I-TIC CHIRIBAYA I-TIC CHIRIBAYA I-TIC CHIRIBAYA i

SEX M F N/A M F N/A M F N/A M F N/A M F N/A M F N/A
~
(fl
BAG N 6 1 3 1 1 2 8 too3

SIZE 0.97 1.0 0.8 0.8 2.0 0.7 0.39 -.-

(II

\0
TUNIC N 14 15 34 2 1 3 8 18 1 2 7 \0

. SIZE 0.96 0.97 0.7S

~
. 1.0 .1.0 0.6 1.98 0.72 2.0 1;5 0.87

UNKNOWN N 12 13 62

SIZE 1.01 1.11 097

EL ALGODONAL CIURIBAYAALTA YARAL

I-TIC CHIRIBAYA I-TIC CHIRIBAYA I-TIC CHIRIBAYA

SEX M F N/A M F N/A M F N/A M F N/A M F NIA M F N/A

BAG N 6 1 3 2 1 2 8
SIZE 0.93 1.0 0.6 1.0 1.0 0.55 0.25
TUNIC N .14 15 34 2 1 3 8 18 1 2 7
SIZE 0.72 0.66 0.74 0.6 0.8 0.3 0.41 0.49 1.7 1.0 0.69
UNKNOWN N 12 13 62
.
SIZE 0.73 0.97 0.71

(Top) Table 3. Weft thickness (average thickness in millimeters).

(Bottom) Table 4. Warp thickness (average thickness in millimeters). ~
00
 W
01:>-
\0
I
EL ALGOOONAL CIURIBAYAALTA . YARAL
I-TIC CIDRIBAYA I.T/C c:IURIBAYA I-TIC CIURIBAYA
SEX M F N/A M F N/A M F NlA M F NIA M F NIA M F N/A
BAG N 5 2. 7 2 2 3 7
SIZE 156.2 108 144 112.5 82.5 170 194
TUNIC ,., 16 17 34 4 I 3 5 17 I I 7
SIZE 91.9 101.2 97.9 83.2 112 167.3 127 164 15 IS 99.1
UNKNOWN N IS 16 86 2 1
SIZE 86.6 89.5 98.8 71.5 187

EL ALOOOONAL CHIRIBAYA ALTA YARAL

I.T/C CHIRIBAYA I-TIC CHIRIBAYA I.T/C CHIRIBAYA

WARP FACE 105 9 3 35 I 7

WARPFACE+ 23 3 1 I 4
REINFORCMANT WARP

MIXED WI WARP FACE 12 6

MIXEDW/OUTWARP 1
FACE . ~
~
BASKET (212) 1 ~
:"!
SEMI BASKET (211) 4
5?
INTERCOURSING 1
~
~
(fop) Table 5. Yam count (weft x warp per centimeter).
(Bottom)Table 6. Main weavingtechniques.
~
:::to
~
~
 EL ALGODONAL CHIRIBAYA ALTA YARAL

I-TIC CHIRIBAYA I-TIC CHIRIBAYA ' I-TIC CHIRIBAYA i

SEX M F N/A TO M F N/A .TO M F NIA TO M F N/A TO M F N/A TO M F N/A TO ~
rn
~
NO WEFT FINISH

NO WARP FINISH 6 3
1

7
1
16 1 1 S 12 17 1 1 1 3 4
--
CA

\0
\0
REGULAR 1 1 1 1 '-"
00
EMBROIDERY
-
DENSE 1 S 2 8 1 1 1 S 6 1 2 3
EMBROIDERY

LOOP STITCH: 2 2 1 1
ONE COLOR "

LOOP STITCH: 1 1 2
MULTI COLOR

BOUBLELOOP 1 1 1 3 2 2 4 1 2 3
STITCH:COLOR
IN PATTERN

BOUBLELOOP
STITCH:COLOR
IN TRAINGLE

LOOP STITCH
STRIPE PART OF
FINlSH

EMBROIDED
LINES IN
2 2
.
SLlVEINECK .

I
W
Table 7. Sleeve finishing techniques. U\
o
 w
VI
.....
I
EL ALGODONAL CHIRIBA YA ALTA YARAL
I-TIC CHIRIBAYA I-TIC I-TIC I-TIC CHIRIBAYA
SEX M F N/A TO M F N/A TO M F N/A TO M F N/A TO M F N/A TO M F /A TO
NO WARPFINISH 3 3 4 10 2 2 1 l. 2 2 4 22
REGULAR
EMBROIDERY

DENSE 1 2 3 6 1 1 3 3 1 1 1 2 3
EMBROIDERY

LOOP STITCH. 6 8 8 22 1 1 1 1
ONE COLOR

LOOP STITCH. 2 2
MULTI COLOR

DOUBLELOOP 3 1 4 2 2 2 3 5
STITCH-COLOR
IN PATTERN

DOUBLELOOP
STITCH-COLOR
IN TRIANGLES

LOOP STITCH 1 1
STRIPEPART OF
FINISH

EMBROIDED 2 4 6 1 1
LINES IN
,
SLEEVEINECK . ~
~:"!
~
~
~
~
::t
Table 8. Neck fmishingtechniques. -:
~
 I-TIC
EL ALGODONAL

. CHIRIBAYA
CHIRIBA YA AL TA

I-TIC I-TIC I-TIC

YARAL

CHIRIBAYA
i
~
SEX M F N/A TO M F N/A TO M F N/A TO M F N/A TO M F N/A TO M F N/A TO ~
f:/.!
NO WARP FINISH 4 3 34 41 1 1 4 4 1 1
-
IooiJ
(.A
REGULAR 14 19 41 74 6 6 1 1 2 8 17 25 1 1 3 7 10 .....
EMBROIDERY \0
\0
00'
DENSE 3 1 4 1 1 1 1 --
EMBROIDERY

LOOP STITCH- 2 2 4
ONE COLOR
LOOP SmCH-
MULTI COLOR

OOUBLE LOOP 3 2 5 3 3
smCH-COLOR
IN PATTERN

OOUBLE LOOP 1 1
SmCH-COLOR
IN TRIANGLES

LooPSmCH 3 7 10 2 2 3 3
STRIPEPARTOF
FINISH

Table 9. Side fmishingtechniques. w,

u.
t-.)
 w
UI
W
I
EL ALGODONAL CHIRIBAYA ALTA YARAL

I-TIC CHIRIBAYA I-TIC I-TIC I-TIC CHIRIBAYA

SEX M F N/A TO M F N/A TO M F N/A TO M F N/A TO M F N/A TO M F N/A TO
NO WEFT FINISH 1 1
WEFT 18 27 46 91 6 1 1 8 2 '2 8 1 19 28 1 1 3 2 S
REINFORCMENT

WEFT 3 1 3 7 1 1
REINFORCMENT
+LOOP S'ITfCH

WEFT 2 2 IS 19 1 1 2 2 S 7 1 1
REINFORCMENT
+EMBROIDERY

DENSE 1 1
EMBROIDERY

LOOP STITCH- 1 1 1 . 1
ONE COLOR

LOO SMCH- 3 3
MULTI COLOR

DOUBLE LOOP 4 4
STITCH-COLOR
IN PATTERN
.

~
~
~
:'!
&
~
~

Table 10. Shirt bottom and bag top fmishingtechniques.

~
:::to
~
~
 SEX M
I-TIC

F
EL ALGODONAL

N/A TO M
CHIRIBAYA

F N/A TO M F
CHIRIBAYA

I-TIC

N/A TO M
ALTA

CHIRIBAYA

F N/A TO M
I-TIC

F N/A
YARAL

TO M
CHIRIBAYA

F N/A TO
i~
(I)
WEAVED COLOR
STRIPES
2 3 22 27 5 5 1 1 7 8 15 1 4 5
-...-
too3
<.18

DIVIDEDLINE IN
SEQUENCE

EMBROIDERY
4 I' 18

1
23

1
1

1
1

1
3 3

.
-
\0
\0
00

UNDER.THE-NECK
ONLY

EMBROIDERY. 2 3 8 13 3 3
SIDES+NECK

EMBROIDERY- 2 2
NARROW
NECK+WIDE SIDES

BAG STRIPES- 1 2 2 5 1 1 2 2
EQUAL WIDTH

BAG WlTII WIDER 4 3 7 2 2

CENTRAL STRIPE

BROCADE 3 4 7

Table 11. Decoration techniques and location on cloth.

~
 w
VI
VI
I
NECK SIDE STRIPE

I-TIC. CHIRIBAYA I-TIC CHIRIBAYA I-TIC CHIRIBAYA

SEX M F N/A TO M F N/A TO M F N/A TO M F N/A TO M F N/A TO M F N/A TO

roE A /\n.. 1 2 1 4

TYPEBAyA. 1 1 2 ' 4

HH
TYPE C
ED " 1 1
1 1

TYPE E e.ra 1 1 1 1 2

TYPE F ....m 4 4
1.1 2 6
TYPEG .. tit 5 13

TYPE A+B 1 1 1 2 3
..
TYPE E+F r 1

ELABORATED 1 1 2
FORMS

ZIG.ZAG 4 4

ZIO ZAG WI 1 1
CHANGING COLORS

REINFORCEDLINES 2 2 4 2 .1 3
UNDERNECK
b:I
~
ii
:"!

&
~
~

Table 12. Embroidered motif types by sex ~d culture at EI Algodonal.

~
:::t
~
~
 /"
NECK SIDE STRIPE
I-TIC CHIRIBAYA I-TIC CHIRIBAYA I-TIC CHIRIBA YA i
SEX M F N/A TO M F N/A TO M F N/A TO M F N/A TO M F N/A TO M F N/A TO ~
(I)
nuN COLORLINE 2 15 F ~
ON THE SIDE

WIDESTRIPES-NO 2 1 7 10 ./
-.-
tit

\0
CHANGE \0

WIDE STRIPE-
--
00
S 5
CHANGE BETWEEN
SIDES

DIVIDED LINE 4 1 18 23 1 1
PATTERN
SEQUENCE

Table 13. Woven motif types by sex and culture at El Algodonal. I

~
VI
0\