Ecology, Settlement and History in the Osmore Drainage, Perú

Edited by

Don S. Rice, Charles Stanish and Phillip R. Scarr

part ii BAR

International Series 545(ii)1989 


This paper is presented as it appeared in the 1989 publication with the exception that the graphics havebeen improved.



The Collapse of the Tiwanaku State:A View from the Osmore Drainage

Marc Bermann
Museum of Anthropology
University of Michigan
[Currently: Department of Anthropology, University of Pittsburgh]

Paul Goldstein
Department of Anthropology
University of Chicago
[Currently: Department of Anthropology, Dartmouth College]

Charles Stanish
Department of Anthropology
Field Museum of Natural History
[Currently: Department of Anthropology, University of California, Los Angeles]

Luis Watanabe M.
Director, Programa Contisuyu
Moquegua, Peru




The collapse of a pre-industrial empire can have a wide range of effects on provincial or peripheral areas. Possible post-imperial changes can range from large scale abandonment of the region, to a "balkanization" into small, competing communities, to the regional administration breaking off intact, while retaining the internal social, political and ideological structures of the once-ruling imperial system. Preliminary investigations in the Osmore drainage suggest that the passage of the region from Tiwanaku control (ca. A.D. 1000) resulted in a complex settlement shift that is still incompletely poorly understood.

By A.D. 700, the Tiwanaku state, centered at the site and valley of the same name in modern Bolivia, was the dominant political power in the Titicaca Basin. Characterizes by an urban capital, smaller satellite urban centers and massive rural sustaining population, Tiwanaku constituted one of the earliest examples of Andean imperial statecraft, establishing patterns for later trans-regional systems, including the Inca empire and the much smaller Titicaca Basin polities such as the Lupaqa and the Colla. Tiwanaku sites have long been famous for their public architecture and sculpture. Recent research has demonstrated that they also were true urban centers (Kolata 1982, 1983, 1986; Kolata et al. 1987). By A.D. 900, the largest of these centers, Tiwanaku itself, had a total estimated occupational area exceeding 4 km and may have reached a population of 30-40,000 (Parsons 1968; Ponce Sangines 1972), and possibly even larger (Stanish 1987).

The urban concentration at Tiwanaku was part of a much larger settlement system supported by massive raised field agricultural areas along the edge of Lake Titicaca and Rio Desaguadero (Kolata 1982; Smith et al. 1968). Sites such as Lukurmata, Pajchiri and probably Khonko Wancané (Wancani), among others, were associated with large areas of raised fields which provided the economic production necessary for supporting this vast settlement system (Kolata 1982; Smith et al. 1968). Recent investigations at Lukurmata have documented a pre-Tiwanaku (Tiwanaku III) occupation suggesting that this pattern of population concentration and intensive lakeside and marsh riverine land use had Formative Period antecedents (see Kolata 1982; Smith et al. 1968; Bermann 1987; Wise 1988).

These local, labor-intensive agricultural systems were augmented by substantial camelid herding (Lynch 1983a) and the creation of inter-regional economic networks throughout the South Central Andes (Browman 1980, 1984; Kolata 1983; Mujica et al. 1983; Núñez and Dillehay 1979). The vast reserves of camelids served not only as sources of meat and wool, but also were used as pack animals to transport goods over the rugged terrain of the region. Colonies were founded in western slope valleys (Goldstein 1985; Kolata 1983) and exchange networks developed between the extant complex societies throughout the region (Berenguer et al. 1980).

The Tiwanaku state developed gradually over a millennium. At its height, the Tiwanaku political economy had coalesced into a vast and complex system incorporating an intensive local economy based upon 1) raised field agriculture (Smith et al. 1968; Kolata 1972), 2) rainfed tuber production (Ponce Sangines 1972), 3) intensive animal husbandry (Lynch 1983a), 4) the establishment of distant colonies (Kolata 1983; Mujica et al. 1983; Foccaci 1969), 5) the development of extensive regional exchange networks (Berenguer et al. 1980) and 6) the possible presence of market systems in the core region itself (Browman 1980, 1984). By approximately A.D. 900 Tiwanaku influence had reached an economic and political apogee both within and outside of the Titicaca Basin and constituted a true pre-industrial, imperial system.

The accumulated evidence to date suggests that the collapse of the Tiwanaku state was a long process characterized by a gradual, but temporally and geographically uneven, decline in geo-political control throughout the South Central Andes. Recent C-14 determinations from initial Chiribaya context at the site of Yaral indicate a maximum date for Tiwanaku ceramics at around A.D. 1000 (Rice and Conrad personal communication) and almost certainly reflect a final date for Tiwanaku influence in the Moquegua Valley as a whole. The decline of Tiwanaku control may have been even earlier in Northern Chile (Foccaci 1969) where the late Tiwanaku ceramic assemblage, known as Maitas, has been dated to A.D. 700 and 980. This style shows decided deviations from typical Tiwanaku 5 assemblages towards the post-Tiwanaku Tricolor del Sur tradition (Lumbreras 1974a; Ponce Sangines 1972) including the development of bowl forms, eight-point stars and changes in the use of space and motif placement on the vessels themselves.

These data from the peripheral zones contrast with the core territory itself where the distinctive Tiwanaku ceramic styles appear to have continued on to at least A.D. 1200, if not slightly later (Ponce Sangines 1972; Oswaldo Rivera personal communication). In gross terms, a distinction between core and periphery in prehistoric empires serves to model the uneven collapse of Tiwanaku through time. That is, peripheral or colonial areas fall out of the imperial political economic orbit well before the core. Core areas, in contrast, tend to maintain the imperial system of political and economic control after the regional system has collapsed.

By A.D. 1200, both the periphery and core areas of the Tiwanaku empire had disintegrated and the ethnohistorically known Titicaca Basin señoríos (kingdoms) were developing in the former core area of the empire. Several contemporary and interrelated events marked this collapse. Most dramatic was the abrupt end of the manufacture of distinctive Tiwanaku iconography on ceramics, textiles and wooden objects (Ponce Sangines 1972). The characteristic ceramic forms, such as the kero, disappeared completely and the motifs associated with the state at its political and economic zenith were also largely abandoned by artisans throughout the areas of former imperial control. Many of the iconographic elements commonly utilized for almost a millennium by Tiwanaku artisans were never seen again in the South Central Andes.

A more-or-less rapid abandonment of the major Tiwanaku 5 administrative/ritual centers, including its capital, coincided with the disappearance of Tiwanaku-style ceramics. While a complete demographic collapse of these core area sites is not indicated by current research (at least at the site of Lukurmata ñ see Wise and Stanish 1988), there was considerable population decline at these sites with concomitant abandonment of the major ceremonial/elite sectors on these formerly urbanized and complex settlements.

A third significant event to mark the collapse of the Tiwanaku empire was the apparent cessation of the raised field-based endogenous economy which supported the core area of the imperial system. Large areas of field systems in the Titicaca Basin fell into disuse by the initial post-Tiwanaku periods (Kolata 1986). These intensive and highly productive agricultural systems were either replaced by the more extensive hillside terrace systems still seen throughout the region today, became inefficient as intensive camelid grazing increased in importance or a combination of these two processes. In any event, the mainstay of the core area economy was abandoned in favor of more extensive systems of agricultural production and a greater relative reliance upon camelids.

A fourth event, or series of events, is the uneven abandonment of Tiwanaku colonial holdings throughout the South Central Andes, such as Osmore. As indicated above, this process was not a sudden and short-term event, but appears to have been the result of a gradual process beginning toward the end of the 10th century A.D.


Tiwanaku in the Osmore drainage

That the Moquegua region contains substantial Tiwanaku remains has been documented in the unpublished work of Ghersi, Vecelius, and in other published materials (Disselhoff 1968a; Geyh 1967; and see Mujica et al. 1983). The extent of this presence was only recently revealed, however, through the efforts of the Programa Contisuyu beginning in 1981. Seventeen sites in the Osmore drainage can be linked to Tiwanaku on the basis of artifactual and architectural evidence. The earliest, well documented Tiwanaku sites in Moquegua date to the Tiwanaku 4 Period (and included M-31, and the M-12 and M-16 components at the site of Omo ñ Goldstein 1985). None of these sites have associated monumental architecture.

The majority of the Osmore Tiwanaku sites have yielded material corresponding to the Tiwanaku 5 or expansive period, and include the massive cemeteries at Chen Chen (M-1) in Moquegua and the Loreto Viejo (C-33) in the Ilo Valley and the residential and administrative site of Omo (M-10) in the mid-Moquegua Valley (Disselhoff 1968a; Goldstein 1985).

The expansion of the Omo occupation was a major development of the Tiwanaku 5 period. The Omo complex, covering nearly 4 km , consists of five separate components. The M-10 (Tiwanaku 5) component displays the most striking architecture in the valley: stone and adobe platform measuring 120 m on a side (Goldstein 1985:31-33). Like the altiplano Tiwanaku pyramids, the tiered platform was built in stages. It is abutted by a raised platform containing only cist tombs. The Omo site exhibits marked spatial patterning in architecture and the distribution of craft-produced artifacts. Room blocks are divided in plaza groupings (Goldstein 1985:31-33) and artifact distributions suggest the existence of specialized craft workers, perhaps under a centralized administration.

The evidence documenting the increased Tiwanaku interaction with Moquegua during the Tiwanaku 5 period suggests that this interaction took the form of directly administered occupations. Significant changes between the Tiwanaku 4 and 5 periods in Osmore included a population increase, construction of public architecture, and increased standardization in ceramic forms and decoration (Goldstein 1985). The Tiwanaku 5 period also saw the introduction of a distinctive altiplano motif characterized by a stylized flamingo.

As with the altiplano Tiwanaku sites, the disappearance of Tiwanaku-style pottery from the Osmore drainage coincided with the abandonment of the major sites themselves. The M-11 component on Omo, the last Tiwanaku occupation in the region, is completely enclosed by a defensive wall. This pattern of fortification was commonly found on post-Tiwanaku pottery or other diagnostics have been found on these late Tiwanaku sites in the Osmore drainage.

The abandonment of these Tiwanaku centers however, was not part of a general abandonment of the region and did not result in the complete collapse of complex forms of political integration. The remainder of this chapter will focus on the passage of Moquegua from the imperial Tiwanaku orbit in the context of what is known of Tiwanaku decline in other regions.

Given the extent of Tiwanaku interaction with the Osmore population, a major question is: how did the population reorganize after Tiwanaku control over the area ceased? In 1985, one of us (Goldstein) proposed the new Moquegua Tiwanaku VI phase, which was subsequently renamed the Tumilaca Period (Goldstein and Feldman 1986), after the U4 or La Chimba site at Tumilaca. This site was first studied by R. Pari, by the University of Tokyo Andes Expedition (Fujii 1980; Ishida 1960) and by members of Programa Contisuyu (see Bawden this volume). The term "Tumilaca" has been adopted to describe the period of perhaps sudden disintegration of Tiwanaku political and economic hegemony, accompanied by a more gradual waning of cultural contact and influence.

Archaeologically, the Tumilaca period is characterized by the disappearance of monumental architecture and hierarchical site arrangement. Settlements became more widely distributed and uniformly sized, and were usually either fortified or strategically located. Tumilaca Period ceramics display increasingly little identity with the altiplano prototypes of Tiwanaku V. Moreover, the Tumilaca Period does not demonstrate a unified "Moqueguan" approach to ceramic technology, choice of form or decorative style. It is, instead, defined by a diversity that suggests extremely local, rather than regional, patterns of production and distribution.

Plainwares: It is difficult to generalize on plainwares from this phase because of the lack of site uniformity between Tumilaca sites. Compared to Tiwanaku 5 period assemblages, ceramics from excavations at Omo site M-11 tend to be less compact and have a very sandy temper and relatively thin vessel walls. Some plainwares popular in this period do demonstrate some similarities with late Tiwanaku ceramics elsewhere. The Punctate Shoulder Band Jar is characterized by one or more raised bands on the shoulder that are often decorated by circular punctate markings made by repeated impression with a hollow reed. There were no rim handles, though body handles are a possibility. Sherds of similar vessels were found by Rydén at Tiwanaku, Wankani and Iktonami (1947, figs 6 A, 4l, I, J, K and 66 H ), all in what were described as Decadent Tiwanaku contexts. Similar jars with punctate, appliqué decoration are illustrated by Tschopik (1946:24, fig 9 a-d, h) and identified as Collao Black-on-Red, a post-Tiwanaku ware. Kolata (personal communication) reports similar vessels in Tiwanaku 4 and 5 contexts suggesting a longer duration for this ware. In Moquegua, however, its association with the Tumilaca Period remains to be tested in the future.

Another notable Tumilaca plainware is the Nubbed One-handle Pitcher, a type that differs entirely from earlier plainware vessels. The vessel’s neck rises almost vertically to the rim, with a rim-to-body strap handle projecting perpendicularly from the lip. Often, a "Thumb nub" protuberance is present, usually incised with crosses or parallel lines.

Tumilaca Period Finewares: Enough details are shared among Tumilaca fine ceramics to indicate common cultural heritage and some continued contact with the altiplano, though direct imports are rare. Red-slipped wares predominate and the main vessel forms ñ keros, flaring sided bowls and small one-handle pitchers ñ continue to be present in the same proportions as during the Chen Chen Phase (see Goldstein this volume). Such traits would indicate a general continuity in cultural and domestic functions and categories.

Some formal innovations such as keros with bulbous upper bodies, became common in the Tumilaca assemblage. Vertical protuberances appeared on the rims of keros and above red-slipped pitcher handles. These often were modeled in the form of lizards, dogs, or other animals. In the Tumilaca Period, the use of cross hatching to fill in figures became for the first time a common technique for painted decoration. On the other hand, versions of many motifs, such as the step-stair and the flamingo birds carried over from the Chen Chen Period and demonstrate some iconographic continuity.

One Tiwanaku 4 and 5 motif that is noticeably absent from the Tumilaca assemblage is the Front Faced or Gateway God figure. If we associate this figure with a Tiwanaku state religion, its disappearance among people of clearly Tiwanaku heritage may reflect the Tumilaca period’s rejection of that state ideology.


Settlement Distribution

Tumilaca Period sites are found over a much wider altitude range than Omo and Chen Chen Phase settlements (see Goldstein this volume). This represents an extension of settlement beyond the agriculturally optimal mid-valley lands and into more defensible marginal zones. Based upon ceramic criteria, particularly diversification of form and style, three discrete sub-areas are suggested during this period:

1) María Cupina focus: A major area of Tiwanaku-descended population developed at the lower end of Moquegua’s middle valley, near the point where the Osmore River disappears into subterranean channels. Steep residential terracing of this period covers much of the western slope of the valley from the site of María Cupina (M-9) to the site of Yaral (M-8) (see Rice, this volume). Ceramic and site data strongly suggest a gradual development from the María Cupina Tiwanaku-derived types to the Chiribaya tradition at the sites of Yaral and Algorrobal in the Ilo Valley.

2) Mid valley focus: Bermann, Goldstein and Watanabe have identified many occupations in the mid-valley zone which are dates to the Tumilaca Period. The larger of these sites include Chen Chen (M-1), Echenique (M-2, M-4), Cerro Los Enriques (M-43), Omo (M-11), Trapiche (M-7), and Maria Cupina (M-9). Settlement at Omo shifted from M-10 to the walled village of M-11, perhaps in response to an episode of violent site destruction.

3) Upper valley focus: The type site (U4) is located 2000 m.a.s.l. at the point where the Tumilaca River first emerges from its highland canyon. Two other sites (T8, T9) are located on the opposite side of Cerro Baul, facing the fertile Torata Valley. In a systematic survey (Stanish 1985), Stanish and Watanabe located the upper Osmore drainage (Otora Valley) site of P-5 (see Figure 1). In the following pages, we will examine P-5, as representative of Tumilaca Period sites in the upper Osmore drainage.

Figure 1.


The Tumilaca Phase in the Otora Valley

An intensive surface reconnaissance located 17 sites in this northernmost tributary of the Osmore drainage. Using ceramic, architectural, land use, geographical and funerary data, the sites were dated and placed into a 500+-year prehispanic sequence. The Otora Valley sequence (Figure 2) has already been described in detail in Stanish 1985, 1987a, 1987b and will not be repeated here (see Stanish this volume).

Figure 2.


The first agricultural settlement in the entire Valley was the Tumilaca phase site of "Kilometro 8" of P-5 (Figure 3). It is a small, undefended hamlet of roughly three-five domestic units built on a small rise on the southern side of the Otora Valley flanks. The basic architectural pattern is one of a series of adjacent rooms, built on a low, artificial terrace. Walls were constructed of single row field stones, mortared with clay and ceramic fragments. There was no apparent attempt to select stones for size, a technique common in later sites in the valley. There was no quarried stone used in the construction of the rooms, also a technique used in the later Estuquiña period site of P-1 where some walls utilized a local rhyolite brought in from a nearby quebrada.


There is no evidence of corporate architecture or planning of the settlement. Likewise, there was no differentiation in wall architecture on the site, except for a solitary room at the top of the hill with superior construction. Individual room groups however, did display some surface differences. There were, for example, two distinct terrace groups, one with three rooms in a tightly packed set with one round room and several solitary structures.

Several rooms were excavated at the site. These excavations indicate a residential occupation with internal features such as hearths, ovens, well-developed floors, extensive middens etc. Also indicated is a general pattern of room function variability with a domestic unit composed of several adjacent rooms. There appears to be a functional segregation of rooms between 1) cooking/heating and 2) eating/sleeping based upon the distribution of utilitarian versus red slipped, decorated vessels. Such differentiation occurs among rooms of the same group, but not between the groups themselves. This resembles later Otora and Estuquiña Period patterns where domestic complexes show minimal differences between each other, but individual rooms within groups are highly variable.

Room #7, located at the top of the site, was one exception to this otherwise uniform, rustic architectural pattern. The walls of this room were constructed out of large, flat slabs set side by side, as opposed to the more casual technique in the construction of the other structures. The effect of this technique was to produce very straight walls. The doorway was placed directly in the south-east corner and not in the center of the room. This contrasted sharply with every other excavated room on the site.

Camelid bone was also recovered artifacts gave no indications of coastal contact. A methodological difficulty here is that contemporary coastal ceramics were stylistically similar to Tumilaca Period ones and therefore difficult to distinguish. On the other hand, one class of artifact is easily detectable: marine fish bone (particularly vertebrae) and scales. At P-5, there were no fish bones or spines in the middens. This contrasts sharply with the later site of P-1, where substantial quantities of fish bone were recovered from finely screened hearth and/or midden contexts. Differential preservation between the early P-5 site and P-1 was not a factor because the fine guinea pig ribs were preserved at P-5.

This hilltop room also had the only prepared floor in any of the rooms excavated in Otora. The plaster was approximately five cms thick, quite smooth and extended throughout the entire room. Under the plaster on the northwest side of the room was a complete Tiwanaku tazon or cup (Figure 4) placed upside down, undoubtedly as an offering before or during the construction of the floor. A similar flat slab wall construction and doorway placement is reported by Rydén (1947:179 map 18) dated to a "Decadent Tiahuanacu" period site at Sollkatiti, Bolivia.


In P-5, the linear grouping of functionally differentiated rooms on artificial terraces was a domestic pattern that continued on into later occupations at P-7 and modified slightly at P-4. Minimally, there are four "households" represented at the site; there is no evidence whatsoever of perishable structures outside of the area of the stone walls, such as concentrated artifact accumulations or cane building materials. A cut into the large open terrace area in the northwest side of the site did not encounter any midden accumulation or evidence of non-stone foundation structures. Significantly, there were no tombs either in the rooms nor in any identifiable segregated cemetery area.

The ceramics from P-5 (Figure 5) fit Goldstein’s (1985) typology for the Tumilaca phase developed independently in the mid Moquegua Valley. The one complete vessel recovered from the site has a feline motif. A distinctive strap handle with protubing nubs molded in the form of a cross on several utilitarian vessels has been reported by Rydén in a number of Tiwanaku and multi-component sites in the Bolivian side of the Titicaca Basin (Rydén 1957: 75 figures 56-10; 91, figure 65-p) and is common elsewhere in Moquegua (see above).


The extant agricultural remains associated with the site of P-5 indicate an extensive and sophisticated system of water control and land use for the Tumilaca Period. P-5 lies at the distal or far end of the longest and oldest prehispanic canal in the Otora Valley (C-1). Approximately 15 hectares of terraced fields, scattered in five sections, are associated with P-5. Furthermore, the first of several reservoirs in Otora was built in the P-5 fields during the Tumilaca phase. P-5 was the first major agricultural settlement in the valley and the first to construct the extensive terrace and canal systems. As the earliest and sole Tumilaca phase site in the whole valley, kilometer 8 represents a pioneering population responsible for the construction of the original canal in the valley which was utilized up to the Estuquiña or Inca periods. (See Stanish 1987b, for an extended discussion of the agricultural land-use dynamics of the Otora sequence).


The Implications of the Tumilaca Period Sites

In the mid- and lower valleys, observable patterns in domestic architecture, burial practices and cranial deformation seem to indicate a maintenance of ethnic, religious and symbolic identity with altiplano tradition even after the disintegration of Tiwanaku 5 provincial administration. Except for the new upper valley settlements, where environmental constraints dictated stone construction, domestic architecture continues to use the same wall trench and cane wall techniques as in the Chen Chen Phase, though houses were often located on steep residential terraces. Unlike subsequent groups, Tumilaca Period peoples never located burials in houses, continuing the use of separate cemetery areas.

Steep and inaccessible site location, site walls and small forts seem to have appeared for the first time in the Tumilaca Period in the lower and mid-valleys. The site of P-5 is located in a previously unpopulated valley well away from contemporary settlements. The founding of this site in an isolated region may have been dictated by defensive considerations, although no actual defensive walls or offensive weapons were found on the site. These data support a scenario of increasing competitive tensions within the Osmore drainage, a process which correlates to the collapse of Tiwanaku 5 state control.

The Tumilaca ceramics found at P-5 are characterized by an iconography obviously derived from Tiwanaku, and in some cases, virtually identical to their Tiwanaku 5 prototypes. On the other hand, the appearance of rounded bowls, coarser tempers, poorly executed designs, and lower firing temperatures (all non-Tiwanaku characteristics) in the ceramic assemblage indicates local ceramic production. This contrasts sharply with the Tiwanaku 4 and 5 fineware ceramics which were most likely altiplano imports. The ceramic data suggest that the political control of non-utilitarian ceramic manufacture for this region shifted from the Titicaca Basin to the Osmore drainage between Tiwanaku 5 and Tumilaca times.

If this geographical shift in ceramic manufacture paralleled changes in the overall political relationship between the Titicaca Basin and Moquegua, then the Tumilaca Period settlement can be characterized as an autonomous political system. Other data support this model. In particular, the fact that the major Tiwanaku 5 administrative centers in the mid-Moquegua Valley had been abandoned by this period, strongly supports the model of one or more regionalized political entities during the Tumilaca Period. That is, the over-arching Tiwanaku control which had politically integrated the Moquegua area collapsed, leading to the reorganization of this region into one or several smaller polities. As the Osmore moved from the Tiwanaku orbit, the administrative sites such as Omo were abandoned, and settlements moved to different areas of the valley. Concomitantly, as the apex of the administrative system collapsed, so did the formal political and economic links to the Titicaca Basin, leading to a reorganization of inter-regional ties.

The nature, duration and extent of this regionalized, post-Tiwanaku political structure remains problematic. The fact that the local population continued to utilize Tiwanaku-derived iconography in spite of an apparent break in former political control is very significant and may indicate political structures based on still-powerful Tiwanaku ideological and socio-political principles. In both plainwares and finewares, there are some strong continuities between the Tiwanaku 5 and Tumilaca Periods. This suggests population stability, or the absence of any major disruptions in the ethnic composition of the Osmore Valley population. In short, the Osmore data suggest a major political rupture without concomitant population changes.

P-5 is a pioneering settlement in the upper reaches of the Osmore drainage. The Otora Valley was therefore first populated in this post-Expansive period and reflects a successful attempt by the lower Moquegua Valley Tumilaca Period population to gain control of this higher zone. At this time we can only speculate on the determinants of this settlement pattern shift. Population pressure does not seem to be a factor as there was considerable land available in the lower Valley, at least as indicated by unsystematic (though extensive) surface reconnaissance. More likely is that with the rupture of the Tiwanaku system of regional resource control, a local polity sought to reestablish control of camelid routes, maize growing areas and strategic access points to the puna region. Otora is an ideal valley for establishing such "vertical" control. P-5 would therefore represent an attempt by one or more small, Tiwanaku-descended polities from the lower and/or mid-Moquegua Valley or from the coast near Ilo, to reassert control of highland access routes and/or products.

Further investigation of the Tiwanaku remains in Moquegua will greatly increase our understanding of Tiwanaku provincial organization and administration, and in turn allow us to further elevate how the Osmore drainage was reorganized after Tiwanaku collapse. The preliminary data suggest that the Tiwanaku collapse led to the formation of one or more small polities in the drainage. Still to be investigated are the abandonment sequences of the Tiwanaku period sites in the valley. Were all the Tiwanaku sites abandoned at the same time? Were the post-Tiwanaku sites of the same size and did the same range of activities take place at each?

In addition we can also only speculate on the size of the Tumilaca Period polity or polities until additional settlement pattern data can be obtained from the Osmore drainage and beyond. Do the three regions--upper, mid and lower valley-- their apparently distinctive ceramic assemblages represent three distinct political units, or is there an over-arching structure uniting the entire drainage? Ceramic stylistic comparisons and analyses of other classes of artifacts between different Tumilaca Period sites throughout the drainage will help clarify the relationship between contemporary sites in the region. The degree of similarity or differences in these data will serve as a baseline for modeling the nature and extent of the post-Expansive Tiwanaku political landscape in the region.




This paper benefited from the comments and assistance of many people in Chicago, Ann Arbor, Lima, Moquegua, and Arequipa. These include Don Rice, Michael Moseley, Luis Lumbreras, Robert Feldman, Joyce Marcus, Robert Mc Adams, Jeffrey Parsons, Victor Barua, Father Francisco Fahlman, Manuel Garcia, Gloria Salinas, Jesus Gordillo, Brian Bauer, Henry Wright, Patricia Dodson, Edmundo de la Vega, and the Archaeology Program of the Universidad Catolica de "Santa María" of Arequipa.

The Otora research was supported by The Henry and Grace Doherty Foundation, The National Science Foundation, The University of Chicago, Robert and Irene Pritzker, Victor Barua, Lucy Barua, Field Museum of Natural History, The Instituto Nacional de Cultura (Peru), The Museo Peruano de Ciencias de la Salud and The Tinker Foundation. The Omo Project was supported by a U.S. Department of Education Fulbright-Hays Dissertation Fellowship. Research in Bolivia received the support of the Social Science Research Council, the National Science Foundation and Proyecto Wila Jawira, directed by Alan Kolata.

Copyright 1989 Charles Stanish (

See Chip's web page for further publications and other information about his work.