The discovery of Cerro Baúl, a Wari site in Moquegua, Peru, raised new questions about the strength of the Wari polity and opened a new chapter of research examining interactions between- Wari and Tiwanku. The initial model described Cerro Baul as a Wari outpost, a colony, populated by a contingent from the Wari heartland in Ayacucho, perhaps positioned on the frontier to stave off further Tiwanaku expansion (see Moseley et al 1991). In this paper I present our current data and a plausible scenario for the Wari colonization of Moquegua.

From the beginning, Cerro Baul was viewed as an anomaly because it was so far from known areas of Wari occupation. Moseley and Feldman recognized that a site like Baul could not be alone and would have required significant infrastructure to sustain. They looked for the canal to get a better grip on the scope of the colony. And Feldman conducted the first excavations on Cerro Baul in 1989.

Bruce Owen’s systematic survey of the upper Moquegua drainage identified other Wari sites and showed that settlement clustered around Baul. Very few Wari remains were identified up valley, however Wari sites had been identified using decorated Wari pottery, which assumed that people living in all types of sites would have access to such pretty pots, which are slipped red, and painted predominantly in white and black. Unfortunately the same could be said for Tiwanaku pottery and later Inka pottery, as well as some Chiribaya, and San Miguel. Therefore in Moquegua one must often have a good portion of a motif, a rim, a kero base or something equally distinctive to know which style is represented by a surface find. This fact made Owen’s job very difficult and he often chose to err on the side of caution. I must say that Owen’s survey data is amazing and we would not be able to discuss the Wari colonization of Moquegua without it!

Nevertheless, Owen’s surveyed the valley in 1994 and 1995. Since then there have been eleven seasons of excavation on Cerro Baúl and Cerro Mejía, which provides other lines of evidence to identify Wari activity, but at the beginning we had very little to go on. When I started, I assumed that the difference between the summit of Cerro Baul and those living on Cerro Mejia was only a difference of status, and that all the Wari settlers were from Ayacucho. I expected that domestic items would be similar, but that the types and quantities of prestige goods would differ. After two seasons on Mejia in 1999 and 2000, followed by two seasons on Cerro Baúl in 2001 and 2002 there seemed to be a problem.

Cerro Mejia had very few decorated ceramics. Even in the patio group residence on the summit we found only one sherd. Serving wares on Cerro Mejia came in a number of forms and did not resemble those recovered on Cerro Baul. Both Ulli Green and Kirk Costion have attributed some of the pottery on Mejia to the Huaracane style, however that does not account for the diversity observed. Focusing exclusively on bowls, there are at least 5 different forms on Mejia, forms on Baul are not as diverse, but are completely different. This does not exclude the hypothesis that the only difference between the two populations is one of class, but it forced me to consider alternative explanations.

Maybe the colonists are not all from Ayacucho, maybe they were from different places, and practiced distinct traditions. At present the sample from Baul is huge. We have found pottery in both Chakipampa and Ocros styles, chevrons, volutes, and face neck jars, however we have also noted Nasca related wares, such as examples of what Johny Isla called Loro style bowls.

In the company of such a bowl we also found a large, Nasca-style drum. It has very thin vessel walls, and paste resembling other pottery on Baul. I suggest it was made on site, rather than being a long-distance import. 


Cerro Baul, on the far edge of the empire, may have received elaborate and oversized imports, necessary for ritual performance, such as the drum, but other goods found on Cerro Mejia are far less likely to have been imported. Instead I would suggest that they were made by imported people using local materials. This seems to be the case for some ritual objects on Cerro Mejia. They were made with local rocks. Both have applied paint, although faded and poorly preserved. These objects are consistent with placas pintadas found in the Majes and Chuquibamba region to the north beyond Arequipa. These rocks are from outcrops on Mejia and are clearly not imports.

Placas pintadas are mortuary offerings and one of these was found above a burial in Un 17 on Cerro Mejia. This finding is perhaps the strongest evidence supporting the idea that the Wari colony was peopled by groups from different areas in the expanding empire and not all migrants originating in Ayacucho.

Based on this find we might conclude that the colony consisted of elites from the Wari heartland and a subordinate group from Majes or Chuquibamba, however that would not explain the diversity of materials and the diversity of other daily practices, such as the organization of house structures, garbage disposal habits, and domestic ritual practices.

It is possible that the Wari colony was a multi-ethnic affair. People from several areas of the empire may have joined elites from Ayacucho to colonize Moquegua. We may assume that the Wari forcibly moved these subordinate groups, however given the frontier location I would suggest that the people who trekked to Moquegua were probably willing and loyal Wari subjects. Although we do not have a precise date for Wari intrusion into Moquegua, if it happened to coincide with one of the droughts between 570 and 650, it is possible that volunteers may have gladly left struggling water scarce regions, to settle an area with an ample untapped supply.

Regardless of whether subjects were willing or not, it is clear that the enterprise was sponsored by a strong state-level society. A group of grass-roots migrants would not have needed to invest so much effort in engineering a canal system to reach Baúl. They certainly would not have invested the thousands (perhaps millions) of hours, required to build the monumental structures, brew vast quantities of beer on the summit, and maintain an elite population on Baul’s impractical summit.

In other words if you look up conspicuous consumption in the dictionary…you might find a picture of Cerro Baul. The scale of which we should ascribe to a large state, rather than a ‘wanna be’ chiefdom.

If we consider what it took to establish the brewery on the summit of Cerro Baul we can gain an idea of the labor investment expended to develop the colony as a whole.  The most important factor was to bring water to the base of the hill. This required the construction of the main canal, which is 14km long. To feed laboring colonists, irrigation agriculture had to be established because valley bottom land would not have been enough. Colonists had to live somewhere, therefore more labor was invested to terrace hills and ridges above irrigated areas, build houses, and make tools, pots, cloth, and other domestic essentials. Work on Cerro Baul could not begin until this infrastructure was established.


This list underestimates the necessary labor needed to get the ball rolling in Moquegua, but it’s a start. If we examine the number of houses at Wari sites identified by Owen, the labor inputs required seem to exceed the potential of such a small population. Based on house ruins I estimate that at most 2200 people lived in the colony. If we exclude the elites living on Baul, that leaves 1700 people. However, population estimates based on housing should be reduced by 30% because a portion of houses would be vacant at any one time. That leaves 1190. If another 30% were children then we only have 833 people. Simply put, this is not enough people to have reshaped the landscape, built a huge canal system and constructed the monumental center of Cerro Baul. There must be more sites with more people to have built the colony and its infrastructure

Given that excavations at Cerro Mejia uncovered so few decorated Wari ceramics I thought it possible that Middle Horizon components at other sites may have been overlooked. Using Owen’s survey I selected a group of sites near the projected course of the Wari canal, with a style of plainware pottery he termed “abuelitos” and obsidian present on the surface. The list included some 40 sites, all along the right or southern bank of the river. Some were residential others were agricultural. Unfortunately many of the sites had been destroyed by recent expansion of agriculture in the region.

Nevertheless, based on architecture and the domestic pottery on the surface, I estimate that at least 8 more Middle Horizon sites were present. These settlements do not exhibit the same preservation as Baul and Mejia, but the foundations of houses and some walls remain. At three different sites masonry styles and wall superposition suggest that a Middle Horizon occupation may have been larger or positioned differently so that parts of the LIP occupation overlap the MH component but do not cover it completely.

Other sites may be completely covered. A recent road cut revealed an earlier structure wall, which was built at a different angle than the overlying Inka terrace. The profile also revealed sediment and carbon that appears to be a domestic floor. Since the Inka system is so extensive it is possible that more than one community was located in this area. Perhaps positioned just above the Wari canal but falling under the course of the later Inka system.

Testing small villages attributed to the Late Intermediate period may help to locate other people. It is also important to determine if there are some remnants of Formative occupation in the area. Although small, there are several patches of valley bottom land that may have been in production when the Wari arrived.