The tropical rainforest blanketing the eastern Andes is rich in resources that have attracted human interest. Vibrant bird-feathers, durable hard chonta wood, chilli peppers, cotton, and coca leaves are but a few of the commodities that were prized by the peoples populating the adjacent highlands. Of these groups, the Incas are known to have placed exploitation of certain exotic goods on a state footing. As far as they were concerned, coca was the single most-important commodity coming from the rainforest. Guaranteeing supply was complicated by the coca tree being intolerant to both frost and drought, so plantations were established in regions with a suitably benign climate to the east and north of Cuzco, the Inca capital. Underlining its interest in coca, the Inca state even established a monopoly over its production and distribution.

To the Incas, coca was a highly esteemed plant known as coca mama, or ‘mother coca’. Its significance was such that it played a part in every ritual celebration, while coca leaves were often placed as offerings at shrines and revered landmarks, such as mountains. People would also pack coca leaves into their mouths, and when the conquering Spaniards asked why, the reply came that it ‘prevents them from feeling hungry and gives them vigour and strength’. This feeling of well-being was not reserved for the living, and even the dead could be sent to their grave with coca leaves padding their mouths. But, despite this enthusiasm for rainforest produce, the Incas were not the first highlanders to develop an interest in its exotic resources, and new research is raising questions about how such commodities were first exploited on a large scale.
Research in the Apurímac Valley
In the late 1960s and early ’70s, Scott Raymond of the University of Calgary in Canada carried out archaeological research along the lower Apurímac River of eastern Peru. This survey provided the first tantalising evidence for highland colonisers establishing rainforest outposts. All these intrusive sites were confined to the western margin of the Apurímac River, except for one at Palestina. Raymond noted that the highland outposts were found at elevations suitable for coca cultivation, thus strongly suggesting that the chief motive behind colonising the tropical lowlands was coca production. He also observed that the highland colonisation of the region occurred at the height of the expansion of a pre-Inca state known as Wari, which flourished c.600-1000 AD. The Wari state was centred on a sprawling capital city at Huari, in the Ayacucho Valley of the Peruvian central highlands. At its height, Wari influence expanded over much of modern Peru, during a period that is known to scholars as the Middle Horizon.

Since Raymond’s pioneering study, archaeologists have detected a Wari presence at Vilcabamba, a site that lies further east of the Apurímac River, demonstrating that the highland colonisers could penetrate deeper into the tropical rainforest. Indeed, the distribution of these sites presents an intriguing possibility. As the Apurímac Valley occupies an intermediate position between the capital city Huari and Vilcabamba, and as Palestina is the Wari site closest to this eastern outpost, it is feasible that the establishment of Vilcabamba was carried out from Palestina. If so, it naturally raises the possibility that an ancient route once linked Huari, Palestina, and Vilcabamba.
That the Wari took a keen interest in rainforest resources chimes with recent research at their provincial centre of Pikillaqta. This remarkable complex was established near Cuzco – about 300km from Huari – and consists of a rectangular compound with walls still standing up to 12m high. Within lies a highly ordered grid of roads, plazas, buildings, and storage facilities. Work at Pikillaqta has produced miniature figurines depicting Wari warriors, who are shown with conspicuously bulging cheeks. Ethnographically, it has been observed that coca-chewers hold the leaves between their gums and their cheeks, making such swellings characteristic of the practice. As such, the Pikillaqta figurines provide strong indirect evidence that coca-chewing was already an accepted part of life during the heyday of the Wari state, long before the rise of the Inca empire. This is backed up by excavations in the northern part of the Ayacucho Valley, which has produced actual coca leaves that should – on the basis of associated pots – date to the threshold between the Early Intermediate Period (c.100-600 AD) and the Middle Horizon (c.600-1000 AD). This discovery – as well as the presence of Wari coca bags – further corroborates the use of coca leaves during Wari times.

Colonists at Cedrocucho?
Over the last three decades, deforestation as a result of agricultural expansion and the cultivation of cash crops such as coffee and cacao has had a massive impact on the Apurímac Valley. A side effect of this loss is that archaeological sites once shrouded by dense foliage are becoming easily accessible for the first time in the modern era. One such site is Cedrocucho, which lies to the west of the Apurímac River, near the modern town of Santa Rosa. While much of the site is still obscured by thick vegetation, making it difficult to determine both the full extent and overall layout of Cedrocucho, study of the visible portions suggests that this was a highly significant complex.
Cedrocucho was founded near the headwaters of the Santa Rosa River, one of the tributaries of the Apurímac River. The site lies 1,300m above sea level, just below the cloud forest on the eastern slope of a mountain carpeted in vegetation. Most of the trees are at least 30m tall, while the forest floor is obscured by shrubs. Only the south-eastern edge of the site was substantially stripped of its vegetation, although the low-lying shrubs were cleared over a greater area. The resulting increase in visibility has made it possible to gain a better sense of the structures present at the site. According to locals, though, about 8ha of land on the eastern flank of the mountain is covered with terraces; if so, the inspected area may represent a mere 10% of the overall complex. While the true size of its footprint must await verification, on the basis of what has been securely identified alone, Cedrocucho is already the single largest archaeological site ever reported in the Apurímac Valley.

Taking a tour of the visible remains at Cedrocucho conveys a sense both of how this complex was organised and of the intentions of its builders. If you approach from lower ground to the south-east of the site, the first apparent feature is a stout, 1m-wide stone wall, which appears to mark the site boundary. This wall runs north–south and connects to a steep cliff that defines the southern edge of the site, while its northern end reportedly terminates at the edge of a ravine, although this lies in an area that remains hard to access due to the forbidding vegetation. Another wall shadows the edge of the southern cliff, meaning that natural and artificial obstacles were skilfully combined to hinder access from that direction. All told, then, the lower portion of the site was protected by two walls and a cliff, strongly suggesting that ensuring security was a key concern for those who founded and occupied Cedrocucho.
The first sign of buildings can be found near the southern end of the boundary wall, where two rectangular structures comprising walls of irregular, mud-bonded stones were built facing each other across an open patio. Ascending upslope from there, while following the line of the southern cliff-edge, brings you to a series of six terraces, which occupy much of the lower portion of the site. Above them is a third rectangular building, which once again lies near the cliff and is associated with an open space, although one of more-modest dimensions than the first. Beyond that another sequence of terraces commences, covering the lion’s share of the visible portion of the site.

Moving further up the slope brings you to a break in the successive terraces, which takes the form of an area of open space, stretching about 23m by 10.4m and girded by a low wall. This probably acted as the main site plaza, and provided a fine vantage point from which the wider complex could be viewed. Once again, a link is apparent between buildings and open spaces, as rectangular structures lay at the southern and northern ends of this possible plaza. In places, the walls of these buildings still stand over 2m high, while in both cases the entrances open on to the area of open space.

Continuing uphill, a further 16 terraces line the slope, but these are once again interrupted by a sizable open space that has all the makings of a second, smaller plaza. Surveying this section of the site was complicated by it coinciding with the end of the area cleared of shrub, making it much more difficult to gain a coherent picture of the associated archaeology. Even so, it was possible to establish that the arrangement adopted for the lower plaza was replicated here, with two rectangular buildings placed at either end of the open space. Equally, it is clear that this did not mark the end of the complex, as further terraces are apparent on the slope as it leads upwards, but the conditions on the ground were sufficiently challenging that it was not judged feasible to pursue them.
Farming the rainforest
It is clear from our survey that the most extensive features at Cedrocucho are its terraces, with only seven actual buildings recorded so far. It is, of course, entirely possible that more await detection further uphill, but this possibility must await verification. Either way, though, what is known allows a number of deductions to be made about the nature of this site. With only one exception, our buildings were arranged in pairs and positioned in association with purpose-built open spaces. On the assumption that what we have seen is broadly representative of Cedrocucho as a whole, and that these rectangular buildings served as residences of some kind, it seems unlikely that the settlement was intended to house large numbers of colonists. Instead, the quantity of terraces strongly suggests that Cedrocucho was created as an agricultural complex that most likely specialised in the cultivation of a specific crop. If so, the rectangular structures perhaps housed the workers who initially built the terraces, and then the farmers who subsequently cultivated and harvested whatever produce merited such lavish use of resources.

The presence of large open spaces next to the rectangular structures would fit with the crop cultivated on the terraces requiring some kind of on-site processing before being transported to the highlands. As such, it is intriguing to note that comparative studies make it clear that after coca leaves are harvested, they need to be dried under the heat of the sun, making open spaces where they can be spread out essential. When seen from this perspective, the location of Cedrocucho is nothing short of ideal. The complex occupies an east-facing slope, exposing it to sunshine for the best part of the day, which would have been a boon if the terraces were dedicated to coca-leaf crops. If so, who was responsible for establishing this sophisticated agricultural complex within the rainforest?

Although certainty is currently impossible, there are grounds to doubt that these structures were the handiwork of local groups already living within the tropical rainforest. At present, there is no conclusive evidence that the established inhabitants of the lowlands built structures such as terraces and stone buildings that required considerable communal labour. Likewise, it is unknown whether lowlanders attempted intensive agricultural exploitation of specific crops. By contrast, systems of terracing and stone buildings were commonplace in the highland region from at least as early as the Early Intermediate Period (c.100-600 AD). The strongest parallels for a construction project of the scale and ambition of the Cedrocucho complex currently lie in either the uplands, or can be reasonably linked to the exploitation of rainforest resources by highlanders. On that basis, Cedrocucho can also stake a claim to being an intrusive outpost that was purposely established in the rainforest by a highland power, most likely to cultivate a specific crop or crops, for which there was a growing appetite in the uplands.
If Cedrocucho was indeed established by highland colonists, the challenges associated with founding the complex must have been intimidating. Not only did a sizable labour force need to be mobilised, but they would also have been labouring in a frontier territory where an unfamiliar climate prevailed. As has been discussed, we cannot yet be certain about the number of residential structures established at Cedrocucho, but it can be speculated that such structures housed farmers who were perhaps sent to the lowlands on a rotational basis. It also seems reasonable to suppose that security personnel were present to ensure the safety of the agricultural workers. After all, if the complex was imposed by outsiders, it could well have antagonised local groups. A sense of vulnerability by whoever established Cedrocucho would certainly explain why care was taken to use both natural obstacles and artificial barriers to protect the site – and its inhabitants.

A Wari site?
Sadly, our archaeological investigation of Cedrocucho did not recover any diagnostic artefacts capable of establishing when the complex was constructed. For now, therefore, we remain dependent on architectural evidence, and parallels from elsewhere in the region. As the Incas are known to have taken a keen interest in the rainforest in general, and coca in particular, they offer an obvious suspect. Their structures, though, typically feature distinctive trapezoidal doorways, windows, and niches. Despite the survival of building walls to a height of almost 2m at Cedrocucho, Inca-style architectural features are absent. This fits with previous research within the Apurímac Valley, which has yet to turn up any Inca artefacts. Thus, the chances of Cedrocucho being an Inca site currently appear remote. Intriguingly, though, Raymond had previously reported another rectangular structure in the vicinity of Santa Rosa, at the site of Vista Alegre. It was associated with ceramics that resemble Wari vessels.

Could it be that highland agriculturalists were exploiting the Ayacucho Valley as early as the Wari period? Terracing elsewhere in the Ayacucho Valley can be dated back to at least the Early Intermediate Period, while the increasing use of such features would fit with Wari treatment of conquered territories elsewhere. There, the state often installed terraces to boost regional productivity, as was the case, for instance, at Pikillaqta. It is also worth noting that, in contrast to Inca walls, Wari structures made abundant use of irregular stones, just like those at Cedrocucho. A Wari outpost at the site would fit, too, with Raymond’s reading of the general behaviour exhibited by such sites in the tropical rainforest. He saw them as a phenomenon largely confined to the western margin of the Apurímac River, which is precisely where Cedrocucho was founded. Furthermore, Raymond asserted that all Wari outposts lay at an elevation conducive for coca cultivation. On that score, the modern coca plantations commencing directly below the lower boundary wall at Cedrocucho elegantly testify to the suitability of the site for cultivating this produce.


As for whether it was feasible for the Wari to operate in the locality, Cedrocucho is only 65km distant from the Wari capital at Huari. That, though, is as the bird flies, and it should be stressed that the presence of two intervening mountain ranges considerably complicates trips for terrestrial travellers. Even so, it would have been possible to complete the journey in three days at most. That at least someone made the trip is underscored by the presence of ceramic sherds from the municipality of Santa Rosa, which on stylistic grounds can be identified as parts of Wari pots. Although this is the only diagnostic proof of a highland connection, it is intriguing that stone axes form a conspicuous component of the artefact assemblage from the locality. Naturally, such tools would have been vital to clear the forest in advance of establishing an agricultural complex. It should also be noted that there is an abundance of sherds and pots that were probably produced by local rainforest groups. If outsiders were responsible for Cedrocucho, then, it seems clear enough that the complex risked encroaching on resources already being used by others.

To test the possibilities outlined here, it is vital to carry out systematic research at Cedrocucho. Securing associated artefacts, as well as radiocarbon dates, would permit certainty about who established Cedrocucho. Pollen analysis could aid determining whether the terraces were indeed used for coca cultivation, while LiDAR survey would clear up the uncertainty concerning the full footprint of the site. Options to pursue these possibilities are currently being considered. Whatever the results, Cedrocucho promises to open a new chapter in our understanding of human exploitation of the Andean rainforest.
FURTHER READING
J S Raymond (1979) ‘A Huari ceramic tapir foot?’, Ñawpa Pacha 17: 81-86.
J S Raymond (1988) ‘A view from the tropical forest’, in R W Keatinge (ed.) Peruvian Prehistory (Cambridge: Cambridge University Press), pp.279-300.
J S Raymond (1992) ‘Highland colonisation of the Peruvian montaña in relation to the political economy of the Huari empire’, Journal of the Steward Anthropological Society 20: 17-36.
L M Valdez (2011) ‘Wari e Inca: el significado de Vilcabamba’, Arqueología Iberoamericana 10: 3-5.
ACKNOWLEDGEMENTS
The archaeological investigation of Cedrocucho was carried out in June 2018 following the initiative and financial assistance of the municipality of Santa Rosa. My thanks are extended to Telésforo Ochoa Taguada, the former mayor of Santa Rosa, who – along with J Neil Cavalcanti – facilitated this archaeological study.