Stéphen Rostain1 & Antoine Dorison2
1 French National Center for Scientific Research (CNRS), ‘Archaeology of Americas’ laboratory, France
2 Newcastle University, Newcastle-upon-Tyne, UK
What is a city? To a Western mindset, the term immediately evokes the idea of a group of buildings, often arranged along streets and, in older examples, usually surrounded by defensive walls or moats. However, the concept has widened considerably, and archaeology has played its part in this development. As a result, we now recognize multiple expressions of cities around the world, especially in the Northern Hemisphere. Although recognition of this diversity is now quite well-accepted, it took several more decades for scholars to incorporate into the concept of city examples that differed from an often glorified classical model, such as Angkor Wat in Cambodia, the Mayan centers of the Yucatan Peninsula, and the cities of ancient Peru. Even then, these are often referred to as ‘pre-urban’, implying evolution and, thus, an unfinished development process. By such criteria, only a few prestigious tropical forest societies are thereby placed on the so-called path of progress. The situation intimates that the construction of sustainable cities by indigenous peoples in the Amazon would be relocated to the realm of fiction, rather than history! And yet...
Figure 12. Artificial mounds in the Upano Valley, Ecuadorian Amazonia. Photo by S. Rostain.
In 2024, Science published the discovery of original urban manifestations in the upper Ecuadorian Amazon (Rostain et al. 2024). Although the existence of artificial earth mounds in the Upano Valley associated with dug paths had been known for some time (Porras 1987; Rostain 1999) (Figure 12), no one suspected the considerable extent and number of these developments. The exceptional importance of this garden urbanism was revealed by the interpretation of LiDAR images covering 600 km² along the raised terraces which thrust up from the valley floor (the Upano River is deeply dug in between cliffs that measure over 100m; Rostain et al. 2024). This discovery led to a profound reconsideration of what was previously known about the ancient indigenous cultures of this part of the Amazon.
Figure 13. Archaeological map of the Upano Valley and extension of the Lidar survey covering. Map by A. Dorison and S. Rostain.
A gradual discovery
Unlike some archaeological projects where field investigations only developed following discoveries revealed by LiDAR, research in the Upano went the other way round. See Figure 13. It all began in the late 1970s when Father Juan Botasso, stationed in the colonial town of Macas in the same valley, was visiting his indigenous parishioners when he noticed several rectangular mounds of earth that appeared to be man-made (Saulieu et al. 2020). Later, he welcomed his colleague Pedro Porras from Quito, and took him to visit these unusual mounds. Porras then returned to Quito to announce that he had just discovered a new archaeological site, which he named Sangay, after the volcano overlooking the valley. Although Porras had little experience in excavation, he organized a dig and a mapping survey with his university students at the site. These revealed a 100-ha wide settlement with over 170 earthen platforms organized in small groups of four to six, all connected by dug footpaths, leaving little doubt as to the urban nature of the establishment. See Figure 14. However, when Porras published the first monograph, he interpreted the site as primarily—if not exclusively—ceremonial (Porras 1987).
Figure 14. Main excavation made by Pedro Porras in the Sangay site (after Porras 1987).
A professional archaeological project was launched in 1996 by Stéphen Rostain, initiating a seven-year-long Franco-Ecuadorian collaboration to conduct research in the Upano Valley. See Figure 15. Excavations and surveys were carried out with international specialists from various disciplines (botanists, soil scientists, geologists, volcanologists, etc.). Several new sites were discovered. See Figure 16. Beside test-pits and given the size of the sites and the artificial mounds (often over 20 m long), Rostain and his team undertook the very first large-scale excavations in open areas in the Ecuadorian Amazon. See Figure 17. Following preliminary stratigraphic test-pits, two groups of mounds were excavated manually at two different sites to uncover each layer of occupation over a large area. See Figure 18. Well-preserved domestic floors were discovered at the top of the mounds. The methods used for their construction were also revealed, which also shed light on the chronology of the valley’s occupation (Rostain 2006, 2023). The Kilamope and Upano archaeological cultures were, thus, identified. See Figures 19-20. The groups belonging to the Kilamope Culture (the older of the two) seem to have begun building earthen mounds around 1000 BC. However, it was likely the Upano groups—contemporaries or successors of the former—who were responsible for the intense modification of the valley’s landscape from 400-100 BC to 400-600 AD. After a long hiatus, some mounds were re-occupied by groups from the so-called Huapula Culture between 800 and 1200 AD (Rostain 1999, 2023). See Figure 21.
Figure 15. Artificial earth platforms found during the 1996 surveys. Photo by S. Rostain.
Figure 16. Extraction of a Huapula Culture jar from the upper occupation floor of a platform at the Sangay site. The first author is in the center. Photo by S. Rostain.
Figure 17. Large-scale excavation at the top of a residential platform, Sangay site. Photo by S. Rostain.
Figure 18. Kitchen floor at the top of a residential platform of the Huapula Culture in the Sangay site. At right, the millstone and its grinder seen in the background of photo at left. Photos by S. Rostain.
Figure 19. Typical pottery bowl with corded print outside and negative painting inside of the Kilamope Culture. Photo by S. Rostain.
Figure 20. Typical red-banded and incised pottery recipients of the Upano Culture. Photo by S. Rostain.
Figure 21. Large jars for the cassava beer of the Huapula Culture. Photo by R. Jones and S. Rostain.
Figure 22. Map of the Sangay site. Map by A. Dorison.
The regional survey revealed the existence of other mound clusters on the high terraces bordering both sides of the Upano River, although none equaled the size of Sangay. See Figure 22. These ancient societies, therefore, occupied a territory larger than this single settlement. However, we had no idea at the time that we had only just scratched the surface of the true extent of this mound building phenomenon.
The sites seen from the sky
Excavations by Rostain’s team in the Upano Valley ended in 2003. No major excavations were carried out thereafter, as national Ecuadorian archaeology tended to focus on the coastal and Andean regions, which were considered more prestigious by the public. However, the Upano Valley was not forgotten and, twelve years later, advances in remote sensing technology provided a more complete picture of its ancient occupation. In 2015, the Ecuadorian Institute of Cultural Heritage (INPC) commissioned a LiDAR survey covering a 600-km² area ranging from the Upano Valley to the headwaters of the Pastaza River (Technoproject 2015). This technology, now well known to archaeologists, provides a very accurate model of the terrain under the vegetation cover, revealing a landscape that cannot be observed directly either on the ground or from the air. Although an article summarizing the data and interpretations of the technical report was published locally (Sánchez-Polo and Álvarez Litben 2023), with no associated fieldwork and little exposure within the archaeological community, let alone internationally. In 2021, the INPC independently entrusted Rostain’s team with the raw LiDAR data0F in order to start a new analysis from scratch. Thanks to the collaboration of the second author of the present paper, the results of this work were published two years later and revealed to the world (Rostain et al. 2024).
Figure 23. Domestic (left) and civic (right) platforms. Image by A. Dorison and S. Rostain.
Where previously, only a few artificial mounds and dug paths could be seen, a myriad of anthropogenic features appeared on the digital image. The first things that can be detected in the field or on LiDAR images are the artificial earth mounds. Remote sensing has made it possible to count more than 6,700 of them! Most of these are rectangular platforms ranging from 10 to 140 meters in length, 3 to 20 meters in width, and 2 to 8 meters in height (Rostain 2012a, b). Their layout follows a standardized pattern, consisting of a square plaza or patio surrounded by peripheral mounds, with a central platform in some cases. See Figure 23. The most common of these so-called ‘complexes’ measure 40 meters on each side and are interpreted as residences based on excavations (Rostain 1999). See Figure 24. However, there are also much larger complexes—both in terms of surface area and platforms—which probably bear a civic and ceremonial function. The largest complex (at Kilamope) covers ten hectares and includes a platform measuring 140 by 40 meters. See Figure 25. The average density is 16.6 platforms/km², but some areas have densities exceeding 100 elements/km² (Rostain et al. 2024).
Figure 24. Complex of residential platform in the Kilamope site. Photo by S. Rostain.
Figure 25. Large civic mound in the Domono site. Photo by S. Rostain.
Not only was the site of Sangay much more impressive than previously imagined, but there were also several other similar cities in the surrounding area where only a handful of isolated complexes had been detected hitherto.
Amazonian urbanism
The sites in the valley turned out to be much more than just clusters of mounds, but some are actual urban centers (Rostain et al. 2024). In fact, almost all of the settlements were organized along one or two perfectly straight streets—or avenues—that materialized along an axis which crossed the site from one end to the other. See Figure 26. Stretching over a kilometer in length, these avenues were dug up to 5 meters deep and up to 15 meters wide. Side streets branching off from the central avenue (which were dug less deeply and which connected numerous domestic complexes) formed a network that organized the entire settlement into a more or less regular grid pattern. See Figure 27. The larger complexes (the ones believed to have had a civic or ritual function) were located at one or both ends of the avenues.
Figure 26. Central avenue of the Junguna site. Image by A. Dorison and S. Rostain.
Figure 27. One dug street in the Sangay site. Photo by S. Rostain.
Figure 28. Archaeological road network in the Upano Valley. Map by A. Dorison.
What is even more striking is that this modification and organization of the landscape as defined by a need for axiality and orthogonality continues outside the clustered areas through a dense road network on the regional scale. Though less dense, numerous other sites consisting of residential and civil buildings and even isolated platform complexes form nodes in the gridded pattern thereby created. Avenues extend beyond the agglomerated cities limits toward other settlements. They thus turn into roads, dug to varying depths cross the region, extending over 10, 15, and even more than 25 kilometers. These are characterized by running in a straight line, without deviation in spite of natural irregularities in the terrain. These roads cut through the slightest elevation, cross valleys, and climb hills. This type of dispersed settlement pattern is reminiscent of other American examples of “low-density urbanism” (Isendahl and Smith 2013), but the almost obsessive orthogonality of the Upano case is unprecedented. See Figure 28.
Indeed, such straight roads are completely unnecessary in a forest environment. The inhabitants of the region could have travelled perfectly well without such labor-intensive infrastructures. It is therefore reasonable to assume that these roads were not simply economical features. Based on the way current indigenous groups create footpaths in other parts of Amazonia (e.g. Heckenberger 2020), we argue that the Upano roads can be interpreted as physical imprints of social connections between settlements on the landscape. It is plausible that the inhabitants had thereby designed an anthropized landscape with high social and symbolic value in order to facilitate interactions between the various communities in place.
Figure 29. Drained fields in the Yukipa site. Image by A. Dorison and S. Rostain.
Furthermore, beside the roads and the residential and civic features, numerous agricultural remains were also detected. Thus, the seemingly empty spaces between the platforms or roads were often dedicated to fields. On the flat, up-lifted terraces bordering the Upano River, most of these plots are drained fields laid out in vast grids connected with the road network, which allowed excess water to be drained away into the numerous ravines of the agricultural hydrographic system. See Figure 29. The ancient inhabitants also took advantage of the concave slopes—especially at the head of ravines—by building cross-channel terraces. On gently sloping terrains, they raised long ridges. All of this demonstrates a keen understanding of the local geomorphology which prioritized making the most of available space and suggests that the roads also had a role to play in a regional-scale water management program. See Figure 30.
Figure 30. Map of the Kilamope site with the detail of the drained fields between the platforms and along the dug streets. Map by A. Dorison
A domesticated landscape
Archaeological excavations and remote sensing attest to the intimate interaction of indigenous peoples with their environment and with their distant neighbors. The Upano peoples invested intensely in their environment by creating what we like to call ‘green urbanism’, in reference to urban theories from the last century (Howard 2007). The integration of buildings into the landscape made them almost imperceptible, even though they have remained standing for three millennia. See Figure 31.
With its preference for a grid plan and an orthogonal road system, the Upano Valley stands out clearly from other settlement planning choices in the Amazon. In the south and southwest of the region, ancient peoples preferred to organize their settlements in a star pattern with radiating roads. These could take the shape of a huge central mound surrounded by earthen structures and raised roads, as was the case in the Llanos de Mojos in Bolivia (Prümers et al. 2022), or of large houses forming a ring around a huge plaza from which slightly sunken roads radiate, as in the state of Mato Grosso in Brazil (Heckenberger et al. 2008; Heckenberger 2020).
Figure 31. Artistic view of the Sangay site ca. 2000 years ago. Watercolour by S. Rostain.
In any case, it is very likely that, given the amount of labor invested, the Upano Valley was densely populated during the first to centuries AD. The road connections between the c. 20 clustered settlements identified suggest that at least a part of these anthropogenized landscape was contemporary. The occasional presence of earthen fortifications nevertheless indicate that their history is far from linear, with probable periods of increased exchanges and others of tension. A prominent feature of this urban planning is the intertwining of distinct functions (civic-ceremonial and residential at the very least), not only within the cities but also dispersed across the landscape.
Implementing such large-scale projects with sophisticated engineering across an entire valley is difficult to imagine for leaderless societies, as we know them today in the Amazon. It must be acknowledged that some ancient peoples may have been less egalitarian, with a hierarchical system. This monumental public work must have been organized by elites and supervised by specialists, particularly in land surveying. The very existence of this large-scale agrarian urbanism, bringing together housing, ceremonies, ritualized movement, trade, and agriculture in the same space, encourages us to rethink the socio-political and economic organization of the first inhabitants of the Amazon.
Thus, this discovery tell us, once more and contrary to a common misconceptions, that the world’s largest tropical rainforest is far from pristine, but the result of human modifications which took place over thousands of years. Furthermore, the urban system thereby revealed—which probably extended beyond the 600 km² studied—was on a scale which has never before been documented in the South American lowlands. This last forces us to fundamentally reconsider the human past of the upper Amazon.
Bibliography
Heckenberger, Michael J. 2020. ‘Xingu Garden Cities. Amazonian Urban Landscapes, or What?’ Landscapes of Preindustrial Urbanism, edited by G. Farhat, Dumbarton Oaks Research Library and Collection, Washington, DC: 225-262.
Heckenberger, Michael J., Christian Russell, Carlos Fausto, Joshua R. Toney, Morgan J. Schmidt, Edithe Pereira, Bruno Franchetto, and Afoba Kuikuro 2008. ‘Pre-Columbian urbanism, anthropogenic landscapes, and the future of the Amazon’ Science, 321: 1214-1217.
Howard, Ebenezer 2007 [1902]. Garden Cities of Tomorrow, Routledge, London.
Isendahl, Christian, and Michael E. Smith 2013. ‘Sustainable agrarian urbanism: the low-density cities of the Mayas and Aztecs’ Cities, 31: 132-143.
Porras, Pedro 1987. Investigaciones arqueológicas a las faldas del Sangay: Tradición Upano, Centro de Investigaciones Arqueólogicas, PUCE, Quito, Ecuador.
Prümers, Heiko, Carla Jaimes Betancourt, José Iriarte, Mark Robinson, and Martin Schaich 2022. ‘Lidar Reveals Pre-Hispanic Low-Density Urbanism in the Bolivian Amazon’ Nature, 606 (7913): 325-328.
Rostain, Stéphen 1999. ‘Secuencia arqueológica en montículos del valle del Upano en la Amazonia ecuatoriana’. Bulletin de l’Institut français d’études andines, 28 (1), Lima: 53-89.
Rostain, Stéphen 2006. ‘Etnoarqueología de la casa Huapula y Jívaro’ Bulletin de l’Institut Français d’Études Andines, 35(3), Lima: 337-346.
Rostain, Stéphen 2012a. ‘Between Sierra and Selva: pre-Columbian landscapes in the upper Ecuadorian Amazonia’ Quaternary International, 249, Elsevier: 31-42.
Rostain, Stéphen 2012b. Islands in the rainforest. Landscape management in pre-Columbian Amazonia, Left Coast Press, Walnut Creek.
Rostain, Stéphen 2023. ‘Tolas, terrazas y casas: arqueología del valle del Upano’ Strata, 1(1), Quito.
Rostain Stéphen, Antoine Dorison, Geoffroy de Saulieu, Heiko Prümers, Jean-Luc Le Pennec, Francisco Mejía Mejía, Anna M. Freire, Jaime R. Pagán-Jiménez, and Philippe Descola 2024. ‘Two Thousand Years of Garden Urbanism in the Upper Amazon’ Science, 383(6679): 183-189.
Sánchez-Polo, Alejandra, and Rita Álvarez Litben 2023. ‘Un paisaje monumental prehispánico en la Alta Amazonía ecuatoriana: primeros resultados de la aplicación de Lidar en el valle del Upano’ Strata, 1(3), Quito.
Saulieu, Geoffroy de, Stéphen Rostain, Anne-Christine Taylor, and Philippe Descola 2020. ‘Entre châtaigner et palmier chonta. Hommage à Juan Bottasso (1936-2019)’ Journal de la Société des Américanistes, 106(1), Paris: 175-180.
Technoproject 2015. ‘Informe del levantamiento aéreo y del análisis preliminar e identificación de las anomalías del relieve relacionadas con las estructuras arqueológicas. Superficie Upano-Mera’ Ms. Unpublished field report, INPC, Quito.
Go back to top