Agricultural and Social Earthworks in the Guianas

  • Stéphen RostainEmail author
Living reference work entry

State of Knowledge and Current Debates

For few years, pre-Columbian agriculture in Amazonia begins to be better understood, thanks to more accurate studies, exempt of the classical, and too simple dichotomy between the poor interfluvial lands and the fertile floodplains. William Denevan (2001) published a useful synthesis on indigenous agriculture in South America outlining that farming patterns observed today are likely to be very different from those that existed before the European Conquest and the introduction of iron tools. In fact, many pre-Columbian agricultural techniques do not exist anymore today in the region. For that reason, ethnographical analogy is not very useful to understand them. Other data must be taken into consideration.

Pre-Columbian Amerindians transformed deeply the Amazon, whether the vegetal cover, the topsoil, the nature of the sediments, and even the modeling of soil. These changes, volunteers or not, were observed in the first place by geographers, anthropologists, or archaeologists some 30 years ago (Posey 1985). Rainforest covering the region is then less natural and less exuberant as thought. During millennia, the former inhabitants weeded, planted, multiplied, crossed, associated, or improved species. For instance, William Balée (1987) estimated that about 12% of the Amazon basin was manipulated in its vegetation by the indigenous. A close observation of the Amazon vegetation leaves no doubt about the importance that had the ancient human intervention in its current state (Clement et al. 2015). Recently, it had been shown the long term and strong influence of pre-Columbian societies on Amazonian vegetation by revealing the relative abundance and richness of domesticated species near archaeological sites (Levis et al. 2017).

In a comparable way, scholars brought to light soils partially created by man. The terras pretas or “Amazonian dark earth” (ADE) are compounds, dark and fertile soils associated with remains of implementations and naturally enriched with debris of occupation, coal, and ash. Many have wondered if they are the result of long or successive occupations on a site and if they are the result of voluntary creative acts or not. In any case, it is clear that they are islands of surprising fertility among the tropical acid and poor soils. It is difficult to believe that this remarkable quality has been not noticed by the first inhabitants of Amazon, although it is still complicated to prove that these soils have been used for the agriculture by pre-Columbian Amerindians.

Finally, pre-Columbian inhabitants not only modified the topsoil and the nature of the soil but also the morphology of the surface in many places. Amerindians dug, turned around, transported, lifted, and built thousands of cubic meters of earth, transforming the modeling of its territory more or less in a radical way. A net of permanent dug routes, channels, and ditches crossed pre-Columbian Amazonia. Moreover, built causeways connected artificial mounds, and dikes delimited pools and reservoirs. Finally, various plants were cultivated on raised fields with a large variety of forms, dimensions, and arrangements (Rostain 2012).

Indeed, the discovery of immense archaeological settlements and the revelation of complex societies by the new generation of archaeologists working in the Amazon since the 1980s provoked radical revision of our assumptions about pre-Columbian peoples. We became aware of the size and density of the earthworks from the Llanos de Moxos of Mojos, Bolivia (Erickson 1993). Similarly, contemporaneous archaeological works at the Marajoara mounds of Marajó Island revealed that the existence of complex societies in the region was not the result of an ancient migration from the Andes (Meggers and Evans 1957) but rather the product of a long-term, local development (Roosevelt 1991). However, if recent investigations revealed the importance of landscape management by pre-Columbian societies, it must be stressed that such human impact has been heterogeneous and did not affect the entire Amazonia but only specific areas (Bush and Silman 2007; Barlow et al. 2012; Balée 2013).

Recently, Eduardo Neves and James Petersen discovered and studied several enormous sites along the Middle Amazon, which often contained settlement mounds (Lima et al. 2006; Neves and Petersen 2006).

At the same time, in the Upper Xingu of Brazil, Michael Heckenberger (2005; Heckenberger et al. 2008) excavated the site of Kuhikugu, which major development occurred around 1510 AD. With a surface area of 50 ha, it is the largest residential site of the region. Its organization is similar to that of the nearby modern Kayapó villages, consisting of a ring-shaped settlement oriented around an immense central plaza. In the archaeological site, the large communal houses were located on the periphery and were protected by two rows of ditches. Very wide roads originated from the village in a radial pattern to connect it with secondary sites. A variety of fairly large earthworks marked the landscape: docks, bridge foundations, etc. The territory as a whole was organized in a system of principal centers and secondary settlements, indicating the existence of a stratified society. Today, the indigenous populations of the Upper Xingu still live in this type of large, ring-shaped village, connected to one another by wide, straight roads.

Pre-Columbian earthworks are varied, but they share three principal functions. The first one is for settlement on artificial mounds, which includes funerary functions since burials have frequently been discovered in these residential mounds, as well as ceremonial uses. Certain mounds were dedicated solely to funerary functions, such as the San Jorge basin of Colombia (Plazas and Falchetti de Saenz 1981). The second function is for defense, through ditches excavated on the periphery of settlements. The third function is agriculture on raised fields. Such pre-Columbian earthworks still can be observed in the Guianas (Rostain 2010), particularly along the coast.

The Context

The Guiana Shield forms an “island” of approximately 1,800,000 km2 bordered on all its periphery by water: the Amazon and Negro rivers, the Casiquiare Canal, the Orinoco River, and the Atlantic Ocean (Fig. 1). Today, this region includes five coastal countries: Venezuelan Guiana, Guyana, Suriname, French Guiana, and Amapá in Brazil. As in the rest of Amazonian rainforest, many different natural environments can be distinguished. However, there are three main landscapes that had strong influence on the pre-Columbian peopling: the grassy savannas in the center, the inland rainforest covering most of the area, and the coastal swampy plain.
Fig. 1

Map of the Guianas and raised-field map of the Guianas coast with the sectors of the different types of raised-field organization. Raised-field areas are in red, the coastal plain in yellow, and the forested inland in green. The name of each sector is indicated vertically. (Drawing by S. Rostain)

The pre-Columbian cultural evolution of the Guianas is divided into four main phases that represent different food procurement strategies: nomadic hunter-gatherers from ca. 11,000 years BP, semisedentary fisher-gatherers from 6,000 years BP, the first farmers employing slash-and-burn agriculture and agroforestry from 3,000 years BP, and raised-field farmers employing permanent agriculture from 1,700 years BP.

It is very probable that several human groups using these different strategies lived simultaneously. However, a clear boundary can be seen along the coast, marked by the Cayenne Island in modern French Guiana, few centuries after the beginning the Christian era. Southeastern communities, connected with cultural traditions of the Lower Amazon, continued the technique of slash-and-burn agriculture and small-scale food procurement strategies up to the arrival of the Europeans. At the same time, Western populations, derived from the Barrancoid and Arauquinoid traditions of the Orinoco Valley, intensified their agriculture using the raised-field technique and transforming enormous surfaces of flooding savannas and swamps. Most of these different agricultures disappeared from 1500 AD because of the effects of the European Conquest.

The human history of the Guianas began 13,000–10,000 years ago, during a dry and cold climatic phase. Paleolithic hunter-gatherers quickly began to experiment plants and to induce change in flora composition and distribution. However, major human impacts on the tropical rainforest occurred some millennia later when agriculturalists transformed soil composition and morphology of the surface of the ground creating coupled human environmental systems.

Process of domestication and incipient agriculture differ notably in the tropics than in temperate regions of the globe. While most of the Neolithic cradles of the northern hemisphere domesticated one major cereal crop and generally developed one single agricultural technique, Amazonia used tens of domesticated and semidomesticated different plants. At the same time, indigenous invented a wide variety of efficient agricultural systems, including agroforestry, slash and burn, alluvial culture, raised fields, etc. Moreover, certain plants were not domesticated but actively managed or cultivated, especially many species of palms. In various cases, such agricultural techniques were coupled with elaborated management of faunal exploitation, as ponds or parks. In fact, scientists agree today to recognize a large set of pre-Columbian strategies of animal, faunal, and habitat management that is considered as a true “domestication of landscape.” Natives took advantage of each environmental quality of their biotope, and, when it was not sufficient, they created ingenious systems to improve them. One fascinating characteristic of these solutions is the systematic reconstitution of an equilibrated habitat after exploitation. They ran and transformed their surroundings without destroying it, but stimulating regeneration.

The agricultural intensification is dated of the beginning of the Christian era with the increasing population and the multiplication of settlements all over the Guianas and, more generally, Amazonia. For instance, Amazonian dark earth (or terra preta) agriculture is a spectacular system that used fertility induced by the intense transformation of the earth provoked by long and important human occupation in the same place. Another spectacular agricultural system has been the raised fields in swamp or seasonally flooded areas that began to expanse along the Guianas coast and the Venezuelan Llanos from 500 AD. Still today, impressive extended monumental heritage of this pre-Columbian raised-field agriculture can be observed, especially from air, in the flooded savannas of the Guianas (Rostain 1991). The most prominent features of these pre-Columbian agricultural systems are, in first place, the durability of their effects and the permanence of their traces up to now and, in second place, that they resulted from a coupled action of culture and nature, proving the intimate marriage between indigenous intervention and environmental activity in Amazonia (McKey et al. 2010).

In the Guianas, like in other places in Amazonia, pre-Columbian populations greatly modified their landscapes, particularly those areas that were otherwise not well adapted to agriculture. The goal was to control excess water through drainage systems. In the Guianas, raised fields were elongated, square, or round, but the internal organization of these complexes can vary considerably. The principal difference among these systems is whether they articulate with rivers, standing water, or periodically flooded spaces.

Raised Fields

The most impressive pre-Columbian landscape modifications in the Guianas can be observed in the coastal plain that runs on 1,600 km from the Amazon Mouth up to the Orinoco Delta. Thousands of raised fields have been built by pre-Columbian inhabitants on a portion of 600 km long, between the Berbice River in Guyana and the Cayenne Island in French Guiana (Fig. 1). There is a great variability of shapes and distribution of these raised fields. Nevertheless, the diversity of agricultural structures does not appear random but is rather the result of careful strategies adapted to the local hydrology, chronology of construction, and the available technology. The form and size of raised fields would be thus influenced by cultural and technological criteria, which thereby justify the development of a general classification scheme. Four major types of raised fields can be proposed, based principally on size and to a lesser extent on morphology: elongated raised fields, large mounds, medium mounds, and small mounds:
  1. 1.

    Elongated raised fields vary in appearance from place to place, with dominant types apparent in different countries. In French Guiana and eastern Suriname, they can range from 1 to 7 m wide, 5 to 30 m long (although they are rarely longer than 10 m), and 50 cm to 1 m high. They are generally rectangular, although occasionally oval. While sometimes located in flooded depressions, they are more often located along the interface between flooded zones and sand bars. Their distribution, whether parallel or perpendicular to the slope, depends on their elevation. Raised fields at the foot of Quaternary sand bars follow the slope to facilitate drainage. Those located in the nonseasonally flooded zone on top of the sand bars are arranged perpendicular to the slope, to optimize water retention. In the year-round seasonally flooded zones, they are organized parallel to one another, forming a checkerboard that facilitates irrigation.

    In western Suriname, one finds regular and irregular complexes (Boomert 1976). The former exist in various forms: round, oval, square, and rectangular. They measure 3–4 m wide and 4–30 m long, while the canals between the mounds measure 50–100 cm wide. They are arranged around habitation mounds in the marshes of the recent coastal plain or close to the sand ridges to the east. Irregular complexes are rounded or oval rectangles measuring 4–5 m wide and anywhere from 8 to 140 m long, with 30 m being the average. The canals between the mounds are narrower, 1–2 m wide, but erosion has partially destroyed them. They are distributed in compact groups of two to 15 mounds in marshy areas of the recent coastal plain or below the sand ridges. Raised fields are generally smaller and shorter on the sand ridges than on the coastal plain. On the western part of the coast of Suriname, raised fields can be differentiated chronologically based on their morphology and their association with Barrancoid or Arauquinoid habitation mounds (Boomert 1980). Regular (quadrangular) raised fields are associated with Wageningen-1 and 2, Buckleburg-1 and 2, and de Burnside habitation mounds. Irregular raised fields are generally near Hertenrits and Wageningen-1 mounds, indicating an Arauquinoid association (650–1200 AD).

    Several complexes of elongated raised fields in eastern Guyana are visible in aerial view. Generally, these constructions are poorly documented, but those observed are most often distributed perpendicular to a river. Close to Fort Nassau, along the Berbice River, George Simon (cited in Plew 2005) recorded 787 rectangular raised fields that measure 1.7–6.9 m wide, 5–8.2 m long, and 50 cm to 1.7 m high. It is likely that other such complexes, associated with habitation mounds, exist along both sides of the Canje and Berbice Rivers.

  2. 2.

    Large mounds vary in diameter from 2 to 6 m and in height from 50 cm to 1 m (only rarely 2 m high). In French Guiana, they are generally rounded near Kourou and quadrangular (although eroded) around the Maillard Savanna to the east. Some are oval, rectangular, or even polygonal, such as in the Iracoubo sector. The largest mounds are generally located in the most inundated areas, their size diminishing progressively with their distance from the seasonally flooded zone. They tend to be arranged in a tight gridded pattern, except for in Guyana, where the large round mounds are widely spaced. They are clearly visible in aerial photographs.

  3. 3.

    Medium mounds are round or square (more rarely rectangular). They measure 1.5–2 m in diameter and 30–50 high. They are organized either in groups or distributed in a checkerboard pattern. They form tight groups when located in forested settings and are difficult to detect from aerial photographs alone. These mounds are often very eroded but likely are identical to the intact mounds that are located in marshy zones.

  4. 4.

    Small mounds are almost always round, only rarely square. They measure 30–150 cm in diameter and 20–30 cm high. They are found on the edges of sand bars or on savannas that are slightly seasonally flooded but are completely dry in August. They are barely visible on aerial photographs at 1:10000 scale and are detected though pedestrian survey.


Large and medium mounds are the most common types; small mounds are the rarest (or perhaps only the most difficult to detect). Elongated raised fields represent an adaptation to specific geomorphological conditions. In most cases, the complexes are made up of an assortment of different types of mounds. This mixing is not accidental. Stereoscopic interpretation of aerial views suggests an intricate engineering at work, addressing multiple factors and well adapted to local conditions.

Multiple survey methods (stereoscopic interpretation of aerial photographs and aerial reconnaissance combined with pedestrian survey) have been conducted by the author in 1989–1991 and 2004–2010 to produce a precise map of pre-Columbian structures on a 200 km stretch of coast in French Guiana. On the basis of this cartography, it has been calculated that raised field cover 2,856 ha in French Guiana. However, in view of multiple destructive factors, it is reasonable to assume that a portion of the pre-Columbian constructions of French Guiana has disappeared and that the extent of territory transformed by indigenous peoples was greater prior to the Conquest.

Similar mapping work was carried out at a larger scale from 2006 to 2008 by the Dutch forester Frans Bubberman along the coast of Suriname (Rostain 2012). Agricultural constructions are less frequently recorded over a great deal of this coast because they are hidden by colonial polders. However, it is likely that raised fields were equally common as in French Guiana before the Conquest. Combining the ensemble of mapping data available across these regions, it is possible to locate the raised fields of Guiana, from Cayenne (French Guiana) up to the Corantijn River (Suriname), a territory that covers 586 km. The only portion that is currently without data is eastern Guyana, consisting of at least 40 km bordered by the Berbice River.

Form and organization of raised-field complexes vary through time and space. The variability of raised fields in the Guianas is tied to cultural, chronological, and technological differences. Differences in local hydrological and edaphic factors also play a role. In French Guiana and Suriname, raised fields are not distributed along rivers, unlike the elongated raised fields of Guyana and Venezuela. Rather, they are constructed in coastal savannas, where they are organized in groups or grids between the marshes and sand formations. Five principal organization schemas are defined for the Guianas, each reflecting the level and degree of flooding, as well as the local topography:
  1. 1.

    Permanently flooded zones are frequently characterized by large constructions. These are elongated raised fields or large mounds, arranged in a grid pattern (Fig. 2). Such complexes have been recorded in permanently flooded depression in the Macouria and Awala sectors of French Guiana, as well as in eastern Suriname.

  2. 2.

    Seasonally flooded banks along rivers were cultivated in western Guyana. Parallel groups of elongated raised fields are arranged perpendicularly or in chevron patterns along rivers (Fig. 3). They appear to function more often for irrigation than drainage. These structures are the most common type in South America (the most spectacular examples are along the San Jorge River in Colombia) but can also be found in Guyana, although only in the Berbice and Canje basins.

  3. 3.

    Thalwegs are the preferred location for constructing raised fields because they provide highly concentrated fertile soils. They have been almost completely modified throughout the Guianas coast. Medium mounds are the type most commonly found on thalwegs (Fig. 4). These mounds follow the natural line of the thalweg and are distributed in sinuous formations across the landscape. They are abundant in French Guiana and the preferred location for building raised fields around artificial habitation mounds in western Suriname.

  4. 4.

    The edges of sand bars are another important location for raised field construction because the slope allows for adaptations to changing flood levels over the course of the year. These zones are marked by the most complex examples of agricultural distribution; each mound is individually adapted to local hydrologic conditions (Fig. 5). Thus, the shape and size of raised fields vary according to their location. This type of organization is common in French Guiana, particularly in the Kourou sector but also in eastern Suriname.

  5. 5.

    The seasonally flooded savannas, completely dry in August, are the largest spaces cultivated by indigenous peoples. Almost the entire surface is modified; only the highest areas and the sandiest zones are left untouched (Fig. 6). It is also notable that the most seasonally flooded areas were not utilized. Across these vast, flat surfaces, raised fields are very similar and located closely together. Only in eastern Guyana are large, widely spaced mounds reported. But doubts regarding whether they are of anthropogenic origin remain unresolved. Elsewhere, the mounds are generally medium in size, although large and small mounds also may occur.

Fig. 2

Raised fields in a flooded area, Maillard Savanna, French Guiana. (Photo by S. Rostain)

Fig. 3

Raised fields along a river, eastern Guyana. (Photo by G. Simon)

Fig. 4

Raised fields along a thalweg, Maillard Savanna, French Guiana. (Photo by S. Rostain)

Fig. 5

Raised fields along a slope, west of Kourou, French Guiana. (Photo by S. Rostain)

Fig. 6

Raised fields in a flooding savanna, Roche Blanche savanna, and French Guiana. (Photo by S. Rostain)

Thus, the morphology and the distribution of raised fields are linked to topography and water levels, with variability possible within a single complex. Furthermore, the mounds sometimes form homogeneous and independent groups extending on half hectare each. Along the coast of the Guianas, some sectors are dominated by specific types of mound complexes. From west to east, seven principal modes of organization have been distinguished (Fig. 1). Cultural or geographic reasons for these differences are not always apparent. Nevertheless, it seems that each sector corresponds to a specific archaeological culture, associated to a specific environment:
  1. 1.

    Berbice sector: along the eastern part of the coast of Guyana, the dominant pattern is a series of elongated raised fields oriented perpendicular to rivers. Also common are very large mounds, up to 2 m high, widely spaced across the seasonally flooded savannas. Perhaps, pertaining to the Barrancoid period, they would represent a different chronological phase.

  2. 2.

    Hertenrits sector: in the recent coastal plain of western Suriname, complexes of elongated raised fields of varying sizes follow the contour of the thalwegs. The raised fields are either arranged in small concentrations or distributed in a disordered manner.

  3. 3.

    Barbakoeba sector: in the east of the country, from the Suriname River to the Maroni, raised fields (large or medium) are principally distributed around Quaternary sand formations.

  4. 4.

    Awala sector: both banks of the Maroni and also the Lower Mana, these consist of flooded depressions that have been modified by humans by the construction of large mounds or elongated raised fields arranged in grids.

  5. 5.

    Iracoubo sector: between the Organabo and Sinnamary Rivers of French Guiana, these are principally seasonally flooded savannas covered by large and medium mounds. There are also raised fields on the southern thalwegs, south of the route nationale no. 1 (national road that runs east-west along the coast).

  6. 6.

    Kourou sector: from the Sinnamary to the Macouria, elongated raised fields and mounds tend to be fairly randomly distributed along the contours of sand bars. South of the coastal plain, below route nationale no. 1, thalwegs are the preferred location and are generally associated with large and medium mounds.

  7. 7.

    Macouria sector: from the Macouria River to the Cayenne Island, although there is some preference for thalwegs, the savanna and flooded depressions are also used. The mounds are large and medium in size and tend to be square.


The coastal stretch between the Hertenrits and Barbakoeba sectors is named Kwatta, where few raised fields have been detected. It is assumed that indigenous peoples cultivated without using the raised field technique the region’s sandy shell ridges that are supposed more fertile than the coastal savannas.

Structures are generally arranged geometrically in a well-defined grid (although less symmetrical examples are occasionally noted). Our knowledge of the distribution of pre-Columbian earthworks on the coast of Guiana is still incomplete but indicates that a great deal of the seasonally flooded zones to the west of Cayenne was transformed by pre-Columbian peoples.

The elevation of the fields permits successful farming of land that has low agricultural fertility and is subject to periodic, prolonged flooding. The ancient coastal plain of the Guianas is largely composed of marshes, unsuitable for cultivation without modification. The solution adopted by indigenous groups was to construct mounds that reached above the highest flood level. This technique combines two principal functions: drainage of the soil and the concentration of fertile material.

The removal of soil from the matrix for constructing the mounds also produced drainage canals between the raised fields. The canals, in addition to their primary function of drainage, permit the elimination of harmful gases that limit root growth. This practice considerably reduces the risk of rot for the crops, particularly tubers.

The construction of raised fields in the Guianas corresponded to a period of extremely humid climatic conditions. Between 1300 and 800 years BP, intense rains provoked flooding in the Amazonian lowlands (Colinvaux et al. 1985; Colinvaux 1989). Pollen analysis carried out on the artificial habitation mounds of western Suriname demonstrates that they were built during periods of freshwater flooding (Versteeg 1985). The elevation of the surface of habitation mounds well above the present day high flood stage for the region, and that of Marajó Island, may also be evidence that the flood level was higher during this past period. Drained land would therefore have been less available in the past. These conditions likely led to the development of raised field techniques, at least in Suriname.

The second function of elevating mounds is to improve the structure of the soil. The clay soils of the marshes on the coastal plain are not well-suited for farming, but the construction of mounds concentrates the highest quality earth for planting. This technique also aerates the soil, which improves plant growth due to deeper penetration of roots and tubers. Restoring eroded mounds with sediment from between the fields rejuvenates the cultivated soil. Nevertheless, the application of fertilizers is important for increasing productivity. In the eighteenth century, Father Juan Gumilla (1963) observed indigenous peoples constructing raised fields along the Middle Orinoco, in Venezuela. He noted that they included dried vegetal material in the mounds. This same method of fertilization is still practiced by modern Brazilian and Haitian farmers who cultivate on raised fields. It seems that this technique could have been also practiced by pre-Columbian farmers. Recovery of phytoliths characteristic of maize rachis, stems, and leaves in raised fields in the Grand Macoua savanna indicates that crop remains were incorporated in the mounds as fertilizer (McKey et al. 2010; Iriarte et al. 2010).

Pedologists and archaeobotanists carried out sampling and measurements on the Grand Macoua and Organabo savannas in French Guiana (McKey et al. 2010). A comparison of phytolith assemblages and variations in the levels of stable carbon isotopes provides insight into the history of the landscape. Samples were collected from a stratigraphic column to a depth of 70 cm below surface. The phytolith spectrum at different depths provides an image of the dominant vegetal community at different points through time, with the deeper deposits older than the deposits closer to the surface. The analysis reveals two trends that evolved in a parallel manner. The first trend shows the increase of the relative proportion of grasses of the subfamily Panicoideae from the lower clay level to the upper, more organic level and the decrease of the proportion of sedges (Cyperaceae) and herbaceous dicotyledons moving from the lowest level to the surface. The second trend indicates fairly homogeneous vegetation in the lower levels transitioning to more differentiated vegetation between the mounds (grasses) and the matrix close to the surface (Cyperaceae). These results are confirmed by isotopic analysis of the organic material in the soil, which shows a transition from a homogeneous isotopic signature in the lower levels (indicating a higher proportion of C3 plants) to a heterogeneous signature near the surface, with the mounds displaying a strong signal of the contribution of C4 plants (the greatest contributor being the grasses of the Panicoideae subfamily) and the matrix dominated by C3 plants (i.e., herbaceous plants such as Heliconia, Marantaceae (arrow root family) and, a few grass species, along with most sedges (Cyperaceae). The modern vegetation at the site demonstrates the same pattern, with a stronger proportion of C4 plants on the mounds and C3 plants in the matrix. The change in the vegetation communities can be interpreted as the result of a change from naturally flat and flooded savanna to a savanna scattered with mounds, where heterogeneity was introduced by humans (Renard et al. 2011).

The principal physical constraint is access to water and control of its circulation. Indigenous people deliberately located their mound complexes between low-lying areas and sand bars. Nevertheless, this precaution did not always protect the mounds from flooding, and it was sometimes necessary to surround certain groups of fields with drainage canals. The distribution of elongated raised fields in grids also reflects a concern for water control. In several complexes, the fields follow the downhill slope of the sand bars, likely to facilitate drainage. Conversely, the upstream elongated raised fields (in a less humid environment) are distributed perpendicular to the slope to favor water retention.

The mounds serve a double function. First, they protect plants from flooding because they elevate the planting surface above the flood level, and they facilitate drainage because the soil cannot absorb all of the rainfall during the wet season. Second, they concentrate organic material (residues of harvests and weeds) for the more demanding crops, such as maize or yams. The mounds also permit aeration of the soil. Physical and chemical constraints could be easily overcome by minimal fertilization and efficient water control. In any case, locating raised fields on the edge of flooded, seasonally flooded, and dry zones reflects a conscious decision-making by indigenous peoples. Finally, compared with highland slash-and-burn cultivation, it seems that raised fields allow for better regeneration of soils and are more sustainable over the long term.

Dating raised fields directly is challenging because they are constructed from material transported from surrounding areas and thus contain little datable material. The anthropogenic remains that are sometimes recovered from the mounds are generally not useful because their presence does not necessarily correspond to the period of utilization. In addition, they may also consist of material that predates the mound but was brought there during its construction, or conversely, material abandoned on the mound more recently. While these possibilities may be unlikely because the seasonally flooded savannas were little exploited prior to the pre-Columbian earthworks and generally ignored by European colonists, they cannot be dismissed.

There are several indices for deductive dating. The young coastal plain of western Suriname was almost never used by Amerindians, Europeans, or Creoles from the Conquest until the 1950s. Therefore, the earthworks found there are most likely of pre-Columbian origin. Furthermore, the raised fields of this region have always been organized around artificial residential mounds that are clearly indigenous and associated with the Barrancoid or Arauquinoid periods (300–1250 AD).

The discovery of a wooden shovel in the peat of the Arauquinoid site Prins Bernhard Polder, dated to the thirteenth century AD, supports this chronological association. This tool, perfectly adapted for excavating clay to construct mounds, was no doubt abandoned during the period of construction and thus represents a useful chronological marker (Fig. 7). Elsewhere in the tropical lowlands of South America, mound complexes are also associated with constructions that are clearly pre-Columbian and located in areas recently taken over by colonists. Nevertheless, those circumstances represent only indirect evidence of the antiquity of the raised fields and are insufficient for secure dating.
Fig. 7

Wooden shovels: at left, archaeological specimen from western Suriname, dated 1240–1280 AD; at right, Ashuslay shovel from Paraguay, at the beginning of the twentieth century. (Photos by S. Rostain and E. Nordenskiöld 1929)

The only sure dates on Guianas raised fields have been obtained in French Guiana. Organic material extracted from the tops of the paleosols buried under the mound complexes of Bois Diable and K-VIII, west of Kourou, has been dated using 14C. The ages obtained are, respectively, 760 ± 40 BP (calibrated 670–700 BP) and 1010 ± 40 BP (calibrated 920–950 BP) (McKey et al. 2010). Other ages are 1060 ± 30 BP (calibrated 898–1022 AD) in Bois Diable and 620 ± 30 BP (calibrated 1289–1404 AD) in Matiti savanna. These dates correspond to those obtained from the sites of Barbakoeba on the west coast of French Guiana and from the eastern coast of Suriname. It is therefore reasonable to attribute the construction of the raised fields to the Barbakoeba groups that inhabited the sand ridges bordering the mound complexes (Figs. 8 and 9).
Fig. 8

Artificial mound of Hertenrits in the 1950s, western Suriname. (Photo by D. Geijskes)

Fig. 9

Hertenrits territory in western Suriname: the red marks are artificial mounds, the yellow areas are raised fields, the brown lines are dug pathways, and the blue lines are small rivers, sometimes deviated by pre-Columbian inhabitants. (Drawing by S. Rostain)

Canals, Causeways, and Mounds

While raised fields constitute the principal landscape modification employed by indigenous peoples, other transformations to the environment are known. Canals were frequently associated with agricultural constructions in order to improve drainage. However, not all constructions served an agricultural function. Streams were diverted and canals excavated to facilitate their circulation. Roads were elevated to cross safely flooded areas. Finally, enormous clay mounds were constructed as foundations for villages.

Stereoscopic interpretation reveals a tremendous number of excavated features, both natural and anthropogenic. These consist of irregular lines created by water flow or rectilinear canals. It is sometimes difficult to determine which are anthropogenic in origin because these earthworks can be easily confused with features caused by natural runoff. Due to their horizontality, the seasonally flooded savannas are characterized by a dense and disordered hydrologic pattern. However, these natural networks are too random to aid with drainage of the raised fields. Thus, it was necessary for indigenous peoples to excavate ditches to redirect excess water, which they often accomplished by simply modifying segments of the existing natural channels. In other cases, they excavated straight and wide drainage canals. This latter technique was also practiced by European colonists, who constructed long canals for the polders but also as routes for communication. Leroy Canal, which stretches for almost 11 km to the west of Kourou, is an example of these colonial constructions. It is therefore often difficult to identify with certainty the nature and origin of the drainage routes observed. Four types of hydraulic modifications are distinguished: enclosed ditches, canals, pools, and dug paths:
  1. 1.

    Enclosed ditches are linear features that follow an irregular course, perpendicular to the natural flow of water. They often connect two areas of high ground and enclose a group of raised fields. These small, artificial features measure from 1 to 2 m wide and several dozen meters long. They provide protection from flooding during the rainy season. They are relatively narrow and shallow at their extremities but often form a basin at the center point. These ditches therefore not only divert excess water in the rainy season, but they also provide a reservoir of water in the dry season.

  2. 2.

    Canals are rectilinear ditches, wider and deeper than enclosed ditches. They measure between 50 and 600 m long and between 2 and 5 m wide. They often enclose a seasonally flooded space and appear to control its drainage. Beyond their drainage function, the ditches and canals can also serve as reservoirs of water during the dry season and as permanent fishponds. Nevertheless, this type of canal is relatively rare. In French Guiana, in the Kourou sector, they are more frequent to the south of route nationale no.1 than to its north. This distribution is probably due to greater problems with drainage in the southern zone, which is elevated slightly higher than the northern plain. Canals are less common toward the marine shore, where they would be counterproductive because the clay soils become impermeable once they are saturated. Also, the construction of a dense drainage network could provoke a rise of brackish water, which would irreversibly render the soil uncultivable.

    Raised fields are constructed and organized based on consideration of the seasonal levels and circulation of water. Ditches were constructed to control water flow.

  3. 3.

    Pools are excavated areas that function to conserve water and/or for pisciculture. They are not formally recorded in French Guiana but occur to the east of Suriname (Boomert 1976).

  4. 4.

    Dug paths and the channeling of streams are similar to canals, but they are much smaller in their dimensions, and their usage is radically different. Rather than being used for agriculture, they facilitate circulation. The most notable example is associated with the artificial settlement mounds of Hertenrits, where a network of these linear furrows forms a grid 10 km in diameter (Fig. 10). Indigenous people have drastically changed the course of small natural streams that cut across the plain. Most of the time, this involves altering the course of water flow to render it more rectilinear. In other cases, two existing creeks are linked by a passage. Completely human-made channels occur as well. The Hertenrits mound is surrounded by a large spider web pattern of dug paths. Some lengthen the course of natural streams, while others run parallel to creeks. Many extend out from residential mounds and link with raised field complexes or secondary domestic mounds. These sunken paths serve a double function. During the dry season, they mark paths to arrive at a specific location. These ancient networks resemble those associated with modern indigenous villages today. During the rainy season, when the plain is flooded, the same paths can be used as canals for pirogues. The flooding is not very deep across the plain and boats thus risk running aground. The sunken paths serve as canals that permit the navigator to direct his boat safely across the landscape.

Fig. 10

Burnt raised fields by modern fire, Maillard Savanna, French Guiana. (Photo by S. Rostain)

In eastern Suriname, the flow of creeks has also been modified to aid the circulation of boats. Across the Guianas, canal reaches are common on the bends of rivers. These are canals that pass through the forest to facilitate passage of pirogues and shorten the overall route by not requiring the craft to follow the longer, natural course of the river. In some cases, the shallow canals were constructed to connect two points of the river. Therefore, it is possible that some of the canals that connect higher parts of the coastal plain of Guiana served as routes for pirogues.

In French Guiana, two ancient human-constructed canals cross the road no. 22 linking Mana to Awala (Cornette 1987). They pass through the chenier on which the road is constructed and enter the marshes on both sides of the sand formation. The first canal cuts the road at 8.6 km east of the village. Its length is estimated at 2 km, but it is not perfectly straight because there are two slight changes of its orientation from 10° to 30° east. At its intersection with the road, it measures 21.4 m wide at the top of its walls (a slope of 25° to the northwest and 31° to the southeast). It is 4 m at its base and its maximum depth is 3.15 m. It has a concave section and a relatively flat base. When it enters the forest to the south, the width of the canal diminishes to an average of 8 m. The second canal, 12 m wide, crosses the road perpendicularly at 5.2 km from the first and is located several dozen meters east of raised fields. Another canal, 5 m wide and approximately 1 m deep, is noted close to the ancient Kali’na village of Coswine. According to the indigenous people of Awala, these canals were related to the site of Nieku dupo (“where the poison root for fishing is found”), which today no longer exists. (The nivrée is fishing with poison using macerated plant material having bioactive principles that paralyze the gills of fish, causing them to rise to the surface clumsily, from several vines in the Sapindaceae (soapberry) and Fabaceae (bean) families or from two cultivated shrubs, one in the Fabaceae family and the other in the Asteraceae (sunflower) family (William Balée, 2012, personal communication).)The myth of Sibalalimbo (“the man dressed in iron”) tells that the Kali’na of this village had dug sunken paths to quickly assemble their warriors during conflicts with the English during the Colonial period.

The inverse of the dug paths is the raised paths or causeways. Stereoscopic interpretation has identified artificial earthen levees in several highly flooded areas of French Guiana, close to the modern shore. Some constructions of this type have been found in French Guiana. These levees cross the marshes, perpendicular to the shoreline. Three link together two cheniers, while the final departs from a chenier and ends in the mangroves along the shore. They measure 430 m long and 8 m wide, 330 m long and 5 m wide, and 450 m long and 3 m wide.

The most remarkable earthworks are the habitation mounds. There are almost no cheniers in the recent coastal plain of western Suriname and eastern Guyana. Thus, indigenous peoples had to adapt in order to build settlements. Due to the lack of sufficient dry land for their villages, they constructed immense, circular clay mounds above the flood level. Arauquinoid populations were accustomed to wetland environments and skilled in their modification. Several artificial mounds and other earthworks are found within the Arauquinoid sphere of influence in the Venezuelan llanos (Redmond and Spencer 2007). Versteeg’s (1985) doctoral thesis at the University of Leiden in the Netherlands presented a study of mounds in western Suriname. In Suriname, of the 15 potentially anthropogenic mounds recorded by van der Heide (1973), seven are clearly ancient indigenous villages, two do not contain archaeological materials, two were circular mounds produced by the growth of Avicennia mangroves, and the remaining four were not visited (Versteeg 1985). The mounds in eastern Guyana are less well-known due to a lack of archaeological work, but the majority of them appear to belong to the Abary culture, a Barrancoid tradition, defined by Clifford Evans and Betty Meggers (1960).

The Barrancoid peoples were the first to construct settlement mounds in Guiana during the first few centuries of the Christian era. Analysis and dating of pollen from peat layers indicate that between 300 and 1000 AD, freshwater conditions prevailed on the western plain of Suriname (Versteeg 1985). It was at this time that Barrancoid groups began mound construction. The mounds were constructed in stages, as is evidenced by the succession of light and dark layers in the stratigraphy. The former correspond to the foundational fill, while the latter represent the actual remains of the occupation.
  • Buckleburg-1 is a circular mound that measures 139 × 137 m in diameter. Its mean altitude is 2.7 m asl or 1.5–1.9 m above the level of the surrounding marsh. Humans arrived in the area around 1835 BP, and construction began at 1735 BP. The mound was abandoned by the Barrancoid people shortly after 1315 ± 55 BP.

  • Buckleburg-2 is slightly larger and taller than Buckleburg-1, located 2.1 km to the west/southwest. The two sites share similar ceramics. Furthermore, the two mounds are built along the same river.

  • Wageningen-1 is a slightly oval mound that measures 130 × 100 m in diameter and has an altitude of 2.25 m asl or close to 1 m above the level of the marsh. Its construction began at 1315 ± 20 BP and its occupants left after 1130 ± 45 BP. The ceramic style is defined as Early Hertenrits, which corresponds to the Early Arauquinoid period.

  • Wageningen-2 is a small oval mound located 1.6 km southeast of Hertenrits and is associated with this culture. It measures 60 × 40 m in diameter and is 40–50 cm high.

  • Wageningen-3 is situated 2.8 km to the northwest of Hertenrits. It measures 80 × 60 m in diameter and 40 cm high. Like the two previous mounds, it is located in the territorial network of the Hertenrits culture.

  • Burnside is the easternmost mound in the region. Oval in shape, it measures approximately 100 m in diameter and rises 50–75 cm above the level of the marsh. The ceramics collected from the mound are associated with the Hertenrits culture.

  • Joanna, on the banks of the Canje River, is the only well-described mound in eastern Guyana (Goodland 1964). Surrounded by a ditch from which the construction material was derived, it measures 90 m in diameter and 2.5 m high.

  • Hertenrits, 6 km south of the modern shoreline, is the largest of all the known residential mounds and has the longest occupation. It was named “sandy ridge of the deer” (hert-en-rits in Dutch) by the hunters who used it in the past. Oval in shape, it measures between 200 and 320 m in diameter and averages 2.9 m above the level of the marsh (Boomert 1980). The ceramic sherds recovered in the different levels are divided into two successive styles, Early and Late Hertenrits (Versteeg 1985). The mound was inhabited between 1265 BP and 935 ± 35 BP, but the stylistic change at the end of the sequence is interpreted as the result of the arrival of a new influx of population from the Orinoco around 1000–900 BP.

Besides the artificial mounds, often associated with paths and raised fields, as well as reservoirs and fishponds, the other notable earthworks found in the Amazon are the peripheral ditches. To date, they have been found in the east of Guiana and the Upper Amazon.

In French Guiana, east of Suriname and the eastern edges of Amapá, more than 40 sites with peripheral ditches have been reported, but it is probable that many more still are hidden by the forest. The Saramaka call them “crowned mountains.” These are generally small hills, close to a river, ringed at the summit by a ditch 3–15 m wide and approximately 1 m deep, generally filled in one, two, or three places to allow crossing. They measure on average 100 m in diameter. Some reach up to 800 m long, 8 m wide, and 5 m deep. In one case, the ditch was carved directly into the rock (Éric Gassies, 2009, personal communication). Sometimes, particularly on the high banks of the river, the ditch only closes a hillfort. Little archaeological work has been carried out on these structures. These sites are particularly frequent in the interior of French Guiana and rare in Amapá and Suriname. They are dated from the beginning of the Christian era to the early Colonial period (sixteenth century).

To the northeast of Suriname, the peripheral ditch of Pondo Kreek-2 was excavated in 1980 (Versteeg 1981). The oval ditch has a diameter of 120 m by 95 m and measures on average 4.2 m wide, with an original depth of 3 m. Surface collections yielded 264 quartz tempered sherds (four of which were painted or 1.5% of the sample) and a quartz polishing stone for ceramics. Three test trenches cut across the ditch produced very little archaeological material (Aad Versteeg, 1994, personal communication). A piece of wood charcoal, collected at 175–180 cm below the base of the ditch, was dated to 1180 ± 70 BP (Versteeg 1981). The exact function of these sites is unknown, but it seems likely that they were fortified villages or ceremonial sites. The thinness of the anthropogenic level and the low percentage of phosphates recorded at Pondo Kreek-2 (Versteeg 1981), as well as the rarity of the archaeological material discovered, do not suggest a long occupation or a permanent settlement.

The ditched hill of Crique Yaou, on the Upper Maroni of French Guiana, excavated in 1992, revealed two archaeological levels and furnished a number of ceramic sherds. The two 14C dates obtained on wood charcoal range between 200 BC and 130 AD. Dating carried out at other ditched hill sites throughout French Guiana have returned ages between the beginning of the Christian Era and the European Conquest (Mazière et al. 1997; Éric Gassies, 2009, personal communication).

The oral tradition of the Wayampi of Oyapock attributes these sites to the ancient Karane, who inhabited villages with defensive structures: “throughout the region, there were many Kalana. They constructed villages that they protected with ditches two meters wide and one meter deep. Stakes were planted at the bottom. Those who did not know fell in and died. There are still remains of these ditches around the mountains close to Camopi and elsewhere” (Captain Norbert cited by Grenand 1982: 270). Moreover, at the end of the eighteenth century, on the Inirida tributary of the upper Orinoco, the Puinave (warriors who practiced cannibalism) lived in fortified villages surrounded by ditches filled with poisoned stakes (Dreyfus 1992). At this time, the Wayana villages of French Guiana also presented elaborate defensive systems (Leblond 1813). However, these data are insufficient to prove the military use of pre-Columbian ditched sites. Thus, if the defensive function cannot be discarded, it is more probable that they had a ceremonial use.

Built Landscape

Stereoscopic interpretation, archaeological excavation, and the analysis of the natural and anthropogenic environments of the coast of Guiana have revealed significant impact by pre-Columbian populations on their landscape. These transformations were accompanied by rigorous management of the territory and the organization of many small parcels. Furthermore, the sociopolitical scope of different settlements seems to have varied, with some exercising more power than others.

One of the most significant criteria to determine the existence of chiefdoms is the hierarchization of sites based on their political importance. This aspect has been, to date, little studied in Amazonian archaeology. I have thus attempted to evaluate whether Arauquinoid sites fall into a hierarchical model.

In the floodplain of the Guianese coast, the cheniers offer ideal locations for villages, which are arranged along the length of these sand formations, extending parallel to the coast and connected to one another by paths. While each of these settlements was limited by the narrow width of the sand bars (less than 200 m), they spread over several hundreds of meters, at times, even over many kilometers.

The Barbakoeba site of Sable Blanc, on the western part of the coast of French Guiana, stands out because it combines three different functions: residential, funerary, and agricultural, this last one being located some hundreds of meters south of the settlement. The residential zone is located at the top of the sand bar, with the houses constructed at the highest and driest point. The settlement was arranged in a linear form, punctuated by open spaces, along the length of the chenier. The site is found on the left bank of the Iracoubo River, just south of its mouth. This location corresponds to a recurring model evident for the ancient coastal Arauquinoid villages. The largest sites were systematically located a short distance to the west of a large river and close to the coast. This is the case for the sites of Bois Diable and Crique Jacques in French Guiana (Rostain 2012) and Kwatta-Tingiholo and Peruvia-2 in Suriname (Versteeg 1985). Access to the ocean was important because the farmers of raised fields also depended heavily on marine resources. The southern slope of the chenier was reserved for funerary activities, and the absence of domestic structures supports this interpretation. Two distinct groups of urns, of similar dimensions, appear in this sector, and it is likely that other such groups can be found in the rainforest farther to the west. These concentrations of urns, simple but with different associations of burials and offerings, could correspond to clans or families within the village. In this case, the cemetery would have been organized very precisely. The drainage canals that cross the necropolis indicate that it was necessary to manage an excess of water at certain times of the year. An artificial embankment 4 m wide and 1.5 m high surrounds the cemetery. It could have served to protect the necropolis from flooding, when the level of the southern marsh rose during the rainy season. In the lowest portions to the south and west, less than a kilometer from the site, there are raised-field complexes that cover the savannas. As in the other sites of Quaternary sand bars, the preferred placement of raised fields is in the marshes south of the settlement.

Thus, Arauquinoid sites of the cheniers demonstrate a specific organization of the settlement and its surroundings. The boundaries are well-defined, and the communities make the most advantage possible of their space. It is difficult to believe that these modifications to the territory resulted from small, independent groups. In addition, notable nonagricultural earthworks were sometimes created. These include embankments, causeways that cross the marshes, and canals or reservoirs (possible fisheries).

These findings are even more convincing for the artificial mound sites of the western part of the coast of Suriname. Pollen samples collected around these constructions show that they were located at the interface of freshwater and saltwater environments (Versteeg 1985). The areas around the mounds were flooded with freshwater at the end of the dry season. This permitted the inhabitants to exploit resources from marine, brackish, and freshwater environments, without going far from their village.

The two Barrancoid artificial mounds of Buckleburg-1 and 2 were constructed on the banks of the same river, about 2 km apart (Versteeg 1985). This likely represents a schema of settlements planned along a river where the principal site is found close to the coast, while the secondary site is established slightly more to the south. It is easy to access one mound from the other, either by boat or on foot.

Hertenrits, in western Suriname, is the most obvious example of carefully defined boundaries. Stereoscopic interpretation of this zone indicates precise management and organization of the territory. The Arauquinoid site of Hertenrits is the largest of the known mounds, covering close to 4 ha. Boomert (1980) estimates that its construction would have required the equivalent of 14,000 loads from a medium size truck. The mound is surrounded by a depression 20–100 m wide, created to extract the clay from which the mound was built. The perimeter is crossed by five canals with traces of moorings, indicating places used as docks. Several channels from water runoff were modified, some perhaps even created by indigenous peoples. The channels that cross the area have often been straightened or lengthened to connect one to another or with the canals. A complex network of straight, shallow canals crosses the landscape.

A number of connections were thus created among the different mounds, the raised-field complexes, and the streams. During the rainy season, the zone was crossed in canoes, while pedestrian paths prevailed in the dry season. Raised-field complexes, consisting of irregular groups of less than 20 elongated fields, occupied the surrounding thalwegs. Two smaller satellite mounds were constructed opposite to one another around Hertenrits. Wageningen-1 is 4 km to the southeast, and Wageningen-3 is 3 km to the northwest. Canals radiating from Hertenrits lead directly to these secondary settlements. The distribution and the connecting features demonstrate that the mounds were contemporaneous. They were constructed around 650 AD and occupied for several centuries before their abandonment.

The inhabitants of Hertenrits, Wageningen-1, and Wageningen-3 constructed their settlement following a precise model. They carefully divided the territory to take advantage of the full range of resources their biotope offered.


Pre-Columbian populations in Amazonia dug and accumulated earth to build many structures and to adapt their environment to their necessities. Although they offer a biotope rich in various resources, the Amazonian marshy savannas were imperfectly adapted to the needs of their pre-Columbian occupants.

As in forested Amazonia, a late-Holocene expansion – and possibly an intensification – of farming also seems to have occurred in seasonally flooded savannas around the rim of Amazonia in Bolivia, Venezuela, and coastal Guianas (Rostain 2012). To respond to the constraints of these environments, farmers had to develop a monumental agriculture to take advantage of these spaces. This management strategy consisted of constructing mounds above the level of flooding. However, this increase was sometimes insufficient to protect the crops from submersion. Canals, ditches, and levees completed the system to overcome the variability of the water level.

Tens of thousands of hectares of floodplain lands have been developed for agricultural purposes in three major Amazonian regions: the Bolivian llanos to the west, the Venezuelan savannas to the north, and the Guianas marshy coast to the east. Raised fields affect a wide variety of shapes, sizes, and layouts. There is no universal model in elevated fields, each region, and even each site, showing particularities.

There is archaeobotanical evidence (phytoliths and pollen) for cultivation of maize and of squash on raised fields in French Guiana (Iriarte et al. 2010; Charconac and Rostain 2015). Microfossil remains of several other crop plants – yam (Dioscorea sp.), sweet potato (Ipomoea batatas), yuca (Manihot esculenta), and chili pepper (Capsicum) – have been found in villages associated with raised-field farmers and were probably also grown on raised fields. Based on the current diet of Amazonian groups, it has often been considered that the diet of pre-Columbian populations must also be based on cassava. Recent researches in the Amazon region, however, shows that this “civilization of manioc” was, in fact, perhaps less preponderant than it is today.

Other archaeobotanical data have been recently obtained in an older settlement in French Guiana based on starch grain analysis (Pagán-Jiménez et al. 2015). In a layer associated to the Late Archaic period, with dates that range from 6090–5740 cal. BP to 4150–3920 cal. BP, of the site of Eva 2, grindstones have been used for the processing of maize (Zea mays), sweet potato (Ipomoea batatas), chili pepper (Capsicum sp.), achira (Canna sp.), legumes (Fabaceae), and yams (Dioscoreaceae).

But the indefatigable of pre-Columbian earthmover farmers did not only build these magnificent agricultural chessboards. They also applied their skills to build a multitude of land monuments with various functions. For instance, they erected enormous mounds to settle their villages, to carry out their ceremonies, and to bury their deceased. They also built long roads and dykes, the latter sometimes serving as dams to trap fish. Elsewhere, they dug basins, reservoirs, peripheral ditches, roads, canals, and diverted streams using excavations. A whole hydraulic system therefore came in addition to a set of raised earthworks above the water.

Ceremonial life was also crucial in the pre-Columbian world. There may be too much tendency to see all these various earthworks as having a strict technological and practical purpose, whereas the Amerindian world had an extensive conception of the interaction of man with nature, associating the mythological universe with reality palpable. Some earthworks, for instance, the geoglyphs of Acre State in Brazil, could have served for rituals.

Slash-and-burn agriculture still is largely used in Amazonia by indigenous, so ethnological observations can be made to help to understand this technique. It is not the case with the raised-field agriculture that has completely disappeared in Amazonia after the European Conquest. As other abandoned agricultural system, its understanding is completely dependent on inferences from the observation of the raised fields themselves and archaeological studies. Such partial data do not authorize to fully evaluate the complete diet of the ancient inhabitants because farming on raised fields was probably only one component of multi-activity subsistence systems (McKey and Rostain 2015). It is necessary to define a wider picture of the sustainability of this raised field, including intensity and productivity, to get a more accurate overview of the ancient indigenous economy.

Because of the absence of similar systems today in South America, a useful path can be to study modern raised-field farmers in other continent. For that reason, Doyle Mckey conducted ethno-ecological research in Africa where many groups still use a similar technic (McKey et al. 2014). Such approach suggested answers to contentious questions about pre-Columbian raised-field agriculture. For example, the integration of fallow periods appears to be a constant.

In sum, agricultural and social earthworks constitute a crucial aspect of the pre-Columbian period in the Guianas. Their study opens wide files of data and conclusions on the ancient inhabitants. Tropical soils can instruct us in many aspects.



  1. Balée, W. 1987. Cultural forest of the Amazon. Garden 11 (6): 12–14.Google Scholar
  2. Balée, W. 2013. Cultural forests of the Amazon: A historical ecology of people and their landscapes. Tuscaloosa: University of Alabama Press.Google Scholar
  3. Barlow, J., T.A. Gardner, A.C. Lees, L. Parry, and C.A. Peres. 2012. How pristine are tropical forests? An ecological perspective on the pre-Columbian human footprint in Amazonia and implications for contemporary conservation. Biological Conservation 151 (1): 45–49.CrossRefGoogle Scholar
  4. Boomert, A. 1976. Pre-Columbian raised fields in coastal Surinam. In Proceedings of the 6th international congress for the study of the Pre-Columbian cultures of the Lesser Antilles, Gainesville, 134–144.Google Scholar
  5. Boomert, A. 1980. Hertenrits: An Arauquinoid Complex in North West Suriname. Journal of the Walter Roth Museum of Archaeology and Anthropology 3 (2): 68–104. Georgetown.Google Scholar
  6. Bush, M.B., and M.R. Silman. 2007. Amazonian exploitation revisited: Ecological asymmetry and the policy pendulum. Frontiers in Ecology and the Environment 5 (9): 457–465.CrossRefGoogle Scholar
  7. Chacornac, M., and S. Rostain. 2015. Les plantes cultivées sur les champs surélevés précolombiens. In Archéologie de l’Amazonie. Les premiers habitants de la Guyane côtière, BAR international series 2758, Paris monographs in American archaeology 44, ed. S. Rostain, 141–155. Oxford: Archaeopress.Google Scholar
  8. Clement, C., W. Denevan, M. Heckenberger, A.B. Junqueira, E.G. Neves, W. Woods, and W. Teixeira (2015). The domestication of Amazonia before European conquest. Proceedings of the Royal Society B 282: 20150813. Scholar
  9. Colinvaux, P.A. 1989. The past and future Amazon. Scientific American 260: 102–108.CrossRefGoogle Scholar
  10. Colinvaux, P.A., M.C. Miller, L. Kam-Biu, M. Steinitz-Kannou, and I. Frost. 1985. Discovery of permanent Amazon lakes and hydraulic disturbance in the upper Amazon Basin. Nature 313: 42–45.CrossRefGoogle Scholar
  11. Cornette, A. 1987. Quelques Données sur l’Occupation Amérindienne dans la Région Basse Mana, Bas Maroni, d’après les Sources Ethno-archéologiques. Equinoxe 24: 70–99. Cayenne.Google Scholar
  12. Denevan, W.M. 2001. Cultivated landscapes of native Amazonia and the Andes. Oxford: Oxford University Press.Google Scholar
  13. Dreyfus, S. 1992. Les réseaux politiques indigènes en Guyane occidentale et leurs transformations aux XVIIe et XVIIIe siècles. L’Homme 32 (122–124): 75–98. Paris.Google Scholar
  14. Erickson, C. 1993. The social organization of prehispanic raised field agriculture in the Lake Titicaca basin. In Economic aspects of water management in the prehispanic new world, Research in economic anthropology, Supplement 7, ed. V.L. Scarborough and B.L. Isaac, 7369–7426. Greenwich: JAI Press.Google Scholar
  15. Evans, C., and B.J. Meggers. 1960. Archeological investigations in British Guiana, Bureau of American Ethnology, bulletin. Vol. 177. Washington, DC: Smithsonian Institution.Google Scholar
  16. Goodland, E.A. 1964. The mound. Journal of the British Guiana Museum and Zoo 39: 9–17.Google Scholar
  17. Grenand, P. 1982. Ainsi parlaient nos ancêtres: essai d’ethnohistoire Wayapi. Paris: ORSTOM.Google Scholar
  18. Gumilla, J. 1963. The Orinoco illustrated and defended. Caracas: Biblioteca de la Academia Nacional de Historia. (original work published 1791).Google Scholar
  19. Heckenberger, M.J. 2005. The ecology of power: Culture, place and personhood in the Southern Amazon, AD 1000–2000. New York: Routledge.Google Scholar
  20. Heckenberger, M.J., J.C. Russell, C. Fausto, J.R. Toney, M.J. Schmidt, E. Pereira, B. Franchetto, and A. Kuikuro. 2008. Pre-Columbian urbanism, anthropogenic landscapes, and the future of the Amazon. Science 321: 1214–1217.CrossRefGoogle Scholar
  21. Iriarte, J., B. Glaser, J. Watling, A. Wainwright, J.J. Birk, D. Renard, S. Rostain, and D. McKey. 2010. Late Holocene Neotropical agricultural landscapes: phytolith and stable carbon isotope analysis of raised fields from French Guianan coastal savannas. Journal of Archaeological Science 37 (12): 2984–2994.CrossRefGoogle Scholar
  22. Leblond, J.-B. 1813. Description abrégée de la Guyane française. Paris.Google Scholar
  23. Levis, C., et al. 2017. Persistent effects of pre-Columbian plant domestication on Amazonian forest composition. Science 355: 925–931.CrossRefGoogle Scholar
  24. Lima, H., E.G. Neves, and J.B. Petersen. 2006. La fase Açutuba: um novo complexo cerâmico na Amazônia Central. Arqueología Suramericana 2 (1): 26–52.Google Scholar
  25. Mazière, G., et al. 1997. L’archéologie en Guyane. Cayenne: édition APPAAG.Google Scholar
  26. McKey, D., and S. Rostain. 2015. Farming technology in Amazonia. In Encyclopaedia of the history of science, technology, and medicine in non-Western cultures, ed. H. Selin. Dordrecht: Springer/Business Media. Scholar
  27. McKey, D., S. Rostain, J. Iriarte, B. Glaser, J.J. Birk, I. Holst, and D. Renard. 2010. Pre-Columbian agricultural landscapes, ecosystem engineers and self-organized patchiness in Amazonia. PNAS 107 (17): 7823–7828.CrossRefGoogle Scholar
  28. McKey, D., D. Renard, A. Zangerlé, J. Iriarte, K.L.A. Montoya, L.E.S. Jimenez, A. Solibiéda, M. Durécu, M. Comptour, S. Rostain, and C. Raimond. 2014. New approaches to pre-Columbian raised-field agriculture: Ecology of seasonally flooded savannas, and living raised fields in Africa, as windows on the past and the future. In Amazonía. Memorias de las conferencias magistrales del 3er Encuentro Internacional de Arqueología Amazónica, ed. S. Rostain, 91–136. Quito: MCCTH/SENESCYT/3EIAA.Google Scholar
  29. Meggers, B.J., and C. Evans. 1957. Archaeological investigations at the mouth of the Amazon, Bureau of American ethnology, bulletin. Vol. 167. Washington, DC: Smithsonian Institution.Google Scholar
  30. Neves, E., and J.B. Petersen. 2006. The political economy of pre-Columbian Amerindians: Landscape transformation in Central Amazonia. In Time and complexity in historical ecology, ed. W. Balée and C.L. Erickson, 279–309. New York: Columbia University Press.Google Scholar
  31. Nordenskiöld, E. 1929. Analyse ethno-géographique de la culture matérielle de deux tribus indiennes du Gran Chaco. Paris: éditions Genet.Google Scholar
  32. Pagán-Jiménez, J.R., R. Rodríguez-Ramos, B.A. Reid, M. van den Bel, and C.L. Hofman. 2015. Early dispersals of maize and other food plants into the Southern Caribbean and Northeastern, South America. Quaternary Science Reviews 123: 231–246.CrossRefGoogle Scholar
  33. Petitjean Roget, H. 1991. 50 sites de montagnes en Guyane française: contribution à l’inventaire archéologique d’Émile Abonnenc. In Comptes rendus du XIIe Congrès de l’Association Internationale d’Archéologie des Caraïbes, Martinique, 241–258.Google Scholar
  34. Plazas, C., and A.M. Falchetti de Saenz. 1981. Asentamientos prehispanicos en el Bajo Río San Jorge. Bogotá: Fundación de Investigaciones Arqueológicas Nacionales, Banco de la República.Google Scholar
  35. Plew, M.G. 2005. The archaeology of Guyana, Bar international series. Vol. 1400. Oxford: Archeopress.Google Scholar
  36. Posey, D.A. 1985. Indigenous management of tropical forest ecosystems: The case of the Kayapo Indians of the Brazilian Amazon. Agroforestry Systems 3 (2): 139–158.CrossRefGoogle Scholar
  37. Redmond, E.M., and C.S. Spencer. 2007. Archaeological survey in the High Llanos and Andean Piedmont of Barinas, Venezuela, Anthropological papers of the American Museum of Natural History. Vol. 86. New York: American Museum of Natural History.Google Scholar
  38. Renard, D., J. Iriarte, J.J. Birk, S. Rostain, B. Glaser, and D. McKey. 2011. Ecological engineers ahead of their time: The functioning of pre-Columbian raised-field agriculture and its potential contributions to sustainability today. Ecological Engineering 45: 30–44.CrossRefGoogle Scholar
  39. Roosevelt, A.C. 1991. Moundbuilders of the Amazon: Geophysical archaeology on Marajó Island, Brazil. New York: Academic.Google Scholar
  40. Rostain, S. 1991. Les champs surélevés amérindiens de la Guyane, collection La Nature et l’Homme. Cayenne: ORSTOM.Google Scholar
  41. Rostain, S. 2010. Pre-Columbian earthworks in coastal Amazonia. Diversity 2 (3): 353–369. Special issue long-term anthropic influences on the diversity of Amazonian landscapes and biota, edited by W. Balée.CrossRefGoogle Scholar
  42. Rostain, S. 2012. Islands in the rainforest: Landscape management in pre-Columbian Amazonia. Walnut Creek: Left Coast Press.Google Scholar
  43. van der Heide, J. 1973. De Hertenrits: Een Bodemkunding/Archeologisch, Rapport Dienst Bodemkartering, 43, Paramaribo.Google Scholar
  44. Versteeg, A.H. 1981. A fortified pre-columbian village in East Suriname? Mededelingen Surinaams Museum 32: 39–56. Paramaribo.Google Scholar
  45. Versteeg, A.H. 1985. The prehistory of the young coastal plain of West Suriname. Ber Rijksdienst Oudheidkundig Bodemonderzoek 35: 653–750.Google Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  1. 1.CNRS (French National Center for Scientific Research), UMR 8096 “Archaeology of the Americas”ParisFrance

Section editors and affiliations

  • Dorian Q. Fuller
    • 1
  1. 1.Institute of ArchaeologyUniversity College LondonLondonUK