1. Department of Anthropology, University of Pennsylvania, 33rd and Spruce Streets, Philadelphia, Pennsylvania 19104-6398, USA

Correspondence to: Clark L. Erickson Correspondence and requests for materials should be addressed to C.E. (e-mail: Email: cerickso@sas.upenn.edu).

The zigzag structure is a particular form of artificial earthwork, found in the seasonally inundated savanna of Baures, Bolivia (Province of Iténez, Department of the Beni) (Figs 1 and 2). Zigzag structures are linear segments of raised earth (1–2 m wide and 20–50 cm tall) that change direction every 10–30 m (Fig. 3). Shrubs, palms, and termite mounds cover the structures. Many zigzag structures cross the savanna from one forest island to another, distances of up to 3.5 km; others terminate 5,000–1,000 m from the forest edge (Fig. 4). Funnel-like openings, 1–3 m long and 1–2 m wide, are present where the structures form a sharp angle (Fig. 5). The structures are associated with small circular ponds. Dense networks of interconnected zigzag structures form enclosures of 10–80 ha. A total of 48.431 linear kilometres of weirs were measured in a sample area of 16.755 km ² of savanna, a density of 2.891 linear km per km².

Figure 1: Map of the Baures prehispanic hydraulic complex.

It is located between the San Joaquín river to the west and the San Martín river to the east and between 13^o 30' and 14^o 20' latitude South.

High resolution image and legend (35K)

Figure 2: The fish weirs (zigzag structures and ponds).

They are concentrated in the North Block (447 km² of savanna excluding forest islands) and the South Block (77 km ²). The structures are restricted to savannas near forest islands that are flooded by shallow water.

High resolution image and legend (45K)

Figure 3: Oblique photograph of a fish weir and artificial ponds between forest islands in the savannas of Baures.

Fish weirs are the zigzag structures, lower left to upper right; artificial ponds are the circular features surrounded by palms (approximately 20 m in diameter). The diagonal feature (upper left to lower right) is a contemporary path.

High resolution image and legend (51K)

Figure 4: Map of fish weirs (irregular lines) and causeways (straight lines) in Baures.

Based on aerial photographs.

High resolution image and legend (53K)

Figure 5: Plans of fish weirs (zigzag structures).

Small parallel openings in the weirs are present every 50–200 m.

High resolution image and legend (21K)

The zigzag structures are artificial constructions created by raising earth removed from the adjacent savanna. Canals or barrow pits flank some of the zigzag structures. Although overlapping in distribution, the zigzag structures are distinct from the long, wide and straight causeways and canals that cross the savanna between forest islands (Fig. 4). The narrow and irregular zigzag structures would be inefficient for transportation. The zigzag structures do not appear to have functioned as check dams or berms for flood-recessional farming. There is no evidence of crop furrows or field platforms between the structures.

On the basis of location, form, patterning, associations and ethnographic analogy, I identify the zigzag structures as fish weirs. Fish migrate to and spawn in the seasonally inundated savannas of Baures during the wet season⁹. Many fish are trapped in water bodies as the floodwaters recede. The zigzag structures provided a means to manage and harvest these fish. The zigzag structures are similar to fish weirs built by native peoples in Bolivia^{10, 11, 12, 13, 14} and throughout the Americas^{15, 16, 17}. Two characteristics shared by fish weirs include construction of barriers across shallow bodies of water and V-shaped openings where fish are trapped. Weirs, ranging in length from several to hundreds of metres, are constructed of soil, rock, reed, branches, logs, aquatic vegetation and/or basketry. Superstructures of perishable materials or a dense wall of vegetation probably covered the earthen fish weirs of Baures.

There are important differences between the zigzag structures of Baures and contemporary fish weirs. Most ethnographic fish weirs are ephemeral and rebuilt each season. Traditional weirs are constructed in rivers, streams or permanent bodies of water. In contrast, the zigzag structures are permanent earthworks built across a seasonally flooded savanna. They are also more numerous, longer and more densely placed than ethnographic fish weirs. In addition to controlling and harvesting fish within the savanna, the weirs and large causeways may have been used for water management. The earthworks could have extended the period of inundation by capturing the first rains and holding floodwaters into the dry season¹⁸.

The savanna fisheries of the Bolivian Amazon are productive. Estimates of 100,000 to 400,000 fish have been calculated for a single hectare of abandoned river channel in the savannas⁹. Yields of 1,000 kg per hectare per year have been recorded for shallow ponds in tropical savannas¹⁹. Large numbers of Pomacea gigas snails are found beside the zigzag structures. In addition to fish, these edible snails may have been managed and raised in the weir structures and ponds. In the past, these snails were eaten in Baures and the gastropods are found in precolumbian sites in Bolivia and Brazil^{20, 21}. The nutritional status of Pomacea is probably similar to other tropical snails, low in calories and protein²². Pomacea gigas reproduce and grow at an impressive rate and an average of 23.8 snails per m³ is recorded in Bolivian wetlands²³. The artificial fisheries of Baures potentially produced hundreds of tonnes of edible snails as a secondary food source.

The most common vegetation associated with the fish weirs and ponds is the palm Mauritia flexuosa^{24, 25, 26} (Fig. 3). A single tree can produce up to 5,000 edible fruits each year and a single hectare yields 10–60 t of fruit. The fruits are high in vitamins C and A, oil (12%) and protein (4–5% dry weight). The ground tissue produces large amounts of edible starch. Edible larvae of the palm beetle thrive in the decomposing trunks. In addition, the palm is a favoured food of game animals and fish. The fibres of the fronds and trunks are used for basketry, mats, hammocks, bowstrings, thatch and roof beams. The palm may have been encouraged or even cultivated on the earthworks.

Artificial ponds overlap with the distribution of zigzag structures (Fig. 3). These measure 0.5–2 m deep and 10–30 m in diameter; the largest hold water year round. The ponds teem with fish such as buchere (Hoplosternum sp.), yallu, cunaré (Cichla monoculos), palometa (Serrasalmus sp.), sábalo (Prochilodus nigricans) and bentón (Erythrina sp), snails, birds, reptiles and amphibians. Contemporary hunters stalk the game animals and birds that congregate at the ponds. Artificial ponds provided a way to store live fish and snails until needed.

The complex of fish weirs and ponds of Baures is a form of intensive aquaculture. The earthworks did not necessarily involve the mobilization of large amounts of labour. I estimate a total of 1,515 linear kilometres of fish weirs in Baures based on a sample of aerial photographs. Using labour estimates for experimental construction of raised fields (5 m³ of earth per person per 5-hr day and weirs measuring 2 m wide and 0.5 m tall), the weirs required 300,000 person-days of labour or the equivalent of 1,000 people working 30 days a year over a period of 10 years. Small groups of kin or communities constructed and managed the weirs recorded in ethnographic accounts^{10, 11, 12, 13, 14}. Similar social groups may have been responsible for the weirs and ponds of Baures.

The weirs also show some evidence of integration at a higher scale. Individual zigzag structures often cross the savanna from one forest island to another (Fig. 4). Assuming each forest island was an autonomous settlement, weir construction may have involved inter-community cooperation. Although individual weirs could operate independently of other weirs, the presence of an integrated network of causeways, canals, weirs and ponds suggest a higher order of water management^{8, 18}. As a permanent food-producing infrastructure, the weirs must have been valuable real estate. The networks of causeways and canals may have promoted communication and alliances between individual communities exploiting the fish weirs. Groups in the Colombian Amazon jealously protect and guard riverine fisheries, valuable resources that are owned and inherited by clans and chiefly lineages¹⁶. The presence of moated, and presumably palisaded, settlements on many of the forest islands suggests potential tension over the fisheries and other resources²⁷.

Colonial accounts describe the use of causeways in Baures for communication and transportation between settlements^{7, 10, 12, 27}. Weirs in lakes and streams are described^{10, 12}, but there is no mention of the zigzag structures in the savanna. To date the fish weirs, I excavated a large causeway directly associated with zigzag structures²⁸. Burned wood from the base of the causeway fill was radiocarbon dated to 335 years bp (before present) 20 (OS-17293) or an uncalibrated calendar date of AD 1615 (AD 1595–1635). The corrected date at 68.2% confidence is AD 1490 (0.26) AD 1530; AD 1560 (0.74) AD 1630. Depending on the context, the sample may date or predate the original construction. The Spanish did not control the Baures region until 1708; thus, the earthwork probably predates European occupation.

The earthworks of Baures are an example of creation and active management of an anthropogenic landscape by native peoples. The linear causeways and canals were a sophisticated means of regulating water levels within the savannas to enhance and manage seasonal aquatic resources. The network of fish weirs provided a means of controlling and harvesting fish, in addition to enhancing the habitat and availability of aquatic and terrestrial fauna. The artificial ponds were a means of concentrating and storing live fish, providing drinking water and improving game habitats. Palms growing on earthworks provided additional foodstuffs and materials. Using this simple, but elegant, technology, the people of Baure converted much of the landscape into an aquatic farm covering 500 km². Rather than domesticate the species that they exploited, the people of Baure domesticated the landscape. The fish weirs and ponds produced abundant, storable, and possibly sustainable yields of animal protein. Thus, they were able to sustain large dense populations in what many would consider a marginal environment.

References

1. Posey, D. & Balée, W. (eds) Resource Management in Amazonia: Indigenous and Folk Strategies (Advances in Economic Botany Vol. 7, New York Botanical Gardens, Bronx, 1989).
2. Balée, W. (ed.) Advances in Historical Ecology (Columbia Univ. Press, New York, 1998).
3. Stahl, P. W. Holocene biodiversity: An archaeological perspective from the Americas. Annu. Rev. Anthropol. 25, 105–126 (1996). | Article | ISI |
4. Denevan, W. M. The pristine myth: The landscape of the Americas in 1492. Ann. Ass. Am. Geog. 82, 369–385 (1992).
5. Piperno, D. & Pearsall, D. Origins of Agriculture in the Lowland Neotropics (Academic, San Diego, 1998).
6. Roosevelt, A. C. et al. Paleoindian cave dwellers in the Amazon: The peopling of the Americas. Science 272, 373– 384 (1996). | ISI | ChemPort |
7. Denevan, W. M. The Aboriginal Cultural Geography of the Llanos de Mojos of Bolivia (Univ. California Press, Berkeley, 1966).
8. Erickson, C. L. in Archaeology in the American Tropics: Current Analytical Methods and Applications (ed. Stahl, P.) 66–95 (Cambridge Univ. Press, Cambridge, 1995).
9. Hanagarth, W. Acerca de la geoecología de las sabanas del Beni en el noreste de Bolivia 134 (Instituto de Ecología, La Paz, 1993).
10. Eder, F. J. Breve Descripcion de las Reducciones de Mojos (Historia Boliviana, Cochabamba, 1985 [1772]).
11. Jones, J. Conflict between Whites and Indians on the Llanos de Moxos, Beni Department Dissertation. Univ. Florida (1980).
12. de Castillo, J. in Documentos para la Historia Geográfica de la República de Bolivia:Las Provincias de Mojos y Chiquitos (ed. Ballivián, M. V.) Vol. 1 308–317 (J. M. Gamarra, Pa Paz, 1906 [1676]).
13. Riester, J. Los Guarasug'we: Crónica de sus últimos días (Editorial Los Amigos del Libro, La Paz, 1977).
14. Ortiz Lema, E. Los Matacos Noctenes de Bolivia (Editorial los Amigos del Libro, La Paz, 1986).
15. Beckerman, S. in Amazonian Indians from Prehistory to the Present: Anthropological Perspectives (ed. Roosevelt, A.) 177–200 (Univ. Arizona Press, Tucson, 1994).
16. Chernela, J. M. The Wanano Indians of the Brazilian Amazon: A Sense of Place (Univ. Texas Press, Austin, 1993).
17. Nordenskiöld, E. The Ethnography of South America seen from Mojos in Bolivia Comparative Ethnographical Studies 3 86–102 (Goteborg, 1924).
18. Erickson, C. L. in Vías Precolombinas: Los Caminos, Los Ingenieros Y Los Viajeros (eds Herrera, L. & Cardale de Schrimpff, M.) (Editorial Abya Yala y El Instituto Colombiano de Antropología, Bogotá, Colombia, in the press).
19. Garson, A. Comment upon the economic potential of fish utilization in riverine environments and the potential archaeological biases. Am. Antiquity 45, 62–567 (1980).
20. Dougherty, B. & Calandra, H. Relaciones de la Sociedad Argentino de Antropología Vol. 16, 37– 61 Ambiente y arqueología en el oriente boliviano: La Provincia de Iténez del Departmento Beni. (1984–5).
21. Schmitz, P. I. , Rogge, J. , Rosa, A. & Beber, M. Aterros indígenas no Pantanal do Mato Grosso do Sul (Pesquisas Antropologia no. 54, Instituto Anchietano de Pesquisas, Rio Grande do Sul, 1998).
22. Moholy-Nagy, H. The utilization of Pomacea snails at Tikal, Guatemala. Am. Antiquity 43, 65–73 ( 1978).
23. Swing, C. K. , Isley, W. & Lutz, M. W. Notas sobre la ecología reproductiva del caracol. Pomacea gigas. Ecología en Bolivia 10, 19–31 (1987).
24. Clement, C. R. in Tropical Forests, People and Food (eds Hladik, C. M. et al. ) 139–154 (The Parthenon, Paris, 1993).
25. Beckerman, S. The abundance of protein in Amazonia: A reply to Gross. Am. Anthropol. 81, 533–561 ( 1989).
26. Smith, N. J. H. The Amazon River Forest: A Natural History of Plants, Animals and People (Oxford Univ. Press, New York, 1999).
27. Lee, K. Complejo Hidráulico de las llanuras de Baures (Area a ser protegida), Provincia Iténez, Departamento del Beni, República de Bolivia (unpublished report). (CORDEBENI, Trinidad, 1995).
28. Erickson, C. L. , Winkler, W. & Candler, K. Las investigaciones arqueológicas en la región de Baures en 1996 (unpublished report). (Instituto Nacional de Arqueología de Bolivia, La Paz, 1997).

Acknowledgements

Fieldwork and analysis were supported by the NSF, the Heinz Charitable Trust Foundation, the American Philosophical Society, the Research Funds of the University of Pennsylvania Museum, and Corporación del Beni. K. Lee originally reported the earthworks of Baures in the late 1950s. A.Vranich, O. Saavedra and F. Bruckner did the first study of the Baures earthworks in 1995. I thank the Bolivian Dirección Nacional de Arqueología y Antropología, the Prefectura and Alcaldia of the Department of the Beni, authorities of Baures and Bella Vista, W. Winkler (project co-investigator), K. Lee, H. Schlink, R. Bottega, R. Pinto Parada, E. Bruckner, A. Bruckner, C. Bruckner, O. Rivera, R. Langstroth, W. Denevan, A. Vranich, P. Stahl, R. Dunn and D. Brinkmeier.