Ancient food for thought

Article metrics

Archaeological evidence of unexpected modes of food production in the tropics of lowland Central and South America carries lessons for modern farmers and students of crop-plant evolution.

In the debate about how to make rational, sustainable use of the American tropics, some of the more enlightened governments are taking a hard look at indigenous farming practices. Two papers (one on page 190 of this issue1, the other in the 19 October issue of Nature2) show that archaeologists have a great deal to contribute by revealing crop regimes and forms of land management that have no counterparts today.

The conventional view that seasonally flooded savannas are worthless for cultivation and are suitable only for cattle ranching has been overturned by Erickson's work1 in the Bolivian Amazon (Fig. 1). He has used aerial photographs and ground surveys to map a 'lost' prehispanic landscape of man-made earthworks, with settlement mounds, causeways, raised and drained fields, and also 500 km2 of weirs and artificial ponds — which, he argues, were used for fish harvesting on a massive scale.

Figure 1: Central and northern South America today.

Erickson's paper1 describes investigations in the Bolivian Amazon and the identification of prehispanic raised fields, as well as weirs and artificial ponds. Piperno et al.2 report evidence of starch grains 5,000–7,000 years old from the Aguadulce rock shelter, Panama. Taken together with other evidence, these findings point to the early and independent development of root-crop agriculture in the lowland American tropics.

Erickson does not say whether the Bolivian government plans to reactivate the fish weirs, but experiments have already begun with his raised fields3. The archaeological sites yielded pollen of edible tubers (Xanthosoma species), and both maize and manioc (cassava) did well on the modern experimental plots. Several older farmers told Erickson it was the first time they had seen the pampas produce agricultural crops. Similar, and successful, experiments have been carried out with high-altitude raised fields in the Titicaca basin.

This may be just the start of things. In the savannas of Caribbean Colombia there are 5,000 km2 of abandoned prehispanic fields and canals, unknown until the 1960s, and there are other, smaller expanses of raised fields in Mexico, Guatemala, Belize, Venezuela, Surinam, Guyana, Ecuador and Peru. Written documents from post-conquest times say nothing about the technology of these systems. But if further experiments show that the ancient agricultural landscapes can be reconstituted, and the savannas can be made productive, we may have an alternative to reckless felling of the rainforest.

Archaeological data of this kind help to resolve the dispute between historical anthropologists who accept European reports of large populations along the Amazon floodplain and in the savannas, and those members of the calorie-counting school who claim that protein scarcity imposed a limit on population growth in the lowland tropics until the introduction of European livestock. There is good evidence that a combination of game, fish, maize and palm fruits in seasonally flooded areas provides abundant protein4. Instead of worrying about protein, we should perhaps now think more about where the starch came from.

Taking the long archaeological view, this is not a silly question, especially for the initial stages in the evolution of agriculture. In the 1970s and 1980s archaeologists began to recover pollen, phytoliths (microscopic bodies of silica found in plant cells) and carbonized plant tissue from tropical sites and, more recently, to identify tuberous-plant starch grains that are not normally preserved. Piperno and colleagues2 report the finding of starch grains from cultivated manioc, yams, arrowroot and maize on milling stones excavated at the Aguadulce rock shelter, Panama, and dated to between 5,000 and 7,000 years ago. This is the earliest evidence for root-crop cultivation in the Americas, and raises a series of questions.

The first generation of domesticated plants will look exactly like the last generation of wild ones, so botanists will never be able to recognize the initial experiments with cultivation. In Panama, the cultivated plants had already begun to diverge from their wild prototypes. The presence of maize and manioc is also significant, as these are not native to Panama and must have been brought into cultivation elsewhere. The genetic history of maize is still controversial5 but the consensus is that the botanical centre of origin is somewhere in Mesoamerica (Mexico and northern Central America), and most plausibly in Mexico. Manioc was probably domesticated in southwest Brazil6. These plants must therefore have reached Panama by a process of secondary dispersal, as domesticated crops and after a period of genetic adaptation to new ecological conditions.

The implication here is that we should be looking for a still earlier stage of tropical agriculture, without maize or manioc (the principal sources of starch today), and characterized by experimentation with whatever was locally available. There are hints of this in the pollen and phytolith cores from the lake of La Yeguada, not far from the Aguadulce site7. Here, below levels with maize pollen dated to about 5,700 years ago, there is a sudden and large rise, around 11,000 years ago, of particulate charcoal and of plants typical of forest gaps. This phenomenon may well represent small-scale forest clearance for horticulture.

Like Erickson's fish farming, the earliest forms of tropical cultivation may be unlike anything that can be seen today. But recent excavations all over the lowlands point to the importance of tubers such as yams, arrowroot, sweet potato and Xanthosoma, and also emphasize the vital role of palm products. Arrowroot, a neglected crop today, is beginning to emerge as a significant early domesticate, with finds from highland Colombia dated to 10,000–9,000 years ago8 and from coastal Ecuador to around 3,600 years ago9. All this tends to confirm what Carl Sauer10 suggested as early as 1952: that there was an early and independent development of root-crop agriculture in the lowland American tropics, comparable in importance to the maize–beans–squash agriculture of Mesoamerica or the high Andean combination of potatoes, seed crops (quinoa) and llamas.

Perhaps we should now bury for ever the concept of 'centres of origin' for American agriculture. Unlike plants and animals, ideas are not constrained by ecological conditions. Archaeology is beginning to demonstrate the huge variety of agricultural practices in pre-European America, and also to suggest that people everywhere began by experimenting with the cultivation of plants they were already collecting in the wild.


  1. 1

    Erickson, C. L. Nature 408, 190–193 ( 2000).

  2. 2

    Piperno, D. R., Ranere, A. J., Holst, I. & Hansell, P. Nature 407, 894–897 (2000).

  3. 3

    Erickson, C. L. in Archaeology in the American Tropics: Current Analytical Methods and Applications (ed. Stahl, P.) 66–95 (Cambridge Univ. Press, 1995).

  4. 4

    Beckerman, S. Am. Anthropol. 81, 533–560 (1979).

  5. 5

    MacNeish, R. S. & Eubanks, M. W. Latin Am. Antiquity 11, 3–20 (2000 ).

  6. 6

    Olsen, K. M. & Schall, B. A. Proc. Natl Acad. Sci. USA 96, 5586–5598 (1999).

  7. 7

    Piperno, D. R. in Archaeology in the American Tropics: Current Analytical Methods and Applications (ed. Stahl, P.) 130–153 (Cambridge Univ. Press, 1995).

  8. 8

    Gnecco, C. & Mora, S. Antiquity 71, 683–690 (1997).

  9. 9

    Pearsall, D. M. in Archaeology in the American Tropics: Current Analytical Methods and Applications (ed. Stahl, P.) 126 (Cambridge Univ. Press, 1995).

  10. 10

    Sauer, C. O. Agricultural Origins and Dispersals (Am. Geog. Soc., New York, 1952).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.