Stable isotope evidence for pre-colonial maize agriculture and animal management in the Bolivian Amazon

Abstract

Over the past decade, multidisciplinary research has seen the Amazon Basin go from a context perceived as unfavourable for food production and large-scale human societies to one of 'garden cities', domestication, and anthropogenically influenced forests and soils. Nevertheless, direct insights into human interactions with particular crops and especially animals remain scarce across this vast area. Here we present new stable carbon and nitrogen isotope data from 86 human and 68 animal remains dating between CE ~700 and 1400 from the Llanos de Mojos, Bolivia. We show evidence of human reliance on maize agriculture in the earliest phases before a reduction in the dietary importance of this crop between CE 1100 and 1400. We also provide evidence that muscovy ducks (Cairina moschata), the only known domesticated vertebrate in the South American lowlands, had substantial maize intake suggesting intentional feeding, or even their domestication, from as early as CE 800. Our data provide insights into human interactions with Amazonian ecosystems, including direct evidence for human management of animals in pre-colonial contexts, further enriching our understanding of human history in what was once considered a 'counterfeit paradise'.

Fig. 1. Llanos de Mojos and the studied sites.

Left: map of Llanos de Mojos, generated using QGIS 3.34 (https://qgis.org/) with layer OpenTopoMap (https://opentopomap.org/) under CC BY-SA 3.0. Graphic by T. Hermenegildo. Right: map of the studied sites, adapted from ref. Fig. 9, coordinates in UTM, zone 20S, graphic by H. Prümers.

Fig. 2. Human and fauna stable isotope values from Salvatierra and Mendoza.

Top: δ¹³C and δ¹⁵N values from Salvatierra and Mendoza humans, and from Salvatierra fauna divided according to dietary niches. Bottom: Bayesian-inferred ellipse of the same values. Snake samples not included in Bayesian-inferred ellipse analysis as the sample size (n = 2) is too small to draw reliable inferred values. Groups: ungulates (deer [Mazama sp.] and tapir [Tapirus terrestris]), rodents (agouti [Dasyprocta sp.] and capybara [Hydrochoerus hydrochaeris]), armadillos (Dasypus novemcinctus and Euphractus sexcinctus), muscovy duck (Cairina moschata), and riverines (eels [Lepidosiren paradoxa and Synbranchus spp.] and caimans [Caiman sp.]). The standard ellipse area of isotopic niches represents an estimated 40% of the population. The standard ellipse of the ungulate group omits two outliers (results in Supplement 2) as the goal of this group is to create a reliable baseline of terrestrial C₃-consuming herbivores.

Fig. 3. δ¹³C and δ¹⁵N values from the Salvatierra and Mendoza populations divided according to ceramic phase.

a, Inferred ellipses for the human populations living during the different ceramic phases of Salvatierra and Mendoza. b, Same results excluding exceptional individual LS1218a from phase 5, showing no ellipse overlap between phases 2 and 5. The standard ellipse area of isotopic niches represents an estimated 40% of the population.