In the ongoing debate about the role of agriculture in the development and decline of Classic (250–950 CE) Maya lowland societies, the archaeological site of Río Bec has a special place. It is one of the few settlements in which, years before the use of LiDAR to survey the Yucatán peninsula, a multi-scalar geoarchaeological study (Río Bec Project 1, 2002–2010) allowed us to define a field system—that is, landed domains of distinct fields organized around their local networks of household units. The dimensions of the fields, their correlation to the size of their corresponding households, and their long period of use (200 years, 700–900 CE), during Río Bec's apogee, raise questions about the fundamental issues of land use practices (intensification, diversification, and specialization) and their sustainability over what was a relatively long period. Those issues are at the heart of a new research program at Río Bec, which we began in 2019, combining an archaeobotanical approach with land use and settlement pattern modeling, ultimately aimed at correlating local agricultural production with demography (Río Bec Project 2, 2019–2022). Based on the knowledge of field and settlement systems in their environment that we accumulated during our past research, the paper starts with a description of the field system in one sector of the Río Bec site, then focuses on the diversity of its plots on the household scale and proceeds to discuss the preliminary correlation that can be established between agricultural intensification and demography at Río Bec at the neighborhood scale and the settlement scale.

The Mexican Central Highlands have been a major cluster for human settlement since pre-Hispanic times and its soils provide much of the food for modern Mexico. Rainfed cultivation yielded and still yields most of the agricultural products. However, pre-Hispanic rainfed cultivation has been less documented than other practices. Moreover, textual and ethnographic records, mostly postdating the deep modifications made by the Spanish conquerors, have long been prevalent in studies on pre-Hispanic farming and have tended to bias our conception of the latter. Archaeology provides new key information but struggles to address rainfed techniques, which leave few remains behind in some landscapes. To that regard, spatial approaches considering geoecological parameters are helpful. Furthermore, remote sensing techniques and airborne laser scanning (LiDAR), above all, offer increasing potential for feature detection and provides new ways to address fossilized landscapes at both archaeological and environmental levels. This paper offers new insights on pre-Hispanic rainfed cultivation through an interdisciplinary approach. It focuses on archaeological settlements on malpaís landforms (young and rugged lava flows) in the Malpaís de Zacapu, in western Mexico. There, a method combining fieldwork and remote sensing in archaeology and soil science was developed to reassess pre-Hispanic farming. After presenting the method and this study's main results, I discuss the latter in light of examples of ethnohistorical and ethnographical uses of malpaís landforms. They suggest that widespread conceptions about agricultural soils of the Mexican Central Highlands held by external observers differ from Indigenous and local farmers' notions, which seem partly inherited from pre-Hispanic times.

Despite an attempt at intercontinental synthesis by Denevan and Turner (1974) almost 50 years ago, studies of agricultural raised fields (RF) in the Neotropics and in Africa and New Guinea are separate research traditions, with almost no communication between them. Neotropical studies refer to “raised-field agriculture” and almost exclusively concern archaeological systems in wetlands. Studies in Africa and New Guinea refer to “mound” or “ridge” cultivation and concern mostly present-day systems (in Africa) or both present-day and archaeological systems (in New Guinea) in both uplands and wetlands. Ethnographic studies of present-day systems provide insights into questions about past systems that are inaccessible using archaeological methods alone. Our review suggests that the Neotropical focus on RF agriculture as an exclusively wetland adaptation is misleading. We argue that the most widespread purpose of building RF, in both wetland and upland environments, is to concentrate topsoil and organic matter, enabling creation of fertile patches in infertile and low-biomass grassland environments. Avoiding flooding is an important function of RF built in wetlands and wetland margins. We further show that Old World RF are often not perennial, but are short-lived structures that rotate over the landscape, being torn down and rebuilt nearby in successive cycles. Short fallow periods are allowed (or even favored) by methods of managing fertility. Finally, we argue that the restriction—on all continents—of archaeological raised fields to wetland and wetland-margin environments is, in part, a result of their better preservation from erosion in wetland than in upland environments.

In the tropics, the reduction in fallow periods in shifting rice cultivation and deforestation have led to soil degradation. How crop diversity is managed by farmers to adapt to this change remains poorly studied. Our study in Madagascar focuses on the management of 38 species in agroforestry, a practice that was gradually adopted by Betsimisaraka farmers to replace shifting rice cultivation. We describe how farmers perceived changes in the soil and how they adapted their farming practices to these changes, and analyze their knowledge of the interactions between the different plant species and the soil (soil-agrobiodiversity interactions) that underlie these adaptations. The farmers' perceptions of changes in the soil, their causes and consequences, were recorded in 19 individual interviews and three focus groups. Farmers' knowledge of soil-agrobiodiversity interactions was recorded in 84 individual surveys and one focus group. Farmers grouped soils in two main classes according to the topography and used four criteria to characterize them. The main change in the soil they observed was the decline in soil fertility due to deforestation. In response to these changes, farmers changed their crop species and management practices (e.g., spatial organization of crops, fertilization, species associations). These adaptations are based on shared knowledge of the soil requirements of crop species and their effect on fertility, despite the recent adoption of agroforestry. Our study highlights the dynamic and holistic dimension of farmers' knowledge of the soil and its interactions with different plant species.

Semi-domesticated plants are good resources for agroecology, because they can survive without human intervention and can be managed together with non-domesticated plants. To explore the ecological know-how underlying farmer management of semi-domesticates at the landscape level, we compared how Yanacona Indigenous People in the southwestern Colombian Andes manage two semi-domesticated species of Phaseolus beans, both locally named cacha (Phaseolus coccineus and P. dumosus), and a full domesticate (the common bean Phaseolus vulgaris). We investigated what functional traits of Phaseolus beans farmers identify and how farmers link these traits with the ability of beans to develop in different cultural-ecological niches within the landscape. In 60 semi-structured interviews conducted with Yanacona farmers, we found that (i) farmers compare functional traits of the different bean species, including perenniality, phenology, resistance to pests and diseases, and ability to compete with surrounding vegetation; (ii) farmers recognize and use the variation generated by hybridization between the two cacha species; (iii) farmers take advantage of the traits of cacha to grow them under a wide array of niches at the landscape scale, including low-intervention spaces, such as hedges and swidden plots in fallow, and have detailed knowledge of cacha beans' interactions with trees; and (iv) cacha traits contribute to the resilience of agroecosystems through the management practices and social relationships with which they are associated. Emic and etic perspectives must be combined to gain full understanding of the roles of functional traits of cultivated plants.

Understanding the social and ecological contexts of past agricultural systems in complex societies requires expansive and nuanced data sets that recognize the role of all players in the production system. Such data sets are not often available and thus, there is a tendency to generalize across polities and ecosystems and to homogenize place- and time-specific variation. Here, we bring together ethnohistoric, ethnographic, and archaeological data to explore Mā‘ohi commoner and elite involvement in the production systems of the Society Islands at the time of European contact (AD 1767). We focus our analysis on the archaeological records of five polities located in four different watersheds on the islands of Mo‘orea and Ra‘iātea. We divide the polities into those that are elite- vs. commoner-centric and those that are located in productive versus marginal agricultural landscapes. We find that elites have a greater presence and closer association with agricultural production in productive ecological settings than in the more marginal ones. Although the archaeological expression of the agricultural systems look superficially the same in all contexts, maintaining productivity in the marginal contexts would have required different knowledge and more effort on the part of the Mā‘ohi farmer than in the more productive settings. In contrast to previous summaries of Mā‘ohi agriculture that focus on elite control and seasonal shortages, we highlight the place-based knowledge of Mā‘ohi commoners that was foundational to the centuries-old production systems that provisioned both the elite and commoners alike.

In several parts of the world, forests have been safeguarded by Indigenous and traditional people, whose plant food production is mostly for self-consumption and for sale based on the promotion of forest biological processes and crop diversification. In the Brazilian Atlantic Forest, Quilombola groups are black communities that have protected and managed agrobiodiversity since the fifteenth century. Although the dynamics of use, production, and donations of Quilombola food plants are still poorly understood, these processes can help us to understand the vulnerability to food insecurity. We analyzed the food security related to food availability in a Quilombola community in southern Brazil (São Roque Pedra Branca), focusing on their dependence on locally produced food plants. Today, São Roque families depend on agriculture, government benefits, and urban low-paid civil construction and general helper jobs. We evaluated the interactions between Quilombola farmers and the cultivated species and varieties and the role of farmer families within the community. The hypothesis is that farmers who produce, manage, and conserve more local species and varieties of plants also contribute to less food vulnerability within the community. The small-scale agriculture carried out by the Quilombolas implies the management of high inter- and intraspecific diversity. During the 2019 agricultural year, 42 species were cultivated in plots and gardens, comprising 83 varieties. Most farmers cultivate a subset of major species and varieties for self-consumption and donations. The farmers who grow the most varieties are the ones who donate the most. Some vulnerability to food insecurity was observed in 53% of the family units. Exchanges and donations between families contribute to the community's food security, but there are still families in a situation of food insecurity. Local agrobiodiversity and established exchange networks strengthen Quilombola food sovereignty.