John H. Walker's book-length study of the central-western Llanos de Mojos, a region of the Bolivian Amazon covering some 10,000 square kilometers, is a welcome expansion and extension of his previously published archaeological research. The book reveals an anthropogenic landscape of forest islands, raised fields, ring ditches, and communities arrayed along and between the Iruyañez, Omi, Yacuma, and Rapulo Rivers. Walker summarizes the results of his excavations of five forest islands, including the impressive 100-hectare site of El Cerro, and integrates these results with those from his analyses of a corpus of aerial and satellite imagery. The results are interpreted on multiple scales, and contribute to a nuanced understanding of large-scale, complex agricultural systems in the marked absence of a prehistoric state.

Walker places his theoretical and conceptual framework on two main bases: influential theory-based histories concerning intensive agriculture and sociopolitical complexity, and the long tradition of Amazonian archaeology (Chapter 2). He charts a different path, grounded in historical ecology instead of arguments about environmental determinism, building on the work of Tim Ingold, Clark Ericson, and others. Landscape archaeology in this manifestation begins with local histories of agricultural work, placed within “taskscapes,” 7a concept from Ingold that considers landscape features from the perspective of community-level tasks. Walker's taskscapes are groups of related activities across seven defined landscapes that include farming, constructing and maintaining agricultural earthworks, hunting, water control, fire control, and transportation (Chapter 4). In Chapters 3, 4 and 5, Walker describes a variety of landscape features, both anthropogenic (for example, groups of raised fields) and non-anthropogenic (for example, rivers and areas of the savannah that seasonally flood), in relation to each other and to evidence for habitation sites. In this way, Llanos de Mojos is brought to life as a region, one that over the course more than two millennia housed a complex and dynamic set of peoples from (likely) different language families and cultural practices.

A chronological framework based on a sequence of 40 radiocarbon dates and paleobotanical data from pollen cores is summarized in Chapter 3. However, the phase boundaries cannot be correlated with changes in material culture and include a notable gap from 4000 to 1200 cal BC, and future research is clearly needed to refine the region's chronology. A second area where Walker's work reveals the need for more research in Mojos concerns demographic estimates. A large majority (39 of 51) of forest islands that have been surveyed by archaeologists since 1996 show evidence of pre-Columbian occupation. The sizes of forest islands, and their spatial relationships with groups of raised fields, are used to roughly calculate a regional population between 10,000 and 100,000. However, this masks the problematic, or at least critically unexamined, assumptions about the contemporaneity and duration of the use (perhaps unavoidable at this early stage of our understanding of this region's complex history) of both agricultural features and, to a lesser extent, domestic remains, coupled with inherent difficulties in dating natural and agricultural features. As Walker rightly notes, much work remains to be done in this area.

Walker's book lays a thoughtful foundation for future Amazonian archaeologists to build upon. This compact volume nicely summarizes his work to date in the region, including a conceptual approach to a complex landscape with a long history and with abundant evidence of human and non-human occupation and modifications through time. The book should serve as a call to other archaeologists to learn what we can, while we still can, from anthropogenic landscapes in the Amazon under dire threat by economic extraction and deforestation. The lessons about long-term landscape and resource management that the Mojeños have to offer us are ones we desperately need to learn.