We agree with Dolores Piperno and Crystal McMichael that more paleo-botanical and archaeological data are necessary to build a more complete view of the impact of pre-Colombian populations in interfluvial areas. Our study, however, presents a novel (for Amazonia) botanical and ecological approach to investigate the historical human impact on forested landscapes of interfluvial areas in Central Amazonia. Using these methods, in addition to some archaeological methods, we provide new information to better understand the legacy of past human management on modern vegetation beyond occupation sites. This approach was also used in the Maya lowlands (Campbell et al. 2006 and Ross 2011), France (Dambrine et al. 2007) and Central Africa (van Gemerden et al. 2003), where results similar to ours contributed to understanding former landscape domestication. Before addressing Piperno and McMichael’s specific criticisms, we feel it is important to clarify that we did not “claim to present archaeological evidences for prehistoric human impacts”, rather we presented botanical and ecological evidence, with some archaeological evidence, that we interpret as being due to past human activities in forested landscapes. Here after we present our thoughts about each of the three criticisms made by Piperno and McMichael:
1) “No archaeological evidence was provided for the interfluvial areas without indications of Amazonian Dark Earth (ADE).”
Across the continuum of domesticated landscapes the degree of human modification varies from intensive (occupation sites and cultivated areas) to less intensive (managed areas) to not intensive (promoted areas) (see Clement (1999), who followed Harris’ (1989) model with different terminology). The archaeological evidence requested by Piperno and McMichael is abundant in occupation sites and cultivated landscapes, rare in managed landscapes and mostly absent in promoted landscapes. Hence, beyond occupation sites (ADE) and cultivated areas we did not expect to find ADE or significant amounts of charcoal and phytolith records of crop species, but we expected to find signs of past human management by assessing the distribution and abundance of useful woody species, as found in the Maya lowlands, Central Africa and France. For this reason, we provide not only archaeological evidence, but principally information regarding the modern forest’s composition that seems to be related to past human use and management. Using relative abundance, richness, and basal area of useful woody species, we found a strong pattern between these variables and the distance from the rivers (a proxy for distance from occupation sites). Piperno and McMichael considered that “these evidences do not add anything in this regard”; however, they give no alternative explanation for the pattern we found, other than to request vegetation reconstruction (below). We agree that archaeological evidences are essential to understand human settlements in occupation sites and cultivated areas, but we do not agree that botanical information is uninformative where the archaeological record is rare or absent. Perhaps Piperno and McMichael are expecting too much from the less intensively modified parts of the domesticated landscape continuum.
2) “…the charcoal data is largely un-interpretable with respect to past fire disturbance because none of it was dated.”
We recognize that dating the charcoal provides more precise information, but we do not agree that it is “largely un-interpretable”. Piperno and McMichael argue that “the uppermost 20 cm in soil profiles is often modern or just a few hundred years-old”, however, almost half of all charcoal dated in the uppermost 20 cm in soil profiles of the southern part of the Purus-Madeira interfluve (McMichael et al., 2012, Table S2) is pre-Columbian. Some of our charcoal samples were collected not more than 40-60 km from their sites with superficial pre-Columbian charcoal. Is it likely that our charcoal will be so different, given that it is so close geographically? To find out, we will collaborate with the new Radioecology Laboratory at the Fluminense Federal University (UFF), the first such laboratory in Brazil, for dating the charcoal of this study. In a forthcoming article, we will present a fire chronology of our sites and compare them more completely with McMichael et al.’s and other archaeological studies from Central Amazonia.
3) “…other interpretations made of prehistoric land usage are equivocal because vegetation reconstructions were not attempted.”
The lack of vegetation reconstructions does not make other interpretations equivocal, given the similarity of our results and interpretations to those from the Maya lowlands, Central African and France, and the fact that new methods do not invalidate earlier methods. Since the majority of the interfluve landscape was not converted into cultivated fields, the footprint of old settlements and fields are not visible today, but the less intensive manipulation of the forest composition is evident on the modern useful woody community, as we reported. We fully agree with Piperno and McMichael that the combination of archaeological methods to reconstruct past vegetation (phytolith studies) and a dated fire chronology, along with modern floristic studies such as we presented, will allow a more complete understanding of prehistoric landscape modification persisting in the Amazonian forest today. We intend to expand our research in this direction. However, contrary to Piperno and McMichael’s wide-ranging assertion that “phytoliths provide information on non-agricultural forms of landscape management through high visibility of species thought to have been artificially concentrated, such as palms, along with good representation of a variety of tree and other species whose frequencies would be expected to decrease in phytolith records upon their selective removal or human enrichment of others”, their own data set (McMichael et al., 2012, Suppl. Mat., p. 5) simply groups sets of tree-rich families, suggesting that the combination of modern floristic and phytolith data will be essential to interpret human modification of managed and promoted landscapes. In addition some of the major fruit species, such as Bertholletia excelsa (Brazil nut) and Theobroma (wild cacao), that we used as indicators of anthropogenic forests, do not leave a phytolith record (McMichael et al., 2012, Suppl. Mat., p.9; Piperno, 2006, Table 1.1, p.7). Even palms, which enjoy somewhat better phytolith discrimination, would need a more refined sampling methodology to permit the analysis that Piperno and McMichael suggest, as widely scattered test pits, such as those used by McMichael et al. (2012), are insufficient to provide the landscape level detail that Piperno and McMichael request here. For example, in 47 of 50 samples of their 13 locations in the Purus-Madeira interfluve forests, conical phytoliths were absent (e.g., Bactris gasipaes (peach palm) and Astrocaryum), including sites 120-125 with no phytoliths at all (McMichael et al., 2012, Suppl. Mat., p.9). The lack of Bactris-Astrocaryum phytoliths in general is surprising, since species of these genera are common understory palms throughout the interfluve (see Henderson, 1995). The absence of phytolith records does not mean that the species are not present in the landscape, since for the same area that McMichael et al. (2012) did not find Astrocaryum in the phytolith record (ancient or modern) we register more than two thousand individuals of Astrocaryum in a 3 ha of palm inventory (T Emilio, personal observation). In addition, we found individuals of Astrocaryum aculeatum in one of our study sites (M6) that is not far from the 121 site of their study. Thus, a combination of phytolith and modern floristic data seems to be a fruitful approach to follow. We will be delighted to collaborate with Piperno and McMichael to develop this more multidisciplinary approach in Amazonia.

Campbell DG, Ford A, Lowell KS, Walker J, Lake JK, et al. (2006) The feral forests of the eastern Petén. In: Balée W, Erickson CL, eds. Time and complexity in historical ecology: Studies in the Neotropical lowlands. New York: Columbia University Press. pp. 21–55.
Clement CR (1999) 1492 and the loss of Amazonian crop genetic resources. I: The relation between domestication and human population decline. Econ Bot 53: 188–202.
Dambrine EJL, Dupouey J-L, Laüt L, Humbert L, Thinon M, et al. (2007) Present forest biodiversity patterns in France related to former Roman agriculture. Ecology 88: 1430–1439.
Harris DR (1989) An evolutionary continuum of people-plant interaction. In: Harris DR, Hillman GC, eds. Foraging and farming: The evolution of plant exploitation. London: Unwin Hyman. pp. 11-26.
Henderson A (1995) The palms of the Amazon. New York: Oxford University Press.
McMichael CH, Piperno DR, Bush MB, Silman MR, Zimmerman AR, et al. (2012) Sparse pre-Columbian human habitation in Western Amazonia. Science 336: 1429–1431.
Piperno DR (2006) Phytoliths: A comprehensive guide for archaeologists and paleoecologists. Lanham, MD: AltaMira.
Ross, NJ (2011) Modern tree species composition reflects Ancient Maya ‘forest gardens’ in NW Belize. Ecol Appl 21(1): 75-84.
van Gemerden BS, Olff H, Parren MPE, Bongers F (2003) The pristine rain forest? Remnants of historical human impacts on current tree species composition and diversity. J Biogeogr 30: 1381–90.