‘Polynesians’ in the Brazilian hinterland? Sociohistorical perspectives on skulls,genomics,identity,and nationhood

Introduction

Since 2007, chicken bones recovered from an archaeological site on the Chilean coast and human skulls from the collection of a natural history museum on the Brazilian Atlantic coast have provided archaeological and bioanthropological evidence adduced for a past ‘Polynesian’ presence in South America.

DNA analyses of the Gallus gallus bones dating from 2900 to 500 years BP were described in detail in 2007 by a team led by Alice Storey (Storey et al., 2007, 2011). Since the chickens could not have arrived independently in the New World – ‘the possibility of natural dispersal has been explored and repudiated’ – the scientists argued that the birds must have been brought to South America by trans-Pacific human migrants from Polynesia. Subsequently, Anna-Sapfo Malaspinas and colleagues (Malaspinas et al., 2014a) outlined their investigation of DNA extracted from skulls held in the ‘Botocudo’ collection of the Museu Nacional (National Museum) in Rio de Janeiro (De Mello e Alvim, 1963; Santos, 2012). Originally discovered in the north of Minas Gerais and Espírito Santo States in southeastern Brazil, the skulls were lost in the fire that devastated the Museu Nacional in September 2018, destroying most of the anthropological collections (Phillips, 2018). The temporality in this case differed, since the skulls were identified as belonging to Indigenous individuals who lived ‘prior to the beginning of the 19th century’. The study concluded that two of the skulls were genomically of ‘Polynesian’ origin, although available data gave no clue as to how they reached Brazil: ‘[They] could either represent genomic evidence of Polynesians reaching South America during their Pacific expansion, or European-mediated transport’ (Malaspinas et al., 2014a).

These chicken bones and human skulls are implicated in heated exchanges in the history of science: on the pre-Colombian peopling of the American continent; and on possible interaction between Indigenous populations in the Americas and elsewhere during both the precolonial and the colonial periods. As a central theme in archaeology and physical anthropology since the 19th century, the peopling of the Americas has recently become of key interest to population geneticists working with genomics (Adovasio, 2002; Dillehay, 2000; Meltzer, 2009; Salzano and Callegari-Jacques, 1988). Among the broad set of hypotheses on early human occupation of the continent, one of the most common is that the first waves entered around 14,000–15,000 years ago from Asia via the Bering Strait, northeast of North America. However, other models and migration routes have been proposed, together with the likelihood of past interactions between Indigenous peoples in the Americas and those in other parts of the world such as Polynesia (Jones et al., 2011; Salzano and Callegari-Jacques, 1988). The archaeologist David J. Meltzer (2009: x) summarized the complex historical trajectory of such research since its emergence ‘in modern form’ in the 1860s: ‘multiple site discoveries, conceptual breakthroughs, pivotal moments that have propelled and guided research, and cycles of bitter controversy and grudging, short-lived periods of peace’.

Focusing on human beings, this article critically examines a linked series of debates: on ‘the Botocudo’; on the peopling of South America; on the genetic constitution of the modern Brazilian nation; and on the possibility of past ‘Polynesian’ presence on the American continent. We do not entirely ignore the chickens, since they are implied in the arguments of geneticists who have studied the ‘Botocudo’ skulls. However, the claim that the chicken bones found in Chile support the premise of trans-Pacific human migration in the distant past has since been seriously challenged (Gongora et al., 2008). A further report (Thomson et al., 2014b) argues that ‘previous studies have been impacted by contamination with modern chicken DNA’, and that consequently ‘there is no evidence for Polynesian dispersal of chickens to pre-Columbian South America’. The contending parties have continued to debate the issue publicly (Storey and Matisoo-Smith, 2014; Thomson et al., 2014a).

The case for precolonial interactions between Polynesian and South American populations rests largely on the presumption that human beings introduced the sweet potato (Ipomoea batatas), a South American domesticate, to pre-European eastern Polynesia (Ballard et al., 2005; Denham, 2013; Kirch, 2010: 141–2), though genetic evidence is uncertain (Roullier et al., 2013: 2205–6). Advocates thus enthusiastically embraced DNA analysis of the Chilean chicken bones as providing ‘the first conclusive evidence’ for such contacts (Jones et al., 2011; Matisoo-Smith and Ramirez, 2010: 77). Not only has doubt since been cast on the Polynesian origin of the Chilean chickens, but a recent paper (Muñoz-Rodríguez et al., 2018) argues from genomic analysis of Ipomoea batatas and its wild relatives that the sweet potato was dispersed ‘by natural means’ (wind, water, or birds) from South America many millennia before people occupied Polynesia. Press reports of this paper’s conclusions have been challenged, especially the claim that they ‘question the existence’ of pre-European contacts between Polynesians and South Americans (Fox, 2018; Zimmer, 2018).

We do not endorse the relative plausibility or ‘truth’ of particular hypotheses, but offer a sociohistorical examination of the investigative routes that produced the recent shift to a hyphenized ‘Botocudo-Polynesian’ identity for two crania held at the Museu Nacional. In terms of genetic science, the procedure was relatively direct: sequencing of DNA extracted from the skulls’ teeth pointed to the existence of two genomically ‘Polynesian’ specimens in a collection presumed to have been ‘purely’ ‘Botocudo’. Our interest, though, is less in this final stage than in the intermediary phases. Informed by ideas and debates in social studies of science (see, for example, M’charek, 2005; Pálsson, 2007; Radin, 2017; Reardon, 2005; Simpson, 2000; Sommer, 2016), we probe sequential interconnections between the social and biological domains that resulted in new narratives linking the biological ancestry of the ‘Botocudo’ skulls to ‘Polynesians’. This work redeploys theoretical apparatus employed in Santos’s earlier partnership with anthropologist Michael Kent (Kent and Santos, 2014), which analysed the ‘social life’ of another field study of human biological diversity in Brazil: genetic studies of the Charrúa, an Indigenous group in the country’s south (see also Santos, Kent, and Gaspar Neto, 2014). The result is a multidisciplinary, multiregional history of more than two centuries of knowledge-making in a shifting kaleidoscope of intersecting sciences – anthropology, craniology, biology, archaeology, genetics, and genomics.

Our core argument is organized in five parts. The first analyses the socio-scientific context of genomic research on ‘the Botocudo’ in Brazil in the early 2000s. This work emerged from studies undertaken on global human biological diversity since 1990, notably in conjunction with research programmes like the Human Genome Diversity Project (HGDP). The following two sections trace how the ‘Botocudo’ skulls, which entered the physical anthropology collection of the Museu Nacional in the heyday of racialized physical anthropology in the late 19th century, became key samples in early 21st-century human genomics and morphological inquiries into the biological formation of the Brazilian population. From historical and anthropological perspectives, the fourth section contextualizes the complex, multiphased technoscientific processes underpinning the shift from ‘Botocudo’ to ‘Botocudo-Polynesian’ identity of two of the Museu Nacional skulls. In the final section, a semantic history of the seemingly transparent, but racialized concept of ‘the Polynesians’ concentrates a key argument of the paper: that the fertile alliance of intersecting sciences that is revolutionizing understandings of deep human pasts must be informed by sensitivity to the deep histories of terms, classification schemes, and the disciplines themselves.

Crisis and genesis: The laboratory as microcosm of the nation

The Brazilian geneticist Sérgio Pena is an author of the two studies of ‘Botocudo’ skulls that detected and further examined evidence of ‘Polynesian’ ancestry (Gonçalves et al., 2013; Malaspinas et al., 2014a). Since 2000, Pena and his team, based at the Federal University of Minas Gerais in Belo Horizonte, Brazil, have published numerous works analysing the biological formation of the Brazilian population through genomic markers (Alves-Silva et al., 2000; Carvalho-Silva et al., 2001; Parra et al., 2003; Pena et al., 2011). Initially, they used mostly samples from urban populations in diverse regions of the country. While researching the regional occurrence and distribution of markers present in mitochondrial DNA (mtDNA), the Y chromosome, and the nuclear genome, the geneticists encountered high frequencies of gene sequences classified as ‘Amerindian’. This stimulated their interest in analysing the ‘Botocudo’ material then held by the Museu Nacional.

In 2010, Pena gave an interview in Portuguese to Santos and fellow anthropologist Michael Kent detailing how he had begun genomic studies of the Brazilian population. ¹ Pena stated that in the early 1990s, he was involved in the Human Genome Project (HGP) and was responsible for organizing the South-North Human Genome conference held in Ouro Preto, Brazil, in 1992. Participants included the geneticist Luigi Luca Cavalli-Sforza, who was then attempting to implement the HGDP (Cavalli-Sforza et al., 1991). While the HGP aimed to define general patterns of human DNA sequences through analysis of a small set of genomes, the HGDP’s proponents called for research projects to describe the genomic diversity of the whole species. Biological samples of Indigenous populations were identified as material of central and even primordial importance, since their genomic characteristics were seen as a potential source of information about general human evolutionary history (Radin, 2017; Reardon, 2005; Santos, 2002).

Pena recalled that he accepted Cavalli-Sforza’s invitation, and began structuring a network of investigators to implement the Latin American component of the HGDP:

We began discussing and trying to coordinate research focused on Amerindians in South America. During this period, however, a number of polemical issues began to emerge concerning the study of genetics in native populations. To cut a long story short, after a time we concluded that the HGDP in Latin America should not work with native populations but with cosmopolitan populations instead.

During the interview, Pena described successive events, within and beyond his laboratory, that definitively shaped his work on Brazilian population genetics. In this account, an everyday episode – taking a visiting foreign scientist to the airport in Belo Horizonte – proved decisive when they detoured to visit a major tourist attraction:

In this interim [after it had become clear that conducting the HGDP would be impossible], something highly improbable happened. Hajime Sakamoto, one of the world’s top population geneticists, with whom I had collaborated on several projects, came to visit me in Belo Horizonte. On the day he was leaving, I decided to take him to visit the Lapinha Cave, alongside Confins Airport. When we arrived at the cave, we came across the Lapinha Archaeological Museum, which houses archaeological findings from the Lagoa Santa region. In the exhibition there was something on Lagoa Santa Man, discovered in the 19th century by the Danish naturalist Peter Lund, who had indeed been cited in Darwin’s On the Origin of Species. The thought came to me: if we can’t work with Amerindians in the present, why don’t we do something with their precursors, including Lagoa Santa Man perhaps?

Pena then discussed how his insight became scientific practice in his laboratory. First, samples of prehistoric human bone remains from the Lagoa Santa region were obtained through contacts with archaeologists and bioanthropologists working in the region. Then, to ascertain whether this ancient material still contained DNA sufficiently intact to be sequenced, a partnership was established with a research group in Germany. However, the results were discouraging. When tests conducted in Leipzig proved ‘that there was no chance of any DNA existing’ in the prehistoric human bones, the team ‘abandoned the idea of doing paleogenetics with the Lagoa Santa material’. After this second ‘crisis’, the elements emerged that would delineate the horizons of Pena’s subsequent inquiry into the biological diversity of Brazil’s population:

While verifying the possibility of contamination, a very important matter in paleogenetic studies, Juliana [Alves-Silva, a doctoral student in the project] began to analyze the DNA of people in the laboratory. That was when she discovered that many of them possessed mitochondrial DNA of Amerindian origin.…It was a chance discovery, because nobody at that time expected the fact, at least as far as I know. Nobody had dreamt it possible.

From genetic narratives about the nation to ‘Botocudo’ skulls

While the diversity observed in samples taken from members of Pena’s laboratory served as a kind of ‘test tube’, the next step involved obtaining materials more broadly representative of the Brazilian population in general. So the team decided to investigate samples obtained from laboratories that carried out paternity tests located in various regions of the country. Based on a set of approximately 250 samples of men from different places who self-identified as ‘white’, Pena and his colleagues published several papers after 2000 (see overview in Pena et al., 2011). The Brazilian press gave the results widespread coverage, and the scientists themselves popularized their discoveries. For instance, Pena himself described his research as an attempt to comprehend the processes involved in the genetic formation of ‘Homo brasilis’ (Pena, 2002; Pena et al., 2000).

The decision to take samples from ‘white’ men was not random, since, if DNA analyses revealed high levels of miscegenation in this segment, then even higher levels could be expected in the country’s general population (Pena et al., 2011). Accordingly, special significance was attributed to the discovery of high frequencies of mtDNA markers of Amerindian origin in samples of urban ‘white’ men. One research paper suggested that if ‘approximately one-third of the mtDNA lineages of self-identified “white” cosmopolitan Brazilians had Amerindian origin’, then ‘a naïve extrapolation’ would suggest that ‘roughly 60 million Brazilians’ in a total population of around 190 million would be expected to do likewise (Gonçalves et al., 2010, 2).

In recent years, numerous publications have contextualized this scientific production in population genetics in terms of debates on race, identities, and nation in Brazil (De Souza and Santos, 2014; Kent, Santos, and Wade, 2014; Santos, 2012; Santos, Kent, and Gaspar Neto, 2014). They have located recent genomic research on human biological diversity within varied theoretical and methodological configurations in physical/biological anthropology. From anthropometric studies by physical anthropologists based at the Museu Nacional from the 1870s to the 1920s, to investigations into blood groups in the emerging field of human genetics in the 1950s to the 1980s, or work on DNA sequencing and renewed craniometric analysis over the last two decades, a common narrative thread identifies a high level of miscegenation in the Brazilian population – mixtures arising from the interactions of Indigenous people, Africans, and Europeans, especially Portuguese.

While Pena and his team concentrated first on the high level of biological mixture in Brazil’s population, the question of the ‘Amerindian contribution’ eventually became a specific focus. A paper by Gonçalves et al. (2010) was the first explicitly to analyse samples of DNA taken from the ‘Botocudo’ skull collection at the Museu Nacional. These authors emphasized ‘recovery’ of Indigenous genetic markers through comparative analysis of Indigenous and non-Indigenous samples. They hoped that, by focusing on ‘extant populations that have always lived in small regions once inhabited by specific Amerindian nations’, they might be able ‘to reconstitute the mtDNA haplotype profile of many extinct original native populations’.

The project involved two main sets of samples. The first came from approximately 170 individuals residing in the small rural town of Queixadinha, in the north of Minas Gerais State, ‘homeland’ of the ‘now virtually extinct Botocudo nation’ until their violent suppression by expanding colonization in the 19th century. Previous investigation of these samples by geneticists had refuted any association between skin colour and genomic ancestry in Brazil (Parra et al., 2003). The work of Gonçalves et al. (2010: 2, 4) confirmed the earlier pattern detected in studies of samples from ‘white’ men (Alves-Silva et al., 2000): Those from Queixadinha also presented significant genetic markers of Amerindian ancestry. This result provided the context for analysis of DNA extracted from teeth belonging to 14 skulls held in the Museu Nacional collection and ‘thought to be remains of Botocudo Amerindians’ – ‘one of the best known’ of the ‘once numerous’ Indigenous populations of eastern Brazil. The research identified ‘some mtDNA haplotypes that had not hitherto been described in any other human population studied’, and proposed them as ‘candidate Botocudo haplotypes’.

This inquiry logically implemented the research trajectory imagined by Pena following his visit to the Lapinha Cave in the 1990s. Born of his difficult access to biological samples from living Indigenous populations, his methodological epiphany was materialized in a study that sought ‘to retrieve the genetic lineages of peoples who are now extinct’ by connecting samples from ‘the present-day population’ with those from old Indigenous specimens obtained at a natural history museum (Gonçalves et al., 2010: 7). So, while Pena and his colleagues initially studied non-Indigenous samples and aimed to comprehend the formation of the Brazilian population, the work of Gonçalves and her team not only retained a central interest in the history of the national genetic composition, but also investigated Indigenous population genetics, focusing on supposedly extinct groups considered important in comprehending regional questions.

If in 2010 it could plausibly be claimed that ‘almost nothing’ was known about the ‘genetic makeup’ of the Indigenous occupants of eastern Brazil (Gonçalves et al., 2010: 1), this is no longer so, as questions about the genetics of ‘the Botocudo’ have become increasingly central to scientific research in Brazil.

‘The Botocudo’ in the body of Homo Brasilis

From a historical perspective, it is ironic that the Museu Nacional’s ‘Botocudo’ skull collection has recently become so salient in research on the biological formation of the Brazilian population and its diversity. A century ago, ‘the Botocudo’ were routinely dismissed as ‘primitive’, ‘backward’, an impediment to the country’s development, and destined to vanish. The Exposição Antropológica Brazileira (Brazilian Anthropological Exhibition) held at the Museu Nacional in 1882 vividly exemplified their standing in Brazil in the late 19th and early 20th centuries (Figure 1) (Andermann, 2004; Paraíso, 1992; Sánchez Arteaga, 2017; Santos, 2012; Schwarcz, 1993). Moreover, organization of the exhibition prompted a significant effort to expand the institution’s Indigenous collections, including the addition of 22 ‘Botocudo’ skulls (Reis et al., 2014: 1.1–1.2). João Batista de Lacerda, a physical anthropologist at the Museu Nacional, averred in the Revista da Exposição, a popular magazine produced for the exhibition, that ‘From the moral and intellectual viewpoint, the Botocudo are the expression of the human race at its most inferior level’ (De Lacerda, 1882a). Years later, ‘recollections’ by De Lacerda (1905: 56–7), now director of the museum, described how ‘some indigenes of both sexes, belonging to the Botocudo and Cherente tribes’ were brought to Rio de Janeiro in 1882 ‘with the aim of enhancing the exhibition’, and were variously objectified by scientists. They and others ‘were depicted in oils on large canvases by two eminent Brazilian painters. The Cherente were cast in plaster, life-size, by a foreign sculptor who was in Rio de Janeiro at the time’. De Lacerda himself ‘took anthropometric measurements of each one’ (cf. Bancel et al., 2002; Sánchez Arteaga, 2010).

OPEN IN VIEWER

Figure 1.

‘Botocudos da Exposição Antropologica’, photogravure (Mello Moraes Filho, 1882b: 6).

Despite – or perhaps because of – prevailing pessimism about the prospect of ‘Botocudo’ survival, their skulls and other biological material became valuable specimens in the nascent physical anthropology practiced in Brazil from the second half of the 19th century (De Castro Faria, 1952; Santos, 2012). Marília Carvalho de Mello e Alvim (1963: 1), who curated the physical anthropology collection of the Museu Nacional for several decades from the late 1950s, suggested that few Brazilian Indigenous groups had been ‘so extensively explored in literature, whether generally or specifically, that is anthropologically, as the “Botocudo” Indians’. Apart from numerous reports on their customs and traditions, ‘very often laden with prejudice’, their observation by European and other scholars had produced a ‘rich documentation’ (see also De Castro Faria, 1952; De Lacerda and Peixoto, 1876: 49–53).

‘Botocudo’ skulls were studied by prominent scholars in Europe and North America from the early 19th century. The German comparative anatomist-anthropologist Johann Friedrich Blumenbach (1820: 15–16; Figure 2) published a Latin description of a ‘Botocudo: Brazilian cannibal’, illustrated by an engraving of a single cranium extracted from a grave in 1816 by his protégé, the German traveller-naturalist Prince Maximilian of Wied-Neuwied (1820–1: Vol. 1, 351–2). The United States anatomist Samuel George Morton (1839: 138–40) published a lithograph of the same skull (which he had not seen), along with an English translation of Blumenbach’s text. It notably included the judgement that the skull’s ‘general aspect approaches nearer to that of the Orang Outang [Linnaeus’s Simia satyrus] than any other skull from a barbarous nation’ in Blumenbach’s extensive collection – though Morton also noted Wied’s comment on the ‘admirable exterior conformation’ of ‘the Botocudos’.

OPEN IN VIEWER

Figure 2.

‘Botocudi’, engraving (Blumenbach, 1820: Plate 58), Göttinger Digitalisierungszentrum, available at: https://gdz.sub.uni-goettingen.de/id/PPN641125399?tify={%22pages%22:[39],%22view%22:%22export%22}.

Charles Darwin did not personally encounter ‘the Botocudos’ during his voyage on HMS Beagle, but mentioned them in The Descent of Man (Darwin, 1871: Vol. 1, 247) as ‘closely’ resembling ‘the Fuegians’, whom he had seen in 1834 and ranked at ‘a lower state of improvement than [man] in any other part of the world’ (Darwin, 1839: 235, note *). Darwin (1871: Vol. 1, 181, 246–7) doubly invoked this alleged racial resemblance as evidence that humanity had undergone ‘steady advancement through natural selection’ in the ‘struggle for existence’: to refute the anti-transmutationist concept that ‘all savages’ were the product of ‘degradation’ from an originally civilized state; and to validate his denial that ‘characteristic’ racial differences resulted from the ‘direct action’ of climate and lifestyle – since ‘the Botocudos’ inhabited ‘the finest parts of Brazil’ while ‘the Fuegians’ occupied ‘inhospitable country’ in the continent’s far south. Darwin’s (ibid.: Vol. 1, 153, 200–1, 238) ‘principle of evolution’ nonetheless dictated that the ‘civilised races of man’ would ‘almost certainly exterminate and replace throughout the world the savage races’ – who included ‘the Botocudo’. Mello e Alvim (1963: 1) noted that ‘the type of the Botocudo race’ was thus constructed as ‘a true relic of a living group with features of heightened primitiveness’.

After 1870, ‘Botocudo’ materials were increasingly studied or referenced by European anthropologists and ethnologists, including Rudolf Virchow (1874, 1875), Philippe-Marius Rey (1880), Armand de Quatrefages (De Quatrefages, 1881; De Quatrefages and Hamy 1882, Vol. 1, 476–7; Vol. 2, Plate 74), Augustus Henry Keane (1884), and Paul Ehrenreich (1887). ‘Botocudo’ skeletal remains were also central in publications by Brazilian researchers working in the 1870s and 1880s to establish physical anthropology at the Museu Nacional – notably De Lacerda (De Lacerda and Peixoto, 1876) and his assistant João Rodrigues Peixoto (1885). Such works sometimes implied miscegenation involving ‘the Botocudo’, generally with reference to a distant, pre-Columbian past on a scale of thousands of years. De Lacerda (1882b: 23), for example, argued in the Revista da Exposição that ‘the Botocudos form one of the links in an [ancient] extensive ethnic chain, which fractured in relatively distant times’.

Deep ambiguities in debates on the consequences of miscegenation in Brazil at the end of the 19th century, before and after the abolition of slavery in 1888, are patent in a contribution to the Revista da Exposição by Ladislau Netto (1882), director of the Museu Nacional. He gave full rein to scientific racism’s harsh, but often contradictory terminology in discoursing on the ‘atavistic symptoms in mestizo persons, very often already of a perfectly white color’. Distinguishing diverse varieties of mestizos, depending on ancestral racial stocks, he differentiated black admixtures from Indigenous in terms favourable to neither, but particularly negative about the ‘ferocious animality’ of white-black fusions. In contrast, he attributed ‘some kind of perfectibility of moral character’ and a degree of ‘intellectual development’ to the products of Indigenous-white intermixing.

However, the potential for ‘perfectibility’ was not attributed equally to all Indigenous groups, as evidenced in a disquisition on ‘extinction of the Indians’ by Alexandre José de Mello Moraes Filho (1882a), editor of the Revista. He claimed that ‘intestine wars…were slowly annihilating the indigenous races of Brazil, in the presence of two new races, the Portuguese and the Negro’. He acknowledged that ‘the great epidemics imported by the European’ were destroying ‘the wild aborigines’, while ‘slavery and violence’ were driving them out. However, he also attributed this alleged demise to ‘habits’ incompatible with civilization and to innate racial characteristics – notably, in the ‘northern provinces’, the failure of successful ‘hybridism’ between ‘Indian’ and either ‘black’ or ‘white’ due to excessive racial distance. The ‘Botocudo’ individuals brought to Rio de Janeiro for public display at the exhibition provided ‘proof’ for this brutal extinction narrative: a month’s absence from home had purportedly induced ‘profound changes in their health’, with ‘functional disturbances of key organs’ that ‘could well be fatal’. Mello Moraes drew a harshly Darwinian conclusion: ‘It is the law of successive transformations working in their mysterious way; it is the bundling together of the predisposing and determinant causes, making adaptations difficult’.

We do not aim to produce an in-depth history of research in physical anthropology on ‘the Botocudo’. Moreover, their denigration as exemplars of racial ‘inferiority’ largely vanished in the second half of the 20th century, reflecting more general transformations in anthropology (Lindee and Santos, 2012). Nonetheless, besides ongoing elision of ‘Botocudo’ participation in the biological formation of the modern Brazilian population through miscegenation, they are still objectified as ‘extinct’, ‘primordial’, or even unrelated to modern Amerindians. Thus, Gonçalves et al. (2010: 6–7) characterized ‘the Botocudos’ as a ‘now extinct’ group, supposedly ancestral to the Queixadinha individuals whose DNA samples they studied while seeking Amerindian markers in the ‘genetic admixture’ of a rural population in the Brazilian interior. On morphological grounds, Mello e Alvim (1963) challenged De Lacerda and Peixoto’s (1876: 72) assertion that the ‘Botocudo’ skulls in the Museu Nacional collection ‘very much’ resembled one of its ‘jewels’, an ancient fossil skull from Lagoa Santa. However, since the 1990s, several papers resulting from a series of research initiatives on the peopling of the American continent have combined craniometric measurement with sophisticated multivariate statistical analysis to infer biological relations between ‘the Botocudo’ and the so-called ‘Paleoindian population’ inhabiting South America 6000–7000 years ago (Hubbe et al., 2015; Neves and Atui, 2004; Neves and Hubbe, 2005). In a craniometric study of 32 ‘Botocudo’ crania held at the Museu Nacional, André Strauss et al. (2015: 202, 213–14) argued that ‘the Botocudo represent a case of late survival of ancient Paleoamerican populations, retaining the morphological characteristics of ancestral Late Pleistocene populations from Asia’. In other words, as ‘biological descendants’ of ancient Homo sapiens who migrated from Asia to the Americas thousands of years ago, they allegedly lack ‘visible affinities’ with recent Asians and most other Amerindians.

Over the last century and a half, the production of knowledge by physical anthropologists based on the Museu Nacional’s collection of ‘Botocudo’ skulls has epitomized global transformations and continuities in the discipline’s theoretical and methodological scaffolding (Lindee and Santos, 2012). As genomics has become increasingly important to comprehending the anthropology of the past – including human occupation of the Americas – renewed analysis of bone material has reconstituted these crania as key epistemological elements in generating often ambiguous evidence and proposing new models. This tendency is very evident in recent conjecture about a ‘Botocudo-Polynesian’ biological identity.

‘Polynesians’ in the Museu Nacional collection

The identification of two ‘Botocudo’ skulls as genomically ‘Polynesian’ involved several research stages detailed in a series of papers published over the past decade. The first, already mentioned, is by Gonçalves et al. (2010), who, seeking Indigenous genetic markers in the Brazilian population, partially sequenced mtDNA extracted from the teeth of 14 ‘Botocudo’ skulls held at the Museu Nacional. Their subsequent report (Gonçalves et al., 2013: 6466) explicitly addressed the theme of the past peopling of the American continent, but unexpectedly described ‘the presence of mtDNA haplotypes considered to be typically Polynesian’ in samples obtained from two of the ‘Botocudo’ skulls.

In a third paper, Malaspinas and her team reported use of extended analytic procedures to test the hypothesis that these ‘bona fide Botocudo’ skulls were indeed ‘Polynesian’ (Malaspinas et al., 2014a: R1035–6). Whole genome sequencing of four crania verified that two were genomically ‘Polynesian’, with ‘no detectable Native American ancestry’. In other words, this research did not simply uncover supposedly ‘Polynesian’ markers in specific regions of the genome of the ‘Botocudo’ samples, but identified two individuals whose genomes were shown to be predominantly ‘Polynesian’ in origin. The aforementioned craniometric study led by Strauss did not conduct genomic research on ‘Botocudo’ skulls, but included the ‘two individuals with Polynesian DNA sequences’ (Strauss et al., 2015: 202, 211, 213–14). Though ‘considerably older’, these genetically anomalous individuals were morphologically aligned to the other ‘Botocudo’ specimens. Further complicating the correlation of genomic and craniometric data, a global comparative craniometric analysis carried out by members of Malaspinas’s team associated all the ‘Botocudo’ skulls with a primarily eastern Polynesian cluster (Bernardo and Neves, 2014).

While confirming that ‘at least two Brazilian Botocudo, who likely pre-dated European-Polynesian contact, are of Polynesian genetic ancestry’, these inquiries permitted no definitive conclusions about ‘how they got there’. Each team considered several scenarios (Gonçalves et al., 2013: 6466–7; Malaspinas et al., 2014a: R1036–7; 2014b; Strauss et al., 2015: 212–14). The likelihood that mislabelled skeletal material of ‘Polynesian’ origin might have been confounded with the ‘Botocudo’ collection was considered but discounted on archival grounds (Reis et al., 2014). The possibility that the anomalous ‘Botocudo’ skulls might have resulted from direct precolonial contact between Polynesians and South Americans, though ‘mired in controversy’, was taken seriously, and the ancient Chilean chicken bones cited in its favour – despite the conundrum that the first human genomic evidence was ‘found in Brazil’, rather than along South America’s west coast. The conclusions of Strauss and his colleagues encapsulated the uncertain reasoning about debatable affinities suggested by ambiguous research results: ‘There is no strong evidence from the morphological point of view that these [“Polynesian”] individuals…do not belong to the Botocudo populations’; but ‘it is exceedingly hard to accept, from a genetic point of view’, that they ‘belong to a Native American population’; and ‘a direct connection between Polynesian and Botocudo Indians due to trans-pacific migration is hard to sustain’ (Strauss et al., 2015: 211–12).

Race to Genome: Inventing ‘the Polynesians’

The recurrent quest to know the origins of Native Americans parallels long-held fascination with Pacific Islanders, particularly ‘the Polynesians’, whose remarkable feats of maritime wayfinding earned them the designation ‘Vikings of the sunrise’ (Buck, 1954[1938]). A long-standing, well-founded scholarly consensus favours a route from Asia via coastal New Guinea for the prehistoric discovery and settlement of uninhabited Remote Oceania, which includes Island Melanesia southeast of the main Solomon Islands and all of Polynesia (Anderson and O’Connor, 2008; Kirch, 2010). However, from the late 16th century (De Acosta, 1590: 68), a substratum of myth, fantasy, and scholarship contemplated the prospect of trans-Pacific human migrations, or at least contacts, between Oceania and the pre-Columbian Americas. Never dominant in anglophone discourse, such schemata were sometimes tinged by romantic racialism, which at once idealized the feats of ‘Polynesian’ navigators and avoided negative racial implications of a western origin through Melanesia (the ‘black islands’) by linking ‘the Polynesians’ with a purportedly less ‘primitive’ Indigenous race to the east (Heyerdahl, 1952; Lang, 1834; Langdon, 2001). The significance of the sweet potato in renewed scholarly arguments for precolonial Polynesian interactions with Native Americans has been mentioned, including the paucity of supporting scientific evidence. Since 2000, a productive conjunction of archaeology, biological anthropology, genetics, and molecular biology has been rapidly reformulating scientific understandings of the origins, movements, chronologies, constitution, and affiliations of global populations, including those of Oceania and the Americas (see, for instance, Matisoo-Smith, 2015; Wade, 2018).

To this point, this article has summarized a long history of scientific objectification of ‘the Botocudos’, signalled in the use of the definite article. Vilified as notably ‘inferior’ by the science of race in the 19th century and for half of the 20th, they have become ‘primordial’ in modern physical and biological anthropology. Classed as ‘extinct’, a whole population is reduced to a handful of skeletal remains lodged in museums and subjected to morphological and biological testing. The essentialized status of ‘the Polynesians’ is far less apparent or recognized in the highly competitive, utilitarian intellectual context of global science. Yet ‘the Polynesians’ is no more a natural or Indigenous category than ‘the Botocudos’.

The purely topographic term Polynésie (Polynesia, ‘many islands’) was proposed by a French savant in the mid-18th century to label the ‘multiplicity of islands’ in ‘the vast Pacific Ocean’ (De Brosses, 1756: Vol. 1, 77, 80). The ‘Polynesian race’ and its essentialized human constituents ‘the Polynesians’ were invented by French geographers half a century later to label a ‘tanned’, ‘very fine race of men’ inhabiting the ‘most eastern’ Pacific Islands. They were opposed spatially, physically, and morally to the implicitly inferior ‘black race’ of ‘Oceanic Negroes’ peopling New Guinea, Australia, and the western Pacific Islands (Mentelle and Malte Brun, 1803–5: Vol. 1, 548–9; Vol. 12, 473–4, 577). From 1832, ‘the Polynesians’ constituted the apical element in the French tripartite geo-racial taxonomy of Pacific Islanders, ranked somewhat above ‘Micronesians’ and well above ‘Melanesians’ – who were named for the purportedly ‘black’ colour of their skin, from Greek melas (‘black’), and graded ‘generally very inferior’ (Douglas, 2011; Dumont d’Urville, 1832).

Anglophone terminology, more varied and ambiguous but no less racialized than the French, often collectivized all Pacific Islanders as ‘Polynesians’. For example, British Protestant missionaries differentiated ‘two races of men’ in the Pacific Islands: the ‘black’ ‘Polynesian negroes’ in the west and the ‘copper-coloured Polynesians’ in the east (Brown, 1887: 320–2; Williams, 1837: 503–4, 512–14). Islanders working as indentured labourers on Queensland plantations between 1863 and 1906 were consistently designated ‘Polynesians’, ² though they were recruited mostly from the ‘Melanesian’ archipelagos of the New Hebrides (Vanuatu) and Solomon Islands. They were often distinguished from ‘the true Polynesian’ as ‘Papuan savages’, and consistently denigrated as members of ‘black and inferior races’ (Saunders, 1982: 138; Wawn, 1893: 7, 186).

A norm in the science of race throughout the 19th century, invidious discrimination of ‘the Polynesians’ from ‘the Melanesians’ remained a common trope in professional anthropology well beyond 1950 (Douglas and Ballard, 2008). In the 1930s, the physical anthropologist Peter Buck (Te Rangi Hiroa) – who professed ‘pride of race’ as son of a Māori mother – insisted that ‘the [“Polynesian”] master mariners of the Pacific must be Europoid for they are not characterized by the woolly hair, black skins, and thin lower legs of the Negroids [“Melanesians”] nor by the flat face, short stature, and drooping inner eyefold of the Mongoloids [“Micronesians”]’ (Buck, 1954[1938]: 13–19, 267). Anthropologist Marshall Sahlins’s (1963) ‘historical caricature’ of ‘Political Types in Melanesia and Polynesia’ was implicitly racialist. Euphemized as an evolutionary dichotomy of ‘developed Polynesian and underdeveloped Melanesian polities’, it disparaged the latter as ‘more rudimentary’, ‘constrained’, and ‘truncated’, and the ‘Melanesian’ big-man as ‘thoroughly bourgeois’ in contrast to the ‘regal’ ‘Polynesian’ chief, who was ‘feudal rather than capitalist’ – a peculiar evolutionary history in which bourgeois capitalism preceded and was less ‘advanced’ than feudal aristocracy. Indigenous scholars condemned Sahlins’s paper as ‘clever, thoughtless and insulting’ (Hau‘ofa, 1975: 285–6; Kabutaulaka, 2015: 117). Even in the 1970s, the biological anthropologist William Howells (1973: 7, 232, 261) asserted the reality of ‘races in the form of major population differences: Caucasoid, African, Mongoloid’, all still ‘breeding true’, and lamented the physical anthropologist’s ‘dilemma of getting Polynesians [racially] unmixed through Melanesia’ to Samoa.

Blatantly or tacitly, the term ‘the Polynesians’ is often racially inflected. Yet if ‘the Polynesians’ topped regional racial hierarchies, they were consigned to relatively low strata in global racial and civilizational rankings. In practice, paternalist racial approval or romantic idealization did not save most ‘Polynesians’ from colonial encompassment, exploitation, and more or less drastic depopulation, or some from dispossession and marginality in settler colonies.

These deep histories constitute a largely ignored discursive context in which residues, recursions, or analogies of the racial essentialization of Pacific Islanders inadvertently inhabit some modern studies in genomics and biological anthropology. Two examples in this paper will suffice. In signalling the presence of ‘typically Polynesian’ mtDNA haplotypes in two ‘Botocudo’ skulls, Gonçalves and her colleagues referred to ‘the mtDNA “Polynesian motif”…associated with the Austronesian expansion and settlement of Polynesia and Micronesia’ (Gonçalves et al., 2013: 6466–7). Their primary authority was a paper by Jonathan Friedlaender and his team. Yet these authors qualified the ‘Polynesian motif’ as ‘so-called’, and implied the geopolitical fatuity of calling the pattern ‘Polynesian’ or even limiting it to Austronesian speakers, who inhabit all of Remote Oceania and many coastal areas of Near Oceania (New Guinea, the Bismarck Archipelago, Bougainville, and the main Solomon Islands chain):

The ‘Motif’ is very common in Polynesia, Micronesia, and many parts of Near Oceania, and is absent in the Papuan New Guinea highlands…. [It] becomes very common and almost reaches fixation in some New Ireland and Bougainville groups, but some of these groups speak [much older] Papuan-languages. (Friedlaender et al., 2007: 4, 11; our emphasis)

In our second instance, Strauss and team suggested that the ‘Botocudo’ were ‘biological descendants from early Paleoamerican groups’, and correlated them with ‘recent Australo-Melanesians’, who are by implication equally primordial:

Paleoamericans present the [generalized] morphological pattern that was observed in Southeast Asia by the end of the Pleistocene, when groups carrying this morphology moved northward towards Northeast Asia and subsequently into the Americas, and southward toward Indonesia, Australia, and Melanesia. Later during the Holocene, the morphological differentiation that gave rise to the typical modern East Asian and Amerindian morphology largely replaced traces of the previous morphology in these regions, resulting in visible affinities among Paleoamericans and recent Australo-Melanesians…but not with Asians and most Native Americans. (Strauss et al., 2015: 213–14)

Conclusion

By examining successive junctures between social and biological domains, this article has demonstrated that novel interpretations of the biological ancestry of ‘the Botocudos’, specifically the discovery of ‘Botocudo-Polynesians’ in the Museu Nacional skull collection, did not initially result from a search for ‘difference’ – for links with distant populations in the Southern Hemisphere. Ironically, it emerged during a research project on the Brazilian national ‘self’, a scientific investigation begun in the 1990s to explore the biological origins and formation of the Brazilian population from a genetic perspective, in the context of broader initiatives such as the HGDP. The bioanthropological trajectories of human populations across the world had been a staple of inquiry in physical anthropology since the mid-19th century. However, from the 1990s older procedures like anthropometry and blood group typing were complemented or superseded by DNA analysis (Lindee and Santos, 2012; Sommer, 2016). In this setting, DNA analyses of the collection of ‘Botocudo’ skulls – conceived as representatives of a ‘primitive’, ‘primordial’ Brazil – produced a radically different outcome from that anticipated – namely, biological markers characteristic of two distinct regions of the planet, South America and Polynesia.

A few years before the 1882 Exposição Antropológica, De Lacerda and Peixoto (1876: 71–2) published a detailed craniometric study of the Museu Nacional’s growing collection of Indigenous skeletal remains. Having equated skulls classified as ‘Botocudo’ with ‘the fossil cranium from Lagoa Santa’, they discerned morphological correspondence – purported ‘small cranial capacity’ – between ‘the Botocudos’ and the Indigenous inhabitants of New Caledonia and Australia, ranked ‘among the races most notable for their degree of intellectual inferiority’, with ‘very limited’ skills or capacity ‘to join the path of civilization’. While not suggesting actual biological connection between Brazil, on one hand, and New Caledonia and Australia, on the other, De Lacerda and Peixoto nonetheless inferred broad racial affinities from highly subjective cranial measurements presumed to index intellect (Gould, 1996).

Nearly a century and a half later, Strauss and his team correlated De Lacerda and Peixoto’s morphological conflation of ‘the Botocudo Indians’ and ‘the early inhabitants from Lagoa Santa’ with their own claim of ‘visible affinities among Paleoamericans and recent Australo-Melanesians’ (Strauss et al., 2015: 209–14).

There are multiple ironies in such racial recapitulations and recursions. Nineteenth-century physical anthropologists drew an intellectual analogy between ‘the Botocudo’ and supposed equally ‘inferior’ populations on the other side of the planet. Recent physical anthropologists materialized that analogy as an ancient ‘generalized morphology’, common to ‘Late Pleistocene/Early Holocene groups worldwide’ and enduring in certain (primordial) modern populations such as ‘Australo-Melanesians’. Simultaneously, geneticists suggested a real ‘Botocudo-Polynesian’ physical presence in the interior of Brazil, dated to some centuries ago.

A very recent technical ‘revolution’ in the genomic sequencing of ancient DNA extracted from bones and teeth in human remains of archaeological origin is reformulating population genetics by enabling much deeper, more precise histories of global and regional human movements and interactions (Reich, 2018). For example, more and better genome sequencing of ancient skeletal remains has exploded long-standing claims that ‘Paleoamericans’, including the Lagoa Santa people, are morphologically distinct from other Native Americans. However, this new research also suggests the still-inexplicable presence of ‘Austro-Melanesian’ or ‘Australasian’ genomic signatures in several ancient Brazilian populations, again including the Lagoa Santa people (Moreno-Mayar et al., 2018; Posth et al., 2018; Raghavan et al., 2015; Skoglund et al., 2015). Acknowledging the shifting authority in population studies from craniometric analysis to genomics, Strauss reportedly admitted that ‘skull shape isn’t a reliable marker of ancestrality or geographic origin. Genetics is the best basis for this type of inference’ (Moon, 2018).

Genomic research on ancient DNA has strong media appeal, but also provokes intense controversy. Alongside excited recognition of the potential for innovative knowledge production about human pasts, critics – often archaeologists whose work has been used by geneticists – deplore overambitious interpretive schemes based on limited sample sizes of ancient DNA, as well as a tendency to reapply the discredited biological concept of race to the human species (Kahn et al., 2018; Lewis-Kraus, 2019). New interpretations of genomic sequences cannot be dissociated from at least two centuries of complex scientific history and asymmetrical colonial relations. When museum items leave the technical reserves in which they are held, such as the Museu Nacional, and enter the ‘purified’ spaces of molecular biology laboratories, they remain stained with the past. Neither words – ‘the Botocudos’, ‘the Paleoindians’, ‘the Polynesians’, ‘the Australo-Melanesians’ – nor the ‘Botocudo’ samples themselves are delivered ‘decontaminated’ from the multiple complex layers of classification they defined or to which they were subjected as objects of almost two centuries of scientific inquiry.