AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 00:00–00 (2014)
Oral Health of the Paleoamericans of Lagoa Santa, Central Brazil Pedro Da-Gloria,1* and Clark Spencer Larsen2 1 Laboratorio de Estudos Evolutivos Humanos, Departamento de Genetica e Biologia Evolutiva, Universidade de S~ ao Paulo, CP11461, S~ ao Paulo, Brazil 2 Department of Anthropology, The Ohio State University, Columbus, OH 43210
KEY WORDS America
dental caries; abscess; western hemisphere project; first Americans; South
ABSTRACT The peopling, origins, and early prehistory of the Americas are topics of intense debate. However, few studies have used human remains to document and interpret patterns of health and lifestyle of Paleoamericans. This study provides the first investigation to characterize oral health in a series of early Holocene skeletal remains from Lagoa Santa, Brazil, a locality containing the remains of some of the earliest inhabitants of South America (10,000–7,000 BP). The sample is composed of 949 teeth and 1925 alveoli from an estimated 113 individuals excavated from 17 archaeological sites located in the State of Minas Gerais. We compare dental caries and abscess prevalence at Lagoa Santa to a large sample of human skeletons from the Western Hemisphere Project (WHP) database using both individual and tooth/alveolus count
methods. In addition, antemortem tooth loss and tooth wear were analyzed in Lagoa Santa by sex and age. The results show that Lagoa Santa dental caries and abscess prevalence are significantly higher than observed among other hunter–gatherers included in the WHP database, except when abscess prevalence is considered by individual count. Adult females have less tooth wear coupled with higher prevalence of dental caries and antemortem tooth loss than adult males. These results point to an unexpected record of poor oral health at Lagoa Santa, especially among females. A diet based on a highly cariogenic combination of wild tubers and fruits is suggested as an explanation for the elevated rate, characterizing an early adaptation to a tropical environment in South America. Am J Phys Anthropol 000:000–000, 2014. VC 2014 Wiley Periodicals, Inc.
The place of origin (Kidd et al., 2011; Stanford and Bradley, 2012), route(s) of migration (Mandryk et al., 2001; Jantz and Owsley, 2005), number and size of migration(s) (Schurr, 2004; Neves and Hubbe, 2005; Fagundes et al., 2008), and dates of arrival (Achilli et al., 2008; Waters et al., 2011) of the first inhabitants of the Americas are subjects of intense debate. The Lagoa Santa region of central Brazil has contributed to discussions regarding migration routes and biological origins due to the remarkable preservation of human skeletal and dental remains dated to the Pleistocene/ Holocene boundary (Neves et al., 2003, 2007). To date, research on the human remains recovered from Lagoa Santa has focused on cranial morphology and dating. New radiocarbon dates have recently confirmed that the Lagoa Santa rockshelters were occupied during the early Holocene (Araujo et al., 2012). Until now, few studies have used assemblages of human remains from Lagoa Santa to infer subsistence and living conditions of those early inhabitants (for exceptions see Neves and Cornero, 1997; Neves and Kipnis, 2004). Traditionally, studies of Paleoamerican1 subsistence have focused on Clovis kill sites in North America, emphasizing the importance of big game hunting to the first inhabitants. In the last decade, however, some studies have focused on the documentation of small game and plant remains recovered from Paleoamerican sites (Walker and Driskell, 2007; Kornfeld and Larson, 2008).
All studies of North American Paleoamerican human remains are based on isolated skeletons, nearly all of which are incomplete and fragmentary. In this regard, much of the record pertains to either single individuals (e.g., Dixon et al., 1997; Green et al., 1998) or skeletal samples pooled from a large area (e.g., Powell and Steele, 1994). Owing to the small sample sizes and the heterogeneity of ecological conditions encompassed within pooled samples, conclusions regarding characterization of health, living conditions, and diet are largely preliminary. The Lagoa Santa human skeletal assemblage provides an unprecedented opportunity to undertake a population-based study of oral health of early Holocene inhabitants from a spatially restricted region. Previous studies of oral health at Lagoa Santa found high prevalence of dental caries at Sumidouro (9.0%, 8/88; Neves
Additional Supporting Information may be found in the online version of this article. Grant sponsor: CNPq-Brazil; Grant numbers: 200034/2007-3, 501318/2012-7; Grant sponsor: The Ohio State University. *Correspondence to: Pedro Da-Gloria, Departamento de Gen etica e Biologia Evolutiva, Rua do Mat~ ao, 277 sala 218, Cidade Universit aria, S~ ao Paulo—SP, Brasil. E-mail: [email protected] Received 5 June 2013; accepted 18 December 2013
1
Throughout the article we prefer the term “Paleoamerican” to describe populations of late Pleistocene and early Holocene in the New World. In contrast to the term “Paleoindian,” “Paleoamerican” is not historically associated with the Clovis-first model.
Ó 2014 WILEY PERIODICALS, INC.
DOI: 10.1002/ajpa.22467 Published online 00 Month 2014 in Wiley Online Library (wileyonlinelibrary.com).
2
P. DA-GLORIA AND C.S. LARSEN
Fig. 1. Location of the Lagoa Santa region in central Brazil (adapted from Neves and Hubbe, 2005).
and Cornero, 1997) and Santana do Riacho (8.3%, 12/ 145; Neves and Kipnis, 2004). However, the low sample sizes at these sites prevented assessments by sex and age-at-death, nor were other dental pathological conditions or tooth wear investigated. Another problem faced by earlier studies of human remains recovered from the Lagoa Santa region is the lack of suitable comparative samples. There are no other skeletal samples from the early Holocene in Brazil, and relatively few in the Americas overall. In addition, comparative studies from different investigators usually involve different methods of data collection, thus limiting comparison (Wesolowski, 2006; Cohen, 2007). In this study, we deal with the comparative issue by drawing upon the Western Hemisphere Project (WHP) database (Steckel and Rose, 2002). This database utilized a standard methodology applied to collections throughout the Americas, allowing us to put Lagoa Santa in the broad context of Native American populations. World-wide studies assessing skeletal remains have documented low prevalence of carious lesions in hunter– gatherers relative to agriculturalists (Turner, 1979; Cohen and Armelagos, 1984; Larsen, 1997; Steckel and Rose, 2002; Cohen and Crane-Kramer, 2007). Almost certainly, the early Holocene inhabitants of Lagoa Santa based their diets exclusively on hunting and gathering because there is no archaeological evidence of either plant domestication or consumption of domesticated plants during this period in central Brazil (Neves and Pilo, 2008). Here, we address the hypothesis that these early Holocene hunter–gatherers from Lagoa Santa possess minimal evidence of poor oral health. The WHP American Journal of Physical Anthropology
database allows for a controlled comparison by age and sex. The database reports only caries and abscessing, but we also provide an assessment of tooth wear and antemortem tooth loss (AMTL) in order to provide a more comprehensive understanding of oral health and dietary implications in the early Holocene.
Biocultural context of Lagoa Santa The archaeological region of Lagoa Santa lies within the state of Minas Gerais, located in east-central Brazil some 400 km from the Atlantic coast (Fig. 1). The region is dominated by karst topography, including hundreds of rockshelters in limestone outcrops (Neves and Pilo, 2008). The tropical climate is characterized by high humidity, marked seasonality, an average precipitation of 1,325 mm and a mean annual temperature of 23 C (Patrus, 1998). The main biome of the region is the Brazilian savannah (cerrado), dotted with patches of deciduous and semideciduous forests (IBGE, 1992). Paleoenvironmental reconstruction based on animal remains suggests that a relatively similar climate prevailed throughout the Holocene (Kipnis, 2002; Perez, 2009). On the other hand, investigation of human occupation intensity revealed an apparent “archaic gap” associated with a period of general aridity in the middle Holocene (Araujo et al., 2005). Environmental reconstructions based on recent excavations of open-air sites in Lagoa Santa suggest instead that the middle Holocene period was characterized by climatic instability rather than by general aridity. In contrast, the early Holocene period shows evidence of a stable environment conducive to human occupation (Araujo et al., 2013).
3
ORAL HEALTH OF THE LAGOA SANTA PALEOAMERICANS Studies of early Holocene lithic artifacts from the sites of Lapa do Santo and Lapa das Boleiras reveal assemblages dominated by quartz flakes (Araujo and Pugliese, 2009a,b; Bueno, 2010). Projectile points are rarely found in the region, suggesting that large-mammal hunting was infrequently practiced by the inhabitants of Lagoa Santa. Although radiocarbon dates have shown a temporal overlap of human occupation and megafauna in the region, no evidence of megafauna exploitation by humans has been documented in Lagoa Santa (Neves and Pilo, 2003; Hubbe et al., 2013). Araujo et al. (2008) suggested that sediment deposition at Lapa das Boleiras is predominantly anthropogenic. That is, ash of burned plant material and general organic matter are responsible for the high rates of sediment deposition during the early Holocene period. The same conclusion appears to apply to Lapa do Santo, which shows more than 4 m of deposited sediment. The earliest dates for stable human occupation at Lagoa Santa occurred around 10,000 yBP and are documented by the open-air sites of Coqueirinho and Sumidouro and the rockshelters of Lapa das Boleiras, Lapa do Santo, and Lapa Grande de Taquarac¸u (Araujo et al., 2012). The faunal assemblages recovered from Lapa do Santo and Lapa das Boleiras contain a diverse array of animals, including deer (Mazama sp.), armadillo (Dasypus novemcinctus), paca (Cuniculus paca), peccary (Tayassu sp.), cavy (Cavia sp.), Brazilian rabbit (Silvylagos brasiliensis), reptiles, and birds (Perez, 2009; Kipnis et al., 2010). Application of dietary breadth models to faunal assemblages from sites located in Minas Gerais reflects consumption of animals with low caloric return (i.e., lower-ranked foods). The botanical record from Lagoa Santa sites reveals the presence of wild fruits and seeds, including araticum (Annona classiflora), jatob a (Hymenaea sp.), palm nut (Syagrus flexuosa), pequi (Caryocar brasiliense), and other plant species (Neves and Pilo, 2008; Nakamura et al., 2010). The earliest record of plant domestication in Minas Gerais (maize) dates to 2,500 yBP (Freitas et al., 2003). Thus, it appears that the diets of early inhabitants of the region were largely based on smaller animals and plants (Kipnis, 2002). This finding is consistent with a lifeway based on local, intensive exploitation of animals and plants. Hermenegildo (2009) undertook an isotopic analysis using early Holocene human remains recovered from the sites of Lapa do Santo and Lapa das Boleiras. He reported d15N values ranging from 7.0 to 8.5&. These values reflect the consumption of intermediate to exclusively herbivorous animals such as deer (Mazama sp.; average of 5.1&) and carnivorous/scavengers such as armadillo (Dasypus novecintus; average of 10.0&). These results suggest a generalized diet based on plants and primary consumers such as deer and cavies. Regarding d13C values, Hermenegildo (2009) calculated the indirect C4 plant consumption as 10–15% of their diet. In short, a generalized diet based on small animals and plants was the mostly likely subsistence strategy in the early Holocene of Lagoa Santa.
MATERIALS Human skeletons have been found in the rockshelters of the Lagoa Santa region since the 19th century. The sample employed in this study is composed of a minimum number of 113 individuals and includes 949 permanent teeth and 1,925 alveoli (Table 1). This sample
TABLE 1. Number of individuals, alveolus and teeth used in this study for each site/collection from the early Holocene period in Lagoa Santa Site/collection Lapa das Boleiras Cerca Grandea Harold Walterb Lapa da Amoreira Lapa de Carrancas Lapa do Santo Lapa Mortu aria Santana do Riacho Total
Individuals
Alveolus
Teeth
3 13 53c 10 10 10 1 13 113
17 198 1,079 156 123 206 9 137 1,925
24 122 403 19 63 227 0 91 949
a
This area includes the rockshelters 2, 5, 6, and 7. This collection includes the following archaeological sites: Abrigo de Limeira, Abrigo de M~ ae Rosa, Abrigo de Samambaia, Abrigo do Eucalipto, Abrigo do Galinheiro, Abrigo do Sumidouro, and Lagoa Funda. c Estimated number based on 105 fragments of mandibles and maxillae. b
was excavated during the 20th century and in recent projects between 2001 and 2009 (see Table 3.1 in DaGloria, 2012). The remains are currently housed in three Brazilian institutions: Laboratorio de Estudos Evolutivos Humanos at the Universidade de S~ ao Paulo, Museu de Historia Natural e Jardim Bot^ anico at the Universidade Federal de Belo Horizonte, and Museu Nacional at the Universidade Federal do Rio de Janeiro. The remains come from 17 archaeological sites located within an area of 20 km2. The only exception is Santana do Riacho, located some 60 km northeast of the other sites. The Lagoa Santa assemblage is derived from an area that is geologically and ecologically homogeneous, especially in comparison with other Paleoamerican samples that have been pooled from an array of localities (e.g., Powell and Steele, 1994). About half of the remains from the Lagoa Santa assemblage are commingled. The estimated number of individuals is based on archaeological reports and direct observation of the collection by DaGloria (see details in Da-Gloria, 2012). Direct radiocarbon dates from human bone indicate that the Lagoa Santa remains span a time frame from about 10,000 yBP–7,000 yBP (Araujo et al., 2012). Santana do Riacho (Neves et al., 2003) and Lapa do Santo (Strauss, 2010) are the only sites in which relative dating was used to clarify the chronology of the human remains (see details in Da-Gloria, 2012).
METHODS The oral health of Lagoa Santa is inferred through four dental markers: abscesses, dental caries, tooth wear, and antemortem tooth loss (AMTL). Dental caries is a process characterized by focal demineralization caused by acids produced by bacterial fermentation of dietary carbohydrates (Larsen, 1997). Carious lesions were scored when a stained and irregular-walled cavity was observed through visual inspection (Steckel et al., 2006). Carious lesion position and pulp exposure were recorded. Position of lesions was accessed through five categories: Gross, occlusal, interproximal, root/cement enamel junction (CEJ), and buccal/labial/lingual. Gross carious lesions are large cavities with no identifiable point of initiation. In this analysis, the percentage of each category of carious lesion position was calculated in American Journal of Physical Anthropology
4
P. DA-GLORIA AND C.S. LARSEN
relation to the total number of carious lesions. Proportion of teeth with pulp exposure by caries or by tooth wear was calculated to obtain the caries correction factor for comparison of caries prevalence by sex (Lukacs, 1995). Abscess prevalence was calculated using minimum and maximum values. The minimum prevalence refers to clear drainage passages in the alveolar bone from the tooth root(s) to the external surface of either maxilla or mandible (Steckel et al., 2006). The maximum value includes cases in which taphonomic processes obliterate the accurate visualization of an eventual margin of the sinus, even though indisputable apical granulomas or cysts were observed (Dias and Tayles, 1997; Ogden, 2008). The scoring methodology for dental caries and abscessing was chosen to be comparable to the Western Hemisphere Project (WHP) database (Steckel et al., 2002). Tooth wear was scored from one to eight using the method of Smith (1984). AMTL was scored in cases in which the alveolus was resorbed to the extent of not being confused with eventual periodontal or periapical diseases. Tooth wear and AMTL were only analyzed by sex and age within the Lagoa Santa sample, since they were not recorded in the WHP database. Sex estimation of adult Lagoa Santa skeletons was based on five morphological traits of the cranium and seven of the pelvis (Steckel et al., 2006). Subadult age estimation relied on dental development (Ubelaker, 1989) and the extent of epiphyseal closure (White and Folkens, 2005). Adult age estimation was based on cranial suture closure (Steckel et al., 2006), pubic symphysis morphology (Brooks and Suchey, 1990), and auricular surface morphology (Lovejoy et al., 1985; Buckberry and Chamberlain, 2002). In addition, a tooth-wear aging method adapted from Miles (1963) was applied in this study. Thirty subadult individuals (under 20-years old) from Lagoa Santa were used as the baseline for this adapted method. For this estimation, tooth wear was scored using the method of Scott (1979), largely because it provides more grades of wear (from 4 to 40) than Smith (1984). Best fit regression curves were used to infer rates of wear among the subadults, which were extrapolated to estimate adult age (see Da-Gloria, 2012). Because the majority of the Lagoa Santa remains are fragmentary, different combinations of these methods were used to sex and age the remains. Age-at-death was divided into four categories: under 20, 20–35, 35–50, and 501 years old. The first two categories represented the majority of the Lagoa Santa sample. The WHP database was used to compare the prevalence of dental caries and abscesses at Lagoa Santa to other prehistoric samples across the Americas. The database encompasses the remains of 6,565 prehistoric individuals from 35 samples, with 3,250 of these individuals coming from North America (49.50%), 1,173 from Mesoamerica (17.87%), and 2,142 from South America (32.63%). Only skeletal collections that antedate 450 BP were included in the prehistoric samples2. In addition, 168 individuals from Windover Pond, Florida, were added to the database (dated to 8120–6990 BP; Dickel, 2002). This assemblage was analyzed with the same methodology employed in the samples included within the WHP database (Wentz, 2006).
2 Dating ranges of some skeletal collections include both prehistoric and historic periods. In these cases, average age was used to compose the prehistoric sample.
American Journal of Physical Anthropology
The WHP database is divided into two subsistence strategies: agriculturalists and hunter–gatherers. The first refers to populations with any plant domesticate (maize, potatoes, beans, squash, among others), while the second refers to populations with an absence of domesticates. Using only two categories (hunter–gatherers and agriculturalists) has the advantage of simplifying the analyses and increasing the sample size within each category, even though each category encompasses considerable dietary diversity (Smith, 2001). Calculating central tendencies and variation within subsistence strategies are ways to deal with these broad analytical categories. The WHP database contains 35,754 teeth and 32,346 alveoli from agriculturalists, and 26,932 teeth and 19,522 alveoli from hunter-gatherers. Sex estimations in the WHP database were originally divided into “certain” and “uncertain.” Here, we considered both types as positive estimations. Two methods of determining prevalence are used in this investigation: tooth/alveolus and individual counts (Lukacs, 1989). In the first method, the prevalence is computed dividing total of teeth/alveoli affected by the total sample of teeth/alveoli available for study. To avoid preservation bias, samples of the WHP database encompassing fewer than 50 teeth/alveoli were excluded from the study. In the second method, only individuals with at least 10 observable teeth/alveoli were considered. Samples within the WHP database with less than five individuals were excluded from the analysis. To maximize the sample sizes, comparisons between the samples found within the WHP database and the assemblage from Lagoa Santa were performed for sex and age categories based on the tooth/alveolus count method. Two analytical approaches are used in this study. The first is a direct statistical comparison of the prevalence of dental lesions in Lagoa Santa assemblage to the corresponding prevalence by subsistence strategy among the samples included in the WHP database. In the cases of nominal-scale data, chi-square (N-1 adjusted) is used to test heterogeneity in 2 3 2 tables as long as all the expected values are above one (Campbell, 2007). Multiple comparisons tests for proportions were applied to statistical tests with two degrees of freedom (3 3 3 tables of contingency). This test is not applied if one of the proportions is zero (Zar, 2009: p 555–557). Fisher Exact test was applied only when expected frequency is below five in more than 20% of the cases. T tests were applied to continuous data (ratio-scale) when two independent samples were compared. Levene’s test was applied for testing equality of variances, and binomial tests were used to test for equality of occurrence of two categories (McKillup, 2005). Alpha was set at 0.05, and the statistical tests were run in SAS 9.2 and SPSS 18.0. The second approach is a position analysis of Lagoa Santa within the spectrum of variation for the two categories of samples included within the WHP database. Dental pathology prevalence among the 36 WHP population samples is plotted in a box plot graph, separating them by subsistence strategy into agriculturalists and hunter-gatherers. Box plots use Tukey’s hinge percentiles, showing the 25th, 50th, and 75th percentiles in the horizontal bars. The maximum length of the whiskers was 1.5 times the interquartile range (IQR). Values between 1.5 and 3.0 IQRs above or below the median were depicted as circles, while values above or below 3.0 IQRs were depicted as asterisks. This approach takes into account the variability within each subsistence
5
27 17 62.96a 1,048 331 31.58 1,520 941 61.91 917 72 7.85a,b 26,932 965 3.58 35,754 4,084 11.42 405 36 8.89a 6,039 266 4.40 13,941 1,565 11.23 329 20 6.08a 2,713 84 3.10 Sample Affected %
Ag 5 agriculturalists; Hg 5 hunter-gatherers; LS 5 Lagoa Santa. Data are divided by age, sex and unit of analysis. a Statistically different from hunter–gatherers (P < 0.05). b Statistically different from agriculturalists (P < 0.05).
7,766 350 4.51 111 24 21.62a 8,777 412 4.69 15,142 2,081 13.74 287 15 5.232 8,567 334 3.90
LS Hg Ag LS Hg Ag LS Hg Ag
13,699 1,660 12.12
LS Hg
Individual count
Ag LS Hg Ag
Hg
LS
Ag
Total 20–35
Oral Health of the Paleoamericans of Lagoa Santa, Central Brazil Pedro Da-Gloria,1* and Clark Spencer Larsen2 1 Laboratorio de Estudos Evolutivos Humanos, Departamento de Genetica e Biologia Evolutiva, Universidade de S~ ao Paulo, CP11461, S~ ao Paulo, Brazil 2 Department of Anthropology, The Ohio State University, Columbus, OH 43210
KEY WORDS America
dental caries; abscess; western hemisphere project; first Americans; South
ABSTRACT The peopling, origins, and early prehistory of the Americas are topics of intense debate. However, few studies have used human remains to document and interpret patterns of health and lifestyle of Paleoamericans. This study provides the first investigation to characterize oral health in a series of early Holocene skeletal remains from Lagoa Santa, Brazil, a locality containing the remains of some of the earliest inhabitants of South America (10,000–7,000 BP). The sample is composed of 949 teeth and 1925 alveoli from an estimated 113 individuals excavated from 17 archaeological sites located in the State of Minas Gerais. We compare dental caries and abscess prevalence at Lagoa Santa to a large sample of human skeletons from the Western Hemisphere Project (WHP) database using both individual and tooth/alveolus count
methods. In addition, antemortem tooth loss and tooth wear were analyzed in Lagoa Santa by sex and age. The results show that Lagoa Santa dental caries and abscess prevalence are significantly higher than observed among other hunter–gatherers included in the WHP database, except when abscess prevalence is considered by individual count. Adult females have less tooth wear coupled with higher prevalence of dental caries and antemortem tooth loss than adult males. These results point to an unexpected record of poor oral health at Lagoa Santa, especially among females. A diet based on a highly cariogenic combination of wild tubers and fruits is suggested as an explanation for the elevated rate, characterizing an early adaptation to a tropical environment in South America. Am J Phys Anthropol 000:000–000, 2014. VC 2014 Wiley Periodicals, Inc.
The place of origin (Kidd et al., 2011; Stanford and Bradley, 2012), route(s) of migration (Mandryk et al., 2001; Jantz and Owsley, 2005), number and size of migration(s) (Schurr, 2004; Neves and Hubbe, 2005; Fagundes et al., 2008), and dates of arrival (Achilli et al., 2008; Waters et al., 2011) of the first inhabitants of the Americas are subjects of intense debate. The Lagoa Santa region of central Brazil has contributed to discussions regarding migration routes and biological origins due to the remarkable preservation of human skeletal and dental remains dated to the Pleistocene/ Holocene boundary (Neves et al., 2003, 2007). To date, research on the human remains recovered from Lagoa Santa has focused on cranial morphology and dating. New radiocarbon dates have recently confirmed that the Lagoa Santa rockshelters were occupied during the early Holocene (Araujo et al., 2012). Until now, few studies have used assemblages of human remains from Lagoa Santa to infer subsistence and living conditions of those early inhabitants (for exceptions see Neves and Cornero, 1997; Neves and Kipnis, 2004). Traditionally, studies of Paleoamerican1 subsistence have focused on Clovis kill sites in North America, emphasizing the importance of big game hunting to the first inhabitants. In the last decade, however, some studies have focused on the documentation of small game and plant remains recovered from Paleoamerican sites (Walker and Driskell, 2007; Kornfeld and Larson, 2008).
All studies of North American Paleoamerican human remains are based on isolated skeletons, nearly all of which are incomplete and fragmentary. In this regard, much of the record pertains to either single individuals (e.g., Dixon et al., 1997; Green et al., 1998) or skeletal samples pooled from a large area (e.g., Powell and Steele, 1994). Owing to the small sample sizes and the heterogeneity of ecological conditions encompassed within pooled samples, conclusions regarding characterization of health, living conditions, and diet are largely preliminary. The Lagoa Santa human skeletal assemblage provides an unprecedented opportunity to undertake a population-based study of oral health of early Holocene inhabitants from a spatially restricted region. Previous studies of oral health at Lagoa Santa found high prevalence of dental caries at Sumidouro (9.0%, 8/88; Neves
Additional Supporting Information may be found in the online version of this article. Grant sponsor: CNPq-Brazil; Grant numbers: 200034/2007-3, 501318/2012-7; Grant sponsor: The Ohio State University. *Correspondence to: Pedro Da-Gloria, Departamento de Gen etica e Biologia Evolutiva, Rua do Mat~ ao, 277 sala 218, Cidade Universit aria, S~ ao Paulo—SP, Brasil. E-mail: [email protected] Received 5 June 2013; accepted 18 December 2013
1
Throughout the article we prefer the term “Paleoamerican” to describe populations of late Pleistocene and early Holocene in the New World. In contrast to the term “Paleoindian,” “Paleoamerican” is not historically associated with the Clovis-first model.
Ó 2014 WILEY PERIODICALS, INC.
DOI: 10.1002/ajpa.22467 Published online 00 Month 2014 in Wiley Online Library (wileyonlinelibrary.com).
2
P. DA-GLORIA AND C.S. LARSEN
Fig. 1. Location of the Lagoa Santa region in central Brazil (adapted from Neves and Hubbe, 2005).
and Cornero, 1997) and Santana do Riacho (8.3%, 12/ 145; Neves and Kipnis, 2004). However, the low sample sizes at these sites prevented assessments by sex and age-at-death, nor were other dental pathological conditions or tooth wear investigated. Another problem faced by earlier studies of human remains recovered from the Lagoa Santa region is the lack of suitable comparative samples. There are no other skeletal samples from the early Holocene in Brazil, and relatively few in the Americas overall. In addition, comparative studies from different investigators usually involve different methods of data collection, thus limiting comparison (Wesolowski, 2006; Cohen, 2007). In this study, we deal with the comparative issue by drawing upon the Western Hemisphere Project (WHP) database (Steckel and Rose, 2002). This database utilized a standard methodology applied to collections throughout the Americas, allowing us to put Lagoa Santa in the broad context of Native American populations. World-wide studies assessing skeletal remains have documented low prevalence of carious lesions in hunter– gatherers relative to agriculturalists (Turner, 1979; Cohen and Armelagos, 1984; Larsen, 1997; Steckel and Rose, 2002; Cohen and Crane-Kramer, 2007). Almost certainly, the early Holocene inhabitants of Lagoa Santa based their diets exclusively on hunting and gathering because there is no archaeological evidence of either plant domestication or consumption of domesticated plants during this period in central Brazil (Neves and Pilo, 2008). Here, we address the hypothesis that these early Holocene hunter–gatherers from Lagoa Santa possess minimal evidence of poor oral health. The WHP American Journal of Physical Anthropology
database allows for a controlled comparison by age and sex. The database reports only caries and abscessing, but we also provide an assessment of tooth wear and antemortem tooth loss (AMTL) in order to provide a more comprehensive understanding of oral health and dietary implications in the early Holocene.
Biocultural context of Lagoa Santa The archaeological region of Lagoa Santa lies within the state of Minas Gerais, located in east-central Brazil some 400 km from the Atlantic coast (Fig. 1). The region is dominated by karst topography, including hundreds of rockshelters in limestone outcrops (Neves and Pilo, 2008). The tropical climate is characterized by high humidity, marked seasonality, an average precipitation of 1,325 mm and a mean annual temperature of 23 C (Patrus, 1998). The main biome of the region is the Brazilian savannah (cerrado), dotted with patches of deciduous and semideciduous forests (IBGE, 1992). Paleoenvironmental reconstruction based on animal remains suggests that a relatively similar climate prevailed throughout the Holocene (Kipnis, 2002; Perez, 2009). On the other hand, investigation of human occupation intensity revealed an apparent “archaic gap” associated with a period of general aridity in the middle Holocene (Araujo et al., 2005). Environmental reconstructions based on recent excavations of open-air sites in Lagoa Santa suggest instead that the middle Holocene period was characterized by climatic instability rather than by general aridity. In contrast, the early Holocene period shows evidence of a stable environment conducive to human occupation (Araujo et al., 2013).
3
ORAL HEALTH OF THE LAGOA SANTA PALEOAMERICANS Studies of early Holocene lithic artifacts from the sites of Lapa do Santo and Lapa das Boleiras reveal assemblages dominated by quartz flakes (Araujo and Pugliese, 2009a,b; Bueno, 2010). Projectile points are rarely found in the region, suggesting that large-mammal hunting was infrequently practiced by the inhabitants of Lagoa Santa. Although radiocarbon dates have shown a temporal overlap of human occupation and megafauna in the region, no evidence of megafauna exploitation by humans has been documented in Lagoa Santa (Neves and Pilo, 2003; Hubbe et al., 2013). Araujo et al. (2008) suggested that sediment deposition at Lapa das Boleiras is predominantly anthropogenic. That is, ash of burned plant material and general organic matter are responsible for the high rates of sediment deposition during the early Holocene period. The same conclusion appears to apply to Lapa do Santo, which shows more than 4 m of deposited sediment. The earliest dates for stable human occupation at Lagoa Santa occurred around 10,000 yBP and are documented by the open-air sites of Coqueirinho and Sumidouro and the rockshelters of Lapa das Boleiras, Lapa do Santo, and Lapa Grande de Taquarac¸u (Araujo et al., 2012). The faunal assemblages recovered from Lapa do Santo and Lapa das Boleiras contain a diverse array of animals, including deer (Mazama sp.), armadillo (Dasypus novemcinctus), paca (Cuniculus paca), peccary (Tayassu sp.), cavy (Cavia sp.), Brazilian rabbit (Silvylagos brasiliensis), reptiles, and birds (Perez, 2009; Kipnis et al., 2010). Application of dietary breadth models to faunal assemblages from sites located in Minas Gerais reflects consumption of animals with low caloric return (i.e., lower-ranked foods). The botanical record from Lagoa Santa sites reveals the presence of wild fruits and seeds, including araticum (Annona classiflora), jatob a (Hymenaea sp.), palm nut (Syagrus flexuosa), pequi (Caryocar brasiliense), and other plant species (Neves and Pilo, 2008; Nakamura et al., 2010). The earliest record of plant domestication in Minas Gerais (maize) dates to 2,500 yBP (Freitas et al., 2003). Thus, it appears that the diets of early inhabitants of the region were largely based on smaller animals and plants (Kipnis, 2002). This finding is consistent with a lifeway based on local, intensive exploitation of animals and plants. Hermenegildo (2009) undertook an isotopic analysis using early Holocene human remains recovered from the sites of Lapa do Santo and Lapa das Boleiras. He reported d15N values ranging from 7.0 to 8.5&. These values reflect the consumption of intermediate to exclusively herbivorous animals such as deer (Mazama sp.; average of 5.1&) and carnivorous/scavengers such as armadillo (Dasypus novecintus; average of 10.0&). These results suggest a generalized diet based on plants and primary consumers such as deer and cavies. Regarding d13C values, Hermenegildo (2009) calculated the indirect C4 plant consumption as 10–15% of their diet. In short, a generalized diet based on small animals and plants was the mostly likely subsistence strategy in the early Holocene of Lagoa Santa.
MATERIALS Human skeletons have been found in the rockshelters of the Lagoa Santa region since the 19th century. The sample employed in this study is composed of a minimum number of 113 individuals and includes 949 permanent teeth and 1,925 alveoli (Table 1). This sample
TABLE 1. Number of individuals, alveolus and teeth used in this study for each site/collection from the early Holocene period in Lagoa Santa Site/collection Lapa das Boleiras Cerca Grandea Harold Walterb Lapa da Amoreira Lapa de Carrancas Lapa do Santo Lapa Mortu aria Santana do Riacho Total
Individuals
Alveolus
Teeth
3 13 53c 10 10 10 1 13 113
17 198 1,079 156 123 206 9 137 1,925
24 122 403 19 63 227 0 91 949
a
This area includes the rockshelters 2, 5, 6, and 7. This collection includes the following archaeological sites: Abrigo de Limeira, Abrigo de M~ ae Rosa, Abrigo de Samambaia, Abrigo do Eucalipto, Abrigo do Galinheiro, Abrigo do Sumidouro, and Lagoa Funda. c Estimated number based on 105 fragments of mandibles and maxillae. b
was excavated during the 20th century and in recent projects between 2001 and 2009 (see Table 3.1 in DaGloria, 2012). The remains are currently housed in three Brazilian institutions: Laboratorio de Estudos Evolutivos Humanos at the Universidade de S~ ao Paulo, Museu de Historia Natural e Jardim Bot^ anico at the Universidade Federal de Belo Horizonte, and Museu Nacional at the Universidade Federal do Rio de Janeiro. The remains come from 17 archaeological sites located within an area of 20 km2. The only exception is Santana do Riacho, located some 60 km northeast of the other sites. The Lagoa Santa assemblage is derived from an area that is geologically and ecologically homogeneous, especially in comparison with other Paleoamerican samples that have been pooled from an array of localities (e.g., Powell and Steele, 1994). About half of the remains from the Lagoa Santa assemblage are commingled. The estimated number of individuals is based on archaeological reports and direct observation of the collection by DaGloria (see details in Da-Gloria, 2012). Direct radiocarbon dates from human bone indicate that the Lagoa Santa remains span a time frame from about 10,000 yBP–7,000 yBP (Araujo et al., 2012). Santana do Riacho (Neves et al., 2003) and Lapa do Santo (Strauss, 2010) are the only sites in which relative dating was used to clarify the chronology of the human remains (see details in Da-Gloria, 2012).
METHODS The oral health of Lagoa Santa is inferred through four dental markers: abscesses, dental caries, tooth wear, and antemortem tooth loss (AMTL). Dental caries is a process characterized by focal demineralization caused by acids produced by bacterial fermentation of dietary carbohydrates (Larsen, 1997). Carious lesions were scored when a stained and irregular-walled cavity was observed through visual inspection (Steckel et al., 2006). Carious lesion position and pulp exposure were recorded. Position of lesions was accessed through five categories: Gross, occlusal, interproximal, root/cement enamel junction (CEJ), and buccal/labial/lingual. Gross carious lesions are large cavities with no identifiable point of initiation. In this analysis, the percentage of each category of carious lesion position was calculated in American Journal of Physical Anthropology
4
P. DA-GLORIA AND C.S. LARSEN
relation to the total number of carious lesions. Proportion of teeth with pulp exposure by caries or by tooth wear was calculated to obtain the caries correction factor for comparison of caries prevalence by sex (Lukacs, 1995). Abscess prevalence was calculated using minimum and maximum values. The minimum prevalence refers to clear drainage passages in the alveolar bone from the tooth root(s) to the external surface of either maxilla or mandible (Steckel et al., 2006). The maximum value includes cases in which taphonomic processes obliterate the accurate visualization of an eventual margin of the sinus, even though indisputable apical granulomas or cysts were observed (Dias and Tayles, 1997; Ogden, 2008). The scoring methodology for dental caries and abscessing was chosen to be comparable to the Western Hemisphere Project (WHP) database (Steckel et al., 2002). Tooth wear was scored from one to eight using the method of Smith (1984). AMTL was scored in cases in which the alveolus was resorbed to the extent of not being confused with eventual periodontal or periapical diseases. Tooth wear and AMTL were only analyzed by sex and age within the Lagoa Santa sample, since they were not recorded in the WHP database. Sex estimation of adult Lagoa Santa skeletons was based on five morphological traits of the cranium and seven of the pelvis (Steckel et al., 2006). Subadult age estimation relied on dental development (Ubelaker, 1989) and the extent of epiphyseal closure (White and Folkens, 2005). Adult age estimation was based on cranial suture closure (Steckel et al., 2006), pubic symphysis morphology (Brooks and Suchey, 1990), and auricular surface morphology (Lovejoy et al., 1985; Buckberry and Chamberlain, 2002). In addition, a tooth-wear aging method adapted from Miles (1963) was applied in this study. Thirty subadult individuals (under 20-years old) from Lagoa Santa were used as the baseline for this adapted method. For this estimation, tooth wear was scored using the method of Scott (1979), largely because it provides more grades of wear (from 4 to 40) than Smith (1984). Best fit regression curves were used to infer rates of wear among the subadults, which were extrapolated to estimate adult age (see Da-Gloria, 2012). Because the majority of the Lagoa Santa remains are fragmentary, different combinations of these methods were used to sex and age the remains. Age-at-death was divided into four categories: under 20, 20–35, 35–50, and 501 years old. The first two categories represented the majority of the Lagoa Santa sample. The WHP database was used to compare the prevalence of dental caries and abscesses at Lagoa Santa to other prehistoric samples across the Americas. The database encompasses the remains of 6,565 prehistoric individuals from 35 samples, with 3,250 of these individuals coming from North America (49.50%), 1,173 from Mesoamerica (17.87%), and 2,142 from South America (32.63%). Only skeletal collections that antedate 450 BP were included in the prehistoric samples2. In addition, 168 individuals from Windover Pond, Florida, were added to the database (dated to 8120–6990 BP; Dickel, 2002). This assemblage was analyzed with the same methodology employed in the samples included within the WHP database (Wentz, 2006).
2 Dating ranges of some skeletal collections include both prehistoric and historic periods. In these cases, average age was used to compose the prehistoric sample.
American Journal of Physical Anthropology
The WHP database is divided into two subsistence strategies: agriculturalists and hunter–gatherers. The first refers to populations with any plant domesticate (maize, potatoes, beans, squash, among others), while the second refers to populations with an absence of domesticates. Using only two categories (hunter–gatherers and agriculturalists) has the advantage of simplifying the analyses and increasing the sample size within each category, even though each category encompasses considerable dietary diversity (Smith, 2001). Calculating central tendencies and variation within subsistence strategies are ways to deal with these broad analytical categories. The WHP database contains 35,754 teeth and 32,346 alveoli from agriculturalists, and 26,932 teeth and 19,522 alveoli from hunter-gatherers. Sex estimations in the WHP database were originally divided into “certain” and “uncertain.” Here, we considered both types as positive estimations. Two methods of determining prevalence are used in this investigation: tooth/alveolus and individual counts (Lukacs, 1989). In the first method, the prevalence is computed dividing total of teeth/alveoli affected by the total sample of teeth/alveoli available for study. To avoid preservation bias, samples of the WHP database encompassing fewer than 50 teeth/alveoli were excluded from the study. In the second method, only individuals with at least 10 observable teeth/alveoli were considered. Samples within the WHP database with less than five individuals were excluded from the analysis. To maximize the sample sizes, comparisons between the samples found within the WHP database and the assemblage from Lagoa Santa were performed for sex and age categories based on the tooth/alveolus count method. Two analytical approaches are used in this study. The first is a direct statistical comparison of the prevalence of dental lesions in Lagoa Santa assemblage to the corresponding prevalence by subsistence strategy among the samples included in the WHP database. In the cases of nominal-scale data, chi-square (N-1 adjusted) is used to test heterogeneity in 2 3 2 tables as long as all the expected values are above one (Campbell, 2007). Multiple comparisons tests for proportions were applied to statistical tests with two degrees of freedom (3 3 3 tables of contingency). This test is not applied if one of the proportions is zero (Zar, 2009: p 555–557). Fisher Exact test was applied only when expected frequency is below five in more than 20% of the cases. T tests were applied to continuous data (ratio-scale) when two independent samples were compared. Levene’s test was applied for testing equality of variances, and binomial tests were used to test for equality of occurrence of two categories (McKillup, 2005). Alpha was set at 0.05, and the statistical tests were run in SAS 9.2 and SPSS 18.0. The second approach is a position analysis of Lagoa Santa within the spectrum of variation for the two categories of samples included within the WHP database. Dental pathology prevalence among the 36 WHP population samples is plotted in a box plot graph, separating them by subsistence strategy into agriculturalists and hunter-gatherers. Box plots use Tukey’s hinge percentiles, showing the 25th, 50th, and 75th percentiles in the horizontal bars. The maximum length of the whiskers was 1.5 times the interquartile range (IQR). Values between 1.5 and 3.0 IQRs above or below the median were depicted as circles, while values above or below 3.0 IQRs were depicted as asterisks. This approach takes into account the variability within each subsistence
5
27 17 62.96a 1,048 331 31.58 1,520 941 61.91 917 72 7.85a,b 26,932 965 3.58 35,754 4,084 11.42 405 36 8.89a 6,039 266 4.40 13,941 1,565 11.23 329 20 6.08a 2,713 84 3.10 Sample Affected %
Ag 5 agriculturalists; Hg 5 hunter-gatherers; LS 5 Lagoa Santa. Data are divided by age, sex and unit of analysis. a Statistically different from hunter–gatherers (P < 0.05). b Statistically different from agriculturalists (P < 0.05).
7,766 350 4.51 111 24 21.62a 8,777 412 4.69 15,142 2,081 13.74 287 15 5.232 8,567 334 3.90
LS Hg Ag LS Hg Ag LS Hg Ag
13,699 1,660 12.12
LS Hg
Individual count
Ag LS Hg Ag
Hg
LS
Ag
Total 20–35
