https://orcid.org/0000-0002-4595-4487
Figures
Abstract
The degree of homogeneity in the diet of archaeological populations associated with Brazilian shellmounds (sambaquis) is an ongoing debate. Isotopic studies have the potential to document both intra- and inter-group dietary variability, especially when paired with quantitative tools, such as Bayesian Mixing Models. In this study, we investigate intra-site variability on diet at two shellmounds in southeastern Brazil: Piaçaguera (7151–5668 yBP), characterized by two distinct burial clusters, and Moraes (6791–5590 yBP). We analyzed δ13C and δ15N on human and fauna bone collagen (n = 43) and on human dentine collagen (n = 5 teeth). Our data were complemented by previously published isotopic data from these sites (n = 88) and from other Brazilian shellmounds (n = 51). To quantify the consumption of different food sources, we adapted the Bayesian MixSIAR model to integrate both bone and dentine values without compromising sample independence. Our results indicate that marine fish contributed approximately 30–50% to the diet of individuals from Piaçaguera, and terrestrial animals accounted for 16–45% of the diet of individuals from Moraes. We observed a slight difference in the consumption of marine and terrestrial sources between the burial groups in Piaçaguera and a higher intake of freshwater fish among subadults in Moraes. New radiocarbon dating in Piaçaguera indicates that the two burial groups did not belong to distinct chronological periods. These findings highlight the importance of a constant dialog between archaeological and isotopic data. The adapted MixSIAR model validated and increased the robustness of our results, proving to be effective in delineating dietary patterns in past groups from distinct archaeological contexts.
Citation: Di Giusto M, Fumis A, Bastos M, Fossaluza V, Stern R, Mendes Cardoso J, et al. (2025) Dietary variability in Middle Holocene South American shellmounds: Insights from isotopic analysis and an adapted Bayesian MixSIAR model. PLoS One 20(12): e0335680. https://doi.org/10.1371/journal.pone.0335680
Editor: Federico Lugli,, University of Modena and Reggio Emilia: Universita degli Studi di Modena e Reggio Emilia, ITALY
Received: June 16, 2025; Accepted: October 14, 2025; Published: December 3, 2025
Copyright: © 2025 Di Giusto et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data supporting this study are available in the Supporting Information and in the GitHub data repository (https://github.com/AFumis/Adapted_MixSIAR_diet_analysis).
Funding: This work was supported by the Coordination of Superior Level Staff Improvement (CAPES, Brazil) (MDG, AF) and the ERC ARCHEIS project (803676) (KJ).
Competing interests: The authors have declared that no competing interests exist.
Introduction
The hunter-gatherer-fisher groups responsible for building shellmounds along the Atlantic coast have long attracted the attention of South American archaeologists. Known as sambaquis in Brazil, these sites date to the Holocene (ca. 8,000–1,500 yBP) and were intentionally constructed by the accumulation of marine mollusk shells mixed with sediments and fish bones [1–3]. These mounds held a significant symbolic value, shaping the identity of a stable and organized society, and providing evidence of human occupation across diverse coastal environments [1,4]. Similar mounds – riverine shellmounds – mainly made of terrestrial mollusk shells are also found inland [5].
Over the past three decades, significant strides have been made in understanding the dietary practices of shellmound-building groups, who exploited both aquatic and terrestrial ecosystems, especially from south-southeastern Brazil. Coastal groups consumed marine and terrestrial resources in varying proportions [6–8], while riverine groups consumed terrestrial and freshwater resources [5,9]. Stable isotopic analysis of carbon (δ13C) and nitrogen (δ15N) have confirmed these dietary patterns [10–15].
Despite these advancements, few studies have attempted to quantify the proportion of different food sources in their diets [11,12,15,16], which could contribute to a more robust discussion on dietary variability on regional/temporal scales. Moreover, intra-site variability on diet remains underexplored, as the study on diet of shellmounds from beyond Brazilian southern coast, especially Santa Catarina. Within this context, the Piaçaguera and Moraes sites, both in southeastern Brazil and recently studied for childhood diet [17], offer compelling case studies for investigating intra-site variability.
At Piaçaguera, stratigraphic evidence points to two distinct burial groups [18], which have never been analyzed separately. Investigating these subgroups may reveal potential differences in lifestyle within the same community, as seen in other shellmounds in southern Brazil [19,20]. At Moraes, there is a dietary difference between subadults up to five years old and adults, with a higher intake of freshwater fish by subadults [17]. This age-related dietary variability highlights the complexity of subsistence strategies in this group.
A detailed analysis of intra-site dietary patterns is essential to understanding the social and cultural dynamics of these precolonial groups. Bayesian Mixing Models (BMMs) are particularly useful for estimating the proportional contributions of different food sources to diet. However, they typically operate with one independent isotopic value per individual (e.g., bulk collagen), since their statistical framework depends on the independence of observations. Integrating multiple dependent values (e.g., sequential dentine samples) is challenging, since they are repeated observations from the same individual and do not provide the same informational gain as adding new individuals. Nonetheless, overcoming this limitation could enable more detailed insights into diet over an individual’s life and comparisons across age groups.
This paper aims to investigate dietary variability at the Piaçaguera and Moraes sites by combining δ13C and δ15N values from dentine and bone collagen of subadults from 5 years of age, adults, and fauna. For this purpose, new data was produced and integrated with previously published data, and a new adapted version of the Bayesian model MixSIAR was developed to integrate both bone and dentine values, allowing for the estimation of food group proportions across the life course.
By considering the archaeological and funerary contexts when performing isotopic analyses and presenting a new version of MixSIAR, this study contributes to building a more complex understanding of the lifestyle of precolonial South American groups and provides a valuable tool for future research.
The studied sites
Piaçaguera and Moraes are located in São Paulo state, southeastern Brazil (Fig 1). Piaçaguera is on the coast of Santos city (UTM: 23J 0380184/735713), at the bottom of an ancient paleobay near mangroves, sandbanks, the Atlantic Forest, and the Atlantic Ocean, environments rich in biodiversity [6]. Excavations conducted in the 1960s recovered a minimum of 90 human remains, dated between 7,151 and 5,668 yBP [21]. Sex and age analyses were conducted by Fischer [18], based on the criteria proposed by Buikstra and Ubelaker [22] and Schaefer et al. [23].
Compiled from open data Brazilian sources: Brazilian Institute of Geography and Statistics (IBGE, 2019, https://www.ibge.gov.br/geociencias/organizacao-do-territorio/malhas-territoriais.html), National Water Agency (ANA, 2019, https://metadados.snirh.gov.br/geonetwork/srv/por/catalog.search#/home), and Department of Highways of the State of São Paulo (DER-SP, https://www.der.sp.gov.br/WebSite/Servicos/ConjuntoDados.aspx?tema=Sistema_Rodoviario_Estadual&conjunto=).
Previous δ13C and δ15N data from adult bone collagen suggest a diet primarily based on both marine and terrestrial protein sources for the Piaçaguera group [24,25], complemented by C3 plants, mollusks, and freshwater fish [12]. Isotopic analysis of subadult bone collagen and dentine collagen suggests that weaning likely ended between two and four years of age, and that pregnant women may have consumed slightly more terrestrial resources [17].
An important feature of Piaçaguera is the presence of two human burial clusters slightly apart from each other both in horizontal space and in stratigraphic sequence: one between 10.87 and 12.13 meters above sea level (here namely Group I), and another between 9.97 and 10.78 meters above sea level (here namely Group II) (Fig 2) [18]. Both clusters include subadults and adults of both sexes.
The Moraes shellmound is located about 25 km inland, in Miracatu city (UTM: 23J 0256908/7313340). Excavations conducted in the 2000s recovered a minimum of 56 individuals, including subadults and adults of both sexes [18], dated between 6,791 and 4,971 yBP [21].
Previous δ13C and δ15N analysis of adult bone collagen indicate a diet based on terrestrial animals for the Moraes group, with no significant difference between sexes [14,24,25], complemented by C3 plants and mollusks [12]. Isotopic analysis of subadult bone collagen and dentine collagen indicates that weaning likely ended between two and five years of age and that children (and/or pregnant women) may have consumed more freshwater fish [17]. Analyses of starch grains, phytoliths, and lipids indicate an important intake of plant-based resources, including yams (Dioscorea sp.) and sweet potatoes (Ipomoea batatas), at higher levels than in coastal mound groups in general [27,28].
The use of δ13C, δ15N and mixing models in archaeology
Bulk collagen δ13C and δ15N analysis is a widely used method for investigating past diets. The δ13C is useful for distinguishing between plant groups with different photosynthetic pathways (C3, C4, and CAM) and between marine and terrestrial diets, while δ15N reflects trophic levels within food webs, increasing by ~3‰ per level, being useful for identifying dietary preferences across the food web [29,30].
In archaeology, δ13C and δ15N for dietary reconstruction often relies on bone and dentine bulk collagen. Bone collagen reflects an average of the protein part of the diet over the last years of life, although this temporal resolution varies depending on the bone element [31] and the individual’s sex [32] and age [33]. In contrast, dentine provides dietary information during specific periods of tooth formation and sequential dentine analysis has emerged as a valuable approach for tracking individual diets across different life stages [17,34–38].
More recently, Stable Isotope Mixing Models (SIMMs) have improved dietary interpretations in past societies [39–41]. Originally developed for ecological research, these models offer a statistical framework to estimate the contribution of different food sources to a consumer’s diet [42]. Among them, BMMs stand out for their capacity to account for the complexities of human subsistence practices, as well as the uncertainties and biases inherent to archaeological samples, particularly those arising from small sample sizes [43].
BMMs operate with independent isotopic inputs, such as bone values, and their application remains limited when applied to dependent data, such as sequential dentine values. Treating dependent values as independent can bias the results and lead to inaccurate estimates, requiring an adaptation to the model that accounts for repeated measures within the same individuals.
One of the most widely used models in archaeology is MixSIAR, an open-source R package [44–49] that, unlike other BMMs, incorporate continuous effects. However, attempts to use this feature in MixSIAR have often failed to produce informative results [39], highlighting its limitations in handling dependent data. Addressing these limitations to incorporate dependent sequential dentine data would allow for the study of diet across an individual’s lifespan, facilitating both inter- and intra-individual comparisons.
Materials and methods
All necessary permits were obtained for the described study, which complied with all relevant regulations. To date, no indigenous groups in Brazil have recognized the shellmound-builders as their direct ancestors or claimed ethnic and/or cultural connection, and the authorization to analyze the human remains in this study was granted by the National Historical and Artistic Heritage Institute in Brazil (Official letter Nº 733/2021/GAB PRESI/PRESI-IPHAN of February 20, 2021, Process nº 01506.000026/2021–62) and the Museum of Archaeology and Ethnology at the University of São Paulo, which houses these remains.
This study used δ13C and δ15N values from new and previously published analysis of human bones (n = 67), human teeth (n = 28, totaling 190 dentine samples), and faunal bones (n = 92), totaling 349 samples.
Newly analyzed materials include human bones from individuals from six years of age to adulthood (n = 19; 5 subadults and 14 adults, including 5 females, 6 males, and 3 adults with indeterminate sex), human teeth with formation periods spanning late childhood (6–12 years) and adolescence (13–18 years) (n = 40 dentine slices from 5 teeth; 1 first molar, 1 second molar, 2 third molars and 1 second premolar), and faunal bones from terrestrial (n = 18) and aquatic species (n = 6) recovered from both sites. Faunal species were identified by one of the authors (JMC) through comparative anatomical analysis, using reference zooarchaeological collections, digital databases, and bibliographic sources [50–53].
Previous analyzed samples include human bones (n = 48; 29 subadults and 19 adults, including 9 females, 6 males, and 4 adults with indeterminate sex), human teeth formed during early childhood (birth-6 years) (n = 150 dentine slices from 23 teeth; 13 first molars, 3 second molars, 2 third molars, and 5 deciduous teeth), and faunal bones from terrestrial (n = 16) and aquatic species (n = 1) recovered from both sites [14,17,24].
To complement the fauna dataset, we also included faunal bone values from other Brazilian shellmounds (n = 51), selecting those species that are also found at Piaçaguera and Moraes (n = 41) or in riverine shellmounds near Moraes (n = 10), based on the premise that the riverine sites were part of a culturally connected network [5,9].
At Piaçaguera, individuals were divided between the two burial clusters (Group I and Group II) to investigate potential differences in diet.
The number of samples used in this study is summarized in Table 1, with further details provided in S1 Table in S1 File for human samples and S2 Table in S1 File for faunal samples.
Laboratory procedures for isotopic analysis
Bones and teeth were weighed, measured, and photographed using a digital camera (Canon EOS Rebel T5, EF-S 18–55 mm lens). Dental morphological characteristics were described in a brief report, occlusal surfaces were molded with high-precision silicone, and teeth were scanned at CIRIMAT, Université de Toulouse, France. Dental calculus was collected following the protocol described by Wesolowski et al. [54].
Collagen extraction was conducted at the Laboratoire Méditerranéen de Préhistoire Europe Afrique (LAMPEA – UMR 7269), Université Aix-Marseille. Bone collagen extraction followed the modified Longin method [55]. Bone chunks (~500 mg) were demineralized in 0.5M HCl at 4°C for four days, treated with 0.125M NaOH, and heated at 75°C in 10-3M HCl (pH 3) for 48h. The solution was filtered using a graduated polypropylene tube with a sintered polyethylene filter (Ezee Filter), frozen at −64°C, and freeze-dried for 48h.
For teeth, a 2.5 mm thickness central slice was cut [56,57] and demineralized following Method 2 from Beaumont et al. [34]. Any secondary dentine and the dentine surrounding the pulp chamber and radicular canal were removed. One side of the central slice (mesial or distal) was sampled in sequential horizontal slices with 1 mm thickness, from crown cusp to root apex, following the anatomical location for age-alignment described by Czermak et al. [57]. Each slice was placed in a 1.5 ml microtube with pH 3 solution, centrifuged, frozen at −64°C, and freeze-dried.
Isotopic measurements used 0.5 mg of bone collagen (yield > 8 mg/g) and 0.3–0.5 mg of dentine collagen. Slices with insufficient weight were combined with adjacent slices. Measurements were conducted at IsoAnalytical Laboratory (England) using a Europa Scientific 20−20 IRMS mass spectrometer. Reference materials included IA-R068 (soy protein, δ13CV-PDB = −25.22‰, δ15NAIR = 0.99‰) and quality control samples included IA-R068 and IA-R038 (L-alanine, δ13CV-PDB = −24.99‰, δ15NAIR = −0.65‰), IA-R069 (tuna protein, δ13CV-PDB = −18.88‰, δ15NAIR = 11.60‰), and a mixture of IAEA-C7 (oxalic acid, δ13CV-PDB = −14.48‰) and IA-R046 (ammonium sulfate, δ15NAIR = 22.04‰) (S3 Table in S1 File).
Based on Szpak et al. [58] equation, for both series, precision was ± 0.07‰ for δ13C and ±0.08‰ for δ15N, accuracy was ± 0.1 for δ13C and δ15N, and analytical uncertainty was ± 0.1‰ for δ13C and δ15N. Collagen quality followed the criteria stablished by DeNiro [59], Ambrose [60], and Guiry & Szpak [61], namely, carbon and nitrogen content values above 13% and 5%, respectively, and a C/N ratio between 2.9 and 3.6. To verify the absence of correlation between δ13C or δ15N values and other variables (carbon content, nitrogen content, C/N ratio, and sample weight), Spearman’s non-parametric correlation test was applied (R Commander package). The non-parametric Mann-Whitney test was used to assess statistically significant differences between group values (sex, age categories, and burial groups).
Input values in MixSIAR
MixSIAR incorporates isotopic values from both consumers and food sources (divided into food source groups). Consumer values are both δ13C and δ15N values measured on human remains from Piaçaguera and Moraes, obtained in this and previous studies (S4 and S5 Tables in S1 File). Adult bone values were used to estimate the overall diet, while subadult bones and dentine slices from 5 years of age onwards were used to estimate diet during childhood. Both subadult bone and dentine samples under 5 years were excluded to avoid potential breastfeeding-related bias in δ15N, as weaning lasted until 4 years of age in Piaçaguera and 5 years in Moraes [17].
Food source values are both δ13C and δ15N values from C3 plants and from faunal remains from Piaçaguera and Moraes, obtained in this and previous studies (S6 Table in S1 File), as well as values measured from other Brazilian shellmounds whose species were identified at Piaçaguera and Moraes or at riverine shellmounds near Moraes (S7 Table in S1 File). The faunal values include marine fish (n = 31), freshwater fish (n = 12), mammals (n = 38), and reptiles (n = 3). Four main group sources were modeled: marine fish, freshwater fish, terrestrial animals, and C3 plants (Table 2). C4 plants and mollusks were not included, since there is no significant evidence of C4 plants’ consumption by these groups [27] and incorporating mollusks into the Bayesian model presents significant challenges: the organic matter typically constitutes only a minor component of carbonate shells, nitrogen concentrations are typically low, and diagenetic alteration may compromise isotopic integrity [62].
C3 plants values were obtained from modern fruits (n = 30), roots (n = 5), and heart of palms (n = 13) collected between 2010 and 2013 in southeastern Brazil’s Atlantic Forest national parks (−29.2 ± 3‰ for δ13C and 1.1 ± 2‰ for δ15N), whose δ13C values were corrected for the Suess effect (+2‰) using atmospheric air value from 2010 (−8.4‰) [15,16,63]. Carbon and nitrogen concentration values for each source were obtained from Chinique de Armas et al. [64] and Newsome et al. [40].
Human bone and dentine values were worked together and grouped into age intervals defined by the authors and based on dentine-age growth anatomical areas [17,57,65] (Table 3). These dentine age intervals were used as reference to define broader age categories, starting at 5 years of age. Due to missing values in some age intervals, we merged adjacent age groups and created four broader age-categories used in our analysis: 5–9 years, 9–15 years, 15–18 years, and adults (Table 4; S4 and S5 Tables in S1 File).
MixSIAR adaptation to sequential dentine samples
An adaptation of MixSIAR to incorporate correlated values was developed at the Institute of Mathematics and Statistics, University of São Paulo, Brazil. We included in the MixSIAR program a priori independent additive random error terms at both individual and sample levels for isotopic measurements, modeling deviations of each individual measurement from the population mean and, when applicable, deviations of repeated measurements from the same individual. This hierarchical approach accommodates both inter- and intra-individual variability, essential for modeling repeated measurements.
Each isotopic value is assumed to follow a conditionally normal distribution, where the average is determined by the weighted average of the tracer values from food sources, adjusted by a trophic enrichment factor. The weights are based on consumption proportions and nutrient concentration variations per food source.
For bone samples, the error from the conditional average includes only the individual-level effect, as there is one measurement per individual. For dentine samples, the error term includes both individual and sample-level effects. The individual-level effect has the same variance as in bone samples and is consistent across multiple samples from the same individual, while the sample-level effect reflects deviations between the sample measurement and the individual’s average values.
We tested the normality assumption for δ13C and δ15N values from adult bone samples in each site, since these are the only subgroups with independent and sufficiently large sample size. Lilliefors’ test indicated non-normality for δ13C values at Moraes (p < 0.001). As a robustness check, we repeated the analysis excluding two samples from Moraes with outlying δ13C values (MO-07 and MO-08) and normality was not rejected without those samples (p > 0.2). The comparative analysis found no major changes and we chose to keep the full dataset for the main results. Details of the statistical tests performed and the robustness analysis are in the S4 File.
Parameter estimation was performed using Bayesian inference, with informative prior distributions for all parameters except the main parameter of interest, the consumption proportions, that were chosen as the uniform prior for each subgroup (e.g., age categories) over the 3-simplex, the Dirichlet(1, 1, 1, 1) distribution. This prior is not marginally uniform over each food source, since the consumption proportions must sum to 100%. Each dimension of a Dirichlet(1, 1, 1, 1) distribution follows a marginal Beta(1, 3) distribution, which is asymmetric and decreases over the (0, 1) interval.
Estimation was conducted jointly across all subgroups within the same site, with shared parameters for error variances and mean food source isotope values. Posterior distribution sampling was conducted using the no-U-turn-sampler Markov Chain Monte Carlo algorithm, implemented in Stan software and interfaced through the R programming language. The R and Stan language code is available in the GitHub repository (https://github.com/AFumis/Adapted_MixSIAR_diet_analysis).
Inference on quantities of interest from posterior samples was performed by computing empirical frequencies when possible. For density plots and highest posterior density (HPD) intervals, we used the kernel density estimator (KDE) from the stats R package, and incorporated the reflection technique for boundary bias correction [66] when KDE is applied within a limited interval, such as the (0,1) interval for consumption proportions.
Radiocarbon dating
Since all available radiocarbon dates from Piaçaguera correspond to individuals from Group I, in this study two individuals from Group II were dated at Beta Analytic (USA). Bone samples (~400 mg of ribs) from individuals PI-32 and PI-56 were analyzed using Beta’s in-house NEC accelerator mass spectrometer and four Thermo IRMS units.
Both new and previously published radiocarbon dates from Piaçaguera were calibrated on calib.org platform (8.2 version) considering the corrections for the marine reservoir effect: a ΔR value of −110 ± 25 years, based on ΔR values reported for the region of Santos [67,68], and a mixed marine curve Marine20 and the atmospheric curve SHCal20, adjusted according to the results on marine protein intake identified by the adapted MixSIAR model.
Individual PI-32, aged approximately 15 years old at death, was assigned to the “9–15-year” age category in the MixSIAR model (S4 Table in S1 File). As its radiocarbon date was obtained from rib bone, whose relatively fast turnover rate reflects dietary input from up to the last five years of life [69], we used the MixSIAR result for 9–15-years age category to calibrate this date. For the others dated individuals, all adults, we used the results for adult age categories.
Results
δ13C and δ15N in bone and dentine collagen
All samples presented in this study meet the quality criteria proposed for archaeological collagen samples (S4-S6 Tables in S1 File). The probability of a correlation between δ13C or δ15N values and other variables is low or non-existent at both sites (Spearman’s test: p > 0.05; rho < 0.6) (S8 Table in S1 File). The δ13C and δ15N values for human bone and dentine and fauna bone are presented in S4-S6 Tables in S1 File and summarized in Table 5.
In Piaçaguera, the fauna values dispersion is broad (coefficients of variation of 33% for both δ¹³C and δ¹⁵N), reflecting the local ecosystem diversity. As shown in Fig 3, trophic level differences are observed: sharks (Elasmobranchii) have the highest δ¹⁵N values (12.8‰ to 17.3‰), followed by most marine and estuarine fish (Teostei) (10.9‰ to 15.5‰), with a statistically significant difference between these groups (Mann-Whitney [MW]: p = 0.001). Smaller marine fish (e.g., sea catfish [Ariidae]), show lower δ¹⁵N (8.6‰ to 10.1‰), significantly different from sharks (MW: p = 0.0001) and other marine/estuarine fish (p = 0.0004).
Terrestrial fauna has the lowest δ¹⁵N values (4.9‰ to 14.6‰), with trophic levels aligning with dietary behavior: carnivorous (e.g., Euphractus sexcintus and Lontra longicaudis), omnivores (e.g., Nasua nasua, Didelphis aurita, Procyon cancrivorus, and Tayassu sp.), and herbivores (e.g., Alouatta sp., Dasyprocta leporina, and Mazama sp.). The otter (Lontra longicaudis), whose diet includes fish, sits highest among terrestrial taxa.
Human δ¹⁵N values in Piaçaguera fall within the range of most marine and estuarine fish (MW: p = 0.05), being lower than shark values (MW: p = 0.009) and higher than small marine fish (MW: p = 0.007) and terrestrial animals’ values (MW: p = 0.0001). Considering an average increase of 3‰ in δ¹⁵N and 1‰ in δ¹³C per trophic level [29,30], human values are within the range of terrestrial and marine animals, suggesting a diet primarily based on marine/estuarine fish and terrestrial mammals, in line with previous studies [6,24].
Humans from Group I present lower δ13C and δ15N values than those from Group II (Fig 3), with statistically significant differences among adults (MW: δ13C: p = 0.001; δ15N: p = 0.004) and subadults for δ13C (MW: p = 0.004). No significant sex-based differences were found within Group I (MW: δ13C: p = 0.1; δ15N: p = 0.06) or Group II (MW: δ13C: p = 0.1; δ15N: p = 0.2).
Fig 4 shows the distribution of bone values for adults and subadults and dentine values (representing the values from childhood and adolescence), in both Groups I and II. Subadult values were divided by age (under and over 5 years old), since breastfeeding may influence them up to that age.
Subadults are divided by age (until and after 5 years) and adults by sex.
In both Groups I and II, dentine values closely align with subadult bone values and the δ15N values until 5 years are higher than those of adults (Fig 4; S1.A – S1.H Figs in S2 File for individual dentine profiles), probably reflecting breastfeeding.
In Moraes, human bone δ13C and δ15N values indicate a predominantly terrestrial-base diet (Fig 5). No statistically significant differences in δ13C were found between sexes (MW: p = 0.27) or between adults and subadults (MW: p = 0.09). However, a significant difference in δ15N was observed between adults and subadults (MW: p = 0.0002) (Fig 6).
Subadults are divided by age (until and after 5 years) and adults by sex.
Dentine values closely align with subadult bone values (Fig 6), with higher variability compared to adult values (Fig 6; S2.A – S2.M Figs in S3 File for individual dentine profiles). δ15N values until 5 years are significantly higher than those of adults, probably reflecting breastfeeding potentially combined with a higher freshwater fish intake by children and/or lactating women (17). Even in the > 5 years old group, some subadults show higher δ15N values than the adults (MW: p = 0.0001) (S2.A – S2.M Figs in S3 File).
MixSIAR and diet modeling
For Piaçaguera, two models were developed: Model 1 grouped all individuals, while Model 2 separated them into Group I and Group II. The 15–18 years age category was excluded from the analysis due to insufficient sample size (Table 4).
In Model 1, marine fish is the primary dietary source across all age groups, whereas C3 plant contributes the least (Fig 7). As shown in S9 Table in S1 File, freshwater fish and terrestrial animals have modes at or near zero across all age categories, but with higher dispersion than C3 plants. A slight variation is observed in the 9–15 years category, where HPD interval limits and quartiles for marine fish are approximately 4 percentage points higher, and those for terrestrial animals about 3 percentage points lower than in other age categories.
Some individuals could not be assigned to either Group I or II due to the lack of burial stratigraphy data, increasing the sample size in Model 1. Nonetheless, this did not affect the results, with dietary proportions remaining consistent.
In Model 2, Group I and II show the posterior distributions that indicate different primary food sources. Group I shows terrestrial animals as the primary source for subadults and marine fish for adults, while C3 plants as the lowest source for all age categories (Fig 7). Group II shows marine fish as the primary source and C3 plants as the lowest for all age groups. When comparing the posterior distribution between Groups I and II, Group I has higher estimates for terrestrial animal intake across all age categories, indicating higher consumption of terrestrial animals in Group I than in Group II.
Group I shows lower values for marine fish consumption than Group II for adults (Group I quartiles: 0.185, 0.292, 0.387; Group II quartiles: 0.395, 0.488, 0.569) and for 5–9 years category (Group I quartiles: 0.161, 0.262, 0.360; Group II quartiles: 0.295, 0.388, 0.474). For 9–15 years category, both Group I and II show similar values for marine fish (Group I quartiles: 0.272, 0.373, 0.466; Group II quartiles: 0.256, 0.352, 0.441) (S10 and S11 Tables in S1 File). Considering these posterior distributions and using the highest third quartile between age groups as a benchmark, we can estimate that the proportion of marine fish may reach up to 39% for adults and 46% for subadults in Group I, and up to 57% for adults and 47% for subadults in Group II.
The values for terrestrial animal consumption are higher in Group I than Group II for adults (Group I quartiles:0.113, 0.212, 0.340; Group II quartiles: 0.056, 0.127, 0.243), for 5–9 years category (Group I quartiles: 0.170, 0.304, 0.459; Group II quartiles: 0.101, 0.203, 0.337), and for 9–15 years category (Group I quartiles: 0.210, 0.359, 0.513; Group II quartiles: 0.165, 0.285, 0.420) (S10 and S11 Tables in S1 File). It can be estimated that the proportion of terrestrial animals in the diet may reach up to 34% for adults and 51% for subadults in Group I, and up to 24% for adults and 42% for subadults in Group II.
When comparing subadults between Model 1 and 2, for both the 5–9 years and 9–15 years categories, the overall sample (Model 1) shows higher estimates for marine fish consumption and lower estimates for terrestrial animal consumption than Groups I and II (Model 2) (Fig 7; S9-S11 Tables in S1 File).
In Moraes, terrestrial animals are the main food source across all age categories, followed by freshwater fish and C3 plants (Fig 8). Terrestrial animal intake is highest among adults, particularly when compared to the 5–9 years and 9–15 years categories, with the highest lower and upper limits 50% HPD interval (0.091, 0.377) and quartiles (0.163,0.301,0.459), matching the 15–18 years category in the third quartile (0.459) (S12 Table in S1 File). The posterior distribution shifts rightward with age, indicating increased terrestrial animal source proportion (Fig 8). It can be estimated that the proportion of terrestrial animals in the diet of Moraes may reach up to 46% for adults and 43% for subadults up to 15 years old.
Conversely, freshwater fish consumption has the highest lower and upper limits of the 50% HPD interval for the 5–9 years and 9–15 years categories (0.024, 0.314 for 5–9 years and 0.056, 0.345 for 9–15 years), compared to the 15–18 years (0, 0.279) and adult categories (0, 0.287) (S12 Table in S1 File). It can be estimated that the proportion of freshwater fish may reach up to 45% for adults and 46% for subadults up to 15 years old.
C3 plant intake is consistently lower than that of terrestrial and freshwater sources across all age categories, but higher than in Piaçaguera. For instance, adults in Moraes show higher quartiles (0.108, 0.213, 0.341) than Piaçaguera’s Model 1 (0.052, 0.108, 0.182) and Model 2 (Group I: 0.044, 0.098, 0.175; Group II: 0.071, 0.142, 0.226) (S9-S12 Tables in S1 File). The estimated proportion of C3 plant in the diet may reach up to 20% for Piaçaguera (Groups I and II) and 30% in Moraes.
Interestingly, marine fish consumption is minimal across all categories in Moraes, with 50% HPD limits typically below 13% and unimodal distributions centered near zero (S12 Table in S1 File), suggesting minimal or no consumption of marine fish.
Radiocarbon dates
The adapted MixSIAR model estimated marine protein intake of up to 40% for adults from Group I, 60% for adults from Group II, and 45% for the 9–15 years age group from Group II (S10 and S11 Tables in S1 File). These values were applied during calibration using a mixed Marine20 and atmospheric SHCal20 curve.
Table 6 presents all the radiocarbon dates for Piaçaguera (divided in Group I and II) and Moraes. The new dates published here regarding Group II suggest a possible contemporaneity between individuals from Group I and Group II.
Discussion
The diet in Piaçaguera shellmound group
Isotopic values from adult bone collagen at Piaçaguera indicate a diet based on the consumption of marine and terrestrial animals, aligning with previous isotopic and zooarchaeological studies [6,24]. Similar values are found in other archaeological groups with mixed marine/terrestrial diets, such as those from Argentine coast of Patagonia [73–75].
However, Piaçaguera δ13C and δ15N values are lower than those from other Brazilian coastal shellmounds where marine/estuarine fish intake was more significant (e.g., Morro do Ouro, Rio Comprido, and Enseada I sites in northern Santa Catarina [10,15,16]), and even lower than in groups with predominantly marine diets (e.g., Sambaqui do Moa in Rio de Janeiro state and Jabuticabeira II, Capivari I, Cabeçuda, Armação do Sul, and Galheta IV sites in central and southern Santa Catarina [7,11–13,24,76,77]. This reinforces the significant contribution of terrestrial resources to the diet in Piaçaguera.
Zooarchaeological studies suggest a diet based on small marine and estuarine fish and local terrestrial fauna [6]. Our results confirm this, pointing to low-trophic-level fish from shallow or estuarine waters (e.g., catfish [Genidens barbus], black drum [Pogonias cromis], snook [Centropomus sp.], and whitemouth croaker [Micropogonias furnieri]) as major protein sources. Terrestrial sources included small mammals from the Atlantic Forest (e.g., coatis [Nasua nasua], raccoons [Procyon cancrivorus], guinea pigs [Cavia aperea], and howler monkeys [Alouatta sp.]) (Fig 3). This local-resource pattern is consistent with other isotopic studies from Brazilian shell mounds [10,13,15,16], and Piaçaguera is no exception.
Some species found at the site, such as alligators, sharks, and open-sea fish (e.g., tripletail [Lobotes surinamensis] and angelfish [Pomacanthus paru]), may have been consumed opportunistically or during special occasions. Borges [6] suggests that higher sea levels at the time may have brought open-sea animals closer to the settlement, increasing their availability. Klökler [78] also notes the frequent use of shark teeth as tools and ornaments in burials, suggesting that their occasional consumption may have been intentional.
The adapted MixSIAR model confirms marine fish as the main protein source, contributing around 30–50% of the diet. Terrestrial animals and freshwater fish accounted for 10–30%, while C3 plants contributed around 10%. Since bulk collagen reflects mostly the protein component of the diet, the contribution of plant-based foods may be underestimated. Protein content in collagen is mainly assessed through nitrogen, which plants supply at relatively low levels, particularly due to the limited digestibility of a substantial portion of the nitrogen sources they contain [79,80].
Due to this limitation, it is important to consider other methods for inferring plant consumption in a population, such as the identification of plant microremains in dental calculus or, indirectly, through patterns of dental wear and the presence of caries. Although carbohydrate intake was not directly studied in Piaçaguera, nor was dental wear analyzed for dietary purposes, dental caries are rare in this group [81]. Caries etiology is complex and multifactorial, but frequent intake of fermentable carbohydrates is a key contributor [82–84]. Their low prevalence in the group suggests that plant-based foods were probably less important than in other shell mound groups from southern Brazilian coast [84,85].
When grouped by burial location, Group II individuals show slightly higher δ13C and δ15N values than those in Group I (Table 6; Figs 3 and 4), with a statistically significant difference for δ13C (MW: p = 0.001). Both groups had mixed marine/terrestrial diets, but Group I appears to have consumed more terrestrial resources. This is supported by the adapted MixSIAR model, which shows higher density values for terrestrial animal intake in Group I (Fig 7). Two hypotheses may explain these dietary differences:
Scenario 1 – Environmental change: The Holocene sea-level regression along the Brazilian Atlantic coast [86,87] may have increased terrestrial fauna availability, leading to higher consumption of this resource by Group I. A similar scenario has been proposed for Sambaqui do Moa, Rio de Janeiro [11]. If this were the case, Group I would represent a more recent occupation of the site. However, new radiocarbon dates presented here for two individuals from Group II, ranging between 6,288 and 5,927 yBP, show that Group I and II were contemporaneous, suggesting that the slight change in diet was more probably due to cultural factors than environmental changes.
Scenario 2 – Cultural factors: Affinity groups based on varying degrees of kinship, neighborhood associations, or other social principles has been proposed for the shellmound Jabuticabeira II, located on Brazil’s southern coast [88,89]. Social structures potentially similar to those proposed for Jabuticabeira II, such as affinity groups, may also have existed at Piaçaguera and could explain both the distinct burial areas and differences in diet, with each group having differential access to terrestrial or marine resources.
Among Group I individuals, females (n = 5) showed slightly lower isotopic values than males (n = 6) (Figs 3 and 4), although this difference was not statistically significant (MW: δ13C: p = 0.1; δ15N: p = 0.06). This may reflect a slightly higher intake of terrestrial resources by females, possibly due to dietary restrictions/preferences or different access to food resources. While osteological stress markers associated with physical activities in Piaçaguera found no significant sex-based differences [90,91], different tasks – that could involve similar physical efforts – may have led to differential food access, for example, males responsible for fishing and females for hunting. Furthermore, the presence of individuals from outside the group cannot be excluded.
In both Groups I and II, subadult isotopic values are generally similar to those of adults, except for δ15N in children under five, which may be influenced by breastfeeding [17]. From age six onward, δ15N values align closely with those of adults (S1.A-S1.H Figs in S2 File), indicating that children and adolescents also had a mixed marine/terrestrial diet. The adapted MixSIAR model supports these findings, showing similar proportions of resource consumption between subadults and adults in both groups (Fig 7).
Notably, the higher intake of terrestrial resources observed in Group I adults was also observed in Group I subadults from age six onward (Fig 7; S10 Table in S1 File). Within scenario 2 hypothesis, which proposes that social structures like affinity groups influenced diet, these groups would have been established early in life, meaning that social organization and its impact on diet were already present from childhood.
Minor individual variations in isotopic values during childhood and adolescence in both Groups I and II (S1.A-S1.H Figs in S2 File) may reflect slight variabilities in the intake frequencies of terrestrial/marine resources and/or punctual periods of metabolic alterations, which could influence δ13C and δ15N values [92,93].
Human skeletal series dating over millennia are often treated as homogeneous groups, as was the case of Piaçaguera. However, such series are not representative of the entire population [94], and radiocarbon dating can help contextualize the individuals temporally. While financial constraints may restrict access to such analysis, archaeological information (e.g., spatial distribution of burials) is also essential to contextualize the individuals and reduce errors associated with homogenizing entire series. In this study, dividing human burials into clusters based on their stratigraphy revealed significant isotopic variations in diet, highlighting the importance of a constant dialogue between archaeological and isotopic data.
The diet in Moraes shellmound group
Isotopic values from adult bone collagen at Moraes suggest a predominantly terrestrial diet, in line with previous isotopic and zooarchaeological studies [14,24]. Unlike Piaçaguera, no distinct burial groups were identified at Moraes [95], and all radiocarbon-dated individuals show similar isotopic values (Table 5 and S4 Table in S1 File), suggesting certain stability in diet within the group.
The δ13C and δ15N values from Moraes are similar to other archaeological groups that relied primarily on terrestrial animals (e.g., Argentine Patagonia [96]), but higher than those of groups relying mainly on terrestrial herbivores (e.g., Archaic-period hunter-gatherers from Chile [97]), C3 plants (e.g., groups from Alto Jararaca II and Gruta do Rio dos Altos sites, southern Brazil [98]), and C3 plant horticulturalists (e.g., CAI period in Chile [97]).
Considering (a) the estimated trophic enrichment in ¹⁵N and ¹³C between consumer and food source values [29,30], (b) the zooarchaeological record from Moraes [14], and (c) the isotopic values of humans and fauna from the site (Fig 5), the main protein sources for the Moraes group were most likely small- and medium-sized terrestrial mammals from the Atlantic Forest. These include lowland paca (Cuniculus paca), South American coati (Nasua nasua), peccary (Tayassu sp.), and howler monkey (Alouatta sp.). These findings align with evidence of low long-range mobility for this group [99] and with zooarchaeological data from nearby riverine shell mounds [9].
Some species, such as tapir (Tapirus terrestres), alligator (Caiman latirostris), and ocelot (Leopardus sp.), may have been consumed opportunistically or in specific cultural contexts. Freshwater fish also contributed to the diet but likely played a second role compared to terrestrial mammals.
The adapted MixSIAR model support these findings, indicating that terrestrial animals contributed approximately 16–45% of total protein intake, freshwater fish accounted for 14–45%, and C3 plants contributed 10–34%. The C3-plant intake is also supported by the high prevalence of dental caries in the group [14] from results provided by one of the authors [VW]), and by the presence of starch/phytoliths in dental calculus [99], associated with the consumption of yams (Dioscorea sp.) and sweet potatoes (Ipomoea batatas) [27].
Although the adapted MixSIAR model points to a minimal marine fish intake in Moraes, this result requires careful interpretation. Evidence of contact between Moraes and coastal groups is proved by the presence of marine mollusk shells and perforated shark teeth (Carcharinus plumbeus) at the site associated with some human burials [14]. However, this interaction does not appear to have been extensive enough to include, for example, intergroup marriage [21,100].
Marine species could have been brought to Moraes for consumption, or individuals from the group may have traveled to the coast and brought to the site marine materials as symbolic artifacts, but their presence in the dietary modeling does not necessarily confirm their regular consumption in Moraes. δ15N values from some estuarine/marine fish (e.g., Archosargus probatocephalus) can overlap with those of large carnivores’ freshwater fish, such as freshwater catfish (Actinopterygii) and trahira (Hoplias spp.) (Fig 5; S6 and S7 Tables in S1 File), potentially leading to ambiguous results. Therefore, BMM results in archaeological studies require cautious interpretation, relying on contextual information about the analyzed population.
Subadult isotopic values also indicate a terrestrial-based diet. However, while δ13C values remain similar between adults and subadults of all ages, δ15N shows a different and interesting pattern. For children under five years old, δ15N values exceed the typical 3‰ increase with breastfeeding (Fig 6; S2.A-S2.M Figs in S3 File). As proposed by DiGiusto et al. [17], this elevation could also result from a higher intake of freshwater fish (which present higher δ15N values than terrestrial animals) by young children and/or nursing women, not only breastfeeding. The Moraes group exhibits a high neonatal mortality rate, and freshwater fish could have served as a source of nutritional support during childhood [17].
From the age of five onwards, δ15N values remain higher than those of adults (Fig 6; S2.A-S2.M Figs in S2 File), as confirmed by the adapted MixSIAR model, that indicates higher proportions of freshwater fish in the diet of subadults than in adults. As subadults’ age increases, the proportions of freshwater fish decrease and terrestrial animals increase (Fig 8). These findings highlight a clear dietary difference between adults and subadults at Moraes and reflect a community strategy aimed at ensuring the survival of subadults.
No significant differences in isotopic values were found between females and males in the group, indicating no dietary differences between sexes. However, osteological stress marker analysis suggests possible differences in activity patterns in Moraes. Males showed higher osteoarthritis frequencies in the upper limbs, indicating more intense and/or frequent use of the upper limbs by them, and auditory exostosis (bony growths of the external auditory meatus, frequently associated with aquatic activities involving diving in cold water) was observed exclusively in males [99].
Considering the terrestrial-based diet of the group and a possible higher freshwater fish intake by subadults and/or nursing women, these activity markers may be related to male-dominated fishing activities, possibly involving diving and sailing. This suggests a male food provisioning system for fish resources for subadults and/or nursing women. Male provisioning for maternal and childhood diets has been proposed for some South American hunter-gatherer groups [101] and is documented in ethnographic studies of the Hazda and Hiwi ethnic groups in Africa [102,103]. A similar dynamic may have been present in the Moraes group.
An evaluation of MixSIAR adaptation for use bone and dentine collagen isotopic values
The introduction of additive error terms in the model’s likelihood function allows for the inclusion of repeated samples from the same individual without violating the independence assumptions between samples. This approach mitigates information loss that would result from using intra-individual averages instead of the full set of measurements, and enables the simultaneous analysis of bone samples (one observation per individual) and dentine samples (multiple observations per individual).
Similar to the original MixSIAR, this adapted model may become overly dependent on prior choice when the number of food sources exceeds the number of isotopic measurements by more than one. This was the case in our study. Stock et al. [104] discuss best practices in such situation, recommending that all food sources be kept in the model and aggregating posterior consumption proportions if needed.
This dependency on the prior is a limitation that raises challenges for consistency in estimates when subdividing the sample, since the number of measurements per group is reduced – another situation where the prior tends to dominate. This is illustrated by the slightly higher marine fish consumption for subadults in Piaçaguera in Model 1 (without divide the group in Group I and II) in comparison with Model 2 (where the groups divided in Group I and II). The higher total sample size in Model 1 allows to obtain a posterior distribution more distinct from the prior than either of the subpopulations separately in Model 2.
Another constraint concerns the identification of food sources with overlapping isotopic values, such as small marine or estuarine fish and large freshwater fish, as observed in the Moraes model. However, this limitation can be mitigated by archaeological contextual information, which aids in interpreting BMM results.
Despite these limitations, BMMs remain a strong alternative to enhance dietary interpretations based on isotopic analysis.
The model presented here was capable to produce distinct posterior distributions for the analyzed human groups from the same prior, despite potentially small sample sizes in certain categories. Its agreement with qualitative analysis of archeological remains reinforces its utility and serves as a useful tool to measure the consumption of different food sources through the life of individuals.
Conclusion
In Piaçaguera, the data revealed a mixed marine and terrestrial diet, with notable differences in diet between the individuals from Groups I and II. Based on the new radiocarbon dates, the higher intake of terrestrial resources by Group I may be attributed to social structures within the community, that could have influenced resource consumption since childhood. This highlights the importance of integrating isotopic and archaeological data to achieve a comprehensive understanding of past dietary practices.
In Moraes, the group had a predominantly terrestrial diet with significant age-related differences. Subadults showed a higher consumption of freshwater fish, reflecting a community strategy aimed at ensuring the nutritional necessities of subadults and nursing women, with a possible male provisioning system for fish resources.
The adapted MixSIAR model, as well as the original MixSIAR, has inherent complexities. Despite this and the limitations of the samples themselves (such as the total number of samples available and the lack of samples at shorter age intervals), this adaptation has proven effective in delineating dietary patterns within archaeological groups, by increasing the number of measurements in the model, adding bone and dentin isotopic values together, and being able to present the sources of food consumed by individuals throughout their lives at different ages. Its ability to incorporate both bone and dentine collagen values at the same time without compromising the independence of the samples increases its robustness. While dependence on prior distributions remains a challenge, the model’s ability to align with archaeological evidence and identify dietary patterns in distinct archaeological contexts validates its utility.
Future research into other isotopic elements has the potential to further illuminate social dynamics within these groups, elucidating the factors influencing their dietary practices.
Supporting information
S1 File. S1-S12 Tables.
Materials, isotopic results, and MixSIAR results.
https://doi.org/10.1371/journal.pone.0335680.s001
(XLSX)
S2 File. S1.A-S1.H Figs
: Individual dentine δ13C and δ15N isotope ratio profiles for Piaçaguera.
https://doi.org/10.1371/journal.pone.0335680.s002
(PDF)
S3 File. S2.A-S2.M Figs
: Individual dentine δ13C and δ15N isotope ratio profiles for Moraes.
https://doi.org/10.1371/journal.pone.0335680.s003
(PDF)
Acknowledgments
The authors thank the Center for Applied Statistics (University of São Paulo) for the statistical assistance in this project, which led to this collaboration. We also thank the staff of the Laboratoire Méditerranéen de Préhistoire Europe Afrique for the technical support, especially Guy André, the Museum of Archaeology and Ethnology (University of São Paulo) for providing access to the archaeological samples, Steve Brookes from Iso–Analytical Ltd for elemental and isotopic measurements, and Victor de Lima (IME/USP) for the support in the development of the adapted MixSIAR model.
References
- 1. DeBlasis P, Gaspar M, Kneip A. Sambaquis from the Southern Brazilian Coast: Landscape Building and Enduring Heterarchical Societies throughout the Holocene. Land. 2021;10(7):757.
- 2. Kneip A, Farias D, DeBlasis P. Longa duração e territorialidade da ocupação sambaquieira na laguna de Santa Marta, Santa Catarina. rsab. 2018;31(1):25–51.
- 3. Villagran XS, Klokler D, Nishida P, Gaspar MD, DeBlasis P. Lecturas estratigráficas: arquitectura funeraria y depositación de residuos en el sambaquí Jabuticabeira II. Latin Am antiq. 2010;21(2):195–216.
- 4. Gaspar MD. Considerations of the sambaquis of the Brazilian coast. Antiquity. 1998;72(277):592–615.
- 5. Figuti L, Plens CR, DeBlasis P. Small Sambaquis and Big Chronologies: Shellmound Building and Hunter-Gatherers in Neotropical Highlands. Radiocarbon. 2013;55(3):1215–21.
- 6.
Borges C. Analyse archéozoologique de l’exploitation des animaux vertébrés par les populations de pêcheurs-chasseurs- cueilleurs des sambaquis de la Baixada Santista, Brésil, entre 5000 et 2000 BP. France: Muséum National d’Histoire Naturelle. 2015.
- 7. Klokler D, Gaspar MD, Scheel-Ybert R. Why clam? Why clams? Shell Mound construction in Southern Brazil. Journal of Archaeological Science: Reports. 2018;20:856–63.
- 8. Scheel-Ybert R, Boyadjian C. Gardens on the coast: Considerations on food production by Brazilian shellmound builders. Journal of Anthropological Archaeology. 2020;60:101211.
- 9.
Tognoli A. Zooarqueologia dos sambaquis fluviais (Caraça, Estreito, Tatupeva e Lageado IV): uma leitura da paisagem sambaquieira da região de Itaoca, Vale do Ribeira de Iguape. Brazil: University of São Paulo. 2016.
- 10. Bastos M, Lessa A, Rodrigues-Carvalho C, Tykot R, Santos R. Análise de isótopos de carbono e nitrogênio: a dieta antes e após a presença de cerâmica no sítio Forte Marechal Luz. Revista do Museu de Arqueologia e Etnologia. 2014;24:137–51.
- 11. Bastos M, Guida V, Rodrigues-Carvalho C, Toso A, Santos R, Colonese A. Elucidating pre-columbian tropical coastal adaptation through bone collagen stable isotope analysis and bayesian mixing models: insights from Sambaqui do Moa (Brazil). Revista de Antropología del Museo de Entre Ríos. 2022;7(1):1–10.
- 12.
Cardoso J. Perspectives on food and movement: shellmounds occupation in the southern and southeastern Brazil through multiple isotopic proxies. Brazil: University of São Paulo. 2024.
- 13. Pezo-Lanfranco L, DeBlasis P, Eggers S. Weaning process and subadult diets in a monumental Brazilian shellmound. Journal of Archaeological Science: Reports. 2018;22:452–69.
- 14. Plens C. Animals for humans in life and death. Revista do Museu de Arqueologia e Etnologia. 2010;20:31–51.
- 15. Toso A, Hallingstad E, McGrath K, Fossile T, Conlan C, Ferreira J, et al. Fishing intensification as response to Late Holocene socio-ecological instability in southeastern South America. Sci Rep. 2021;11(1):23506. pmid:34873216
- 16. Pezo-Lanfranco L, Eggers S, Petronilho C, Toso A, da Rocha Bandeira D, Von Tersch M, et al. Middle Holocene plant cultivation on the Atlantic Forest coast of Brazil?. R Soc Open Sci. 2018;5(9):180432. pmid:30839761
- 17. DiGiusto M, Herrscher E, Jaouen K, Bastos M, Mion L, Wesolowski V. Early childhood in precolonial South America: Breastfeeding, infant mortality and stable isotopes analysis in southeastern shellmound builders during the Middle Holocene. Int J Osteoarchaeol. 2025;35:e3381.
- 18.
Fischer P. Os moleques do morro e os moleques da praia: estresse e mortalidade em um sambaqui fluvial (Moraes, vale do Ribeira de Iguape, SP) e em um sambaqui litorâneo (Piaçaguera, Baixada Santista, SP). Brazil: University of São Paulo. 2012.
- 19. Di Giusto M. Lifestyle on the south coast of Brazil: considerations about shell mound (sambaqui) builders through bone and dental analysis. Bol Mus Para Emílio Goeldi Cienc Hum. 2023;18(1).
- 20.
Estevam R. (Re)começando do princípio: o que a arqueografia de uma área funerária do Sambaqui de Cabeçuda pode nos ensinar sobre práticas funerárias sambaquieiras?. Brazil: University of São Paulo. 2020.
- 21. Fidalgo D, Wesolowski V, Hubbe M. Biological affinities of Brazilian pre-colonial coastal communities explored through bootstrapped biodistances of dental non-metric traits. Journal of Archaeological Science. 2022;138:105545.
- 22.
Buikstra J, Ubelaker D. Standards for data collection from human skeletal remains. Arkansas Archaeological Survey. 1994.
- 23.
Schaefer M, Black S, Scheuer L. Juvenile osteology: A laboratory and field manual. Schaefer M, Black S, Scheuer L, editors. Elsevier Inc. 2009.
- 24. Colonese AC, Collins M, Lucquin A, Eustace M, Hancock Y, de Almeida Rocha Ponzoni R, et al. Long-term resilience of late holocene coastal subsistence system in Southeastern South america. PLoS One. 2014;9(4):e93854. pmid:24718458
- 25.
DiGiusto M. Para a infância e mais além: uma perspectiva bioarqueológica da alimentação, estresse e morte em duas sociedades sambaquieiras. Brazil: University of São Paulo. 2023.
- 26.
Uchôa D, Alvim M, Gomes J. Demografia esqueletal dos construtores do sambaqui de Piaçaguera São Paulo, Brasil. Dédalo. 1989;455–70.
- 27.
Boyadjian C. Microfósseis contidos no cálculo dentário como evidência do uso de recursos vegetais nos sambaquis de Jabuticabeira II (SC) e Moraes (SP). Brazil: University of São Paulo. 2007.
- 28. Colonese AC, Farrell T, Lucquin A, Firth D, Charlton S, Robson HK, et al. Archaeological bone lipids as palaeodietary markers. Rapid Commun Mass Spectrom. 2015;29(7):611–8. pmid:26212278
- 29. Hedges REM, Reynard LM. Nitrogen isotopes and the trophic level of humans in archaeology. Journal of Archaeological Science. 2007;34(8):1240–51.
- 30. O’Connell TC, Kneale CJ, Tasevska N, Kuhnle GGC. The diet-body offset in human nitrogen isotopic values: a controlled dietary study. Am J Phys Anthropol. 2012;149(3):426–34. pmid:23042579
- 31. Matsubayashi J, Tayasu I. Collagen turnover and isotopic records in cortical bone. Journal of Archaeological Science. 2019;106:37–44.
- 32. Hedges REM, Clement JG, Thomas CDL, O’connell TC. Collagen turnover in the adult femoral mid-shaft: modeled from anthropogenic radiocarbon tracer measurements. Am J Phys Anthropol. 2007;133(2):808–16. pmid:17405135
- 33. Lapillonne A, Picaud J, Glorieux F, Salle B. Bone turnover assessment in infants. Acta Paediatrica. 2000;89(7):772–4.
- 34. Beaumont J, Gledhill A, Lee‐Thorp J, Montgomery J. Childhood Diet: A Closer Examination Of The Evidence From Dental Tissues Using Stable Isotope Analysis Of Incremental Human Dentine*. Archaeometry. 2012;55(2):277–95.
- 35. Beaumont J, Atkins E-C, Buckberry J, Haydock H, Horne P, Howcroft R, et al. Comparing apples and oranges: Why infant bone collagen may not reflect dietary intake in the same way as dentine collagen. Am J Phys Anthropol. 2018;167(3):524–40. pmid:30187451
- 36. Czermak A, Schermelleh L, Lee‐Thorp J. Imaging‐assisted time‐resolved dentine sampling to track weaning histories. Intl J of Osteoarchaeology. 2018;28(5):535–41.
- 37. Ganiatsou E, Georgiadou A, Souleles A, Aidonis A, Protopsalti T, Tzevreni S, et al. Application of machine learning on isotopic data from tooth microsections for reconstructing weaning patterns and physiological stress. Journal of Archaeological Science: Reports. 2023;47:103765.
- 38. Scharlotta I, Goude G, Herrscher E, Bazaliiskii VI, Weber AW. “Mind the gap”-Assessing methods for aligning age determination and growth rate in multi-molar sequences of dietary isotopic data. Am J Hum Biol. 2018;30(5):e23163. pmid:30129288
- 39. Cheung C, Szpak P. Interpreting Past Human Diets Using Stable Isotope Mixing Models. J Archaeol Method Theory. 2020;28(4):1106–42.
- 40. Newsome SD, Phillips DL, Culleton BJ, Guilderson TP, Koch PL. Dietary reconstruction of an early to middle Holocene human population from the central California coast: insights from advanced stable isotope mixing models. Journal of Archaeological Science. 2004;31(8):1101–15.
- 41. Tsutaya T, Yoneda M. Quantitative reconstruction of weaning ages in archaeological human populations using bone collagen nitrogen isotope ratios and approximate Bayesian computation. PLoS One. 2013;8(8):e72327. pmid:24015234
- 42. Phillips DL, Koch PL. Incorporating concentration dependence in stable isotope mixing models. Oecologia. 2002;130(1):114–25. pmid:28547016
- 43. Fernandes R, Grootes P, Nadeau M-J, Nehlich O. Quantitative diet reconstruction of a Neolithic population using a Bayesian mixing model (FRUITS): The case study of Ostorf (Germany). Am J Phys Anthropol. 2015;158(2):325–40. pmid:26175210
- 44. Anastasios Z, Angeliki G, Elissavet G, Panagiota X, Constantina K, Christina P. Variations in dietary patterns in the ancient Greek colony of Abdera: insights from isotopic evidence and Bayesian modelling. Archaeol Anthropol Sci. 2025;17(7).
- 45. Bantavanou P, Ganiatsou E, Souleles A, Georgiadou A, Xanthopoulou P, Aidonis A, et al. An Investigation of enamel hypoplasia and weaning through histomorphological analysis and Bayesian Isotope Mixing Models. Dental Anthropology: A Publication of the Dental Anthropology Association. 2024;37(1):20–43.
- 46. Chinique de Armas Y, González Herrera UM, Reyes Serrano I, Buhay WM, Skelton S, Rodríguez Suárez R, et al. Multiproxy paleodietary reconstruction using stable isotopes and starch analysis: The case of the archaeological site of Playa del Mango, Granma, Cuba. Journal of Archaeological Science: Reports. 2022;46:103671.
- 47. Choy K, Yun HY, Fuller BT, Mannino MA. Enhanced dietary reconstruction of Korean prehistoric populations by combining δ13C and δ15N amino acids of bone collagen. PLoS One. 2024;19(3):e0300068. pmid:38536809
- 48. Feranec RS, Hart JP. Fish and maize: Bayesian mixing models of fourteenth- through seventeenth-century AD ancestral Wendat diets, Ontario, Canada. Sci Rep. 2019;9(1):16658. pmid:31723157
- 49. Hart JP. Human and dog Bayesian dietary mixing models using bone collagen stable isotope ratios from ancestral Iroquoian sites in southern Ontario. Sci Rep. 2023;13(1):7177. pmid:37137965
- 50.
Burgess W. Suborder Acanthuroidei. Ephippidae. In: Carpenter K, editor. The living marine resources of the Western Central Atlantic. Bony fishes part 2. 2002. p. 1375–2127.
- 51.
Carneiro M, Castro P, Tutui S, Bastos G. Micropogonias furnieri (Desmarest, 1823): Southeast stock. In: Cergole M, Ávila-da-Silva A, Rossi-Wongtschowski C, Editors. Análise das principais pescarias comerciais da região Sudeste-Sul do Brasil: dinâmica populacional das espécies em explotação. São Paulo: Instituto Oceanográfico da USP. 2005.
- 52.
Figueiredo L, Menezes N. Manual de peixes marinhos do sudeste do Brasil. São Paulo: Museu de Zoologia da USP. 1980.
- 53. Kozuch L, Fitzgerald C. A guide to identifying shark centra from southeastern archaeological sites. Southeastern Archaeology. 1989;8(2):146–57.
- 54. Wesolowski V, Ferraz Mendonça de Souza SM, Reinhard KJ, Ceccantini G. Evaluating microfossil content of dental calculus from Brazilian sambaquis. Journal of Archaeological Science. 2010;37(6):1326–38.
- 55. Brown TA, Nelson DE, Vogel JS, Southon JR. Improved Collagen Extraction by Modified Longin Method. Radiocarbon. 1988;30(2):171–7.
- 56. Cheung C, Fernández-Crespo T, Mion L, Di Giusto M, Goude G, Macdonald RA, et al. Micro-punches versus micro-slices for serial sampling of human dentine: Striking a balance between improved temporal resolution and measuring additional isotope systems. Rapid Commun Mass Spectrom. 2022;36(21):e9380. pmid:35986908
- 57. Czermak A, Fernández-Crespo T, Ditchfield PW, Lee-Thorp JA. A guide for an anatomically sensitive dentine microsampling and age-alignment approach for human teeth isotopic sequences. Am J Phys Anthropol. 2020;173(4):776–83. pmid:32779777
- 58. Szpak P, Metcalfe JZ, Macdonald RA. Best practices for calibrating and reporting stable isotope measurements in archaeology. Journal of Archaeological Science: Reports. 2017;13:609–16.
- 59. DeNiro MJ. Postmortem preservation and alteration of in vivo bone collagen isotope ratios in relation to palaeodietary reconstruction. Nature. 1985;317(6040):806–9.
- 60. Ambrose SH. Preparation and characterization of bone and tooth collagen for isotopic analysis. Journal of Archaeological Science. 1990;17(4):431–51.
- 61. Guiry E, Szpak P. Improved quality control criteria for stable carbon and nitrogen isotope measurements of ancient bone collagen. J Archaeol Sci. 2021;132:105416.
- 62. Graniero LE, Gillikin DP, Surge D, Kelemen Z, Bouillon S. Assessing δ15N values in the carbonate-bound organic matrix and periostracum of bivalve shells as environmental archives. Palaeogeography, Palaeoclimatology, Palaeoecology. 2021;564:110108.
- 63. Hellevang H, Aagaard P. Constraints on natural global atmospheric CO2 fluxes from 1860 to 2010 using a simplified explicit forward model. Sci Rep. 2015;5:17352. pmid:26611741
- 64. Shokryazdan P, Faseleh Jahromi M, Liang JB, Ramasamy K, Sieo CC, Ho YW. Effects of a Lactobacillus salivarius mixture on performance, intestinal health and serum lipids of broiler chickens. PLoS One. 2017;12(5):e0175959. pmid:28459856
- 65. AlQahtani SJ, Hector MP, Liversidge HM. Brief communication: The London atlas of human tooth development and eruption. Am J Phys Anthropol. 2010;142(3):481–90. pmid:20310064
- 66.
Silverman B. Density estimation for statistics and data analysis. Chapman and Hall. 1986.
- 67. Alves E, Macario K, Souza R, Pimenta A, Douka K, Oliveira F, et al. Radiocarbon reservoir corrections on the Brazilian coast from pre-bomb marine shells. Quaternary Geochronology. 2015;29:30–5.
- 68. Alves EQ, Macario KD, Spotorno P, Oliveira FM, Muniz MC, Fallon S, et al. Nineteenth-century expeditions and the radiocarbon marine reservoir effect on the Brazilian coast. Geochimica et Cosmochimica Acta. 2021;297:276–87.
- 69. Xia Y, Zhang J, Yu F, Zhang H, Wang T, Hu Y, et al. Breastfeeding, weaning, and dietary practices during the Western Zhou Dynasty (1122-771 BC) at Boyangcheng, Anhui Province, China. Am J Phys Anthropol. 2018;165(2):343–52. pmid:29131307
- 70. Filippini J, Pezo-Lanfranco L, Eggers S. Estudio Regional Sistemático De Treponematosis En Conchales (Sambaquis) Precolombinos De Brasil. Chungará (Arica). 2019;(ahead):0–0.
- 71.
Figuti L. Investigações arqueológicas e geofísicas dos sambaquis fluviais do vale do Ribeira de Iguape, Estado de São Paulo. 2004.
- 72. Posth C, Nakatsuka N, Lazaridis I, Skoglund P, Mallick S, Lamnidis TC, et al. Reconstructing the Deep Population History of Central and South America. Cell. 2018;175(5):1185-1197.e22. pmid:30415837
- 73. Borrero LA, Barberena R. Hunter‐Gatherer Home Ranges and Marine Resources. Current Anthropology. 2006;47(5):855–68.
- 74.
Gómez-Otero J. Isótopos estables, dieta y uso del espacio en la costa atlántica centro-septentrional y el valle inferior del río Chubut (Patagonia, Argentina). In: Morello F, Martinic M, Prieto A, Bahamonde G, Editors. Arqueología de Fuego: Patagonia levantando piedras, desenterrando huesos…y develando arcanos. Punta Arenas, Chile: Ediciones CEQUA. 2007. p. 151–61.
- 75. Politis GG, Scabuzzo C, Tykot RH. An approach to pre‐Hispanic diets in the Pampas during the Early/Middle Holocene. Intl J of Osteoarchaeology. 2009;19(2):266–80.
- 76. Mendes Cardoso J, Merencio F, Villagran X, Wesolowski V, Estevam R, Fuller BT, et al. Late shellmound occupation in southern Brazil: A multi-proxy study of the Galheta IV archaeological site. PLoS One. 2024;19(3):e0300684. pmid:38512858
- 77.
Oppitz G. Coisas que mudam: os processos de mudança nos sítios conchíferos catarinenses e um olhar isotópico sobre o caso do sítio Armação do Sul, Florianópolis/S. Brazil: University of São Paulo. 2015.
- 78.
Klökler D. The worked bone and tooth assemblage from Piaçaguera: Insights and challenges. In: Wild M, Thurber B, Rhodes S, Gates St-Pierre C, Editors. Bones at a Crossroads: Integrating worked bone research with archaeometry and social zooarchaeology. Leiden: Sidestone Press. 2021. p. 133–50.
- 79. Milton K, Dintzis F. Nitrogen-to-protein conversion factors for tropical plant samples. Association for Tropical Biology and Conservation. 1981;13(3):177–81.
- 80. Ambrose SH, Butler BM, Hanson DB, Hunter-Anderson RL, Krueger HW. Stable isotopic analysis of human diet in the Marianas Archipelago, western Pacific. Am J Phys Anthropol. 1997;104(3):343–61. pmid:9408540
- 81. Uchôa D, Francisco M. Desgastes dentários em populações brasileiras extintas. O sambaqui de Piaçaguera. Arquivos de Anatomia e Antropologia. 1980;38:147–57.
- 82. Beighton D, Adamson A, Rugg-Gunn A. Associations between dietary intake, dental caries experience and salivary bacterial levels in 12-year-old English schoolchildren. Arch Oral Biol. 1996;41(3):271–80. pmid:8735013
- 83. Campain AC, Morgan MV, Evans RW, Ugoni A, Adams GG, Conn JA, et al. Sugar-starch combinations in food and the relationship to dental caries in low-risk adolescents. Eur J Oral Sci. 2003;111(4):316–25. pmid:12887397
- 84.
Wesolowski V. Cáries, desgaste, cálculos dentários e micro-resíduos da dieta entre grupos pré-históricos do litoral norte de Santa Catarina: é possível comer amido e não ter cárie?. Brazil: Oswaldo Cruz Fondation (FIOCRUZ). 2007.
- 85. Pezo-Lanfranco L. Evidence of variability in carbohydrate consumption in prehistoric fisher-hunter-gatherers of Southeastern Brazil: Spatiotemporal trends of oral health markers. Am J Phys Anthropol. 2018;167(3):507–23. pmid:30159869
- 86. ANGULO R, LESSA G, SOUZA M. A critical review of mid- to late-Holocene sea-level fluctuations on the eastern Brazilian coastline. Quaternary Science Reviews. 2006;25(5–6):486–506.
- 87. Scheel-Ybert R, Afonso MC, Barbosa-Guimarães M, Gaspar MD, Ybert J-P. Considerações sobre o papel dos sambaquis como indicadores do nível do mar. ABEQUA. 2009;1(1).
- 88.
Bendazzoli C. O processo de formação dos sambaquis: uma leitura estratigráfica do Sítio Jabuticabeira II. Brazil: University of São Paulo. 2007.
- 89.
Gaspar MD, DeBlasis P, Fish SK, Fish PR. Sambaqui (Shell Mound) Societies of Coastal Brazil. The Handbook of South American Archaeology. Springer New York. 2008. p. 319–35. doi: https://doi.org/10.1007/978-0-387-74907-5_18
- 90. Okumura MMM, Boyadjian CHC, Eggers S. Análise da exostose do meato auditivo externo como um marcador de atividade aquática em restos esqueletais humanos da costa e do interior do Brasil. Rev Mus Arqueol Etnol. 2006;(15–16):181.
- 91.
Stabile R. Ossos do ofício: análise de marcadores de estresse ocupacional em séries esqueléticas de sambaquis da Baixada Santista-SP. Brazil: University of São Paulo. 2017.
- 92. Beaumont J, Montgomery J. The Great Irish Famine: Identifying Starvation in the Tissues of Victims Using Stable Isotope Analysis of Bone and Incremental Dentine Collagen. PLoS One. 2016;11(8):e0160065. pmid:27508412
- 93. King CL, Halcrow SE, Millard AR, Gröcke DR, Standen VG, Portilla M, et al. Let’s talk about stress, baby! Infant-feeding practices and stress in the ancient Atacama desert, Northern Chile. Am J Phys Anthropol. 2018;166(1):139–55. pmid:29355900
- 94.
Waldron T. Counting the dead: the epidemiology of skeletal populations. Chichester: John Wiley & Sons. 1994.
- 95. Plens CR. O papel dos amontoados de conchas no sambaqui fluvial. SAB. 2009;22(2):77–93.
- 96. Flensborg G, Martínez G, Tessone A. First approach to the paleodiet of hunter-gatherers through stable isotopes (δ13C and δ15N) in the eastern Pampa-Patagonia transition during the Middle Holocene. Journal of Archaeological Science: Reports. 2018;17:571–80.
- 97. Tykot RH, Falabella F, Planella MT, Aspillaga E, Sanhueza L, Becker C. Stable isotopes and archaeology in central Chile: methodological insights and interpretative problems for dietary reconstruction. Intl J of Osteoarchaeology. 2009;19(2):156–70.
- 98. Loponte D, Carbonera M. From the Atlantic coast to the lowland forests: Stable isotope analysis of the diet of forager–horticulturists in southern Brazil. Intl J of Osteoarchaeology. 2021;31(6):1237–46.
- 99. Eggers S, Petronilho CC, Brandt K, Jericó-Daminello C, Filippini J, Reinhard KJ. How does a riverine setting affect the lifestyle of shellmound builders in Brazil?. Homo. 2008;59(6):405–27. pmid:19027113
- 100. Filippini J, Eggers S. Distância biológica entre sambaquieiros fluviais (Moraes – Vale do Ribeira-SP) e construtores de sítios litorâneos (Piaçaguera e Tenório-SP e Jabuticabeira IISC). Rev Mus Arqueol Etnol. 2006;(15–16):165.
- 101. Loponte D, Mazza B. Breastfeeding and weaning in Late Holocene hunter-gatherers of the lower Paraná wetland, South America. Am J Phys Anthropol. 2021;176(3):504–20. pmid:34338320
- 102. Hawkes K, O’Connell JF, Blurton Jones NG. Hunting and Nuclear Families. Current Anthropology. 2001;42(5):681–709.
- 103. Hurtado AM, Hill K, Hurtado I, Kaplan H. Trade-Offs between female food acquisition and child care among hiwi and ache foragers. Hum Nat. 1992;3(3):185–216. pmid:24222429
- 104. Stock BC, Jackson AL, Ward EJ, Parnell AC, Phillips DL, Semmens BX. Analyzing mixing systems using a new generation of Bayesian tracer mixing models. PeerJ. 2018;6:e5096. pmid:29942712