1 INTRODUCTION

While historians and archaeologists have known about the broad patterns of the Atlantic slave trade, including those areas in Africa from which trading countries exported enslaved individuals, specific origins for skeletal remains recovered from archaeological sites in the United States can rarely be determined. A particular African cultural group is sometimes suggested by housing, foods, religious and personal artifacts, preserved historic documents, mortuary practices, and other cultural indicators if such information is available (Blakey, 1998; DeCorse, 1999; Fennell, 2011; Ogundiran & Falola, 2007; Singleton, 1995, 1999). Traditionally, features of the skull and dentition have been used to assess the ancestral origins of skeletal remains (Blakey and Rankin-Hill, 2009; Gill & Rhine, 1990; Jantz & Ousley, 2005; Spradley, 2006), but these methods do not distinguish first generation Africans from those born in North America of African parents. Even DNA is inadequate in this respect since subsequent generations of Africans born in North America may not have genetic admixture with non-African groups. Intentional dental modification has arguably been the only other means used to determine recent arrival to the Americas on the basis of skeletal remains. This practice has been noted in studies from North and Central America, and the Caribbean (Blakey, 1998; Blakey & Rankin-Hill, 2009; Handler, Corruccini, & Mutaw, 1982; Handler, 1994; Ortner, 1966; Price et al., 2012; Schroeder, Haviser, & Price, 2014; Stewart & Groome, 1968; Tiesler, 2002). However, the historic occurrence of distinctive dental modification patterns overlaps across African regions, varies through time, and has not been studied adequately to allow a definitive determination of regional origin. Furthermore, the practice has not been proven to be exclusive to first-generation Africans in the Americas (Rivero de la Calle, 1973; Roksandic, Alarie, Suárez, Huebner, & Roksandic, 2016).

This study examines stable carbon, nitrogen, and oxygen isotope values from the cosmopolitan coastal site of Elmina, Ghana in West Africa and compares these data to values from historic sites in North America. To date, isotope studies of historic period remains in North America have focused on origins and demographic factors within the Americas because no comparable African data existed from this time period. Stable isotope values from Elmina allow comparisons to be made and appear unique in some respects from isotope values of individuals from the Mid-Atlantic region of the United States. In essence, Elmina values can serve as a proxy for recent West African arrivals to the Mid-Atlantic, quantitatively distinguishing them from African Americans born in this region.

Stable isotopes help determine resource use by, and thus regional origins of, historic period North American populations, including enslaved individuals (Bruwelheide et al., 2019; France, Owsley, & Hayek, 2014; Raynor & Kennett, 2008; Ubelaker & Owsley, 2003). Carbon isotope ratios reflect regional vegetation availability to both humans and animals, and can separate region of origin only so far as certain regions tend to rely on different grain sources. Carbon is incorporated into the hydroxyapatite carbonate (δ¹³C_carbonate), as well as the collagen protein within bone and tooth dentin (δ¹³C_collagen). Values are reported in standard delta notation where δ¹³C = [(¹³C/¹²C_sample − ¹³C/¹²C_standard)/¹³C/¹²C_standard] ×1,000; units are in permil (‰), and the standard is Vienna Pee Dee Belemite. Carbon isotopes fractionate differently depending on whether a plant employs the C3 or C4 photosynthetic pathway. C3 plants (wheat, barley, rice, trees, shrubs, and temperate/cool climate grasses) show more negative δ¹³C values (approximately −33‰ to −24‰), and C4 plants (maize, millet, sorghum, sedges, sugarcane, and warm/dry climate grasses) show relatively higher values (approximately −16‰ to −10‰) (Heaton, 1999; O'Leary, 1988; Smith & Epstein, 1971). Carbon in hydroxyapatite carbonates primarily reflects the carbohydrate and lipid carbon isotope dietary input, while collagen carbon primarily reflects the protein dietary input (Ambrose & Norr, 1993; Fernandes, Nadeau, & Grootes, 2012; Jim, Ambrose, & Evershed, 2004; Krueger & Sullivan, 1984). Although carbon isotopes fractionate during incorporation into tissues, the C3/C4 pattern is still observable (Balasse, Bocherens, & Mariotti, 1999; Hedges, 2003; Kohn & Cerling, 2002; Passey et al., 2005; van der Merwe, 1982). Therefore δ¹³C_carbonate values largely indicate the types of plants and grains consumed directly by an individual, while δ¹³C_collagen values are heavily influenced by the plant and grain fodder for consumed animals.

Nitrogen isotopes, incorporated exclusively into collagen in bone and dentin, reflect the amount of protein input or marine resources in the diet. Values are reported in standard delta notation (δ¹⁵N_collagen) where the ratio of interest is ¹⁵N/¹⁴N, and the standard is atmospheric air. Both increased protein and/or marine resources result in higher δ¹⁵N_collagen values (Bocherens & Drucker, 2003; DeNiro & Epstein, 1981; Fogel, Tuross, Johnson, & Miller, 1997; Schoeninger & DeNiro, 1984) and tend to indicate localized availability of food within a region or population. Nitrogen values are not necessarily correlated with broad regional origins.

Oxygen isotope values have proven the most effective at determining region of origin within North America because there are significant differences between northern and southern areas. Oxygen isotopes are incorporated into bone, tooth enamel, and tooth dentin in carbonates within the mineral hydroxyapatite. Values are reported in standard delta notation (δ¹⁸O_carbonate) where the ratio of interest is ¹⁸O/¹⁶O, and the standard is Vienna Standard Mean Ocean Water. The primary pool of oxygen integrated into hydroxyapatite during mineralization is body water, which directly reflects oxygen isotope ratios in drinking water (Bryant & Froelich, 1995; D'Angela & Longinelli, 1990; Daux et al., 2008; Kohn, 1996; Levinson, Luz, & Kolodny, 1987; Longinelli, 1984; Luz & Kolodny, 1985; Luz, Kolodny, & Horowitz, 1984). Prior to the 20th-century, drinking water in North America and Africa was obtained largely from local meteoric water with oxygen isotope values correlated to latitude and general region of origin (Bowen & Wilkinson, 2002; Dutton, Wilkinson, Welker, Bowen, & Lohmann, 2005; Kendall & Coplen, 2001; Landwehr, Coplen, & Stewart, 2014). Oxygen isotopes will fractionate during the incorporation into hydroxyapatite carbonates (i.e., δ¹⁸O_carbonate), but this fractionation is expected to be constant across the human species. As such, observed differences between oxygen isotopes in North America versus Africa will be maintained in archaeological remains. Some overlap in δ¹⁸O meteoric water values exists between North America and Africa, specifically between the southern United States and northern Africa. The δ¹⁸O values of North American meteoric water range from approximately −21 to −1‰. Values in Africa range from approximately −11 to +4‰ (Figure 1). While this may be a complicating factor in interpreting data from archaeologically recovered African American remains from southern locales and the Caribbean, isotope values from the Mid-Atlantic should be distinct.

2 MATERIALS

The single site of Elmina holds four centuries of burials representing both free and enslaved Africans of diverse heritage. Effective comparison with North American remains utilizes eleven different sites in Mid-Atlantic States, and one in New Mexico (a battlefield burial for Confederate soldiers from Texas; Table 1, Supporting information Table S1).

3 METHODS

Both bone and tooth samples were included in subsequent analyses. Selection of material was based largely on sample availability. When both bone and teeth were available from an individual, a tooth was selected for this study. Approximately 24% of the remains had available teeth; the majority of teeth are from the Glorieta Pass site (Table S1). Teeth do not remodel after mineralization and provide a better assessment of childhood origins compared to bone which remodels throughout life. The inclusion of bone from adult individuals in the comparative North American population reduces the visibility of recently arrived African individuals in the data set. However, the goal of this study is to provide a comparative data set that reflects a population of people that were most likely born in North America or spent significant time there. If a set of mixed North American bone and teeth show isotopic distinction from the Elmina individuals, it is likely that recent arrivals from the latter (and other African localities) will appear as outliers compared to the former.

Solid bone and tooth dentin samples for collagen isotope and elemental analyses were obtained by coring with a rotary tool or extraction with pliers. Powdered bone, dentin, and enamel samples for carbonate isotope and Fourier transform infrared (FTIR) spectroscopy analyses were obtained by crushing with a mortar and pestle or drilling with a rotary tool. Collagen and carbonates were extracted from bone, dentin, and enamel according to methods detailed in France et al. (2014). Briefly summarized, collagen in bone and dentin was extracted via a standard acid–base–acid method including demineralization in cold hydrochloric acid, humic and fulvic acid removal with sodium hydroxide, denaturing the collagen in weak hot hydrochloric acid, and isolation of collagen via lyophilization. Carbonates were isolated by removal of organic material with sodium hypochlorite followed by buffered acetic acid to remove secondary diagenetic carbonates.

All samples were analyzed on a Thermo Delta V Advantage mass spectrometer (Thermo Fisher Scientific, Waltham, MA) in continuous flow mode at the Smithsonian MCI Stable Isotope Mass Spectrometry Laboratory. Collagen samples were weighed into tin cups and combusted on a Costech 4010 Elemental Analyzer (Costech Analytical Technologies Inc., Valencia, CA). The purified N₂ and CO₂ gases were transferred to the mass spectrometer via a Conflo IV interface and measured for δ¹⁵N_collagen and δ¹³C_collagen values. Raw values were corrected to an acetanilide house reference material and Urea-UIN3 (Schimmelmann et al., 2009), both calibrated to USGS40 and USGS41 international reference materials. Weight % N and weight % C values were calibrated using the acetanilide. Carbonate samples were weighed into exetainer vials and flushed with pure helium. Samples were acidified with concentrated phosphoric acid (SG > 1.92) for 24 hr at 25°C. The released CO₂ was purified and transferred to the mass spectrometer via a Thermo GasBench interface and measured for δ¹³C_carbonate and δ¹⁸O_carbonate values. Raw values were corrected to LSVEC and NBS-19 international reference materials. Errors for δ¹⁵N_collagen, δ¹³C_collagen, δ¹³C_carbonate, and δ¹⁸O_carbonate are ±0.2‰ (1σ).

Raw bone powders were analyzed using attenuated total reflectance FTIR spectroscopy (ATR-FTIR) on a Thermo Nicolet 6700 FTIR (Thermo Fisher Scientific, Waltham, MA) with Golden Gate ATR (diamond crystal, single bounce, 45°) equipped with a DTGS detector. Spectra were collected for 128 scans from 450 to 4,000 cm⁻¹ with a resolution of 4 cm⁻¹. All baseline corrections and ratio calculations were automated via TQAnalyst EZ version 8. Measured parameters include phosphate peak (ѵ₄) heights at 565 and 605 cm⁻¹ and the associated valley at ~590 cm⁻¹, phosphate peak (ѵ₁) height at ~960 cm⁻¹, phosphate peak (ѵ₃) height at 1,035 cm⁻¹, carbonate peak (ѵ₃) heights at 1,415 and 1,455 cm⁻¹, and ѵ₁PO₄ position. Using these parameters, the following peak height ratios were calculated: infrared splitting factor (IRSF) ([565 + 605]/590), carbonate/carbonate (1,455/1,415), and carbonate/phosphate (1,415/1,035).

Potential postmortem diagenetic alteration of collagen isotope values was examined using collagen yields obtained during extraction and elemental yields obtained during mass spectrometry analyses. Collagen extracted from well-preserved bones and teeth should constitute ~2–20% whole bone/dentin weight, show a weight % N of ~10–15%, and an atomic C:N ratio of 2.8–3.6 (Ambrose, 1990; DeNiro, 1985; Jorkov, Heinemeier, & Lynnerup, 2007; McNulty et al., 2002). Potential diagenetic alteration of bone hydroxyapatite carbonate isotope values was examined using the calculated peak height ratios from ATR-FTIR spectra. Based on previous ATR-FTIR studies and suggested conversions from the more common transmission FTIR methods (Beasley, Bartelink, Taylor, & Miller, 2014; Garvie-Lok, Varney, & Katzenberg, 2004; Lebon et al., 2010, 2011; Snoeck, Lee-Thorp, & Schulting, 2014; Thompson, Gauthier, & Islam, 2009; Thompson, Islam, Piduru, & Marcel, 2011; Wright & Schwarcz, 1996), well-preserved hydroxyapatite should have an IRSF of <4.4, carbonate/carbonate ratio (C/C) of ~0.9, carbonate/phosphate ratio (C/P) of ~0.3, and a ѵ₁PO₄ position <962.5 cm⁻¹. Collagen and hydroxyapatite samples yielding these values were considered well preserved and included in further analyses. A paucity of FTIR data in the literature precludes direct analysis of tooth enamel as the different mineralization process of enamel is expected to produce different peak height ratios than those described above. However, enamel mineralization is generally more dense than that of bone and dentin, and the former is often considered more resistant to diagenesis than the latter. Therefore, tooth enamel samples were considered well preserved if the corresponding dentin met the collagen quality criteria outlined above.

Isotope values were compared across all sites using nonparametric Mann–Whitney tests due to small sample sizes and non-Gaussian distributions for some sites. For tooth samples, dentin data were used for all δ¹⁵N_collagen and δ¹³C_collagen comparisons; enamel data were used for all δ¹⁸O_carbonate and δ¹³C_carbonate comparisons unless only dentin data were available for a particular tooth.

4 RESULTS

Isotope data for well-preserved samples are shown in Table 2. Only samples meeting the preservation criteria outlined above are included in further analyses and interpretations; all results from diagenesis testing are included in Table S2. The δ¹³C_carbonate values from Elmina show an average of −5.9‰ (±1.2, 1σ) with a range of −8.9 to −3.9‰ (Figure 2). The study's African Americans show a slightly more negative δ¹³C_carbonate average of −7.1‰ (±2.4, 1σ) with a greater range of −11.5 to −2.8‰. Euro-Americans show a considerably more negative δ¹³C_carbonate average of −8.9‰ (±2.1, 1σ) with the widest range of −14.0 to −4.4‰. If the southwestern site of Glorieta Pass is removed, the remaining Mid-Atlantic Euro-Americans show a lower δ¹³C_carbonate average of −9.4‰ (±1.7, 1σ) and a slightly smaller range of −13.3 to −6.0‰.

The δ¹³C_collagen values show the same pattern as the δ¹³C_carbonate values. Elmina has an average δ¹³C_collagen value of −10.3‰ (±1.6, 1σ) and a range of −13.5 to −7.7‰ (Figure 3). African Americans have a slightly more negative δ¹³C_collagen average of −12.0‰ (±2.6, 1σ) with a greater range of −17.0 to −8.2‰. A considerably more negative average of −14.2‰ (±2.0, 1σ) with a range of −19.7 to −9.2‰ is noted for Euro-Americans. Removing Glorieta Pass results in a Mid-Atlantic Euro-American δ¹³C_collagen average of −14.5‰ (±1.7, 1σ) and a slightly smaller range of −17.8 to −10.5‰.

There is less variation between regions in δ¹⁵N_collagen values (Figure 3). Elmina has an average δ¹⁵N_collagen value of +11.9‰ (±1.1, 1σ) with a range of +8.3 to +14.1‰. Euro-Americans and African Americans show slightly lower averages of +10.7‰ (±0.9, 1σ) and +10.4‰ (±0.7, 1σ), respectively. The range of δ¹⁵N_collagen was only slightly different among groups with Euro-Americans and African American values extending from +8.1 to +13.2‰ and +8.8 to +11.9‰, respectively. Removing Glorieta Pass results in a similar δ¹⁵N_collagen average value for Mid-Atlantic Euro-Americans of +10.7‰ (±1.0, 1σ) with a range of +8.7 to +13.2‰.

The δ¹⁸O_carbonate values from Elmina show an average of +27.6‰ (±1.0, 1σ) with a range of +23.3 to +28.9‰ (Figure 2); most Elmina δ¹⁸O_carbonate values fall in the range of +26.1 to +28.9‰. Both Euro-Americans and African Americans have lower δ¹⁸O_carbonate averages of +26.0‰ (±2.6, 1σ) and +25.3‰ (±1.2, 1σ), respectively. The African Americans have a relatively smaller range of +22.0 to +27.7‰, while the Euro-Americans show the greatest overall range of +15.9 to +34.0‰. If Texans from Glorieta Pass are removed, remaining Mid-Atlantic Euro-Americans show a lower δ¹⁸O_carbonate average of +24.5‰ (±2.3, 1σ) with the same range of +15.9 to +34.0‰.

Site-to-site comparison of isotope averages using Mann–Whitney tests are listed in Table 3. Sites are considered statistically different if p<.05. The δ¹³C_collagen and δ¹³C_carbonate values from Elmina are consistently distinct from Euro-Americans in the study with the exception of the Kincheloe Cemetery. The Elmina δ¹³C_collagen values are similar to three of the five African American sites (i.e., Robinson Cemetery, A.P. Hill, and Pettus). The Elmina δ¹³C_carbonate values are also similar to three of the five African American sites (i.e., Robinson Cemetery, A.P. Hill, and Parkway Gravel). The Elmina δ¹⁵N_collagen values are statistically distinct from all North American sites except the Foscue Plantation.

The δ¹⁸O_carbonate values from Elmina are distinct from every North American site in the Mid-Atlantic. The δ¹⁸O_carbonate values from Texans buried at Glorieta Pass in New Mexico are not statistically different from Elmina, showing there are similarities between certain areas of Africa and southern North America. However, the Glorieta Pass values are also notably distinct from all Mid-Atlantic sites with the exception of Foscue Plantation (although Glorieta and Foscue are statistically distinct at p < .08). In the few North American sites where bone and tooth were analyzed concurrently, the teeth tend to be equal to the bone δ¹⁸O_carbonate values, or slightly higher. However, the differences could not be robustly tested given that all sites with both elements consisted of a majority of either bone or tooth, with ≤3 samples of the alternate type.

5 DISCUSSION

Current knowledge of North American slave origins is limited to a broad understanding of the Atlantic slave trade with exceptions where historical records and archaeological evidence exist. Once enslaved persons arrived in North America, they may have been moved to several different locations, primarily within the mid- or southern colonies or states, making it difficult to track origins. Despite the 1807 Act Prohibiting the Importation of Slaves (effective in 1808), populations of enslaved African Americans increased as children born in North America were incorporated into the existing system of enslavement. New generations of enslaved individuals are difficult to distinguish from their first generation parents. Previous studies have used oxygen, strontium, carbon, and nitrogen isotopes from individuals in Central America, Brazil, Barbados, and the Dutch Caribbean to confirm recent arrival of enslaved persons from Africa and infer dietary resources (Bastos et al., 2016; Laffoon, Espersen, & Mickleburgh, 2018; Laffoon, Valcárcel Rojas, & Hofman, 2013; Price et al., 2012; Schroeder et al., 2014; Schroeder, O'Connell, Evans, Shuler, & Hedges, 2009). These studies focused on origins and demographic factors within the Americas and were unable to compare enslaved individuals directly to African sites due to the lack of comparable isotope data on historic human remains from West Africa. Stable carbon, nitrogen, and oxygen isotope values from Elmina, Ghana in this study are a promising step toward identifying Africans newly arrived to the Mid-Atlantic.

Carbon and nitrogen isotopes in the Elmina population would have been influenced by local food sources. A wide variety of domesticated animals and crops were utilized at Elmina, including millet, sorghum, multiple species of yams and, by the 17th-century, a variety of American cultigens (DeCorse, 2001). Maize, a C4 dietary staple in the Mid-Atlantic, was not native to Elmina or Africa, but was brought there from the Americas. By the 17th-century, cornbread and kenkey, a staple dish made from fermented ground corn, were common. Fishing was a major part of the local economy and marine foods were a significant dietary component at Elmina (DeCorse, 2001).

The Elmina isotope results show higher δ¹³C_carbonate and δ¹³C_collagen values than Mid-Atlantic Euro-Americans. They show some overlap with the values of southern Euro-Americans and African Americans (Figures 2 and 3). This is likely due to the greater prevalence of C4 vegetation in southern North American regions and heavy reliance on maize, especially in the diets of enslaved individuals. The prevalence of C4 grains in African American diets is well documented in the historic and archaeological record (Bowes, 2011; Bowes & Trigg, 2012; France et al., 2014; Franklin, 2001; Mrozowski, Franklin, & Hunt, 2008). Bruwelheide et al. (2019) observed the predominance of C4-based diets in African Americans from Mid-Atlantic sites in Virginia and eastern Maryland. Despite the observed overlap, Elmina results trend toward higher values supporting the historical evidence of significant C4 grains in the coastal Ghanaian diet by the early 17th-century, particularly maize, millet, and sorghum. The observed similarity between Elmina individuals and African Americans suggests the latter group may have had diets limited to certain resources, perhaps due to food availability or economic resources. Additionally, the values may reflect retained cultural preferences for certain foods and cooking styles.

Nitrogen isotope values from Elmina are somewhat distinct from North American sites, but this isotope system does not serve as a good indicator of regional origin. The relatively high δ¹⁵N_collagen values in Elmina support the archaeological evidence for reliance on marine dietary sources. North American sites in this study do not include coastal locations where marine dietary input would be equally high. Without this direct comparison, one cannot conclude that nitrogen isotopes are regionally distinct between Africa and all of North America. Rather, nitrogen isotopes are controlled more likely by local food sources, and to some extent, social class (France et al., 2014).

The results of this study indicate that oxygen isotope values are particularly useful for distinguishing Africans from North Americans. The Elmina population, which represents a mix of West African people, shows oxygen isotope values that are distinct from a majority of North American values in this study. Remains from Glorieta Pass (Figure 2), a military company mustered out of Texas, are the exception. As one of the southern-most points in the United States, this area shows some of the most positive North American δ¹⁸O values in meteoric water. As noted, Elmina essentially represents the lowest expected oxygen isotope values in exported African slaves whereby δ¹⁸O values for other exported slave populations are expected to be higher than or approximately equal to Elmina values. Individuals living for years in the southern-most regions of North America may show some oxygen isotope overlap with recently arrived Africans from Senegambia, Sierra Leone, Liberia, Ghana or the Bight of Benin (Figure 1). However, recently arrived Africans from further south (i.e., Bight of Biafra, western Central Africa, and Southeast Africa) should be isotopically distinct from North American populations. Oxygen isotopes in southeastern regions and all central and northern regions of North America potentially could be used to identify recent arrivals from Africa.

Of particular interest are North American sites with enslaved African Americans from the mid- to late-1700s: A.P. Hill, Pettus, and Robinson Cemetery. Values from the Pettus and Robinson skeletal remains show average δ¹⁸O_carbonate statistical differences from values of several other sites, including both Euro-Americans and African Americans, while the average value for A.P. Hill individuals is statistically similar to almost all North American sites. Most individuals from these sites show δ¹⁸O_carbonate values ~2‰ less than the Elmina average. Similarly, individuals from Parkway Gravel, representing slaves or former slaves in the later 1800s, and Elmina show no overlapping δ¹⁸O_carbonate values. Rather, Parkway Gravel remains show δ¹⁸O_carbonate values typical of its North American location. This observation suggests that these individuals were born in North America or spent the majority of their lives there. It is notable that the two African American sites containing both tooth and bone data (Parkway Gravel and Robinson Cemetery) show similar δ¹⁸O_carbonate values between the two tissue types. This supports the idea that these adult bone isotope values reflect a lifetime spent in North America, rather than values integrated through a forced migration. Teeth from African Americans were limited in this study, but this small subset of data lends credence to the idea that slaves were often born into the system by the early 1800s.

Three notable outliers with relatively higher δ¹⁸O_carbonate values were observed among African Americans in this study: two individuals from Pettus (44JC33-PETTUS-191 and 44JC33-PETTUS-253) and one from A.P. Hill (44CEAPHILL-VAOCME-2). These three individuals show δ¹⁸O_carbonate values that are more similar to the majority of Elmina values (Figure 2). These outliers may have been recent arrivals to the North American Mid-Atlantic from Africa, or alternatively, were born in the furthest southern states (i.e., Florida, Alabama, Georgia, or Louisiana) or the Caribbean. These two sites contain some of the earlier remains likely from the late 1700s (Pettus) or earliest 1800s (A.P. Hill) when the slave trade was still quite active. Others from these sites appear to have been in North America for most of their lives.

This preliminary case study demonstrates how an isotopic approach potentially can explore slavery as a self-sustaining system in North America, identifying a reduced influx of new individuals in the mid- to late-1800s. To comprehensively test this idea, additional isotope testing of African and North American archaeological remains is required, coupled with detailed analysis of age to facilitate interpretation of individual residency duration. While some overlap between African and far southern North American sites may become apparent in future studies, these data offer insight for identifying regional origins and numerical representation of African peoples in the North American diaspora.

While this study did not directly test isotope values from Central American and Caribbean individuals, previous research provides an isotopic comparison for this region which was active in the slave trade to North America. Local δ¹⁸O_carbonate values from human remains in these regions ranges from approximately +25 to +29‰, with most values in the range of about +27 to +29‰ (Laffoon et al., 2013, 2018; Price et al., 2010, 2012; Schroeder et al., 2009). This is significantly higher than the North American values in this study, but it does overlap with the Elmina values. This implies that oxygen isotope values may be able to distinguish recent arrival in the mid-Atlantic region, but distinction between a Caribbean origin and a more northern African origin may be confounded without additional data, such as strontium isotopes. However, origin from more central or southern Africa may be distinguishable, although additional data sets are necessary to test this. Both δ¹³C_carbonate and δ¹³C_collagen values from Central American and Caribbean locals and recent migrants show a range of values including dominantly C3 consumption to dominantly C4 consumption (Bastos et al., 2016; Laffoon et al., 2013, 2018; Price et al., 2012; Schroeder et al., 2009). The relative δ¹³C relationship between local individuals and confirmed migrants from Africa (i.e., which group consumed more C4 plants) varies, likely depending on the dominant food base in the region of birth. With additional future analyses from African archaeological remains, the δ¹³C and δ¹⁸O values may be coupled to provide a more precise region of origin and distinguish between the African and Caribbean slave trading routes, both of which were active until the legal cessation of this practice.

6 CONCLUSIONS

The African diaspora contributed significantly to the cultural and biological make-up of the larger colonial population of the United States of America, with estimates of the number of enslaved persons brought to North America from Africa varying from approximately 208,000 to 370,000 (Curtin, 1969; Lovejoy, 1989; Voyages: The Trans-Atlantic Slave Trade Database, 2017). Exact numbers and knowledge of the precise origins of these individuals are limited to incomplete historical and bioarchaeological records. The unique stable isotope profile of West Africans presented in this study facilitates the identification of recent arrivals in a number of archaeological Mid-Atlantic sites, potentially enabling increased understanding of the slave trade and in turn, African American population dynamics.

This study presents a rare direct comparison between 18th- and 19th-century North Americans and Africans from the relatively cosmopolitan site of Elmina, Ghana to determine whether oxygen isotopes can be used to identify recently arrived Africans, and thus shed light on patterns of arrival into North America. Isotope values of individuals from Elmina serve as a proxy for isotope values of individuals exported from Africa since the values from Elmina are distinct from most North American populations, with the exception of extremely southern locations and the Caribbean. While results are considered preliminary due to available sample sizes, comparison to North American slave population samples from specific archaeological sites suggests that by the mid-1700s the North American slavery system may have been largely self-sustaining, with the majority of individuals born and raised in North America. Further broad comparison of carbon and nitrogen isotope values confirms their utility in discerning dietary resources, specifically a high influence of C4 vegetation and marine resources in Elmina.

Extrapolation of these data to the larger African continent may be possible with additional contributions and future analyses. With the current data it is not possible to discern precisely where in Africa the individuals living in Elmina originated, although it is known that this site attracted immigrants from disparate locations within a restricted geographical region based on the homogeneity of δ¹⁸O_carbonate values with few outliers. Stable isotope data from African archaeological remains are limited for 17th–19th-century individuals, but this study demonstrates the potentially useful applications should such remains become available. With a more comprehensive data set from North America as well, it may be possible to distinguish African regional origins of North American enslaved individuals, which would provide a new window into the cultural components carried to the new world via the Atlantic slave trade.