Teotihuacan was by far the largest urban center of its time in the Western Hemisphere. Situated in a side valley of the Valley of Mexico, with a population of about 100,000 and covering some 20 km2, it dominated central Mexico during the Classic period (200–600 AD) and had a strong influence on the rest of Mesoamerica (Cowgill Reference Cowgill2015; Millon Reference Millon1973, Reference Millon and Sabloff1981; Smith et al. Reference Smith, Abhishek Chatterjee, Huster and Forest2019). The size of the city and its rapid growth raise the question of how it gathered, and retained, such a large population. Archaeologists have come to realize that ancient urban centers did not usually have stable populations, either in terms of numbers or composition (Smith Reference Smith2014; Storey Reference Storey1992). Individuals and groups were frequently mobile, just as they are in modern settings. The analysis of stable isotopes, particularly strontium and oxygen, is now allowing us to verify and track some of that movement into (and out of) Teotihuacan (e.g., Price et al. Reference Price, Manzanilla and Middleton2000; White et al. Reference White, Spence, Stuart-Williams and Schwarcz1998).
Not surprisingly, the developing picture is complex. For the most part it is based on the large multifamily apartment compounds that housed the majority of Teotihuacan's residents (Manzanilla Reference Manzanilla1993; Millon Reference Millon and Sabloff1981; Séjourné Reference Séjourné1966). However, this leaves a significant gap in our understanding: we lack information on the nature, extent, and demographic history of insubstantial architecture in the city. Robertson (Reference Robertson2008) states that there are some 800 insubstantial sites in Teotihuacan, perhaps holding up to 15% of the city's population. To remedy this shortcoming, Cabrera Cortés in 2004 excavated the San José 520 site, a small occupation area near the city's edge that lacked the large apartment compounds that characterized most of Teotihuacan (Cabrera Cortés Reference Cabrera Cortés2005, Reference Cabrera Cortés2011). Our concern here is to learn what we can about the people who occupied San José 520, their origins, and their movements through time. To this end, we conducted stable isotope analyses of the recovered human skeletal materials.
San José 520
San José 520, about 1 ha in size, is located some 500 m beyond the southeast edge of the city (Figure 1). Among other features, the excavation uncovered the remnants of Architectural Feature 2 (AF2), a single-room structure about 2.5 m on a side. Probably the residence of a single family, AF2 had a meager stone foundation with upper walls of adobe or perishable materials. The associated ceramics and radiocarbon dating indicate its occupation through nearly three centuries, including the Early Tlamimilolpa (170–250 AD), Late Tlamimilolpa (250–350 AD), and Early Xolalpan (350–450 AD) phases. An alternative chronology would reduce this span to 220 years from 200 to 420 AD (Manzanilla et al. Reference Manzanilla, Mejía, Jiménez, Schaaf, Lailson, Solís, Morales, Cienfuegos and Manzanilla2012:457). The lower proportion of Xolalpan ceramics associated with the structure suggests that its use might not have extended through the latter part of the Xolalpan period.
Map of Teotihuacan region (map by Ian Robertson).
Three burial features are associated with AF2. Data on these features are presented in Cabrera Cortés (Reference Cabrera Cortés2011), Driscoll (Reference Driscoll2005), and Nado and Cruz (Reference Nado and Cruz2006). Burial 1, located about 2 m south of AF2, is the single burial of an adult of about 30–35 years, individual A. Poor preservation prevents an assessment of sex and also makes it difficult to determine the mode of burial. Some skeletal elements appear articulated, but a number of elements were either absent or too deteriorated for recognition. A large number of ceramic vessels accompanied the burial and indicate a date in the Tlamimilolpa period (Cabrera Cortés Reference Cabrera Cortés2011:236–238).
Burial 2 was in a 75 cm deep pit in the northwest part of the AF2 residence. The pit contained artifacts and the skeletal remains of a number of individuals, deposited in multiple interment episodes over the Early Tlamimilolpa–Early Xolalpan span (ca.170–450 AD). Skeletal analyses indicate that eight adults and one subadult were present (Cabrera Cortés Reference Cabrera Cortés2011:238; Nado and Cruz Reference Nado and Cruz2006). Minimum number of elements (MNE) counts show seven adults, based on seven left humeri and seven left femora.
The repeated entries into the feature had resulted in the displacement and mixture of earlier burials, so that the bones of any particular individual may be scattered among two or three of the five definable clusters (“lots”) in the feature. Many of the skeletal elements cannot be assigned to any particular individual. The final individual to be placed in Burial 2, an adult woman here designated individual F (lot 5), was still complete and fully articulated in a seated position. A ceramic vessel placed with her is of a style known from the Xolalpan period (Cabrera Cortés Reference Cabrera Cortés2011:Figure 6.13; see Rattray Reference Rattray2001:219, 251).
The skeletal analyses by Driscoll (Reference Driscoll2005) and by Nado and Cruz (Reference Nado and Cruz2006) allowed the partial reconstruction of five of the adults in Burial 2: individuals B–F (Table 1). However, some of the sampled teeth and bones cannot be assigned to a particular adult. In addition, with the exception of individual F, the adults cannot be organized in a sequence of deposition in the feature.
Adult Data.
The ninth individual in Burial 2 is a subadult, represented only by a few scattered and fragmentary elements. The sizes of two complete elements, a tibia and a first metacarpal, indicate a neonate. Another neonate, Burial 3, had been placed in a bowl at the southeast edge of AF2. None of the subadult bones were sampled for isotope analysis.
Given the even sex balance among the adult burials (Table 1) and assuming the continuous occupation of AF2 by lineally related families (as suggested by the repeated use of Burial 2), it seems likely that the adults buried there were the couples in residence over the two to three centuries of the structure's use. However, the life table constructed by Storey (Reference Storey1992:159, Table 6-1) for the Tlajinga 33 site suggests that the burials found at AF2 are not numerous enough to account for all those who would have lived and died there over the two to three centuries of occupation. Unfortunately, this sort of mortuary underrepresentation is a constant feature of Mesoamerican sites (Spence and White Reference Spence and White2009:237). In particular, the number of subadults who died at birth or in their first year at AF2 should be considerably higher than the two neonates actually found there (Storey Reference Storey1992:Table 6-1). Some older children would also be expected but, as becomes apparent later in the article, they probably lived, and sometimes died, elsewhere.
Materials and Methods
Twenty-nine bone and dental enamel samples, all from adults, were selected for analysis of bioapatite phosphate oxygen-isotope composition at the Laboratory for Stable Isotope Science at the University of Western Ontario. The enamel samples included the entire crown, thus providing a homogenized measure of the tooth's δ18Op over the 1.5–4.5 years of crown development. Four enamel and bone samples are from individual A of Burial 1, and 20 are from individuals B–F of Burial 2. The remaining five samples are also from Burial 2 but cannot be assigned to particular individuals. An additional five Burial 2 samples, all taken from adult ribs (one from each of the five lots), were analyzed in a separate project at Northern Arizona University (Nado Reference Nado2017; Nado et al. Reference Nado, Zolotova and Knudson2017). These oxygen-isotope compositions were measured for bone structural carbonate, not phosphate, and the results were then converted to phosphate-equivalent oxygen-isotope compositions (Nado et al. Reference Nado, Zolotova and Knudson2017:122–123). With the exception of the sample from lot 5 (individual F), none of these rib samples can be associated with a particular Burial 2 individual. Given that all of the lots except lot 5 contained elements from more than one person, it is even possible that some rib samples may be duplicates from the same individual.
Phosphate oxygen was isolated from samples as silver phosphate (Ag3PO4) following Crowson and Showers (Reference Crowson and Showers1991). About 30 mg of Ag3PO4 were then placed into nickel reaction vessels under dry nitrogen gas. Samples were then heated in vacuo for 2 hours at 300°C, followed by the addition of bromine pentafluoride (BrF5) and reaction at 600°C for 18 hours, following Clayton and Mayeda (Reference Clayton and Mayeda1963), Crowson and Showers (Reference Crowson and Showers1991), Stuart-Williams and Schwarcz (Reference Stuart-Williams and Schwarcz1995), and Tudge (Reference Tudge1960). The oxygen gas that was liberated was converted completely to CO2 by reaction with red-hot carbon and was cryogenically transferred to evacuated gas tubes for analysis using a Micromass Optima dual-inlet stable-isotope-ratio mass spectrometer. Conventional δ-notation was used to report all oxygen-isotope results for the phosphate oxygen (δ18Op) relative to Vienna Standard Mean Ocean Water (VSMOW). Reproducibility was normally better than ±0.2‰.
The δ18O of most biomaterials, including bioapatite in bone and enamel, is derived from the oxygen-isotope composition of environmental (drinking) water (Longinelli Reference Longinelli1984; Luz and Kolodny Reference Luz and Kolodny1985). The δ18O of water is influenced by both geography and climate, including variables such as distance from the sea, latitude, elevation, temperature, and humidity (Stuart-Williams et al. Reference Stuart-Williams, Schwarcz, White and Spence1996; White et al. Reference White, Spence, Longstaffe and Law2000). Combinations of these variables generally result in different δ18O values for different bodies of water. As skeletal and dental mineral develop, the oxygen-isotope composition of the drinking water is incorporated, with some modification, through body water into the bioapatite of the bone and enamel (Longinelli Reference Longinelli1984; Luz et al. Reference Luz, Kolodny and Horowitz1984). Although a possible natural offset between enamel and bone oxygen-isotope compositions could potentially limit the usefulness of such data in tracing movement during an individual's lifetime, such effects have been shown elsewhere to be limited (Luz and Kolodny Reference Luz and Kolodny1985:32; Webb et al. Reference Webb, White and Longstaffe2014;106; Wright and Schwarcz Reference Wright and Schwarcz1998:14–15). Webb and coauthors (Reference Webb, White and Longstaffe2014:101), for example, report a median enamel-bone phosphate oxygen-isotope offset of −0.1‰ for a Maya sample suite (see the later discussion).
Bone remodels throughout life and therefore reflects the last 10 or so years of life. Enamel oxygen-isotope compositions reflect conditions only over the 1.5–4.5 years of formation of the crown. Because the crowns of different tooth categories develop at different times during an individual's early lifespan, the comparison of the δ18O of different teeth in a single dental arcade can help track that individual's movements and changing conditions through time (e.g., White et al. Reference White, Spence, Longstaffe and Law2000). However, this requires that we know the timing of crown development for each tooth category. For the San José 520 sample, we adapted the data of the London Atlas of Tooth Development and Eruption (AlQahtani et al. Reference AlQahtani, Hector and Liversidge2010). In particular, we used the median age columns presented by AlQahtani and colleagues in their Tables 3–8, defining crown formation as the period spanning the “cusp outline complete” (Coc) to “crown complete” (Crc) stages (Table 2). In the few cases where they do not provide ages for those stages, ages were extrapolated from those assigned to the preceding and subsequent stages.
Dental Development Ages.
Note: Adapted from AlQahtani et al. (Reference AlQahtani, Hector and Liversidge2010:Tables 3–8, median columns).
a Cusp outline complete (Coc) to crown complete (Crc).
b 0.7‰ was subtracted from the measured δ18Op to compensate for breastfeeding.
c 0.35‰ was subtracted from the measured δ18Op to compensate for breastfeeding.
An isotopic correction was applied to the earlier developing teeth to compensate for the trophic level increase associated with breastfeeding. Breastfeeding in Mesoamerican societies often continued into the third or fourth year of life (White, Spence, et al. Reference Spence, White, Longstaffe, Rattray, Law, Gallut and Soto2004:393). Given the varying times of crown formation (Table 2), we decided to reduce the δ18Op of the maxillary central incisors (UI1), the mandibular central and lateral incisors (LI1, LI2), and the first molars (UM1, LM1) by 0.7‰ (White, Spence, et al. Reference Spence, White, Longstaffe, Rattray, Law, Gallut and Soto2004:543, 545; White, Storey, et al. Reference White, Spence, Longstaffe and Law2004:177–178; see also Britton et al. Reference Britton, Fuller, Tütken, Mays and Richards2015:231; Wright and Schwarcz Reference Wright and Schwarcz1998:13). Because the maxillary lateral incisors (UI2) and the canines (UC, LC) develop somewhat later (Table 2), we reduced their δ18Op by only 0.35‰. The measured δ18Op of the later forming premolars, second molars, and third molars were left unchanged.
The samples were tested for diagenetic recrystallization using Fourier transform infrared spectroscopy (FTIR; see Webb et al. Reference Webb, White and Longstaffe2014 and references therein). The calculated crystallinity indices, with one exception, fit within the range expected for unaltered bone and enamel. The exception occurred for sample 28, but its value fell below, rather than above, the expected range for apatite-retaining original oxygen-isotope signatures. The significance of a low value is unclear, so the sample was retained in the analysis.
Results
Distribution of Oxygen-Isotope Compositions
There are two distinct clusters of δ18Op, one for bone from +14.2 to +14.7‰ (n = 12; mean and s: +14.4 ± 0.2‰) and one for enamel from +16.2 to +16.8‰ (n = 13; mean and s: +16.6 ± 0.2‰; Figure 2). Between these clusters there are a number of δ18Op whose affiliations are less clear. We believe that the two clusters identified here represent residence in two distinct geographic regions, given the 2.2‰ difference between their means (see Webb et al. Reference Webb, White and Longstaffe2014). The +14.2 to +14.7‰ cluster strongly suggests residence in the Teotihuacan region, which has a range of +14.0 to +16.0‰ (White, Storey, et al. Reference White, Spence, Longstaffe and Law2004). A number of the δ18Op values from Tlajinga 33, a Teotihuacan apartment compound 2.4 km to the northwest of San José 520 (Figure 1), fall in the +14.2 to +14.7‰ range (White, Storey, et al. Reference White, Spence, Longstaffe and Law2004:Table 1). The location of the region producing the cluster of oxygen-isotope data for the enamel is less clear. For want of a better term we refer to it as the “relocation region.”
Oxygen-isotope compositions from San José 520 burials. Additional bone data compiled from Nado and colleagues (Reference Nado, Zolotova and Knudson2017).
Some of the ambiguous oxygen-isotope compositions can be tied to one or another of the two regions by examining the other δ18Op values for those individuals. The +15.2‰ bone value, for example, is from a tibia of individual B (Table 3). Three other bones of individual B have δ18Op of +14.3 to +14.5‰, so the tibia composition may just represent part of the bone cluster's range of variation, and thus residence in the Teotihuacan region. However, the bone δ18Op of +13.3, +15.9, and +16.0‰ are from skeletal elements that cannot be assigned to a particular individual. As such, their geographic reference remains uncertain, although the +15.9‰ and +16.0‰ values are within (but at the high end of) the Teotihuacan range. These last two values are from two of the five ribs analyzed by Nado and colleagues (Reference Nado, Zolotova and Knudson2017). The three other rib δ18Op values are all within the +14.2 to +14.7‰ Teotihuacan cluster (Table 3). Nado (Reference Nado, Zolotova and Knudson2017:187–189, Appendix B) also reports strontium (87Sr/86Sr) values for the five ribs. They form a tight cluster, 0.7046–0.7047, within the Teotihuacan range for 87Sr/86Sr.
Oxygen-Isotope Compositions of Bioapatite Phosphate.
a 0.7‰ was subtracted from the measured δ18Op to compensate for breastfeeding.
b 0.35‰ was subtracted from the measured δ18Op to compensate for breastfeeding.
The five ambiguous enamel δ18Op values can be approached in the same way. The +15.3‰ value represents the lower right second premolar (LRP2) of individual A from Burial 1 (Table 3; Figure 3). It and the +15.7‰ value from the earlier-developing lower first molar crown (LM1) contrast with the +16.5‰ value for the second molar crown. This pattern suggests that the two lower values of δ18Op represent residence in the Teotihuacan region, whereas the second molar developed after movement to the relocation region. The relocation may have occurred around the age of 5.0–5.5 years. There are two other δ18Op values of +15.7‰ for enamel, one for an upper lateral incisor of individual C and the other for a lower central incisor of individual E. In both individuals these δ18Op values contrast with the oxygen-isotope compositions of later developing teeth that are securely in the enamel cluster (Table 3).
Oxygen-isotope compositions for individuals A–F. For individual F, the oxygen-isotope composition for bone was acquired from Nado and colleagues (Reference Nado, Zolotova and Knudson2017).
These four enamel δ18Op of +15.3 to +15.7‰ have a mean of +15.6‰, which is within the Teotihuacan range (White, Spence, et al. Reference Spence, White, Longstaffe, Rattray, Law, Gallut and Soto2004; White, Storey, et al. Reference White, Spence, Longstaffe and Law2004) but 1.2‰ higher than the mean (+14.4‰) of the +14.2 to +14.7‰ Teotihuacan bone cluster. For contemporaneous tissue formation, Webb and coauthors (Reference Webb, White and Longstaffe2014) found a mean bone-enamel offset in δ18Op of ±0.7‰ with a standard deviation of 0.5‰. However, there is considerable inconsistency in both the magnitude and direction of the offsets in their small sample, leaving it unclear whether they represent situational or inherent factors. The authors suggest that differences of less than 1.0‰ are not meaningful and that differences of 1.0 to 1.5‰ require some caution in interpretation (Webb et al. Reference Webb, White and Longstaffe2014:102). We shall assume, then, that these four enamel δ18Op and the bone cluster both represent Teotihuacan residence.
The +16.0‰ enamel value is from sample 28, for the lower right third molar of individual F. Sample 28 has a low crystallinity index but was retained in this analysis. The three other individual F enamel δ 18Op values are in the range of +16.3 to +16.5‰. The +16.0‰ value may represent the lower end of the enamel cluster's range, or it might include some degree of equilibration toward Teotihuacan oxygen-isotope compositions, indicating that individual F returned to Teotihuacan during the formation of her third molar crowns.
Individual Histories
Despite these uncertainties, patterning is visible in the results when individual histories are examined (Table 3; Figure 3). Individual A of Burial 1 has bone and enamel isotopic compositions in the Teotihuacan range but also a second molar crown that formed in the relocation geographic region (Figure 3), suggesting movement there at about 5.0–5.5 years of age.
In Burial 2 we have no samples from the earliest teeth of individual B (Figure 3). She may have moved from Teotihuacan to the relocation region before the development of the UI2 crown, which forms at 1.5–5.5 years and has a relocation region composition. Individual C (Figure 3) may have relocated sometime during UI2 formation.
We have no samples for individual D from his earlier-forming teeth (Figure 3). It is possible that he was not born in Teotihuacan, or he may have moved from Teotihuacan to the relocation region before formation of the UP1 crown; that is, before the age of 3.5–4.0 years. Individual E (Figure 3) may have moved sometime during the formation of the LI1 crown (0.6–3.5 years) and before the development of his LP1 crown (3.0–6.5 years). All of the tested teeth of individual F produced isotopic compositions in the enamel cluster, suggesting that she was living in the relocation region during her second year (Figure 3). Like individuals B and D, it is possible that F was not born in Teotihuacan. Without isotopic evidence from the earliest teeth, the central incisors and first molars, we cannot identify the region of birth of these three individuals.
Potential Limitations of the Method
There are a number of possible reasons for variability in enamel δ18Op. For one, dental crowns commonly take 1.5–4.5 years to form (Table 2), so an enamel sample taken for isotopic analysis may encompass time spent in more than one location, as we suggested for the M3 of individual F. In addition, weaning is a process rather than an event, and its timing can be affected by a number of factors. A competing younger sibling may hasten it, or a lack of a suitable transition food may prolong it. Furthermore, the breastfeeding reductions applied to particular dental categories may not be fully accurate. There is also the possibility that the dental development ages taken from AlQahtani and colleagues (Reference AlQahtani, Hector and Liversidge2010) may require some modification. Ubelaker (Reference Ubelaker1978:46) suggests that Amerindian dental development may have been more rapid than the European standards used by most investigators, including AlQahtani and colleagues.
Some investigators have suggested that enamel oxygen-isotope compositions may be affected by seasonal or other short-term fluctuations (e.g., Britton et al. Reference Britton, Fuller, Tütken, Mays and Richards2015:236; White, Longstaffe, et al. Reference White, Longstaffe and Law2004:243), although these variations would not significantly affect the results for bone (Ayliffe and Chivas Reference Ayliffe and Chivas1990:2603). Use of the whole tooth crown in this study should also have homogenized any limited intra-tooth variation in δ18O values. We also note that the water consumed by the residents of San José 520 probably came from local runoff and perhaps also from the San Lorenzo River, which passes only 1.2 km north of the site. The δ18O values of local runoff would have varied seasonally similar to precipitation, and hence should be reflected in bulk enamel and bone similarly, notwithstanding that the multiyear record is shorter in bulk tooth enamel. The San Lorenzo River is small and slow moving. As such, it may very well have been enriched in 18O relative to precipitation because of evaporation; again, however, this enrichment would have affected the bone and enamel of local inhabitants more or less equally. Local groundwater would have been relatively homogeneous in terms of oxygen-isotope composition (Scherer et al. Reference Scherer, de Carteret and Newman2015) and therefore not a strong factor in producing variation in inter- or intraindividual δ18O values. Some variation in δ18O values might have also been introduced by evaporation if water was stored uncovered over long periods in small reservoirs or ceramic vessels (Scherer et al. Reference Scherer, de Carteret and Newman2015). Again, however, any such variation would presumably have affected the AF2 household uniformly and should not have led to an enamel-bone offset in δ18O values.
Later Postclassic communities in marginal areas like that of San José 520 also used aguamiel and pulque (unfermented and fermented maguey sap products) as potable liquids, particularly in the dry season when they may have become important resources (Evans Reference Evans1990; Parsons and Parsons Reference Parsons and Parsons1990; Robertson and Cabrera Cortés Reference Robertson and Cabrera Cortés2017). Pulque residues have been identified in Teotihuacan period ceramics (Correa-Ascencio et al. Reference Correa-Ascencio, Robertson, Cortés, Castro and Evershed2014). There is the potential for these drinks to become enriched in 18O relative to precipitation during these processes: (1) evaporation of soil water, (2) transpiration during sap development, (3) heating or boiling during their preparation, and (4) fermentation during pulque formation (e.g., McCool and Coltrain Reference McCool and Coltrain2018). If these liquids were consumed in significant quantities, and differentially by subadults and adults, an offset between enamel and adult bone δ18O values could conceivably result. The probability, however, is that such 18O-enrichment would be more likely expressed in adult bone than subadult tooth enamel, thus decreasing rather than increasing the difference between bone and tooth δ18O values.
Pathological conditions might affect metabolism and, by extension, δ18O values (Webb et al. Reference Webb, White and Longstaffe2014:102). An analysis of Nubian skeletons by White, Longstaffe, and colleagues (Reference Spence, White, Longstaffe, Rattray, Law, Gallut and Soto2004:243), however, found no changes, with cases of iron-deficiency anemia and only slightly lower bone δ18O values in females with osteopenia. Pathology at San José 520 is difficult to evaluate, given the poor condition of the remains. The only clear case beyond the usual arthritis and oral pathology is some slight evidence of healed porotic hyperostosis, an anemia-related condition, in the cranium of individual A (Driscoll Reference Driscoll2005).
More problematic is the possibility of an inherent enamel-bone offset in oxygen-isotope compositions for coevally formed tissues, perhaps due to metabolic differences between subadults and adults. Despite the results of some analyses (Luz and Kolodny Reference Luz and Kolodny1985; Webb et al. Reference Webb, White and Longstaffe2014; White, Longstaffe, et al. Reference White, Rebecca Storey, Longstaffe and Spence2004; Wright and Schwarcz Reference Wright and Schwarcz1998:14–15), such an offset cannot yet be entirely dismissed. However, the variability in offset direction (Webb et al. Reference Webb, White and Longstaffe2014) and the wide range of offset values reported elsewhere in the literature cannot be explained solely by an inherent difference. For example, the Tlailotlacan and Tlajinga series of Teotihuacan show almost equal frequencies of higher enamel and higher bone δ18O values (White, Spence, et al. Reference Spence, White, Longstaffe, Rattray, Law, Gallut and Soto2004 White, Storey, et al. Reference White, Rebecca Storey, Longstaffe and Spence2004). A Chapantongo series (see the later discussion) of 25 enamel-bone pairs includes only two in which the bone value is higher, but the mean offset is only 1.6 ± 0.9‰ and the range is large, 0.2‰–4.1‰ (Spence et al. Reference Spence, White, Fournier, Webb and Longstaffe2011). Rather than seeking some inherent cause, it might be more fruitful to look for an explanation in variability introduced through weaning practices, dietary shifts, changes in local water sources (Scherer et al. Reference Scherer, de Carteret and Newman2015) or, as suggested here, geographic relocation.
Discussion
The oxygen-isotope evidence points to a long-standing pattern of individual movement between two distinct geographic regions, the Teotihuacan region (in particular, residence AF2 of the San José 520 settlement) and the relocation region, the latter characterized by oxygen-isotope compositions in the +16.2 to +16.8‰ range. The location of this region is not known. Moreiras Reynaga (Reference Moreiras Reynaga2019) developed a map of δ18Op values for Mesoamerica, showing that values in the relocation region enamel cluster would be expected in the southern highlands. However, +16.2 to +16.8‰ values are also found closer to Teotihuacan. In a series from the Epiclassic site of Chapantongo, some 90 km northwest of Teotihuacan, 10 of the 32 enamel δ18Op values are in that range (Spence et al. Reference Spence, White, Fournier, Webb and Longstaffe2011). Isolated δ18Op values in the range have also been noted from sites in Puebla and Michoacán.
Individuals A, C, and E were born in Teotihuacan and passed their earliest years there, moving to the relocation region sometime between one and six years of age. Perhaps this movement was initiated when the child passed a stage in development that made the journey easier, like weaning or walking. The two neonates in burials 2 and 3, of course, died before any movement was possible. Individuals B, D, and F may also have followed this relocation pattern, but with no samples from their earliest teeth, we cannot be sure of their birth location. They may have been born in Teotihuacan or in the relocation region after the family moved there.
All of the tested individuals passed the rest of their childhood in the relocation region. Almost all of the canine, premolar, and second and third molar oxygen-isotope compositions, and even most of the lateral incisor compositions, fall in the range associated with the relocation region. The +16.0‰ value of individual F's third molar may include some degree of equilibration toward Teotihuacan oxygen-isotope compositions. If so, relocation to Teotihuacan, at least for that individual, would have taken place in adolescence. Most of the bone isotopic compositions are in the δ18Op range of +14.2 to +14.7‰ characteristic of Teotihuacan, with a few exceptions. The +13.3‰ value is below the Teotihuacan range, but because it cannot be assigned to an individual, it cannot be fully assessed. The +15.9‰ and +16.0‰ values are also from unidentified individuals and thus difficult to evaluate. However, both values are in the general Teotihuacan range. and both samples have 87Sr/86Sr values typical of Teotihuacan (Nado Reference Nado2017:Appendix B).
Conclusions
Together, the evidence suggests a pattern, continuing over some centuries, of birth in the area of the San José 520 settlement, movement (presumably as part of a family) to a different region, and then return to San José 520 during the teen or young adult years. This is a highly unusual pattern. Stable isotope analyses in other Teotihuacan sites frequently show the movement of individuals to Teotihuacan (e.g., Manzanilla Reference Manzanilla2012; Spence et al. Reference Spence, White, Longstaffe, Rattray, Law, Gallut and Soto2004; White, Storey, et al. Reference White, Spence, Longstaffe and Law2004) but not the back-and-forth pattern suggested by the San José 520 data. One comparable situation may be the sojourning of Tlailotlacan infants and their parents in other Zapotec communities, although their stays were of much briefer duration (White, Spence, et al. Reference Spence, White, Longstaffe, Rattray, Law, Gallut and Soto2004).
Perhaps a better example of long-term relocation is presented by Estructura 19 (E19), an apartment compound a short distance east of Tlailotlacan. E19 contains evidence of occupants of both Zapotec and Michoacán affiliation (Begun Reference Begun2013; Gómez Chávez Reference Gómez Chávez and Gallut2002). Some 20 m south of the structure is a deep well that contained the articulated skeletons of some 30 people. The dental enamel and bones of eight of these skeletons were analyzed for their oxygen-isotope composition (Spence et al. Reference Spence, White, Gazzola, Chávez and Longstaffe2006). The enamel δ18Op values are for the most part in the Teotihuacan range, whereas the bone values are all below it. The enamel-bone offset has a mean of 3.0‰ but varies widely, from 1.7 to 5.2‰. Oxygen-isotope compositions similar to those of the bone samples are found in the Valley of Oaxaca and the Lake Patzcuaro region of Michoacán. It may be, then, that these individuals were born and grew up in Teotihuacan but then moved to Oaxaca or Michoacán, returning to Teotihuacan only late in their lives.
The relocation region apparently played a very important role in the lives of the people of San José 520 or, at least, the lives of those living in the AF2 residence. Unfortunately, we cannot define that role without knowing where that region is located and something about the lives of the people there. Perhaps it was the original homeland of the San José 520 people, one to which they returned to ensure that their children were immersed in the homeland culture during their formative years. Or perhaps the purpose of this unusual settlement system was to send young adults (with their children) there to maintain some sort of exchange, provisioning, or social relationship with Teotihuacan that was conducted through San José 520. Identification of the relocation/homeland region will help in resolving these questions, but it will require further isotopic research in Mexico, both in and beyond San José 520, and the additional evidence that can be provided by other isotopic systems, such as strontium (Spence and White Reference Spence and White2009:237–238).
Acknowledgments
We are very grateful to Christine White, who initiated the isotopic analysis of the San José 520 series. Dr. White's work on Teotihuacan and other Mesoamerican sites has had a profound impact on our understanding of the structure and volatility of these ancient populations. We also thank the Consejo de Arqueología of the Instituto Nacional de Antropología e Historia for permission to conduct the project, and the INAH-Zona de Monumentos Arqueológicos for their support. We are grateful to the Foundation for the Advancement of Mesoamerican Studies, the Wenner-Gren Foundation, and Arizona State University for their financial support of the San José 520 excavation, as well as the ASU-Teotihuacan Research Laboratory and their staff for facilitating laboratory work. Kristin Nado kindly allowed use of her dissertation data. We also thank the reviewers for their helpful suggestions. The Canada Research Chairs program, Canada Foundation for Innovation, and Ontario Research Fund provided financial support for analytical costs and the infrastructure that supported the isotopic measurements. This is the University of Western Ontario's Laboratory for Stable Isotope Science Contribution #377.
Data Availability Statement
The San José 520 collection is temporarily maintained in the Arizona State University Teotihuacan Research Laboratory, Teotihuacan, Mexico.
