Mesoamerican Preclassic Trade Connections

The combined morphological, osteometric and aDNA evidence confirm the early presence of M. gallopavo at El Mirador in the Maya lowlands. The presence of male, female and subadult turkeys, some with reduced flight capabilities, suggests that the introduced birds were captive reared and/or domesticated. Although we do not yet know the immediate source of the turkey bones, the most reasonable explanation is that a few Mexican turkeys entered the site as exchange goods directly from central/northern Mexico suggesting that Late Preclassic association between El Mirador and contemporary northern Mesoamerican cultures at sites such as Teotihuacan was closer than previously recognized. Most of the evidence for the exchange of goods and ideas between central Mexico and the Maya region dates to the Classic period several centuries later (ca. AD 250–900) [15]. The El Mirador turkeys therefore add to a relatively sparse record of Preclassic cultural and material exchange between the Maya lowlands and northern Mesoamerica [16]–[18]. Prior information on Preclassic exchange comes primarily from non-perishable goods such as obsidian and ceramics so the non-local turkeys at El Mirador also expand our understanding of the types of goods that were exchanged long distances during this early period of Maya history. Although The El Mirador turkey specimens could represent the transport of dried meat or partial carcasses, the presence of associated upper and lower limb bones suggests that the animals were imported whole and possibly live. The imported turkeys further emphasize that El Mirador's vast Late Preclassic trade connections extended some 1000 kilometers north into central Mexico, in addition to the site's better known connections with the Atlantic and Pacific coasts and the Maya highlands [14], [19], [20]. El Mirador's participation in interregional trade and cultural interaction was likely pivotal to the site's accumulation of political and economic power during the Late Preclassic [20], [21].

Timing of Mesoamerican Turkey Domestication and Diffusion

The Late Preclassic presence of M. gallopavo in the Maya region has important implications for documenting the timing of Mesoamerican turkey domestication and diffusion.

To date, morphological and genetic studies have been unable to distinguish between wild and early domestic forms of M. gallopavo. In the absence of morphological and genetic markers, archaeologists have relied on indirect evidence of domestication such as the presence of pen structures, egg shells and neonates or appearance of the species outside its presumed natural geographic range. Previous to our study, all indirect evidence for Mesoamerican turkey husbandry dated to the Classic period or later [8], [22], [23]. Since the Preclassic El Mirador turkeys represent movement of M. gallopavo outside its natural geographic range, the specimens represent the earliest indirect evidence of captive turkey rearing or domestication in Mesoamerica. A Preclassic origin for Mexican turkey domestication has been suggested previously [11], [24], but archaeological evidence has been lacking. The El Mirador turkey specimens confirm that turkey domestication, or at least captive rearing, dates to the Preclassic.

Determining when Mesoamerican cultures started experimenting with turkey rearing and domestication is vital to the larger question of whether the origins of New World turkey husbandry should be attributed to cultures of the American Southwest or Mesoamerica. It was originally believed that the turkey was first domesticated in Mesoamerica and then introduced in domestic form to the American Southwest [1]. More recent archaeological and genetic evidence has overturned this scenario demonstrating that turkeys were independently domesticated in these two regions although the timing of domestication remains unclear [9], [25], [26]. It is possible that the idea for turkey rearing or husbandry prior to domestication also arose independently in the American Southwest and Mesoamerica. However, the wealth of documented cultural and material exchange between the regions supports a model of diffusion for the concept of turkey husbandry as part of early exchange networks. The Late Preclassic El Mirador M. gallopavo specimens provide evidence for the antiquity of Mesoamerican turkey rearing, and support the probable role of interregional turkey exchange in the diffusion of ideas about animal management in the New World.

The presence of Late Preclassic Mexican turkeys at El Mirador also confirms that this non-local species was introduced to the Maya region over one thousand years earlier than previously thought. If the El Mirador turkeys are isolated examples of imported captive-raised/domestic Mexican turkeys, it raises the question of why the Maya did not broadly adopt the idea of turkey rearing or domestication until the Postclassic. One possibility is that turkey domestication was not widespread or common in any part of Mesoamerica until the later half of the Classic period despite its potentially earlier origins. This suggestion is supported by the relative rarity of M. gallopavo specimens in Preclassic/Formative central Mexican faunal assemblages, and their increasing frequency in Classic and Postclassic deposits [11], [22]. An alternative explanation is related to the nature of the Postclassic Maya economy. During the Postclassic, long-distance trade between central Mexico and the Maya area increased with the expansion of maritime trade routes around the Yucatan Peninsula between the Gulf of Mexico and Central America's Caribbean coast [27], [28]. Increased Postclassic exchange throughout Mesoamerica could have facilitated the dispersal of domesticated turkeys to the Maya area through repeated introductions of breeding pairs and transmission of rearing information. In contrast, the rare earlier introduction of the bird might not have been sufficient to fully incorporate the species into the Maya economy.

Although the El Mirador turkeys may represent isolated imports, it is also possible that M. gallopavo has been under-identified in Preclassic and Classic Maya zooarchaeological assemblages since M. gallopavo and M. ocellata can be difficult to differentiate morphologically when preservation is poor. Researchers also may not have considered the possible presence of M. gallopavo in earlier assemblages due to the longstanding belief that they were not introduced to the Maya region until the Postclassic. It is essential to determine whether the Mexican turkey appeared in the Maya region earlier than previously understood because an earlier introduction would have provided a second domestic vertebrate during the Late Preclassic to Classic period of Maya population expansion and increasing social complexity. During the Preclassic, the Maya relied extensively on the domestic dog (Canis lupus familiaris), which they used for both dietary and ritual purposes [29], although perhaps primarily for ceremonies related to elite display and power negotiations [29], [30]. The turkey was another important food and ritual animal among the Maya [31]. Prior models suggest that only local, wild ocellated turkeys were used through the Classic period, but an early demand for domesticated or captive-reared turkey (i.e., M. gallopavo) could have been related to increased elite ceremonial and status-displaying activities as well as the need for meat to feed growing populations during the Late Preclassic to Classic period of population growth and cultural florescence.

Conclusions

Combined zooarchaeological and aDNA analyses identified the earliest non-local Mexican turkey remains in the Maya cultural region at the site of El Mirador. Prior to this discovery, the earliest evidence of M. gallopavo in the Maya area dated to approximately one thousand years later. The El Mirador turkeys may represent rare or isolated imports from central/northern Mexico, but it is also possible that captive/domestic Mexican turkey husbandry was practiced by the ancient Maya much earlier than previously thought. The Maya may therefore have had access to another domestic vertebrate, besides the dog, during the Late Preclassic to Classic period of population expansion and increasing social complexity. Significantly, the El Mirador turkeys also provide the earliest indirect evidence of M. gallopavo captive rearing or domestication in Mesoamerica. Previously, all other indirect evidence of husbandry (e.g., pen structures, egg shells and neonates, or the appearance of the species outside its natural geographic range) dated to the Classic period or later [8], [22], [23].

The early presence of M. gallopavo at Late Preclassic El Mirador demonstrates a need to reassess the timing of turkey domestication and diffusion in Mesoamerica. Understanding when the Mexican turkey was domesticated and when it was introduced to and fully adopted by the ancient Maya has important consequences for understanding Mesoamerica subsistence systems and long-distance trade connections. The topic also has broader ramifications with respect to the process and timing of New World animal domestication, and the culture-specific motivations for incorporating or not incorporating potential domesticates or managed species into ancient social and economic systems.

Zooarchaeology and Osteometrics

The archaeological turkey bones were identified within a larger zooarchaeological assemblage from the site (number of identified specimens = 3470). The sample also contained other bird bones that we could only identify to the level of taxonomic subclass (Aves) because they were undiagnostic elements or poorly preserved. Nearly all of the unidentified bird remains come from large-bodied species, and some of these may represent additional M. gallopavo elements.

Zooarchaeological specimens were identified through comparison with modern skeletons housed in the Florida Museum of Natural History Environmental Archaeology and Ornithology collections (www.flmnh.ufl.edu/museum/collections.htm). Turkey age and sex determinations were based on skeletal element size, osteometrics [6], [32], and tarsometatarsus spur morphology. Archaeological bones were measured using standard osteometric measurements and were compared to published metric data available for M. gallopavo and M. ocellata [6: Tables 10,11,14,15,20,21].

Ancient DNA Analysis

The four archaeological bird bones were processed in the Ancient DNA Laboratory located in the Department of Archaeology at Simon Fraser University. The ancient DNA laboratory is specifically designed for and dedicated to ancient DNA work - no modern DNA samples have ever been processed in the lab. The lab is equipped with a UV filtered ventilation and positive airflow, with dedicated equipment and bench UV lights. Strict contamination control protocols are followed in the lab, including: 1) the use of protective clothing including Tyvex™ suits, gloves, masks, etc.; 2) the separation of the pre- and post-PCR work (located in two buildings with separate ventilation systems); and 3) the inclusion of multiple blank DNA extractions and negative PCR controls.

Two separate DNA extractions were conducted for each bone, with the repeat extractions occurring several months after the initial extractions. For both extractions, the analyzed bone samples weighed approximately 0.5 g. Bone samples were subjected to rigorous chemical decontamination in order to remove possible surface contamination [33]. The samples were immersed in a 6% sodium hypochlorite solution for 7 minutes, followed by immersion in 1 N HCl solution for 30–60 seconds, then immersion in 1 N NaOH for 30–60 seconds, before being rinsed twice in ultra-pure water and UV irradiated in a crosslinker for 30 minutes on two sides. The samples were crushed into powder using a liquid nitrogen grinding mill (6750 SPEX CertiPrep Freezer/Mill). Three additional ancient turkey bones were included in analysis to act as positive controls for both the initial and repeat extractions, as well as the subsequent PCR reaction sets. These three bones were recovered from archaeological sites in Arizona (ca. AD 1100–1300) [9] and were processed separately from the El Mirador samples. DNA extraction was performed using a modified silica-spin column technique [9], [34], and approximately 100 µl of DNA solution was collected for each sample.

PCR amplifications were conducted in a Mastercycler Personal (Eppendorf, Hamburg, Germany) in a 30–50 µL reaction volume containing 50 mM KCl, 10 mM Tris-HCl, 2.5 mM MgCl2, 0.2 mM dNTP, 1.0 mg/ml BSA, 0.3 µM each primer, 3.0–5.0 µl DNA sample and 2.5–3.5 U AmpliTaq Gold™ LD (Applied Biosystems). Primers were designed to target fragments of Meleagris mitochondrial DNA of various lengths. Several different primer sets were tested (Table S3). PCR began with an initial 12 minute denaturing period at 95°C, followed by 60 cycles at 94°C for 30 seconds (denaturing), 52°C for 30 seconds (annealing), and 72°C extension for 40 seconds. Blank extracts and negative controls were included in each of the PCR reaction sets. Ancient positive controls (Arizona archaeological turkey bone extracts) were also tested to ensure the efficacy of the primer sets and PCR conditions.

Five µL of PCR product were visualized via electrophoresis on a 2% agarose gel using SYBR Green™ staining. Successfully amplified PCR products of expected length were purified using MinElute™ purification kits (Qiagen, Valencia, CA). Purified products were sequenced using both forward and reverse primers at the Central Facility of the Institute for Molecular Biology and Biotechnology Laboratory at McMaster University (using an ABI 3100) and at Macrogen, Seoul, Korea (ABI 3730XL). The obtained electropherograms were edited, aligned and compiled using ChromasPro software (www.technelysium.com.au). Consensus sequences were developed based on multiple PCR amplifications and sequencing.

Once the DNA analysis of the ancient samples was completed, DNA was extracted from a modern M. ocellata phalanx collected from Guatemala (FLMNH catalog number Z11050, Table S4). The M. ocellata samples were processed in the SFU Center for Forensic Research in a lab dedicated to DNA analysis of modern or forensic bone samples. Two 0.5 g bone samples were extracted using methods similar to those listed above. The DNA extracts were PCR amplified using primers TK-F2/TK-R405 and TK-F252/TK-R567 (Table S3) and produced amplicons of 254 bp and 338 bp in length, respectively. The PCR products were sequenced from both directions and consensus sequences matched identically with the GenBank M. ocellata reference sequence AF487120.

Ancient DNA Extraction Results and Authenticity

Sequences were obtained for three (631.0209, 631.0173, 631.0206) of the four bone samples. Only one bone yielded replicable sequences (631.0206) using different primer sets and using both the initial and repeated DNA extracts (Table S4). The obtained ancient DNA sequences were BLAST-compared through GenBank to determine if they would match Meleagris sequences and to ensure that they did not match with any other unexpected species or sequences. Multiple alignments of the sample sequences and published Meleagris mtDNA reference sequences were conducted using ClustalW [35] in order to confirm the species identifications.

Two replicable DNA fragments were obtained for sample 631.0206 totaling 71 bp and 55 bp respectively once the primer sequences were removed. The two fragments correspond to positions 15731–15800, and 15858–15913 on the M. gallopavo mitochondrial genome (reference NC010195). Despite their short lengths, the sequences clearly match most closely with M. gallopavo rather than M. ocellata, as demonstrated by the multiple alignment (Table S5).

Due to the antiquity of the samples and the tropical climate from which they were recovered, a low success rate for DNA extraction and amplification is expected, and provides support for the authenticity of the recovered sequences. The successfully reproduced sample, the carpometacarpus fragment, was the largest of the four bone samples and the best preserved morphologically [36]. PCR targeted fragments of mtDNA varying in length from <100 to 400 bp. Only short fragments of DNA could be amplified despite repeat amplifications with longer primer sets designed to detect contamination from modern sources [37]. Multiple blank extractions and negative PCR controls were included in the study, none of which yielded DNA fragments of expected length. Successful amplification of three positive controls (ancient turkey bones from Arizona) demonstrated the efficacy of both the extraction method and PCR primers [9].

The retrieved sequences matched very closely or identically with modern turkey reference sequences in GenBank, and therefore cannot be used to definitively rule out the possibility of contamination from modern sources. However, considering the short length of the retrieved sequences significant differences between the ancient and modern turkeys were not expected. Moreover, the primer sets were designed to target areas maximizing differences between M. gallopavo and M. ocellata, rather than polymorphic sites within the M. gallopavo control region. The retrieved sequences themselves, including those unreplicated sequences, demonstrate significant post-mortem damage characterized by C→T transitions (Fig. S1). These DNA transitions were likely caused by hydrolytic damage and are anticipated to occur in ancient sequences [37], [38]. Finally, the DNA identification of M. gallopavo supports the morphological and osteometric identification of the bones.