Complexity of the Maya Diet

It is well known that the diet of the ancient Maya was derived from the Maya forest within which they thrived (Figure 1). Its diversity is, in part, revealed in the plant and animal remains of the excavated middens, as well as the isotopes of the collagen in human bones of the area. This evidence of plant and animal consumption represents resources from a complex landscape ranging from uplands to wetlands and from the home infields to agricultural outfields (Ardren and Miller Reference Ardren and Miller2020; Emery Reference Emery2007; Emery and Thornton Reference Emery and Thornton2008; Fedick Reference Fedick, Freidel, Masson and Demarest2020; Ford Reference Ford, Hutson and Ardren2020; McNeil Reference McNeil, Hutson and Ardren2020; Slotten et al. Reference Slotten, Lentz and Sheets2020). The varied uses of fields and forests inferred archaeologically is consistent with the observations at the time of the sixteenth-century Spanish conquest and among traditional contemporary Maya (Colunga-Garcia Marín and Zizumbo-Villarreal Reference Colunga-Garcia Marín and Zizumbo-Villarreal2004; Terán and Rasmussen Reference Terán and Rasmussen1995; Varela Scherrer and Liendo Stuardo Reference Varela Scherrer and Stuardo2021).

Figure 1. Map of the Maya area with place names indicated (MesoAmerican Research Center UCSB).

Food resources have distinct stable isotopic values (δ¹³C, δ¹⁵N) that are incorporated into the bones and teeth of humans and animals (Schwarcz and Schoeninger Reference Schwarcz and Schoeninger1992) with seasonal and long-term variations averaged over about 10–20 years. Recognizing that the diet must include both carbohydrates and protein, interpretations of consumption of major food components have been based on isotopic analyses of human skeletal collagen and their relationship to the compositions of the sources of carbohydrate and protein.

Although an enormous variety of plants have been used by the Maya for carbohydrates, maize is considered among the most important and is well documented in the Mesoamerican and Maya diet, past and present. Measured isotopically for carbon, δ¹³C, human bone collagen generally averages around −8 to −10‰, which has been accepted as an indicator of a high degree of consumption of maize, although there is archaeological evidence of diverse plant consumption in the diet of the Maya (Farahani et al. Reference Farahani, Chiou, Harkey, Hastorf, Lentz and Sheets2017; Fedick Reference Fedick, Freidel, Masson and Demarest2020; McNeil Reference McNeil, Hutson and Ardren2020; Terán and Rasmussen Reference Terán and Rasmussen1994; Trabanino Reference Trabanino2014).

Humans require approximately 50–60 g of protein per day (Young and Pellett Reference Young and Pellett1987). Based on skeletal remains at Maya archaeological sites and using the δ¹⁵N values of human collagen, many previous scholars have inferred that the principal source of protein for the Maya was the flesh of terrestrial animals, principally white-tailed deer (Odocoileus virginianus; Emery and Thornton Reference Emery and Thornton2008). Faunal remains of many other animals (ca. 100 species) characteristic of the full range of habitats of home gardens, fields, and forests have also been found that contributed to the protein intake (Emery and Thornton Reference Emery and Thornton2008; Varela Scherrer and Liendo Stuardo Reference Varela Scherrer and Stuardo2021). The isotopic composition of most of these remains would not have differed significantly from deer. Deer—browsing ecotone habitat dwellers—were a common resource at the time of the conquest, and it was noted that they appeared “tame” as part of the varied domesticated landscape of the Maya (Diaz del Castillo Reference Diaz del Castillo1927 [1568]); Landa Reference Landa and Gates1937; Pohl and Feldman Reference Pohl, Feldman and Flannery1982).

The Maya consumed, in addition, a wide variety of animals, birds, and insects. At sites near the coast, fish and other marine foods would have been available. Populations living on rivers or lakes would have had access to aquatic sources such as fish and turtles. All of these would have provided an adequate protein supply when present in sufficient amount (Emery Reference Emery2007; Emery and Thornton Reference Emery and Thornton2008; Vanderwarker Reference Vanderwarker2006).

The Maya today know and use nearly 500 native edible plants (Fedick Reference Fedick, Freidel, Masson and Demarest2020), many of which have been identified in the archaeological record (Dussol et al. Reference Dussol, Elliott Michelle and Nondedeo2017; Thompson et al. Reference Thompson, Hood, Cavallaro, Lentz, Lentz, Dunning and Scarborough2015; Trabanino and Liendo Stuardo Reference Trabanino and Stuardo2016). Only a few, however, would provide significant amounts of protein. Legumes, including beans, containing ~20% dry weight of protein, were cultivated across Mesoamerica and provided about three times more protein content than most other vegetables. Seeds are another important source of protein; for example, tambulo squash seeds contain 40 wt% protein, whereas chia seeds contain 20 wt%.

Chaya, a plant widely cultivated by the modern Maya (Everton Reference Everton2012:104, 135, 225, 427; Sterling Reference Sterling2014), has a notably high-protein content and thus presents a potential contributor of protein to ancient Maya diets. Evidence confirms early domestication, likely in the central Maya Lowlands, and is documented at the time of the conquest (Colunga-Garcia Marín and Zizumbo-Villarreal Reference Colunga-Garcia Marín and Zizumbo-Villarreal2004). Today, it is a common ingredient in contemporary cooking across the Maya area (Kuti and Torres Reference Kuti, Torres and Janick1996; Standley Reference Standley1930; Standley and Steyermark Reference Standley and Steyermark1949; Sterling Reference Sterling2014:128–132). We propose that chaya was an important source of protein in the ancient Maya diet. In this article we examine its isotopic and nutritional composition and suggest how it could have been a contributor to the total protein intake of the Maya at times and places where animal-based protein supplies were limited.

Isotopic Paleodiet Studies

Interpretation of human diet is partly based on the isotopic values of human tissues and, in particular, the δ¹³C and δ¹⁵N values of bone collagen. δ¹³C values tend to be about 5‰ higher than that of the food consumed (whether starch, fat, or protein). About 95% of plant species use the C₃ pathway to fix carbon and as a result have δ¹³C values around −26‰. Maize and amaranths instead use the C₄ pathway to fix carbon, resulting in higher δ¹³C values averaging around −12‰. In addition, the Maya ate succulent plants such as prickly pear that use the CAM pathway to fix carbon and have δ¹³C values intermediate between C₃ and C₄ plants. Variation in the δ¹³C value of human collagen is largely attributable to variations in the relative amount of C₄ versus C₃ plant foods consumed.

In general, the δ¹⁵N values of human collagen tend to be about 3‰ higher than that of the food consumed because of the trophic-level effect (Schoeninger and DeNiro Reference Schoeninger and DeNiro1984; Schoeninger and Moore Reference Schoeninger and Moore1992; Schwarcz Reference Schwarcz, Holland and Turekian2014). Where humans have access to adequate supplies of high-protein foods, nonessential amino acids tend to be routed directly from diet to collagen synthesis (Schwarcz Reference Schwarcz, Holland and Turekian2014), causing the δ¹³C value of collagen to be biased toward that of the high-protein foods. Average δ¹³C and δ¹⁵N values provide clues to the long-term (~10–20 years) foods consumed and can be interpreted based on the flora and faunal collections from archaeological contexts.

Based on research reported by Scherer (Reference Scherer2017) and the wider comparative data provided by Scherer and colleagues (Reference Scherer, Wright and Yoder2007:Figure 6), the average δ¹³C value of human collagen varies from −15 to −7‰ (Figure 2). These relatively high values are attributed to widespread consumption of maize (Kennett et al. Reference Kennett, Prufer, Culleton, George, Robinson, Trask and Buckley2020; Reed Reference Reed1998; Scherer et al. Reference Scherer, Wright and Yoder2007; Tykot et al. Reference Tykot, van der Merwe, Hammond and Orna1996; White and Schwarcz Reference White and Schwarcz1989; White et al. Reference White, Healy and Schwarcz1993), although amaranths and CAM plants also contribute to the higher δ¹³C values.

Figure 2. Average isotopic composition (δ¹³C, δ¹⁵N) of human bone collagen from Maya sites (after Scherer 2007).

The average δ¹⁵N values of human collagen at those archaeological sites discussed by Scherer and colleagues (Reference Scherer, Wright and Yoder2007:91–92, 97) vary from about 8 to 10‰, averaging 9.1 ± 1.0‰. It is noteworthy that this average remains constant across space and from Preclassic to Postclassic times, indicating that the Maya had reliable sources of protein with an average δ¹⁵N of 6 ± 1‰. This measure overlaps with the δ¹⁵N value of the flesh of C₃-consuming animals such as the herbivorous white-tailed deer and red brocket deer (Mazama spp.), as well as the flesh of other terrestrial animals (Emery Reference Emery2007).

Other possible protein sources may have included aquatic and marine fish, amphibians, shellfish, and insects. Human consumers of fish (marine or freshwater) tend to harvest fish with high trophic levels with δ¹⁵N values of 10‰–18‰ (Katzenberg et al. Reference Katzenberg, McKenzie, Losey, Goriunova and Weber2012; Schoeninger and DeNiro Reference Schoeninger and DeNiro1984; Schwarcz et al. Reference Schwarcz, Chisholm and Burchell2014; Williams et al. Reference Williams, White and Longstaffe2009)—with the result that humans whose diets included marine foods could show increased δ¹⁵N values. Considering the sites from which data were compiled by Scherer and colleagues (Reference Scherer, Wright and Yoder2007:91; Figure 2), we can compare δ¹⁵N values for sites of varying distance from a marine coast (Figure 3). There is a very slight trend of increasing δ¹⁵N with decreasing distance from the coast, suggesting only a modest influence of the consumption of marine foods.

Figure 3. δ¹⁵N of average human collagen from Maya sites plotted against distance of the site from the seacoast to assess the influence of marine fish as a nutrient by humans.

Table 1. Isotopic and Protein Data Derived from Samples of Chaya and Residue from Boiling of Chaya.

* Protein = N wt% × 6.25 (Elgar et al. Reference Elgar, Hill, Holroyd and Peddie2020).

Insects range widely in their δ¹⁵N values depending on trophic level and feeding habits, from values as low as 0‰–12‰ (Albers et al. Reference Albers, Schaefer and Scheu2006; Feldhaar et al. Reference Feldhaar, Gebauer and Blüthgen2010; Quinby et al. Reference Quinby, Curtis Creighton and Flahert2020), and it is difficult to estimate what would have been their net effect on human consumer δ¹⁵N values. Most shellfish tend to have low δ¹⁵N values because they are either herbivores or feed on suspended particulate organic matter (Schwarcz et al. Reference Schwarcz, Chisholm and Burchell2014).

Beans were widely consumed by the Maya and are a rich source of protein. However, the average δ¹⁵N value of beans is 3.9 ± 0.8‰ (Warriner et al. Reference Warinner Christina and Tuross2013). This is substantially lower than the average value of the flesh of animals, including deer, as well as the recorded values of Maya bone collagen (Scherer at al. Reference Scherer, Wright and Yoder2007). If beans had been a major source of protein, this would result in δ¹⁵N values for human collagen significantly lower than are observed in ancient Maya populations, ranging between 6 and 12‰, as shown in Figure 2.

Chaya as a Source of Protein

Leaves of the edible “tree spinach” chaya (Cnidoscolus aconitifolius [Mill.] I.M.Johnst.), contain about 30% protein (dry basis: Peregrine Reference Peregrine1983; Sarmiento-Franco et al. Reference Sarmiento-Franco, Sandoval-Castro, McNab, Quijano-Cervera and Reyes-Ramirez2003). This is high compared with other plant-based food (12–25 wt%). Only one common Mesoamerican food plant has appreciably high-protein levels: beans (19.5%–24.8%; Carbas et al. Reference Bruna, Machado, Oppolzer, Ferreira, Queiroz, Brites, Rosa and Barros2020). Importantly, unlike beans, the protein of chaya is complete and contains all essential amino acids in adequate dietary proportions.

Chaya is a member of the drought-resistant Euphorbiaceae family. Indigenous to the Maya Lowlands, the wild variety is characterized by seriously irritating urticating hairs on the stems and leaves that burn at the touch, a reaction that lasts for several days, which is the chief disincentive to its collection. As a result of domestication, a crop species, C. aconitifolius, was developed to express greatly reduced or nonexistent urticating hairs, making the leaves easier to harvest (Figure 4; Ross-Ibarra and Molina-Cruz Reference Ross-Ibarra and Molina-Cruz2002; see also Colunga-Garcia Marín and Zizumbo-Villarreal Reference Colunga-Garcia Marín and Zizumbo-Villarreal2004; Roys Reference Roys1933). Today chaya is considered a favored plant among the Maya (Ardren and Miller Reference Ardren and Miller2020; McNeil Reference McNeil, Hutson and Ardren2020), and there is evidence suggesting that domestication extended over millennia (Colunga-Garcia Marín and May Pat Reference Colunga-Garcia Marín, Pat, Villarreal, Rasmussen, Arias Reyes and Contreras1992; Ross-Ibarra and Molina-Cruz Reference Ross-Ibarra and Molina-Cruz2002).

Figure 4. A young chaya shrub growing from recent cuttings (from MacduffEverton.com). (Color online)

Chaya has been difficult to distinguish prehistorically. However. Miksicek (Reference Miksicek, Turner and Harrison1983:102–103), studying floral remains at the Pulltrouser Swamp in Belize, proposed a tentative identification of five fragments of charred Euphorbiaceae root as chaya. Chaya produces few pollen grains (Ross-Ibarra and Molina-Cruz Reference Ross-Ibarra and Molina-Cruz2002) and is thus unlikely to appear in palynological records. Chemical digestion of chaya leaves failed to yield siliceous phytoliths, but roots have not been examined and deserve consideration and should be tested for them (Morell-Hart, personal communication 2020). Although domesticated chaya rarely produces seeds, it is easily propagated from stem cuttings (Ross-Ibarra Reference Ross-Ibarra2003:289).

The productivity of chaya is remarkably high (Bendaña Reference Bendaña2020). Peregrine (Reference Peregrine1983) estimated that a fully mature C. aconitifolius plant produces 4.9–7.4 kg of fresh leaves per month, with annual yields ranging from 5 to 8 tons per ha⁻¹. Cultivation of chaya in 50–70 ha for a population of 10,000 people would yield about 400 g per day of leaves per person, or about 120 g protein d^-1. Assuming an average human protein requirement of 60 g per day (Young and Pellett Reference Young and Pellett1987), this could more than satisfy human needs. Dispersed within the traditional infield home gardens and outfield areas of Maya communities, the protein-rich chaya would form part of the polyculture milpa forest garden cycle (Ardren and Miller Reference Ardren and Miller2020; Bendaña Reference Bendaña2020; Ford Reference Ford, Hutson and Ardren2020; Ford and Clarke Reference Ford, Clarke, Isendahl and Stump2019; Ford and Nigh Reference Ford and Nigh2015) and would meet the protein requirements of Maya populations.

Chaya flourishes across Mesoamerica as a large shrub with many lush leaves and with multiple sturdy branches that exude a latex sap (see Landa Reference Landa and Gates1937). Recognized as a native crop and identified in the conquest and colonial literature (Colunga-Garcia Marín and May Pat Reference Colunga-Garcia Marín, Pat, Villarreal, Rasmussen, Arias Reyes and Contreras1992; Colunga-Garcia Marín and Zizumbo-Villarreal Reference Colunga-Garcia Marín and Zizumbo-Villarreal2004; Mariaca Méndez Reference Mariaca Méndez2015; Terán and Rasmussen Reference Terán and Rasmussen1994), chaya was then, as it is today, eaten as a green vegetable by the Maya (Landa Reference Landa and Gates1937). It is prepared in soups, tamales, and tortillas, as well as in a fresh green drink and as an infusion (Greenberg Reference Greenberg and Howard2003; Kuti and Torres Reference Kuti, Torres and Janick1996; Standley Reference Standley1930; Standley and Steyermark Reference Standley and Steyermark1949; Sterling Reference Sterling2014:128–132).

The prevalence of chaya in the Maya region today attests to its continued importance. It is found in dooryards, hedges, and around dwellings and is commonly used as a living fence protecting garden plots in the Lowlands of Guatemala, Mexico, and Belize (Everton Reference Everton2012). Its ubiquity in the Lowland Maya area and its incorporation into traditional recipes indicate its deep roots in Maya cooking (Anderson Reference Anderson, Gómez-Pompa, Allen, Fedick and Jiménez-Osornio2003; Coe Reference Coe1994; De Clerck and Negreros-Castillo Reference De Clerck and Negreros-Castillo2000; Everton Reference Everton2012; Fedick Reference Fedick, Freidel, Masson and Demarest2020; Hamman Reference Hamman1998; Lundell Reference Lundell1938; McNeil, Reference McNeil, Hutson and Ardren2020; Steggerda Reference Steggerda1943; Sterling Reference Sterling2014; Terán et al. Reference Terán, Rasmussen and Cauich1998; Williams Reference Williams1981). To determine whether chaya might have contributed protein to ancient Maya diet, we analyzed the δ¹⁵N and δ¹³C of the leaves from modern plants.

Isotopic and Chemical Analysis of Chaya Samples

Chaya leaves were sampled from C. aconitifolius plants growing in home gardens of Santa Elena Town and Santa Familia Village in the Cayo District, Belize. In addition, we purchased fresh chaya leaves in a farmers market in the city of Campeche, Mexico. The leaves were air dried and stored at 25°C in preparation for analysis.

Samples were analyzed both raw and boiled. Three raw leaves from each of the Santa Elena plants were crushed while frozen in a Spex cryomill; the resulting powder was dried at 60°C. Leaves were boiled in water for 10 minutes and dried, reserving the water. The water remaining from the boiling procedure was lyophilized. Leaves were heated five times in a microwave oven for one minute each. The leaves from Campeche and Santa Familia were sun-dried before analysis. To test for intra-leaf variation, pieces were removed from the bottom edge and middle of a leaf, as well as the stem (identified in Table 1 as leaf parts).

All stable isotope measurements were performed using a Costech elemental analyzer (Costech ECS 4010) connected to a ThermoFinnigan DeltaPlus XP. The separated CO₂ and N₂ were carried in a helium stream to the mass spectrometer via a Conflo III coupling. Both carbon and nitrogen were analyzed on the same samples. The results are reported as δ = {(Rsample/Rstandard)-1} × 1,000 where R is either ¹³C/¹²C for δ¹³C, with standard = VPDB, or ¹⁵N/¹⁴N for δ¹⁵N with standard = atmospheric nitrogen (AIR). The precision of both δ¹³C and δ¹⁵N is ± 0.1‰. The mass spectrometric analyses also gave values for the wt% C and N in the samples through measurement of the areas under the peaks for m/e = 44 and 28 (¹²C¹⁶O₂ and ²⁸N₂, respectively). The peak intensities were calibrated by analyses of known amounts of glutamic acid. N wt% values were converted to protein using the conversion factor P = 6.25 N (Elgar et al. Reference Elgar, Hill, Holroyd and Peddie2020).

The isotopic compositions of fresh, boiled, and residual solid extracted by boiling of chaya are given in Table 1. Chaya analyses for δ¹³C exhibit expected levels for a C₃ plant (~ −27‰), but surprisingly high values of δ¹⁵N when compared with values for deer, beans, maize, and amaranth (Figure 5). Untreated raw chaya leaves from Santa Elena are highly enriched in ¹⁵N; boiled leaves are not as enriched: δ¹⁵N = 7.7 (Table 1). Notably, the residue extract from boiling is very rich in protein (45 wt% in dried residue) and is enriched isotopically (δ¹⁵N = 10.5‰). Samples from Santa Familia and Campeche yielded comparable isotopic results for uncooked leaves. Cooked and uncooked, the δ¹⁵N values of chaya are higher than those of the meat of deer (Figure 5), the most common terrestrial source evident in middens (Emery Reference Emery2007).

Figure 5. Isotopic compositions (δ¹³C, δ¹⁵N) of leaves of fresh chaya, boiled, and residue from boiling chaya compared with compositions of other Maya foods, and with the data in Figure 1 translated by the diet-collagen offset: 5‰ for δ¹³C and 3‰ for δ¹⁵N.

The protein content of dried chaya leaves (~30 wt%) is higher than any other plant food known for the Maya and is comparable to the protein content of some animal flesh. The fact that the δ¹⁵N values exceed those of the average terrestrial sources, together with its exceptionally high-protein content, sets chaya apart from other plant foods available to the Maya, past and present.

The Green Deer of the Ancient Maya

Archaeologists have viewed the main source of protein for the Maya as having been the flesh of animals (Emery Reference Emery2007; Scherer Reference Scherer2017; Thornton et al. Reference Thornton, Emery, Speller, Steadman, Matheny and Yang2012). Given the isotopic compositions of the foods consumed by humans (Figure 2), and accounting for the isotopic offsets between collagen and diet due to trophic and biochemical effects by 5‰ for δ¹³C and 3‰ for δ¹⁵N (Schwarcz Reference Schwarcz, Holland and Turekian2014), the average δ¹⁵N value of deer flesh (5.2 ± 0.6‰) is close to known dietary values for δ¹⁵N of humans. The human diet, however, would have been influenced by all food consumed, including low-δ¹⁵N foods (Figure 5), particularly maize and beans (Warinner et al. Reference Warinner Christina and Tuross2013). The average δ¹⁵N value of human collagen of 9 ± 1‰ represents the balance of all sources contributing to the diet. These sources include animal flesh, aquatic and marine foods, low-δ¹⁵N plants, and additional contributions with high δ¹⁵N values to attain the evident averages. We propose that chaya, a plant-based food with high δ¹⁵N, is a potential source of nutrients, including protein, calcium, iron, and vitamin C, for the ancient Maya (Kuti and Kuti Reference Kuti and Kuti1999; Kuti and Torres Reference Kuti, Torres and Janick1996). Chaya would complement terrestrial and aquatic protein sources and match their generally high δ¹⁵N values.

If, as we have suggested, chaya leaves and boiled residue were sources of dietary protein, what effect would this have on the isotopic composition of bone collagen? And, can we distinguish the presence of chaya in human diet from other foods that would have been available to the ancient Maya? As we have noted, variation in the δ¹³C value of human collagen is attributable to variations in the relative amount of C₄ plant foods consumed. The average of −27‰ for chaya is comparable to the general values of all other food plants. Therefore, the δ¹³C values of collagen would not be able to distinguish chaya as a component of the Maya diet.

The δ¹⁵N values of chaya range from 8 to 10‰. As we know, other plant foods known to have been consumed by the Maya, including beans and maize, have low δ¹⁵N values (Warinner et al. Reference Warinner Christina and Tuross2013). Although both beans and chaya were certainly contributors to the protein in the Maya diet, only chaya and fish with high trophic levels could have contributed to the higher δ¹⁵N values in some Maya humans. In fact, a simple mixture of chaya leaves and beans alone could account for the observed δ¹⁵N value in the diet of the Maya (Figure 5). This suggests that chaya could have been a significant component of the Maya diet.

The nutritional properties of chaya support it as a potentially important contributor to the protein supply of ancient Maya communities, just as it is with their contemporary descendants. Beans have always been considered important in the Mesoamerican and Maya diet. Yet, based on their low N isotopic composition, beans could not have been a major protein source for humans. Interestingly, although the Maya had many plant foods available, a diet of beans and chaya could account for the observed range of δ¹⁵N values observed in ancient Maya populations.

Although the extensive data on faunal remains show that Maya were consuming terrestrial animals and some marine and aquatic organisms, we know that the Maya had access to a variety of plant-based foods. New analytical strategies such as aDNA of food residues in the calculus of human teeth (Scott et al. Reference Scott, Power, Altmann-Wendling, Artzy, Martin, Eisenmann and Hagan2021), as well as the expansion of investigations of the roots of chaya, may provide direct data on its use. Undoubtedly, the Maya protein diet contribution would involve a mixture of foods, possibly including meat, chaya, and other plant foods grown from home gardens and agricultural fields that provided subsistence for the population. Managing the complex Maya forest landscape, the Maya clearly devised a stable basis for access to protein and essential amino acids that came from both animals and plants. Chaya is recorded at the time of conquest, and the evidence points to its origins and domestication over thousands of years. This suggests that chaya could have been among the important food sources of the Maya. The domestication and use of this nutritious perennial vegetable would have been consistent with the extensive environmental knowledge of traditional Maya farmers. The ubiquity of chaya in the Maya area today and its obvious potential as a green protein source situate it well as a component of the complex, diverse, and multifaceted diets characteristic of ancient Mesoamericans.