*3.1. Ceramic Paste Sourcing Results*

We discovered that Wari ceramics from Cerro Baúl were chemically distinct from contemporary materials in the local region, including both Tiwanaku and local sherds.They were also distinct from a control group of ceramics excavated at the Wari capital in Ayacucho. Statistical analysis of the INAA data resulted in the definition of three chemical groups from the Wari heartland (Wari-1, 2, and 3), one chemical group concentrated on the Cerro Baúl site summit (Baúl Ref.), and seven groups of local ceramic production (Mejia A-G) [19].In the INAA data, a biplot on the first two principal components calculated from the variance–covariance matrix of the ceramic samples shows the brewery ceramics can be distinguished from both Wari heartland materials and other local ceramics on an axis of variability associated with the second principal component (Figure 4).

**Figure 4.** PC1 v PC2 for INAA data from Baúl, Mejia (local) and Wari heartland ceramics [19].

This component is positively loaded on Na, Sr, Ca, and Th and negatively loaded on Fe, V, Co, Sc, Ti, Sb, Cr, As, and Zn. A biplot of Cr and La distinguishes the Baúl group from the local Mejia pastes (Figure 5).

**Figure 5.** Chromium v Lanthanum plot for INAA data [19].

In the ICP-MS data, the distinction of the Cerro Baúl brewery ceramics is also pronounced, but this time the second principal component, which together with PC1 accounts for 47% of the total variance, is loaded positively on Be, K, Nb, Rb, U, and Th. It is negatively loaded on Fe, Cu, Sr, V, Co, Mg, and Ca [22]. Analysis of local clay sources indicate that the Baúl ceramics match most closely with a clay source across the valley from the site. The local Tiwanaku and Mejia ceramics best match clay sources from geological formations in the Torata and Moquegua valleys themselves [22]. None of the archaeological ceramics from the Moquegua region were chemically similar to the three sources identified from the Wari heartland.

Ceramic samples from the brewery (*n* = 20) represent 25% of the collection of pottery sourced by INAA on the summit of Cerro Baúl (*n* = 80). While the Baúl chemical group accounts for about 70% of the summit assemblage, it accounts for all but two samples from the brewery. Thus, 90% (18/20) of the brewery ceramic vessels are from the Baúl ceramic chemical group (Table 1). Most of those are serving wares constituted by bowls and drinking cups. It is apparent that the brewery was a key consumer of the special ceramic workshop production represented by the Baúl chemical group. The palace and adjacent consumption contexts on the summit of Cerro Baúl near the brewery were also key consumers of the Baúl workshop ceramic wares. Nash [6] identified the probable location of a ceramic workshop in the palace on the summit of the mountain that is the likely locale of production for these specialized ceramic vessels.

Off the summit of Cerro Baúl, the finely made and decorated Baúl group ceramics are very rare [6]. At the adjacent Wari site of Cerro Mejia, only 6% of the samples match the Baúl group, while in local settlements on the slopes of Cerro Baúl, only 2% (a single vessel) of the ceramic assemblage is part of the Baúl group [19]. Outside of these sites immediately adjacent to the Wari administrative center at Cerro Baúl, none of the other Moquegua valley samples match the Baúl chemical group. In fact, greater than 90% of all these ceramic assemblages were produced from Moquegua or Torata valley clay sources with the exception of the elite wares that are part of the Baúl group from the summit of the mountain.

Thus, the archaeometric data indicate that the brewery ceramics were produced from a specialized set of raw materials, dedicated to political consumption events in ritually charged monumental architecture on the summit of Cerro Baúl. We previously identified the likely clay source for that ceramic production on a mountain across the valley from the site [22]. Interestingly, that clay source is uniquely used by the ceramicists at Cerro Baúl and is not used by any other contemporary groups, despite the high quality of the clays. This likely demonstrates an exclusionary control over the access to the clay sources by the Wari elite.

#### *3.2. Residue Analysis Results*

The results of the residue analysis of the ethnographic ceramics by DART-MS are presented in Table 2. Several compounds from the *Schinus* database were found in a control vessel that contained none of the *chicha* recipes and were thus excluded from further results as being not indicative of the presence of *Schinus molle* residues. The base peak of all of the *chicha de jora* samples (even after washing) is observed at *m*/*z* 341.108. Of the possible formulas calculated for this mass was C19H17O6, which would correspond to a number of tetrahydroxyflavone compounds, or C12H22O11, which would correspond to a disaccharide like maltose. While the calculated masses for each of these compounds is different, it differs only in the third decimal place, meaning that within the mass resolution of the AccuTOF mass spectrometer, we cannot differentiate between these compounds. Maltose is the most likely source of that peak.

The *chicha de molle* experimental samples contained few of the compounds that were both found in the literature for *Schinus* species and could be ionized by DART-MS. Fewer yet were identified in the *chicha* with both corn and *molle*. The potentially relevant compounds that could pertain to *molle* which seem to have been identified in ethnographic controls of (1) *chicha de molle* alone and (2) *chicha de molle* with maize were respectively: (1) terebanene, teredenene, or β-spathulene (a sesquiterpenoid); safrole (a phenylpropanoid); and gallic acid (a phenolic acid) and (2) terebanene, teredenene, or β-spathulene (a sesquiterpenoid). None of these compounds alone or altogether as an assemblage affords a firm identification of *molle* to these ethnographic controls, but all are consistent with *molle*. Indeed, the DART mass spectra of the *molle* drupes showed all of these compounds, as well as a compound with *m*/*z* 453.332, which corresponds to the [M−H]<sup>−</sup> ion for the formula C30H46O3. This is most likely either (1) masticadienoic acid (also known as terebinthone), (2) a positional isomer of that compound such as isomasticadienoic acid, or otherwise (3) moronic acid; these compounds are indistinguishable based on the molecular mass. The DART mass spectrum of the drupes from the Cerro Baúl excavation also showed an additional signal most likely arising from β-sitosterol (*m*/*z* 413.380, C29H50O [M−H]−).


**Table 2.** Experimental ceramic residues on vessels from DART-MS.

Identifications of triterpenoids that are chemotaxonomically characteristic especially of certain anacardiaceous resins (*Pistacia* spp., *Rhus* spp., and *Schinus* spp.) are particularly heartening. In J. Henkin's view these triterpenoids are, where present in the DART-MS spectra, the best phytochemical evidence for *chicha de molle* production from the archaeologically recovered ceramics and *molle* drupes analyzed (Figure 6; Table 3). Hydroxymasticadienoic acid, (iso)masticadienoic acid, moronic acid, schinol, and/or simiarenol together as this relevant triterpenoid group underlie a solid chemotaxonomic argument for *molle* processing and *chicha de molle* production at Cerro Baúl. These triterpenoids should be regarded as *molle* biomarkers in further research along these lines (Figure 7). Sesquiterpenoid and triterpenoid biomarkers identified, plus the presence of the phenylpropanoid safrole, support the archaeological evidence for *chicha de molle* production at Cerro Baúl.

**Figure 6.** *Cont*.

**Figure 6.** Example DART-MS spectra for one replicate each of the Cerro Baúl brewery ceramic residue samples.

**Table 3.** Ceramic residues on archaeological sherds from DART-MS (nd = not detected). Each + indicates a positive identification (based on mass for the [M–H]− ion) in one replicate. Thus ++ means that the compound was identified in two replicates. For CB89-1221, only two replicate analyses were carried out.


**Figure 7.** Triterpenoid compounds identified by DART-MS in *molle* samples, diagnostic especially for *Schinus* spp. and its close relatives [26].

The results in Table 3 show only the compounds that, based on literature sources, are indicative of *Schinus molle* and that are not found in the control samples that contained only corn-based *chicha*. In none of the spectra were the identified *Schinus* compounds the most intense signal, referred to as the base peak of the spectrum (Figure 6).For two of the sherds the observed base peak was at *m*/*z* 220.150 (CB04-01-0530 and CB89-1200), a compound that we were unable to identify, but appears to give primary fragments at *m*/*z*148 and 205 when Orifice 1 on the AccuTOF is set to −90V.The intensity of this signal may indicate it is a contaminant from the environment or handling. For the remaining three sherds, the base peak was most likely a fatty acid. These include myristic acid (tetradecanoic acid, *m*/*z* 227.201) in replicates 1 and 2 of CB89-1110, palmitic acid (hexadecanoic acid, *m*/*z* 255.232) in the third replicate of both CB89-1110 and -1233, and the shorter-chain pelargonic acid (nonanoic acid, *m*/*z* 157.123) in both replicates of CB89-1221 and the second of -1223. None of these fatty acids are observed in significant quantities (based on signal intensity) in the control vessels.

Interestingly, the medium-chain fatty acids are absent in the ceramic vessel that contained the ethnographic *chicha de jora* but dominate the spectra of the ones containing *chicha de molle*, even when corn was also present. Caproic, caprylic, and pelargonic acids (hexanoic, octanoic, and nonanoic) are the top three intensity peaks in the DART-MS spectra of all of vessels that contained the ethnographic *chicha* made with *Schinus molle* berries. Medium-chain fatty acids, while not anticipated by the initial chemoinformatic search that has driven the residue analysis aspect of this research into the material culture of Cerro Baúl, do appear prominently in ethnographic and archaeological *chicha de molle* DART-MS spectra. Truly accounting for the fatty acids' prominence in terms of the phytochemistry of *S. molle* itself, the beverage processing, and/or other environmental inputs would be a sensible focus for any follow-up research. These fatty acids, specifically their isotopic composition, are an excellent target for further analyses. Confirmation of the exact identities of the terpenoids and phenolics would further increase confidence in the identification of the *molle* residues.

#### **4. Discussion**

How do unique or rare materials contribute to sustainable governance? In reality, it is the know-how in this case that creates the unique experience, not the material itself. Clay sources utilized to produce brewing vessels may be a limited resource or may be distant or difficult to procure. But in the case of Cerro Baúl, the clay source identified by the chemical data was abundant and located less than half a day's walk from the brewery. Wari ceramic specialists did establish a privileged access to this source and created a recipe for working the material. Yet, they did not overexploit the source nor did they focus on incorporating materials that were not readily available in the local area.

The clay itself was a local resource, and it was particularly reserved by the Wari for their specialized workshop in the region. Wari ceramic specialists did not rely on extra-local materials (except perhaps for some pigments) for their ceramic production, nor did they rely on imported ceramic vessels with their ritually charged iconography. They decorated these vessels themselves with highly significant ritual iconography, representing Wari supernatural beings and geometric motifs. The means of forming and decorating these vessels replicated ways from the Wari heartland and did require specialized knowledge. That knowledge was reproduced locally to create the special vessels for the feasting events. Having a local source for the ceramics made the Baúl brewery a sustainable producer of both the containers for production and consumption of the beer itself since they were not reliant on foreign materials.

Yet, these ornate, regionally produced ceramic vessels did signal strong affinity with the imperial center. They were not dependent on external producers distant from the brewing and feasting locales in order to provide the vessels for these political events. They were not subject to disruptions in long distance supply chains, external political interference, nor disasters in distant production locales for these important constituents of the Wari feast. They were, however, dependent on the Wari specialist potters who knew how to create the Wari vessel forms and to decorate them with the highly stylized ritual iconography. These specialists had to experiment with local resources: tempers, clays, and pigments, in order to transform those raw materials into legitimate and authentic Wari ceramic wares. This local production of highly specialized containers provided a high degree of locally sustainable fabrication of genuine Wari ritual receptacles. That was critical to the reproduction of the Wari political economy, not only at Cerro Baúl, but likely in every major Wari center throughout the empire.

The other material components of crucial importance to the brewery's production were of course the grains and fruits fermented to make the political beverage, selected by the Wari specialist brewers for their recipes. Corn is a fairly water intensive crop to produce, though yields are quite high. Other native grains like quinoa have lower yields (500 kg per acre), but may be less water intensive. In modern farming, corn generally requires more water per field than many other native Andean crops, including other grains, tree fruits, potatoes, and other potential *chicha* producing ingredients. Modern corn, though, does provide high calories per unit area (up to 15 million calories per acre) and thus produces significantly more energy despite its higher water requirements.

The pepper berries are a renewable resource as well, available on trees that naturally form hedgerows and are commonly available on the edges of cultivated fields. Growing to heights up to 15 m, they are also hearty and resistant to drought conditions. In an environment of climatic variability, they are a raw material resource that is resilient in its availability. While the fruits are especially prevalent after extended periods of heavy rain, they are present on trees in the study area year round [9]. Thus, while maize is a fairly water intensive crop, *molle* is a drought resistant tree that consistently produces fruit. In addition to the cultural properties *molle* contributes, it also provides a consistent ingredient for *chicha* that is drought tolerant.

The Middle Horizon (600–1000 CE) was generally a period of climatic stability with some variability in precipitation pronounced in later times. Wari terracing technology provided some mitigation of drought conditions through more efficient water use [28].Towards the end of the period, increased variability in precipitation lead to localized conditions of water deficiency, which in the succeeding centuries became severe and prolonged droughts [29].The use of *molle* as a *chicha* product seems to be pronounced in the final years of the Baúl brewery's production, slightly after 1000 CE, evidenced by the fact that the residue results presented here come from ceramics used in the ceremonial closure of the brewery and that *molle* dumps are associated with the final use of the floor in the brewery. This may reflect the reliability of *molle* as a resource, even in times of drought stress.

Thus, it was both local resourcing of ceramic vessels for brewing and serving the alcoholic beverage and a focus on a set of brewing ingredients that were locally produced and resilient to climatic variations in production that created a sustainable brewing operation for four centuries. These variables were incredibly important to the Wari political economy and to the building of local allegiances that sustained imperial relationships over decades. Archaeometric data confirms that local source material was utilized in both ceramic production and beverage fabrication. Yet, it was the ceramicists' specialized knowledge on how to construct and decorate the elaborate drinking vessels that made these events quintessentially Wari. It was also the brewers' knowledge of the culinary recipes of different *chicha* production strategies across dispersed areas of the Andes mountains that brought that shared political identity to strangers in many lands. Local resourcing joined with shared knowledge over vast areas to create the political unity that Wari represented as the Andes' first empire.

**Author Contributions:** Conceptualization, P.R.W. and D.J.N.; Formal analysis, P.R.W., D.J.N., J.M.H., and R.A.A.; Funding acquisition, P.R.W., D.N., and R.A.A.; Investigation, P.R.W., D.J.N., J.M.H., and R.A.A.; Methodology, P.R.W., J.M.H., and R.A.A.; Writing—original draft, P.R.W., D.J.N., and J.M.H.; Writing—review & editing, J.M.H. and R.A.A.

**Funding:** Field research on the Cerro Baúl brewery was funded by the National Science Foundation (BCS-0074410, BCS-0226791), the National Endowment for the Humanities (RZ-50098), the G. A. Bruno Foundation, and the Grainger Foundation. DART-MS analyses were carried out on EMU's AccuTOF mass spectrometer, funded under NSF MRI-R<sup>2</sup> (0959621).LA-ICP-MS analyses were carried out on TFM's Varian inductively-coupled plasma-mass spectrometer funded under NSF MRI (0320903).

**Acknowledgments:** INAA was conducted at MURR under the direction of Michael Glascock and preliminary interpretation of results was done by Jeff Speakman. LA-ICP-MS was conducted at the Field Museum, with contributions by Nicola Sharratt, Laure Dussubieux, and Mark Golitko. Archaeological materials were exported from Peru under permit from the Ministry of Culture acuerdos 1659/792 on 18 December 2006 and 123 on 2 October 2003.

**Conflicts of Interest:** The authors declare no conflict of interest.
