Introduction

Colonial encounters inevitably unsettle—and often upend—existing patterns of social labor among settler and indigenous communities alike (Lightfoot et al. 1998; Silliman 2001, 2006; Deagan 2003, 2004; Lightfoot 2005; Voss 2008a, 2008b). By tracing changes to the organization of labor from initial contact through the course of the encounter, archaeologists may trace the broader shape of the encounter itself. We use construction practice, specifically the domains of wood selection and wood preparation, to evaluate changes in the organization of labor during the occupation of Fort San Juan de Joara, located at the Berry site in the western piedmont of present-day North Carolina. Joara was a leading ancestral town of the modern Catawba Indian Nation, and in the mid-16th century it had nominal authority over neighboring towns and communities in its vicinity (Beck and Moore 2002; Moore 2002; Beck 2013). In December 1566, Captain Juan Pardo established Fort San Juan at Joara and garrisoned it with 30 men. This fort, together with its adjoining domestic area, was the first European settlement in the interior of the present-day United States (Beck et al. 2006:65).

Much of our prior research at the Berry site has focused on the archaeology of daily life in the domestic contexts of Fort San Juan, a group of five burned structures and associated features that we refer to as the Spanish compound. To date we have excavated two of these buildings, Structures 1 and 5, which together provide data on the techniques and architectural grammars (Glassie 1975) that guided the compound’s construction. The remarkably well-preserved wood remains from Structures 1 and 5, a class of materials rarely available for archaeological analysis, offer unique insights into construction practices and the organization of labor during a very early colonial encounter. Analysis of these remains—posts, timbers, and smaller architectural elements—together with close scrutiny of construction details, can illuminate culturally specific architectural practices. These practices, in turn, offer insights both into how the structures were built and into the identities of the builders themselves. Such an approach allows us to reconstruct initial social dynamics between the two groups, including some sense of the degree of coercion or cooperation between the Spaniards and their hosts, and to evaluate whether and how colonial relations at Joara changed through time.

Joara and Fort San Juan

During the first half of the 16th century, Spanish explorers failed in several efforts to colonize what is now the southeastern United States. Finally, in 1565–1566, Pedro Menéndez de Avilés established a pair of new settlements on the south Atlantic coast—St. Augustine, founded in September 1565 in Florida, and Santa Elena, established in April 1566 on present Parris Island, South Carolina. The latter settlement, Santa Elena, was to be the principal town of Menéndez’s colonial aspirations (Lyon 1976, 1984; Hoffman 1990; Hudson 1990; Paar 1999). When King Philip II learned of Menéndez’s actions, he ordered reinforcements for the new colony. In July 1566, Captain Juan Pardo arrived at Santa Elena with a company of 250 soldiers and began to fortify the settlement. As the Santa Elena colony was not prepared to feed this large contingent of men for very long, Menéndez ordered Pardo to prepare half his army for an expedition into the interior lands that lay behind the Atlantic coast. Pardo’s task was to explore this region, to claim the land for Spain while pacifying local peoples, and to forge an overland path from Santa Elena to the silver mines in Zacatecas, Mexico. Pardo departed with 125 men on 1 December 1566.

Later that month, after traversing the Carolina Piedmont along the Wateree and Catawba rivers, Pardo and his soldiers arrived at Joara, a large Native American town in the upper Catawba Valley near the eastern edge of the Appalachian Mountains (DePratter et al. 1983; Hudson 1990). The leader of Joara, referred to in the accounts as Joara Mico, held some authority over other towns along the upper Catawba and its tributaries (Beck and Moore 2002:201), mico being a native term for regional or multi-community chief (Hudson 1990; Anderson 1994). Pardo renamed this town Cuenca, after his own native city in Spain. At Joara, he built a fort, christened San Juan, that he manned with 30 soldiers. While previous expeditions into the interior had either founded seasonal encampments or had temporarily occupied native towns, Pardo explicitly established Fort San Juan to expand Santa Elena’s colonial reach into the northern frontiers of La Florida (Beck et al. 2006). During the course of a second expedition in 1567, Pardo would go on to establish five more forts between the Atlantic and what is now eastern Tennessee, yet Fort San Juan de Joara was to be the center of his imperial designs (Fig. 1).

Fig. 1
figure 1

Towns visited by Juan Pardo’s second expedition, 1567–1568. (Map by Berry Site Archaeological Project, 2015.)

Full size image

Throughout most of the 18 months that Spanish soldiers lived at Joara, amicable relations existed between the people of this town and their European guests. On at least two occasions, for example, the Spaniards accompanied native warriors in attacks on hostile native chiefs across the Appalachians in Tennessee and Virginia (Martínez 1990; Beck 1997). Likewise, as Pardo was preparing to leave from Joara during his second expedition, he ordered the ensign, Alberto Escudero de Villamar, to “judge and have a care of the conservation of the friendship of the caciques and Indians of all the land” (de la Bandera 1990:278). Yet, in the months that followed, relations between the soldiers at Fort San Juan and the town of Joara took a calamitous turn. By May 1568, word reached Santa Elena that Indians had attacked all six of Pardo’s interior garrisons, including Fort San Juan, and that all were lost (Hudson 1990:176). Several factors must have played a role in this decisive action, but two stand out: the soldiers’ demands for food and their relations—whether coercive or consensual is unclear—with Indian women (Hudson 1990:176). Before leaving Fort Santiago de Guatari, for example, Pardo had ordered his officers “that no one should dare bring any woman into the fort at night ... under pain of being severely punished” (de la Bandera 1990:285). In the end, 130 soldiers and all of Pardo’s isolated forts were lost, and with them Spain’s last attempt to colonize the northern frontiers of La Florida. Indeed, it was more than a century before other Europeans are known to have penetrated as far into the southern Appalachians (Beck 2013).

Building Houses for the Pardo Expedition

During the past 20 years, historical archaeologists have developed a deep appreciation for the significance of labor and labor regimes in the colonial process, from forms of outright slavery to different modes of servitude, concubinage, and conscription (Deagan 1996, 2003; Hackel 1998; Lightfoot et al. 1998; Silliman 2001, 2004, 2006; Lightfoot 2005, 2015; Voss 2008a, 2008b; Jordan 2009; Van Buren 2010). Silliman (2001) has insisted that such labor be regarded as more than merely a matter of economy, but as embedded in a broader context of social relations. He argues that what distinguishes socially situated labor from work is that the former has the potential to be “appropriated and enforced” (Silliman 2001:380). Thinking of labor as a social phenomenon grounded in everyday lived experiences

allows archaeologists to see the ways that administrators, overseers, capitalists, managers, and supervisors structured and often imposed labor and the ways that those laboring accommodated, resisted, made use of, and lived through labor situations. Labor marks a shared, yet multiply interpreted, experience. (Silliman 2006:149)

Drawing on theories of agency and structure, Silliman (2001:381–384) advocates an emphasis on labor as practice—as culturally meaningful work that people enact under particular historical and social conditions—and for linking the study of labor to related domains like gender and power.

We firmly support this approach to labor as practice, yet we caution against the tendency to reduce indigenous agency in colonial labor regimes to either acts of resistance or acquiescence (Silliman 2001:402). This is particularly important in the analysis of early colonial contacts and encounters, when roles of dominance and submission were more easily, and often unexpectedly, reversed (Wilcox 2009; Liebmann 2012). Otherwise, scholars risk assuming that, in any colonial interaction, it was inevitably native peoples who resisted Europeans and, more, that the arrival of Europeans was inevitably disastrous for all native populations. While true in the macronarrative of the Columbian exchange, it is less helpful for understanding any particular context of colonial entanglement and can obfuscate variability in the kinds of labor relations that developed between indigenous groups and Europeans at different times and places. In some cases, such as the Pardo expeditions, the Europeans’ arrival played into the hands of Indian communities engaged in their own local contexts of competition, alliance, and exchange.

At several of the towns that Pardo visited during his first expedition, he specified that the caciques should build houses for the soldiers’ use—presumably as way stations along the planned route to Mexico—and then fill some part of them, such as an elevated room, with maize, or else build a smaller structure like a corncrib nearby. When Pardo retraced his path several months later, his notary Juan de la Bandera observed that large new houses of wood or of wood and earth had been built for the army in all of the towns in which Pardo had issued the command, and even in a few in which he had not. Both Pardo and Bandera interpreted these buildings as clear evidence of Indian submission to the Spaniards’ authority. In his cursory report of the first expedition, Pardo notes that many Indians had gathered to meet him at Canos, where he "made them the customary speech on behalf of God and His Majesty. They were very content and obedient to the service of God and His Majesty" (Pardo 1990:311). In September, when Pardo returned to Canos with his second expedition, Bandera observed that

before now it had been declared and said unto them by the captain how they were to make the house and maize which they had for His Majesty, and that it was suitable for them to turn Christian and be under his dominion ... and ... now that His Grace saw that they actually fulfill the things stated above, it is necessary that in the presence of me, the notary, they ratify and approve it. They ... declared and said that they will be very happy to do that which has been commanded and declared to them by the captain in His Majesty's name and as such, remaining ... under the said dominion, they made the "Yaa," which I, the notary, attest. (de la Bandera 1990:260)

Subsequent events would emphatically reveal the degree to which these Spaniards were mistaken about native submission to “His Majesty” at Canos and other piedmont towns. To better appreciate the reasons for this disastrous misunderstanding, it is necessary to look more broadly at the indigenous political context for house construction (Beck et al. 2011).

The town of Canos—also known as Cofitachequi—was one of the most important places in which Pardo specified that a house be built for the Spaniards’ use. As many as a dozen caciques met with Pardo at Canos during his first expedition (de la Bandera 1990:260), and it was expected that all of these local chiefs, or orata, would assist in constructing the house at Canos and would help to fill a portion of it with maize. When Pardo returned there during the second expedition, he met again with many of these leaders, all of whom wanted credit for having done their part in making the house. One of these, Ylasi Orata, came to inform Pardo that he had helped to build the house at Canos, yet he was holding his maize contribution to the Spaniards in a house that he had made for that purpose in his own town. In 1540, when Soto crossed the same region, Ylasi was subject to Canos (the latter keeping a corncrib there to store its tribute), but by building his own structure and filling it with maize, Ylasi Orata was demonstrating both to Pardo and to Canos Orata that he was no longer obligated to pay tribute at another town. Other towns along Pardo’s path probably had a similar motivation; specifically, making a house for Pardo’s expedition became a means for town and village leaders to demonstrate autonomy from other native towns and polities, and was more a response to contested local relations than a bow to Spanish authority. Indeed, doing so gave native leaders an opportunity to enmesh Pardo and his forces in local politics from a position of confidence or power.

Beyond the opportunity to demonstrate autonomy, building a house and providing it with a store of maize also permitted some caciques to coordinate labor on a regional scale and to show their capacity for such to other native leaders and to the Spaniards. At Canos, Guatari, and Joara, for example, construction and agricultural labor was drawn from multiple towns and villages. As noted, no fewer than a dozen town caciques met Pardo at Canos, seven orata and their associates came to Guatari when summoned by its mico, and at least as many traveled to Joara. The ability to draw from a regional labor pool was not shared by all the chiefs who built houses for Pardo and his soldiers (for example, those from smaller towns like Ylasi, Tagaya, Otari, and Aracuchi). Thus, the act of building a house for the Spaniards also gave some towns—particularly those such as Joara and Guatari that seem to have lacked the status of Canos (Cofitachequi) during the time of the Soto expedition—an opportunity to improve their positions at the regional scale.

Therefore the Pardo expedition can be viewed as a kind of novel resource, one with the potential to be mobilized by indigenous communities toward a range of their own political, economic, and military ends. It is especially important to recognize, we believe, that the peoples of the Carolina Piedmont—from their micos and oratas on down—had no knowledge of the conquests that Spain’s military machinery and colonial institutions had accomplished in other parts of the Americas. As such, they had little reason to treat Pardo and his soldiers as representatives of an extensive and powerful empire or to gauge their own actions accordingly. Instead, they understood the occupation of their communities by these Spaniards in terms of their own cultural knowledge and experience and acted accordingly. The motivations of leaders like Joara Mico are best understood by focusing on the local contexts of their political actions.

Household Archaeology at the Berry Site

Written accounts of the Juan Pardo expeditions, as essential as they are for reconstructing Pardo’s route across the Carolina Piedmont and the political geography of its native peoples, offer the reader almost nothing about daily life at Fort San Juan. To understand how daily practice shaped the contours of this colonial encounter, as well as how these contours changed during the 18 months that Spanish soldiers occupied Joara, the archaeological data must be reviewed. Archaeological and historical evidence (Beck 1997; Beck and Moore 2002; Worth 2016) clearly indicate that the Berry site (31BK22) is the location of Joara and Fort San Juan. Berry is located along Upper Creek, a tributary of the upper Catawba River, in what is now Burke County, North Carolina (Fig. 2). Today the Berry site covers approximately 4.5 ha and is situated along the easternmost margin of a 75 ha alluvial floodplain at the junction of Upper and Irish creeks. Systematic surface collections indicate that Berry was one of the largest Late Mississippian (A.D. 1400–1650) sites in the upper Catawba Valley (Beck and Moore 2002:200; Moore 2002:61). Berry was briefly noted in Cyrus Thomas’s Catalogue of Prehistoric Works East of the Rocky Mountains as a "[m]ound on the west Bank of Upper Creek 8 miles north of Morganton (about 15 feet high and unexplored)" (C. Thomas 1891:151). Both the earthen mound and the surrounding site were regularly plowed, while the mound itself was eventually bulldozed to fill a low-lying area west of the site—probably a borrow pit—that was prone to flooding.

Fig. 2
figure 2

Berry site excavations, 1986–2013, indicating the Spanish compound, mound areas, and the Fort San Juan moat. Note that the north arrow on this and all subsequent maps indicates grid north; the 1986 excavations are oriented toward magnetic north. (Map by Berry Site Archaeological Project, 2015.)

Full size image

Our fieldwork at Berry has extended over 17 seasons (1986, 1996–97, 2001–2014), a total of 97 weeks of excavation and survey. Excavations to date total over 1400 m2 and concentrate on a 0.3 ha area where we have recovered a significant assemblage of 16th-century Spanish ceramics and hardware, along with personal and military artifacts (Beck et al. 2006; Rodning et al. 2016). This area on the northernmost margin of the Berry site is associated with a cluster of five burned structures, dozens of pit features, and hundreds of postholes that we refer to collectively as the Spanish compound (Fig. 3). These remains constitute the material footprint of Pardo’s Cuenca. In 2013, we identified the dry moat of Fort San Juan about 20 m south of the compound, such that the fort was situated between the Spaniards’ domestic zone and the town of Joara (Fig. 2). None of Pardo’s other settlements has yet been discovered.

Fig. 3
figure 3

Plan map of the Berry site Spanish compound with all excavated features numbered. (Map by Berry Site Archaeological Project, 2015.)

Full size image

The Spanish compound contains at least five large, set-pole structures, four of which were built in semisubterranean basins, and all of which were burned. To date we have investigated two of these buildings, Structures 1 and 5, the excavation and analysis of which provide much of our data for understanding construction techniques and practices (Beck et al. 2016). Structure 1, the larger of the two, measured about 7.5 m on a side (56 m2) and was among four structures built in semisubterranean basins. Such basins were typical of Late Mississippian architecture across the South Appalachians (Lacquement 2007b; Hally 2008), and, indeed, much of Structure 1 was typical of indigenous-style architecture across the region. It contained a central hearth, four large and deeply placed interior support posts, and a well-defined entryway with exterior wall trenches at the western corner of the structure (Fig. 4).

Fig. 4
figure 4

Structure 1 excavations, 2008. (Map by Berry Site Archaeological Project, 2015.)

Full size image

Structure 5, with an area of 49 m2, was less typical of native architecture. Although its central hearth was similar to that of Structure 1, it was not constructed in a basin, its entryway was less clearly marked, and its four central posts were placed in very shallow postholes, such that they that offered insufficient support for the structure’s roof; two additional interior posts were later added to stabilize the building (Fig. 5). Compounding this structure’s relative instability was the fact that many of its exterior posts were set into postholes that were likely dug with metal shovels and were, therefore, much wider than the posts themselves (Beck et al. 2016:102). In Structure 1, however, the exterior and center posts were set into postholes excavated by digging sticks and by ramming the posts into place, creating postholes with about the same diameter as the posts. Both structures—as well as the three that remain unexcavated—burned at apparently the same time, and there is no evidence that any was ever rebuilt.

Fig. 5
figure 5

Structure 5 photomosaic, 2008. (Map by Berry Site Archaeological Project, 2015.)

Full size image

Since 1986 we have identified more than 200 features at the Berry site, of which we have excavated more than 60 (Beck et al. 2016). Most of these identified and excavated features are inside the Spanish compound, and here we focus on the sample of features containing European artifacts or copper fragments that probably derive from non-aboriginal sources. To date, project archaeologists have excavated 16 such features inside the compound, and these are subdivided into two distinct clusters referred to as the central features and the western features. Most of these contexts, especially those identified as the central features, probably began as daub-processing pits during the process of house construction, after which they were filled with domestic refuse. Others represent midden deposits near Structure 5 that also contain large amounts of domestic refuse.

We propose two phases of primary construction activity in the Spanish compound. In the first phase of occupation, three structures—Structures 1, 3, and 4—were built in a slightly curving arc near the northern edge of the site (Fig. 6). The beginning of this first phase probably coincides with Pardo’s arrival at Joara in December 1566. Indeed, when Pardo returned to Joara in September 1567, his notary Bandera reported that “he found built a new house of wood with a large elevated room full of maize, which the cacique of the village ... had built by the command of the captain for the service of His Majesty” (de la Bandera 1990:265). The central pits were probably dug at the same time that Structures 1, 3, and 4 were constructed and may have initially served as pits for processing daub. After the initial phase of building construction was complete these pits became the main locus of refuse disposal in the new compound. Feature 76, a circular hearth 66 cm in diameter and about 10 cm deep, was situated at the eastern edge of the central pits and may have served as an open-air kitchen during this first phase of the compound’s use.

Fig. 6
figure 6

Plan map of the Spanish compound, first-phase buildings and features with possible kitchen indicated. (Map by Berry Site Archaeological Project, 2015.)

Full size image

During the second phase of occupation (Fig. 7), Structures 2 and 5 were built together along the same axis about 15–20 m west of the first-phase structures, and both new buildings intrude into the central pits. Structure 2 was built in a basin of about the same depth as Structures 1, 3, and 4. Structure 5 was thus an anomaly, and we propose that it was used as a formal cocina or kitchen during the compound’s second phase, replacing the informal Spanish kitchen associated with Feature 76 (Beck et al. 2016). Alone among those structures in the Spanish compound, Structure 5 was adjacent to active trash pits and a large sheet midden (the western features), the only such deposit identified in the compound area.

Fig. 7
figure 7

Plan map of the Spanish compound, second-phase buildings and features with possible kitchen indicated. (Map by Berry Site Archaeological Project, 2015.)

Full size image

In the following section, we use data on wood selection and preparation from Structures 1 and 5 to evaluate three scenarios of construction practice in the Spanish compound: (1) Structures 1 and 5 were built by people from Joara on behalf of the Spaniards, (2) both structures were built as a joint effort between the two groups, or (3) they were built by Pardo’s soldiers using their own tools and architectural knowledge. Some implications of the three scenarios are clear. If people from Joara and nearby towns constructed both buildings using indigenous tools, knowledge, and practices, then the specific woods used and the methods of their deployment should be consistent with data from precolonial or otherwise exclusively Native American structures. In this case, all tool marks associated with wood and bark reduction should be consistent with native technology, particularly with the stone axes, adzes, and other woodworking implements that were part of the indigenous tool kit (Lacquement 2007a, 2007b; Hally 2008).

However, if the building effort was cooperative, with the native Joarans working together with the Spanish soldiers, then evidence of mixed technologies and practices—including at least some signs of the use of European carpentry tools—might be expected, particularly implements with metal blades or bits. Finally, if these two buildings were constructed by expedition members working alone, then they should contrast with and deviate from native constructions at least to some degree and perhaps considerably. This might take the form of raw-material selection and reduction, joining, tooling marks, and other carpentry details consistent with European practices of the time; particularly relevant here is mention in the Pardo documents of the European carpentry tools, such as metal axes, chisels, nails, and saws, with which the expedition was outfitted (Hudson 1990:45,149–150,152).

Wood Selection and Preparation in Structures 1 and 5

The two sample sets considered here include wood and other organic materials recovered from the two buildings, including 74 samples from Structure 1 (Fig. 8) and 122 from Structure 5 (Fig. 9). Field-identified structural elements include upright posts and other large timbers, as well as a variety of what are likely secondary (rafters and minor wall supports) and tertiary elements (cane wattle siding/framing, bark siding or roofing material, packing and insulation between roofing or timbers) used in construction. On the basis of their size, field-identified specimen type, and integrity, we assigned selected organic remains unique field numbers for which we used the designation OG. Specimens receiving an OG number were removed, wrapped in archival-quality foam, and packed in boxes for transport to the laboratory. Individual specimens were classified according to five form categories, partly in conjunction with field identifications: upright posts, roundwood, lumber, secondary elements, and tertiary elements. Roundwood refers to radial (cylindrical) segments of branch or stem from trees and shrubs; individual segments can be of any length and diameter. All of the upright posts are also classified within this roundwood category, as are many other timbers from the site. Each roundwood specimen was also classified according to its relative completeness with respect to original circumference, whether whole (full circumference), half, or finer segments.

Fig. 8
figure 8

Structure 1 floorplan with all individual organic samples (OGs) plotted. (Map by Berry Site Archaeological Project, 2015.)

Full size image
Fig. 9
figure 9

Structure 5 floorplan with all individual organic samples (OGs) plotted. (Map by Berry Site Archaeological Project, 2015.)

Full size image

All samples were examined for the presence of tooling marks and any other indication of the particular reduction/carpentry techniques used in procurement and preparation. Two basic tool-mark types were recognized. Facets are prismatic-shaped marks, typically longer than wide, that usually slope upwards toward the end of the specimen. Depending on the tool bit used, they can be relatively flat surfaced or concave in form. Bevels are sloping surfaces or edges resulting from straight angular cuts and are commonly wider than long; these, likewise, may be quite flat or slightly concave depending on the tool used. In the following analysis, we use the two criteria of wood selection and wood preparation to identify similarities and differences in Structures 1 and 5, with the aim of evaluating the three scenarios of labor organization and construction practice that we outlined at the end of the previous section (Newsom 2016).

Wood Selection

Seven wood taxa were identified in Structure 1: red oak, American chestnut, black locust, true hickory, pecan hickory, pine, and white oak; the same set of taxa was identified in Structure 5, but with the addition of a single specimen of another hardwood, ash. Overall, red oak proved to be the most commonly used wood for the construction of both buildings, comprising 36% of wood samples in Structures 1 and 47% of those in Structure 5. Pine was the second-most abundant wood taxon, comprising 24% of the Structure 1 assemblage and 15% of the Structure 5 set. White oak was the third-most frequently used wood in both buildings (22% in Structures 1, 13% in Structure 5), followed by chestnut in Structure 1 and true hickory for Structure 5, both at 10% of their respective sample sets. Chestnut also comprised 7% of Structure 5’s wood samples. Black locust and pecan hickory were present in small amounts in both buildings, each comprising 3% of identified wood, which is also the proportion of true hickory in Structure 1. Ash was identified only in Structure 5, from a sample field-designated as “cane and wood mass.” This sample may have entered the building as fuelwood or else represents something other than a construction element.

Excepting the hickories and the ash, all of the abovementioned taxa were used as upright posts in both structures (Tables 1, 2). The 18 analyzed uprights from Structure 1 indicate a preference for pine, which constitutes almost 40% of the post assemblage. This includes three center posts, which were the most substantial such elements (in terms of diameter) used for this building (Table 1). Pine is followed by chestnut (28%) and red oak (17%) within the upright post assemblage. Essentially the same pattern was true of Structure 5 posts, with pine making up the majority (42%) of uprights (n=12), followed by chestnut (17%) (Table 2). One pine upright post from Structure 5 (OG65) was of a similar diameter class as the center posts recovered from Structure 1.

Table 1 Upright posts from Structure 1
Full size table
Table 2 Upright posts from Structure 5
Full size table

To the extent that the data are representative, this shared pattern of wood selection for the major support elements (emphasis on pine and chestnut with but minor use of the other taxa) was also indicated in the age classes of the material selected. Age estimates for six of the Structure 1 posts spanned 15 to 100 years; four of the pine posts were the oldest, individually aged to 56, 73, 86, and 100 years (OG21, OG93, OG32, and OG47, respectively [Table 1]). Only one post from Structure 5 could be reliably age estimated, the previously noted OG65, a 47-year-old pine post. Post diameters were relatively consistent between the two structures, ranging from about 4 to 17 cm, with an average of 10 to 11 cm (Tables 1, 2). To reiterate, the largest upright posts, including the center posts for Structure 1, were pine, which thus served as the primary structural support for both buildings. These largest uprights include the four older ones noted for Structure 1, with diameters that ranged from 13 to 17 cm. The diameter of OG65 (17 cm) was the largest among those recorded for Structure 5, with the girths of the other pine, black locust, and chestnut posts ranging from 8 to 11 cm (Table 2).

In contrast, non-post wood from these two structures—mostly roundwood that presumably represents secondary elements, such as roofing and minor wall elements—was dominated by oak, especially red oak. More than half (52%) the non-post roundwood specimens recovered from Structure 5 (n=110) were of the red-oak anatomical group, while another 13% of that material was from the white-oak group. Pine comprised an additional 13% of the Structure 5 non-post wood remains. A similar pattern was observed for Structure 1, with a non-post wood assemblage that included 42% red oak, 27% white oak, and 20% pine; chestnut and the two types of hickory were also present in non-post assemblages from both buildings and ranged from about 4% to 6% of the respective assemblages.

The builders of Structures 1 and 5, whether the same or different groups, selected similar woods for each building, specifically preferring pine for wall and center posts, and oak for non-post elements. Neither preference is surprising. Pine typically forms long, straight boles that lack horizontal branches along the lower portions, making it ideal for poles and posts. Likewise, oak is one of the strongest readily available hardwoods and—like pine—was often used in the construction of native architecture during the Late Mississippian. What is more surprising is the consistent use of chestnut in both buildings. Across the Eastern Woodlands, Hovey Lake Village (12PO10) in southern Indiana (A.D. 1400–1700) is the only archaeological site with documented evidence of chestnut wood in aboriginal construction (Munson et al. 2009). It is unclear why there is so little evidence of chestnut use in native architecture across the region, particularly given its subsequent widespread use in European American architecture. Yet, the harvesting of mature chestnut trees, some of which grew to enormous proportions, would have imposed technological constraints on native stone-tool technology. The relatively high proportions of this wood in Structures 1 and 5 may thus indicate the availability of European metal tools for wood harvesting.

Any absolute correspondences of detail between these two buildings begin to break down when additional aspects of construction are considered. It is particularly with respect to indications of carpentry practices that the details of the two structures begin to diverge. As might be expected if there were distinctions in the amount of time and care that went into the construction of the two buildings, the non-post roundwood demonstrates some qualitative differences in terms of material proportions, age estimates, and specific details of the way the material was harvested and put to use. Individual samples from Structure 1 ranged from 2.4 to 14 cm in diameter (up to about the size of the upright posts), while those from Structure 5 overlapped in size, but exhibited a smaller range spanning 1.3 to 8.8 cm in diameter. Likewise, there is a general impression of smaller and younger material in association with Structure 5, whether when examining the non-post roundwood separately or including upright posts. The average diameter for non-post roundwood from Structure 5 is 3.3 cm (n=60; s.d.=1.671). In contrast, the average diameter for Structure 1 non-post roundwood is 6.6 cm (n=25; s.d.=3.12)—exactly double that of the other building. This distinction in overall size of construction material also holds true when the upright posts are factored in. Thus, the entire assemblage of roundwood for Structure 5 (including posts) has an average diameter of 4.12 cm (n=66). The average diameter for Structure 1 roundwood is 8.17 cm (n=43), which again is twice that of Structure 5 (Fig. 10).

Fig. 10
figure 10

Histogram comparing roundwood diameters from Structure 1 (n=43) and Structure 5 (n=66). (Map by Berry Site Archaeological Project, 2015.)

Full size image

This discrepancy in roundwood diameter classes is also reflected in the age estimates for the two assemblages (Fig. 11). To reiterate, the posts from Structure 1 were determined to span 15 to 100 years of age. Ages for the non-post class from this building exhibit a wider range, having growth-ring counts of as low as 8 to as many as 132 years (mean=24.3 years, s.d.=26.8) (Table 3). The 132-year-old specimen is a large, plank-like timber from Unit 50 designated OG44; when it is excluded, the average age for non-post roundwood is 19.4 years, with a range of 8 to 70 years (n=22). In contrast, no samples from Structure 5 exceed 50 years of age (and recall that the entire structure has been excavated). One previously noted post, OG65, was counted at 47 years, while the rest of the assemblage has a range of 5 to 28 years, with a mean of 12.6 years (s.d.=6.32) (Table 4). The contrasts in roundwood age and diameter between Structures 1 and 5 are striking and leave little doubt that the latter was a much less durable building, as might be expected for a structure that probably served as a cocina. Elsewhere in La Florida, for example, kitchens are noted to have been of lighter construction than houses and other kinds of permanent dwellings (Saunders 1991:132; D. Thomas 1991:117).

Fig. 11
figure 11

Age classes for all roundwood, including posts, from Structure 1 (n=26) and Structure 5 (n=31). (Map by Berry Site Archaeological Project, 2015.)

Full size image
Table 3 Growth-ring observations and harvest periods for Structure 1 samples
Full size table
Table 4 Growth-ring observations and harvest periods for Structure 5 samples
Full size table

Wood Preparation

A second line of evidence that distinguishes Structures 1 and 5 is post-harvest dressing or finishing of posts and timbers (Table 5 summarizes roundwood-segment data for both structures). These data appear to suggest a fundamental difference between the two structures in the way raw material was processed for construction. The majority (47%) of non-post roundwood from Structure 1 was fully or essentially complete, about 28% were half segments, while the rest was reduced to finer portions. The posts and other roundwood elements from Structure 5 indicate that its builders used younger and smaller wood overall, with diameters being consistently about half of those from Structure 1. Given that so much of the raw material from Structure 5 was of a smaller girth than that used for Structure 1, we might have expected these materials to have been used whole. Instead, only about 30% of the wood from Structure 5 was used intact, with 45% reduced to half segments and the remaining 25% reduced to one-third or even finer splits. This practice would have resulted in a more efficient use of construction materials and was perhaps a time-saving exercise as well, since it would have required the harvesting of less raw material. Yet the operation would also have produced a weaker structure, and combined with the data on roundwood age and diameter it supports the conclusion that this building was assembled in a far more expedient manner than Structure 1.

Table 5 Proportion classes for roundwood
Full size table

Another reason that wood-reduction techniques deserve special attention is the possibility that differential use of metal tools—rather than traditional implements of stone—distinguishes the construction process for Structures 1 and 5 (Table 6). OG21 from Structure 5 contained several specimens that were collected as individual bags or as small groups of elements bagged together due to their proximity to one another. Four of these samples have roundwood or other items that exhibit clear tool marks or otherwise suggest the use of very sharp implements; several others are half or quarter segments for which the separation is crisp or decisive enough to suggest that the specimen was cut with a sharp-edged metal blade rather than simply having been split along the grain. Sample 21c includes a MNI total of 19, among which are 8 larger fragments that also show sharp, sloping bevel cuts (transversely angled across the grain) that would have separated the element from the rest of its branch or trunk member. These are clearly not the result of natural breakage or of the typical chopping motion produced with stone adzes. OG 21h includes a roundwood section with a long, even cut down its midline resembling machete cuts.

Table 6 Tool-mark measurements (in mm)
Full size table

Several additional specimens from Structure 5 merit discussion. Two samples have circular holes that suggest iron nails or other fastenings. OG48b, a particularly good example, is a relatively flat, somewhat prismatic, and slat-like specimen of pine with a small hole (only partially complete) that measures 4.55 mm in diameter and penetrates the wood horizontally across the grain (Fig. 12). The rim around the edge of the hole is discolored relative to the surrounding wood, which is very unusual in charcoal and suggests that something with an overlapping lip—like the head of a small nail—once occupied the hole and was present at the time the wood burned. In addition to this specimen, 12 other slat-like items of pine were recovered from Structure 5, including 1 from Feature 91 (OG54). The specimen from Feature 91 is only partly carbonized—lightly charred on one end—while the others are fully carbonized. All of these slat-like specimens are suggestive of wooden shakes or shingles. The use of shakes—cut and split into fairly standard sizes to cover the roof or sides of a building—may represent an exclusively European practice that differs markedly from the thatch, various-sized bark sheets, or woven cane mats used as wall and roof coverings in Native American architecture.

Fig. 12
figure 12

OG48b (Structure 5), a slat-like wood specimen with a small hole penetrating the wood horizontally across the grain; the lighter rim around the edge suggests the overlapping head of a nail. (Photo by Berry Site Archaeological Project, 2015.)

Full size image

Moving on to consider Structure 1, several samples bear tool marks and other evidence of carpentry activities. A half segment of white-oak roundwood, OG3 (Fig. 13), was worked to a point at one end, which is evident from the presence of four clear facets; these tooling marks are apparent on the convex surface, range in length from 10 to 25 cm, and are about 8.5 to 16 mm wide. One whole, red-oak roundwood specimen, OG20a, has a sharply bevel-cut end, and a half segment of red-oak roundwood that was recovered lying immediately adjacent to it, OG20b, has a long, sharp cut down the centerline where a lateral knot or branch was deliberately removed by a sharp, transverse cut that suggests the use of a metal blade. Two red-oak “plank-like” timbers, OG43 and OG44, are about 12 to 13 cm wide and were partially trimmed by removal of some of the outer-growth increments, resulting in somewhat angular, squared forms that are narrowed toward one end as a result of sharpening or shaping (Fig. 14). The OG44 sample produced a ring count of 132 years. Very similar to these two finished, roughly quadrilateral timbers are the upright pine posts OG32 and OG47. OG32 is roughly square in outline, having been at least crudely shaped (Fig. 15). Structure 1 also had a chestnut post with a square hole penetrating transversely and intersecting five growth increments, suggesting that someone hammered a large metal spike into the top of the timber (OG10) (Fig. 16).

Fig. 13
figure 13

OG3 (Structure 1), with facets at left end. (Photo by Berry Site Archaeological Project, 2015.)

Full size image
Fig. 14
figure 14

OG44 (Structure 1), a worked timber or plank. (Photo by Berry Site Archaeological Project, 2015.)

Full size image
Fig. 15
figure 15

OG32b (Structure 1), a squared post shown in transverse (cross-section) perspective. (Photo by Berry Site Archaeological Project, 2015.)

Full size image
Fig. 16
figure 16

OG10 (Structure 1), an upright chestnut post with a large (1 cm diameter), angular hole (indicated by arrow) penetrating the grain and suggesting the use of an iron spike or bolt. (Photo by Berry Site Archaeological Project, 2015.)

Full size image

A very large chestnut plank, OG2003a, was recovered from Structure 1 in 2003 (Fig. 17); this lumber was minimally 76 cm long, 24 cm wide, and about 3 to 4 cm thick. The relative position of the growth rings makes it clear that this specimen is a tangential plank- or plane-sawn timber. This plank preserves at least 28 years of growth from a tree that must have been several times that old at the time of harvest, given both the positioning of growth increments and the minimal growth-ring curvature. As noted, the method—or, more precisely, the orientation—that was used to create this timber was plane sawing or plank separation along the tangential dimension (Hoadley 2000). This contrasts with quarter sawing and other means of dividing timber. It is important to convey that this plank represents a tangentially oriented separation from the bole, perpendicular to the radius of the tree. These cuts are far more difficult to achieve than quarter cutting, which follows the natural planes of weakness in a tree, such as the ray system.

Fig. 17
figure 17

Excavations in Structures 1, 2003, showing detail of the chestnut plank, OG2003a: (a) overhead view of structure floor and (b) close-up view of the plank. (Photo by Berry Site Archaeological Project, 2015.)

Full size image

If oriented correctly (and, depending on the type of technology used, the size of the trunk and the details of growth increments), plane sawing or splitting may sometimes take advantage of weaker zones within individual growth increments, but, even so, this technique would still require a difficult maneuver, given such a length of plank as that indicated by the Berry site specimen. In addition, this plank is a mature, dense wood with relatively narrow growth increments and minor curvature, indicating that it was harvested from a very large tree and probably from growth areas nearer to the base of the bole (if it was harvested from higher up the stem, then this was indeed a tree of gigantic proportions). We are skeptical that such a large bole could have been effectively managed for cutting planks as large as the Berry site specimen without metal carpentry tools and, specifically, something like a crosscut saw. Saws, axes, wedges, and chisels were indeed among the woodworking tools that accompanied Pardo’s expeditions, and we suggest that this plank, in particular, is a good indication of the presence and use of such tools, as well as European woodworking knowledge, techniques, and practice.

Finally, three large red-oak specimens—an upright (OG53) and two timbers (OG78 and OG81)—are especially revealing (Fig. 18). These have sharp, diagonal cuts across their ends running fully and cleanly across their whole circumferences of 12.8 cm (OG53), 14.14 cm (OG78), and 34.55 cm (OG81) (computed areas of 13.20, 15.90, and 95.03 cm2, respectively). The last, largest timber likewise bears very straight-edged, narrow marks across the cut face that strongly suggest saw marks (Fig. 19). Also, there is no suggestion of the several adze cuts that would have been needed to fell this specimen using indigenous stone implements or even a metal axe. This timber and the aforementioned plank probably offer the strongest evidence for the use of European carpentry tools. In general, it seems that the presence of these essentially lopped-off or sawn specimens distinguish the fundamental carpentry techniques—or at least the procurement of raw materials—used in Structure 1 from those used in Structure 5.

Fig. 18
figure 18

Structure 1 timbers: (a) OG78, 4.5 cm diameter, and (b) OG81, 11 cm diameter, shown in profile, demonstrating very clean, complete separation across the entire circumference of woody stems. (Photo by Berry Site Archaeological Project, 2015.)

Full size image
Fig. 19
figure 19

OG81 (Structure 1) shown in cross-section view with parallel saw marks spanning the full diameter. (Photo by Berry Site Archaeological Project, 2015.)

Full size image

It is unlikely that people from Joara would have used these European tools for carpentry. Although metal axes and adzes had clear analogues in native stone-tool kits, crosscut saws did not. More significantly, native peoples initially emphasized the prestige value of metal tools and other goods, as is evident both from the iron knives that Pardo gave to Joara Mico and from the interest that many native leaders had in receiving such items from the expedition. There is also clear evidence from across the native South that metal goods, such as European tools, circulated within elite exchange networks during the early colonial period, and that many of these materials were ultimately deposited in high-status burials (Smith 1987; Waselkov 1989; Hally 2008). These observations all underscore the point that evidence of metal-tool use in Structures 1 and 5 suggests labor by Spaniards, not native peoples from Joara.

To summarize, most basic design details of Structure 1 conform quite well to precolonial Native American construction practices, lending weight to the inference that native Joarans were involved in the building project. But the use of relatively mature, whole timbers as posts for the building, as well as indications of what are most likely European practices like plane sawing, are details that distinguish it from Native American architecture and instead may reflect the presence of Europeans and the use of metal tools. Such woodworking equipment would have enabled the builders of the structure—whether Europeans, native Joarans, or both working together—to secure larger boles from some relatively dense and strong woods for more substantial posts that were set in the ground as full rounds. Structure 5 differed in several important respects, particularly in the use of smaller and younger wood, and in the predominance of timber segments, rather than complete boles. Moreover, the use of wooden shakes or shingles and the digging of postholes with metal shovels suggest an altogether different conception of how to build such a structure using local materials.

Labor and Construction Practice in the Spanish Compound

All of these data on construction materials, techniques, and practices offer a solid basis for considering how construction labor was organized and how it changed over time. Structure 1 was built prior to Structure 5, near the beginning of the compound’s Spanish occupation. During this early phase, when social relations between the Europeans and their native hosts were at their best, construction labor seems to have been divided between them. Given that its form so closely follows native style and practice, it is reasonable to conclude that people from Joara planned and organized the building of Structure 1 and provided much of the labor required for its construction (including the excavation of its basin pit). Yet, Pardo’s men may have cut and processed much of the raw material that went into its framework. While it is possible that workers from Joara used metal implements with indigenous analogues, such as the axes and adzes carried by Pardo’s men, it seems more likely that the Europeans would have kept these essential and irreplaceable tools in their own possession. It is unlikely, moreover, that native builders could have quickly learned to use unfamiliar tools, such as the crosscut saws that processed some of the larger wood specimens recovered from Structure 1. The probable use of iron spikes to join timbers in this structure also indicates that Pardo’s soldiers may have played an active part in its construction, as well as in the acquisition and processing of its wood. We can therefore suggest that native and European work parties labored together in the building of Structure 1. Indeed, Structure 1 may be the new house that Bandera observed on Pardo’s return to Joara in September 1567.

Structure 5, however, exhibits a different pattern of labor organization. This building was constructed later than Structure 1, probably during what we have referred to as the second period of the Spanish compound’s occupation, and, as noted, there are several features that distinguish it from Structure 1. Some, particularly the use of wooden shakes and the digging of postholes with shovels, suggest an altogether different conception of how to construct such a building with local materials. Other features, such as the shallow central supports—requiring two additional supports to be added—suggest that the laborers who built Structure 5 were less familiar with the process of making a durable building with set-post construction. At the same time, we need not assume that Structure 5’s builders intended that it serve the same purpose as Structure 1; as noted, we believe that Structure 5 served the garrison as a formal kitchen, or cocina. In sum, Structure 5 appears to have been built by Pardo’s men with little input of labor or experience from those native builders who had committed their expertise to Structure 1.

Why might the people of Joara have stopped contributing labor to such construction projects if doing so had earlier conferred political benefits to their town and its chief, Joara Mico? Perhaps the most intuitive answer would be that the soldiers had lost their usefulness to the mico, that whatever status he and his town had gained by hosting the garrison and its soldiers was undermined by the expectation, or insistence, that his people would provision the guests indefinitely. This problem could only have been compounded had the Spaniards run out of materials like cloth, glass beads, iron chisels, and other items to exchange for provisions, or had the soldiers come to understand that they and their Indian hosts—who had shown their willingness to serve as subjects of the Crown—were no longer equal parties in an exchange relationship. Add to this dangerous mix the report of Teresa Martín, an Indian woman taken to Santa Elena during Pardo’s second expedition, who testified before the governor of La Florida in 1600 that Pardo’s men, after waiting “many moons” for his return, had committed improprieties with native women (Hudson 1990:176). We would not be surprised, all things considered, had Joara Mico chosen to withhold his townspeople’s labor as the men of Fort San Juan embarked on new building projects like Structure 5.

Yet, there is another intriguing possibility. In 1573, Spain’s Philip II established a formal set of ordinances for colonial town planning known as the Laws of the Indies. Although Pardo’s Fort San Juan de Joara was destroyed five years previous to Philip’s regulations, many historians interpret these laws as expansions and revisions of previous decrees and codifications of existing colonial practices (Mundigo and Crouch 1977:248); that is, even though Fort San Juan lay in ruins by 1573, those ordinances codified in the Laws of the Indies may have been common practice at the time of the Spanish compound’s reorganization. We argue that the compound, during the second phase of its occupation, like most other early Spanish settlements in the Americas and as later stipulated by Philip’s laws, was reorganized on a formal grid (Beck et al. 2016:143–145). Ordinance 110 commanded that “a plan for the site is to be made, dividing it into squares, streets, and building lots, using cord and ruler, beginning with the main square from which streets are to be run” (Mundigo and Crouch 1977:254). Importantly, according to another stipulation in the Laws of the Indies, Ordinance 137:

While the town is being completed, the settlers should try, inasmuch as this is possible, to avoid communication and traffic with the Indians, or going to their towns ... nor [should the settlers] allow the Indians to enter within the confines of the town until it is built and its defenses ready and houses built so that when the Indians see them they will be struck with admiration and will understand that the Spaniards are there to settle permanently and not temporarily. (Mundigo and Crouch 1977:258)

This ordinance suggests another reason for increasing tensions between the people of Joara and the soldiers of Fort San Juan, one related to the issue of labor as political practice raised at the beginning of this paper.

In colonial contexts it is easy to reduce the presence or absence of native labor to acts of acquiescence or acts of resistance: its presence is evidence of the former, while its absence is evidence of the latter. Either way, native peoples provided or withheld labor merely in reaction to the active political, economic, and military agendas of Europeans. Research at the Berry site offers an alternative perspective that upends this traditional narrative by situating the colonial context of Fort San Juan at the intersection of both Spanish and native labor regimes. Native leaders at Joara and other towns along Pardo’s path contributed labor to the expedition as part of a vital strategy to demonstrate their autonomy relative to one another and their capacity for organizing labor at local and regional scales. Leaders such as Joara Mico had much to gain by hosting Pardo, though they clearly expected the Spaniards to reciprocate their labor investment with trade goods and military support. There can be little doubt that Joara Mico mobilized native labor regimes at Fort San Juan to advance political, economic, and military agendas of his own, namely consolidating and even expanding the chiefdom he controlled.

The architecture of Structure 1 at the Berry site suggests that such a cooperative labor regime characterized the early phase of the Spanish compound’s occupation. Yet it seems that the soldiers of Fort San Juan may have systematically begun to exclude native labor from their formal reorganization of the compound, as was stipulated several years later by Ordinance 137. The architecture of Structure 5, for example, suggests that it was something other than a failed or inferior model of Structure 1, built by Spaniards trying to reproduce a native-style structure, but denied native labor and expertise. Given the use of shakes or shingles in its roofing, it is equally plausible that its Spanish architects intended nothing less than to build a Spanish-style structure from locally available materials. This is not to say, of course, that only one of these scenarios—Indians withholding their labor from Spaniards or Spaniards rejecting it—explains the unravelling of relations between town and garrison. Once denied the opportunity to mobilize labor as political practice, Joara Mico might, in turn, have withheld native labor at other times or in other contexts when the Spaniards may have needed or preferred it. Thus, these two scenarios are less mutually exclusive than mutually reinforcing, and together they may have fueled the growing dissociation between Joara and Fort San Juan.

By tracing changes in the organization of labor, in this case construction labor invested in two buildings at Fort San Juan de Joara, we can evaluate how social relations between settler and host communities unraveled in the course of an 18-month colonial encounter. We set out here to evaluate three scenarios of labor and construction practice in the Spanish compound. Fortunately, both structures contained abundant, well-preserved organic remains, allowing us to use the two domains of wood selection and wood preparation as proxies for construction labor. We conclude that the compound’s initial occupation phase—as represented by Structure 1—was a cooperative effort between the settler and host communities, planned and built at a time when relations between these communities were at their most harmonious. Structure 5, though, was built by Spanish soldiers with little evidence of indigenous support as part of the compound’s spatial reorganization, perhaps during the winter of 1567–1568. By this time, relations between settler and host communities had become strained, and the townspeople of Joara may have refused to commit labor to this project, been denied access to the compound, or both. The growing isolation of the garrison from its host community, apparent in these changes to labor and construction practice, would simultaneously seal its own fate and doom what would prove to be Spain’s final attempt to colonize the interior of La Florida.