Platform Mounds of the Arizona Desert

In the fall of 1989 we began an eight-year project to investigate platform mound communities in the Tonto Basin of central Arizona. The project is being funded by the Bureau of Reclamation as part of an overall plan to study sites that may be affected by proposed modifications to Theodore Roosevelt Dam (Fig. 2; see box). We are particularly interested in how the societies that built the platform mounds were organized, and how this organization might have been related to developments either within or beyond the local area. We began the project believing that the platform mounds of the Arizona desert were pretty much part of a single historic movement that swept the region, but as our work has progressed, we have had to rethink our models for the development and organization of the mound-building societies of Arizona.

Platform mounds were built by the prehistoric Salado and Hohokam peo­ple of southern Arizona from the 13th through the 15th century A.D., the Classic period. They are basically artifi­cial, flat-topped hills on which the rul­ing families of the day built their homes. Additional residences and storage rooms were built around the base of a mound, and the whole was enclosed within a compound wall (Fig. 3).

Each mound was the administrative, ceremonial, and economic center for a small-scale political system, or polity, whose settlements were scattered over 5 to 25 square miles (Fish 1989; Rice et al. 1990). At their maximum extent there were probably about 100 of these little political systems scattered across the Sonoran Desert of Arizona.

Platform mound polities were not evenly distributed across the desert floor nor did they completely “fill” the region. They occurred in groups of as few as four to as many as two dozen, and the groups were limited to the floors of the major valleys and low basins (Fig. 4) where agriculture, com­bined with the gathering of abundant wild Foods, allowed for the growth of large populations. Distances of up to 25 miles might separate one cluster from the next, and much of the Sonoran Desert of central and southern Arizona remained uninvolved in the platform mound polities. There is little doubt that some people continued to live in relatively independent and self-suffi­cient villages in the areas between the polities, but probably in far fewer num­bers than were associated with the plat­form mounds.

These platform mound polities scarcely compare to the Mesoamerican empires controlled by the Toltecs, Aztecs, or Maya, yet it is their lilliputian scale that we find most interesting. Although the Salado had not formed states, their societies were organized on a scale larger and more complex than the indi­vidual village unit. The elites living on top of the platform mounds, who organized the large labor forces needed to construct and maintain the mounds, were much more than tribal elders with temporary authority over a cere­mony or a task group. These were full-time leaders who had the authority and power to make demands on the local population for services and for products. In exchange, they provided the or­ganizational structure needed to protect the community, maintain public facili­ties, and ensure that people were fed.

The large sites in the Chaco Canyon area might indicate another example of a complex society in the prehistoric Southwest. However, some of the researchers working in the area feel that the environment of the Colorado plateau simply could not have generat­ed the levels of surplus food needed to support a complex society (Vivian 1990; Johnson 1989).

On the other hand, the production of a food surplus would not have been a problem for the platform mound peo­ple living in the well-watered valleys of the Sonoran Desert (Rice et al. 1990:31-34), The desert climate pro­vides a long growing season, with opportunities for two or more planting cycles per year. To agricultural produce could be added the rich harvests of nat­ural foods such as cacti fruits, agave (see article by Minnis and Whalen, this issue), and tree legumes. The abun­dance of the desert is attested by Catholic priests writing in the 18th century who found that the Pima (desert descendants of the Hohokam) frequently left surplus harvests rotting in the fields, having planted far more than they could possibly use. In bad years the extra acreage planted in fields would have come in handy. During the era of the platform mounds, such sur­pluses could have been harvested and accumulated in central locations for use by the leaders living on the mounds. The peoples of the platform mounds had taken the first step towards the kind of organizational complexity that functions today throughout much of our industrialized world. Over and over again human societies have taken this road to complexity, and within the last century, human societies have almost completely abandoned the tribal forms of life that had served us so well since the end of the last ice age. As archaeol­ogists we would like to know why and how human societies make these changes, but many of the first (or pris­tine) transitions to complexity took place thousands of years ago, and evi­dence of them has been obscured, if not eradicated, by the subsequent development of even more complex societies. This is the case for many of the original “cradles” of complex soci­eties, like Mesoamerica, the Middle East, and Asia.

But in southern Arizona, prehistoric peoples of the platform mounds took a small step towards complexity, experi­mented with it for the relatively brief span of two centuries, then nearly com­pletely disappeared. Small pockets of population remained, hut they lacked the size to maintain, much less require, a complex organization. The evidence of this early experiment with complexi­ty has not been clouded by a long archaeological record of subsequent states and empires (Rice 1990:157-58). With the arrival of European American settlers in the last part of the 19th cen­tury, however, complex society once again began to spread across the
Arizona desert, and true to historical precedence, our own growth is rapidly destroying the evidence of that first experiment in complexity. Even working at a distance of 100 miles from the nearest city, we find our investigations into the Tonto Basin platform mounds constantly beset by problems caused by vandalism and modern construction.

Dams, Problem-Oriented Research, and Salado Platform Mounds

In the first decade of this century Theodore Roosevelt laid out a vision for the development of the abundant nat­ural resources of the western United States, a vision that had as much to do with the creation of the Yellowstone National Park as it did with the construction of the Theodore Roosevelt Dam, the nation’s first major hydro­electric and water conservation dam at the confluence of the Salt River and Tonto Creek in Arizona. By storing water, generating electricity, and controlling the unpre­dictable floods of the Salt River, the dam made possible the phenomenal growth of modern Phoenix.

The building of the dam took nearly a decade, and brought together Italian stone masons, Irish construction workers, and Apache laborers in what was at that time the largest project ever to have been undertaken by the United States government. The project gave birth to a new agency, the United States Reclamation Service, known today as the Bureau of Reclamation.

In the last decade of this century the Bureau of Recla­mation has returned to the Roosevelt Dam. The height of the dam will be increased and larger spillways will be con­structed. Ironic as it may seem to people not accustomed to life in the Sonoran Desert, the dam is being modified not to store more water but to provide an improved capability for controlling floods that might threaten the Phoenix met­ropolitan area. These planned modifications will result in a larger reservoir, and could potentially involve the inunda­tion of more than 600 archaeological sites in the Salt and Tonto River valleys.

To deal with the impact to these sites, the archaeological staff of the Bureau of Reclama­tion sought to develop a program that fulfilled the spirit of the laws^on historic preservation. Kathy Pedrick, one of the archaeologists with the Bureau of Reclamation, remembers that “the challenge was to develop a mitigation program that would embrace critical research issues for the area.” After nearly a year of study and preparation, a study team of government archaeologists (including Kathryn Pedrick and Thomas Lincoln from Reclamation, Scott Wood from the Tonto National Forest, and representatives from the Arizona State Historic Preservation Office) identified a set of distinct research efforts, each of which was to be undertaken by a different research group.

Those of us on the research team from Ari­zona State University have the task of docu­menting the nature of Salado social organization through the excavation of five platform mound communities (i.e., mounds and their associated sites in the surrounding hinterland) located in three areas within the Tonto Basin (Fig. 5). Desert Archaeology, Inc., a private firm directed by William Dolled, is examining two platform mound communities with a slightly different emphasis, to detail the changes in social organization that took place as the platform mound communities developed from the earlier period of pithouse villages. A research team from Statistical Research, headed by Jeff Alschul, has studied a series of small compounds, field houses, and ter­race systems in order to document the relationships be­tween “rural” sites and the populations living at the platform mounds. Richard Dahlstrom of SWCA, Inc., has conducted a sample survey in the upper reaches of the ba­jadas (alluvial fans) that surround the valley, and has pro­vided valuable information about the contemporaneous populations living in the upland areas. In the spring of 1993 Arizona State University will excavate some of the bajada sites in order to establish the relationship of the up­land populations to those living at the platform mounds. Finally, far to the south, Paul and Suzanne Fish of the Ari­zona State Museum are mapping and excavating a plat­form mound community in the northern Tucson Basin. Some of their research was supported by the Bureau of Reclamation (a large aqueduct was constructed by Recla­mation through a portion of the 15 square mile community complex). By the close of this decade the archaeologists working on these various teams will have developed the first com­prehensive views of the community systems that envelop platform mounds.

Life During the Early Classic Period (A.D. 1200-1325)

For the first two years of the Roosevelt Archaeological Project, our research focused on the excavation of several platform mound communities and a larger number of residential compounds in a portion of the Tonto Basin called Livingston (Fig. 6). Most of the sites in the Livingston area date to the Roosevelt phase (A.D. 1200-1325) and only a few settlements were occu­pied in the later Gila phase (A.D. 1325-1450). We found a number of surprises in the Livingston area that have forced us to continue to rethink our models of Classic period society (A.D. 1200-1450).

During the Roosevelt phase the peo­ple of the Livingston area lived in about 40 small compounds and 5 platform mounds dispersed along a 3-mile stretch of the Salt River valley. We had pored over maps showing similar kinds of dispersed community patterns in the nearby Phoenix basin, but the only extensive excavation of multiple sites (that is, both mounds and compounds) in such a grouping of settlements had taken place more than a century earlier (see Hauiy’s 1945 report on the Hemenway Expedition of 1887 and 1888). Excavation coverage from more recent projects usually focused on only a few compounds or mounds within much larger groupings.

Here in the Livingston area of Tonto Basin was an opportunity to gather fresh data using modern methods and current research orientations. Along with our colleagues from Desert Archaeology (see box), we are paying particular attention to architecture and associated features. Artifacts are being systematically collected from screened samples or from carefully recorded locations on the house floors and com­pound plazas. We are collecting flota­tion samples (soil from which burned seeds and plant pieces are extracted), pollen samples, charcoal fragments, and archaeomagnetie samples (clay samples taken from hearths for dating) needed to resolve archaeological ques­tions concerning subsistence, chronolo­gy, trade, and craft production.

The level of detail in this new information has been gratifying. Inside the walled compounds (Fig. I) we found the circular bases of granaries, open-air hearths, and small pits for mixing the adobe needed for the regular replaster­ing of the adobe walls. Pollen studies in­dicate that the granaries were used for the storage of corn and squash. Pine pollen, usually found at higher eleva­tions, was recovered from the plaza areas, suggesting that pine boughs were used in ceremonial processions through the compounds. Fragments of a local form of barley were found clinging to the walls of storage vessels on the house floors. Outside some of the compounds were large roasting pits that our flotation studies showed were used for the baking of agave. We found that many of the compounds were divided in half by a wall, suggesting that perhaps two closely related families, or two generations of the same family, resided within each site.

In a typical residential compound about a third of the rooms contained hearths and were used as residences; other rooms were used as workshops or for storage. One room might contain small flakes and pieces of shattered stone indicating it had been used as a within workshop. Other rooms contained polishing stones and chunks of pig­ments used for pottery manufacture, along with hardened pieces of raw clay that had never been worked.

Just as the rooms within a com­pound had different functions, so did the compounds within the local area. For example, team member David Jacobs found that compounds that had large pits for roasting agave tended not to have granaries for the storage of corn. Large knives made Prom polished slabs of schist and used for harvesting agave occurred with great frequency at some compounds, while projectile points and obsidian were found in unusual concentrations at others. Yet the metates and manes that were needed to prepare corn and other grains for cooking were found in all off the sites, as were fragments of agave fibers and the bones of deer and rabbit. The individual sites may have been eco­nomically specialized, but the fruit of each family’s labor was made available to other families within the polity.

This evidence for the economic diversity and specialization of house­holds was one of the things we had expected to find if our ideas about the complexity of Classic period society were correct. One of the benefits for individuals living in a complex society is that the elite rulers provide an adminis­trative framework and the facilities tc ensure regular trade and exchange This made it possible for the families who concentrated on planting ark harvesting agave, for instance, to rely on exchange with those families whc concentrated on growing corn of specialized in making the projectile points needed to bring down game.

An Unusual Building

While our research was beginning tc provide answers to some of our ques­tions, it was producing highly unexpect­ed results in other areas. Based on wheat was known about platform mounds it the Salt and Gila Basins, we had expected to find buried beneath man) of our platform mounds evidence of ar earlier and much simpler kind of struc­ture called a dance mound. Dance mounds were frequently circular although some tended towards the rec­tangular. They contained no rooms or top and were little more than a slight]) raised stage on which ceremonies could have been held. Emil Haury drew on comparisons with the 19th century Pima culture to suggest that these early mounds might have been used for dances celebrating military victories (1976). He excavated a particularly elaborate dance mound at the site of Snaketown that included steps, but was no more than a meter in height (Fig. 7). A circular arrangement of post holes probably represents an adobe wall, reinforced by internal upright posts, that formed the facing of a later mound. It eroded but left a covering of dirt to preserve the earlier dance mound.

As we excavated into the interior of our first platform mound in the Livingston area, however, we found not a dance mound but the remains of a very special kind of ceremonial building that we called a “little Casa Grande.” The structure began as a series of two very large rooms (about 38 sq m each) that were entered from a long, narrow courtyard (see figure accompanying “Introduction,” p. 3). The doorways of the rooms and the courtyard opened to the east.

Within each room were a pair of adobe columns measuring nearly a meter in diameter and rising to a height of a little over a meter. A socket at the top of each mound held a post used to support the main beam in the ceiling. Neither room had a hearth, and the site had practically no trash.

During the excavation the director, David Jacobs, realized that the columns and doorways in the two rooms of the Pillar Site were aligned in such a main  that for a period of about 10 days during the winter and summer solstices they block all but a thin sliver of light from reaching the back wall of the room (Fig. 9). Modern Native Americans such as the Hopi and Zuni celebrate the winter solstice in a 16-day ceremony, and the builders of the Pillar Site may have been similarly concerned with identifying a span of days around the general event rather than the precise day of the solstice.

This was, and still is, an impressive building. It was not, however, an elite residence, nor was it necessarily an indi­cator of a complex society. It was a cere­monial facility, but it was constructed by a large group of people for the benefit of the entire group. The rooms within the building were very likely used for ceremonial meetings. No other building like this has been found in the Tonto Basin, but a very similar set of rooms with columns was found beneath a small mound in the Tucson Basin.

The First Platforms Mounds

Platform mounds appeared through­out Tonto Basin during a very short time span around A.D. 1280; five different platform mounds were con­structed within the Livingston area alone. Many of the mounds were built by removing the roofs from the rooms of pre-existing buildings, and filling the rooms with cobbles and dirt to create the platform. Some of the buildings that were modified into platform mounds were large ceremonial struc­tures, such as the one at the Pillar Mound, but others were residential rooms within small compounds. Where some pre-existing building was not con­veniently available, a grid-like structure of walled cells was first constructed and then filled to create the platform mound. The use of rooms or cells served a useful purpose in the construc­tion of large mounds by helping to stabilize lateral movements of fill within the mound.

At the Pillar Mound, the large rooms of the building were filled with dirt and cobbles to a height of about 1.2 meters, and new rooms were constructed on top of the resulting platform. The construc­tion of the platform was halted at inter­vals during which layers of brushy plants were laid over the soil. The pur­pose of these layers is not known, but radiocarbon analyses of samples of these materials have helped date the construction of the mound. Similar layers have been found in other Roosevelt phase platform mounds. Adobe columns and (for the first time) hearths were built in the elevated rooms, but doorways were omitted in favor of entrance through openings in the roof. If the columns continued to be used as solstitial markers, it was with the aid of windows that have long since dis­appeared as the walls of the elevated rooms collapsed. In its final form, the entire building had been transformed into an elevated platform mound with four rooms and a small plaza on top.

Although five platform mounds existed within the Livingston area at the same time, the relationships between mounds may have been more cooperative than competitive. Coopera­tion was needed, for instance, to build and maintain the canal system that traversed the Livingston area. The mounds imply that society had begun to develop important leaders, but as yet there was no single commonly recog­nized leader for the whole polity. Instead, the Livingston polity was composed of five loosely linked groups, each probably representing closely related kinfolk, and each group con­structed a platform mound for its own leader. The Livingston area as a whole, however, was still led by a committee of at least five representatives, one from each of the family groups.

This stage in the development of the society was extremely short-lived. It is exciting to actually have captured it in the archaeological record, for in little more than a generation four of the five Livingston platform mounds were aban­doned and apparently rebuilt around the mound at the Schoolhouse Site.

A Late Platform Mound at the Schoolhouse Site

The Schoolhouse Mound, one of the largest in the Tonto Basin, was excavat­ed under the direction of Owen Lindauer in 1990 and 1991. The archi­tecture of the Schoolhouse Mound con­ tains a striking record of the changes that took place during the Gila phase. The site was first occupied in the Roosevelt phase, and our current think­ing is that it had a small mound at that time. Three additional mounds may have been added at the start of the Gila phase (Fig. 8a). Through time the mounds expanded and additional rooms were built at ground level. More than a century later, at the end of the. Gila phase (A.D. 1450), the mounds had been linked into a nearly continu­ous unit of rooms and raised patio-like areas. The additional mounds and rooms at the Schoolhouse Mound were built by people who moved in from Schoolhouse Mound. This find provid­ed us with badly needed data on the size and construction of these struc­tures. The base of the granary was a cir­cular pedestal of adobe and cobbles, designed to prevent rodents from bur­rowing into it. The upper part was made of branches coarsely woven into a shape roughly like a bee hive. Both the outside and inside walls were then plas­tered with mud. Slabs of rock were laid across the opening at the top to seal the granary. It was about 70 centimeters high, with a diameter of 1.14 meters at the base, and had a capacity of 560 liters (about 15.5 bushels).

Assuming that the people living on top of the mound controlled the food amassed in the central storage rooms, Lindauer estimates that the granaries in those rooms alone contained as much as five years’ worth of surplus food. This surplus would be increased by approxi­mately another 80 percent if the stor­age volume of the large vessels were also included. The granaries in the rooms of the peripheral residential zone might have contained three years’ worth of surplus food. Both of these figures are considerably higher than the maximum estimate of one year’s sur­plus for the small domestic compounds of the preceding Roosevelt phase. The central storage rooms thus contained the surplus “capital” that would be needed by the leaders in order to feed people while they were involved in con­struction activities or services on behalf of the entire polity.

Arleyn Simon, the laboratory direc­tor for the project, has determined that the Schoolhouse Mound was one of a few pottery-producing communities in the valley. The residents of this site also had a disproportionate amount of turquoise, obsidian tools, pigments, projectile points, and even deer meat. A total of four shell trumpets, five stone “batons,” horn cores of mountain sheep, and unusual burial accompani­ments suggests the presence of particu­larly prestigious individuals living at the mound (Figs. 12, 13). Shell trumpets and elaborate headdresses are still used for important ceremonial occasions by present-day Native American groups in the Southwest.

All of this suggests to us that the platform mound leaders had intensified the extent of their control over the economy of the polity. The architecture of the Schoolhouse Mound reflects this concern with economic matters: ritual was undoubtedly still an important function of the platform mounds, but it was no longer as dominant. The emphasis on amassing considerable sur­pluses, the control exercised over the production of craft items such as ceramic vessels, and the concentration of prestige items (including prestige Foods such as deer meat) reflect a far greater level of centralization than we saw at platform mounds during the Roosevelt phase.

Some Closing Thoughts 

The archaeology of the Livingston area has provided new insights into the phenomena of platform mounds. It has Shown in surprising detail how a cen­tralized, powerful leadership emerged through a process of gradual amalgama­tion. The importance of family leaders in the early Roosevelt phase was replaced in the late Roosevelt by a smaller number of important lineage leaders, and in the Gila phase by the authority of a single clan chief. The lower-leveI leaders did not disappear From the society; every family still had a family head, and every lineage still had their respected elders. But a new level of hierarchy developed, with leaders who assumed greater power and authority.

The amalgamation process in Livingston also exhibited an interesting progression from ceremonial authority toward economic authority. Public buildings at the beginning of the Roo­sevelt phase were devoted completely to large communal meeting rooms. By the end of the Gila phase, the platform mounds functioned primarily as elite residences and as storehouses for the society’s wealth.

Unexpectedly, we found that the mounds in the Livingston group devel­oped from a previous architectural tradition of big houses, rather than from the dance mounds found in the Salt and Gila Basins. We will continue to expand our ideas about platform mounds in the next two years of research, and to think about the power­ful forces that were shaping the soci­eties of the Southwest during the 14th and 15th centuries.

Thus, the small dispersed com­pounds of the Roosevelt phase disap­peared. Only a few continued to be occupied during the growth and expan­sion of the Schoolhouse Mound. As the people of the Livingston area collected at Schoolhouse, they effectively aban­doned nearly all of the area that lay upstream (to the east; see Fig. 6).

As it grew, the Schoolhouse Mound became segregated into very distinct areas (see Fig. 8b). There were three concentric zones within the site, with accessibility decreasing toward the center. Much of the population resided in the rooms at the outside base of the mound. These rooms contained cook­ing as well as storage vessels, decorated bowls, tools, and frequently one or more granaries. Within this was a second zone consisting of the platform itself. The rooms on the platform were also residences, and while they con­tained many highly decorated vessels in addition to the various utilitarian arti­facts needed for households, very little space was devoted to storage.

At the absolute center of the plat­form mound, but at ground level, were rooms with massive quantities of stored food and supplies. The central storage rooms were entered through openings in the ceilings, and contained granaries, large storage jars (Fig. 10), ,and a variety of more portable artifacts including axes, agave knives, serving bowls, and decorated vessels. One of the great mysteries of the Schoolhouse Site is why these rooms were not emptied as the site was vacated. The completeness of the assemblages suggests that Schoolhouse may well have been one of the very last sites in the basin to be abandoned.