Agriculture and society in arid lands: a case study

Item Type Dissertation-Reproduction (electronic); text

Authors Fish, Suzanne K.

Publisher The University of .

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 08/10/2021 07:04:03

Link to Item http://hdl.handle.net/10150/191178 AND SOCIETY IN ARID LANDS:

A HOHOKAM CASE STUDY

by

Suzanne Kitchen Fish

A Dissertation Submitted to the Faculty of the

ARID LANDS RESOURCE SCIENCES COMMI1 lEE (GRADUATE)

In Partial Fulfillment of the Requirements for the Degree of

DOCTOR OF PHILOSOPHY

In the Graduate College

THE UNIVERSITY OF ARIZONA

1993 THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE

As members of the Final Examination Committee, we certify that we have read the document prepared by Suzanne Kitchen Fish entitled AGRICULTURE AND SOCIETY IN ARID LANDS: A HOHOKAM CASE STUDY

and recommend that it be accepted as fulfilling the requirements for

the Degree of DOCTOR OF PHILOSOPHY

3

Date . n 0379 3 Date °7)-( D 3 73c4-4 ? Date

1/1 Date

Date

Final approval and acceptance of this document is contingent upon the candidate's submission of the final copy of the document to the Graduate College.

I hereby certify that I have read this document prepared under my direction and recommend that it be accepted as fulfilling the requirement.

a k t 4 (-2:3 79-3 Director Date 3 .

STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgement of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgement the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.

SIGNED 4 .

ACKNOWLEDGEMENTS

The contents of this dissertation reflect scholarly endeavors undertaken with different but overlapping sets of colleagues, over the course of a decade. Acknowledgements must therefore be selective. Foremost are the co-authors of those collaborative works that were completed during my graduate enrollment and appear in the following pages. Not by chance, Paul Fish is the most frequent among these. He and John Madsen were greatly valued partners in the Tucson Basin studies that generated the majority of dissertation ideas and data. Gary Nabhan and Marci Donaldson co-authored individual papers. Acknowledgements for specific research efforts appear in several papers and in Chapter 1. Publishers of appended papers generously granted permission for inclusion here. The original contexts of publication are acknowledged at the beginning of each entry.

Four units within the university provided essential support. The Arid Lands Interdisciplinary Program was the optimal setting for an eclectic and multi-focused orientation. The Arizona State Museum supplied the institutional base for archaeological research. The Department of Anthropology and the Tumamoc Hill Desert Laboratory were further sources of collegial inspiration and material aid. My committee consisted of five distinguished members whose impressive and varied expertise, unflagging interest, and patience made a long graduate career and the ultimate completion of this dissertation a pleasure. Many thanks to Charles Hutchinson, Paul Martin, Norman Yoffee, Raymond Thompson, and Robert Altschul. Fellow graduate students with common interests, Charlie Miksicek, Gary Nabhan, Eric Mellink, and Mary Kay O'Rourke, particularly enriched the development of dissertation topics.

It is impossible to disentangle family from colleagues in acknowledging primary support over the years. I am fortunate to have collaborated with Paul Fish throughout my research career, and he is additionally the co-author of two helpers, Marcus and Allison Fish. My parents also offered encouragement from start to finish. TABLE OF CONTENTS

LIST OF FIGURES 13

LIST OF TABLES 17

ABSTRACT 19

CHAPTER 1 INTRODUCTION 20 Definition of Terms 21 Archaeological Approaches to Southwestern Agriculture and Society 23 The Tucson Basin Study 25 Structure of Tucson Basin Research 26 From Data to Dynamics 28 Dissertation Format with Published Papers 29

CHAPTER 2 PRESENT STUDY 34 Culturally Modified Aspects of the Hohokam Ecosystem 34 Zonal Landuse of Desert Basins 37 The Magnitude of Cultivation and Its Economic Significance 40 Economic Strategies and Trajectories 42 Multisite Communities as Integrative Structures for Diversified Settlement and Agriculture 45

REFERENCES CITED 48 APPENDIX A: THE MODIFIED ENVIRONMENT OF SALT- GILA AQUEDUCT SITES: A PALYNOLOGICAL PERSPECTIVE 58 Title Page 59 Copyright Permission Letter 61 Baseline Vegetation 62 Interpreting the Record of Hohokam Sites 64 An Alternative Perspective on Hohokam Pollen 66 A Palynological View of Hohokam Agriculture 67 Comparisons with Modern Pollen 67 Agricultural Weed Flora Analogues 68 Annuals and Perennials 70 Season of Cultivation 70 Variation in the Modified Vegetation 71 Sedimentary Regime of Irrigated Fields 72 Habitation Samples and Environment 73 Nonirrigated Agricultural Features 73 Conclusion 74 References Cited 75

APPENDIX B. AGRICULTURE AND SUBSISTENCE IMPLICATIONS OF THE SALT-GILA AQUEDUCT POLLEN ANALYSIS 80 Title Page 81 Copyright Permission Letter 83 The Pollen Record of Cultigens 85 Beans 85 Cucurbits 85 85 Corn 86 The Pollen Record of Possible Cultigens 89 Chenopods and 89 Native Grasses 90 Agave 91 92 Cholla 92 Significance of the Manipulated Complex 94 AgriculturalWeeds 95 Weeds as Resources 96 Microhabitats 97 Management of the Modified Environment 98 Gathered 99 Intersite Comparisons 101 Pollen Record of Irrigated and Nonirrigated Agriculture 101 Agricultural Implications of Habitation Samples 104 Comparison of Economic Types 105 Conclusions 110 References Cited 112

APPENDIX C. PREHISTORIC DISTURBANCE FLORAS OF THE SONORAN DESERT AND THEIR IMPLICATIONS 121 Title Page 122 Copyright Permission Letter 124 Abstract 125 Introduction 125 Agricultural Pollen Assemblages in the Sonoran Desert. 126 Methods 126 Modern Pollen and Vegetation 127 Spectra of Houses and Irrigated Fields 127 Field Samples 127 Habitation Samples 129 Spectra of Hillslope Terraces 130 Los Morteros 131 Tumamoc Hill 132 Discussion 132 Archaeological Interpretation 132 Floristic History 133 Environmental Reconstruction 133 References Cited 134

APPENDIX D. PRODUCTION AND CONSUMPTION IN THE ARCHAEOLOGICAL RECORD 137 Title Page 138 Copyright Permission Letter 140 Abstract 141 Defining the Concepts 142 Assumptions and Interpretations 142 The Problem of Redundancy 143 A Multiscale Example From the Tucson Basin 145 The Intrasite Level 147 The Site Level 151 The Regional Level 155 Conclusions 158 References 159

APPENDIX E. PARAMETERS OF AGRICULTURAL PRODUCTION IN THE NORTHERN TUCSON BASIN 162 Title Page 163 Copyright Permission Letter 165 Consultation with Traditional 166 Precipitation and Regional Potential 167 Agricultural Attributes of Basin Settings 167 Mountain Slopes 167 Canyon Bottoms 169 Upper Bajadas 169 Middle Bajadas 171 Lower Bajadas 172 Santa Cruz Floodplain 175 Agricultural Parameters and Regional Land Use 176 References Cited 178

APPENDIX F. ANALYZING REGIONAL AGRICULTURE: A HOHOKAM EXAMPLE 184 Title Page 185 Copyright Permission Letter 187 Introduction 188 Northern Tucson Basin Survey 190 Survey Strategy 190 The Setting 192 Settlement Patterns 193 Agricultural Patterns 201 Zonal Distributions 204 Advantages of Full Coverage 205 Recognition of Scale 206 Selection of Samples for Detailed Study 207 Quantified Estimates from Archaeological Distributions 207 Conclusions 213 References Cited 215 1 0

APPENDIX G. AN ARCHAEOLOGICAL ASSESSMENT OF ECOSYSTEM 220 Title Page 221 Copyright Permission Letter 223 Introduction 224 The Hohokam 225 Populations, Communities, and Replication 227 Tucson Basin Case Study 228 Environmental Setting 228 Definition of Communities 228 Community Dynamics and Evolution 230 Community as Ecosystem 240 Specialization and Exchange 241 Conclusions 244 References Cited 246

APPENDIX H. DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 252 Title Page 253 Copyright Permission Letter 255 Distance and Diversity in Hohokam Basins 258 Hydrological Regimes 259 Geographic Patterns of Climate 262 Vegetation and Habitat Complexity 266 Distribution of Wild Resources 268 Modified Landscapes and Their Resources 269 Water Control for Agriculture 274 Ethnographic Technology 274 Hohokam 275 11

Ethnographic Analogy and Hohokam Environmental Relationships 277 Hohokam Environment in Summary 278 References Cited 279

APPENDIX I. PREHISTORIC LANDSCAPES OF THE SONORAN DESERT HOHOKAM 288 Title Page 289 Copyright Permission Letter 291 Abstract 292 Introduction 292 The Agricultural Basis for Settlement Continuity 293 Conservation and Extractive Practice 295 Productivity of Modified Landscapes 296 The Role of Community Organization 297 A Hohokam Community of the Tucson Basin 298 Zone 5 300 Zone 1 301 Zone 2 301 Zone 3 302 Zone 4 302 Zone 6 302 Conclusion 303 References Cited 304

APPENDIX J. COMPARATIVE ASPECTS OF PARADIGMS FOR THE NEOLITHIC TRANSITION IN THE LEVANT AND THE AMERICAN SOUTHWEST 307 Title Page 308 Copyright Permission Letter 310 Introduction 311 Regional Contrasts 312 1 2

Nature and Rate of Change 313 Paradigms for Preceding Hunters and Gatherers 314 Problems of Analogy 315 Rethinking Variation 316 Paradigms for Transitional Entities 317 Synchronic Models of Variation 319 Diachronic Models of Variation 321 Paradigms and Variability in Settlement Patterns 322 Concluding Observations 324 References Cited 326 13

LIST OF FIGURES

FIGURE Cl, Pollen frequencies in Frogtown field samples at depths > 20 cm 1 28

FIGURE C2, Relative proportions of Ambrosia-type, Cheno-Am, and agricultural weed pollen plotted by depth in Frogtown field trenches 1 28

FIGURE C3, Pollen frequencies in Frogtown house floor samples 1 30

FIGURE C4, Pollen frequencies in samples from Tucson archaeological profiles 1 3 1

FIGURE Dl, Settlement and zonal configuration of the Marana community 1 44

FIGURE D2, Rockpile fields (Zone 2) and excavated sites yielding botanical remains in the Marana community 146

FIGURE El, Locations of weather stations in Tables 4.1 and 4.2 168

FIGURE E2, Braided channel of a drainage on the upper bajada downslope from the Tortolita Mountain pediment in the Marana survey area 1 70

FIGURE E3, Segment of Demo Wash with arable bottom land on the upper bajada in the Marana survey area 172 14

FIGURE E4, Water diversion structure on a tributary of the Rio Sonora near Baviacora, Sonora 1 7 3

FIGURE E5, Earthen embankment for intrafield distribution of water along a tributary of the Rio Sonora near Baviacora, Sonora 1 7 3

FIGURE E6, Canal carrying diverted floodwaters to multiple fields along a tributary of the Rio Sonora near Baviacora, Sonora 1 7 4

FIGURE Fl, Preclassic (ca. A.D. 750-1100) settlement pattern in the northern Tucson Basin 1 94

FIGURE F2, Early Classic (ca. A.D. 1100 to 1300) settlement pattern in the northern Tucson Basin 196

FIGURE F3, Zonal patterns in the northern Tucson Basin Early Classic community 1 98

FIGURE F4, Segment of a large agricultural field with terraces, rock piles, check dams, and roasting pits 202

FIGURE F5, Segment of a large agricultural field with terraces, rock piles, check dams, and roasting pits 203

FIGURE Gl, Important locations within the Hohokam region of southern Arizona 226 1 5 .

FIGURE G2, Survey areas and important site locations in the northern Tucson Basin 230A

FIGURE G3, Preclassic settlement pattern in the northern Tucson Basin 232

FIGURE G4, Classic Period settlement pattern in the northern Tucson Basin 233

FIGURE G5, Zonal divisions in the early Classic Period settlement pattern 235

FIGURE G6, Idealized cross-section of the northern Tucson Basin showing zonal relationships of the early Classic Period Community 236

FIGURE H1, Subdivisions of the Sonoran Desert in Arizona 257

FIGURE H2, Current and reconstructed flow in southern Arizona streams 260

FIGURE H3, Locations of weather stations furnishing precipitation data 263

FIGURE H4, Distribution of saguaro in southern Arizona 270

FIGURE H5, Distribution of mesquite and little-leaf paloverde in southern Arizona 271

FIGURE H6, Distribution of agave and yucca in southern Arizona 272 16

FIGURE Ii, Map showing the extent of the Hohokam with locations for the Marana Study Area and important place names used in the text 294

FIGURE 12, Map of the Northern Tucson Basin depicting major topographic features and configuration of the early Classic Period community 299

FIGURE 13, Idealized cross-section of the Northern Tucson Basin 300

FIGURE ii, (A) Distribution of Natufian and Natufian- like assemblages in the Levant; (B) Distribution of Piman peoples in the Southwest; (C) Locations of Piman subgroups named in the text 3 20 1 7

LIST OF TABLES

TABLE Bi, Agricultural pollen types in Frogtown samples 87

TABLE B2, Summary of pollen distributions in Salt-Gila Project agricultural loci 1 02

TABLE B3, Summary of pollen distributions in Salt-Gila Project habitation sites 1 0 3

TABLE B4, Occurrence of economic pollen types at nonagricultural sites 1 09

TABLE D1, Plant remains and associated artifacts recovered from structures and households at Rancho Derrio, AZ AA:12:3 (ASU) 1 48

TABLE D2, Plant remains and associated artifacts recovered from Locus A, Muchas Casas, AZ AA:12:2 (ASU) 149

TABLE D3, Cholla and saguaro recovered from the Marana sites 1 52

TABLE D4, Corn, agricultural weeds, and ground stone recovered from the Marana sites 1 54

TABLE D5, Agave, cotton, and associated artifacts and features recovered from the Marana sites 1 5 6 18

TABLE El, Annual precipitation measures for stations in the Tucson basin and other Hohokam subareas.... 167

TABLE E2, July, August, and September precipitation for stations in the Tucson Basin and other Hohokam subareas 167

TABLE E3, Mountain watershed sizes for trans-bajada drainages in the Marana survey area 171

TABLE Hl, Precipitation for selected stations in the area of Hohokam occupation 264 19

ABSTRACT

The interplay between social and productive spheres in arid land agrarian societies with non-mechanized technologies is exemplified in a case study of the prehistoric Hohokam Indians of southern Arizona. In addition to chapters unique to the dissertation, ten papers are included that were published during the period of doctoral enrollment. Results from a variety of investigative techniques are combined to characterize Hohokam agriculture and its relationship to societal forms and dynamics. Among these are archaeological survey and settlement pattern analysis, technical studies of prehistoric fields, palynological analysis for reconstruction of agricultural environments, and comparison with methods and concepts employed by historic and modern traditional farmers in the southwestern and northwestern Mexico.

Hohokam farming is examined at sequential scales, beginning with the setting, layout, and yield of individual fields and ending with comparison between Hohokam agricultural configurations and those of other arid land cultures. Topics receiving emphasis within the broader dissertation theme are the nature of Hohokam agrarian landscapes, the recently recognized role of cultivated agave in subsistence systems, and the social and economic framework for agricultural decision-making and strategies. 20

CHAPTER 1 INTRODUCTION

The study of archaeological remains is an indirect and complicated method for examining the dynamic between agrarian societies and low technology agriculture in arid lands. The nature of both society and agriculture in the past must first be reconstructed from a fragmentary material record that has survived into the present; interrelationships between the farmers who left these remains and their subsistence systems require a further level of inference. What, then, can the archaeological record contribute to a broader understanding of the interplay between arid land social and productive spheres?

Archaeological cases add to our knowledge of the variety of cultural and technological responses to the challenges of desert farming. In the absence of relevant historical or observational accounts, this methodology may be the only means for documenting even the relatively recent history of cultivation for a given time and place. Although such particularistic reconstructions are valid goals for archaeological studies of agriculture, practical applications include the demonstration of historically established land and water rights for Native Americans in the southwestern United States (e.g. Ferguson and Hart 1985: 34-39; Colton 1974; Ellis 1974; Hackenberg, 1974). Archaeological cases increase the scope of diachronic studies that seek to define the role of variables related to subsistence in and environmental transformations. Because current agrarian configurations reflect long developmental histories, researchers with wide ranging interests in modern dynamics and their real-world implications benefit from a knowledge of the past. Finally, archaeological instances expand the repertoire of social and technological solutions that might be reinstituted or refurbished to 21 meet the needs of population in arid lands today (for example, Erickson and Candler 1989; Evenari et al. 1982; Denevan et al. 1987; Gomez-Pompa et al. 1982).

The prehistoric inhabitants of the southwestern United States and other non-Andean arid regions of the represent a unique adaptation among past and present low technology agriculturalists of the world's dry regions. They lacked the domesticated animals that provide their and Andean counterparts with a means for efficiently harvesting the dispersed vegetative resources of arid landscapes, supplementing human labor, and extending the radius of subsistence operations. The addition of Old World irreversibly altered the technological and environmental parameters of indigenous farming systems. In spite of the fact that animal inputs may not be equally advantageous in every instance of low technology farming (Netting 1993: 102-122), similarly direct productive and extractive landuse by desert agriculturalists persists nowhere today. Additional contact vectors further realigned prehistoric economies (e.g. Crosby 1972; Dobyns 1981; Ezell 1961; Cook and Lovell 1992; Reff 1991). If pre- Columbian New World examples are to be included in the universe of arid land experience available to scholars, it must be through the discipline of archaeology.

Definition of Terms

The present case study of prehistoric agriculture and society in the desert basins of southern Arizona examines the Hohokam, members of an archaeologically defined culture that is distinguished by red-painted pottery and other stylistic attributes (Haury 1976; Fish 1989; Crown 1990; Gumerman 1991). The Hohokam appellation is usually reserved for pottery-making farmers after A.D. 100, but Late Archaic predecessors without ceramics had planted domesticates for 1000 years or more by that time (Huckell 1990; 2 2

Fish et al. 1990). The Hohokam continued as a recognizable cultural entity until the latter half of the fifteenth century. During this interval, they occupied over 40,000 square kilometers in the central portion of southern Arizona, producing by means of largescale irrigation, floodwater techniques, and the capture of overland runoff. The demise of the Hohokam and the conversion of survivors into the Indian groups of historic times are matters of ongoing archaeological debate.

As a relatively autonomous and productively self sufficient agrarian society (or set of related societies) practicing low technology agricultre in one of the hottest sectors of the southwestern United States, the Hohokam should embody maximum accomodations to their arid setting. An agrarian society is defined in this study as one in which almost all members engage in farming activities. Agriculture is therefore a primary concern of most participants and a major factor in social forms and interactions.

Low technology agriculture implies both a physical simplicity of non-mechanized cultivation techniques and a "non-scientific" cognition of the principles of implementation. Minimal power is obtained from sources other than human and animal labor or natural forces such as wind and water. Materials are primarily unmodified and local; mechanical devices are uncomplicated. Requisite knowledge and skills are widespread among farmers. Only in the construction of the largest canal networks in prehistoric North America or (Doolittle 1990: 80) is it likely that the Hohokam may have departed from this last criterion of a low technology designation. The level of specialized knowledge in the engineering of these networks is unknown.

The dynamic between Hohokam society and agriculture encompasses a myriad of potential interactions, few of which are amenable to absolute cause-and-effect resolution. On the one hand, 2 3 the exigencies of desert farming influence cultural expressions. Although correlation is not causation, the correlation of cultural patterns with environmental variables is basic to the archaeological recognition of significant interrelationships and their subsequent interperetation. Conversely, the organization of society and its economic framework influences the deployment of even the simplest agricultural techniques, in spite of the stringent constraints and restricted opportunities of an arid setting. Interactions of this sort are archaeologically addressed through the cultural correlates of synchronic and diachronic variation in agricultural patterns.

Archaeological Approaches to Southwestern Agriculture and Society

Three approaches are predominant, although not mutually exclusive, among studies of agriculture in . The remains of crops are retrieved from habitation sites, identifying agricultural products and an associated general range of requirements for their cultivation. Site environments can then be assesed in terms of these requirements. The remains of farming activities are a second focus of investigation, sometimes also yielding evidence for crops. These studies reveal the technology of prehistoric agriculture and technical attributes of the environment at locations in which it was practiced. Settlement patterns, defined primarily by distributions of residential sites and only occasionally including agricultural loci, are also analyzed in terms of environmental variables considered pertinent to farming. Changing potentials for agriculture are inferred by comparing environmental attributes of settlement patterns over time against climatic records.

A pervasive paradigm in archaeological investigations of Southwestern agricultural societies perceives these groups as fundamentally reactive to the constraints of aridity. Major changes 2 4 in cultural configurations proceed in response to changes in environmental states. Agriculture is the chief medium through which environmental constraints and perturbations are transmitted to dependent farmers. This paradigm is in keeping with the inherent materialist bias of the archaeological record - in this case a record that lacks textual accounts and extensive representational art to illuminate social or economic factors in agricultural pursuits. This emphasis can also be seen as an outgrowth of the exceptional opportunities in the Southwest to model interactions between precisely monitored climate and well-dated cultural patterns.

The analysis of tree rings has brought prehistoric culture and environment into remarkable conjunction in the northern Southwest above the Mogollon Rim, where conifer timbers were regularly utilized in Puebloan buildings. Preserved beams simultaneously encode calendric and climatic chronologies that can be linked to the builders. These circumstances have stimulated scenarios at many geographic scales of the effects of changing climate on aboriginal agrarian societies. The best examples of research reflecting the reactive paradigm are highly sophisticated, multidisciplinary analyses that variously incorporate settlement patterns, agricultural technology, climatic measurements, geomorphological, hydrological, and other environmental components (e.g. Euler et al. 1979; Wetterstrom 1986; Gumerman 1988; Orcutt et al. 1990; Vivian 1990). This conceptual orientation, however, directs attention toward influences and prime movers exterior to Southwestern societies (Plog 1992). Developments arising within the social sphere, admittedly more difficult to discern in archaeological data, are seldom accorded equivalent force, particularly in matters of subsistence with close and obvious ties to environment.

In a pioneering Hohokam botanical analysis for the site of Snaketown, Vorsila Bohrer (1970, 1971) interpreted differential proportions of wild resources and cultigens in samples as the result 25 of varying success in agriculture during sequential intervals. Although the same botanical sequence was not replicated in later analyses (Gasser 1980), this view of Hohokam farming as a fluctuating endeavor that covaried tightly with climate remains prominent. Local tree ring series have not been available for southern Arizona, but Hohokam archaeologists have advanced climate-based models following northern chronologies. The geographic expansion and contraction of Hohokam cultural boundaries (Weaver 1972) as well as more locationally specific conditions have been posited through interpolation of northern percipitation (e.g. Masse 1979: 179; Miksicek 1984). More recently, Donald Graybill (1989) reconstructed streamflow in the Salt and Gila rivers by means of tree ring records for the upland watersheds, thus providing an immediate rationale for Hohokam response in the form of water supplies and flood damage to canals.

Emil Haury (1950: 546-548) articulated an influential contrast between riverine and desert divisions of the Hohokam, based on the implications of differential regional opportunities for large-scale irrigation. The cultural heartland along the perennial Salt and Gila rivers in the Phoenix Basin derived its primacy from the underlying magnitude of agricultural production. The paucity of archaeological research in outlying regions heightened the perception of a core surrounded by less developed peripheries that were unable to similarily support manifestations of Hohokam high culture such as edifices for public events.

The Tucson Basin Study

The Hohokam of the Tucson Basin, the focus of the present case study, inhabited a peripheral basin drained by an intermittent river. Before 1980, the riverine-desert contrast was a guiding principle in interpretations of Tucson culture history. For example, Paul Grebinger (1976) and Bruce Masse (1979) proposed an influx of 26 population from the Phoenix Basin after A.D. 500 rather than a concurrent development of agriculture lifeways. These immigrants from the core area were thought to have settled on the irrigable floodplain of the Santa Cruz River and only later cultivated other basin zones as a response to population growth. David Doyel (1977) challenged Grebinger's reconstruction of integrative organization in the Tucson Basin, noting that the presence of prehistoric canals was unconfirmed and known instances of large villages with public architecture were few. He suggested that Tucson reliance on alternative farming techniques accounted for an absence of societal complexity comparable to that of the Phoenix core. Subsequent research has substantially expanded archaeological data in the Tucson Basin and contradicted these earlier views of the history of Tucson agriculture and its constraining effects.

Current method and theory in the social sciences offers a multiplicity of frameworks for understanding the dynamic between agrarian societies and low technology agricultural systems (e.g. Geertz 1963; Boserup 1965; Chayanov 1966; Chisholm 1970; Barlett 1980; Shaner et al.1982; Wilken 1987; Netting 1993). More than mechanistic relationships between archaeologically detectable variables such as population size, technology, and environment, these constructs emphasize decision-making among alternatives, long and short term strategies, economic motivations of individuals and groups, and socially defined objectives of production. If archaeological cases are to meaningfully expand the comparative base for agrarian studies, issues of this sort must be addressed in some manner.

Structure of Tucson Basin Research

Exploration of the social and economic context of Hohokam agriculture in the present study is the outcome of a long term collaborative effort in the northern Tucson Basin (Fish et al. 1992). 2 7

This research began in 1981 under my joint direction with Paul Fish and John Madsen. Initial objectives were largely descriptive: What were the patterns of Hohokam settlement during different periods, and how was their farming accomplished? The emerging archaeological record suggested successive avenues of analysis and interpretation. Sources of method and theory are eclectic in the appended publications, evolving over the course of more than ten years.

When research began, relatively little was known about Hohokam occupations in the northern Tucson Basin or the subsistence systems which supported them. Only a few investigations had been undertaken in this area beyond expanding urban limits where archaeological remains were still extensively preserved. We formulated a multiscalar research design intended to bridge the common gap in archaeological reconstructions between broad brush regional trends and fine grained examinations of single locales. To this end, project members surveyed a study area greater than 1800 square kilometers (700 square miles) and performed more intensive localized investigations at dozens of residential and agricultural sites. Methods and results for different segments of study are included in the publications in Appendix A.

Recovery of comprehensive settlement patterns covering large blocks of territory was a first step. Systematic survey on foot to record archaeological remains was guided by three objectives: (1) the recovery of spatial relationships among remains of all sizes, representing agriculture and ephemeral activities as well as habitation; (2) the ability to evaluate settlement pattern against a full range of environmental variation; and (3) the ability to define territorial units of interrelated sites that included relatively low densities and dispersed distributions (Fish et al. 1992: xi). Broad coverage in settlement patterns permitted the selection of representative locations for more intensive investigation of 28 agricultural remains. The alternation of areally expansive and tightly targeted studies over the course of ongoing research refined understanding in turn at each geographic scale.

From Data to Dynamics

The results of northern Tucson Basin research provide the basis for defining relationships among residence, agriculture, and environment in the following publications. Comprehensive and systematically acquired distributions allow the discrimination of economically relevant social units among settlements and fields. Agricultural components of Hohokam occupations are identified through combinations of technology, crops, and environmental attributes.

Social units and agricultural components are analyzed in a variety of economic frameworks. These include processes of diversification, intensification, specialization, subsistence exchange, and risk management. In some cases it is possible to quantify elements of agricultural production for comparison across space and over time. Changes in spatial expressions of societal structure are linked to concurrent changes in productive patterns. Finally, the cultural ecology of the Tucson Hohokam is described by means of the articulation of settlement, environment, and subsistence activities.

Participant elicitations and researcher observations are not possible in archaeological studies. Therefore, ethnographic analogies typically provide the starting point for considering social context and volition in the past, as well as for insights into technical aspects of agricultural features and -practices. The historic and recent agriculture of Native American and other traditional farmers in the southwestern United States and northwestern Mexico is central to interpretation in the present study, but some important divergences are noted here and elaborated in appended papers. Indigenous 2 9 peoples did not occupy large sectors of the Tucson environment during the period of ethnographic record, nor were certain categories of agricultural features ever reported in use. More importantly, magnitudes of regional population and the integrative organization associated with Hohokam central sites and public architecture are not paralleled in post-contact times. Particularly in these realms, interpretations are most directly derived from archaeological patterns and depart most markedly from the documented range of later behavior.

Visits to areas of persisting traditional practices were invaluable for envisioning aspects of prehistoric agriculture that did not tangibly survive in the archaeological record. Farming was observed in greatest detail on the Tohono O'odham reservation west of Tucson and near the towns of Cucurpe and Baviacora in Sonora, Mexico. Gary Nabhan was a guide in the first instance, Tom Sheridan in the second, and Bill Doolittle shared information about the third locality. Other trips were made along the Rio Sonora and other drainages to the west in Mexico, and to Hopi, Zuni, Apache, and Rio Grande Pueblo farming districts.

Traditional Sonoran and O'odham farmers were also consulted regarding agricultural remains and environmental potential. With the assistance of Gary Nabhan, a study area visit was arranged for the Sonoran , Adalberto Cruz, and the O'odham farmer, Delores Lewis. This event was instrumental in translating a static array of environmental variables into an economic landscape offering investment choices and strategy alternatives.

Dissertation Format with Published Papers

Each entry in Appendix A addresses some facet of the dynamic between Hohokam society and agriculture. The publications represent different stages of data acquisition, analysis, and 3 0 interpretation that range from one specialized kind of evidence at a single location to interregional comparisons. Findings from the northern Tucson Basin form the core of research content, but are supplemented by additional investigations with close cultural and geographic ties. All research was completed and published during my graduate enrollment in the Arid Lands Resource Sciences interdisciplinary program.

The Modified Enviroment of the Salt-Gila Aqueduct Project Sites (S. Fish); Agriculture and Subsistence Implications of the Salt-Gila Aqueduct Project Pollen Analyses (S. Fish); Prehistoric Disturbance Floras of the Sonoran Desert and Their Implications (S. Fish)

The first two papers are palynological studies concerning the culturally modified environment and the subsistence system of Hohokam populations in the vicinity of Florence, Arizona. Although adminstratively discrete, these studies coincided with the beginnings of research in the Tucson Basin. Abundant pollen data from residential sites and fields underscored the impact and effectiveness of Hohokam desert agriculture and stimulated the pursuit of Tucson evidence in the same vein. The third paper adds pollen results from agricultural features in the Tucson Basin and generalizes the case for anthropogenic vegetation communities in prehistoric southern Arizona.

Production and Consumption in the Archaeological Record: A Hohokam Example (S. Fish and M. Donaldson)

A paper about Hohokam production and consumption examines the assumptions and logic behind the interpretation of excavated botanical remains. The record of crops and wild resources from several Tucson sites is compared against artifactual patterns and evidence for farming in the site vicinities. The topic of agricultural specialization is evaluated in the light of evidence from different 31 archaeological contexts and spatial scales.

Parameters of Agricultural Production in the Northern Tucson Basin (S. Fish, P. Fish, and J. Madsen); Analyzing Regional Agriculture: A Hohokam Example (S. Fish, P. Fish, and J. Madsen)

The next two papers are discussions of agricultural production in the Marana Community, a set of related settlements surrounding a central site in the northern Tucson Basin. The first describes zonal patterns of agricultural locations and their environmental correlates. The understanding of operational aspects is enhanced by the experience of traditional farmer consultants under similar conditions. The second paper elaborates the widespread cultivation of agave in fields of rockpile features. The economic contribution of this is quantitatively modeled.

An Archaeological Assessment of Ecosystem in the Tucson Basin (S. Fish and P. Fish)

The concepts of cultural ecology are applied to the Marana Community as an example of an archaeological application. This theoretical approach is particularly valuable when the ecological universe of subject populations can be convincingly bounded by means of comprehensive settlement patterns at a regional scale, as in the Tucson case. The role of societal organization in cultural ecology is emphasized.

Desert as Context: The Hohokam Enviroment (S. Fish and G. Nabhan); Prehistoric Landscapes of the Sonoran Desert Hohokam (S. Fish and P. Fish)

The eighth and ninth papers are syntheses of the relationship between culture and environment throughout all of Hohokam territory. The desert of southern Arizona is assessed as a backdrop 3 2 for variations in subsistence systems among populations in different subareas. In the following publication, Sonoran Desert environment and farming are discussed as factors contributing to the characteristics that distinguish the Hohokam cultural tradition from other major prehistoric traditions in the Southwest.

Comparative Aspects of Paradigms for the Neolithic Transition in the Levant and the American Southwest (S. Fish and P. Fish)

The final paper compares influential theoretical positions in studies of the transition to agriculture in the southwestern United States and the similary arid Levant. Inappropriate usages of ethnographic analogs and stage concepts tend to obscure the degree of variation in agricultural configurations that is apparent in Southwestern ethnography and archaeology. Illustrations from southern Arizona and the Southwest suggest areas for caution in the interpretation of unidirectional schemes.

The first three publications in Appendix A are summaries of palynological analyses for which I am solely responsible. The remaining seven are co- authored with colleagues who collaborated in the reported research in all but one case; Marcia Donaldson, co- author in the discussion of production and consumption in the archaeological record, participated only in the writing of the paper. From the inception of the northern Tucson Basin research, I served as co-director with Paul Fish and John Madsen, equally formulating the research design, directing the field work, designing the data analysis, interpreting the results, and writing proposals to obtain funding.

I am senior author of each co-authored entry, indicating a foremost role in synthetic structure and presentation. As the Tucson Basin research director with most extensive botanical training and the broadest interdisciplinary background in technical studies of arid 3 3

land environment. I provided particular competence in those areas. I also took the lead in interpreting the social and economic implications of Hohokam agriculture. Published papers in Appendix A are presented in original format and style, but with consecutive pagination according to order in this dissertation. If references from individually authored papers were merged into a master list in the original publication, entries pertaining to the reprinted paper have been extracted and reassembled following the text. Otherwise, references follow the text in the originally published position and style. 34

CHAPTER 2 PRESENT STUDY

The Hohokam of the Tucson Basin are the subject of a core set of observations and conclusions in the papers of Appendix A. The research reported herein underscores the chronological depth and substantial nature of prehistoric occupation in this region without large scale irrigation. Most significantly, investigations in the northern basin reveal the alternate technologies that underwrote a local developmental sequence and expression of the dynamic between bearers of the Hohokam culture and their agricultural systems. Ancillary studies lead up to the Tucson Basin vantage, and broader perspectives depart from this point.

The most important findings in the following papers are summarized under five themes: (1) culturally modified aspects of the Hohokam ecosystem; (2) zonal landuse of desert basins; (3) the magnitude of agave cultivation and its economic significance; (4) economic strategies and trajectories; and (5) multisite communities as integrative structures for diversified setlement and agriculture.

Culturally Modified Aspects of the Hohokam Ecosystem

The landscape of Hohokam settlement is not accurately portrayed as scattered villages and fields set in natural communities of desert flora and fauna. Long term occupations of individual sites and preferred zones, often over hundreds of years, had immediate and cumulative environmental impacts that only recently have become a topic of inquiry. Agricultural activity was a prime agent in transforming the natural enviroment into the modified version that constituted the effective ecosystem of the Hohokam. 3 5

Traditional farmers of the greater Southwest shape surrounding vegetational communities in intentional and unintentional ways. For example, soil disturbance and supplemental water create a weedy field flora. Field edge vegetation may be managed to serve agricultural functions or to increase resource availability. These effects have been aptly illustrated for current Sonoran Desert farming (e.g. Bye 1979; Rea 1979, 1983; Nabhan and Sheridan 1977; Nabhan et al. 1982). Environmental studies near Florence, Arizona, as part of the Salt-Gila Aqueduct archaelogical project (Fish et al. 1984) were innovative efforts to investigate similar enviromental outcomes of Hohokam agriculture.

Salt-Gila and Tucson pollen analyses, reported in the first three publications of Appendix A, reveal suites of weedy taxa in Hohokam fields that do not resemble those in samples from the modern vegetation zones containing the sites. Contrasts in pollen content between several kinds of agricultural and residential proveniences suggest that levels of distinctive agricultural weed pollen types were influenced by amounts of water supplied to different kinds of fields. Archaeological pollen spectra are interpreted by comparison with weeds in a variety of currently disturbed situations because modern methods of cultivation do not duplicate the microhabitats of prehistoric farming. Weedy pollen patterns vary with enviromental settings and agricultural technology, but consistently suggest a predominantly summer planting season and the enhanced availability of potential plant resources in addition to domesticated crops.

Distributions of pollen types representing non-cultivated resources raise the question of intentional manipulation by the Hohokam. There are many kinds of manipulation to insure or increase supplies of such plants among historic Native Americans, but prehistoric tending or cultivation rather than gathering is 3 6 difficult to confirm. Agave is a case in point. Species identification is not yet possible through the botanical attributes of agave remains. In the absence of confirmation for species known to be cultivated in Mexico, agave was initially characterized as a manipulated plant. Only subsequent evidence for widespread prehistoric cultivation changed its designation to that of a cultigen. The distribution of elevated cholla pollen both inside and outside structures in Hohokam sites and its failure to correspond well with locations of natural growth make it a strong candidate for another drought-adapted plant that was sometimes transplanted and tended. The discussion of a culturally modified environment and its potential for Hohokam resource enhancement in Salt-Gila project reports (Fish et al.1984) was followed by increasing attention to this topic, as summarized in Fish and Nabhan (1991) and Bohrer (1991).

The comprehensive recording of all sorts of archaeological remains over broad areas was the key to glimpsing the true degree to which the Hohokam modified their surroundings (Fish et al.1992). The extent of agricultural features made of cobbles is impressive. These occur both in concentrations of thousands of features and in such diffuse distributions that conventional mapping is not possible. Ephemeral earthen and brush constituents of water management systems are not preserved, but the additional engineering of surface runoff is hinted by Piman ditches, dykes, and diversions that began to concentrate flows as much as a mile upslope from floodwater fields (Castetter and Bell 1942: 168). Hohokam irrigation and stream diversions also created new ribbons and patches of alluvial soils (Dart 1986).

The long term sedentism occasioned by Hohokam agrarian lifestyles left other detectable signs of repetitive and intensive landuse. Annual camping by large parties to harvest saguaro fruit is attested by sites in dense stands of cacti on basin slopes. These lack structures or other permanent facilities. Small extractive sites were 3 7 not spread across all slopes expanses in spite of pervasive opportunities for hunting and gathering. Such sites cluster in general proximity to the boundaries of residential settlements, in part because of convenient access but also in relation to the localized water sources that simultaneously constrain residence. The aquisition and consumption of scarce desert firewood and other wild resources has yet to be modeled in any detail. The network of use rights for households, kin groups, villages, and larger territorial units determines the operative ecosystem for members of human societies.

Zonal Landuse of Desert Basins

The concept of zonal landuse was applied in a number of Hohokam subareas by the early 1980's, when appreciable distributional data could be compiled (e.g. Fish et al. 1984, 1985; Wood and McAllister 1984; Crown 1987; Howard 1987). Settlement types and agricultural remains were found to be correlated with topographic positions in desert basins and other locational characteristics of environment. The Tucson Basin study is unique in its survey coverage of all remains in a broad swath from edge to edge in a basin, providing the most detailed and complete data set for zonal description. Although there are variations in prehistoric zonal landuse within and between basins, comprehensive patterns at this scale are of great value in evaluating more fragmentary distributions elsewhere.

Combinations of residence and subsistence activities define six landuse zones in the multisite Marana Community covering 146 square kilometers (56 square miles) in the northern Tucson Basin. Only one zone lacks evidence of farming. Zones 1 through 4 proceed up the bajada or valley slope from the floodplain. Generally small, dispersed habitation sites and floodwater farming on alluvial fans occur in Zone 1. The slopes of Zone 2 were not inhabited, but rockpile fields were cultivated with surface runoff. Zone 3 sites are 3 8 extractive in function. The largest seem associated with the gathering of saguaro fruit. Zone 4, skirting high mountains on the eastern basin edge, contains abundant habitation sites of various sizes, and supported farming in large drainges, small drainages, and on intervening slopes. Zone 5 on the Santa Cruz River floodplain is marked by irrigation and numerous, continuously occupied settlements. Zone 6, from the western floodplain to the low peaks of the immediately adjacent Tucson Mountains, exhibits attenuated and interdigitating modes of eastern bajada landuse in addition to unique hillside terraces with houses and gardens.

Water is a major determinant of settlement patterns. The need for sustained domestic water is more restrictive than the need for supplemental water for crops, which can mature with transient flows. A willingness to travel surprising distances for household water in dry seasons is apparent in ethnographic accounts of Southwestern groups (e.g. Castetter and Bell 1942: 42-43). Prehistoric wells and impoundments are also undoubtedly under- represented in the archaeological record. Nevertheless, dates for Zone 1 sites at a distance from permanent water show that these sites were occupied in tandem with the construction of long canals to supply floodwater farmers. Expansion of runoff fields in Zone 2 is also linked to this extended radius of domestic water.

One of the most significant findings about Hohokam landuse in the Tucson Basin is the prominence of widespread farming along the flanks of more massive mountain borders. The idea of the Hohokam tradition as a riverine orientation, coupled with relatively fewer upland investigations than near valley bottoms, engendered previous models that emphasized early and permenant settlements almost exclusively along floodplains of the river and the largest tributaries. Not only was farming early in the mountain flank zone, however, but diversified techniques supported large, permanent sites and an independent unit of territorial organization before A.D.1100. 3 9

Ethnographic analogies from the Papagueria west of Tucson during the historic period also proved misleading as guides to zonal landuse in the northern Tucson Basin. Seasonal movement from upland winter habitations to summer farming villages near the basin floor is not supported. Agricultural remains are abundant on upper basin slopes. Furthermore, it is doubtful that individual cultivators could have responded to control rapid runoff after rainfall in more than one location. The positions of more recent fields in Papagueria are not indicative of the full range of even late historic locations that were cultivated here and in Sonora by farmers who invested equal or greater effort in small scale herding (Nabhan 1983, 1986; Doolittle 1984,1988). Land-extensive techniques of prehistoric times were deleted even earlier by Hohokam successors. Hillside terraces were never recorded in use in southern Arizona, for example. Similarly, early twentieth century Pimans associated cobble features like those in rockpile fields with cultivation, but could offer no further explanation of their use (Russell 1908; 88-89).

Much more could be accomplished in the way of technical studies and quantified methods in the description of agricultural technologies and locations. Time, funding, the skills of project members, and the participation of colleagues from other disciplines were limitations in this regard. However, the "ethnoscience" of the traditional farmer consultants was truly impressive, allowing them to quickly and efficiently evaluate critical variables without sophisticated measurement devices or complicated mathematical formulas. Experienced eyes noted residual flood detritus, particle size, and other characteristics of soils along drainages in order to assess the force of flows and the practicality of locally diverting and controlling water. The presence, color, and vigor of woody species inparted information as to the suitability of floodplains for planting and the possibilities for shallow wells. William McGinnes (1988: 66- 67) cited A.W. Kuchler (1967) in concurring that plants and their 4 0 distributional patterns are superior to any other means for gauging the biological potential of an area. The Hohokam probably would have agreed about the efficacy of such holistic categories and methodologies. There are many paths to describing and understanding zonal landuse.

The Magnitude of Agave Cultivation and Its Economic Significance

Botanical evidence from documents widespread usage of agave in the Southwest by late Archaic and early ceramic times. The prehistoric cultivation of this versatile source of food and fiber throughout Mesoamerica is also well known. In spite of the fact that the majority of the aboriginal Southwestern complex of crops originated to the south, the possibility that agave was formerly cultivated north of Mexico was not considered until recently. By the first years of the 1980's, several lines of evidence had prompted suggestions that manipulation or cultivation in the Southwest was probable (summarized in Fish and Nabhan [1991 1). Unlike most crops grown from seeds, however, agave grown from offsets or clones shows few morphological divergences from ancestral wild populations. The inability to discriminate species among charred remains from residential sites and the likelihood that mostly local species were cultivated further added to the diffuculty of archaeologically demonstrating Hohokam cultivation rather than gathering.

Tucson Basin research assembled incontrovertible evidence for agave cultivation and furnished the means for recognizing the magnitude and significance of production. This evidence consists of the well-replicated recovery of charred agave from roasting pits in rockpile fields outside the range of natural populations, the associated presence of distinctive tools for harvesting and processing agave, and an experimentally tested rationale for the agricultural 41 function of rockpile features. The quantified attributes of these fields and their relationships to other components of settlement pattern provide the context for economic interpretation.

The extent and preservation of features within rockpile fields is surprising. Heaps of cobbles- the rockpiles-and more linear but equally informal constructions cover 485 hectares or nearly 5 square kilometers in a Marana Community territory of 146 square kilometers. This estimate of cultivated area is a bare minimum because it does not include all small clusters and very diffuse instances. Chronological evidence supports the contention that almost all of these features were in simultaneous use at some time during the early Classic period between approximately A. D. 1150 and 1350. Thus, size of harvest, dietary contribution, fiber production, and labor requirements can be calculated by reference to numbers of cobble features and their ratio to plants.

Experiments and studies to refine previous estimates are still underway. Vegetative reproduction by in experimental plantings indicates that annual estimates of 40.8 metric tons of edible product and 3.7 metric tons of fiber may be low by a factor of 2 or more. It is also apparent that different strategies might be followed to achieve maximum harvest in a particular year or to sustain lower levels of production over longer intervals. Gauged experiments still in progress show that soil moisture levels in the rockpiles remain higher than in control soils for days or weeks after rainfall.

Rockpile fields in the northern Tucson Basin can be spatially linked with residential sites. In this way it is possible to contrast production magnitudes and to evaluate these in terms of agricultural alternatives. Field size is a more precise measure of production than the proportion of agave among all charred remains in samples from habitation sites. The use of agave wastes for fuel and the need to 4 2 compare burned agave fibers against seeds and other plant parts are complicating factors. The spatial organization of agave cultivation can be seen to differ among community sectors. Inhabitants of the central site farmed in large contiguous fields, used huge pits for roasting agave in common, and traveled farther from residences to cultivate. The disposition of agave harvests is a question that awaits information from residential excavations. Artifacts denoting processing and manufacturing activities such as fiber extraction and spinning may or may not parallel the locational emphases of primary production.

Information about cultivated agave is rapidly accumulating throughout the area of the Hohokam tradition. Recognition of this agricultural practice and its ubiquity could not have been gleaned from the farming systems of later indigenous groups in the same region. It is unlikely that rockpile fields represent the sole means of Hohokam cultivation, but hundreds of locations are now recorded throughout southern Arizona. In many cases where evidence is sufficiently complete, botanical remains, distinctive artifacts, and roasting pits again indicate agave as the crop.

Economic Strategies and Trajectories

The Marana Community of the early Classic period provides an optimal opportunity to examine Hohokam economic strategies and trajectories. Its developmental history is well documented and its final form is a detailed, relatively synchronous configuration of settlement and landuse in all basin situations. At the beginning of the Classic period, two independent riverine and mountain flank communities coalesced into a single enlarged community with a new central site between former settlement concentrations. An influx of population also seems to have occurred. Early Classic habitation sites cover about 6,000,000 square meters, compared to about 2,000,000 square meters for those of prior date. A larger population had to be 4 3 supported in much the same territory, with only modest additions.

Demographic change in the Classic period implies an intensification of agriculture in the sense of greater production from community territory as a whole. Agricultural intensification as defined by Ester Boserup (1965) is based on tropical and temperate agriculture in which land is a limited good and intensification is accomplished by cropping the same land more frequently with shortened fallow cycles. In arid environments like the Tucson Basin, however, agricultural water rather than land is the limited good. Structural improvements to irrigated and floodwater systems might increase yields from existing fields as in Boserup's scheme, but the size of the water source puts absolute limits on this strategy. A related instance in the Marana Community is the construction of a canal that permitted farming on alluvial fans far from domestic water sources. This canal also may have supported additional irrigation on a small scale.

Even in the most deterministic environmental situations, there is a lattitude of economic choice in decisions to increase production. Perceptions of acceptable effort are illustrated in the evaluation of a potential floodwater locale by the Sonoran and Tohono O'odham farming consultants. The Sonoran did not approve of the area because of coarse soils. The O'odham farmer agreed, but also noted the possibility of overcoming this drawback. Diverting floods onto the intended field several years before planting would create a suitable overlay of fine soil. Decisions to invest for future gain are similarly evident in the cultivation of agave. Harvests on newly cropped land would have been delayed for a number of years. Thus, the changing balance between acceptable effort and desired return must have governed the pace of agricultural development in the Marana Community, perhaps resembling the process that William Doolittle (1984) terms incremental growth. 4 4

As a means to raise total production in a desert basin, bringing hydrologically marginal land into cultivation is an alternative to the improved delivery of supplemental water. A dramatic expansion of rockpile fields in the Classic period Marana Community exemplifies this option. Labor investment per yield was concurrently increased by greater travel time than to fields nearer homes. By analogy with the Zuni and other Southwestern groups, it is probable that tenure of secondary land was more flexible and available for entrepreneurial use. Jeffrey Parsons and Mary Parsons (1990: 4) compare agave to Andean camelid herds that enable a more effective use of the agricultural landscape and extend productivity into colder, drier, and less fertile elevations. In Mesoamerica, agave was a factor in the settlement of the cool, dry highlands by sedentary agriculturalists. In the hot desert basins of southern Arizona, agave facilitated larger populations through production on previously unfarmed and poorly watered slopes.

Diversification of productive activities is a common strategy for countering the risks of arid land farming. Annual and seasonal variability in precipitation is typically great. Rainfall from summer thunderstorms is localized and runs off quickly. Although fields tapping small catchments may yield well under favorable circumstances, the risk of insufficient water in any one season is high. Conversely, drainages with large watersheds are prone to destructive flooding. Diversification among several situations with differential risks may take place at the level of the individual farmer, the household, the kin group, the village, or some combination.

The distances separating upland, alluvial fan, and irrigated floodplain zones in the Marana Community would make it difficult for individuals to attend to more than one of these locations at critical times. Residence near fields is an advantage to farmers who must divert and distribute water in at most a few hours after rains. 4 5

The Tohono O'odam responded to summer storms in the evening by torchlight, for example. Rockpile fields on the other hand, might have been more readily combined with other locations because they presented no pressing need to control channelized flow.

A strategy of agricultural specialization and exchange is best exemplified by rockpile fields. Production can be attributed to a single crop and quantitatively compared at the village level. Rockpile cultivation is also most volitional in that it is not dependent on sources of supplemental water in drainages. Distributions of rockpiles show that investment in these techniques was highest where large populations had poor access to well watered land, and was lower near settlements with better opportunties for floodwater farming and irrigation. Specialization in other resources is more inferential, but appears, like specialization in agave, to be a matter of emphasis rather than exclusivity of production (Fish and Donaldson 1991). However, specialization and exchange of even redundant resources may serve to increase overall production, fine-tune crops to habitats, smooth out shortfalls, and maintain relationships for times of greatest need.

Multisite Communities as Integrative Structures for Diversified Settlement and Agriculture

A maximum degree of environmental diversification is evident in the size and configuration of the Marana Community. Stretching from edge to edge of the Tucson Basin, it encompassed every kind of basin setting, every kind of opportunity and risk for desert farmers. The range of environmental zones likewise encompassed the full variety of any specialized production. Diversification at the community scale constituted an important counter to agricultural risk to the extent that subsistence items circulated among members in times of scarcity (Fish and Fish 1990). 4 6

Communities of known size in the Tucson Basin are larger than the 40 square kilometer average in the irrigated Phoenix Basin, although it is unclear how population numbers compare (Fish and Fish 1991). The Marana Community covers 146 square kilometers and a second community near the Picacho Mountains covers 136 square kilometers. The similarity of the Tucson area entities may reflect the need to encompass sufficiently diversified opportunities and risks on the one hand, and to mantain effective communication and a spirit of commonality on the other.

The common bond in Marana and other communities of outlying regions was not the need to cooperate, coordinate, and adjudicate water allocation and maintenance along a massive canal network as in the Phoenix core. Overlapping members from one agricultural user-group to another in Tucson communities might promote a somewhat similar if less pervasive and compelling social cement. For example, the lengthy canal supplying the Marana central site from the river passed through several smaller villages. Inhabitants of more than one of these sites also may have cultivated plots in a common rockpile field.

Hierarchy among Phoenix settlements has been ascribed to a preeminent role in administering a common irrigation system. Again, hierarchy in Tucson communities cannot be ascribed to the same function. Nevertheless, in Tucson communities, a single central site is the focus for public events, contains the most elaborate residential structures, and exhibits greater proportions of high value goods. Furthermore, this expression of material and functional inequality does not arise from greater access to the best productive land. The location of the central site in the Marana Community is marked by little or no potential for irrigation and floodwater land of secondary quality. Domestic water was available only through canal construction. However, three other large sites in a second tier of settlement hierarchy do coincide with prime locales for riverine 4 7 irrigation and upland farming along a major tributary.

It is likely that central site functions in the Marana Community are associated with a preeminent role in subsistence exchange. The sole advantage of site location is centrality in community territory, midway between upland and riverine concentrations of settlement. As specialized producers of agave who had minor opportunities for other crops, central site inhabitants had the highest stake in regular and predictable exchange. The persistence of the community through agricultural failures at different geographic scales likewise may have rested on the strength of such interlinkages. 48

References Cited

Barlett, Peggy F. 1980 Agricultural Decision Making: Anthropological Contributions to Rural Development. Academic Press, New York.

Bohrer, V orsila 1970 Ethnobotanical apsects of Snaketown, a Hohokam village in southern Arizona. American Antiquity 35: 413-430.

1971 Paleoecology of Snaketown. The 36: 11-19.

1 9 9 1 Recently recognized cultivated and encouraged plants among the Hohokam. The Kiva 56: 237-254.

Boserup, Ester 1 9 6 5 The Conditions of Agricultural Growth: The Economics of Agrarian Change Under Population Pressure. Aldine, Chicago.

Bye, Robert 1979 Incipient domestication of mustards in northwest Mexico. The Kiva 44: 237-256.

Castetter, Edward F. and Willis H. Bell 1 942 Pima and Papago Indian Agriculture. University of Press, Albuquerque.

Chayanov, A. V. 1 986 The Theory of Peasant Economy. University of Wisconsin Press, Madison.

Chisholm, M. 1970 Rural Settlement and Land Use. Aldine, Chicago. 49

Colton, Harold S. 1974 Hopi ethnobotany and archaeological history. The Hopi: American Indian Ethnohistory. Garland Publishing Company, New York.

Cook, N. D. and W.G. Lovell, eds. 1 9 92 Secret Judgments of God: Old World Disease in Colonial Spanish America. University of Oklahoma Press, Norman.

Crosby, Alfred W., Jr. 1972 The Columbian Exchange: Biological and Cultural Consequences of 1492. Greenwood Press, Westport.

Crown, Patricia L. 1987 Classic Period Hohokam settlement and land use in the Casa Grande Ruin area, Arizona. Journal of Field Archaeology 14: 147-162.

1990 The Hohokam of the American Southwest. Journal of World Prehistory 4: 223-255.

Dart, Alan 1986 Sediment accumulation along Hohokam canals. Th e Kiva 51: 63-84.

Denevan, William M., Kent Mathewson, and Gregory Kapps (eds.) 1987 Pre-hispanic Agricultural Fields in the Andean Region. British Archaeological Reports International Series No. 359. Oxford.

Dobyns, Henry F. 1 9 8 1 From Fire to Flood: Historic Human Destruction of Sonoran Desert Riverine Oases. Ballena Press Anthropological Papers 20. Ballena Press, Socorro. S 0 .

Doolittle, William E. 1 9 84 Agricultural change as an incremental process. Annals of the Association of American Geographers 74: 124-137.

1 988 Prehispanic Occupance in the Valley of Sonora, Mexico. Anthropological Papers of the University of Arizona 48. The University of Arizona Press, Tucson.

1990 Canal Irrigation in Prehistoric Mexico: The Sequence of Technological Change. The University of Texas Press, Austin.

Doyel, David E. 1 977 Rillito and Rincon Period settlement systems in the middle Santa Cruz River valley: alternative models. The Kiva 43: 93-110.

Ellis, Florence H. 1974 Anthropology of Laguna Pueblo land claims. Pueblo Indians III: American Indian Ethnohistory, pp. 9- 120. Garland Publishing, New York.

Erickson, Clark L. and Kay L. Candler 1 989 Raised fields and in the Lake Titicaca basin. In Fragile Lands of , edited by John Browder, pp. 230-248. Westview Press, Boulder.

Euler, Robert C., George J. Gumerman, Thor N. V. Karlstrom, Jeffrey S. Dean, and Ric hard Hevly 1979 The Plateaus: cultural dynamics and paleoenvironment. Science 205: 1089-1101.

Evenari, Michael, Leslie Shanan, and Naphtali Tadmor 1982 The Negev: The Challenge of the Desert. Harvard University Press, Cambridge. 51

Ezell, Paul 1 9 6 1 The Hispanic Acculturation of the Gila River Pimas. American Anthropologist Memoir 90, Vol. 63 (5), Pt. 2. Menasha.

Ferguson, T. J. and E. Richard Hart 1 985 A Zuni Atlas. University of Oklahoma Press, Norman.

Fish, Paul R. 1989 The Hohokam: 1000 years of prehistory in the Sonoran Desert. In Dynamics of Southwestern Prehistory, edited by Linda S. Cordell and George J. Gumerman, pp. 19-63. Smithsonian Institution Press, Washington, D.C.

Fish, Paul R. and Suzanne K. Fish 1 99 1 Hohokam social organization. In Changing Views of the Hohokam, edited by George J. Gumerman, pp. 84-101. University of New Mexico Press, Albuquerque.

Fish, Paul R., Suzanne K. Fish, and John H. Madsen 1984 Northern Tucson Basin Survey: research summary and recommendations for the Marana complex. A Supplemental Class III Archaeological Survey of the Phase A. Reach 3 Corridor, Tucson Aqueduct, edited by Jon Czaplicki and Adrianne Rankin, pp. 83-89. Arizona State Museum Archaeological Series 165. University of Arizona, Tucson.

Fish, Suzanne K. and Marcia Donaldson 1 99 1 Production and consumption in the archaeological record: a Hohokam example. The Kiva 56: 255-275. 52

Fish, Suzanne K. and Paul R. Fish 1990 An archaeological assessment of ecosystem in the Tucson Basin of southern Arizona. In The Ecosystem Concept in Anthropology, edited by Emilio Moran, pp. 159-190. University of Michigan Press, Ann Arbor.

Fish, Suzanne K., Paul R. Fish, and John H. Madsen 1 9 8 5 A preliminary analysis of Hohokam settlement pattern and agriculture in the northern Tucson Basin. In Proceedings of the 1983 Hohokam Symposium, Part I, edited Alfred E. Dittert and Donald E. Dove, pp. 75-106. Arizona Archaeological Society Occasional Paper 2. Arizona Archaeological Society, Phoenix.

1990 Sedentism and settlement mobility in the Tucson Basin prior to A.D. 1000. In Perspectives on Southwestern Prehistory, edited by Paul E. Minnis and Charles L. Redman, pp. 76-91. Westview Press, Boulder.

1992 The Marana Community in the Hohokam World. Anthropological Papers of the University of Arizona No. 56. The University of Arizona Press, Tucson.

Fish, Suzanne K., Charles Miksicek, and Christine Szuter 1984 Introduction. In Hohokam Archaeology Along the Salt-Gila Aqueduct, Central Arizona Project, Vol. II: Environment and Subsistence, edited by Lynn S. Teague and Patricia Crown, pp. 1-6. Arizona State Museum Archaeological Series 150. University of Arizona, Tucson.

Fish, Suzanne K. and Gary P. Nabhan 1 99 1 Desert as context: the Hohokam environment. In Changing Views of the Hohokam, edited by George J. Gumerman, pp. 35-54. University of New Mexico Press, Albuquerque. 53

Gasser, Robert E. 1980 Exchange and the Hohokam archaeobotanical record. In Current Issues in Hohokam Prehistory, edited by David Doyel and Fred Plog, pp. 72-77. Arizona State University Anthropological Research Papers 23. Arizona State University, Tempe.

Geertz, Clifford 1963 Agricultural Involution. University of California Press, Berkeley.

Gomez-Pompa, A., H. Morales, E. Jimenez, and J. Jimenez. 1 9 8 2 Experiences in traditional hydraulic agriculture. In Pre-hispanic Maya Agriculture, edited by Kent Flannery, pp. 327-342. Academic Press, New York.

Graybill, Donald A. 1 9 8 9 The reconstruction of prehistoric Salt River streamflow. In The 1982-1984 Excavations at Las Colinas: Environment and Subsistence, by Donald A. Graybill, David Gregory, Fred Niais, Suzanne Fish, Robert Gasser, Charles Miksicek, and Christine Szuter, pp. 25-38. Arizona State Museum Archaeological Series 162. University of Arizona, Tucson.

Grebinger, Paul E. 1976 Salado--perspectives from the middle Santa Cruz Valley. The Kiva 42: 39-46.

Gumerman, George P., ed. 1 9 8 8 The Anasazi in a Changing Environment. Cambridge University Press, Cambridge.

1 99 1 Changing Views of the Hohokam. University of New Mexico- Press, Albuquerque. 54

Hackenberg, Robert A. 1974 Aboriginal landuse and occupancy of the Pima- Maricopa Indians. American Indian Ethnohistory I, pp. 288-316. Garland Publishing, New York.

Haury, Emil W. 1950 The Stratigraphy and Archaeology of Ventana Cave, Arizona. University of Arizona Press, Tucson.

1976 The Hohokam: Desert Farmers and Craftsmen. University of Arizona Press, Tucson.

Howard, Jerry 1 9 8 7 The Lehi Canal system: organization of a Classic Period community. In The Hohokam Village: Site Structure and Organization, edited by David E. Doyel, pp. 211-222. Southwestern and Rocky Mountain Division of the American Association for the Advancement of Science, Glenwood Springs, Colorado.

Huckell, Bruce B. 1990 Late Pre-ceramic Farmer-Foragers in Southeastern Arizona: A Cultural and Ecological Consideration of the Spread of Agriculture into the Arid Southwestern United States. Ph.D. Dissertation, Arid Lands Resource Sciences, University of Arizona, Tucson.

Kuckler, A. W. 1967 Vegetation Mapping. Ronald Press, New York.

Masse, R. Bruce 1 97 9 An intensive survey of prehistoric dry farming systems near Tumamoc Hill in Tucson, Arizona. The Kiva 45: 141-186. S S

McGinnes, William G. 1 9 8 8 Climatic and biological classifications of arid lands: a comparison. In Arid Lands: Today and Tomorrow, edited by E. E. Whitehead, C. F. Hutchinson, B. N. Timmermann, and R. B. Varady, pp. 61-68. Westview Press, Boulder.

Miksicek, Charles H. 1 9 84 Historic desertification, prehistoric vegetation change, and Hohokam subsistence in the Salt-Gila basin. In Hohokam Archaeology Along the Salt-Gila Aqueduct, Central Arizona Project, Vol. VII: Environment and Subsistence, edited by Lynn Teague and Patricia Crown, pp. 53-80. Arizona State Museum Archaeological Series 150. University of Arizona, Tucson.

Nabhan, Gary P. 1983 Papago Fields: Arid Lands Ethnobotany and Agricultural Ecology. Ph.D. Dissertation, Arid Lands Resource Sciences, University of Arizona, Tucson.

1986 Ak-chin "arroyo mouth" and the environmental setting of the Papago Indian fields in the Sonoran Desert. Applied Geography 6: 61-75.

Nabhan, Gary P. and Thomas E. Sheridan 1977 Living fence-rows of the Rio San Miguel, Sonora, Mexico: traditional technology for floodplain management. Human Ecology 5: 97-110.

Nabhan, Gary P., Amadeo M. Rea, Karen L. Reichhardt, Eric Mellink, and Charles F. Hutchinson 1982 Papago influences on habitat and biotic diversity: Quitovac oasis ethnoecology. Journal of Ethnobiology 2: 124-143. 56

Netting, Robert 1993 Smallholders, Householders: Families and the Ecology of Intensive, Sustainable Agriculture. Stanford University Press, Stanford.

Orcutt, J. D., E. Blinman, and T. A. Kohler 1990 Explanations of population aggregation in the Mesa Verde region prior to A.D. 900. In Perspectives on Southwestern Prehistory, edited by P. E. Minnis and C. L. Redman, pp. 196-212. Westview Press, Boulder.

Parsons, Jeffrey R. and Mary Parsons 1990 Maguey Utilization in Highland Central Mexico: An Archaeological Ethnography. Museum of Anthropology Anthropological Paers 82. University of Michigan, Ann Arbor.

Plog, Stephen 1 9 9 0 Agriculture, sedentism, and environment in the evolution of political systems. In The Evolution of Political Systems: Sociopolitics in Small-scale Sedentary Societies, edited by Steadman Upham, pp. 177-199. Cambridge University Press, Cambridge.

Rea, Amadeo 1979 The ecology of Pima fields. Environment Southwest 484: 8-15.

1 9 83 Once a River: Bird Life and Habitat Changes on the Middle Gila. University of Arizona Press, Tucson.

Reff, Daniel T. 1991 Disease, Depopulation, and Cultural Change in Northwestern New , 1518-1764. University of Utah Press, Salt Lake City. 57

Russell, Frank 1 9 0 8 The Pima Indians. U.S. Bureau of American Ethnology 26th Annual Report. Washington, D.C.

Shaner, W. W., P. F. Philipp and W. R. Schmehl 1982 Farming Systems Research and Development. Westview Press, Boulder.

Vivian, R. Gwinn 1990 The Chacoan Prehistory of the San Juan Basin. Academic Press, San Diego.

Weaver, Donald E., Jr. 1972 A cultural-ecological model for the Classic Hohokam period in the lower Salt River valley, Arizona. The Kiva 38: 43-52.

Wetterstrom, Wilma 1 986 Food, Diet, and Population at Prehistoric Arroyo Hondo Pueblo, New Mexico. School for American Research Press, Santa Fe.

Wilken, Gene C. 1987 Good Farmers: Traditional Agricultural Resource Management in Mexico and Central America. University of California Press, Berkeley. 58

APPENDIX A

THE MODIFIED ENVIRONMENT OF THE

SALT-GILA AQUEDUCT SITES

by Suzanne K. Fish

1984. In "Hohokam Archaeology along the Salt Gila Aqueduct, Central Arizona Project: Volume VII: Environment and Subsistence," edited by Lynn S. Teague and Patricia L. Crown, pp. 39-52. Arizona State Museum Archaeological Series No. 150. 59

HOHOKAM ARCHAEOLOGY ALONG THE SALT GILA AQUEDUCT CENTRAL ARIZONA PROJECT Volume VII: Environment and Subsistence

Edited by Lynn S. Teague and Patricia L. Crown

CULTURAL RESOURCE MANAGEMENT SECTION ARIZONA STATE MUSEUM THE UNIVERSITY OF ARIZONA

1 94 ARCHAEOLOGICAL SERIES NO. 150 60

HOHoKAM ARCHAEOLOGY ALONG THE SALT-GILA AQUEDUCT

CENTRAL ARIZONA PROJECT

Volume VII: Environment and Subsistence

Edited by

Lynn S. Teague and Patricia L. Crown

Contributions by

Suzanne K. Fish Patricia L. Crown Charles H. Miksicek Russell J. Barber Christine R. Szuter Frank Hull

Submitted by

Cultural Resource Management Division Arizona State Museum University of Arizona

Prepared for

United States Bureau of Reclamation

Contract No. 0 - 07-32-V0101

1984

Archaeological Series No. 150 61

Arizona State Museum The University of Arizona Tucson, Arizona 85721 Telephone (602) 621-4794 FAX (602) 621-2976 BITNET: Archaeo @ ARIZR VAX

October 15, 1993 Suzanne K. Fish 2214 E. 4th St. Tucson, AZ 85719

Dear Suzanne:

Thank you for your letter requesting permission to reproduce two of your chapters, "The Modified Environment of the Salt-Gila Aqueduct Sites" and "Agriculture and Subsistence Implications of the Salt-Gila Aqueduct Project Pollen Analysis" from Hohokam Archaeology along the Salt-Gila Aqueduct, Central Arizona Project, Vol.. VII: Environment and Subsistence, edited by Lynn S. Teague and Patricia L. Crown. Arizona State Museum Archaeological Series 150. Permission extends to microfilming and publication by University Microfilms Incorporated. I understand that University Microfilms may sell single copies of the dissertation containing the copyrighted material.

I am pleased to grant permission for reproduction of the chapters in your dissertation contingent on full credit for the original publication.

Sin

Raymond Thompson, Director 6 2

Chapter 1

THE MODIFIED ENVIRONMENT OF THE SALT-GILA AQUEDUCT PROJECT SITES: A PALYNOLOGICAL PERSPECTIVE

Suzanne K. Fish

Environmental studies for the Salt-Gila Aqueduct (SGA) Project began in the summer of 1980, concurrent with the initiation of fieldwork. The design for this research was created through consideration of the kinds of sites, features, and situations known or anticipated for the project area (Fish and others 1982; Fish and Miksicek 1982).

A very broad definition of the word "environment" was subsumed under the label of "environmental studies." A primary goal of these studies included identifying evidence of the natural environment during the periods of archaeological occupation. In the case of vegetation, evidence usually consists of a climatically controlled, regional climax configuration. However, environmental research on the SGA Project has also been directed toward detecting the form and extent of human modification of that configuration.

The primary pollen data cited in this discussion are presented in appendices to the other SGA volumes. Methodology employed for collection, extraction, and tabulation is described earlier in this volume. Average pollen frequencies for sites and for grouped samples representing a number of categories can be found in Chapter 1, Part III, of this volume.

Baseline Vegetation of Salt-Gila Aqueduct Project Sites

Palynology originated in studies of lake and bog sediments from forested areas. In such situations, relatively homogenous mixes of arboreal pollen are indicators of regional forest type. The authors of a study of modern Sonoran Desert pollen rain concluded that no regional pollen rain could be defined (Hevly and others 1965: 135). By this they meant that pollen in desert surface sediment strongly reflects immediate herbaceous or shrubby communities rather than a consistent mixture of windborne pollen characterizing the region as a whole. A similar 6 3

correlation of pollen types with local vegetation has been reported for desert grasslands (Martin 1963: 15). The nature of the desert pollen record presents a stumbling block to reconstruction of region-wide vegetation and climatic change by introducing response to complex, but geographically limited edaphic conditions. At the same time, it creates a record with the potential spatial sensitivity to allow detection of palynological hallmarks of localized cultural modifications.

Prehistoric vegetation in the SGA Project study area was a variant of the present desert scrub community. None of the SGA Project biological research suggests that the environment during the Hohokam occupation differed radically from the modern environment. The strong probability that grasses may have been more important and creosote less so is discussed in this volume by Miksicek (Chapter 2, Part II, this volume). Substantial occupations spanning the period from Gila Butte to Civano times (roughly 700 years), undoubtedly encompassed climatic variation. However, historic vegetation change in the lower desert of the Southwest has been characterized as less striking and consistent than in the desert grassland and higher zones (Hastings and Turner 1965: 270).

SGA Project sites occur today in vegetation dominated by creosote bush and bur sage. Other elements important locally include palo verde, ironwood, mesquite, saguaro, and cholla. A few sites occur in a transitional zone between creosote-bur sage and the palo verde- saguaro association typical of lower bajada slopes (Lowe 1980: 24). The most frequent modern pollen type from the sites is Ambrosia-type, a low- spined grain produced by members of the Compositae or sunflower family. Bur sage pollen is included in this category, and these shrubs appear to account for the prominent representation of the type. The predominance of low spine Compositae types in creosote and bur sage communities has also been established in other studies (Hevly and others 1965; Schoenwetter and Doerschlag 1971; Gish 1975, 1979a). A study by Schoenwetter and Doerschlag (1971) included samples from the SGA Project area.

Contemporary pollen spectra from SGA Project sites show little variability. The archaeological samples are more diverse, even within the best approximations of a single time horizon. The source of prehistoric heterogeneity in pollen spectra, and presumably in localized vegetation, appears to have resulted from a more intense and heterogenous cultural influence on study area vegetation than at present. This prehistoric variation and its interpretation is discussed at greater length in following sections of this chapter.

Without reference to pollen types that may be closely related to prehistoric modification, some ways in which the past environment of the study area differs from the present are apparent. Pollen of riparian trees, now either very, rare or absent in the study area, appears as trace amounts at many sites. Willow (Salix) is the major riparian type. Some hackberry (Celtis) pollen also may originate from riparian species 6 4

rather than from shrubby forms occurring in mesic but nonriparian locales.

Pollen of cattail (Typha) was recovered at several sites of different periods. The presence of cattail documents the persistence of ponds or very moist habitats along drainages throughout the year. These may be related to the former presence of beaver and to a more regular streamflow prior to damming, heavy , and agricultural water pumping. Cattail pollen was present in both Queen Creek and Gila River sites. Reservoirs and canals do not appear to have been the sources of these plants (no cattail pollen was recovered from these features). The palynological indication of riparian flora during Hohokam occupations accords well with reconstructions by Miksicek (Chapter 2, Part II, this volume).

Interpreting the Record of Hohokam Sites

The difficulty of separating natural from culturally biased patterns in archaeological pollen records has stimulated much comment. In some interpretations, kinds and frequencies of pollen present in sites have been treated as variables almost wholly controlled by climatic fluctuation. The other end of the interpretive spectrum has been the assignment of most herbaceous pollen to an undifferentiated "disturbance" category. The problem is compounded in Hohokam sites by the fact that plants producing the most abundant pollen are characteristic both of undisturbed and disturbed desert vegetation. Pollen morphology within the Compositae and Chenopodiaceae families plus the genus Amaranthus (Cheno-Am) is seldom sufficiently distinctive to allow differentiation of individual species.

When pollen is abundant and large tabulations are possible, selected pollen types can be removed from the sum used to calculate percentages (Faegri and Iversen 1964: 83-85). In this way, distributional patterns in less frequent types become more apparent. Pollen is not usually sufficiently abundant in sediments at Hohokam sites to make this a practical technique. In an instance where sample tabulations were made exclusive of Cheno-Am pollen (McLaughlin 1977), results magnified frequencies of other possible disturbance taxa. Boerhaavia (discussed later as an agricultural weed) and Compositae became more prominent in the record.

Arguments for interpreting Hohokam pollen spectra as records of natural environment have been made by Schoenwetter and Doerschlag (1971). They interpreted modern surface samples from Sonoran Desert scrub as indicating a correlation between common herbaceous pollen types and a relative scale of xericity. In this study, Cheno-Am pollen was found indicative of high effective moisture. In a subsequent discussion of five surface samples from the Lake Roosevelt area (Schoenwetter 1979: 168-170; and one reported by Bohrer [1971] from Snaketown), he found high 6 5

Cheno-Am pollen to reflect xericity compared to more mesic conditions expressed by Tubiliflorae (a division of Compositae). He proposed that the pollen rain reflects the ecological potential of a sampling locale even though frequencies do not correlate with relative abundance of source plants, and anticipated vegetation associations are not present.

Many palynologists have felt unable to discount the probability of important human influence on the pollen records of Southwestern desert archaeological sites, for example, Martin (1963), Jelinek (1966), Bohrer (1970), Spaulding (1974), McLaughlin (1977), and Lytle-Webb (1978). It is difficult to consistently attribute the overwhelming amounts of Cheno-Am pollen in sites of many different periods and in many different geographic locations to natural environmental conditions. In the literature just cited the strong association of chenopods and amaranths with human disturbance has been correlated. It has also been suggested on the basis of modern observations that subgroups of Compositae pollen have increased with disturbance, for example, Bohrer (1970), Gish (1979a), and Fish (1982a).

Most comparisons of modern pollen and source vegetation undertaken as archaeoloO.cal analogues have focused on the least disturbed vegetation available. Thus, it remains difficult to examine the magnitude and expression of cultural influence even at the present time. The few investigations designed to sample particular kinds of cultural disturbance have shown clearly that the pollen record does reflect a bias from human activity. An analysis of over 75 paired samples from fields and adjacent natural vegetation at many elevations in Oaxaca, Mexico, demonstrates a variety of patterned differences in the cultivated loci. The greatest divergence between pairs occurred where adjacent vegetation was least disturbed (Fish 1974, 1983a). Lytle-Webb (1978: 24-25) reports substantial differences in arboreal pollen from fields and control samples at two New Mexico pueblos. On the Papago Indian Reservation near Tucson, Lytle-Webb found over 30 percent more Cheno-Am pollen in habitation and field proveniences than in comparable natural vegetation (Lytle-Webb 1978: 24-25). The same palynological expression of disturbance was repeated in additional samples from Papago fields and near structures (Miksicek and Fish 1981). Palynological bias also can be found at small, limited activity loci in the desert without extended habitation or cultivation. It is not as consistently detectable or nearly so pronounced in effect, however (Gish 1975; Miksicek and Fish 1981).

To simply associate pollen types or frequencies with cultural disturbance provides minimal information. The disturbance pollen assemblage may be a key to particular kinds of disturbance that can be investigated in other ways. A disturbance flora is also related to the regional climax vegetation in that its constituents are determined by locally available plants that can flourish in the microhabitats created or extended by cultural activity.

In this vein, polarity regarding the interpretation of herbaceous archaeological pollen as predominantly reflecting natural 6 6

vegetation or locally modified plant communities may mask the relationship between growth conditions for vegetation and palynological representation. For example, chenopods and amaranths are abundant and are, accordingly reflected in the pollen of riparian habitats in the Southwest (Martin 1963; Hevly and others 1965).

These habitats are areas of natural soil disturbance, enrichment, and increased moisture. The same factors encouraging these plants occur in archaeological sites. Soil disturbance and enrichment are obvious, but moisture availability may also be enhanced in sites. Disturbance of soil allows significantly greater penetration and absorption of water when rains occur. Also, nearby fields may be receiving supplemental water. Although conditions at sites may mimic favorable, mesic natural habitats, a climatic interpretation is not in order.

An Alternative Perspective on Hohokam Pollen

Part of the debate over environment-versus-culture as the major signal in archaeological pollen can be viewed as conceptual. Environment can be narrowly defined as that biotic (or specifically, vegetational) community that would exist in an area if no human beings had ever been present. Such a definition is appropriate for some problems, but does not adequately describe the surroundings in which human beings exist and subsist. Culturally modified vegetation can be treated as an integral part of the total environment, if human beings are viewed as one element of an ecological community (a community whose final form results from the interaction of all components).

Therefore, separation of that part of the prehistoric environment influenced by man from an ideal natural climax type may be an artificial distinction. In terms of human beings, any pollen record is, in a sense, environmental. Human subsistence cannot be understood in terms of either modified or unmodified aspects of the environment alone. With agriculturalists, the culturally shaped portion of the environment assumes an extensive and unique character. The investment of time and the reduced mobility necessary for agricultural success may sharply focus subsistence pursuits in the general vicinity of sites. The flora of Hohokam sites and fields deserves investigation as an unknown and potentially important aspect of a complex subsistence strategy.

The potential for interpreting both environmental and cultural aspects of the SGA Project pollen spectra would be realized if prehistoric patterns could be matched with a particular kind of disturbance in a particular kind of baseline vegetation. Unfortunately, the combinations of environment and culture that might allow this precision of interpretation cannot be fully duplicated. Alteration of New World environments in postcontact times has been great; 6 7

industrialized populations and domestic animals have changed the landscapes familiar to the Hohokam. Even the most traditional Native Americans do not maintain a wholly aboriginal lifestyle. Therefore, many lines of evidence, incorporating both modern analogues and prehistoric patterns must be used to interpret archaeological spectra.

A Palynological View of Hohokam Agriculture

As a means of understanding the modified environment of Hohokam sites, both exploratory efforts and a substantial part of the palynological research were concentrated on contexts directly associated with agricultural activity. Irrigated fields, canals, reservoirs, and nonirrigated agricultural features were sampled to characterize microenvironments associated with sites, to locate production of specific crops, and to obtain evidence of cultivation techniques and strategies. Such primary agricultural contexts are a potential source of information on aspects of subsistence that cannot be interpolated from habitation proveniences.

Irrigated fields were sampled at Frogtown (AZ U:15:61) on Queen Creek. Probable locations of the fields were defined on the basis of proximity and elevational relationships to both primary and secondary canals. Stratigraphic columns were then collected at eight separate locations. At the tops of these columns, samples matched modern distributions. After a transition at less than 3U cm below the surface, a different configuration of pollen types and a few grains of cultigens marked the levels of Hohokam fields (Fish 1983b).

A combination of three types represented a major portion of the pollen record at Frogtown. Thirty-five to 45 percent of all pollen present consisted of spiderling (Boerhaavia-type), globemallow (Sphaeralcea), and Arizona poppy (Kallstroemia). Average frequencies for individual types are presented in Part III of this volume (Table III.1.1). Unusual but somewhat lower frequencies of spiderling, globemallow, and Arizona poppy had been found in residential areas of Frogtown and other sites. The sediments of canals and reservoirs at Frogtown and two other sites contained elevated amounts of these three types as well.

Comparisons with Modrrn Pollen from Undisturbed Vegetation

Today, agricultural pollen types occur in samples from natural vegetation communities in the southern Arizona desert, but never in similar combined frequencies and rarely as individual types. Pollen of the family that includes spiderling (Nyctaginaceae) is often encountered near Tucson on mountain slopes and upper bajadas (the greatest slope value was 8 percent). It is absent in creosote bush-bur sage vegetation 6 8

(Hevly and others 1965: 129). In 15 samples from desert scrub beginning north of Phoenix and ending near Florence in the project area (Schoenwetter and Doerschlag 1971: 217), the highest amount of pollen in this family is 0.25 percent. In modern SGA Project samples a peak of 4 percent was recorded. Even this amount may have been influenced by archaeological deposits eroding to the present ground surfaces. The maximum modern record of the category encompassing spiderling pollen (13 percent) was collected in a creosote-bur sage community near Yuma at a sampling locus near irrigation canals (Hevly and others 1965: 130).

Pollen of Arizona poppy and globemallow follow similar distributions in modern vegetation communities. These types are only sporadically present in creosote-bur sage communities (Hevly and others 1965; Schoenwetter and Doerschlag 1971: 217), but appear fairly often in upper bajada and mountain slope vegetation. A second zone having consistent, if very low, representations of these types is the floodplain. Thus, the triad of major prehistoric agricultural pollen types are most strongly associated today with vegetation communities more mesic than that found at the SGA Project sites. This correlation seems less likely to be evidence of moister-than-present conditions throughout the long span of prehistoric occupations in which they occur, but rather the cultural creation of mesic habitats for herbaceous plants. Furthermore, no natural plant associations of the Lower Sonoran Desert produce (at present) frequencies of these pollen types comparable to Salt-Gila agricultural loci and habitation sites.

Agricultural Weed Flora Analogues

It would be very satisfying if the constellations of prehistoric pollen associated with agricultural proveniences could be matched with spectra from fields today. This unfortunately cannot be done, h , lwever, due to postcontact vegetational and cultural changes. Old World species have become naturalized and are often the most common plants growing under disturbed conditions. In even the most traditional of Native American groups, prehistoric cultivation techniques have not remained unchanged.

In modern fields in the SGA Project study area, weeds are often sparse. Plowing and chemical herbicides discourage weeds in fields supporting cotton, a major local crop. In one corn field of limited size near the Gila River, weeds were allowed to remain and grew prodigiously. Mechanized agriculture involves frequent, thorough, and deep turning of the earth (yearly plowing insures that only annuals will grow). Careless weed, Amaranthus palmeri, was the single species present in any quantity in the cornfield. Pollen of chenopods ard amaranths was much more frequent in some habitation site and canal samples than in Frogtown fields. Amaranths appear to have been less important in the Hohokam fields than in the modern ones. The differences in field flora are probably related to differences between aboriginal and modern cultivation practices. 6 9

In a study by Nabhan and Tiewes (1983), careless weed was found to be abundant in modern Papago floodwater fields. Cheno-Am pollen is likewise by far the most frequent type (Webb 1981; Miksicek and Fish 1981). Although most Papago fields are plowed annually, spiderling and globemallow are present. Spiderling is repeatedly recorded and in one field was the third most abundant weed. It attained the same prominence in grazed and disturbed but unplowed areas adjacent to fields. One plot studied by Nabhan was not plowed, but only hoed. In this field with less uniformity and deep disturbances, both spiderling and Arizona poppy occurred. An account of range conditions and management trials, early in this century, in connection with the University of Arizona Agicultural Experiment Station (Thornber 1910), provides a valuable analogue for SGA Project samples. The report includes descriptions of areas from just south of Tucson to Picacho. At this early date, based on groundwater pumping had yet to occur. Vast tracts of Old World weeds such as Russian thistle did not exist in association with mechanized agriculture. Seed sources for introduced plants were not so widespread as at present.

Thornber (1910: 324) reports effects of soil cultivation on native annuals undertaken to enhance grazing at lower elevations. Summer species mentioned are grasses, Arizona or Mexican poppy, and several kinds of Boerhaavia (spiderling). Because no other species are included, it can be assumed that these are among the most abundant. These plants grew larger, yielded more, and remained green longer in areas that had been harrowed or otherwise cultivated, than in similar undisturbed areas, although the number of individual plants was fewer. "This increased yield of annual plants is due no doubt to the fact that a larger percentage of rainfall remains in the ground longer where it is available for plant growth" (Thornber 1910: 324). Harrowing or disking for range improvement prior to 1910 should have created a shallow, limited disturbance more similar to Hohokam cultivation than the disturbance in modern mechanized and leveled commercial fields. The correspondence between the 1910 range weeds encouraged by cultivation and pollen types from Hohokam fields strengthens the inference of localized abundance of agricultural weeds prehistorically.

Vegetation along poorly maintained roadsides and ditches represents another useful analogue for the flora of Hohokam irrigated agriculture. After initial disturbance, the soil is not turned annually, and herbaceous perennials can develop. Supplemental water beyond rainfall is delivered by road runoff or the ponding effect of ditches. In such situations, floristic matches with Hohokam agricultural pollen can be made. Spiderling and Arizona poppy were prominent along roadsides near Picacho Peak, Arizona (Picacho Peak is 24 km from the SGA Project sites) after heavy summer rains in 1981. Globemallow was another common roadside plant in the vicinity. Between Picacho and the study area, large ditches contained sunflowers (high spine Compositae) and abundant globemallow. 70

Annuals and Perennials

Hohokam cultivation by hoe or digging stick was not likely to have disrupted the weedy vegetation of fields uniformly. Probable tolerance of useful weeds would have resulted in selected plants persisting from year to year. Perennials might then have become established, increasing the diversity of field vegetation. For example, spiderling and globemallow are perennial. Although species cannot be distinguished through distinctive pollen morphology, ones associated with agriculture today are the likely field weeds of the past.

A publication on Arizona , farm, and garden weeds by the University of Arizona Agricultural Extension Service (1958) lists B. coulteri and B. coccinea. The latter, commonly called spiderling, is perennial. The genus Sphaeralcea is mostly perennial (Kearney and Peebles 1964: 540). S. emoryi is one of two species described as favoring disturbed areas by Kearney and Peebles (1964: 542-543). It occurs along roadsides and in fields, and one variety is especially common in irrigated districts. S. angustifolia occurs on roadsides and edges of cultivated fields (Kearney and Peebles 1964: 545).

In an investigation of the possibility of perennials, succession in modern abandoned fields can be compared to the prehistoric field spectra. Here again the search for analogues is complicated. Plowed fields create conditions for field succession unlike prehistoric fields. Plants growing in abandoned fields cease to benefit from supplemental water, in contrast to weeds coexisting with crops. Most importantly, early stages of succession throughout the western United States are dominated by Old World species, Russian thistle being the most notable Southwestern example (Karpiscak 1980).

In a study of southern Arizona field succession encompassing the SGA Project area (Karpiscak 1980), Russian thistle invaded, then gave way to Old World mustards after two to three years. Native chenopods and amaranths, Arizona poppy, and other species may represent similar recently established growth in irrigated fields. Globemallow and perennial spiderling may represent substantial amounts of vegetation persisting in fields from season to season. Globemallow increased in modern abandoned fields after two to three years.

Season of Cultivation

Ecological attributes of source plants for agricultural pollen types have implications for cultivation practices. Seasonal emphasis may be suggested. Two of the three genera are strong indicators of summer growth. Arizona poppy is a summer annual. Kallstroemia grandiflora, today a weed of irrigated fields in valleys and along roadsides, appears mostly from July through August (University of Arizona Agricultural Extension Service 1958: 176). Near Picacho, 71

Arizona, it was observed to proliferate on roadsides following summer rains. The genus Boerhaavia, spiderling, is reported by Kearney and Peebles (1964: 275) as flowering in late summer and autumn. B. coulteri is an annual weed, especially common and annoying in southern Arizona in late summer (University of Arizona Agricultural Extension Service 1958: 112). B. coccinea, a weed of gardens, fields, and roadsides, blooms mostly from August to October (University of Arizona Agricultural Extension Service 1958; Kearney and Peebles 1964: 276).

Globemallow or Sphaeralcea is less securely tied to one season. There are many species, a number of which flower and produce pollen in the spring and again after the summer rains (Kearney and Peebles 1964: 54). S. emoryi is said to flower mostly in the spring (Kearney and Peebles 1964: 542-543). S. angustifolia flowers from May to October (Kearney and Peebles 19o4: 545). Globemallow was seen in the study area blooming profusely in 1983 after much winter rain, and throughout the summer as well.

The three weedy plants discussed do not exhaust the list of species in and around the fields sampled at Frogtown. Yet all three are low contributers to the modern pollen records. Furthermore, frequencies for none of these types can be matched in natural vegetation. Therefore, these three genera must have been among the most abundant in the fields. It seems apparent that summer irrigation encouraged thick stands of spiderling and Arizona poppy. No equally clear palynological markers for early spring planting is well represented. Even though such types might also be rare, a similar opportunity for irrigated proliferation and magnified pollen production would have existed during spring cultivation. Elevated globemallow pollen might result from a spring crop, but the situation is ambiguous. Certainly, major emphasis on summer farming seems indicated.

The volume of seasonal flow in the Salt River and the size of Phoenix-area canals implied emphasis on spring irrigated crops to Masse (1981). The pollen of fields at Frogtown suggests a contrasting seasonal strategy on this section of Queen Creek. Differences in streamflow, frost risk, or other factors may have guided the inhabitants of this site in their agricultural decisions. Unfortunately, no SGA Project samples were available from fields irrigated from the Gila River or other smaller drainages. Comparison of such sets of samples remains an exciting avenue for future research.

Variation in the Modified Vegetation

A fourth pollen type has not been discussed with the preceding three because a unique field asociation is not quite so clear. Pollen of the Evening Primrose family (Onagraceae) resembling that of the genus Oenothera, appeared in a number of Frogtown field samples in amounts below 3 percent. Such representation is conceivable as a background 7 2

level. In canal samples, however, as much as 7 percent was found, and appearance of the type was more consistent. The genus Oenothera and most others in the family are spring indicators. It seems quite possible that these plants were fairly specific to canal habitats. Canals might catch and retain moisture in the channels during periods when irrigation water was not present, so that plant growth and cultivation season would not be entirely synchronous. Indeed, these weeds might have been growing when the cultivated fields were cleared in preparation for irrigation.

There are some additional patterned differences between samples from the fields at Frogtown and canal and reservoir samples at this and other sites. In the latter two contexts, spiderling, Arizona poppy, and globemallow also are again prominent. Globemallow has an even higher representation (12.3 percent for all canal samples and 10.5 percent for reservoirs versus 8.1 percent for the fields). Like Onagraceae species, globemallow may have been even more abundant along canals and in reservoirs than in fields. Greater persistence of moisture, favoring perennials, may have favored globemallows. Cheno-Am pollen is also somewhat higher in canals, but elevated occurrences are sporadic. Differences between fields, canals, and reservoirs define microhabitats existing within the culturally modified surroundings of Hohokam sites. These variations would have diversified immediate resources just as variation in natural vegetation created diverse opportunities on a wider scale.

Sedimentary Regime of Irrigated Fields

The distinctive pollen of Hohokam fields indirectly adds detail to another facet of the agricultural environment. Samples were collected from stratigraphie columns in trenches through suspected fields. Field levels were anticipated by elevational reference to adjacent canals, but pollen confirmed cultivation horizons not visibly detectable. Only the top 10 cm to 20 cm resembled modern spectra. Between 10 on and 3u cm in different columns, the agricultural configuration appeared. An interesting by-product of this study on vertical pollen distributions was that it provided a means of comparing deposition rates during and after prehistoric agriculture. The agricultural types remained prominent in some columns for 4,, cm tn 50 cm. Deposition appears to have been more rapid during field cultivation than in the period after prehistoric occupation until the present. The sediment load of Queen Creek irrigation water is the likeLy source for accelerated accumulations. As today in traditional floodwater farming (Sheridan 197 8; Nabhan and Tiewes 1983), this waterborne material must have enriched and renewed the Hohokam fields. 7 3

Habitation Samples and Environment

Environmental and climatic reconstructions have been based on pollen spectra from habitation sites, as these contexts have furnished the great majority of samples. Although spatially removed to some degree from the bias of field weeds, site samples also present a culturally filtered view of the environment. The three prominent agricultural pollen types appear in residential areas of Frogtown and other SGA Project sites in appreciable quantites, although less strongly than in the fields (Table III.1.1). Their anomalous presence has also been widely noted in other Hohokam sites, as discussed in connection with subsistence (Chapter 1, Part III, this volume). On the other hand, Cheno-Am pollen is higher in habitation proveniences than in the fields (Tables 111.1.2 and 111.1.3). This type, so abundant in modern Papago fields (Lytle-Webb 1978; Miksicek and Fish 1981), is more an "urban" weed at Frogtown. At some other project sites, Cheno-Am pollen is more abundant still. Disturbed and organically enriched residential precincts supported their own suite of weedy plants. Pollen spectra reflect this source and the flora of nearby fields.

Nonirrigated Agricultural Features

Samples were analyzed from the environs of rock features investigated for nonirrigated agriculture at the Jones Ruin (AZ U:15:48), Smiley's Well (AZ U:14:73), the Ellsworth Site (AZ U:15:57), and at a series of loci designated AZ U:15:46. Features of this nature appear to date from at least the Santa Cruz phase, and perhaps earlier, to the Classic period. The triad of pollen prominent in irrigated fields was repeated in the runoff features, but in lesser amounts (Table 111.1.2). Runoff plantings should, of course, receive less supplemental water and, therefore, less dense weedy growth. The average combined total for the types in runoff farming samples is 8.5 percent compared to 35 to 45 percent for irrigation. This figure is distinctive from most modern samples in undisturbed vegetation, but it may not adequately portray typical frequencies in such features.

It was difficult or impossible for excavators to identify precisely prehistoric planting surfaces associated with nonirrigated features. Both rock features and adjacent areas were sampled. Some samples may pertain to the correct prehistoric level, but many may be associated with other time periods. Some samples were virtually free of the types in question, while others exhibited combined frequencies as high as 28.5 percent. The 8.5 percent average reflects this hit-and- miss sampling to an unknown degree. Summer planting in nonirrigated plots seems to be indicated. 7 4

Conclusion

A unique aspect of the SGA Project study has been the implementation of a research design oriented toward the greatest variety of cultural contexts. Pollen spectra were not consistent across the cultural landscape, and this inconsistency has proved the most productive basis for cultural interpretation. Pollen deposited in Hohokam sites is a dynamic mixture of windblown increments from baseline vegetation, anthropogenic plant communities, and even from plants carried in by the inhabitants. To characterize Hohokam pollen spectra as a record of natural environment is only a partial truth; to treat them as pure records of natural vegetation and climate would be misleading.

The SGA Project pollen evidence suggests that agricultural habitats and habitation areas were characterized by different species of weeds. It appears likely that on Queen Creek, major agricultural components consisted of plants responding to summer moisture and perennial elements persisted along with ephemerals. The geographic variability of culturally modified vegetation throughout the Hohokam domain remains to be explored. Even within the SGA Project study area, differences among sites suggest greater variability in cultivation practices and a host of other factors.

The environment depicted by SGA Project pollen analysis was richer and more varied than the creosote and bur sage vegetation of the modern study area or any prehistoric variation of this type. It contained plant resources in a proximity, diversity, and abundance that was shaped by human activity. This culturally modified landscape, a product of both conscious and unconscious manipulation, offers a new perspective on the meaningful environment of the Salt-Gila Hohokam. 75

References Cited

Bohrer, Vosila L. 1 970 Ethnobotanical aspects of Snaketown: a Hohokam village in southern Arizona. American Antiquity 35 (4): 413- 43 O.

1971 Paleoecology of Snaketown. The Kiva 36 (3): 11-19.

Faegri, Knut, and Johannes Iversen 1964 Textbook of Pollen Analysis. New York: Hafner Publishing Company.

Fish, Suzanne K. 1 9 7 4 Palynology of some cultural contexts. Paper presented at the 73rd Annual Meeting of the American Anthropological Association, Mexico City.

1982a Palynology of the Modified Vegetation of the Salt-Gila Hohokam Sites. Paper presented at the 47th Annual Meeting of the Society for American Archaeology, Minneapolis. MS, Arizona State Museum Library, University of Arizona, Tucson.

1983a Culture in Environment: Palynological Interactions. Paper presented at the 48th Annual Meeting of the Society for American Archaeology, Pittsburgh. MS, Arizona State Museum Library, University of Arizona, Tucson.

1 9 8 3 Pollen from agricultural features. In the "Hohokam Archaeology along the Salt-Gila Aqueduct, Vol. III: Specialized Activity Sites," edited by Lynn S. Teague and Patricia L. Crown. Arizona State Museum Archaeological Series 150(3): 575-603. Tucson: University of Arizona. 7 6

Fish, Suzanne K., and Charles H. Miksicek 1982 Sampling guidelines for pollen and flotation. Appendix C in "Hohokam Archaeology along the Salt-Gila Aqueduct, Central Arizona Project, Vol. I: Research Design," edited by Lynn S. Teague and Patricia L. Crown. Arizona State Museum Archaeological Series 150(1): 169-173. Tucson: University of Arizona.

Fish, Suzanne K., Russell J. Barber, and Charles H. Miksicek 1982 Environment and subsistence studies. In "Hohokam Archaeology Along the Salt-Gila Aqueduct, Central Arizona Project, Vol. I: Research Design," edited by Lynn S. Teague and Patricia L. Crown. Arizona State Museum Archaeological Series 150(1): 129-140. Tucson: University of Arizona.

Gish, Jannifer W. 1975 Preliminary report on pollen analysis from Hecla I, II, and III. In "Hecla II and III: An Interperspective Study of Archaeological Remains from the Lakeshore Project, Papago Reservation, South-Central Arizona," by Albert C. Goodyear, III. Arizona State University Anthropological Research Paper 9. Tempe: Arizona State University.

1979a Palynological research at the Pueblo Grande Ruin. The Kiv a 44(2-3): 159-172.

Hastings, James R., and Raymond M. Turner 1965 The Changing Mile. Tucson: University of Arizona Press.

Hevly, Richard H., P. J. Meringer, Jr., and H. G. Yocum 1965 Studies of the modern pollen rain in the Sonoran Desert. Journal of the Arizona Academy of Science 3(3): 123- 125.

Jelinek, Arthur J. 1966 Correlation of archaeological and palynological data. Science 152: 1507-1509. 7 7

Karpiscak, Martin M. 1980 Secondary Succession of the Abandoned Field Vegetation in Southern Arizona. MS, doctoral dissertation, University of Arizona, Tucson.

Kearney, Thomas H., and Robert H. Peeples 1964 Arizona Flora, 2nd edition. Berkeley: University of California Press.

Lowe, Charles H. 1980 Arizona's Natural Environment, 2nd edition. Tucson: University of Arizona Press.

Lytle-Webb, Jamie L. 1978 Pollen analysis in Southwestern archaeology. In Discovering Past Behavior: Experiments in the Archaeology of the American Southwest, edited by Paul Grebinger, pp. 13-28. New York: Gordon and Breach.

1 98 1 Pollen studies at Las Colinas. In "The 1968 Excavations at Mound 8 Las Colinas Ruins Group, Phoenix, Arizona," edited by Laurens C. Hammack and Alan P. Sullivan. Arizona State Museum Archaeological Series 154. Tucson: University ofArizona.

Martin, Paul S. 1963 The Last 10,000 Years, a Fossil Pollen Record of the American Southwest. Tucson: University of Arizona Press.

Masse, W. Bruce 1981 Prehistoric irrigation systems in the Salt River Valley, Arizona. Science 214 (4519): 408-415.

McLaughlin, Diane 1977 Analysis of pollen from the Baca Float sites. In "Excavations in the Middle Santa Cruz Valley, Southeastern Arizona," by David E. Doyel. Arizona State Museum Contribution to Highway Salvage Archaeology in Arizona 44: 170-177. 7 8

Miksicek, Charles H., and Suzanne K. Fish 19 81 Understanding Aboriginal Subsistence Activities in Southern Arizona. Paper presented at the Fourth Biology Conference, Columbia, Missouri.

Nabhan, Gary P., and Helga Tiewes 19 83 Tepary beans, O'odham farmers, and desert fields. Desert Plants 5: 15-37.

Schoenwetter, James 1979 Initial assessments of the palynological record: Gila Butte-Santan Region. In "An Archaeological Test of Sites in the Gila Butte-Santan Region, South-Central Arizona," by Glen , David Wilcox, Kevin Rafferty, and James Schoenwetter. Arizona State University Anthropological Research Papers 18: 164-177. Tempe: Arizona State University.

Schoenwetter, James, and Larry A. Doerschlag 1971 Surficial pollen records from central Arizona, I: Sonoran Desert scrub. Journal of the Arizona Academy of Science 6(3): 216-221.

Sheridan, Thomas E. 1978 Traditional Agriculture in the Southwest: Past, Present, and Future. MS, Center for Documentary Relations of the Southwest, Arizona State Museum, University of Arizona, Tucson.

Spaulding, W. G. 1974 A preliminary statement on the pollen analysis from the Escalante Ruin Group. In "Excavations at the Escalante Ruin Group, Southern Arizona," by David E. Doyel. Arizona State Museum Archaeological Series 37: 262- 268. Tucson: University of Arizona. 7 9

Thornber, J. J. 1910 The grazing ranges of Arizona. University of Arizona Agricultural Experiment Station Bulletin 65. Tucson: University of Arizona.

University of Arizona Agricultural Extension Service 1958 Arizona ranch, farm, and garden weeds. Agricultural Extension Service Circular 265. Tucson: University of Arizona. 80

APPENDIX B

AGRICULTURE AND SUBSISTENCE IMPLICATIONS OF THE

SALT-GILA AQUEDUCT PROJECT POLLEN ANALYSIS

by Suzanne K. Fish

1984. In "Hohokam Archaeology along the Salt Gila Aqueduct, Central Arizona Project: Volume VII: Environment and Subsistence," edited by Lynn S. Teague and Patricia L. Crown, pp. 111-138. Arizona State Museum Archaeological Series No. 150. 81

HOHOKAM ARCHAEOLOGY ALONG THE SALT GILA AQUEDUCT CENTRAL ARIZONA PROJECT Volume VII: Environment and Subsistence

Edited by Lynn S. Teague and Patricia L. Crown

CULTURAL RESOURCE MANAGEMENT SECTION ARIZONA STATE MUSEUM THE UNIVERSITY OF ARIZONA

1984 ARCHAEOLOGICAL SERIES NO. 150 82

HOHoKAM ARCHAEOLOGY ALONG THE SALT-GILA AQUEDUCT

CENTRAL ARIZONA PROJECT

Volume VII: Environment and Subsistence

Edited by

Lynn S. Teague and Patricia L. Crown

Contributions by

Suzanne K. Fish Patricia L. Crown

Charles H. Miksicek Russell J. Barber

Christine R. Szuter Frank Hull

Submitted by

Cultural Resource Management Division Arizona State Museum University of Arizona

Prepared for

United States Bureau of Reclamation Contract No. 0-07-32-V0101

1984

Archaeological Series No. 150 83

Arizona State Museum The University of Arizona Tucson, Arizona 85721

Telephone (602) 6214794 FAX (602) 621-2976 BITNET: Archaeo ARIZRVA X

October 15, 1993 Suzanne K. Fish 2214E. 4th St. Tucson, AZ 85719

Dear Suzanne:

Thank you for your letter requesting permission to reproduce two of your chapters, "The Modified Environment of the Salt-Gila Aqueduct Sites" and "Agriculture and Subsistence Implications of the Salt-Gila Aqueduct Project Pollen Analysis" from Hohokam Archaeology along the Salt-Gila Aqueduct, Central Arizona Project, Vol.. VII: Environment and Subsistence, edited by Lynn S. Teague and Patricia L. Crown. Arizona State Museum Archaeological Series 150. Permission extends to microfilming and publication by University Microfilms Incorporated. I understand that University Microfilms may sell single copies of the dissertation containing the copyrighted material.

I am pleased to grant permission for reproduction of the chapters in your dissertation contingent on full credit for the original publication.

Sin

Raymond H Thompson, Director 84

Chapter 1

AGRICULTURE AND SUBSISTENCE IMPLICATIONS OF THE SALT-GILA AQUEDUCT POLLEN ANALYSES

Suzanne K. Fish

More pollen samples were analyzed in the course of the Salt-Gila Aqueduct (SGA) Project investigations than in any previous Hohokam project. The question should be asked as to whether the data specifics and the general insights derived from 400 samples were sufficient to justify such numbers, and whether a similar amount of information might not have been derived from fewer. The intersite comparisons made later in this paper are a partial answer to these questions. Contrasts of this nature have seldom before been either possible or attempted. Even on less integrative levels, however, the most productive aspect of SGA Project interpretation has been the opportunity to contrast pollen distributions among the widest range of archaeological proveniences.

The SGA Project research design for environmental and subsistence studies (Fish and others 1982) stressed the diversity of anticipated contexts and detailed a strategy for systemtically sampling them (Fish and Miksicek 1982). When the time came for final selections, thousands of samples had to be eliminated. The ultimate total, large in absolute numbers though it was, had to be divided among many sites and a variety of research interests emphasized by individual investigators. As with most stimulating research, a handful of issues received substantial resolution, while many new possibilities were raised by suggestive patterns requiring further data. As an effort to extend the comparative potential of Hohokam palynology, the SGA Project analysis has generated new interpretations and new bodies of inconclusive but intriguing results.

Some aspects of subsistence technology and stategy have been discussed under the heading of environment. As will be emphasized again, it has proven difficult to compartmentalize categories of environment and subsistence from the perspective of SGA Project pollen data and by inference, from the perspective of the SGA Project Hohokam. Information involving cultigens will be presented first, as a category best retaining its traditional significance. 85

The Pollen Record of Cultigens

Beans

The record of palynologically rare cultigens will be discussed first, since there is little that can be interpreted beyond presence and absence. Beans are least amenable to interpretation for several reasons. The Leguminosae comprise a palynologically heterogenous family. Keys have not been published for the larger numbers of New World and Southwestern species related to cultivated types. The cultivated legumes commonly called beans are produced by more than one genus. For instance, jack beans are Canavalia ensiformis and most other common types are in the genus Phaseolus. Even within the latter genus there is considerable morphological diversity in pollen. In view of the uncertainities of identifying cultivated bean pollen, none has been separated from a more inclusive category of Leguminosae in the SGA Project analysis. Even those palynologists who differentiate Phaseolus report it rarely. For all these reasons, the rarely preserved macrofossils of beans are by far the better source of information for SGA Project use.

Cucurbits

Indications of cucurbits are also relatively infrequent in both macrotossil and pollen recovery. in a small number of instances, cucurbit pollen was identified. Sites yielding the pollen included Las Posas (AZ U:15:19), Smiley's well (AZ 0:14:73), the Gopherette Site (AZ 0:15:87), the Jones Ruin (AZ U:15:48), the Siphon Draw Site okZ U:10:6), and the Reservoir Site (AZ 0:15:98). The greatest representation in a single sample was 10 grains from a house floor at the Jones Ruin. All the grains are of a type bridging one cultigen species and wild cucurbits native to the study area. pollen with spined opercula may be produced by C. pepo (related varieties of and squash) and by wild cucurbits called buffalo . Since the wild cucurbits were used ethnographically, the pollen could be from either source. Cucurbit pollen in SGA Project samples fits a general pattern of sporadic occurrence and overall rarity in Hohokam samples. The record should not be interpreted as indicating limited use of cultivated forms by study area inhabitants.

Cotton

Equally or even more rare than cucurbit pollen has been the pollen of cotton, Gossypium. The few grains found in SGA Project samples have greatly increased the number of palynological documentations for the entire Southwest. This is in contrast to the 86

more common recovery of macrofossil remains in the SGA Project sites and other Hohokam sites. Gasser (1980: 75) found cotton to be one of the six most common species in a consideration of eight sites. Pollen was recovered from the Tonka Site (AZ AA:3:22), Las Fosas, the Gopherette Site, the Chiadag Site (AZ U:15:99), the Jones Ruin, and AZ U:15:46. The unprecedented total of six grains was encounterd on the floor of a field house structure at the Chiadag Site.

There is an ethnographic precedent for trading cotton fibers containing seeds (Russell 1975: 94). Although this cultigen probably was widely grown, pollen may be a stronger confirmation of local production than seeds alone. It is possible that pollen could adhere to traded material. With so rare a pollen type, the probability of introduction into site contexts decreases with processed and transported material; the changes for introduction should be much greater at sites where cotton was grown. The reason for the relatively substantial record of pollen (14 grains) from SGA Project analysis may be twofold. First, there are simply more samples and a better chance to identify rare types. Second, a large proportion of Hohokam samples have come from large sites. Such sites may have received cotton raised or initially processed at outlying, smaller sites.

Most SGA Project sites appear to be closely oriented to agricultural production. Macrofossil and pollen evidence points to cotton as having been a common crop. Cultivation was indicated by pollen recovery both in a field house (the Chiadag Site), in irrigated fields, and in two sets of nonirrigated features (at AZ U:15:46 and the Jones Ruin). Cotton is closely linked with irrigation in modern farming, but the Hohokam achieved production in floodwater or runoff fields as well.

Corn

Corn pollen appears more consistently than other cultigen types in Hohokam sites. It was tabulated in most SGA Project sites and in a wide range of frequencies. Because of its more consistent record, contrasts between proveniences and frequencies are amenable to interpretation beyond presence and absence. Table 111.1.1 illustrates the widest level of contrast in SGA Project samples: fields versus habitation proveniences.

Modern frequencies of corn pollen in corn fields are variable, but almost always quite low (Martin and Schoenwetter 1965; Berlin and others 1977). An atypical finding of 40 percent corn pollen in one Hopi sample (Scott 1981) is unlike other reported results. In samples from 70 fields in highland Mexico, frequencies ranged from absence to 16 percent (average 2.6 percent) in 300-grain tabulations (Fish 1974, 1983a). Corn pollen in three Papago floodwater fields constituted 0.5 percent of all types (Miksicek and Fish 1981). 8 7

Table III. 1.1

AGRICULTURAL POLLEN TYPES IN FROGTOWN SAMPLES

Q) w w a a a in co a +..., a a -4 ,--. __, a a c a a c a a a a Q) a C a) E E •-• .--. a- a .--. cct,--. m m a a a ,u a -. cn --. CO cn E E .--.4 o -o o m m O. 7. O. 7. C c7. 7 .--. __, 7 M M o ...-1 ct m mE w mE w -155 u u a a m c m a u c Q---? 7 C 7 e-c=.' 0 C 0 03 , 17 C ., a u a u...., --4 u u u --. -, al ...-n u) 7 sa o a 03 ii--.P. 03 0) 03 O a ,.0 a3 a) U u a L L , L . a L ,r) 03 4.) 0 .4 •.-. C • .4 ,-4 O ,..., WO ,-, UC ,-. =C-.1 W3.-I a 34 00.0 AI ..0 0 II 0 II 0 II --so a W , m (13 03 Vegetation type Zz-ZUZ 8--2UZ irRUZ- =< Z < 2 = Z =

Evening Primrose Family 36.4 62.5 62.5 7.0 1.2 23.0 2.5 Onagraceae

Globemallow 100.0 100.0 87.0 26.5 10.8 51.5 6.5 Sphaeralcea

Spiderling 100.0 100.0 95.0 42.0 21.9 20.8 8.7 Boerhaavia

Arizona Poppy 91.0 75.0 57.5 20.0 3.3 6.5 1.5 Kallstroemia

In general, prehistoric agriculture might be expected to entail lower frequencies of corn pollen than modern ones. Hand cultivation and tolerance, or useful weeds, would account for greater pollen production by noncultigens within fields, proportionally decreasing the representation of cultigen types as a percentage of total pollen. SGA Project field samples bear out this expectation. Corn pollen averaged 0.04 percent in Frogtown (A2 U:15:61) irrigated fields. In samples from nonirrigated features, this type ranged from absent to 0.5 percent. Corn may not have been raised in every field, but aboriginal mixed cropping practices and the ubiquity of corn pollen and macrofossils in habitation contexts suggests that it was nearly universal.

The opportunity to contrast corn frequencies in fields with residential samples clarifies interpretation of the latter. The suggestion has been made on more than one occasion that low frequencies of corn pollen in habitation features may reflect windblown 8 8

contributions from nearby fields. This possibility is not supported by experimental data (Raynor and others 1970, 1972), nor is it supported by the pollen record of fields. Comparing frequencies in SGA Project fields and habitation sites, the temptation is to suggest that corn pollen blew into the fields from the houses. Concentrations of corn pollen on floors, in pits, in vessels, and on groundstone appear quite clearly to be the result of purposeful human activity.

A second level of contrast, within residential precincts, further strengthens this conclusion. At Las Fosas, samples from plazas furnished a set of 11 extramural spectra without corn pollen. A total of 87 grains was recovered from 20 floor samples. Although, plant materials in structures must have been transported across the plazas, activities concentrating pollen in houses created a record easily differentiated by frequencies.

To interpret greater or lesser abundance of corn pollen in specific features or structures again requires multiple proveniences for comparison. Room function seems likely to have affected the representation of cultigens. SGA Project sites all seem closely aligned with agricultural production. However, it is not easy to identify structures with exclusive functions.

Frogtown structures were differentiated by Sires (1983a) into types, including categories with primarily habitation or storage function. The greatest abundances of economic pollen types on house floors did not neatly pattern along these same lines. Both structure types contained elevated amounts of the most common resources, suggesting that function in these cases may have been more a matter of emphasis than discrete divisions. On the other hand, three periodically reoccupied structures had strong records of summer weeds and high corn pollen, perhaps indicating seasonal use by agricultural helpers. At a Classic compound on the Gila, Las Fosas (Sires 1983b), one sector of the complex showed a clustering of corn and cattail pollen that seems to have persisted through remodeling, episodes. Patterned localization of resource use within the compound rooms is likely. Furthermore, pollen of agricultural weeds was heightened in room and plaza samples in one area of the site.

Another factor controlling the representation of corn pollen may be the nature of plant materials present. It has been shown by experi- mental washing (Bohrer 19U8; Gish and Scott 1983) that pollen does adhere to husked corn, and can be recovered even from kernels. On the other hand, outer parts such as husks and leaves are more directly exposed to pollen dispersal and have rough surfaces to trap pollen. Floor samples from six Mexican village structures had corn frequencies below 10 percent in an ethnographic study. Only in two sheds storing unhusked ears did lloor samples yield approximately 30 percent corn pollen (Fish 1983a). In 11 samples collected within and around five historic Papago houses, no corn pollen was recovered (Lytle-Webb 1978). 89

Field houses or more permanent structures surrounded by fields might be expected to reflect primary processing of harvested corn. Initial stages of processing or storage involving outer parts might be associated with such houses. In the SGA Project sites, some of the most abundant corn pollen in floor samples was associated with isolated structures, as at Casas Pequeffas (AZ U:15:97, Locus A) there was 12 percent, Rancho Sin Vacas (AZ U:15:62) had 23 percent, and the Ellsworth Site (AZ U:15:57) had 12 percent. As previously mentioned, contrasts between structures are not so clear in the SGA Project series, since all sites seem to have been inhabited by agriculturalists close to their fields. At the hamlet of Frogtown, surrounded by fields and intersected by canals, a few individual structures matched these field houses in corn pollen frequencies.

The most elevated occurrence of corn in an SGA Project sample (92 percent) occurred in Frogtown, Structure 53 beneath a broken vessel. Frequencies approaching this amount were reported beneath sherds on the floor (72 percent) and in the fill of a bowl (78 percent) in a Soho structure at Pueblo Grande (Gish 1979a: 166). Less high but atypical percentages in SGA Project samples were occasionally associated with vessels, ground stone, or pits, in proveniences where the greatest cultural bias might be expected. More data from experiment and analogy are needed to fully understand the implications of cultigen distributions.

The Pollen Record of Possible Cultigens

Several species became candidates for semicultigens or manipulated plants over the course of SGA Project analyses. Some have been discussed in this role before; in these cases SGA Project research provided additional evidence. In a few instances, the suggestion of human manipulation has originated with remains recovered by the project or through examination of SGA Project contextual patterns. Macrofossil evidence was stronger for manipulation of some species and pollen evidence was stronger for others. Usually, the two lines of investigation were complementary.

Chenopods and Amaranths

Miksicek (1963) has described morphological changes in SGA Project amaranths commensurate with domestication. Pollen of most species in the Chenopodiaceae or Goosefoot Family cannot be distinguished from that of the genus Amaranthus. The combined pollen type (Cheno-Am) has often been so overwhelmingly abundant in Hohokam sites as to mask the diversity and distributions of all other types. No confirmation can be found in pollen morphology for human manipulation. 90

The ecological characteristics of chenopods and amaranths also make interpretation of pollen frequencies uncertain. These plants are colonizers of disturbed ground, and would increase in the vicinity of human habitation. They are today, abundant weeds of plowed fields in the study area. No indisputable palynological feature separates cultigens, wild gathered plants, uninvited or encouraged weeds. Aggregates uncommon in airborne pollen, uneven distribution on house floors, and atypical quantities associated with pits and groundstone are indications of probable use found in almost every SGA Project site.

Native Grasses

Seeds of various Southwestern grasses were used by historic Indian groups, and archaeological specimens extend the record prehistorically. Charred macrofossils from SGA Project sites include little and panic grass. The utility of pollen is limited in documenting use of grasses. The Gramineae or Grass Family is notable for palynological homegeneity; separation of many genera is impossible under ordinary microscopy. Large size is one attribute that has been used to distinguish cultivated grassses such as corn (Zea) and the subgroup Cerealia which contains many Old World cultigens (Faegri and Iverson 1964: 196).

Grass pollen in the size range of 40 to 60 microns was encountered in SGA Project samples and has been noted elsewhere in Hohokam and Sonoran sites (Gish 1981b); Halbirt and Diggs 1981; Rankin 1981). One possibility is that these might be immature or abnormal grains of corn pollen. Various palynologists employ a lower limit in diameter size for corn from about 60 to 70 microns (Faegri and Iversen 1964: 197; Kapp 1969: 74; Martin 1963: 49). However, measurements as low as 55 microns have been reported for modern hyrid corn (Gish 1979a: 162). Complicating the picture are native grasses that could produce pollen within this range. Oryzopsis, Indian rice grass (Fish 1982a) and Andropogon (blue stem) are two such Arizona species. The former is a more northern species, but Andropogon may have occurred in the study area.

Since larger pollen size of cultigen grasses would represent a morphological change associated with domestication, the source of these grains might be a semidomesticated plant. Two candidates for domesti- cation are offered by repeated macrofossil recoveries of Hordeum and Panicum. The first, little barley, was discussed by Bohrer (1978) as an incipient domesticate on the basis of macrofossil characeristics in archaeological material from the New River area. Little barley seeds also lacking bracts have been found at Snaketown, La Ciudad (Gasser 1982: 220) and now at SGA Project sites. Although Old World Hordeum species have pollen greater than 40 microns (Faegri and Iverson 1964: 196-197), Bohrer (1978) reports the native barley (Hordeum pusillum) to have a mean diameter of 31.1 microns. However, modern reference 9 1

materials and prehistoric semidomesticates might produce pollen differing in size.

The second possibility for large grass pollen is panic grass. Panic grass was a Yuman staple (Castetter and Bell 1951) and was used widely elsehwere in the Greater Southwest (Nabhan and others 1982). Cultigen status has been argued for a distinctive race of Panicum sonorum in consideration of its morphology and associated agricultural practices among aboriginal groups (Nabhan and others 1982). Relatively large pollen is characteristic even of nondomesticated species. For example, a diameter of about 38 microns is cited for Panicum dichotomiflorum (Ikuse 1956: 42). Genetic trends in domestication might have increased pollen size.

Large grass pollen is rare in SGA Project habitation sites relative to the occurrence of barley and panic grass macrofossils. It was present in only one sample from a potential field locale, at the Ellsworth Site. Panic grass and barley seeds and large grass pollen were found in houses at Frogtown, but the pollen was not identifeid in adjacent fields. This contrast suggests an interesting possibility. Enthnographically, grass manipulation was often designed to enhance abundance in natural habitats. For example, Paiutes diverted water in ditches onto natural stands (Forde 1963: 35), while tribes along the Colorado broadcast seed on damp margins of the floodplain (Castetter and Bell 1951: 161-178). Similarly, the SGA Project Hohokam may have recognized the floodplains of the Gila, Queen Creek, or Siphon Draw as optimal locations in which to augment production of grass seed. Irrigated fields may have been used for other crops. Initial processing near harvest locales would have been likely to reduce transported bulk. Much pollen could have been dislodged in winnowing or destroyed by parching before harvests were carried to sites, resulting in infrequent pollen evidence but common charred seeds.

Agave

Cultivation of agave persists in parts of Mexico adjacent to the southwest. Planting for edible crowns and roots occurs in about half of Pima Bajo house gardens (Pennington 1980: 177), for example. Cultivation by the Pima Bajo can be traced to early historic times (Castetter and others 1938: 50). Prehistoric cultivation of Agave murpheyi between the Tonto Basin and northern Sonora has been proposed on the basis of modern distributions (Gentry 1972: 99-101). Likewise, disjunct distributions of Agave parryi have been interpreted as evidence of more northerly cultivation (Minais and Plog 1976). Although there is no clear historic record for the Hohokam area, a strong case can be made for precontact cultivation.

Agave remains have been recovered in substantial quantities from many SGA Project sites. All parts of the plant have been found 92

(Miksicek, Chapter 2, Part II, this volume). Transport of so much bulky, unprocessed material from higher elevations seems doubtful.

The pollen evidence for agave is almost nonexistent. One grain was identified at El Polvor6n (AZ U:15:59). The same amount was identified in Escalante samples (Spaulding 1974: 268). The great rarity of pollen is not at all inconsistent with cultivation in the immediate vicinity. Agaves intended for harvest would not be allowed to reach a stage of blooming and pollen production. These plants store energy throughout their lifespan, awaiting a final stage of florescence and fruiting. The optimum moment for harvest is before a flower stalk has begun to emerge, dispersing the concentrated carbohydrates. Blooming and setting of seed is not even necessary to perpetuate the crop. Propagation is accomplished through transplanted offshoots of more mature plants.

Yucca

The possibility of second planted species very similar to agave is intriguing, but less supported by historic parallels. Like agave, yucca grows successfully as an ornamental when planted today in environmental settings like those of Hohokam sites. Like agave but in much lower amounts, yucca macrofossils were repeatedly recovered from SGA Project proveniences, at some distance from higher elevation habitats. The pollen record of yucca is more abundant than that of agave. Trace amounts of this naturally infrequent type were identified at the Gopherette Site, the Jones Ruin, the Siphon Draw Site, AZ U:10:4, and Las Fosas. Even more significant, a grain was recorded in scanning a sample from a rock feature of probable agricultural function at AZ U:15:72. Pollen of yucca also was recovered from Hohokam terraces at the site of Los Morteros in the northern Tucson Basin (Fish and others 1984), again elevationally below the boundaries of its current distribution.

The somewhat more frequent recovery of yucca than agave pollen may be related to manner of use. Fruits of some yucca species were eaten, and would require prior pollination. Even the flower stalk is edible. Utility of yucca leaves and fibers for craft materials, such as cord and basketry, would not be affected by blooming. Thus by allowing completion of the reproductive cycle, the yucca plant could be made to yield edible items, important craft supplies, and (incidentally) pollen.

Cholla

Cholla has been classed as strictly a gathered resource, and gathering always must have tapped a bountiful supply. The pollen has been identified at a majority of sites where samples have been analyzed. Pollen of cholla can be separated from pad cactus or prickly types 9 3

(the two divisions of the genus Opuntia). The highest figure reported for undisturbed modern plant communities, including both Opuntia types, is 2.3 percent (Hevly and others 1965; Schoenwettter and Doerschlag 1971). Ethnographic use of buds and fruits illustrate the means of artificial pollen concentrations in sites. Pollen is contained within buds which may be prepared by grinding, and pollen might adhere to the ripening fruits. Cholla is a resource documented more consistently by pollen than by macrofossils, as soft, perishable parts are used. Frequencies for Cylindropuntia in SGA Project field, canal, and reservoir samples averaged less than one percent, with a single exception of 1.3 percent at AZ U:15:72. By comparison, site frequencies are obviously the result of cultural introduction (Table III.1.1).

Intrasite distributional patterns at Las Fosas suggest that cholla may have been transplanted within the site. The economic pollen types of corn, cattail, and cholla were found in samples from structures. Corn pollen was absent from plaza surface samples, as was cattail pollen. Cholla pollen, on the other hand, appeared even more consistently in the plaza, occurring in every sample. Part of the cholla pollen may have resulted from outdoor processing, but intentional planting seems equally probable. If plazas were largely free of vegetation, plants allowed or encouraged to grow around them would contribute an artificially elevated proportion of local pollen production. Frequencies as high as 5.5 percent occurred in plaza samples.

Southwestern ethnographic analogues for growing cactus are absent. In postcontact times of reduced population, it may have been less efficient to bother with manipulation than to rely completely on natural stands. Some hints are still available, however, that propagating may have been done. An account of Pima cholla bud gathering and preparation is presented by Greenhouse (Greenhouse and others 1981). Gathering was done from natural communities. However, one woman recalled that a Papago relative would have enough buds from cholla growing around her house to fill a backyard pit oven (Greenhouse and others 1981: 228). Cholla becomes established from disjointed stems as well as from seeds, so that transplanting would have been a simple task.

The planting of prickly pear cactus in yards is common among present-day hispanic populations of southern Arizona and Mexico. The plants may be used as fencing and also utilized for nopalitos (edible pads) or fruits. In highland Mexico, the growing of prickly pear attains regional commercial production, as in the Valley. These practices are doubtfully of historic origin, since cacti are New World taxa. Pennington (1969: 91) describes manipulation among the Tepehuan of Chihuahua. Two Opuntia species with preferred fruits are occasionally grown by both upland and lowland groups in garden plots or at any suitable spot near habitations. The Tepehuan live in an area remote from industrialized populations. Natural stands should be available to them as to the Hohokam, but convenience is evidently a motivation for improving upon the natural supply. 9 4

Planting for courtyard gathering is a possibility; use of cholla to emphasize boundaries of cultural space is another. As already mentioned, cacti and other spiny plants such as ocotillo are used to delimit traditional Mexican homesites. Fences of spiny plants protect Tepehuan (Pennington 1969: 70), Pima (Castetter and Bell 1942: 131), and Papago fields. Baskets of corn stored in pits were protected by branches of a very spiny cholla among the Papago (Castetter and Bell 1942: 184). Cholla plantings may have fitted into such concepts, as well, fulfilling a dual purpose for the Hohokam. In the Classic period, and possibly earlier (Tuthill 1947; Di Peso 1958) more substantial Hohokam "fencing" is preserved. Low walls of adobe around compounds are the usual Classic form.

Significance of the Manipulated Complex

Native species proposed as having been manipulated, by all the lines of evidence, can be classified into two groups. First are the desert herbs, mostly annuals, and some more woody perennials that grow and reproduce in rapid response to seasonal rains. Such plants avoid the problems of aridity by completing their life cycles quickly and producing seeds to await the next episode of adequate moisture. Amaranths, little barley, panic grass, and other ethnographically manipulated grasses belong in this group. Other possible members include additional preferred and encouraged field weeds, discussed separately in the following section.

The second category of plants consists of succulent desert perennials attuned to survival through uneven periods of deficient moisture. Conservation and storage of water, energy, and variable growth rates are involved in successful adaptation. Agave, yucca, cholla, and by ethnographic analogy, prickly pear, are potential manipulated plants of this type.

Together, the two categories can be seen as injecting a measure of diversity and risk management into the Hohokam subsistence strategy. Although corn varieties as closely suited as possible to desert farming were developed, corn plants required higher levels of sustained moisture than Sonoran Desert natives. With beans, teparies represented a cultigen able to produce under harsher conditions of aridity than other types (Nabhan 1979). Annuals setting seed rapidly and succulent perennials would have been means of capturing the otherwise wasted potential of limited amounts of moisture. Furthermore, these plants might have yielded in those times when more important crops failed through canal washouts and seasonal or prolonged drought. During such periods, the harvest of manipulated desert plants might have been a welcome addition to stored foods and gathered resources. 95

Agricultural Weeds

The pollen configuration of irrigated and nonirrigated SGA Project fields suggests the presence of many herbaceous plants in addition to the major cultigens, some perhaps being perennial (Fish, Chapter 5, Part II, this volume). Summer moisture, and by inference summer irrigation, was emphasized by the pollen types. The three dominant field taxa were spiderling, Arizona poppy, and globemallow. Evening Primrose Family pollen also appeared in fields, but more consistently in canal samples. These pollen types do not exhaust the list of agricultural weeds, but allow insight into the setting for secondary subsistence resources.

The four pollen groups with SGA Project agricultural associations have previously been recognized as unusually well represented in Hohokam sites. For example, Webb (1981: 343-344) noted that pollen of Onagraceae, Nyctaginaceae, and Malvaceae occurred in 75.7, 86.5, and 97.3 percent respectively of Las Colinas samples. The latter two plant families are those containing spiderling (Boerhaavia) and globemallow (Sphaeralcea), with most Las Colinas pollen campatible with these two genera (Webb 1981: 350-351). At Pueblo Grande, Gish (1979a: 164-165, 168, 171) commented upon abundances of these three types and Arizona poppy (Kallstroemia) as well.

Taxa producing the four pollen types are insect-pollinated and produce relatively little pollen compared with plants pollinated by wind. Furthermore, these types are infrequent in samples from modern vegetation zones in which source plants are growing. In view of natural occurrence, the consistent appearance and sporadic abundance of these pollen types in sites has been discussed as evidence of probable cultural bias. These interpretations have referred to ethnographic use of plants.

SGA Project samples from irrigated and nonirrigated fields and canals indicate that some of these four types of pollen in habitation proveniences results from the proximity of fields. Spiderling, globemallow, and Arizona poppy constituted 35 to 45 percent of all pollen in irrigated field sediments at Frogtown. Agricultural features not associated with canal irrigation at a number of sites yielded lower, but consistent frequencies. Onagraceae pollen appeared in field samples, but was most regularly present in those from canals. Although all four pollen types can be characterized as large and poorly suited for wind dispersal, habitations in the midst of fields are likely to have received increments of pollen in this manner. Another means of introduction would be pollen adhering to crop plants brought into structures. 96

Weeds as Resources

Field and canal samples reveal one source of spiderling, globemallow, Arizona poppy, and Onagraceae pollen in habitation contexts. The opportunity to compare agricultural and habitation spectra also provides a means of examining frequency bias associated with economic plant use. Table 111.1.1 presents the percentage of field, canal, and habitation samples at Frogtown containing each agricultural pollen type, as well as the means and highest percentage in agricultural and habitation contexts. Of the four types in Table 111.1.1, only Arizona poppy is without analogue as an edible resource. Only medicinal uses for this plant are known among the Pima Bajo (Pennington 1980: 272) and Tewa Indians (Robbins and others 1916: 56-57). It is also the only type appearing in an appreciably lower percent of habitation samples. The highest frequency of Arizona poppy in agricultural samples is more than three times higher than in habitation samples.

By contrast, the three agricultural types used ethnographically for food in the Southwest occur as often or nearly as often in residential areas as in fields. The difference between the highest agricultural value for Onagraceae, 7 percent, and the habitation counterpart of 23 percent from the interior of a vessel is a convincing demonstration of cultural bias. Likewise, 51.5 percent globemallow in a pit shows the effect of cultural concentration when compared to the highest field frequency of 26.5 percent.

Spiderling pollen appears in nearly every habitation sample, although frequencies reach their height in agricultural proveniences. In this case, the stage of maturity at which the plant was used may account for introduction of limited pollen. A Southwestern analogue for spiderling use is as a green or pot herb. Among the Seri Indians, Boerhaavia coulteri was one of the only three plants recorded as being eaten in this form (Felger and Moser 1961). Preferred plants would probably be young and tender, prior to pollen formation and flowering. Seed use has been cited by Bohrer (1981) for African species of spiderling, but no such use has been recorded in the Southwest. The abundant availability of seeds in fields may have prompted their use by the Hohokam, however, particularly in difficult times.

Consumption of globemallow is restricted in the ethnobotanical literature, but seeds are considered edible by the Navajo and have been found in Hohokam sites. The Hopi chewed the stems (Castetter 1935: 31) and recorded Piman use is medicinal (Russell 1975: 79) with similar application among other groups such as the Hopi (Whiting 1939) and Apache (Gallagher 1977: 97). The Tewa also use the roots to make a face paint (Robbins and others 1916: 61).

The roots of Oenothera bieenis and probably the related Arizona species O. hookeri in the Evening Primrose Family are edible (Kearney and Peebles 1964: 593). Of considerable importance to the Sand Papago 97

were the greens of O. trichocalyx (Castetter and Bell 1942: 62). Greens of another species, Camissonia claviformis, also were eaten by the Sand Papago (Felger 1980). Still another species, Epilobium angustifolium, furnishes leaves and stems for a potherb (Sweet 1976: 44; Kearney and Peebles 1964: 591). The Apache ate the fruits of one genus (Castetter 1935: 17). Use of greens or fruit is more likely to have introduced pollen than use of roots, unless the entire plant was uprooted in collecting. The only one of these edible species found in the study area, O. hookeri, grows in damp habitats (Kearney and Peebles 1964: 595), fitting a greater ubiquity in canal than in field samples.

Microhabitats

A strong linkage of the four pollen types with Hohokam agriculture is not necessarily obvious; other plants are more common as field weeds today (Fish, Chapter 1, Part I, this volume). At Casa de Loma, Gish (1979b) suggested that spiderling might have been gathered along canals on the basis of high frequencies in one canal sample and in two samples from ground stone artifacts with canal associations. However, the strong linkage of these types with agricultural disturbance becomes apparent only through a comparative base in agricultural as well as habitation proveniences. SGA Project sites were optimal for examining these palynological relationships. Analyses involving numerous samples often have focused on large, complex sites, whereas a settlement such as Frogtown was small, had agricultural production as a primary function, and was located in the midst of fields.

The four most prominent agricultural weeds in SGA Project samples are not the only ones suggested by pollen (macrofossils also add many to the list), but their ubiquity provides insight into the degree of cultural modification of vegetation surrounding Hohokam sites and its human resources. Furthermore, the existence of microhabitats for gathered plants within this modified vegetation has been suggested. Fields and perhaps field borders supported bountiful supplies of spiderling and globemallow. By contrast, a comparison of frequencies show chenopods and amaranths sometimes to have been more available as plants around houses, perhaps including the weeds of kitchen gardens. The moist precincts of canals added a further set of edaphic conditions, offering the most abundant supplies of plants such as evening primrose.

SGA Project analyses have shown that prehistoric disturbance floras and microhabitats, with attendant potential for gathering, cannot be predicted well from modern parallels. Weedy growth in modern irrigated fields in the study area is dominated by Amaranthus palmeri and various Old World introductions. Canals are now characterized by such plants as thornapple (Datura), knotweed (Polygonum), dock (Rumex), and Johnson grass (). Historic vegetation change and differences in technology affecting plant growth are probable causes of these differences. Remains from archaeological contexts are the only means of exploring the resource diversity of the Hohokam cultural environment. 9 8

Management of the Modified Environment

In industrialized societies with mechanized agriculture and market economies, a sharp dichotomy tends to be drawn between the desired products of cultivation, the cultigens, and undersirable competitors, the weeds. Applying this distinction literally in interpreting Hohokam subsistence may obscure an important suite of resources and deflect attention from the skill of Hohokam farmers in manipulating their arid environment. Constellations of plants growing in the fields with formal cultigens may have been shaped to some degree by conscious effort rather than simply gathered when fortuitously present.

It is difficult to judge relative importance of such resources and very difficult to reconstruct any practices associated with them. Manipulation of disturbance plants is not likely to have been noted by early Spanish travelers, and categorization of plants by European observers would seldom elicit information pertaining to them. The role of weedy plants has become a topic of increasing interest among ethnobotanists, however, and ethnographic literature contains many instances of intentional manipulation, ranging from selective removal to planting of seeds and transplanting. It seems improbable that the Hohokam, lacking many alternatives of historic groups, were less adept in this regard.

As discussed in the section on native grasses, panic grass was obtained from natural stands and, consequently, broadcast seeds. The tending of available plants without manipulation is clearly illustrated by this case. Other examples show optimizing management of disturbance plants similar to that likely by the SGA Hohokam. An edible Old World mustard, Brassica, is planted by the Tarahumara around homesites in selected microhabitats, such as the enriched environs of domestic animals. Another kind of weed management is of a more permissive nature. The Hopi allow beeweed (Cleome serrulata) to mature and seed in cornfields, insuring a source of gatherable plants for food and pottery paint in the following year (Whiting 1939). Zapotec Indians of Oaxaca have been recorded as practicing mixed strategies with gatherable plants, including planting from seed and transplanting near habitations and selective weeding in fields (Messer 1978: 82-88).

Traditional farmers of Cucurpe, Sonora, feel that the effort of making a temporal field is justified by the weedy yield, even though cultigens may not recieve sufficient moisture to mature (Sheridan 1978). The modern value of these plants is for animal forage, but insurance against crop failure is also illustrated, if humans could eat the yield. Native species in the disturbance flora were adapted to completing growth cycles under arid conditions.

The diversity and abundance of vegetation in modern Pima fencerows, fields, and canals (Rea 1979; Nabhan 1979) has probable parallels in Hohokam agriculture. Crosswhite (1981: 64) has emphasized 9 9

the element of intentional management by Pima cultivators. He has called the growth of desired weedy plants in excess water from irrigated fields a "second garden." Desirable plants were collected, undesirable competitors were removed as time permitted, and mature plants were left to insure seeding.

Ethnographic management of disturbance vegetation in cultural environments can suggest ways in which the Hohokam increased and diversified resources and insured gathered foods in poor years. Candidates for manipulated species can be derived from sampling habitation and agricultural proveniences. The degree and specific practices of management remain questions for future research. Resources recognized in SGA Project sites are, undoubtedly, a small part of a large repertoire of manipulated plants, varying with time, natural environment, and cultural preference.

If the demonstrated relationship of SGA Project agricultural weeds to utilized resources has raised issues of management without final solution, certain aspects of Hohokam agriculture have, nonetheless, been highlighted. In the development of irrigated, floodwater, and runoff fields, the Hohokam created more than the means for producing formal cultigens. They provided habitats in which herbaceous species from a variety of natural associations could be concentrated within easy reach. Plants of mesic locales within the immediate baseline vegetation and those more common at higher elevations could prosper. To the native regional species could have been added agricultural weeds entering southern Arizona along with cultigens of Mesoamerican origin. Through selective removal of encouragement, and by tolerance of both annuals and perennials, prehistoric fields may well have been the richest, most varied, and most predictable source of "wild" edible foods. Plant remains from SGA Project sites imply exploitation of just such diversity.

Gathered Plants

Recognition of the extent of noncultigen resources in the modified vegetation of Hohokam sites makes it difficult to define boundaries for the category of gathered resources. Useful plants in disturbance communities probably were gathered in natural vegetation as well. In fact, response to irrigation water may have altered times of growth and fruiting compared to wild stands, seasonally extending prospects for gathering some species.

A number of herbaceous plants appear in the spectra of fields, but in low quantities which cannot be distinguished with certainty from natural background levels. These include ( Family, including nightshade, groundcherry, and ), Eriogonum (wild buckwheat), Cruciferae (Mustard Family including tansy mustard), Euphorbia (spurge), Plantago (Indian ), and Gramineae (grasses). 100

All these groups include edible species and all occur as charred macrofossil remains. Most also occur in sporadic high frequencies and aggregates in habitation pollen samples such as pits or vessel contents, suggesting economic use. These plants are gathered ethnographically from wild communities, but fields and disturbed areas near settlements may have supplied convenient and abundant harvests as well.

Agave, yucca, and cholla have been discussed as potentially manipulated plants. Again, this possiblity does not preclude gathering of wild plants. Both agave and yucca would have entailed travel to upland zones for procuring untended plants. These are among only a few resources not immediately available to SGA Project sites, and the macrofossil evidence for cultivation of agave is convincing by ubiquity and range of materials (Miksicek, Chapter 2, Part II, this volume).

Pollen of prickly pear and Cereus-type cacti was found repeatedly, with prickly pear occasionally in aggregates. Cereus-type pollen is produced by saguaro, hedgehog, and other related cacti. This type occurred in trace amounts, and could indicate proximity in the natural vegetation as well as resource use.

The leguminous trees furnish many useful products, but are more fully documented by flotation results than by pollen. Mesquite, palo verde, and ironwood occurred, often signalling nearby trees for exploitation. Only mesquite (and acacia in one instance) was found in apparent economic concentrations.

Riparian vegetation with specialized resources was indicated more by inference than by recovery of economic pollen. Occasional willow (Salix) pollen suggested a more vigorous riverine community on the Gila and probably stretches of the secondary drainages as well. Hackberry (Celtis) pollen may have come from this source or from shrubby forms in more mesic desert locales. Cattail (Typha) was the only certain riparian resource shown by pollen, but is a clear indication of permanently damp habitats supporting other localized species and probably abundant fauna. Cattail occurred in Santa Cruz through Classic proveniences over the span of SGA Project occupations. In some instances, cattail pollen may have resulted from gathering of edible roots, shoots, pollen, or seeds. At least at El Polvol6n, association with roof fall suggested use in construction of pit house superstructures.

It has been suggested that Hohokam canals may have acted to extend the riparian habitat away from water courses, increasing the availability of more mesic resources including leguminous trees (Plog 1978; Stein 1979). Such an effect can be seen in the dense strips of trees paralleling the modern unlined canal in the study area. Water seepage has similarly encouraged lush "riparian" strips along traditional Piman canals on the Gila Indian Reservation (Rea 1983). No evidence of response by perennial vegetation was indicated in SGA Project pollen samples from canals, although a specialized herbaceous flora was apparent. However, all canal samples analyzed were from 101

canals near Frogtown, fed by Queen Creek. While canals supplied by this secondary drainage apparently did not contain water sufficiently long to encourage woody species, vegetational effects remain to be examined for SGA canals drawing water from the Gila River.

Intersite Comparisons

Intersite comparison of pollen results is poorly developed. The great majority of reports in the literature interpret pollen spectra from the perspective of single sites. The methodology of comparison includes a few models. Furthermore, analyses published to date involving as many sites and as many samples as the SGA Project study are almost nonexistent. Problems exist in interpreting variation in economic types from site to site, since many of these types are rare and the degree to which a resource has been processPd may affect pollen abundance. Sample size also presents problems--an initially large number of samples divided among many sites does not insure adequate numbers from appropriate kinds of proveniences at each site. Variation due to geographic position and local environment is difficult to separate from variation introduced by cultural activity.

Pollen Records of Irrigated and Nonirrigated Agricuture

Inherent levels of functional contrast are limited among SGA Project sites. In part this may be due to Hohokam cultural patterns, but also to the nature of the sample in the study area. The more elaborate public or ceremonial architecture known for the Hohokam does not occur. All habitation sites are modest in size and probably oriented toward agricultural production. Given a degree of overlap in the nature of sites, however, intersite comparisons of pollen distributions reflects contrasts in subsistence organization and activity.

Tables 111.1.2 and 111.1.3 summarize descriptive statistics for selected pollen types in SGA Project agricultural and habitation sites. Types identified with agricultural precincts are included for the insight they provide into kind and location of farming compared to habitation loci. Cheno-Am pollen is included as a disturbance-related category with distributions at least partially independent of agricultural activity. Quantitative measures are provided for corn and cholla as the two most consistently present and unequivocal subsistence products registered by pollen. Prickly pear provides a useful contrast with cholla, although frequencies were so low that means were not calculated. Other cultigens and cattail are noted only for presence, since low frequency and sporadic occurrence reduce the utility of quantification. In habitation sites, only floor samples were considered in calculating site percentages. Although floors are by no means free 102

Table 111.1.2

SUMMARY OF POLLEN DISTRIBUTIONS IN SALT-GILA AQUEDUCT PRO.IECT AGRICULTURAL LOCI

>3

. .... W E M = u W c Q -0 m M 3.. m O C o , 'N00m a, -, m >,. L. m o a a adME E T 3 a N 3 a .. m 1., VI u w 0 0 0 L. 00 C .2 . .,-, - . O. ; .10 , 1. e -. 0 --, mo S M W u m M M o0 o 7 W L ., O C e!ECC a C00.n-, .0 c o GI 0 n-n 0 0 4 -cia I' N CON C C C C 3 0 - m 0 .... . o L. m w ,z, u T1 I-. 0.- L. 0 .. 0 .21 0 0 0 Agricultural loci W ..(S' < X 0 < 41 0 0 )U.... u 3

Irrigated Frogtown Fields 100.0* 100.0 91.0 38.7 36.4 100.0 9.1 0.1 1-1 .11 24.8 5.1 0.7 7.0

Frogtown Canals 100.0 100.0 75.0 34.2 62.5 100.0 6.3 0.1 - N.16 19.9 12.2 2.1 1.6 14.1

Runoff

Jones Ruin Fields 93.8 33.3 8.7 6.7 100.0 - - 9 .15 8.4 0.3 0.1 7.9

AZ 11:15:72 Fields 75.0 25.0 6.2 100.0 9.4 5.8 0.4 19.9

AZ 11:15:73 Fields 100.0 12.0 100.0 9 .11 12.0 17.4

AZ U1546 Fields 56.2 37.5 12.5 1.9 100.0 6.3 0.1 - 9 .16 1.5 0.3 0.1 4.5

' Count Percent - Noted in scanning

of culturally skewed representations, it was felt that they better reflected site-wide pollen deposition and cumulative resource use than did pit and vessel fills and artifact washes.

One contrast immediately apparent in Table 111.1.2 is the greater abundance of agricultural indicator types in irrigated fields and canals, than in nonirrigated loci. Summary statistics from nonirrigated features exaggerate these differences. It was difficult to precisely locate prehistoric surfaces associated with the use of runoff farming features and difficult to know where to sample rockpiles. Thus, low mean values for agricultural indicators incorporate samples pertaining to deposition before or after feature use of some temporal admixture. For example, sums for spiderling, globemallow, and Arizona poppy in individual samples from nonirrigated contexts were as high as 103

Table 111.1.3

SLMIARY OF POLLEN DISTRIBUTIONS 1N SALI-5140 AQUEDUCT PROJECT HABITATION SITES

., • n ...•

Site and Sire Type Sample Drainage Size

Hamiets

Frogtown 1 23 Queen Creek 30 96.1 90.0 60.0 lU 63 100.0 73.3 11.6 90.0 13.3 X 5.5 1.8 2.1 15.1 7.9

Las Posas Gila River 17 100.0 35.3 23.5 17.1 17.6 100.0 29.4 1.8 82.4 11.8 X 16.3 0.6 0.2 0.3 31.0 5.6

The Junkyard Stte 2 Gila River 100.0 100.0 100.0 25.2 50.0 100.0 100.0 2 75.0 50.0 17.9 5.3 2.0 0.4 20.1 5.4

El PoLvor en I 7 Queen Creek .20 100 0 90.0 60.0 12.6 20.0 100.0 15.0 0.3 15.0 15.0 X 2.5 0.6 0.1 22.5 0.1

Farmsteads

Rancho Sin Var/os 1 Queen Creek 8 e2 100.0 100.0 100.0 6.6 100.0 100.0 100.0 14.8 50.0 0.75 4.8 1.0 0.73 19.3 2.5

Seller's Well 3 Queen Creek -7 100.0 100.0 85.7 5.5 - 100.0 43.0 1,1 14.2 57.0 1.3 2.6 1.6 19.1 0.07

The Dust Bow: :e I 1 Gila River 4 100.0 100.0 25.0 10.6 50.0 100.0 50.0 1 75.0 25.0 9.1 1.4 0.1 2.6 41.6 2.9

Saguaro Site 1 1 Silo River UrS 100.0 100.0 40.0 19.2 0.0 100.0 40.0 5.1 80 .0 16.3 2.7 0.2 0.0 19.2 3.7

Jones Ruin 0 5 Gila River 7.,i5 86.7 20.0 6.7 4.87 - t00.0 40.0 40.0 33.3 X X 4.6 0.2 .07 29.2 1.6

The Gopherette SLce I 5 CiLa River 9.70 100.0 100.0 70.0 24.5 20.0 100.0 60.0 60.0 0 X 15.7 7.5 1.3 0.1 29.3

FietdhoLses

Casa. Pequea1n1 I 5 Queen Creek .6 100.0 100.0 50.0 31.4 - 100.0 66.7 2,.2 17.8 12.7 0.9 32.8

The Cntadag 5, 7 e I 1 Tree, Creek 1 100,0 100.0 - 25.5 100.0 100.0 22.0 3.5 - 29.0

The Ellsworth Site I 1 Queer, Creek 0-.3 100.0 100.0 66.7 23.0 - 100.0 100.0 15.8 13.8 8.7 0.5 2.3

AZ a.7 ,3.22 R 1 Gila River 3 66.7 100.0 - 7.7 - 100.0 66.7 5.5 2.2 8.5 1.7

AZ aq 5:20 R 1 Gala River 1 100.0 100.0 8.5 - 100.0 5.0 3.5 - 6.5

. Count R. Percent - weted in scanning ' tn question

NOTE: 1 s irrigation canal: F fields; R . runoff 104

10 percent at AZ U:15:46, 14.5 percent at the Jones Ruin, 18 percent at Smiley's Well, and 23.5 percent at AZ 0:15:72. Agricultural samples at the Ellsworth Site are not included, since the relationship of proveniences to irrigated or nonirrigated fields is unclear. In spite of mixture problems, agricultural indicator values for irrigation are clearly higher.

Irrigated fields and nonirrigated farming locales share the palynologically prominent weeds. Higher frequencies under irrigation probably reflect densities responding to more abundant water. Paralleling ethnographic spacing of crop plants in concert with water supply, secondary harvests of weedy species were probably greater in irrigated fields. In addition, spiderling appears to have constituted a greater proportion of combined weeds in most nonirrigated locales than in irrigated ones.

Agricultural Implications of Habitation Samples

Differential pollen production of irrigated and nonirrigated agriculture are reflected in associated residential areas as well as in field sediments (Table 111.1.3). Habitation sites adjacent to rockpile fields, watered by runoff or diversion of small channels, show low combined values for spiderling, globemallow, and Arizona poppy. These are the Jones Ruin, the Tonka Site, and the Oidag Site. Rancho Sin Vacas is another site lacking irrigated fields. Possible floodwater fields supplied by a minor drainage would have been nearby, but not adjacent. Again, agricultural indicators have low values.

The single remaining site with a mean frequency for agricultural types below 10 percent is Smiley's Well. Unlike the sites near runoff agriculture, however, all three types appear in virtually every sample. This ubiquity stems from the tact that Smiley's Well is a farmstead located in the vicinity of irrigated fields, although removed some distance from the canal. Demonstrating the palynological effects of both proximity and functional types is the Ellsworth Site. The two sites are contemporary or nearly so, but the Ellsworth Site is adjacent to the canal (and presumably irrigated fields). In addition, the structure at the Ellsworth Site appears to be a field house in contrast to a more substantial farmstead occupation at Smiley's Well. The field house, located in or near fields, and presumably most directly associated with agricultural activities, has a mean value of 23 percent for agricultural types; the more removed farmhouse has 5.5 percent.

Structures at two other sites in the midst of irrigated fields have the architectural and artifactual characteristics of field houses. The Chiadag Site near Frogtown, and Casas Pequeffas have among the highest values for agricultural types. Frogtown is a larger hamlet with more varied activities, and is intersected by canals. Average values are somewhat lower than in the similarly located field houses, but still reflect the immediate proximity of irrigated fields. A later hamlet 105

occupation at nearby El Polvorbn may have been more separated from its fields, with somewhat lower means for agricultural types, but similar ubiquity.

Sites already discussed as being near irrigated fields have all been on Queen Creek. Canals in this area could be investigated directly, and some smaller branches located. Along the Gila, knowledge of canals is supplied largely by maps of earlier workers. Historic canals and agriculture have removed opportunities to define details of prehistoric networks and branches. The relationshp of contemporary canals and fields to excavated sites is more suppositional than on Queen Creek. Recorded canals are downslope from SGA Project sites, between the sites and the Gila floodplain. Topography suggests that fields would have been elevationally below canals on the side toward the river, not immediately adjacent to the habitation sites (the Dust Bowl Site [AZ U:15:761; the Saguaro Site [AZ U:15:771; the Junkyard Site [AZ U:15:831; the Gopherette Site; and Las Fosas.

The first two sites are farmsteads, approximately 0.5 km apart and as far from the nearest canals. Values of combined agricultural types for both sites are distinctive from habitations associated with runoff features, but the Saguaro Site has almost twice the represen- tation. Whether this difference might reflect different positions of fields during the two periods of occupation or differential introduction of plants into the sites is not known.

The remaining three Gila sites are aligned along the river, with approximately 2 km separating the farthest. Although the recorded canal passes the sites at a distance of about 1 km, the abundance and ubiquity of agricultural pollen types is as great as sites with immediately adjacent fields. The repetition of these findings at all three sites suggests that an undetected secondary branch watered fields in the nearby vicinity.

Comparison of Economic Types

There are a number of difficulties in comparing and interpreting economic pollen types among sites. Economic types are often rare and may appear very sporadically within any set of samples as a sampling phenomenon. When many samples are available from a single site, this periodicity can be evaluated. For sites such as isolated field houses, however, one or a few samples from a single structure may have been analyzed. Thus the percentages calculated for ubiclity in Table 111.1.3 are only meaningful for sites with larger numbers of samples or when considered in the aggregate for a type of site.

It also cannot be assumed that abundance of pollen is relatPd in a simple manner to abundance of the resource. Different parts of a plant may introduce differential amounts of pollen. For example, gathering of edible cattail pollen should have an effect on the pollen 106

content of a provenience distinctive from that associated with use of roots or shoots. Another factor to be considered is degree of processing. Corn pollen might be expected in decreasing order of abundance in a structure if the introduced form were unhusked ears, husked ears, kernels, or ground meal. In an ethnographic pollen study of house floors in Oaxaca, Mexico, for instance, storage sheds containing unhusked ears had several higher orders of magnitude for corn pollen than did habitation structures in which more processed forms were present (Fish 19b3b). The distribution of corn pollen among SGA Project sites illustrates some of the problems of intersite comparison, while offering insights into patterns of resource use.

The mean value for corn pollen at Frogtown is appreciably heightened by a few samples with exaggerated amounts (that is, a sample with 590 grains). At the same time, ubiquity at Frogtown is impressive- -corn occurred in 73.3 percent of all samples. Relatively high means as well as ubiquity are matched by two field house sites and an isolated farmstead, where agricultural products in initial stages of processing might be expected. The amount of corn in several houses at Frogtown is similar to pit houses at these smaller sites, suggesting that some structures in this hamlet surrounded by fields could duplicate functional aspects of field houses. Two houses in the Santa Cruz occupation of Frogtown appear to have been reoccupied periodically, and both exhibited high frequencles af corn and summer weed pollen. These two structures may have been most analogous to field houses, seasonally occupied by farmers or extra farm laborers, paralleling historic Papago workers in Pima fields.

Contrasts in subsistence activities or in organization of resource use can be seen in comparison between larger sites with more numerous samples. For the Junkyard Site, comparisons are uncertain; only four samples from two structures furnish data. Differences between Frogtown and the two Classic hamlets are less equivocal. The three contain roughly similar levels of agricultural pollen types, but show very different patterns for corn pollen. At Frogtown, 73.3 percent of floor samples contain corn pollen, compared to 29.4 percent at Las Fosas and 15 percent at tl Polvor6n.

These distributional differences are amenable to alternative interpretations. Corn could have been a more common crop at Frogtown. Macrofossil recovery does not support this inference; higher percentages of flotation samples at the two Classic hamlets yielded corn remains than at Frogtown. Two other explanations therefore seem more plausible. Corn could have been husked or shelled outdoors at fieldside or under insubstantial ramadas by inhabitants of the Classic sites. More processed forms brought into habitation structures might then have contributed to the pollen record less frequently. Even trace amounts of pollen enter into the ubiquity percentage, however, and a third alternative seems most plausible.

Classic organization of functional space may have been more elaborated than in the earlier hamlet. At Las Fosas and El Polvor6n, 107

the appearance of corn in a lower percentage of samples could reflect a more restricted localization of storage and/or initial processing. The distribution of corn pollen within the compound at Las Fosas accords with well-defined spatial categories, since this type was concentrated in several adjacent proveniences. Localization of resoure materials reflected in pollen patterns could correlate with Classic concepts of architectural differentiation or with more centrally located communal stores.

Four samples from two field houses near runoff fields produced corn pollen only in the scanning of additional material following tabulations. Samples from features watered by runoff indicate that corn was grown by this technique as well as by irrigation, although yield per area and ratio of corn to other crops may have been lower. Three field houses with irrigated fields have corn pollen in 8 of 10 samples, and two of the sites have high mean frequencies. Differences underscored by combining the lower numbers of samples from each site do not apply equally in comparison of individual structures, however. One grain of corn pollen in the single sample from the Ellsworth Site, amid irrigated fields, is not significantly different from the record of the two field houses lacking irrigation. Multiple samples from numbers of structures in each category in future Hohokam investigations may refine the interpretability of comparisons such as these.

Comparison of cholla, the most common noncultigen economic pollen type in SGA Project sites, again presents reliable contrasts and more questionable patterns based on small sample size. Cumulatively, field houses have a much lower record of cholla pollen than do sites with evidence of substantial occupations. This would be the opposite of expected findings, if field house occupants were envisioned as apart from large settlements and conveniently located near less intensively exploited natural stands. A farmstead site which should have had access to optimal natural sources, the Jones Ruin, also fails to match or' surpass a number of other farmsteads and hamlets in ubiquity and mean frequencies. The Jones Ruin is located on a gravel-covered ridge overlooking the Gila floodplain, surrounded today and probably prehistorically by peak cholla densities for the study area.

Occupation of field houses during inappropriate seasons for gathering does not seem to account for absence of cholla pollen. Buds, the best source of pollen and an ethnographic staple, were observed to span late spring through summer in the study area. The relative concentration of cholla pollen at larger and more permanent sites, in contrast to the field houses and a farmstead site with optimal access, suggests that cultural patterns rather than natural distributions of plants are involved. Indeed, natural cholla availability may predict inverse evidence in sites. At seasonal sites and ones adjacent to dense stands, the impetus to enhance immediate supplies would have been absent.

As discussed in an earlier section, cholla pollen was consistently present in Las Fosas plaza samples although other economic 108

types found in structures were absent. Intentional planting of cholla about the site to provide convenient sources of buds and/or to delineate intrasite boundaries was suggested and may apply to other sites with similar records of cholla pollen. If inhabitants of the field houses and the Jones Ruin made equal use of cholla, the absence of artificially concentrated plantings and on-site preparation may have introduced proportionally less pollen.

Three of the four hamlets have abundant cholla pollen by frequency and ubiquity measures. The fourth, El Polvor6n, does not. This difference appears to make a much reduced emphasis on the resource, again almost certainly a cultural decision rather than an environmental restriction. At nearby Frogtown, Santa Cruz through Sacaton samples consistently contain cholla. It is doubtful that this widespread desert plant would have been displaced during the Classic phase despite climatic oscillations, but it may have been no longer tended.

The distribution of prickly pear pollen in Table 111.1.3 shows the occurrence record for another relatively common noncultigen type. Frequencies for prickly pear were typically so low that average values per site have not been presented. The lesser appearance of prickly pear compared to cholla may be influenced by lower pollen production or by selective emphasis on forms less likely to introduce pollen, such as fruits rather than buds. Nevertheless, reliance on prickly pear in SGA Project sites was probably less than on cholla. For the three hamlets with many samples, ubiquity percentages are similar at 11.8, 13.3, and 15 percent of all samples. Fairly stable use of this cactus in time (Santa Cruz through Classic) and space (Queen Creek and Gila) in the study area seems indicated. The third hamlet, with prickly pear pollen in two of four samples, cannot be reliably judged as having a similar or dissimilar record of use.

The lack of correlation between natural availability of cholla and relative abundance of cholla pollen at the Jones Ruin is not repeated with prickly pear. The gravelly ridge location of this site supports concentrated prickly pear as well as cholla; spectra from this most thoroughly sampled of farmstead sites have the most abundant record of prickly pear pollen. Ten samples from the Gopherette Site, a farmstead at the edge of the Gila floodplain, in vegetation with widely scattered prickly pear, contained none of this pollen type. Comparative abundance at other sites with fewer samples is open to question.

The columns in Table 111.1.3 for cucurbits, cotton, and cattail register only presence. Most occurrences are a single grain. These economic types illustrate the unmeasured role of sampling accident in shaping observed distributions, and the attendant uncertainties of intersite comparison. Presence is significant, but absence may be misleading, and particularly so for sites with few samples. Presence alone can provide useful information, as with the fact that cotton was grown in both irrigated and runoff agricultural locales. 109

The distribution of cattail appears to be patterned, but sporadic occurrence in low amounts complicates interpretation. Cattails occur in all four hamlets, on both the Gila and Queen Creek. A single scanning identification was made elsewhere. It may be that cattails were used in the construction of more substantial or functionally differentiated architecture or that some other form of resource use was more prominent in the larger sites. The cumulative total of samples from all nonhamlet sites (57 versus 71 from hamlets) should be sufficient to dismiss sampling as the source of the overall pattern, yet the absence of cattail in individual small sites can still be challenged on the basis of sample size.

Table 111.1.4 is a compilation of presence data for economic types from all nonagricultural sites. For habitation sites, it will be seen that the number of resources revealed by pollen correlates with large sites and many samples. Undoubtedly longer occupations and the

Table 111.1.4

OCCURRENCE OF ECONOMIC POLLEN TYPES AT NONAGRICULTURAL SITES

.. T.: 1 2:. — ..!: t'o 1 i l r 2 ,t,.. ii i' `2 .

1-; Z . o, :2 2 2 .T. 2 l' 5 t4 : ".' " 1 2 t.,' g, : :; go % 4° Site and Site Type ci 3 5 1 .1. 5 :, g 5 4 2 <., 5 5 ,:'. .2 >. t li' 2 2; 2 J',

Hamlets The Siphon Dray Site X X XXXXXXXXX X X X X Frogcoun X X XXXX X X X X X X Las Fosas XXXXX X X X XXXX X X El Polvor6n X X X X X X X X X X X

Farmsteads Rancho Sin Saros X X X Smiley's Well X X X The Dust Bowl Site X X X Saguaro Site X X X X Jones Ruin XXXX XXXXX The Copherette Site X X X X X X X X X X

Fieldhouses Casas PequeAns X X X The Chiadag Site X X The Ellsworth Site X The Tonka Site X X X X The °Wag Site X

Small Non—habitation Sites AZ Uil0:14 X X X X U:1594 X Oil5,102 X 110 cumulative activites of more people enter into this pattern as well. For the small nonhabitation sites, a mixture of resources was shown by both pollen analysis and flotation. Many of these small sites were marked by facilities such as roasting pits that required initial labor investment, but could be reused many times. At no great distance from habitation sites, these small processing stations may have been employed for a variety of seasonally gathered plants.

Conclusions

The environment of the SGA Project study area is naturally redundant. The range in topographic variety and in natural vegetation as measured by modern distributions is limited. Although concentrations of human resources vary from segment to segment, only riparian species appear to have been truly localized. The pollen record of SGA Project sites suggests that this natural homogeneity was accentuated by accident and/or intention by the Hohokam. Disturbance of residential areas and the disturbance and supplemental water of agriculture created vegetation around sites sharing many of the same elements.

By its very nature, agriculture homogenizes the environment from the perspective of human resources. Selected cultigen species are modified and moved from zones of natural occurrence to convenient sources in fields. In this sense, the SGA Hohokam were able to successfully accomplish this desired agricultural repetition in a number of settings and with multiple technologies. Overlap is shown by pollen between crops which could be harvested in irrigated and in runoff fields. Indirect pollen evidence from Gila hab-tation sites, where field samples are lacking, shows similarity between these sites and those along Queen Creek in ability to produce quantities of corn. A dichotomy between riverine and nonriverine agricultural potential is not indicated for specific locales, although the magnitude of riverine canal systems may have been greater.

Repetition in disturbance floras from site to site points out a further effect of occupations in creating a predictable sort of modified vegetation. A variety of evidence suggests that the configuration of plants accompanying Hohokam sites was not totally fortuitous. At the least, site inhabitants benefited from the culturally altered vegetation. Relevant evidence includes the comparative distribution of pollen in different proveniences, economic concentrations of pollen of disturbance taxa, and macrofossils recovered from sites. Ethnography provides records of such intentional manipulation involving a much greater range of plants than formal cultigens. These implications create an appreciation for the skills and success of the SGA Hohokam in maintaining themselves in their difficult environment. The combination of disturbance plants and those which seem to have been tended would have extended available resources and increased flexibility in dealing with periodic droughts or agricultural disasters. 111

None of these findings negate the possibilities of speciali- zation and exchange in subsistence resources. Rather, the potential for duplication of resources among sites implies motivations for exchlnge in subsistence items beyond ordinary consumption needs. Likewise, differences in resource use as with cholla, between the nearby sites (Frogtown and El Polvor6n), seem better interpreted as cultural emphasis than as consequences of natural availability. The SGA Hohokam were able to produce or acquire subsistence staples and most of the documented plant resources from the low desert in which they lived. Specialization and exchange may have been culturally legislated elements adding to the success of a long-lived subsistence system.

Pollen in SGA Project sites incorporates components from regional vegetation, disturbance communities, and human activities involving plants. To obtain reliable information on any of these, it is necessary to evaluate the contribution of each component to the total palynological configuration. Analogue data on modern pollen and vegetation and on ethnographic use of plants provide starting points for distinguishing among alternatives. The broad range of proveniences and large number of samples in the SGA Project analysis added the critical opportunity to extend comparisons between meaningful units of Hohokam remains. 112

References Cited

Berlin, G. Linnis, J. Richard Ambler, Richard H. Hevly, and Gerald G. Schraber 1977 Identification of a Sinagua agricultural field by aerial thermography, soil chemistry, pollen/plant analysis, and archaeology. American Antiquity 42(4): 588-600.

Bohrer, Vorsila L. 1968 Paleoecology of an Archaeological Site Near Snowflake, Arizona. MS, doctoral dissertation, University of Arizona, Tucson.

1978 Prehistoric Plant Remains from the Upper Bajadas and the Mesa-canyon Complex of the Central Arizona Ecotone Project. MS, Eastern New Mexico University, Portales.

1981 Methods of recognizing cultural activity from the pollen in archaeological sites. The Kiva 46(3): 135-142.

Castetter, Edward F. 1935 Ethnobiological studies in the American Southwest I: Uncultivated native plants used as sources of food. University of New Mexico Bulletin, 266, Biological Series 4(1): 1-62. Albuquerque: University of New Mexico.

Castetter, Edward F., and Willis H. Bell 1942 Pima and Papago Indian agriculture. Inter-American Studies 1. Albuquerque: University of New Mexico Press.

1951 Yuma Indian Agriculture: Primitive Subsistence on the Lower Colorado and Gila Rivers. Albuquerque: University of New Mexico Press. 113

Castetter, Edward F., Willis H. Bell, and Alvin R. Grove 1938 Ethnobiological studies in the American Southwest VI: The early utilization and the distribution of agave in the American Southwest. University of New Mexico Bulletin 335, Biological Series 5(4). Albuquerque: University of New Mexico.

Crosswhite, Frank S. 1981 Desert plants, habitat and agriculture in relation to the major pattern of cultural differentiation in the O'odham people of the Sonoran Desert. Desert Plants 3(2): 47-76.

Di Peso, Charles C. 1958 The Reeve Ruin of southeastern Arizona: A study of prehistoric Western Pueblo migration into the middle San Pedro Valley. The Amerind Foundation 8. Dragoon, Arizona: The Amerind Foundation.

Faegri, Knut, and Johannes Iverson 1964 Textbook of Pollen Analysis. New York: Hafner Publishing Company.

Felger, Richard S. 1980 Vegetation and flora of the Gran Desierto, Sonora, Mexico. Desert Plants 2(2): 87-114.

Felger, Richard S., and Mary B. Moser 1976 Seri Indian food plants: Desert subsistence without agriculture. Ecology of Food and Nutrition 5: 13-27.

Fish, Suzanne K. 1974 Palynology of some cultural contexts. Paper presented at the 73rd Annual Meeting of the American Anthropological Association, Mexico City.

1982a Palynology of the Modified Vegetation of Salt-Gila Hohokam Sites. Paper presented at the 47th Annual Meeting of the Society for American Archaeology, Minneapolis. MS, Arizona State Museum Library, University of Arizona, Tucson. 114

1983a Culture in Environment: Palynological Interactions. Paper presented at the 48th Annual Meeting of the Society for American Archaeology, Pittsburgh. MS, Arizona State Museum Library, University of Arizona, Tucson.

1983b Pollen from agricultural features. In "Hohokam Archaeology Along the Salt-Gila Aqueduct, Central Arizona Project, Vol. III: Specialized Activity Sites," edited by Lynn S. Teague and Patricia L. Crown. Arizona State Museum Archaeological Series 150(3): 575-603. Tucson: University of Arizona.

Fish, Suzanne K., and Charles H. Miksicek 1982 Sampling guidelines for pollen and floatation. Appendix C in "Hohokam Archaeology along the Salt-Gila Aqueduct, Central Arizona Project, Vol. I: Research Design," edited by Lynn S. Teague and Patricia L. Crown. Arizona State Museum Archaeological Series 150(1): 169-173. Tucson: University of Arizona.

Fish, Suzanne K., Russell J. Barber, and Charles H. Miksicek 1982 Environment and subsistence studies. In "Hohokam Archaeology Along the Salt-Gila Aqueduct, Central Arizona Project, Vol. I: Research Design," edited by Lynn S. Teague and Patricia L. Crown. Arizona State Museum Archaeological Series 150(1): 129-140. Tucson: University of Arizona.

Fish, Suzanne K., Paul R. Fish, and Christian Downum 1984 Hohokam terraces and agricultural production in the Tucson Basin. In "Agricultural Strategies in the Southwest," edited by Suzanne K. Fish and Paul R. Fish. Arizona State University Anthropological Research Papers 33: 55-71. Tempe: Arizona State University.

Forde, C. Daryll 1963 Habitat, Economy, and Society. New York: E. P. Dutton and Company. 115

Gallagher, Marsha V. 1977 Contemporary ethnobotany among the Apache of Clarkdale, Arizona area, Coconino and Prescott National Forests. Archaeological Report 14. Albuquerque: U.S.D.A. Forest Service, Southwestern Region.

Gasser, Robert E. 1980 Exchange and the Hohokam archaeobotanical record. In Current "Issues in Hohokam Prehistory: Proceedings of a Symposium," edited by David Doyel and Fred Plog. Arizona State University Anthropological Research Papers, No. 23: 72-77. Tempe: Arizona State University.

1982 Hohokam use of desert food plants. Desert Plants 2(4): 216-235.

Gentry, H. S. 1972 The Agave family in Sonora. Agricultural Handbook 399. Washington, D.C. U.S. Department of Agriculture.

Gish, Jannifer W. 1 979a Palynological research at the Pueblo Grande Ruin. The Kiva 44 (2-3): 159-172.

1979b Cave Buttes alternative dam mitigation, Arizona: a palynological perspective. In "Archaeological Investigations in the Cave Creek Area, Maricopa County, South-Central Arizona," by T. Kathleen Henderson and James B. Rodgers. Arizona State University Anthropological Research Papers 17: 158-179. Tempe: Arizona State University.

1 98 1 Pollen Results from the Adobe Dam Site, Central Arizona. MS, Museum of Northern Arizona, Flagstaff. 116

Gish, Jannifer W. and Linda J. Scott 1983 Experimental Palynology: Garden Plots and Washes of Ethnobotanic Plants. Paper presented at the 48th Annual Meeting of the Society for American Archaeology, Pittsburgh. MS, Arizona State Museum Library, University of Arizona, Tucson.

Greenhouse, Ruth, Robert E. Gasser, and Jannifer Gish 1 98 1 Cholla bud roasting pits: an ethnoarchaeological example. The Kiva 46(4): 227-242.

Halbirt, Carl D. and Robert D. Diggs 1 98 1 Analysis of pollen samples from La Ciudad. In "Archaeological Investigations, Testing at La Ciudad (Group III), West Papago--Inner Loop (I-10), Maricopa County, Arizona," by Ronald K. Yablon and others. MS, Museum of Northern Arizona, Flagstaff.

Hevly, Richard A., P. J. Mehringer, and H. G. Yocum 1965 Studies of the modern pollen rain in the Sonoran Desert. Journal of the Arizona Academy of Science 3(3): 1 23 - 1 3 5 .

Ikuse, Masa 1 956 Pollen Grains of . Tokyo: Hirokawa Publishing Company.

Kapp, Ronald O. 1969 How to Know Pollen and Spores. Dubuque, Iowa: William C. Brown.

Kearney, Thomas H., and Robert H. Peebles 1964 Arizona Flora, 2nd edition. Berkeley: University of California Press.

Lytle-Webb, Jamie L. 1 978 Pollen analysis in Southwestern archaeology. In Discovering Past Behavior: Experiments in the Archaeology of the American Southwest, edited by Paul Grebinger, pp. 13-28. Gordon and Breach, New York. 117

1 9 8 1 Pollen studies at Las Colinas. In "The 1968 Excavations at Mound 8 Las Colinas Group, Phoenix, Arizona," edited by Laurens C. Hammack and Alan P. Sullivan. Arizona State Museum Archaeological Series 154. Tucson: University of Arizona.

Martin, Paul S. 1 9 63 The Last 10,000 Years, a Fossil Pollen Record of the American Southwest. Tucson: University of Arizona Press.

Martin, Paul S., and James Schoenwetter 1965 Arizona's oldest cornfield. Science 132: 33-34.

Messer, Ellen 1 97 8 Zapotec plant knowledge: Classification, uses, and communication about plants in Mitla, Oaxaca, Mexico. Memoirs of the Museum of Anthropology. University of Michigan 10(2).

Miksicek, Charles H. 1983 Plant remains from agricultural features. In "Hohokam Archaeology Along the Salt-Gila Aqueduct, Central Arizona Project, Part III: Specialized Activity Sites," edited by Lynn S. Teague and Patricia L. Crown. Arizona State Museum. Archaeological Series 150(3). Tucson: University of Arizona.

Miksicek, Charles H., and Suzanne K. Fish 1981 Understanding Aboriginal Subsistence Activities in Southern Arizona. Paper presented at the Fourth Biology Conference, Columbia, Missouri.

Minnis, Paul E., and Stephen E. Plog 1976 A study of the • site specific distribution of Agave parryi in east-central Arizona. The Kiva 41(3-4): 299-308. 118

Nabhan, Gary P. 1979 The ecology of floodwater farming in arid southwestern North America. Argo-Ecosystems 5: 245-255.

Nabhan, Gary P., Amadeo M. Rea, Karen L. Reichhardt, Eric Mellink, and Charles F. Hutchinson 1 982 Papago influences on habitat and biotic diversity: Quitovac oasis ethnoecology. Journal of Ethnobiology 2: 124- 143 .

Pennington, Campbell W. 1 969 The Tepehuan of Chihuahua, Their Material Culture. Salt Lake City: University of Utah Press.

1980 The Pima Bajo of Central Sonora, Mexico, Vol. I: The Material Culture. Salt Lake City: University of Utah Press.

Plog, Fred 1 97 8 Explaining Culture Change in the Hohokam Preclassic. Paper presented at the 43rd Annual Meeting for the Society for American Archaeology. MS, Arizona State Museum Library, University of Arizona, Tucson.

Rankin, Adrianne G. 1 9 8 1 Paper presented at the 46th Annual Meeting of the Society for American Archaeology, San Diego. MS, Arizona State Museum Library, University of Arizona, Tucson.

Raynor, Gilbert S., Janet V. Hayes, and Eugene C. Ogden 1970 Experimental data on dispersion and deposition of timothy and corn pollen from known sources. Brookhaven National Laboratory 957(T-398), Brookhaven National Laboratory, Upton, New York.

Raynor, Gilbert S., Eugene C. Ogden, and Janet V. Hayes 1972 Dispersion and deposition of corn pollen from experimental sources. Journal 64: 420-427. 119

Rea, Amadeo M. 1979 The ecology of Pima fields. Environment Southwest 484: 8-13.

1 9 83 Once a River: Bird Life and Habitat Changes on the Middle Gila. Tucson: University of Arizona Press.

Robbins, Wilfred W., John P. Harrington, and Barbara Friere- Marreco 1916 Ethnobotany of the Tewa Indians. Smithsonian Institution, Bureau of American Ethnology, Bulletin 55.

Russell, Frank 1975 The Pima Indians. (Originally published in 1908.) Tucson: University of Arizona Press.

Schoenwetter, James A., and Larry A. Doerschlag 1971 Surficial pollen records from central Arizona, I: Sonoran Desert scrub. Journal of the Arizona Academy of Science 6(3): 216-221.

Scott, Linda J. 1983 Transects Near First Mesa. Paper delivered at the 2nd Annual Ethnobiology Conference, Flagstaff, Arizona.

Sheridan, Thomas E. 1978 Traditional Agricultural in the Southwest: Past, Present, and Future. MS, Center for Documentary Relations of the Southwest, Arizona State Museum, University of Arizona, Tucson.

Sires, Earl W., Jr. 1983a Excavations at Frogtown (AZ U:15:61). In "Hohokam Archaeology Along the Salt-Gila Aqueduct, Central Arizona Project, Vol. IV: Village Sites on Queen Creek and Siphon Draw," edited by Lynn S. Teague and Patricia L. Crown. Arizona State Museum Archaeological Series 150(4). Tucson: University of Arizona. 120

1983b Excavations at Smiley's Well (AZ U:14:17, Locus A): A transitional (Sedentary-Classic Period) Hohokam farmstead along the lower Queen Creek drainage, central Arizona. In "Hohokam Archaeology Along the Salt-Gila Aqueduct, Central Arizona Project, Vol. V: Small Habitation Sites on Queen Creek," edited by Lynn S. Teague and Patricia L. Crown. Arizona State Museum Archaeological Series 150(5). Tucson: University of Arizona.

Spaulding, W. G. 1974 A preliminary statement on the pollen analysis from the Escalante Ruin Group. In "Excavations at the Escalante Ruin Group, Southern Arizona," by David E. Doyel. Arizona State Museum Archaeological Series 37: 262- 268. Tucson: University of Arizona.

Stein, Pat H. 1979 Archaeological Investigations Along the Salt-Gila Aqueduct. Phoenix: Arizona Projects Office, United States Bureau of Reclamation.

Sweet, Muriel 1976 Common Edible and Useful Plants of the West. Happy Camp, California: Naturegraph Publishers.

Tuthill, Can The Tres Alamos site on the San Pedro River, southeastern Arizona. The Amerind Foundation 4. Dragoon, Arizona: The Amerind Foundation.

Whiting, Alfred R. 1939 Ethnobotany of the Hopi Indians. Museum of Northern Arizona Bulletin 15. Flagstaff: Museum of Northern Arizona. 121

APPENDIX C

PREHISTORIC DISTURBANCE FLORAS OF THE LOWER

SONORAN DESERT AND THEIR IMPLICATIONS

by Suzanne K. Fish

In Late Quaternary Vegetation and Climates of the American Southwest, edited by Bonnie F. Jacobs, Patricia L. Fall, and Owen K. Davis, pp. 77-88. American Association of Stratigraphic Palynologists Contribution Series No. 16. 122

American Association of Stratigraphic Palynologists

Contributions Series

Number 16

LATE QUATERNARY VEGETATION AND CLIMATES OF THE AMERICAN SOUTHWEST

EDITED BY

Bonnie F. Jacobs

Shuler Museum, Southern Methodist University, Dallas, Texas 75275

Patricia L. Fall and Owen K. Davis

Department of Geosciences, University of Arizona, Tucson, Arizona 85721

December, 1985 123

AMERICAN ASSOCIATION OF STRATIGRAPHIC PALYNOLOGISTS, INC. The Association was founded in 1967 to promote the science of palynology, especially as it relates to stratigraphie applications and biostratigraphy; to foster the spirit of scientific research among its members; and to disseminate information relating to palynology. In line with its goals, the Association encourages scientific collection, study, and publication of palynological data.

The AASP Contributions Series is published on an irregular basis by the American Association of Stratigraphie Palynologists Foundation for distribution to persons interested in the field of palynology. Opinions expressed herein are those of the writers and do not necessarily represent the views of the AASP Inc., AASP Foundation, or the editorial staff. Articles published in this series are subject to a review process prior to being accepted for publication.

AASP INC. OFFICERS 1984-1985

President: VAUGHN M. BRYANT, JR. President-E/ect: RAYMOND A. CHRISTOPHER Secretary-Treasurer: KENNETH M. PI EL Managing Editor: DOUGLAS J. NICHOLS Directors at Large: LUCY E. EDWARDS VIRGIL D. WIGGINS JUDITH K. LENT1N E. REED WICANDER

AASP FOUNDATION TRUSTEES 1977-1985

Chairman: NORMAN J. NORTON ROBERT T. CLARKE RICHARD W. HEDLUND

MEMBERSHIP AND PUBLICATIONS

Membership in the AASP is for the calendar year. Ducs arc $20.00 (U.S.) per year for individuals and $30.00 (U.S.) per year for institutional members. All members of the AASP Inc. receive Palynology which is published annually, the AASP Newsletter mailed out four times a year, and an annual membership directory. In addition to Palynology the AASP Foundation also publishes a monograph series called the AASP Contributions Series. Back issues of Palynology and the ilitSP Contributions Series arc listed on the inside back cover.

" Copyright by the American Association of Stratigraphie Palynologists Foundation, 1985 124

American Association of Stratigraphic Palynologists Foundation

October 1, 1993

Suzanne Fish Arizona State Museum University of Arizona Tucson, Arizona 85719

Dear Suzanne,

Approval is hereby given for you to use any and all information in your article in the American Association of Stratigraphic Palynologists Foundation publication Contributions Series No. 16. AASP Foundation is the copyright holder. Please give proper citation and/or credit to the original publication in your current work.

Sincerely,

- Robert T. Clarke Treas. AASP Foundation

Trustee a Charm Trustee & Treasurer: Trustee St Ediux: Trustee é Secretary: NOTZ11111 I. Norton Robert T. Clarke Richard W. Hedlund Vaughn M. Litirmaiy Chevron U.S.A.. Inc. Mobil Etr-P Tech. Center 6923S. Knoxville Ave. FalYnologY P.O. Box 1635 P.O. Box 650232 Tulsa, OK 74130 Texas A&M Univertity Houston, TX 77251 Dallas, 7X 75265-0232 tel. (918) 492-3302 College &mica. 1X77843-4352 tel. (713) 754-3848 teL (714) 951-2971 tel. (409) 845-5242 FAX (713) 754-2016 FAX (214) 951-4415 FAX (409) 845-4070 125

PREHISTORIC DISTURBANCE FLORAS OF THE LOWER SONORAN DESERT AND THEIR IMPLICATIONS

Suzanne K. Fish Arizona State Museum University of Arizona Tucson, Arizona 85721

Abstract Pollen spectra derived from archaeo- nantly non-arboreal desert vegeta- logical sites represent a mixed rec- tion. Second is the greater influ- ord of local disturbance species and ence of localized plant communities regional vegetation. Analysis of on pollen assemblages in surface over 400 samples from sites in the soils. Sonoran Desert of southern Arizona produced non-arboreal pollen types Many taxa are not wind pollinated and frequencies atypical of disturbed in desert scrub biotic communities of and undisturbed vegetation in the the Lower Sonoran Desert (Lowe, area today. The introduction of Old 1964). Numerically important species World weedy species and historic such as creosote bush (Larrea), the patterns of environmental disturbance cacti, and leguminous trees disperse have removed the opportunity to di- so little pollen that only occasional rectly identify pre-contact distri- grains are encountered. The few butions of herbaceous taxa. Differ- arboreal forms tend to be small, and ences between archaeological and most pollen is released near the modern disturbance floras suggest ground. General patterns of pollen that cultural intervention may have frequencies can be recognized in influenced regional herbaceous con- surface soils for widespread vegeta- figurations before the arrival of tion associations such as creosote

Europeans. bush - bursage (Larrea - Franseria). However, adjacent herbaceous or shrubby communities dominated by different species can result in sur- INTRODUCT ION face samples with widely divergent frequencies. A striking example of In low-elevation deserts of the change in surface spectra over short southwestern United States, the ma- horizontal distances is given by jority of pollen analyses have dealt Martin (1963, p.15) for desert grass- with terrestrial sediments, often in lands of southern Arizona. Samples archaeological sites. Two factors beneath dense stands of floodplain cause pollen distributions in such ragweed contained 70 to 80% Com- sediments to differ from the tradi- positae pollen; 120 m away, 90% tionally studied lake and bog depos- Cheno-Am (encompassing the Cheno- its of forested regions. First is podiaceae and the genus Amaranthus; the dissimilar pollen production and Martin, 1963, p. 401) pollen was dispersal from sparse and predomi- recorded.

AASP Contribution Series No. 16:77-88 (1985) 126

After studying the relationship of natural and culturally disturbed desertscrub vegetation and surface communities. A prime example is the pollen in transects near Tucson and Cheno-Am category. High values for Yuma, Arizona, Hevly et al. (1965, p. Cheno-Am pollen in desertscrub zones 135) concluded that no regional pol- may occur in conjunction with ripar- len rain could be defined. Important ian situations, alkaline soils, and pollen producers affecting sample the environs of human disturbance. frequencies were controlled by local Cultivated and wild edible species edaphic factors. The nature of pol- could further heighten pollen fre- len assemblages in desert soils pre- quencies in site sediments. Elevated sents an obstacle to reconstructing Cheno-Am percentages commonly encoun- region-wide vegetation and climate by tered in Southwestern archaeological introducing responses to complex and spectra have been variously inter- geographically limited conditions. preted as indicating vegetational At the same time, such assemblages response to climate, the weedy flora afford the spatial sensitivity to of human habitations, or use of re- detect localized and specific cultur- source species. al modifications in the prehistoric landscape. Examples illustrating the magni- tude of cultural influence on archae- ological pollen configurations have ARCHAEOLOGICAL POLLEN ASSEMBLAGES been drawn from recent analysis of IN THE SONORAN DESERT over 400 samples from three locations in southern Arizona. Values for Archaeological pollen assemblages infrequent pollen types that would be of the low deserts incorporate sever- anomalous for natural associations of al components: airborne pollen from the Lower Sonoran Desert can be link- regional vegetation, and more local- ed to disturbance communities in ized increments from both natural several prehistoric contexts. The associations and the disturbance composition of disturbance floras communities at a site. Introduction aids in understanding the culturally of pollen-bearing plant materials by modified environment and subsistence site inhabitants may add a fourth activities of site occupants, and component of artificially concen- suggests broader implications for trated types. For higher elevation Holocene studies in the Southwest. Southwestern environments, such tech- niques as arboreal ratios (Hevly, 1981) and principal components analy- METHODS sis (Fall et al., 1981) have been employed to distinguish environmental Samples used in this discussion or climatic patterns from culturally were collected between 1981 and 1983 biased ones. However, the problem from excavations near Florence and has not been fully resolved by cur- Tucson, Arizona. Sixty cc of sedi- rent methodology, and the relative ment were processed to insure ade- importance of each component is not a quate recovery of pollen from rela- matter of consensus among research- tively coarse and low-yielding desert ers. soils. After deflocculation in hy- drochloric acid, initial matrix re- Interpretation of archaeological duction was accomplished by a single spectra is further hampered by the swirl and settling step as described fact that abundantly dispersed and by Mehringer (1967, p. 136). Zinc commonly encountered pollen types may bromide of 2.0 density was then used be equally characteristic of both for further separation. Hydrofluoric 127

Prehistoric disturbance floras

and hydrochloric acid and water SPECTRA OF HOUSES AND rinses completed the extraction pro- IRRIGATED FIELDS cedure. In view of the low organic content of the soils and to minimize Vegetation encompassing the site potential damage to poorly preserved of Frogtown, near Florence, Arizona, grains, acetolysis, nitric acid, and is dominated by creosote bush (Earrea KOH treatments were not employed. tridentata) and bursage (Franseria The extract was mounted in glycerol deltoidea). Occasional trees are and stained for viewing. Tabulation mostly foothills palo verde (Cercid- of a standard 200 grain sum was made, ium microphyllum). The site was exclusive of cultigen types. This occupied from about A.D. 800 to 1100 sum was legislated by the difficulty by the Hohokam. Inhabitants lived in of recovering and concentrating pol- a loosely organized cluster of semi- len from site sediments, but a 200- subterranean pithouses, never total- grain sample size has been shown to ing more than 20 to 30 structures. produce reliable frequencies of com- Fields immediately adjacent to Frog- mon types in Southwestern spectra town were irrigated by small canals (Martin, 1963, pp. 30-31). Frequen- from Queen Creek, a seasonally flow- cies of pollen types are computed as ing tributary of the Gila River. percentage of all types present, Supplemental water was an absolute excluding cult igens. necessity for crops in this region with its present average annual pre- cipitation near 230 mm. MODERN POLLEN AND VEGETATION

Modern surface spectra at all Field Samples three study locations are dominated by more than 50% Ambrosia - type pol- Prehistoric irrigated fields were len. The present site communities located by proximity to both primary and their pollen records cannot be canals and lateral branches. Exact assumed to precisely duplicate pre- levels corresponding to prehistoric historic ones. However, vegetation planting surfaces were difficult to has never been cleared in post- define. Identification of crops was contact time for farming or other attempted by a diffuse vertical and purposes. Domestic animals have been horizontal sampling strategy from excluded from one Tucson site, Tuma- trench profiles at several separate moc Hill, for approximatley 80 loci (Text-Figures 1 and 2). Spectra years. Published transects near from upper levels matched modern Tucson (Hevly et al., 1965) and near distributions. After a transition at Florence (Schoenwetter and Doer- less than 30 cm, a different configu- schlag, 1971) confirm the typical ration of pollen types marked the dominance of Ambrosia - type pollen in levels of Hohokam fields. creosote bush-bursage (Larrea- Franseria) associations and often in A combination of three types ac- palo verde-saguaro (Cercidium- Cereus) counted for a major segment of each associations. These two associations spectrum; 35 to 45% of all pollen share many species. In both cases, consisted of Boerhaavia - type bursage is the major producer of (spiderling), Sphaeralcea (globe-

Ambrosia - type pollen. Frequencies of mallow), and Kallstroemia (Arizona this type are presented as approxima- poppy), with a corresponding reduc-

tions for the relative proportion of tion in Ambrosia - type pollen. Fre- archaeological pollen contributed by quencies of the three types vary natural vegetation. within individual samples. Maximum 128

Modern Surface

40-50 cm • Trench 50-60cm 197 60-70 cm 90-80 cm 20-30 cm Trench 20-30 cm

183 20-30 cm 70-80 cm

Trench 209 - 20-30cm „ 1 0% 75% 0 20516 0 40% 0 30% 0 25140 50%

Text-Figure 1. Pollen frequencies in Frogtown field samples at depths > 20 cm.

Depth in cm below Surface 0-10 I=INE COMM

[=:] Ambrosio-type • Boerhaavia Kallstroemia Sohaeralcea Cheno- Am

o 100%

Text-Figure 2. Relative porportions of Ambrosia - type, Cheno-Am, and agricul- tural weed pollen plotted by depth in Frogtown field trenches.

values were 37.5% for Boerhaavia, wetter and Doerschlag, 1971), but 26.5% for Sphaeralcea, and 20% for never in similar combined frequencies Kallstroemia. The agricultural pol- and never approaching these maximum len types typically occur today as a values. Agricultural weeds in Frog- few percent in surface samples from town fields, like the cultigens, natural vegetation communities in appear to have benefited from supple- southern Arizona deserts (Martin, mental moisture. Irrigation per- 1963; lievly et al., 1965; Schoen- mitted an artificial density of weedy 129

Prehistoric disturbance floras

herbaceous species in the midst of seeds, medicines, and a raw material the most xeric desert association. for paints among a number of South- western groups (Russell, 1908, p. 79; These three weedy genera, mention- Robbins et al., 1916, p. 61; Whiting, ed above, provide information con- 1938, p. 85; Elmore, 1943, p. 104; cerning Hohokam agricultural prac- Vestal, 1952; Gallagher, 1977, p. tices. Season of cultivation is not 97). obvious in an environment with a growing season spanning most of the A third kind of evidence concern- year and peak riverf low in the spring ing agriculture at Frogtown stems following winter rains and snow- from the vertical distributions of melt. Common Sphaeralcea species in pollen (Text-Figure 2). High the vicinity of Florence, Arizona, Ambrosia- type frequencies occur in have been observed to be biseasonally the upper 30 cm or less below the opportunistic. Weedy species of present ground surface. Deposition Boerhaavia are characteristically in the 800 years since occupation can abundant from midsummer until October therefore be bracketed. By contrast, (Parker, 1958, pp. 110-112). Species during a maximum 350 years of irriga- of Kallstroomia may bloom as early tion, at least 40 to 50 cm of sedi- as May (Parker, 1958, p. 176), but ment were deposited. An accelerated appear in the study area today in the rate associated with irrigation is later summer months. The combination consistent with floodwater farming points to summer as the foremost practices among historic Indian season of irrigation and concomitant groups. Canals filled during high weedy growth. seasonal flows deliver a suspended sediment load to fields, simulta- Most SphaeraZcea species are pe- neously moistening and enriching rennial (Kearney and Peebles, 1964, desert soils poor in organic content p. 540). In modern succession in (Nabhan, 1979; Doolittle, 1984). abandoned desert fields, globe mallow increases after two or three years (Karpiscak, 1980). These character- istics suggest that some herbaceous Habitation Samples plants may have escaped removal in an agricultural technology lacking Environmental and climatic recon- draught animals and plows or that structions have been based on pollen they were allowed to persist in the spectra from habitation areas, since fields from year to year as gather- these contexts have furnished the able resources. Ethnographically, great majority of Southwestern Southwestern Indians remove weeds archaeological samples. Residential selectively, supply water to desir- samples also offer a culturally fil- able volunteers, or even sow seeds in tered view of vegetation. Village order to insure the availability of and fields are immediately adjacent useful species within the weedy com- at Frogtown. Text-Figure 3 presents munities of their fields (Whiting, spectra from Frogtown house floors. 1939; Crosswhite, 1981; Bye, 1979). The three prominent agricultural pol- It is suggestive that Boerhaavia len types appear in residential areas coulteri is eaten as a potherb by the in appreciable quantities, although Seri Indians (Felger and Moser, 1976, less strongly than in the fields p. 15), that Kanstroemia has been (combined field mean = 36%; combined used medicinally (Robbins et al., habitation mean = 16.7%). The pres- 1916, pp. 56-57), and that ence of these types in unusual fre- Sphaeralcea has furnished edible quencies has been widely noted in ▪

130

Modern Surface

Structure 24 Structure 18 Structure 2 Structure 3 Structure 3 Structure 7 Structure 20 Structure 25 Structure 30 MIN Structure 38 Structure 38 Structure 39 Structure 57 Structure 15 Structure 17 Structure 14 WEI Structure 14 Structure 1 Structure 9 111=1 Structure 4 • Structure 23 Structure 37 MF

o 50% 0 25%0 20% 15%0

Text-Figure 3. Pollen frequencies in Frogtown house floor samples.

other Hohokam habitation sites cincts supported a distinctive suite (Bohrer, 1981; Gish, 1979; Webb, of weedy plants. 1981).

On the other hand, Cheno-Am is SPECTRA OF RILLSLOPE TERRAGES higher in Frogtown residential prove- niences than in the fields. Cheno- Stone walls, terraces, and struc- pods and amaranths are more "urban" tures characterize a category of weeds at Frogtown. A mean frequency hillslope sites in southern Arizo- for habitation samples is 15.1%, na. Two such sites near Tucson, Los compared to 6.8% from the fields. At Morteros and Tumamoc Hill, furnish some nearby Hohokam sites also sam- palynological evidence for distur- pled in this study, Cheno-Am pollen bance floras overlapping but not is more abundant still in habitation identical to those at Frogtown. Both contexts, comprising up to 33% of the occur on hills of less than 300 m total. Disturbed and enriched by height. Vegetation is typical of the domestic refuse, residential pre- palo verde-saguaro association, with

131

Prehistoric disturbance floras

7 o

Surface I

Hohokam tr l° cm 24 cm 42. cm ? Late Archaic{ 51 cm MINN 65 crn Non-Cultural — 90cm

Surfaceo Limmiow • Hohokami25cm • 45cm 0 90%0 50%0 40%0 35%0 55%

Text-Figure 4. Pollen frequencies in samples from Tucson archaeological terrace profiles.

creosote bush and bursage also numer- accounts for nearly half of all pol- ically prominent elements. At each len in the lower part of the ter- site, the modern spectrum contains race. In two of the three non-modern

more than 50% Ambrosia - type pollen. samples, Zea (corn) pollen is pres- ent.

Edaphic conditions in hillslope Los Morteros terraces do not duplicate those in irrigated fields, and not surprising- A number of terraces were trenched ly, the prominent weeds are not iden- at Los Morteros in order to investi- tical. Supplemental water supplied gate function. One class of wider by the terrace catchment was probably terraces supported structures and a more limited. Weeds may have been narrower set appears to have been more carefully controlled on re- cultivated (Fish et al., 1984). stricted terrace planting surfaces. Spectra from a stratigraphic column The highest Boerhaavia frequencies in a trench profile are shown in are elevated compared to modern un- Text-Figure 4. disturbed vegetation, but not nearly so pronounced as at Frogtown. The Uppermost in the terrace is soil isolated occurrence of Tidestromia that accumulated since prehistoric may indicate some unique agricultural

use. At 10 cm, Ambrosia- type pollen episode or weedy growth between the is still the dominant type. At time of terrace construction and 25 cm, this type has decreased to 40% initial cultivation. The seasonal and corresponds with the highest restriction of Tidestromia species to value for Boerhaavia. The lowermost middle and late summer is even more terrace sample at 45 cm reveals pronounced than for Boerhaavia Tidestromia (tidestromia) as a second (Kearney and Peebles, 1964, p. 268; prominent terrace weed. This type Parker, 1958, p. 106). 132

Tumanoc Hill crease. In both of these samples, Tidestromia totals nearly 10%. Tumamoc Hill supports a vegeta- Values for this type peak in the tional community similar to that at lowest Archaic spectrum. The highest

Los Morteros. Again, Ambrosia- type combined frequency for the three pollen is the most abundant type in a disturbance taxa, 70%, is found in surface sample. A Hohokam terrace this sample. investigated at the site was con- structed atop earlier deposits con- Higher Cheno-Am and tromia taining Late Archaic remains (Fish et and lower Ambrosia - type values dis- al., in press). The Archaic occupa- tinguish Archaic samples from those tion dates just prior to the appear- of the Hohokam terrace. The nature ance of ceramics in the region, but of disturbance as reflected by the includes cultigens. A trench inter- weedy response would appear to differ sected both occupations, and pollen between the two occupations. An samples represent a discontinuous interpretation of Archaic habitation sequence from the modern surface and Hohokam agriculture would be through Hohokam, Archaic, and cultur- compatible with both pollen and arti- ally sterile horizons (Text-Figure fact distributions. Below both cul- 4). tural horizons, a sterile and rocky basal layer of unknown age is Three weedy types with non- found. Its spectrum is unlike any synchronous vertical maxima are re- known at present for the Tucson area, vealed by spectra in stratigraphic with Cheno-Am and pine values both order. The upper 25 cm of the sam- near 30%. pling column is contained within the Hohokam terrace. At 10 cm below surface, a minor decrease in DISCUSSION

Ambrosia - type pollen frequencies is the only departure from the surface Archaeological Interpretation spectrum. At 24 cm, the bottom of the Hohokam level, the summer weed In the foregoing examples, the Boerhaavia shows a sharp increase. composition and contribution of archaeological disturbance floras The next three lower samples ex- could be easily distinguished because hibit variability within Late Archaic the abundant types are infrequent in disturbance communities. It cannot undisturbed vegetation. In the case be determined whether agriculture was of widely encountered Cheno-Am pol- practiced directly at the site during len, correspondence with particular this occupation, but charred corn kinds of disturbance could be estab- cupules and kernels were recovered lished only by contrasting spectra from this level. A diverse artifact from different prehistoric con- assemblage suggests more than passing texts. If disturbance taxa which visits by the Archaic peoples. disperse relatively little pollen can influence archaeological spectra so In the uppermost of the three greatly, interpretation of environ- Archaic samples, ChenoLAm pollen ment and climate from more prolific rises and Boerhaavia retains a producers should proceed with great heightened value similar to the adja- caution. cent Hohokam sample. The next lower spectrum shows only slightly reduced It is clear that prehistoric dis- frequencies of Boerhaavia, while turbance floras of the Lower Sonoran Cheno-Am pollen continues to in- Desert are not all equal and that 133 Prehistoric disturbance floras

variability in weedy species may be Intentional and unintentional human applied to cultural interpretations transport of seeds and creation of such as cropping season, field depo- cultural habitats may have similarly sition rate, and residential versus affected pre-Columbian distribu- agricultural function. In particu- tions. Unusual associations of par- lar, pollen configurations from ticular taxa with archaeological fields may provide a means for inves- sites have been noted (Yarnell, tigating prehistoric manipulation of 1965), and it seems plausible that disturbance floras to enhance desir- some field weeds may have followed able elements. cultigens to the Southwest from areas of geographic origin in Central or The potential for interpreting . both past environment and cultural activity would be realized if prehis- At the upper end of the time toric patterns could be matched with scale, previous disturbance species a particular analog activity in a may be disappearing. Local extirpa- particular kind of baseline vegeta- tion of Kallstroemia on the Colorado tion. Unfortunately, the necessary plateaus of Arizona and New Mexico combinations of environment and cul- has been suggested by Bohrer (1978) tural practice no longer can be fully on the basis of modern absence in duplicated. Alteration of New World contrast to multiple occurrences in environments in postcontact times has archaeological contexts. Similarly, been great; industrialized popula- Betancourt and Van Devender (1981) tions, Old World plants, and domestic have recorded Boerhaavia and animals have changed the landscape. Kanstroemia in late Holocene packrat The most common weeds on modern study middens at Chaco Canyon. Again, the area are naturalized species plants are absent today. Such fac- such as tumbleweed (Salsola kali) tors as heavy grazing by domestic (Karpiscak, 1980). Even the most animals or introduced pests may have traditional Native Americans do not exerted pressure on these taxa, but maintain a wholly aboriginal life- it is equally likely that loss of style. For example, the weedy suite optimal habitats and seed sources in of Hohokam fields in this study, Indian fields figures in historic produced by cultivation without reduction. plows, is not identical to that in currently plowed Papago Indian flood- water fields near Tucson where ama- Environmental Reconstruction ranths predominate. Therefore, both modern analogs and internal pattern- Unlike more readily encountered ing must be combined to interpret categories such as Compositae and archaeological spectra. Cheno-Am, pollen types with limited dispersal may appear only rarely in regional lacustrine spectra. How- Floristic History ever, in cases of small catchment, infrequent types may form meaningful Prehistoric disturbance communi- cumulative distributions within core ties have implications for floristic sequences (Davis et al., 1985). The as well as cultural history. Analy- detection of less common disturbance sis of adobe bricks in Spanish mis- types should signal the possibility sions has demonstrated how quickly that culturally biased patterns have Old World weeds became established been obtained. Where more local and common in the Southwest (Hendry records are involved, as with alluvi- and Kelly, 1925; Hendry, 1931). al profiles and packrat middens, the 134

ability to discriminate disturbance a redundant list, suites of prominent components is particularly critical weeds should be expected to vary with for valid environmental reconstruc- such factors as elevation, topogra- tion. phy, soils, climate, and human activ- ity. Prehistoric disturbance floras Taxa in the foregoing discussion revealed by pollen assemblages repre- are only illustrative of disturbance sent a repository of cultural and species that can be compiled from ecological information awaiting archaeological spectra. Rather than future directed study.

References Cited BETANCOURT, J.L., and VAN DEVENDER, DOOLITTLE, W.E. T.R. 1984 Agricultural change as an 1981 Holocene vegetation in Chaco incremental process. Annals Canyon, New Mexico. Science, of the Association of Ameri-

214: 656-658. can Geographers, 74: 124 - 137.

BOHRER, V.L. ELMORE, F.H. 1978 Plants that have become lo- 1943 Ethnobotany of the Navajo. cally extinct in the South- The University of New Mexico west. New Mexico Journal of Monograph Series, 1(7), Albu- Science, 18: 10-18. querque.

1981 Methods of recognizing cul- FALL, P.L., KELSO, G., and MARKGRAF, tural activity from pollen in V. archaeological sites. The 1981 Paleoenvironmental recon- Kiva, 46: 135-142. struction at Canyon del Muerto, Arizona, based on BYE, R.A. principal-component analy- 1979 Incipient domestication of sis. Journal of Archaeolog-

mustards. The Kiva, 44: ical Science, 8: 297 - 307. 237-256. FELGER, R.S., and MOSER, M.B. CROSSWHITE, F.S. 1976 Seri Indian food plants: 1981 Desert plants, habitat, and desert subsistence without agriculture in relation to agriculture. ecology of Food

the major pattern of cultural and Nutrition, 5: 13 - 27. differentiation in the O'odham people of the Sonoran FISH, P.R., FISH, S.K., LONG, A., and Desert. Desert Plants, 2: MIKSICEK, C.H. 3-61. In Early corn remains from Tuma- Press moc Hill, southern Arizona. DAVIS, O.K., HEVLY, R.H., and FOUST, American Antiquity. R.D., JR. 1985 A comparison of historic and FISH, S.K., FISH, P.R., and DOWNUM, prehistoric vegetation change C. caused by man in central 1984 Hohokam terraces and agricul- Arizona. (This volume.) tural production in the Tuc- 135 Prehistoric disturbance floras

son Basin. In: Fish, S., Arizona, Tucson. and Fish, P. (eds.), Prehis- toric Agricultural Strategies KEARNEY, T.H., and PEEBLES, R.H. in the Southwest. Arizona 1964 Arizona Flora. University of State University Anthropolog- California Press, Berkeley. ical Research Papers, 33: 55-71. LOWE, C.H. 1964 Arizona's Natural Environ- GALLAGHER, M.V. ment. University of Arizona 1977 Contemporary ethnobotany Press, Tucson. among the Apache of the Clarkdale, Arizona, area, MARTIN, P.S. Coconino and Prescott Nation- 1963 The Last 10,000 Years. Uni- al Forests. USDA Forest versity of Arizona Press, Service, Southwestern Region, Tucson. Albuquerque, New Mexico, Archaeological Report, 14. MEHRINGER, P.J. 1967 Pollen analysis of the Tule GISH, J.W. Springs area, Nevada. In: 1979 Palynological research at Wormington, H.M., and Ellis, Pueblo Grand ruin. The Kiva, D. (eds.), Pleistocene Stud- 44: 159-172. ies in Southern Nevada. Nevada State Museum Anthropo- HEN DRY, G.W. logical Papers, 13: 130 - 200. 1931 The adobe brick as a histor- ical source. Agricultural NABHAN, G.P. History, 5: 110-127. 1979 The ecology of floodwater farming in arid southwestern HENDRY, G.W., and KELLY, M.P. North America. Agro-ecosys- 1925 The plant content of adobe tems, 5: 245-255. brick. California Historical Society Quarterly, 4: PARKER, K. 361-373. 1958 Arizona ranch, farm, and garden weeds. University of HEVLY, R.H. Arizona Agricultural Exten- 1981 Pollen production, transport sion Service Circular, 265. and preservation: potentials and limitations in archaeo- ROBBINS, W.W., HARRINGTON, J.P., and logical palynology. Journal FREIRE-MARRECO, B. of Ethnobiology, 1: 39 - 54. 1916 Ethnobotany of the Tewa Indi- ans. Smithsonian Institution HEVLY, R.H., MEHRINGER, P.J., and Bureau of American Ethnology YOCUM, H.G. Bulletin, 55. 1965 Modern pollen rain in the Sonoran Desert. Journal of RUSSEL, F. the Arizona Academy of Sci- 1908 The Pima Indians. Bureau of ence, 3: 123-135. American Ethnology Annual Report, 26: 3-390. KARPISCAK, M.M. 1980 Secondary succession of aban- SCHOENWETTER, J., and DOERSCHLAG, doned field vegetation in L.A. southern Arizona. Ph.D. 1971 Surficial pollen records from dissertation, University of central Arizona, I: Sonoran 136

desert scrub. Journal of the 8, Las Colinas Ruins Group, Arizona Academy of Science, Phoenix, Arizona. Arizona 6: 216-221. State Museum Archaeological Series, 154: 337-354. VESTAL, P.A. 1952 Ethnobotany of the Ramah WHITING, A.F. Navajo. Papers of the - 1939 Ethnobotany of the Hopi. body Museum of Archaeology Museum of Northern Arizona and Ethnology, 40(4). Bulletin, 15.

WEBB, J.L. YARNELL, R.A. 1981 Pollen studies at Las 1965 Implications of disturbance Colinas. In: Hammack, L.C., flora on pueblo ruins. Amer- and Sullivan, A.P. (eds.), ican Anthropologist, 67: The 1968 excavations at Mound 662-674. 137

APPENDIX D

PRODUCTION AND CONSUMPTION IN THE

ARCHAEOLOGICAL RECORD: A HOHOKAM EXAMPLE

by Suzanne K. Fish and Marcia Donaldson

1991. The KIva, Vol. 56, No. 3, pp. 255-275. 138

KIYA, Vol. 56, No. 3, 1991

KIVA

Quarterly Journal of the

ARIZONA ARCHAEOLOGICAL AND HISTORICAL SOCIETY

Volume 56 Number 3 1991

HOHOKAM ETHNOBIOLOGY

CONTENTS

Preface Robert E. Gasser and Scott J. Kwiatkowski 189 A Look at Archaeological-Ethnobiological Collaboration in Hohokam Studies J Simon Bruder 193 Regional Signatures of Hohokam Plant Use Robert E. Gasser and Scott]. Kwiatkowski 207 Recently Recognized Cultivated and Encouraged Plants Among the Hohokam Vorsila L. Bohrer 227 Current Perceptions, Recent Discoveries, and Future Directions in Hohokam Palynology Jainufer W. Gish 237 Production and Consumption in the Archaeological Record: A Hohokam Example Susanne K. Fish and Marcia Donaldson 255 Hunting by Hohokam Desert Farmers Christine R. Suter 277 Hohokam Paleonutrition and Paleopathology: A Search for Correlates T. Michael Fink and Charles F. Merbs 293 Dietary and Parasitological Analysis of Coprolites Recovered from Burial 5, Ventana Cave, Arizona Karl J. Reinhard and Richard H. Herly 319 139

INSTRUCTIONS FOR AUTHORS

)0.'8 Is published quarterly by the Arizona Archaeological and Historical Society, an affiliate of the Arizona State Museum, The University of Arizona, Tucson. Kiva welcomes un- solicited, original papers relating to the prehistoric and historic archaeology, ethnology, history, and ethnohistory of the southwestern United States and northwestern Mexico. All manuscripts submitted are reviewed by scholars familiar with the subject matter. Pub- lished papers become the copyright of Kiva and the Society. CORRESPONDENCE. Correspondence concerning manuscripts should be addressed to the Editor, Kiva, Arizona State Museum, The University of Arizona, Tucson, AZ 85721. MANUSCRIPTS. Manuscripts must be typed, double-spaced, on one side of 81/2 x 11 in. bond paper. An original and two copies must be submitted, each accompanied by photo- copies of all illustrations. Manuscripts should be paginated consecutively beginning with the title page. Content of the paper should be indicated by appropriate headings and sub- headings. A brief abstract of 10010 150 words must accompany the paper. The fisting of ref- erences should include only those cited in the text. TABLES, FIGURES, AND ILLUSTRATIONS. Kiva printed page is 5 x 7 in. Authors should scale photographs and line art to this ratio (5:8). Line drawings should be drawn in black Ink no larger than 9 x 12 in. and be suitable for 50% reduction. Photographs should be glossy prints no larger than 8 x 10 in., also suitable for 50% reduction. All figures and tables must be cited in the text, be numbered consecutively, and have descriptive titles. Each figure should be labeled on the back with the author's name and the figure number. The space used by the photographs, line art, and tables is included in the page allowance for manuscript length (see below). Authors are requested to keep all original illustrations until these materials are requested by the Editor. PAGE CHARGES. Manuscript length should not exceed an aggregate length of 30 pages, including typescript text, photographs, line art, tables, and references. A charge of $25.00 per Kiva page will be applied to papers in excess of 15 printed Kiva pages. As a general rule, two double-spaced pages of typescripts equal one printed pageot Kiva. (In other words, 30 manuscript pages will be about 15 printed Kiva pages). • GALLEY PROOFS AND OFFPRINTS. Galley proofs only will be sent to authors. Proofs should be read, corrected, and returned to the Editor within 48 hours of receipt. Minor alter- ations of the text in galley proofs are permitted, but major alterations will be charged to the author. Each author will receive 10 complimentary offprints and one copy of the Kiva in which the paper appears. STYLE. Manuscripts should be prepared according to the current Style Guide for Ameri- can Antiquity (4812):429-442, April 1983), and to The Chicago Manual of Style (13th ed., The University of Chicago Press, 1982). Manuscripts submitted in an inappropriate style willte returned.

COPYRIGHT ARIZONA ARCHAEOLOGICAL AND HISTORICAL SOCIETY, INC. TUCSON, ARIZONA 1991 ALL RIGHTS RESERVED ISSN:0023-1940

Cover design by Jacki Breheney 140

KIVA The Journal of Southwestern Anthropology and History Office of the Editor Arizona Archaeological and Historical Society Arizona State Museum, University of Arizona Tucson, Arizona 85721

1 July 1993

Suzanne K. Fish 2214 E. 4th St Tucson, AZ 85719

Dear Ms. Fish,

This note is in response to your request to reprint "Production and Consumption in the Archaeological Record: A Hohokam Example" from Kiva Vol. 56, No. 3, pp. 255-276.

The Arizona Archaeological and Historical Society grants permission for the reprinting of this article in your doctoral dissertation in the Arid Lands Resource Sciences Interdisciplinary Program at the University of Arizona.

Sincerely,

,

Gayie Harrison Hartmann Editor 141

KIVA, Vol. 56, No. 3, 1991 PRODUCTION AND CONSUMPTION IN THE ARCHAEOLOGICAL RECORD: A HOHOKAM EXAMPLE

SUZANNE K. FISH Arizona State Museum University of Arizona Tucson, AZ 85721 MARCIA DONALDSON Department of Anthropology Arizona State University Tempe, AZ 85287

ABSTRACT Analysis of botanical materials from archaeological sites provides basic data in the form of quantified resource tabulations. Although biases are acknowledged in the formation of this record, it is usually assumed that sufficiently large sam- ples provide insight into representative resources for a provenience. It is a further inferential leap to equate recovered resources with either production or consumption, or with local or extra-local acquisition. We must attempt to dis- tinquish evidence of production from consumption in the archaeological record if we are to effectively investigate problems such as resource specialization, trade, and redistribution. These distinctions can be approached through direct evidence of production and consumption activities and through circumstantial and comparative argument. Illustrations are drawn from early Classic period Hohokam contexts in the Tucson Basin.

RESUMEN El arzdlisis de restos botanicos provenientes de sitios arqueol6gicos propor- ciona datos bdsicos en forma de tablas con cantidades de recursos. Aun cuando se reconocen sesgos en la formaci6n de este registro, normalmente se asume que muestras suficientemente grandes permiten visualizar los recursos rep- resentativos en una unidad de proveniencia dada. Equiparar recursos re- cuperados con producci6n o consumo, o con adquisici6n local o extra-local , constituye aun otro salto inferencial. Debemos tratar de distinguir la evidencia de producci6n de la evidencia de consumo en el registro arqueol6gico afin de indagar en forma eficiente problemas tales como la especializaci6n en ciertos recursos, el intercambio y la redistribuci6n. Estas distinciones pueden ser en- caradas a frayés de evidencias directas de actividades de producci6n y con- sumo, as! coma mediante evidencias circunstanciales y argumentas corn- parativos. Se ofrecen ejemplos derivados de comextos Hohokam del Perlodo Cldsico en la Cuenca de Tucson.

Hohokam botanical analyses have accumulated at an accelerated rate over the past 10 years. The means of disseminating and digesting these data have barely kept pace. Difficulties in compilation and comparison are being encountered as a consequence of nonstandardized recovery methods. Lagging furthest behind, perhaps, is the development of comparative techniques to fully exploit the mas- sive data sets and conceptual frameworks for applying results to broader economic and organizational issues of current archaeology. 142

FISH AND DONALDSON

We will focus on the concepts of production and consumption. The relation- ships between plant remains from archaeological sites and the body of assump- tions involving these concepts will be examined in a case study from the northern Tucson Basin. The discussion addresses interpretation of plant remains because this is the realm of our joint experience, but we believe that the examination of in- ference applies more broadly to biotic remains and often to nonbiotic ones as well.

DEFINING THE CONCEPTS Production and consumption can be thought of as sequential segments of the more inclusive process of human resource use. We define production as the origi- nation process by which materials are made available for future consumption. This definition admits both production through the direct action of cultivation and production in the sense of cultural acquisition and accumulation of wild resources. Consumption is the disposition of a product, after which potential for similar dis- position or use is exhausted. These processes may be repetitive in the history of a particular resource; the production of a refined product may involve consumption of a raw material or a prior processed form. Because these are analytical con- structs, the boundary between production and consumption may sometimes be blurred. For instance, intermediate storage might be regarded as simultaneously linked to both processes. Archaeobotanical analyses generate lists of recovered taxa from individual and cumulative proveniences. Quantified by a variety of techniques, these lists can be manipulated to characterize patterns in subsistence remains for features, sites as a whole, and ultimately regions. Even the simplest interpretive statements regarding recovery patterns at each level require assumptions about production and consumption, which all too often are neither explicit nor linked to an adequate body of supporting evidence.

ASSUMPTIONS AND INTERPRETATIONS To illustrate this situation, we will use a straightforward example of two sites yielding only the same two taxa, agave and corn. Complicating factors such as disparate seasons of occupation and multi-stage processed products are excluded. We assume that recovery and reporting techniques are equivalent and that a simi- lar range of features was proportionally sampled. At site A, remains consist of 80 percent corn and 20 percent agave, and at Site B, 20 percent corn and 80 percent agave. At least three interpretations of the data can be made: (1) Quantities reflect balanced production and consumption by site inhabitants. Much more corn was both produced and consumed at Site A and much more agave was both produced and consumed at Site B. (2) These recovery percentages reflect greater surplus production of corn at Site A and agave at Site B. At both sites, a more balanced 143

Production and Consumption in the Archaeological Record ratio of actual consumption was achieved by exchange. (3) Each site is produc- tively self-sufficient for both resources at an equivalent level; locational biases differentially affect quantitative recovery. Initial field-side processing of corn at Site B and of agave at Site A account for a reduction in production residues of the respective resource within each site boundary. These three equally plausible in- terpretations of the basic quantitative data are not exhaustive but aid in illustrating implicit assumptions and the need for supporting evidence. Some level of corn and agave consumption at each site is virtually the only in- terpretation possible from these data without assumptions concerning production and consumption. (Situations in which the social unit producing a resource would not also participate in its consumption seem imcompatible with the level of dif- ferentiation in Hohokam society.) The most conservative of the three alternatives presented above is clearly the first. It is usually assumed that quantities of remains reflect some balanced combination of both production and consumption com- pleted within the recovery site. Our experience suggests that this interpretation is likely to be considered acceptable without further confirmation. The second and third interpretations would be received more skeptically in the absence of addi- tional, substantiating evidence. The propensity to regard production and consumption as self-contained within the household or at most the site is undoubtedly influenced by ethnographic analogs. Ethnographic plant use is widely documented in the Southwest, and is es- sential to our understanding of archaeological remains in most respects, but may restrict it in some instances. In the case of the Hohokam, several spheres of subsis- tence interpretation can be identified for which Piman analogy is almost certainly inadequate. These include uniquely prehistoric technologies such as the largest canal networks and rockpile fields, the role of plants such as agave that were culti- vated less widely in historic than prehistoric times, landuse over broad areas of Hohokam settlement unoccupied by later Pimans, and the economic implications of prehistoric organizational modes reflected in ballcourts, compounds, platform mounds, and the higher range of population densities. Balanced production and consumption cannot be accepted as the parsimo- nious or normative interpretation of abundances of plant taxa without simultane- ous rejection of alternatives. The same canons of argument or evidence should be sought in all cases. Furthermore, imbalances between production and consump- tion in specific contexts are of particular relevance for current issues in Hohokam research such as specialization, redistribution, and trade. Thus, the ability to ef- fectively distinguish relative magnitudes of production and consumption in the archaeological records is critical.

THE PROBLEM OF REDUNDANCY The majority of plant species used by the Hohokam are recovered at most adequately studied sites. Much of the duplication in gathered staples such as mes-

144

FISH AND DONALDSON

CZ/ 0 o

D :

a o "0 d /itt''''', • rf'

° 0 -- • .L.Is 0c'' .. ZONE / 4° 44. /

.--,011 Arai. ' ceK 92' e° rlif't)r .., ZONE 3 .-6(4,• • f ''''' /. .

ZONE 2 .•°' ed` 1 •

ZONE /% 1

co Archaeological Sites

6.• Early Classic Period Sites

• 12tL'

Figure 1. Settlement and zonal configuration of the Marana community. 145

Production and Consumption in the Archaeological Record quite, cholla, or saguaro is attributable to widespread natural distributions and ac- cessibility within most inhabited zones of the Hohokam culture area. Redundancy in crops is the result of successful cultivation through a variety of floodwater and runoff technologies, as well as through irrigation along major rivers. Recent re- search has not supported the notion that the Hohokam, and particularly the non- riverine populations, vacillated periodically between cultivated and wild products as a function of crop failure and dietary stress (for example, Fish 1984; Gasser 1988). Rather, broadscale redundancies in Hohokam cultigens and gathered re- sources can be understood as the outcome of balanced and sustainable mixtures of wild, semi-cultivated, and cultivated resources. Variability in Hohokam subsistence is characterized by differential degrees of emphasis rather than by specialization approaching exclusivity of production; pres- ence alone therefore imparts little insight into meaningful patterns of consumption or production for any plant resource. Production, consumption, and imbalances be- tween these processes must be inferred, with reference to qualitative attributes, quantitative distributions, and contextual associations of botanical remains (for example, Dennell 1974, 1976; HaIly 1981; Hastorf 1988; Hillman 1981, 1984). The reliability of such inferences is enhanced by supporting evidence in the form of facilities and artifacts related to gathering, cultivation, processing, storage, food preparation, or manufacture, and by environmental attributes of the recovery locus. Additional pertinent measures might include the proportional amount of space devoted to production processes (Tosi 1984), and other comparisons of non- biotic or environmental correlates.

A MULTISCALE EXAMPLE FROM THE TUCSON BASIN The geographic or cultural scale at which imbalances in production and con- sumption occur is of major concern for broader economic interpretation. Contrasts may apply at the level of households, sites, settlement zones, multi-site com- munities, regions, or culture areas. Our ability to detect and identify differential levels of production and consumption may not be equal at each of these scales and may vary in this respect for differenct resources. Present understanding of the archaeological units at which production and consumption can best be demonstrated and quantitatively compared is limited. We believe that the most powerful insights can be derived through cumulative evidence from a sequence of levels and multiple classes of remains. To explore a multiscale approach, we attempt to examine production and consumption from the smallest unit upward and the largest unit downward in a segment of a Classic Period, multi- site community near Marana, Arizona (Figure 1). Two small sites, and three loci of similar size that are encompassed by a single site designation (Figure 2), represent constituents of a band of rancherfa settlement paralleling the Santa Cruz floodplain along the lower bajada or valley slope. These settlements are contained within an early Classic period (about A.D. 146

FISH AND DONALDSON

o

o 2 km

c=p o Archaeological sites omo• Rockpile fields C1Z;) Site yielding botanical remains • Plat form mound

Figure 2. Rockpile fields (Zone 2) and excavated sites yielding botanical remains in the Marana community.

1100-1300) multi-site community, for which the central mound site lies nearby to the north. The five clusters of structures are typical of the dispersed, small-scale settlements occupied by a substantial segment of Hohokam populations. Although occupation may have begun somewhat earlier at one site and persisted later at the other two, chronometric evidence suggests major, simultaneous occupations at each from the Rincon-Tanque Verde phase transition through an early portion of the Classic period predating the appearance of Gila Polychrome pottery (Hender- son 1987a). Sequential structures at the sites appear to represent continuity in 147

Production and Consumption in the Archaeological Record households and courtyard groups over time and presumably maintenance of pro- ductive orientations. The Marana sites were excavated as a part of the Central Arizona Project miti- gation sponsored by the Bureau of Reclamation (Henderson 1987c; Rice 1987c). Full coverage survey of 350 square kilometers that defined the configuration of the Classic community was accomplished in a separate study supported by this agency, the State Historic Preservation Office, and the National Science Founda- tion (S. Fish and others 1989, 1990; P. Fish and others 1984). These two high and low resolution data sets permit examination at a variety of scales. Cultural and environmental patterns resulting from regional survey are uncommonly com- prehensive for Hohokam studies. Horizontal exposures at each site are relatively extensive. Botanical data are represented by pollen and flotation samples as well as by individual charcoal specimens recovered during excavation (Fish 1987a, 1987c; Miksicek 1987a, 1987b). The resources considered in this discussion are corn, agave, cholla, saguaro, and cotton. These represent primary food and fiber sources exploited by the Hohokam and by historic Pima and Papago within south- ern Arizona. In the first section, we will review the distribution of plant remains, relevant facilities, and artifacts to characterize production and consumption at the level of intrasite proveniences. These processes will then be examined at the larger scales of site, multi-site community, and region.

The Intrasite Level Kisselburg (1987) noted a degree of craft specialization in the excavated Marana sites based on artifact distributions, and Rice (1987b) has compared this differentiation with household specialization in chiefdoms. Households consist of several residential structures with associated storage and other features. Structures appear to have a range of functions, from strictly residence or storage, to admix- tures of these uses with craft production (Rice 1987a). All houses may not have been inhabited simultaneously, but clustered dating (Henderson 1987a) supports overlapping and generally contemporary occupation. Evidence for differential production and consumption at the intrasite level will be evaluated by comparing distributions among structures and proposed households. Assessing relative proportions of production and consumption within the Marana sites involves a number of problematic factors. It can only be assumed that common location permitted equal access to resources by all occupants with- out significant social restrictions. Other potentially complicating factors include post-habitational processes, the possibility that many household activites took place outside the residential environs of structure or household, and inconsisten- cies in various aspects of sampling. For intrasite comparison, we focus on the more thoroughly sampled Rancho 148

FISH AND DONALDSON

Table 1. Plant Remains and Associated Artifacts Recovered from Structures and Households at Rancho Derrio, AZ AA:12:3(ASU).

Household 1 Household 2 PH19(R) PH21(S) AS29 PH18(R) PH24(R) Whorls/Blanks 4 0 0 0 1 Tabular Knives 0 0 5 4 0 Manos 1 1 0 4 0 0 0 0 1 0 Corn — / + ° +1+ +/— +1— +/NA Agave —/+ —/— —/— —/— —/NA Cotton —1+ — / — — / — —/— Cholla + / — —/— +/— +/— +/NA Saguaro —/— —/— —/— —/— —/NA a. Plant resources are indicated as present (+) or absent (—) according to recovery technique in the following order: pollen analysis/flotation. Key: PH = Pit house; AS = Activity Surface; R = Residential; S = Storage; NA = Not Analyzed. Data on spindle whorls and whorl blanks, and tabular knives are from Rice (1987a: Table 5.9). Data on manos and metates are from Bernard-Shaw (1987b: Table 9.2). Pollen results are from Fish (1987a: Appendix 7.2). Flotation results are from Miksicek (1987a: Table 9.2).

Derrio, AZ AA:12:3(ASU). and Muchas Casas, AZ AA:12:2(ASU), Locus A (Figure 2). Stripping of large blocks in these loci assures comprehensive recovery of artifacts and features. Botanical sampling and analytical coverage is somewhat variable within these stripped areas due to the usual strictures of time and money, but encompasses the majority of structures and a variety of extramural features (Miksicek 1987a, 1987b; Fish 1987a, 1987c). Plant remains and related artifacts are presented in Tables 1 and 2. Due to dif- ferences in critical factors of formation process such as burning, and variability in numbers of samples from each architectural unit, plant resource data has been "normalized" to simple presence or absence. In part, incongruencies between flo- tation and pollen results from the same provenience reflect the differential detecta- bility of particular resources by the two techniques. Designations of structure function as primarily residential or storage were assigned by Rice (1987c:79) ac- cording to the presence or absence of hearths. Evidence for cholla and saguaro production and consumption is limited to distributions of plant remains. Probable artifact correlates can be associated with corn and agave, and cotton to a lesser degree. Admittedly ground stone can be used with a host of other resources, but is an essential element of corn consump- tion. Spindle whorls and unperforated circular sherds, regarded as whorl blanks, were reported as a combined category from structures (Rice 1987a) and therefore cannot be considered separately. These artifacts types could pertain to cotton as well as agave, although some unperforated disks may have alternative uses. The

149

Production and Consumption in the Archaeological Record

n-n 0 g) o 4. - ct -6 CD y-) , c" (--n

0.) O r--• -0 -... erN +1+ 11 c. 'ccl. c,S CO CD CD 0 .72) CO .,i) t....)) (..) ' t E . 7, a) a) 0 " CG ri) -0 :-'4 0 >, ++1 Ii II g CD CD rq m (-4 o o '-' II a c21 L,u el +11+1 + x a) ,..,• CO G-2, a . c) ..o 4t •-s• ,,,- - as fx .0 '7/ c. ot cU 0 0 4; --", C.) 'CI ct LI. < < < < < + cd G (... • 0, Z Z ZZ Z •-•./ t a: L ''..-..‘

+ I +++ tct, Cr9P.. G 1'r- 00- Xr.: œi Oc L5 2 .5. 0... .0 — a 11 72. e9 g. CO O.- 0, 73 < = rx < O cl) + II CD CD CD c'S E 0 un 1:3 n c--. I I + I - m= .----. ,- cco, m 0 ci.) - .- ct, .,..., 0 m 0 ..= CO 0 = z z zcu = 3 2.%) L.T.: 0 c,, c..) a .14 4.) a> CO O 1.-2 - E,...., 2 0 H o.. 0 ,.., Pa ,..cu 0 ‘... c0 .5.1 0 a cu CO ,.., 0- 0 -o.) 0 a ,a `a' 0., 0 .0 VI E > o = = (3-•CI ° = = a :.--., o bi, ., V.,, al 0 CO ,3 15 o 0.1) o ..= 4) oel cl) rn -(._)

FISH AND DONALDSON tabular knife category subsumes both types with irregular or sharp edges appro- priate for detaching agave leaves during harvesting and types with smooth edges and, in some cases, handle projections for extracting fiber. Only the sharp-edged type has been recovered in Marana agave fields. Tables 1 and 2 record the presence of resources and numbers of relevant arti- facts within the 10-cm level defined as a floor association.' Information is arranged by structure and activity surface. These in turn are combined into proposed court- yard groups or households based on Henderson (1987a). A few tentative distinctions can be made among structures and households. Pit houses 226 and 214 at Locus A are similar in being devoid of all botanical re- mains except modest amounts of cholla pollen, while containing fiber-related arti- facts. Within households, it appears possible to identify structures such as these which have a linkage with fiber production but minimal remains pertaining to raw materials or foods. Such structure patterns could reflect functional specialization toward fiber production. At each site the structures with the strongest combined evidence for fiber activities are not necessarily these just described, however, but ones with more plentiful resource records. For example, Pit House 19 at Rancho Demo and Pit House 211 at Locus A yielded the higher range of whorls and blanks in conjunction with remains of multiple resources including fiber species. Both cotton and agave were detected in Pit House 19 and cotton in Pit House 211. At the scale of individual structures, artifacts associated with either corn or agave tend to dominate the abbreviated assemblage dealt with here. That is, struc- tures with relatively abundant agave or fiber processing artifacts have little or no ground stone. (An exception is Pit House 18 at Rancho Derrio where relatively numerous artifacts associated with both activities are found together.) This pattern suggests that fiber production (detachment of agave leaves, extraction of agave fiber, and spinning of agave or cotton fiber) and activities linked with preparation of corn for consumption tended to be concentrated within different structures in the two settlements. Although magnitudes of production and consumption of particular resources may be significantly different among structures and households in the same site, reliably representative quantifications are difficult to achieve. The physical nature of a resource and the manner in which it is processed influence the likelihood of preservation in a manner detectable by a particular recovery technique; formation processes incorporating such evidence into the archaeological record of structures are not necessarily systematic or uniform (for example, Hally 1981). Pollen col- lected directly from floor sediments that accumulated and were compacted over time represents a substantial interval of deposition during structure occupation and, with adequate sampling, may provide comparable rosters for architectural units. Charred remains on floors of structures frequently reflect a more temporally compressed period of time near the end of an occupation or subsequent trash dis- posal. For these materials, contents of burned and unburned houses are not likely 151

Production and Consumption in the Archaeological Record to reflect similar behavioral patterns in an equivalent manner. Inferences can be drawn from botanical remains alone, but added evidence in the form of artifacts and facilities is confirmatory and often critical in converting the botanical pre- sence of a resource into evidence for production or consumption. Among structures and even households, numbers of artifact indicators at the Marana sites are small for convincing differentiation and comparison, particularly in view of the fact that comprehensiveness of floor assemblages is quite variable. As with plant remains, even implements near the floor may represent inclusions in post-occupational trash in some instances. These problems are notably com- pounded by the typical Hohokam mixture of burned and unburned structures, with and without extensive floor artifacts. An adequate sample of securely associated extramural facilities such as ramadas or other activity surfaces and pits for each household might have sharpened discrimination of production and consumption. Sufficient samples of these feature types were not available for this approach, and we move on to more amenable data at the site level.

The Site Level Lumping of data within sites should serve to smooth bias in individual prove- niences stemming from uneven sampling, modes of assemblage origin, and post- occupational processes. Secondary trash deposits become valid associations. In the present study, site level analysis also offers the advantage of incorporating ex- tramural features, which were not so uniformly sampled as to permit comparison among structures or households. Site level comparison is confined to mechani- cally stripped areas in order to insure representative recovery of the various fea- ture types. Distributions of cholla and saguaro, lacking clear correlates in artifacts or facilities, will be examined first among the five Marana settlements (Table 3). Pollen and flotation results are available with the exception of Rancho Bajo, AZ AA:12: 1(ASU), where only pollen was analyzed. Rancho Derrio clearly yields the weakest record, with the lowest values for cholla and minor evidence for saguaro. At Rancho Bajo, less cholla pollen was encountered than at any Muchas Casas locus, but equal saguaro. Quantities recovered by both pollen and flotation concur in identifying Muchas Casas loci as the settlements with greater cactus abundance. Amounts of cholla are greatest at Loci D and E. Locus E is also nota- ble for saguaro, as is Locus A. It can be argued that cholla and saguaro remains should be generated most profusely by production. Pollen should he more plentifully dispersed by raw materials than refined plant products. Similarly, saguaro and cholla seeds, spines. and occasional soft tissue residues are most likely to be burned in initial process- ing by cooking. Accidental burning of processed food may take place during meal preparation or subsequent disposal, but the majority of durable parts would have 152

FISH AND DONALDSON

,•••

.2

O

N •lr 0 0 sr3 c•i Z

. . . --•

C‘I < 'tt •n,1- z

S.C) Q C'") •-•-• f•I 153

Production and Consumption in the Archaeological Record been previously removed. It is probable that Loci D and E produced more cholla and that Loci A and D produced more saguaro. Even quantitatively low amounts indicate the two resources were consumed at all sites. The possibility of exchange occasioned by differential production cannot be resolved with these data alone. In historic times, however, the Papago are known to have regularly traded saguaro syrup to the Pima (Russell 1975:93), although both groups produced it. Production of corn can be assessed through the combination of botanical dis- tributions for corn and agricultural weeds (Table 4). Each technique emphasizes different weedy species. For flotation, this category includes globe mallow, spurge, spiderling, lupine, several grasses, purslane, trianthema, and clammy weed (Miksicek 1987a:210). For pollen, spiderling, globe mallow, Arizona poppy, chenopods, and amaranths are included (Fish 1987c:163). Both pollen analysis and flotation reveal low amounts of agricultural weeds at Rancho Derrio. Lesser amounts of such weeds appear to reflect comparatively minor field acreage near this site, a finding in keeping with its poorer situation for floodwater farming than the others. Corn pollen is also lowest as would be ex- pected with a minimum of freshly harvested crops. Muchas Casas, Locus D has a similar but less pronounced pattern for agricultural weeds and corn pollen. In contrast, to indications from weed remains and corn pollen, ubiquity of charred corn at Rancho Derrio is not less than at other sites (Table 4). Ground stone, a marker for consumption, also bears out the likelihood of equivalent dietary use. Ground stone occurrences per excavated cubic meter approximate quantities at the other sites when fragments of secondary function in roasting pits are excluded (Table 4). It appears that inhabitants of Rancho Demo consumed as much corn as neighboring settlements while producing less. Muchas Casas, Locus D exhibits lowest corn ubiquity in flotation and ground stone in the general site range. Exchange would offer the means to correct imbalances between corn produc- tion and consumption. Ratios of recovered kernels to cupules may reflect ex- change due to the nature of resource transfer. Cupules are structures on cobs that contain the kernels and are typically charred when cobs are burned as fuel. As edi- ble parts, kernels are more directly related to food consumption. The higher ratio of kernels to cupules at Rancho Demo (Table 4) than at the more plentifully pro- ducing sites would be expected if shelled corn rather than whole cobs was often exchanged. In this case , fewer cobs would be available at Rancho Demo for fuel, producing a lower proportion of charred cupules. Muchas Casas, Locus D likewise exhibits a high ratio of kernels to cupules. Distributions of agave, cotton, and related artifacts and facilities are pre- sented in Table 5. Agave pollen is deleted due to rarity, as agave is typically har- vested prior to flowering and pollen production. Dispersed rarely from the flower, cotton pollen is registered simply as present or absent. At Ranch Derrio, a reciprocal product for corn is commensurate with botani- •

154

FISH AND DONALDSON

5 • • rn eg

Ca

• co rn cc. < Ni 71: s.0 z •-• a, E E -61 . cu ', 74 ii; :-, c r.,, Ca E -c,:,) C4 cr, -, --CCC o E- • - -' O c. 0 Q ;.; 10 .-1:i - 1.> -Cr.= • c. . _ E .... , '0 .- co co N cn ,. > • c.. .:- - 0 m 00 N- LI. t•-; '0 .0 0 0 \ • G (-,1 71' en G CC; c c c c c c.) n--, ..-n ...., -4 Ca .,.., , ,4-• 4,7 M 0.) '''' .r) 1-8 5 t) ,...= cn.4 Hoej co,,,t 0 - cc., ?...) 5 1... .., c to co to - - - tl:) = 4) t:-: = a , c. • N ON .4C 15 -,-;) 0 z; Û) 2')a) ... ca ON 1.., ..a G.) 4) c.) 0 0 0 -' G .0 4-, C-. eLl -0 1) 01 E° to c I --u H Z 0 a 0 aC •--,Q fa.,0 -8 1 Z 8 :=5 0 ¢ 3 0 0.. .0 .0c_, c. c. _:, ._ a• _ f.)- 0 0 Ln >-.,,, ? ,L) a x = <,) co ay.., • N --. I) 4) 0 cc' 0 :.0 0 4) - C.) 1 :34 z G. 714) G. X 6.. -. : 1 0 sn (-41 tr) 03 0 s- -o ,octl (1 C:1 œl elE ->, . G 0 a ii-.) L.) -a E.-n ,.) cn ., 4 )-9 4c7i N C' 'à' 2? "g_ E E 422. .-..;' 47 :',-., 6 clô 0 E z: >--, c "P., a "- . - 0N : ! ,..,...' -:. •,--4,,,, : 4 <-) .H.....0 -5 > 15 40 V-) CO V) ,.(;) Z z' CC) •,T, C) E H c-› H c4 = 2 ..--' H 7" cu I I a. .6 ,:7 '.à4 4,-4 cq _° a) 0 c .. < N `if, "5 ca4) .7--.' 7; E' ' Z .4 -0a _ c E , cl, c r--: E _ 0 a) 7,4 0 ,.. ---• -._.. H -o .73. L> -0 ° >'a 0 ci'‹) r) Cei Q'' •-• Q- t-- CC> n0 C•1 < '-' . c 0 -o '4,.., g:0". . 0 CC) 8 ° ,4 ',7/' cn Z M îfL 2 g c. 8 .,,,13 cr, r__: 0 ,,q 75. >, `a" r, r.) c. , 0 CL o a E t ...1,..? • c. 0 -0 0 a E -0 E ..'-. cl N 1.)u. c.i 1?I ) - d E 6" czct ' 2 E vl a./ 0C O 3" '6-06 ‘4. 0 E ..., H -.. c .....r-, ,. 0L.) ,.„- 1_, as cct i2. 't , ‘.0 0 4-) I " 'cil c' u el [-.-' cj en 8 p, 8 -a ..,_1 E 5 '8 ,,, ki cc ue, z- 73 CL 0 a) E v al r- • CL C.) .--• a G 'ia' '-_ Q CO -' a ) .r.-; Q n-. 0 a.) ,-, I 0. oo "S i›A . E 0° « ., ". 1it ', ,i F, 7 OQ 0 . -‘.• - ‘'' ' . - ' > • ..0 C. ,..o..c o ., t.) „.., = ,..,.z , z-,- n--' '-- ca '' -. c = cu „_, > c .,c 0 ..cct r4 - = cu '-' -c' 0£1 e' = .-' .0 .. 0 . cr•-. Qu, .9 E c N P. ;e. 0 - -S `u0. "5 2 0' 5 (-D 0 - ...... g mou '6 : t% -0". ) ",0- -II x‘1 I g •ci E-.ez. V.. '2 '5 " ' u 0 > ' ce u ° U "L'; 1.1 -6 (I. < 2 < g 6 >, cz 2 '" °`° Fi/ ci 4 'a 6 8 -6 3 5, 6 `.1.‘ 1 5

Production and Consumption in the Archaeological Record cal, artifactual, and feature distributions. Agave is a highly probable surplus for exchange at that site. Ubiquity of charred remains is equaled only at Muchas Casas, Locus E where use as an architectural material introduces a known bias (Miksicek 1987a:213). Production evidence is found in tabular knives and roasting pits. For site to- tals, knives were reported in two categories (Kisselburg 1987:144). Whether sharp or smooth, concave-edged knives were classified as agave harvesting imple- ments. Straight or convex-edged knives were considered implements for fiber working and combined with spindle whorls and plummets. For the category in which knives can be evaluated separately, numbers recovered per excavated cubic meter are notably highest at Rancho Demo. Although roasting pits in residential sites may have served for multiple re- sources, numbers per 100 square meters of exposure are outstanding at Rancho Demo. The largest roasting feature by several magnitudes, termed an homo, oc- curs here and is unique among the settlements. This feature approaches the size of roasting facilities in agave fields for post-harvest processing (S. Fish and others 1985, 1990). The combined evidence indicates a strong emphasis on raw material production for fiber or edible use relative to the other sites. Artifacts for fiber processing and manufacture at Rancho Derrio, on the other hand, are within the common range. Some of the agave fiber for craft products at the other settlements may have originated at Rancho Derrio; longer distance exchange of such a portable material is also probable. Because spindle whorls could indicate working of cotton as well as agave, finished cotton products may be additionally reflected in artifact totals at some sites. Botanical evidence clearly links primary production of cotton with Muchas Casas loci and particularly with Locus A.

The Regional Level Interpretations of production and consumption at the site level are both strengthened and made comprehensible as economic behavior by reference to community and regional settings (S. Fish and others 1989, 1990). The early Classic period Marana community encompasses 146 square kilometers and a multiplicity of site types and locations. Six zones have been defined with differing subsistence potential and differing settlement evidence for productive activities (Figure 1). The distinctive nature of these zones implies that exchange was an important ele- ment in the economic soundness of the community as a whole. The marginal ag- ricultural potential of much community land lacking supplemental water also should have fostered circulation of more moisture dependent annual crops and drought-adapted ones. Potential fields for the five excavated sites fall within a single lower bajada zone where alluvial fans supported floodwater farming (Fi- gure 2). Adjacent upslope in a second zone, rockpile fields producing agave de- pended solely on surface runoff in one of the driest community environments.

• > •

156

FISH AND DONALDSON

u 7.6 C '1:i f-. C?., 0 ,r) . O 'L) 4' " • 5 ''' .... CD. po = cd . c ,; i._; 47 o. C a) C ..r. 14) c.., ...... 0 .-. •ct. ,--.-0 Z 7,7 E 2 eu 0.0..,.) 0)0va.) 1-. .,.., • . .4 0 C.) Ei. an • = ,...0 00- 4' 7t VD 00 2 ....= = . ,1" 0 CD. CR CR ..cv. - -po *ô -CA co ;-, C.) C cd Ai) a.) .-i= cn aD = C ...... 0 77• .....C • o m .7.; •- a.)k E ;, v v 1...a) cct 0 an cs L.I., a) :74. -v 5 A • 07 0 an N 0., v •c",1 ....-az.. Eco -cr ca ,,a 4-- a) 4 0 va = cd a> .0 64 ... O ,-.c a =??.. mi .Ln ..mi O 0 0 .0 ,1 EC 0 c , >9e E II .6 00 sc, cq a> c E ° • vn •-• Z V- 5 -cii '75 * °° - Cl en > -6 (7' 's'au) 5 a.) 4...4. ,.., CO - o -c N 7 -C ,—.4 NI (1) >4, 0 -CD .m. 0 —4 I • C CO x a) ..0 '—' 0 = . an .a-4 ...t ..0 >CO ' a) . • t 4.. a.) < -cj .... X Z v. cn. r. ,--, 0 a) Co „.. Ti. 's E 2 Z = g E 5 v:D (-4 c-a -t!C c°c 0 II E 41- f cfl-at z c.nas ,-. c) 0.) . .! . E z' CO • OT,-,, ,..,a.) a) ><

""o' ,,,7°0) 0`4 ak :"'ta. 1 ;(L) a 5 . c x cn a.) CL. 5 a) a) = 7-...; ,„ a (.) x cn c..) .4") I.6.,-. CO =...., CO ,... - C II La) < -. en cd 5. -c1 4) c ..,...,,.»,14 . = 7:1 .., Es- .. Ci• O ,c.12 2 • ;__.. 00 CO CO CO ... . •- ,..= ..i.-, 6.-. c., C O. a) Cr —4 .41 --5 ei,2 0.0 .4-. "--• >., 0 M ai •-•• c ,..at .) ..7' 4 ‘A j) ° be r4 o '.7. • -. Cr' C MI = CO CZ 0 il' 1) .1:4 t) t CO c o C -8. a. '0.1 cal> u o 4.'7i o. <4 Il ci) •-.O c ...., a..a..t, t; x . a. o t,-, .r. zi, o...., LI g '2 ' ° ''' 4 Lr.. C.) 1:4 —4)' < ci,>. ,,LI1 ,"e.., -S,,, ci .i c;) •5 .6 4 a..) 157

Production and Consumption in the Archaeological Record

Each site has ready access to the zone hydrologically suited to floodwater farming, but individual locations have differing potential for agricultural success (S. Fish and others 1985; Field 1985; Waters and Field 1986). Alluvial fan condi- tions influencing water availability and successful control become less optimal from south to north. Rancho Bajo is in a very favorable situation, the Muchas Casas loci in intermediate ones, and Rancho Demo in a comparatively poor set- ting. These conditions correlate with progressively lower agricultural weed pollen from a high at Rancho Bajo to a minimum at Rancho Demo and similarly lower flotation values at Rancho Derrio than at Muchas Casas (Table 4). In view of com- parative farming potential, it is reasonable that inhabitants of Rancho Demo would produce less corn as site evidence suggests, and acquire some portion of the dietary total through exchange with more favored sites in the same floodwater zone or from irrigators along the Santa Cruz River. Patterns of primary agave production in relation to the five settlements are unlike those for corn. Environmental opportunity for rockpile fields is virtually unlimited uphill from each, although slopes offer somewhat more mesic condi- tions to the south. The distribution of rockpile fields and their sizes in relation to the sites can be seen in Figures 1 and 2. One large field occurs above Rancho Bajo, a few small ones are above the Muchas Casas loci, and a series of very large con- tiguous ones are near Rancho Demo. Rancho Demo is also distinctive in its more immediate proximity to Zone 2 fields. The placement of this site suggests a greater intensity of involvement in rockpile field cultivation, as did materials from the excavated pro veniences . Investment in rockpile field construction and cultivation can be seen as an economic alternative, the pursuit of which was balanced against potential for floodwater farming. Rancho Bajo possessed the most favorable settings for both kinds of fields. With good floodwater land, inhabitants of this locale chose to con- struct only one large rockpile field, possibly shared with several unstudied sites nearby. Situated in intermediate floodwater lands, Muchas Casas residents culti- vated several small rockpile fields and possibly a part or all of one large complex to the north. With poor floodwater potential, Rancho Demo was located at the edge of a group of large rockpile fields, and its occupants apparently concentrated their productive activities in this context. Differential site evidence for cacti associated Muchas Casas Loci E and D most strongly with cholla and Loci A and D with saguaro. These patterns seem to reflect emphasis prompted by economic decisions regarding gathered resources as well as cultivated ones, although a body of data suggests cholla may have been cultivated by the Hohokam (for example, Fish 1984; Bohrer 1987). Environmen- tal access to cacti is not greater at Muchas Casas. However, heightened emphasis on cacti at these loci may have supplemented secondary production levels of both corn in floodwater fields and agave in rockpile complexes. Densest cacti occur today on slope ridges above the alluvial fans. Rancho Demo is directly adjacent to 158

FISH AND DONALDSON such ridges. Nevertheless, Rancho Derrio inhabitants seem to have concentrated their productive activities on agave without equivalent attention to saguaro and cholla. Regional patterns at a still broader scale enhance understanding of the economic implications of differentiated production in the overall subsistence sys- tem of the early Classic period (Fish 1987b). South of the five excavated sites, ad- ditional settlement in the Marana community is concentrated near the Santa Cruz at the north end of the Tucson Mountains (Figure 1). This area offers good flood- water potential along the lower bajada in proximity to maximum surface flow in the river for irrigated farming. Eastern bajada slopes opposite the Tucson Moun- tains afford opportunities for rockpile fields similar to those above Rancho Derrio, Muchas Casas, and Rancho Bajo, but only one small field was constructed here. Broadscale reciprocal spheres of economic emphasis are suggested by more ex- tensive agave production at northern community sites with access to lesser ag- ricultural water, and greater concentration on moisture-dependent annual crops in southern sites near the river.

CONCLUSIONS The preceding exercise illustrates the challenge of discriminating production and consumption in the Hohokam archaeological record. Even with the rare lux- ury of adequate sampling, quantities of plant taxa may not clarify the role of each process in generating the archaeobotanical assemblage. At the level of structures and other individual proveniences, the greatest potential exists for recognizing the multitude of occupational and post-occupational factors influencing quantities of recovered biological remains. Variable bias, small numbers of items, and unclear association may simultaneously render comparisons of individual features incon- clusive. These problems pertain to artifactual evidence for production and con- sumption as much or sometimes more than to plant remains. At the level of the multi-structure household, consideration can potentially be extended to a larger variety of proveniences. Particularly in the absence of super- positioning or other indications of non-contemporaneity, extramural facilities can be more confidently included. However, only with equally complete exposure of household areas can optimal data sets be assembled. For the component or site, individual provenience bias is "smoothed" and association is less problematic. Infrequent items may be more reliably compared with increased sample numbers. At this level, attributes concerning environmental and cultural settings can be contrasted. If regional settlement patterns are avail- able, site level subsistence data can be translated into a framework of economic relationships. Implicit assumptions in interpreting plant remains from Hohokam and other archaeological sites should be resisted. Equivalent levels of production and con- sumption as well as conclusions concerning imbalance should be supported. The 159

Production and Consumption in the Archaeological Record task of isolating evidence for these two processes and assessing the relative con- tribution of each to the archaeobotanical record is the key to successfully address- ing some of our most interesting archaeological questions.

Acknowledgments. Settlement pattern and subsistence data from the Northern Tucson Basin Sur- vey result from field work supported by the Arizona State Museum, the Arizona State Historic Preser- vation Office (SP104 / 1981 and SP9314/ 1983), the National Science Foundation (BNS-8408141), and the Bureau of Reclamation and Arizona State Land Department (4-CS-30-01280). Arizona State University excavations at Muchas Casas, Rancho Derrio, and Rancho Bajo were funded by the Bureau of Reclamation (5-CA-30-02570).

NOTE 1. Quantities of floor-associated artifacts for the same structure differ in some cases between tables in site summary chapters of Henderson (1987c) and tables in Rice (1987a). When possible, these data were taken from Rice (1987a) because artifact totals for sites are given in the same volume, and it was assumed that quantification procedures would be internally consistent. Ground stone quantities for features and, cumulatively, for sites were available only in Henderson (1987c), however, requiring use of that source for these artifact categories.

REFERENCES Bemard-Shaw, Mary 1987a Descriptive Summary of the Excavation Results at Locus A, AZ AA:12:2 (ASU), Muchas Casas. In Field Investigations at the Marana Community Complex, edited by T. Kathleen Henderson, pp. 61-84. Anthropological Field Studies No. 14. Depart- ment of Anthropology, Arizona State University, Tempe. 1987b Descriptive Summary of the Excavation Results at Locus I, AZ AA:12:3 (ASU), Rancho Demo. In Field Investigations at the Marana Community Complex, edited by T. Kathleen Henderson, pp. 159-176. Anthropological Field Studies No. 14. Depart- ment of Anthropology, Arizona State University, Tempe. Bohrer. Vorsila L. 1987 The Plant Remains from La Ciudad, a Hohokam Site in Phoenix. In La Ciudad: Specialized Studies in the Economy, Environment, and Culture of La Ciudad, Part 3, edited by JoAnn Kissleburg, Glen Rice, and Brenda Shears, pp. 67-178. An- thropological Field Studies No. 20. Department of Anthropology, Arizona State Uni- versity, Tempe. Bostwick. Todd W. 1987 Descriptive Summary of the Excavation Results at Locus D, AZ AA: 12:2 (ASU), Muchas Casas. In Field Investigation at the Marana Community Complex, edited by T. Kathleen Henderson, pp. 85-121. Anthropological Field Studies No. 14. Depart- ment of Anthropology, Arizona State University, Tempe. Dennell, Robin 1974 Botanical Evidence for Prehistoric Crop Processing Activities. Journal of Ar- chaeological Science 1:275-284. 1976 The Economic Importance of Plant Resources Represented on Archaeological Sites. Journal of Archaeological Science 3:229-247. Field, John 1985 Depositional Facies and Hohokam Settlement Patterns on Holocene Alluvial Fans, Northern Tucson Basin. Unpublished Master's thesis, Department of Geosciences, University of Arizona, Tucson. Fish, Paul. Suzanne K Fish, and John Madsen 1984 Northern Tucson Basin Survey: Research Summary and Recommendations for the Marana Complex. In A Supplemental Class III Archaeological Survey of the Phase A, Reach 3 Corridor, Tucson Aqueduct, Central Arizona Project, edited by Job Czap- 160

FISH AND DONALDSON

licki and Adrianne Rankin, pp. 83-90. Archaeological Series No. 165. Arizona State Museum, University of Arizona, Tucson. Fish, Suzanne K. 1984 Agriculture and Subsistence Implications of the Salt-Gila Aqueduct Project Pollen Analyses. In Environment and Subsistence, pp. 111-138. Hohokam Archaeology along the Salt-Gila Aqueduct, Central Arizona Project, vol. 7, edited by Lynn Teague and Patricia Crown. Archaeological Series No. 150. Arizona State Museum, University of Arizona, Tucson. 1987a An Evaluation of Subsistence and Specialization at the Marana Sites. In Studies in the Hohokam Community of Marano, edited by Glen Rice, pp. 235-248. Anthropologi- cal Field Studies No. 15. Department of Anthropology, Arizona State University, Tempe. I987b Marana Pollen Analysis. In Studies in the Hohokam Community of Marana, edited by Glen Rice, pp. 171-170. Anthropological Field Studies No. 15. Department of An- thropology, Arizona State University, Tempe. 1987e Percentage of Pollen Types at the Marana Sites. In Studies in the Hohokam Commun- ity of Marano, edited by Glen Rice, pp. 273-278. Anthropological Field Studies No. 15. Department of Anthropology, Arizona State University, Tempe. Fish, Suzanne K., Paul Fish, and John Madsen 1985 A Preliminary Analysis of Hohokam Settlement and Agriculture in the northern Tuc- son Basin. In Proceedings of the 1983 Hohokam Conference, edited by Alfred E. Dit- tert and Donald E. Dove, pp. 75-100. Occasional Paper No. 2. Phoenix Chapter, Arizona Archaeological Society, Phoenix. 1989 Classic Period Hohokam Community Integration in the Tucson Basin. In Sociopoliti- cal Structure of Prehistoric Southwestern Societies, edited by Steadman Upham, Kent Lightfoot, and Roberta Jewett, pp. 237-268. Westview Press, Boulder. Fish, Suzanne K., Paul Fish, Charles Miksicek, and John Madsen 1985 Prehistoric Agave Cultivation in Southern Arizona. Desert Plants 7:102- 112. Gasser, Robert 1988 Flotation Studies. In Hohokam Settlement along the Slopes of the Picacho Moun- tains: Environment and Subsistence, vol. 5, edited by Donald Weaver, pp. 143-235. Research Paper No. 35. Museum of Northern Arizona, Flagstaff. Hackbarth, Mark I987a Descriptive Summary of the Excavation Results at Locus E, AZ AA:12:2 (ASU). Muchas Casas. In Field Investigations at the Marano Community Complex, edited by T. Kathleen Henderson. pp. 121-150. Anthropological Field Studies No. 14. Depart- ment of Anthropology, Arizona State University, Tempe. I987b Descriptive Summary of the Excavation Results at Locus I. AZ AA:12: I (ASU), Rancho Baja. In Field Investigations at the Marana Community Complex, edited by T. Kathleen Henderson, pp. 177-184. Anthropological Field Studies No. 14. Depart- ment of Anthropology, Arizona State University, Tempe. Hally, David 1981 Plant Preservation and the Content of Paleobotanical Samples: A Case Study. Ameri- can Antiquity 46:732-742. Hastorf, Christine 1988 The Use of Paleoethnobotanical Data in Prehistoric Studies of Crop Production, Pro- cessing, and Consumption. In Current Paleoethnobotany: Analytical Methods and Cultural Interpretations of Archaeological Plant Remains, edited by Christine A. Hastorf and Popper, pp. 119-144. University of Chicago Press, Chicago. Henderson, T. Kathleen 1987a Ceramics, Dates, and the Growth of the Marana Community Complex. In Studies in the Hohokam Community of Marano, edited by Glen Rice, pp. 49-78. Anthropologi- cal Field Studies No. 15. Department of Anthropology, Arizona State University, Tempe. 1987b Excavation Procedures and Field Summary. In Field Investigations at the Marano Community Complex, edited by T. Kathleen Henderson, pp. 49-60. Anthropological 161

Production and Consumption in the Archaeological Record

Field Studies No. 14. Department of Anthropology, Arizona State University, Tempe. I987c Field Investigations at the Marana Community Complex. Anthropological Field Studies No. 14. Department of Anthropology, Arizona State University, Tempe. Hillman, Gordon C. 1981 Reconstructing Crop Husbandry Practices from Charred Remains of Crops. In Farm- ing Practice in British Prehistory, edited by R. Mercer, pp. 123-162. Edinburgh Uni- versity Press, Edinburgh. 1984 Interpretation of Archaeological Plant Remains: The Application of Ethnographic Models from Turkey. In Plants and Ancient Man: Studies in Paleoethnobotany, edited by W. Van Ziest and W. Casparie, pp. 1-41. A.A. Balkema, Rotterdam. Kisselburg, Jo Ann 1987 Economic Specialization in the Community System at Marana. In Studies in the Hohokam Community of Marana, edited by Glen Rice, pp. 143-160. Anthropologi- cal Field Studies No. 15. Department of Anthropology, Arizona State University, Tempe. Miksicek, Charles 1987a Late Sedentary-Early Classic Period Hohokam Agriculture: Plant Remains from the Marana Community Complex. In Studies in the Hohokam Community of Marano, edited by Glen Rice, pp. 197-216. Anthropological Field Studies No. 15. Department of Anthropology, Arizona State University, Tempe. 1987b Plant Remains at the Marana Sites. In Studies in the Hohokam Community Marana, edited by Glen Rice, pp. 287-309. Anthropological Field Studies No. 15. Department of Anthropology, Arizona State University, Tempe. Rice, Glen 1987a Floor Assemblages of the Marana Community Complex. In Studies in the Hohokam Community of Marano, edited by Glen Rice, pp. 79-162. Anthropological Field Studies No. 15. Department of Anthropology, Arizona State University, Tempe. 1987b The Marana Community Complex: A Twelfth Century Hohokam Chiefdom. In Studies in the Hohokam Community of Marano, edited by Glen Rice, pp. 249-250. Anthropological Field Studies No. 15. Department of Anthropology. Arizona State University. Tempe. 1987e Studies in the Hohokam Community of Marano. Anthropological Field Studies No. 15. Department of Anthropology, Arizona State University, Tempe. Russell, Frank 1975 The Pima Indians. Re-edition with Introduction, Citation Sources, and Bibliography by Bernard L. Fontana. The University of Arizona Press, Tucson. Originally pub- lished 1908. Smithsonian Institution, Washington, D.C. Waters, Michael, and John Field 1986 Geomorphic Analysis of Hohokam Settlement Pattem on Alluvial Fans along the Western Flank of the Tortolita Mountains, Arizona. Geoarchaeology I: 329-345. 162

APPENDIX E

PARAMETERS OF AGRICULTURAL PRODUCTION

IN THE NORTHERN TUCSON BASIN

by Suzanne K. Fish, Paul R. Fish, and John H. Madsen

1992. In "The Marana Community in the Hohokam World," edited by Suzanne K. Fish, Paul R. Fish, and John H. Madsen, pp. 41-51. Anthropological Papers of the University of Arizona No. 56. The University of Arizona Press, Tucson. 163

ANTHROPOLOGICAL PAPERS OF THE UNIVERSITY OF ARIZONA NUMBER 56

The Marana Community in the Hohokam World

Suzanne K. Fish, Paul R. Fish, and John H. Madsen, Editors

CONTRIBUTORS

John J. Field Paul R. Fish Suzanne K. Fish John H. Madsen Charles H. Miksicek James M. Skibo Alan R Sullivan III Christine R. Szuter Mary Van Buren

THE UNIVERSITY OF ARIZONA PRESS TUCSON 1 992 164

About the Editors Contributors

SUZANNE K. Fist is a Research Archaeologist at the Arizona John J. Field Stale Museum, University of Arizona. She has specialized in Department of Geosciences, University of Arizona, lùcson I lohokam archaeology for more than ten years; other areas of professional emphasis include ethnobotany, settlement patterns, Charles H. Miksicek and Mesoamerican prehistory. Recent publications include 13ioSystems Analysis, Inc., Santa Cruz, California articles and books on the advantages of full-coverage archaeo- logical survey, subsistence issues, demography, and 1 lohokam James M. Skitx) cultural ecology and social organisai Department of Anthropology, University of Arizona, Tucson

PAUL R. FISH is Curator of Archaeology at the Arizona State Alan P. Sullivan Museum and Research Professor in the Department of Anthro- Department of Anthropology, University of Cincinnati, Ohio pology, University of Arizona. I lis research focuses on the pre- history of the deserts in southern Arizona and northern Mexico. Christine R. Szutcr Recent publications address archaeological classification, settle- University of Arizona Prms, Tucson ment patterns, political and social organization, factors influenc- ing regional abandonment, and cultural resource management. Mary Van Buren Colorado State University, Fort Collins JOIN H. MADSEN is a Research Archaeologist at the Arizona State Museum, University of Arizona. Over the past twelve years he has worked closely with the Arizona State Land Department as a cultural resource management advisor. His research interests include prehistoric and historic land use in the Sonoran Desert. In recent publications he has written atxxit prehistoric agave use, raw material procurement strategics associated with stone tool manufacturing, full-coverage archaeological survey, and Hohokam political and social organization.

Cover Early Classic period Hohokam settlements in the Manna survey area north of 'Ilicsott, Arizona (see Figure 3.2).

uNtvERsrry OF ARIZONA

Copyright 1992

The Arizona Board of Regents All Rights Reserved

This book was set in 10/12 Dutch Roman TIns bc)ok is printed on acid-free, archival-quality paper. Manufactured in the United Slates of America.

96 95 94 93 92 6 5 4 3 2

Library of Congress Cata-loging-in-Publication Data

The Manna Community in the Hohokam world /Suzanne K. Fish, Paul R. Fish, and John H. Madsen, editors : contributors, John J. Field ..4e1 ail. p. cm. — (Anthropological papers of the University of Arizona; no. 56) Includes bibliographical references and index. ISBN 0-8165-1314-7 (acid-free paper) 1. Hohokam culture—Social conditions. 2- Hohokam culture— Irrigation. 3. Hohokam culture—Agriculture. 4. Social archaeology— Santa Cruz River Basin (Ariz. and Mexico). S. Santa Cruz River Basin (Ariz. and Mmico)—Antiquities. 6. Arizona—Antiquities. I. Fish, Suzanne, K. IL Fish, Paul R. III. Madsen, John It. (John Henry), 1947 — . IV. Field, John J. 1961 — . V. Series. E99.1-168M35 1992 979. I 7901--dc20 92-8510 CII' 165

THE UNIVERSITY OF

The University of Arizona Press ARIZONA 1230 N. Park Avenue, Suite 102 TUCSON ARIZONA Tucson, Arizona 85719-4140 (602) 621-1441

July 6, 1993

Suzanne K. Fish 2214 E. 4th St. Tucson, AZ 85719

Dear Suzanne Fish:

Thank you for your letter of June 26 requesting permission to reproduce your chapter, "Parameters of Agricultural Production in the Northern Tucson Basin," in The Marana Community in the Hohokam World by Fish, Fish and Madsen in your Ph.D. dissertation for The University of Arizona's Arid Lands Studies Program and in a reprint by University Microfilms.

We are pleased to grant the permission contingent on full credit to the authors, the book title, the series name and number, The University of Arizona Press and copyright 1992.

Thank you for your interest in our publication.

Sincerely yours,

Robyn Forkos Editorial Assistant 166

CHAPTER FOUR

Parameters of Agricultural Production in the Northern Tucson Basin

Suzanne K. Fish, Paul R. Fish, and John H. Madsen

As a microcosm of environmental variation in the and storm-fed watersheds of bajadas. The hydrologic portion of the Sonoran Desert inhabited by the system includes surface flow and a second component of Hohokam, the 'Meson Basin is an ideal location for less rapidly mobile groundwater. Distributions of investigating a diversified but integrated system of agricultural remains across the varied zones of the Hohokam agricultural production. Regional demography, Marana Community offer a uniquely comprehensive tes- settlement, social organization, and exchange were tament to past productive relationships between noncore shaped by specific resource needs, technological capa- Hohokam and their desert basin environments. bilities, and environmental potential for production in an agricultural economy. The productive landscape of the CONSULTATION WITH TRADITIONAL Marana Community and of the larger study area can be FARMERS examined in detail by reference to environmental vari- ables, settlement pattern, and agricultural features, and Ethnographic practices of traditional farmers in by comparison with the practices of traditional farmers environments similar to the study area furnish insights in historic times. into natural and cultural factors influencing agricultural A number of excavation and survey reports document production. Piman analogies have been applied to the localized occurrences of Hohokam agricultural features interpretation of llicson archaeological patterns, particu- and complexes. At a higher synthetic level, extensive larly with reference to floodwater farming or the diver- canal networks have been analyzed by integrating historic sion of short-term flows in ephemeral drainages (Wilson records, aerial photography, and settlement data (for 1985; Field 1985, Chapter 5 in this volume; Waters and example, Haury 1976; Masse 1981, 1987, 1991; Nicholas Field 1986). However, historic Piman floodwater farming and Neitzel 1984; Nials and others 1989; Ackerly and (ak chin farming) is a method with few remaining practi- others 1987). Combinations of canal irrigation and runoff tioners on isolated remnants of previously productive devices also have been described for a few substantial acreage (Reichhardt and Nabhan 1982; Nabhan 1986a, study areas (Gumerman and Johnson 1971; Crown 1987; 1986b). Although well documented, the ethnographic Doyel 1984). However, in none of these cases has a instances do not duplicate prehistoric situations that systematic regional sample of non-canal features been involved higher population densities and, necessarily, available. Archaeologists attempting to deal with agri- more intensive land use. Other prehistoric techniques cultural patterns at this scale (Masse 1979, 1991; Doyel such as the northern basin forms of riverine irrigation 1984) have had to interpolate from geographically and cultivation in rockpile fields lack analogs altogether. restricted and noncomparable survey findings. ltaditional farmers from two cultural and agricultural The most important qualification of the northern backgrounds in the Sonoran Desert were consulted in 'Meson Basin for a regional perspective is the presence order to include multiple approaches, to broaden the of undisturbed land in all environmental zones. Granitic range of associated population and land use intensity, mountain ranges on the east reach approximately 1325 and to explore agricultural alternatives and decision m (4300 feet). 'lb the west, rough basalt peaks form a making. Visits to the study area by a traditional Tbhono lower chain mostly below 860 m (2700 feet). Elevational O'odham (Papago) farmer from an area southwest of diversity is repeated on opposing sides of the basin. Tbcson and a farmer from Cucurpe in northern Sonora, Major and minor Santa Cruz River tributaries carry Mexico, were arranged with the aid of Gary Nabhan of runoff from both orographic rainfall in the mountains the Phoenix Botanical Garden, who participated in all 167 Suzanne K Fish, Paul R. Fish, and John H. Madsen

Table 4.1. Annual Precipitation Measures for Stations in the Tucson Basin and Other Hohokam Subareas

Elevation Mean annual Range of annual Percent (ASL) Years of precipitation precipitation Variability of years Station (m) (feet) record (mm) (inches) (mm) (inches) index > 10 inches

Gila Bend 275 900 77 145 5.8 52-340 2.1-13.6 34.3 7 Phoenix Airport 335 1100 34 175 7.1 60-335 2.4-13.4 29.1 16 Sacaton 365 1200 72 195 7.9 47-392 1.9-15.7 27.9 26 Florence 460 1500 69 250 9.9 103-470 4.1-18.8 24.6 27 Casa Grande National Monument 425 1400 59 220 8.8 97-480 3.9-19.2 27.1 26 Cortaro (Marana Study Area) 670 2200 27 280 11.2 158-463 6.3-18.5 23.9 65 Tucson, University of Arizona 700 2300 82 275 11.1 143-408 5.7-16.3 23.7 60

Sources: Range of annual precipitation from Green and Sellers 1964; other data from Sellers and others 1985.

phases of consultation. Both individuals engage in abundance and predictability of local rainfall corres- variants of floodwater farming, and the Cucurpe farmer ponding to differential risks and returns for fields more irrigates with gravity-fed canals from a small Sonoran directly dependent on this source for water than irrigated river. Evaluations and comments by these consultants are land (Tables 4.1, 4.2; Fig. 4.1). Substantial investment in cited in relevant sections of this chapter. such fields in the lbcson Basin may be explained as much by comparatively favorable precipitation regimes as by fewer possibilities for riverine canals. AND REGIONAL PRECIPITATION Mean annual precipitation for the Marana study area POTENTIAL at 280 mm (11.2 inches) is among the highest for basin Ilicson Basin agriculture must be understood in the interiors in all of Hohokam territory. For agriculturalists context of local risks and opportunities. The potential in arid environments, annual variability may be as critical for riverine canals is lower than along the Salt and Gila as average amounts. lhble 4.1 confirms that higher rivers near Phoenix, based on volume and duration of annual "lbcson precipitation corresponds with lower flow in the Santa Cruz River, terrace morphology, and variability, as measured by departures from the long-term extent of topographically irrigable basin floor. Alterna- mean. Likewise, precipitation minima in dry years are tive prehistoric technologies for securing agricultural higher in the northern llicson Basin than at other sta- water were concomitantly more prominent. Precipitation tions. Tlicson riverine irrigators as well as farmers using measures for Hohokam subareas reveal variation in the channelized floodwater and overland flow depended on relatively abundant local precipitation. Because upland snow melt is too limited to generate high flows in the

Table 4.2. July, August, and September Precipitation for Santa Cruz and its tributaries during the spring and Stations in the Tucson Basin and Other Hohokam Subareas because summer precipitation predominates in south- central Arizona, rainfall in this season is the major factor Mean no. of in total agricultural production. As in the case of annual summer days amounts, summer precipitation for the Tucson Basin is Mean summer with > 0.25

Years of precipitation inches precip- more favorable than for much of the territory inhabited

Station record (mm) (inches) itation by the Hohokam (Table 4.2).

52 2.1 2 Gila Bend 77 AGRICULTURAL ATTRIBUTES Phoenix Airport 34 57 2.3 3 Sacaton 72 90 3.6 5 OF BASIN SETTINGS Florence 69 87 3.5 5 Casa Grande Mountain Slopes National Monument 59 77 3.1 3 Cortaro At higher elevations, orographic rainfall delivers a (Marana Study Area) 27 135 5.4 7 more abundant and predictable water supply for crops Tucson dependent on direct precipitation than at lower eleva- University of Arizona 82 145 5.8 6 tions. The Tucson Mountains on the west edge of the Source: Sellers and others 1985 Tucson Basin (Zone 6 of the Marana Community) exert 168

Parameters of Agricultural Production

57.—

-

AR IZONA

- 35 . —

- -

• Phoenix Gila Bend •

• • Florence - 33° — SOCOtOrl • Casa Grande Notional Monument

Cortara•

Tucson• -

, 3/.—

Figure 4.1. Locations of weather stations in Tables 4.1 and 4.2. lesser influence on precipitation patterns than the more between first and last freezes for a winter averaged 36 massive ranges on the east (Figs. 3.8, 3.9). However, the days on the hill and 157 days below. Low temperatures benefits of orographic rainfall are balanced against two were of shorter duration on the hill as well. These eleva- drawbacks of mountain slope cultivation. Parcels of suit- tional contrasts would he critical for spring crops planted able soil depth and flatness are restricted. At altitudes sufficiently early to benefit from winter rains. above 925 m (3000 feet), frost hazards increase, although 1vo mountain slope situations appear most advanta- lower mountain flanks may be warmer than valley floors geous for agricultural pursuits. One occurs on the flat exposed to cold air drainage. land of ridge tops or plateaus at higher elevations where In the Ilicson Basin, temperature inversion has signif- more abundant rainfall is available. Alternately, gentle icant elevational consequences for early or late season slopes at intermediate elevations with less orographic planting. The impact of cold air flowing down slope is rainfall have less frost, better soil accumulations, and illustrated by five years of weather records for a station offer opportunities for water catchment and concentra- on the shoulder of a hill in the 'Dimon Mountains and tion through such features as terraces and checkdams. for a second station on the basin floor below, near the The lbrtolita mountain slopes have been covered less Santa Cruz River (Hastings and Minter 1965: 17). The fully in survey than other zones. The best coverage is in stations are separated by a horizontal distance of only 0.8 the northeast sector of the study area (Fig. 3.2). Here, a km (0.5 mile) and by a vertical distance of only 100 m number of habitation sites are located away from can- (330 feet). A difference of 10.1° C (20° F) was recorded yons on flatter land at elevations as high as 1128 m on some nights. Over the five-year period there was a (3700 feet). It seems probable that orographie rainfall total of only 38 freezing nights on the hill in contrast to and relatively small catchments supplied water for agri- 263 nights on the floodplain. The frost-danger period culture. Survey transects (Hewitt and Stephen 1981) 169

Suzanne K Fish, Paul R. Fish, and John H. Madsen

located agricultural terraces near sites at still higher Tort°lita canyon sites tend to be located along the elevations on the eastern Tbrtolita slopes facing into the floodplain edges, no doubt above the contour of most adjoining basin. floods. The only recorded habitation sites are small.

The -Meson Mountains are much less massive than Farming in the canyon bottoms may often have been the lbrtolitas and were fully surveyed. The relative carried out on a daily basis by inhabitants of settlements sparsity of sites in the Thcson Mountains (Zone 6) may on the adjacent upper bajada. Deeper soils are typically be related to preferred habitation on the immediately associated with shallow, meandering channel segments adjacent Santa Cruz floodplain (Zone 5), but potential from which water could be diverted. It is likely that use of western mountain land also seems limited. lbr- diversion structures, agricultural features, and perhaps raced slope sites in southern Arizona of the type found even habitation remains on canyon floodplains have been in the Tlicson Mountains correlate almost exclusively removed or buried by the larger floods occurring be- with the kind of dark volcanic substrates occurring in tween prehistoric and modern times. The absence of sites this range; locations of such sites in the Marana Com- in canyons of the Cochie and Cottonwood drainages munity additionally coincide with large populations on (Fig. 3.2) reflects survey coverage rather than a departure the adjoining floodplain. from the general distributional pattern in the northern

Four clusters of stone terraces or trincheras features -Meson Basin. were constructed in the 'Meson Mountains within the study area. The two largest coincide with dense occupa- Upper Bajadas tions around the northern end of the mountain chain. Many of these terraces appear to have been agricultural West of the river in the southern community (Zone in function, and corn pollen has been recovered from 5), the width of the bajada is greatly compressed between several (S. Fish, P. Fish, and Downum 1984). The agri- the llicson Mountains and the Santa Cruz, increasingly cultural benefits of terraces on mountain slopes include so toward the northern end of those mountains (Fig. water concentrated from slope catchments above and 3.9). Relatively few sites have been recorded on the slope between terraces, protection from freezing winter tem- below these mountains and only small ones with evi- peratures threatening lower elevations through inversion, dence for habitation are located here. Farmers with and relief from high summer temperatures by preferen- floodwater fields along the mountain edge may have

tial use of north and east exposures. 11- incheras culti- resided in nearby settlements along the river. vation probably ranged from kitchen gardens around the Ilvo situations for agriculture predominate on the houses built on some terraces to clusters of features of upper bajada below the Tintolitas (Zone 4), where sub- solely agricultural function (Downum and others 1985; stantial habitation sites of all periods are found. The first Downum 1991) as described for a large site in the involves areas between major trans-slope washes that Robles Community to the northeast (Fig. 1.8). originate in the mountains. Secondary and less incised drainages of various sizes in these areas also carry water from more localized upland watersheds. Precipitation at Canyon Bottoms the base of the mountain front, heightened by the initial Tbrtolita canyon bottoms represent a topographic uplift of moisture-laden air passing over the peaks, adds class related to mountain slopes, but with distinctive to flow in the smaller streams. Numerous instances of agricultural potential. Canyons carry streams with large low-density but extensive arrays of simple stone align- uphill watersheds. Effluent water continues to be dis- ments indicate the agricultural use of overland runoff in charged into these stream beds for extended periods and broad, gentle swales as well as water diverted from supports perennial flow in some cases. The quantity of channels. effluent stream flow may be reduced in the summer, lb the southeast of the Marana Community, a dis- however (Chapter 5). Flat bottomland with adequate soil persed pattern of small habitations and other sites on the depth is often quite narrow. Since high-energy floods bajada occurs in the vicinity of Canada Agua (Fig. 3.9). occur, smaller tributaries to main canyon drainages may These sites are primarily oriented toward moderate-sized have been easier to divert in some cases. Cold air, which drainages. Farther from the mountains here, slopes flat- flows down canyons, represents one environmental draw- ten over deeper valley fill, and many smaller channels back for early crops. Wider canyon bottomland in the become shallow, braided, and easier to divert (Fig. 4.2). lbrtolitas is atociated with a number of settlements in A continuous scatter of isolated artifacts corresponds to uppermost Zone 4 of the Marana Community in contrast the vicinity of braided channels even where sites arc to the lluson Mountains (Zone 6), where canyons arc lacking and appears to represent remains resulting from smaller and drain restricted slope areas. farming activities. Sites at the southeastern edge of this 170 Parameters of Agricultural Production

Figure 4.2. Braided channel of a drainage on the upper bajada downslope from the Tortolita Mountain pediment in the Marana survey area. (Photograph by Helga leiwes.) area are located outside the Marana Community and are upper bajada. Contours of approximately 875 m (2700 affiliated with Preclassic or Classic period communities feet) delimit the southern sites and the same elevation with centers farther east. bounds similar sites within the Marana Community. It is difficult to assess the relative agricultural reliance Agricultural parameters probably figure in this pattern. on primary and secondary drainages on the upper bajada Near the mountain front, sediment over pediment bed- within the Marana Community. In the Classic period and rock is typically no deeper than a few meters; water earlier, both large and small sites tend to cluster along tables in drainages are correspondingly close to the the major Cottonwood, Derrio, and Guild washes. surface. As drainages continue downhill across the Smaller habitation sites are located between these on bajada, surface flow tends to diminish or disappear in secondary channels, but distances would have permitted channels through infiltration into increasingly deep valley additional land in such situations to be cultivated by fill. The lower elevational limit of sites on small drain- residents of settlements along the major watercourses. ages likely marks the downslope extent of significant sur- Near the upper reaches of Guild Wash as it leaves the face flow from all but the largest precipitation events. Ibriolitas, the placement of dispersed small habitations Another factor in agricultural production involves the on minor drainages resembles site locations about opportunity for early cropping at elevations above the Canada Agua. Agricultural features such as rockpiles and level of cold air drainage. Winter frontal systems pass alignments are also concentrated here. Both within the from the west and north toward the east across southern Marana Community and to the south near Canada Agua, Arizona. Therefore, the greatest precipitation benefits remains oriented toward secondary drainages extend a from orographic uplift are experienced during this season limited distance beyond the mountain front onto the along the western edge of the -Ibrtolitas. In most years, 171

Suzanne K Fish, Paul R. Fish, and John H. Madsen

winter rains cease well before May. The ensuing fore- Table 4.3 Mountain Watershed Sizes for Trans -balada summer drought lasts under high temperatures until July, Drainages In the Marana Survey Area preventing satisfactory maturation of mid to late spring plantings. lb effectively use winter moisture, the earliest Watershed Area Drainage (square km) (square miles) possible planting date would have been necessary. A relationship between upper bajada settlement near small Canada Agua 4.22 1.63 drainages and the elevational limits of inversion is illus- Guild 6.76 2.61 trated by the fact that urban smog can be observed to Ruelas 8.26 3.19 Cochie hover in the valley bottom just below these sites. Crops 14.06 5.43 Wild Burro 17.95 6.93 may have been planted sufficiently early on these warmer Cottonwood 18.78 7.25 slopes for spring harvests as well as summer ones. Derrio 27.07 10.45 Water in the large trans-bajada washes supported the Prepared by Matis Myhrman second major farming orientation in Zone 4 on the upper bajada. Bottomland with high agricultural poten- tial is not evenly distributed along the major washes, but for water diversion from shallow channels onto fields at varies with factors such as width and morphology of the the side above flood limits. Tlenching in Derrio and floodplain, water table depth, watershed size, and drain- Cottonwood floodplains (Chapter 5) revealed intact age gradient. The importance of such acreage for sup- features such as hearths in these floodplain edge situa- porting relatively dense populations is indicated by the tions at depths less than a meter. The greater upland locations of large habitation sites along those stretches watersheds of Derrio and Cottonwood (Table 4.3) may of the major drainages suitable for floodplain fields. have supplied effluent flow from winter rains into the A proliferation of both large and small settlements on spring season as far downslope as these large sites or Guild, Derrio, and Cottonwood washes occurred during within short distances from them. Peak flow from the Classic period. lb some extent, this density may summer rains in the two drainages would also reflect reflect cultural preference: the desire of expanding pop- watershed size through frequency and volume of flow. ulations to locate new habitations within community Derrio and Cottonwood floodplains were judged to be boundaries and near sites of origin. However, the largest agriculturally desirable by the traditional farming consul- upper bajada sites of Preclassic times also occur here, tants on the basis of vegetation indicators. The Mexican indicating a long-term productive advantage. informant noted the size of mesquite trees. The Tohono Compared to Derrio and Cottonwood drainages, O'odham farmer additionally commented on the intense Guild, Canada Agua, Ruelas, Wild Burro, and Cochie green color of palo verde trees, the presence of catclaw washes are characterized by small watersheds (Table 4.3), (Acacia greggiz), and a potential for hand-dug floodplain lesser volume and frequency of flow, and narrower flood- wells. In situations resembling Derrio and Cottonwood plains. The greater length of Guild than these others floodplains, highly productive fields are still cultivated along the mountain front, where bedrock depth is shal- today along major tributaries in the Rio Sonora valley of lowest, and its course across a gentler slope likely Mexico (Doolittle 1984: 124-135). Series of adjacent account for larger sites along the upper reaches. Canada fields share diverted water from canals of moderate Agua, with only one large site, and Guild, with numerous length. Such fields are also cultivated within tributaries large and small ones, traverse less steep portions of the of the San Miguel River in Sonora by means of short upper bajada on which water diversion and control of diversion ditches and floodwater techniques (Nabhan and flow was easier. Evidence of the strength of flow in Wild Sheridan 1977). Burro VVash was considered disadvantageous by the Tbhono O'odham and Mexican traditional farmers. Wild Middle Bajadas Burro, Ruclas, and coche have steeper courses, and sites along them extend for only short distances beyond Middle reaches of the valley slope, representing large the mountains. areas in Zones 2 and 3 (Figs. 3.8, 3.9) on the eastern Compared to those washes, Derrio and Cottonwood side of the basin, were not a significant factor in agricul- cross flatter topography and are lined by clusters of large tural production prior to the Classic period. Drinking sites farther downhill on the upper bajada. Bedrock vis- water is inconveniently distant at the river or above at ible in bank cuts indicates a favorably high water table. the mountain edge. Potential agricultural water appears These large sites coincide with segments of wide, arable as brief flows in the major trans-bajada drainages follow- floodplain and lush riparian vegetation within the wash ing only the largest storms. Water from lesser precipi- bottoms (Fig. 4.3). Width of the floodplains is sufficient tation events typically generates local floods, which are 172

Parameters of Agricultural Production

Figure 43. Segment of Derrio Wash with arable bottomland on the upper bajada in the Marana survey area. (Photograph by Helga Tbiwes.)

not sustained over long distances and infiltrate channels Lower Bajadas over deep valley fill. These drainages are incised beyond a depth for easy diversion even when water is available. Alluvial fans composed of outwash sediment from the Many small drainages with bajada rather than moun- uplands coalesce on the lower bajada in Zone 1 on the tain catchments are sufficiently shallow for successful east and Zone 6 on the west (Figs. 3.8, 3.9). In portions diversion. However, such water would have been avail- nearer the floodplain, gentle slopes providing an active able only in cases of thunderstorms directly over the depositional environment and controllable water flow watershed, a relatively unpredictable event compared to were favored by cultivators of every period. In these situ- higher elevation precipitation triggered by uplift of air ations, floodwaters following storms provided both mois- over the mountains. The use of small drainages by means ture and simultaneous enrichment for crops in the form of earthen checkdams in Zone 2, which likely was of suspended nutrients and organic detritus. Dispersed attempted only in seasons of more promising rainfall, settlements of farmers rather than the remains of agri- correlates exclusively with the Classic period prolifera- cultural activities register the reliance on lower bajada tion of an agricultural technology fed by surface runoff. cultivation. Analogy with historic Sonoran Desert culti- Simple mulches of piled cobbles, or rockpiles, enhanced vators suggests brush, earth, and stone diversion struc- and conserved soil moisture for drought-adapted crops of tures on watercourses (Fig. 4.4), intrafield constructions agave in vast fields (see Chapter 7). Overland runoff and of similar materials for water distribution (Fig. 4.5), and direct rainfall were the sole water sources on gentle ditches or canals of moderate length (Fig. 4.6), all of middle bajadas that were too marginal for annual crops which would rarely leave evidence in the archaeological such as corn, beans, and squash. record.

173

Suzanne K Fish, Paul R. Fish, and John H. Madsen

"4,,,qcs„,:• -„ .1,-, --• - :„.., 4...1reif • • n .: •r• •••••••••%"••••-• -f-'-'...... Ti•t•:;„.••_1•.;•„„. 1., ;; ..2, ..,." .',1=',..;.„„••,1,,-;.••• ••,...... „„,.._,.., •nn=it,.....„,„ —•-•-....•0•••• " 1 -•••••••••Apt.••••• ,.. • s•-.1"::-1.•.•-••••14.•14.•-•••• 47 Cr• --• 'n•' "*4111.r - lk • •,.1,„•;,-. - .....- .0'. ....• ,.„,,,:o— • • ,• ..,,„:•...-:..,••:---.-r;-...±....„--• - •••.14•• ...... 1.•...,•„;;...... „„,..-4,47„, rat,,r--_,.. . - ..4.&,„.„,,_ ...... --; • -----1, :.- ' —`71.f•z- --',"-- -it';W"' 'rt.-4 hP„A:, - ---,,k..e.r.,,tr:44.r:...... ;*:.`,ve--... -zii....^.._•-,7%.:...----(.5;,;,",-13`4.4".:4. - ;7- ;-..-. :;"7-1,-..:1:: :....".aii7;:'2:'--.::- -T'; .t. . -7.-4,...,, ... : : it,....tw.. ,....42.-....7...,2L.,—•• : „.,.4,41. -;.;:r 4.,t1t.:7;,....,..,„:„.fil.a.:•±, ...... -..i....., - ,. wi•rx;,,,--.-Q::, t ••••"1":,.....,;,45%,-;...... ,-,.. „, , ,=.t. .sr.„-.._, r;._., -:„. „,...Na••• ..--.N.•-•-1•1';;;;;„*" rle.',.••;4..._. _ 's%,11 1,:-;.••?;7•-7* 5..... • . •••„;.,:.?••••,,., - • •.;=,. - • tersro,... — .41"."' '-'" •.1 , ....." , 1-. , • • ...... s. . •...„.„ •,„0„...-1,' • , ...... ---TZ6,„•-.,--A.1.- • .v. , -• :-...rs.:,,, ••••:•• •,` • • -- , ...z.„.„ • ..... ,7;,..,._,;`,...2,--J44%,- • .-.-,.• ••,„!•:„.i.k...., . - -•-• •• n•• - _ `W•••tr• .1-..1.4.4.:.- .rt5..- q" .1614- : ::•-** '•n• . -....41 •••Le‘illal • • .. ''-'• • .-..^. , •r•••.W. X•f , ..:: .._ ..._ir.sove _. :„.'?f,„1:7„r .. .1t .,,z _ .

Figure 4.4. Water diversion structure on a tributary of the Rio Sonora near Baviacora, Sonora.

Figure 4.5. Earthen embankment for intraficrd distribution of water along a tributary of the Rio Sonora near Baviacora, Sonora. 174

Parameters of Agricultural Production

Figure 4.6. Canal carrying diverted floodwaters to multiple fields along a tnbutary of the Rio Sonora near Baviacora, Sonora.

Tbhono O'odham ak chin or floodwater farming in mountain front, also correlated with fan size, as in Zone late historic time has provided the foremost model for 1 further enhances the accessibility of agricultural water. understanding strategies of land use on alluvial fans. The However, these advantages on small fans, as in the ideal location of fields coincides with down-fan positions Tlicson Mountains of Zone 6 must be balanced against in which floodwater overflows increasingly shallow chan- the lower chance for thunderstorms over watersheds of nels and spreads laterally over the fan surface, delivering more limited extent. water of sufficiently low force to avoid disruption of Late nineteenth century and more recent observations plantings. The closer the correspondence between these emphasize lbhono O'odham field placement with re- conditions and field location, the less labor investment spect to points of natural water spreading (Bryan 1925, would be necessary to divert water and construct ditches 1929), but ethnographic accounts recall more intensive or protective barriers. Locations of optimal conditions past practices. Constructions are reported for runoff shift on fans over time as hydrological activity changes in concentration in watershed areas upsiope from fields and response to continuing geomorphological processes. for water delivery from drainages to conjoined series of Geomorphological variables affecting floodwater fields (Underhill 1939; Castetter and Bell 1942; Clous farming in the study area are discussed by Field (1985; 1915, 1917). Collective efforts were involved in con- Chapter 5 in this volume), Waters (1987, 1988), and structing canals up to I km (0.5 mile) in length and even Waters and Field (1986). Advantages of small fans over longer walls to divert runoff to plots or reservoirs. large ones include higher proportions of fine-grained Nabhan (1986a, 1986b) notes that such efforts suggest a soil, lower thresholds of overbank flow, and less erosive broader range of former field locations than at loci of force of flow when it occurs. Greater distance from the natural water spreading, and he documents more varied 175

Suzanne K Fish, Paul R Fish, and John H. Madsen situations of late historic lbhono O'odham cultivation. in Zone 1 to the north of Wild Burro Wash. Observa- Downcutting and channel incision of major washes in tions of sediment deposition and organic flood detritus southern Arizona in the late 1800s may have prevented that were made over several years document frequent their diversion, increasing subsequent use of smaller flow in the larger secondary drainages of this area. This tributaries and watersheds with less predictable flood- kind of evidence and vegetational indications convinced water flow (Nabhan 1986a: 74). Higher densities of both traditional farming consultants that this zone was farmers in prehistoric times and greater dependence on among the most desirable settings for farming in the their own agricultural yields may have occasioned other Marana Community. A second likely factor in concen- departures from idealized 'Mono O'odham practices of trated settlement of the lower bajada is the prolonged the late historic period. availability of domestic water. Portions of Zone 1 with Possible evidence for more intensive methods of water substantial early settlement are those close to the high management on alluvial fans has been found in excavated water table along the Santa Cruz near the end of the cross sections of buried drainages. "'Tenches dug on the Tlicson Mountains. lower bajadas in Zone 1 of the Marana Community Irrigation by canal from the river may have (Katzer and Schuster 1984) intersected several secondary augmented agricultural production on the eastern channels that differ in morphology from current natural segments of the lower bajada. Thrrace height diminishes drainages and that may have been constructed to direct and width of the floodplain increases rapidly downstream overland flow. Compared with modern drainages, these from the Tlicson Mountains, topographically permitting channels are shallower and less concave, are unrelated to canal paths to diverge from the river and traverse the present drainage patterns, exhibit unusual lateral and lower edges of "Ibrtolita alluvial fans. A series of vertical continuity, and contain indications of introduced prehistoric canals paralleling historic ones has been flow greater than that to which the fluvial system was identified in aerial photographs and by surface remains, adjusted (Katzer and Schuster 1984). They appear to be passing near or through a number of sites (Fig. 3.2). It contemporary with adjacent Hohokam occupations. Ca- appears that these canals supplied drinking water for nals revealed by excavations on a fan west of the ellicson permanent residence and also may have supported some Basin (Withers 1973) represent another alternative to re- irrigated acreage. liance on purely natural water-spreading processes. Ca- diversion from large nals would have permitted upstream Santa Cruz Floodplain drainages with substantial watersheds that were less prone to natural overbank flow and floodwater spreading. Risks and opportunities for Hohokam agriculturalists Late historic floodwater farming of isolated fields in Zone 5 on the river floodplain cannot be judged allowed ideal positioning for minimal labor investment precisely by present conditions (Figs. 3.8, 3.9). His- and relocation whenever events such as massive floods torically, perennial surface flow in the Santa Cruz has altered favorable conditions. Use of fan areas by larger been absent north of Tlicson. Pre-Columbian runoff numbers of farmers prehistorically would have restricted would have been less rapid due to wooded stream relocations and prompted cultivation of less easily courses and the grass cover heavier before the appear- farmed locales. In a variety of situations with divertible ance of livestock. Floods would have been poorly con- water within the Marana Community, both the Mexican tained when the river channel was less incised. Current and "Mono O'odham consultants considered soils too entrenchment and associated lowering of water tables sandy to achieve high yields. They suggested that finer- began prior to this century (Cooke and Reeves 1976; grained and more water-retentive planting mediums Betancourt and Tlirner 1988). could be created by water' diversion and resultant silt It is not clear whether episodes of channel incision deposition for one to several years prior to cultivation. apply equally and simultaneously to the Santa Cruz Such improvements might encourage further modifica- throughout its length in the Tticson Basin. An instance tions toward field permanence, such as ditches or canals of downcutting prior to the Classic period in the of moderate size from upstream diversions. Along the southern Tbcson Basin apparently caused the abandon- Rio Sonora, Doolittle (1984) describes the incremental ment of riverine settlements dependent on irrigation growth of such improved and locationally stable field (Waters 1988: 217). The lack of later occupations at systems over the course of long-term use. several Preclassic settlements along the river near the Particular site locations on the lower bajada were 'southern boundary of the Marana Community (Figs. 3.1, used over long intervals of time, indicating commen- 3.2) may also have followed the disruption of canal surate stability in general farming locations. A dense intakes through vertical or horizontal shifts of the river concentration of lower bajada sites of all periods occurs channel. 176

Parameters of Agricultural Production

Historically, localized areas of high water table and are poorly defined. As a second method of prehistoric sustained flow of surface water were created in the Thc- floodplain cultivation in this area, a large set of rockpilcs son Basin by igneous intrusions related to the mountain on an undisturbed stretch of a low west terrace just masses. Without such impervious barriers, water infil- north of the mountains may have functioned to control trates the porous riverbed in other stretches and flows the dispersion of shallow floods. underground after short-term floods. In the northern basin, a volcanic intrusion, higher water table, and more AGRICULTURAL PARAMETERS prolonged flow occur near the end of the lbcson Moun- AND REGIONAL LAND USE tains in Zone 5. Accessibility and duration of flow create the best situation for canal intakes, and a variety of A perspective of regional scale provides richer historic lines headed here. insights into land-use patterns than would be possible Compared to the Phoenix Basin, floodplain width and through a compilation of isolated evidence. With the terrace morphology in the "Meson region significantly emergence of an overall settlement configuration in the restrict the extent of irrigable land. Nevertheless, northern "Meson Basin, the insufficiency of ethnographic irrigated fields along the floodplain supported the analogy for understanding the breadth of Hohokam sub- densest populations in the study area. Intensive produc- sistence has become apparent. Some agricultural tech- tion of annual crops by irrigation likely played a signifi- nologies and environmental situations were similar to cant role in community-wide population levels. Riverine those used by Piman Indians, but other components of settlement in both Preclassic and Classic periods was Hohokam production are without analogs even in the greatest in the area of high water table surrounding the earliest Spanish accounts. There are no descriptions of mountain end. agricultural patterns similar to those in upper and Floodplain surfaces of the Santa Cruz have been middle bajada zones. Ethnographic parallels are also alternately scoured and buried. Lateral channel move- lacking for productive modes that were capable of under- ment that has obscured prehistoric activity is indicated by writing the higher range of Hohokam population den- archaeological materials eroding from a west bank cut at sities and the occupation of settings that were historically the 'Meson Mountain terminus. Remains of prehistoric abandoned. canals are not visible on the surface of this zone and The availability of regional distributions sharpens the have only been identified in excavation. Evidence of discrimination of environmental variables affecting land- buried canals has been encountered on the west side of use patterns. Elements of the technological repertoire the river just south of the study area (Kinkade and Fritz represent incomplete information without reference to 1975) and within the Marana Community near the end the regional variety of associated environmental contexts. of the mountains (Bernard-Shaw 1988). Due to overlying For example, rockpile devices occur in more than one deposits, it is doubtful that the extent of Hohokam relationship to water and topography. In most instances, irrigation along the Santa Cruz can ever be compre- rockpilcs were constructed on middle bajada ridge tops hensively documented. where runoff follows the general elevational trend from North and downstream from the lbcson Mountains, uplands to valley floor. Other rockpiles were located on the river floodplain widens rapidly and terrace barriers the broad, gently angled sides of bajada drainages. Here, to the lateral extension of canals diminish. At the same overland flow of water is at right angles to the downhill time, the channel becomes more poorly defined and sur- slope toward the valley bottom and coincides with the face flow disappears underground except during large path of runoff as it joins streams laterally. Still another flood events. Floodplain expanses are subject to exten- setting for rockpiles is on a low terrace of the Santa sive shallow flooding over fine-grained alluvial soils. Cruz River. In this location, water originated from Canal headings are not feasible on this stretch of the shallow overbank flooding or canals. river, but irrigated fields were supplied historically by From one viewpoint, stability in Hohokam land use lines with intakes near the end of the mountains. Gravity characterizes the northern Thcson Basin. A dual site canals on the east side of the river extended as far north distribution, with one band along the floodplain and as the modern town of Marana (Roskruge 1896a, 1896b). adjacent lower bajada and one band paralleling the It is likéiy that Hohokam canals irrigated more land in flanks of the eastern mountains, was established early in this area prior to historic river channel incision. The the sequence and remained constant. Harshness of the broad floodplain is under cultivation today, and surface Hohokam environment has been viewed as imposing a indications of prehistoric canals are preserved only at modicum of settlement stability through the topographic higher elevation along the lower edge of the bajada. Sim- requirements for canal systems and labor investments in ilarly, the outlines of settlements in these modern fields them (Haury 1976: 354; Nicholas and Neitzel 1984; 177

52 Suzanne K Fish, Paul R. Fish, and John H. Madsen

Masse 1981, 1991). Restricted opportunities to concen- Site illustrates a dramatic agricultural reorientation trate water for floodwater farming or surface runoff based on preexisting technology (see Chapter 7). management likely crystallized land use in additional An understanding of environmental opportunity and topographic situations. Technological conservatism is available technology are not sufficient to predict pre- reflected by construction of the same feature types for historic decisions as to the form and extent of implemen- hundreds of years. In the Tucson Basin, the long- tation in agricultural production. Additional natural but standing relationship between settlement patterns and nonagricultural variables may impinge, such as the co- hydrological opportunity suggests continuity in agri- occurrence of suitable domestic water. More importantly, cultural approaches. agricultural production is the outcome of economic These observations do not imply, however, that the decisions on the part of individuals and groups, who system of Hohokam production was static or without consider cost, risk, expectations, and cultural values in identifiable trajectories. Dynamic elements were formulating their responses. During successive periods, interjected by changing economic needs and aspirations agricultural parameters of the northern Tucson Basin of Basin inhabitants over time. The Classic period influenced economic behavior in changing contexts of proliferation of rockpile fields above the Marana Mound regional settlement and demography. 178

References Cited

Acklery, Neal, Jerry Howard, and Randall H. McGuire 1987 La Ciudad Canals: A Study of Hohokam Irrigation Systems at the Community Level. Arizona State University Anthropological Field Studies 17. Tempe: Department of Anthropology, Arizona State University.

Bernard-Shaw, Mary 1988 Hohokam Canal Systems and Late Archaic Wells: The Evidence from the Los Morteros Site. In "Recent Research on Tucson Basin Prehistory," edited by William H. Doelle and Paul R. Fish. Institute for American Research Anthropological Papers 10: 153-174. Tucson: Institute for American Research.

Betancourt, Julio L., and Raymond Turner 1988 Historic Arroyo-cutting and Subsequent Channel Changes at the Congress Street Crossing, Santa Cruz River, Tucson, Arizona. In Arid Lands: Today and Tomorrow, edited by E. Whitehead, C. Hutchinson, B. Timmerman, and R. Varady, pp. 1353-1373. Boulder: Westview Press.

Bryan, Kirk 19 25 The Papago Country, Arizona: A Geologic Reconnaissance with a Guide to Desert Watering Places. U.S. Geological Survey Water Supply Paper 499. Washington.

1929 Floodwater Farming. Geographical Review 19: 444-456.

Castetter, Edward F., and Willis H. Bell 1942 Pima and Papago Indian Agriculture. Albuquerque: University of New Mexico Press.

Clotts, H. V. 1915 Report on Nomadic Papago Surveys. Report prepared by the U. S. Indian Services. MS, Arizona State Museum Library, University of Arizona, Tucson. 179

1917 History of the Papago Indians and History of Irrigation, Papago Indian Reservations, Arizona. Report prepared by the U.S. Indian Services. MS, Arizona State Museum Library, University of Arizona, Tucson.

Cooke, Ronald U., and Richard W. Reeves 1976 Arroyos and Environmental Change in the American Southwest. Oxford: Clarendon Press.

Crown, Patricia L. 1987 Classic Period Hohokam Settlement and Land Use in the Casa Grande Ruin Area, Arizona. Journal of Field Archaeology 14(2): 147-162.

Doolittle, William E. 1984 Agricultural Change as an Incremental Process. Annals of the Association of American Geographers 74: 124- 137.

Downum, Christian E. 1991 The Los Robles Survey: Archaeological Investigations of the Lower Santa Cruz Basin, from Marana to Red Rock, Arizona. MS on file, Arizona State Museum Library, University of Arizona, Tucson.

Downum, Christian E., John E. Douglas, and Douglas B. Craig 1985 Community Structure and Agricultural Strategies at Cerro Prieto (AZ AA:7:11). In "Proceedings of the 1983 Hohokam Symposium, Part II," edited by Alfred E. Dittert, Jr., and Donald E. Dove. Arizona Archaeological Society Occasional Papers 2: 545-556. Phoenix: Arizona Archaeological Society. 180

Doyel, David E. 1984 Sedentary Period Hohokam Paleo-economy in the New River Drainage, Central Arizona. In "Prehistoric Agricultural Strategies in the Southwest," edited by Suzanne K. Fish and Paul R. Fish. Arizona State University Anthropological Research Papers 33: 35-52. Tempe: Department of Anthropology, Arizona State University.

Field, John J. 1 985 Depositional Facies and Hohokam Settlement Patterns on Holocene Alluvial Fans, Northern Tucson Basin, Arizona. MS, Master's thesis, Department of Geosciences, University of Arizona, Tucson.

Fish, Suzanne K., Paul R. Fish, Christian E. Downum 1 984 Terraces and Hohokam Agricultural Production in the Tucson Basin, Arizona. In "Prehistoric Agricultural Strategies in the Southwest," edited by Suzanne K. Fish and Paul R. Fish. Arizona State University Anthropological Research Papers 33: 55-72. Tempe: Department of Anthropology, Arizona State University.

Green, Christine R. and William D. Sellers 1964 Arizona Climate. Tucson: University of Arizona Press.

Gumerman, George J., and R. Roy Johnson 1971 Prehistoric Human Population Distribution in a Biological Transition Zone. In "The Distribution of Prehistoric Population Aggregates," edited by George J. Gumerman, pp. 83-101. Prescott College Anthropological Reports 1. Prescott: Prescott College.

Hastings, James R., and Raymond M. Turner 1965 The Changing Mile: An Ecological Study of Vegetation Change with Time in the Lower Mile of an Arid and Semi-arid Region. Tucson: University of Arizona Press, Tucson. 181

Haury, Emil W. 1976 The Hohokam: Desert Farmers and Craftsmen. Tucson: University of Arizona Press.

Hewitt, James M., and David Stephen 1 98 1 Archaeological Investigations in the Tortolita Mountains Region, Southern Arizona. Archaeological Field Report 10. Tucson: Pima Community College.

Kinkade, Gay M., and Gordon L. Fritz 1975 The Tucson Sewage Project: Studies of Two Archaeological Sites in the Tucson Basin. Arizona State Museum Archaeological Series 64. Tucson: Arizona State Museum, University of Arizona.

Masse, W. Bruce 1979 An Intensive Survey of Prehistoric Dry Farming Systems near Tumamoc Hill in Tucson, Arizona. Th e Kiva 45 (1-2): 141-186.

1 9 8 1 Prehistoric Irrigation Systems in the Salt River Valley, Arizona. Science 214: 408-415.

1 987 Archaeological Investigations of Portions of the Las Acequias-Los Morteros Irrigation System. Arizona State Museum Archaeological Series 176. Tucson: Arizona State Museum, University of Arizona.

1 99 1 The Hohokam Quest for Sufficiency and Civilization in the Sonoran Desert. In Chaco and Hohokam: Prehistoric Regional Systems in the American Southwest, edited by Patricia L. Crown and W. James Judge, pp. 195-224. Santa Fe: School of American Research Press.

Nabhan, Gary P. 1 986 Ak-chin "Arroyo Mouth" and the Environmental Setting of the Papago Indian Fields in the Sonoran Desert. Applied Geography 6 (2): 61-75. 182

Nials, Fred L., David A. Gregory, and Donald A. Graybill 1 989 Salt River Streamflow and Hohokam Irrigation Systems. In "The 1982-1984 Excavations at Las Colinas: Environment and Subsistence," by Donald A. Graybill, David A. Gregory, Fred L. Niais, Suzanne K. Fish, Robert E. Gasser, Charles H. Miksicek, and Christine R. Szuter. Arizona State Museum Archaeological Series 162: 59-78. Tucson: Arizona State Museum, University of Arizona

Nicholas, Linda, and Jill Nietzel 1 984 Canal Irrigation and Sociopolitical Organization in the Lower Salt River Valley: A Diachronic Analysis. In "Prehistoric Agricultural Studies in the Southwest," edited by Suzanne K. Fish and Paul R. Fish. Arizona State University Anthropological Research Papers 33: 161-178. Tempe: Department of Anthropology, Arizona State University.

Reichhardt, Karen, and Gary P. Nabhan 1 982 Application of Remote Sensing in Evaluating Floodwater Farming on the Papago Reservation. MS, Office of Arid Land Studies, University of Arizona, Tucson.

Rosknige, George J. 1986a Township 12 South, Range 12 East, Gila and Salt River Meridian Arizona. Survey General's Office, Tucson, Arizona, November 28th, 1896.

1986b Township 11 South, Range 12 East Gila and Salt River Meridian Arizona. Survey General's Office, Tucson, Arizona, November 28th, 1896.

Schuster, J. E. and Keith Katzer 1984 The Quaternary Geology of the Northern Tucson Basin, Arizona and Its Archaeological Implications. MS, Master's thesis, Department of Geosciences, University of Arizona, Tucson. 183

Sellers, William D., Richard H. Hill, and M. Sanderson-Rae 1985 Arizona Climate: The First Hundred Years. MS, Institute of Physics, University of Arizona, Tucson.

Underhill, Ruth M. 1 939 Social Organization of the Papago Indians. New York: Columbia University Press.

Waters, Michael R. 1 987 Holocene Alluvial Geology and Geoarchaeology of AZ BB:13:14 and the San Xavier Reach of the Santa Cruz River, Arizona. In "Archaeology of the San Xavier Bridge Site (AZ BB:13:14), Tucson Basin, Southern Arizona," edited by John C. Ravesloot. Arizona State Museum Archaeological Series 171: 39-60. Tucson: Arizona State Museum, University of Arizona.

1 988 The Impact of the Fluvial Processes and Landscape Evolution on Archaeological Sites and Settlement Patterns along the San Xavier Reach of the Santa Cruz River, Arizona. Geoarchaeology 3(3): 205-219.

Water, Michael R., and John J. Field 1 986 Geomorphic Analysis of Hohokam Settlement Patterns on Alluvial Fans along the Western Flank of the Tortolita Mountains, Arizona. Geoarchaeology 1(4): 329- 345.

Wilson, John P. 1 985 Early Piman Agriculture: A New Look. In "The Southwestern Culture History: Collected Papers in Honor of Alfred H. Schroeder," edited by Charles H. Lange. Tii Archaeological Society of New Mexico 10: 129-138. Santa Fe: Archaeological Society of New Mexico.

Withers, Arnold M. 1 973 Excavations at Valshni Village. The Arizona Archaeologist 7. Phoenix: Arizona Archaeological Society. 184

APPENDIX F

ANALYZING REGIONAL AGRICULTURE:

A HOHOKAM EXAMPLE

by Suzanne K. Fish, Paul R. Fish, and John H. Madsen

1990. In The Archaeology of Regions: A Case for Full - Coverage Survey, edited by Suzanne K. Fish and Stephen A. Kowalewski, pp. 189-219. The Smithsonian Institution Press, Washington, D.C. 185

AYChaeology ofRegions A Case for Full-Coverage Survey

Edited by Suzanne K. Fish and Stephen A. Kowalewski

Smithsonian Institution Press Washington, D.C. London Copyright 0 1990 by Smithsonian Institution All rights reserved

Editor: Robin A. Gould

Library of Congress Cataloging-in-Publication Data The Archaeology of regions : a case for full-coverage survey / edited by Suzanne K. Fish and Stephen A. Kowalewslci. p. cm. Bibliography: p. ISBN 0-87474-404-0 1. Archaeological surveying—Congresses. I. Fish. Suzanne K. II. Kowalewski, Stephen A. CC76.3.A74 1989 930. 1 028 —dc20 89-6249

British Library Cataloguing-in-Publication Data available Manufactured in the United States of America 10 9 8 7 6 5 4 3 2 1 98 97 96 95 94 93 92 91 90 89 co The paper used in this publication meets the minimum requirements of the American National Standard for Performance of Paper for Printed Library Materials Z39.48-1984.

For permission to reproduce illustrations appearing in this book, please correspond directly with the owners of the works, as listed by chapter. The Smithsonian Institution Press does not retain reproduction rights for these illustrations or maintain a file of addresses for contributors. 187

Smithsonian Institution Press

r I I I

July 12, 1993

Suzanne K. Fish 2214 4th Street Tucson, AZ 85719

Dear Dr. Fish:

Thank you for your letter requesting permission to reprint your work "Analyzing Regional Agriculture: A Hohokam Example," in your book The Archaeology of Regions. We'd be happy to grant your request and supply the supply credit line below.

Reprinted from The Archaeology of Regions: The Case for Full-Coverage Survey edited by Suzanne K. Fish and Stephen Kowalewslci (Washington, D.C.: Smithsonian Institution Press) 189-218, by permission of the publisher. Copyright 1989 Smithsonian Institution.

Thanks for contacting me. Good luck with your work.

Cheryl Anderson Editor/Acquisitions and Permissions

:,'N -•'-;',7;'•"•• 1111 \ \ \ --; I ' • 188

8. Analyzing Regional Agriculture: A Hohokam Example Suzanne K. Fish, Paul R. Fish, and John H. Madsen

Introduction Full coverage is a relative term in the nomenclature of archaeo- logical survey. Intensity of coverage is almost infinitely ref in- able by increasing the precision of recorded detail and by decreasing the distance between adjacent observations. How- ever, for detectable items at any level of coverage intensity, full- coverage surveys are distinguished from sampling surveys by the degree of completeness in recovered spatial patterns. It is this as- pect of coverage that most affects the potential for interpreta- tion beyond inventory statements. The complexity of relation- ships that can be examined among spatially patterned items varies directly with the completeness of the pattern in which those items are arrayed. Insofar as relationships between items reflect cultural and ecological interactions, pattern recovery 189

also governs the complexity of research questions that can be ad- dressed with survey results. The role of full-coverage data in studying central place func- tions and hierarchy is a primary one (Blanton et al. 1979 ; Sanders et al. 1979). With respect to items relevant for these problems, pattern segments may be differentially weighted, and the ab- sence of segments with unique items may be critical. An under- standing of subsistence is usually not perceived as similarly devolving upon unique elements ; the utility of full-coverage sur- vey may be less obvious. Food production characteristically in- volves repetitious activities across large expanses of landscape. Correlation between classes of subsistence remains and environ- mental variables or zones may be detected by adequate stratified or systematic samples. Yet even for comparatively simple agri- cultural societies, such low-order relationships are among the least complex issues. Research orientations derived from cul- tural evolution and other current paradigms involve systemic frames of interaction. Preservation of agricultural features and other subsistence- related remains is often unequal within and between study areas. Limited remains have stimulated innovative methods for indirect measurement of subsistence phenomena. In areas with textual resources, social, technical, and spatial parameters of prehistoric production may sometimes be drawn from docu- ments (Jacobsen 1958 ; Calnek 1972). Catchment analysis struc- tures the modeling of extractive relationships between sites and hinterlands. Excavated materials and ethnographic analogy allow more finely tuned estimates of sustaining area (Flannery

1976 ; Adams 1981). Analyses also have been freed from model- ing constraints of individual sites by considering potential pro- ductivity of spatial units (Kowalewski 1980, 1982). Derivation of patterns in agricultural production on a re- gional scale from archaeological distributions is less common in the literature. Large canal networks provide the most compre- hensive investigation of this magnitude (e.g., Adams 1981 ; Nicholas and Neitzel 1984). The potential of smaller, more diffuse remains associated directly with cultivation is less fully explored. Results of survey in southern Arizona provide a back-

Chapter 8 1 9 0

drop for discussing detection, analysis, and regional interpreta- tion of such features.

Northern Tucson Basin Survey Survey Strategy The decision to undertake full-coverage survey in the Tucson Basin was guided by a variety of theoretical and practical consid- erations. Research goals focused on spatial differentiation and integrative mechanisms at a regional scale. Settlement patterns of this magnitude based on full coverage, particularly ones in- corporating environmental variation, were unknown within the Hohokam tradition. The opportunity for regional pattern recov- ery was temporally finite in the study area, for Tucson's urban expansion posed an immediate threat. If full and intense cover- age was not attempted, the research value of such data could not be tested later. At the outset of the present survey, settlement configurations in the Tucson Basin were poorly known. Hohokam site densities away from river floodplains were generally assumed to be low, and lower in the Tucson Basin than in the Salt and Gila River heartland (Doyel 1977). A previous random stratified sample constituting less than 3 percent coverage of 240 sq mi (McCarthy 1982) encompassed portions of the proposed study area. Quadrats totaling 6.78 sq mi had predicted approximately five sites per square mile in each of four environmental strata, but had recorded only 29 sites. Five ceramic periods with inter- nal phase divisions were anticipated. Therefore, site numbers sufficient to examine functional and environmental relation- ships in any given period were likely to involve survey of ex- tremely large areas. Sampling the low density distributions would further increase areal coverage necessary to obtain ade- quate numbers while producing less complete patterns. In view of these considerations, optimal pattern recovery would be en- hanced by a corresponding advantage in procedural efficiency. Full coverage was not an option based on high funding levels. The Northern Tucson Basin Survey was significantly under- funded by standards of contemporary North American mitiga-

Fish, Fish, and Madsen P age Missing 1 n Original Volume

MissingPageTarget.tif 192

tion archaeology. Only a special set of circumstances allowed contemplation of this methodology. Application of the Arizona State Museum mandate to inventory state lands provided an in- stitutional framework and basic logistical support. Proximity of the study area to the University of Arizona made it possible for investigators and students to participate, without abandoning other obligations, during the academic year. Immediate access by road and large expanses without cultural barriers also prom- ised cost efficiency. The Bureau of Reclamation furnished sup- plemental support to this previously initiated survey in order to augment a regional framework for interpreting archaeological remains impacted by construction of the Central Arizona Proj- ect Aqueduct.

The Setting The Southwestern United States is considered ideal for recovery of detailed settlement configurations, including ephemeral sur- face remains. Arid vegetation obscures little, prehistoric popula- tion was seldom so dense as to eliminate early components, and modern disturbance is comparatively restricted. Within the Southwest, expression of these desirable conditions varies. The northern Tucson Basin in southern Arizona represents an opti- mal combination. As a microcosm of Sonoran Desert environments, the Tucson Basin is a particularly appropriate setting for study of diversi- fied arid-land agricultural production. Annual rainfall is just under 300 nun. Mountains bounding the basin to the east rise to a height of approximately 9,000 ft. To the west, a lower chain typifies the topography of more recent volcanic activity. The seasonally flowing Santa Cruz River drains the basin and ulti- mately joins the Gila. Major and minor intermittent drainages cross the valley slopes or bajadas. Some carry runoff from orographic rainfall in the mountains, and some originate in storm-fed bajada watersheds. The full range of environmental variation exists within an average horizontal distance of 15 mi between mountains on the east and west. No internal barriers are present to inhibit travel, communication, and exchange. Long-standing cultural boundaries conforming to the basin are

Chapter 8 1 9 3 suggested by a local ceramic sequence distinguishable from par- allel ones in the Hohokam tradition. Agricultural features are preserved and visible on the surface in all environmental zones of the northern Tucson Basin. Signif- icant modern settlement of the study area is late, mainly after 1915, and clustered along a few roads. The dominant economic activity has been raising. Modern agriculture is confined to localized and fragmented strips along the Santa Cruz and one major tributary. These conditions are in marked contrast to other Hohokam regions such as the Gila and Salt River valleys, where urban sprawl and highly mechanized, leveled-field agri- culture have obliterated major segments of prehistoric settle- ment patterns.

Settlement Patterns Survey at intervals of 30 m has resulted in aerial photo plotting of more than 700 sites (multicomponent sites have been tabu- lated only once for this total) and thousands of scatters and iso- lated artifacts. Distributions of these remains show that major patterns of landuse began early and remained constant. Sites of all ceramic periods beginning with the initial Pioneer period par- allel the Santa Cruz River. This concentration includes sites at the edge of the floodplain itself, as well as those on the lower edge of the adjoining bajada. Along the flanks of the eastern Tortolita Mountains, occupation is clearly indicated from the succeeding Colonial period onward. Long-standing preferred lo- cations within both river-oriented and mountain flank concen- trations are attested by numerous multicomponent sites and clusterings of temporally discrete ones (see Fig. 47). The upland and riverine settlement groups appear to represent equally substantial and permanent occupations. Ballcourts and large trash mounds occur at the largest sites. A range in site sizes and topographic settings in each cluster suggests differentiation in functions and productive capacities. Agricultural complexes of stone features are present in both upper bajada and riverine contexts. The variety of overall artifact classes is duplicated, in- cluding ground stone and shell. Year-round springs are the basis of sustained domestic water in the upland community ; locales of

Fish, Fish, and Madsen 194

Figure 47. Preclassic (ca. A.D. 750-1100) settlement pattern in the northern Tucson Basin.

Chapter 8 195

high water table on the floodplain provided opportunities for small reservoirs and shallow wells to supplement surface flows.

The Classic Community. At about A.D. 1 1 00, a dynamic settlement element is added to the long-term and persisting bands of occu- pation in riverine and mountain edge zones. The new locational orientation is evident in substantial remains upslope from a site containing a platform mound (Fig. 48). Dense sites span the pre- viously unused middle portion of the bajada and are continuous to mountain flanks. Settlement also increases on the eastern lower bajada and adjacent floodplain (S. Fish, Fish, and Madsen 1989). Ballcourts fall into disuse at this time. A platform mound is constructed approximately midway between the two Preclassic villages with courts, and the new inclusive Classic Marana community encompasses both earlier site clusters. The new configuration appears to be a temporally restricted phenome- non (ca. AD. 1 1 00- 1 300) fully developed in the early Classic and terminating prior to the advent of Gila or Tonto Polychrome in the later Classic. The new community incorporates multiple sites and environ- mental zones, extending from valley bottom to both eastern and western mountain flanks. A central site with platform mound and compounds near the present town of Marana, lesser sites with compounds, sites without compounds in a range of sizes, a trincheras or terraced hillside site, large communal agricultural fields, small agricultural fields, and a variety of specialized ac- tivity sites are components of this highly differentiated com- plex. Classic reorganization is marked by a sharp increase in site density, in the elaboration of functionally and locationally spe- cialized sites, and in an innovative, intensified use of the mid- bajada. The population level evident in the community seems only partially duplicated by settlements identified for preceding or succeeding periods.

Fish, Fish, and Madsen 196

4*4

A 0 I 2 NIL"

I 2 .0 Z.01.7177

CD o Archaeological Sites • Early Classic Period S Canal

Figure 48. Early Classic (ca. A.D. 1100-1300) settlement pattern in the northern Tucson Basin.

Chapter 8 197

Zonal Patterns in the Marana Community. Six zones of settlement can be defined within the larger Classic community of the Marana complex (Fig. 49). The first four pertain to the Tortolita bajada. Zone 1 sites, including the platform mound, occur in a more or less continuous band along the lower bajada. Coalescing alluvial fans in this zone create an active depositional environ- ment. Abundant sherd, ground stone, and lithic scatters across these fan surfaces correlate with dispersed, rancheria sites of various size. The density of Zone 1 settlement is likely the result of oppor- tunity for floodwater farming in an area receiving runoff from the full expanse of bajada slopes. Geomorphological studies and trenching (Katzer and Schuster 1984 ; Field 1985 ; Waters and Field 1986) have shown that central portions of large fans were less favored than fan edges and small fans associated with secon- dary drainages. Easily diverted and controlled flows in small channels would have renewed fields with rich loads of sus- pended sediment in the course of supplying water. The most desirable floodwater farming situations are concen- trated in the southern third of Zone 1 (S. Fish 1987:236-238). This is the segment with earliest settlement, predating study area reorganization in the Classic. Later expansion to the north in Zone 1 is almost certainly correlated with dependable domes- tic water through canal construction from the river to the mound vicinity. Although Zone 1 inhabitants may have realized some agricultural benefits from this canal, topographic place- ment suggests maximal irrigation potential for floodplain in- habitants of sites in Zone 5. Zone 2 is uphill from the mound and lower bajada edge sites. Dominant remains in this zone consist of huge complexes of ag- ricultural features without habitations. Rock piles are the prom- inent feature type, accompanied by low cobble terrace align- ments, check dams, and roasting pits (S. Fish, Fish, Miksicek, and Madsen 1985). Over 485 ha or more than 2 sq mi of large (10-50 ha) rock pile fields have been located.

Fish, Fish, and Madsen 198

Marano Community Zones

t 0 n 2 2: OPLCS

le•NrERS cc .1( t

lAtd - Bajaeo Fields

• Woe, 80p00 041 61

.7c Arcna.o4ic8 Sites

Figure 49. Zonal patterns in the northern Tucson Basin Early Classic community.

Chapter 8 199

A third settlement zone (Zone 3) in the middle portion of the bajada interfingers with the upper bajada nearer the foothills of the Tortolitas. A few small sherd and lithic scatters are widely dispersed in a rancheria pattern and located in optimal situa- tions for water diversion and utilization. Like Zone 2, Zone 3 is characterized by a scarcity of substantial habitation remains and the occurrence of unique and specialized sites. Surface scatters of ceramics with few or no other artifact classes comprise the only other Zone 3 sites. Sometimes huge (approaching 1.0 km in length), these tend to be linearly ar- ranged along ridgetops. Relatively dense distributions of sherds number as high as the tens of thousands. The obviously specialized function of these sites is unknown, but location in high densities of saguaro cacti makes long-term seasonal re- source gathering one possibility. Rock rings, usually associated with saguaro fruit procurement (Goodyear 1975 ; Raab 1973), are the only surface features and offer some support for this hypothesis. Zone 3 settlement patterns intergrade with a fourth zonal type nearest the mountains and between the major Cottonwood, Derrio, and Guild drainages. Unlike the lower zones underlain by deep colluvial basin fill, Zone 4 corresponds with mountain pediment, where shallow bedrock prevents deep percolation of water originating on the Tortolita slopes. A relatively high and accessible water table is therefore maintained in the drainages. The proliferation of large and small habitation sites undoubt- edly reflects the availability of water. Both the three major drain- ages and secondary ones appear to have supported cultivation. Agricultural features such as terraces, rock piles, and checkdams occur in substantial numbers in Zone 4 in conjunction with large sites, small sites, and isolated structures, but never inde- pendent of habitation as in Zone 2. The largest Zone 4 sites (a few approach or exceed 1 sq km) are found on ridges overlooking Derrio and Cottonwood flood- plains. Structural remains at these include a number of architec-

Fish, Fish, and Madsen 200

tural forms: compounds with house mounds, compounds exhibiting only cobble room outlines, dry-laid masonry struc- tures, and isolated cobble outlined structures. The density of ar- chitectural remains at large sites indicates a high probability of pithouses as well. The floodplain and terraces of the Santa Cruz constitute Zone 5. From the southern boundary of the study area to the end of the Tucson Mountains, the river channel and floodplain are more concisely delimited. Igneous intrusions near the end of the mountains force underground flow to the surface, creating an el- evated water table and more persistent surface water. Large and small sites of all periods occur on both sides of the river as it par- allels the mountains. Two of the largest long-term sites in the community, Los Morteros and the Huntington site, are located on either side of the mountain terminus. A ballcourt at Los Morteros and a probable one at the Huntington site (Huntington 1910) identify focal sites for southern pre-Classic settlement (Fig. 47). To the north, surface flow in the river is more infrequent and of briefer duration. Terraces become poorly defined and the floodplain broadens substantially. Precision of site definition decreases on the broad northern floodplain due to extensive modern agriculture. However, maximum site size appears to be smaller and overall densities reduced. The highly foreshortened bajada between the Tucson Moun- tains and the Santa Cruz compresses zonal topography equiva- lent with that in the eastern trans-bajada axis. Inhabitants of southern Zone 5 undoubtedly diverted flow from Tucson

Mountain watersheds ; some sites at the western floodplain edge seem oriented toward such floodwater situations. Riverine ca- nals undoubtedly account for the denser populations and con- sistently favored locales. Multiple canals and shallow wells are reported from excavations at the site of Los Morteros (Bernard- Shaw 1988). Also drawing upon the optimal water table of this general locale was the intake of the canal to mound vicinity. A number of historic gravity canals heading here and irrigating fields in the broad floodplain to the north probably duplicate earlier land use.

Chapter 8 201

The Tucson Mountains define Zone 6. These form a low chain with maximum study area height of less than 130 m above the floodplain. Dark volcanic hills are covered with a variety of cacti and leguminous trees, providing immediate access to upper bajada resources for inhabitants of the river edge. Trincheras sites, characterized by terraces and walls of dry- laid masonry, occur on Tucson Mountain billslopes as large and

small groups of features (Wallace and Holmlu-nd 1982). Some of the 250 terraces in the largest concentration yield evidence of agricultural function, and excavated pithouses in several other terraces at this site have all the appearances of permanent habita- tions (S. Fish et al. 1984). A cobble-outlined compound and a few masonry surface structures are also present.

Agricultural Patterns In the Northern Tucson Basin, as in most other regions, all facets of prehistoric agricultural activity have not left equally tangible or easily retrieved records. Santa Cruz floodplain canals are bur- ied by active alluvial and colluvial deposition. Inviting bottom- lands of larger tributaries also exhibit few surface features. Some ethnographic forms of cultivation with probable Hohokam an- tecedents involve small-scale brush diversions, low earthen ridges, and short ditches. Such ephemeral structures often are obliterated in single planting seasons and are unlikely to persist archaeologically. Farming of this sort can only be inferred from settlement location and associated topographic opportunity. With the exception of floodplains, unburied surface features are relatively well preserved throughout the study area. The bulk of such remains are small agricultural devices informally con- structed from unshaped local rock. These kinds of remains are fragile in the face of later prehistoric or modern surface distur- bance, and are therefore absent or more spottily preserved in many regions. Distributional characteristics hinder systematic recording even where preservation is optimal. The features are areally diffuse, but are so widespread in the aggregate that boundary definition presents problems. Associated artifacts are usually sparse. These factors challenge standard survey proce- dures. Nevertheless, in the Northern Tucson Basin, such re-

Fish, Fish, and Madsen 202

mains best illustrate the advantages of full-coverage survey for detecting and interpreting regional patterns of production. The ubiquitous unit of related agricultural remains in the study area has been termed a rock pile field after its most fre- quent feature. Rock piles or rounded heaps were constructed from cobbles scattered on the basin slopes. The rock piles rarely exceed 1.5 m in diameter and 75 cm in height. Short contour ter- races and checkdams of one to several cobble courses are often interspersed in small fields and are always present in large com- plexes. Segments of two typical fields are mapped in Figures 50 and 51. Field size varies widely ; the smallest are less than 1 ha and the largest is greater than 50 ha.

/ 1 --,---- / f /- 1 z 1 / y -- ( )r„. / z - » _.,---„, , , ? ,, 7 z ../

/ .._

NO

0 100

TCRS

Figure 50. Segment of a large agricultural field with terraces, rock piles, check dams, and roasting pits.

Chapter 8 203

AZ AA - 12470 (ASH)

—

• Roc lpol•

\. 2 A •••.

Genoa., 1•n••rve1 2 1••1

100

Figure 51. Segment of a large agricultural field with terraces, rock piles, check dams, and roasting pits.

Rock Piles and Planting. Rock pile complexes have been recorded throughout Hohokam territory, but have received little prior di- rected study. Multiple lines of evidence indicate that the rock piles as well as terraces and checkdams are facilities of agricul- tural production. They are improbable as residuals of clearing activity. Some occur in the midst of dense concentrations of sur- face rock and some appear to have necessitated importation of rock for construction. A function of clearing the surface to in- crease runoff directed to more arable land (Evenari et al. 1971) also seems unlikely. Many are topographically situated so as to receive optimal runoff from ridge catchments, and diversion de- vices for directing water to other sectors are not present in any field.

Fish, Fish, and Madsen 204

Rock piles as well as the more obviously agricultural terraces and checkdams with which they occur should enhance the plant- ing environment. The uneven, porous surface of the piles allows penetration of rainfall in contrast to surrounding hard-packed and impermeable ground surfaces. The rocks then act as a mulch, slowing evaporation of soil moisture by blocking capil- lary action and preserving higher moisture levels beneath. This effect of rocks in desert soils has been established experimen- tally in the literature (Evenari et al. 1971) and by soil moisture measurement in the present study. The response of currently growing plants to the microhabitat of rock piles is demonstrated by comparatively dense concentra- tions of annuals, strong preference by perennials, and the pres- ence of lichens and mosses requiring substantial moisture. Modern plant response has been quantified in the current study by comparing root biomass in soil directly beneath rock piles and adjacent controls. Root weight in rock pile soil was an aver- age of 80 percent higher than in the controls (S. Fish et al. 1985). These observations further strengthen an interpretation of rock piles as moisture-enhancing facilities for crop plants.

Zonal Distributions Although a few of the 112 rock pile loci in the study area are ear- lier, most fields and all the large ones date to the early Classic. A single rock pile field on the upper edge of a Santa Cruz terrace demonstrates occasional construction of these features in non- bajada environments. This field may be a remnant of more ex- tensive instances. Deposition and flooding on the terraces makes prehistoric distribution speculative. Forty rock pile fields on the upper bajada cover less than 2 ha each. Here, cobble features always occur in proximity to habita- tion. Some fields are adjacent to residential sites, but others con- tain evidence for one or several isolated structures within their boundaries. Topographic location is varied, including both ridge tops and broad intervening areas drained by minor unen- trenched channels. Middle bajada rock pile fields are homogeneous in locational attributes. All occur on ridges of dissected Pleistocene alluvial

Chapter 8 205

fans. Seventy-one loci occur in Zone 2 of the Classic community or in similar settings. These fields occur within a band less than 2.5 km wide, between elevations of 625 and 670 DI above sea level. Large-scale fields (from 10 to over 50 ha) account for 485 ha or almost 5 sq km on the middle bajada. Major complexes derive a linear outline from their ridge-top situations (Fig. 49). Cobble features are infrequent between ridges on bottomlands of secon- dary drainages. At least some of these intervening areas were farmed with the aid of earthen water control devices, as attested by occasionally preserved berms. Artifacts are thinly scattered across the large fields. No indications of fieldhouse structures or concentrated domestic refuse occur within them. Absence of habitation remains also characterizes the numerous smaller fields of this zone. Roasting pits filled with ash and fire-cracked rock are found among the cobble features of many small fields. Large fields usu- ally contain multiple pits. Stone artifact assemblages are marked by steep-edged core tools and knives made from tabular stone with flaked and ground edges. The higher density of stone tools and multiple roasting pits suggest a production level in large middle bajada rock pile fields that is not matched in those of other zones.

Advantages of Full Coverage Full coverage proved critical at all stages of investigation of re- gional agriculture. It allowed recognition of large-scale feature complexes as coherent units. It provided a reliable basis for com- paring distributions across environments and through time. Quantified reconstructions based on concrete archaeological distributions rather than modeled estimates also rest on full- coverage data. Alternative survey methodologies would have produced less precise description of the archaeological record of agriculture. Moreover, comprehensive pattern recovery was the key to understanding more inclusive aspects of Hohokam eco- nomic behavior.

Fish, Fish, and Madsen 206

Recognition of Scale Three previous surveys intersecting the study area either failed to detect the fields or proved inadequate for offering even esti- mates of size. The earliest, a transect restricted to a powerline right-of-way, crossed two of the large middle bajada fields with densest and most elaborate features (Rozen 1979). At this time the features were noted and recorded as a site, but only acknowl- edged to be of substantial size. In a succeeding stratified random quadrat survey, two 40-acre quadrats contained segments of large fields (McCarthy 1982). The presence of a few rock piles was mentioned with no further treatment. A second transect survey (Czaplicki 1984) consisted of a 300 m swath crossing al- most 3 km of the middle bajada area with abundant fields. The features were undetected. A number of factors can be cited that contribute to the inade- quacy of lesser coverage surveys in dealing with remains of this nature. Expectations based on compact habitation sites and con- centrated artifacts may lead to perception of diffuse features as insignificant, isolated occurrences. When transects or quadrats intercept edges of distributions, this perception is magnified. Full coverage reduces the probability of dismissal for unfamiliar diffuse items because the overall shape of the distribution be- comes readily apparent. As with the large fields, diffuse dis- tributions of unanticipated magnitude are least likely to be recognized without full coverage. Moreover, if 50 ha field boundaries were to be identified in the course of formal sam- pling surveys, significant departure in the areal size of sample units probably would have been necessary. Systematic boundary definition of large-scale as well as more finite phenomena is an important advantage of full coverage, and particularly so for understanding rock pile fields in a re- gional context. Precise size ranges and comparative acreage by environmental zone or by time can be derived from such data. Statements regarding items associated with fields and the pres- ence or absence of adjacent habitation are also more reliable with complete pattern recovery.

Chapter 8 207

Selection of Samples for Detailed Study Full coverage demonstrated clearly that roasting pits were highly correlated with rock pile fields. In the 350 sq km surveyed, roasting pits outside residential precincts coincided with the fields in all but one instance. Survey inventory also aided selec- tion of representative pits for additional study. Flotation of more than 600 liters of fill from 20 excavated roasting pits in 16 fields produced charred specimens of agave in each. Other eco- nomic plants were virtually absent. Cultivation of agave in the fields is strongly implied by the fact that natural stands are re- stricted to higher elevations in mountains bounding the Tucson Basin and are absent on bajadas. The largest excavated pit was 35 m in diameter and 1.5 m in depth. In this and similar instances, the term roasting area is probably more correct. Discrete pit shapes could not be easily distinguished among the intrusions and accretion of seasonal reuse over many years. Smaller, well-defined pits occurring al- most exclusively at small fields were less than 3 m in diameter. Stone artifacts from more intensive field collections following initial survey are commensurate with ethnographic Southwest- ern agave processing. Broad, flat tools made on tabular raw ma- terials are prominent. Such specialized tool types were used to sever agave leaves from the hearts (Castetter, Bell, and Grove 1938). In collections from middle bajada fields, these imple- ments (called agave knives) account for 19.2 percent of re- touched tools. Steep-edged core tools, - often designated scraping or pulping planes, have been associated ethnographically and ex- perimentally with fiber removal from agave leaves (Rogers 1939 ; Osborne 1965 1. Together the two artifact types account for 53.9 percent of retouched tools, an assemblage emphasis unique among study area sites.

Quantified Estimates from Archaeological Distributions Quantification of the components of prehistoric agriculture rep- resents a valid goal of archaeological inquiry as well as a resource for higher order inference. Many variables remain at the level of

Fish, Fish, and Madsen 208 informed speculation, supplied by reference to immediate or universal ethnographic example or experiment. Nevertheless, the shape, density, and extent of archaeological distributions can provide measures of scale and tangible bases for comparison. Full-pattern recovery is necessary to reliably compare quantified estimates among pattern segments.

Labor of Construction. Numbers of rock piles and meters of cobble alignments were counted in multiple 50 m squares at fields se- lected to represent variation in density. These figures were then used to calculate a total for the area covered by middle bajada fields. An estimate of 42,000 rock piles and 120,000 m of align- ments emerges. Experimental rock pile construction gave an ap- proximate range of 20 minutes to one hour, depending on size and availability of cobbles. An averaged figure of 40 minutes re- sults in initial construction labor of 28,000 person-hours or 14 person-years with five-day work weeks. Experimental building of comparable cobble terrace alignments gave a figure of 1.65 m per hour (Fish and Fish 1984). For the terraces and checkdams in the fields, an initial investment on the order of 72,500 person- hours or slightly over 36 person-years would have been required. Farmers could be expected to have expended a total of over 100,000 person-hours or more than 50 person-years in con- structing cobble features prior to the demands of planting, tend- ing, harvesting, and processing crops. Any perishable compo- nents of the system and maintenance of field facilities would have been additional. Intriguing but unanswered questions con- cerning labor investment include completion of fields during single construction episodes or by accretion during use, and the degree to which overall field layout may have been a centrally co- ordinated effort.

Yields. In view of consistent agave recovery from roasting pits and the viability of agave in such low moisture situations, we consider middle bajada fields with roasting pits and associated stone tools to be predominantly loci of agave production. Corn

Chapter 8 209

pollen recovered from subsurface samples in the vicinity of sev- eral cobble features suggests at least one additional crop. Mexi- can analogy relates agave to stone features in field systems where annual crops are also planted in better watered field seg-

ments or in seasons of favorable rainfall (K. Johnson 1977 ;

Messer 1978 ; Wilken 1976 ; Sanders et al. 1979). Agaves often line terrace walls for soil stabilization and runoff control, and a similar association seems likely for rock piles. Distributions of these features are therefore used to estimate potential agave yield. For a minimal estimate, a single plant is posited per rock pile, and one for every 2 m of alignment. A total of 102,000 plants could have been growing at one time. Harvest comes at plant maturity, for which an average 10-year figure is used. Thus only one-tenth or 10,200 plants would have been available in a given year. Yield in food and fiber is based on small Southwestern species most suited to lower elevations. Precise identification from macrofossils has not been possible, but fragments suggest smaller plants than the upper range of Mexican cultivars. If it is found in the future that the Hohokam used agaves of Meso- american origin, like other types of crops, estimates would be distinctly minimal. Yields for two end-products are calculated independently ; use of the edible heart should not have precluded fiber extraction from the leaves. Agaves harvested in a single year should have supplied approx- imately 365 gm of fiber per plant (Charles Miksicek, personal communication) or 3.72 metric tons. At an average of 4 kg, agave hearts would have produced 40.8 metric tons of edible product. Hearts furnish 347 calories and 4.5 gm of protein per 100 gm (Ross 1944). Using FAO recommendations, agaves in the middle bajada fields could have supplied the equivalent of annual calo- ric requirements for 155 individuals and protein requirements for 110. The probability of multiple agaves per rock pile sug- gests that actual production was even higher. While specific var- iables of estimation are subject to refinement or change, it is clear that these fields could have added a significant increment to Classic Hohokam diet and economy.

Fish, Fish, and Madsen 210

Economic Inference

Post-Harvest Activities. It is difficult to identify artifacts convinc- ingly linked to edible uses of agave. With future excavation of habitation sites in the study area, it may be possible to compare frequency of plant remains against full-coverage patterns of pro- duction. Fiber extraction and processing have artif actual corre- lates that are more amenable to post-harvest analysis. Stone knives with serrated edges pertain by morphology and ethno- graphic analogy to the cutting of leaves from hearts. Smoothed- edge knives, by similar argument, are best suited for scraping pulp from the leaves, freeing fibers without tearing them. Both types occur in the fields, indicating that some extraction proba- bly took place there after roasting. The relative concentration of these implements and appropriate spindle whorls in habitation sites might provide information on circulation of harvest within the Classic community or specialization at stages of processing. Existing survey collections would have to be supplemented for such inquiry, but results of intensive retrieval require knowl- edge of regional production for fullest interpretation.

Organization of Production. The regional distribution of rock pile fields has implications for the organization of labor and land tenure. Large-scale complexes would have necessitated commu- nal labor or common recognition of intrafield boundaries. In one large middle bajada field, a small boulder with petroglyphs resembles field boundary markers of the Hopi and Zuni (Forde

1931:233, 235 ; Cushing 1920:153). Like a number of South- western native people, kin-based or other corporate groups may have controlled arable land, with use-rights assigned to individ- ual plots. The patterning of fields with respect to habitation also can be compared by time and setting. Small Classic fields of the upper bajada and the few pre-Classic ones in the study area occur in conjunction with habitation. Only Classic fields on the mid- dle bajada are situated at a distance from any residential site, sig- naling a system of agricultural tenure not based on proximity. Large-scale fields undoubtedly tapped the labor of many culti- vators, whether through individual incentive or more central-

Chapter 8 211

ized direction. Indicative of communally scheduled efforts in these complexes are the huge roasting facilities, up to 35 m in di- ameter. While such features show repetitive use, excavation pro- duced little evidence of discrete, small firings. It appears that numbers of cultivators roasted their harvests together. This would have been an efficient use of woody fuel in a desert envi- ronment, but would have entailed coordination of harvest and preroasting preparation of the plants. By contrast, farmers using the modest pits of smaller fields could process their harvests ac- cording to individual convenience.

Production and Intensification. Recognition of the middle bajada field configuration as a temporal innovation rests on access to full-coverage data from the larger study area. These confirm that the technology and the crop were not new. Only the locations, the remarkable size of individual complexes, the expansion of total acreage in rock pile fields, and the concomitant emphasis on their yield are unique to the early part of the Classic. Reorga- nization of production at that time seems best understood in light of demographic conditions. The community containing large-scale fields has the highest site densities for any period of study area occupation. Middle bajada fields can be interpreted as a form of agricultural intensification in response to increased levels of demand. In an environment where aridity circumscribes agricultural activity, opportunities to expand production are limited. Agave cultivation on marginal bajada slopes would have offered an op- timal solution. These plants are adapted to low and unreliable moisture to a greater degree than many annual crops. Poorer land could therefore be used to satisfy growing needs for food- stuffs. In addition, fibers could support craft specialization and furnish highly portable raw materials and finished products for trade. Cultivation of middle bajada areas with little previous use represents an arid land version of intensification through expan- sion of extensive land use practices rather than more frequent cropping of land already cultivated. Intensification is also indi- cated by an increased labor cost in travel time per unit of produc-

Fish, Fish, and Madsen 212

production (G. Johnson 1977:490). Farmers of the middle bajada fields had to invest more time going to and from their fields than did farmers whose fields and habitations were adjacent. Kowalewski (1980, 1982) has tested the correlation of agricul- tural potential with location of population concentrations in the Valley of Oaxaca. The complexity and hierarchical nature of this Mesoamerican society are seen as involved in the lack of fit be- tween most productive land and population distributions. By contrast, the Hohokam of the Tucson Basin represent a simple level of social complexity in an environment with more limited water. Their settlement patterns might be expected to conform more closely to restricted environmental opportunity. Examination of full-coverage patterns reveals that distribu- tions of Hohokam remains most directly linked to agricultural production were shaped equally by cultural circumstance. The technology and crops of rock pile fields were available before the Classic, as was an unutilized abundance of bajada slopes. The

ability to raise crops on such land was clearly not an issue ; the decision to do so must have been an economic one based on re- source demands in a culturally perceived framework of invest- ment and return.

Scale and Replication. At 350 sq km, the Northern Tucson Basin Survey is large by Southwestern standards. Yet even at this scale, full coverage does not guarantee replication of all pattern elements needed to evaluate associations and trajectories. Re- organization of land use in the early Classic through elabora- tion of rock pile fields cannot be tested for representativeness against contemporary communities in the study area. Innova- tive expansion of production on terraces in the low western mountains can be documented (Fish et al. 1984), but a trans- bajada settlement complex is not duplicated. Sites of this time spanning similar zones are only known 20 mi to the south and 8 mi to the north. However, full coverage provides a model that can be examined for regional validity without equally complet, data in all instances.

Chapter 8 213

Conclusions Preservation of unburied, ephemeral, or diffuse but widespread remains is often uneven across a region. In these circumstances perhaps more than under ideal conditions, full coverage is the optimal methodology for perceiving and interpreting shapes of distributions. It becomes critical that less disturbed instances and ones with intact boundaries not be missed. Sample units may fail to intercept low-frequency occurrences of exceptional preservation that permit well-founded interpolation in pattern reconstruction. Fragmentary distributions providing empirical bases for such interpolation are even more likely to go unde- tected in sampling. The kinds of agricultural features represented by Tucson rock pile fields have parallels in other classes of archaeological re- mains. Diffuse features and/or artifacts of significance in under- standing regional settlement occur at all levels of social complexity. Included are remains of Great Basin hunting and gathering activities (Thomas 1975) and dispersed, continuous Aztec settlement in the Valley of Mexico (Sanders et al. 1979). Such remains present difficulties of recording and analysis in a survey paradigm of compact sites separated by empty space (Wobst 1983:66). Dispersed but related remains often fit the definition given by Plog et al. (1978:389) to a nonsite area as ". • . a potentially interpretable but not spatially discrete locus of cultural materials." These materials are further characterized as being so limited in quantity or covering so broad an area, or both, that meaningful boundaries cannot be defined by normal survey methods. However, boundaries can be systematically de- fined by full-coverage survey, and that is one of the unique assets of this approach. Even with well-defined boundaries, a number of current an- alysis methods fail to provide for maximum integration of full-coverage data. Many geographic procedures operate on spa- tially discrete representations with point symbols as the basic unit of settlement patterns. These approaches do not adequately represent diffuse and large-scale phenomena. Development of methods overcoming these constraints has been identified as a

Fish, Fish, and Madsen 214

current need in archaeological survey research (Ammerman 1981:80). In the present study, dating of agricultural remains posed no major problems. Although widely scattered, adequate diagnos- tic ceramics were recovered from the surface and were later con- firmed by others in roasting pit fill. A paucity of diagnostic artifacts prevents direct placement of many diffusely distrib- uted remains, however. In such cases examination of chronology at all settlements in the vicinity of distributions is a more reli- able basis for interpolation than is determination from frag- ments of the surrounding pattern. Full-coverage data also present the best opportunity to test proposed chronological associations by repetition and consistency throughout a study area. Some degree of homogeneity must be assumed in projecting the results of sample survey to unobserved portions of a region. Subsistence pursuits may be marked by redundancy with respect to environmental variables, but expression of these activities may differ significantly among appropriate environmental strata. Environmental opportunity is necessary but not suffi- cient as a predictor of archaeological distributions. Rare and unique items in settlement pattern are outcomes of hierarchy and other nonsymmetrical structure in human society ; to the de- gree that economic activity is influenced by such structure, eco- nomic remains will also acquire unique aspects of distribution. Concentrations or absences of subsistence activities responding to political boundaries, market locations, demographic thresh- olds or a host of other cultural conditions are most likely to be understood in a context of full-coverage patterns.

Acknowledgments Funding for the Tucson Basin study has come from the National Science Foundation (BNS-8408141), the Bureau of Reclamation and the Arizona State Land Department (4-CS-30-01380), and the Arizona State Historic Preservation Office (SP8315-70). Charles Miksicek conducted the plant macrofossil analysis re- sulting in the identification of agave from field roasting areas.

Chapter 8 215

References Cited

Adams, Robert McC. 1981 Heartland of Cities. University of Chicago Press, Chicago. Ammerman, Albert J. 1981 Surveys and Archaeological Research. Annual Review of Anthropology 10:63-88. Bernard-Shaw, Mary 1988 Hohokam Canal Systems and Late Archaic Wells: The Evi- dence from Los Morteros. In Recent Research on Tucson Basin Prehistory: Proceedings of the Second Tucson Basin Conference, edited by William Doelle and Paul Fish, pp. 153-174. Institute for American Research Anthropological Paper No. 10. Blanton, Richard E., Jill Appel, Laura Finster, Stephen Kowalewski, Gary Feinman, and Eva Fisch 1979 Regional Evolution in the Valley of Oaxaca, Mexico. Jour- nal of Field Archaeology 6:369-390. Calnek, Edward E. 1972 Settlement Pattern and Chinampa Agriculture at Tenoch- titlan. American Antiquity 37:104-114. Castetter, E., W. Bell, and A. Grove 1938 Ethnobiological Studies in the American Southwest: The Utilization and Distribution of Agave. University of New Mexico Bulletin 335. Cushing, Frank N. 1920 Zuni Breadstuff. Museum of the American Indian, Heye Foundation, Indian Notes and Monographs 8. Czaplicki, Jon (compiler) 1984 A Class III Survey of the Tucson Aqueduct Phase A Corri- dor. Arizona State Museum Archaeological Series 165. Doyel, David 1977 Rillito and Rincon Period Settlement Systems in the Mid- dle Santa Cruz River Valley. The Kiva 43:93-110. Evenari, Michael, Leslie Shana, and Naphtali Tadmor 1971 The Negev: The Challenge of a Desert. Harvard Univer- sity Press, Cambridge. Field, John 1985 Depositional Facies and Hohokam Settlement Patterns on Holocene Alluvial Fans, Northern Tucson Basin. M.S. the- sis, Department of Geosciences, University of Arizona, Tucson. Fish, Paul R., and Suzanne K. Fish 1984 Agricultural Maximization in the Sacred Mountain Basin. In Prehistoric Agricultural Strategies in the Southwest,

Fish, Fish, and Madsen 216

edited by S. K. Fish and P. R. Fish, pp. 147-160. Arizona State University Anthropological Research Papers 33. Fish, Suzanne K. 1987 An Evaluation of Subsistence and Specialization at the Marana Sites. In Studies of the Hohokam Community of Marana, edited by Glen Rice, pp. 235-248. Arizona State University Anthropological Field Studies 15. Fish, Suzanne K., Paul R. Fish, and Christian Downum 1984 Hohokam Terraces and Agricultural Production in the Tucson Basin. In Prehistoric Agricultural Strategies in the Southwest, edited by S. K. Fish and P. R. Fish, pp. 55-72. Arizona State University Anthropological Research Papers 33. Fish, Suzanne K., Paul R. Fish, and John Madsen 1985 A Preliminary Analysis of Hohokam Settlement and Agri- culture in the Northern Tucson Basin. In Proceedings of the 1983 Hohokam Symposium, edited by A. E. Dittert and D. E. Dove, pp. 75-100. Arizona Archaeological Soci- ety Occasional Paper 2. 1989 Classic Period Hohokam Community Integration in the Tucson Basin. In The Sociopolitical Structure of Pre- historic Southwestern Societies, edited by S. Upham, K. Lightfoot and R. Jewett. Westview Press, New York. Fish, Suzanne K., Paul Fish, Charles Miksic,ek, and John Madsen 1985 Prehistoric Agave Cultivation in Southern Arizona. Desert Plants 7:102-11. Flannery, Kent V. 1976 Empirical Determination of Site Catchments in Oaxaca and Tehuacan. In The Early Mesoamerican Village, edited by K. V. Flannery, pp. 103-116. Academic Press, New York. Forde, C. Daryll 1931 Hopi Agriculture and Land Ownership. Journal of the Royal Anthropological Institute of and Ire- land 61:357-407. Goodyear, Albert 1975 Hecla II and III: An Interpretive Study of Archaeological Remains from the Lakeshore Project, Papago Indian Reser- vation, South-Central, Arizona. Arizona State University Anthropological Research Paper 9. Huntington, Ellsworth 1910 Notebook 10A: The Journal of Ellsworth Huntington. Manuscript on file, Yale University Library, New Haven. Jacobsen, T 1958 Waters of the Ur. Iraq 22:174-185.

Chapter 8 217

edited by S. K. Fish and P. R. Fish, pp. 147-160. Arizona State University Anthropological Research Papers 33. Fish, Suzanne K. 1987 An Evaluation of Subsistence and Specialization at the Marana Sites. In Studies of the Hohokam Community of Marana, edited by Glen Rice, pp. 235-248. Arizona State University Anthropological Field Studies 15. Fish, Suzanne K., Paul R. Fish, and Christian Downum 1984 Hohokam Terraces and Agricultural Production in the Tucson Basin. In Prehistoric Agricultural Strategies in the Southwest, edited by S. K. Fish and P. R. Fish, pp. 55-72. Arizona State University Anthropological Research Papers 33. Fish, Suzanne K, Paul R. Fish, and John Madsen 1985 A Preliminary Analysis of Hohokam Settlement and Agri- culture in the Northern Tucson Basin. In Proceedings of the 1983 Hohokam Symposium, edited by A. E. Dittert and D. E. Dove, pp. 75-100. Arizona Archaeological Soci- ety Occasional Paper 2. 1989 Classic Period Hohokam Community Integration in the Tucson Basin. In The Sociopolitical Structure of Pre- historic Southwestern Societies, edited by S. Upham, K. Lightfoot and R. Jewett. Westview Press, New York. Fish, Suzanne K., Paul Fish, Charles Miksicek, and John Madsen 1985 Prehistoric Agave Cultivation in Southern Arizona. Desert Plants 7:102-11. Flannery, Kent V. 1976 Empirical Determination of Site Catchments in Oaxaca and Tehuacan. In The Early Mesoamerican Village, edited by K. V. Flannery, pp. 103-116. Academic Press, New York. Forde, C. Daryll 1931 Hopi Agriculture and Land Ownership. Journal of the Royal Anthropological Institute of Great Britain and Ire- land 61:357-407. Goodyear, Albert 1975 Hecla II and III: An Interpretive Study of Archaeological Remains from the Lakeshore Project, Papago Indian Reser- vation, South-Central, Arizona. Arizona State University Anthropological Research Paper 9. Huntington, Ellsworth 1910 Notebook 10A: The Journal of Ellsworth Huntington. Manuscript on file, Yale University Library, New Haven. Jacobsen, T. 1958 Waters of the Ur. Iraq 22:174-185.

Chapter 8 218

Johnson, Gregory A. 1977 Aspects of Regional Analysis in Archaeology. Annual Re- view of Anthropology 6:479-508. Johnson, Kirsten 1977 Disintegration of a Traditional Resource-Use Complex: The Otomi of the Mezquital Valley, Hidalgo, Mexico. Eco- nomic Geography 53:364-367. Katzer, Keith, and Jeanette Schuster 1984 The Quaternary Geology of the Northern Tucson Basin. M.S. thesis, Department of Geosciences, University of Ar- izona, Tucson. Kowalewski, Stephen A. 1980 Population-Resource Balances in Period I of Oaxaca, Mex- ico. American Antiquity 45:151-164. 1982 Population and Agricultural Potential: Early I Through V. In Monte Albém's Hinterland, Part I: The Prehispanic Settlement Patterns of the Central and Southern Parts of the Valley of Oaxaca, Mexico, by R. E. Blanton, S. Kowalewski, G. Feinman, and J. Appel, pp. 149-180. Memoirs of the Museum of Anthropology, University of Michigan 15. Ann Arbor. McCarthy, Carol H. 1982 An Archaeological Sample Survey of the Middle Santa Cruz River Basin. Picacho Reservoir to Tucson, Arizona. Arizona State Museum Archaeological Series 148. Messer, Ellen 1978 Zapotec Plant Knowledge: Classification, Uses, and Com- munication about Plants in Mitla, Oaxaca, Mexico. Mem- oirs of the Museum of Anthropology, University of Mich- igan 10(2). Nicholas, Linda, and Jill Neitzel 1984 Canal Irrigation and Sociopolitical Organization in the Lower Salt River Valley. In Prehistoric Agricultural Strate- gies in the Southwest, edited by S. K. Fish and P. R. Fish, pp. 161-178. Arizona State University Anthropological Research Paper 33. Osborne, C. 1965 The Preparation of Yucca Fiber. American Antiquity 31:45-50. Plog, Stephen, Fred Plog, and Walter Wait 1978 Decision Making in Modern Surveys. In Advances in Ar- chaeological Method and Theory, vol. 1, edited by Michael B. Schiffer, pp. 383-412. Academic Press, New York.

Fish, Fish, and Madsen 219

Raab, L. Mark 1973 AZ AA:5:2, a Prehistoric Cactus Camp in Papagueria.

Journal of the Arizona Academy of Science 8:116 - 118. Rogers, Malcolm 1939 Early Lithic Industries of the Colorado River and Adjacent Desert Areas. San Diego Museum Papers 3. Ross, Winifred 1944 The Present Day Dietary Habits of the Papa go Indians. M.S. thesis, Department of Home Economics, University of Arizona, Tucson. Rozen, Kenneth 1979 Archaeological Survey of the Northern Tucson 138 KV Transmission Line System. Arizona State Museum Archae- ological Series 132. Sanders, W. T, J. R. Parsons, and R. S. Santley 1979 The Basin of Mexico, Ecological Processes in the Evolu- tion of a Civilization. Academic Press, New York. Thomas, D. H. 1975 Nonsite Sampling in Archaeology: Up the Creek Without a Paddle? In Sampling in Archaeology, edited by J. W. Mueller. University of Arizona Press, Tucson. Wallace, Henry, and James Holmlund 1982 The Mortars, Petroglyphs, and Trincheras on Rillito Peak. The Riva 48:137-246. Waters, Michael, and John Field 1986 Geomorphic Analysis of Hohokam Settlement Patterns on Alluvial Fans Along the Western Flank of the Tortolita Mountains, Arizona. Geoarchaeology 1:29-345. Wilken, G. C. 1976 Traditional Slope Management: An Analytical Report. In Hill Lands, edited by J. Luchok, J. D. Cawthon, and M. J. Breslin, pp. 416-421. West Virginia University, Morgantown. Wobst, H. Martin 1983 We Can't See the Forest for the Trees: Sampling and the Shapes of Archaeological Distributions. In Archaeological Hammers and Theories, edited by James A. Moore and Arthur S. Keene, pp. 37-85. Academic Press, New York.

Chapter 8 220

APPENDIX G

AN ARCHAEOLOGICAL ASSESSMENT OF ECOSYSTEM IN THE

TUCSON BASIN

by Suzanne K. Fish and Paul R. Fish

1990. In The Ecosystem Approach in Anthropology: From Concept to Practice, edited by Emilio F. Moran, pp. 159-187. The University of Michigan Press, Ann Arbor. 221

The Ecosystem Approach in Anthropology From Concept to Practice

Emilio F. Moran, Editor

Ann Arbor The University of Michigan Press 222

Copyright © by the University of Michigan 1990 All rights reserved Published in the United States of America by The University of Michigan Press Manufactured in the United States of America

1993 1992 1991 1990 4 3 2 1

Distributed in the United Kingdom and Europe by Manchester University Press, Oxford Road, Manchester M13 9PL, UK

Library of Congress Cataloging-In-Publication Data

The Ecosystem approach in anthropology : from concept to practice/ Fmilio F. Moran, editor. p. cm. Rev. ed. of: The Ecosystem concept in anthropology. Includes bibliographical references and index ISBN 0-472-10178-1 (cloth : alk. paper). —ISBN 0-472-08102-0 (pbk. : alit. paper) 1. Anthropology—Methodology. 2. Human ecology—Research. 3. Anthropology—Philosophy. 4. Environmental archaeology— Research. 5. Biotic communities--Research. L Moran, Fmilio F. H. Ecosystem concept in anthropology. GN33.F:27 1990 301' .01-4c20 90-11306 CIP

British Library Cataloguing in Publication Data The ecosystem approach in anthropology : from concept to practice. 1. Anthropology. Research. Methodology L Moran, Fmilio F. 301

ISBN 0-472-10178-1 ISBN 0-472-08102-0 pbk 223

2214E. 4th St. Tucson, AZ 85719 June 26, 1993

Copyright Permissions Department The University of Michigan Press The University of Michigan Ann Arbor, Michigan

To Whom It May Concern:

1 am writing to request permission to reprint the following chapter:

Fish, Suzanne K. and Paul R. Fish (1990). An Archaeological Assessment of Ecosystems in the Tucson Basin of Arizona. In The Ecosystem

Approach in Anthropology edited by Emilio Moran, pp. 159 - i 87. The University of Michigan Press.

This material is to appear, in the same form as originally published, in my Ph.D. dissertation in the Arid Lands Resource Sciences Interdisciplinary Program at the University of Arizona. The dissertation, entitled Agriculture and Society in Arid Lands: A Hohokam Case Study," has been prepared under an option for the inclusion of published papers. This format requires require letters of permission for copyrighted materials. The entire dissertation, including the chapter cited above and copyright permission, will be reprinted by University Microfilms.

If you have any question, please contact me at your convenience.

lnc‘\" 224

CHAPTER 6

AN ARCHAEOLOGICAL ASSESSMENT OF ECOSYSTEMS IN THE TUCSON BASIN OF SOUTHERN ARIZONA

Suzanne K. Fish Paul R. Fish

Introduction

The ecosystem concept in archaeology has been aptly characterized as an heuristic device (Moran 1984:10; Jochim 1984:87). Compared to biological and human ecological studies of living populations, archaeological application necessarily entails a more generalized approach. Precision diverges from levels attainable in contemporary investigations because past environmental and cultural variables cannot be directly measured. Reconstruction is an intervening step to analysis. A more restricted range of variables can be identified or observed, and their significance in past ecosystems is inferential. Interrelationships among populations and other variables must be tested indirectly. In spite of the operational difficulties, ecological orientations are pervasive in current archaeology, fostered by concern with the physical environment of prehistoric populations. A benefit of explicit ecosystem analysis is a rigorous examination of underlying assumptions about interrelationships among cultural and environmental factors in past systems. As a generalized frame of reference for organizing and understanding data (Netting 1984/31-232), ecosystem concepts may even be invoked ex post facto in archaeology, if relevant observations have been gathered. Indeed, the process of identifying archaeological systems may legislate such a sequence because data revealing system shape are either acquired simultaneously with detail for particular components, or in reverse order.

159 225

160 Suzanne K. Fish and Paul R. Fish

Definition of meaningful boundaries is a preoccupation of ecosystem studies (Moran 1984, this volume; Ellen 1984, this volume; Butzer, this volume). Some level of system familiarity is requisite for drawing even the most preliminary boundaries from which to proceed to more detailed examination. The problem is particularly acute for archaeological systems because evidence for components is selectively preserved. The present case study of Hohokam remains in the Tucson Basin of southern Arizona illustrates the operational requirements of ecosystem studies in archaeology and the interplay between data collection and analysis.

The Hohokam

Cultigens and an agricultural lifestyle were adopted in the Sonoran Desert of south-central Arizona (Figure 6.1) by about 1000 B.C. Decorated pottery styles that define the Hohokam archaeological culture appeared in the first few centuries A.D. Cultural identity was maintained thereafter for over 1000 years, until or just preceding the 16th century. Throughout this interval, the Hohokam lived in pithouse structures of wattle-and-daub construction, with the addition of adobe surface structures in the later centuries of the sequence. Elements of ceremonial expression and public architecture such as ballcourts and mounds of earthen construction were shared with Mesoamerican cultures to the south, although in less elaborated forms. Archaeological study of the Hohokam has been concentrated in the contiguous lower valleys of the Salt and Gila Rivers. This area surrounding Phoenix, Arizona, termed the Hohokam core, is marked by massive irrigation networks totalling more than 400 km of primary canals (Masse 1980) and the highest population densities, estimated in the range of 8 to 25 persons/km' (Haury 1976:356; Doyel, in press; Fish and Fish, in press). Other regions of the Hohokam sphere lacked similar perennial rivers with large upland watersheds outside the desert. Farming in these other areas was accomplished through smaller-scale irrigation and a variety of techniques tapping intermittent and ephemeral watercourses. Until recently, archaeological study in non-core regions was limited. Settlement patterns among more dispersed populations were virtually unknown, and it was widely held that expressions of Hohokam high culture were attenuated or absent (See Fish 1989; Doyel 1980a; McGuire and Schiffer 1982). 226

Village

0 50 O•ii•• - 8o0oeonsow 60 ails••••• ° P'= L o " • Towns Los Tnneneros .. O archaeological Sites HohOliaen Lena,

Figure 6.1 Important locations within the Hohokam region of southern Arizona. 227

162 Suzanne K. Fish and Paul R. Fish

In the Phoenix Basin, irrigation systems and larger sites were mapped in the early part of this century before intensive modern development, permitting the recognition of interdependent site clusters along a shared canal network. Units of more closely interrelated sites along the same network have been inferred from the presence of prominent sites with ballcourts and mounds at fairly regular intervals (Doyel 1980b; Wilcox and Sternberg 1983; Gregory and NiaIs 1985; Crown 1987). Ballcourts are the common form of public architecture during the earlier Hohokam sequence, superseded in later times by mounds. Integrative functions for settlements without public architecture in an adjacent radius are ascribed to the sites with these structures. At the start of the present study, instances of public architecture in the Tucson Basin and other noncore regions implied similar functions of multi-site integration. However, configurations of related settlement surrounding preeminent sites with this architecture were unknown in areas lacking irrigation. No basis existed for even estimating magnitudes and distributions of populations in the Tucson Basin. The goal of research was to recover a sufficiently comprehensive settlement pattern to address such issues and to document the nature of the underlying subsistence base. Topographic diversity across the linear basin would have created variability in subsistence strategies, suggesting that integrated territories centered on mounds and ballcourt sites might incorporate differentiated productive activities. In order to characterize subsistence diversity and investigate economic relationships among sites, it would be necessary to record archaeological distributions across broad expanses of the basin, in conjunction with environmental data.

Populations, Communities, and Replication

Ecosystem analysis necessitates the partitioning of a continuous reality. Boundary definition of a unit appropriate for study is based on a discernible level of systemic closure, wherein interactions among components are more strongly oriented within than without. For some interactions, human populations may overcome physical separation by long intervening distances, and concepts such as "nested zones" or "graded boundaries" (Ellen 1984: 195-198) may be necessary for different realms of transaction or systemic inclusiveness; the role of extralocal materials in an archaeological 228

Tucson Basin Ecosystem 163

system is largely evaluated through quantities and contexts of extralocal occurrence. Nevertheless, following the general orientation of Rappaport (1968), an analytically relevant ecosystem for an archaeological population may be defined by the distributional extent of remains associated with its basic sustenance and reproduction. The integrity of an ecosystem defined in this manner may be indicated by spatial separation from other contemporary systems which also possess internal capacities for satisfaction of subsistence needs. Segments of the productive potential inherent in regional environment are bounded and integrated through cultural principles by human populations. According to the preceding definition of ecosystem, replications of such culturally-prescribed segments would be independent for the bulk of subsistence needs and would therefore encompass the resources and interactions fundamental to population support and continuity. Archaeological criteria for such a segment or unit would include: (1) environmental opportunity for internal production of most subsistence materials known to have been used; (2) remains of such productive activities; and (3) evidence for territorial cohesiveness and segregation from similar contemporary units. Among the Hohokam, boundary criteria for ecosystems of this nature do not seem best satisfied by single sites or small groups of sites in areas of largescale irrigation or in those emphasizing alternative technologies. The productive unit replicated across the landscape is a community of economically interdependent sites, whose identity and integration is symbolized by shared participation in public functions associated with ballcourts and platform mounds at a central site.

Tucson Basin Case Study

Environmental Setting

As the drainage basin of, a major river, the Tucson Basin is a typical focus of regional research in archaeology. Mountains rimming the basin and dividing it from other drainage systems form physical barriers promoting a degree of both natural and cultural closure. Annual rainfall, between 225 and 300 mrnj supports a relatively diverse desert vegetation dominated by shrubs, small, leguminous trees, and cacti, including arborescent forms. Somewhat more than half the rainfall occurs in summer as thunderstorms with diameters of 229

164 Suzanne K. Fish and Paul R. Fish

less than 5 km. Rapid runoff and high evaporation rates in this season increase the need to concentrate direct rainfall for successful cropping. The northern basin is the setting of the present study. Topographic diversity is mirrored to either side of the Santa Cruz River between the floodplain and mountains at the basin rim. The river runs near the base of the low Tucson Mountains (maximum elevation = 825 m) on the west, creating a foreshortened valley slope. The more massive Tortolita Mountains (maximum elevation = 1375 m) rise at a greater distance from the river to the east, and the eastern valley slope occupies a majority of the basin interior. The Santa Cruz is intermittent in the study area, but surface flow persists for extended periods near the end of the Tucson Mountains. Here, igneous intrusions of bedrock prevent water infiltration into deep valley fill, creating locally high water tables that might be tapped by shallow wells. Permanent or extended water sources also occur as springs and seeps in drainages near the base of the Tortolita Mountains.

Definition of Communities

A longterm program of regional survey between 1981 and 1986 eventually encompassed a study area of 2100 sq km (Figure 6.2). The basis of the present discussion is a contiguous survey block of 390 sq km near the town of Marana (S.Fish et al. 1989a). This area was covered by surveyors at intervals of 30 meters, permitting the identification of small, isolated loci of extractive activities and evidence of agricultural fields, as well as the more substantial remains of habitation sites. The resulting array of over 700 sites and thousands of small artifact scatters reflects only minor preservation bias. Ancillary geomorphological studies reveal a low degree of site burial for the Hohokam sequence, and historic land disturbance is confined to limited segments near the Santa Cruz River. The elements of settlement pattern constitute the basic variables of archaelogical ecosystem analysis, but could be segregated into meaningful spatial and interactive units only through examination of distributions at a regional scale. The nature of such configurations and their territorial extent was unanticipated, emerging only after the initial stages of data acquisition. Boundary definition, in this case, proceeded from higher to lower levels. Broader territorial 23 OA

Brady Tom Mix Big House ' !

\Tr:, • Brady Wash

I Suffering •• Wash • Olsen Wash

N.

Tortoilla

Mountains

4 O 5 miles 1E;11T126@ / O 8 klometers

• Salcoun • Pladotm Mound • Trincheras — Survey Transact WA 100% Survey

Figure 6.2 Survey areas and important site locations in the northern Tucson Basin. 230B

166 Suzanne K. Fish and Paul R. Fish

integration was recognized through settlement clustering and the positioning of central sites with public architecture as integrative nodes. Within these multi-site communities, local zones of direct environmental interaction for inhabitants of individual sites were then discerned through patterning of residence, productive activity, and environmental variables.

Community Dynamics and Evolution

Archaeology provides the opportunity for viewing dynamics of human ecosystems over tin c The temporal precision at which this can be realized rests on the ability to discriminate the chronological association of relevant remains. The shortest time increments, corresponding to the finest subdivisions of ceramic style, are difficult to isolate in Hohokam settlement patterns. Restricted occurrence of water for seasonal agriculture and permanent domestic supplies in the desert basins encouraged continuous reoccupation of optimal locales. Later occupations disturbed earlier ones, and ceramics of sequential phases are mixed on the surface. Non-residential sites reflecting hunting and gathering activities or field cultivation commonly contain small numbers of widely dispersed artifacts, few of which may be temporally diagnostic Ecosystems of a generally synchronous nature are therefore inclusively defined for the Preclassic period (circa AD. 700 to 1050) and the early part of the Classic period (circa. A.D. 1050 to 1300). Change over time in cultural ecosystems is apparent in the northern Tucson Basin, in spite of locational constraints imposed by the desert environment and an essentially static productive technology over the period in question. However, there are also aspects of long-term continuity in patterns of land use, which begin with settlements of the earliest farmers. Sites of all periods parallel the Santa Cruz River. This concentration includes both sites at the edges of the floodplaill itself and sites situated on the adjacent lower edge of the valley slope. A second concentration is found along the flanks of the Tortolita Mountains. Preferred locations within river and mountain flank bands of settlement are attested by numerous sites with superimposed occupations and clusterings of discrete ones. During the earliest phases (A.D. 300 to 700) defined by ceramic styles, scattered sites occur in both bands of settlement, but do not form cohesive clusters. Linkages among sets of sites are unclear for 231

Tucson Basin Ecosystem 167 this time. Later in the Preclassic as site numbers increase, two loose clusterings of residential sites can be identified in the immediate study area near Marana (Figure 6.3). These clusters satisfy the criteria of broad self sufficiency in subsistence for communities as units of ecosystem analysis. Both the cluster along the Santa Cruz river at the end of the Tucson Mountains and the one to the north on the upper valley slope of the Tortolita Mountains contain focal sites with ballcourts, suggesting independently integrated territorial units. Both incorporate permanent sources of potable water, diversity in productive activities, and a range of site types. Major habitation sites in each group appear to represent equally permanent occupations, as indicated by comparable assemblages of items difficult to transport such as grinding stones, and exotic goods such as shell ornaments. Each community is surrounded by areas lacking substantial habitation sites and with sharply reduced densities of other classes of remains. The two Preclassic communities are of equivalent territorial scale. The Tortolita flank community covers 57 sq km and the one along the Santa Cruz covers 70 sq km. Oval, earthen banked ballcourts at the central sites are widely assumed to represent ball game arenas as in Mtsoamerica, although use for ceremonies or dances has also been suggested. Competitions could have involved players from sites within the community or from different communities. Occasions of games likely would have served secondary purposes of exchange and social interaction (Wilcox and Sternberg 1983). Preclassic ballcourt communities are replicated across the Tucson Basin, with intervening areas of minimal settlement. For example, the ballcourt site of a second riverine community lies 16 km to the south of the one in the study area. Yet another ballcourt 19 km southeast along the mountain flank is similarly located with respect to its study area counterpart. Ecosystem boundaries are transformed in the early Classic Period by merger of the two previous communities into a single territorial unit encompassing 146 sq km (Figure 6.4). Settlement and societal reorganization at that time appears to have been accompanied by additions to population from outside the preceding communities. Residence and land use expanded into the former intervening area. A mound constructed at a large, new site in this area designates the focal point of the Classic community.

232

. v v•-z) , t1., \\ r. ,...... 01._,... ,.0_913::, . ) ..,-, ,- :: '-': k 1 i \ 8oficonyi

g ,,,, ° y co . L O°°1°' TORTOLITA ,-- '— 4 s.'" -- ) 4, ) .14: _I 0 rJ'‘,..e , c „ <,- ,_c? . ?..1 `-' P MOUN TA INS r.i, a ° .2, dsboec, „lz Stz,— n -\'' k ''' 1; ° i, A V.75L41:. r f /- 1 01, °

4

0 eo e 017 40,„

• fLo

Figure 63 Preclassic settlement pattern in the northern Tucson Basin.

233

° cz,

o/_ ° 0...-.. ° • v..' °•' o / ,oe- ••-• -,. ,.? f . MO UN TA I N S a e `,-/ .. or..° „, ...... '° Ic.aa. . 40.16.... c4 ty, ).' o o . . ° o 4,---, •eP s ., aft s . f ,,. ill09.:e t o ..../. o .0 0

^ o • ° ‘ ii. . •• I . •ad . r - . II .4°4 - q/j1. ,t;.'.., '... •

/...r. P• • • .° .... 1 •?... l

• t/sv °•.•. ° 5 ' two d -- *O. o C.-' ; 5* i* z ...6. :. 0 . .,. -- . ,,:. • / CJ.° — ' • " °.% /AL ev (4w / • / ... ..

Co Archaeological Site elb• Early Classic Period Site _ CD CI Platform Mound o o — Canal OC' co ,,.:. .

0 0 0 4.8 e . •0 ro -. IN.c' i a

% 11.• • 0 o

Figure 6.4 Classic Period settlement pattern in the northern Tucson Basin. 234

170 Suzanne K. Fish and Paul R. Fish

The mound site covers approximately 1,000,000 square meters. Inhabitants lived in adobe structures enclosed by compound walls of the same materiaL The mound, serving as a platform for rooms on the top, and adjacent structures of presumably specialized function were also enclosed by a compound walL In the Preclassic, observances of ceremony and ritual are thought to have been dispersed because objects such as figurines and censors are widely distributed among and within sites. During the Classic Period, such objects of individual use are no longer manufactured; evidence for ceremonial events is concentrated at mound loci. Differentiated personnel are amcciated with the residences and other rooms of mound compounds, in contrast to the unenclosed precincts of Preciassic ballcourts without conjoined structures. Centralization of ritual at focal sites is believed to indicate the stronger integration of Classic communities. Site hierarchies based on architectural elaboration and patterns of consumption evident in artifacts can also be distinguished at this time (S. Fish et al. 1989a; Howard 1987).

Zonal Patterns in the Classic Community

Gatherable resources in the Tucson Basin exhibit considerable spatial redundancy in the face of topographic diversity. There is species overlap across sequential elevations, although densities may differ. Short trans-valley distances in the linear basin allow convenient access to those items that are environmentally more localized. Charred plant (Miksicek 1988) and pollen (Fish 1988) evidence for crops raised in different locations also registers a redundant list. However, productive capacities, technologies, and risks are differentially correlated with hydrological regimes. Environmental variability in productive potential seems to have been further magnified through cultural patterns of economic emphasis. Differentiated production within northern Tucson Basin communities can be described in greatest detail for the Classic Period. Individuals, households, and inhabitants of particular sites did not interact routinely and directly with the ecosystem as a whole but with more immediate subdivisions and discrete components (Ellen 1984:168). Six zones have been defined within the Classic community (Figures 6.5 and 6.6). Differential patterns of residential and productive remains, and environmental variables characterize these relatively homogeneous zones.

235

Marano Community Zones \, t oo sties' )\\ / t5e

7 1.,—'-, z ..::'"-g"" A .:x -2., \ z 7,1 • mid-eaiada Melds ZONE 5 • upper Boiodo Fields 2 (:9 0 o -,A 77: Cm Archaeoloqical Sites !..,

tl'/I a * ';0 \ CI 2

ZONE 6 1--,,,,,N 0 ,,, , .(.0:0). 6

, / / :t iir,7% . , 1.1'N IV'', )i, ''' o ' \ ' D -7f -....— 'S,( — 1.v, b4 , \ , ./ / \ ", ( ". / k \

Figure 6.5 Zonal divisions in the early Classic Period settlement pattern. 236

O 0 0 o O 0 0 0 O ,r) 0 ,n ✓ K) r0 cv I I I I

- 0 OC - tm 00 I-5 ts.1

z. . o çr E-2-tgi w DO 3tct- Z =‘,. ..= -6 f, 0 N m114!ctmo o

re) 'br- .8=8: z 0 Eaa2a. N

w Z 3*.!!'s.';! 0 f:2c, NI 4 c u-4

I - i 1 i V) o Ln 0 I- in ni 0 7.. 0 a) co 237

Tucson Basin Ecosystem 173

Zone 1. Zone 1 residential settlements, including the mound site, occur in a more or less continuous band along the lower edge of the Tortolita valley slope. Coalescing alluvial fans create an active depositional environment. The density of remains in this zone is likely the result of opportunities for floodwater farming in an area receiving runoff from the full expanse of the valley slope. The most desirable floodwater farming situations are concentrated in the southern third of the zone. More shallow, easily diverted channels carry water after storms to moisten fields and fertilize them with organic detritus (Nabhan 1986). Permanent water is nearby in high riverine water tables. This is the segment of Zone I with earliest settlement, predating Classic period reorganization in the study area. Later expansion to the north and away from the river is correlated with a new source of domestic water made available through canal construction from the river to the mound vicinity. Zone 2. Zone 2 is uphill from the mound and lower valley slope sites. This zone lacks any sources of drinking water. Agricultural water from flow in drainages is less abundant and more difficult to divert than in zones above and below. Residential sites are virtually absent. Dominant remains consist of huge complexes of agricultural features dependent on surface runoff. Low, rounded piles of rock are the prominent feature type, accompanied by cobble terrace alignments, checkdams, and roasting pits. Perennial crops were planted in the rock piles, which act as water-conserving mulches and create barriers to rodent predation of crop bases and roots (S. Fish et al. 1985: 110). Large rock pile fields, each measuring 10 to 50 ha in size, cover more than 485 ha. A zonal total of 42,000 rock piles and 120,000 m of linear terrace alignments is estimated. This magnitude represents a dramatic increasç in the Classic Period over a few earlier instances of small fields. The principal crop of Zone 2 fields was the agave or century plant (Agave sp.) This drought-adapted succulent is widely cultivated in Mesoamerica, where it is called "maguey." Although wild stands were known to have been extensively exploited in the southwestern United States, recognition of agave as a major Hohokam cultigen is recent. Carbohydrates stored in the plant base over its lifetime for a sole and final flowering episode provide a sweet and nutritious food after pit-roasting. Stiff leaves produce a fiber for cordage crafts and coarse textiles. 238

174 Suzanne K. Fish and Paul R. Fish

Scattered stone tools in rock pile fields consist predominantly of knife and scraper types ethnographically related to agave harvesting and processing. Huge—up to 35 m diameter--roasting "areas" in all large fields appear to have been communally reused over many years. Excavated pits have yielded charred agave. Zone 2 fields occur on ridges between secondary drainages on gentle mid-valley slopes. Corn pollen recovered from sediments of the small drainage bottoms suggests occasional cultivation of this annual crop in seasons of particularly favorable rainfall. Such opportunistic supplements to dependable harvests of drought-resistant agave through interspersed plantings are also found in contemporary Mexican fields (e.g. Wilken 1979; Johnson 1977; West 1968). Zone 3. A third zone is located in the middle portion of the eastern valley slope. A few artifact scatters that may represent isolated or temporary structures are dispersed in optimal situations for water diversion and utilization. However, like Zone 2, Zone 3 is characterized by a scarcity of substantial habitation remains and the occurrence of unique and specialized sites. Sites marked by abundant pottery and few or no other artifact types extend up to one km in length and tend to be arranged linearly along the ridgetops. The obviously specialized function of these sites has not been directly confirmed, but locations coincide with high densities of saguaro cacti. Longterm seasonal gathering by large groups is probable. Rock rings, thought to have served as supports for large conical-bottomed baskets, are usually associated with procurement of saguaro fruits (Goodyear 1975; Raab 1973). These rings are the only surface features. In this waterless zone, pottery accumulations would result from containers for imported drinking supplies and for resource processing. Zone 4. Zone 3 settlement patterns intergrade with a fourth zonal type nearest the mountains and between three major drainages. Unlike the lower zones underlain by deep colluvial basin fill, Zone 4 corresponds with mountain pediment, where shallow bedrock prevents deep percolation of water originating on the Tortolita slopes. A relatively high and accessible water table is therefore maintained in the drainages. The proliferation of large and small habitation sites undoubtedly reflects this availability. The three major drainages and secondary ones all appear to have supported cultivation. Features such as terraces, rock piles, and checkdams are numerous in Zone 4 in conjunction with large sites, small sites, and isolated structures. Agricultural complexes are 239

Tucson Basin Ecosystem 175 never located at a distance from habitations as in Zone 2, nor do they approach the size of large fields common to that zone. Zone 5. The floodplain and terraces of the Santa Cruz River constitute Zone 5. Large and small residential sites of all periods occur on both sides of the river as it parallels the mountains in response to the elevated water table and more persistent surface flow. Two of the largest, most continuously inhabited sites in the community are located on either side of the mountain terminus. From the southern boundary of the study area to the end of the Tucson Mountains, the river channel and floodplain are concisely delimited. To the north, the river flows more infrequently and for briefer durations. Terraces become poorly defined and the considerably broadened floodplain is subject to flooding after major precipitation. Residential sites are smaller and overall densities are reduced. The highly foreshortened valley slope between the Tucson Mountains and the Santa Cruz River compresses zonal topography equivalent to that of the eastern valley slope. Inhabitants of southern Zone 5 undoubtedly diverted flow from Tucson Mountain watersheds; some sites at the western floodplain edge seem oriented toward such floodwater situations. Riverine canals undoubtedly account for the denser populations and consistently favored locales, however. Zone 6. The Tucson Mountains form a low chain with a maximum study area height less than 130 m above the floodplain. These dark volcanic hills that form Zone 6 are covered with a variety of cacti and leguminous trees. Inhabitants of the river edge therefore had immediate access to these characteristic resources of upper valley slopes. Terraces and walls of dry-laid Masonry occur on Tucson Mountain hillslopes in both large and small groups. The largest concentration consists of over 250 terraces. Some excavated terraces have yielded evidence of agriculture while others contained foundations for what appears to have been permanent habitations (S. Fish et al. 1984; Downum 1986). Intervening Areas. The most concentrated extractive activities took place in zones within community boundaries. Widely scattered artifacts in the areas lacking settlement between communities attest to more occasional use for hunting, gathering, or collection of raw materials. Intervening territory may have buffered potential conflict between communities, but it is likely that it was 240

176 Suzanne K. Fish and Paul R. Fish

also essential for acquisition of resources exhausted by longterm use near habitation sites. In particular, hunting of larger game and firewood gathering in the desert environment may have required travel beyond community boundaries. In the most productive zones along the river and the edge of the uplands, spaces between adjacent communities are less extensive and sharply defined. Broad expanses of the middle valley slopes at appreciable distances from community settlements contain virtually no indications of exploitive visits, in spite of abundant seasonal resources such as cacti. The burden of water transport likely discouraged use of this otherwise available bounty. During the prolonged summers when temperatures routinely exceed 100 ° F (37.7 ° C), the weight of water for extractive trips of any distance or duration may have been prohibitive.

Community as Ecosystem

According to the definition of ecosystem in this study, inhabitants of sites in the productively differentiated zones of the Clnsic community must be interlinked by more than shared ritual or political identity. Economic relationships basic to community subsistence and continuity cannot be documented directly from the archaeological record, but must rest on cumulative, inferential evidence. Diversity in zonal resources, and agricultural opportunity, become significant to the degree that economic integration can be demonstrated. Under average or better circumstances, the zones of long-term, dense habitation along the river and mountain flanks might have been largely independent for both cultivated and noncultivated subsistence needs. Indeed, the viability of separate community units was apparent during the Preclamic Period. However, community continuity requires strategies for successfully circumventing intervals of poorer-than-average conditions for desert farmers. The ability to counteract frequent and unpredictable adversity is affected by the ratio of population to stored or alternative resources. Higher levels of population in the Classic Period may have negated the options of the smaller, independent Preclassic communities. The magnitude of contrast between earlier and later populations can be appreciated by a comparison of cumulative totals for areas within sites devoted to residential function: 2,300,000 square meters in Preclassic sites compared to 6,100,000 square meters in the Classic ones. 241

Tucson Basin Ecosystem 177

The two Preclassic communities faced reciprocal environmental hazards common to the deserts of the southwestern United States (Lightfoot and Plog 1984; Abruzzi 1989). Storms that flooded riverine zones and destroyed canal intakes might ensure bountiful harvests on valley slopes. Conversely, rains sufficient for irrigated crops and posing no floodplain threat might produce scanty harvests in upper fields. Seasonal limits to cultivation exhibit a similar dual pattern. Although predictable yields per area would be highest on the irrigated floodplain, pronounced cold air drainage and temperature inversions in desert basins seriously inhibit winter crops on valley floors. Winter rains could be used for an early crop on upper valley slopes free of frost. Strategies of diversification were followed historically by many Southwestern groups to counteract environmental threats. In the Classic Period community, diversification was shifted further above the level of individuals, sites, and zones to the broad productive spectrum of the expanded boundaries. Localized risks were defused to the extent that fortunes and resources were shared within the larger entity. Overlapping biota among zones and a parallel suite of staples at sites yielding botanical evidence might imply limited routine exchange. Yet, differing magnitudes, emphases, and organization of production within the community suggest resource circulation.

Specialization and Exchange

Fine-scale examination supports subsistence exchange even within single zones. The most complete intra-zone data come from six excavated sites to the south of the mound in Zone 1 (Rice 1987). Artifacts, facilities, and botanical remains emphasize agave production at the northernmost of these sites, which is located nearest the large Zone 2 rock pile fields. Potential for floodwater farming is lowest at this site, with correspondingly few indications of adjacent cultivation of corn or other annual crops. Grinding stones document substantial corn consumption at the site, but it is doubtful that much of it was grown locally. Other patterns among the six excavated sites lack obvious environmental correlates. For example, the same northernmost site with poor land for corn had best access to plentiful saguaro. Nevertheless, botanical remains do not show that saguaro gathering was intensified. A second site with adjacent floodwater fields 242

178 Suzanne K. Fish and Paul R. Fish

yielded evidence for the most intensive saguaro processing found. This site similarly emphasized cholla cactus, a uniformly available resource. To the extent that production and consumption can be discriminated from archaeological remains, differential productive emphasis and exchange appear to have both broadened and homogenized consumption at individual Zone 1 sites (Fish and Donaldson, in press). The strongest expression of subsistence specialization is the class of remains unequivocally linked to agave production. Specialization in this case is coincident with agricultural expansion onto marginal lands in the vicinity of high Classic population densities. Zone 2 rock pile fields represent a means to support the inhabitants of the mound vicinity, a location lacking situations for floodwater farming comparable to other segments of Zone 1. Based on quantities of stone agricultural features and a plant-to-feature ratio, the minimal estimate of annual yield is 40.8 metric tons of edible product. This amount would supply the total caloric needs of 155 people and protein for 110, or an appreciable fraction of these dietary requirements for a much larger population (FAO/WHO 1973). Agave specialization above the mound has implications for the organization of production as well as its extent. Huge roasting areas that entailed communal harvest and processing schedules occur only in large Zone 2 fields. Massive complexes with thousands of contiguous stone features necessitated coordination in apportioning plots and maintaining boundaries. These fields, separated spatially from habitations, must have created greater need for land tenure consensus than small fields in proximity to residences, as in Zone 4. Finally, large Zone 2 fields required greater labor input per unit of production than the ones adjacent to dwellings. Fields at a distance from living areas demand an additional investment in travel and transport (Chisholm 1970; Bentley 1987;42-44), and reduce the ease of monitoring. Concentrated labor pools with limited alternatives may have encouraged pursuit of this option. Water supplies could not support intensified cropping of corn and other high-moisture annuals near the mound. In an arid land variation on agricultural intensification, as defined by Boserup (1965), total production was increased by expansion onto marginal land rather than by more intensive cultivation of existing fields. Specialized production of raw materials is often accompanied by specialization in manufactured products. Locations of agave crafts need not duplicate those of primary production, but dense populations 243

Tucson Basin Ecosystem 179

with poorer agricultural land might be drawn toward both activities. The artifactual correlates of fiber processing are prominent at the six excavated Zone 1 sites. Stone knives for fiber extraction and spindle whorls of a type suitable for agave products were recovered at each site. The estimated minimum annual yield of 3.72 metric tons of agave fiber from Zone 2 fields (S.Fish et al. 1985: 112; S. Fish et aL 1989b) should have provided a surplus of conveniently portable goods for exchange, perhaps both within and outside community boundaries. As with subsistence resources, part of the basis for community exchange in other items is differential opportunity. Chipped stone materials occur in the volcanic rock of the Tucson Mountains rather than in the granitic Tortolita range. Circulated raw materials and finished products from these localized sources were essential for everyday extractive and processing equipment Almost 90 percent of stone tools in sites of the Tortolita slopes were made from raw materials requiring exchange. The abundance of circulated items is commensurate with specialized acquisition by residents near the sources and reciprocal exchange on the part of consumers elsewhere in the community. Exchange outside the Classic community was necessary for other materials equally vital for subsistence activities. Knives used in agave production and the majority of grinding implements originated to the north and west. Quarries for both have been identified in another contemporary community 30 km from the mound site. Importation of finished products is indicated by manufacturing debris at sites near those sources but not within the community under study. Some community surplus must have been offered in return. Nonutilitarian raw materials and manufactured products also circulated widely. Many categories such as shell, turquoise, red pipestone, imported ceramics, and copper originated outside the community, often at considerable distances. Some exotic materials such as shell were worked into finished forms at a variety of locations in the community, while working of other materials, such as red pipestone for ornaments, has been documented only at the mound site. It is probable that both long distance and internal exchange of these items also involved subsistence resources, but its extent is unknown. Several lines of evidence suggest a close relationship between resource circulation, hierarchical structure, and principles of political or religious integration in the Classic community. Routine 244

180 Suzanne K. Fish and Paul R. Fish use of nonlocal raw materials and a degree of differentiated and specialized subsistence production created a widespread base for exchange. Geographical centrality of the mound site within the community bestowed an advantage for performers and regulators of transfer activities. In fact, few other advantages can be cited for the newly settled location, with no natural source of domestic water and secondary agricultural potential As specialized producers from marginal land, inhabitants of the mound site had high stakes in a comprehensive and dependable system of exchange. Strong association of nonlocal ceramics with this site, and the three others marked by compound architecture, suggests the prominence of long distance exchange. Rice (1987) has proposed that Classic community exchange resembled redistribution in a chiefdom context. However, additional possibilities such as periodic market-like gatherings, ritually organized exchange among relatively equal social units as in the Pueblos (Ford 1972), or some mixture of these according to resource type are consistent with the evidence available.

Conclusions

In archaeological ecosystem studies, where relationships among variables cannot be elicited from actors or directly observed, correlations between environmental factors and cultural remains provide a starting point. Such correlations reveal combinations of environmental opportunity and productive activity that underlie the interactions of human populations in an ecosystemic context. Settlement patterns at a regional scale were crucial in the present study for an understanding of ecosystem dynamics beyond this elemental level and for the designation of boundaries. Regional replication also confirms the multi-site community as an appropriate unit, or level of analysis, in ecosystem studies. Although patterns of environmental correlation can be distinguished within the Classic community, broader consideration of the community and the northern Tucson Basin underscores their insufficiency for predicting many aspects of prehistoric economic behavior. Choices by community inhabitants among a range of economic alternatives seem to have shaped both synchronic settlement expression of ecosystem interactions and ecosystem dynamics over time. 245

Tucson Basin Ecosystem 181

With present archaeological data, the role of economic decisions can best be evaluated at the level of individual sites. For instance, potential investment in rock pile fields on middle valley slopes was virtually unlimited. Topographic and hydrologic variables similar to those of the largest fields are duplicated widely in proximity to the Classic community. As judged by the relative lushness of natural vegetation across the middle slope, segments with somewhat greater moisture availability were used less intensively than drier ones. The extent of rock pile fields in the vicinity of particular sites can best be understood as the outcome of investment decisions that took into account other kinds of agricultural opportunity and population densities. The distribution of Zone 2 rock pile fields in relation to Zone 1 habitation sites can be seen in Figure 4. The largest fields and the greatest total acreage occur above Zone I sites with highest population densities and poorest potential for floodwater farming. These fields are located in a slightly drier portion of the valley slope and at a greater distance from habitations. Even though it likely entailed higher labor input per unit return, expansion of slope cultivation was economically advantageous under local circumstances. To the south in Zone 1, less dense population and small sites occur in a more promising area for floodwater farming, adjacent to better vegetated middle slopes. Nevertheless, inhabitants of these sites constructed fewer and mostly small rock pile fields. Further south still, fields are absent on slopes opposite heavy settlement along the river at the Tucson Mountains. The agricultural labor of river dwellers could evidently be absorbed more effectively in irrigated production. Agave cultivation in rock pile fields also exemplifies the primarily cultural dynamic of ecosystem change from Preclassic to Classic configurations. Botanical and geomorphological evidence reveal no significant environmental change during this time. The requirements of higher populations were met by expanded production on marginal land, without new technologies or new cultigens; scattered, small fields producing agave are documented in the Preclassic Period. Innovation occurs only in culturally determined variables of cultivation, such as the extent, distance from residence, and communal nature of the rock pile fields. With regard to the critical factor of domestic water, location of the preeminent mound site and new, dense Classic settlement can also be seen as the result of cultural rather than environmental 246

182 Suzanne K. Fish and Paul R. Fish criteria. Similar canal technology was implemented along the river during the Preclassic (Bernard-Shaw 1988), but construction of the 10 km canal to the mound site coincides with development of the Classic community. Thus, unlike earlier sites, the focal site of this period was dependent on distant water sources and its canal lifeline was vulnerable to interruption by human and natural forces. The polemic between culture and environment as prime cause of social patterns and diversity has formed the basis for intellectual development of an ecological anthropology (Moran 1984:3). Archaeological investigations provide an opportunity to assess the interplay of culture and environment through time and to test conclusions derived from synchronic perspectives. Not only must tendencies to assume ecosystem homeostasis be overcome in archaeological studies, but differential roles for culture and environment can be evaluated in particular cases of change. Ecosystems that include human components are subject to uniquely cultural sources of instability through changing perceptions of alternative behaviors. Hohokam settlement in the northern Tucson Basin illustrates the extent to which the ecosystem of a human population may be culturally defined even under restrictive conditions of an arid environment, Neolithic economy, and moderate level of political complexity. Classic sociopolitical integration encompassed greater enviromnental diversity within expanded community boundaries. The Classic Period community differed from its Preclassic precursors in overall numbers, localized densities, and geographic locations of population. Desert production was intensified in large part through expanded cultivation of a drought-adapted crop on marginal land. Notwithstanding an essentially static technology, Hohokam populations generated the motivations and possessed the ability to bring about ecosystem change.

References Cited

Abruzzi, William S. 1989 Ecology, resource redistribution, and Mormon settlement in northeastern Arizona. American

Anthropologist 91:642 - 655. Bentley, Jeffrey W. 1987 Economic and ecological approaches to land fragmentation: in defense of a much-maligned

2 4 7

Tucson Basin Ecosystem 183

phenomenon. Annual Review of Anthropology 16:31-67. Bernard-Shaw, Mary 1988 Hohokam canal systems and late Archaic wells: the evidence from the Los Morteros site. Recent Research on Tucson Basin Prehistory, edited by William Doelle and Paul Fish, pp. 153-173. Tucson: Institute for American Research Anthropological Paper 10. Boserup, Ester 1965 The Conditions of Agricultural Growth: The Economics of Agrarian Change under Population Pressure. Chicago: Aldine. Chisholm, M. 1970 Rural Settlement and Land Use: An Essay in L,ocation . Chicago: Aldine. Crown, Patricia L. 1987 Classic Period Hohokam settlement and land use in the Casa Grande Ruins area. Journal of Field Archaeology. 14:147-162. Downum, Christian E. 1986 The occupational use of hill space in the Tucson Basin: evidence from Linda Vista Hill. The Kiva 51:219-232. Doyel, David E. 1980a The prehistoric Hohokam of the Arizona desert. American Scientist 67:544-554. 1980b Hohokam social organization and the Sedentary to Classic transition. Current Issues in Hohokam Prehistory, edited by David Doyel and Fred Plog, pp. 23-40. Arizona State University Anthropological Research Paper 23. In Press Hohokam cultural evolution in the Phoenix Basin. Changing Views on Hohokam Archaeology, edited by Geage Gurwmall Albuquerque The University of New Mexico Press. Ellen, Roy.F. 1984 Trade, environment, and the reproduction of local systems in the Moluccas. The Ecosystem Concept in Anthropology, edited by Emilio Moran, pp. 163-204. AAAS Selected Symposium 92. Boulder: Westview Press.

248

184 Suzanne K. Fish and Paul R. Fish

FAO/WHO 1973 Energy and protein requirement: report of a joint FAO/WHO ad hoc expert committee. World Health Organization Technical Report Series 522. Fish, Paul R. 1989 The Hohokam: 1000 years of prehistory in the Sonoran Desert. Dynamics of Southwestern Prehistory, edited by George Gumerman and Linda Cordell, pp. 19-64. Washington, D.C.: Smithsonian Institution Press. Fish, Paul R. and Suzanne K. Fish In Press Hohokam political and social ogranization. Changing views on Hohokam Archaeology, edited by George aunennan. Albuquerque The University of New Mexico Press. Fish, Suzanne K. 1988 Environment and subsistence: the pollen evidence. Recent Research on Tucson Basin Prehistory, edited by William H. Docile and Paul R. Fish, pp. 31-38. Tucson: Institute for American Research Anthropological Papers 10. Fish, Suzanne K., Paul R. Fish, and Christian E. Downum 1984 Hohokam terraces and agricultural production in the Tucson Basin. Prehistoric Agricultural Strategies in the Southwest, edited by Suzanne K. Fish and Paul R. Fish, pp. 55-71. Tempe: Arizona State University Anthropological Research Paper 33. Fish, Suzanne K., Paul R. Fish and John Madsen 1989a Differentiation and integration in a Tucson Basin Classic Period Hohokam community. The Sociopolitical Structure of Prehistoric Southwestern Societies, edited by Steadman Upham, Kent Lightfoot, and Robert Jewett, pp. 237-267. Boulder: Westview Press. 1989b Analyzing prehistoric agriculture: a Hohokam example. The Archaeology of Regions: The Case for Full Coverage Survey, edited by Suzanne Fish and Stephen Kowalewslci, pp. 189-219. Washington, D.C.: Smithsonian Institution Press. Fish, Suzanne K. Paul R Fish, Charles Miksicek, and John Madsen 1985 Prehistoric agave cultivation in southern Arizona. Desert Plants 7:107-112. 249

Tucson Basin Ecosystem 185

Fish, Suzanne K. and Marcia Donaldson In Press Production and consumption in the archaeological record: a Hohokam example. The Kiva. Ford, Richard I. 1972 An ecological perspective on the Eastern Pueblo. New Perspectives on the Pueblos, edited by Alfonso Ortiz, pp. 1-17. Albuquerque: University of New Mexico Press. Goodyear, Albert C. 1975 Hecla II and III: An Interpretive Study of Archaeological Remains from the Lakeshore Project, Papago Indian Reservation, South-Central Arizona. Tempe: Arizona State University Anthropological Research Paper 9. Gregory, David A. and Fred L. Niais 1985 Observations concerning the distribution of Classic Period Hohokam platform mounds. Proceedings of the 1983 Hohokam Symposium, edited by A.E. Dittert, Jr. and D. E. Dove, pp. 373-388. Phoenix: Arizona Archaeological Society Occasional Paper 2. Haury, Emil 1976 The Hohokam: Desert Farmers and Craftsmen. Tucson: The University of Arizona Press. Howard, Jerry B. 1987 The Lehi Canal System: organization of a Classic Period community. The Hohokam Village: Site Structure and Organization, edited by David E. Doyel, pp. 211-222. Glenwood, Colorado: Southwestern and Rocky Mountain Division of the American Association for the Advancement of Science. Jochim, Michael 1984 The ecosystem concept in archaeology. The Ecosystem Concept in Anthropology, edited by Emilio Moran, pp. 87-102. AAAS Selected Symposium 92. Boulder: Westview Press. Johnson, Kirsten 1977 Disintegration of a traditional resource use the Otomi of the Mezquital Valley, Hidalgo, Mexico. Economic Geography 53: 364-367. Lightfoot, Kent G. and Fred Plog 1984 Intensification along the north side of the Mogollon

2 S 0

186 Suzanne K. Fish and Paul R. Fish

Rim. Prehistoric Agricultural Strategies in the Southwest, edited by Suzanne K. Fish and Paul R. Fish, pp. 179-195. Tempe: Arizona State University Anthropological Research Paper 33. Masse, W. Bruce 1981 Prehistoric irrigation systems in the Salt River Valley, Arizona. Science 214:408-415. McGuire, Randall H. and Michael B. Schiffer 1982 On the threshold of civilization: the Hohokam of southern Arizona. Archaeology 35: 22-29. Miksicek, Charles H. 1988 Rethinking Hohokam paleoethnobotanical assemblages a progress report from the Tucson Basin. Recent Research on Tucson Basin Prehistory, edited by William FL Docile and Paul R. Fish, pp. 47-56. Tucson: Institute for American Research Anthropological Papers 10. Moran, Emilio F. 1984 Limitations and advances in ecosystems research. The Ecosystem Concept in Anthropology, edited by Emilio E Moran, pp. 3-32 AAAS Selected Symposium 92. Boulder: Westview Press. Nabhan, Gary P. 1986 Ak-chin "arroyo mouth" and the environmental setting of the Papago Indian fields in the Sonoran Desert. Applied Geography 6:61-75. Netting, Robert McC. 1984 Reflections on an Alpine village as ecosystem. The Ecosystem Concept in Anthropology, edited by Emilio F. Moran, pp. 225-235. AAAS Selected Symposium 92. Boulder: Westview Press. Raab, L. Mark 1973 AZ AA:5:2, a prehistoric cactus camp in Papagueria. Journal of the Arizona Academy of Science 8:1-9. Rappaport, Roy A. 1968 for the Ancestors: Ritual in the Ecology of a New Guinea People. New Haven: Yale University Press. Rice, Glen E. 1987 Studies in the Hohokam Community of Marana. 251

Tucson Basin Ecosystem 187

Arizona State University Anthropological Field Studies 15. West, R. 1968 Population densities and agricultural practices in Pre-Columbian Mexico, with emphasis on semi-terracing. Verhandlungen des XXXVIII Internationalen Amerikanistenkongresses, Band II, pp. 361-369. Munich. Wilcox, David and Charles Sternberg 1983 Hohokam Ballcourts and Their Interpretation. Tucson: Arizona State Museum Archaeological Series 150. Wilken, G.C. 1970 The ecology of gathering in a Mexican farming region. Economic Botany 24:286-295. 252

APPENDIX H

DESERT AS CON IEXT: THE HOHOKAM ENVIRONMENT

by Suzanne K. Fish and Gary P. Nabhan

1991. In Exploring the Hohokam: Prehistoric Desert Peoples of the American Southwest, edited by George J. Gumerman, pp. 29-60. The University of New Mexico Press, Albuquerque. e

253 EXPLORING THE •HOHOKAM•

PREHISTORIC DESERT PEOPLES OF THE AMERICAN SOUTHWEST

Edited by George J. Gumerman

An Amerind Foundation Publication, DRAGOON, ARIZONA University of New Mexico Press, ALBUQUERQUE Library of Congress Cataloging-in-Publication Data Exploring the Hohokam : prehistoric desert peoples of the American Southwest / edited by George J. Gumerman. — 1st ed.

p. an. — (Amerind Foundation New World studies series; no. 1) Papers of a seminar held during Feb. 1988 at the Amerind Foundation, Dragoon, Ariz., and sponsored by the Bureau of Reclamation. Includes bibliographical references and index. ISBN 0-8263-1228-4 1. Hohokam culture—Congresses. 2. Sonoran Desert—Antiquities--Congresses. 3. Southwest, New—Antiquities—Congresses. I. Gumerman, George J. IL United States. Bureau of Reclamation. ilL Series. E99.H68E94 1991 979'.01—dc20 90-22509 CIP

Design by Susan Gutnik. • © 1991 by the Amerind Foundation, Inc. All rights reserved Second printing, 1992 255

THE UNIVERSITY OF NEW MEXICO PRESS

1720 Lomas Boulevard, N.E Albuquerque, New Mexico 87131-1591 U.S.A. (505) 277-2346 FAX: (505) 277-9270

July 7, 1993

Suzanne K. Fish 2214 E. 4th Street Tucson AZ 85719

Dear Ms. Fish:

Thank you for requesting permission to reproduce your chapter, "Desert as Context: The Hohokam Environment" from Exploring the Hohokam: Prehistoric Desert People of the American Southwest. Because this title was edited, you hold the rights to your essay, and no permission is required from us. Good luck with your dissertation!

Sincerely,

Cecilia E. Coats Permissions Assistant 256

•2• DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT

Suzanne K. Fish Gary P. Nabhan

he northern Sonoran Desert (Fig. 2.1) provided resources and opportuni- ties that figured prominently in Hohokam cultural identity. Broad geographic and biotic similarities contributed to widespread commonalities in lifestyle throughout the Hohokam range and underwrote the distribution of shared material culture by which these prehistoric peoples are recognized. From a complementary perspective, a level of secondary but significant environ- mental variability can be distinguished within the cultural domain. The presence of cultivators in southern Arizona is well established for a time prior to emergence of the Hohokam tradition (Huckell 1988; S. Fish et al. 1990), if this name is reserved for the first makers of pottery. Their ceramic successors can be regarded as primarily agriculturalists whose subsistence also included a broad range of natural products, many of which were abun- dant in culturally modified habitats. Groups with predominantly hunting and gathering strategies may have coexisted with the Hohokam in a relationship resembling that between the historic Sand Papago and more agriculturally

29 257

30 Suzanne K. Fish and Gary P. Nabhan

uN n TED STATES -m-ExICO

ARIZONA SONORA

0 (20_3 ISOLINES OF NORMAL ANNUAL PRECIPITATION (mm )

50 km LOWER COLORADO RIVER VALLEY

Figure 2.1 Subdivisions of the Sonoran Desert in Arizona.

oriented Pimans. Nevertheless, remains of such groups have not been ar- chaeologically identified as Hohokam (e.g., Hayden 1967), although they may have been linguistically or genetically related. In the sense that they circumscribe the distribution of unequivocally and distinctively Hohokam assemblages, two factors can be considered key en- vironmental determinants of Hohokam culture. These are location within the northeast Sonoran Desert and, with few exceptions, an elevational range below 3,500 ft (1,065 m). Residence and subsistence activity was concentrated on the floors and slopes (bajadas) of basins, and frequently extended into moderate elevations of adjacent desert mountains. Even where larger moun- tain masses occur, as in the Santa Catalinas, utilization of higher slopes and valleys was minor when compared with upland habitation and farming by other Southwestern cultures. The Hohokam inhabited the Sonoran Desert without extending into its driest fades to the west and southwest or into the more tropically influenced ones to the south. To the north and east, boundaries for the Hohokam largely coincide with those for this desert type. In keeping with their elevational tendencies, the Hohokam were rarely out of contact with arborescent desert plants characterizing the Sonoran biome. Such conjunction can be illustrated 258

DESERT AS CONTEXT: THE HOF i0KAM ENVIRONMENT 31

by a correlation between the great majority of Hohokam occupations and distributions of two conspicuous elements of this vegetational assemblage, the saguaro and mesquite. The interplay of environmental factors shaping Sonoran Desert biota were also central forces in the patterns of Hohokam adaptation.

DISTANCE AND DIVERSITY IN HOHOKAM BASINS

That portion of the Basin-and-Range Province (Fenneman 1931) settled by the Hohokam consists of parallel northeast-southwest trending basins bor- dered by relatively low and often discontinuous mountain ranges of limited mass. Intervening basins typically are broader than their mountain borders and are filled with detrital sediments to great depth. Along the flanks of the mountains, rock pediments formed by erosional recession of the mountain front are overlain by a thinner soil mantle, typically less than 2 m in depth. Streams with upland catchments deposit soil as they leave the mountains, creating alluvial fans on the upper bajada or valley slope. Further downslope, outwash materials from individual fans coalesce on the lower bajada. Terraces frequently border floodplains of major axial drainages, which are positioned at the juncture of bajadas from opposing mountain ranges. The structure of Sonoran Desert basins has implications for Hohokam subsistence patterns as a whole. Localized environmental conditions create a gradient of differentiated habitats along the transect from mountain peak to floodplain. These are duplicated in a generalized progression on both valley slopes. Only in the few cases of higher ranges do mountains on one side of a basin afford substantial resources unavailable on the other side or on the upper reaches of bajadas. Within Hohokam territory, the level of biotic uni- formity occasioned by the lower elevation Sonoran Desert biome is intensified by this intravalley redundancy. Typically, widths of well-defined basins range from 6 to 20 mi (10 to 32 km), and average about 12 mi (19 km). Even the broadest expanses, associated with junctures of major drainages as in the Phoenix area or with an absence of linear mountain chains as in the Santa Cruz Flats, rarely exceed 30 mi (48 km). Since the majority of resources could be acquired on a single slope without necessitating travel across the full valley width, access was not ex- cessive from most points in any basin. Most hunting and gathering forays from habitations near domestic and agricultural water could be completed within a single day. The six to seven quarts of water needed by active persons at high summer temperatures (Adolph 1947:113) could be reasonably transported, with the weight of gathered prod- ucts replacing consumed water for the return. The linear conformation of most Hohokam basins permitted practical foraging radii that were essential

259 -

32 Suzanne K. Fish and Gary P. Nabhan

to routine acquisition of resources from the waterless bajada. Furthermore, while there might be labor-scheduling conflicts during the busiest agricultural periods, permanency of residence for farming did not preclude access to a full array of wild resources in season. Although the basins of southern Arizona share a structural pattern, geo- morphic and topographic attributes significant to hunters, gatherers, and farmers exhibit variability between basins as well as locally within a single basin. In general, the mechanics of outwash transport create coarser soils on upper bajada slopes and finer soils below. However, parent materials differ among basins and among ranges bordering the same basin. Bajada widths influence the details of slope morphology. Degree of dissection affects both the extent of riparian communities and agricultural opportunities. Slope as- pect differs with directional trend of each basin, between opposing slopes, and even for individual intravalley locations. Such differences create variety in plant densities and distributions within broader vegetation zones and may be critical elements in differential crop success under conditions of agricultural stress. Variation among basin cross sections also has consequences for subsistence practices. Degree of slope affects runoff that supports vegetative resources and permits floodwater farming in favorable situations. Valley morphology associated with major axial drainages is an important variable controlling irrigation potential. Successful diversion during any period depends on dian- nel incision and stability. Width of the floodplain, the abruptness of the valley slope at the floodplain margin, and terrace structure determine the extent of irrigable acreage. The angle of the valley slopes as it intersects the floodplain controls not only the lateral extent of canals from the river, but also the force of flow in tributary drainages over which canals must pass.

HYDROLOGICAL REGIMES

Characteristics of two major divisions of desert watercourses correspond to a division between primary emphases in agricultural orientation among the Hohokam. Perennial rivers fed by watersheds outside the desert provided the basis for subsistence systems centered on large-scale riverine irrigation (Fig. 2.2). Intermittent streams and smaller ephemeral washes supported a range of more restricted irrigation and floodwater cultivation. The exogenous Gila and Salt rivers heading on the Mogollon Rim to the north and east, carry water from uplands of substantially higher rainfall into the lower desert. The Verde River shares many of the characteristics of the Salt, of which it is a tributary. Watersheds tap upland areas of winter snowfall, with resultant high flow in late winter and early spring. The Salt below the Verde confluence in the upper Phoenix Basin drains 12,900 sq mi (33,400 sq km) and the Gila drains 20,615 sq mi (53,400 sq km) (U.S. Geological Survey 260

DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 33

- Perennial , unregulated flow 0 50 Mi. Formerly perenn iai Regulated flow 0 75km Ephemeral

Figure 2.2 Current and recontructed flow in southern Arizona streams.

1954). Before modem damming, these rivers provided perennial flow for the Phoenix Basin and, prehistorically, for areas along much of their joint course to the Colorado River. Historic travel by flatboat was possible at times between the confluence of the two rivers and the Colorado (Castetter and Bell 1942:13). Vast, higher rainfall watersheds resulted in sustained flow and seasonal periodicity largely independent of precipitation in the desert basins. Flow peaked by the early spring in the Salt River, with a considerably lesser summer increase. Early spring and late summer peaks are documented for the upper and middle Gila. Bimodal streamflow regimes could have supported more than one crop in a growing season if other conditions permitted. Along the Salt and Gila, gently sloping terrace and basin floor morphology allowed canal networks to be extended laterally from the riverbeds, creating wide expanses of irrigable land. 261

34 Suzanne K. Fish and Gary P. Nabhan

Intermittent Hohokam rivers and other larger drainages originate in lower, less massive uplands within the Sonoran Desert or along its northern and eastern edge below the Mogollon Rim. Watershed sizes are highly variable. The Santa Cruz River, with the largest, drains 8,524 sq mi (22,100 sq km) at Laveen (U.S. Geological Survey 1954). Rainfall and runoff in the watersheds is lower than in the Mogollon highlands; snowmelt does not figure as sig- nificantly or at all in annual supply. Consequently these watercourses, almost all of which ultimately join the Salt and Gila network, carry water only sea- sonally over parts or all of their courses. Flow in major streams south and west of the Phoenix Basin peaks in conjunction with summer rains rather than in the spring and is not well timed for extensive early crops. Valley slopes rise more steeply from floodplains, restricting the width of irrigated acreage. While canal systems of secondary size could be constructed in some locales as along New River (Doyel 1984) and the Santa Cruz (Bernard-Shaw 1988), groups living along these watercourses placed greater emphasis on alternative farming technologies. Much of the moisture moving through Sonoran Desert basins flows below the surface in deep valley fill and generally follows gradients of surface water channels. Where this underflow encounters a bedrock obstacle, it is forced toward the surface. Igneous intrusions beneath basin floors are therefore correlates of sustained surface flow in intermittent streams. An example is along the Santa Cruz in the vicinity of Tucson. Water is forced upward near San Xavier, at the end of the Tucson Mountains, and again near the modern town of Red Rock (Smith 1910:176-178). Shallow wells might be filled from high water tables in such areas during drier intervals when surface flow had ceased. Similar geological intrusions created maximal discharge and channel stability for canal headings on the permanently flowing Salt and Gila (Lee 1905; Pewe 1978; Graybill and Nials 1989). Instances occur on the Gila River just above Snaketown and on the Salt in the vicinity of the Tempe Buttes. Current and reconstructed flow in southern Arizona streams according to Hendrickson and Mincldey (1984) is shown in Figure 2.2. During drier sea- sons, water along segments of intermittent rivers might be sufficient for do- mestic supply, if not for irrigation. Such stretches would have been continuously preferred for occupation. However, locations of perennial flow varied some- what over time even in favored locales. Well-developed cienegas have dis- appeared in recent times after downcutting by major floods and lowering of water tables (Betancourt and Turner 1988; Hastings and Turner 1965; Hen- drickson and Mincldey 1984; Dobyns 1981). An example of prehistoric incision with similar consequences is well documented near San Xavier on the Santa Cruz (Waters 1988). In addition to axial drainages, long-term surface water in the Sonoran Desert can be found in the lower canyons of mountains bordering basins and on their pediments. Upland catchments receive more regular precipitation and greater quantities than lower basins. As streams leave the mountains, 262

DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 35

rocky basements beneath pediment channels maintain flow at or near the surface. Floodplains of drainages along the mountain edge offer potential for dry season water sources in hand-dug wells and impoundments. From larger mountains such as the Catalinas or Bradshaws, persistent or perennial stream- flow may be found on the mountain flanks and may occur locally in channels for some distance onto the upper valley slope. Flanks of even moderate-sized desert ranges also give rise to occasional permanent springs and rock-lined tanks such as those supplying the historic Tohono O'odham (Papago) well villages (Underhill 1939:57). As drainages with upland sources pass over deep valley fill beyond the pediments water infiltrates porous channels and disappears from the surface. Lesser drainages finger into smaller channels as gradient diminishes on lower bajada alluvial fans. Typically, only a few larger tributaries maintain contin- uous transbajada courses and carry water to the rivers after storms. Where the floodplains of such large tributaries or arroyos are sufficiently wide, sea- sonal floods could be diverted onto fields beside channels. Numerous ephemeral drainages head on bajadas or on small ranges. Larger channels may join through-going streams crossing the valley slope. Minor ones cut and fill cyclically as they intercept overland runoff following storms, without forming continuous drainage networks. Potential water sources for floodwater farming on alluvial fans include both tributaries originating in the mountains and larger ephemeral drainages heading on the bajadas.

GEOGRAPHIC PATTERNS OF CLIMATE

With exceptions on the north and east fringes, present annual precipitation in the area inhabited by the Hohokam (Figs. 2.1, 2.3 and Table 2.1) averages less than 15 in (400 mm). The extremes for yearly rainfall within the area vary by a factor of two, with the low end of the range to the west (Sellers et al. 1985). Totals from 7 to 12 in (175 to 300 mm) are widespread, but a few locales such as Gila Bend receive less than 6 in (150 mm). The areas of highest summer temperatures coincide with annual rainfall below 9 in (225 mm) in the vicinity of Phoenix and southwest to Gila Bend. Lower annual precipitation is also correlated with greater variation about the mean. Harsher climatic conditions reduce both wild resources and the potential for farming with local rainfall. These disadvantages were offset along the Salt and Gila by harvestable plants of rich riparian communities and by abundant water for riverine irrigation. In addition to variation in yearly amounts, seasonal precipitation exhibits geographical patterns. The percentage of total precipitation occurring in the summer months decreases from east to west (Hastings and Turner 1965:14), while evapotranspiration increases. Proportionally lower summer rainfall oc- curs in conjunction with lower total precipitation and high annual coefficients 263

36 Suzanne K. Fish and Gary P. Nabhan

/1 40 /112 0 1/10 0

37° -

AR I ZONA

J5° -

366 • 56• • 26 28 • • 265 • 149 • 239 012 047 PHOENIX

• 191 59 , •114 33° - • I28

• 262 • 2

• 349 TUCSON • 299 0 50 Miles " 0 50 Kilometers

4\ H

I — 5/° - Figure 2.3 fixations of weather stations furnishing precipitation data. of variation to the west, producing maximal summer drought as a limiting factor for vegetation. A bimodal distribution of rainfall in the Hohokam area contrasts with winter dominance to the west and summer dominance to - the east. This bal- ance has been associated both with the greater structural diversity of Sonoran Desert vegetation and with the arborescent character of many of its perennials (Turner and Brown 1982:182), compared to shrub dominance in the Chihua- huan, Great Basin, and Mohave deserts. Distinctively large Sonoran Desert lifeforms include such important economic plants as paloverde, ironwood, and saguaro. Less seasonally balanced rainfall and lower annual amounts correspond with diminished desert arboreals. Higher precipitation in a broad arc to the north, east, and south of the greater Phoenix Basin is correlated with increased density and diversity in these key food species. 264

DESERT AS CONTEXT . THE HOHOKAM ENVIRONMENT 37

Table 2.1 Precipttation Data for Selected Stations in the Area of Hohokam Occupation Average Years Average Annual July-September Weather Station of Record Precipitation (inches) Precipitation (inches) 366 Wickenburg 71 11.15 4.32 56 Carefree 13 12.26 4.04 26 Bartlett Dam 43 12.70 3.88 265 Roosevelt 76 15.74 4.74 12 Apache Junction 13 9.53 2.80 28 Beardsley 28 7.36 2.61 239 Phoenix Airport 34 7.50 2.44 191 Maricopa 51 7.47 3.20 59 Casa Grande Ruins 59 8.79 3.11 114 Florence 69 9.93 3.52 262 Red Rock 44 9.72 4.21 349 Tucson Airport 34 11.09 5.91 299 Sells 26 12.26 6.66 2 Ajo 68 8.83 4.20 128 Gila Bend 77 5.82 2.14 47 Buckeye 80 7.43 2.65 149 Harquahala Plains 24 6.17 2.42 (data compiled from Sellers et al. 1985, table 11)

Winter precipitation originates largely over the Pacific Ocean and is cy- clonic or frontal in nature. Broad geographic expanses of rain at relatively gentle rates are common in this season. Sources for most summer moisture are storms developing in the Gulf of Mexico. Thunderstorms of restricted size and rapid delivery are typical, but may be followed by several hours of gentler rain (Turnage and Mallory 1941). In spite of the areally extensive nature of winter frontal systems, summer rainfall is less variable in total amount and exhibits a more seasonally predictable onset (McDonald 1956). For Hohokam farmers lacking opportunities for riverine irrigation, sum- mer rainfall was a particularly pivotal factor in annual crop production. Sum- mer rainfall is also a significant source of soil moisture for irrigated crops. Higher amounts decrease the need for supplemérdal water or, conversely, raise the potential for successful harvests with minimal irrigation. Abundance and predictability of summer rains were critical for irrigators along many watercourses other than the Salt and Gila, since high and sustained flows occur as a result of precipitation in these months. All along the western border of the Hohokam territory, precipitation in the summer months of July, August, and September (Sellers et al. 1985) is less than 4 in (100 mm). This amount also is not exceeded in a more central area about the Phoenix Basin and extending to Wickenburg and Bartlett Dam on the north, beyond Florence on the east, and as far south as Red Rock. In 265 -

38 Suzanne K. Fish and Gary P. Nabhan most of the Phoenix Basin, summer rainfall is below 2.5 in (60 mm), increasing to about 3.5 in (90 mm) in the vicinity of Florence. To the south and east, the middle Santa Cruz and San Pedro valleys receive 6 to 7 in (150 to 175 mm) of rain in the summer. These higher quantities extend across the Tucson Basin and to the west beyond Fresnal (6.5 in or 165 mm) on the Tohono O'odham Reservation. Advantages of predictable onset and quantity were greater for Hohokam cultivators in the areas of higher summer rainfall, since coefficients of seasonal variation generally decrease as precipitation totals increase. In the Tucson area, coefficients of variation (Raymond Turner, 1989 personal communication of unpublished data) for June through August precipitation are near .40. Values increase to near .60 at Florence and Casa Grande, and exceed .70 in the central Phoenix Basin. High values continue north of Phoenix as far as Wickenburg and Bartlett Dam before decreasing. In the vicinity of Sells in recent times, summer storm runoff events of 1 in (25 mm) or more rainfall allow floodwater production. Runoff initiation events of sufficient magnitude failed to occur in some years between 1941 and 1973, suggesting that this cultivation technique would not be successful every year (Nabhan 1983). Rapid rainfall and runoff during summer cloudbursts, their spotty distri- bution across the landscape, and most frequent occurrence from afternoon until midnight (Turnage and Mallory 1941) would have allowed Hohokam cultivators little opportunity to reach distant fields for diversion and distri- bution of water. Manipulations to avoid water loss must be accomplished quickly. For example, nighttime torchlit irrigation by the 0 1odham reduced wastage of temporary streamflow following summer storms. Overabundances in fields must be dispersed equally rapidly to prevent washouts and damage to growing crops. Protective constructions serve these purposes, but the timely presence of farmers for adjustment would be most critical where summer rainfall was a central factor in annual food production. The benefits of field proximity may have particularly encouraged dispersed patterns of settlement among the Hohokam in areas not served by the Salt and Gila canals. Local effects in some cases exert substantial influence on the expression of climatic factors. Mountains usually create higher rainfall on the adjacent windward basin and a rain shadow of reduced precipitation on the adjacent leeward basin. Low desert mountains of small mass may have little effect in this regard, while ones of greater bulk and height have a pronounced influ- ence. On the broad, low, and hottest valley floors, cold temperatures present minimal restriction of growing season. Major expanses with little yearly frost in the Salt and Gila valleys coincide with peak river flow in the early months of the year and extended availability of water. Spring peaks also characterize tributaries originating along the Mogollon Rim. Elsewhere in Hohokam ter- ritory, conflict in the scheduling of spring crops would have arisen between the need to plant sufficiently early in the year to benefit from winter rains, 266

DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 39

and the need to avoid desiccation of maturing crops during the late spring- early summer drought. Since both a greater proportion of annual rainfall and high fl ow occur in these areas during summer, spring crops would have been limited to particularly favored locales. Situations advantageous to early crops would be found in the best-watered fields at higher elevations on the basin edge, above levels of cold air inversion on the valley floor (S. Fish et al. 1984).

VEGETATION AND HABITAT COMPLEXITY: FUNCTIONAL VARIATION IN THE SONORAN DESERT

The Sonoran Desert includes almost 120,000 sq mi (310,000 sq km) of desertscrub and enclaves of upland vegetation types within Arizona, Sonora., southeastern California, and Baja California (McGinnies 1981:41). Within the Hohokam range in Arizona (Fig. 2.1), two vegetational subdivisions of the Sonoran Desert biome predominate: the Arizona Upland on the northeast edge, and the western-lying Lower Colorado Valley (Turner and Brown 1982). At the extreme east of the Hohokam range in the San Pedro drainage, the Sonoran Desert interdigitates with depauperate elements of the Chihuahuan Desert to form a "generic desert" of creosote (Larrea), mesquite (Prosopis), acacias (Acacia), and (Yucca), while lacking saguaros (Carnegiea gigantea). At slightly higher elevations (above 3,500 ft/1,200 m), semidesert grassland, interior chaparral or madrean, and evergreen woodlands are encountered, all but the latter within 60 mi (100 km) of any place within Hohokam country. Within a 120 mi (200 km) reach can be found Rocky Mountain subalpine conifer forest above 6,000 ft (1,900 m), a vegetation formation that the Ho- hokam may have drawn upon to obtain the long coniferous vigas at Casa Grande (Wilcox and Shenk 1977) and elsewhere. Although wetland and riparian areas make up far less than 0.1 percent of the Sonoran Desert lands today, oasis marshes and riparian forests (bosques) contributed disproportionately to habitat complexity and overall diversity of available plant species. This is illustrated by a recent riparian classification - (Asplund et al. 1988) of great utility in assessing habitat heterogeneity along watercourses. The resource richness of these zones for human hunting and gathering cannot be underestimated. Archaeological correlations with habitat heterogeneity frequently suffer from inappropriate comparisons of communities or series with higher order units (biomes, formations) as if they were equivalent entities. When habitat heterogeneity is assessed, as by the proximity of physiognomic vegetation formations (desert, grassland, woodland, forest) along an elevational or lat- itudinal gradient, comparison of elements within the same level of a classi- fication hierarchy is important for consistency. Within Sonoran Desert 267

40 Suzanne K. Fish and Gary P. Nahhan

thornscrub, at least four mappable communities can be located in Arizona's Lower Colorado Subdivision, and three within the Arizona Upland Subdi- vision. These communities or series (Lowe and Brown 1982) are the vegetation units with which archaeological variation can and should be correlated. A brief description of each community follows. The paloverde-cacti mixed scrub series is the Arizona Upland subunit best known on upper bajadas in the Tucson Basin, the "classic" Sonoran Desert landscape. Structurally complex with a diversity of lifeforms, from towering saguaros to a variety of tree legumes, halfshrubs, and herbaceous annuals and root perennials, 90 percent of this series appears on bajadas and sloping plains. It is the best-watered, least desertlike series in the Arizona desertscrub formation. The jojoba-mixed shrub series is often at the Sonoran Desert's upper limits, particularly in the northeastern reaches. Although it physiog- nomically appears like chaparral, its species do not mix with chaparral species per se. At the northern reaches, a crucifixion thorn (Canotia holocantha) co- dominates with creosote at moderate elevations in the Verde, Gila, and Has- sayampa valleys, but it mixes with chaparral at slightly higher elevations. Near Safford, this crucifixion thorn-creosote series also mixes with Chihua- huan elements to form generic desertscrub. At lower elevations, particularly in the west, the creosote-bursage series forms the dominant community in the Lower Colorado Subdivision. Struc- turally simple, and poorer in perennial species diversity compared to the Arizona Upland, this community can nevertheless be rich in ephemeral an- nuals that flower, seed, and die after rains. On lowlands subject to flooding and salinization, saltbush (Atriplex) dominates with other compact shrubs. The saltbush series extends up the Gila River at least as far as Snaketown. Close to the Lower Colorado River itself, a creosote-big galleta grass (Hilaria rigida) series occurs on sandy soils. The most complex series in the Lower Colorado Subdivision is a mixed scrub series that occurs on the granitic and volcanic outlying mountains. Although little-leaf paloverde and saguaros may be rare or absent, other tree legumes, desert lavender (Hyptis), tree Nolinas, Agave, and jojoba may be present. This vegetation series is present near waterholes and ruins in the Tinajas Altas. Along a gradient of increasing aridity in the Sonoran Desert, ephemerals make up a larger percentage of the total flora, as perennials decrease in diversity. Wildlife densities and diversity also decrease on this gradient (Van- der Wall and MacMahan 1984). However, many of the ephemerals produce energy-rich seeds (Felger and Nabhan 1976), which were important food sources, and the shrubs are well endowed with secondary compounds (ter- penes, etc.) of medicinal value. To understand the full complement of environmental variability with which the Hohokam dealt prehistorically, it must be noted that at times they drew upon areas outside the Sonoran Desert proper for their sources of raw ma- terials, foods, and refined products. Guayule (Parthenium argentaturn) and 268

DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 41 perhaps even peyote (Lophophora williamsii) have been historically and/or pre- historically traded into this region. Ethnographically, chiltepines ( annum), wild beans (Phaseolus spp.), and bellotas (Quercus emoryi) have been traded in or sought in higher elevations, mostly to the south of Hohokam territory. Exogenous resources may have been used by the Hohokam occasionally to buffer themselves from seasonal failures in desert plant production. Higher elevation resources within Hohokam territory also may have served this pur- pose in addition to providing dietary variety. Without animal transport, how- ever, acquisition in bulk of both exogenous and high elevation resources would have been more difficult, probably less regular than in historic times, and likely restricted to high-value items such as materials for medicinal, psy- chotropic, or ceremonial use. The strength of the primary orientation in Ho- hokam subsistence toward basin interiors and the low slopes of bordering mountains is attested by the infrequent archaeological recovery of extraneous biotic remains and by low quantities in those instances when they are present (Gasser 1981). In a geographically diversified compilation of 1,963 flotation- samples (Milcsicek, in preparation), a significant record of resources unob- tainable in these settings was present only among remains from the Santa Rita Mountains, at atypically high elevations on the edge of Hohokam ter- ritory.

DISTRIBUTION OF WILD PLANT RESOURCES

Compared to the Chihuahuan Desert and the northern tropics of Mexico, the Sonoran Desert has a higher ratio of food plants to medicinal plants in its flora (Nabhan 1988). Perhaps one ecological explanation for this is that the Sonoran Desert's high percentage of ephemeralized annual plants (many of them winter-blooming) generally lack bitter or toxic secondary compounds to protect them from insect and vertebrate predators. Because they mature and set seed so rapidly, they escape much of the predation with which longer lived plants must contend. In the tropics and the Chihuahuan Desert, plants with terpenes and alkaloids as feeding deterrents make up a higher percentage of their respective floras; even in the Sonoran Desert, such plants are the typical sources of medicine. Within the Sonoran Desert region north of Mexico, there are more than 250 native plant species that have served as food for Native Americans, ac- cording to ethnographic documentation summarized by Hodgson (1982, in press). It is remarkable to many observers that even in the hottest, driest portions of the Sonoran Desert (e.g., Tiburon Island, the Pinacate), edible calories or available food energy was not so much a limiting factor as potable water. Desert ephemerals have extremely high ratios of reproductive to veg- 269

42 Suzanne K. Fish and Gary P. Nabhan

etative biomass; in other words, they efficiently translocate much of their productivity into nontoxic, energy-rich seeds (Felger and Nabhan 1976). Not surprisingly then, the seeds of desert plants are the most frequently documented plant product used as food ethnohistorically. More than 100 species in the northern Sonoran Desert produce edible and harvestable seeds (including grass caryopses or "grains"); more than 50 species produce flesh fruit; and more than 75 species produce leaves, used primarily when immature (Hodgson, in press). Roots, rhizomes, flower stalks, flowers, flower buds, and stem exudes were also used as food by native cultures of the Sonoran Desert in prehistoric and historic times. Although the ethnographically documented plants of the region constitute an inventory far larger than the number of species found to date within contexts of use in Hohokam sites, such an inventory can be considered as a baseline of potential resources accessible to the Hohokam. Why they may have failed to use plants amply utilized by the O'odhara or by Yuman cultures, or vice-versa, constitutes an interesting question with regard to archaeological recovery, and cultural continuity and change. Rea (1981a) has used contrasts of culture-specific foods and tabooed plants to test in a preliminary way hypotheses regarding a Piman (Osodham) versus Yuman continuum with the Hohokam. Most foods attributed to the Hohokam are also eaten by both Piman and Yuman speakers today, and the number of exceptions encountered to date are in no way definitive. Aside from cultural differences in plant utilization, there is geographic variation in the distribution of potential resources within Hohokam country. Figures 2.4, 2.5, and 2.6 show distributions of saguaro, mesquite, and palo- verde and the more restricted occurrence of agave and yucca at higher ele- vations. Three trends emerge from overall patterns for major Hohokam resources: (1) the majority of these resource plants are most densely and continuously distributed along riparian corridors; (2) to the southwest and with decreasing rainfall, increasing temperature, and increasing unpredicta- bility of precipitation, the major resource plants become more patchy in dis- tribution and less available overall; and (3) to the northeast,the Mogollon Rim and other uplands most abruptly limit the distribution of these resource plants. Thus greater similarity would be expected between the subsistence of the Hohokam and desert dwellers to the south and west, than with those above the Mogollon Rim.

MODIFIED LANDSCAPES AND THEIR RESOURCES

In the introduction to this chapter, the role of the Hohokam as farmer- gatherers in anthropogenic vegetation was emphasized. In large part, insight regarding this behavior derives from ethnographic studies of plant use by 44

270

DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 43

1/4 ° 370_

330_ ARI ZONA • • • •• • • ••• • • • •• • • • • • • • • • :• ••• • • • • • • ••• • • • • Rivera • • O••O••• • • ••• •• •• • • • • •••• • • • colt ••• • • • • ''' • •• • • • • • • •• IV •• • • • •• • • • • 067/0 • • • f 33o_ • • • i • • • • ...i. • a • GO •• • • • ••• • •• • •• • • .13 • • • • • • • • • • •• • •• • • • -.1 ••••• •• •• •••• • •• •• •••••• ••• • • ••• • 5;1 woo • •• • F. • • • 13 • • • • • • • . • , • • Z-,-i

31 °- - Figure 2.4 Distribution of saguaro in southern Arizona. groups who intensively harvest wild plants (as well as crops) in field, ditch, hedgerow, abandoned field, and dooryard garden microhabitats (Rea 1981b, 1983; Crosswhite 1981; Nabhan et al. 1983). However, none of these plants concentrated in anthropogenic contexts is found exclusively in Native Amer- ican fields; they are also common where natural disturbance occurs along watercourses or on valley plains. Current distributions of utilized species are not sufficient indicators of prehistoric availability if the Hohokam capacity is acknowledged for trans- porting, transplanting, irrigating, burning, or otherwise managing wild plants for their desirable products. While the presence of particular weedy or semi- cultivated species in the archaeobotanical record is not definitive evidence for origin in anthropogenic communities, patterns of quantitative contrast and

271

44 Suzanne K. Fish and Gary P. Nabhan

/4'

37 ° -

o

35 ° - 00 0 0 xx x 000 00 X o to X X xiqt St XX x 0 00 o 00 x x9 S2° xR _ x sto x 0pi° 2 2 2 °x x x x x RR R o otxx xxxx * xxv x*x * x 0 x X XX )0XX 00 0 0 X AWer * X X • x xx R 0 0 o X *X*X 00 xxs x so v salt o x R x x 900 xv x x o * o ô Rio x )0 x)o o x,0 61/0 0 33 °- X 0xX x X.4 ( tV4 X o• o • oo Gli 0 XX 0 X *XX* • o )P XS„o o X, .. X o o o xo o o o X ° XR ei) ,o» X .-, x X o o o X' o fpflO 0 x000 o o Neu o RR ; R R °x o o IP OX • 11 0 0 RR R oo R o 0 00 — xx R .?.) X vo xt o z.7 . oo o • O o 00 o o • o oo o o Prosopis velplino 00 • Prosopis pubescens • 000 o o X Cercicliam microphyllum o

1 I I I _ 3/°- I I Figure 2.5 Distributions of mesquite and little-leaf paloverde in southern Arizima.

of contextual correspondence have been increasingly identified in recent stud- ies. These provide the basis for reconstructing environmental alterations that may have included both intentional manipulation and unintentional en- hancement of species other than cultigens. The nature and degree of Hohokam modification of natural landscapes would have varied according to local environmental attributes and cultural practice. Agricultural activity was likely a primary factor in the most far- reaching modifications, and would have produced a wide range of effects (Bohrer 1970; Gasser 1982; Fish 1984, 1985; Miksicek 1984, 1988). Runoff farm- ing on floodplains, locationally tied to the environs of drainages, might be expected to create the least divergence from surrounding plant communities 272

DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 45

/140 /12° uci 37°-

X

ARIZONA

X X X 35 °- X X X x x XXXx X xx o X X 0 0000 80 X o 016X oo oo 0 0 River o o Salt o o

33° -

o Agave clown • Agave ochnsri X Yucco boccota

3/ 0-

Figure 2.6 Distributions of agave and yucca in southern Arizona. in naturally disturbed riparian habitats. Even in these situations, manipula- tion through such documented aboriginal practices as selective removal of unwanted species, reseeding and tending of utilized ones, or introduction of nonlocal taxa could have altered distributions toward advantageous ends. Canals transporting water laterally as well as downstream for miles beyond drainage sources had the greatest potential for creating biotic conditions dis- similar to locally prevailing natural ones. Riverine canals traversed flatter portions of lower basins among xeric saltbush or creosote communities that contrasted with irrigated fields in the absence of both supplemental water and surface disturbance. Surface preparation to increase infiltration, concen- tration of surface runoff, and diversion of ephemeral drainages in terrace, bajada, or hillslope agricultural complexes also generated contrastive growth 273

46 Suzanne K. Fish and Gary P. Nabhan

conditions, although augmenting moisture less than canal irrigation. Even without management, weedy plants of these latter two agricultural contexts would have differed at least in densities from adjacent uncultivated land. A role for fire in landscape alteration and management has also been suggested (Miksicek 1984; Bohrer 1971). Pollen assemblages that differ from natural analogs in the distribution of weedy taxa, and suites of similar species in charred remains have been rec- ognized as reflecting anthropogenic plant communities in diverse residential and agricultural provenances (e.g., Fish 1984, 1985, 1987; Bohrer 1984; Mik- sicek 1987, 1988; Gish 1987). The modified plant communities of prehistoric occupations have been linked to corresponding faunal distributions, partic- ularly for those mainstays of Hohokam cuisine, the lagomorphs (Szuter 1984, 1986). Animal as well as plant resources appear to have been abundantly harvested in fields and settlement margins. Classification of archaeobotanical remains into the exhaustive dual cate- gories of wild plants and cultigens may provide inadequate insight into an important continuum among species and activities. Active intervention in- volving a variety of plants beyond those recognized as cultigens is probable in Hohokam agriculture. Proposed species include mesquite (Prosopis), cholla (Opuntia spp.), hedgehog cacti (Echinocereus engelmannii), wolfberry (Lycium spp.), hog potato (Hoffmanseggia densi flora), (Amaranthus paltnert), c.henopods (Chenopodium berlandieri, C. murale, Monolepis nutalliana, Atriplex zvrightii), little barley (Hordeum pusillum), tobacco (Nicotiana trigonophylla), and spiderlirtg (Boerhaavia spp.). Numerous combinations of happenstance and design are probably em- bodied in the array of noncultigens now thought to have been productively enhanced or concentrated in cultural contexts by the Hohokam. Among this group are plants such as mesquite that may have been differentially spared in field clearing, become dense in hedgerows, thrived on canal seepage, or been tended and selectively harvested in adjacent natural settings. Species such as cholla may have been transplanted to dooryard gardens or fields as a "crop," to out-of-the-way spots among habitations and fields, or employed as residential fencing. Chenopods and spiderling are representative of weedy herbaceous plants that may have received focused attention or none at all in fields and in other culturally disturbed habitats. All species suspected of increasing in abundance in Hohokam-managed fields and vegetation could also be found in quantity in some undisturbed locales. Most exhibit no morphological characters attributable to intervention or domestication. Quantitative or qualitative variance has been examined in a few instances for amaranths (Amaranthus hybridus) (Miksicek, in preparation) and little barley (Hordeum ptaillum) (Bohrer 1984; Adams 1987). However, evidence for even active cultivation as in the case of agave (Fish et al. 1985) may not be reflected in distinctive attributes of associated plant remains. 274

DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 47

Replicable quantitative and contextual data and innovative approaches are needed to refine current understanding of this sphere of Hohokam subsistence and environment.

WATER CONTROL FOR AGRICULTURE

Recent attempts to define the kinds of agricultural systems within arid zones, and to classify those of the Sonoran Desert accordingly (e.g., Vivian 1974; Lawton and Wilke 1979; Nabhan 1979), suffer from either ambiguous use of terms, or from noncomprehensivenas (e.g., Nabhan 1985; Rankin and Katzer, in press). Rankin and Katzer (in press) have progressed furthest in straightening out the current tangle of terms by considering four variables: (1) whether a field receives an augmented water supply over and above the precipitation that falls directly on its surface; (2) whether an area above or within a field has had surface preparation to generate more runoff or con- versely to increase infiltration; (3) the geomorphic setting (alluvial fan, swale, floodplain terraces) of field location; and (4) whether a crop has had water diverted to it, or is in the direct path of natural inundation. A fifth variable is alluded to in their discussion but does not fully enter into their classification: the duration of waterflow along a stream course (perennial, intermittent, or ephemeral). In addition, Rankin and Katzer correlate field surface areas with watershed or catchment sizes, and find a positive relationship between these two variables.

ETHNOGRAPHIC TECHNOLOGY Because all aspects of productive technology are not preserved archaeo- logically, ethnographic practices amPlify direct evidence for Hohokam culti- vation. If a multidimensional matrix of interaction were reconstructed from the five variables in the preceding dassification system, an ethnographic or prehistoric example of nearly each combination could be found within the Southwest. Perhaps the only set of variables for which a good ethnographic example is not known in the Sonoran Desert proper is that of dry farming (a nonaugmented water supply) in any of the geomorphic habitats (Nabhan 1979). Although modem farmers in the Avra Valley have been known to harvest a wheat crop by winter dry farming on normally irrigated fields, most ethnographic examples show some effort at augmenting soil moisture by diversion or surface preparation. Cooke and Reeves (1976) contend that his- toric Papago (Tohono O'odham) did both, by clearing vegetation to increase runoff in watershed catchment areas upstream from fields, then channelizing and diverting this augmented supply onto their annual crops. Water control features of traditional Sonoran Desert farmers include ca- nals, shallow ditches, rock or brush diversion weirs, rock or brush water 275

48 Suzanne K. Fish and Gary P. Nabhim spreaders, living cottonwood and willow fencerow silt traps, multicourse cobble terraces, and single-course alignments. The same materials may be used for different functions in different geomorphic locations; mesquite posts with interwoven brush are still used by the Tohono O'odham as diversion weirs from watercourses to fields, water spreaders at the ak-chin location just above fields, water spreaders in fields, and baffles just above fields. With all runoff agricultural systems, the greater the watershed size about a field system, the greater the probability of sufficient soil moisture being provided (Hack 1942). However, social organization for water control must be sophisticated, and even on ephemeral watercourses supplying more than 200 ha of fields, the Tohono O'odham had ditch bosses coordinating irrigation crews much like those of the Pima on the Gila River (Spicer 1943). Thus, there was a gradient in sophistication of water control from ephemeral through intermittent to perennial streams, not a dear-cut separation of system sizes and functions.

HOHOKAM AGRICULTURAL TECHNOLOGY Agricultural technology represents a major arena of interaction between the Hohokam and their Sonoran Desert environment. Boundaries encom- passing canal networks in the broad valleys of perennial rivers have long been perceived as a cultural watershed, dividing large-scale irrigators from their less fortunate contemporaries. These concepts are embodied in the terms riverine and nonriverine or desert Hohokam, with the implication that social organization as well as productive capacity was linked with the requirements of irrigation. The dichotomy has become less sharply drawn as ongoing investigations in nonriverine areas have revealed evidence of a previously unrecognized degree of cultural elaboration (e.g.., Wilcox and Sternberg 1983; Gregory 1987; Fish et al. 1989; Rice 1987; Wilcox and Ciolek-Torrello, 1988). Similarly, research has confirmed the presence of substantial canals on larger intermittent wa- tercourses (Bernard-Shaw 1988; Kinkade and Fritz 1975; Scantling 1940), and documented cases of multimile systems as on the Santa Cruz (Fish et al. 1989), New River (Doyel 1984), and Queen Creek (Dart 1983). Technical understanding of Hohokam irrigation has been refined for the large networks of the Salt and Gila. To basic knowledge of the systems (e.g., Woodbury 1960, 1961; Haury 1976; Midvale 1965, 1968; Masse 1981; Herskovits 1981) has been added a new wealth of detail. System outliers from pre-urban maps have been augmented by aerial photographs, documentary study, and correlation with archaeological settlement pattern (e.g., Upham and Rice 1980; Crown 1987; Masse 1987; Howard 1987; Nicholas and Neitze11984; Cable and Mitchell 1988). Two recent analyses (Ackerly et al. 1987; Nials et al. 1989) of overall system capacity in the Phoenix Basin concur in a total for irrigated acreage, down- 276

DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 49

scaled from some previous estimates, to a range between 30,000 and 60,000 acres (12,100 to 24,200 ha). Multiple excavated examples of canal interruption, abandonment, and rebuilding have increased appreciation of the dynamic nature of these systems and cautioned against assumptions for contemporary use of all detectable segments, a point strongly made by Ackerly et al. (1987). Knowledge is rapidly accumulating concerning environmental correlates of canal construction, engineering attributes, and ancillary structures such as headgates, settling ponds, and canal junctures (e.g., Niais and Fish 1988; Ackerly et al. 1987; Masse 1987; Dart 1986). Knowledge of irrigation systems should be dramatically extended by a number of ongoing excavation projects

in the Phoenix ar— ea. Appreciation is emerging for the role of nonriverine farming in the total pattern of Hohokam production. Runoff technologies across widely diverse topography, reported formerly in anecdotal fashion, are now being studied as segments of larger subsistence systems and in relation to other aspects of settlement pattern (e.g., Crown 1987; Doyel 1984; Fish et al. 1985; Spoerl and Gumerman 1984; Rankin and Katzer, in press). Fine-grained study of tech- nological detail and systematic recording of component or system extent pro- vide a basis for estimating such parameters as acreage, labor, and yield (Fish et al. 1985; Fish et al. 1989; Crown 1987; Doyel 1984). Recognition of envi- ronmental variables affecting cultivation has clarified productive opportu- nities, limitations, and Hohokam strategies in these areas (e.g., Field 1985; Waters and Field 1986; Fish et al. 1985; Wilson 1985; Rankin and Katzer, in press). Regional orientations and systematic recording or compilation of settle- ment reveal integrated subsistence activities spanning environmental diver- sity. A zonal concept of settlement and landscape organization with sociopolitical implications has been described for inclusive systems in nonriverine regions (Fish et al. 1989; Rice 1987; Doye11984) and along the Gila River (Crown 1987). Agricultural technologies, crops, and natural resources varying with envi- ronmental gradients, particularly across basin profiles, appear to have func- tioned as components of broadly based subsistence and settlement units rather than as independent and localized entities. Zonal organization entails an appreciable degree of intrabasin productive differentiation, exchange, and integrative interaction. Such organization appears as central to the cohesion of nortriverine multisite communities, as canal networks are to their riverine counterparts on the Salt and Gila. It is becoming increasingly clear that the maximum variety of agricultural technologies and settings were utilized in prehistoric times. Reduction during the historic period is likely the result of population loss and rearrangement, alternative economic pursuits, and the absence of farmers in regions and zones of archaeological occupation, particularly upland locales. Adequate analogs are lacking for agricultural complexes on bajadas, hillslopes, and smaller drainages that combine terraces, rock piles, checkdams, and diver- 277

50 Suzanne K. Fish and Gary P. Nabhan

sions. The greatest density and variety of features has been encountered in areas of higher potential for such rainfall-dependent technologies. Although influenced by intensity of archaeological investigation, distributions tend to correlate with increasing summer and annual precipitation to the north, east, and south of the Phoenix Basin. At present, chronological trends in subsistence and technology cannot be analyzed against independent sequences of comprehensive, long-term en- vironmental evidence derived within the boundaries of Hohokam territory. Dendrochronologically sensitive timbers are irregularly incorporated in Ho- hokam dwellings, pollen distributions are strongly shaped by herbaceous types responsive to cultural habitats, and other plant remains in sites have passed through a cultural filter. Trajectories of change initiated by post-contact forces illuminate process dynamics of change (e.g., Hastings and Turner 1965; Dobyns 1981; Rea 1983; Cooke and Reeves 1976), but cannot characterize sequential environmental patterns over archaeological time. Recent studies spanning substantial intervals represent encouraging ex- ceptions to the paucity of systematic prehistoric data. Settlement patterns in adjacent portions of the riverine environment have been tied to change over time in channel morphology and hydrological characteristics of the Santa Cruz (Waters 1988). In a major contribution, Graybill (1985, Graybill and Niais 1989) has analyzed relevant dendrochronological information from the Salt River watershed and established a link between conditions in these areas and strearnflow in the Phoenix Basin. An absolute chronology of streamflow, in- fluencing irrigation and flooding potential, provides a temporal resolution unavailable with other techniques and spans much of Hohokam occupation. Studies of geological evidence for flooding in the Phoenix area concur in the general timing of major events and add detail to local extent (Ely and Baker 1985, 1988).

ETHNOGRAPHIC ANALOGY AND HOHOKAM ENVIRONMENTAL RELATIONSHIPS

Interactive connections between environmental variables and archaeolog- ical remains cannot always be bridged by reference to ethnographic analogy. In cases lacking historic parallels, interpretation must be generated largely from patterning in these two classes of evidence. Divergence between eth- nographic and prehistoric expressions is commonly acknowledged for the largest irrigation networks. Recognition of the magnitude and diversity of runoff agriculture and its importance in regional economies has been delayed by a lack of adequate analogs for pertinent methods of cultivation and as- sociated features. Conversely, the opportunity to actively learn from ethno- graphic practice is far from exhausted as indicated by a number of recent studies (e.g., Greenhouse et al. 1981; DoeIle 1976; Nabhan et al. 1983; Doolittle 278

DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 51

1987; Rea 1983; Felger and Moser 1985). Many earlier historic documents also await research. Hohokam use of agave illustrates the need for interpretive elaboration beyond immediate analogy. Many records for the O'odham describe acqui- sition through the harvesting of wild stands. This mode was commonly as- sumed to account for Hohokam remains, although agave had been recovered in Hohokam roasting pits at a distance from natural populations (Haury 1945:39; Fewkes 1912; Hayden 1957:103) and is widely cultivated in Mexico. Culturally affected distributions and suggestions concerning cultivation were advanced by Gentry (1972, 1982), Ford (1981), Mirtnis and Flog (1976), and Crosswhite (1981). Distributions of plant parts supporting an inference of cultivation, and agricultural features in conjunction with agave remains were subsequently noted (Miksicek 1984; Gasser and Milcsicek 1985; P. Fish et al. 1984). Conclusive evidence for production rather than gathering was estab- lished by the replicable association between agricultural features and plant remains, processing facilities, and appropriate artifactual assemblages (Fish et al. 1985). The identity of a cultivated species, Agave murpheyi, and possibly additional ones, has been greatly strengthened by discovery of relict popu- lations among agricultural features in the Tonto and New River basins (G. R. Delamater, 1989, personal communication). A second example of the need for caution in resorting to literal analogy is illustrated by the relationship of Tohono O'odham settlement to potable water. In the post-1850 period of most extensive recording, annual movement between two settlements was a dominant pattern. Residence at villages with ak-chin fields near valley bottoms could be sustained only for the duration of potable water in reservoirs following summer rains. Water was also a limiting factor on group size at settlements; Underhill (1939:58-59) notes that devel- opment of sustained water sources decreased annual movement. At the time of ethnographic observation and for many years previously, the Tohono O'od- ham had possessed cattle and . A population of approximately 5,000 persons was estimated to own 20,000 cattle (Underhill 1939:29), also requiring water. Comparison of consumption for cattle and people in summer (Adolph 1947; McKee and Wolf 1963; Tom Wegner, 1989 personal communication) reveals that total water requirements were multiplied by a factor of 45 or more in order to support both cattle and people. Thus, the relationship between O'odham settlement and water may diverge significantly from that of the Hohokam in similar environments, with concomitant differences in strategies for agriculture and wild resource procurement.

HOHOKAM ENVIRONMENT IN SUMMARY

The Sonoran Desert occupied by the Hohokam is one of the truly rich areas for the gatherer in North America in terms of multiseason abundance, 279

52 Suzanne K. Fish and Gary P. Nabhan diversity, and storability of plant foods. Bimodal rainfall in southern Arizona supports a distinctive array of productive arboreal and succulent perennials in addition to a wide variety of seedy annuals. Linear valleys create elevational diversity within transverse distances. Most resources could have been ac- quired on a routine basis from settlements near long-term water at river or mountain edge. Diminished desert arboreals in drier central and western regions are countered in the Phoenix Basin and further downstream along the Gila by concentrated resources in riparian borders of perennial rivers. The Hohokam created further diversity in the anthropogenic plant com- munities surrounding their settlements. The degree of management toward desired configurations of wild plants is currently under investigation. How- ever, it appears that the culturally modified landscape furnished a significant increment of gathered species as well as small game. Two major divisions in Hohokam agricultural orientation have long been recognized. Irrigators from perennial rivers had access to a more extended regime of agricultural water than did farmers with irrigated and nonirrigated fields dependent on seasonal rainfall. Topographic and streamflow prereq- uisites for large-scale riverine irrigation coincided with lower precipitation along desert reaches of the Salt and Gila. In surrounding areas to the north, south, and east, higher summer and annual rainfall created a distinct advan- tage for alternative technologies that also supported a Hohokam lifestyle. Documentation of the variety and productivity of agricultural complexes away from the major rivers has increased rapidly in recent years. These modes of cultivation can now be seen as parts of larger subsistence systems, integrating production across diverse environments.

ACKNOWLEDGMENTS

We gratefully acknowledge Raymond M. Turner of the United States Geo- logical Survey for shared data and insights on the Sonoran Desert.

REFERENCES CITED

Ackerly, Neal, Jerry B. Howard, and Randall H. McGuire 1987 La Ciudad Canals: A Study of Hohokam Irrigation Systems at the Community Level. Arizona State University Anthropological Field Studies 17. Adams, Karen 1987 Little Barley (Hordeum pusillum Nutt.) as a Possible New Wild Domesticate. In Specialized Studies in the Economy, Environment, and Culture of La Ciudad, 280

DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 53

edited by J. Kisselburg, Glen Rice, and B. Shears, pp. 203-238. Arizona State University Anthropological Field Studies 20. Asplund, K., F. Baucom, W. Bayham, T. Cordery, D. Haywood, W. Hinter, A. Laurenzi, K. Reichhardt, and R. Szaro 1988 Riparian Classification for Arizona. Arizona Riparian Council, Riparian Clas- sification Inventory Committee, Phoenix. Adolph, E. F. 1947 Physiology of Man in the Desert. Interscience Publishers, Inc., New York. Bernard-Shaw, Mary 1988 Hohokam Canal Systems and Late Archaic Wells: The Evidence from the Los Morteros Site. In Recent Research on Tucson Basin Prehistory: Proceedings of the Second Tucson Basin Conference, edited by W. Doelle and P. Fish, pp. 153-174. Institute for American Research Anthropological Papers 10. Betancourt, Julio and Raymond Turner 1988 Historic Arroyo-Cutting and Subsequent Channel Changes at the Congress Street Crossing, Santa Cruz River, Tucson, Arizona. Arid Lands: Today and Tomorrow, edited by E. Whitehead, C. Hutchinson, B. Timmerman, and Robert Varady, pp. 1353-1372. Westview Press, Boulder. Bohrer, Vorsila L. 1970 Ethnobotanical Aspects of Snaketown, A Hohokam Village in Southern Arizona. American Antiquity 35:413-430. 1971 Paleoecology of Snaketown. The Kiva 36:11-19. 1984 Domesticated and Wild Crops in the CAEP Study Area. In Prehistoric Cul- tural Development in Central Arizona: Archaeology of the Upper New River Region, edited by P. Spoerl and G. Gumerman, pp. 183-259. Southern Minois University Center for Anthropological Investigations Occasional Paper 5. Cable, John, and Douglas Mitchell 1988 La Lomita Pequefia in Regional Perspective. In Excavations at La Lomita Pequefia: A Santa Cruz/Sacaton Phase Hamlet in the Salt River Valley, edited by Douglas Mitchell, pp. 395-446. Soil Systems Publications in Archaeology 10. C_astetter, Edward F., and Willis H. Belt 1942 Pima and Papago Indian Agriculture. University of New Mexico Inter-Amer- ican Studies 1. Cooke, Ronald U., and Richard W. Reeves 1976 Arroyos and Environmental Change. Oxford Research Studies in Geography. Clarendon Press, Oxford. Crosswhite, Frank 1981 Desert Plants, Habitat and Agriculture in Relation to the Major Pattern of Cultural Differentiation in the O'odham People of Southern Arizona. Desert Plants 3:47-76. Crown, Patricia 1987 Classic Period Hohokam Settlement and Land Use in the Casa Grande Ruins Area, Arizona. Journal of Field Archaeology 14:147-162. Dart, Allen 1983 Prehistoric Agricultural and Water Control Systems along the Salt-Gila Aqueduct. In Hohokam Archaeology along the Salt-Gila Aqueduct, Central Ari- zona Project, vol. 3, Specialized Activity Sites, edited by L. Teague and P. Crown, pp. 345-573. Arizona State Museum Archaeological Series 150. 281

54 Suzanne K. Fish and Gary P. Nabhan

1986 Sediment Accumulation along Hohokam Canals. The Kiva 51:63-84. Dobyns, Henry 1981 From Fire to Flood: Historic Human Destruction of Sonoran Desert Riverine Oases. Ballena Press, Socorro. DoeIle, William 1976 Desert Resources and Hohokam Subsistence: The Conoco Florence Project. Arizona State Museum Archaeological Series 103. Doolittle, William 1987 Pre-Hispanic Occupance in the Valley of Sonora, Mexico: Archaeological Confir- mation of Early Spanish Reports. Anthropological Papers of the University of Arizona 48. Doyel, David 1984 Sedentary Period Hohokam Paleo-Economy in the New River Drainage, Central Arizona. In Prehistoric Agricultural Strategies in the Southwest, edited by S. Fish and P. Fish, pp. 35-52. Arizona State University Anthropological Research Papers 33. Ely, Lisa, and Victor Baker 1985 Reconstructing Paleoflood Hydrology with Slackwater Deposits: Verde River, Arizona. Physical Geography 6:103-126. Ely, Lisa, J. E. O'Connor, V. R. Baker 1988 Paleoflood Hydrology of the Salt and Verde Rivers. In Friends of the Pleis- tocene, Rocky Mountain Cell, Fall Fieldtrip Guidebook, assembled by L. An- derson and L. Piety. Arizona State University, Tempe. Felger, Richard and Mary Moser 1985 People of the Desert and Sea: Ethnobotany of the Seri Indians. University of Arizona Press, Tucson. Felger, Richard and Gary Nabhan 1976 Deceptive Barrenness. Ceres/FAO Journal 9:34-39. Fenneman, Nevin 1931 Physiography of the Western United States. McGraw-Hill, New York. Fewkes, Jesse Walter 1912 Casa Grande, Arizona. 28th Annual Report of the Bureau of Ethnology, 1906— 1907. Washington, D.C. Field, John 1985 Depositional Facies and Hohokam Settlement Patterns on Holocene Alluvial Fans, Northern Tucson Basin. M.S. thesis, Department of Geosciences, University of Arizona, Tucson. Fish, Suzanne K. 1984 The Modified Environment of the Salt-Gila Aqueduct Project Sites: A Paly- nological Perspective. In Hohokam Archaeology along the Salt-Gila Aqueduct, Central Arizona Project, vol. 7: Environment and Subsistence, edited by L. Teague and P. Crown, pp. 39-52. Arizona State Museum Archaeological Series 150. 1985 Prehistoric Disturbance Floras of the Lower Sonoran Desert and Their Implications. In Late Quaternary Vegetation and Climates of the American South- west, edited by B. Jacobs, P. Fall, and O. Davis, pp. 77-88. American Association of Stratigraphic Palynologists Contribution 16. 1987 Marana Sites Pollen Analysis. In Studies in the Hohokam Community of Mar- ana, edited by Glen Rice, pp. 161-170. Arizona State University Anthro- pological Field Studies 15. 282

DESERT AS CONTEXT: THE HOHOKAIA ENVIRONMENT 55

Fish, Suzanne K., Paul R. Fish, and Christian Downum 1984 Hohokam Terraces and Agricultural Production in the Tucson Basin. In Prehistoric Agriculture Strategies in the Southwest, edited by S. K. Fish and P. R. Fish, pp. 55-71. Arizona State University Anthropological Research Papers 33. Fish, Suzanne K., Paul R. Fish, and John Madsen 1985 A Preliminary Analysis of Hohokam Settlement and Agriculture in the Northern Tucson Basin. In Proceedings of the 1983 Hohokam Symposium, ed- ited by A. E. Dittert and D. E. Dove, pp. 75-100. Arizona Archaeological Society Occasional Paper 2. 1989 Classic Period Hohokam Community Integration in the Tucson Basin. In The Sociopolitical Structure of Prehistoric Southwestern Societies, edited by S. Upham, K. Lightfoot, and R. Jewett, pp. 237-268. Westview Press, Boulder. 1990 Sedentism and Settlement Mobility Prior to A.D. 1000 in the Tucson Basin. In Proceedings of the Southwest Symposium, edited by Paul Mixuds and Charles Redman. Westview Press, . Fish, Suzanne K., Paul R. Fish, Charles Miksicek, and John Madsen 1985 Prehistoric Agave Cultivation in Southern Arizona. Desert Plants 7:107-112. Ford, Richard L 1981 Gardening and Farming before A.D. 1000: Patterns of Prehistoric Cultiva- tion North of Mexico. Journal of Ethnobiology 1:6-27. Gasser, Robert 1981 The Plant Remains from the Escalante Ruin Group. In Late Hohokam Pre- history in Southern Arizona, by David E. Doyel. Gila Press Contributions to Archaeology 2. 1982 Hohokam Use of Desert Food Plants. Desert Plants 2:216-235. Gasser, Robert and Charles Miksicek 1985 The Specialists- A Reappraisal of Hohokam Exchange and the Archaeo- botanical Record. In Proceedings of the 1983 Hohokam Symposium, edited by A. E. Dittert and D. E. Dove, pp. 483-498. Arizona Archaeological Society Occasional Paper 2. Gentry, Howard S. 1972 The Agave Family in Sonora. U.S. Department of Agriculture Handbook 399. 1982 Agaves of Continental North America. University of Arizona Press, Tucson. Gish, Jannifer 1987 Structured Diversity on the Resource Base of a Hohokam Village: The Pollen Evidence from La Ciudad. In La Ciudad: Specialized Studies in Economy, En- vironment, and Culture of La Ciudad, edited by V. Kisselburg, G. Rice, and B.Shears, pp. 1-67. Arizona State University Anthropological Field Studies 20. Graybill, Donald A. 1985 Paleodirnate of the Hohokam Area: Problems and Prospects. In Proceedings of the 1983 Hohokam Symposium,' edited by A. Dittert and D. Dove, pp. 29— 44. Arizona Archaeological Society Occasional Paper 2. Graybill, Donald A., and Fred L. Niais 1989 Aspects of Climate, Strearnflow, and Geomorphology Affecting Irrigation Systems in the Salt River Valley. In The 1982-1984 Excavations at Las Colinas: Environment and Subsistence, edited by C. A. Heathington and D. A. Greg- ory, pp. 39-58. Arizona State Museum Archaeological Series 162. 283

56 Suzanne K. Fish and Gary P. Nabhan

Greenhouse, Ruth, R. Gasser, and J. Gish 1981 Cholla Bud Roasting Pits: An Ethnoarchaeological Example. The Kiva 46:227— 242. Gregory, David 1987 The Morphology of Platform Mounds and the Structure of Classic Period Hohokam Sites. In The Hohokam Village: Site Structure and Organization, edited by David Doyel, pp. 183-210. Southwestern and Rocky Mountain Division of the American Association for the Advancement of Science, Glenwood. Hack, John T. 1942 The Changing Physical Environment of the Hopi Indians of Arizona, Reports of the Awatovi Expedition I. Papers of the Peabody Museum of American Ar- chaeology and Ethnology 35. Hastings, James R., and Raymond M. Turner 1965 The Changing Mile: An Ecological Study of Vegetation Change with Time in the Lower Mile of an Arid and Semiarid Region. University of Arizona Press, Tucson. Haury, Emil 1945 The Excavation of Los Muertos and Neighboring Ruins in the Salt River Valley, Southern Arizona. Papers of the Peabody Museum of American Archaeology and Ethnology 24. 1976 The Hohokam: Desert Farmers and Craftsmen. University of Arizona Press, Tucson. Hayden, Julian 1957 Excavations, 1940, at University Indian Ruin. Southwestern Monuments As- sociation Technical Series 5. 1967 A Summary Prehistory and History of the Sierra Pinacate, Sonora, Mexico. American Antiquity 41:274-289. Hendrickson, Dean A., and W. L. Minckley 1984 Cienegas: Vanishing Climax Communities of the American Southwest. Desert Plants 6:131-175. Herskovitz, Robert M. 1981 AZ U:9:46--A Dual Component Hohokam Site in Tempe, Arizona. The Kiva 47:1-98. Hodgson, Wendy C. 1982 Edible Native and Naturalized Plants of the Sonoran Desert North of Mexico. M.S. thesis, Arizona State University, Tempe. University Microfilms, Ann Arbor. In Press Some of My Best Friends Are Edible Plants: Food Plants of the Sonoran Desert. University of Arizona Press, Tucson. Howard, Jerry B. 1987 The Lehi Canal System: Organization of Classic Period Community. In The Hohokam Village: Site Structure and Organization, edited by David Doyel, pp. 211-222. Southwestern and Rocky Mountain Division of the American Association for the Advancement of Science, Glenwood Springs, Colorado. Huckell, Bruce 1988 Late Archaic Archaeology of the Tucson Basin: A Status Report. In Recent Research on Tucson Basin Prehistory: Proceedings of the Second Tucson Basin 284

DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 57

Conference, edited by William Docile and Paul Fish, pp. 57-80. Institute for American Research Anthropological Paper 10. Kinkade, Gay M., and Gordon Fritz 1975 The Tucson Sewage Project: Studies of Two Archaeological Sites in the Tucson Basin. Arizona State Museum Archaeological Series 64. Lawton, H., and P. Wilke 1979 Ancient Agricultural Systems in Dry Regions. In Agriculture in Semi-Arid Environments, edited by A. Hill, G. C_annell, and H. Lawton. Springer- Verlag, New York. Lee, Willis T. 1905 Underground Waters of the Salt River Valley, Arizona. U.S. Geological Survey Professional Paper 252. Lowe, Charles, and David Brown 1982 Introduction. Desert Plants 4:17-24. McDonald, James E. 1956 Variability of Precipitation in an Arid Region: A Survey of Characteristics for Arizona. University of Arizona Institute of Atmospheric Physics Technical Report 1. McGinnies, William 1981 Discovering the Desert: Legacy of the Carnegie Desert Botanical Laboratory. Uni- versity of Arizona Press, Tucson. McKee, Jack E., and H. Wolf 1963 Water Quality Criteria. U.S. Department of Health, Education, and Welfare, Division of Water Supply and Pollution Control, Sacramento. Masse, W. Bruce 1981 Prehistoric Irrigation Systems in the Salt River Valley, Arizona. Science 214:408-415. 1987 Archaeological Investigations of Portions of the Las Acequias—Los Muertos Irri- gation System. Arizona State Museum Archaeological Series 176. Midvale, Frank 1965 Prehistoric Irrigation of the Casa Grande Ruins Area. The Kiva 30:82-86. 1968 Prehistoric Irrigation in the Salt River Valley, Arizona. The Kivu 34:28-34. Miksicek, Charles 1984 Historic Desertification, Prehistoric Vegetation Change, and Hohokam Subsistence in the Salt-Gila Basin. In Hohokam Archaeology along the Salt Gila Aqueduct, Central Arizona Project, vol. 7: Environment and Subsistence, ed- ited by L. Teague and P. Crown, pp. 53-80. Arizona State Museum Ar- chaeological Series 150. 1987 Late Sedentary—Early Classic Period Hohokam Agriculture: Plant Remains from the Marana Community Complex. In Studies in the Hohokam Community of Marana, edited by Glen Rice, pp. 197-216. Arizona State University Anthropological Field Studies 15. 1988 Rethinking Hohokam Paleoethnobotanical Assemblages: A Progress Report for the Tucson Basin. In Recent Research on Tucson Basin Prehistory: Proceedings of the Second Tucson Basin Conference, edited by William Doelle and Paul Fish, pp. 47-56. Institute for American Research Anthropological Papers 10. In Prep. Plant Remains from the Land of the Stone Hoe: A Century of Hohokam Archaeo- botany. Ph.D. dissertation. Office of Arid Land Studies, University of Ari- zona, Tucson.

285

58 Suzanne K. Fish and Gary P. Nabhan

Minnis, Paul, and Stephen Flog 1976 A Study of the Site Specific Distribution of Agave parryi in East-Central Arizona. The Kiva 41:299-308. Nabhan, Gary P. 1979 The Ecology of Floodwater Farming in Southwestern North America. Agro- Ecosystems 5:245-255. 1983 Papago Fields: Arid Lands Ethnobotany and Agricultural Ecology. Ph.D. disser- tation. Office of Arid Land Studies, University of Arizona, Tucson. 1986 Ak-chin "Arroyo Mouth" and the Environmental Setting of the Papago Indian Fields in the Sonoran Desert. Applied Geography 6:61-75. 1988 Ethnobotany of the Biosphere Reserves of the U.S./Mexico Borderlands: Preliminary Assessment of Trends. Pager presented at the First Conference on the El Cielo Biosphere Reserve, Ciudad Victoria, Tamaulipas. Nabhan, Gary P., A. M. Rea, K. L. Reichhardt, E. Mellink, and C. F. Hutchinson 1983 Papago Influences on Habitat and Biotic Diversity: Quitovac Oasis Eth- noecology. Journal of Ethnobiology 2:124-143. Niais, Fred and Suzanne Fish 1988 Canals and Related Features. The 1982-1984 Excavations at Las Colinas: The Site and Its Features, by D. Gregory, W. Deaver, S. Fish, R. Gardiner, R. Layhe, F. Niais, and L. Teague, pp. 275-306. Arizona State Museum Ar- chaeological Series 162. Niais, Fred, David Gregory, and Donald Graybill 1989 Salt River Streamflow and Hohokam Irrigation Systems. In The 1 982-1 984 Excavations at Las Colinas: Environment and Subsistence, edited by C. Heath- ington and D. Gregory, pp. 59-78. Arizona State Museum Archaeological Series 162. Nicholas, Linda, and Jill Neitzel 1984 Canal Irrigation and Sociopolitical Organization in the Lower Salt River Valley: A Diachronic Analysis. In Prehistoric Agricultural Strategies in the Southwest, edited by S. Fish and P. Fish, pp. 161-178. Arizona State Uni- versity Anthropological Research Papers 33. Pewe, Troy L. 1978 Terraces of the Lower Salt River Valley in Relation to the Late Cenozoic History of the Phoenix Basin, Arizona. In Guidebook to the Geology of Central Arizona, edited by D. M. Burt and T. Pewe, pp. 1-46. State of Arizona Bureau of Geology and Mineral Technology Special Paper 23. Rankin, A. and K. Katzer In Press Agricultural Systems in the ACS Waddell Project Area. Archaeological Con- sulting System Archaeological Report. Rea, Amadeo 1981a Resource Utilization and Food Taboos of Sonoran Desert Peoples. Journal of Ethnobiology 1:69-83. 1981b The Ecology of Pima Fields. Environment Southwest 484:8-15. 1983 Once a River: Bird Life and Habitat Changes on the Middle Gila. University of Arizona Press, Tucson. Rice, Glen 1987 The Marana Community Complex: A Twelfth Century Hohokam Chief- dom. In Studies in the Hohokam Community of Marana, edited by Glen Rice, pp. 249-254. Arizona State University Anthropological Field Studies 15. 286

DESERT AS CONTEXT: THE HOHOKAM ENVIRONMENT 59

Scantling, F. H. 1940 Excavations at the Jackrabbit Ruin, Papago Indian Reservation, Arizona. M.A. thesis, Department of Anthropology, University of Arizona, Tucson. Sellers, William D., Richard H. Hill, and Margaret Sanderson-Rae 1985 Arizona Climate: The First Hundred Years. University of Arizona Press, Tuc- son. Sheridan, Thomas, and Gary P. Nabhan 1978 Living with a River: Traditional Farmers of the Rio San Miguel. The Journal of Arizona History 19:1-17. Smith, G. E. P. 1910 Groundwater Supply and Irrigation in the Rillito Valley. University of Arizona Agricultural Experiment Station Bulletin 64. Spicer R. B. 1943 An Outline of Papago Society in the Baboquivari District. Unpublished manuscript, Arizona State Museum Library, University of Arizona, Tucson. Spoerl, Patricia, and George Gumerman 1984 Prehistoric Cultural Development in Central Arizona Archaeology: Archaeology of the Upper New River Region. Southern Illinois University Center for Ar- chaeological Investigations Occasional Paper 5. Szuter, Christine 1984 Faunal Exploitation and the Reliance on Small Animals among the Ho- hokam. In Hohokam Archaeology along the Salt Gila Aqueduct, Central Arizona Project, vol. 7: Environment and Subsistence, edited by Lynn Teague and P. Crown, pp. 139-170. Arizona State Museum Archaeological Series 150. 1986 Lagomorph and Artiodactyl Exploitation among the Inhabitants of the West Branch Site. In Archaeological Investigations at the West Branch Site: Early and Middle Rincon Occupation in the Southern Tucson Basin, edited by F. Hun- tington, pp. 273-288. Institute for American Research Anthropological Pa- per 3. Turnage, W. V., and T. D. Mallory 1941 An Analysis of Rainfall in the Sonoran Desert and Adjacent Territory. Carnegie Institution of Washington Publication 519. Turner, Raymond, and David E. Brown 1982 Sonoran Desertscrub. Desert Plants 1982:181-222. Underhill, Ruth M. 1939 Social Organization of the Papago Indians. Columbia University Press, New York. U.S. Geological Survey 1954 Compilation of Records of Surface Waters of the United States through September 1950, Part 9, Colorado River Basin. U.S. Geological Survey Water Supply Paper 1926. Upham, Steadman, and Glen Rice 1980 Up the Canal without a Pattern: Modelling Hohokam Interaction and Ex- change. In Current Issues in Hohokam Prehistory: Proceedings of a Symposium, edited by D. Doyel and F. Plog, pp. 78-105. Arizona State University Anthropological Research Papers 23. Vander Wall, S. B., and J. A. MacMahan 1984 Avian Distribution Patterns along a Sonoran Desert Bajada. Journal of Arid Environments 7:59-74. 287

60 Suzanne K. Fish and Gary P. Nabhan

Vivian, R. Gwinn 1974 Conservation and Diversion: Water-Control Systems in the Anasazi South- west. In Irrigation's Impact on Society, edited by T. E. Downing and M. Gibson, pp. 95-112. University of Arizona Anthropological Papers 25. Waters, Michael 1988 The Impact of Fluvial Processes and Landscape Evolution on Archaeolog- ical Sites and Settlement Patterns along the San Xavier Reach of the Santa Cruz River, Arizona. Geoarchaeology 3:205-219. Waters, Michael, and John Field 1986 Geomorphic Analysis of Hohokam Settlement Patterns on Alluvial Fans along the Western Flank of the Tortolita Mountains, Arizona. Geoarchaeology 1:29-35. Wilcox, David, and Richard Ciolek-Torello 1988 Hohokam Settlement along the Slopes of the Picacho Mountains, Synthesis and Conclusions, Tucson Aqueduct Project. Museum of Northern Arizona Research Papers 35. Wilcox, David, and Lynette Shenk 1981 The Architecture of the Casa Grande and Its Interpretation. Arizona State Mu- seum Archaeological Series 115. Wilcox, David, and Charles Sternberg 1983 Hohokam Ballcourts and Their Interpretation. Arizona State Museum Archae- ological Series 160. Wilson, John P. 1985 Early Piman Agriculture: A New Look. In Southwestern Culture History: Collected Papers in Honor of Albert H. Schroeder, edited by Charles H. Lange, pp. 129-138. The Archaeological Society of New Mexico 10. Woodbury, Richard B. 1960 The Hohokam Canals at Pueblo Grande, Arizona. American Antiquity 26267- 270. 1961 A Reappraisal of Hohokam Irrigation. American Anthropologist 63:550-560. 288

APPENDIX I

PREHISTORIC LANDSCAPES OF THE SONORAN DESERT HOHOKAM

by Suzanne K. Fish and Paul R. Fish

1992. Population and Environment, Vol. 13, No. 4, pp. 269-284. s

289

POPULATION AND ENVIRONMENT

A JOURNAL OF INTERDISCIPLINARY STUDIES

CONTENTS

VOLUME 13, NUMBER 4 SUMMER 1992

Editorial: Lessons from the Past 235

Guest Editorial: The Prehistory of Sustainability 237 Timothy A. Kohler

Early Neolithic Exploitation Patterns in the Levant: Cultural Impact on the Environment 243 Gary O. Rollefson and Ilse Köhler-Rollefson

Prehistoric Human Impact on the Environment in the Upland North American Southwest 255 Timothy A. Kohler

Prehistoric Landscapes of the Sonoran Desert Hohokam 269 Suzanne K. Fish and Paul R. Fish

Management in the Andean Highlands: Prehistoric Landscape Raised Field Agriculture and Its Environmental Impact 285 Clark L. Erickson

PRIVATE BAG: Two Approaches to Sustainability 301

Why Excess Immigration Damages the Environment 303 Population-Environment Balance, Inc.

Departing from Resource-Intensive Lifestyles: Problems and Possibilities in Industrialized Societies 313 Lincoln H. Day 290

POPULATION AND ENVIRONMENT: A JOURNAL OF INTERDISCIPLINARY STUDIES is a journal that focuses primarily upon the linkages between demographic and environmental variables. MANUSCRIPTS should be submitted in duplicate, double-spaced throughout, with a 400- word abstract to Dr. Virginia Abernethy, AA-2206 MCN, Department of Psychiatry, Vander- bilt Medical School, Nashville, TN 37232. Style should conform to the APA Publication Manual, 3rd Edition (1983), which can be purchased from the American Psychology Associa- tion, 1200 17th Street, NW, Washington, DC 20036. Except for an introductory footnote, footnotes are discouraged. Tables and figures must be submitted as originals. Authors should seek written permission from the copyright owner for the use of tables, illustrations, or exten- sively quoted material which has previously appeared in another publication. Include title of manuscript, your name, affiliation, title and mailing address on face sheet. Further informa- tion may be obtained from the Editor. SUBSCRIPTION inquiries and subscription orders should be addressed to the publisher at Subscription Department, Human Sciences Press, Inc., 233 Spring Street, New York, N.Y. 10013-1578 or faxed to the Subscription Department at its number (212) 807-1047. Subscip- tion rates: Volume 13, 1991-1992 14 issues) $145.00 (outside the U.S., $170.00). Price for individual subscribers certifying that the journal is for their personal use, $42.00 (outside the U.S., $50.00). ADVERTISING inquiries should be addressed to the Advertising Sales Representative, Dan- iel S. Lipner, Weston Media Associates, P.O. Box 1110, Greens Farms, Connecticut 06436— telephone (203) 261-2500 and fax (203) 261-0101. INDEXED in Sociological Abstracts, Social Work Research and Abstracts, Safety Science Abstracts, Development and Welfare (), Current Index to Journals in Education, Popula- tion Index Current Contents/Social and Behavioral Sciences, Social Sciences Citation Index, Excerpta Medica, Environment Abstracts, Environment Index, CICRED (France), Psychological Abstracts, Guide to Social Science and Religion in Periodical Literature, and International Bibliography of Periodical Literature, Biosis-Biological Abstracts, Sociological Abstracts, Au- tomatic Subject Citation Alert, Institute for Scientific Information and Environmental Periodi- cals Bibliography. PHOTOCOPYING: The appearance of a code line at the bottom of the first page of an article in this journal indicates the copyright owner's consent that copies of the article may be made for personal or internal use. However, this consent is given on the condition that the copier pay the flat fee of $6.50 per copy per article (no additional per-page fees) directly to the Copyright Clearance Center, Inc., 27 Congress Street, Salem, Massachusetts 01970, for all copying not explicitly permitted by Sections 107 or 108 of the U.S. Copyright Law. The CCC is a nonprofit clearinghouse for the payment of photocopying fees by libraries and other users registered with the CCC. Therefore, this consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale, nor to the reprinting of figures, tables, and text excerpts. 0199-0039/92 $6.50. COPYRIGHT 1992 by Human Sciences Press, Inc. Published quarterly in the fall, winter, spring and summer. Second-class postage paid at New York, N.Y., and at additional mailing offices. Postmaster: Send address changes to Population and Environment, Human Sciences Press, Inc., 233 Spring Street, New York, N.Y. 10013-1578.

LC77-77352 ISSN 0199-0039 PENVOK 13(4)231-326(1992) 291

2214 E. 4th St. Tucson, AZ 85719 June 26, 1993

Human Sciences Press, Inc. 233 Spring St. New York, New York 10013-1578

To Whom It May Concern:

I am writing to request permission to reprint the following article:

Fish, Suzanne K. and Paul R. Fish (1992). Prehistoric Landscapes of the Sonoran Desert Hohokam. Population and Environment: A Journal of Interdisclplimary Studies, Vol 13, No 4, pp. 269 - 284.

This material is to appear, in the same form as as originally published, in my Ph.D. dissertation in the Arid Lands Resource Sciences Interdisciplinary Program at the University of Arizona. The dissertation, .entitled Agriculture and Society in Arid Lands: A Hohokam Case Study," has been prepared under an option for the inclusion of published papers. This format requires require letters of permission for copyrighted materials. The entire dissertation, including the article cited above and copyright permission, will be reprinted by University Microfilms.

If you have any question, please contact me at your convenience.

Sincerely, 292

Prehistoric Landscapes of the Sonoran Desert Hohokam

Suzanne K. Fish Paul R. Fish University of Arizona

ABSTRACT: The Hohokam of southern Arizona are noted for greater duration of settlement than other major agricultural traditions in the archaeological record of the southwestern United States, including the Anasazi and Mogollon. The 40,000 square mile area inhabited by the Hohokam is marked by low elevation desert basins, but encompasses a range of topographic and climatic variability that shaped opportunities for prehistoric farming technol- ogies. Irrigation from rivers was frequently associated with the longterm persistence of indi- vidual sites, while floodwater farming along ephemeral drainages was more often correlated with continuous occupation of hydrologically favored zones. Renewal of fields by waterborne nutrients and efficient practices in the use of natural resources countered the limited mobility options afforded by the Hohokam environment. In spite of a restrictive agricultural setting and an essentially static suite of productive technologies over many centuries, relationships among population, settlement, and landuse were redefined in evolving social and economic configurations. An example from the Tucson Basin illustrates differentiated patterns of settlement and agriculture arising in conjunction with increased levels of population and territorial integration in the late prehistoric period. Community organization among interrelated settlements incorporated a diversity of topo- graphic zones and agricultural technologies in this high-risk context for prehistoric cultivators.

INTRODUCTION

Accounts of prehistoric peoples of the southwestern United States fre- quently end with a mysterious and sudden departure. The Hohokam who inhabited the desert basins of southern Arizona are no exception.

Please address correspondence to Suzanne Fish, Arizona State Museum, University of Arizona, Tucson, AZ 85721.

Population and Environment: A Journal of Interdisciplinary Studies Volume 13, Number 4, Summer 1992

© 1992 Human Sciences Press, Inc. 269 293

270

POPULATION AND ENVIRONMENT

Hohokam territory was largely circumvented by the earliest Spanish ex- plorers, but when reliable observations were recorded near the end of the seventeenth century, speakers of several different languages occupied the area. None of these successor groups equaled the Hohokam in geographic extent, regional population densities, size of largest settlements, or com- plexity of social and political organization. The Hohokam cannot be iden- tified in the archaeological record with any certainty after A.D. 1450; the transition to post-Columbian groups and the potential role of contact phe- nomena such as Old World disease is virtually unknown. However, by that time, the Hohokam had maintained a cultural identity as desert farmers for more than 1,000 years—not a bad record on a New World scale of Neolithic persistence. Perhaps the manner in which they achieved such continuity is more interesting than the poorly understood circum- stances of their disappearance. The Hohokam inhabited approximately 120,000 sq km (45,000 sq mi) of the Sonoran Desert in the south-central portion of Arizona (Haury 1976; P. Fish 1989). Within this vast territory (Figure 1), significant axes of envi- ronmental variability are the related factors of climate and hydrology. Al- though annual precipitation throughout is less than 380 mm (15 in), the highest average amounts exceed the lowest by a factor of two. In the cen- ter of the Hohokam world, the broad, flat Phoenix Basin experiences the lower range of precipitation at 180 mm (7.5 in) and the hottest tempera- tures, with over 90 days above 56°C (100°F) Sellers et al 1985). An arc of somewhat higher elevation, lower temperatures, and increased rainfall sur- rounds the Phoenix Basin to the north, east, and south. The harsher climate of the Phoenix core is ameliorated by two peren- nial rivers arising in mountain watersheds outside the low desert. The Salt and Gila Rivers supplied the massive irrigation systems for which the Hohokam are noted (e.g. Masse, 1981), and other technologies were less frequently invoked. Spring and summer planting seasons were possible. In the surrounding arc of less extreme climate, the rivers are intermittent. Hohokam populations here depended primarily on summer precipitation falling within the desert for the major agricultural season. Small-scale irri- gation from river segments with more permanent flow, floodwater farming from ephemeral drainages, and diversions of surface runoff supported crops.

THE AGRICULTURAL BASIS FOR SEULEMENT CONTINUITY

Environmental diversity across Hohokam territory was reflected in dif- ferential subsistence opportunities and technological responses, yet cul- 294

2TI

SUZANNE K. FISH AND PAUL R. FISH

FIGURE 1. Map showing extent of Hohokam with locations for the Marana Study Area and important place names used in the text.

Nt:CV.

tural development throughout progressed broadly in tandem, including so- cial investments in nonutilitarian expressions such as public architecture. Both in regions of largescale irrigation and in those lacking it, agricultural products are uniformly prominent among archaeologically recovered food remains (Fish & Nabhan, 1991). At widespread locales, agriculture and a sedentary way of life persisted over hundreds of years of continuous occu- pation. A common repertoire of basic cultural behaviors and technologies must have permitted highly sustainable local adaptations under a variety of arid conditions. The nature of agricultural water use by the Hohokam was central to the achievement of population and settlement continuity. In other South- 295

272

POPULATION AND ENVIRONMENT

western regions with sufficient rainfall for dry farming or where clear up- land streams furnished supplemental water, progressive soil exhaustion may have induced sequential clearing of short-term fields or prolonged fallowing, and the eventual abandonment of arable locales. Although wa- ter availability is circumscribed in the low desert, sources supplying Hohokam cultivation renewed soil fertility with each wetting. Continuity of individual sites over long intervals is most closely associated with irriga- tion, while continuous settlement of favored zones in sites of shorter dura- tion is a hallmark of alternate technologies. Suspended sediment in the undammed flow of the Salt and Gila rivers was carried by canals to the perimeters of cultivation, up to 10 km (6 mi) inland from the river channel and for linear distances up to 30 km (20 mi). Some of this rich soil and detritus was delivered directly to fields through irrigation and some appears to have been spread beyond the banks of canals in the course of periodic channel cleanings. Ribbons of distinctive soil types of waterborne origin still mark the paths of both Hohokam canals tapping perennial rivers and ones filled by seasonal runoff (Dart, 1986). Floodwater farmers of alluvial fans on lower basin slopes, broader ar- royo bottoms, and upper basin pediments simultaneously fertilized and watered their crops by diverting storm flows onto fields at the sides of shallow drainages. Organic debris concentrated in these waters by rapid runoff following seasonal rains offset a major deficiency of desert soils (Bryan, 1929; Nabhan, 1979; 1983). Even the texture of coarse soils could be improved for water retention by the addition of fine particles in water- laid deposits. As active channels shifted across the surfaces of alluvial fans and new fans became hydrologically active, Hohokam settlement also shifted within the boundaries of topographically defined zones of deposi- tional enrichment. Complexes of agricultural features intercepting surface runoff on basin slopes benefited from similar annual renewal.

CONSERVATION AND EXTRACTIVE PRACTICE

Highly localized sources of domestic and agricultural water in Hohokam basins must have created a strong impetus for extractive prac- tices that conserved the longterm productive potential of the environment. Situations suitable for water diversion or canal headings, and investment in the construction of canals (Doyel, 1976; 1980) and runoff features (Fish, Fish, & Madsen, 1985) added to locational constraints. Movement to new locations was not a simple alternative if the local environment became depleted or degraded, and mobility options decreased as populations 296

273

SUZANNE K. FISH AND PAUL R. FISH

grew. For example, within the 2000 sq km (750 sq mi) of the Phoenix Basin, late prehistoric populations are estimated between 40,000 and 100,000 (Haury, 1976; Schroeder, 1960; Doyel, 1991). Several cultural patterns appear to have contributed to the Hohokam ability to maintain stable occupations in their fragile arid environment. Un- like early farmers in the Old World, the Hohokam and other prehistoric North Americans were direct gatherers and consumers of desert vegetation, without domestic animals as intermediaries. As a result, ground covers were not extensively destroyed, nor was damage to perennial woody vege- tation as serious. Compared to prehistoric groups of the northern South- west (cf. Kohler, this volume), the Hohokam had a greatly reduced need for heating fuel, with long months of hot weather and average annual tem- peratures between 10 and 15°C (50 to 64°F) (Sellars et al., 1985). Destruction of leguminous trees, the most common desert species, was likely minimized in deference to the dietary importance of their abun- dant and nutritious beans. Among historic Piman Indians, geographic suc- cessors and probable descendants of the Hohokam, desirable trees such as mesquite were often left standing in cleared fields and grew densely in hedgerows benefiting from agricultural water (Rea, 1981; Castetter & Bell, 1942). Identified seeds and charcoal from long prehistoric occupations at- test to the consistent availability of such trees. Wide ranging sources of additional fuels included riparian trees, drift wood, and woody desert shrubs. Pit-ovens or hornos were fuel-efficient cooking facilities of Hohokam villages. Food was placed in these pits along with coals and heated stones, then coyered and left to cook slowly over long periods. Shared homo us- age by several households was a typical pattern (e.g. Sires, 1987, p. 180) that would have further minimized fuel consumption. Ceramics were fired at relatively low temperatures. Calcium carbonate additives to plaster re- ceived no prior thermal treatment (compare with Rollefson & KOhler-Rol- lefson, this volume). Pit house, wattle and daub, and adobe architectural styles utilized a minimum of structural wood. Materials from alternative species of desert plants such as saguaro cactus and ocotillo were frequently substituted.

PRODUCTIVITY OF MODIFIED LANDSCAPES

The Hohokam created distinctively anthropogenic landscapes about their settlements with culturally altered distributions of plants and animals (Fish et al., 1984). Supplemental water in fields substantially increased bio- 297

274

POPULATION AND ENVIRONMENT mass production over surrounding vegetation. Pollen samples from prehis- toric agricultural contexts reveal a rich weedy flora responding to these enhanced conditions alongside crops (Fish, 1985). The Hohokam likely followed ethnographic Southwestern practices of permitting desirable weeds to remain in fields and sometimes even scattering seeds to insure a sufficient supply (Whiting, 1939; Crosswhite, 1981; Bye, 1979). Among weedy plants of Hohokam fields were species such as chenopods, am- aranths, and spiderling that furnish edible seeds and greens. In seasons when water was inadequate to mature cultigens, these secondary resources of fields may have constituted a lesser but welcome harvest. Modified environments created by the Hohokam also featured trans- planted desert species receiving more directed attention. Tepary beans are one probable Sonoran Desert addition to corn, bean, and squash varieties acquired from areas of origin in Mexico (Ford, 1981; Felger & Nabhan, 1976). Recently, the agave or century plant has been recognized as a com- monplace product of Hohokam cultivation. Remains of this plant do not 'exhibit morphological changes distinguishing true domesticates from their wild progenitors, but fields where agaves were tended have been con- clusively identified (Fish, Fish, Miksicek, & Madsen, 1985; Fish et al., 1990). Agaves furnish both edible and fiber products. Less conclusive evi- dence suggests that additional desert plants in this tended category were cholla and prickly pear cacti (Fish, 1984; Bohrer, 1987). Prickly pear is commonly cultivated in Mexico. Such plantings in poorer fields, field bor- ders, and fallow ground would have offset pressure on natural stands near longterm residence. In the case of agave, this drought-adapted crop en- abled agricultural production on large tracts of otherwise marginal land. A substantial proportion of animals hunted by the Hohokam also con- forms to the culturally modified environs of their settlements (Szuter & Bayham, 1989). Bones of large animals such as deer are infrequent com- pared to quantities for smaller species. Rabbits and rodents, characteristic of agricultural habitats, were mainstays of Hohokam cuisine. Animals at- tracted to the vegetation and water in fields could be conveniently trapped or hunted during agricultural tasks. At sites occupied for long intervals, trends in consumption reflect growing percentages of species preferring open vegetation. In multiple instances, an earlier predominance of jack- rabbit was superceded by greater reliance on cottontail.

THE ROLE OF COMMUNITY ORGANIZATION

By A.D. 1000, territorial entities integrating multiple sites can be rec- ognized. In each of these units of interrelated settlement, called commu- 298

275

SUZANNE K. FISH AND PAUL R. FISH nities, central sites are distinguished by the presence of public architecture. In the earlier Hohokam sequence, earthen-banked ballcourts were the common form of public architecture (Doyel, 1980; Wilcox & Sternberg, 1983), but in the Classic Period after about A.D. 1100, platform mounds predominated (Gregory & Nials, 1985; Gregory, 1987). Communities are thought to have provided a framework for allocating water among sites on irrigation networks of the Phoenix vicinity and for mobilizing and coor- dinating labor in the construction and maintenance of canals. In regions without perennial rivers such as the Tucson Basin, communities linked sites in diverse basin segments, thus encompassing a variety of productive potentials and risks. Functional emphases of community organization may have differed between the irrigated Phoenix core and other regions; nev- ertheless, this flexible integrative structure appears to have provided the basis for cooperation, specialization, and risk management throughout the challenging environment of the Hohokam tradition.

A HOHOKAM COMMUNITY OF THE TUCSON BASIN

A comprehensively investigated study area in the northern Tucson Ba- sin near Marana, Arizona, exemplifies the typical duration of Hohokam settlement, varied components of the subsistence base, community devel- opment and structure, and responses to increasing population (Fish et al., 1989). A community reaching maximum size in the early Classic Period (ca. A.D. 1100 to 1300) integrated zones of reciprocal annual threats to agriculture in Southwestern basins: floodplains of the primary drainages flooded destructively with too much rain, and upland slopes yielded poorly with too little (Lightfoot & Plog, 1984; Abruzzi, 1988). A degree of productive specialization and exchange within this Marana community, coinciding with the highest population levels of the prehistoric sequence, would have served to diffuse localized effects of low and unpredictable precipitation. Annual rainfall is between 225 and 300 mm in the Tucson Basin. Mountains rim the basin and divide it from adjacent drainage regimes. Elevational diversity is repeated on either side of the floodplain of the Santa Cruz River, but the eastern valley slope occupies much of the northern basin interior as it rises toward the Tortolita Mountains (Figure 2). A fore- shortened slope occurs to the west below the smaller Tucson Mountains. Full-coverage archaeological survey of 390 square kilometers in this basin segment has revealed settlement patterns from the Late Archaic adoption of agriculture through all subsequent ceramic periods. One of two sources for permanent water is found in springs and can- yon streams along the flanks of the Tortolita Mountains. Although the Santa

299

276

POPULATION AND ENVIRONMENT

FIGURE 2. Map of the Northern Tucson Basin depicting major topographic features and configuration of the early Classic Period community.

/ •

c: • • '6 •'S •

• ;0. eti:i n 0 L. . •10.e, p o...... "tr- • .- Js'.. i. T OR TOLITA ..e..r. , v: •• q -- — • • .114,11. 0 / • $ ' . ...e••.! . .,,Z., f ) 0 • : • a T.'...t. MOUNTAINS

S‘' . '. r.— :•.. eirl:-,in. • • A :, .:, •. .• • to,: • • , (...)4...... ;,:d< , 1".:,..- -... 5.,..... :.' • Ai % /,'' • e.',) r / ...de AI 1:4•4-ty, .: a , -.4.,., .; " "11 4 ,% • : • "L. r 1 „NI' (i • l'e • • ,/ v ,,,, f. , ._., ...4, 1...1.-t,...• c 4. j 1 4 . ....r.,;* 1 e .• 4 ' r'-f%li./ : -4 v 5 i,

, • P.• PI•

• • 1 •./.

••••:/.. • 00/• • e- o 1 • ft t • . •01.

i 1 it•eser•S Oa Archoeological Site 1 411/• Early Classic Period Site 0 Plot form Mound — Canal IA 'at

ll"4.--' 4_,1...• .•..k. N.. -.. 300

277

SUZANNE K. FISH AND PAUL R. FISH

Cruz is intermittent in the study area, a second location of persistent sur- face flow occurs near the end of the Tucson Mountains where intrusive bedrock maintains high water tables. Throughout the Hohokam occupa- tion, both areas were continuously preferred for settlement. In the Pre- classic Period prior to A.D. 1100, independent communities existed in each of these two hydrologically optimal locales. The riverine and moun- tain flank communities covered 56 and 70 sq km (22 and 27 sq mi), re- spectively. Developments in the Classic Period illustrate the flexibility of commu- nity organization in integrating larger and denser populations. At this time, the two earlier communities coalesced into a single larger one incorporat- ing 146 sq km (56 sq mi). A platform mound was constructed in a new central site at the juncture of Preclassic axes of settlement. Additional sites appeared in the previously intervening area, and substantial population growth is registered by an increase from approximately 2,000,000 sq m of habitation area in Preclassic sites to 6,000,000 sq m in Classic ones. The magnitude of increase suggests that some Classic community members were drawn from elsewhere in the Tucson Basin. Within the enlarged later boundaries, six zones can be defined on the basis of residence patterns, productive activities, and environmental variables (Figures 2 and 3). Envi- ronmental opportunities and limitations were different in each of these zones, as were methods and consequences of Hohokam landuse.

Zone 5

The floodplain and terraces of the Santa Cruz River constitute Zone 5. Dense populations lived here consistently, concentrated along the FIGURE 3. Idealized cross-section of the Northern Tucson Basin.

WEST C AST

ZONE 5 ZONE I ZONE 2 ZONE 3 ZONE 4 MITIRS ZONE 6 fin Ism ..01••n••1 manna IN... V.... • .0.0.0 SO .81110M COMMIS •••••••••••• an... Are... n un- •••• •.• woo. v..* • ...If C....0m t•: Mew •,••••••• •••••••••• wow um. Y.. 1n••• •••••••• •••••••• l.m.1

- 3600

- X0J0

MgoaR1•141. isola Cou My. WIN./ Slope 301

278

POPULATION AND ENVIRONMENT

stretches of high water table. Irrigation would have made this zone the foremost producer of crops like corn with high moisture requirements. Due to alluvial deposition on the Santa Cruz floodplain, canals have been iden- tified only through excavation (Bernard-Shaw 1988), and the extent of irri- gation networks is unknown. In an instance where canals leave the active floodplain, paths can be traced for more than 10 km (6 mi) north to the Classic mound site. Such canals would have extended fields with crops and assemblages of weedy summer annuals into shrubby creosotebush and bursage associations. Channels of intermittent rivers tend to shift with large floods. Early Anglo canals in Tucson initiated destructive headcutting (Be- tancourt and Turner 1988), but prehistoric intakes likely washed away in floods before creating similar erosional damage. Where canals could be filled from high water tables along the river, relatively minor investment is apparent in other methods of farming.

Zone 1

In Zone 1 at the lower edge of the basin slope, the gradients of tribu- tary steams decease and the deposition of suspended soil forms alluvial fans. Shallow channels were easily tapped by the short ditches of floodwa- ter farmers. Ethnographic data suggest that better floodwater fields yielded as bountifully as irrigated ones (Doelle 1980: 67-75), although at greater risk from spotty rainfall. Village locations over time correspond with depo- sitionally active fans, but no permanent water occurs in Zone 1. Preclassic settlement was within daily travel distances for domestic water from the river. At the beginning of the Classic Period, a new canal permitted sites farther from the Santa Cruz.

Zone 2

Permanent water is even more distant from Zone 2, and large drain- ages are too entrenched for diversion. This zone was utilized primarily for hunting and gathering prior to the Classic Period. Higher Zone 1 popula- tions at that time transformed large segments of these dry slopes into fields for agave, a dual source of food and fiber. Ninety field locations, up to 50 ha in extent, covered a total of 485 ha (1200 acres) in Zone 2. Piles of rock served as water-conserving mulches and low contour terraces and check dams concentrated surface runoff for agaves planted below their natural elevational limit. A drought resistant succulent, agave could mature with less moisture than annual crops such as corn and survive seasons of defi- cient rainfall. Crop identity is confirmed by charred remains in roasting pits 3 0 2

279

SUZANNE K. FISH AND PAUL R. FISH

for fieldside processing. In large fields, pits up to 30 m in diameter cooked the communal harvests of many farmers (Fish, Fish, Miksicek, & Madsen, 1985; Fish et al., 1990). Although a few smaller rockpile fields for agave are of Preclassic date, Classic populations expanded this cultivation onto broad tracts of previ- ously unused and marginal land. Zone 2 fields are the most striking evi- dence for response to the heightened resource requirements of Classic in- habitants. Fields of largest size were constructed in the vicinity of dense settlements lacking better land for irrigated or floodwater farming. Thus, this form of cropping appears to represent an arid land version of spe- cialized and intensified production.

Zone 3 Domestic and agricultural water is also absent in Zone 3. Residential sites are not found. The common site type consists of huge scatters of bro- ken pottery and few other artifacts. The Hohokam seem to have camped in locales of abundant saguarO cactus year after year, harvesting the fruits in the early summer. Vessels for drinking water and fruit boiling were broken in the process. The only constructions preserved at these Zone 3 sites are rings of rocks, used ethnographically to support the broad, conical-bot- tomed baskets in which saguaro fruits were collected.

Zone 4 Shallow bedrock insures high water tables in Zone 4 drainages. The three largest streams originating in the mountains had floodplains of suffi- cient width for floodwater farming on bottomland, and were correlated with clustered sites of all periods on surrounding ridges. Smaller drainages and surface runoff were also utilized, with low terraces and checkdams at scattered locations. Sets of rock piles and terraces for agave were culti- vated in Zone 4, but did not compare in size or number with fields in Zone 2 below.

Zone 6 The compressed basin slope west of the Santa Cruz forms Zone 6. Small alluvial fans supported on-ly minor floodwater farming. In the adja- cent Tucson Mountains, rock-walled terraces for dwellings and small gar- dens were constructed on hillsides. Volcanic soils in this low range are high in clay and moisture-retentive. Above the elevation affected by cold 303

280

POPULATION AND ENVIRONMENT

air drainage and winter freezes on the basin floor, small spring crops could precede major summer harvests. Agave recovered from excavated struc- tures suggests a planting pattern still followed in Mexico, with these plants lining terrace Walls and annual crops on the remainder of the planting surfaces (Fish et al., 1984).

CONCLUSION

The Hohokam are noted for more enduring settlement than the two other major cultural traditions of the prehistoric southwestern United States, the Anasazi and Mogollon. Renewal of fields through natural, wa- terborne additives permitted a sustainable agriculture based on a range of technologies from the most impressive canal networks of aboriginal North America to simple diversions of flooding streams. Domesticated crops were supplemented by a series of tended and weedy indigenous species whose productivity was enhanced in the modified landscapes about sites. The effects of intensive landuse were concentrated in hydrologically active locales, where captured water might otherwise have escaped from desert ecosystems in rapid downstream runoff. Because water and therefore set- tlement was localized, surrounding expanses of uninhabited land would have offered generous reserves of fuel, craft materials, and edible wild resources. Hohokam continuity cannot be fully explained, however, by a ver- satile suite of productive and extractive techniques in physical settings of desert basins. The integration of residents of multiple sites into community organizations played a critical role in their ability to meet both short and longterm challenges to subsistence. Population sizes in individual commu- nities must have represented local balances between adequate and sustain- able production under arid conditions and the numbers necessary to main- tain canals or to spread agricultural risks over a sufficient number of environmental zones. Internal and external interdependencies that per- mitted specialization and countered temporary shortfalls were likely medi- ated through community structures. When growth and change occurred, as in the Tucson Basin case, new fits between population and environment appear more an outcome of innovation in organization than technology. The basic elements of Hohokam agricultural technology can be docu- mented well before A.D. 1000, including large riverine irrigation net- works, floodwater diversions from ephemeral drainages, and complexes of rockpiles, terraces, and other devices capturing overland runoff. Despite continued dependence on this repertoire in the ensuing centuries, con- s

304

281

SUZANNE K. FISH AND PAUL R. FISH

straints between populations and desert environment did not preclude the means or motivation for social and economic innovation. Changing config- urations of population, settlement, and landuse continued to appear through new combinations and deployments of existing technological ele- ments by dynamic and evolving social entities. Along the perennial rivers of the Phoenix area, the largest, most densely inhabited sites and the most extensive canal irrigation systems date to the later part of the prehistoric timespan. In the Tucson Basin, the Marana Community of the Classic Pe- riod illustrates the potential for simultaneously concentrating population and intensifying production within an enlarged and more environmentally diversified territorial base. Neither botanical remains from Marana sites nor reconstructions from tree-ring records in Arizona uplands suggest that the developmental history of this community was correlated with climatic amelioration. Circumstances surrounding the disappearance of the Hohokam tradi- tion are marked by a paucity of archaeological evidence and a lack of scholarly consensus. The nature of this process and its precise timing within the interval following the early fifteenth century cannot be specified with available information. Deleterious effects of landuse practices do not figure prominently in current speculation, and no single environmentally- based scenario could apply uniformly to locations across southern Arizona that encompass considerable variability in topography, seasonality of pre- cipitation, and agricultural water sources. A role for introduced disease and other disruptions occasioned by European contact is a matter of de- bate, but recent evaluations of chronological data support this possibility (Dean, 1991; Eighmy & McGuire, 1988). However, the impressive record of Hohokam persistence surely holds more lessons for a general under- standing of relationships between population and arid environments than do the final episodes of collapse and disappearance.

REFERENCES

Abruzzi, William S. (1989). Ecology, resource redistribution and Mormon settlement in north- eastern Arizona. American Anthropologist, 91, 642-655. Bernard-Shaw, Mary (1988). Hohokam canal systems and Late Archaic wells: The evidence from the Los Morteros site. In William H. Doelle and Paul R. Fish (Eds.). Recent research on Tucson Basin prehistory, pp. 153-170. Tucson: Institute for American Research An- thropological Paper No. 10. Betancourt, julio L. & Turner, Raymond M. (1988). Historic arroyo-cutting and subsequent channel changes at the congress Street Crossing, Santa Cruz River, Tucson, Arizona. In Emily E. Whitehead, Charles F. Hutchinson, Barbara Timmerman, and Robert G. Varady (Eds.). Arid lands: Today and tomorrow, pp. 1353-1372. Boulder: Westview Press. 305

282

POPULATION AND ENVIRONMENT

Bohrer, Vorsila (1987). The plant remains from La Ciudad, a Hohokam site in Phoenix. In JoAnn Kisselburg, Glen Rice, and Brenda Shears (Eds.). Specialized studies in the envi- ronment and culture of La Ciudad, pp. 67-238. Tempe: Arizona State University Anthro- pological Field Studies 20. Bryan, Kirk (1929). Floodwater farming. The Geographical Review, 19, 444-456. Bye, Robert (1979). Incipient domestication of mustards in northwestern Mexico. The Kiva, 44, 237-256. Castetter, Edward F. & Bell, Willis H. (1942). Pima and Papa go Indian agriculture. Albuquer- que: University of New Mexico Inter-American Series I. çrosswhite, Frank (1981). Desert plants, habitat, and agriculture in relation to the major pattern of cultural differentiation in the O'odham people of southern Arizona. Desert Plants, 3, 47-76. Dart, Alan (1986). Sediment accumulation along Hohokam canals. The Kiva, 51, 63-84. Dean, Jeffrey S. (1991). Thoughts on Hohokam chronology. In George J. Gumerman (Ed.). Exploring the Hohokam: Prehistoric desert peoples of the American Southwest, pp. 61-150. Albuquerque: The University of New Mexico Press. DoeIle, William H. (1980). Past adaptive patterns in Western Papagueria: An archaeological study of nonriverine resource use. Tucson: Ph.D. dissertation, Department of Anthropol- ogy, University of Arizona. Doyel, David (1976). Classic Period Hohokam in the Gila Basin. The Kiva, 42, 27-38. Doyel, David (1980). Hohokam social organization and the Sedentary to Classic transition. In David Doyel and Fred Plog (Eds.). Current issues in Hohokam Prehistory, pp. 23-40. Tempe: Arizona State University Anthropological Papers 23. Doyel, David (1991). Hohokam cultural evolution in the Phoenix Basin. In George J. Gumer- man (Ed.). Exploring the Hohokam: Prehistoric desert peoples of the American South- west, pp. 231-278. Albuquerque: The University of New Mexico Press. Eighmy, Jeffrey L. & McGuire, Randall H. (1988). Archaeomagnetic dates and the Hohokam phase sequence. Fort Collins: Colorado State University Archaeomagnetic Laboratory Technical Series 3. Felger, Richard S. & Nabhan, Gary P. (1976). Deceptive barrenness. Ceres, 9, 34-39. Fish, Paul R. (1989). The Hohokam: 1,000 years of prehistory in the Sonoran Desert. In Linda Cordell and George Gumerman (Eds.). Dynamics of Southwest Prehistory, pp. 16-64. Washington, D.C.: The Smithsonian Institution Press. Fish, Suzanne K. (1984). Agriculture and subsistence implication of the Salt-Gila Aqueduct pollen analysis. In Lynn S. Teague and Patricia Crown (Eds.). Hohokam Archaeology along the Salt-Gila Aqueduct, Central Arizona Project, Volume VII: Environment and Subsistence, pp. 111-138. Tucson: Arizona State Museum Archaeological Series 150. Fish, Suzanne K. (1985). Prehistoric disturbance floras of the Lower Sonoran Desert and their implications. In Bonnie Jacobs, Patricia Fall, and Owen Davis (Eds.). Late Quaternary vegetation and climates of the American Southwest, pp. 78-88. Houston: American As- sociation of Stratigraphic Palynologists Contribution 16. Fish, Suzanne K., Fish, Paul R., & Downum, Christian (1984). Hohokam terraces and agri- cultural production in the Tucson Basin. In Suzanne K. Fish and Paul R. Fish (Eds.). Prehistoric agricultural strategies in the Southwest, pp. 55-71. Tempe: Arizona State Uni- versity Anthropological Research Paper 33. Fish, Suzanne K., Fish, Paul R., & Madsen, John (1985). A preliminary analysis of Hohokam settlement and agriculture in the northern Tucson Basin. In A. E. Dittert, Jr. and D. E. Dove (Eds.). Proceedings of the 1983 Hohokam Symposium, pp. 75-100. Phoenix: Ari- zona Archaeological Society Occasional Paper 2. Fish, Suzanne K., Fish, Paul R., & Madsen, John (1989). Classic Period Hohokam community integration in the Tucson Basin. In Steadman Upham and Kent Lightfoot (Eds.). Socio- political structure of prehistoric Southwestern societies, pp. 237-268. Boulder: Westview Press. 306

283

SUZANNE K. FISH AND PAUL R. FISH

Fish, .Suzanne K., Fish, Paul R., & Madsen, John (1990). Analyzing regional agriculture: a Hohokam example. In Suzanne K. Fish and Stephen A. Kowalewski (Eds.). The archaeol- ogy of regions: A Case for full-coverage survey, pp. 189-218. Washington, D.C.: Smithsonian Institution Press. Fish, Suzanne K., Fish, Paul R., Miksicek, Charles, & Madsen, John (1985). Prehistoric agave cultivation in southern Arizona. Desert Plants, 7, 107-112. Fish, Suzanne K. & Nabhan, Gary P. (1991). Desert as context: the Hohokam environment. In George J. Gumerman (Ed.). Exploring the Hohokam: Prehistoric desert people of the American Southwest, pp. 29-60. Albuquerque: The University of New Mexico Press. Ford, Richard (1981). Gardening and farming before A.D. 1000: patterns of prehistoric culti- vation north of Mexico. Journal of Ethnobiology, 1, 6-27. Gregory, David (1987). The morphology of platform mounds and the structure of Classic Period Hohokam sites. In David E. Doyel (Ed.). The Hohokam village: Site structure and organization, pp. 183-210. Glenwood Springs, Colorado: Southwest and Rocky Moun- tain Division of the American Association for the Advancement of Science. Gregory, David & Niais, Fred (1985). Observations concerning the distribution of Classic Period Hohokam platform mounds. In A. E. Dittert, Jr. and D. E. Dove (Eds.). Proceed- ings of the 1983 Hohokam Symposium, pp. 373-388. Phoenix: Arizona Archaeological Society Occasional Paper 2. Haury, Emil (1976). The Hohokam: Desert farmers and craftsmen. Tucson: The University of Arizona Press. Lightfoot, Kent G. S, Plog, Fred (1984). Intensification along the north side of the Mogollon Rim. In Suzanne K. Fish and Paul R. Fish (Eds.). Prehistoric Agricultural Strategies in the Southwest, pp. 179-195. Tempe: Arizona State University Anthropological Research Pa- per 33. Masse, R. Bruce (1981). Prehistoric irrigation systems in the Salt River Valley, Arizona. Sci- ence, 214, 408-415. Nabhan, Gary P. (1979). The ecology of floodwater farming in southwestern North America. Agro-ecosystems, 5, 245-255. Nabhan, Gary P. (1986). Ak-chin "arroyo mouth" and the environmental setting of the Pa- pago Indian fields in the Sonoran Desert. Applied Geography, 6, 61-75. Rea, Amadeo (1981). The ecology of Pima fields. Environment Southwest, 484, 8-15. Schroeder, Albert Henry (1960). The Hohokam, Sinagua, and Hakataya. Madison: Society of American Archaeology, Archives of Archaeology. Sellers, William D., Hill, Richard H., & Sanderson-Rae, Margaret, (1985). Arizona climate: The first hundred years. Tucson: The University of Arizona Press. Sires, Earl W. (1987). Hohokam architectural variability and site structure during the Seden- tary-Classic transition. In David E. Doyel (Ed.). The Hohokam village: Site structure and organization, pp. 171-182. Glenwood Springs, Colorado: Southwest and Rocky Moun- tain Division of the American Association for the Advancement of Science. Szuter, Christine R. & Bayham, Frank E., (1989). Sedentism and animal procurement among desert honiculturalists in the North American Southwest. In Susan Kent (Ed.). farmers as Hunters: The Implications of Sedentism. Cambridge: Cambridge University Press. Whiting, Alfred E. (1939). The ethnobotany of the Hopi. Flagstaff: Museum of Northern Ari- zona Bulletin 15. Wilcox, David & Sternberg, Charles (1983). Hohokam ballcourts and their interpretation. Tucson: Arizona State Museum Archaeological Series 160. 307

APPENDIX J

COMPARATIVE ASPECTS OF PARADIGMS FOR THE NEOLITHIC

TRANSITION IN THE LEVANT AND THE AMERICAN SOUTHWEST

by Suzanne K. Fish and Paul R. Fish

1991. In Perspectives on • the Past: Theoretical Biases in Mediterranean Hunter-Gatherer Research,. edited by Geoffrey A. Clark, pp. 396-410. The University of Pennsylvania Press, Philadelphia. 308

Perspectives on the Past Theoretical Biases in Mediterranean Hunter-Gatherer Research

Edited by Geoffrey A. Clark

UNIVERSITY OF PENNSYLVANIA PRESS Philadelphia 309

Copyright © 1991 by the University of Pennsylvania Press All rights reserved Printed in the United States of America

U.S. Library of Congress Cataloging-in-Publication Data Perspectives on the past : theoretical biases in Mediterranean hunter-gatherer research / edited by Geoffrey A. Clark. p. cm. Includes bibliographical references and index. ISBN 0-8122-8190-X 1. Paleolithic period—Mediterranean Region. 2. Mesolithic period— Mediterranean Region. 3. Hunting and gathering societies—Mediterranean Region. 4. Archaeology—Research—Mediterranean Region. 5. Mediterranean Region—Antiquities. I. Clark, Geoffrey A. GN772.25.A113 47 1991 909'.09822—dc20 91-16444 CI? 310

2214 E. 4th St. Tucson, AZ 85719 June 26,1993

Copyright Permissions Department The University of Pennsylvania Press University of Pennsylvania Philadelphia, PA 19104

To Whom It May Concern:

I am writing to request permission to reprint the following chapter:

Fish, Suzanne K. and Paul R. Fish (1991). Comparative Aspects of Paradigms for the Neolithic Transition in the Levant and the American Southwest. In Perspectives on the Past: Theoretical Biases in

Mediterranean Hunter - Gatherer Research, edited by G. A. Clark, pp. 396-410, The University of Pennsylvania Press.

This material is to appear, in the same form as originally published, in my Ph.D. dissertation in the Arid Lands Resource Sciences Interdisciplinary Program at the University of Arizona. The dissertation, entitled Agriculture and Society in Arid Lands: A Hohokam Case Study," has been prepared under an option for the inclusion of published papers. This format requires require letters of permission for copyrighted materials. The entire dissertation, including the chapter cited above and copyright permission, will be reprinted by University Microfilms.

If you have any question, please contact me at your convenience.

e i -5 S: 0 6,'ii cIe P Sincerely, a lac as SiGe 47,4 4.4„

, tr. /1/10."1/ Suzanne K. Fish 7x,dx,„ esg 7'3 an-d 0-r eepirl.111/deia ?/CSS" 311

Chapter 21 Comparative Aspects of Paradigms for the Neolithic Transition in the Levant and the American Southwest Suzanne K. Fish and Paul R. Fish

Introduction The paradigms that govern the prehistory of a region tend to become reified and attain the status of basic assumptions from which all further interpretations proceed. Parochial frameworks for reconstruction are countered in current archaeology through resort to concepts drawn from broader anthropological observations and theory. However, re- gional paradigms may come to relate even these concepts in such crystallized combinations or sequences that flexibility and rearrange- ment become unlikely. Interregional comparison of paradigms that structure similar spheres of inquiry therefore offer a source of fresh perspectives and identify common points of weakness. We will examine some aspects of influential paradigms that shape our understanding of the origins of domestication economies in the Levant and the southwestern United States—Old and New World regions with which we are most familiar. Climatic, topographic, and other physical similarities in the two areas foster convergences in ana- lytical approaches to the variables affecting the transition from hunt- ing and gathering to food production. Parallels can also be seen in cultural patterns and trajectories attending this transition. Regional developments of a similar nature are disjunct in absolute time, but appear more synchronous in terms of elapsed time. For example, there are trends in both the Levant and the Southwest from earlier circular or oval pit house structures to later rectilinear, contiguous room com- plexes, culminating in aggregated settlements during the Neolithic and later prehistoric times respectively. It is probably not an accident 3 1 2

S. K. Fish and P. R. Fish 397

that a number of circum-Nfediterranean archaeologists studying the transition, including several contributors to this volume, have also worked in the Southwest.

Regional Contrasts With regard to the establishment of domestication economies, an im- mediate contrast is between the Levant as a region of primary domes- tication and the Southwest as an instance of secondary acquisition of extra-regional domesticates. This disparity is critical only if the focus of inquiry is on circumstances surrounding the first appearance of the morphological changes producing domesticates. However, if domes- tication economies are viewed as alternatives adopted under certain ranges of natural and cultural conditions and with particular implica- tions for practitioners. then indigenous or non-indigenous origin of domesticates is a minor issue (Earle 1980, Christenson 1980. Clark and Yi 1983). From this perspective, questions concerning combinations of variables under which agriculture was embraced and the nature of its consequences are equally appropriate for primary and secondary sit- uations. Distributions of wild ancestors generally encompass early domestica- tion loci in the Levant, although localized development and later trans- port to adjacent zones figure in some reconstructions. Among these are models of Braidwood and Howe (1960) for and Hole (1984) for and . In the Southwest, botanical markers for cultivation have always been extra-local and clearcut, a benefit not without draw- backs. With cultigen status beyond dispute for even the earliest corn, beans, and squash, inadequate attention has been devoted to identify- ing environmental and cultural preconditions more compatible with initial adoption in some Southwestern locales than in others. Because domestication in the Levant has been expected to entail a temporal continuum from wild to domesticated forms, predispositions such as intensified foraging, dependence on a few productive species, and incipient plant and animal manipulation have been proposed, sought and analyzed in greater depth. The presence of domesticated herd animals in the Levant and their absence in the Southwest is among the most portentious of regional contrasts. Animals provide an efficient means for harvesting the dis- persed vegetative resources of arid landscapes—with the exception of turkeys in the Southwest, animal products were obtained directly by hunting. On an interregional scale, animals undoubtedly account for dissimilar organization of seasonal activity and divergences in path- ways to sedentary and aggregated settlements (Redman 1978). The 313

398 Neolithic Transition in the Levant and American Southwest

potential of animals for capturing energy, expanding human labor, and extending scopes of operation also must have generated regionally different rates of progression toward and beyond neolithic cultural con figurations.

Nature and Rate of Change A renewed scientific interest and dialogue on the role of gradual versus punctuated equilibrium models within evolutionary schemes of all sorts is echoed in the archaeology of the neolithic transition. Gradualist views are being supplemented in Old and New World studies by mod- els based on chronologically compressed parameters of change. This tendency can be seen in recently advanced scenarios for the emergence of domesticate morphology from that of wild progenitors. Iltis (1983) has proposed a catastrophic genetic change to account for the differen- tiation of corn from ancestral teosinte. Likewise, Hillman and Davies (1990) use accumulating knowledge about wild cereals in the Near East to construct a model of successful human selection for a rare mutation within 20 to 200 years. They speculate that if the morphological changes creating domesticates occurred this rapidly and were rapidly disseminated thereafter, it might be impossible to find archaeological contexts encapsulating a botanical record of transition (Hillman and Davies 1990:200-201). Thus, the diffusion of cereals with domesti- cate morphology in the Levant, without a detectable prelude of inter- mediate types, may resemble the dissemination of exogenous corn in the Southwest to a greater degree than older reconstructions would suggest. Refinement of archaeological data has also led to models of more punctuated change. In the Southwest, inaccurate dating of the pre- sumed earliest corn created the appearance of up to 3,000 years during which this domesticate was present but created little visible effect on the economies or organization of Archaic populations. This impres- sion encouraged perceptions that its introduction was a "monumental nonevent" (Minnis 1983:310, also Irwin-Williams 1973, Ford 1981, Cordell 1984). These dates and their interpolations further pinpointed a prime highland environment from which agriculture was thought to have spread only gradually to less favored Southwestern locales (e.g., Haury 1962, Woodbury and Zubrow 1979, Ford 1981, Minnis 1985). Reexamination has now placed the anomalously dated corn later in time (Berry 1982, Wills 1988) and brought it in line with earliest instances in other parts of the region, including basin, plateau. moun- tain, and desert settings from central Utah to southern Arizona and New Mexico (e.g., Fish et al. 1986, Huckell 1990, Simmons 1986, 3 1 4

S. K. Fish and P. R. Fish 399

Wilde and Newman 1989, Wills 1988, Upham et al. 1987). Further- more, where data are sufficient to judge representation in these earliest contexts, direct measures of diet such as coprolite contents and bone chemistry (e.g., Minnis 1989, Matson 1990) and indirect ones such as ubiquity among charred remains (e.g., Huckell 1990) have identified corn as a major resource. The presence of domesticated cereals or livestock for any substantial period in the Levant prior to Neolithic horizons also has failed to be borne out by the archaeological record. Sickle blades, grinding imple- ments, substantial houses, and other Natufian attributes considered to be probable indicators of a food producing society have not been associated with domesticates in subsequent analyses (see Henry, Ols- zewski, this volume). Indeed, a particularly promising set of botanical remains thought to represent residues of Epipaleolithic plant cultiva- tion prior to the appearance of domesticates (Moore 1982, 1983). has since been reinterpreted as gathered resources (Hillman et al. 1989). The possibility exists for future evidence of cultivation, herding, or even domesticates in the poorly known latest part of the Natufian era. However, a pattern of relatively abrupt but geographically widespread appearance of cultivation and domesticates, lacking recognized pre- cursors, now appears to characterize both the early Neolithic of the Levant and the Late Archaic of the Southwest.

Paradigms for Preceding Hunters and Gatherers Concepts of social and economic modes characterizing sequential stages in cultural evolution enhance the comprehensibility of vast bod- ies of data and summarize meaningful trends, but at the expense of acknowledging or evaluating variability. In the basic schematization, innovations underlying food production appear, diffuse, and propel hunters and gatherers into a new neolithic stage in cultural develop- ment. Changes coincident with domestication economies are so dra- matic that the temptation to overlook the significance of variability in both preceding and subsequent stages may be reinforced. The tendency to regard hunters and gatherers as a primarily unitary phenomenon is inherent in the stage concept of cultural evolution, but is also fostered by a pervasive focus of current anthropological atten- tion on a particular subset of hunter-gatherer behavior. According to a prevailing, generalized template for hunters and gatherers preceding the transition, small bands with seasonally mobile lifestyles "map onto" spatially discrete and sequentially available resources, most of which can support only modest group sizes in single locations. Although archaeological entities directly subsuming the transition, and particu- 315

400 Neolithic Transition in the Levant and American Southwest

larly the Natufian of the Levant, are widely acknowledged to diverge to a lesser or greater degree from this model (e.g., Bar-Yosef 1981, 1983; Cauvin 1978, Henry 1985, 1989; Hillman et al. 1989, Moore 1982, Perrot 1966), its strictures are seldom seriously challenged for the immediately preceding Kebaran. Likewise, variants of this lifestyle are almost uniformly assumed for Archaic groups up to and usually in- cluding those possessing the earliest Southwestern domesticates (Wills 1988). In both the Levant and Southwest, the predominant model for hunters and gatherers is heavily influenced by a few ethnographies of bands exhibiting highly mobile foraging economies. Descriptions of the Paiute Indians in the Great Basin of the United States and the Bushmen of the Kalahari Desert in Africa can be singled out as par- ticularly influential in current perceptions (see Mueller-Wille and Dick- son, this volume). Not only have foremost students of the Paiute and Bushmen, notably Julian Steward and Richard Lee, been instrumental in shaping anthropological theory concerning hunters and gatherers, but the arid environments of these groups invite comparisons with the Southwest and the Levant.

Problems of Analogy In the last several years, questions have been raised as to the economic or political independence of modern groups still following a hunting and gathering lifestyle, and their representativeness of earlier hunters and gatherers (e.g., Wobst 1978, Bailey et al. 1989, Headland and Reid 1989, Schrire 1984). The Bushmen have been a particular target in this regard. It has been pointed out that Bushmen groups have maintained relationships with neighboring farmers and herdsmen for hundreds of years; individuals used to typify universal characteristics of hunters and gatherers participate on occasion in mixed economies (e.g. Wilm- sen 1989). The Paiute analogue is likewise derived from hunters and gatherers amidst an aggressive and ultimately dominant population of farmers and ranchers. Prior to the interval of ethnographic documentation, settlers had preempted better watered locations in Paiute territory, and their livestock were in the process of transforming the natural land- scape and removing grasses and other vegetation important for human subsistence (Winter 1976:426). These pressures discouraged or dimin- ished small scale plantings made by some bands. Many Paiute of the ethnographic interval interacted regularly with European colonizers and with adjacent groups of Native American cultivators. Portions of a broad belt between the Great Basin and Colorado Plateaus, inhabited 316

S. K. Fish and P. R. Fish 401

historically by mobile Paiute, were occupied prehistorically by Anasazi agriculturalists. While a case can be made that the Bushmen, Paiute, and other modern groups embody some essential qualities of hunters and gather- ers (e.g., Solway and Lee 1990, Silberbauer 1981), arguments can also be advanced that the totality of ethnographic patterns do not duplicate the probable range of variability of the late Pleistocene and earlier Holocene, and may include some uniquely recent configurations. With the exception of Australia, the historically documented sample of hunters and gatherers is environmentally biased toward extremes lim- iting agriculture and other kinds of marginality, a caution noted by Leslie Freeman (1968) and others but sometimes disregarded in cross- cultural studies. Peoples with high latitude (cold), maritime, jungle, and desert orientations make up an overwhelming majority of the hunter and gatherer sample. There is a virtual absence of analogues from more desirable locations for food production, which might have constituted prime territories for prehistoric foragers. Proximity and ongoing relationships with food producers may in some cases affect the expression of variables such as population size and density, sedentism, and storage behavior that are cited as figuring in the transition to agriculture. The balance between population and resources is one example (Cohen 1975, 1977). It is difficult to judge the independent viability of desert bands following a seasonal round, when resort is made to agricultural neighbors and wage labor in years of pronounced scarcity, as among some Bushmen (e.g. Headland and Reid 1989, Denbow 1984, Wilmsen 1983, 1989). It seems possible that geographic adjustments, greater storage, different population distri- butions, or other alternatives might have been necessary to counter starvation in the face of recurring droughts in earlier times.

Rethinking Variation If a probable breadth of variability among prior hunters and gatherers is acknowledged, questions regarding the emergence of modes that become more common during neolithic times can be framed in a manner other than as the departure from a previous norm. Rather than asking how triggering mechanisms might have pushed or pulled earlier groups away from "standard" hunter-gatherer behavior, it can be asked under what conditions such alternatives as greater sedentism or larger group size might have been advantageous and selected more frequently over time. A continuum exists between residential duration in base settlements of foragers practicing mobility (Binford 1980) and a degree of seden- 317

402 Neolithic Transition In the Levant and American Southwest

tism conducive to harvesting cultivated plants. In spite of drawbacks including waste accumulation and fuel depletion, a variety of eth- nographic hunters and gatherers have opted for this less mobile form of economic organization. A key variable in these patterns seems to be the availability of abundant staples that can be stored under given states of environment and technology. These are exemplified by the large game frozen by arctic hunters, the dried fish of Northwest Coast Indians, and the acorns of temperate California groups. It is not im- probable that storable staples. capable of supporting localized subsis- tence economies of this nature, are represented by foxtail in the Tehuacan Valley (Callen 1967:287), mammoth at Doini Vestonice (Soifer 1985), wild cereals and acorns in portions of the Levant (Henry (1989:93-94), or Indian rice grass in the northern Southwest and riparian mesquite groves in the south (Fish et al. 1990). In drier re- gions, the combination of a storable staple, permanent water, and nearby topographic diversity should predict the location of prior pat- terns of logistical mobility, relatively greater sedentism, and subse- quent early agriculture. Along with sites consisting of more ephemeral remains, items such as residential architecture and storage pits appear sporadically in the Upper Paleolithic and earlier Epipaleolithic of the Old World and pre- agricultural Archaic of the New World. In the Levant, multiple in- stances of Kebaran settlements with dwellings are known, in one case co-occurring at Ein Gev with Natufian counterparts (Bar-Yosef 1980:122). While similar elements have frequently been invoked in re- constructions of sedentary life or more complex organizational forms for the Natufian, qualitative equivalence is seldom accorded to occur- rences of these features in preceding times. If the model of hunters and gatherers as small mobile bands did not color perceptions so strongly, this variability in the archaeological record might better be interpreted to reflect the economic, social, and environmental variety among those populations who collectively gave rise to neolithic suc- cessors.

Paradigms for Transitional Entities As with preceding hunters and gatherers, stage concepts also encour- age monolithic treatment of archaeological manifestations that bracket the appearance of domesticates. The Natufian in the Levant and the terminal Cochise stages in the southern Southwest exhibit a number of elements typical of the ensuing neolithic periods. There are theoretical bases for understanding the origin and cohesion of these archaeo- 318

S. K. Fish and P. R. Fish 403 logical entities over their relatively broad geographic distributions. For example, sequential but overlapping interactions over wide areas among band and tribal peoples for mate acquisition, information ex- change, and risk sharing have been discussed by Wobst (1974, 1977) and others (e.g., Johnson 1978, Sahlins 1976). Widespread interac- tions of these kinds have even been applied in contexts of domestica- tion economies in the northern Southwest to explain the greater spatial extent of earlier (vs. later) design styles (Plog 1980, Braun and Flog 1982:513-515). The Natufian is regarded as a unitary phenomenon at some level (e.g., Henry 1973a, Byrd 1987:246-247, 1989; Bar-Yosef 1980:124, 1983:24; Smith 1972), in spite of internal variability in terms of en- vironmental setting (e.g., Henry 1985b, Cauvin 1978, Cauvin and Cauvin 1983:45), qualitative and quantitative assemblage characteris- tics (e.g., Henry 1973a, 1977; Olszewski 1984, Bar-Yosef 1981a, b), faunal remains (e.g., Cauvin 1978, Cauvin and Cauvin 1983:45; Ed- wards 1989), settlement pattern (e.g., Bar-Yosef 1983, Bar-Yosef and Goren 1973, Byrd 1987, 1989), and degree of sedentism (e.g., Henry 1985b, 1989; Bar-Yosef 1983, L. Binford 1968). Although relatively vast in size, the territorial magnitude of the Natufian is not so great as to preclude a quasi-ethnic conceptualization at implicit if not explicit levels of interpretation (see Bar-Yosef, this volume). In view of its distributional contiguity and shared stylistic and technological attri- butes, the unifying structure of the Natufian could also be compared to an interaction sphere or an environmentally based commonality in economic orientation. An alternative to explication by reference to synthetic concepts is insight derived from analyses of ethnographic situations. The ability to envision Southwestern archaeological manifestations against a rich and varied ethnographic backdrop is a unique regional advantage. A greater range of cultural variables relevant to the transition can be observed among Native American groups than among peasant vil- lagers and herders of the Near East. Hunters and gatherers, simply organized cultivators, groups with predominantly herding orienta- tions, ones with mixed subsistence economies, and aggregated agricul- turalists are represented in Southwestern ethnohistoric and ethno- graphic accounts. As analytical models for interaction among subsets of variables rather than as holistic patterns, these analogues can also enhance an understanding of archaeological manifestations beyond the Southwest. In particular, the spectrum of aboriginal societies in the Southwest offers an optimal vantage point for examining archaeologi- cally tangible aspects of cultural variability. 319

404 Neolithic Transition in the Levant and American Southwest

Synchronic Models of Variation An instance will be analyzed which might entail similar levels of both unity and variability embodied in the Natufian. By observing the syn- chronic distributions of variables of interest in Natufian research, some suggestions can be made concerning their articulation and the manner in which they might aid in posing future questions. Recurrent issues in the Levantine archaeological literature involve the degree of sedentism in the Natufian and the closeness with which Natufian subsistence practices approach incipient forms of cultivation. A Southwestern il- lustration of what can be termed a subsistence mosaic has implications for the pursuit of such problems. A Piman example from southern Ari- zona and adjacent portions of the Mexican state of Sonora serves not as a literal analogue for the Natufian phenomenon. but rather as a source of models for interrelationships among elements judged significant by archaeologists. Anthropologists emphasize unity among Pimans primarily as a lin- guistic phenomenon (Miller 1982:120, Kroeber 1934, Sauer 1934). Mutually intelligible dialects were spoken within an overall territory of roughly 120,000 km 2 (Miller 1982:114, Figure 1). This territorial size compares closely to that encompassing the Natufian distribution ac- cording to Bar-Yosef (1983:33, Figure 1). Ethnographically recognized Piman subgroups include the Pima, Papago, Sand Papago, and Lower Pima (Figure 21.1). These peoples considered themselves to pertain to an inclusive Piman ethnic classification, although there were more separately named divisions than the four large ones noted here. Inter- marriage from group to group was frequent. Trade in both consum- able items and durable materials such as shell was routine. Risk sharing among large and small subdivisions was institutionalized to the extent that formalized procedures existed for initiating intergroup food do- nation and delayed reciprocation (Russell 1975:93-95, Underhill 1939:100-107). A degree of technological and artifactual unity among Pimans is exemplified by shared ceramic attributes including modes of vessel manufacture, certain shapes, and decorative treatments. In spite of linguistic intelligibility and abundant social interaction, subsistence orientations and environmental settings could hardly have been more diverse. This diversity is of the nature of a subsistence mosaic, a series of somewhat overlapping but distinctive economic patterns forming a geographically contiguous network. The Piman subsistence mosaic encompasses almost the entire range of economic orientations documented for the Greater Southwest. The Lower Pima were high- elevation cultivators in the forested Sierra Madre of central Sonora. The Sand Papago emphasized a mobile hunting and gathering lifestyle 320

FIGURE 21.1. (A) Distribution of Natufian and Natufian-like assemblages in the Levant (after Bar-Yosef 1983:33, Fig. 1 and Byrd 1989:62, Fig. 1); (B) Distribution of Piman peoples in the Southwest (Miller 1982:120, Fig. 1); (C) Locations of Piman subgroups named in the text. 321

406 Neolithic Transition in the Levant and American Souttrwcit

in small groups in the driest lowland desert of western Sonora and Arizona. The Papago subsisted by a combination of floodwater farm- ing and intensive foraging in the more mesic desert zones of south- western Arizona. Most favorably situated along perennial desert rivers, the sedentary Upper Pima practiced riverine irrigation as well as flood- water farming, with foraging a more supplemental endeavor. The Piman analogy is not meant to imply that the element account- ing for Natufian identity among contemporary Epipaleolithic peoples of the Levant was necessarily linguistic. Its relevance lies in insights concerning the articulation of variables at different levels of cultural commonality. Variables pertinent to sedentism and resource patterns tend to co-vary with divisions of what has been termed the subsistence mosaic, which exists at a more restricted demographic and geographic scope than the higher order Piman linguistic unity. The Piman model indicates a scale for questions concerning subsistence practices, and suggests that the Natufian manifestation as a whole might have pro- duced a number of simultaneous answers. The structure of a subsistence mosaic is likely to play a major role in shaping archaeological distributions because of the technological and economic nature of much preserved material culture. However, dis- tributions for individual elements need not be isomorphic with one another or with subsistence modalities. Technological and stylistic commonalities in Pima and Papago water jars are among characteris- tics that extend beyond and cross-cut differing subsistence orientations of these two Piman subgroups (Fontana et al. 1962).

Diachronic Models of Variation

Synchronic variety within the Piman subsistence mosaic encompasses categories such as "hunter and gatherer" and "intensive agricultural- ist" which are usually ordered sequentially (i.e., temporally) in evolu- tionary schemes. While current anthropological theory does not legis- late that these categories or stages progress in mutually exclusive, unilineal series, regional paradigms influence perceptions toward that mold. The concept of a subsistence mosaic is also relevant IO diachronic patterns in the Southwest, whether a cultural unit or a geographic area is the focus of interest over time. Southwestern instances of exceptions to unilineality in temporal progression are particularly instructive. For an area as the analytical unit, the boundary between the Colo- rado Plateaus and Great Basin in the northern Southwest exemplifies an expression of a subsistence mosaic in the temporal dimension. Archaic hunters and gatherers and preceramic cultivators were suc- ceeded by agricultural peoples of the puebloan Anasazi and Fremont 322

S. K. Fish and P. R. Fish 407

ceramic traditions, and in turn were followed by the historic Paiute. These latest inhabitants were seasonally mobile foragers but included groups who tended stands of wild grasses or planted minor quantities of cultigens at more permanent bases. Subsistence reversals in this area occurred within highly abbreviated intervals relative to the Natufian timespan. For a cultural category as the analytical unit, historic Athabascans illustrate fluidity in economic orientations and mobility over the short term. Affected by post-contact pressures and opportunities, economic lifestyles among various Apache groups frequently shifted from hor- ticulture to mobile hunting and gathering and back again. Large seg- ments of the Navajo switched to a herding emphasis after the introduc- tion of sheep, often interacting with surrounding agriculturalists in a manner reminiscent of circum-Mediterranean mutualism. It is proba- bly not coincidental that later arriving Athabascans rather than long- established Puebloans exhibited more pronounced flexibility in life- styles in the historic Southwest. The potential for subsistence shifts away from sedentary and inten- sified tendencies would likely be greatest among groups whose social and demographic patterns had been least shaped by long traditions of such experience. During initial portions of the transition to domes- tication economies in both regions, commitments to emerging food producer stances are likely to have been more ephemeral, and multi- directional shifts more commonplace. Donald Henry (1985b:380, 1989) suggests deteriorating climate as an impetus for such a reversal in Natufian subsistence trends. In the marginal Negev and Sinai des- erts, intense involvement with grasses during a mesic interval was replaced by more diversified hunting and gathering near the end of the Natufian period. Under such circumstances, remains of rela- tively sedentary groups could occur stratigraphically sandwiched be- tween layers of mobile band occupations or be deposited in other unanticipated combinations.

Paradigms and Variability in Settlement Patterns Aspects of archaeological settlement patterns are likely to diverge from expectations of uniformity evoked by stages in cultural evolution. Lithic scatters of limited size and density are commonly equated with extractive activities of temporary duration. Where diagnostics indicate chronological placement within the general time frame of transition, such remains and activities may more readily be attributed to mobile hunters and gatherers than to incipient agricultural groups, a conven- tion and interpretational bias arising from regional paradigms. 323

408 Neolithic Transition in the Levant and American Southwest

In the northern Southwest, an automatic, traditional assignment of several modest lithic scatters to nonagricultural Late Archaic groups would have obscured some of the earliest evidence for cultigens. Three sites studied by Alan Simmons (1982. 1986) yielded lithic assemblages totaling between fifty and several hundred artifacts after excavation. Persistence in thoroughly investigating these sites beyond surface in- dications led to recovery of both pollen and macrobotanical indicators of corn. Initially, contradiction of paradigmatic expectations occa- sioned strong skepticism in the archaeological community that was overcome only with additional field investigation and confirmatory evidence (Simmons 1986). In the low deserts of southern Arizona, several unanticipated associations have likewise been found between remains of corn and similarly unimpressive Archaic sites (P. Fish et al. 1986, Dart 1986, S. Fish et al. 1990). It is significant that these Arizona sites are now known to be parts of preceramic settlement systems which also include sites with diverse artifact assemblages, varied trade goods, massive midden deposits, storage pits, substantial habitation struc- tures, and cemeteries (Huckell 1990) comparable to Natufian counter- parts at Hayonim Cave (Bar-Yosef and Goren 1973) or Ein Mallaha (Perrot 1966). It seems easiest to misassign and misinterpret the small-scale ele- ments of settlement patterns most apt to lack diagnostic artifacts. As- semblages in the small Southwestern lithic sites yielding corn were insufficient per se for assignment to settlement systems with or without agriculture; the presence of cultigen remains constituted the effective criterion. Wild resources are more strongly associated with models of hunter-gatherer strategies than with ones involving domesticates; a lack of diagnostics may particularly hamper recognition of extractive sites indicating important components of domestication economies. Problems of recognition loom larger in the Levant, where approx- imately 2,000 years of the early Neolithic antedate the appearance of pottery as an abundant and highly visible marker. In terms of elapsed time following the initial appearance of regional domesticates, late prehistoric societies of the Southwest near the time of European con- tact would be roughly equivalent to ones in the Levant during the first several centuries of the ceramic Neolithic. Because Southwestern pot- tery appears from 1000 to 1500 years after cultigens, even small extrac- tive sites with indistinctive lithics can routinely be linked to relatively early agricultural timespans, and often to particular ceramic phases. Near Tucson, Arizona, remains of the most intensive extractive ac- tivity at a distance from water sources correlate with densest settlement during the Hohokam Classic Period between A.D. 1100 and 1350 (Fish et al. 1989). This agricultural period, marked by site size hierarchies 324

S. K. Fish and P. R. Fish 409

and mounds as public architecture, is separated in time from the earliest local farming by an interval approximating that for the Pre- pottery Neolithic B in the Levant. Mobile or lower density populations in arid environments can emphasize hunting and gathering near water sources, but with increasing sedentism and population growth, con- veniently located wild resources may be depleted. It has been sug- gested by Cohen (1977), Hayden (1982:529), and others (e.g., Conklin 1969:229-230) that pressure on preferred or easily acquired and pro- cessed resources may lead to broader spectrum diets and generally intensified gathering among , a process commensurate with this Hohokam example. In the absence of ceramics, misinterpre- tation of the chronological affiliation of such activities would reduce understanding of subsistence systems in both earlier and later portions of the cultural sequence.

Concluding Observations Regional paradigms serve to organize results of previous study in a logical order and provide a basis for shared understanding. Influential synthetic concepts and analogues that structure our perceptions about the neolithic transition in the Southwest and Levant suffer from a common failure to provide frameworks for describing and evaluating variability at multiple scales. Not only can overly monolithic percep- tions hamper our detection and communication of the variation inher- ent in regional archaeology, but they can also circumscribe the field of inquiry and channel our questions into well worn pathways. The nature and significance of variation should not be overlooked in the era of dramatic change in human history from predominantly foraging modes to ones based on food production. Among native peoples of the Southwest, the transition was long term, with persistent exceptions. Foraging remained a prominent component of subsistence systems, and nonagricultural lifestyles did not disappear. The mainte- nance of Southwestern variation can be approached in the context of alternatives or options followed under particular conditions, whether from negative standpoints such as environmental deterioration or pos- itive standpoints such as the energetic efficiencies of hunting and gathering when resources are abundant and populations are corre- spondingly low. An eventually more comprehensive shift to domestication econo- mies was realized in the Levant, although multiplicity also emerged in these forms. The additional domestication of animals in the Old World is undoubtedly a critical factor in this regional contrast. Without live- stock, Southwestern Indians transformed the natural environment to a 325

410 Neolithic Transition in the Levant and American Southwest lesser degree, a process which accelerated rapidly in post-contact times. Progressive elimination of previous demographic and economic alter- natives due to the historic introduction of herd animals must mirror similarly irreversible effects in the Neolithic Levant. Interregional studies offer opportunities for conceptual cross- fertilization. Comparison of primary zones of food production as in Mesoamerica and the Levant are instructive; inclusion of secondary instances should reveal parallels and contrasts at scales above the out- comes of unique or localized domestication processes. Secondary ac- quisition of domesticates is unlikely to have been a process limited to singular classes of isolated elements. Other facets of early agricultural economy and organization influenced by donor groups may have af- fected the locations, timing, and rate at which those transitions pro- ceeded. Comprehensive comparisons of elapsed time between initial appearance of domesticates and other economic and social thresholds in both primary and secondary contexts are particularly intriguing subjects for further study. 326

References Cited

Bailey, R., G. Head, M. Jenike, B. Owen, R. Rechtam, and E. Zechenter 1989 Hunting and gathering in tropical rain forest: is it possible? American Anthropologist 91: 59-82.

Bar-Yosef, O. 1980 Prehistory of the Levant. Annual Review of Anthropology 9:101-133.

1 98 1 a The epipaleolithic complexes in the southern Levant. In Prehistoire du Levant (J. Cauvin and P. Sanlaville, eds.), pp. 551-570. Paris: Editions du CNRS.

1 98 1 b The "Pre-Pottery Neolithic" period in the southern Levant. In Prehistoire du Levant (J. Cauvin and P. Sanlaville, eds.), pp. 551-570. Paris: Editions du CNRS.

1 983 The Natufian in the southern Levant. In The Hilly Flanks and Beyond (T.C. Young, Jr., P. Smith, and P. Mortenson, eds.). Chicago: Oriental Institute Studies in Ancient Oriental Civilization No. 36.

Bar-Yosef, O., and N. Goren 1973 Nautufian remains in Hayonim Cave. Paleorient 1:49- 68.

Berry, M. S. 1982 Time. Space. and Transition in Anasazi Prehistory. Salt Lake City: University of Utah Press.

Binford, L. R. 1968 Post-Pleistocene adaptations. In New Perspectives in Archaeology (S. and L. Binford, eds.), pp. 313-341. Chicago: Aldine. 327 1980 Willow smoke and ' tails: hunter-gatherer settlement systems and archaeological site formation. American Antiquity 45: 4-20.

Braidwood, R. J., and B. Howe 1960 Prehistoric Investigations at Iraqi Kurdistan. Chicago: Oriental Institute Studies in Ancient Oriental Civilizations No. 31.

Braun, D., and S. Plog 1982 Evolution of "tribal" social networks: theory and prehistoric North American evidence. American Antiquity 47: 504-525.

Byrd, B. 1987 Beidha and the Nautufian: Variability in the Levantine Settlement and Subsistence. Ph.D. dissertation. Department of Anthropology, University of Arizona, Ann Arbor, University Microfilms.

1989 The Natufian: settlement variability and economic adaptation in the Levant at the end of the Pleistocene. Journal of World Prehistory 3: 159-198. Callen, E. O. 1 9 6 7 Analysis of the Tehuacan coprolites. In The Prehistory of the Tehuacan Valley (Vol. 1): Environment and Subsistence (D. S. Byers, ed.), pp. 261-289. Austin: University of Texas Press.

Cauvin, J. 1978 Les premiers villages de Syrie-Palestine du IXeme-au VIIememillenaire avant. J. C. Lyon : Maison de l'Orient Mediterraneen. 328

Cauvin, J., and M-C. Cauvin 1983 Origines de l'agriculture au Levant: facteurs biologiques et socio-culturels. In The Hilly Flanks and Beyond (T. C. Young Jr., P. Smith, and P. Mortenson, eds.). Chicago: Oriental Institute Studies in Ancient Oriental Civilization No. 36.

Christenson, A. 1980 Change in the human niche in response to population growth. In Modeling Change in Prehistoric Subsistence Economies (T. Earle and A. Christenson, eds), pp. 31-72. New York: Academic Press. Clark, G., and S. Yi 1983 Niche width variation in Cantabrian archeofaunas: a diachronic study. In Animals and Archaeology I: Hunters and Their Prey (J. Clutton-Brock and C. Grigson, eds.), pp. 325-350. Tempe: Arizona State University Anthropological Research Papers No. 36. Cohen, M. 1975 Archaeological evidence of population pressure in pre- agricultural societies. American Antiquity 40:471-475.

1977 The Food Crisis in Prehistory. New Haven, CT: Yale University Press. Conklin, H. 1 969 An ethnoecological approach to shifting agriculture. In Environment and Cultural Behavior (A. Vayda, ed.), pp. 36-49. New York: Natural History Press. Cordell, L. 1984 Prehistory of the Southwest. Orlando, FL: Academic Press. 329

Dart, A. 1 9 8 6 Archaeological Investigations at La Paloma: Archaic and Hohokam Occupations in the Northeastern Tucson Basin. Tucson, AZ: Institute for American Research Anthropological Paper No. 4.

Denbow, J. R. 1984 Prehistoric herders and foragers of the Kalahari: the evidence for 1500 years of interaction. In Past and Present in Hunter-Gather Studies (C. Schrire, ed.), pp. 175-193. New York: Academic Press.

Earle, T. 1980 A model of subsistence change. In Modeling Change in Prehistoric Subsistence Economies (T. Earle and A. Christenson, eds.), pp. 1-30. New York: Academic Press.

Edwards, P. C. 1989 Revising the broad spectrum revolution: its role in the origins of southwest Asian food production. Antiquity 63: 225-243.

Fish, P., S. Fish, A. Long, and C. Miksicek 1986 Early corn remains from Tumamoc Hill, southern Arizona. American Antiquity 51: 563-572.

Fish, S., P. Fish, J. Madsen 1989 Classic Period Hohokam community integration in the Tucson Basin. In Sociopolitical Structure of Prehistoric Southwestern Societies (K. Lightfoot, S. Upham, and R. Jewett, eds.). Boulder: CO: Westview Press.

1990 Sedentism and settlement mobility in the Tucson Basin prior to A.D. 1000. In Perspectives on Southwest Prehistory (D. Minnis and C. Redman, eds.), pp. 76-91. Boulder, CO: Westview Press.

Fontana, B., W. Robinson, C. Cormack, and E. Leavitt 1962 Papago Indian Pottery. Seattle: University of Washington Press. 330

Ford, R. I. 1981 Gardening and farming before A.D. 1000: patterns of prehistoric cultivation north of Mexico. Journal of Ethnobiology 1: 6-27.

Freeman, L. G. 1968 A theoretical framework for interpreting archaeological materials. In Man the Hunter (R. Lee and I. Devroe, eds.), pp. 262-267. Chicago: Aldine.

Haury, E. 1962 The greater American Southwest. In Courses Toward Urban Life: Some Archaeological Considerations of Cultural Alternates (R. J. Braidwood and G. R. Willey, eds.), pp. 106-131. New York: Viking Fund Publications in Anthropology No. 32.

Headland, T. N., L. A. Reid 1989 Hunter-gatherers and their neighbors from prehistory to the present. Current Anthropology 30: 43-66.

Henry, D. O. 1 9 7 3 a The Nautifian of Palestine: Its Material Culture and Ecology. Ph. D. dissertation. Department of Anthropology, Southern Methodist University, Dallas.

1977 An examination of the artifact variability in the Nautifian of Palestine. In Eretz-Israel, Vol. 13 (B. Arensburg and O. Bar-Yosef, eds.), pp. 229-239. Jerusalem: Hebrew University.

1985b Preagriculture sedentism: the Natufian example. In Prehistoric Hunter-Gatherers: The Emergence of Cultural Complexity (T. D. Price and J. A. Brown, eds.), pp. 365- 384. New York: Academic Press.

1989 From Foraging to Agriculture: The Levant at the End of the Ice Age. Philadelphia: University of Pennsylvania Press. 331

Hillman, G. C., S. Colledge, and D. Harris 1989 Plant food economy during the Epi-Paleolithic period at Tell Abu Hureyra, Syria: dietary diversity, seasonality, and modes of exploitation. In Foraging and Farming: The Evolution of Plant Exploitation (D. Harris and G. Hillman, eds.), pp. 240-268. London: Unwin and Hyman.

Hillman, G. C., and M. S. Davies 1990 Measured domestication rates in wild and barley under primitive cultivation, and their archaeological implications. Journal of World Prehistory 4: 157-222.

Hole, F. 1984 A reassessment of the . Paleorient 10: 49-60.

Huckell, B. 1990 Agriculture and Late Archaic Settlements in the River Valleys of Southern Arizona. Ph.D. dissertation. Department of Arid Lands Studies, University of Arizona. Ann Arbor: University Microfilms.

Iltis, H. 1983 From teosinte to : the catastrophic sexual transmutation. Science 222: 886-894.

Irwin-Williams, Cynthia 1973 The Oshara Tradition: origins of Anasazi culture. Eastern New Mexico Contributions to Anthropology 5: 1- 30.

Johnson, Gregory A. 1978 Information sources and the development of decision- making organizations. In Social Archaeology: Beyond Subsistence and Dating (C. Redman, M. Berman, E. Curtin, W. Langhorne, N. Versaggi, and J. Wanser, eds.), pp. 87- 112. New York: Academic Press. 332

Kroeber, A. L. 1934 Uto-Aztecan languages of Mexico. Ibero-Americana 8: 23 -3 7 .

1 93 9 Cultural and Natural Areas of Native North America. Berkeley University of California Publications in American Archaeology and Ethnology No. 38.

Matson, R. G. 1990 Dietary Change in the Upland Southwest: The Coming of Agriculture. Paper presented at the 55th annual meeting of the Society for American Archaeology, Las Vegas.

Miller, W. 1 982 Uto-Aztecan languages. In Handbook of North American Indians, Vol. 10: Southwest (A. Ortiz, ed.). Washington, D.C.: Smithsonian Institution.

Minnis, Paul 1985 Domesticating plants and people in the Greater American Southwest. In Prehistoric Food Production in North America (R. I. Ford, ed.), pp. 309-340. Ann Arbor: Museum of Anthropology, University of Michigan, Anthropological Papers No. 75.

1989 Prehistoric diet in the northern Southwest: macroplant remains from Four Corners feces. American Antiquity 54: 543-563.

Moore, Andrew M. T. 1982 Agricultural origins in the Near East--a model for the 1980's. World Archaeology 14: 224-236.

1983 The first farmers in the Levant. In The Hilly Flanks and Beyond (T. Young, P. Smith, and P. Mortensen, eds.), pp. 91-112. Chicago: Oriental Institute Studies in Ancient Oriental Civilization No. 36. 333

Olszewski, Deborah I. 1984 The Early Occupation at Tell Abu Hureyra in the Context of the Late Epipaleolithic of the Levant. Ph.D. Dissertation, University of Arizona. Ann Arbor: University Microfilms.

Perrot, Jean 1966 Le Gisement Natoufien de Mallaha (Eynan), Israel. L'Anthropologie 70: 437-483.

Plog, Steve 1980 Stylistic Variation in Prehistoric Ceramics. Cambridge: Cambridge University Press.

Redman, Charles L. 1978 The Rise of Civilization. San Francisco: W. H. Freeman.

Russell, F. 1975 The Pima Indians. Tucson: University of Arizona Press.

Sahlins, Marshall 1976 Tribesmen. Englewood Cliffs, N.J.: Prentice-Hall.

Sauer, C. 1 934 The distribution of aboriginal tribes and languages in northwestern Mexico. Ibero-Americana 5: 13-21.

Schrire, C., ed. 1984 Past and Present in Hunter-Gatherer Studies. New York: Academic Press.

Silberbauer, G. B. 1981 Hunter and Habitat in the Central Kalahari Desert. Cambridge: Cambridge University Press.

Simmons, Alan 1 982 Prehistoric Adaptive Strategies in the Chaco Canyon Region,Northwestern New Mexico. Window Rock: Navajo Nation Papers in Anthropology No. 9. 334

1986 New evidence for the early use of cultigens in the American Southwest. American Antiquity 51: 73-88.

Smith, P.E.L. 1972 Diet and nutrition in the Natufians. American Journal of Physical Anthropologist 37: 233-238.

Soffer, O. 1985 The Upper Paleolithic of the Central Russian Plain. New York: Academic Press.

Solway, J. and R. Lee 1990 Foragers, genuine or spurious? Situating the Kalahari San in history. Current Anthropology 31: 109-146.

Underhill, R. 1 939 Social Organization of the Papago Indians. New York: Columbia University Press.

Upham, Steadman, R. MacNeish, W. Galinat, and C. Stevenson. 1987 Evidence concerning the origin of Maiz de ocho. American Anthropologist 89: 410-419.

Wilde, J. and D. Newman 1989 Late Archaic corn in the eastern Great Basin. American Anthropologist 91: 712-719.

Wills, W. 1 988 Early Prehistoric Agriculture in the American Southwest. Santa Fe, N.M.: School of American Research.

Wilmsen, E. 1983 The ecology of illusion: anthropological foraging in the Kalahari. Reviews in Anthropology 10: 9-20.

1989 Land Filled with Flies: A Political Economy of the Kalahari. Chicago: University of Chicago Press. 335

Winter, J. 1976 The process of farming diffusion in the Southwest and Great Basin. American Antiquity 41: 421-429.

Wobst, H. M. 1974 Boundary conditions for paleolithic social systems: a simulation approach. American Antquity 39: 147-178.

1977 Stylistic behavior and information exchange. In For the Director: Research Essays in Honor of James B. Griffin (C. Cleland, ed.). Ann Arbor: University of Michigan, Museum of Anthropology, Anthropological Papers No. 61.

1978 The archaeo-ethnology of hunter-gatherers, or the tyranny of the ethnographic record in archaeology. American Antiquity 43: 303-309.

Woodbury, R.B. and E. Zubrow 1979 Agricultural beginnings, 2000 B.C. - A.D. 500. In Handbook of North American Indians. Vol. 9: Southwest (A. Ortiz, ed.), pp. 43-60. Washington: Smithsonian Institution Press.