EIGHT
Hohokam Household Organization, Sedentism, and Irrigation in the Sonoran Desert, Arizona
INTRODUCTION
Studies of Hohokam households have most often focused on identifying and describing them as elements of the distinctive social structure that emerged in the low desert areas of the Southwest. Such studies have generally taken a static view that emphasizes continuity through time and space. Recent anthropological theories, however, take a more dynamic view of household structure and organization, linking changes to processes such as increased agricultural dependency and sedentism (Flannery 1972; Netting 1990; Wilk and Rathje 1982). According to these theories, agricultural dependence influences the degree of sedentism, the makeup of the units of production, and the systems of land tenure in which households participate (see Neff, this volume). As is envisioned by many of the authors in this volume, the household is the elemental social and economic unit in preindustrial societies. It is isomorphic with the basic unit of consumption and is distinguished from the reproductive unit commonly known as the family (Ciolek-Torrello and Reid 1974; Goody 1972; Wilk and Rathje 1982; also see Douglass and Gonlin, this volume). The family as a set of individuals and the household to which these individuals may belong are not necessarily the same (Buchler and Selby 1968:21–23).
Two basic models describe the transition to a sedentary, agricultural way of life in the prehistoric Southwest (Gilman 1997:4; for discussion of the transition through time of groups in other areas, see chapters by Beaule, González Fernández, Henderson, and McCormack, this volume). One model suggests that agricultural dependence and residential sedentism came quickly following the introduction of agriculture, occurring as early as the last centuries BC in the Late Archaic period. In accordance with this model, a distinctive pattern of domestic organization developed early and persisted essentially unchanged throughout the prehistoric period despite substantial changes in architectural forms and materials. This model has been of great appeal to Hohokam archaeologists, who have noted the abundance of wild resources along the ancient desert rivers, the ubiquity of corn in Late Archaic and Early Formative contexts, and the persistence of settlements over periods spanning hundreds of years (Fish, Fish, and Madsen 1992; Nelson and Schachner 2002:190–191).
By contrast, the second model states that agricultural dependence and residential sedentism came much later, perhaps over 1,000 years later, because people initially used agriculture only to maintain a hunting and gathering subsistence system (Gilman 1987, 1997; Wills 1991). Proponents of the second model have argued that agricultural dependence and sedentism, which are closely linked in the first model, may function independently (Graham 1994; Rocek 1995). They argue that in the Southwest, intensive agricultural production may have preceded sedentism (see Douglass and Gonlin, this volume). The second model emphasizes the differences between early pithouse architecture and later aboveground masonry and adobe architecture, along with substantial changes in material culture, demography, diversification of agriculture, and the development of intensive agricultural technologies that occurred late in prehistory as important indicators of changes in economy and mobility patterns. Gilman (1997), for example, finds it difficult to explain such substantial changes if sedentism and agricultural dependency occurred early in the prehistory of the Southwest (see also Ciolek-Torrello 1995, 1998a; Whittlesey and Ciolek-Torrello 1996). Although the notion of mobile or non-sedentary farmers has been readily accepted by many archaeologists working in the Colorado Plateau and Mogollon regions (Gilman 1997; Nelson and LeBlanc 1986; Powell 1983), it has found few adherents among those studying the Hohokam. This situation is due in great part to widely held views regarding the greater stability and persistence of Hohokam culture in comparison to other Southwestern cultures (Nelson and Schachner 2002:190–191).
In this chapter, I apply these recent anthropological theories and models of agricultural dependency and sedentism to examine changes in household structure in the Sonoran desert region occupied by the Hohokam from the Late Archaic to the Classic period. I infer household organization from architecture and site structure and relate these patterns to the development of canal systems and other indicators of agricultural intensification.
AGRICULTURAL DEPENDENCE AND RESIDENTIAL MOBILITY
Two variants of the second or gradual development model are useful for examining the relationship among household organization, agricultural dependence, and sedentism. In a model originally presented in 1972 (and revisited in 2002), Flannery argues that intensification of agricultural strategies influences the degree of sedentism, the makeup of the units of production, and the systems of land tenure in which households participate. In the archaeological record, intensification is manifested by the construction of agricultural features requiring investment of labor and the development of new technologies. Changes in the degree of sedentism and units of production are reflected in household and settlement structure. Flannery (1972:325) relates mobility patterns, organization of production, and control of key resources to different types of household and settlement structure. Using ethnographic data, he argues that circular houses are found more often in relatively mobile societies that exhibit a low degree of agricultural dependence. By contrast, more modular, rectangular architecture is found among groups that move less frequently and invest greater effort in food production. In a recent revision of this model, Flannery de-emphasizes these architectural differences while placing greater emphasis on changes in resource sharing and the location of storage facilities. More recently, Flannery (2002:421) argues that his main distinction is between settlements where small huts are occupied by individuals and storage is shared, and settlements where larger houses are occupied by whole nuclear families and storage is private (see also Gilman 1987).
According to Flannery’s model, settlements formed of small, circular houses reflect a fluid and flexible social structure similar to that of hunter-gatherer bands that aggregate during periods of local resource abundance. The entire group is the unit of production in the settlements formed by these temporary aggregations, and resource sharing is necessary. This informal domestic arrangement is generally associated with low-intensity land-use patterns characterized by low population densities, long-fallow cycles, shifting cultivation, and low labor investment in agricultural production (see also Boserup 1965; Howard 1993; Wills 1991). Systems of land tenure are informal as well, and exercise of ownership rights to residential and agricultural lands is relatively lax. The archaeological correlates of this economic and organizational pattern are the arrangement of dwellings in circles, common ownership of storage facilities and their location in extramural areas, and occupation of houses by individuals, not households (see chapters by Neff and Wiewall, later in this volume, for other discussions of land tenure).
Flannery contrasts this type of settlement with those having larger, rectangular architecture and a modular structure; the members of each household reside together in houses that may be subdivided into rooms, and each has its own private storage facilities. The modular structure of the settlement gives it a planned look and allows for accretionary growth without disrupting the residential unit (Flannery 2002:425). The basic economic unit in these settlements is the nuclear household, which produces for its own use in accordance with its own perceived needs. He relates this type of settlement structure to increased investment of labor in food production, reduction of fallow cycles, increased sedentism, and the development of private landholdings (see also Howard 1993).
Emphasizing risk and privatization of storage space, Flannery (2002:421) maintains that these settlements have a more “closed” site plan to avoid jealousy and conflict that might arise from differences in food and wealth. He adds that such a “closed” site plan might also be reflected in the enclosure of extramural work spaces with mud walls. Under such circumstances, there exists more incentive to intensify production, since any resulting surplus does not have to be shared.
Flannery (2002:423–425) adds a third type of settlement composed of larger social and residential units. He argues that in some regions, villages composed of nuclear households were replaced by larger, multigenerational extended households containing multiple hearths, kitchens, and storage facilities. This third type of settlement structure suggests even greater residential permanence to Flannery, as it resulted from offspring remaining attached to parental households rather than moving away when they reached adulthood and married. Larger households may have been a response to the need to divide labor among various simultaneous tasks because resources or field systems were dispersed or because of the greater demands for pooled labor to construct and maintain intensive irrigation systems (see Netting, Wilk, and Arnould 1984; Pasternak, Ember, and Ember 1976; see also detailed discussions of small and large social and residential units in Henderson, this volume).
The second variant of the gradualist model relating sedentism and agricultural dependence focuses on architecture and de-emphasizes household arrangements. Using ethnographic and archaeological data, several archaeologists have suggested that people living in pithouses did not inhabit them year round but instead were residentially mobile for at least part of the year (Diehl 1992; Gilman 1987, 1997; Rocek 1995; Varien 1999). Rocek (1995:205), however, warns that archaeological examples of pithouse architecture cannot automatically be interpreted as residences of mobile groups. A corollary of this argument is that residents of pithouses were not fully dependent on agriculture (Gilman 1987:560; 1997:2). Ethnographic studies reveal that pithouses are most common in regions with at least one cold season per year; in cultures with at least biseasonal settlement patterns, where food is concentrated during one season; and, most important, where such concentrated food supplies can be stored (Gilman 1987, 1997). If agriculture is part of the subsistence system, much of the stored food may be in the form of agricultural products (Gilman 1987:546). Gilman (1997:6) also suggests that pithouses are not generally occupied during the hotter summer months, because of the higher seasonal temperatures or because of greater residential mobility during this time of the year. Instead, arguing for the greater thermal efficiency of subterranean structures, Gilman (1987:542) maintains that pithouses were usually occupied in winter, when their residents were dependent on stored food supplies. It must be made clear, however, that we are not suggesting that pithouses were occupied by highly mobile groups. Rather, pithouse settlements reflect a shift to a substantially reduced level of seasonal and interannual mobility or a mixture of settlement strategies, in contrast to the ephemeral settlements of more mobile hunter-gatherers (Rocek 1995:210–211; see also Gilman 1997:6).
McGuire and Schiffer (1983), for example, have proposed that the pit-house-to-pueblo transition in the Anasazi region of the Southwest was closely related to the shift to a more sedentary life style. They argue that pithouses are relatively quick and easy to build, especially by people who are residentially mobile, but they are difficult to maintain and have relatively short use-lives. Pithouses have a maximum use-life of about twenty-five to thirty years (Ahlstrom 1985; Cameron 1990; Craig 2000; Diehl 1992). By contrast, McGuire and Schiffer (1983:292–294) maintain that pueblos have longer use-lives and are more versatile to use but are more difficult to build (cf. Wilshusen 1988). They assume that as the Anasazi became more sedentary, they were willing to invest more effort in construction to get houses with a longer use-life and lower maintenance costs. In a similar view, Diehl and Gilman (1996) contend that ephemeral structures are low cost and briefly occupied, pithouses are moderate in cost and are inhabited for intermediate lengths of time, and masonry pueblos are most costly to build but are occupied for the longest periods. From such arguments, Gilman (1997:3) proposes that pit structures are in the middle of a continuum between ephemeral structures and the more substantial stone and adobe pueblo structures that replaced them in the Southwest.
In an argument that closely parallels Flannery’s, Gilman (1987:556) suggests that the primary factor in the shift from pithouses to aboveground pueblos was increasing dependence on agriculture, specifically, the storage, preparation, and cooking of food. Increased time devoted to each of these activities and the need for greater storage capacity resulted in the increasing differentiation of household space into discrete rooms or structures devoted to individual activities (Gilman 1987:558). (Byrd [2000:65] interprets similar changes in interior domestic space during the transition to sedentary food production in southwest Asia as a reflection of increasing household autonomy and changes in the way nuclear households interacted with one another.) In Gilman’s view, storage space became an increasingly critical factor not only because of greater dependence on agricultural products but also because of the need to store surpluses for more than one year as a hedge against crop failure and to retain sufficient seeds for planting (see Van West and Altschul 1994). Furthermore, a shift from underground and outdoor storage to aboveground, indoor storage resulted from the need to protect vital food supplies from dampness, vermin, and thievery for longer periods of time (Gilman 1987:559). Ethnographic data suggest that underground storage features are indicative of food caching associated with seasonal site abandonment and are most effective when they are sealed for prolonged periods and accessed intermittently (Rocek 1995:207). Storage pits would not be effective if they were accessed continuously by a resident population. Large aboveground ceramic containers were perhaps the least costly and most effective facilities for protecting and accessing stored foods, especially agricultural cereals (Gilman 1997:8; Whittlesey and Ciolek-Torrello 1996). Similarly, the increased time devoted to processing and cooking cereals to make them more digestible also would have required a greater number and variety of containers.
TRANSITION TO SEDENTARY FARMING SETTLEMENTS
Although these models have been used to describe various aspects of the pithouse-to-pueblo transition in the Anasazi and Mogollon regions of the Southwest, they are relevant to interpreting changes in prehistoric household and settlement structure in Sonoran desert areas, such as the Phoenix Basin (Figure 8.1). Here, too, the span of time from the Late Archaic period, around 1500–1000 BC, to the Classic period, between AD 1200 and 1450 (Figure 8.2), witnessed the transition from a residentially mobile, broad-based foraging/farming society to one of large sedentary, agriculturally dependent villages in the river valleys of the region.
Figure 8.1. Map of south-central Arizona
Late Archaic and Early Formative Period Beginnings
The first farming settlements appeared in the Southwest during the Late Archaic period between 1100 and 600 BC (Mabry 2000:39). These settlements contained small, ephemeral pithouses with intramural and extramural storage pits and thick midden deposits indicating an intensive occupation. Larger settlements with numerous small pithouses occurred in better-watered areas in the Cienega phase (600 BC to AD 100), the terminal phase of the Late Archaic. These larger settlements contained clusters of houses in discrete residential groupings (Figure 8.3). Some of the structures in these house groups reflected specialized storage or integrative functions. Mabry (2000:39) interprets this development as indicative of a trend toward increasing sedentism and the presence of villages composed of multiple extended families.
Figure 8.2. Chronology for the Sonoran desert
Architectural changes in the following Agua Caliente–Red Mountain phase (ca. AD 100–500), the initial phase of the Early Formative period, include the construction of larger, deeper, and more substantial structures and a shift to bean-shaped and subrectangular house forms, pointing to further increases in sedentism (Ciolek-Torrello 1995; Mabry 2000; Figure 8.4). These changes are associated with the appearance of small ceramic cooking and storage vessels and a shift to an expedient stone tool technology that came to be associated with later desert cultures.
Figure 8.3. Late Archaic period houses
Figure 8.4. Pioneer period house cluster at the Eagle Ridge site (after Elson and Lindeman 1994:fig. 3.25)
By the first centuries AD, mobility patterns appear to have been reduced to seasonal movements between floodplain settlements and smaller upland campsites. Floodplain settlement was based largely on expedient or casual cultivation of maize using floodwater farming and exploitation of wild riparian resources. In the Phoenix Basin proper, these farming methods may have been aided by small irrigation canals tapping the Salt River near the active floodplain (Ackerly 1989; Howard 1991). A biseasonal settlement pattern may have followed. Villages were distributed along canals and in the floodplain in the fall, winter, and spring seasons and dispersed into summer farmsteads in the floodplain and surrounding upland areas (Cable and Doyel 1987; Howard 1991; Wilcox 1979).
Despite the early occurrence of canals, it remains unclear whether these early settlements were maize dependent and sedentary. With the exception of the canals, food-production and -processing technology and the organization of settlements and domestic life appear to have remained geared to a more mobile lifestyle and generalized mode of subsistence. Whittlesey and Ciolek-Torrello (1996) argue that the use of storage pits, the small number and size of ceramic vessels, and the exploitation of plant and animal resources from a variety of locales during the Early Formative period indicate continued residential mobility and a lack of dependence on agricultural products. The construction of relatively small pithouses with few interior features also fits Gilman’s criteria for residential mobility, although at a much more reduced level than in the preceding Late Archaic period.
These early agricultural settlements fit well Flannery’s model for the fluid and flexibly structured settlements associated with a low degree of agricultural dependence. Circular and oval house forms predominated. Extramural storage pits were common, suggesting shared storage. Floor area averaged just over eight square meters (Ciolek-Torrello 1995; Ciolek-Torrello, Klucas, and Whittlesey 2000:table 1; Mabry 2000:51), a figure that is generally considered too small to have housed a nuclear household. Although the mean values vary considerably (Byrd 2000:80), a figure of about ten square meters per person, suggested by cross-cultural studies of per capita floor area (Cook and Heizer 1968; Naroll 1962), is the most widely recognized amount of space required for a nuclear household.
By the Vahki phase (AD 500–600), at the end of the Early Formative period, house size increased to about eleven square meters in the Phoenix Basin and doubled to an average of sixteen square meters in the contemporary Tortolita phase in the Tucson Basin. Rectangular houses became the dominant forms, and the presence of formal entryways and extensive use of plaster in wall and floor construction indicates greater investment in house construction. Superpositioning of houses was rare, however, suggesting that domestic space shifted over time and rules of land tenure were not well developed. Houses were often clustered with entryways oriented parallel to one another and had a common easterly orientation, which appears to reflect a culturally prescribed norm (Elson and Lindeman 1994; Gregory 1995; Mabry 2000; see Figure 8.4). The focal points of these early settlements appear to have been large communal houses, up to ninety square meters in area, which may have originated among the smaller, less formal communal houses in Cienega phase settlements. The communal houses of the Phoenix Basin contrasted with the circular and bean-shaped houses from other periods and regions of the Southwest (Ciolek-Torrello 1998a; Figure 8.5). In the Phoenix Basin, communal houses were represented by the large subrectangular pithouses originally identified at Snaketown (Gladwin et al. 1937). These large houses are common in Flannery’s (1972:336) sample for the less structured type of settlement and apparently organized a social unit on the scale of a band (Ciolek-Torrello 1995; Gregory 1995).
Figure 8.5. Hohokam communal houses from the Pioneer period (from Ciolek-Torrello 1998a:fig. 100)
Preclassic Period Developments
At the end of the Pioneer period, around AD 750, the Hohokam emerged as a distinctive cultural system. The undisputed high degree of maize use among the Hohokam is reflected in technological developments, although canal construction appears to have started much later than in the neighboring Tucson Basin. For example, the Price Road canal, the oldest known irrigation feature in the Phoenix Basin, was constructed between 130 BC and AD 275 (Henderson 1989:194–196). In a major treatise on irrigation in North America, Doolittle (1990) asserts that this canal represented a monumental engineering enterprise that was more technologically advanced than early Mesoamerican canals. He estimates it irrigated an area of up to 750 hectares. It also drew its water from the Salt River, one of the most flood-prone rivers of the Southwest (Welch 1994), not the smaller streams or upland runoff that fed early canal systems in the Tucson Basin (Doolittle 2002; Ezzo and Deaver 1998; Mabry and Holmlund 1998).
The Pioneer period canals of the Phoenix Basin were elaborated into the largest and most technologically sophisticated irrigation systems known in North America (Doolittle 1990:79–80). In their final form, these systems were a complex web of canals, several over sixteen kilometers in length (Howard and Huckleberry 1991), and covered an area of about 128,000 acres (Figure 8.6). Using Graybill’s (1989) paleo-streamflow data for the Salt River, Howard and Huckleberry (1991:table 5.1) estimate that at its peak the largest system could have watered more than 24,000 acres. Using a different method, Schroeder (1943:380–381) estimates that as much as 9,000 hectares could have been irrigated at any one time. Larger and more productive varieties of maize also became available, and cotton, beans, squash, agave, and a variety of other plants constituted a broad and diverse agricultural complex (Gasser and Kwiatkowski 1991).
Early reconstructions of Hohokam agricultural and social developments provide strong support for Flannery’s model. Based on surface information from early maps and aerial photography, a number of investigators (Haury 1976; Neitzel 1987; Upham and Rice 1980; Wilcox and Shenk 1977; Woodbury 1960) have argued that Hohokam canal systems were constructed gradually through a process of accretionary growth that culminated in the large, complex systems of the Classic period. The development of runoff-control technology in upland areas is also generally considered to be a late Sedentary or Classic period response to increasing population size or environmental degradation (Doyel and Elson 1985; Wood and McAllister 1984).
COURTYARDS TO COMPOUNDS
Many researchers viewed the transition from courtyard groups to compounds as another developmental process toward greater settlement size and social complexity (for detailed studies of architectural differences among households in other areas, see chapters in this volume by Beaule, González Fernández, and Snow). Courtyard groups are generally regarded as the basic structural elements of Hohokam settlements from at least the Colonial period until their transformation into the residential compounds of the Classic period (Sires 1987; see also Ciolek-Torrello 1988; Henderson 1987; Howard 1985; Wilcox, McGuire, and Sternberg 1981). Early Hohokam settlements, however, often lack evidence of courtyard groups (Klucas, Ciolek-Torrello, and Riggs 1998; Wilcox, McGuire, and Sternberg 1981); prior to the appearance of courtyard groups, Hohokam houses were distributed in loosely structured arrangements (Craig 2000:147).
Figure 8.6. Prehistoric canal systems in the Phoenix Basin (from Turney 1929)
According to Wilcox, McGuire, and Sternberg (1981), the typical Hohokam courtyard group consists of two or more contemporaneous pithouses oriented in a semicircular arrangement with their entryways opening onto a common living and work area. Discrete refuse middens, cemeteries, and food-processing areas are usually on the periphery, leaving the courtyard area itself devoid of features or trash. The courtyard group is generally interpreted as the residence of an extended household or lineage segment (Wilcox, McGuire, and Sternberg 1981). Habitation structures tend to be rectangular or subrectangular forms, whereas circular and oval forms are largely restricted to smaller special-function structures (see the chapter by Beaule, this volume, for a discussion of architectural shape and its correlation with household unit function in the Bolivian Andes).
The development of hereditary land ownership in these settlements is suggested by the persistence of individual courtyard areas over several generations. The sequence of superimposed houses commonly found in Hohokam courtyards is considered to represent several generations of a social unit occupying the same space and replacing houses as they deteriorated or as the needs of the social unit changed (Wilcox, McGuire, and Sternberg 1981). The association of discrete cemetery areas with these courtyards suggests the presence of corporate units larger than households that maintained descent and controlled land (Wilcox, McGuire, and Sternberg 1981; Wills 1992). These cemeteries enhanced the symbolic relationship among these corporate groups, the land upon which their residences were built, and associated fields (Varien 1999:210).
Small courtyard groups comprising two to three contemporary houses are common in smaller Hohokam settlements such as Scorpion Point, a Preclassic period Hohokam village in the Lower Verde Valley at the northeastern edge of the Phoenix Basin. This site may have contained as many as 300 pithouses, of which 60 were investigated. Only a small fraction of these houses, however, were occupied at any one time during the approximately 400-year span of time in which this village existed. The earliest occupation occurred in the late Pioneer–early Colonial period, when a scatter of single houses was built in three of the six residential loci identified at the site (Deaver 1998a). The main occupation occurred in the late Colonial period (Santa Cruz phase), when house density increased dramatically and two small ballcourts (see below) were constructed in the largest residential locus. Trash mounds and two distinct cremation areas were present in this locus, along with a cluster of hornos (pit ovens) in a pattern typical of the Hohokam residential units (Howard 2000). Several of the courtyard groups investigated exhibited superpositioning of houses and multiple instances of remodeling, indicative of extended and multigenerational use.
Courtyard groups at Scorpion Point contain three types of pithouses: large (mean ~24 m2, range 20–37 m2, n = 24) subrectangular to rectangular habitation structures with formal hearths and ramped entryways; smaller (mean ~13 m2, range 8–16 m2, n = 20) oval to rectangular habitation structures with formal hearths, usually with entry ramps; and smallest (mean ~10 m2, range 6–15 m2, n = 15) oval to rectangular structures lacking hearths and entry ramps but often containing storage pits (Ciolek-Torrello, Klucas, and Whittlesey 2000:83–85; Figures 8.7 and 8.8).
Figure 8.7. Preclassic period house types and sizes from the Lower Verde Valley; N = 46 (from Ciolek-Torrello, Klucas, and Whittlesey 2000:fig. 5)
During the Colonial period, house size and function were most diverse, and courtyards groups were composed of roughly equal frequencies of all types of structures. In the Gila Butte phase (early Colonial period), courtyard groups were often not apparent and houses were arranged in an eastward-facing orientation (Figure 8.9). Typical courtyards appeared in the Santa Cruz phase (late Colonial period). By the following Sacaton phase (Sedentary period), however, large habitations were much more frequent, comprising 73 percent of all the houses in courtyard groups, while frequencies of small habitations and storage structures were dramatically reduced (Ciolek-Torrello, Klucas, and Whittlesey 2000:table 2). The reduced variation in house size and function evident in the Sacaton phase indicates subtle changes in the composition of the social groups that inhabited them. The diversity of houses in the Colonial period is consistent with a previously defined extended household pattern in which family members resided in a large habitation and either a smaller habitation or storage structure (Ciolek-Torrello 1996; Huntington 1986). The concentration of domestic activities within large habitations in the Sedentary period may reflect a shift to nuclear household organization or a shift in the permanence or periodicity of settlement. Similar changes have been observed in house size and function between the Colonial and Early Classic period house clusters in the Picacho Mountain area at the southern edge of the Phoenix Basin but have not been replicated in the heart of the Phoenix Basin (Klucas, Ciolek-Torrello, and Riggs 1998).
Figure 8.8. Preclassic period house types at Scorpion Point Village in the Lower Verde Valley: (a) large habitation; (b) small storage structure; (c) small habitation (from Ciolek-Torrello, Klucas, and Whittlesey 2000:fig. 4)
Howard (1985), among others (see also Doelle, Huntington, and Wallace 1987; Sires 1987), suggests courtyard groups like those found in small settlements are the elemental, modular residential groups in larger settlements. For example, village segments comprise clusters of discrete courtyard groups placed back to back and side to side (Howard 1985; Wilcox, McGuire, and Sternberg 1981:figure 40; see Figure 8.9). More recently, Howard (2000:169) has identified village segments in the Phoenix area as large residential areas centered on large open plazas averaging about 100 meters in diameter. The only features found within these plazas are hornos, which Howard interprets as communal-use facilities. Clustered around the periphery of these plazas were several residential units, each comprising several contemporary courtyard groups. Discrete trash disposal areas also occurred between the residential units along the periphery of the plazas. Howard (2000:187) also notes the presence of discrete cemetery areas east of the residential units.
Figure 8.9. Preclassic period house clusters at Locus A, Scorpion Point Village (from Ciolek-Torrello, Klucas, and Whittlesey 2000:fig. 6)
Sires (1987) argues that the trend toward modular, rectangular house arrangements in some Preclassic period settlements culminates in the Classic period residential compound around AD 1250 (Figure 8.10). In many cases, compounds were built directly upon the remains of older courtyards and appear to represent a continuation of the same social unit (Doyel 1978; Sires 1987). Clark (1995) argues that small compounds are composed of similar sized and structured domestic groups as the earlier courtyards (see also Ciolek-Torrello 1994), whereas large compounds are composed of multiple sets of these domestic groups.
When viewed from the long-term perspective of the more than 3,000 years of agricultural production in the Sonoran desert, it is apparent that changes in the structure of households closely parallel increased sedentism and agricultural dependence. Changes in architecture, domestic arrangements, and settlement structure appear to occur in step with the development of intensive agricultural technologies and associated material culture, supporting the various models that propose a causal relationship among these processes.
Figure 8.10. Classic period compound and courtyard groups at AZ U:15:3, Escalante Ruins (after Doyel 1981:fig. 4)
Recent research, however, suggests a much more complicated picture. Contrary to the views held by many Hohokam archaeologists, the transition to a sedentary agricultural lifestyle was not a steady progression that inexorably culminated in the large irrigation communities of the Classic period. Rather, as I will document in the remainder of this chapter, the archaeological record reveals considerable instability throughout the lengthy period that agricultural settlements were found in the region. Such instability is reflected in landscape changes, canal construction, and shifting locations of canals and associated settlements.
IRRIGATION SYSTEM DYNAMICS
It has been a long-held view that Hohokam canal systems were constructed gradually through a process of accretionary growth (Haury 1976; Neitzel 1987; Upham and Rice 1980; Wilcox and Shenk 1977; Woodbury 1960). Recent research, however, reveals that individual canal alignments exhibit a high degree of instability with numerous instances of abandonment. Canals had a short use-life and were constantly rebuilt, often in response to catastrophic flooding (Ackerly 1989; Greenwald and Ciolek-Torrello 1988; Howard 1993). Henderson (1989:198–199) estimates that the average use-life of canals in one system was roughly thirty-five years. According to Doolittle (2002:408), “the entire Salt River Valley for the period extending from AD 0 to 1450 was a dynamic landscape of canals of various sizes and locations undergoing constant renovation and relocation. . . . [T]he valley can be best characterized in terms of the irrigated landscape as a constantly changing mosaic.”
Howard (1993) has developed a detailed chronological and hydraulic reconstruction of Canal System 2, the largest integrated system in the Phoenix Basin. Between AD 500 and 1450, roughly fifty main canals were constructed, fanning out in a series of parallel channels closely following topographic contours. At any one time, however, only about nine of these canals were active. Perhaps the most interesting aspect of his reconstruction is the intensity of growth exhibited during the Colonial period. Instead of gradual expansion over the Sedentary and Classic periods, the greatest growth of Canal System 2 occurred early in the sequence. Howard (1991; 1993:297) demonstrates Canal System 2 had reached its greatest spatial extent by about AD 1000, the end of the Colonial period, and was capable of irrigating more land than at any subsequent time.
Expansion of Hohokam Farming Systems in the Colonial Period
Between the late Pioneer period and the early Colonial period, Hohokam methods of floodwater farming and irrigation were spread to their geographical limits in the Phoenix Basin and surrounding areas. Small farming communities spread into the smaller tributary valleys along the Agua Fria, New River, Cave Creek, and Lower Verde River along the northern edge of the Phoenix Basin and the Queen Creek and Buttes Dam area along its southeastern edge. These areas lacked the large expanses of arable alluvium present along the Salt and Gila Rivers at the center of the Phoenix Basin, but the lower flows of the smaller drainages in these peripheral valleys may have been more easily managed. Furthermore, these valleys were located in more upland areas that provided a great variety of important wild plants that could be exploited and encouraged. Game was also abundant, and these valleys may have been important sources of protein for the large settlements along the Salt and Gila Rivers (Abbott 2000).
In the Phoenix Basin, many of the Pioneer and Colonial period canals were built near the active floodplain (Figure 8.11), where they could be taken easily out of the Salt River (Ackerly 1989:335; Doyel 1991:247; Greenwald and Ciolek-Torrello 1988; Henderson 1989; Howard 1988). Although these canals exhibited reasonably good engineering principles, their location along relatively steep topographic slopes near the active floodplain caused considerable problems. For example, the Dutch Canal Ruins consist of a series of early canals located in an area of Canal System 2 that witnessed major flooding during the historical period. The gradients of the canals in the Dutch Canal Ruins are estimated at about 2.35 meters per kilometer, more than twice as high as recommended for an earthen ditch (Haury 1976; see also Katzer 1989). The high water velocities resulting from such steep gradients are reflected in the coarse sediments contained within their fill, the high incidence of channel migration within individual canal alignments, and numerous examples of basal channel and bank erosion (Greenwald and Ciolek-Torrello 1988:84). In reference to Canal System 1, Ackerly (1989:342) also points to considerable evidence for flood-caused destruction of head gates, deposition of sediments, breaching of canal berms, and erosion of fields.
Erosion problems were undoubtedly aggravated by the streamflow regime and high-magnitude discharges that characterized the Colonial period. Graybill’s (Nials, Gregory, and Graybill 1989) paleoclimatic reconstruction, based on regional tree-ring records, reveals that this was a time of high streamflow variability with periods of extremely high discharge and severe droughts (Graybill 1989). Using Graybill’s data, Ackerly (1989) shows that this approximately 200-year time period was characterized by seven intervals of droughts and floods with a mean of only twenty-five years between intervals. By contrast, the following 150 years of the Sedentary period were characterized by only three droughts and floods with a mean of forty-one years between intervals. High discharges during the Colonial period were also of much higher magnitude than in the Sedentary period. The southern portion of Canal System 2, which was closest to the river and the most susceptible to damage by even low magnitude floods, was abandoned by the end of the Colonial period, and new canals and settlements were constructed on a higher contour away from the river. Located on the higher Pleistocene terraces overlooking the river, the new canals were not only better protected from flooding, but their new positions led to a reduction in gradient and the resulting erosion damage from high water velocity and discharge and an increase in the acreage that could be irrigated downslope of the canals (Howard 1991:4.16).
Figure 8.11. Pioneer and Colonial period canals and villages in the lower Salt River (adapted from Howard 1993)
Landscape Changes and Hohokam Retrenchment in the Sedentary Period
Waters and Ravesloot (2001) suggest a different cause for the changes that occurred in the Sedentary period based on a geomorphic reconstruction of the Gila River and its floodplain. They (2001:290) note that during the early part of the Preclassic period, the Gila River was characterized by a narrow channel with a broad floodplain. Flow was perennial and confined to the channel except during floods when the water would overtop its banks and inundate the adjacent lowlands, resulting in overbank deposition and vertical aggrading of the floodplain. Although early canal systems may have been damaged by flooding, these floods would also have renewed nutrients to the fields. This stable and productive landscape changed radically between AD 1020 and 1160 (Waters and Ravesloot 2001:291). The channel of the Gila River downcut and widened significantly, eroding the banks, destroying riparian forests, and generally disrupting nearly a millennium of floodplain stability. This wider channel had a braided streambed and the main flow channel shifted over the streambed with each large flow, forcing Hohokam farmers to rebuild diversion dams and to move their head gates repeatedly. Although this reconstruction was developed along the Gila River, Waters and Ravesloot (2001:294) believe similar events occurred along the Salt River because, as a major tributary of the Gila River, the channel of the Salt River would have responded to changes in the larger river.
The Cashion site, one of the largest Preclassic period settlements along the Salt River, appears to have been one of the casualties of these changes. Located at the downstream end of the Salt River near its confluence with the Gila River, the farmers at Cashion drew their water from Canal 12, a small canal downstream of Canal System 2. Antieau (1981) suggests that Cashion was abandoned at some time during the Sedentary period, when changes in upstream use left insufficient water to supply Canal 12. The population of Cashion apparently shifted a short distance downstream, establishing several smaller settlements and building new canals tapping the flow of the Agua Fria River, which empties into the Gila River a few miles downstream of the Salt River confluence.
The landscape changes proposed by Waters and Ravesloot (2001) provide an explanation for the abandonment of Cashion. The heads of most Sedentary period canals were concentrated at Pueblo Grande, a point immediately downstream of where the Salt River flows between two large bedrock outcrops, Papago and Tempe Buttes (Figure 8.12). The presence of these bedrock outcrops would have reduced channel downcutting and bank erosion in this location, providing greater stability for the canals that had their source close to these outcrops. Canals and associated fields were also shifted from the floodplain to the higher Pleistocene terraces during the Sedentary period. The locations of settlements followed suit. During the Colonial period, canals and settlements were scattered throughout the basin. By contrast, during the Sedentary period, most canals and settlements were located above 1,110 feet (ASML; see Figure 8.12). Canal systems, such as Canal 10 and the Cashion canal (Canal 12), which headed further downstream and irrigated low-lying floodplain areas, were abandoned by the end of the Sedentary period.
The displacement of canals and population to higher ground combined with less variability in streamflow to create optimal conditions for irrigation in the Phoenix Basin during the Sedentary period. According to Howard (1993:300–301), the Sedentary period is distinguished by much less rebuilding. He (Howard 1993:table 5) estimates that only about half as much earth (400,000 m3) was moved to build canals compared to the Colonial period (800,000 m3). Despite half the construction effort, irrigated acreage was only about 20 percent lower than the Colonial period. Although Ackerly (1989:340) sees little improvement in canal engineering over time, improved location of canals and head gates and more stable climatic conditions made Canal System 2 much less prone to failure during the Sedentary period.
Renewed Expansion of Irrigation Systems in the Classic Period
The Classic period is characterized by a new wave of canal abandonments, more major settlement shifts, and the construction of new canals on still higher ground in the northern portion of Canal System 2. At the same time, canals and settlements in the central and distal portions of the system were abandoned (Figure 8.13). The amount of canal construction in Canal System 2 increased dramatically, (850,000 m3 of earth moved; Howard 1993), exceeding slightly the Colonial period levels, but irrigated acreage at any one time was 10 percent lower. Canals were extended to the northernmost limit of the system and some of the farmlands that had fallen into disuse in the floodplain were reclaimed (Greenwald and Zyniecki 1996; Howard 1993). Canal System 1, upstream of Papago Buttes and south of the Salt River, was also greatly expanded and an entirely new system, the Lehi Canal System (Canal 11), was constructed further upstream on the south side of the Salt River in the area of the modern city of Mesa (Abbott 2000:33; see Figure 8.6). Canal Systems 13 and 14 were also greatly expanded in the vicinity of modern Scottsdale well upstream of Canal Systems 1 and 2. The expansion of these systems was monumental in scale and required considerable effort and coordination (Abbott 2000:28).
Figure 8.12. Sedentary period canals and villages in the lower Salt River (adapted from Howard 1993)
Figure 8.13. Classic period canals and villages in the lower Salt River (adapted from Howard 1993)
While irrigation networks were greatly expanded or newly created in the center of the Phoenix Basin, the small river valleys along the northern edge of the basin and the upper Queen Creek and Buttes Dam areas southeast of the basin were abandoned by the beginning of the Classic period (Wilcox, Robertson, and Wood 2001). New Hohokam canals were constructed along the Verde River a short distance above its confluence with the Salt River later in the Classic period. The upper reaches of the Lower Verde Valley, however, from Sycamore Creek to Horseshoe Dam remained abandoned by the Hohokam. From as early as the late Pioneer period, this area had sustained numerous large Hohokam settlements, including Azatlan, considered by some to be the largest known Preclassic period Hohokam town (Ciolek-Torrello 1998b).
Along the Salt River, streamflow variability and high flows increased during the Classic period, although they were not nearly as high as during the Colonial period. Unreliable water flows and shortages plagued Hohokam farmers throughout the region during many years (Ackerly 1982; Howard 1991). The residents of Canal System 2, the largest canal system and the farthest downstream, were particularly hard hit by these problems (Abbott 2000:207). Dietary problems developed as a result of an overdependence on cultigens and a lack of wild plant foods and animal protein (Abbott 2000:197).
Despite these problems, conditions may have improved along the major desert waterways to facilitate the tremendous growth that characterized the Classic period. Waters and Ravesloot (2001:292) observe further landscape changes along the Gila River during the Classic period. The river channel narrowed again, the riparian zone was likely restored, and the floodplain stabilized. If such conditions can be extrapolated to the Salt River, then floodplain canals and head gates downstream of Papago Buttes were probably viable once again. The concentration of Classic period canals on higher ground and the proximity of head gates to Papago Buttes, however, made them less vulnerable to flooding and low water flows and probably resulted in more dependable crop yields than during the Colonial period.
Implications of Irrigation System Changes
Recent research counters the widely accepted notion that the Hohokam quickly mastered the rivers of the low deserts, gradually expanding their canals into one of the largest systems known in prehistory. Instead, the archaeological, paleoclimatic, and geomorphic records reveal that the Hohokam were often at the mercy of highly variable streamflow conditions and changing landscapes and were forced to continuously rebuild and move their canal systems to adjust to these changing conditions. Canal System 2 reached its greatest capacity during the Colonial period, but erosional problems led to a pattern of repeated rebuilding and abandonment because of the location of many canals in flood prone areas, their steep gradients, and high streamflow variability at this time. Much of the floodplain zone was abandoned during the Sedentary period, as canals and settlements were shifted to higher ground. During the Classic period, existing canals were expanded, while new ones were once again constructed in the floodplain. Canal systems were also greatly expanded and new ones constructed upstream of Papago Buttes, perhaps leading to water shortages along Canal System 2, especially at the “tail ends” of the canals, during low river flows (Abbott 2000:33, 46, 198, 207; Ackerly 1982). At the same time, large tracts of land that had been farmed for hundreds of years in the smaller valleys on the northern and southeastern edges of the Phoenix Basin were abandoned (Ciolek-Torrello 1998b).
This evidence has at least two important implications: first, agricultural intensification did not occur through the gradual expansion of individual canal systems; and second, the constant rebuilding and relocation of canals resulted in a pattern of shifting land use during the Colonial period followed by major settlement reorganization during the Classic period. The reorganization involved the expansion of canal systems and associated settlements in the central Phoenix Basin and abandonment of outlying areas. Although Canal System 2 was at its greatest capacity during the Colonial period, destruction of canals and fields undoubtedly resulted in repeated crop failure. Expansion of farming and settlement into the smaller outlying valleys during this time may represent an attempt toward diversification of the greater Hohokam agricultural system as a hedge against failure of one or more of its parts. Such diversification was apparently less necessary during the Sedentary period, when canals were concentrated in optimal locations and climatic conditions were generally more stable. At this later time, however, flood-prone areas of Canal System 2 were abandoned along with many of the smaller systems in outlying valleys. During the Classic period, Canal System 2 was rebuilt once again and expanded greatly along with the other major canal systems along the Salt River and the Gila River. By contrast, the northern edge of the basin became a no-man’s-land between the Hohokam and new immigrants into the uplands surrounding the basin (Ciolek-Torrello 1998b; Doyel and Crary 1995).
Clearly, the early development of large canal systems did not provide the rich and stable resource base that prehistorians have traditionally believed sustained the Hohokam culture. Rather, Hohokam farmers continually faced numerous challenges from flooding, changing landscapes, and unreliable water flows. They were forced to build and repair their canals constantly and to shift their locations along with their settlements to entirely new areas.
RESPONSES TO CHANGING SYSTEMS OF PRODUCTION
The changes in Hohokam domestic organization and settlement structure, discussed above, must be viewed from the perspective of changing landscapes, fluctuations in the productivity, and shifting locations of irrigation systems rather than an inexorable evolutionary process toward greater sedentism and complexity. Hohokam settlement reached its greatest spatial extent during the Colonial period, expanding into the smaller river valleys surrounding the Phoenix Basin, north to the Colorado Plateau, east to the mountains of central Arizona, south into the Tucson Basin and southeastern Arizona, and west to the Gila Bend area. During this time, villages and hamlets composed of courtyard groups identical to those in Phoenix Basin settlements appeared in many of these areas along with small-scale irrigation systems (Ciolek-Torrello 1998b; Doyel and Elson 1985; Gregory 1995). These settlements also shared a similar cultural pattern evidenced by common use of Hohokam Buff ware ceramics and their local varieties, similar architectural styles, and similar ritual paraphernalia and mortuary patterns (Wilcox 1979).
Public Architecture and Regional Integration
Ballcourts, platform mounds, and the activities associated with these structures were key elements of Hohokam settlements and communities. Ballcourts, low earthen embankments surrounding an oval depression, were centrally located within many of the larger Preclassic period settlements of central and southern Arizona and are believed to have been used for playing a game similar to the Mesoamerican ballgame (Wilcox 1991). Unrestricted access to these ballcourts indicates that the games held within them were viewed by the members of the surrounding community. A continuous spatial distribution of ball-courts linked Hohokam settlements together into a regional system (Wilcox and Sternberg 1983). Archaeologists generally infer that this was a network of geographically separate and, most likely, unrelated populations dependent on one another through their shared participation in ritual and social events, as well as the exchange of a great variety of goods and services (Abbott 2000:28, 31).
At some time during the Sedentary period, a series of events began that transformed Hohokam society and the manner in which it was integrated. According to Abbott (2000:192), the distribution and composition of settlements were completely rearranged across the landscape by the end of the period. The social, ritual, and economic ties that integrated much of central Arizona for centuries had fallen into disarray as the vast network of ballcourts was abandoned. By the Soho phase, the early part of the Classic period, a much smaller and more nucleated Hohokam regional system appeared with an entirely different look. Further dramatic changes occurred to the Hohokam system during the Civano phase, the latter part of the Classic period, before Hohokam society ceased to exist as a recognizable entity at the end of the fourteenth century.
By AD 1200, use of ballcourts had virtually ceased. In their place appeared large earthen and rock mounds that initially served as platforms for public ceremonies and eventually became residences of presumably elite households. Abbott (2000:32) uses the distributions of ballcourts and platform mounds to document the extent of regional interaction during the Preclassic and Classic periods. In contrast to ballcourts, platform mounds were restricted to the center of the Phoenix Basin and a few of the larger irrigated valleys to the south and east. Notably, the lower Verde Valley and upper Queen Creek areas at the northeastern and southeastern edges of the Phoenix Basin had contained a concentration of ballcourts in the Preclassic period, but no platform mounds in the Classic period. Nor were any found in the Agua Fria, New River, and Cave Creek valleys along the northern edge of the basin, reflecting a substantial reduction in the Hohokam interaction sphere.
Furthermore, ballcourts and platform mounds functioned in very different ways. Ballcourts are believed to have integrated diverse populations from distant locations, facilitating the exchange of commodities from different environments. By contrast, platform mounds were symbolic edifices of inward-focused, boundary-maintaining communities. They were part of a highly structured settlement system closely linked to the Classic period canal networks. Spaced at regular intervals along canals, they were ideally positioned to facilitate communication and decision making with respect to hydraulic management (Gregory 1987). Platform mounds also may have been the focus of elaborate ritual systems that served as a spiritual means of coping with environmental and economic uncertainty (Bostwick 1992) or as communal storage facilities (Lindauer 1992) that could have been used to ameliorate spatial and temporal variability in food production (Altschul and Van West 1992). Another theory suggests that platform mounds were prominent symbols reflecting the establishment of a new system of land tenure in which canal systems or segments of the larger systems became closely associated with individual descent groups. Abbott (2000:204–205) proposes that platform mounds were constructed by a descent group for the worship of its ancestors, who had constructed the fields and the canals that supplied water to them (see also Ciolek-Torrello and Whittlesey 1996). Each platform mound symbolized the link between the canals and cultivated fields within its view and a closed group of direct descendants who controlled access to these canals and fields.
Reorganization of Hohokam Society in the Classic Period
As the Preclassic period Hohokam interaction sphere collapsed and ball-courts were replaced by platform mounds, population swelled along the Salt and Gila Rivers and the residents of small sites aggregated into larger villages. At Pueblo Grande, one of the largest Classic period settlements in the Phoenix Basin, residence group size increased dramatically, and several new residence groups were established, which together nearly doubled the size of the population (Abbott 2000:32). A significant immigration of population is suggested by the rapid scale of growth. Abbott (2000:195) suggests that the massive growth of Los Muertos in Canal System 1 may have been related to the immigration of Hohokam settlers, who had abandoned large Preclassic period communities such as Snaketown along the Gila River. He also suggests that the establishment of the Lehi Canal System (Canal System 11) may have been the product of a wholesale resettlement of groups from other areas of the Phoenix Basin. Emigrants from the smaller valleys along the northern edge of the Phoenix Basin may have been the primary source of these settlers (Ciolek-Torrello 1998b). Many large new Hohokam settlements also were built along the Gila River and in the Picacho Mountains along the southern edge of the Phoenix Basin in the Classic period (Ciolek-Torrello 1988; Doyel 1981). It is more likely that the many Hohokam settlements that were abandoned in the Lower Verde and other smaller valleys along the northern edge of the Phoenix Basin were the primary source of population swelling the villages along the Salt River.
Changing Household Structure and Residential Patterns
The reorganization of Hohokam society during the Classic period as evidenced by population resettlement, aggregation, changing public architecture, and possibly changing systems of land tenure is reflected in changing domestic structure. Preclassic period courtyard groups composed of clusters of pithouses were gradually transformed into groups of adobe-walled surface structures contained within rectangular compounds partitioned from other residential groups by walls that may have stood two or more meters high. Most prehistorians infer that these walled compounds represent an unaltered social pattern from Preclassic period courtyards (Clark 1995; Doyel 1991). In a view clearly reminiscent of Flannery, Abbott (2000:175), however, suggests that compounds represent a “departure from an open pattern of social intercourse unencumbered by stringent social distinctions to one where such differences were clearly apparent.” He links this change to the increased importance of agricultural land and its produce, emphasizing that the “loosely aligned and less-organized families” of Preclassic period settlements “coalesced into socially exclusive and closely cooperating units that effectively protected their agricultural yields and water rights.”
Thus, when viewed from the perspective of concurrent changes in irrigation systems and the reorganization of Hohokam society, the shift from courtyards to compounds probably reflects a fundamental change in lifeways rather than merely a modification of preexisting residential patterns. The residents of Preclassic Hohokam courtyard groups exploited an ever-changing landscape with shifting irrigations systems and residential locations. These courtyard groups represented much more loosely structured and more mobile residential groups than previously believed. Although the courtyard group, with its evidence for multigenerational use, has usually been regarded as the hallmark of deep sedentism in the Southwest (B. Nelson 1990), better chronological controls of the occupational histories of individual courtyards reveal much greater instability and a high degree of mobility (Ciolek-Torrello, Klucas, and Whittlesey 2000; Deaver 1998b; Vanderpot, Klucas, and Ciolek-Torrello 1999). Rather than continued growth and expansion over long periods of time, Hohokam courtyards often exhibit patterns of intermittent occupation, unexpected decline, and shifting location. McGuire and Schiffer (1983:286) characterize the early Hohokam residential pattern as one of mobile households occupying permanent settlements. According to this pattern, settlements may have been permanently occupied, but households moved frequently within and between them (see Wilk and Rathje 1982:633–637).
For example, Henderson and Rice (1987:61) describe changing residential patterns in one residential unit at La Ciudad, one of the largest Preclassic period settlements in Canal System 2. Between AD 750 and 875 only one or two houses were occupied at a time, with a new house built only after the previous one had been abandoned. This was followed by a thirty-year period in which five houses were occupied simultaneously. The area was then abandoned for the next thirty years, but reoccupied by groups of four to eight houses until AD 1000. The occupation then reverted to a solitary house.
This pattern of shifting occupation is better illustrated at Scorpion Point Village. Between 28 and 50 percent of the dated houses at the site were occupied during the approximately 100-year span of the Santa Cruz phase (depending upon how many of the general Colonial period houses can be attributed to the Santa Cruz phase) (Ciolek-Torrello, Klucas, and Whittlesey 2000:table 2). Craig’s (2000) method for estimating the periodic occupation—average use-life of 25 years per pithouse, a 100-year span for the phase, and total of 84 to 150 houses (28–50 percent of 300)—provides a rough estimate of between twenty-one and thirty-eight houses that may have been occupied at any one time during Santa Cruz phase. Analysis of domestic structures at the site (Ciolek-Torrello, Klucas, and Whittlesey 2000; Klucas, Ciolek-Torrello, and Riggs 1998) indicates that the Colonial period population was distributed among extended households occupying at least two houses—a large habitation and either a small habitation or storage structure (see also Ciolek-Torrello and Greenwald 1988; Huntington 1986:44; Sires 1984). Applying this figure to our calculations suggests that between ten and nineteen extended households resided at Scorpion Point at any one time during the Santa Cruz phase.
Both ballcourts were abandoned sometime during the Santa Cruz phase, as indicated by the intrusion of Santa Cruz phase cremations. By the end of the phase, the entire site appears to have been abandoned. The settlement was reoccupied following a fifty-year hiatus between the Santa Cruz and Sacaton phases (Sedentary period) (Deaver 1998b). Sacaton phase settlement, however, was structured in an entirely different manner. Approximately 25 percent of the houses at Scorpion Point date to the Sacaton phase occupation, which probably lasted only fifty to seventy-five years. Using the same calculations as for the Santa Cruz phase, approximately twelve houses (300 houses × 0.25; 75-year span of occupation; 25-year house use-life) were occupied at any one point of time during the Sacaton phase. In contrast to the Santa Cruz phase, the Sacaton phase population was distributed in nuclear households residing primarily in large habitation structures, which comprised almost 75 percent of the houses during this phase (Ciolek-Torrello, Klucas, and Whittlesey 2000:table 2). Thus, roughly nine households were occupied at any one point in time during the Sacaton phase. These data indicate that the number of households at the site was either halved or approximately equal, again depending on whether we can use the larger part of the range for the Santa Cruz phase occupation. Considering the much smaller size of Sacaton phase households, however, the population size was reduced dramatically.
Despite the difference in size of households, houses continued to be arranged in courtyard groups at Scorpion Point during the Sacaton phase. In fact, courtyard groups with at least three houses are slightly more common and these groups are more tightly clustered than their Santa Cruz phase counterparts (Klucas et al. 1998:513). Chronological data, however, indicate that the houses in individual Sacaton phase courtyards at Scorpion Point were not used at the same time. The contrast between Santa Cruz and Sacaton phase courtyard groups at Scorpion Point is striking. The earlier houses exhibit considerable superpositioning and remodeling, indicating multigenerational use of the same domestic area. By contrast, Sacaton phase houses lack evidence of superpositioning and remodeling (see Figure 8.9). If the Sacaton phase houses were occupied contemporaneously, it would suggest the presence of a more aggregated and structured village, albeit short-lived. The evidence for noncontemporaneity of Sacaton phase courtyards at Scorpion Point, however, suggests a short-term and more intermittent occupation (Deaver 1998b).
DISCUSSION
Howard and Wilcox (1988) have argued that abandonment or changes in settlement locations were closely related to shifts in the location of canals in the Phoenix Basin. As canals within Canal System 2 shifted northward onto the higher terraces during the Colonial and Classic periods, a progression of site types was constructed. Temporary field houses were initially constructed while Hohokam farmers tended their new fields. These were replaced by small settlements occupied by families that budded from the original villages. As the network continued to expand northward, a large village was eventually constructed in the new location.
This type of progression, however, does not account for the type of fluctuations evident at La Ciudad and Scorpion Point. It is more likely that the changing residential patterns evident at these sites reflect the relocation of individual households in response to economic fluctuations. Floods that damaged fields and irrigation systems or droughts that reduced water flows are the most likely culprits for these fluctuations. Households may have moved to other portions of a canal system, to a different canal system, or even to different river valleys that were not affected in the same way, only to return when conditions improved.
The construction of pithouses by the Hohokam must also be viewed from the perspective of a mobile residence pattern. Although Hohokam pithouses were not constructed in pits as deep as their Anasazi and Mogollon counterparts, materials and construction techniques were similar. Use-lives of Hohokam pit-houses are also believed to be of a similar span as Anasazi and Mogollon pit-houses. Thus, like most cultural groups known in the ethnographic and archaeological record, the Hohokam built pithouses during the Preclassic period because they probably remained relatively mobile agriculturalists.
This relationship between architecture and mobility has not gone unrecognized by Hohokam archaeologists. Most prehistorians regard the Polvoron phase as a terminal portion of the Classic period, when Hohokam society collapsed in the late 1300s with the destruction of the irrigation systems that sustained it. The Hohokam reverted to an earlier, less structured society that occupied pithouses once again (Crown and Sires 1984).
Henderson and Hackbarth (2000) present an alternative view of the Polvoron phase. They argue that the Civano and Polvoron phase manifestations are contemporary and represent the normal range of variability expected in complex late Classic period communities. Civano phase manifestations are represented by compounds in large, permanent, aggregated riverine villages, whereas Polvoron phase manifestations occur in small, seasonal farmsteads in outlying areas. Using the ethnographic analogy of Tohono O’odham people, who seasonally or temporarily resided in the riverine villages of their Akimel O’odham neighbors, Henderson and Hackbarth (2000) suggest that the occurrence of Polvoron manifestations in the large riverine villages represents the presence of transient populations that moved between the outlying farmsteads and riverine villages. The less permanent status of these transient populations is reflected in the construction of less substantial jacal pithouses on the periphery of the large villages as well as in outlying areas. Henderson and Hackbarth (2000) explicitly state that in constructing pithouses labor investment was minimized in the face of an uncertain residential status. It is highly likely that this type of residential pattern had a long history that may have extended as far back as the beginnings of Hohokam culture.
The Preclassic period ballcourt system also should be viewed from the perspective of a more mobile Hohokam society during the Preclassic period. The network of Preclassic period ballcourts was associated with the movement of people and goods between riverine villages and outlying regions. Ballcourts were the focal points of communities where people from different settlements and regions probably came together to conduct games or rituals; to exchange news, goods, and services within a reciprocal framework; and even, perhaps, to arrange marriages. It would not be a significant stretch of the imagination to propose that the relationships established at the events conducted at ballcourts facilitated residential shifts among settlements within the same canal network, among different networks, and even among different river valleys. People did not shift residence at random or into unknown areas. Rather, they undoubtedly moved to areas where they had established relationships with existing residents. Interestingly, the bulk of the immigrants to Pueblo Grande during the Classic period joined existing residence groups rather than establishing new ones (Abbott 2000:198), suggesting that they knew and were welcomed by the existing residents.
The replacement of the widespread ballcourt network by the smaller network of more inwardly focused platform mounds during the Classic period marks a radical departure from the more open, wide-reaching, and fluid Preclassic period system. That this change was associated with the expansion and improvement of irrigation systems within the Phoenix Basin, abandonment of outlying regions, and the replacement of pithouses by aboveground, adobe-walled houses enclosed by compound walls tends to support the models proposed by Flannery and Gilman regarding the relationship among agricultural intensification, more rigid systems of land tenure, increased sedentism, and architectural change.
Expected changes in domestic structure, however, are not as readily apparent. Domestic organization clearly responded to increasing agricultural dependency and sedentism between the late Archaic and Preclassic periods. Larger, more substantial houses arranged into courtyard groups replaced loose clusters of smaller, ephemeral pit structures. Hohokam courtyard groups and larger residential units, however, exhibit much greater organizational structure and residential stability than contemporary Anasazi and Mogollon settlements, despite similarities in architectural forms. I argue, however, that the continued use of pit-house architecture combined with the evidence for shifting settlement and irrigation systems during the Preclassic period indicates greater residential mobility, less agricultural dependency, and the existence of a system of land tenure based more on usufruct than during the Classic period. It was only during the Classic period that the Hohokam could be considered fully sedentary and agriculturally dependent. Whether they were hemmed in by enemies or tethered to an immense and highly productive irrigation system cannot be determined as yet, but their degree of dependence on agricultural products may have reached the point of dietary stress and disease as suggested by the data from Pueblo Grande (Abbott 2000).
SUMMARY AND CONCLUSIONS
In this chapter I have attempted to evaluate the relationship among domestic organization, sedentism, and agricultural dependency in the Phoenix Basin using models developed through cross-cultural studies by Flannery, Gilman, and others. My focus has been on how changes in household architecture and organization were related to changes in settlement and farming systems. Specifically, I have tried to demonstrate that a shift occurred from extended households in less structured and more mobile communities in the Late Archaic and early Preclassic periods to nuclear families in more structured and sedentary communities in the later Preclassic and Classic periods. These changes, in turn, were related to the development of more structured rules of land tenure associated with agricultural intensification involving increasing dependence of large-scale irrigation systems. Close parallels among changes in irrigation systems, settlement, and domestic organization tend to support Flannery’s model overall. A true test of this model would seek evidence for greater dependability in food production during the Sedentary and Classic periods. For example, maize frequencies or ubiquity values should be much higher or more consistent. Sedentary and Classic period courtyard groups and compounds also should exhibit greater residential stability; they should display extended occupations with gradual and continued growth rather than the intermittent occupations evident in many Colonial period settlements.
The simple evolutionary schemes for increasing sedentism, changes in domestic organization, and development of irrigation systems used by many Hohokam archaeologists, however, are not supported by the evidence presented here. I have tried to demonstrate that Preclassic period Hohokam culture was not the icon of “deep sedentism” that most Southwestern archaeologists have considered. While many of the large Preclassic period villages and towns were occupied over several centuries of time, individual households most likely shifted their residences on a regular, if not seasonal, basis, and entire settlements or loci may have been abandoned periodically. This pattern of shifting residence was a response to environmental fluctuations, unreliability of floodplain farming, and a broad landscape that offered a variety of other farming opportunities.
The larger and more complexly organized villages and communities of the Classic period were not merely the product of steady and inexorable growth and development from Preclassic period roots, but rather a response to a changing physical and social landscape, as well as the movement of canals to locations that provided more stable and reliable agricultural production.
Although the focus of this chapter has been on the relationship among environment, farming technologies, and domestic organization in the Sonoran desert, it should be clear that Hohokam prehistory provides a rich source of data to evaluate theoretical models concerning these relationships. This chapter is intended as a case study but has important implications for cross-cultural studies.