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 portion of the Sonoran Desert inhabited by the Hohokam, the Tucson Basin is an ideal location for investigating a diversified but integrated system of Hohokam agricultural production. Regional demography, settlement, social organization, and exchange were shaped by specific resource needs, technological capabilities, and environmental potential for production in an agricultural economy. The productive landscape of the Marana Community and of the larger study area can be examined in detail by reference to environmental variables, settlement pattern, and agricultural features, and by comparison with the practices of traditional farmers in historic times.

A number of excavation and survey reports document localized occurrences of Hohokam agricultural features and complexes. At a higher synthetic level, extensive canal networks have been analyzed by integrating historic records, aerial photography, and settlement data (for example, Haury 1976; Masse 1981, 1987, 1991; Nicholas and Neitzel 1984; Nials and, others 1989; Ackerly and others 1987). Combinations of canal irrigation and runoff devices also have been described for a few substantial study areas (Gumerman and Johnson 1971; Crown 1987; Doyel 1984). However, in none of these cases has a systematic regional sample of non-canal features been available. Archaeologists attempting to deal with agricultural patterns at this scale (Masse 1979, 1991; Doyel 1984) have had to interpolate from geographically restricted and noncomparable survey findings.

The most important qualification of the northern Tucson Basin for a regional perspective is the presence of undisturbed land in all environmental zones. Granitic mountain ranges on the east reach approximately 1325 m (4300 feet). Tb the west, rough basalt peaks form a lower chain mostly below 860 m (2700 feet). Elevational diversity is repeated on opposing sides of the basin. Major and minor Santa Cruz River tributaries carry runoff from both orographic rainfall in the mountains and storm-fed watersheds of bajadas. The hydrologic system includes surface flow and a second component of less rapidly mobile groundwater. Distributions of agricultural remains across the varied zones of the Marana Community offer a uniquely comprehensive testament to past productive relationships between noncore Hohokam and their desert basin environments.


Ethnographic practices of traditional farmers in environments similar to the study area furnish insights into natural and cultural factors influencing agricultural production. Piman analogies have been applied to the interpretation of Tucson archaeological patterns, particularly with reference to floodwater farming or the diversion of short-term flows in ephemeral drainages (Wilson 1985; Field 1985, Chapter 5 in this volume; Waters and Field 1986). However, historic Piman floodwater farming (ak chin farming) is a method with few remaining practitioners on isolated remnants of previously productive acreage (Reichhardt and Nabhan 1982; Nabhan 1986a, 1986b). Although well documented, the ethnographic instances do not duplicate prehistoric situations that involved higher population densities and, necessarily, more intensive land use. Other prehistoric techniques such as the northern basin forms of riverine irrigation and cultivation in rockpile fields lack analogs altogether.

Traditional farmers from two cultural and agricultural backgrounds in the Sonoran Desert were consulted in order to include multiple approaches, to broaden the range of associated population and land use intensity, and to explore agricultural alternatives and decision making. Visits to the study area by a traditional Tohono O'odham (Papago) farmer from an area southwest of Tucson and a farmer from Cucurpe in northern Sonora, Mexico, were arranged with the aid of Gary Nabhan of the Phoenix Botanical Garden, who participated in all

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

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

phases of consultation. Both individuals engage in variants of floodwater farming, and the Cucurpe farmer irrigates with gravity-fed canals from a small Sonoran river. Evaluations and comments by these consultants are cited in relevant sections of this chapter.


Tucson Basin agriculture must be understood in the context of local risks and opportunities. The potential for riverine canals is lower than along the Salt and Gila rivers near Phoenix, based on volume and duration of flow in the Santa Cruz River, terrace morphology, and extent of topographically irrigable basin floor. Alternative prehistoric technologies for securing agricultural water were concomitantly more prominent. Precipitation measures for Hohokam subareas reveal variation in the

Table 4.2. July, August, and September Precipitation for
Stations In the Tucson Basin and Other Hohokam Subareas

Source: Sellers and others 1985

abundance and predictability of local rainfall corresponding to differential risks and returns for fields more directly dependent on this source for water than irrigated land (Tables 4.1, 4.2; Fig. 4.1). Substantial investment in such fields in the Tucson Basin may be explained as much by comparatively favorable precipitation regimes as by fewer possibilities for riverine canals.

Mean annual precipitation for the Marana study area at 280 mm (11.2 inches) is among the highest for basin interiors in all of Hohokam territory. For agriculturalists in arid environments, annual variability may be as critical as average amounts. Table 4.1 confirms that higher annual Tucson precipitation corresponds with lower variability, as measured by departures from the long-term mean. Likewise, precipitation minima in dry years are higher in the northern Tucson Basin than at other stations. Tucson riverine irrigators as well as farmers using channelized floodwater and overland flow depended on relatively abundant local precipitation. Because upland snow melt is too limited to generate high flows in the Santa Cruz and its tributaries during the spring and because summer precipitation predominates in southcentral Arizona, rainfall in this season is the major factor in total agricultural production. As in the case of annual amounts, summer precipitation for the Tucson Basin is more favorable than for much of the territory inhabited by the Hohokam (Table 4.2).


Mountain Slopes

At higher elevations, orographic rainfall delivers a more abundant and predictable water supply for crops dependent on direct precipitation than at lower elevations. The Tucson Mountains on the west edge of the Tucson Basin (Zone 6 of the Marana Community) exert

Figure 4.1. Locations of weather stations in Tables 4.1 and 4.2.

lesser influence on precipitation patterns than the more massive ranges on the east (Figs. 3.8, 3.9). However, the benefits of orographic rainfall are balanced against two drawbacks of mountain slope cultivation. Parcels of suitable soil depth and flatness are restricted. At altitudes above 925 m (3000 feet), frost hazards increase, although lower mountain flanks may be warmer than valley floors exposed to cold air drainage.

In the Tucson Basin, temperature inversion has significant elevational consequences for early or late season planting. The impact of cold air flowing down slope is illustrated by five years of weather records for a station on the shoulder of a hill in the Tucson Mountains and for a second station on the basin floor below, near the Santa Cruz River (Hastings and Turner 1965: 17). The stations are separated by a horizontal distance of only 0.8 km (0.5 mile) and by a vertical distance of only 100 m (330 feet). A difference of 10.10 C (200 F) was recorded on some nights. Over the five-year period there was a total of only 38 freezing nights on the hill in contrast to 263 nights on the floodplain. The frost-danger period between first and last freezes for a winter averaged 36 days on the hill and 157 days below. Low temperatures were of shorter duration on the hill as well. These elevational contrasts would be critical for spring crops planted sufficiently early to benefit from winter rains.

Two mountain slope situations appear most advantageous for agricultural pursuits. One occurs on the flat land of ridge tops or plateaus at higher elevations where more abundant rainfall is available. Alternately, gentle slopes at intermediate elevations with less orographic rainfall have less frost, better soil accumulations, and offer opportunities for water catchment and concentration through such features as terraces and checkdams.

The Tortolita mountain slopes have been covered less fully in survey than other zones. The best coverage is in the northeast sector of the study area (Fig. 3.2). Here, a number of habitation sites are located away from canyons on flatter land at elevations as high as 1128 m (3700 feet). It seems probable that orographic rainfall and relatively small catchments supplied water for agriculture. Survey transects (Hewitt and Stephen 1981) located agricultural terraces near sites at still higher elevations on the eastern Tortolita slopes facing into the adjoining basin.

The Tucson Mountains are much less massive than the Tortolitas and were fully surveyed. The relative sparsity of sites in the Tucson Mountains (Zone 6) may be related to preferred habitation on the immediately adjacent Santa Cruz floodplain (Zone 5), but potential use of western mountain land also seems limited. Terraced slope sites in southern Arizona of the type found in the Tucson Mountains correlate almost exclusively with the kind of dark volcanic substrates occurring in this range; locations of such sites in the Marana Community additionally coincide with large populations on the adjoining floodplain.

Four clusters of stone terraces or trincheras features were constructed in the Tucson Mountains within the study area. The two largest coincide with dense occupations around the northern end of the mountain chain. Many of these terraces appear to have been agricultural in function, and corn pollen has been recovered from several (S. Fish, R Fish, and Downum 1984). The agricultural benefits of terraces on mountain slopes include water concentrated from slope catchments above and between terraces, protection from freezing winter temperatures threatening lower elevations through inversion, and relief from high summer temperatures by preferential use of north and east exposures. Trincheras cultivation probably ranged from kitchen gardens around the houses built on some terraces to clusters of features of solely agricultural function (Downum and others 1985; Downum 1991) as described for a large site in the Robles Community to the northeast (Fig. 1.8).

Canyon Bottoms

Tortolita canyon bottoms represent a topographic class related to mountain slopes, but with distinctive agricultural potential. Canyons carry streams with large uphill watersheds. Effluent water continues to be discharged into these stream beds for extended periods and supports perennial flow in some cases. The quantity of effluent stream flow may be reduced in the summer, however (Chapter 5). Flat bottomland with adequate soil depth is often quite narrow. Since high-energy floods occur, smaller tributaries to main canyon drainages may have been easier to divert in some cases. Cold air, which flows down canyons, represents one environmental drawback for early crops. Wider canyon bottomland in the Tortolitas is associated with a number of settlements in uppermost Zone 4 of the Marana Community in contrast to the Tucson Mountains (Zone 6), where canyons are smaller and drain restricted slope areas.

Tortolita canyon sites tend to be located along the floodplain edges, no doubt above the contour of most floods. The only recorded habitation sites are small. Farming in the canyon bottoms may often have been carried out on a daily basis by inhabitants of settlements on the adjacent upper bajada. Deeper soils are typically associated with shallow, meandering channel segments from which water could be diverted. It is likely that diversion structures, agricultural features, and perhaps even habitation remains on canyon floodplains have been removed or buried by the larger floods occurring between prehistoric and modern times. The absence of sites in canyons of the Cochie and Cottonwood drainages (Fig. 3.2) reflects survey coverage rather than a departure from the general distributional pattern in the northern Tucson Basin.

Upper Bajadas

West of the river in the southern community (Zone 5), the width of the bajada is greatly compressed between the Tucson Mountains and the Santa Cruz, increasingly so toward the northern end of those mountains (Fig. 3.9). Relatively few sites have been recorded on the slope below these mountains and only small ones with evidence for habitation are located here. Farmers with floodwater fields along the mountain edge may have resided in nearby settlements along the river.

Two situations for agriculture predominate on the upper bajada below the Tortolitas (Zone 4), where substantial habitation sites of all periods are found. The first involves areas between major trans-slope washes that 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 the base of the mountain front, heightened by the initial uplift of moisture-laden air passing over the peaks, adds to flow in the smaller streams. Numerous instances of low-density but extensive arrays of simple stone alignments indicate the agricultural use of overland runoff in broad, gentle swales as well as water diverted from channels.

To the southeast of the Marana Community, a dispersed pattern of small habitations and other sites on the bajada occurs in the vicinity of Caņada Agua (Fig. 3.9). These sites are primarily oriented toward moderate-sized drainages. Farther from the mountains here, slopes flatten over deeper valley fill, and many smaller channels become shallow, braided, and easier to divert (Fig. 4.2). A continuous scatter of isolated artifacts corresponds to the vicinity of braided channels even where sites are lacking and appears to represent remains resulting from farming activities. Sites at the southeastern edge of this

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 Teiwes.)

area are located outside the Marana Community and are affiliated with Preclassic or Classic period communities with centers farther east.

It is difficult to assess the relative agricultural reliance on primary and secondary drainages on the upper bajada within the Marana Community. In the Classic period and earlier, both large and small sites tend to cluster along the major Cottonwood, Derrio, and Guild washes. Smaller habitation sites are located between these on secondary channels, but distances would have permitted additional land in such situations to be cultivated by residents of settlements along the major watercourses.

Near the upper reaches of Guild Wash as it leaves the Tortolitas, the placement of dispersed small habitations on minor drainages resembles site locations about Caņada Agua. Agricultural features such as rockpiles and alignments are also concentrated here. Both within the Marana Community and to the south near Caņada Agua, remains oriented toward secondary drainages extend a limited distance beyond the mountain front onto the upper bajada. Contours of approximately 875 m (2700 feet) delimit the southern sites and the same elevation bounds similar sites within the Marana Community. Agricultural parameters probably figure in this pattern. Near the mountain front, sediment over pediment bedrock is typically no deeper than a few meters; water tables in drainages are correspondingly close to the surface. As drainages continue downhill across the bajada, surface flow tends to diminish or disappear in channels through infiltration into increasingly deep valley fill. The lower elevational limit of sites on small drainages likely marks the downslope extent of significant surface flow from all but the largest precipitation events.

Another factor in agricultural production involves the opportunity for early cropping at elevations above the level of cold air drainage. Winter frontal systems pass from the west and north toward the east across southern Arizona. Therefore, the greatest precipitation benefits from orographic uplift are experienced during this season along the western edge of the Tortolitas. In most years,winter rains cease well before May. The ensuing foresummer drought lasts under high temperatures until July, preventing satisfactory maturation of mid to late spring plantings. To effectively use winter moisture, the earliest possible planting date would have been necessary. A relationship between upper bajada settlement near small drainages and the elevational limits of inversion is illustrated by the fact that urban smog can be observed to hover in the valley bottom just below these sites. Crops may have been planted sufficiently early on these warmer slopes for spring harvests as well as summer ones.

Water in the large trans-bajada washes supported the second major farming orientation in Zone 4 on the upper bajada. Bottomland with high agricultural potential is not evenly distributed along the major washes, but varies with factors such as width and morphology of the floodplain, water table depth, watershed size, and drainage gradient. The importance of such acreage for supporting relatively dense populations is indicated by the locations of large habitation sites along those stretches of the major drainages suitable for floodplain fields.

A proliferation of both large and small settlements on Guild, Derrio, and Cottonwood washes occurred during the Classic period. To some extent, this density may reflect cultural preference: the desire of expanding populations to locate new habitations within community boundaries and near sites of origin. However, the largest upper bajada sites of Preclassic times also occur here, indicating a long-term productive advantage.

Compared to Derrio and Cottonwood drainages, Guild, Caņada Agua, Ruelas, Wild Burro, and Cochie washes are characterized by small watersheds (Table 4.3), lesser volume and frequency of flow, and narrower floodplains. The greater length of Guild than these others along the mountain front, where bedrock depth is shallowest, and its course across a gentler slope likely account for larger sites along the upper reaches. Caņada Agua, with only one large site, and Guild, with numerous large and small ones, traverse less steep portions of the upper bajada on which water diversion and control of flow was easier. Evidence of the strength of flow in Wild Burro Wash was considered disadvantageous by the Tohono O'odham and Mexican traditional farmers. Wild Burro, Ruelas, and Cochie have steeper courses, and sites along them extend for only short distances beyond the mountains.

Compared to those washes, Derrio and Cottonwood cross flatter topography and are lined by clusters of large sites farther downhill on the upper bajada. Bedrock visible in bank cuts indicates a favorably high water table These large sites coincide with segments of wide, arable floodplain and lush riparian vegetation within the wash bottoms (Fig. 4.3). Width of the floodplains is sufficient

Table 4.3 Mountain Watershed Sizes for Trans-bajada
Drainages In the Marana Survey Area

Prepared by Matts Myhrman

for water diversion from shallow channels onto fields at the side above flood limits. Trenching in Derrio and Cottonwood floodplains (Chapter 5) revealed intact features such as hearths in these floodplain edge situations at depths less than a meter. The greater upland watersheds of Derrio and Cottonwood (Table 4.3) may have supplied effluent flow from winter rains into the spring season as far downslope as these large sites or within short distances from them. Peak flow from summer rains in the two drainages would also reflect watershed size through frequency and volume of flow.

Derrio and Cottonwood floodplains were judged to be agriculturally desirable by the traditional farming consultants on the basis of vegetation indicators. The Mexican informant noted the size of mesquite trees. The Tohono O'odham farmer additionally commented on the intense green color of palo verde trees, the presence of catclaw (Acacia greggii), and a potential for hand-dug floodplain wells. In situations resembling Derrio and Cottonwood floodplains, highly productive fields are still cultivated today along major tributaries in the Rio Sonora valley of Mexico (Doolittle 1984: 124-135). Series of adjacent fields share diverted water from canals of moderate length. Such fields are also cultivated within tributaries of the San Miguel River in Sonora by means of short diversion ditches and floodwater techniques (Nabhan and Sheridan 1977).

Middle Bajadas

Middle reaches of the valley slope, representing large areas in Zones 2 and 3 (Figs. 3.8, 3.9) on the eastern side of the basin, were not a significant factor in agricultural production prior to the Classic period. Drinking water is inconveniently distant at the river or above at the mountain edge. Potential agricultural water appears as brief flows in the major trans-bajada drainages following only the largest storms. Water from lesser precipitation events typically generates local floods, which are

Figure 4.3. Segment of Derrio Wash with arable bottomland on the upper
bajada in the Marana survey area. (Photograph by Helga Teiwes.)

not sustained over long distances and infiltrate channels over deep valley fill. These drainages are incised beyond a depth for easy diversion even when water is available.

Many small drainages with bajada rather than mountain catchments are sufficiently shallow for successful diversion. However, such water would have been available only in cases of thunderstorms directly over the watershed, a relatively unpredictable event compared to higher elevation precipitation triggered by uplift of air over the mountains. The use of small drainages by means of earthen checkdams in Zone 2, which likely was attempted only in seasons of more promising rainfall, correlates exclusively with the Classic period proliferation of an agricultural technology fed by surface runoff. Simple mulches of piled cobbles, or rockpiles, enhanced and conserved soil moisture for drought-adapted crops of agave in vast fields (see Chapter 7). Overland runoff and direct rainfall were the sole water sources on gentle middle bajadas that were too marginal for annual crops such as corn, beans, and squash.

Lower Bajadas

Alluvial fans composed of outwash sediment from the uplands coalesce on the lower bajada in Zone 1 on the east and Zone 6 on the west (Figs. 3.8, 3.9). In portions nearer the floodplain, gentle slopes providing an active depositional environment and controllable water flow were favored by cultivators of every period. In these situations, floodwaters following storms provided both moisture and simultaneous enrichment for crops in the form of suspended nutrients and organic detritus. Dispersed settlements of farmers rather than the remains of agricultural activities register the reliance on lower bajada cultivation. Analogy with historic Sonoran Desert cultivators suggests brush, earth, and stone diversion structures on watercourses (Fig. 4.4), intrafield constructions of similar materials for water distribution (Fig. 4.5), and ditches or canals of moderate length (Fig. 4.6), all of which would rarely leave evidence in the archaeological record.

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

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

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

Tohono O'odham ak chin or floodwater farming in late historic time has provided the foremost model for understanding strategies of land use on alluvial fans. The ideal location of fields coincides with down-fan positions in which floodwater overflows increasingly shallow channels and spreads laterally over the fan surface, delivering water of sufficiently low force to avoid disruption of plantings. The closer the correspondence between these conditions and field location, the less labor investment would be necessary to divert water and construct ditches or protective barriers. Locations of optimal conditions shift on fans over time as hydrological activity changes in response to continuing geomorphological processes.

Geomorphological variables affecting floodwater farming in the study area are discussed by Field (1985; Chapter 5 in this volume), Waters (1987, 1988), and Waters and Field (1986). Advantages of small fans over large ones include higher proportions of fine-grained soil, lower thresholds of overbank flow, and less erosive force of flow when it occurs. Greater distance from the mountain front, also correlated with fan size, as in Zone 1 further enhances the accessibility of agricultural water. However, these advantages on small fans, as in the Tucson Mountains of Zone 6 must be balanced against the lower chance for thunderstorms over watersheds of more limited extent.

Late nineteenth century and more recent observations emphasize Tohono O'odham field placement with respect to points of natural water spreading (Bryan 1925, 1929), but ethnographic accounts recall more intensive past practices. Constructions are reported for runoff concentration in watershed areas upslope from fields and for water delivery from drainages to conjoined series of fields (Underhill 1939; Castetter and Bell 1942; Clotts 1915, 1917). Collective efforts were involved in constructing canals up to 1 km (0.5 mile) in length and even longer walls to divert runoff to plots or reservoirs. Nabhan (1986a, 1986b) notes that such efforts suggest a broader range of former field locations than at loci of natural water spreading, and he documents more varied situations of late historic Tohono O'odham cultivation. Downcutting and channel incision of major washes in southern Arizona in the late 1800s may have prevented their diversion, increasing subsequent use of smaller tributaries and watersheds with less predictable floodwater flow (Nabhan 1986a: 74). Higher densities of farmers in prehistoric times and greater dependence on their own agricultural yields may have occasioned other departures from idealized Tohono O'odham practices of the late historic period.

Possible evidence for more intensive methods of water management on alluvial fans has been found in excavated cross sections of buried drainages. Trenches dug on the lower bajadas in Zone 1 of the Marana Community (Katzer and Schuster 1984) intersected several secondary channels that differ in morphology from current natural drainages and that may have been constructed to direct overland flow. Compared with modern drainages, these channels are shallower and less concave, are unrelated to present drainage patterns, exhibit unusual lateral and vertical continuity, and contain indications of introduced flow greater than that to which the fluvial system was adjusted (Katzer and Schuster 1984). They appear to be contemporary with adjacent Hohokam occupations. Canals revealed by excavations on a fan west of the Tucson Basin (Withers 1973) represent another alternative to reliance on purely natural water-spreading processes. Canals would have permitted upstream diversion from large drainages with substantial watersheds that were less prone to natural overbank flow and floodwater spreading.

Late historic floodwater farming of isolated fields allowed ideal positioning for minimal labor investment and relocation whenever events such as massive floods altered favorable conditions. Use of fan areas by larger numbers of farmers prehistorically would have restricted relocations and prompted cultivation of less easily farmed locales. In a variety of situations with divertible water within the Marana Community, both the Mexican and Tohono O'odham consultants considered soils too sandy to achieve high yields. They suggested that finergrained and more water-retentive planting mediums could be created by water diversion and resultant silt deposition for one to several years prior to cultivation. Such improvements might encourage further modifications toward field permanence, such as ditches or canals of moderate size from upstream diversions. Along the Rio Sonora, Doolittle (1984) describes the incremental growth of such improved and locationally stable field systems over the course of long-term use.

Particular site locations on the lower bajada were used over long intervals of time, indicating commensurate stability in general farming locations. A dense concentration of lower bajada sites of all periods occurs in Zone 1 to the north of Wild Burro Wash. Observations of sediment deposition and organic flood detritus that were made over several years document frequent flow in the larger secondary drainages of this area. This kind of evidence and vegetational indications convinced both traditional farming consultants that this zone was among the most desirable settings for farming in the Marana Community. A second likely factor in concentrated settlement of the lower bajada is the prolonged availability of domestic water. Portions of Zone 1 with substantial early settlement are those close to the high water table along the Santa Cruz near the end of the Tucson Mountains.

Irrigation by canal from the river may have augmented agricultural production on the eastern segments of the lower bajada. Terrace height diminishes and width of the floodplain increases rapidly downstream from the Tucson Mountains, topographically permitting canal paths to diverge from the river and traverse the lower edges of Tortolita. alluvial fans. A series of prehistoric canals paralleling historic ones has been identified in aerial photographs and by surface remains, passing near or through a number of sites (Fig. 3.2). It appears that these canals supplied drinking water for permanent residence and also may have supported some irrigated acreage.

Santa Cruz Floodplain

Risks and opportunities for Hohokam agriculturalists in Zone 5 on the river floodplain cannot be judged precisely by present conditions (Figs. 3.8, 3.9). Historically, perennial surface flow in the Santa Cruz has been absent north of Tucson. Pre-Columbian runoff would have been less rapid due to wooded stream courses and the grass cover heavier before the appearance of livestock. Floods would have been poorly contained when the river channel was less incised. Current entrenchment and associated lowering of water tables began prior to this century (Cooke and Reeves 1976; Betancourt and Turner 1988).

It is not clear whether episodes of channel incision apply equally and simultaneously to the Santa Cruz throughout its length in the Tucson Basin. An instance of downcutting prior to the Classic period in the southern Tucson Basin apparently caused the abandonment of riverine settlements dependent on irrigation (Waters 1988: 217). The lack of later occupations at several Preclassic settlements along the river near the southern boundary of the Marana Community (Figs. 3.1, 3.2) may also have followed the disruption of canal intakes through vertical or horizontal shifts of the river channel.

Historically, localized areas of high water table and sustained flow of surface water were created in the Tucson Basin by igneous intrusions related to the mountain masses. Without such impervious barriers, water infiltrates the porous riverbed in other stretches and flows underground after short-term floods. In the northern basin, a volcanic intrusion, higher water table, and more prolonged flow occur near the end of the Tucson Mountains in Zone 5. Accessibility and duration of flow create the best situation for canal intakes, and a variety of historic lines headed here.

Compared to the Phoenix Basin, floodplain width and terrace morphology in the Tucson region significantly restrict the extent of irrigable land. Nevertheless, irrigated fields along the floodplain supported the densest populations in the study area. Intensive production of annual crops by irrigation likely played a significant role in community-wide population levels. Riverine settlement in both Preclassic and Classic periods was greatest in the area of high water table surrounding the mountain end.

Floodplain surfaces of the Santa Cruz have been alternately scoured and buried. Lateral channel movement that has obscured prehistoric activity is indicated by archaeological materials eroding from a west bank cut at the Tucson Mountain terminus. Remains of prehistoric canals are not visible on the surface of this zone and have only been identified in excavation. Evidence of buried canals has been encountered on the west side of the river just south of the study area (Kinkade and Fritz 1975) and within the Marana Community near the end of the mountains (Bernard-Shaw 1988). Due to overlying deposits, it is doubtful that the extent of Hohokam irrigation along the Santa Cruz can ever be comprehensively documented.

North and downstream from the Tucson Mountains, the river floodplain widens rapidly and terrace barriers to the lateral extension of canals diminish. At the same time, the channel becomes more poorly defined and surface flow disappears underground except during large flood events. Floodplain expanses are subject to extensive shallow flooding over fine-grained alluvial soils.

Canal headings are not feasible on this stretch of the river, but irrigated fields were supplied historically by lines with intakes near the end of the mountains. Gravity canals on the east side of the river extended as far north as the modern town of Marana (Roskruge 1896a, 1896b). It is likely that Hohokam canals irrigated more land in this area prior to historic river channel incision. The broad floodplain is under cultivation today, and surface indications of prehistoric canals are preserved only at higher elevation along the lower edge of the bajada. Similarly, the outlines of settlements in these modern fields are poorly defined. As a second method of prehistoric floodplain cultivation in this area, a large set of rockpiles on an undisturbed stretch of a low west terrace just north of the mountains may have functioned to control the dispersion of shallow floods.


A perspective of regional scale provides richer insights into land-use patterns than would be possible through a compilation of isolated evidence. With the emergence of an overall settlement configuration in the northern Tucson Basin, the insufficiency of ethnographic analogy for understanding the breadth of Hohokam subsistence has become apparent. Some agricultural technologies and environmental situations were similar to those used by Piman Indians, but other components of Hohokam production are without analogs even in the earliest Spanish accounts. There are no descriptions of agricultural patterns similar to those in upper and middle bajada zones. Ethnographic parallels are also lacking for productive modes that were capable of underwriting the higher range of Hohokam population densities and the occupation of settings that were historically abandoned.

The availability of regional distributions sharpens the discrimination of environmental variables affecting landuse patterns. Elements of the technological repertoire represent incomplete information without reference to the regional variety of associated environmental contexts. For example, rockpile devices occur in more than one relationship to water and topography. In most instances, rockpiles were constructed on middle bajada ridge tops where runoff follows the general elevational trend from uplands to valley floor. Other rockpiles were located on the broad, gently angled sides of bajada drainages. Here, overland flow of water is at right angles to the downhill slope toward the valley bottom and coincides with the path of runoff as it joins streams laterally. Still another setting for rockpiles is on a low terrace of the Santa Cruz River. In this location, water originated from shallow overbank flooding or canals.

From one viewpoint, stability in Hohokam land use characterizes the northern Tucson Basin. A dual site distribution, with one band along the floodplain and adjacent lower bajada and one band paralleling the flanks of the eastern mountains, was established early in the sequence and remained constant. Harshness of the Hohokam environment has been viewed as imposing a modicum of settlement stability through the topographic requirements for canal systems and labor investments in them (Haury 1976: 354; Nicholas and Neitzel 1984; Masse 1981, 1991). Restricted opportunities to concentrate water for floodwater farming or surface runoff management likely crystallized land use in additional topographic situations. Technological conservatism is reflected by construction of the same feature types for hundreds of years. In the Tucson Basin, the longstanding relationship between settlement patterns and hydrological opportunity suggests continuity in agricultural approaches.

These observations do not imply, however, that the system of Hohokam production was static or without identifiable trajectories. Dynamic elements were interjected by changing economic needs and aspirations of Basin inhabitants over time. The Classic period proliferation of rockpile fields above the Marana Mound Site illustrates a dramatic agricultural reorientation based on preexisting technology (see Chapter 7).

An understanding of environmental opportunity and available technology are not sufficient to predict prehistoric decisions as to the form and extent of implementation in agricultural production. Additional natural but nonagricultural variables may impinge, such as the cooccurrence of suitable domestic water. More importantly, agricultural production is the outcome of economic decisions on the part of individuals and groups, who consider cost, risk, expectations, and cultural values in formulating their responses. During successive periods, agricultural parameters of the northern Tucson Basin influenced economic behavior in changing contexts of regional settlement and demography.

Next: Chapter 5 | Previous: Chapter 3

UAP Online Books  |  UAP Home Page

The University of Arizona Press, 2/10/01 5:03PM