The Dairy Site: Occupational Continuity
on an Alluvial Fan

Paul R. Fish, Suzanne K. Fish, John H. Madsen,
Charles H. Miksicek, and Christine R. Szuter

During the period of early ceramic occupation, when subsistence patterns in the Marana area can first be described in detail, nearly the full range and typical proportions of later Hohokam cultigens are present. By this time, the transition to agriculture appears to have been completed, consistent with evidence for the prior importance of domesticates in Late Archaic sites of the Tucson Basin and surrounding areas (Chapter 2). Studies at the Dairy Site (AZ AA:12:285) document subsistence activities from an initial interval of plain and red ware pottery into the early Colonial period (Fig. 2.1).

The Dairy Site is situated on an alluvial fan of the lower bajada (in Zone 1 of the later Marana Community; Figs. 2.3, 3.9). Floodwaters reaching the fan after storms nourished the fields of its residents and periodically covered the remains of their houses, all the while supplying sediments that continued to build the fan in height and in downslope extent. Features exposed in a cross section of the Dairy fan provide a powerful illustration of floodwater farming and persistent zonal land use, inferred elsewhere from distributions of sites visible on fan surfaces. Occupations spanning more than 600 years between the third and ninth centuries A.D. exemplify the continuing correspondence between agricultural technology and settlement locations throughout the chronological sequence.


Settlement history at the Dairy Site is derived not from the present surface of the alluvial fan but from remains in a profile 500 m long and 3 m high created by the Shamrock Dairy along its property line (Fig. 6.1). Cultural deposits were exposed 50 cm or more below the present ground surface along all but the western portion of the profile (Fig. 6.2). Intersected features represent a sample of fan occupations through time and space. The Dairy profile was selected as a study site for investigating floodwater farming as part of research on prehistoric agricultural strategies in the northern Tucson Basin, supported by the National Science Foundation. The profile was cleaned by students in an undergraduate field methods class at the University of Arizona and a detailed profile map was made by John Field, project geomorphologist. Artifact collections and samples of subsistence remains were obtained by first cutting into the profile to a depth of 25 cm to 40 cm above intact deposits in each cultural feature. Excavations then proceeded down from the initial cut through the fill to the floor or bottom of the exposed feature.

Late Archaic Occupation

Initial use of the Dairy fan for agriculture appears to represent a preceramic occupation, although earliest ceramics elsewhere have been dated within the same general time range. The profile intersected only one welldefined feature of probable Late Archaic affiliation. Features 12 and 14, concentrations of charred wood and ash, and a hearth, Feature 34, occur on separate segments of surfaces in the profile for which similar stratigraphic position implies previous continuity (Fig. 6.2). These surface segments are from 1 m to 50 cm below ceramic features in the respective profile sectors. Feature 34 produced a mesquite charcoal date with a midpoint in the late third century A.D. (Table 6. 1) and a pollen sample containing corn and distributions of weedy species that suggest agriculture (Table 6.2). Exposed preceramic remains do not necessarily represent the earliest farming activity at this fan location. Determinations on cultigens from several Late Archaic sites in the northern Tucson Basin predate Feature 34 at the Dairy Site by a number of centuries (Table 2.1).

Figure 6.1. View downslope along a section of the Dairy Site profile.

Early Ceramic Occupation

Fifteen pit houses and a variety of other features of the early ceramic sequence were encountered. Profile cross sections suggest large houses, analogous to the Pioneer period structures at the well-studied site of Snaketown in the Phoenix area (Haury 1976). Some floors are well plastered; others are not. Exposed intramural firepits are clay lined. Bell-shaped storage pits are present in the floors of some structures. Two inhumations in pits within houses and two cremations were exposed by profile cleaning, but numbers of burials have been encountered by Dairy employees during a decade of fill removal in front of the existing profile. Dairy employees also reported that numerous areas of dark soil and a variety of artifacts were observed during fill removal from an area covering 72 ha (180 acres). The site extends for an unknown distance behind the present profile.

Decorated ceramics represent types from the Phoenix core and somewhat later types from the Tucson Basin sequence, including Estrella Red-on-gray, Sweetwater Red-on-gray, Snaketown Red-on-buff, Cañada del Oro Red-onbrown, Gila Butte Red-on-buff, and Santa Cruz Red-on-buff (Table 6.1). A polished Tucson red ware unlike the thin Vahki Red at Snaketown is present in all Pioneer period contexts and is classified as Tortolita Red following Bernard-Shaw (1990a: 210). This red ware accounts for up to seven percent of the pottery in some features but only two percent of the total ceramic assemblage. Pottery in the local red-on-brown tradition first appears in the Cañada del Oro style of the early Colonial period. Only two percent of the 7,000 sherd total are painted or incised.

The lack of superpositioning and wide spacing of most features along the profile (Figs. 6.2, 6.3) suggest relatively rapid burial of each unit and a restricted time span for postoccupational fill above the floors. Because there is some overlap between occupied areas of the site during sequential phases and because of the rarity or absence of diagnostic artifacts in individual features, radiometric dates, stratigraphic relationships, and ceramic associations (Table 6.1) have been combined to identify the chronology of profile segments shown in Figures 6.2 and 6.3. Neither dates nor ceramics indicate occupation of the Dairy fan after the early Colonial period.

Figure 6.2. Dairy Site profile, showing locations of features.

Figure 6.3. Central portion of the Dairy Site profile, showing stratigraphic
relationships of chronologically significant features.


As the result of opportunities for floodwater farming, late Holocene alluvial fans on the lower bajada are areas of high site density. Interaction of several factors influences the floodwater potential on these alluvial fans, including areal extent of sheetflooding on active fan lobes and minimum discharge needed to induce overbank flooding by drainages. Areas frequently inundated by sheetfloods are highly conducive to floodwater farming techniques because water evenly wets the entire active surface and deposited sediments have relatively high silt content (Chapter 5). At present, the active and most easily farmed portion of the 1.35-square-kilometer (0.75square-mile) Dairy fan is more than 20 percent of the total surface, compared to the approximate 15 percent average for the study area as a whole.

Occupants of the Dairy Site changed settlement locations on the fan surface as it aggraded and as active portions shifted laterally and along the length of the fan over time. It is likely that contemporary habitation occurred adjacent to active reaches and that remains were subsequently buried by locational shifts in sedimentation. Nearby structures of differing date in the central profile suggest that some fan locales remained relatively stable during occupations of sequential phases. The absence of features postdating the ninth century A.D. and the early Colonial period implies that the profile portion of the fan or the entire fan became inactive and unamenable to farming after this time.


The Dairy fan is advantageously located for efficient exploitation of a variety of plant and animal resources in several topographic zones as well as for receiving floodwaters. The head of the alluvial fan is only 3 km (1.5 miles) from the Santa Cruz River and its riparian habitats. Dense cacti and leguminous trees with edible beans grow in mid and upper basin vegetation immediately above the fan. Resources from both bajada and riparian zones are consistently found among subsistence remains from the Dairy Site.

Table 6.1. Feature and Chronology Information for the Dairy Site (AZAAA2:285)

Note: Features yielding only plain wares and red wares are included in the Early Pioneer
period; feature numbers 35 and 38 were not used. * Period designation based
on stratigraphic relationship only.
a = corn, b = mesquite, c = mesquite seeds.

Table 6.2. Percentage of Pollen Types in Samples from the Dairy Site (N = 200)

Cultigen pollen, corn and cotton, was excluded from a standard sum of 200 pollen
grains per sample that was used to calculate percentages of all other types.
Cultigen pollen was tabulated in addition to this standard sum and is
quantified as the number of occurrences encountered during tabulation
of the 200 grains

Pollen Analysis

Pollen distributions at the Dairy Site (Table 6.2) suggest continuity in land use over the entire course of occupation into the Colonial period. Even in the earliest features, pollen configurations reflect substantial modification of natural vegetation by residential disturbance and surrounding floodwater fields. High chenopod and amaranth (Cheno-Am) percentages are not characteristic of the natural vegetation of the Dairy Site environment (Hevly and others 1965) but are hallmarks of recent Tohono O’odham floodwater fields (S. Fish 1984a, 1985). Cheno-Am is the dominant pollen type at the Dairy Site. Frequencies are comparable with those at early Classic period sites in the most favorable farming locations farther north along the lower bajada and are greater than at later sites in less agriculturally favored fan situations (S. Fish 1987a). Additional species (spiderling, globe mallow, and Arizona poppy) in an agricultural weed category (S. Fish 1985) are also well represented in Dairy Site samples.

Cultigen values throughout the Dairy profile are also comparable with later Hohokam sites of the study area and Tucson Basin (S. Fish 1988). Again, equivalence is

Table 6.3. Numbers of Flotation Samples Containing
Carbonized Plant Remains at the Dairy Site

with sites having the best floodwater potential or even with sites in riverine settings. Corn pollen is widely distributed among the earliest features with plain wares and red wares. Variation in amounts of cultigen pollen does not form consistent temporal trends, but rather appears to have resulted from differential activities involving resources on a feature-by-feature basis during each time segment.

Pollen evidence for use of wild products, best exemplified by cholla and other cacti, is found throughout the occupation but at a moderate level compared to other Tucson sites. Greater emphasis on cacti can be documented at some Zone 1 sites of the early Classic period (S. Fish 1987a, 1987b; S. Fish and Donaldson 1991). At least occasional use of upland resources is apparent in the pollen of yucca in one pit house. The occurrence of cattail pollen from earliest to latest contexts shows access to permanently damp habitats, probably along the Santa Cruz River, over the duration of occupation.

Flotation Analysis

Flotation samples further reveal a diverse subsistence base for the Pioneer and early Colonial periods (Table 6.3). Corn, cotton, squash, and agave remains were identified from features dating to the earliest occupation with

Table 6.4 Faunal Remains from the Dairy Site

Note: Nonmammalian remains include: Feature 20, Kinosternon sp., mud turtle
(1 fragment); Feature 21, Anas sp. Or Mareca Americana, duck or widgeon
(1 fragment); Feature 23, Phasianidae, quail (1 fragment); Feature 29 - hasianidae,
quail (1 fragment); Feature 33, Crotaphytus (= Gambelia) wislizeni, leopard lizard
(1 fragment). A = 1 specimen worked, b = 2 specimens worked

plain ware ceramics. The only additional cultigens, bottle gourd and common bean, were encountered in the same Late Pioneer feature. Seeds from weedy plants that may represent purposely encouraged species or weeds of agricultural contexts are common throughout the occupational sequence. Recovery rates for mesquite, saguaro, and other cacti suggest an important wild plant component in Dairy Site diet. The frequencies of corn, mesquite, and saguaro are in the upper range of proportions recorded from most later Sedentary and early Classic period sites in the Tucson Basin (Miksicek 1988).

Identification of a range of plant parts (spines, fibers, leaf and heart fragments) in significant quantities extends the possibility of Hohokam manipulation and cultivation of agave back into the Pioneer period. Except for a few species that may have been introduced into the Hohokam crop repertoire during the Sedentary period (tepary beans, jack beans, and Amaranthus hypochondriacus), the data from the Dairy Site suggest Hohokam subsistence patterns were firmly established in the Marana study area by the early part of the ceramic horizon.

Faunal Analysis

As with most Hohokam sites, lagomorphs comprise the majority of identifiable bones (Table 6.4). The faunal assemblage, however, is comparatively diverse. That is, the quantity of distinct taxonomic groups is relatively high for the small sample of identified remains. Fourteen different taxa, including mud turtle, duck or widgeon, leopard lizard, quail, jack rabbit and cottontail, mule deer and bighorn sheep, coyote, ground squirrel, pocket gopher, kangaroo rat, and cotton rat were identified. Compared to other Hohokam faunal assemblages of similar size from settlements occupied during the Sedentary or Classic periods, the Dairy Site assemblage exhibits somewhat greater taxonomic richness.

Additionally, the fauna is characterized by a relatively greater abundance of artiodactyl remains from a variety of features. Artiodactyls comprised 13 percent of the identifiable fauna, were recovered from more than a third of the features, and their skeletal elements represent portions of the entire body. The relative abundance of artiodactyl remains, their widespread spatial distribution, and entire skeletal representation are unique characteristics for Tucson Basin Hohokam sites. Sedentary and Classic period lowland Hohokam sites tend to have few artiodactyl remains and these few represent the cranial and podial portions of the skeleton.


The nature of the Late Archaic presence on the Dairy fan is unclear, although pollen of corn and weedy species suggests floodwater farming had begun. In a similar zonal setting 4 km (2 miles) to the north, trenching at a Late Archaic site revealed pits containing corn with midpoint dates in the fourth century B.C. (Roth 1989). Just south of the study area, structures were preserved at another Late Archaic fan site that yielded corn and a still earlier date in the eleventh century B.C. (Mabry 1990). The earliest feature in the Dairy profile, in the third century A.D., marks an occupation that already postdated more than 1000 years of floodwater farming on alluvial fans of the northern Tucson Basin.

Diagnostic artifacts in houses of the Dairy fan cross section are commensurate with continuous occupations through the early Colonial period. overall numbers of structures cannot be reconstructed for any specific phase or through time from the profile sample. However, distributions of houses and the presence of ancillary features and burials suggest a settlement structure little different in size and arrangement from the dispersed clusters of dwellings in excavated sites of the Classic period on Zone 1 alluvial fans (Henderson 1987a; G. Rice 1987a).

Domestic water for Dairy Site inhabitants undoubtedly came from the river at a distance of 3 km (1.5 miles). Occupations along the Santa Cruz at the toe of the Dairy fan and at other nearby locations are generally contemporary, according to ceramic attributes and absolute dates. Summer cropping would have predominated on the fans and the river floodplain, prohibiting effective participation in both floodwater farming and riverine irrigation during critical episodes of rainfall, rapid runoff, and peak flows. Nevertheless, fan residents may have increased the labor force that could be tapped for swift repairs of canal intakes following summer floods.

The happenstance of profile creation at the Dairy Site offered a uniquely vertical perspective on settlement continuity that could otherwise be seen only in its horizontal dimensions. Floodwater farming on lower bajada alluvial fans of the northern Tucson Basin was capable of supporting long-term sequential occupations, in this case illustrated by 600 years of sustainable agriculture. The record of such occupation and production in the early years of the Hohokam sequence is surely underrepresented in settlement patterns based on surface visibility. Continuing fan deposition and the scarcity of diagnostic artifacts in early ceramic assemblages makes it difficult to distinguish these phases where settlement is registered by sparse remains. With the apparent demise of favorable hydrological conditions, the Dairy fan was eventually abandoned, but not the zone or this enduring mode of farming.

Next: Chapter 7 | Previous: Chapter 5

UAP Online Books  |   UAP Home Page

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