Water Control and Subsistence

by Melissa J. Churchill

The people of Woods Canyon Pueblo grew crops, hunted animals, gathered plants, collected and managed water, and used a variety of resources in their environment to build and heat their homes, cook food, and make tools and clothing. In this chapter, I discuss where the residents obtained domestic water, how they managed water runoff, and where they grew their crops. I also draw on the results of faunal (animal bone) and archaeobotanical (plant) analyses to reconstruct the diet as well as the hunting, gathering, and agricultural practices of the people of Woods Canyon Pueblo.

Water Resources and Water Management

Water was essential for drinking, cooking, washing, building homes, watering plants, and making pottery. The proximity of Pueblo III sites to natural water sources and the common occurrence of Pueblo III water-control features demonstrate that access to, and management of, water was a primary concern of the people living in the Mesa Verde region during this time (Haase 1985*1:19, 20, 24). Reservoirs and stone alignments were the most common water-conservation devices of the Pueblo III period (Haase 1985*1:20), and both are present at or near Woods Canyon Pueblo.

Natural Water Sources

Springs and seeps are located in the canyon bottom south of, and upstream from, the site. Currently, there is water in the canyon bottom year-round, which may be partly the result of modern irrigation. Even so, it is likely that the springs and seeps provided a year-round supply of water when the pueblo was occupied. Water that was captured in the nearby reservoir appears to have been a supplemental source of domestic water.

Woods Canyon Reservoir (Site 5MT12086)

Woods Canyon Reservoir is located .75 km northeast of Woods Canyon Pueblo, near a tributary of Woods Canyon (Database Map 337). Water was not intentionally diverted from existing drainages into the reservoir, nor was the reservoir fed by a spring. Rather, rainfall and water runoff from the surrounding area was impounded by a horseshoe-shaped dam constructed of earth and some stone. The predominance of white ware jar sherds found in the vicinity of the reservoir indicates that the water in the reservoir was used for domestic, rather than agricultural, purposes; white ware jars would have been appropriate for transporting water the short distance to the site (Wilshusen et al. 1997*1:674, 675). Furthermore, the shallow depth of the soil near the reservoir and the absence of a canal system argue against the impounded water having been used to irrigate fields, although pot irrigation is a possibility (Wilshusen et al. 1997*1:675).

The available evidence suggests that the reservoir was built earlier than the pueblo, although their periods of use overlapped. A detailed analysis of the dating of the pueblo is presented in "Chronology"; in that chapter, a case is made for construction beginning in the mid–A.D. 1100s, with occupation continuing through the late 1200s. A comparison of the formal white ware pottery types from both the pueblo and the reservoir (Table 1) reveals that the assemblage from the reservoir is dominated by Mancos Black-on-white and McElmo Black-on-white, which suggests that this feature was constructed and used sometime between A.D. 1125 and 1200. The small amount of Mesa Verde Black-on-white, a later type that predominated in the mid- to late 1200s, indicates only limited use of the reservoir during this period. The white ware assemblage from the pueblo, on the other hand, contains significant proportions of all three types but has a much higher percentage of Mesa Verde Black-on-white than does the reservoir, which suggests that occupation of the pueblo continued after the reservoir was no longer used or was used only occasionally.

It is unlikely that people would have abandoned a water source, or that a pool of water would have gone unused during the late Pueblo III period. One possible explanation for why the reservoir might not have been used, or not used very much, during this period is that there was seldom any water in it, which in turn could indicate that there was a drought during the middle to late 1200s. Another possibility is that, over time, the reservoir filled with so much silt that it was no longer able to hold water.

I should note that sample size might be skewing the results, because the pottery assemblage from the pueblo is almost 10 times larger than the assemblage from the reservoir. It is also possible that Mesa Verde Black-on-white sherds are underrepresented at the reservoir—the assemblage is characterized by a high percentage of jar sherds, which as a group tend to be decorated with designs that make their assignment to formal type difficult, and a low percentage of bowl and mug sherds, which are more likely to exhibit typical Mesa Verde Black-on-white stylistic elements and therefore are more likely to be assigned to that category (see Wilshusen et al. 1997*1:674). The ratio of "typed" white ware bowl sherds to "typed" white ware jar sherds is only approximately 1:1 in the assemblage from the reservoir, whereas in the assemblage from the pueblo it is 4:1. A comparison of type frequencies for jar and bowl sherds at the reservoir shows little difference between the two (Table 2). Most notable is the low frequency of both bowl and jar sherds typed as Mesa Verde Black-on-white at the reservoir. There is much more variation in the frequencies of jar and bowl sherds at the pueblo. When the typed white ware sherds from the early and late components at the pueblo are subdivided by form, the reservoir assemblage still appears to be earlier than the pueblo assemblage. These data seem to reflect a true scarcity of Mesa Verde Black-on-white at the reservoir.

Thus, I conclude that the reservoir was built before Woods Canyon Pueblo, probably by people living at several earlier mesa-top sites near the reservoir (Dexter, the Bass Site Complex, and the Albert Porter preserve; see Database Map 337). It is likely that the reservoir was associated with a long-lived community rather than with Woods Canyon Pueblo exclusively (Wilshusen et al. 1997*1:679).

Unlike the springs and seeps in the canyon bottom, the reservoir did not provide a constant source of water to the residents of Woods Canyon Pueblo. Water would have been contained in the reservoir only after rains or during spring snowmelt, when water would have been captured behind the dam. Even though a reservoir of this size could have held 40,000 to 50,000 gallons of water (Wilshusen et al. 1997*1:673), it probably did so only sporadically during the summer monsoon season. The springs and seeps in the canyon bottom undoubtedly were a more reliable source of water, and they are closer to the pueblo than is the reservoir. The distance between the reservoir and the pueblo is almost half a mile. The most direct route to the reservoir from the pueblo was from a side canyon just east of the site. Although the route to and from the reservoir is not arduous, carrying pottery vessels full of water back to the pueblo would have made the trip more difficult.


Linear alignments of stones are visible in the main drainage that bisects the site, in several peripheral drainages, on the talus slope, and at the base of the slope near the canyon bottom. Most of these alignments are believed to be checkdams, that is, constructions used specifically for water management and soil conservation (see Hayes [1964*1], Rohn [1963*1], and Winter [1978*1] for more-detailed discussions of the possible functions of checkdams), as opposed to dams, which are used for water impoundment.

Most checkdams at Woods Canyon Pueblo are located in the main drainage that bisects the site. The checkdams extend across the drainage and are clustered in groups of two to three. Six sets of checkdams, not all of which were mapped, have been identified in the main drainage; most are located above the canyon rim. Two of the sets above the cliff were partly exposed during test excavations in Nonstructure 9-N and are discussed in paragraphs 12–19. Below the cliff face, the main drainage is more entrenched, and rubble from nearby buildings made the identification of checkdams more difficult.

Checkdams in Nonstructure 9-N

Two sets of checkdams were excavated in Nonstructure 9-N, located above and near the edge of the canyon rim, in the main drainage of the site (see Database Map 270). These checkdams were constructed of large, unshaped rocks. The walls were dry-laid and uncoursed. The three checkdams visible on the modern ground surface range from 6 m to more than 20 m long. In contrast, walls identified during the Chapin Mesa survey at Mesa Verde National Park average 4.5 to 5.5 m in length (Rohn 1963*1:442), and the walls at Point of Pines average 10 m, rarely exceeding 20 m (Woodbury 1961*1:12). Thus, some of the walls in Nonstructure 9-N at Woods Canyon are longer than the walls in either of those other locations.

Most of the checkdams in Nonstructure 9-N are one stone wide and a little less than 50 cm thick. One wall (Nonstructure 9.1-N, Feature 3) is 1.36 m thick, although this measurement may include fallen stones. With the exception of one unusually tall checkdam (Feature 3), the average existing height of the checkdams in Nonstructure 9-N is .5 m, and most stand only a few courses high. The tall checkdam is comparable in height to most of the checkdams on Chapin Mesa (Rohn 1963*1:442). It is probable that the shorter walls at Woods Canyon originally stood taller and have collapsed over time.

The northern set of checkdams consists of two parallel walls (Nonstructure 9.1-N, Feature 1, and Nonstructure 9.2-N, Feature 2) that lie close to each other and perpendicular to the drainage (Database Map 299). The southern wall is the most substantial of the two; it rests on bedrock and was built first. The northern wall was built on a layer of sediment that had accumulated behind the southern wall. The fact that the sediment had not been removed before the second checkdam was built suggests that the purpose of the first wall was to promote sediment accumulation. The proximity of the two checkdams, and the order in which they were built, suggests that the later wall might have been built to replace the earlier wall or to help it slow the flow of water.

Three walls (Nonstructure 9.1-N, Features 3 and 4, and Nonstructure 9.3-N, Feature 5) make up the southern set of checkdams in Nonstructure 9-N (Database Map 270). They were built very close to the cliff edge and close to one another. The northernmost (Feature 3) and southernmost (Feature 5) walls are quite long and extend across both branches of the main drainage. The north wall and middle wall (Feature 4) were built first, directly on bedrock. Later, the south wall was built on a layer of construction fill. Construction sediment downslope from the north wall and on both sides and the top of the middle wall might have helped these constructions retain water. It seems unlikely that walls made exclusively of stone could have held water for long.

The checkdams in Nonstructure 9-N might have served several purposes. They probably were intended to slow runoff rather than to channel or impound water. If they had been designed to channel water, they would have been neither straight (Doug Ramsey, personal communication 1996) nor oriented perpendicular to the drainage. The linear configuration of these walls contrasts sharply with the curved dam at Woods Canyon Reservoir and would have rendered them ineffective at retaining water for any length of time. The three checkdams near the canyon edge would have prevented water from cascading over the 35- to 40-ft cliff, which in turn would have prevented damage to the buildings below and controlled erosion both above and below.

The most massive checkdam exposed in Nonstructure 9-N (Feature 3) might have impounded water, although it seems unlikely that it was designed for that purpose. The wall stands 1.12 m tall, which falls within the range of the estimated height of the Woods Canyon Reservoir dam (Wilshusen et al. 1997*1:673). Sediment intentionally placed on the downslope side of the wall would have improved the water-retaining capability of this dam. The sediment might have been placed as part of the construction of the checkdam or added later to reinforce the stone wall; it also could have been excess accumulated sediment that was removed from the behind the wall and deposited on the downslope side during routine cleaning and maintenance. Distinctive laminae behind the wall are evidence that sediment settled out of slow-moving or standing water. All these circumstances indicate that water did in fact accumulate behind the checkdam, but the presence of additional checkdams farther down the drainage suggests that the impoundment of water may have been incidental.

Researchers have suggested that crops might have been planted in the sediment that accumulated behind this type of checkdam (Hayes 1964*1; Rohn 1963*1; Smith 1987*2; Winter 1978*1; Woodbury 1961*1). The checkdams would have trapped or slowed runoff, allowing silt to settle out of the water and create fertile terraces that could have been farmed (Rohn 1963*1:441, 442). Ernest M. Vallo, Sr., of the pueblo of Acoma in New Mexico and a member of Crow Canyon's Native American advisory group, offered his own perspective on the possible functions of these constructions. He believes that the residents of Woods Canyon Pueblo might have planted crops to the side of the main channel, where water could have seeped or been carried to the plants. Alternatively, the nutrient-rich sediments that accumulated behind the checkdams may have been collected and transported to nearby fields, a strategy that was practiced in the "old days" at Acoma (Ernest M. Vallo, Sr., personal communication 1996).

There is evidence that sediment accumulated behind at least one of the walls (Feature 1) during the occupation of Woods Canyon Pueblo. The fine sediments that accumulated behind other checkdams (Features 2 and 3) also might have been deposited during occupation. The sediment originated from the mesa top above the site and could have traveled to the canyon edge only when an erosional environment existed. Such an environment was probably present during the occupation of the site, when vegetation had been cleared from the area (Doug Ramsey, personal communication 1996).

Other Checkdams at Woods Canyon Pueblo

Two possible checkdams were identified toward the southern end of the site, near the base of the talus slope (Database Map 330). The upper checkdam might be an extension of a massive retaining wall located along the east side of the main drainage. The lower checkdam is separate from, but close to, the retaining wall. These two features might have been built to prevent soil erosion from around the bases of the buildings to the south. Alternatively, the walls might have diverted water out of the main channel to irrigate the relatively flat area that lies east of the drainage and north of the buildings. Ernest Vallo, Sr., believes that the main purpose of these ancient checkdams was to prevent erosion, slow down water runoff, and trap rainwater (Ernest M. Vallo, Sr., personal communication 1996).

Several checkdams are also located along the drainage that roughly marks the west edge of the site. Two prominent alignments of boulders are on the canyon edge. Farther down the slope, there are several rough stone alignments that might have been checkdams. There may be a series of checkdams along a peripheral drainage that empties into an area that could have been a garden or field (Nonstructure 1-N). Most of these possible checkdams were not mapped.

The tallest exposed wall on the site is located midway down the talus slope, immediately west of the main drainage. The wall is 20 to 25 m long and reaches a height of 2 m above the modern ground surface. Its east end forms a corner with another wall. There is no kiva depression behind this wall. It is possible that the wall was constructed to impede the flow of water, and the flat area behind the wall could have been used for a garden or field. Two walls in almost the same configuration and constructed of massive stones were identified farther upslope just west of the drainage and might have served a similar purpose. Further testing would be required to make a final determination on the function of these walls, and none is labeled a checkdam on Database Map 330.

Portions of nine stone alignments were identified in a large, flat, open area designated Nonstructure 1-N. These walls promoted the accumulation of sediments and therefore might have created areas where crops could have been planted. These walls are discussed in greater detail in paragraphs 28–36.

Agricultural Areas

Identification of Agricultural Areas

Several locations at or near the site would have been suitable for farming or gardening, including the mesa tops, the canyon bottom, and the checkdam terraces. The mesa tops surrounding Woods Canyon are believed to have been used as farmland by the earlier residents of the community who lived at the mesa-top sites of Albert Porter, Bass, Dexter, and 5MT4700. It is likely that the people living at Woods Canyon Pueblo also planted crops on the mesas, even though their homes were farther away from the fields than were the earlier sites. The thick, eolian soils on the mesas are ideal for dry-land farming because they retain moisture from rainfall and melted snow.

The people also might have farmed in the broad, flat canyon bottom of Woods Canyon, which contains rich floodplain deposits and several springs and seeps. However, cold-air drainage—that is, the tendency for cold air to "pool" in low-lying areas—might have affected the length of the growing season in some years, occasionally limiting agricultural production in this area.

Checkdams in the main drainage (Nonstructure 9-N; see paragraphs 12–19) and in several peripheral drainages at the site might have trapped sediment, creating terraces of nutrient-rich soil that could have been used as fields or garden plots. Similarly, flat terraces behind walls built on, or at the bases of, slopes (Nonstructure 1-N) might have been farmed. In dry years, terraces such as these could have been critical farming or gardening areas, because, unlike fields on open mesa tops, they would have been watered by runoff (see Rohn 1963*1:447; Woodbury 1961*1:12). Even so, the crops from small terraces at Woods Canyon Pueblo probably only supplemented the crops from the mesa top and the canyon bottom because the terraces would not have been large enough to grow crops for the entire population of the pueblo (see Smith [1987*2:29] for interpretations of similar terraces at Mesa Verde National Park).

Modern Hopi farmers use a wide variety of agricultural fields, many designated for the planting of specific crops (Rohn 1963*1:447). Similarly, farmers at the pueblo of Acoma use terraces as special fields for the cultivation of ceremonial plants, a practice that may have been employed prehistorically at Woods Canyon Pueblo as well (Ernest M. Vallo, Sr., personal communication 1996).

Possible Agricultural Terraces in Nonstructure 1-N

Nonstructure 1-N is an open, relatively flat area in the canyon bottom (Database Map 330). The area is located at the foot of the steep talus slope and measures approximately 375 m2 (about one-tenth of an acre). Architectural units surround Nonstructure 1-N on its north, west, and east sides. In two places, there are gaps in the architecture where two small drainages enter the area. The drainage in the northwest corner lies between two kiva suites and is spanned by several possible checkdams. The drainage near the northeast corner cuts between two boulders. The southern boundary of this possible terrace is marked by a stone alignment. The area to the south of this alignment slopes gently to the south.

On the basis of surface indications, we first defined Nonstructure 1-N as a possible plaza or a great kiva. The results of testing indicate that neither a plaza nor a great kiva is present. Instead, Crow Canyon's excavations exposed a kiva (Structure 9-S), midden deposits from the surrounding buildings, crude stone walls, small extramural pits, and extramural surfaces in this area. Although our testing was limited in scope, we concluded that Nonstructure 1-N was probably used for multiple purposes—as a midden, as a possible outdoor work area, and as a possible garden or a small agricultural field—over a long period of time. Here, I focus on its possible use as an agricultural area.

Portions of nine extramural walls were documented in Nonstructure 1-N and in the north end of Nonstructure 4-N, an adjacent study area (see Database Map 258). The walls in Nonstructure 4-N are similar to the walls in Nonstructure 1-N. In fact, one section exposed in Nonstructure 4-N (Nonstructure 4.4-N, Feature 3) is believed to be a section of a wall that was also partly exposed in Nonstructure 1-N (Nonstructure 1.20-N, Feature 7). With one exception (Nonstructure 1.8-N, Feature 3), the walls are oriented east-west, the ideal orientation for slowing water runoff from the small drainages at the north end of Nonstructure 1-N. The walls consist of rough alignments of large, unshaped stones. All of the stones in each wall rest at approximately the same elevation. Two-thirds of the walls are only one course high; the remaining one-third are composed of three to four roughly stacked courses. Two walls are known to extend beyond the excavation units in which they were exposed because they are visible on the modern ground surface; they measure 6 and 18 m long, respectively. The average thickness of these walls where both faces are visible is 39 cm, which suggests that the walls were not thicker than masonry room walls.

The stratigraphic position of the walls indicates that some were probably used in conjunction with other walls; others, however, were no longer in use when other nearby walls were built. Three walls are interpreted as having been built early in the occupation of the pueblo, and the remainder are interpreted as being later constructions.

Three possible stone alignments were constructed during the early occupation of this area (Nonstructure 1.5-N, Feature 5; Nonstructure 1.22-N, Feature 8; and Nonstructure 1.19-N, Feature 12) (Database Map 261). These walls were found .5 to 1.0 m below the modern ground surface and rest on undisturbed native sediment. They are clustered together at the south end of Nonstructure 1-N.

Seven stone alignments are inferred on the basis of their location on or near the modern ground surface to have been built relatively late in the occupation of the pueblo; these are Nonstructure 1.6-N, Feature 2; Nonstructure 1.8-N, Feature 3; Nonstructure 1.20-N, Feature 7; Nonstructure 1.16-N, Feature 10; Nonstructure 1.12-N, Feature 11; Nonstructure 4.6-N, Feature 1; and Nonstructure 4.4-N, Feature 3 (Database Map 259 and Database Map 260). With the possible exception of Feature 3 in Nonstructure 4.4-N, these walls rest on cultural deposits. In most cases, they are located in midden deposits, and they probably prevented the refuse from eroding downslope.

Some of the fill at the south end of Nonstructure 1-N (Segment 1) might have been deposited intentionally to create farming or gardening terraces. It was clear that these sediments had not been deposited naturally, but they did not appear to be trash deposits either, which was the most common type of cultural deposit at the site. A couple of natural strata that might have been altered by agricultural activities were also identified at the south end of Nonstructure 1-N (Segment 2).

The presence of occupational and postoccupational alluvial deposits indicates that water moved through this large terraced area both during and after the occupation of the pueblo, probably directed there by the two northern drainages. Runoff likely would have deposited sediment behind the walls and might have helped to irrigate the terraces after they were established.

Before Crow Canyon's excavations began at the site, a Native American consultant from Hopi remarked that Nonstructure 1-N would have been an excellent garden area (Varien et al. 1995*1:8). The presence of stone alignments and stratigraphic deposits that appear to be consistent with agricultural activity supports the hypothesis that this area might have been a garden or small field. The stone alignments appear to have been constructed to reduce erosion and promote the accumulation of sediment. The naturally deposited sediments might have been supplemented by intentionally deposited fill.

Plant and Animal Resources

The results of the analysis of plant remains recovered from Woods Canyon Pueblo (see "Archaeobotanical Remains") provide information about the diet, agricultural practices, and wood-procurement strategies of the village inhabitants. The abundance of maize (corn) in the assemblage clearly documents the importance of this plant as a food source. Even though beans and squash were not recovered in the analyzed samples from Woods Canyon, they were staples of the Pueblo diet in ancient times and likely were consumed by the Woods Canyon residents as well. The presence of maize stalks and cobs in the assemblage suggests that the agricultural fields were close enough to the village that fairly large portions of the plants—not just the kernels—could be transported to the site. Local wild plants, including chenopodium-amaranth and groundcherry, were also collected by the inhabitants of the pueblo. In "Archaeobotanical Remains," Rainey and Jezik argue that the high proportions of maize and weedy plant remains, relative to less desirable or higher-cost foods such as wolfberry and cactus fruit, suggest that food was relatively plentiful. Juniper, pine, and sagebrush were the most common types of fuelwood identified in the Woods Canyon Pueblo assemblage, and juniper and some pinyon were used to construct kiva roofs.

The results of animal bone analysis are reported in "Faunal Remains." For the most part, the various animals represented by the remains were consumed as food. The assemblage indicates that the inhabitants of the pueblo relied on turkey and rabbits (cottontails and jackrabbits) as their primary sources of animal protein and fat. Turkey remains are more than twice as abundant as rabbit (lagomorph) remains. Artiodactyl (deer, sheep, or pronghorn antelope) remains are extremely rare at the site. The low percentage of artiodactyl remains is typical of many Pueblo III sites in the Mesa Verde region (Driver 2002*1). In addition, there are other sites in the region where turkey remains outnumber rabbit remains by at least two to one and artiodactyl remains make up 1 percent or less of the faunal assemblages, which leads Driver to infer that people were not hunting large game (see discussion in "Faunal Remains"). Comparisons of faunal remains from the four major sections of the site did not reveal any significant differences in species represented. This suggests that use of faunal resources was similar across the site and throughout the occupation.

The human skeletal remains examined from Woods Canyon Pueblo indicate that the residents were relatively well nourished and healthy (see "Human Skeletal Remains"). The people were average to above average in height and exhibited a low incidence of anemia. Dental enamel hypoplasia, an indicator of malnutrition and/or infection, was pervasive but not pronounced. Concurrent stress markers indicative of serious or prolonged episodes of stress were absent from the individual skeletal remains.


The people of Woods Canyon Pueblo obtained water from seeps and springs in the canyon bottom and supplemented their supply with water from a nearby reservoir. They controlled water runoff by building checkdams in the main drainage, in peripheral drainages, and on steep slopes of the site. The residents probably grew crops on the nearby mesa tops; they also might have planted gardens and small fields within the pueblo, especially in level areas at the bases of the steep slopes and on terraces that formed behind checkdams. Botanical and faunal evidence points to a diet of domesticated plants (maize) and domesticated animals (turkeys), supplemented by a variety of wild plants and animals (primarily rabbits). And, if the small sample of human skeletal remains analyzed during test excavations is representative of the population of the village as a whole, then the people of Woods Canyon were a relatively well fed and healthy group.

References cited | To borrow, cite, or request permission

Copyright © 2002 by Crow Canyon Archaeological Center. All rights reserved.