Go to Table of Contents.
About This Publication
List of Tables
List of Illustrations
Research Design
Population Estimates
Faunal Remains
Archaeobotanical Remains
Human Skeletal Remains
Rock Art
Yellow Jacket Pueblo as Community Center

Archaeobotanical Remains

by Shawn S. Murray and Nicole D. Jackman-Craig


Archaeobotanical remains were collected from a variety of contexts during Crow Canyon's test excavations at Yellow Jacket Pueblo (Site 5MT5). The analyses presented in this chapter were undertaken specifically to provide insight into the use of plants by the prehistoric inhabitants for food, fuel, and construction. In addition, the results of archaeobotanical analysis are used to infer the seasons of site occupation, the proximity of agricultural fields, and the nature of the past environment, as well as to address questions of resource depletion and food stress. The archaeobotanical data and interpretations presented in this chapter are based on the analysis of reproductive plant parts and wood charcoal from a total of 47 flotation and 444 macrofossil samples. The temporal framework used in the discussion section is from Kuckelman and Ortman's reconstruction of the history of site occupation as presented in "Chronology."

The plant remains from Yellow Jacket Pueblo were recovered in one of three ways: (1) collection by hand of plant remains found during excavation, (2) screening of sediment through ¼-in mesh in the field, and (3) water-processing of flotation samples in the laboratory. Macrofossils (that is, larger plant remains such as charcoal, maize [corn], and beans) were generally collected using the first two methods, whereas microfossils (smaller seeds and charcoal bits) were recovered through flotation. Together, these three methods of collection provide a better representation of plant remains at a site than does any single method used alone.


Flotation Samples


Ninety-five flotation samples were collected during Crow Canyon's excavations at Yellow Jacket Pueblo; of these, 47 were selected for processing and analysis. Because analyzing flotation samples of standard, or nearly standard, original volume reduces the chance that sample size will bias the results, we measured all selected samples and, when there was sufficient volume, processed 1 liter of sediment for each. For the majority (64 percent) of the 47 flotation samples reported here, it was possible to process 1 liter of sediment; any sediment in excess of 1 liter was curated. Twenty-six percent of all samples processed were ¾ liter in volume, and the rest (10 percent) contained less than ¾ liter of sediment. All samples of less than 1 liter were processed in their entirety.

The flotation process uses water to separate lighter, more buoyant organic remains, called the "light fraction," from heavier organic and inorganic materials, or the "heavy fraction." The measured sediment of each flotation sample was poured into a 5-liter bucket of lukewarm water and gently agitated by hand or with a wooden stirring rod. The organic matter that floated to the surface was slowly poured onto a flotation cloth (mesh size 0.355 mm) nested inside a screen. The bucket was again filled with water, the sample agitated, and the organic remains poured off. This process was repeated a minimum of three times or until no organic matter remained floating on the surface of the water. The light fraction collected in this manner was then allowed to dry in the flotation cloth, away from direct sunlight, for a minimum of two days. When dry, the light fraction was gently sorted through a series of geological sieves (mesh sizes 4.75 mm, 2.80 mm, 1.40 mm, 0.71 mm, 0.25 mm, and less than 0.25 mm). Sorting of flotation light fractions serves two purposes: (1) it allows us to sort taxa by size and (2) it reduces eye strain by minimizing changes in eye focal distance for analysts examining plant remains through a microscope (Bohrer and Adams 1977*1:40). Although the flotation cloth consists of 0.355-mm mesh, particles smaller than 0.355 mm sometimes adhere to larger particles when wet. Once dry, these smaller materials separate from the larger particles and are caught in either the 0.25-mm sieve or the sieve smaller than 0.25 mm. Materials smaller than 0.25 mm are not examined, because seeds of this size generally do not occur in the region, and because tiny, fragmented items are difficult to identify. After the light fraction was sorted by size, each Yellow Jacket sample was labeled and bagged according to provenience and particle size. The sediment remaining at the bottom of the bucket, the heavy fraction, was also dried thoroughly, then bagged, labeled, and stored.


All materials caught in the largest sieve (4.75 mm) were examined. Remains of this size are relatively easy to sort and may include unique items. For the remainder of the particle sizes, a subsampling strategy was employed to ensure that the analyst identified the largest number of taxa without having to completely examine all materials within each particle-size group. This subsampling strategy assumes that taxa are randomly distributed within each particle-size group and that identifying the number of different taxa present in a sample provides reliable, useful information about ancient plant use (Adams 1993*1; Popper 1988*1). Somewhat different subsampling strategies were applied to the two categories of flotation materials: (1) charred reproductive plant parts and other nonwood materials and (2) wood charcoal (the term "charcoal" is used throughout this chapter to indicate the burned remains of, specifically, the woody parts of plants).

When analyzing charred reproductive parts and other nonwood materials, we use a sample's own taxon diversity to decide when the sorting of a particular particle-size group is complete. This strategy was adapted from the "species area curve" concept developed by ecologists (see Mueller-Dombois and Ellenberg 1974*1:52–53; Pianka 1974*1). The species area curve plots the curve of the increasing number of taxa detected in a sample against the number of standard subsample volumes examined. The curve flattens when no new taxa are recovered in succeeding, standardized subsamples; it is at this point that we stop examination of that particle-size group (see Adams 1993*1).

We use this approach when analyzing materials with particle sizes smaller than 4.75 mm. A standard subsample has been defined by Bohrer and Adams (1977*1:40) to be "the volume of material that can be packed, but not piled, contiguously under a [microscope's] field of view for each particle size." For open-air archaeological sites in the Southwest, standard volumes have been developed for each particle-size group analyzed at 7X magnification: 1.8 ml for particles larger than 2.80 mm, 0.9 ml for particles larger than 1.40 mm, 0.4 ml for particles larger than 0.71 mm, and 0.3 ml for particles smaller than 0.71 mm (Adams 1993*1). For each of these particle-size groups, the first 6 ml is analyzed in subsamples of the stated volumes (for example, four 1.8-ml subsamples of the 2.80-mm particle-size group, seven 0.9-ml subsamples of the 1.40-mm particle-size group, and so on); any given particle-size group that measures less than 6 ml is sorted in its entirety. If new taxa are still being discovered at the end of the first 6-ml sort, we continue sorting until three consecutive subsamples have yielded no new taxa. It is likely that this sampling scheme results in our failure to detect some rare taxa.

A standard subsampling strategy was also developed for the analysis of wood charcoal. Here, we examine 20 pieces of charcoal from the 4.75-mm and 2.80-mm particle-size groups. Charcoal pieces are selected on the basis of their potential to be identified (for example, they are of sufficient size to allow identification) and on their varying appearances, so that as many different kinds of charcoal as possible are included. If fewer than 20 pieces are available in the 4.75-mm particle-size group, the remaining number are taken from the 2.80-mm group. Wood charcoal smaller than 2.80 mm is not analyzed, because reliable recognition of defining characteristics on such small specimens is difficult. Each charcoal specimen is identified by examining the cross section, but not the tangential or radial sections; this results in quicker analysis, with identification usually to genus level only (Bohrer 1986*1). In most cases, it is necessary to break the archaeological specimens to obtain clean cross sections. The cross sections are examined under magnifications of 20X to 45X. Identification of all archaeobotanical specimens is based on direct comparison to a collection of charred and uncharred plants obtained within the Mesa Verde region.

Macrofossil Sample Analysis

The macrofossil samples from Yellow Jacket Pueblo were first scanned for any charred nonwood items such as fruit or seeds, maize kernels or cob parts, and bean cotyledons. When present, these items were segregated and analyzed. Then, as for flotation samples, a subsampling strategy was employed for the wood charcoal. Twenty pieces of wood charcoal were selected and analyzed, unless fewer than 20 pieces were present, in which case all were examined. The same methods of identification were used for the macrofossil wood charcoal as were used for wood charcoal present in flotation samples.

Identification to Type

The reference flora for common and scientific plant names used in this chapter is A Utah Flora, by Welsh et al. (1987*1). Because taxa indigenous to the Southwest are sometimes very similar to one another in appearance and therefore are difficult to distinguish, we append the word "type" to most of our family-, genus-, and species-level identifications. This signifies that the specimen closely resembles the taxon named, but that other taxa in the area might also have similar-looking parts (see Adams 1993*1:197). When the designation "type" is used, it is understood that it encompasses all identifications made to the taxonomic level in question and to those below that level as well. For example, "Cercocarpus-type" includes all items identified to the genus Cercocarpus and to the species Cercocarpus montanus. For ease of use, the word "type" is designated only in the tables of this chapter; it is omitted from text references to the same taxa. The one exception to the rule is maize (Zea mays), which is easy to identify and occurs as only one species and therefore is never reported as "type."

Finally, some archaeobotanical specimens have characteristics that resemble two closely related genera—for example, the wood charcoal of either chokecherry or rose. In such cases, the genus names of both plants, separated by a slash, are used to indicate that a finer level of identification could not be achieved. In the foregoing example, the scientific name would be given as Prunus/Rosa in text and as Prunus/Rosa-type in tables.


A large number of flotation and macrofossil samples (n = 107) were collected from secondary refuse (materials discarded away from their area of use) found in various pits and middens, in subterranean structures, and on extramural surfaces (Table 1). Fewer, but more interpretable, samples (n = 24) came from de facto (intentionally left, but useable, materials) and primary (material remaining in its area of use) refuse deposits in hearths and on intact surfaces. De facto and primary refuse deposits are believed to preserve the best record of plant use by humans, and it is on samples from these contexts that much of the following discussion will focus. The majority of macrofossil samples, together with some flotation samples (n = 325), were recovered from mixed deposits, construction deposits, and collapsed structural deposits, as well as deposits disturbed in modern times by nonprofessional digging. Because remains from disturbed deposits are less interpretable than those from undisturbed deposits, these samples and contexts will not be discussed further.

Charred remains of a minimum of 35 plant taxa were identified in the analyzed archaeobotanical assemblage from Yellow Jacket Pueblo (Table 2). This estimate was derived by conservatively combining taxonomic levels, when appropriate. For example, for all analyses that follow in this chapter, Pinus edulis bark scales, charcoal, and cone scales are counted as a single taxon, and the same is done for all Pinus ponderosa parts. It is reasonable to assume that Pinus remains not identified to the level of species likely represent either Pinus edulis or Pinus ponderosa. Therefore, for purposes of assessing taxon diversity in a given context, Pinus remains found in association with P. edulis remains are counted as P. edulis, and those found in association with P. ponderosa remains are counted as P. ponderosa. Pinus is counted as a separate taxon only when its remains are found in association with the remains of both, or neither, of the two named species.

At least 18 of the 35 taxa in the Yellow Jacket Pueblo archaeobotanical assemblage are present as wood charcoal, whereas 23 taxa were recovered in various nonwood forms. The diversity of plant parts represented within some taxa leads us to infer that many of these plants had multiple uses. For example, maize kernels were eaten and cobs were used as fuel. It is also probable that some of these taxa entered the archaeological record incidentally, especially as macrofossils—for example, cones likely were carried in on tree branches that were burned as fuel, and other items may have been introduced into site deposits by animals or wind. It is for these reasons and others that archaeological plant remains rarely, if ever, provide an unambiguous record of prehistoric plant use (Minnis 1981*1:143).

Uncharred plant specimens were also recovered in the flotation and macrofossil samples at Yellow Jacket Pueblo (Table 2 and Table 3). Although Minnis (1981*1) states that, under most conditions of preservation, unburned organic materials degrade within about a century, it is possible at many ancient Pueblo sites to find unburned plant remains and partly burned construction beams preserved by extraordinarily arid and protective circumstances. Unburned remains may be inferred to be ancient if they occur in contexts that are clearly ancient: numerous seeds, for example, in a basket within a sealed context, or partly burned or unburned wood that is contained in an intact constructional deposit. However, because it is often difficult to distinguish ancient uncharred plant remains from more-recent uncharred remains, we usually consider unburned remains to be modern. We have therefore excluded from our discussion three plant taxa (Lappula redowski, Marrubium vulgare [introduced from Europe], and Rumex) that occur only in an uncharred state in the Yellow Jacket assemblage, as well as several occurrences of unburned wood and seed types that were also found in charred condition. Partly burned wood, however, we consider to be ancient.

The complete archaeobotanical database for Yellow Jacket Pueblo may be accessed through Crow Canyon's research database. Two additional on-line publications provide further information relevant to the analysis and interpretation of plant remains recovered from sites excavated by Crow Canyon. The criteria used to identify the various plant taxa and parts are described in the Plant Identification Criteria, by Shawn Murray and Karen Adams, and ethnographically documented uses of various plants identified in the assemblages are reported in the Ethnographic Uses of Plants, by Katharine Rainey and Karen Adams.



The ancient plant remains found at Yellow Jacket Pueblo indicate that the inhabitants grew and consumed the domesticated plants maize (Zea mays), beans (Phaseolus vulgaris), and squash (Cucurbita), but also exploited wild plants such as cheno-ams (Chenopodium and/or Amaranthus) and various other small-seeded plants (Table 4). Much of the evidence for food preparation and cooking was found in primary (16 samples) and de facto (three samples) refuse in hearths and other pits, features typically associated with cooking. Additional food refuse was recovered as secondary refuse in middens (eight samples). Some de facto refuse came from collapsed roofing material; maize was probably dried and processed for consumption on some roofs. Other contexts contained a mix of food and other debris, but because of mixing, these contexts are excluded from further discussion.

Maize was a food item found often at Yellow Jacket Pueblo. Maize was most commonly recovered in the form of cupules, the dense, cup-shaped structures that hold two kernels within their associated bracts. The cupule is the toughest, and usually the best-preserved, part of the cob, which probably explains why so many were recovered at Yellow Jacket Pueblo.

Maize cupules and larger cob parts were found in many samples from primary and de facto refuse (Table 5). In particular, they were found in the collapsed roofing material of Structure 1214, a bi-wall room of the great tower. The maize was probably refuse from a hearth in Structure 1222, the room above Structure 1214, that fell into the roofing material of Structure 1214 when the upper story collapsed. Maize cupules and cob parts were also found in several hearths and pit features in Architectural Blocks 200 and 2600, and in Structure 1201 (the bi-wall tower kiva). In these hearths, cobs were probably used as fuel after the kernels were removed, a practice observed in historic pueblos (see Elmore 1944*1; Robbins et al. 1916*1). The presence of cob parts in two pit features (Features 1 and 2) on an extramural surface (Nonstructure 2606) may indicate maize storage or processing; alternatively, the cobs might have been refuse that was discarded in the pits. In Feature 1, a complete bowl containing maize cupules and charcoal was also found. Numerous other cobs were contained in various midden deposits at the site and in several deposits of mixed refuse.

In contrast to cupules and cob fragments, maize parts such as kernels, shank (the short stem below the ear) parts, and main stalk (stem) fragments were found in low numbers and in few contexts. The small number of shank and stalk parts could indicate that mostly ears (cobs with kernels) were brought to the village, with the remainder of the plant generally being left in the fields. This practice would have reduced transport costs by minimizing weight and the handling of unwieldy plant parts. Alternatively, it is possible that these less-sturdy stalks and stems were transported to the village but were used as fuel in hearths. This latter possibility might explain the presence of several charred, but disfigured, plant parts that were identifiable only as monocotyledons (see paragraphs 29-33).

Seeds and cotyledons (half seeds) of domesticated beans (Phaseolus vulgaris) were found in seven samples. The majority of these bean parts were recovered from midden deposits in Structures 704 and 903 and in Nonstructure 102. The collapsed roofing material in Structure 1214 contained beans, together with maize and cheno-ams, in deposits interpreted as de facto refuse. As discussed in paragraph 19, Structure 1214 is the lower story of a two-story structure, and the beans probably fell from the second-story floor into the roofing material of the first-story room when the second-story structure collapsed. In general, beans are poorly represented in the archaeological record; they are less likely to be preserved than maize, because they are usually boiled before consumption, rather than roasted or parched, and this renders them soft and degradable.

Rind fragments of a member of the squash genus, Cucurbita, were the only other domesticated plant remains recovered at Yellow Jacket Pueblo. Two Cucurbita species, Cucurbita pepo and C. moschata, are edible squashes that have been found in sites in northern New Mexico dating from as early as 950 B.C. (Simmons 1986*1). The rind fragments recovered from midden deposits in Architectural Block 400 and in deposits above a disturbed burial pit (Feature 1, Nonstructure 603) were probably food refuse.

The inhabitants of Yellow Jacket Pueblo also relied on wild plants for a portion of their diet. Of the wild plant remains recovered, cheno-am seeds are the most ubiquitous, occurring in 27 samples. Cheno-am seeds are so-called because of the difficulty in distinguishing the tiny, black, spherical seeds of Chenopodium from those of Amaranthus, especially when they are charred and degraded. Cheno-am seeds were recovered from several samples from primary and de facto refuse. In particular, cheno-am seeds were found with maize cobs, beans, and numerous types of wood charcoal in de facto refuse in roof-fall deposits in bi-wall Structure 1214 (part of the great tower). Cheno-am seeds were also preserved in the hearth of Structure 1201, which is the oversize kiva in the great tower. This hearth contained abundant maize parts, many types of wild seeds, and a wide range of charcoal types, all of which suggest that this hearth was used for a variety of purposes, including cooking. Several other features outside the great tower complex also contained cheno-am seeds. Cheno-ams in de facto refuse were associated with Nonstructure 506, an extramural surface, and were also found in the primary refuse of three firepits in Architectural Blocks 200 and 2600.

It is likely that cheno-am seeds were an important wild food resource for the inhabitants of Yellow Jacket Pueblo. There are extensive postcontact records of Southwestern groups harvesting Chenopodium and Amaranthus seeds for consumption (Adams 1988*2). In postcontact times, the leaves were boiled and eaten as greens (summer and fall); the seeds were harvested (late summer and fall) and ground into flour for mush (Adams 1988*2:204).

A variety of charred wild seeds was consistently found in thermal features and middens across the site, despite the substantial disturbance of some of these contexts. Seeds or fruit of the most commonly recovered wild genera—cheno-am, groundcherry (Physalis), datil yucca (Yucca baccata), purslane (Portulaca), bulrush (Scirpus), ricegrass (Stipa hymenoides), and big sagebrush (Artemisia tridentata)—were found in hearths, firepits, pit features, and middens more often than in any other context. This pattern suggests that these seeds were the remains of food-related activities, rather than remains of plants introduced incidentally during occupation, though the sagebrush achenes might have entered on branches brought in as fuel.

Several other types of wild seeds were recovered from only a few samples each. In a hearth (Feature 7) in Structure 1201 (the oversize kiva in the great tower), hedgehog cactus (Echinocereus fendleri) seeds and oak (Quercus) nutshell were found, along with the slightly more common purslane, groundcherry, and bulrush, suggesting that this kiva hearth was used for several purposes, including cooking. The seeds of many other taxa were preserved in only one sample each (Table 4). It is difficult to know whether these rare plants were used by the inhabitants of the village or were introduced accidentally into the Yellow Jacket Pueblo deposits. Most of these taxa have known ethnographic uses and might have been used by the inhabitants, but they will not be discussed further in this chapter.

Inferring Seasonality

It is highly unlikely that Yellow Jacket Pueblo was occupied only seasonally—stored food resources could have permitted year-round occupation. Nonetheless, it is important to identify those specific seasons for which we have direct archaeobotanical evidence of human presence at the site. Individual seasons of occupation can be inferred by identifying the times of the year when the plant parts present in the assemblage would likely have been available (Adams and Bohrer 1998*1). Many of the plant parts found at Yellow Jacket Pueblo were in fruiting form, which allows us to make inferences about the season in which they were gathered and, by extension, the seasons when areas of Yellow Jacket Pueblo were occupied. This strategy, however, has its limitations. First, differences in elevation, latitude, and rainfall can make it difficult to accurately predict when fruits will mature at a given location. Second, fruits are often gathered and stored for later consumption; therefore, the presence of fruits, though an indicator of season of harvest, does not necessarily indicate the season of use (Adams and Bohrer 1998*1). To infer seasonality, it is best to use plants with short and predictable fruiting seasons, as observed in the fruiting of modern plants in the vicinity of the site. It is also important to understand that we cannot infer nonoccupation on the basis of plant remains, in part because some seasons (most particularly, winter) have few, if any, plant resources available for harvest.

At a minimum, the inhabitants of Yellow Jacket Pueblo were present during the late spring sowing and fall harvesting of maize, and one can argue that they were likely present throughout the summer to protect and weed their crops (see Adams and Bohrer [1998*1] for a discussion of the scheduling requirements of growing maize in the Southwest, including a discussion of the implications for the presence of people at given locations during different seasons). We can also infer seasonality by the presence of other plant parts recovered from the site, including ricegrass caryopses (grains), which usually ripen in late spring to early summer, and groundcherry, purslane, cheno-am, bulrush, and datil yucca seeds, which are available for harvest mid- to late summer and into the fall (Adams 1988*2, 1993*1). The occurrence, in two firepits (in Nonstructures 904 and 2601), of remains of at least three of the above-mentioned, mid- to late-summer-fruiting species, suggests that these firepits were used during the summer months. Similarly, the presence of many of these seed types in the hearth of Structure 1201, the oversize kiva in the great tower, implies that this hearth was last used in mid- to late summer or early fall. Because Structure 1201 is thought to have been one of the last structures occupied at Yellow Jacket Pueblo, this same season might have been the season of last use at the site.

Fuels and Construction Materials

Burned wood or woody plant parts account for the greatest quantity of archaeobotanical remains found at Yellow Jacket Pueblo—at least 15 different trees, shrubs, and other plant types are represented (Table 5). Although some of this charcoal is the remains of wood that was originally brought into the village for construction and then was eventually burned, either intentionally or accidentally, most charcoal at the site probably derives from fuelwood. Inferences regarding fuel choice were drawn from charcoal found in hearths, ashpits, firepits, and middens. By far the most ubiquitous fuelwood was Juniperus, which was recovered in 43 of 84 (51 percent) of these contexts. In thermal features alone, the recovery rate (58 percent) was similar. In particular, juniper charcoal was found in de facto refuse in a pit feature (Nonstructure 2606, Feature 1) and as primary refuse in three firepits (in Nonstructures 203, 2601, and 2605) and two hearths (Structures 1201, Feature 7, and Structure 1217, Feature 1). Other woods commonly used for fuel were sagebrush (Artemisia), pine (Pinus), oak (Quercus), serviceberry/peraphyllum (Amelanchier/Peraphyllum), and mountain mahogany (Cercocarpus). Wood types more rarely recovered from the site include rabbitbrush (Chrysothamnus), cottonwood/willow (Populus/Salix), cliff-rose/bitterbrush (Purshia), wolfberry (Lycium(1)), Mormon tea (Ephedra), saltbush (Atriplex), and chokecherry/rose (Prunus/Rosa). Most of these trees and shrubs are common on the landscape around Yellow Jacket Pueblo today.

In this discussion of wood fuels we also include the cupules, cob fragments and segments, stalk segments, and stem sections (culm) of Zea mays (see Table 5). Although the kernels and kernel embryos were probably consumed as food, the remainder of the cob might have been used as fuel or tinder (see Elmore 1944*1; Robbins et al. 1916*1). Maize cob fragments, cob segments, and cupules were found in 48 of 120 contexts at Yellow Jacket Pueblo. Many of these contexts were middens—both secondary refuse and recently disturbed deposits (28 samples). Primary refuse from hearths, ashpits, and firepits (nine samples) also contained charred cupules, cob fragments, and cob segments, as did some roof-fall and wall-fall contexts (11 samples). Although the presence of maize parts in roof fall and wall fall probably resulted from the processing of maize on rooftops, the hearths and middens likely contained maize refuse from the burning of cobs as fuel. Repeated cleaning of the hearths would have resulted in the buildup of cob parts in midden deposits. It is also possible that the cobs were consumed as food during periods of food stress (see Buskirk 1986*1). Evidence of cob consumption as food might be seen archaeologically as an abrupt decrease in the number of cob parts recovered, accompanied by a continued presence of edible and nonedible (for example, shank and stalk) parts. Evidence from human coprolites, absent at Yellow Jacket, would best address such an argument (Minnis 1991*1).

The most useful contexts for documenting the use of plant materials in construction are roof fall and wall fall. These two contexts were heavily sampled in the great tower complex (Architectural Block 1200), and most of this discussion is relevant only to that block. Overall, juniper (Juniperus) was the most commonly recovered construction wood, occurring in 56 percent of the analyzed samples from roof fall and wall fall (Table 5), and it was the preferred wood in most of the structures tested in the great tower complex. Sagebrush (Artemisia) and pine (Pinus) were found in 36 percent and 25 percent, respectively, of the analyzed roof-fall and wall-fall samples. The presence of these species indicates that pinyon-juniper and sagebrush woodland existed nearby. The presence of pine bark scales and juniper twigs also suggests that the trees were located not far from the village, as unnecessary parts would likely have been stripped had the logs been transported long distances.

Charred wood from trees that are absent from the modern landscape surrounding Yellow Jacket Pueblo were found in five structural contexts in the great tower complex. Ponderosa pine (Pinus ponderosa) and Douglas fir (Pseudotsuga menziesii) are both tall, straight trees; beams from these types of trees were found in the fill of three kivas (Structures 1201, 1206, and 1209) and two rooms (Structures 1208 and 1213), but were not found in hearth, roof-fall deposits, or wall-fall debris. Rather, these specimens were found in fill redeposited during excavations by the Museum of Western State College in 1931; thus it is possible that they were from construction beams that became mixed with other fill. Ponderosa pine and Douglas fir have been recovered from structures at sites on Mesa Verde, where they grow today in the higher or cooler elevations. Large stands are currently available on Sleeping Ute Mountain, 25 km south of Yellow Jacket Pueblo (Adams 1993*1). Smaller stands are also found in the Dolores River canyon, about 8 km from Yellow Jacket Pueblo; it is likely that ponderosa pine trees also grow in the cooler sections of Yellow Jacket Canyon. It is possible that, in the past, ponderosa logs that floated down the Dolores River during high rains and floods were carried overland to Yellow Jacket Pueblo for use as construction beams.

Construction needs might also have been met by other woods. Oak (Quercus) and serviceberry/peraphyllum (Amelanchier/Peraphyllum) were relatively common in roof-fall contexts; mountain mahogany (Cercocarpus) and cottonwood/willow (Populus/Salix) were found in only a few contexts. Additional information about wood use at Yellow Jacket Pueblo is presented in the following discussions of plant remains found in different kinds of deposits.

Plants in Selected Contexts with Primary and De Facto Refuse

This section focuses on plant remains found in flotation and macrofossil samples from selected hearths, pit features (including firepits), and extramural surfaces with primary and de facto refuse. Analysis of these remains provides evidence that may be used to infer feature and structure use (Table 6 and Table 7). It is important to note that materials found in these features are more likely to represent the last use or uses, rather than the longer history and varied plant uses represented by the remains found in middens. Hearths and firepits were periodically cleaned out, though perhaps incompletely, such that it is unlikely one would find evidence of all the different activities that may have been associated with those features throughout their entire use lives.

The hearth (Feature 7) of Structure 1201, the oversize kiva in the great tower, was deep, circular, and lined with masonry, and it contained at least three distinct strata. The two lower strata (Strata 3 and 4) in this hearth apparently had been left intact when overlying layers of ash were periodically cleaned from the pit; these lower strata thus represented earlier uses of the hearth than did the upper fill. The lowermost (earliest) layer (Stratum 4) contained various types of seeds, such as hedgehog cactus (Echinocereus fendleri), groundcherry (Physalis), cheno-am (Chenopodium/Amaranthus), purslane (Portulaca), and bulrush (Scirpus), but only three types of wood—juniper (Juniperus), sagebrush (Artemisia), and an unknown, diffuse porous-type wood. Stratum 3 contained some fine sand and four types of seeds: cheno-am, purslane, and two unknown types. Six types of charcoal were found in this stratum: serviceberry/peraphyllum (Amelanchier/Peraphyllum), sagebrush, mountain mahogany (Cercocarpus), pine (Pinus), cliff-rose/bitterbrush (Purshia), and oak (Quercus). The only reproductive parts from the uppermost stratum of ash (Stratum 2) were oak nutshell and an unknown type of nutshell. However, more charcoal types were recovered from this level than from any other stratum in this hearth: juniper, chokecherry/rose (Prunus/Rosa), and maize (Zea mays) cob parts were identified, in addition to the charcoal types found in the earlier levels. Given the quantity and diversity of seeds recovered, we infer that this hearth was used, at least in part, for food preparation, perhaps for boiling or parching seeds and for roasting of maize. The wide range of charcoal recovered may indicate that this hearth had a special use or function.

Southwest of Structure 1201 was an extramural surface (Nonstructure 1217) containing a very steep-sided hearth (Feature 1) that was partly lined with rock. The excavated portion of this feature contained primary refuse of sagebrush and juniper charcoal. A single seed of unknown type was also found. The lack of reproductive parts recovered leads us to infer that the last use of this hearth was probably for heat and light, rather than for cooking food. This does not rule out use of this possible communal feature in activities that were unrelated to plants, such as meat preparation or pottery firing—activities best evaluated by examining other data sets.

In Architectural Block 200, an extramural surface (Nonstructure 203) that contained a firepit (Feature 1) was exposed. This surface was used sometime after A.D. 1180. The firepit was shallow (a maximum depth of 10 cm) but contained seeds of cheno-am, groundcherry, and globemallow (Sphaeralcea). The wood charcoal was identified as juniper, pine, and oak. The last use of this firepit is likely to have been for the preparation of food, but the fire would have also provided heat and light.

Nonstructure 506 was an extramural surface (with a pit feature and an intact jar) used for an unknown length of time after A.D. 1100. De facto refuse containing cheno-am seeds, maize cupules, and serviceberry/peraphyllum charcoal was associated with the surface, and primary refuse containing juniper and pine charcoal was located just above the surface. Although no thermal features were found within the excavation unit, one or more could have been present nearby. Alternatively, this surface could have been used as a work space for processing food items.

Nonstructure 2403 was an extramural surface located just north of the roomblock in Architectural Block 2400. This surface was probably used sometime after A.D. 1150. No hearth or firepit was found within the excavation unit, but on the surface was primary refuse containing sagebrush, juniper, and cottonwood/willow (Populus/Salix) charcoal. No seeds or other reproductive plant parts were identified in the sample collected from this surface. It is possible that the charcoal was either secondary refuse or the partial remains of a collapsed roof from an adjacent burned structure.

Two firepits and one pit feature containing primary and de facto refuse were found in Architectural Block 2600 and date from sometime after A.D. 1100. A large firepit (Feature 1), which showed archaeological evidence of heavy use and contained primary refuse, had been excavated into an extramural surface (Nonstructure 2601). In the lower portion of the firepit were three varieties of seeds—groundcherry, cheno-am, and ricegrass (Stipa hymenoides). The firepit also contained sagebrush, rabbitbrush (Chrysothamnus), Mormon tea (Ephedra), juniper, and cottonwood/willow charcoal, as well as abundant maize parts ranging from cupule and cob fragments to shank segments. A firepit (Feature 1) on Nonstructure 2605, another extramural surface, was not carefully constructed and apparently was used very little. The primary refuse in this shallow firepit contained purslane and cheno-am seeds, and serviceberry/peraphyllum, juniper, wolfberry (Lycium), and cottonwood/willow charcoal, as well as maize cupules. A pit feature (Feature 1) with no evidence of thermal alteration had been excavated into an extramural surface (Nonstructure 2606). Inside this pit was a complete pottery vessel containing maize cupules and serviceberry/peraphyllum charcoal interpreted as possible primary refuse. Surrounding the bowl was construction fill containing primarily juniper and sagebrush charcoal. The pit apparently had been dug before the construction of the roomblock in Architectural Block 2600 and might have been used for storage or refuse disposal.

Our interpretation of the recovered plant remains suggests slightly different uses for the three documented nonstructures in Architectural Block 2600. Because of evidence of heavy use and numerous types of charred plant parts recovered from the firepit of Nonstructure 2601, we suggest that this pit was probably used many times for food preparation and to provide heat and light. The firepit associated with Nonstructure 2605 was probably used only a few times for cooking and other activities, as evidenced by the shallowness of the pit and the apparent expedient construction. The purpose of the pit feature associated with Nonstructure 2606 is more difficult to ascertain. The maize and wood charcoal in the vessel may indicate ritual or ceremonial activities, or these remains may have been refuse deposited in the bowl after use.

Patterning of plant species among these features is not clear; however, some taxa are more common than others. If one considers the three stratigraphic levels of the hearth in Structure 1201 as a single context, maize was present in 75 percent of the hearths, firepits, and other pit features discussed in this section. Although some of the maize cob remains could have been leftover from the summer roasting of green ears for food, much of the maize refuse likely resulted from the burning of cobs as fuel. Cheno-ams, representing food refuse, were present in 63 percent of all features included in the study, whereas groundcherry seeds were present in 38 percent. All other seed types were found in only one or two contexts. The most commonly recovered charred wood in these features was juniper (found in 100 percent of the samples); other types were sagebrush (63 percent), serviceberry/peraphyllum (50 percent), pine (38 percent), and cottonwood/willow (38 percent).

The use of different excavation strategies makes it difficult to compare the presence or absence of taxa between the great tower complex and the other tested areas. Observed differences, such as the presence of several taxa in one part of the site but not in another, can probably be attributed to small sample size and to the different kinds of contexts sampled. Temporal comparisons between areas are also difficult to make because the chronological resolution of Yellow Jacket Pueblo is not especially fine-grained (see "Chronology"). Small sample size is also an obstacle to detecting changes in plant use over time at this site.

Overall, the uses of the thermal features discussed in this section seem oriented toward general activities such as food preparation and burning of wood for warmth and light. Although other possible functions might include ritual activity or medicinal preparation, these kinds of activities are difficult to identify from plant remains. No exotic or rare plant species were recovered from the excavations.

Plant Remains in Collapsed Roof and Wall Debris

Plant remains recovered from collapsed roofs and walls are good indicators of the kinds of wood used in construction and the kinds of activities, such as food processing, carried out on rooftops. These materials can also yield information about items that were suspended from interior or exterior beams. The contexts examined here include collapsed roof and wall deposits inside structures, as well as materials found on kiva benches and on some extramural surfaces (Table 8).

Our analysis indicates that certain species of trees were used more often than others to provide the heavy beams used in structural support at Yellow Jacket Pueblo. Juniper (Juniperus) was recovered as charcoal in many constructional contexts and was the main wood used for roof beams. Pine (Pinus) charcoal was found in Structure 704 (subterranean structure, type unknown) and in Structure 1214 (a bi-wall room of the great tower). In these structures, pine probably was used in conjunction with juniper, but not where strong, structural roof support was needed, as pine tends to fracture more easily than juniper.

In addition to the large, load-bearing beams used in roof construction, smaller plant materials that provided fill and support for the outer layers of plaster or adobe also would have been incorporated into the roofs of most structures. At Yellow Jacket Pueblo, it appears that these smaller materials most often consisted of smaller branches from shrubby plants like serviceberry/peraphyllum (Amelanchier/Peraphyllum), sagebrush (Artemisia), mountain mahogany (Cercocarpus), and occasionally rabbitbrush (Chrysothamnus). In sites throughout the Southwest, many roofs were lined with straight and narrow branches of willow (Salix) that were secured with strips of willow or yucca (Yucca) fiber (Gallagher 1977*1; Robbins et al. 1916*1; Whiting 1966*1). Cottonwood (Populus) and/or willow were found in only two structures (Structures 1214 and 704), and yucca was not recovered in any of the contexts discussed here.

Juniper charcoal and a maize (Zea mays) stalk segment were found in roofing debris that had collapsed, intact, onto the bench surface of a blocked-in, aboveground kiva (Structure 1204) in the great tower complex. This structure is thought to have been the lower story of a two-story structure. It is possible that the charcoal and stalk segment originated either in the roofing material of this structure or on the floor of the upper structure.

Above a bi-wall room (Structure 1214) of the great tower, there was probably a second story (Structure 1222) containing a heavily used hearth; there was no evidence that either structure burned (Kuckelman 1997*1). It was probably from the upper-story hearth that much of the charcoal in the fill of the lower room derived (numerous taxa, including Zea mays, were represented). West of Structure 1214 was an outdoor surface (Nonstructure 1219) that contained collapsed roof and wall debris. All the charcoal types in this context were also found in the samples from collapsed roof and wall material in Structure 1214, and they likely derive from wood that was originally part of that structure.

The collapsed roof and wall debris in Structure 204, a masonry room, contained several maize cob parts and a wide range of wood charcoal (five different taxa). The cob parts might have been discarded on the intact roof, or they might have been thrown into the collapsed roofing debris after the roof was dismantled. The five wood types (juniper, serviceberry/peraphyllum, sagebrush, mountain mahogany, and oak) recovered from this structure were probably part of the various layers of construction, though some of the charcoal might have been introduced as refuse.

Little is known about subterranean Structure 704. Tree-ring dating suggests that this structure was built sometime after A.D. 974. The types of pottery sherds found on the floor indicate that the structure was last used an unknown length of time after A.D. 1100. The collapsed roof and wall debris contained several different kinds of wood charcoal (probably roofing materials) and maize kernel and cob parts. The maize was probably dried and shelled on the roof, and the refuse discarded there, or the ears were, perhaps, suspended from the roof beams. Alternatively, any or all of this material could have been discarded into this depression as secondary refuse after the roof collapsed.

The collapsed roofing debris in Structure 903, an earth-walled pit structure, contained only maize cob fragments. As in Structure 704, these remains could have been refuse associated with the processing of maize on the rooftop, or they could have been refuse deposited after the roof collapsed. The lack of charcoal in the roof debris might suggest that the wood beams and support material were scavenged for use in other structures.

A handful of plant foods was also recovered from collapsed roof and wall contexts at Yellow Jacket Pueblo. Although maize parts were found in every context presented in Table 8, almost all other reproductive parts (fruits or seeds) were found only in Structure 1214, the great tower bi-wall room located below Structure 1222. Structure 1214 contained material—including, probably, the contents of a hearth—that had collapsed from Structure 1222 above. The uneven distribution of reproductive plant parts as revealed in Table 8 leads to several inferences: (1) that a hearth in Structure 1222 had been used for processing plants such as cheno-ams, beans (Phaseolus vulgaris), maize, purslane, and bulrush; (2) that the roofs of Structures 1204, 204, 704, and 903 might have been used for the processing of maize but not of other foods; (3) that Structure 1222 had some special use, perhaps as a public space for cooking or as a place for ritual or medicinal preparation; and (4) that perhaps Structures 1204, 204, 704, and 903 belonged to individual families rather than to a larger group. The last inference assumes that maize and structures were considered personal or family property and that families would have dried maize on their own rooftops more often than on public ones.

Proximity of Agricultural Fields

The proximity of agricultural fields to an ancient village can be inferred by the kinds of plant parts found in the village. Hypothetically, if fields were located far from a village, then parts that were heavy or bulky, such as maize shanks or stalks, would more often be removed and left behind in fields than carried back for processing. If, on the other hand, fields were located nearby, then the remains of these bulkier plant parts should be found at the site more often. Once such materials were transported to villages, they might have been burned as waste or used as tinder or fuel. Historically, the burning of cobs as fuel was common (see Elmore 1944*1; Robbins et al. 1916*1), and this activity no doubt also occurred in prehistory.

At Yellow Jacket Pueblo, maize shank and stalk parts were found in nine contexts, suggesting that the fields were located nearby. If only kernels and cobs had been recovered, we might infer transport over a longer distance, with waste products left in the fields. Proximity to fields is also suggested by Yellow Jacket Pueblo's location on a relatively flat point that is surrounded by arable land and by the location nearby of at least two reliable springs. The rolling uplands that surround the site on three sides would probably have provided enough farmland to support the occupants of this pueblo (Kristin Kuckelman, personal communication 1997). North and northeast of the site are well-drained sagebrush flats that would have been easy to access; second-choice locations for farming would have been across the draw to the west and across the canyon to the east. Although the canyons that border the site to the east, west, and south are rocky and steep in places, these second-choice locations are relatively easy to access. Possible third-choice farming locations would have included the more gradual slopes of the canyon walls and perhaps the talus slopes below the site, which could have been terraced by the people of Yellow Jacket Pueblo. The first- and second-choice locations are commercially farmed today.

One definite and four possible dams, as well as a large depression that probably was a reservoir, may have been used in part for agricultural purposes. The reservoir and one of the possible dams are located just north of Architectural Block 2400. At the west edge of the reservoir is a somewhat poorly constructed berm that was probably a dam built to keep runoff from draining out of the basin into the canyon. The reservoir appears at present to collect water only rarely, but this could be because the dam has been breached. The other four water-control features are located near the east end of the great tower complex. This series of one definite and three possible dams is associated with a natural drainage and with a spring that today is productive even in the driest of seasons (Kristin Kuckelman, personal communication 1997). Although some of this water might have been used to irrigate small garden plots located near the pueblo, most of it was probably used for domestic purposes.

Resource Depletion and Food Stress

In this section, we explore the possibility that resource depletion and food stress increased during the latter (late Pueblo III) part of the occupation of Yellow Jacket Pueblo. Kohler and Matthews (1988*1) proposed local forest depletion for early Pueblo sites in the Dolores River valley, resulting, perhaps, in greater residential mobility. Adams and Bowyer (2002*2) found minimal evidence of food stress when they examined the changes in plant-food choice by residents of the Sand Canyon locality in the century before the late-thirteenth-century regional depopulation.

Assessing changes in resource use for Yellow Jacket Pueblo is difficult because of the low chronological resolution, the disturbed nature of some of the archaeological deposits, and the different excavation strategies employed in different areas of the site (that is, the great tower complex vs. the other architectural blocks). It is possible, however, to broadly examine plant diversity within the flotation record, comparing structures dating from the late Pueblo II period with those in the late Pueblo III great tower complex (Table 9). These specific contexts, representing a diversity of depositional situations (for example, primary refuse, secondary refuse, and wall fall), were chosen because they are the ones most securely limited to the time periods in question. It is important to note that differences between the plant remains found in the great tower complex and those found in other structures or architectural blocks may not indicate changes over time in resource availability but, rather, differences in how various areas of the village were used. The great tower complex, in particular, might have been a ceremonial or special-use area. Therefore, activities conducted in this section of the village might have differed considerably from those conducted in the parts of the village dating from the late Pueblo II period. It is also possible that any dissimilarities in plant taxa between the two time periods are due to the different numbers of samples representing each of the time periods.

Differences in plant use between the late Pueblo II and late Pueblo III time periods at Yellow Jacket Pueblo are not dramatic. The reproductive plant parts in late Pueblo II and late Pueblo III contexts are similar, as is the diversity of charcoal. A ranking of the different taxa indicates that juniper (Juniperus) and sagebrush (Artemisia) were the two most commonly used woods throughout the tested areas of the village and for the duration of the occupation. This suggests that the availability of wood in the surrounding environment was probably not significantly different just before the last use of the site than when Yellow Jacket Pueblo was first established and that fuel use probably did not change much over time.

One potentially interesting difference between samples from contexts dating from the late Pueblo II and late Pueblo III periods is the noticeable drop in the number of late Pueblo III contexts containing maize (Zea mays) cob cupules, cob fragments, and kernels. Though cupules and cob fragments are, technically, maize reproductive parts, they can probably be considered as fuel because their presence in certain contexts is likely the result of intentional burning of maize cobs. This would suggest that either cobs were burned or disposed of more rarely in the great tower complex than in the late Pueblo II structures or the inhabitants of the great tower did not have as much access to maize late in the occupation (which might indicate that crops were poor). However, evidence of maize cupules, cob fragments, and cob segments is preserved in three macrofossil samples (not listed in Table 9) from the great tower complex (Structures 1201 and 1214), documenting some burning of cobs in hearths during the late Pueblo III period.

Although, overall, the results of this comparison of plant remains across time periods are inconclusive, the observed difference between the late Pueblo II and late Pueblo III structures in recovery of maize cob parts is interesting. These differences may represent changes in the availability or use of maize through time. Alternatively, they may indicate that the great tower complex was used in a different manner than were other parts of the village.

Season of Last Use

The types of remains found in hearths can indicate the season(s) when those hearths were last used. This evidence may be used to address the question of the season of final use of Yellow Jacket Pueblo. Tree-ring evidence indicates that the great tower complex was one of the last architectural blocks to be built at Yellow Jacket Pueblo and that it was constructed and used during the late Pueblo III period (see "Chronology"). The wide range of seed types found in one of the hearths (Feature 7, Structure 1201) sampled in the great tower complex represents more than one season (Table 6). Of the taxa listed in Table 6, groundcherry (Physalis) can be collected by midsummer, but its seeds are often not mature until later in the summer and early fall. The majority of plants identified in the samples from this hearth mature in late summer and fall. It therefore seems likely that this hearth in the great tower complex was last used sometime during the fall, although some of the plants or plant parts could have been stored from an earlier season. Ricegrass (Stipa hymenoides) is an early summer grain-producer, and its presence in a nonstructure firepit hints that, at some time, people were present in the area earlier in the growing season.

Nature of the Past Environment

The environment around Yellow Jacket Pueblo during the prehistoric occupation was probably very similar to what it is today. Although it is difficult to reconstruct the relative proportions of different plants on the ancient landscape, the species found in archaeological contexts at Yellow Jacket Pueblo closely approximate those found in the area of the site today. A notable exception is the presence of ponderosa pine (Pinus ponderosa) and Douglas fir (Pseudotsuga) in the archaeological record, as discussed previously. Although today much of the land surrounding the site is farmed, the site itself is covered with sagebrush (Artemisia) and oak (Quercus). Juniper (Juniperus) and pine (Pinus) are also present, as are patches of serviceberry/peraphyllum (Amelanchier /Peraphyllum), rabbitbrush (Chrysothamnus), and chokecherry (Prunus). Saltbush (Atriplex), cliff-rose (Purshia), and ephedra (Ephedra) are scattered along the canyon edge.

Weedy wild plants found in the ancient deposits of Yellow Jacket Pueblo are also common around the site today. Goosefoot (Chenopodium), pigweed (Amaranthus), globemallow (Sphaeralcea), purslane (Portulaca), and groundcherry (Physalis) grow particularly well in disturbed soils and were probably abundant prehistorically around habitation sites, especially if agricultural fields were near. We have discovered, through modern experimental gardening on the Crow Canyon Archaeological Center campus, that many of these species thrive in places where water collects, even in small amounts, and ancient farmers might have used various strategies to take advantage of this fact. Less-weedy plants such as datil yucca (Yucca baccata), hedgehog cactus (Echinocereus fendleri), needle-and-thread grass (Stipa comata), and ricegrass (Stipa hymenoides) are also widely found throughout the area, whereas the rushes (Scirpus, for example) prefer moister ground around springs or watercourses; many of these plants were available in the ancient environment as well.


In this chapter, we interpret the archaeobotanical remains found in 47 flotation and 444 macrofossil samples from Yellow Jacket Pueblo, the largest ancestral Pueblo site in the Mesa Verde region. We found that the inhabitants of Yellow Jacket Pueblo farmed maize, beans, and squash and collected many different types of wild plant foods, including cheno-ams, purslane, groundcherry, and yucca. Because the season of fruiting and harvesting of these plants is known, we can infer from the archaeobotanical record that ancestral Pueblo peoples were present at Yellow Jacket from at least late spring through early fall.

Juniper, sagebrush, pine, oak, serviceberry/peraphyllum, and, to a lesser extent, cottonwood/willow provided both fuel and construction materials to the people of Yellow Jacket Pueblo. Juniper was probably the main wood used for load-bearing beams. Woods used only occasionally by the inhabitants for fuel or construction were ponderosa pine, Douglas fir, mountain mahogany, and the woody parts of shrubby plants like cliff-rose/bitterbrush and rabbitbrush. The ponderosa pine and Douglas fir trees possibly grew some distance from the village, as they do today.

Analysis of primary and de facto refuse from hearths, firepits, and pit features revealed that most thermal features were probably used for parching, boiling, or other cooking of plant reproductive parts, mainly seeds. It is likely that these thermal features were also used for burning wood for heat and light. It appears that the hearth in the oversize kiva (Structure 1201) in the great tower complex served multiple purposes, including the preparation or cooking of edible seeds. This use contrasts with the use of a hearth on a surface (Nonstructure 1217) just outside Structure 1201 that contained no archaeobotanical evidence of food preparation. Other firepits (in Nonstructures 203, 2601, and 2605) are likely to have been used for cooking, heating, and lighting. Two surfaces (Nonstructures 506 and 2403) might have been work areas, and a pit feature (on Nonstructure 2606) outside the roomblock in Architectural Block 2600 may have been used as a storage pit. Samples from collapsed roof and wall deposits suggest that maize might have been processed on rooftops. The notable diversity of seeds found in a great tower bi-wall room (Structure 1214) may indicate that this room, or—more likely—the one above it (Structure 1222), served some special use, perhaps as a public space where people had access to a wide variety of foods.

On the basis of the site's location and the characteristics of the modern landscape, we believe that ancient agricultural fields were located close to Yellow Jacket Pueblo. We found no evidence for or against resource depletion or food stress over time, but our studies were hampered by small sample size and the lack of fine chronological resolution for the site. Finally, it is probable that the great tower complex was last used sometime in the fall, during or after a time when maize and many wild plants were harvested.

1The identification of wolfberry (Lycium) is tentative, based on characteristics that only generally resemble modern Lycium specimens.

References cited | To borrow, cite, or request permission

Copyright © 2003 by Crow Canyon Archaeological Center. All rights reserved.