DONATE TODAY
17
Pollen Results
Jannifer W. Gish
Contents
- Interpretations of Pollen Data
- Reconstructing Past Environments
Reconstructing Ethnobotanic Plant Use
-
Pollen Results
Introduction
Thirty-two pollen samples were evaluated from nine sites in the Sand Canyon drainage and on adjacent mesas
(Table 17.1).(1) There were two main objectives to the pollen analysis. The first goal was to characterize the
pollen compositions at these nine sites in terms of their environmental and ethnobotanic implications, with a
particular emphasis on kiva floor contexts. A second goal was to contrast paleoenvironmental conditions and
paleoethnobotanical evidence among sites in different topographic settings:
1. Sites in lower Sand Canyon (near McElmo Creek): 5MT262 (Saddlehorn Hamlet) and 5MT1825 (Castle Rock Pueblo)
2. Upper Sand Canyon talus-slope sites: 5MT3967 (Catherine's Site) and 5MT10246 (Lester's Site)
3. Upper Sand Canyon mesa-top sites within 2 km of the head of the canyon: 5MT3930 (Roy's Ruin), 5MT3936 (Lillian's
Site), 5MT3951 (Troy's Tower), 5MT5152 (Kenzie Dawn Hamlet), and 5MT11338 (G and G Hamlet)
It was expected that the contrasts in topographic/ecological settings among the sites would be apparent in the pollen records, and that these results would produce contrasting information on both resource potential of the different site locations and concomitant resource exploitation by the prehistoric occupants. Modern vegetation has been studied for the three groups of sites (Table 17.2). These data pertain to on-site and near-site plant composition; McElmo drainage plant composition is not included. The extent of local/regional vegetation contrast is hence underrepresented in Table 17.2, and patterns by topographic setting are not strongly evident. Certainly, the plants occurring along the drainage and on the McElmo floodplain could have impacted the pollen rain at nearby sites, but they were less likely to impact pollen rain at more distant (talus-slope and mesa-top) sites. Evaluations of temporal change in the prehistoric pollen assemblages had to take into account the potential variability by contrasting the topographic settings of the sites.
Methods and Data Presentation
All 32 pollen samples were processed chemically by Dr. John G. Jones of the Texas A and M University Palynology Laboratory. A zinc bromide flotation technique was used in the pollen extraction.
Pollen was moderately abundant in most samples. Only one sample from 5MT3930 and two samples from 5MT11338 failed to yield sufficient pollen for either 100-grain or 200-grain counts of combined arboreal (AP) and nonarboreal (NAP) pollen. Pollen identifications were made at 400× magnification. A moderate diversity of taxa was recorded (Table 17.3).
In addition to counts, all 32 slides were scanned at 100× magnification to detect additional large-sized taxa that might have an ecological or ethnobotanic significance. Occurrences are marked by an "X" on the pollen data tables. Large-sized (60 microns or greater in diameter) ethnobotanic taxa include maize and squash.
Occurrences of pollen aggregates were recorded during both counting and scanning. Aggregates are clumps of the same pollen type that usually indicate short-distance pollen dispersal or actual cultural introduction of pollen into a context. Pollen introduction, whether in aggregate or nonaggregate form, may be related to use of buds/flowers, fruits, or other plant parts through retention of pollen; direct use of pollen is only rarely implied. In the coding of the aggregates, "a" signifies one aggregate and "b" signifies two to five aggregates. Thus, the aggregate designation "Ju-3b" means that between two and five aggregates of juniper pollen were observed during the count, and the largest aggregate contained three grains. Aggregates observed in scanning are designated by parentheses. In this case, only the largest aggregate observed is recorded; no letter designation is used.
In the raw data tables, the pollen results are presented as percentages of the combined arboreal and nonarboreal pollen counts. The total number of grains counted (100 or 200) is listed at the bottom of the column of data for each sample. Pollen results are presented individually for each site by site group. Mesa-top sites are discussed first, followed by upper canyon talus-slope sites, and then the lower Sand Canyon sites.
The text and tables of this chapter use a two- or three-part sample code (e.g., 324-1-16) to identify a pollen sample. The first number in the sample code is a provenience indicator (PD), the second number is a laboratory catalog number (FS), and the third number, when assigned, is a point location number (PL). Together these numbers identify the exact location of a given pollen sample at a site.
Pollen Deposition
In order to fully evaluate the differences among the sites, it is important to review the types of information revealed in pollen samples in general. Any given sample, whether surface or subsurface, is a composite of local, regional, and extraregional pollen taxa. "Local pollen taxa" would derive from plants in the immediate site vicinity and adjacent environs. "Regional pollen taxa" would derive from plants in the surrounding environs of Sand Canyon. "Extraregional pollen taxa" would derive from plants beyond the ecological amplitude (potential floristic compositions) represented by the sites and surrounding region. These latter occurrences would have to be related to transport of pollen into a site through natural wind- or water-transport mechanisms or cultural means. An example of an extraregional occurrence would be spruce pollen recovered in a surface sample from a sagebrush community. There are no set distances for the terms "local," "regional," and "extraregional" pollen rain. Rather, they depend on the individual conditions of each site's topographic and geographic setting. These concepts are still useful, however, for assessing possible culturally related pollen types in the prehistoric pollen rain.
In archaeological contexts, there is an additional cultural overlay that adds further to the complexities of interpretation. Wind-pollinated taxa usually dominate the records unless moderated by unique depositional circumstances or cultural impacts. Any single sample usually represents more than one year of deposition.
The pollen record of a structure floor, for example, can be impacted by numerous factors. Sampling sediments used in construction (for example, the sediment into which a structure was excavated or the sediment used to construct the floor) can result in the recovery of pollen that predated the occupation of the structure. Human disruption of local vegetation and the use of plant materials in construction can result in the introduction of plant material/pollen into the substrate or manufactured floor surface (all structures in the current study had prepared floors). The effects of occupation throughout the use life of a structure will also be represented in a pollen record from a structure floor. Pollen is continually introduced into structures through food and nonfood use of plants. Incidental introduction of pollen can also occur when plant material adheres to clothing or items brought into a structure. The continued natural influx of local, regional, and extraregional pollen and a variety of special circumstances, such as structure remodeling or a hiatus in occupation, can also affect the pollen rain. Different kinds of structure abandonment can affect the degree of preservation of pollen assemblages in floor samples.
Any particular structure sample, then, will be a composite of natural pollen rain and cultural pollen rain. In general, the longer a structure or feature, such as a hearth, is continuously used, the more distinctive the expression of cultural activity in the pollen records. Also, the more protected the cultural signatures are at abandonment (by artifacts left in place on structure floors, for example), the more distinct and better preserved the cultural evidence. An example of an optimum sample would be one collected beneath a metate lying near the hearth of a structure whose roof collapsed at abandonment. Such a protected context would be highly likely to yield informative results on plant use in the structure, as well as plant use directly associated with the hearth. These considerations guided the pollen-sampling strategy in this study, and all are factors that are evaluated in the pollen interpretations.
Sampling Strategy
The 32 pollen samples representing nine sites (Table 17.1) were selected by the archaeologists to optimize recovery of definitive pollen evidence for plant use in structures. All structures, with the exception of Structure 401 at Castle Rock Pueblo, are masonry-lined pit structures traditionally classified as kivas. Structure 401 is a masonry tower. Floor samples were evaluated from all structures; whenever possible the floor scrape samples were collected near hearths. Since hearths are often the focus of food processing, hearth-area samples often yield substantial ethnobotanic evidence. When possible, the floor scrapes were collected under sandstone slabs or artifacts lying directly on the floors. Such conditions contribute to the high integrity of the pollen assemblages since the contexts were sealed. Sealed contexts prevent the continuing influx of pollen into a pollen assemblage after abandonment and help ensure that the pollen assemblages being studied relate to the use life of the structure.
For most structures, upper control pollen samples also were evaluated. These were collected from collapsed roof fall (burned or unburned) found immediately above the floors. The roof fall lies directly on top of the floor in each structure, and was interpreted as being deposited at the time the structures were abandoned. The upper control pollen samples were collected to serve as comparative records for the floor samples. They were expected to yield assemblages that reflected roof construction materials and also evidence of postabandonment deposits that began to accumulate in the remains of the structure. Comparing the upper controls with the floor samples was expected to help identify diagnostic ethnobotanic attributes of the floor pollen assemblages. Such attributes as high percentages or unique occurrences of certain pollen taxa, present in floor records but absent in control samples, could thus be interpreted as directly relating to the use of the structure.
Interpretations of Pollen Data
In the discussion of pollen deposition presented earlier, it is apparent that the composition of any one pollen sample is complex. The interpretations of the samples involve an even greater hierarchy of considerations that are rarely discussed in the framework of archaeological pollen reports (Gish 1993a). The type of sampling strategy used in this project, however, illustrates the very specific nature of pollen information that can be obtained when careful sample selection and collection are employed. Thus, the database provides a good framework for highlighting the interpretive considerations that comprise pollen evaluations of past environment or plant use by the inhabitants of the sites.
Reconstructing Past Environments
Environmental reconstructions involve many factors. First, the specifics of the pollen taxon are considered. This includes the identification level (broad or narrow) of the pollen taxon (family, genus, species group, or species). For example, Gramineae (grass family) is differentiated at the family level (excluding Zea, maize), while Typha latifolia (cattail), is a pollen taxon differentiated at the species level because it produces distinctive tetrad pollen. Other factors include the dispersal mechanisms of the plants subsumed by the pollen taxon (wind- vs. animal- or self-pollinated), and the related significance of pollen aggregates. Preservation potential (fragile vs. robust grains) is another factor that is important for interpreting the pollen taxa representations. Second, modern plant community analogues are used in assessing pollen records. This involves considerations of (1) pollen dominance patterns, (2) pollen taxa variety (local, regional, and extraregional), and (3) highly diagnostic taxa (riparian, nonnative). Third, depositional circumstances (fluvial, alluvial, colluvial, eolian, or archaeological context) are evaluated.
The interpretations of the pollen data for reconstructing past environments depend on the relative frequencies of certain taxa, particularly of wind-pollinated taxa. The occurrences of large or abundant pollen aggregates (which vary in significance depending on the taxon), combined occurrences of high frequencies and significant aggregates, and occurrences of highly diagnostic taxa (such as riparian taxa) are also considered.
Reconstructing Ethnobotanic Plant Use
Ethnobotanic interpretations of the cultural use of plants for food and nonfood purposes take into account the factors listed above and numerous other factors. Considerations of dispersal, for example, involve an understanding of the types of inflorescences of plants subsumed by the pollen taxon and the potential for pollen introduction through cultural activity. The presence of flowers or persistence of flower parts on gathered or harvested plant material or pollen trapped in the coarse bark of fuelwood, for example, all represent mechanisms for pollen introduction into site contexts. Cultural activities incorporate a broad spectrum of factors.
First, the specifics of the utilized plants are considered. These include (1) known use established by ethnographic observation, (2) potential use established by botanical evaluations of the edible nature of plants and other traits of plants that can be exploited by humans, (3) utilized plant part (bud, flower, fruit, stem, leaf, root, or tuber), (4) stage of development at harvest, and (5) availability. Second, plant processing, refuse disposal, and storage are considered. These include (1) level of reduction, (2) processing method (sun dry, parch, roast, boil, grind, thresh, sieve, winnow), (3) location of processing (on-site vs. off-site), (4) location of refuse disposal, (5) location of use, and (6) location of storage. Third, the less tangible attributes of preference and selection are considered.
The actual interpretations of ethnobotanic significance from the pollen results again depend on the relative frequencies of certain pollen taxa, occurrences of large or abundant pollen aggregates (which vary in significance depending on the taxon), combined occurrences of high frequencies and significant aggregates, and occurrences of highly diagnostic taxa, such as maize.
Ethnobotanic inferences based on pollen also take into account many nonpollen factors. An understanding of provenience and sample context is crucial to the interpretation. Complementary flotation sample analysis is helpful. The full range of site depositional and abandonment processes must be understood. This includes the setting of the unit under study in relationship to the natural environs (slope, soil or sediment type, and so on). The extent to which the unit was protected or exposed to natural pollen influx, both during and after its use life, is another factor. In addition, the kind of abandonment--systematic (planned) vs. catastrophic (unplanned)--is taken into account. This involves the presence or absence of well-sealed pollen trap contexts and the potential for preservation of last-use ethnobotanic pollen assemblages. Other considerations are site re-use and disturbance.
By assessing these factors, it is possible to draw conclusions about several topics of interest to archaeologists. These topics include subsistence plant use, such as the diversity of exploited plants and the reliance on agriculture vs. gathering, and nonsubsistence plant use, such as plants gathered for fuel and construction use. Pollen results also provide insights into specific site characteristics, such as (1) structure, feature, and site function; (2) occupational seasonality; (3) spatial differences in plant use; (4) changes in subsistence base through time; (5) local resource zone opportunity, exploitation, and manipulation of plants; (6) geographic range of resource acquisition; and (7) plant resource zone deterioration and conditions upon human abandonment of an area.
By looking at several sites from different settings in a region, as in this study, it is possible to address broader topics of settlement pattern and abandonment. In addition, evidence can be accumulated for local and regional resource availability and for deterioration of environments through time.
Pollen Results
The pollen results are presented by site (Table 17.4, Table 17.5, Table 17.6, Table 17.7, Table 17.8, Table 17.9, Table 17.10, Table 17.11, and Table 17.12). The results from the upper Sand Canyon mesa-top sites are discussed first, followed by an evaluation of the upper canyon talus-slope sites, and then the lower Sand Canyon sites. The paleoenvironmental significance of the samples at each site is described first, then the ethnobotanical significance. An overview is then presented addressing both spatial and temporal contrasts among the nine sites.
Upper Sand Canyon Mesa-Top Sites
The upper Sand Canyon mesa-top sites are situated at elevations generally above 6800 ft. Predominant plant communities are pinyon-juniper woodlands with sagebrush understories (5MT3936 and 5MT11338) or sagebrush scrublands with scattered pinyon and juniper in the vicinity (5MT3930, 5MT3951, and 5MT5152). Some of the sites, such as 5MT3936 and 5MT11338, were chained in the 1960s. Floristic associates at many of the sites include various shrub and herbaceous composites, mustards, goosefoot, globemallow (Sphaeralcea sp.), borages, herbaceous legumes, and grasses. More occasional associates include cholla (5MT3930, 5MT3936, and 5MT11338), prickly pear (5MT5152), and joint-fir (5MT3930 and 5MT11338).
G and G Hamlet: 5MT11338
Three samples were evaluated from this mesa-top site. All were subsurface samples from Structure 1, an unburned kiva. Sample 324-1-16 was collected as an upper control in the unburned roof fall. Samples 255-2-18 and 260-4-8 were floor scrape samples collected beneath rocks near a hearth. Pollen preservation was poor at this site, and only the upper control sample yielded a count. The two samples that did not produce sufficient pollen for counts did contain such taxa as pine, juniper, Cheno-am, sagebrush, and grass. The reasons for the poor preservation are unclear.
The upper control record exhibits codominance of juniper and Cheno-am pollen. Moderately low values of pinyon-type pine, Low-spine Compositae, and sagebrush pollen were also recorded. No diagnostic ethnobotanic taxa were observed. The Cheno-am value appears high, unless saltbush or other shrub or herbaceous chenopodiaceous plants were present in the past (saltbush is absent today). Some cultural impact on the chenopodiaceous pollen rain could be in evidence, but this is difficult to infer based on the single count. Many herbaceous species subsumed in the taxon are weedy. Ruderal invasion in the postabandonment fill of the kiva, with percolation of pollen down into the roof fall horizon, for example, could have enhanced the Cheno-am value. On the whole, however, the pollen record could be essentially natural, indicating that a juniper woodland community with a shrub understory (of sagebrush and potentially saltbush) characterized the site vicinity in the past.
Lillian's Site: 5MT3936
Five pollen samples were evaluated from Lillian's Site. These included a surface control pollen sample, an upper control sample from Structure 1, a kiva, and three floor samples from the kiva, all beneath slabs.
In the single surface sample, pinyon-type pine and juniper pollen codominate (5.0 percent difference in values). A moderate sagebrush value was recorded in the NAP, but other NAP (and AP) percentages are low, including Cheno-am. Greasewood pollen, which is morphologically distinct from the Cheno-am taxon, also was recovered. Overall, the pollen record is suggestive of a pinyon-juniper woodland community with a sagebrush understory, consistent with the community that characterizes the site today. The site was chained in the 1960s, temporarily reducing the arboreal cover, but this seems to have had little impact on the surface/recent pollen rain.
The upper control subsurface record (139-5-100) is dominated by juniper pollen. One floor sample (156-1-3) is dominated by Cheno-am pollen, while another (156-2-4) exhibits codominance of pinyon-type pine and Cheno-am pollen, and the third (172-1-10) exhibits codominance of juniper and Cheno-am pollen. The variable picture of pollen dominance in the floor samples is probably the result of cultural activities within the structure. The Cheno-am taxon is the category that is most likely affected. Any natural background pollen rain expression of pinyon-juniper woodland would be suppressed relative to cultural introduction of Cheno-am pollen through use of chenopodiaceous plant parts. Many shrubs and herbaceous plants in the cheno-am taxon yield edible greens and seeds. Hence, the environmental interpretation that can be made from the structure samples is not independent of ethnobotanic considerations. In this structure, the Cheno-am values in all three floor samples are sufficiently high (24.5 to 36.0 percent) and sufficiently distinct from the surface sample (4.5 percent) to suggest some cultural impact. Thus, the natural arboreal values are presumed to be suppressed relative to the culturally enhanced Cheno-am values. The arboreal values are still sufficiently high (pinyon-type pine 8.5 to 20.5 percent, juniper 17.5 to 30.0 percent) in comparison with the surface sample (29.5 percent pine, 34.5 percent juniper), however, to suggest local presence of both pinyon and juniper in the past. On the whole then, a pinyon-juniper woodland community with a sagebrush understory appears to be natural for the site location, and there are no strong indications of a more open setting in the past.
Ethnobotanic taxa from the kiva floor samples include maize, the Cheno-am category, beeweed, and purslane (Portulaca). The purslane occurrence is associated with the highest maize value (7 percent in 156-2-4) in the four samples. Purslane seeds and greens are edible. The plants are weedy and often volunteer in disturbed soils. Purslane is a common weed in experimental traditional crop gardens being studied by Crow Canyon Archaeological Center staff (personal observation). As an edible resource, it might have been exploited in conjunction with agricultural activities as a field weed. Purslane seeds, along with maize and other economics, were recovered in the flotation analysis of the hearth fill in this structure, and thus direct food use is a likely explanation for the pollen evidence.
Roy's Ruin: 5MT3930
Three samples were evaluated from a kiva at this mesa-top site. One (133-20) was an upper control from unburned roof fall, another (176-1-3) was collected beneath wall mortar lying on the floor in the structure, and the third (176-2-44) also was collected from the floor of the kiva, beneath a slab. Pollen preservation at this site was not as good as elsewhere, and only the two floor samples produced counts.
Juniper pollen dominates in 176-2-44, whereas juniper and the Cheno-am taxon codominate in 176-1-3. Moderate pinyon-type pine and sagebrush values were recorded, and the records are suggestive of a mixed setting comparable to today's situation of a pinyon-juniper woodland community with a shrub understory. Greasewood pollen was observed in 176-2-44, which suggests some long-distance wind transport of pollen of this anemophilous plant from floodplain settings elsewhere, or potentially incidental cultural transport and introduction of pollen into the site.
The greasewood occurrence warrants additional discussion. The nearest modern locations for greasewood are the McElmo drainage and probably Yellow Jacket Creek, both four miles or less from Roy's Ruin. Greasewood pollen was evident in the surface pollen rain in this study at 5MT262 near the McElmo drainage but also at 5MT3936 on the mesa top. So, minor occurrences from wind transport can occur naturally at sites away from the floodplain habitats. For subsurface occurrences, incidental cultural transport of pollen could result if plant materials from the floodplain, such as maize in floodplain fields, were transported into a site. Pollen from greasewood plants could have blown onto maize plants (plant-to-plant pollen transfer) and then been brought into a site incidentally with harvested maize ears. Experimental pollen washes done by this author of numerous ethnobotanic plants (Gish 1993b) have shown that plant-to-plant pollen transfer is a common mechanism of incidental pollen transport. Hence, a cultural mechanism of pollen transport could be evident here, although far more intensive studies of modern distributions of greasewood plants and associated pollen rain are needed to reliably address the meaning of the subsurface occurrences.
Ethnobotanic taxa in the kiva at 5MT3930 are beeweed and maize. The Cheno-am value in 176-1-3 also might be slightly enhanced. Only a small portion of this kiva has been excavated, and the ethnobotanic complement probably does not reflect the full extent of plant use within the structure.
Kenzie Dawn Hamlet: 5MT5152
Two subsurface samples were evaluated from the kiva at this mesa-top site. Both were floor scrapes collected from beneath rocks near the hearth in Structure 6.
Both floor samples are dominated by Cheno-am pollen. Sagebrush pollen is of secondary value in both samples, but juniper and pinyon-type pine pollen are represented by low to moderate values. Ethnobotanic taxa are beeweed and maize in both samples, along with cholla in sample 332-1-15. Thus, both samples clearly show cultural influences, and the moderate to high Cheno-am values (27.0 and 41.5 percent) are probably also culturally related. Considering this, it is problematic to use the results to reconstruct past vegetation. It can be noted, however, that the moderate sagebrush values (15.0 and 20.0 percent) that occur here are not a characteristic of many of the sites evaluated in this study. This tentatively suggests that the site setting might have been open (less wooded) in the past and that the predominant plant community might have been sagebrush/saltbush shrub. Stands of pinyon-juniper woodland, however, were probably present at no great distance. The modern vegetation setting is shrub and herb dominated (Table 17.2), so continuity between the past and modern environments is suggested.
Troy's Tower: 5MT3951
Three subsurface samples from Structure 2, a kiva, were evaluated from this mesa-top site. One (75-8-3) was an upper control from unburned roof fall, and the other two were floor samples. Neither of these was sealed by artifacts, although 77-5-12 was collected from beneath unidentified "gray material" on the floor.
Pinyon-type pine, juniper, and sagebrush pollen all codominate in the upper control record. Juniper pollen dominates (27.0 percent) in 77-5-12, while pinyon and juniper pollen codominate (25.0 and 24.0 percent, respectively) in 150-1-9. The Cheno-am values are all below 20.0 percent in these samples, which suggests low cultural impact on the pollen rain. Sagebrush values are moderately high in two of the samples (23.0 and 18.5 percent in 75-8-3 and 77-5-12, respectively). The past environmental setting appears to have been pinyon-juniper woodland with a well-developed sagebrush understory.
Ethnobotanic taxa include beeweed, purslane, and maize. Although the economic evidence is sparse, there is some suggestion of intramural patterning. The most significant ethnobotanic evidence (a 3.0 percent maize value and associated purslane pollen) was recorded for 150-1-9, which was collected near the hearth. A greater intensity of food preparation in the hearth vicinity, in contrast to lower intensity elsewhere in the structure, could account for this pattern.
Upper Sand Canyon Talus-Slope Sites
The two upper Sand Canyon talus-slope sites are situated below the rim of Sand Canyon at elevations just under 6800 feet. The predominant plant community at Lester's Site (5MT10246) is pinyon-juniper woodland with a sagebrush understory. Associated plants include joint-fir, prickly pear, and yucca, with oak in the vicinity. The predominant plant community at Catherine's Site (5MT3967) is pinyon-juniper woodland with a saltbush understory. Associated plants include joint-fir and cholla. Herbaceous associates at both sites include various composites, mustards, globemallow, and borages; grasses are common.
Lester's Site: 5MT10246
Three pollen samples were evaluated from this site. These were an upper control from burned roof fall and two floor scrape samples from Structure 1; no modern surface sample was studied. The Surface 1 sample (125-4-21) was collected from beneath a rock, and the Surface 2 sample (213-2-2) was sealed by the overlying Surface 1.
The upper control sample is dominated by pinyon-type pine pollen, with juniper of secondary value. Cheno-am and sagebrush pollen are about equally represented by moderately low values in the NAP. Together, these AP and NAP percentages suggest a pinyon-juniper woodland community, with probably both sagebrush and saltbush as significant elements in the understory. Alternatively, some roof fall plant material may be reflected. Ethnobotanic taxa in this sample are beeweed and maize, and thus the record is not totally devoid of cultural influences. Because this upper control sample was collected in burned roof fall, some storage of food items by suspension from roof rafters or food preparation activities on top of the roof are possible explanations for the ethnobotanic pollen evidence.
The count from the Surface 1 sample (125-4-21) is dominated by sagebrush pollen, with Cheno-am pollen of secondary value. Pinyon-type pine pollen is moderately well represented in the AP (12.0 percent). Ethnobotanic taxa are also significant; beeweed, cholla, and maize were recorded. The 6.5 percent maize value is notable and probably relates to the well-sealed nature of the sample (beneath a rock). The overlying burned roof fall seems to have helped protect a strong pollen signature of last-use occupational activities. It should be noted that the high sagebrush percentage (31.0 percent) might also be an artifact of cultural activity. Elsewhere, at the Duckfoot site (Adams 1993a), macrobotanical analyses demonstrated that sagebrush was probably added to structures to promote burning, and the pollen results at Lester's Site could be reflecting such an activity. Even with these cultural impacts, the moderate pinyon-type pine value continues to suggest a wooded habitat in the past.
In contrast to Surface 1, the Surface 2 record is very similar to the upper control count from the roof fall. Pinyon-type pine pollen strongly dominates in the count (51.5 percent), with juniper pollen of secondary value (23.0 percent). Concomitantly, NAP percentages are consistently low. Only a 0.5 percent maize value is suggestive of cultural influences, despite the well-sealed nature of the context. The pollen rain then, appears to be essentially natural and diagnostic of a pinyon-juniper woodland during the Surface 2 occupation.
It is interesting that the Surface 2 record is more similar to the upper control roof fall record than it is to the intervening Surface 1 record. This pattern appears also to be expressed elsewhere in this study (see discussion of Catherine's Site), although different pollen taxa are involved. Essentially, the lower floor pollen record from Structure 1 at Lester's Site appears to be a "roof fall" record. As mentioned earlier, such roof fall records could incorporate evidence of both roof construction materials and postabandonment deposition of fill in the structure. Hence, the Surface 2 record could reflect roof fall and a reassertion of the natural background pollen rain of a surrounding pinyon-juniper woodland setting. This would imply a significant hiatus in occupation of the structure. Alternatively, it is possible that the roof was removed and rebuilt at the same time that the floor was remodeled, but this could still expose the structure interior to pollen influx from the surrounding environs. Such remodeling (and also the later abandonment of the structure) possibly occurred during the peak early summer pollinating periods of pinyon and juniper (note aggregates). The construction of prepared Surface 1 could have trapped this evidence on Surface 2. Such a model would not necessarily imply a hiatus in occupation, although it does have seasonality implications for construction and abandonment activities. In any case, it seems that some form of construction/structure use-life pattern is demonstrated in the pollen expressions.
In summary then, the Structure 1 records at Lester's Site provide both informative paleoenvironmental and ethnobotanic data, as well as suggestions of a particular set of cultural practices in construction of architectural units. The predominant plant community throughout the occupation seems to have been pinyon-juniper woodland. The greater proportion of pinyon-type pine pollen relative to juniper suggests that pinyon might have been the dominant arboreal at this location, although other factors (such as simply closer position of a source tree) could account for the high pinyon-type pollen values. Cultigens, as well as gathered wild plants, are indicated for the site. As will be seen, this ethnobotanic complement is comparable to that at 5MT3967 (Catherine's Site) but slightly less diverse than that at the two sites in the lower Sand Canyon drainage.
Catherine's Site: 5MT3967
Four pollen samples were evaluated from 5MT3967, located in the upper Sand Canyon drainage. These included a surface control sample and three samples from Structure 2. Of the three subsurface samples, sample 227-26-18 was an upper control from unburned roof fall above Floor 1, sample 204-1-1 was a floor scrape from under a rock on Floor 1, and sample 206-2-2 was a floor scrape from Floor 2. The latter was sealed by Floor 1, and the rock on Floor 1 was covered with about 50 cm of unburned roof fall.
The surface control sample is strongly dominated by AP. Juniper (51.5 percent) is the dominant category, with pinyon-type pine pollen (25.0 percent) of secondary value. Other AP taxa are represented by only low values (2.5 percent or less) and reflect regional or extraregional taxa (oak, ponderosa-type pine, and Douglas fir). In the NAP, a low value (8.0 percent) of Cheno-am pollen was recorded. Low values (2.0 percent or less) of composites, grass, and joint-fir were also recorded. Together, the AP and NAP values suggest a pinyon-juniper woodland with some saltbush in the understory.
Of the three subsurface samples, the upper control and Floor 2 records exhibit codominance of juniper and Cheno-am pollen, while the Floor 1 sample is dominated by pinyon-type pine pollen with juniper of secondary value. All three records are strongly diagnostic of a past woodland community, probably with some saltbush in the understory. All three samples also include ethnobotanic taxa (maize, cholla, or beeweed), and it is likely that the relatively high Cheno-am values in the upper control sample (31.0 percent) and Floor 2 sample (26.0 percent) are at least partly the result of human activity. As in Structure 1 at Lester's Site, there is greater similarity in the pollen assemblages between the upper control and Floor 2 records than in the intervening Floor 1 record. This is expressed primarily in the pinyon-type pine, juniper, and Cheno-am percentages, rather than in the ethnobotanic values, which are low in all three samples, with the exception of beeweed (3.5 percent). Again, a construction/remodeling pattern appears to be in evidence, although, in this case, the cultural impacts are more subtle since the pollen records predominantly reflect natural background pollen rain.
Lower Sand Canyon Sites
The two lower Sand Canyon drainage sites are situated at elevations ranging between 5500 and 5800 ft, substantially lower than the sites previously discussed. The predominant plant community at Saddlehorn Hamlet (5MT262) is pinyon-juniper woodland with an understory that includes saltbush and sagebrush. The predominant plant community at Castle Rock Pueblo (5MT1825) is also pinyon-juniper woodland with saltbush and sagebrush. Winterfat (Eurotia sp.--another Cheno-am source), prickly pear, and joint-fir all occur in the vicinity of Castle Rock Pueblo. Herbaceous associates at the two sites include various composites, mustards, and globemallow; grasses are common. Saddlehorn Hamlet overlooks an intermittent tributary of McElmo Creek, and Castle Rock Pueblo is adjacent to the McElmo floodplain. Pollen rain from riparian plants could have potentially impacted both sites.
Saddlehorn Hamlet: 5MT262
One surface control sample and two subsurface samples from Structure 1, a kiva, were evaluated from this site. Of the two subsurface samples, one (115-2-33) is an upper control sample from roof fall covering the floor; the other sample (101-4-4) is a floor scrape sample collected from beneath a mano near the hearth. The stratigraphy of the fill sequence in the kiva suggests that the roof over the main chamber collapsed at abandonment due to intentional burning. Hence, the floor sample was protected from postabandonment pollen influx not only by the artifact, but also by a thick layer of immediately deposited roof material.
The surface sample is strongly dominated by arboreal pollen. Juniper pollen (40.0 percent) dominates the record, with pinyon-type pine pollen (29.0 percent, note aggregates) of secondary value. The other AP taxa, which are represented by low percentages, probably reflect wind transport of pollen from regional and extraregional sources. In the NAP, moderately low values of low-spine Compositae, sagebrush, and Cheno-am probably reflect local sources. Other NAP taxa represented by low values in the sample are high-spine Compositae, greasewood, grass, joint-fir, and prickly pear. All of these are known to occur in the local site vicinity. On the whole, although the pollen record includes regional and extraregional as well as local taxa, the pollen assemblage is strongly diagnostic of the local pinyon-juniper woodland plant community. Riparian taxa are not evident, with the exception of the high groundwater indicator, greasewood.
The two subsurface samples from the kiva are characterized by markedly different pollen assemblages that strongly reflect cultural activity. The upper control sample is dominated by Cheno-am pollen (66.0 percent); the floor sample is dominated by maize pollen (36.0 percent). In the case of the upper control sample from burned roof fall, several explanations could be offered for the Cheno-am value. It is possible, for example, that shrub or herbaceous materials used in roof construction could be in evidence. Alternatively, rodent caching of chenopodiaceous plant material in the roof fall stratum could have influenced the sample, although care was taken to collect the control sample away from burrows. A third possibility is that pollen from weedy, chenopodiaceous species that pioneered in the fill over the burned and collapsed roof percolated down into the roof fall stratum. Finally, food or nonfood items stored by suspension from roof rafters or kept externally on top of the roof may have resulted in the incorporation of Cheno-am pollen in the roof fall stratum. Although it cannot be clearly ascertained which of these possibilities is the correct explanation for the pollen expression, it seems clear that the assemblage is not entirely natural. Maize and beeweed pollen were also noted in the sample, further indicating cultural effects on the pollen record. Disregarding for the moment the high Cheno-am value, it is still apparent in the count that juniper (13.5 percent) and pinyon-type pine pollen (5.5 percent) are sufficiently present in the pollen rain to tentatively suggest that a pinyon-juniper community predominated in the locality in the past as it does today. The enhancement of the Cheno-am category clearly masks the natural pollen rain. Riparian taxa are not evident in the count.
The floor sample (101-4-4) bears out the earlier stated premise that protected, high-integrity contexts near the hearth area can be highly productive of plant use evidence. This is borne out not only by the 36.0 percent maize value (note aggregates), but also by the presence of squash (Cucurbita) pollen. A 1.0 percent Umbelliferae (parsley or carrot family) value (note aggregate) is probably ethnobotanically significant as well, because members of the Umbelliferae family are low pollen producers and it is rare to see the pollen, let alone an aggregate, represented in the natural background pollen rain. Numerous species of the Umbelliferae family produce edible roots, with some also producing useful flowers, seeds, leaf stalks, leaves, and stems (Kirk 1970:117-125). Members of the Umbelliferae family seem to have been commonly utilized by the ancestral Puebloans, and the taxon appears in other sites in the Sand Canyon watershed (Gish 1990). The variety of utilized plants revealed in this floor context is thus squash, maize, and parsley/carrot family, which suggests that substantial food-preparation activities took place inside the kiva. Although the floor sample is heavily modified by cultural pollen, AP values of pinyon-type pine and juniper are still fairly notable, along with sagebrush pollen in the NAP. As with the upper control sample, the suggested natural vegetation setting is pinyon-juniper woodland with sagebrush in the understory flora. Again, riparian taxa are not evident.
In summary then, it can be inferred that the predominant plant community in the past at this lower Sand Canyon drainage site was pinyon-juniper woodland, as it is today. The single floor sample revealed a significant focus on cultivated crops, both maize and squash. Squash pollen has not been commonly found elsewhere at sites in the Sand Canyon drainage (Gish 1988, 1990). The occurrence here raises the interesting possibility that the lower Sand Canyon drainage region (including the nearby McElmo Creek floodplain) might have been particularly well suited for the cultivation of squash as well as maize. From this perspective, additional study of lower Sand Canyon sites might be especially useful for addressing questions of regional abandonment, since any changes in conditions in optimal agricultural areas of the Sand Canyon drainage could be expected to have a major impact on the regional subsistence network as a whole. This topic of changing conditions could also be addressed through geo-palynological studies of alluvial/fluvial sequences in the Sand Canyon and McElmo drainages, where shifts in the surrounding environment might be distinctive in the more natural background pollen rain.
Castle Rock Pueblo: 5MT1825
Six pollen samples were evaluated from this site. These included one surface control sample, three samples from Structure 206, one sample from Structure 302, and one sample from Structure 401.
The single surface sample is strongly dominated by arboreal pollen. Juniper is the most abundantly represented taxon (36.0 percent); pinyon-type pine is of secondary value (23.0 percent, note aggregates). Other AP taxa represent long-distance wind transport of pollen from regional and extraregional sources. In the NAP, sagebrush and Cheno-am are the two most abundantly represented taxa. These evidently reflect locally occurring shrubs (Table 17.2). Sources for the Cheno-am pollen could be either saltbush or winterfat (Eurotia sp.), as both shrubs occur in the area. Other NAP taxa are represented by only very low values, but include Low-spine Compositae, grass, mustard family, joint-fir, filaree (Erodium), and prickly pear. In most cases, plants occurring locally today account for these representations. The filaree pollen is notable because it probably reflects the nonnative species, and this is diagnostic of the recent/modern nature of the pollen rain. Maize (note aggregate) pollen was also recorded in the surface sample. Modern corn fields occur about 100 m from the site, which probably accounts for the surface maize pollen occurrence. On the whole, the AP and NAP representations strongly correlate with the pinyon-juniper woodland community with a shrub understory that predominates in the area today.
Of the three samples from Structure 206, a kiva, one (552-2) was an upper control sample from unburned roof fall immediately above Floor 1, and two others (549-1-5 and 553-1-2) were floor scrapes. Neither of the two floor samples was collected from beneath artifacts.
The upper control sample, 552-2, is strongly dominated by Cheno-am pollen (59.0 percent). Because the presence of maize and an outstanding beeweed (Cleome) value (14.0 percent) reflect cultural activity, the high Cheno-am percentage is probably the result of cultural activity as well. The cultural pollen could relate to food or nonfood items having been stored in (or suspended from) roof rafters, upward mixing from on-floor assemblages, or other causes. In any case, the strong cultural influences on the pollen rain make it problematic to use this record for paleoenvironmental information. It can be noted, however, that the upper control sample still yielded a 10.0 percent juniper value and 4.0 percent pinyon-type pine value. Although these AP values are greatly suppressed relative to the culturally influenced pollen taxa, they still hint at a pinyon-juniper community in the past. Greasewood pollen was also recorded in sample 552-2. Since Castle Rock Pueblo overlooks McElmo Creek, this probably reflects vegetation on the floodplain.
The two floor records are also strongly dominated by Cheno-am pollen, with values higher in both samples in comparison with the upper control sample. Both samples exhibit obvious cultural impacts, through the presence of maize and low to high beeweed values. One sample (549-1-5) also contained Umbelliferae pollen. The ethnobotanic complement for this structure thus appears to be Cheno-am, beeweed, parsley family, and maize, with both floor samples characterized by very similar pollen assemblages.
The single sample (464-2-2) evaluated from Structure 302, a kiva, was collected from the base of the hearth. The record is very similar to sample 549-1-5 from Structure 206. Cheno-am pollen is the strong dominant; obvious ethnobotanic taxa include beeweed, Umbelliferae, and maize. Again, the Structure 302 count is heavily influenced by cultural activity, evidently reflecting foods prepared at the hearth. Fairly low juniper, pinyon-type pine, and sagebrush values, however, are still suggestive of a pinyon-juniper woodland setting with a scrub understory in the past.
The one sample (437-1-39) from Structure 401, a tower, was collected from beneath an abrader on the original floor surface (Floor 2) of the structure, which was subsequently sealed by Floor 1. Again the count is dominated by Cheno-am pollen (46.0 percent). Diagnostic ethnobotanic taxa include purslane and maize. The high maize pollen value of 24.5 percent testifies to the sealed nature of the context. Also of note in the sample is a 3.0 percent value of greasewood pollen, which reflects the proximity of the McElmo floodplain. In the AP, low values of pinyon-type pine and juniper still hint at a surrounding pinyon-juniper community.
In summary, the pollen records from the structures at Castle Rock Pueblo show a high degree of cultural influences. The presence of the greasewood pollen indicates that the site location is sufficiently close to the McElmo drainage to potentially pick up evidence on the nature of plant communities that characterized the drainage in the past. Further pollen studies of the pueblo might provide a better picture of these floodplain communities.
Summary and Conclusions
The pollen results met the major goals of the analysis. It was possible to characterize the pollen compositions at the nine sites in terms of environmental and ethnobotanic implications. It also was possible to identify contrasts in environmental conditions and ethnobotanic complements among the three groups of sites.
From a paleoenvironmental perspective, the nine sites seem to have been characterized by slightly different vegetative settings in the past. At least two of the mesa-top sites possibly had more open vegetative settings--juniper/sagebrush (or juniper/sagebrush/saltbush) at 5MT11338 and sagebrush/saltbush (with scattered pinyon and juniper in the vicinity) at 5MT5152. The three other mesa-top sites, 5MT3930, 5MT3936, and 5MT3951 were apparently dominated by pinyon-juniper communities with understories that included sagebrush. The two talus-slope sites, 5MT10246 and 5MT3967, were probably also dominated by pinyon-juniper woodland, but with understories of sagebrush/saltbush or saltbush. Similarly, both of the lower Sand Canyon sites, 5MT262 and 5MT1825, probably had mixes of pinyon, juniper, and some understory shrubs.
These inferences about past plant communities must be seen against differences in occupational impact on the sample sets evaluated for each site. Cultural effects on the pollen rain at Catherine's Site (5MT3967), for example, are not as strongly demonstrated as cultural effects on the pollen rain at Castle Rock Pueblo. Concomitantly, arboreal pollen values are high at Catherine's Site and low at Castle Rock Pueblo, but both sites are inferred to have been characterized by some presence of conifers (pinyon and juniper) in the vicinity, although 5MT3967 was possibly more wooded.
The different site settings apparently all had similarities in floristic composition. In fact, the pollen results (particularly in terms of taxa diversity) are more notable for their similarities than their dissimilarities (Table 17.2). At the same time, the topographic settings of the sites were probably sufficiently distinct overall to offer their own spectra of plant resource opportunities. On the basis of today's plant distributions, it is inferred that ancient plant availability on talus slopes and in the lower canyon, in particular, was probably more diverse. Such resource differentiation, however, is not well expressed in the ethnobotanic pollen assemblages at the various sites. Perhaps an agricultural emphasis in subsistence leads to uniformity in plant use.
The full variety of anthropogenic/ethnobotanic taxa in the mesa-top sites appears to be Cheno-am, beeweed, purslane, cholla, and maize. The range of culturally related taxa in the upper Sand Canyon talus-slope sites is Cheno-am, beeweed, cholla, maize, and sagebrush (fuel use). In lower Sand Canyon sites, the variety appears to be Cheno-am, beeweed, purslane, parsley/carrot family, maize, and squash. Possibly somewhat greater field weed use is evident in these sites in conjunction with more intense agricultural efforts on the floodplain. The two lower Sand Canyon sites are in the group of sites that date after A.D. 1250, but it is not clear if temporal placement is any more important than location and resource opportunity in the slightly different ethnobotanic assemblages associated with these sites.
It is interesting that greasewood pollen occurs in association with relatively high or very high maize values in samples from two upper canyon sites (5MT3951 and 5MT10246) and one lower canyon site (5MT1825). Sample 150-1-9 from 5MT3951 has 3.0 percent maize and 0.5 greasewood values; sample 125-4-21 from 5MT10246 has 6.5 percent maize and 0.5 percent greasewood values; and sample 437-1-39 from 5MT1825 has 24.5 percent maize and 3.0 percent greasewood values. The correlation in the Castle Rock Pueblo sample is especially strong. Greasewood pollen was also recorded in subsurface samples at two mesa-top sites, 5MT3936 and 5MT3930. At 5MT3936, greasewood pollen was recorded in an upper control sample (139-5-100), and another sample from the same structure (156-2-4) had a high maize value (7.0 percent). At 5MT3930, sample 176-2-44 had low values (0.5 percent) of both maize and greasewood pollen in the same sample. With the exception of Castle Rock Pueblo, there were presumably no local or nearby sources for the greasewood pollen at most of the sites. Since greasewood is wind pollinated and was found in surface control samples at 5MT262 and 5MT3936, it is possible that these greasewood/maize pollen associations simply could be fortuitous. It also seems possible, however, that the greasewood pollen in the subsurface samples is the result of incidental transport of pollen (from plant-to-plant pollen rain) on harvested ears of maize brought into the sites from fields on the McElmo Creek floodplain. If the McElmo floodplain was a focus of agricultural activities by the prehistoric populace in this region, then the greasewood pollen might signify interaction with the floodplain field system.
In reviewing earlier studies (Scott and Aasen 1985; Gish 1988, 1992), there appear to be some other associations between greasewood and maize representations. For example, the combination of maize, greasewood, and willow pollen seen in a sample from Structure 208 at Sand Canyon Pueblo in upper Sand Canyon (Gish 1988) could be more than just a coincidental association of long-distance wind-transported riparian and floodplain taxa with maize. It could be a diagnostic signature for plants cultivated on the McElmo floodplain and transported up-canyon to the pueblo. A general pollen signature for crops transported into sites from McElmo floodplain fields could thus be greasewood pollen or notable occurrences of riparian taxa, such as willow along with maize. It could prove useful to examine modern distributions of greasewood across the greater Sand Canyon region, along with site distributions, and compare these with subsurface occurrences of riparian and floodplain pollen taxa and evidence for agricultural activities.
The agricultural strategy used in the region was probably at least twofold. People living in mesa-top farmsteads could have used the more naturally open or intentionally cleared areas of level ground for extensive maize fields. People who farmed large fields on the McElmo floodplain could have taken advantage of greater water availability for cultivation of squash and maize. Plant materials transported into sites from mesa-top fields would not incidentally introduce any diagnostic taxa of the mesa-top setting, since the total taxa variety in the mesa-top records appears to be subsumed by the Sand Canyon pollen taxa variety. Plant materials transported in the other direction from floodplain fields, however, might be detected through occurrences of riparian or floodplain taxa. Such taxa could become associated with transported materials due to plant-to-plant pollen rain or other mechanisms of incidental contact. This kind of evidence could help define the overall nature of the subsistence strategy in the greater Sand Canyon region.
1. Throughout this chapter, the contexts from which pollen samples were collected are reported using the same terms that appear in the corresponding site-report chapters (Chapters 2-14). Variation in how different authors refer to similar contexts (for example, one author may use the term Floor 1, whereas another may use Surface 1) account for the apparent inconsistencies in provenience terminology in this chapter.