Petroglyphs in the Horseshoe Reservoir Area of the Lower Verde Valley, Central Arizona Henry D. Wallace Report Submitted to Statistical Research, Inc. P.O. Box 31865 Tucson, Arizona 85751 Center for Desert Archaeology 3975 North Tucson Boulevard, Tucson, Arizona 85716 ! October 1996 TABLE OF CONTENTS List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Petroglyph Recording Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Variables Coded . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photography and Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 3 5 8 RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 The Crash Landing Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Nearby Petroglyph Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Horseshoe Mesa, AZ U:2:27/01-187 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Site AZ U:2:96/01-259 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Site AZ U:2:188/01-73 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Site Survey Data for Other Nearby Sites with Petroglyphs . . . . . . . . . . . . . . . . . . . . . . . 14 REGIONAL COMPARISONS, STYLE, AND CULTURAL IDENTITY . . . . . . . . . . . . . . . . . . . . Horseshoe Reservoir Area Rock Art Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparisons with the Phoenix Basin and Beyond . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 16 17 20 ROCK ART FUNCTION AND INTERPRETATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 REFERENCES CITED . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 iii LIST OF FIGURES Figure 1. Map of the Horseshoe Reservoir area, showing the location of petroglyph sites visited during this study (map prepared by Statistical Research, Inc.). Figure 2. Coding index for metric data on boulders and panels. Figure 3. Coding index for petroglyph design element data. Figure 4. Petroglyph design element categories and coding numbers. Figure 5. Coding form for LVAP petroglyphs (boulder/panel data). Figure 6. Coding form for LVAP petroglyphs (Class II data). Figure 7. Petroglyph design elements on Panel A-1a, Crash Landing site. Figure 8. Petroglyph design elements on Panels A-1b and A-1c, Crash Landing site. Figure 9. Petroglyph design elements on Panel A-2a, Crash Landing site. Figure 10. Petroglyph design elements on Panel IS-1a, Crash Landing site. Figure 11. Petroglyph design elements on Panel IS-1b, Crash Landing site. Figure 12. Petroglyph design elements on Panel IS-1c, Crash Landing site. Figure 13. Petroglyph design elements on Panel IS-2a, Crash Landing site. Figure 14. Petroglyph design elements on Panel IS-2b, Crash Landing site. Figure 15. Petroglyph design elements on Panel IS-2c, Crash Landing site. Figure 16. Photograph of petroglyph Panel A-1a, Crash Landing site. Figure 17. Photograph of petroglyph Panel A-2a, Crash Landing site. Figure 18. Photograph of petroglyph Panel IS-1a, Crash Landing site. Figure 19. Photograph of petroglyph Panel IS-2-a, Crash Landing site. Figure 20. View to the northeast from the top of Horseshoe Mesa toward Horseshoe Dam. The Crash Landing site is visible on either side of the large road cut at top left of center. iv Figure 21. View to the southeast from the top of Horseshoe Mesa toward site AZ U:2:188 on ridge tip flanked by road in left center. Figure 22. Large,complex petroglyph panel at the Horseshoe Mesa site. Figure 23. Large,complex petroglyph panel at the Horseshoe Mesa site. Figure 24. Connected scroll petroglyph designs at the Horseshoe Mesa site. Figure 25. Anthropomorphic and zoomorphic petrogoyph designs at the Horseshoe Mesa site. Figure 26. Zoomorphic petroglyph designs the Horseshoe Mesa site. Figure 27. Petroglyph designs the Horseshoe Mesa site. Panel with large deer and adjacent vandalized panel. Figure 28. Petroglyph designs the Horseshoe Mesa site. Note hand-holding anthropomorphs to the right. Figure 29. Petroglyph designs at the Horseshoe Mesa site. Probable scene with dancing anthropomorphs, unusually place zoomorphic figures, and sheep horns to far right. Figure 30. Petroglyph designs on panel atop the Horseshoe Mesa site, adjacent to rock-walled enclosure. Figure 31. View from masonry compound and petroglyph site AZ U:2:96 east-northeast toward Horseshoe Mesa site. Figure 32. Petroglyph boulder with grinding slip on top at site AZ U:2:96. LIST OF TABLES Table 1 Letter designations and short descriptions of the lower Verde River area petrofacies. Table 2 Point count parameters used in the petrographic analysis of lower Verde River area sand samples and sherd thin sections (after Kamilli 1994; Lombard 1987; Miksa 1992). Table 3 Recalculated point count parameters used in correspondence analysis and discriminant analysis. Table 4 Lower Verde River area sand sample correspondence analysis ranked parameter optimal scores. v PETROGLYPHS IN THE HORSESHOE RESERVOIR AREA Henry D. Wallace This report is focused on the rock art present at a small ridge top agricultural locality in the lower Verde Valley near Horseshoe Dam known as the Crash Landing site, AZ U:2:78/01-278. Four boulders that exhibited over 24 petroglyph design elements were found at this site, as well as numerous other cultural features including a two-room isolated masonry field house and a large agricultural complex with rock piles, contour terraces, and boundary walls. The research design for the Lower Verde Archaeological Project (LVAP) (Ciolek-Torrello et al. 1992) presented three primary research questions pertaining to the rock art at the Crash Landing site: 1. Can the petroglyphs be characterized according to a specific known style or tradition of petroglyph art? 2. Can the petroglyphs be provisionally dated? 3. What is the context of the petroglyphs? That is, what are the locational, archaeological, and environmental parameters? In addition to the desire to classify the rock art, the research design focuses on the issue of cultural affiliation. Who made the petroglyphs? Can the styles present provide the necessary clues to help elucidate this issue? This report will document the rock art at the Crash Landing site, and discuss it in relationship to rock art reported elsewhere in southern and central Arizona. Discussion of three other nearby rock art sites — Horseshoe Mesa (AZ U:2:27/01-187), AZ U:2:96/01-259, and AZ U:2:188/01-73 — is also provided to help assess rock art in the study area and how it may bear on issues of cultural identity and rock art function. (Note: all LVAP site numbers are ASM numbers; unless otherwise noted, all other site numbers are ASU numbers). Definitive conclusions are not possible from an investigation of a small rock art site such as the Crash Landing site, and without a detailed analysis of the other major sites in the region, the conclusions presented in this report must be considered preliminary. Nevertheless, it was possible to tentatively date the designs at the site and make an argument for contemporaneity with other larger sites in the vicinity. The settings of these sites provide important clues to the role of rock art in the prehistory of the Verde Valley. Figure 1 provides an overview of the region and the location of the four sites in relation to one another. 1 METHODS The methods applied in this study are built on those developed for the recording of the Picacho Mountains petroglyphs sites (Wallace and Holmlund 1986) and the Painted Rocks Reservoir sites (Wallace 1989). The ridge upon which the rock art was located had been intensively surveyed for all types of cultural features, including rock art, prior to my field efforts, and no attempt was made to search for additional unreported boulders. A detailed topographic map with cultural features plotted had been prepared by Statistical Research, Inc. (SRI), prior to my field work and this was used for spatial control. In order to obtain comparative data I visited three nearby rock art sites: Horseshoe Mesa, AZ U:2:96/01-259, and AZ U:2:188/01-73. No formal recording was conducted on these sites, but photographs were taken and notes made. I first visited the sites on May 21, 1992, accompanied by Steve Shelley of SRI. Because of inclement weather, I was unable to do more than photograph the panels and make notes regarding the sites. This initial set of photographs was used to prepare most of the drafted illustrations seen in this report. A second visit to the Crash Landing site was made on October 14, 1992 to finish recording the rock art at the site. I also made a second trip to the Horseshoe Mesa site to obtain additional photographs and take a more careful look than had been possible during the lightning storm on the previous less productive, but more exciting, trip. Following this field effort, the Bureau of Reclamation, using a helicopter and sling, moved Boulder A-1 at the Crash Landing site to a safe haven on the ridge, approximately 50 m to 60 m west-northwest of its former location, where construction activities would not damage it. In the early fall of 1993, an additional petroglyph boulder (A-2), formerly hidden beneath a large cholla and other vegetation, was discovered by a Reclamation construction inspector near the previously recorded Boulder A-1 at the Crash Landing site. On September 2, 1993, I returned once more to the site, in the company of Reclamation archaeologist Jon Czaplicki, to record this boulder and move it from the path of construction activities upslope to the ridge where Boulder A-1 had been moved. 2 Petroglyph Recording Procedures The rock art at the Crash Landing site was recorded using many of the procedures documented in Wallace (1989). Some of the techniques are not relevant to such a small site. They are used here in the interest of maintaining comparability in approaches and to facilitate the construction of larger regional databases. The techniques applied are discussed below. Petroglyphs in central Arizona commonly occur in groups or clusters consisting of anywhere from two to over a hundred separate boulders. This is taken into account in the hierarchical numbering system used here, which separates glyphs into cluster, boulder, panel, and element numbers. Clusters are defined in the field as groups of two or more boulders, or boulders and outcrops, that exhibit petroglyphs, regardless of chronological age. They are defined by field evaluations of spatial distribution and, although they probably can be considered meaningful cultural units, they are not necessarily the only possible arrangements. The boulder numbering system is designed to facilitate spatial control over the distribution of petroglyph panels. Therefore, each boulder with glyphs within a cluster is given an integer number between 1 and n for each cluster. Each cluster is assigned an upper-case letter designation. Boulders are distinguished from outcrops by being physically separated from bedrock. Each boulder face with petroglyphs is treated as a panel. Thus, a panel is a discrete side of a boulder, which exhibits a roughly comparable inclination and declination across its surface. Panels are numbered consecutively on each particular boulder by lower-case letters (a-zz). Particular petroglyph design elements or motifs on a panel are numbered consecutively (1-n). Thus the format for coding Cluster A, Boulder 2, Panel a, Element Number 1 would look like this: A-2a:1. Isolated boulders are given the separate designation “IS” in place of the cluster designation and numbered consecutively by boulder for each site (e.g. IS-1, IS-2, IS-3, etc.). Prior to any petroglyph recording, an assessment of patination and repatination was necessary. As the process of repatination can provide relative dating for petroglyphs, a careful consideration of the process in relation to the local lithology is required. Of particular concern is the potential for different types and rates of chemical weathering. Rock varnish, distinguished here from chemical and mechanical weathering of rock surfaces, can also be useful for assessing the relative age of petroglyph designs; although the techniques used to quantitatively derive dates, such as cation-ratio dating and direct AMS radiocarbon dating, are controversial (Dorn 1983, 1989, 1994a; Dorn et al. 1986; Krinsley et al. 1990). Dorn, in a long series of articles, has effectively addressed 3 most of the criticisms that have been leveled against the approaches (e.g. Bierman and Gillespie 1991; Harry 1992) and freely points to the weak points (Dorn 1992, 1994a, 1994b; see especially Whitley et al. 1996). Additional verification is needed for the AMS approach to evaluate the conditions under which ancient carbon can be incorporated into the varnish and inadvertently lead to dates older than the age of the varnish. Several factors can contribute to such a result, and the issue is by no means settled as to how to effectively and confidently address and control for them. Caution is warranted and additional testing is advisable. It is assumed in this study that the findings of Dorn and Oberlander (1981) and Dorn (1991) are correct: that rock varnish is a product of mixotropic manganese-oxidizing microorganisms. Detailed discussions and considerations of varnish formation are not provided here as none of the designs investigated exhibited rock varnish on the petroglyphs. The focus is placed here on chemical and mechanical weathering. As in previous work in the Picacho Mountains (Wallace and Holmlund 1986:76-79) and the Painted Rock Reservoir area (Wallace 1989:84-85, 119-121), the level of repatination of each petroglyph design was coded as one of a series of possible patina classes. The mineralogy is different for the project petroglyph boulders than for the Picacho Mountain sites, and the small sample of petroglyph panels present for evaluation seriously limited the observations that could be made concerning the potential for relative dating in the vicinity. The rock art at the Crash Landing site was pecked onto boulders made of basalt, as is the case for the other nearby sites. Although basalt can exhibit clear patina distinctions for surfaces of different ages, it does not offer distinctions that are as clear-cut as some other rock types, such as the granodiorites and latites found in the Picacho Mountains. This is true for several reasons. First, rock varnish is generally not present on the rock art of the vicinity due at least partly to the relatively recent ceramic period age of the designs. Second, although rock varnish is relatively indiscriminate in formation, provided suitably porous rock surfaces are available, the visibility of chemical weathering is more dependant on the mineralogy of the rock involved. Certain minerals oxidize and degrade more rapidly and visibly than others. Where a clear color shift in exposed hornblendes is readily discernible in the felsic rocks of the Picachos, no such trend is readily identifiable at a macroscopic level on the project area basalts. The rate of varnish formation on the Verde Valley basalts is believed to be relatively slow compared to that in the western desert or Picacho Mountains, based on petroglyph design style comparisons. The relative amount of aeolian-deposited clays and other minerals on rock surfaces 4 is thought to be correlated to the development of chemical weathering (particularly hematite staining), and to the formation of rock varnish; the more clay, the more rapid the development of chemical weathering and rock varnish (Wallace 1989:84). Speculating here, I suspect there is a much lower relative level of aeolian deposition on the rock surfaces in the confined, well-vegetated, narrow valley of the lower Verde River than in the wind-swept, open country of the western desert or Picacho area. The glyphs on the Crash Landing site boulders and those on the other nearby sites exhibited comparable patina development. No marked distinctions in my brief inspections could be made. Because I have observed clear distinctions in different age designs on basaltic rocks in other regions (e.g., Coso Mountains in California, Signal and Tumamoc Hills in Tucson, Gillespie Dam near Buckeye), not all of which are explicable by differences in aeolian deposition, it is possible that a chronometric system could be established in the future. The lack of discernible distinctions within the sites observed in the lower Verde area could be taken to mean rough contemporaneity, though perhaps at the extremely gross level of 200-800 years. Without significant effort, it cannot be honed down any finer. The petroglyphs at the Crash Landing site were recorded using the coding indices provided in Figures 2, 3, and 4, forms similar to those in Wallace (1989:187-188) (Figures 5 and 6), black-andwhite and color photography, and scaled drawings. I was assisted in this endeavor by Steve Shelley and Jon Czaplicki for some measurements; although because I did all the attribute coding and illustrations, they are only guilty by association. Variables Coded The selection of variables to be recorded was dictated by the findings of several previous studies, including Wallace and Holmlund (1986), Bruder (1983), Wallace (1983, 1989), Ferg (1979), and White (1965). Figures 2 and 3 provide summary descriptions of the selected variables. Information pertinent to each boulder, panel, and individual glyph are recorded. Many of the variables considered and examined at the project site were not relevant or present on the panels or elements recorded. The fact that a particular variable is listed on the coding indices, but not listed among the raw data, is an indication that this particular variable did not apply or no cases were observed at the recorded site. The variables are generally self explanatory. Those requiring clarification are briefly discussed 5 below. Element Elaboration Some design elements, such as scrolls, squares, and diamonds, may be elaborated by additional lines, hatchure fill, and so on. For those elements that can, by definition, be elaborated in some way beyond their basic form, this variable is used to indicate if such an elaboration is present or absent. Those elements that are defined as having the potential for elaboration are marked with asterisks on the design element coding index in Figure 4. Patination Please see the brief discussion above. Period/Style This variable was applied in a previous study (Wallace and Holmlund 1986:63) based on a rigorous set of standards pertaining to repatination. It was intended in that study to be strictly chronological in application. Stylistic factors were intentionally avoided in order that an independent measure of stylistic change could be assessed. Enough is now known concerning the definition and chronological placement of the principal southern Arizona styles so as to permit the inclusion of style as one avenue to assigning a chronological and stylistic label to a particular design or set of designs. Execution of Element Most of the attributes for this variable are self evident. Abraded designs are distinguished from those that are ground, or ground and pecked, by exhibiting a rubbed surface that still displays some of the original rock texture and surface within the limits of the design. Although an abraded design could be called lightly ground, there is a clear dichotomy between those in the field that are well ground and those that are merely abraded or lightly ground. Some panels exhibited designs that graded from being composed of well-pecked, clear designs to being composed of scattered 6 peck marks (dints). These were generally coded as being “pecked.” Superimposition This is only applied to cases where it is clear which is (are) the overlying and which is (are) the underlying elements. If it is not clear, or if the elements are thought to be contemporaneous, they are treated as a case of panel integration. No cases of superimposed designs were recorded. Panel Integration This term is used to describe the combination of distinct elements into an overall design. This does not apply to simple motifs such as concentric circles or an elaborated spiral. It is only applied to cases where the overall composition of the design incorporates multiple elements (or motifs). Combinations of elements within a case of panel integration were separated out and recorded individually, if possible, and it was noted that they were tied together. Feature Type and Number Determining whether particular features were culturally and temporally associated with particular petroglyph panels is often a difficult task. Functional association is determined through proximity on the basis of the feature type and local terrain. At the Crash Landing site, a large group of agricultural features is located near the petroglyphs, as are a possible trail and a clearing. The potential relationship of these features to the rock art is discussed further below. Design Element Typology and Application The typology developed and applied to the LVAP petroglyphs evolved from the work of Wallace (1989) and Wallace and Holmlund (1986:64-67), which in turn had developed from the initial efforts of Ferg (1979) and Wallace (1983). See Wallace and Holmlund (1986:64-67) for a complete description of the rationale behind the system applied here and the basic typology. The present typology has not been changed from that seen in Wallace (1989:87, Figure 5.5), with the exception that the element type coded as a “torah,” a mistake in the original system, is now called 7 a “candelabra,” to fit its appearance (with no intent of saying that it represents such an object!). Photography and Mapping All petroglyph panels were photographed using both 35-mm color slide film (Kodachrome 64® and Fujichrome 100®) and 35-mm black-and-white print film (Kodak Plus-X®). Additional photographs were taken to provide an idea of the setting for petroglyph clusters and isolated panels. All photographs were taken by the author. Each panel was photographed as a unit, with occasional supplementary photographs taken to illustrate particular features or elements. Scales were included in all photographs, enabling future researchers to work with element size. Whenever possible, photographs were taken in plan view for each panel. Natural light was used for all photographs. Some panels were shaded to optimize the lighting available. Color slides of most of the panels at the Crash Landing site were taken during my initial field visit under suboptimal conditions due to wetting of some of the boulder surfaces from a light rain. This initial set of photographs was useful for projection to permit scaled, highly accurate drawings of each panel. These drawings were taken out to the field on my second visit and field checked and corrected as necessary under the improved field conditions of the second visit. A new set of photographs, both in color and black-and-white, were taken, and used as the archival set for the project given their much greater legibility. Boulder A-2, recorded during a later visit to the site, was sketched in the field and then cross-checked for scale and accuracy via the photographs taken. RESULTS Discussions of the sites visited in the project vicinity are offered here. Most attention is focused on the Crash Landing site; but brief summary descriptions of the other three rock art sites are also provided. Comparisons with sites and regions further afield are offered following this section. The Crash Landing Site The focus of this study is a small ceramic period complex of agricultural features, a two-room masonry field house, and four rock art boulders located on a low flat-topped basaltic ridge 800 m 8 west of Horseshoe Dam. For a more complete discussion of this site, see Volume 2. I will concentrate on the aspects of the site directly pertinent to the petroglyph discussion. Dosh and Henderson (1990) documented what is labeled here as Boulder A-1 and noted the presence of a possible “partial enclosure,” formed in part by the petroglyph boulder. They also noted one petroglyph boulder in the portion of the ridge where most of the agricultural features are located; what I recorded here as Boulder IS-2. The site is situated in a prominent location, with excellent views across the Verde Valley to the north, east, and southeast. Large habitation sites ranging throughout the ceramic period and canalirrigated fields are located in the valley below the site. Immediately north of the larger ridge upon which the site is situated, is the mouth of Lime Creek and the concentration of large Classic period habitation sites found there. This larger ridge is what forms the constriction of the valley that was used to facilitate the construction of Horseshoe Dam. Large zones of arable land are found north and south of the site along the Verde River. The river was formerly half a kilometer distant at its closest point northeast of the site and water was also potentially available from springs less than a kilometer distant to the southwest. The small ridgetop where the site is located is variably strewn with cobbles and occasional boulders, though much of the rock cover has been modified for agricultural or other prehistoric pursuits. Small boulders and cobbles also litter the slopes. Petroglyph Boulder IS-2 (Feature 19) is situated partway down the rocky northeast slope of the ridge, whereas Boulders IS-1 (Feature 267) and Cluster A (Features 294 and 295) are situated atop the ridge on relatively level terrain. Boulder IS-2 (Feature 19) is the largest of the boulders in its vicinity on the rocky slope. No cultural features were found nearby. The designs on Boulder IS-2, Panel A may have been visible for some distance downslope if the slope was clear of vegetation below it. At present, it is obscured. Although agricultural features are present within 20 m of Boulder IS-1 (Feature 267), no direct relationship can be drawn such as would be possible if the boulder was incorporated within the rock pile field. Petroglyph Cluster A is situated immediately adjacent to a modern trail that follows the very narrow ridgetop at this point and then heads down off the toe of the ridge. The geomorphic surface of the ridge is too unstable to conclude that this was necessarily also a prehistoric trail, but the setting would seem to support such an interpretation. As noted by Dosh and Henderson (1990), there is a sort of “partial enclosure” formed just southwest of Boulder A-1 (Feature 294) by the boulder itself and several others nearby downslope. I could find no evidence that any of the 9 other boulders in the vicinity had been moved in prehistory (though this could not be conclusively demonstrated), leaving the interpretation of a feature here problematic. It appears to be a natural clearing, perhaps accentuated by some rock removal. The presence of a lithic core and a few lithic flakes nearby suggest activities in the vicinity, but no clear association of the artifacts could be drawn. The 25 petroglyph design elements are situated on 11 panels distributed among four moderatesized boulders of basalt. Tables 1 and 2 provide the raw data recorded on them. The boulders and panels vary in size, shape, height, orientation, and surface texture, although all are prominent and among the largest or most obtrusive in their areas (discounting the effects of modern vegetation). Boulders range in maximum height from 23 cm to 52 cm and from 65 cm to 290 cm in maximum horizontal dimension. Panels range from 20 cm to 90 cm in width and 20 cm to 64 cm in height. By token of the nature of the boulders, all panels are on sloping surfaces, though they range in inclination from 12 to 75 degrees from horizontal (see Table 1). Because of the rounded nature of some of the rock surfaces, the panel designations used here are sometimes somewhat arbitrary. With angular-shaped rocks, panels are clearly demarcated by rock morphology; when more rounded rocks are involved, points of inflection on a rock’s sloping surface may be used to split out panels or more arbitrary sections of a rock may be identified as a panel in order to facilitate mapping and photography if no points of demarcation are readily apparent. The characteristics of the rock, rather than a need to place designs with particular orientations, seems to be the guiding factor for design placement. In each case the rock surfaces that could be most easily pecked are the ones with the designs. In some cases, even the smallest panels on the rocks were used (Panels IS-1b, A-1c), indicating a preference for these particular boulders rather than other boulders nearby that had larger surfaces. In general, there were not many other boulders on the ridge of comparable size to Boulders IS-1 and IS-2, but many boulders were available that were the same size or larger than Boulder A-2. In most cases, petroglyph panels at the site are not oriented in particular directions or towards particular cultural features that could be discerned. Because all but one of the boulders had multiple panels, it is clear that although some of the panels may have been intentionally placed to face a particular cultural feature or activity space, this could not have been the case for all panels. In the case of Boulder A-1, the principal designs were situated on a relatively horizontal surface and were situated immediately adjacent to a natural travel corridor, suggesting that they were meant to be seen when coming up onto the ridge or when leaving it to head down towards the 10 river. As noted above, the most obtrusive designs on Boulder IS-2 may have been visible for some distance to individuals looking up the slope of the ridge, provided vegetation did not obscure it. Boulder IS-1’s designs are not markedly noticeable from a distance and are unlikely to have served a signaling function to anyone other than someone already familiar with its shape and general location. All petroglyphs exhibited some chemical and mechanical weathering and no rock varnish formation. As such, they were recorded as Patina Class 2. The petroglyphs recorded are illustrated in drafted form in Figures 7 through 15 to elucidate the designs which often are not clear enough to see in a black-and-white image. Many of the petroglyph designs present were not well-pecked and were not readily identifiable as seen in these illustrations. Many areas of amorphous, unidentifiable pecking were present and it was sometimes difficult to decide what should be pulled out as separate designs. As such, the number of design elements coded here must be considered a best guess rather than a hard count. Photographs of some of the panels are provided in Figures 16 through 19. Table 3 summarizes the design element tally for the site as a whole. The designs are relatively equally split between life forms, geometric designs, and designs that could not be classified. The curvilinear, meandering quality of Panel A1a, the presence of variations on the circle (bulls-eye and tailed circle), and the use of stick figure anthropomorphs are all common features of the Hohokam (or Gila) Petroglyph Style (Schaafsma 1980; Thiel 1994; and Wallace and Holmlund 1986; ). Of particular interest is the anthropomorphic figure IS-2a:3, illustrated in Figures 13 and 19. The depiction of the head on this figure seems to indicate a particular hairdo or mask I have not seen before in southern Arizona. The depiction might be an indication that this figure is really a horned lizard, and that these are some of the cephalic “horns.” Arguing against such an interpretation is the fact that the “arms” bend to the rear rather than to the front as they do on living horned lizards. As the depiction is not necessarily even of something in the natural world, it would be unwise to say much more in this regard. Nearby Petroglyph Sites Three sites near the Crash Landing site were visited to provide comparative data and aid in establishing the chronological and cultural setting of the rock art. Drawings and notes from two additional sites in site survey records supplement the sites visited. Brief descriptions of the sites 11 are provided here. Additional data are incorporated in subsequent sections of the report. Horseshoe Mesa, AZ U:2:27/01-187 This is a spectacular site situated on a high, flat-topped mountain located approximately 2 km southwest of Horseshoe Dam on the west side of the Verde River (see Volume 2). The mountain rises some 120 m above the Verde River floodplain to an elevation of 2,400 feet (731.5 m) above sea level. The following description is excerpted from Schoonover (1979): The mountain is capped by a basalt flow roughly 10 feet thick, which has produced a remarkably level surface of perhaps a couple of acres in extent. The basalt cap has also served to protect the underlying less resistant pyroclastics and poorly consolidated gravels from erosion, thus preserving the distinctive topographic feature. Jointing in the basalt, resulting from cooling of the molten lava, has produced many smooth surfaces, some up to 15 square feet in size, which have facilitated the inscribing of the petroglyphs. Weathering along the periphery of the cap has dislodged angular blocks of basalt, some of considerable size, from the main mass. These blocks have slumped part way down the precipitous slopes, remaining in unstable positions. The petroglyphs at the site are concentrated on the basaltic cliff faces on the eastern exposure of the mountain top facing the river, and on the dislodged boulders that form a talus slope below the cliff faces. A series of small additional clusters are found on the western gentle slope of the mesa top. The views from this site are spectacular, encompassing a full panorama of the Verde Valley (Figures 20 and 21). A total of 168 petroglyph elements were recorded at the site by an Arizona Archaeological Society rock art class under the direction of Grace Schoonover (Schoonover 1979). Although many other elements are thought to be present, the data recorded are a useful indicator of the element categories present. The information is reproduced here in Table 4, using the categories of this study. Many panels have multiple elements and many of the designs are quite elaborate and well pecked. Figures 22 through 30 provide representative examples from the site. Located on the level top of the mountain on the northern edge is a masonry-walled circular enclosure 25 m in diameter. The rubble from the wall of this enclosure is substantial, averaging nearly a meter in height and 6 m in width. Although I use the terms “wall” and “rubble” here, these may be inaccurate. There is no evidence of coursed masonry or a carefully laid foundation and it is possible that there never was a formal wall. Attached to the exterior of the enclosure is 12 a masonry structure. Virtually no artifacts are present in association with the enclosure or structure. Several petroglyph panels are present on the mountaintop, including the dramatic panel seen in Figure 30. Also present atop the hill are several rock alignments and a disjunct set of two masonry structures. Several other localities on the benches down the slopes of the mountain also have masonry structures, none of which were examined during the course of this study. Site AZ U:2:96/01-259 This is a small masonry compound with at least six rooms and a plaza, situated on a low bench below the eastern face of the mesa on which AZ U:2:27 /01-187 is located (Figure 31), and just west and above the Cow Wallow site (AZ U:2:61/01-1152), excavated by SRI as part of the LVAP. The bench is formed from a small basaltic outcrop and the compound walls are made of basalt boulders. The bench, approximately 30 m above the floodplain at an elevation of 1,960 feet (597.4 m) above sea level, commands an excellent view of the Verde Valley floodplain. Petroglyphs were seen on a handful of panels both on isolated boulders and on portions of the basaltic outcrop. Some of the panels are interspersed within the architecture of the compound, but most are on the outcrop to the north. Design elements included a significant percentage of anthropomorphic designs, as well as zoomorphs and a well-pecked spiral (Figure 32). The boulder seen in Figure 32 had a minimally ground metate slick on the top. No more than 30 elements are present; I did not tabulate actual numbers or counts of motifs. Site AZ U:2:188/01-73 This is a large petroglyph site situated on the basaltic rim of a high mesa top overlooking the Verde Valley 2.5 km south of Horseshoe Dam on the west side of the Verde River. Its geological setting is similar to that of site AZ U:2:27/01-187 described above, though noticeably lower in elevation at 2,040 feet (621.8 m) above sea level, and shares a spectacular view of the valley. Its prominent views are to the north and east, and it is immediately across the river from site AZ U:2:129/01-176 (the McCoy Mesa ball court site) and a string of sites of later time periods. Petroglyphs are found primarily on the basaltic cliff faces at the northeastern tip of the mesa. The designs are fewer in number and less elaborate than those seen at AZ U:2:27/01-187, though the range of motifs is comparable: well over a hundred designs are present on at least 57 panels. A 13 sizable masonry pueblo ruin is located on a ridgetop a short distance northwest of the site. Four masonry rooms, three of which could be part of a compound, are located on the mesa top immediately adjacent to the glyphs. The site has suffered significant vandalism, with some panels removed and others damaged with graffiti. The road access to the site and the fact that it is marked on the USGS 7.5 minute topographic map of the area has no doubt contributed to this problem. Site Survey Data for Other Nearby Sites with Petroglyphs A total of 27 additional sites in the SRI study area are reported to have petroglyphs according to site survey records at ASU, ASM, and the Tonto National Forest. The majority of these sites are located in the Horseshoe Reservoir area and most have associated cultural features such as evidence of habitation and agricultural features. Most notable among these records is the repeated association of masonry architecture on mountain and ridge tops with petroglyph designs. For the ten sites that are tentatively dated, nine are assigned to the Classic period and only one is given a pre-Classic date range. The data available on the rock art at these sites is generally very limited. Records from seven of the sites included sketches or mapped drawings of design elements, and photographs are available for several (but were not seen by the author). The seven sites with data are: AZ U:2:28 (ASU)/01-188, AZ U:2:39/01-745, AZ U:2:120/01-801, AZ U:2:133/01-810, AZ U:2:249/01-261, AZ U:2:147 /01-264, and AR-03-12-01-1240 (TNF). Several of these sites are located very near to the sites discussed above. Based on the site survey records, all of the sites have scroll or scroll-related designs and six of the seven have anthropomorphic figures. Complex connected scroll designs are present at two of the sites. Also notable are full-body anthropomorphs at three sites. REGIONAL COMPARISONS, STYLE, AND CULTURAL IDENTITY The comparative database for the rock art at the Crash Landing site is extremely limited for neighboring regions. Although a wide range of sites from the Verde Valley, other than the Crash Landing site, have been recorded in the site files of various institutions, only the Horseshoe Mesa site (Schoonover 1979), discussed above, and the Orme Reservoir vicinity at the confluence with the Salt River (Kearns 1975) have received serious attention from the archaeological community. The Orme Reservoir survey located 36 rock art sites, the general setting and characteristics of which 14 are reported by Kearns (1975:313-331). Not enough data are provided to allow for serious comparisons of style. Some limited work has been accomplished in the neighboring Cave Creek drainage to the west. Sites in that area have been recorded by Rodgers (1977a, 1977b, 1985), Bruder (1983), and Lindauer (ed., 1992), though none of these authors provide published illustrations or detailed descriptions of the material in the area. Recent work along Roosevelt Lake has resulted in the documentation of the petroglyphs at the Bass Point site (Loendorf et al. 1994), but other than preliminary sketches, the results of this analysis were not yet available for consideration in this study. Several panels were documented at the Meddler Point site, AZ V:5:4/01-26, as a result of the work by Desert Archaeology, Inc. (Craig and Clark 1994), but they were small and of little comparative value. Somewhat further afield, Fernstrom (1984), Bruder (1983), and Rodgers (1985) have reported on petroglyph sites in the New River, Skunk Creek, and Agua Fria areas. In the Phoenix area, detailed studies have also been limited, despite the extensive distribution of sites in the area. Bostwick’s (1989) work at Moon Hill and Weaver’s (1985) work in Hieroglyphic Canyon in the Superstition Mountains are notable exceptions. Other data from the region are limited to various synthetic articles such as Snyder’s (1966) work at South Mountain, Schaafsma’s (1980) summary of the Hohokam Petroglyph Style (Gila Style), Weaver and Rosenberg’s (1978) study of the Sierra Estrella, and Hamann and Hedges’ (1986) work along the Gila River. I have visited a significant number of sites in the South Mountains and along the Gila River, but have limited direct knowledge of areas north and west of Phoenix and have not personally visited the sites that have been located along the Verde Valley other than those discussed here. To the north, a range of rock art sites have been reported on Perry Mesa and sketches and photographs of some of the panels have been made available to the author from Scott Wood and Patricia Whitley. A recent survey near the town of Prescott by Andrew Christenson (1994) netted 11 sites with over 273 petroglyph-bearing boulders and cliff areas. Although overlap is seen in the illustrations of the rock art from these sites with the Hohokam and Western Archaic styles, there are also distinct differences. Taken with the illustrations provided on other Prescott area sites by Barnett (1981:24-32), there are more similarities in the rock art of this region to Patayan sites along the lower Gila River than there are to Gila Style petroglyph sites in central and southern Arizona. Much more work in the area will be required to evaluate this proposition. The rock art in the Payson/Starr Valley area is sparse and cannot offer much in the way of comparative stylistic material (Owen Lindauer, personal communication 1994). 15 The limited nature of the data available precludes a comprehensive discussion of the Horseshoe Dam area sites in their regional context. What can be offered is a gross-level discussion of how the sites relate to rock art in the Phoenix/New River/Cave Creek areas, and to rock art on Perry Mesa. The starting point in this endeavor is an examination of the Crash Landing site in relation to the other sites that were investigated. Horseshoe Reservoir Area Rock Art Comparisons The four rock art sites examined in the Horseshoe Reservoir area all share significant characteristics. None of the sites exhibit evidence of designs that would fall under the Western Archaic style (Hedges 1982; Schaafsma 1980; Thiel 1994; Wallace and Holmlund 1986), and there is very little evidence that would indicate significant time depth at the sites. This is not unexpected given the setting of the sites. The rock art at the Crash Landing site is placed within and around what is likely to have been relatively short-term Classic period agricultural activity areas on the ridge. At site AZ U:2:96/01-259, the rock art is associated with a small Classic period masonry pueblo. The two mountain top sites would be the most likely to be multicomponent, but even in their case, Classic period occupations are located adjacent to the rock art. At the Horseshoe Mesa site, the rock art may be directly associated with the activities occurring on the mountain top. There is no definite evidence that the sites discussed are contemporaneous. Rock art researchers commonly turn to locally produced decorated pottery design styles for dated stylistic analogs. The paucity of locally produced decorated pottery precludes ceramic cross-dating as a feasible dating strategy. To a certain degree, there are broad regional patterns in the evolution of stylistic expression in central and southern Arizona and these can be referenced. The interlocking scroll complexes are suggestive of those seen on Sedentary and Classic period ceramics in the Phoenix Basin, but the curvilinear aspects of many designs, such as those on Boulder A-1 (Feature 294) at the Crash Landing site would, going by gross trends in Phoenix Basin ceramic design, suggest a Colonial period association. My research at other sites and areas in southern Arizona has led me to be cautious in making such direct interpretations. Rock art design conventions do not follow the same sequence as that seen on pottery (Wallace and Holmlund 1994), and unless particularly diagnostic motifs are present (and if they are present, I did not find them), one cannot draw meaningful conclusions. The best that can be said right now concerning the sites is that the majority of designs appear 16 similar stylistically (though larger and more complex panels are present on Horseshoe Mesa); the designs tend to only be minimally repatinated, with none observed bearing true rock varnish; and all, by association with other datable cultural remains, are likely to have a significant Classic period component. Another feature supporting contemporaneity, and by inference, stylistic cohesiveness, is the presence of anthropomorphic figures with “full” bodies (i.e., thickened lines or filled-in oval or round pecked areas, such as that seen in Figures 7, 8, 9, 13, 14). Although present at many other sites in southern Arizona, this form is usually vastly outnumbered by stick figures and figures with expanded midbodies. The anthropomorph with full body and curved-down arms almost touching the body (Element IS-2b:3 seen in Figure 14) observed at the Crash Landing site is a form repeated many times at the Horseshoe Mesa site. This is a relatively rare figure type elsewhere in the state and could indicate that the same people were actually producing designs at these two nearby sites. The presence of similar full-bodied figures at nearby sites AZ U:2:28/01-188, AZ U:2:133/01-810, and AZ U:2:249/01-261 may indicate the presence of a local stylistic tradition. Similar differences in figure-type expression have been observed between major rock art site complexes in different mountain ranges within south-central Arizona (Wallace and Holmlund 1986:121-125). Comparisons with the Phoenix Basin and Beyond Assuming that the Crash Landing site is at least roughly contemporaneous with the three other nearby rock art sites visited as part of this study, and that the limited data available from the site supports the interpretation that it is stylistically similar, it is now possible to draw some general comparisons on a larger regional scale. The Crash Landing site alone is too small to draw meaningful comparisons, but taken with the wealth of intricate panels at the Horseshoe Mesa site, and the information from the other sites investigated and the seven sites with data in the site survey records, some interpretations are possible. The most distinctive qualities of the Horseshoe Reservoir area sites may be summarized as follows: 1. Full-bodied anthropomorphs are common and full-bodied anthropomorphs with curveddown arms touching or nearly touching the body are frequent. 2. Repeated panels with “scenes” of rows of anthropomorphs holding hands (3 panels on Horseshoe Mesa) occur. 17 3. Zoomorphs with antlers (presumably deer) are frequently depicted to the near exclusion of sheep-like designs. 4. Elaborated scrolls, including interlocking, connected, and running forms are common. 5. Pipettes and dint patterns (see Wallace and Holmlund 1986:149-151, 225), both of which are commonly found on large sites in the Phoenix and Tucson areas, are absent. These qualities mark the sites as distinct and different from those currently known in the Phoenix area and sites further south. Whereas none of the traits cited are diagnostic in and of themselves, taken together, they present a package that has a slightly different look from that seen to the south and southwest. The differences are subtle, not dramatic, and the design styles present are closely allied to that seen in the Phoenix area and northern periphery. In the Phoenix area, panels with anthropomorph “scenes” are present, albeit rare; there are many complex scroll depictions, though they are not generally as common, and there are generally many more circular and curvilinear designs of other types. The large, well-pecked lizards at Horseshoe Mesa do occur elsewhere in the Phoenix and Tucson areas, though again, they are quite rare. The Hedgpeth site, located 53 km to the west-southwest in the Skunk Creek tributary of the New River drainage, is the nearest large, well-recorded site to those investigated here (Bruder 1983). It bears very little resemblance to the Horseshoe Reservoir sites, due for the most part, to a lack of contemporaneity. Contrary to Bruder’s initial assessment, the majority of designs at the site are believed to fall under the Western Archaic style, dating no later than A.D. 800 (Wallace and Holmlund 1986:86; Wallace et al. 1995). This interpretation is based on the design styles present, a reconsideration of the repatination data from the published report, and through field inspection. The other sites recorded by Rodgers and Bruder (Bruder 1983:62-82) in the Cave Creek, New River, and Agua Fria drainages west-southwest of the Horseshoe Reservoir site are of greater interest. Bruder provides brief site descriptions, summary descriptions of the petroglyphs present, and tallies of the motifs. The lack of published photographs of the designs limits the interpretations possible here, but even with the data available, it can be seen that there may be similarities to the Horseshoe Reservoir sites. Two sites, AZ U:1:23 and AZ U:1:24, together containing six petroglyph-bearing boulders, are closest to the project sites. They are situated near Elephant Butte in the mountainous country forming the headwaters of Cave Creek, less than 25 km to the southwest of Horseshoe Reservoir. The element tally listed by Bruder includes nearly as many deer as sheep and a nearly identical high percentage of scrolls (13 percent) to that recorded at 18 Horseshoe Mesa (see Table 4). The dating of the two sites is likely to be roughly contemporaneous with the project sites, as both are listed as Classic period components. Differing from the Horseshoe Reservoir sites are the presence of birds and a higher frequency of sheep than deer. These data leave open the question of stylistic association, and it would be necessary to see the sites or photographs of the designs to assess specific stylistic traits. At the other sites discussed by Bruder (1983:62-86), general similarities are present, as they are with most sites in the Phoenix area, but none have the high percentages of deer present at the Elephant Butte sites. The other sites do have relatively high percentages of scrolls, averaging 8.1 percent. This compares to 5.1 percent for the South Mountain database compiled by Snyder (1966) and the much lower percentages seen at most sites further south. Of the recorded sites in the Phoenix and Tucson areas with more than 30 design elements (Snyder 1966; Wallace and Holmlund 1986; Weaver 1985), only Signal Hill, AZ AA:12:63 (ASM), in Avra Valley exhibits comparably high values (13.6 percent; Wallace and Holmlund 1986:211). The significance of scroll percentages on a regional basis is difficult to assess because there is not a ready means to establish contemporaneity. The scroll symbol is virtually universal in rock art and does not by itself denote particular cultural affinities. The most that can be said is that the data may be indicating a greater affinity in rock art depictions among sites on the northern periphery of the Phoenix Basin than between these sites and the Phoenix Basin proper. Before concluding, I also want to briefly address other nearby regions that could potentially be relevant. The rock art at Perry Mesa in the upper Agua Fria drainage is very distinctive and very different from that seen in the Horseshoe Reservoir area (see Wood 1985). Birds, anthropomorphs with footprint feet, complex geometric forms, and shield-like framed motifs are all markedly different from the material seen either to the south or in the Verde sites. The Horseshoe Reservoir sites also bear no resemblance to the rare reported sites in the Mogollon country to the east (Burton 1988; Jernigan 1992; Schaafsma 1980; Thiel 1994; Weaver 1991a, 1991b, 1991c), or to the Anasazi sites to the north and northeast (Grant 1978; Schaafsma 1980; Thiel 1994). They also cannot be mistaken for styles present at the sites in the Prescott area illustrated by Franklin Barnett (1981:24-32), some of which bear a strong resemblance to the Patayan Sears Point style (Thiel 1994). 19 Discussion There are several critical missing pieces to this discussion that will require future field research to fill in. Almost nothing has been published regarding the rock art found in the Verde Valley north of Horseshoe Lake. It would be reasonable to develop hypotheses of cultural affiliation to be tested with rock art data in the Horseshoe Reservoir area if the area were well known archaeologically; but few data-oriented studies have been accomplished in recent years and little consensus on the culture history of the region has been achieved. Plus, theories of stylistic change and how it might relate to changing cultural scenarios are only weakly developed. Post-A.D. 1200 cultural manifestations in the middle and upper Verde Valley have been attributed to one of several scenarios: (1) Migration of Sinaguan groups into the area (Colton 1946), perhaps involving the absorption of indigenous populations (Schroeder 1960); (2) alteration of trade relationships among indigenous groups resulting in the appearance of changing populations (Fish and Fish 1977); or (3) pan-regional changes associated with the onset of the Classic period affecting indigenous populations (Tagg 1986). Perhaps most useful of the scenarios presented is the possibility that there may be strong cultural ties to the north and northeast in the middle Verde, leading to the possibility that design styles might be influenced in this direction (or at least be different internally from that seen further south in the Phoenix Basin). Supporting this possibility is Scott Wood’s (1987:50) observation that the middle Verde Valley-produced ceramic type Verde Red-on-buff bears motifs in the Classic period design style (Style 3, Wallace 1995), with some use of elements similar to those present on Flagstaff Black-on-white. As the motifs on Flagstaff Blackon-white are not clearly demarcated from those seen in Classic period style types in the Phoenix, Tonto, and Tucson basin areas, Wood’s suggestion must be treated with caution. The best that can be deduced from the information at hand is that if we were to look for a style that might characterize the middle and upper Verde regions, it would most likely be one that shares traits with the Phoenix Basin to the south and the Sinagua area to the northeast. These conclusions leave us standing in fairly murky water. In the Horseshoe Reservoir area almost no locally produced Classic period pottery is present, setting the area somewhat apart from other areas such as the middle Verde Valley, Tonto Basin (Wallace 1995), Phoenix Basin, Tucson Basin, Papaguería, lower San Pedro Valley, and Safford Valley, all of which were producing ceramics with the Classic period design style typified by the types Tanque Verde Red-on-brown and San Carlos Red-on-brown. The resemblances to the Phoenix Basin rock art seen in the 20 Horseshoe Reservoir area are probably significant, but how much the differences might relate to styles potentially originating further north or northeast is unknown. My guess is that the rock art of the Horseshoe Reservoir area is drawing from cultural affinities or religious beliefs that are originating further north in the Verde Valley. How much of this might simply be a local cultural development cannot presently be determined. It does seem safe to say that the motifs present are not simply those one would expect from migrants moving up from the Phoenix Basin. ROCK ART FUNCTION AND INTERPRETATION A variety of functions have been ascribed to rock art sites in southern and central Arizona, many of which can be readily discounted for the sites discussed. In previous studies, James Holmlund and I (Wallace 1983; Wallace and Holmlund 1986) discuss a range of potential functions and roles rock art may have played for sites in the Phoenix and Tucson areas. Readily excluded for the Horseshoe Reservoir sites are doodling or graffiti, mnemonic devices, decoration, and clan or totem symbols. Briefly, doodling or graffiti is excluded because of the lack of consistent direct association with habitation areas, the hidden nature of many panels, and the effort placed on some panels, which exceeds that which seems reasonable to expect for idle doodling (Busby et al. 1978; Wallace 1983:236). Were mnemonic devices present, it should be possible to identify their syntax; a task not proven possible in studies searching for it. The ethnographic data on Hopi clan symbol rock art (Colton and Colton 1931) suggests that repetitive groupings of certain totemic symbols would be anticipated if this was a primary motivation for rock art production in the lower Verde Valley. Although some panels do exhibit repeated designs, sometimes to the extent of occurring repeatedly on several adjacent boulders at the Horseshoe Mesa site, this is not a pattern seen at the Crash Landing site, nor is it the dominant feature of design element distribution at Horseshoe Mesa. Alan Ferg (1979:116) proposed that the petroglyphs on Tumamoc Hill in Tucson were made “for the enhancement of the petroglyph maker’s immediate surroundings.” If such a decorative function were foremost in the minds of the prehistoric artisans, then it would be expected that they would correlate very highly with activity areas and areas of habitation. Further, in areas with masonry architecture like the lower Verde Valley, one would expect stones used in construction or located nearby to be used on a regular basis for rock art depictions. Although the petroglyphs at site AZ U:2:96/01-259 are near a masonry compound, the designs at the other sites are not 21 directly associated with architectural remains and it cannot be said that they could be serving a purely decorative function. Indeed, most of the rock art at the sites with masonry architecture in the Horseshoe region with sufficient data to evaluate are located on cliff faces not directly visible from the architectural features located nearby. Many researchers have suggested that the rock art of central Arizona might be functioning as a signaling device, perhaps marking the location of trails; commemorating historical, mythical, astronomical, or cosmological events (or marking the location of such events); commemorating other ceremonial activities; or perhaps marking boundaries important in systems of land tenure (Bostwick 1989; Bruder 1983; Thiel 1994; Wallace 1983; Wallace and Holmlund 1986). One could include calendrical functions here. Not enough detailed recording work has been accomplished at Horseshoe Mesa and site AZ U:2:188/01-73 to fully evaluate each of these possibilities for those sites, but some preliminary observations are possible. The Crash Landing site is discussed separately below. In previous studies of rock art in the Phoenix and Tucson areas, a clear association of rock art and ritual or ceremonialism has been identified. At sites in the Samaniego Hills, Wallace and Holmlund (1986:174) report on an association of rock art with talus-slope burial cysts. Such an association is a strong indication of a ritual and/or commemorative role. In the Picacho Mountains at the Shelter Gap site (Wallace and Holmlund 1986), petroglyphs were found inside and at the mouth of small, difficult-to-access shelters. The private nature of the small shelters and difficulty involved in making designs led Wallace and Holmlund (1986:136-138) to argue that the shelters were sacred spaces that may have been used for a variety of possible functions including purification, the acquisition of personal power, and the storage of ritual paraphernalia. The presence of glyphs at the mouths of these small shelters was interpreted as possible boundary maintenance devices, indicating the private nature of the localities. Some evidence for ritual activity comes from the designs themselves. At a wide variety of sites in the Tucson and Phoenix regions, anthropomorphic figures are occasionally depicted with elaborate headdresses, holding items of material culture often assumed to relate to ritual activities, or engaged in dances or group activities that are likely to be related to ritual activity. Another indication that ritual, or at least social activities involving more than just the petroglyph artisan, were occurring at rock art sites in the Phoenix and Tucson areas is the presence of “bell” rocks. At some sites with the appropriate geological circumstances, large boulders are present that resoundingly “ring” when hammered with a small stone. Different pitches and 22 volumes may be obtained on different parts of some of these rocks. That the rocks were utilized in prehistory for this function is evidenced by the presence of battered margins and, in some cases, what have been called cupules or cupule/glyphs. Based on the chronology developed in the Picacho Mountains, they were utilized throughout the sequence. Their presence suggests that not only the “ringer,” but also an audience was present while they were being “rung.” It is not difficult to imagine ceremonies at the sites where the remarkably loud, clear chimes from these rocks could have made a major impact on the participants and the general atmosphere of the event. Rock art serving as a signaling device or boundary marker may also function in other roles, as seen in the case of the Shelter Gap site noted above. Sites where rock art is present in association with agricultural features such as canals and rock pile fields are common in the Phoenix area, and have been documented in the lower Verde region near Fort McDowell (Kearns 1975). Rock art sites in southern and central Arizona also commonly occur in association with wild-plant-food processing stations (Wallace 1983; Weaver 1985; Weaver and Rosenberg 1978), with lithic quarries (Wallace and Holmlund 1986; White 1965), and along trails (Bostwick 1989; Hayden 1972; Snyder 1966; Wallace 1983, 1989). The relationship of rock art to the types of activities implied by the cultural associations could have a range of possible interpretations. That it was likely serving a signaling role is relatively obvious in many cases, but the fact that many designs are often not visible from the nearby trail or processing area suggests other functions as well. Given the strong evidence for a ritual role for rock art at many sites and the evidence from ethnographic accounts in North America (Russell 1975; Whitley 1987), Africa (Lewis-Williams 1981), and Australia (Layton 1992) that overwhelmingly point to a ritual context for rock art in the societies that produce it, I suggest that even in these apparently secular settings, the rock art may be serving a function that relates to such diverse ritual themes as marking sacred localities, relating part of a relevant myth or cosmological event that has some bearing on the activities performed nearby, or requesting assistance from the cosmos to improve the harvest or facilitate the work to be performed. At the Horseshoe Reservoir sites, specific indications of ritual are evident at the Horseshoe Mesa site. There, the three panels with lines of hand-holding anthropomorphs are distinctive and almost certainly relate to ceremonial activities. The fact that the most distinctive of these panels, seen in Figure 30, is located on a relatively isolated boulder atop the mountain adjacent to the large walled enclosure may be an indication that this structure was serving a ritual role for people living in surrounding settlements. The mesa, given that it is the highest such peak in the region bordering the river, and that it is suited for mountain-top activities due to its flat top, would be a 23 prime candidate for a sacred site in prehistory. The range of architectural features present on the mountain are of some interest if this interpretation is correct. The lack of artifactual remains support an interpretation of limited or very short-term activity. This hypothesis could be tested through excavations in the structures atop the mountain. Note that this interpretation does not preclude a defensive motivation for the construction of the massive-walled enclosure. Ritual and warfare are inseparable in tribal societies. Hopi informants who visited the site and viewed the enclosure and rock art panel with the hand-holding figures, stated that this was a wuwtsim site, supporting a ritual interpretation (Scott Wood, personal communication 1996). The rock art at sites AZ U:2:188/01-73 and AZ U:2:96/01-259 is more problematic in terms of specific clues to their cultural role. By token of the mesa top setting for the former site, one could argue a similar function to the Horseshoe Mesa site, but this is purely speculative. For the latter site, general functional possibilities might include something related to ritual, marking historical events, and boundary maintenance. A ritual function is suggested by the presence of the grinding slick atop the boulder illustrated in Figure 32. These features are relatively ubiquitous on petroglyph sites across southern and central Arizona (Schaafsma 1980:96-97; Wallace 1989:123), and the current evidence strongly supports a function for them directly tied to the production or use of the petroglyph designs. Based on the presence of a group of these features on a precipitous cliff face with petroglyphs in the Gila Bend Mountains, I inferred that they were sometimes used in a group activity, and that only a token amount of grinding was necessary for the function sought (Wallace 1989:123). At the well-known Painted Rocks site, there is a boulder with a remarkably well-ground, oval flat-bottomed depression several centimeters deep that suggests that, on occasion, very formal grinding was performed at the sites. In both cases, the obvious specialfunction grinding stone at Painted Rocks and the innumerable slicks present at other rock art sites, indicate a ritual function. This may be nothing more than making a small food offering, or smearing a small quantity of pigment on a particular design to renew its “power.” In any event, the presence of this feature at site AZ U:2:96/01-259 offers supportive though not conclusive evidence for a ritual component in the use of the site and the production of the rock art. The petroglyphs at the Crash Landing site, could conceivably have served a variety of functions. As remarked by SRI project personnel as soon as maps became available, the petroglyph boulders neatly demarcate the limits of the agricultural component of the site on the north, south, and east sides. Because these are the three directions from which one would most likely approach the ridge, it could be argued that they are boundary signaling devices. As petroglyphs on the 24 boulders would have been potentially visible from the logical access routes onto the ridge, this might account for the presence of glyphs in these locations. Furthermore, the glyphs in Cluster A are situated along what was probably a prehistoric trail that led down the ridge towards the river. The interpretations are complicated by the fact that only a moderate subset of the designs on the four boulders would likely be visible to a casual visitor. Why place designs on portions of the rocks that would not be readily visible if the only concern was boundary maintenance? One possibility might be that although the designs do serve to signal the presence of the field system, thereby determining their placement, the designs chosen might relate to fertility or ritual requests for a good harvest, or perhaps to symbols that denote personal power of one sort or another. The specific process of design selection and illustration is therefore potentially a very different affair from the process of design placement. I know of no current means to deduce the actual processes involved in design selection and illustration other than to note that the cultural template involved appears to have been similar to that involved for illustrating the designs on Horseshoe Mesa and sites AZ U:2:96/01-259 and AZ U:2:188/01-73. A complementary interpretation for the Crash Landing site petroglyph boulders that would account for the limited visibility of the designs and yet support the boundary maintenance hypothesis is that the designs are primarily oriented towards marking property rights should disputes arise. This would be akin to the often difficult-to-see property markers used today. It is not necessary to see such a marker on a day-to-day basis, nor is it important that every visitor sees the boundary markers. What is important is the evidence of property rights established by the markers if the land is contested. The use of the largest boulders available might be anticipated if this interpretation is correct, as they could not be readily moved with the idea of surreptitiously changing property boundaries. SUMMARY AND CONCLUSIONS The four rock art boulders at the Crash Landing site were found to be stylistically similar to those found at Horseshoe Mesa, AZ U:2:96/01-259, and AZ U:2:188/01-73. Together, the sites exhibited no indications of superpositioning or distinct patination differences that would be indicative of significant time depth. Taken with the presence of Classic period architecture and, in the case of the Crash Landing site, a dry-farming field system, most of the designs at the sites are thought to date sometime in the A.D. 1100 to 1450 range. No evidence for protohistoric period 25 designs was observed, and no repecking was documented. This observation is based on the lack of definable historic design elements rather than a clear conception of what a protohistoric style might be in the area due to the lack of previous research in this regard. The rock art is stylistically most similar to the Hohokam Petroglyph style of the Phoenix and Tucson areas in terms of named and defined styles (Thiel 1994); however, some differences were observed that suggest a different local tradition was in operation. The presence of other sites in the Horseshoe Reservoir area and the nearby upper Cave Creek Valley that share some of the most distinctive stylistic traits identified in the visited sites supports the possibility of a localized stylistic tradition. It is currently unknown whether some influences from the north or northeast are present due to the lack of published data on the rock art in the critical areas. It is recommended that a reconnaissance of other known rock art sites in the area and particularly in the valley to the north and northeast be pursued to document the styles and associations present in this region. Quick field visits combined with photography would serve to begin this process. Such a study would help considerably to evaluate the style of designs seen in the sites discussed here. Rock art in the region is thought to function primarily in a ritual context, though other specific functional roles may often be involved. The documentation of historic events and the possibility of a boundary or signaling role may be involved for some of the localities observed. Regardless of the true function of the designs at the Crash Landing site, they were likely related to the agricultural pursuits seen on the ridge top. Acknowledgments I would like to thank the archaeologists at Statistical Research, Inc., for giving me the opportunity to work on the rock art of the Lower Verde. Special thanks are due Steve Shelley, who shuttled me in to the sites and took me on the grand tour of the area, and to Stephanie Whittlesey and Richard Ciolek-Torrello for strongly supporting this rock art research. I would also like to thank Jon Czaplicki, archaeologist at the Bureau of Reclamation, for taking me out to record a rock art boulder that one of the construction workers located at site AZ U:2:78 (ASM). He helped record the panel and provided a strong back to help move the boulder from the path of the earthmovers. Several individuals provided useful comparative information and shared unpublished manuscripts with me. These include Grace Schoonover, Owen Lindauer, David Jacobs, Scott Wood, and Andrew Christenson. I would also like to thank Jim Holmlund, David Jacobs, Scott 26 Wood, Sharon Urban, and Andrew Christenson for discussing the rock art of the region and generally providing useful insights. Review comments were received from Jon Czaplicki, Scott Wood, Stephanie Whittlesey, and Bill Doolittle. All were helpful and appreciated. The production and support staff at Desert Archaeology, Inc., made the generation of this report as painless as possible. Thank you Donna Breckenridge, Kara Myrick, Elizabeth Black, Catherine Gilman, and Chip Colwell for helping me get this together. REFERENCES CITED Barnett, Franklin 1981 These Were the Prehistoric Prescott Indians; A History of the Tenure of these Pioneers in Arizona. Arizona Archaeological Society, Yavapai Chapter, Prescott. Bierman, P., and A. Gillespie 1991 Accuracy of Rock-Varnish Chemical Analyses: Implications for CR Dating. Geology 19:96–199. Bostwick, Todd W. 1989 The Greenway Road and 17th Avenue Petroglyph Site (AZ T:8:102 [ASU]). Report No. PGM-8819. Pueblo Grande Museum and Cultural Park, Phoenix. Bruder, J. Simon 1983 Archaeological Investigations at the Hedgpeth Hills Petroglyph Site. Research Report No. 28. Museum of Northern Arizona, Flagstaff. Burton, Jeffery F. 1988 Prehistoric Rock Art of the Southeast Arizona Uplands: A Formal Record of 53 Rock Art Sites on the Coronado National Forest. Trans-Sierran Archaeological Research, Tucson. Reprinted in July. Originally published February 1988. Submitted to USDA Forest Service, Contract Nos. 40-8197-6-321 and 40-8197-7-268. Copies available from Coronado National Forest Office, Tucson. Busby, C., R. Fleming, R. Hayes, and Karen Nissen 1978 The Manufacture of Petroglyphs: Additional Replicative Experiments from the Western Great Basin. In Four Rock Art Studies, edited by C. W. Clewlow, Jr., pp. 89–108. Publications on North American Rock Art No. 1. Ballena Press, Socorro, New Mexico. Christenson, Andrew 1994 Archaeological Survey of the Prescott Lakes Project Area, Prescott, Arizona. Ms. prepared for the M3 Companies, Phoenix, Arizona. In possession of author, Prescott, Arizona. Ciolek-Torrello, Richard S., Stephanie M. Whittlesey, John R. Welch, Steven D. Shelley, and Jeffrey A. Homburg 1992 The Lower Verde Archaeological Project: Draft Research Design. Submitted to U.S. Department 27 of Interior, Bureau of Reclamation, Phoenix, Contract No. 1425-2-CS-32-01870. Statistical Research, Tucson. Colton, Harold S. 1946 The Sinagua: A Summary of the Archaeology of the Region of Flagstaff. Bulletin No. 22. Museum of Northern Arizona, Flagstaff. Colton, Mary Russell F., and Harold S. Colton 1931 Petroglyphs: The Record of a Great Adventure. American Anthropologist 33(1):32–37. Craig, Douglas, and Jeffery Clark 1994 The Meddler Point Site (AZ V:5:4/26). In The Roosevelt Community Development Study: Vol. 3: Meddler Point, Pyramid Point, and Griffin Wash Sites, by M. Elson, D. Swartz, D. Craig, and J. Clark. Anthropological Papers No. 13. Center for Desert Archaeology, Tucson. Dorn, Ronald I. 1983 Cation-Ratio Dating: A New Rock Varnish Age Determination Technique. Quarternary Research 20:49–73. 1989 Cation-ratio dating of rock varnish: A geographical perspective. Progress in Physical Geography 13:559–596. 1991 Rock Varnish. American Scientist 79:542–553. 1992 Comment on accuracy of rock-varnish chemical analyses: Implications for cation-ratio dating. Geology 20:470–471. 1994a Dating Petroglyphs with a Three-tier Rock Varnish Approach. In New Light on Old Art: Recent Advances in Hunter-gatherer Rock Art Research, edited by D. S. Whitley and L. L. Loendorf, pp. 13–36. Institute of Archaeology, University of California, Los Angeles. 1994b Surface Exposure Dating with Rock Varnish. In Dating in Exposed and Surface Contexts, edited by C. Beck, pp. 77–113. University of New Mexico Press, Albuquerque. Dorn, Ronald I., D. B. Bamforth, T. A. Cahill, J. C. Dohrenwend, B. D. Turrin, D. J. Donahue, A. J. T. Jull, A. Long, M. E. Macko, E. B. Weil, D. S. Whitley, and T. H. Zabel 1986 Cation-ratio and Accelerator Radiocarbon Dating of Rock Varnish on Mojave Artifacts and Landforms. Science 231:830–833. Dorn, Ronald I., and T. M. Oberlander 1981 Microbial Origin of Desert Varnish. Science 213:1245–1247. Dosh, Steven G., and T. Kathleen Henderson 1990 Site survey record for site AZ U:2:78. On file, Arizona State Museum, University of Arizona, Tucson. 28 Ferg, Alan 1979 The Petroglyphs of Tumamoc Hill. The Kiva 45(1–2):95–118. Fernstrom, Katherine 1984 Petroglyphs. In Prehistoric Cultural Development in Central Arizona: Archaeology of the Upper New River Region, edited by P. M. Spoerl and G. J. Gumerman, pp. 323–338. Occasional Paper No. 5. Center for Archaeological Investigations, Southern Illinois University at Carbondale, Carbondale, Illinois. Fish, Paul R., and Suzanne K. Fish 1977 Verde Valley Archaeology: Review and Prospective. Research Paper No. 8. Museum of Northern Arizona, Flagstaff. Grant, Campbell 1978 Canyon de Chelly, Its People and Rock Art. University of Arizona Press, Tucson. Hamann, Diane, and Ken Hedges 1986 Topographic Distribution of Hohokam Petroglyph Sites. Rock Art Papers 21:77–86. Harry, Karen G. 1992 Lithic Procurement and Rock Varnish Dating: Investigations at CA-KER-140, A Small Quarry in the Western Mojave Desert. Submitted to Bureau of Land Management, California Desert District, Contract No. N651-C1-3016. Copies available from Statistical Research, Tucson. Hayden, Julian D. 1972 Hohokam Petroglyphs of the Sierra Pinacate, Sonora, and the Hohokam Shell Expeditions. The Kiva 37(2):74–83. Hedges, Ken 1982 Great Basin Rock Art Styles: A Revisionist View. American Indian Rock Art 7:205–211. Heizer, Robert F., and Martin A. Baumhoff 1962 Prehistoric Rock Art of Nevada and Eastern California. University of California Press, Berkeley. Jernigan, E. Wesley 1992 Hour-Glass Rock Art Figures of Southeastern Arizona. Publication No. 4. Eastern Arizona College Museum of Anthropology, Thatcher, Arizona. Kearns, Timothy M. 1975 Rock Art. In An Archaeological Survey of the Orme Reservoir, assembled by V. Canouts, pp. 94–97, 313–331. Archaeological Series No. 92. Arizona State Museum, University of Arizona, Tucson. Krinsley, David H., Ronald I. Dorn, and Steven W. Anderson 1990 Factors that Interfere with the Age Determination of Rock Varnish. Physical Geography 11(2):97–119. 29 Layton, Robert 1992 Australian Rock Art: A New Synthesis. Cambridge University Press, New York. Lewis-Williams, J. D. 1981 Believing and Seeing: Symbolic Meanings in Southern San Rock Paintings. Academic Press, New York. Lindauer, Owen, (editor) 1992 Spur Cross Ranch Surveys. In The Archaeology of Sur Cross Ranch, Cave Creek, Arizona, edited by C. L. Redman and P. E. Minnis, pp. 17–31. Anthropological Field Studies No. 23. Office of Cultural Resource Management, Arizona State University, Tempe. Loendorf, Chris, David Jacobs, and Glen E. Rice 1994 Petroglyph, Grinding Slicks, and Cupules of the Rock Island Complex: U:8:392/862. In Where the Rivers Converge: Report on the Rock Island Complex, Roosevelt Platform Mound Study (Draft Report), edited by O. Lindauer. Roosevelt Monograph Series No. 4, Anthropological Field Studies No. 33. Office of Cultural Resource Management, Arizona State University, Tempe. Rodgers, James B. 1977a Archaeological Investigation of Granite Reef Aqueduct, Cave Creek Archaeological District, Arizona. Anthropological Research Papers No. 12. Office of Cultural Resource Management, Arizona State University, Tempe. 1977b A Comparative Petroglyph Analysis of the Northern Salt River Valley, Central Arizona. Paper presented at the 4th annual Symposium of the American Rock Art Research Association, Tempe. Ms. on file, Department of Anthropology, Arizona State University, Tempe. 1985 Rock Art Analysis and Regional Comparison of the Terrace Garden Site (AZ T:8:19). In Hohokam Settlement and Economic Systems in the Central New River Drainage, Arizona, edited by D. E. Doyel and M. D. Elson, pp. 823–845. Publications in Archaeology No. 4. Soil Systems, Phoenix. Russell, Frank 1975 The Pima Indians. Reprinted. University of Arizona Press, Tucson. Originally published 1908, Twenty-sixth Annual Report of the Bureau of American Ethnology, 1904–1905, U.S. Government Printing Office, Washington, D.C. Schaafsma, Polly 1980 Indian Rock Art of the Southwest. University of New Mexico Press, Albuquerque. Schoonover, Grace, editor 1979 A Rock Art Site Located in the Tonto National Forest Southwest of Horseshoe Dam, Maricopa County, edited by Grace Schoonover. Department of Certification, Arizona Archaeological Society, Rock Art Recording Class. Ms. on file, Tonto National Forest, Arizona. 30 Schroeder, Albert H. 1960 The Hohokam, Sinagua and Hakataya. Archives of Archaeology No. 5. Society for American Archaeology, University of Wisconsin Press, Madison. Snyder, Ernest 1966 Petroglyphs of the South Mountains of Arizona. American Antiquity 31:705–709. Tagg, Martyn D. 1986 The Tuzigoot Survey and Three Small Verde Valley Projects: Archaeological Investigations in the Middle Verde Valley, Arizona. Publications in Archaeology No. 40. Western Archaeological and Conservation Center, National Park Service, U.S. Department of the Interior, Tucson. Thiel, J. Homer 1994 Rock Art in Arizona. Technical Report No. 94-6. Center for Desert Archaeology, Tucson. Wallace, Henry D. 1995 Decorated Buffware and Brownware Ceramics. In The Roosevelt Community Development Study: Vol. 2: Ceramic Studies, Draft, edited by J. Heidke and M. Stark. Anthropological Papers No. 14. Center for Desert Archaeology, Tucson. 1983 The Mortars, Petroglyphs, and Trincheras on Rillito Peak. The Kiva 48(3). Wallace, Henry D. 1989 Archaeological Investigations at Petroglyph Sites in the Painted Rock Reservoir Area, Southwestern Arizona. Technical Report No. 89-5. Institute for American Research, Tucson. Wallace, Henry D., William H. Doelle, and James M. Heidke 1995 Hohokam Origins. Kiva 60(4):575–618. Wallace, Henry D., and James P. Holmlund 1986 Petroglyphs of the Picacho Mountains, South Central Arizona. Anthropological Papers No. 6. Institute for American Research, Tucson. Weaver, Donald E., Jr. 1985 Hieroglyphic Canyon. Monograph No. 1. American Rock Art Research Association, El Toro, California. 1991a Documentation and Evaluation of Four Rock Art Sites at Fools Hollow Lake on the ApacheSitgreaves National Forests Navajo County, Arizona. Plateau Mountain Desert Research, Flagstaff, Arizona. 1991b Documentation and Evaluation of Two Rock Art Sites at Chevelon Crossing and Mormon Crossing on the Apache-Sitgreaves National Forests Coconino County, Arizona. Plateau Mountain Desert Research, Flagstaff, Arizona. 1991c Documentation and Evaluation of Two Rock Sites along Pigeon Creek on the Apache-Sitgreaves National Forests Greenlee County, Arizona. Plateau Mountain Desert Research, Flagstaff, Arizona. 31 Weaver, Donald E., Jr., and Bettina H. Rosenberg 1978 Petroglyphs of the Southern Sierra Estrella, A Locational Interpretation. American Indian Rock Art 4:108–123. White, Cheryl Ann 1965 The Petroglyphs of Saguaro National Monument, Tucson, Arizona. Ms. on file, Arizona State Museum Library, University of Arizona, Tucson. Whitley, David S. 1987 Socio-religious Context and Rock Art in East Central California. Journal of Anthropological Research 6:159–188. Whitley, David S., R. I. Dorn, J. Francis, L. L. Loendorf, T. Holcomb, R. Tanner, and J.Bozovich 1996 Recent Advances in Petroglyph Dating and their Implications for the Pre-Clovis Occupation of North America. In Proceedings of the Society for California Archaeology, Vol.9, pp. 92–103. Society for California Archaeology. Wood, J. Scott 1987 Checklist of Pottery Types for the Tonto National Forest: An Introduction to the Archaeological Ceramics of Central Arizona. The Arizona Archaeologist No. 21. Arizona Archaeological Society, Phoenix. 1995 Winter Solistics Expedition to Perry Mesa. Ms. on file, Desert Archaeology, Tucson, and Tonto National Forest, Phoenix. 32
© Copyright 2026 Paperzz