Sources of Obsidian for the Sinagua of Northern Arizona (USA

Sources of Obsidian for the Sinagua of Northern Arizona (USA) John C. Whittaker Department of Anthropology, Grinnell College, Grinnell, IA 50112 USA*, Daniel Dong Keun Lee Grinnell College, Grinnell, IA 50112 USA, Lee Sharpe Department of Chemistry, Grinnell College, Grinnell, IA 50112 USA, and Jeffrey R. Ferguson University of Missouri Research Reactor Center, Columbia, MO, 65211 USA Preprint manuscript version for posting Academia.edu Published: Journal of Archaeological Science: Reports 21 (2018) 486–495 https://doi.org/10.1016/j.jasrep.2018.08.020 *Corresponding author John C. Whittaker, Department of Anthropology, Grinnell College, Grinnell, IA 50112 USA E-Mail address: whittake@grinnell.edu Abstract Obsidian was an important resource for prehistoric people in the North American Southwest. Elemental analysis of obsidian samples from Southwestern archaeological sites has been widely successful in identifying the raw material sources used by the people in the region. X-ray fluorescence (XRF) is the most commonly used method for sourcing obsidian artifacts because it is a rapid, non-destructive analytical technique with high discriminating ability. In this study, we analyzed over 450 pieces of obsidian from several Sinagua sites near Flagstaff, AZ. Obsidian debitage and unfinished points are common on these sites. The majority of points, and almost all the debitage, are from Government Mountain, the expected primary source. A few other sources have a minor presence, and it is likely that any points made of obsidian from these sources reached the sites as completed arrowheads, rather than being made there. Our limited data from survey finds and early points on late sites suggest that the pre-Sinagua Archaic populations were more mobile and used a wider variety of stone sources. Keywords: obsidian, sourcing, XRF, Sinagua, projectile points, Southwest, debitage 1.0 Introduction Identification of obsidian sources through chemical characterization has been a productive avenue of archaeological research for some 50 years. In the American Southwest, prominent obsidian sources were exploited over the whole span of regional prehistory, and their products widely spread. Our archaeological projects in northern Arizona focus on the prehistoric culture known as the Sinagua, who occupied a zone of cultural contacts not far from heavily-exploited obsidian sources. In sourcing a large number of obsidian artifacts from our sites, we hoped to see patterns of change through time and reflections of the Sinaguas’ place in regional systems of exchange and cultural contact, as previous work in the region has suggested (Brown, 1991; Horn-Wilson, 1997; Rondeau, 1979; Shackley, 2005). 1.1 The Sinagua One of the numerous archaeologically-defined cultures in the prehistoric North American Southwest, the Sinagua people occupied a relatively small but culturally central zone in Northern Arizona near the modern city of Flagstaff (Colton, 1946; Cordell and McBrinn, 2012; Kamp, 1998). Sites are defined as Sinagua on the basis of characteristic brown-ware pottery and early pithouse forms. Later Sinagua sites are mostly small masonry pueblo villages with pithouses continuing as part of the architectural suite. The Sinagua sequence begins somewhat before the eruption of Sunset Crater in the two decades after 1060 A. D. (Elson et al., 2007). The two centuries after the eruption saw the height of Sinagua population and culture (Pilles, 1988, 1996). The Sinagua were dry farmers, growing the usual Southwestern suite of crops in a volcanic landscape with little water (Kamp and Whittaker, 1999). Their agriculture was always marginal, and a climatic downturn after the 1250s led to the abandonment of most of the scattered small villages, and retrenchment of the population into a few larger sites. By about 1300, these sites too were abandoned, and the people appear to have moved south and west to large sites on Anderson Mesa and among the ancestral Hopi (Kamp and Whittaker, 2009; Pilles, 1988, 1996; Whittaker and Kamp, 2012). Several major archaeological cultures intersected in the Sinagua region of northern Arizona. The Sinagua received goods and influence from the Hohokam to the south, including raw shell and cut shell ornaments, and rare copper bells and macaws passed on from even further south. The Sinagua enjoyed a lively trade with the Anasazi (Ancestral Pueblo) communities to the north and east. About 10% of the pottery in our Sinagua sites is painted wares from the Kayenta and Little Colorado Anasazi (Kamp and Whittaker, 1999). The Anasazi appear to have had little interest in Sinagua brown-ware pottery, even though it was finely made, so something else was moving north in exchange for Anasazi pots. The Sinagua may have traded as middle-men in the exchange of shell and other southern goods, and perhaps also exchanged cotton, yarn, or cloth. Cotton seeds are common in our Sinagua sites (Hunter et al.,1999). The area north and west around the San Francisco Peaks was occupied by Cohonina groups (Garcia, 2004; Hanson, 1999; McGregor, 1951). Some ceramic vessels traveled both ways, but the most important obsidian sources were in the Cohonina cultural zone, at Government Mountain and RS Hill, approximately 30km or a hard day’s walk from the sites in our study. Obsidian is another potential exchange good for the Sinagua. It has long been recognized (Jack, 1971; Schreiber and Breed, 1971) that Government Mountain obsidian was one of the most widely used sources in the southwest, reaching Hohokam sites such as Snaketown and Palo Verde Ruin in southern Arizona (Shackley, 2005:156, 166), and into the Anasazi area as far north as Black Mesa (Parry, 1987:31). 1.2 Southwestern Obsidians Obsidian is a volcanic glass which may contain crystalline structures whose formation is dependent on the rate of cooling of the lava and its chemical composition. Low moisture (less than 2%), high silica (up to 77%), and alumina (up to 15%) are required for optimal obsidian formation (Shackley, 2005:14). Fast cooling produces obsidian, while slower cooling allows time for crystal growth, producing rhyolites and other volcanic rocks, some of which are sufficiently homogeneous and cryptocrystalline to also be exploited for stone tool manufacture. The best obsidian is glassy and homogenous from rapid cooling. Although trace elements vary usefully in obsidians, the quality of the obsidian is determined primarily by physical characteristics, rather than elemental composition. Both the flaking quality of the stone and the size and quantity of available pieces influenced prehistoric use. Shackley (2005) provides a thorough survey of obsidians in the Southwest. Only a few of the many known sources are relevant to us. The Sinagua lived close to two extensive volcanic fields, the San Francisco Volcanic Field, and the Mount Floyd Volcanic Field. Both of these produced obsidian from several chemically-distinctive sources (Figure 1). The San Francisco Volcanic Field includes the most important sources for the Sinagua. The obsidian at sources close to the San Francisco Peaks such as at O’Leary Peak and Robinson Crater is generally of poor quality with many phenocrysts, and was little used. Further west from Flagstaff some 20-25 miles (32-40 km), the Government Mountain and RS Hill sources were extensively exploited. The obsidians at Government Mountain and RS Hill are visually similar. They are not as glassy as some of the small marekanite (“Apache Tear”) obsidians from southern Arizona, but are of excellent flaking quality and occur in relatively large pieces easily collected from the surface or shallow diggings (Hanson, 1999). Shackley (2005) reports large nodules, up to 30cm, but these are exceptional. The largest pieces Whittaker has observed on the surface over the last 30 years are in the 10-20 cm size range, quite adequate for both early dart points and Sinagua arrow points. Even after prehistoric use and considerable modern collecting, surface obsidian is plentiful at these sources. The Government Mountain material occurs all over the substantial rhyolite dome at the source, and is widespread in the surrounding flats and washes. RS Hill obsidian is just as good as obsidian from Government Mountain, but is less common in most archaeological sites, probably because much less material is available in a more limited area on the slopes of that dome. Government Mountain obsidian was an especially popular tool stone, and has been found in Archaic through Late period contexts as far east as Chaco Canyon and as far south as the Mexican border (Shackley, 2005:21). Shackley suggests this is due to the large nodule sizes of the raw material, the overall quality relative to the nodule sizes, and good relationships between prehistoric groups. <Figure 1 about here> Even farther west of our sites is the Mount Floyd Volcanic Field, where the Black Tank and Partridge Creek/Round Mountain sources provide obsidian that is often glassier than the Government Mountain material and almost as large (Lesko, 1989; Shackley, 2005). The Presley Wash source includes grey and black glassy material as well as black subvitreous rhyolite. These sources are even more extensive than reported by Shackley (2005:30), and include pieces larger than from any other source in the region. The visually distinctive opaque but glassy grey obsidian occurs rarely in Sinagua sites, as does the black rhyolite. Further south in Arizona and western New Mexico, there are sources of good quality obsidian, such as Vulture and Superior in central Arizona, and Mule Creek in New Mexico. At many of these sources, especially the southern Arizona ones, the obsidian is very glassy but only available as marekanites, seldom as large as 10 cm diameter. The Mule Creek source area in western New Mexico was less extensively used than the large sources farther north in the Valle Grande, but also produced obsidian that shows up in Arizona sites. 1.3 X-Ray Fluorescence (XRF) X-Ray Fluorescence (XRF) is the most commonly used method for sourcing obsidian projectile points and debitage because it is a rapid, non-destructive analytical technique with high discriminating ability (Graves, 2005; Mitchell and Shackley, 1995;). Most lava flows exhibit unique trace element signatures that allow one to match a piece of obsidian to its geologic source. Identifying source locations for artifacts, combined with a basic technological analysis, gives information on behaviors of the people who lived on the site, such as whether they collected the resources themselves or if the material reached them as worked pieces. Once preferences for specific sources are identified, conclusions regarding patterns of lithic raw material procurement and exchange can be made. 1.4 Instrument and analytical methods X-ray fluorescence data was collected using a bench top ThermoFisher ARLTM QUANT'X EDXRF spectrometer located at the Missouri University Research Reactor at the University of Missouri, Columbia. The instrument uses a rhodium X-ray tube source which was held at 35 keV and typically 1.43 mA (the current varies somewhat in order to obtain an ideal dead time) with the beam collimated to a diameter of 3 mm. The resulting X-rays emitted by the samples were measured with a thermoelectrically-cooled silicon-drift detector. Artifacts were placed into either the 10 or 20 position auto sampler and irradiated for two minutes. The quantitative values were generated using a calibration developed at MURR (Glascock and Ferguson, 2012). The calibration uses 37 obsidian samples from sources all over the world chosen to encompass the range of values for the primary elements of interest (Mn, Fe, Rb, Sr, Y, Zr, and Nb) as well as to include varying elemental ratios. Elemental concentrations for the standards were based on analyses by neutron activation analysis (NAA), microwave digestion inductively coupled plasma mass spectrometry (MD-ICP-MS), laser-ablation inductively coupled plasma mass spectrometry (LA-ICP-MS), and XRF at other laboratories. The elemental concentration data was compared to data from source samples from all major and almost all minor obsidian sources known from the American Southwest. While there are numerous approaches to comparing compositional data from unknown artifacts to known source samples, we primarily rely on visual inspection of elemental scatterplots. Using a simple process of elimination it is possible to narrow down the potential source matches for artifacts in the Southwest to only one source for almost all samples. A plot of yttrium (Y) vs. niobium (Nb) concentrations (Figure 2) allowed us to clearly separate the source obsidians and thus our artifacts. <Figure 2 about here> 2.0 Archaeological Samples The material analyzed in this study came from four archaeological sites excavated by the Grinnell College Archaeological Field School under the direction of John Whittaker and Kathryn Kamp between 1984 and 2006 (Kamp and Whittaker, 1990, 1999, 2009; Whittaker, 1992; Whittaker and Kamp, 2012). All the sites are within a two-mile (3 km) radius, about 7 miles (11 km) east of Flagstaff, Arizona (Figure 1, Figure 3), in the Coconino National Forest. Artifacts from our projects are curated at the Museum of Northern Arizona. <Figure 3 about here> Lizard Man Village was completely excavated between 1984 and 1988. It was a multi-period habitation site with about a dozen pithouses of the Angell/Winona and Padre Phases (ca. 1060-1100, 1100-1150 A.D.) and two small blocks of later masonry rooms of the Elden Phase (1150-1250 A.D.). There were never more than a few families living there over the 200-year span of its occupation, but the midden deposits are as dense as any we have seen at a small site in the area. We excavated all or part of about 8 pueblo rooms and 15 pithouse structures, as well as numerous test trenches in exterior areas. Fortress Hills Pueblo is a similar site not far away, visible on its hillside location from Lizard Man Village. On the north edge is an early component, of which one Angell-Winona Phase pithouse was excavated. The rest of the site dates to the Padre and Elden Phases, and the pottery type proportions suggest a slightly later occupation than at Lizard Man Village. Four rooms of the single small roomblock were excavated, and one isolated masonry pithouse. The density of trash is similar to that at Lizard Man. Both these sites were chosen for excavation because they represent the typical small habitation sites of the Elden Phase Sinagua. New Caves Pueblo is a complex of single rooms and room blocks scattered over the slope of a prominent cinder cone mountain and in cavate rooms lining the crater. There are about 200 rooms all told. Bench Pueblo is part of this complex, on a bench half-way up the slope of the mountain. Our excavations sampled several of the New Caves roomblocks (about 15 rooms excavated), and 5 of the roomblock rooms and associated pit structures at Bench Pueblo. The New Caves complex dates from the end of the Elden Phase and into the succeeding Turkey Hill phase, ending before 1300 AD. In addition to our excavations, the field school has conducted intensive pedestrian survey of about 1670 hectares of the Coconino National Forest lands surrounding our sites (Figure 3), during most of which we collected points found as isolated occurrences and on unexcavated sites we recorded during survey. 2.1 Projectile Point Assemblage We worked with a projectile point sample of 501 specimens collected from all four sites and the surface survey work. All the obsidian points (343) were analyzed for this study. The assemblage of points from our four excavated sites came from both surface and excavated contexts, mostly from fill within rooms and middens (Table 1). In addition, a sample of 271 pieces of obsidian debitage from excavated contexts at the four sites was also analyzed and will be considered after the points. <Table 1 about here> All of the 343 obsidian points (68.5 % of a total of 501 points recovered) from our sites were analyzed. The other 158 points were made of a variety of materials, primarily high quality chert, chalcedony, and other crypto-crystalline materials (see Kamp and Whittaker, 1999:83-98; Kamp et al., 2016:14-16). These materials tend to be less localized, and although attempts have been made for a long time to chemically characterize sedimentary tool-stone sources by a variety of techniques (Hoard et al., 1993; Luedtke, 1992; Mehta et al., 2017; Newlander and Lin, 2017; Ray, 2007:46) these have shown mixed success and are not currently applicable in our area. 3.0 Results and Interpretations 3.1 Projectile Point Analysis We divided the points into 3 major groups: the small late arrow points of generally triangular form, larger older points, and points that were untyped. The untypable points included small fragments and many unfinished small arrow points. Most of the projectile points found at our sites are small late arrow points of a variety of forms, which we refer to as the Small Triangular Point Complex. Some of these point forms have chronological or cultural significance. The bow and arrow is expected to have replaced the atlatl and dart in the Southwest around 500-700 AD. There are suggestions of earlier bow and arrow use (Reed and Geib, 2013), but there is no evidence that atlatls survived into the period of our Sinagua sites (Whittaker, 2012). The larger, earlier points from our excavated sites were probably collected by later Sinagua people, and are often pre-ceramic forms that are recognizably much older than the occupation of our sites. These early large points more often show wear and damage, evidence that they were collected rather than made or lost on site, and it is likely that they were used symbolically, as are collected prehistoric points in ethnographic times (Whittaker and Kamp, 2016). In contrast, many of the small arrow points show typical manufacture breaks or are intact, and there are good examples of obsidian points in the process of manufacture at all of our sites (Kamp and Whittaker, 1999). As Table 1 shows, the majority of the obsidian points in our sample (310/343) were made of obsidian from the Government Mountain source. As Government Mountain is the closest source of plentiful good quality obsidian, this is not surprising. RS Hill obsidian is often visually similar to that from Government Mountain, and we expected RS Hill to be another popular source, as it is similar in distance and quality to Government Mountain. Although RS Hill was the second most used source, only nine points (2.6%) came from there, probably because less material was available. As expected, Government Mountain obsidian dominated our sample; however, obsidian variability was less than expected. Out of the 33 points that are not from Government Mountain, eight could not be assigned to a source. Another eight are from Presley Wash in northern AZ, while six are from Partridge Creek, formerly known as Round Mountain (Lesko, 1989). Both of these are northern Arizona sources in the Mount Floyd Volcanic Field, further west from our site area. One point came from the Superior source in Southern AZ and another from Mule Creek in New Mexico. The low variability in the assemblage, with only a small number of points from sources other than Government Mountain makes it difficult to identify clear trends, although several can be suggested. 3.1.2 Point Types and Obsidian Sources We classified the small arrow points into a number of different types (Figure 4, Figure 5; Table 2), using a system developed in collaboration by Whittaker and others working on sites in the Coconino National Forest (Kamp et al., 2016). Some of the categories used in this paper are inclusive of a more detailed type list used elsewhere (Kamp et al., 2016). Only the obsidian points are considered here. In addition to the 222 small obsidian arrow points contemporary with our site occupations, 30 larger preceramic points of obsidian were recovered, and 91 specimens that could not be assigned to a type. <Figures 4, 5 about here> <Table 2 about here> Of the later obsidian arrow points, 92% were made of Government Mountain obsidian, while only 70% of the obsidian pre-ceramic points were made of this material. We suspect that this reflects greater mobility on the part of the earlier people in the region. <Table 3 about here> Table 3 shows the number of obsidian points of each type, by site and obsidian source. Obsidian was plainly the preferred material for Sinagua projectile point manufacture. Some types, like our Sinagua Triangular, the Short Serrated group, and Sinagua Side-Notched, are local types that were rarely made of anything else. The Elden Side-Notched type is another common form, but in our sites these are made of fine cherts as often as obsidian. This form is frequently better made than the others and associated with burial and cremation contexts (Kamp et al., 2016). The Sinagua Triangular and Sinagua Side-Notched forms are also common in Cohonina sites to the west. The Red Lake group, which is a long slender point, often serrated, and sometimes with side-notches, has traditionally been associated with the Cohonina culture that occupied the area around the San Francisco Volcanic Field obsidian sources (Justice, 2002; McGregor, 1951), and these points are very common on sites in this district (Horn-Wilson, 1997), as well as in sites near ours (Rondeau, 1979). Both in the Cohonina territory and on Sinagua sites, they are mostly made of Government Mountain obsidian, and some were made on small natural tabular pieces of obsidian which occur at that source. Points made on these tabular pieces frequently retain cortex on both faces. The Red Lake and Sinagua forms could have arrived at our sites as finished points, but we believe that we have unfinished specimens broken in manufacture at our Flagstaff area sites, and the tabular pieces of obsidian also occur in our debitage. Although our ability to make fine chronological distinctions is weak within the excavated contexts, it has been possible to divide rooms and fills into phases, and we hoped to see temporal trends in obsidian use. The limited source variability and the small size of chronological subsamples restricts the potential for significant conclusions. Red Lake points are expected to be seen early in Sinagua sites, because the Cohonina culture area of which they are characteristic was supposed to have been depopulated by the mid-1100s (Garcia, 2004). Such a change in culture or population would have had implications for Sinagua access to obsidian, as all the good sources, especially Government Mountain, are in Cohonina territory or beyond it. The Red Lake forms occur on all our sites, including the latest. New Caves and Bench Pueblo in particular should have had little or no occupation before the 1200s, yet both have Red Lake points. It has been suggested that the Cohonina were an ecological specialization of part of the Sinagua population, or as a more likely alternative, that they were an offshoot of the Kayenta Anasazi populations around the Grand Canyon. In this view, when climatic downturns led to depopulation of the Cohonina heartland, the people moved east and blended into the Sinagua, especially east of the San Francisco Mountains and north of our sites, where there had always been a zone of cultural contact (Garcia, 2004; Hanson, 1999; Weintraub et al., 2006). Farther south, the Hohokam received obsidian from the San Francisco and Mount Floyd volcanic fields (Peterson et al., 1997). During the Sacaton Phase (900-1150 AD), the Hohokam site of Palo Verde followed the northern Hohokam Solares tool-making tradition defined by Hoffman (1997), and yielded a number of Red Lake points made of Government Mountain obsidian. Shackley (2005:164) considers this the “best evidence yet for the procurement of finished obsidian tools or arrows with hafted points in Sacaton contexts; in this case, tools from 200 km north.” Our sites produced four points that we classified as Hohokam Serrated. These have relatively large and careful serrations, unlike those typical of the Red Lake serrated variant, and our Short Serrated class. One is made of an unusual chalcedony, and the smallest is made of almost transparent obsidian from the Superior source in south-central Arizona. It closely resembles small Hohokam “Snaketown Serrated” points being made of that material and excavated from sites near the source (Shackley, 2005:156-161). It is possible that others among our Short Serrated category, although made of Government Mountain material, are of Hohokam manufacture. One Short Serrated point represents our most distant obsidian source, Mule Creek in New Mexico. It is an aberrant form, with two serrations on one edge and none on the other. At the end of the Sinagua occupation of the Coconino Plateau, much of the population seems to have moved west and south to several large sites on Anderson Mesa. Here they were in contact with a group of proto-Hopi sites around Homolo’ovi, near present Winslow, Arizona, and they may have become part of the developing ancestral Hopi population there. Both the Anderson Mesa and the Homolo’ovi sites used obsidian, primarily from Government Mountain (Harry, 1989; Young and Harry, 1989). Brown (1991), working in the Anderson Mesa area, believes that Government Mountain obsidian was an important trade good for these sites. As noted, the large pre-ceramic points showed more wear and damage, and were probably collected by people at our much later sites. Although there should be pre-ceramic habitation sites in our area, these have not been documented and even ceramic sites dating before the eruption of Sunset crater in the last half of the 1000s are rare (Elson 2011). A similar situation of widespread early points without definable habitation sites has been documented by Lyndon (2005) working in the Kaibab National Forest west of our study area. The distribution of materials among our old points is different from that of the late arrow points. A much higher percent of them (40 out of 70, or 57%) were made of non-obsidian materials, compared to 65 of 287 or about 22% of the small arrow points (Table 2). More of the obsidian pre-ceramic specimens (9 out of 30, or 30%) were made of obsidians other than that from Government Mountain, compared to only 17 out of 222, or 8% of the small points. Three of the 9 points of RS Hill material are early forms. The pre-ceramic populations are expected to have been more mobile, and perhaps had access to more sources (Shackley, 2005). The results of a sourcing study (Roberts et al., 2014) of 274 Archaic and Paleoindian points from the Kaibab National Forest suggest a similar pattern. The Government Mountain and RS Hill sources, as well as the Mount Floyd Volcanic Field sources are all in the present Kaibab National Forest. Pre-ceramic populations in this area also preferred the Government Mountain obsidian over all other igneous materials, but the Mount Floyd Volcanic Field sources were used much more often than in our area. Perhaps the most unexpected points in our sample are the Eden points. These are expected to date around 7440 to 6450 BC (Justice, 2002), and are generally associated with Plains bison hunters of this period located well to the east and north of our area. Nevertheless, one of our examples is an absolutely characteristic Eden mid-section, with small regular collateral flake scars and a diamond cross-section, made of a red chert. A less-exemplary tip is made of similar material, and the last is a well-made mid-section with one surviving ground shoulder and a severe impact fracture. This one is made of heavily patinated Government Mountain obsidian, and was a surface find in our surveys. Although we had none in our sample, Clovis points made of Government Mountain and RS Hill obsidians (Roberts et al., 2014), and of an unsourced material that is surely banded Presley Wash obsidian/rhyolite (Downum, 1993) are also known in our region, so these sources have been exploited as long as people have lived or traveled in northern Arizona. 3.2 Debitage Analysis Small flake tools are not uncommon on our sites, and about a third of them are made on obsidian flakes. Retouched tools other than projectile points are all quite simple and show little regularity in form, whatever their material. We did not include any in our obsidian analysis, but think it unlikely that they would show any different pattern. Our visual assessment was that almost all obsidian flake tools were made of Government Mountain obsidian (Kamp and Whittaker, 1999: 94-95). Obsidian made up about 11% of the debitage at Lizard Man Village. The 168 specimens analyzed are about 12% of the obsidian debitage. Comparable information is not available for our other sites. Obsidian debitage to be analyzed (N = 271) was selected from bulk lithic collections from Lizard Man Village, Fortress Hills Pueblo, New Caves, and Bench Pueblo. The debitage was chosen from as many as possible of the more controlled contexts excavated at each site, while attempting to get examples of all the visible variability and different technological types. Due to time limitations and the difficulty of sorting through collections now curated at the Museum of Northern Arizona, no statistically random sampling scheme was attempted. We remain confident that the results present a realistic picture. <Table 4 about here> The obsidian debitage from the sites was even less variable than the points, confirming our past visual assessments (Kamp and Whittaker, 1999:95). Of 271 specimens, there are only two specimens from Lizard Man Village, one from Fortress Hills, one from New Caves, and 2 from Bench Pueblo that were sourced to RS Hill. All the rest of the debitage came from the Government Mountain source. Table 5 shows the distribution of technological types of debitage at the sites, which is consistent with previous analyses (Kamp and Whittaker, 1999) and less quantified observations. At Lizard Man Village, obsidian made up about 11% of all debitage, with a slight trend toward becoming more common in later contexts. Biface thinning flakes are not uncommon, but of course the tiny waste from pressure flaking small points is under-collected. The flakes analyzed are all in the 1-4 cm range. Most of the flakes are irregular hard hammer flakes derived from the unsystematic knapping of small cores typical of all materials in our sites. Two of the Government Mountain pieces from Lizard Man in our analysis were the thin natural tabs of obsidian from that source that were sometimes used as blanks for the Cohonina Red Lake type points. The fairly high proportion of flakes and core fragments from bipolar knapping is intriguing. Bipolar reduction is performed by setting a small piece of stone on a hard anvil and striking it with a hammerstone, producing characteristic opposed scars. It is hard to control flake size and form using this technique, although it is considered a way to produce usable flakes from small pieces of stone that are difficult to work otherwise. Sinagua stone workers had access to better pieces of obsidian, as shown by the size of some discarded flakes, including some that could have been made into small points, but perhaps the bipolar pieces indicate some felt need to work obsidian to the uttermost. Other materials are only rarely subjected to bipolar knapping. <Table 5 about here> The lack of debitage from sources other than Government Mountain and RS Hill again suggests that obsidian from other sources arrived at our sites as finished points, rather than unworked obsidian pieces. 4.0 Conclusions Rather than travel large distances or engage in long-distance exchange for obsidian, the Sinagua exploited the closest sources, especially the plentiful and good quality obsidian at the Government Mountain source, the closest to our sites. They also used the RS Hill source, a neighboring source in the San Francisco Volcanic Field, to a lesser extent because it was more difficult to collect obsidian at this source. Sinagua relations with the Cohonina, the archaeological culture around the obsidian sources, are currently not well understood. They probably changed through time, but as far as we can tell, the supply of obsidian and the sources used did not change greatly. Although Government Mountain obsidian is the most common on Sinagua sites, and not just ours (Brown, 1991; Horn-Wilson, 1997; Rondeau, 1979), there were a few points in our sites that originated from other sources. Presley Wash and Partridge Creek, two sources that are further west, are represented, but were more often used by the earlier pre-ceramic populations. We suspect that the greater variety of obsidians in the early points reflects greater mobility and access to resources of the pre-ceramic folk. The single points of unusual form made of Superior and Mule Creek obsidian were likely acquired through exchange. Figures Figure 1. Map showing Sinagua sites and obsidian sources. Figure 2. Plot of yttrium (Y) vs niobium (Nb) concentrations showing source references (ovals) and artifact specimen source assignments. Only those sources with assigned artifacts in this study are shown. Figure 3. Map of study area showing our excavated sites, other prominent local sites, and Grinnell College Archaeological Field School survey area. Figure 4. Small arrow point types from Sinagua sites. The obsidian examples are typical. Figure 5. Further small triangular point types from Sinagua sites. The Red Lake, Sinagua Triangular, and Short Serrated forms are usually made of obsidian. Tables Table SEQ Table \* ARABIC 1. The number of points from each site; obsidian and non-obsidian specimens. Table 2. Projectile point types and their materials. XRF Source Total GM RSH PC PW Sup MC UNID Lizard Man Red Lake 12 1 13 Sinagua Triangular 30 1 1 32 Short Serrated 25 1 26 Hohokam Serrated 1 2 Long Triangular 5 5 Mogollon Triangular 1 1 Elden Side-Notched 9 1 10 Sinagua Side-Notched 11 1 12 Anasazi Side-Notched 9 1 10 Sinagua Corner-Notched 2 1 3 Gypsum Cave 1 1 Sinagua Ovate 3 3 Pinto 1 1 San Pedro 4 1 1 6 Cienega 1 1 Elko 5 5 Small Stemmed 1 1 Untypable 52 1 1 1 55 Total 171 2 2 5 1 1 4 186 Fortress Hills Red Lake 3 1     4 Sinagua Triangular 10     10 Short Serrated 3     3 Hohokam Serrated 1     1 Long Triangular 1     1 Elden Side-Notched 6     6 Sinagua Side-Notched 7     7 Anasazi Side-Notched 3     3 Other Side-Notched 2 1     3 Sinagua Corner-Notched 1 1     2 San Pedro 1 1     2 Bajada 1     1 Untypable 14     1 15 Total 53 1 2 1     1 58 New Caves Red Lake 3       3 Sinagua Triangular 7       1 8 Short Serrated 3       3 Hohokam Serrated 1       1 Long Triangular 5       5 Elden Side-Notched 4       4 Sinagua Side-Notched 4 1       5 Anasazi Side-Notched 2       2 Other Side-Notched 3 1       4 Sinagua Corner-Notched 1 1       2 San Pedro 1       1 Untypable 6   1     7 Total 40 3   1     1 45 Bench Pueblo Red Lake 1           1 Sinagua Triangular 7           7 Long Triangular 3           3 Elden Side-Notched 1           1 Sinagua Side-Notched 7           7 Anasazi Side-Notched 2           2 Other Side-Notched 2           2 Pinto   1         1 Untypable 10           10 Total 33   1         34 Survey Red Lake 3     3 Sinagua Triangular 1 1     2 Short Serrated 1     1 Long Triangular 1     1 Sinagua Side-Notched 1     1 Sinagua Corner-Notched 1     1 Basketmaker 1     1 2 Ovate Pinto 1   1 Pinto 1     1 San Pedro 1 1     2 Eden 1     1 Untypable 2 1     1 4 Total 13 3 1 1     2 20 Table 3. Obsidian projectile point types by site, with obsidian source determinations. Table 4: The obsidian debitage sample, composed of Government Mountain and RS Hills material. Table 5. Distribution of debitage pieces by technological type at each site. Acknowledgements Kathryn Kamp produced Figure 3. We are grateful to the many students who participated in the fieldwork that produced our data. Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References Cited Brown, G.M., 1991. Embedded and Direct Lithic Resource Procurement Strategies on Anderson Mesa. Kiva 50 (4), 359-384. Colton, H.S., 1946. The Sinagua: A Summary of the Archaeology of the Region of Flagstaff, Arizona. Museum of Northern Arizona Bulletin 22. Cordell, L., McBrinn, M., 2012. Archaeology of the Southwest, third ed. Routledge, New York. Downum, C.E., 1993. Evidence of a Clovis presence at Wupatki National Monument. Kiva 58 (4), 487-494. Elson, M.D., Ort, M.H., Anderson, K.C., Heidke, J.M., 2007. Living with the Volcano: The 11th Century AD Eruption of Sunset Crater, in: Grattan, J., Torrence, R. (Eds.), Under the Shadow: Cultural and Environmental Responses to Volcanic Eruptions. Left Coast Press, Walnut Creek, California. Elson, M. D., ed., 2011. Sunset Crater Archaeology: The History of a Volcanic Landscape: Prehistoric Settlement in the Shadow of the Volcano. Anthropological Papers No. 37, Center for Desert Archaeology, Tucson. Ferguson, J.R., 2012. X-Ray Fluorescence of Obsidian: Approaches to Calibration and the Analysis of Small Samples, in: Shugar, A.N., Mass, J.L. (Eds), Handheld XRF for Art and Archaeology. Leuven University Press. pp. 400-421. Frahm, E., 2016. Can I get chips with that? Sourcing small obsidian artifacts down to microdebitage scales with portable XRF. J. Archaeol. Sci. Rep. 9, 448-467. Garcia, D., 2004. Prehistoric Ceramic Boundaries in the Flagstaff Region of Northern Arizona. Unpublished MA thesis, Northern Arizona University, Flagstaff. Glascock, M.D., Ferguson, J.R., 2012. Report on the Analysis of Obsidian Source Samples by Multiple Methods. Manuscript on File at MURR. Graves, W., 2005. Obsidian Procurement among the Jumanos Pueblos, New Mexico, A. D. 1300-1670s. Kiva 71 (1), 7-35. Hanson, J., 1999. Cultural Affiliation Assessment: Cohonina. Unpublished report, USDA Kaibab National Forest, Williams, Arizona. Harry, K.G., 1989. The Obsidian Assemblage from Homol’ovi III: Social and Economic Implications. Kiva 54 (3), 285-296. Hoard, R.J., Bozell, J.R., Holen, S.R., Glascock, M.D., Neff, H., Elam, J.M., 1993. Source Determination of White River Group Silicates from Two Archaeological Sites in the Great Plains. Am. Antiq. 58 (4), 698-710. Hoffman, C.M., 1997. Alliance Formation and Social Interaction During the Sedentary Period: A Stylistic Analysis of Hohokam Arrowpoints. Unpublished PhD dissertation, Arizona State University. Horn-Wilson, A., 1997. Projectile Points of the Cohonina: A Pilot Study. MA Thesis, Northern Arizona University. Hunter, A., Kamp, K.A., Whittaker, J.C., 1999. Plant Use, in: Surviving Adversity: The Sinagua of Lizard Man Village, University of Utah Anthropological Papers Number 120. University of Utah, Salt Lake City, pp. 139-151. Jack, R., 1971. The Source of Obsidian Artifacts in Northern Arizona. Plateau 43 (3), 103-114. Justice, N.D., 2002. Stone Age Spear and Arrow Points of the Southwestern United States. Indiana University Press, Bloomington. Kamp, K.A., 1998. Life in the Pueblo: Understanding the Past Through Archaeology. Waveland Press, Prospect Heights. Kamp, K.A, Whittaker, J.C., 1990. Lizard Man Village: A Small Site Perspective on Sinagua Social Organization. Kiva 55 (2), 99-125. Kamp, K.A., Whittaker, J.C., 1999. Surviving Adversity: The Sinagua of Lizard Man Village. University of Utah Anthropological Papers Number 120. University of Utah: Salt Lake City. Kamp, K.A. Whittaker, J.C., 2009. A Sinagua Acropolis: Architectural Adaptation at New Caves, Arizona. Kiva 74 (3), 281-304. Kamp, K.A., Whittaker, J.C., Bryce, W., 2016. The Magician of Ridge Ruin: Insights from Bows and Arrow Points. Lithic Technology 41 (4), 313-334. Lesko, L.M., 1989. A Reexamination of Northern Arizona Obsidians. Kiva 54 (4), 385-400. Luedtke, B., 1992. An Archaeologist's Guide to Chert and Flint. Cotsen Institute of Archaeology Press, Los Angeles. Lyndon, M. G., 2005. Projectile Points as Indicators of Preceramic Occupation of the Coconino Plateau. Unpublished MA thesis, Northern Arizona University. McGregor, J.C., 1951. The Cohonina Culture of Northwestern Arizona. University of Illinois Press, Urbana. Mehta, J.M., McCall, G., Marks, T., Enloe, J., 2017. Geochemical source evaluation of archaeological chert from the Carson mounds site in northwestern Mississippi using portable X-ray fluorescence (pXRF). J. Archaeol. Sci. Rep. 11, 381–389. Mitchell, D.R., Shackley, M.S., 1995. Classic Period Hohokam Obsidian Studies in Southern Arizona. J. Field Archaeol. 22 (3), 291-304. Newlander, K., Lin, Y., 2017. Integrating visual and chemical data to source chert artifacts in the North American Great Basin. J. Archaeol. Sci. Rep. 11, 578–591. Parry, W.J., 1987. Sources of Chipped Stone, in: Parry, W.J., Christenson, A.L. (Eds.), Prehistoric Stone Technology on Northern Black Mesa, Arizona. Southern Illinois University Center for Archaeological Investigations, Occasional Paper No. 12, Carbondale, pp.21-42. Peterson, J., Mitchell, D., Shackley, M.S., 1997. The Social and Economic Contexts of Lithic Procurement: Obsidian from Classic-Period Hohokam Sites. Am. Antiq. 62 (2), 231-259. Pilles, P., 1988. Flagstaff Province Entries, in: Jelks, E., Jelks, J. (Eds.), Historical Dictionary of North American Archaeology. Greenwood Press, Westport. Pilles, P., 1996. The Pueblo III Period along the Mogollon Rim: The Honanki, Elden, and Turkey Hill Phases of the Sinagua, in: Adler, M.A. (Ed.), The Prehistoric Pueblo World A.D. 1150-1350, University of Arizona Press, Tucson, pp. 59-72. Ray, J.H., 2007. Ozarks Chipped Stone Resources: A Guide to the Identification, Distribution, and Prehistoric Use of Cherts and Other Siliceous Raw Materials, Missouri Archaeological Society Special Publication 8, Springfield, MO. Reed, P.F., Geib, P.R., 2013. Sedentism, Social Change, Warfare, and the Bow in the Ancient Pueblo Southwest. Evol. Anthropol. 22 (3), 103-110. Roberts, T.M., Skinner, C.E., Bryce, W.D., 2014. Hunter-gatherer Igneous Toolstone Procurement in Northern Arizona: A Geochemical Study of Projectile Points and Raw Material Sources. Kiva 80 (2), 193-212. Rondeau, M.F., 1979. Projectile Point Analysis for the Kahorsho Site: NA 10,937, Central Arizona. Unpublished MA thesis, California State University, Sacramento. Schreiber, J.P., Breed, W.J., 1971. Obsidian Localities in the San Francisco Volcanic Field, Arizona. Plateau 43 (3), 115-119. Shackley, M.S., 1988. Sources of Archaeological Obsidian in the Southwest: An Archaeological, Petrological, and Geological Study. Am. Antiq. 53 (4), 752-772. Shackley, M.S., 1995. Sources of Obsidian in the Greater American Southwest: An Update and Quantitative Analysis. Am. Antiq. 60 (3), 531-551. Shackley, M.S., 2005. Obsidian: Geology and Archaeology in the North American Southwest, University of Arizona Press, Tucson. Skoog, D.A., Holler, F.J., Crouch, S.R., 2007. Principles of Instrumental Analysis, sixth ed. Thomson Brooks/Cole. Weintraub, N., Sorrell, D., Hanson, J., Christensen, D., 2006. A GIS Model for Late Pueblo II/Cohonina Multi-directional Migration from the Coconino Plateau. Paper presented at the Annual Meeting, Society for American Archaeology, May 18, 2006. Whittaker, J.C., 1992. Hard Times at Lizard Man. Archaeology 45 (4), 56-58. Whittaker, J.C., 2012. Ambiguous Endurance: Late Atlatls in the American Southwest? Kiva 78 (1), 79-98. Whittaker, J.C., Kamp, K.A., 2012. The Troubled End of Pueblo Life in the Sierra Sin Agua: Evidence from the Site of New Caves, in: Downum, C.E. (Ed.), Hisat’sinom: Ancient Peoples in a Land Without Water. School for Advanced Research Press, Santa Fe, pp. 148-154. Whittaker, J.C., Kamp, K.A., 2016. Flint from the Ancestors: Ritualized Use of Stone Tools in the Prehistoric Southwest, in: Sullivan, A., Olszewski, D. (Eds.), Archaeological Variability and Interpretation in Global Perspective. University Press of Colorado, pp. 267-290. Young, L.C., Harry, K.G., 1989. A preliminary analysis of temporal changes in the Homol'ovi III chipped stone assemblage. Kiva 54 (3), 273-284. 30