Journal of Southwestern Anthropology and History. ISSN: (Print) (Online) Journal homepage:

Transcription

1 KIVA Journal of Southwestern Anthropology and History ISSN: (Print) (Online) Journal homepage: Implications for Migration and Social Connections in South-Central New Mexico Through Chemical Characterization of Carbon-Painted Ceramics and Obsidian Jeffrey R. Ferguson, Karl W. Laumbach, Stephen H. Lekson, Margaret C. Nelson, Karen Gust Schollmeyer, Toni S. Laumbach & Myles Miller To cite this article: Jeffrey R. Ferguson, Karl W. Laumbach, Stephen H. Lekson, Margaret C. Nelson, Karen Gust Schollmeyer, Toni S. Laumbach & Myles Miller (2016) Implications for Migration and Social Connections in South-Central New Mexico Through Chemical Characterization of Carbon-Painted Ceramics and Obsidian, KIVA, 82:1, To link to this article: Published online: 06 Jul Submit your article to this journal View related articles View Crossmark data Full Terms & Conditions of access and use can be found at Download by: [Myles Miller] Date: 06 July 2016, At: 14:57

2 kiva, Vol. 82 No. 1, March, 2016, Implications for Migration and Social Connections in South-Central New Mexico Through Chemical Characterization of Carbon-Painted Ceramics and Obsidian Jeffrey R. Ferguson, 1 Karl W. Laumbach, 2 Stephen H. Lekson, 3 Margaret C. Nelson, 4 Karen Gust Schollmeyer, 5 Toni S. Laumbach, 6 and Myles Miller 7 1 University of Missouri Research Reactor Center, Columbia, MO, USA fergusonje@missouri.edu 2 Human Systems Research, Las Cruces, NM, USA KLAUMBACH@comcast.net 3 University of Colorado, Boulder, CO, USA Lekson@Colorado.EDU 4 Arizona State University, Tempe, AZ, USA mnelson@asu.edu 5 Archaeology Southwest, Tucson, AZ, USA karen@archaeologysouthwest.org 6 New Mexico Farm and Ranch Heritage Museum, Las Cruces, NM, USA toni.laumbach@state.nm.us 7 Versar, Inc., Springfield, VA, USA epmyles@aol.com Magdalena Black-on-white ceramics from two sites (Gallinas Springs and Pinnacle Ruin) in west-central and southwestern New Mexico have been interpreted as evidence of a migration of Northern Pueblo groups from the Four Corners region into southwestern New Mexico during the thirteenth century. They also appear to be linked to sites with similar carbon-painted ceramics on the Rio Puerco of the east and beyond. An additional site (Roadmap Village) reveals import of Magdalena Black-on-white ceramics produced at Gallinas Springs as well as possible local production. Limited quantities of Copyright 2016 Arizona Archaeological and Historical Society. All rights reserved. DOI /

3 IMPLICATIONS FOR MIGRATION AND SOCIAL CONNECTIONS 23 carbon paint ceramics have been found on El Paso Phase sites in south-central New Mexico that have previously been attributed to contemporaneous carbonpainted pottery produced at communities in the Galisteo Basin and the upper Rio Grande. Recent compositional analysis of carbon-painted ceramics from the Gallinas Springs, Pinnacle, and Roadmap sites has identified characteristic chemical signatures that suggest local production of carbon paint ceramics at all three sites and distribution of carbon paint ceramics from Gallinas Springs to Pinnacle and Roadmap in the eastern Black Range of southwestern New Mexico. Analysis of carbon paint ceramics from Madera Quemada, an El Paso Phase site in the Tularosa Basin indicates that the carbon paint wares found in El Paso Phase sites were acquired through trade connections from the Black Range rather than from more northern sources. The overall Magdalena Black-on-white production patterns are contrasted with the obsidian procurement data from the same sites to reveal a complex and divergent pattern. La presencia de cerámica del tipo Magdalena negro-sobre-blanco proveniente de dos sitios (Gallinas Springs y Pinnacle Ruin) en el centro y suroeste de Nuevo México ha sido interpretada como evidencia de la migración de grupos Pueblos norteños provenientes de la región de Four Corners ( cuatro esquinas ) de Estados Unidos hacia el suroeste de Nuevo México durante el siglo XIII. Este tipo de cerámica también parece estar vinculado con cerámica con pintura de carbono que se encuentra en sitios al este del Río Puerco y más allá. Un sitio adicional (Roadmap Village) revela la importación de cerámica Magdalena negro-sobre-blanco producida en Gallinas Springs así como posible producción local. Cantidades limitadas de cerámica con pintura de carbono han sido encontradas en sitios de la fase El Paso en el centro-sur de Nuevo México y han sido previamente atribuidas a cerámica con pintura de carbono contemporánea producida en comunidades de la cuenca de Galisteo y del Río Grande superior. Recientemente, un análisis de la composición de cerámica con pintura de carbono de los sitios de Gallinas Springs, Pinnacle, y Roadmap identificó marcos químicos característicos que sugieren que la cerámica fue producida localmente en cada uno de los tres sitios. Sugiere también una distribución de cerámica con pintura de carbono desde Gallinas Springs hasta Pinnace y Roadmap en la Sierra del Diablo este (también llamada Sierra Negra en inglés), en el suroeste de Nuevo México. El análisis de cerámica con pintura de carbono de Madera Quemada, un sitio de la fase El Paso en la cuenca del Tularosa indica que la cerámica con pintura de carbono hallada en sitios de la fase El Paso fue adquirida a través del intercambio con la Sierra del Diablo y no con fuentes más al norte. Los patrones generales de la producción de Magdalena negro-sobre-blanco están contrastados con datos sobre la adquisición de obsidiana en los mismos sitios para revelar un patrón complejo y divergente

4 24 JEFFREY R. FERGUSON ET AL. keywords NAA, Ceramic exchange, Migration, Magdalena ceramics, Obsidian provenance Introduction Evidence for fourteenth-century migrations in the American Southwest varies from very clear to circumstantial. In southern Arizona, site unit intrusions mark very clear migrations of Kayenta groups to the Gila River valley, and beyond (Clark et al. 2014, Haury 1958, Stone 2000). These sites repeat in detail the kivas, domestic architecture, and decorated pottery of the Kayenta region, far to the north near Hopi. They contrast markedly with local, Classic Hohokam sites. At the other end of the Southwest on the upper Rio Grande, evidence for migration is more circumstantial: including a possible marked spike in population in the early fourteenth century. Certain ceramic types parallel broader ceramic changes noted for white wares produced across the Southwest, including those that parallel stylistic change noted in the Mesa Verde region, but also reflect a sequence of local development within the Northern Rio Grande. The principal evidence for migration is demographic, not material (see Ortman 2010 for additional lines of evidence of this migration). The Magdalena phase of west-central and southwestern New Mexico falls between the clarity of the Kayenta migrations into the Gila River valley and the more circumstantial suggestions for migrations into the upper Rio Grande. Magdalena pottery and architecture stand in marked contrast to local traditions, but many of the key hallmarks of Mesa Verde society keyhole kivas, mugs, and other characteristic artifacts are absent or not yet discovered (Lipe 2010). We hypothesize that this range of evidential clarity reflects social processes: ethnic enclaves of Kayenta peoples in southern Arizona, smaller groups deliberately blending in to existing villages in the upper Rio Grande, and the possibility that ethnic enclaves of carbon paint producing populations extended south to the eastern slopes of the Black Range in southwestern New Mexico. This latter group is interpreted as the primary source for carbon paint ceramics in the El Paso phase sites of southern New Mexico. We examine assemblages from three large late Pueblo III period sites with carbonpainted ceramics. From north to south these sites include Gallinas Springs on a tributary of the Rio Salado, Pinnacle Ruin on the upper Rio Alamosa, and Roadmap Village on the lower reaches of the Rio Palomas south of Truth or Consequences. We also examined a small sample of Magdalena Black-on-white sherds from Madera Quemada Pueblo, far to the southeast in the southern Tularosa Basin (Figure 1). When the compositional analysis of the carbon paint ceramics is compared to obsidian procurement patterns, divergent patterns of cultural contact among the site populations are revealed. Magdalena Black-on-white ceramics likely produced at Gallinas Springs Pueblo are found at all sites included in this study, however this movement does not appear to be reciprocal: that is, Magdalena Black-on-white ceramics likely produced at Pinnacle Ruin and Roadmap are not found in the sample from Gallinas Springs. The small sample of Magdalena Black-on-white from Madera Quemada Pueblo was likely produced at Gallinas Springs; however, the mechanism of exchange was most likely through contact

5 IMPLICATIONS FOR MIGRATION AND SOCIAL CONNECTIONS 25 figure 1. Locations of sites included in this study. with populations at Pinnacle and Roadmap in the Black Range. While some ceramic evidence indicates a link to Gallinas Springs, there is little evidence of exchange in obsidian. The Gallinas Springs obsidian assemblage is dominated by material from northern New Mexico, while Roadmap Village and the Magdalena component at Pinnacle Ruin reveal an obsidian assemblage predominantly from western New Mexico sources.

6 26 JEFFREY R. FERGUSON ET AL. Background Description and Discussion of Magdalena Black-on-white Magdalena Black-on-white was first recognized and described by Emma Lou Davis (1964, n.d.) from samples she examined at LA1178, the Gallinas Springs Site, which is a large masonry pueblo located about 60 miles northeast of the Pinnacle Ruin. It was she who first suggested a similarity with Mesa Verde pottery types and the possibility of immigrants from that region settling at Gallinas Springs. Helen Warren described an organic painted variety labeled as southern McElmo (Warren:1974). Gomolak and Knight also examined and reported on the Gallinas Springs Magdalena Black-on-white. Knight s (1981:2) initial interpretation of Magdalena Black-on-white was that it represents a local tradition developed over some length of time. Later, Gomolak and Knight (1990:8 24) changed their interpretation and attributed Magdalena Black-on-white at Gallinas Springs to proto-laguna and post Mesa Verde immigrants, referencing Laguna migration legends documented by Ellis (1974). Although some sherds from Gallinas Springs have design attributes of Mesa Verde Black-on-white (Knight 1981:55 57), the designs on Magdalena Black-on-white from Pinnacle Ruin are most similar to McElmo Black-on-white or Santa Fe Black-on-white from the middle Rio Grande. Like McElmo Black-on-white, Magdalena Black-on-white (Figure 2) is characterized by a creamy white to gray white slip that frequently appears crackled, an organic pigment, and polishing over both slip and pigment. Magdalena Black-on-white from the Pinnacle Ruin shares the McElmo design style that exhibits bold solids with opposed diagonal lined hatch lines. Multiple parallel lines, multiple stepped lines, appended dots from linear elements, and checkerboard patterns are common motifs. Curvilinear motifs are absent on Magdalena Black-on-white. Designs on both bowls and jars are framed within a body band that encircles the vessel. Like McElmo Black-on-white, rim ticking is rare to absent on sherds of Magdalena Black-on-white from the Pinnacle Ruin. Also similar to some of the pottery classified as McElmo Black-on-white is the use of rock temper. The Pinnacle Ruin Magdalena Black-on-white is tempered with rhyolitic detritus. At Roadmap Village, Magdalena Black-on-white sherds also commonly show multiple parallel lines, appended dots from linear elements, and checkerboard patterns, although sherd size there is too small for more detailed stylistic analysis. Neutron activation analysis (NAA) of Magdalena Black-on-white from the Pinnacle Ruin has shown that the majority of this pottery was made at the Gallinas Springs Site (Ferguson, et al. 2016). NAA data also provide evidence that there was limited local production at the Pinnacle Ruin and possibly at Roadmap Village. In support of Gomolak and Knight s early view of Magdalena Black-on-white as representing a local tradition, Bertram (1990:3 4) suggested that Magdalena Black-on-white could be considered to be a local version of Tularosa Black-on-white and its variants. However, the design motifs and overall design style of Magdalena Black-on-white and McElmo Black-on-white are quite dissimilar to that of Tularosa Black-on-white, a contemporary of both types. The hallmark designs for Tularosa Black-on-white are the massed joined hatched and solid element unit or the hatched unit with complex edges that is interlocked with a

7 IMPLICATIONS FOR MIGRATION AND SOCIAL CONNECTIONS 27 figure 2. Magdalena Black-on-white sherds from Pinnacle Ruin. Photo courtesy of John Fitch. solid unit. Interlocked motifs are most often curvilinear and hatching is rendered parallel to the framing line. Such features are not part of either a McElmo or Magdalena Black-on-white design system. The dominant design style of carbon-painted Magdalena Black-on-white at the Pinnacle Ruin is that of McElmo Black-on-white and not that of Mesa Verde Black-on-white. McElmo Black-on-white began to be made around A.D while Mesa Verde Black-on-white began at A.D (Ortman et al. 2005:5 15). McElmo Black-on-white was dominant from 1150 to 1200 and continued to be produced well into the 1200s (Ortman et al. 2005:5 15; Breternitz et al. 1974:41 43), and thus, the production of Magdalena Black-on-white seems to reflect the continuation of earlier design elements as compared to those represented by Tularosa or Mesa Verde Black-on-white produced at the same time. By the early to middle thirteenth century, sites in the Mesa Verde region have a ceramic assemblage that is dominated by Mesa Verde Black-on-white pottery, although McElmo Black-on-white also continued to be produced in diminishing quantities (Breternitz et al. 1974:41 43). Based on carbon dates from the stratigraphic positions of the McElmo style Magdalena Black-on-white, the data suggests two things. First, the Mesa Verde immigrants arrived by at least A.D at the Pinnacle Ruin, and they abandoned the site by A.D if not before. Second, the immigrant community was isolated at the Pinnacle Ruin and was somehow cutoff from those cultural trends that could have influenced the transition from a McElmo design style to a

8 28 JEFFREY R. FERGUSON ET AL. Mesa Verde design style. Given the dominance of McElmo style Magdalena Black-on-white in the Pinnacle Ruin assemblage, the presence of shaped and coursed masonry, and the presence of a well-developed midden, the defensive location of the site, and the marked contrast in all of these traits from the adjacent Victorio Site, there is little doubt that populations from the San Juan culture area had something to do with the occupation of the Pinnacle Ruin (Lekson et al. 2002; Clark and Laumbach 2011: ). Spatial Distribution of Carbon-Painted Magdalena Black-on-White Ceramics The distribution of Magdalena Black-on-white starts in the north with Gallinas Springs, although an argument could be made that the carbon paint wares found at Prieta Vista, Hummingbird and other sites on the Rio Puerco of the East could be included. The local carbon paint on the Rio Puerco is referred to as Loma Fria Black-on-white and is described by Baker and Durand (2003:90 95, ) as technologically and morphologically similar to Mesa Verde Whiteware. The type is tempered with sherd and/or sand, is self-slipped and is decorated with a variety of Pueblo III Band and other, less standardized, styles, reminiscent of both Santa Fe and McElmo Black-on-whites. Baker and Durand (2003: ) place this type in the middle 1200s and suggest that it developed due to an influx of population from the McElmo producing areas to the north in the early 1200s and was solidified by yet another influx of population post A.D Roney (1996: ) comments on the lack of apparent contact between carbon paint pueblos on the Rio Puerco and the contemporary Kowina (Tularosa) Phase sites on the Cebolleta Mesa immediately to the west, stating that the spatial separation between the Tularosa and Mesa Verde traditions implies the existence of a social boundary and that it can be inferred that at least two broad regional social networks were operative in the eastern San Juan Basin and the Acoma-Laguna areas during the 12th and 13th centuries. Magdalena Black-on-white has so far been noted in assemblages recovered from Gallinas Springs, Pinnacle, and Roadmap in some quantity. Beyond that, there are no other known strong concentrations. Interestingly, carbon-painted wares are found in small quantities in El Paso phase sites on both sides of the San Andres Mountains, on the Rio Grande to El Paso, and in the southern and eastern portions of the Tularosa Basin (Alves n.d.; Lekson and Rorex 1987:17; Dennis 1931; Lehmer 1948:55; Miller and Graves 2009:257; Browning et al. 1993:51; O Laughlin 2001:125). In general, the distribution pattern is heavily weighted to the east of the Black Range with small quantities found on late sites across south-central New Mexico. To further complicate matters, there are a number of large sites on the Rio Grande between Socorro and Truth or Consequences with Elmendorf Black-on-white, from which yet another carbon-painted type has been defined from components that seem to date to the same late 1200s and early 1300s time frame. Little or no excavation has been done at these sites and the surveyors might not have recognized Magdalena

9 IMPLICATIONS FOR MIGRATION AND SOCIAL CONNECTIONS 29 Black-on-white in the midst of all the Elmendorf Black-on-white. Elmendorf is similar to Magdalena Black-on-white in that it is thick walled, thick slipped, and displays McElmo-like designs, yet there are several distinguishing characteristics between the types such as Elmendorf ceramics include sherd temper, while Magdalena ceramics are exclusively rock tempered. A few sherds of Elmendorf Black-on-white (n=3) have been recovered from Pinnacle Ruin indicating at least limited contact with the Rio Abajo (southern Rio Grande). The examination of carbon-painted pottery produced across a number of distinct regions largely during the thirteenth century has resulted in the definition of a number of different types and a full discussion of these types is beyond the scope of this paper. The relationship between Elmendorf and Magdalena Black-on-white is an aspect that clearly should be considered in the future. Sites and Samples Gallinas Springs (LA 1178): The largest of the Magdalena phase sites, Gallinas Spring Pueblo (Figure 3) has as many as 300 slab masonry rooms and a half-dozen kivas (based on surface indications and limited excavations) and dates to the thirteenth and fourteenth centuries. The carbon-painted ceramics caught the attention of Florence Hawley Ellis (1959:326) and Emma Lou Davis (1964), who concluded that it was in the Mesa Verde ceramic tradition. Two unpublished university field schools (by the University of New Mexico and Western Michigan State University) and a CRM project created extensive collections from the site and several useful gray literature reports (Knight 1981; Bertram, et al. 1990, Tainter, n.d.). Magdalena Black-on-white is by far the dominant painted ware, making up 74% of the 1977 data (Knight 1981:8) and an even more robust 80% of the 1987 data figure 3. View of the Gallinas Springs site.

10 30 JEFFREY R. FERGUSON ET AL. figure 4. View of Pinnacle Ruin from the east. (Bertram et al. 1990: ). After plain brown, Indented Corrugated is the dominant utility ware. The brown indented corrugated is a Mogollon tradition ceramic and is common to Victorio, Pinnacle and Gallinas Springs. The flattened corrugated (Seco) is common at Roadmap and at Pinnacle but maybe not at Gallinas Springs as no comprehensive analysis is available (Laumbach and Laumbach 2013). While all are thought to be Mogollon brownwares, it is interesting to note that the Cibola Gray Corrugated found on both the Kowina (Tularosa) Phase sites on the Cebolleta Mesa and the carbon paint sites on the Rio Puerco is essentially a grayware version of Reserve Indented Corrugated (Roney 1996). A single sherd of Indented Cibola Corrugated was recovered from Pinnacle Ruin. Pinnacle Ruin (LA2292): Pinnacle Ruin (Figure 4) is a late thirteenth-century and fourteenth-century, 200-room masonry pueblo in Cañada Alamosa, approximately 50 km northwest of Truth or Consequences on the upper Rio Alamosa which empties into the Rio Grande just north of Elephant Butte. It is situated on a defensible rocky uplift only ½ mile upstream from the Victorio Site, a large Tularosa Phase community. In addition to a later Early Glaze Period (ca. A.D ) occupation of the site, there is a component dating to ca. A.D wherein Magdalena Black-on-white composes 48% of the painted ware assemblage. The next most numerous painted wares are El Paso Polychrome (16%) and White Mountain Red Ware (14%). Other than plain brown, the dominant utility wares are Reserve Indented Corrugated and Seco (Obliterated) Corrugated. The site also boasts shaped slab masonry resembling that of both Gallinas Spring and the Mesa Verde region. Excavation and extensive survey of Cañada Alamosa have been undertaken by Human Systems Research since 1999 (Clark and Laumbach 2011; Laumbach 2000; 2002; 2005; 2006; 2007; 2009; 2010; 2011a; 2011b; 2011c; 2012; 2013) As partners in the Cañada Alamosa Project, teams from the University of Colorado at Boulder worked on the Pinnacle Ruin from 2000 through 2004

11 IMPLICATIONS FOR MIGRATION AND SOCIAL CONNECTIONS 31 and again in 2008 (Lekson et al. 2001; Lekson et al. 2002; Laumbach 2005; 2006; 2007). Ceramics from three other sites near Pinnacle Ruin (Kelly Canyon, Victorio, and Montoya) are included in the broader study of the ceramics in the region (Ferguson et al. 2016), but they are only included here to provide evidence of ceramic production within the Cañada Alamosa. In tandem these three sites contain pithouse and pueblo components beginning in the eighth century and continuing well into the thirteenth century with the Tularosa Phase component of the Victorio Site. Radiocarbon and cross-ceramic dating indicate that the Victorio and Pinnacle populations were contemporary during the second half of the thirteenth century, thus creating a microcosm of the relationship between the Tularosa and Mesa Verdean social networks as described by Roney (1996: ). There are a very few sherds of Tularosa Black-on-white at Pinnacle (n=18) and even fewer Magdalena sherds at Victorio (n 10). Pinnacle and Victorio have Reserve Indented Corrugated common while obliterated Seco Corrugated is ubiquitous at Pinnacle and rare at Victorio. Roadmap Village (LA45157): Roadmap Village (Figure 5) is located along a perennial stretch of Palomas Creek, which flows southeast from the Black Range to the Rio Grande (Schollmeyer, Russell, and Nelson 2010). The site consists of at least 80 cobble masonry pueblo rooms arranged in two room blocks around two partially enclosed plazas. The pueblo dates to the late thirteenth and early fourteenth centuries, based on ceramics found on the surface and in excavated rooms. The decorated pottery assemblage consists of 41% indeterminate red wares. The dominant figure 5. Plan map of Roadmap Village. Figure by Will Russell.

12 32 JEFFREY R. FERGUSON ET AL. identifiable decorated type is El Paso Polychrome (26% of the decorated assemblage of 5098 sherds), followed by Magdalena Black-on-white (9%), Chupadero Black-on-white (5%), and Cibola Whitewares (<1%). The dominant utility ware is plain brown, followed by flattened corrugated and El Paso brown. Unlike Pinnacle Ruin and Gallinas Springs, masonry at Roadmap Village consists of courses of unshaped cobbles held together with mud plaster; chipped stone and sherds are used as chinking material in some walls. Although the height of the walls (nearly 2 m of standing wall height in some places, plus additional courses of wall fall) suggests many rooms were two stories tall, no definite indications of second-story floors were found during excavation. Walls were also quite narrow for their height, just a single cm course wide, and rest either on foundations of larger cobbles or (less commonly) on adobe footings. This construction style does not resemble the slab masonry of Gallinas Springs or Pinnacle Ruin, but its marbles in mud style has much in common with the Classic Mimbres architecture found in earlier time periods in the area. Madera Quemada Pueblo (LA91220): We have included five samples of carbonpainted ware from Madera Quemada Pueblo (LA 91220), a fourteenth-century Jornada Mogollon (El Paso Phase) pueblo located in the southern Tularosa Basin, southeast of the Organ Mountains on Ft. Bliss (Miller and Graves 2009). One area of the extensive site contained an adobe room block with 13 rooms (Figure 6). The site appears to have been ritually burned at abandonment. Turquoise and other materials are interpreted as intentional offerings left at abandonment. The site was occupied for a brief period sometime between A.D and This site is typical of fourteenth-century El Paso Phase sites in south-central New Mexico in that it includes a small number of carbon-painted sherds. In the past archaeologists have referred to these sherds as affinis Galisteo suggesting that their origin was in the Galisteo Basin or northern Rio Grande (Lekson and Rorex 1987:17). figure 6. Aerial view of Madera Quemada Pueblo. Courtesy of Myles Miller.

13 IMPLICATIONS FOR MIGRATION AND SOCIAL CONNECTIONS 33 Summary Comparison of the Four Sites: The four sites differ in significant ways. Of the four, Gallinas Springs and Pinnacle Ruin have the most in common with similar ceramic assemblages and shaped slab architecture produced from local latite slabs, but Gallinas Springs is a considerably larger site. Separated by 50 aerial miles and a mountain range, Pinnacle appears as a smaller version of Gallinas Springs. The decorated ceramic assemblages at both sites are dominated by Magdalena Black-on-white. Roadmap Village, located two drainages and 27 miles southeast of Pinnacle Ruin is quite different from either Gallinas Springs or Pinnacle, boasting cobble masonry architecture supplemented by adobe. In form it resembles the large adobe pueblos common across south-central and southwestern New Mexico during the late thirteenth and fourteenth centuries. El Paso Polychrome (26%) and Magdalena Black-on-white (9%) are the most common identifiable painted wares. Madera Quemada is a linear adobe pueblo common to the El Paso Phase and is located approximately 90 miles east of Roadmap. The dominant painted ware is El Paso Polychrome. Results Comparative NAA Data NAA of ceramics, clays, and adobe from Cañada Alamosa, Gallinas Springs, Roadmap, and Madera Quemada have been ongoing since 2007 with a combined total of over 800 specimens analyzed. The methods of sample preparation, data collection, and statistical analyses have been well documented elsewhere (Baxter and Buck 2000; Bishop and Neff 1989; Glascock 1992; Harbottle 1976; Neff 2002), figure 7. Bivariate plot of scandium and zinc showing the separation of Groups 3, 4, 6, 9, 10, and 11. Ellipses represent 90% confidence intervals for membership in the group.

14 Traditional production area TABLE 1 DISTRIBUTION OF COMPOSITIONAL GROUPS BY TYPE AND SITE FOR ALL SAMPLES FROM GALLINAS SPRINGS AND CAÑADA ALAMOSA. Type Site a 5b Unassigned Outlier totals N. Pueblo Magdalena B/W Gallinas Springs (Rio Salado and Magdalena B/W Cañada Alamosa Rio Puerco) Kiatuthlana B/W Cañada Alamosa San Marcial B/W Cañada Alamosa Socorro B/W Cañada Alamosa Cibola St John s Cañada Alamosa Polychrome (Zuni South & Cebolleta Mesa) Tularosa B/W Cañada Alamosa N. Pueblo/East Chupadero B/W Cañada Alamosa Mogollon Mimbres Classic B/W Cañada Alamosa (Mimbres, Gila Mimbres Boldface B/W Cañada Alamosa & Rio Grande) Mogollon R/Br Cañada Alamosa N. Pueblo Los Lunas Smudged (Rio Salado) Cibola (Cebolleta Mesa Zuni South) Pitoche Rubbed Ribbed Magdalena unpainted Reserve Indented Corr. Reserve Plain Corr. Cañada Alamosa Cañada Alamosa Gallinas Springs Cañada Alamosa Cañada Alamosa Continued 34 JEFFREY R. FERGUSON ET AL.

15 Traditional production area Mogollon: (Mimbres, Gila, and Rio Grande) TABLE 1 CONTINUED. Type Site a 5b Unassigned Outlier totals Mimbres Corr. Cañada Alamosa Mogollon Seco Corr. Cañada Alamosa Totals Assigned Production Area Cañada Alamosa Rio Salado Gila Mimbres/ Gila Rio Grande Unknown South Unknown North Unknown IMPLICATIONS FOR MIGRATION AND SOCIAL CONNECTIONS 35

16 36 JEFFREY R. FERGUSON ET AL. and will not be further addressed here. A representative sample of Magdalena Black-on-white was obtained from each of four sites. The assemblage included 40 sherds from five proveniences at Gallinas Springs (LA 1178), 37 sherds from five proveniences at Pinnacle Ruin (LA 2292), 20 sherds from five proveniences at Roadmap (LA 45157) and 5 sherds from five proveniences at Madera Quemada (LA 91220). Cañada Alamosa and Gallinas Springs Most of the compositional groups identified in the Cañada Alamosa and Gallinas Springs samples were identified in elemental bivariate plots, and, when sample sizes were sufficient, confirmed by calculations of membership probabilities using Mahalanobis distance calculations. The primary groups of concern for the Magdalena ceramics include 3, 4, 6, 9, 10, and 11 (Ferguson et al. 2016). Figure 7 shows the separation of all groups containing Magdalena Black-on-white samples from Pinnacle and Gallinas Springs. The compositional groups show strong correlations by type (Table 1). For example, Group 5 (including 5a and 5b) almost exclusively consists of painted ceramics presumed to have come from the north. Groups 1, 2, and 14 include all of the assigned Mimbres painted ceramics. Group 4 includes almost all of the presumably locally made Reserve, Mimbres, and Seco Corrugated wares from the four sites in Cañada Alamosa, as well as some of the Magdalena Black-on-white sherds from Pinnacle Ruin. Groups 3 and 10 include the majority of the Magdalena Black-on-white samples from Gallinas Springs as well as most of the corrugated wares from the same site. Magdalena Black-on-White NAA Data from All Sites The Magdalena Black-on-white samples from Roadmap Village and Madera Quemada Pueblo are projected against the established groups from Pinnacle and Gallinas Springs. The majority of the Roadmap and Madera Quemada samples fit into established groups, however we have identified two additional small clusters of samples (Groups 15 and 16) from Roadmap. Figure 8 is a plot of all the assigned Magdalena Black-on-white samples and the sites represented are listed for each compositional group. Gallinas Springs Samples: Magdalena black-on-white from Gallinas Springs (n=40) includes members of Groups 6, 11, 3, and 10. Groups 6 and 11 do not include any unpainted wares from the site, and may be related to other carbonpainted wares from the sites to the north on the Rio Puerco (Roney 1996) which appear to be part of the migration from the San Juan culture area. Group 3 is found in both the painted and unpainted sherds at Gallinas Springs, suggesting the possibility of local production. The samples are compositionally very similar to Group 10 except for a currently unexplained increase in zinc concentration (Figure 9). Group 3 does not occur at any other sites even though it is relatively abundant at Gallinas Springs. Group 10 is found at all sites, and is the dominant

17 IMPLICATIONS FOR MIGRATION AND SOCIAL CONNECTIONS 37 figure 8. Bivariate plot of zinc and thorium showing all assigned Magdalena Black-onwhite samples. Ellipses represent 90% confidence intervals for membership in the group. chemical group for the Magdalena painted ceramics. There are a number of unpainted examples from Gallinas (but no other sites) in Group 10 suggesting production at or near Gallinas Springs. Eight samples of unpainted wares belonging to Group 10 were found at a site very near Pinnacle (Kelly Canyon) that has been interpreted as the result of a twelfth-century community movement from the Rio Salado to the Rio Alamosa (Laumbach and Laumbach 2006; Clark and Laumbach 2011). The unpainted sherds belonging to Group 10 are distinctive corrugated types commonly associated with Socorro Black-on-white in the Rio Salado and were likely imported from that drainage. Pinnacle Ruin Samples: The Magdalena sherds from Pinnacle fall into three compositional groups. Group 10 samples may have been produced at Gallinas Springs, but the sherds in Groups 4 and 9 were probably locally made. Group 4 also includes the vast majority of the unpainted samples from the assemblages of all four sites in Canada Alamosa. There are a couple possible clay matches from raw clay from the Victorio Site (CAP452) and two adobe samples from Pinnacle (CAP436 and 437) for Groups 4 and 10, however the group membership probabilities are very low and we have yet to sample clays from the area around Gallinas Springs. The links between clay and adobe samples and the ceramics is an area of ongoing research. Roadmap Village Samples: Roadmap exhibits a pattern very similar to Pinnacle Ruin in that there are sherds from Group 10 as well as what may be smaller locally made samples (Groups 15 and 16). We have yet to analyze unpainted samples from Roadmap in order to get a better understanding of possible local production. Madera Quemada Pueblo Samples: The five samples from Madera Quemada Pueblo do not match El Paso Brownware compositional groups from the Jornada

18 38 JEFFREY R. FERGUSON ET AL. figure 9. Bivariate plot of zinc and cesium showing the Group 3 and 10 samples from Gallinas Springs and Pinnacle Ruin. Ellipses represent 90% confidence intervals for membership in the group. region (see Figure 10), but do belong to Magdalena Group 10. It is quite likely that the samples from Madera Quemada Pueblo were procured from sites located in the Black Range including either Pinnacle or Roadmap. The same could be said of the Obliterated (Seco) Corrugated) common in El Paso Phase assemblages and present at Madera Quemada. A piece of manganese ore was recovered from a pit at Madera Quemada Pueblo that most likely came from the Lake Valley mining district located 40 km to the south of Roadmap indicating at least some material exchange between the regions (Miller and Graves 2009:405). Summary of the Magdalena Production Model Figure 11 represents the best approximation of Magdalena Black-on-white production and distribution in central and southern New Mexico. It does not necessarily reflect the exchange pathways. There appears to be a one-way movement of ceramics out of Gallinas Springs. The only group found at more than one site was likely produced at Gallinas Springs, while the smaller groups that were likely produced at Pinnacle Ruin and Roadmap are not found at Gallinas Springs. Comparative Obsidian Source Data We include a brief discussion of the obsidian source data from most of the same archaeological assemblages in order to compare procurement and exchange patterns

19 IMPLICATIONS FOR MIGRATION AND SOCIAL CONNECTIONS 39 figure 10. Bivariate plot of chromium and iron showing the separation Madera Quemada Pueblo Magdalena Black-on-white samples from some of the major compositional groups in the El Paso area. Ellipses represent 90% confidence intervals for membership in the group. across material types. Figure 12 shows the archaeological site location relative to the relevant obsidian sources in New Mexico and Arizona. Gallinas Springs: Of the 70 samples analyzed (Figure 13), there was not a single artifact from any of the western New Mexico sources. The assemblage is dominated by northern material (the Mount Taylor and Jemez Mountains sources). This pattern may reflect a lingering cultural connection with groups to the north, or it may simply reflect economical choices that could include collection of secondary gravels from the Rio Grande. Pinnacle Ruin: The Magdalena contexts at Pinnacle (Figure 14) have the lowest percentage of obsidian from Northern sources (Jemez and Mt. Taylor) of all the assemblages at the four sites in Cañada Alamosa. The Tularosa component from Victorio is shown for comparison (Figure 15). The assemblage is dominated by material from the west, which is typical of most sites in the region except for the pre-a.d. 900 pithouse period where northern obsidian makes up over 60 percent of the assemblage. There is some potential for obsidian from pithouse (Figure 16) contexts contaminating the Tularosa component because many of the pueblo rooms were built on or near pithouse period structures and use areas. As a result obsidian from the pithouse component was probably utilized by the Tularosa phase population as well as inadvertently mixed with adobe during pueblo construction. Interestingly, an earlier movement of population from the Rio Salado to the Rio Alamosa during the 1100s documented at the Kelly Canyon Site also resulted in a loss of contact with northern obsidian sources even though it is clear that both Socorro Black-on-white and the primary corrugated wares were imported from

20 40 JEFFREY R. FERGUSON ET AL. figure 11. Proposed production model for Magdalena Black-on-white ceramics in central and southern New Mexico. Numbers refer to the ceramic compositional group. the north (Laumbach and Laumbach 2006; Ferguson et al. 2016). It is difficult to make confident statements about obsidian procurement for the Magdalena component at Pinnacle because only 24 obsidian artifacts have been recovered from clear Magdalena occupation contexts and we cannot rule out scavenging of previous occupations in the area as a significant source of material. Roadmap/Las Animas Village: The obsidian assemblage from Roadmap Village (Figure 17) very closely matches that of the Magdalena component from Pinnacle Ruin, in that there is very little northern material, and a dominance of obsidian from sources in Western New Mexico. We included a sample from a slightly later

21 IMPLICATIONS FOR MIGRATION AND SOCIAL CONNECTIONS 41 figure 12. Map of western New Mexico and Eastern Arizona showing the locations of the archaeological sites and sources mentioned in the text. Map adapted from Shackley (2005: Figure 1.1) site located on Animas Creek a few kilometers southwest of Roadmap Village (Las Animas Village, LA3949, Figure 18) that does not include Magdalena Black-on-white ceramics. The Las Animas Village obsidian assemblage is almost identical to that of Roadmap Village and the Magdalena component from Pinnacle. figure 13. Obsidian source frequency among artifacts from Gallinas Springs.

22 42 JEFFREY R. FERGUSON ET AL. figure 14. Obsidian source frequency among artifacts from the Magdalena component at Pinnacle Ruin. figure 15. Victorio. Obsidian source frequency among artifacts from the Tularosa component at figure 16. Victorio. Obsidian source frequency among artifacts from the pithouse component at

23 IMPLICATIONS FOR MIGRATION AND SOCIAL CONNECTIONS 43 figure 17. Obsidian source frequency among artifacts from Roadmap Village. figure 18. Conclusions Obsidian source frequency among artifacts from Las Animas Village. While we do not yet fully understand the exchange relationships involving Magdalena Black-on-white pottery, there are compositional data supporting interaction between sites in central and southern New Mexico yielding Magdalena Black-on-white ceramics. The movement of Magdalena Black-on-white from its primary production area at Gallinas Springs to the Pinnacle Ruin and Roadmap Village is easily inferred and may, at least in the case of Pinnacle Ruin, represent long-term connections between migrant communities. Roadmap Village stands in marked contrast, despite the clear compositional link with Pinnacle and Gallinas Springs. The Magdalena ceramics at Roadmap may be the result of interactions between local populations and other Magdalena communities (perhaps others beyond Pinnacle and Gallinas Springs), or perhaps they incorporated small numbers of individual migrants as potentially indicated by a production signature for Magdalena Black-on-white that only occurs in the Roadmap assemblage. Although the samples from Madera Quemada match the production groups from Gallinas Springs it seems much more likely that those carbon paint vessels were

24 44 JEFFREY R. FERGUSON ET AL. acquired from the geographically closer Roadmap Village. Supporting this notion is the consistent presence of a smudged obliterated corrugated ware in El Paso Phase sites in south-central New Mexico (Miller and Graves 2009:569; Lekson and Rorex 1987:17; Browning et al. 1993:52; Dennis 1931). Named Seco Corrugated by Wilson and Warren from its presence on Seco and Animas Creeks in the eastern Black Range, this type is associated with carbon paint wares at both Roadmap and Pinnacle, but not Gallinas. The likely production of this type on a site located on the lower Animas Creek just south of Roadmap Village has been documented by Laumbach and Laumbach (2013:87 100). The site also contains small quantities of Magdalena Black-on-white. Seco Corrugated, often referred to as Obliterated Corrugated, is considered an intrusive ware on El Paso Phase sites and it seems probable that both Magdalena Black-on-white and Seco Corrugated were obtained from the Black Range sites and taken to the Late El Paso Phase sites in south-central New Mexico. This scenario is much more likely than either a direct connection to the more northern Gallinas Springs or the earlier inference that the carbon paint wares on El Paso Phase sites originated in the Galisteo Basin or the Northern Rio Grande (Lekson and Rorex 1987:17; Lehmer 1948:55). While the evidence for ceramic provenance is far less geographically precise than the data from the obsidian source determinations, the two material types exhibit contrasting overall patterns. There is some evidence of local production of Magdalena Black-on-white at Pinnacle and possibly Roadmap, but the majority of the samples at all sites, including those from Madera Quemada in south-central New Mexico, appear to have been produced at Gallinas Springs and transported to the other sites. In contrast, the obsidian procurement patterns at Roadmap and Pinnacle differ dramatically from that at Gallinas Springs. There is no western New Mexico obsidian present in the sample from Gallinas Springs, with the majority originating from the Mt. Taylor sources. The assemblages from both Pinnacle and Roadmap reveal strong connections to the western New Mexico sources and very little northern obsidian. The Roadmap obsidian procurement pattern very closely matches the Las Animas Village assemblage which lacks a carbon-painted ceramic component. Is this a function of different economic spheres for obsidian and ceramics or possibly distinct gender-based interaction and procurement patterns? Perhaps Gallinas Springs was still close enough to Mt. Taylor and the Rio Puerco carbon paint sites to maintain relationships with that area. Pinnacle and Roadmap are much smaller and much more isolated, and may have interacted more through local exchange relationships, a concept that matches the inferred east west movement of carbon paint and corrugated wares. In addition, both Pinnacle and Roadmap are situated next to or on top of local communities where obsidian had already been gathered. The western obsidian source use in the Pinnacle and Roadmap assemblages may have resulted from scavenging discarded local pieces. It is surprising that Magdalena occupants at Pinnacle would maintain some material connection to Gallinas Springs, as demonstrated by the ceramics, yet not with obsidian. Other components and sites in Cañada Alamosa generally have some moderate percentage of Mt. Taylor obsidian and each of the earlier components has a ceramic-based connection to the north.

25 IMPLICATIONS FOR MIGRATION AND SOCIAL CONNECTIONS 45 The majority of the Magdalena Black-on-white was likely produced at Gallinas Springs. Pinnacle, and possibly Roadmap, had some local production of the type and, unlike Gallinas Springs, were likely producing Seco Corrugated. The presence of both Magdalena Black-on-white and Seco Corrugated in El Paso Phase sites across south-central New Mexico make it extremely likely that Roadmap, Pinnacle Ruin and contemporary sites in near proximity were the source for those ceramic types. Thus, it is likely that the source for carbon paint ceramics on El Paso Phase sites in south-central New Mexico was not from the Galisteo Basin and Northern Rio Grande but rather from the Gallinas Springs site, which was extended through supply of carbon-painted pottery to Pinnacle and Roadmap to the south. While the majority of the carbon paint ceramics from Pinnacle, Roadmap and Madera Quemada can be traced to production at Gallinas Springs, the obsidian procurement patterns vary by site. This deviation may reflect basic economic choices of raw material availability in the greater use of sources to the south and west, but, in the case of Pinnacle, the complete shift away from Northern sources contrasts with both their presumed relatives to the north at Gallinas Springs and their local contemporaries at the Victorio site. When considering what these patterns might indicate about the migrants, it is important to consider the degree of social and economic isolation of the migrant communities. It seems reasonable to assume that there were strong social networks between the migrant communities who were producing Magdalena Black-on-white and social distance between the migrant communities and local populations. Following Roney s (1996) interpretation of interaction between the Rio Puerco carbon paint sites and the Cebolleta Mesa Tularosa (Kowina) Phase sites, not only were migrants and locals in distinct social networks but there was likely a language difference as well. References have been made in this paper to the possibility of additional migrant communities that might have been producers of Magdalena Black-on-white. It is important to note that available survey data from the Black Range and points north do not contain evidence for such sites, leaving only a few unsurveyed locations where additional migrant sites might be found. Thus it seems reasonable that migrant sites south of the Rio Salado that produced Magdalena Black-on-white are limited to Gallinas Spring, Pinnacle Ruin and perhaps one or at the most, two, others. This scenario depicts isolated migrant communities whose interaction with local populations is limited rather than extensive. Such isolation would also have limited movement and acquisition of resources. A comparison of faunal data between the Tularosa Phase component at Victorio and the Magdalena component at Pinnacle reveals that while the local Victorio population was accessing bison, mountain sheep and antelope, none of these game animals are represented at Pinnacle (Wylde 2014). Instead, Pinnacle has a wealth of turkeys compared to the Victorio Site. One interpretation is that the migrant population had neither the knowledge of the area or the social network to access certain game animals. In this paper we have argued that Magdalena Black-on-white ceramics represent at least three distinct sources of carbon-painted ceramic production and exchange in SE New Mexico. Gallinas Springs and Pinnacle represent primary migrant communities dominated by migrants from far to the north. Other primary migrant

26 46 JEFFREY R. FERGUSON ET AL. communities may have existed and could possibly explain some of the additional Magdalena compositional groups found at Roadmap village if there was no local primary production. The Magdalena Black-on-white ceramics from Roadmap are a very small portion of the assemblage from a site that otherwise looks like a typical local community in terms of ceramic assemblage, obsidian source use, and architecture. The mechanism of carbon-painted ceramic acquisition is uncertain but might have included small-scale migration, intermarriage, or regional economic interaction. The Magdalena sherds at Madera Quemada is a fairly common example of long-distance regional exchange into heart of Jornada Mogollon and is commonly seem in the Southwest with long-distance movement of specific types (such as St. Johns Polychrome) that does not necessarily indicate long-distance movement of people. Acknowledgments This research could not have been completed without the assistance of the following individuals/institutions: Dennis and Trudy O Toole, the New Mexico Archaeological Council, the NSF Archaeometry Program (including BCS , BCS , and BCS to MURR), NSF Anthropology program (BCS to University of Colorado), Southwestern Anthropological Research Fund at the CU Museum of Natural History, the ASU Archaeological Field School partially funded by Arizona State University, the owners of the A Spear Ranch and the Ladder Ranch, Ken and Ruth Wright, John Fitch, David Phillips and the Maxwell Museum of Anthropology, Neal Ackerly, Bruce Kaiser, Chris Oswald, Mike Glascock, Kelly Donahue, Will Russell, Earthwatch Institute, Earthwatch volunteers, the students and staff of the University of Colorado and Arizona State University field schools, and volunteers for Human Systems Research, Inc. Investigations at Madera Quemada Pueblo were sponsored by the Environmental Division of Fort Bliss Military Reservation. References Alves, Mrs. n.d. LA 707: Site form on file at Laboratory of Anthropology, Museum of Indian Arts and Culture, Santa Fe, New Mexico. Baker, Larry L., and Stephen R. Durand (editors) 2003 Prehistory of the Middle Rio Puerco Valley, Sandoval County, New Mexico with contributions by Larry L. Baker, Linda Brett, Stephen R. Durand, Winston B. Hurst, Cynthia Irwin-Williams, and Fred L. Nials. A Publication of the Archaeological Society of New Mexico. Baxter, Michael J., and C. E. Buck 2000 Data Handling and Statistical Analysis. In Modern Analytical Methods in Art and Archaeology, edited by Enriquo Ciliberto and Guiseppe Spoto, pp John Wiley and Sons. Bertram, J In Excavations in the South Room Block of Gallinas Springs Pueblo (LA 1178), A Large Town of the Gallinas Mountain Phase (Late Pueblo III Early Pueblo IV) on the Mogosazi Frontier, edited by J. B. Bertram, A. R. Gomolak, S. R. Hoagland, T. L. Knight, E. Garber, and K. J. Lord, 8:1 55. Chambers Group, Albuquerque, NM. On file at USDA Forest Service, Albuquerque.