AMISTAD American Indian Tribal Affiliation Study Phase I: Ethnohistoric Literature Review

The Amistad NRA is situated within an area known as the Lower Pecos Archeological Region (see Figure 4), our micro-region. The region is dominated by the drainages of the Devils and Lower Pecos rivers, and includes all of Val Verde County as well as the southern portions of Crockett and Sutton counties, eastern Terrell County, and western Edwards and Kinney counties. The region also stretches south of the Rio Grande into Coahuila, but its southernmost extent is poorly known due to the very limited archeological explorations that have been reported for the area (Labadie et al. 1997; Turpin 1991:2). Pecos River rock art, the defining characteristic of the region (see Figure 10), is known to extend at least 90 miles south of the Rio Grande (Sayther 1998:90), indicating that the archeological region continues this far south into Mexico.

The Lower Pecos and our micro-region are further defined by its aridity. As described in the preceding chapter, it is part of the Chihuahuan Desert Biotic Province and its low growing, xeric, and thorny plants dominate the landscape. Across the area, rolling tablelands are deeply dissected by the Rio Grande, Pecos, and Devils rivers and their tributaries. Steep cliffs, formed during the eons of down cutting, frame the sides of these drainages. Rock shelters are found in these cliff faces and some have remarkable accumulations of deep, stratified archeological remains and/or contain the spectacular polychrome pictographs for which the region is well known (see Figure 10).

These same cliffs, however, have proven an impediment to travel (Figure 23). Berroteran (Hadley et al. 1997:200; Ayer 1729), a presidial commander who traveled through the area in 1729—one of the few Spaniards to do so—provided one of the most eloquent descriptions of this impediment:

In [1729, I traveled] from the presidio of San Juan Bautista, along the course of the Rio del Norte. . . . by way of the watering hole of Santo Domingo and the San Rodrigo, San Antonio, and San Diego de las Vacas rivers. Having gone as far as the last without finding a ford across it, I turned back along the south bank and traveled for two days to return to [the Rio Grande]. After crossing to the north bank [likely near Del Rio], I walked for four or five days, slowed by the lack of water for either horses or men. I saw that it was necessary to travel [west] along the south bank because the mountains on the north side impeded our passage . . . . I sent our seven Indian scouts to search the hills and mountains for water holes and a route by which we could continue our march. After seven days, two of them returned with the news that they had found neither a watering hole nor a route, and that they had not had anything to drink for two days. They had seen water, but at such a great depth that it took them four days to find a way to get down to it.

Figure 23. Cliff in the area of the Rio Grande within the Amistad NRA (photo courtesy Texas Department of Transportation).

Similar problems were experienced by the Castaño de Sosa party that set out from Monclova (then Villa Almaden) in July of 1590 seeking a new route to New Mexico (Sosa 1871). Although the party crossed the Rio Grande with little difficulty, likely in the region of Cuidad Acuña and Del Rio, their route north through the lands between the Salado (Pecos) and the Devils rivers was exceedingly difficult (Sosa 1871:197-204). Despite multiple tries, they found few places where horses or men could cross the Pecos and no place where the wagons could negotiate a crossing, forcing them to remain east of the river. Other than the river, the expedition members found that water was scarce to non-existent; near the river, it had to be hauled up in containers.

Another notable aspect of these two journeys was the few Native Americans seen in either journey. The Sosa (1871:196) party met some natives (called Jocome) on the Rio Salinas, well to the south of the Rio Grande but during the month of difficult travel north, as they paralleled the Pecos, none were encountered. Instead, only a few signs (rastros) of Native Americans were seen from time to time. Berroteran (Ayer 1729) did encounter Apache in the region, but, even then, their total number was quite small and, when asked if they knew of other natives in the area, the Apache replied that they did not. As will be shown, the archeological data from the Lower Pecos has a similar paucity of material from the Late Prehistoric and Historic periods. The scarcity of Native Americans seen by the two expeditions may reflect a cultural reality. On the other hand, the paucity of native groups in the Lower Pecos in 1590 and 1729 may be coincidental. In each case, the Spanish were present only for brief periods of time and neither party was especially familiar with the terrain or the native groups. Since the Native Americans of the region were hunters and gatherers, they simply may have been occupied elsewhere. It is equally feasible that the native groups deliberately tried to avoid the Spanish. Castaño de Sosa is believed to have been involved in the native slave trade (Labadie 1994:11-18). While serving as Lieutenant Governor of Nuevo León, he likely had hunted for slaves in this general region. Any Native Americans aware of Sosa's history would not have wished to provide an opportunity for him to hunt them again as slaves. Berroteran was a Spanish presidial soldier and the natives in the general region in the early eighteenth century were generally hostile to the Spanish (Kenmotsu 1994:203). In sum, the scarcity of Native Americans encountered in the region in 1590 and 1729 may or may not accurately represent the quantity of people living in the region.

In the remainder of this chapter we present an overview of the archeology pertaining to the last 1500 years of Native American habitation in the Lower Pecos Archeological Region, our micro-region. The discussion begins with a summary of the Late Prehistoric and Historic periods. It is followed by a summary of known sites in the micro-region that contain artifacts and/or rock art that are believed to date to these periods and, when possible, a guess about the groups responsible for those material remains.

OVERVIEW OF THE ARCHEOLOGY

Several overviews of the archeology of the Lower Pecos have been completed by Turpin (cf. 1982, 1984, 1985, 1986, 1991, 1995), who has a long history of research related to the archeology and rock art of the Lower Pecos. Other recent summaries relevant to this affiliation study include those by Bement (1989) and Mehalchick and Boyd (1999). Regardless of which summary the reader selects, he/she should recognize that the Late Prehistoric period for the Lower Pecos is poorly known and the Historic period is even less well known. With the completion of the analysis of data from the 1999 Field School of the Texas Archeological Society, held at the Amistad NRA and focused on these periods, it is likely that new information on these two periods will be revealed (Collins et al. 2000).

Turpin (1991:33-37) divided the Late Prehistoric period into two phases. She assigned the name Flecha to the oldest phase since it is during this time that the bow and arrow (flecha means arrow in Spanish) made their appearance in the archeological record of the Lower Pecos. She dates the Flecha phase from A.D. 630-1500 based on the presence of arrow points in Stratum 2a at Arenosa Shelter. A hearth, present in this stratum, has been radiocarbon dated and calibrated to A.D. 619-673 (Turpin 1991:36, and Table 1.12; Dibble 1967:30). Given the presence of Scallorn arrow points in the same stratum, she marks this as the beginning of the Flecha phase. Scallorn arrow points are generally regarded as one of the diagnostics of the early portion of the Late Prehistoric in Central and adjacent parts of Texas (Prewitt 1995:83-173; Turner and Hester 1985:169). In the micro-region, unstemmed arrow points are sometimes considered to be in association with these stemmed points (McClurkan 1966), and Turpin (1991:35) believes that unstemmed points dominate the Late Prehistoric assemblages south of the Rio Grande (for a slightly different view, see discussion of these unstemmed points below). It is not uncommon to find Ensor dart points also present in the same excavated stratum with diagnostics (Scallorn and others) from the Flecha phase. Since Ensor points are part of the inventory of the preceding Late Archaic period, Turpin concludes that their presence with arrow points indicates that there was temporal continuity and that the new technology was not adopted unilaterally. Bement (1989:59, and Table 1) lists Perdiz, Toyah, and Livermore arrow points as index markers of the Flecha phase possibly because Perdiz arrow points were also recovered from Stratum 2A at Arenosa. However, Dibble (1967:34) noted that this stratum was mixed, and most researchers (Mehalchick and Boyd 1999; Turpin and Robinson 1998; Johnson 1994) would place those diagnostics, along with Cliffton (Perdiz performs) and Harrell arrow points with the final portion of the Flecha phase, e.g., A.D. 1300-1500. Turpin (1991:35, 1995:550-552) assigns a variety of site types to the Flecha period, including ring middens, crescentic scatters or piles of burned rock, and cairns. Cairns have been interpreted to represent mortuary features (Turpin 1982:148, 160), but the only one that has been investigated in the micro-region (41VV364) did not have human remains. [1] A new weaving technique (mats made of threaded and twined bulrush) was also introduced during the Flecha phase (Turpin 1995:550).

The Red Monochrome rock art style is considered another aspect of this phase as it contains human figures carrying bows and arrows (Turpin 1991:35, 1986). Unlike the earlier, more abstract, polychrome figures in the rock art of the Lower Pecos, these human figures are full bodied and represented by a naturalistic form painted in hues of red and orange (Kirkland and Newcomb 1967:81). Some human figures have arrows protruding from them, suggesting that warfare was not unknown. Although the major concentration of the known Red Monochrome sites is located near the mouth of the Pecos River, some of these paintings have been recorded as far west as the Big Bend region and as far northeast as the Dry Devil's drainage (Turpin 1995:551; Jackson 1938:239). It is not known if the quantity of these sites near the mouth of the Pecos reflects a cultural phenomenon or simply a lack of intensive archeological investigations elsewhere. Bement (1989:68) speculates that:

the appearance of such a fully developed art style [indicates] the intrusion of an outside group. The relative scarcity of sites with this style [however,] may indicate that the intrusive group inhabited the area for only a short time.

While we agree with the broad outline of the Flecha phase, we take a slightly more conservative approach to several aspects of it because the data are not clear-cut. First, the integrity of the archeological assemblages at many sites dating to this phase has often been compromised. Of those sites that have been radiocarbon dated to the period A.D. 630 to 1500 (see Turpin 1991:Tables 1.2 and 1.12 for list of dates), not all have yielded diagnostic arrow points while others have yielded a variety of projectile point styles from Archaic through the Late Prehistoric period. For example, Turpin (1991:36, and Table 1.12) uses the calibrated Stratum 2A date from Arenosa as the beginning date for the Flecha phase. The date, based on the principal investigator's (Dibble 1967) assessment of its context, seems reliable. Recovered with it, however, were Perdiz, Cliffton, and Harrell arrow points—diagnostics typically associated with absolute dates several centuries later (cf., Mehalchick and Boyd 1999; Johnson 1994). Moreover, Dibble (1967:34) stated that this deposit was mixed, indicating that, while the date is valid, individual artifacts may or may not be related to each other. In another example, a second core Flecha dates comes from stratum D at the Sotol site (41CX8) in Crockett County where at least one plain brownware sherd was recovered. Ceramics are nowhere common in the Lower Pecos, but they are considered one of the diagnostics of the subsequent Infierno phase (Turpin 1991:37). This same stratum at the Sotol site also contained two Garza, one Perdiz, and one Fresno arrow points—again, diagnostics that would fit more comfortably in time periods after Flecha. Similar confusion exists regarding the unstemmed points recovered from two rock shelters in northern Mexico: La Calsada (NL 103, Nance 1992) and Cueva de la Zona de Derrumbres (NL 92) (McClurkan 1966). Nance (1992:69-71, 171) has typed the unstemmed points at these two sites as Fresno, Toyah, Starr, and miscellaneous unstemmed points (Table 3; see also Appendix 7) and Turpin (1991:35) concludes that they represent Flecha phase materials. Again, Fresno and Toyah, the two dominant arrow point types recovered from the sites, have been associated with absolute dates or archeological assemblages that range from A.D. 1300-1790 (Johnson 1994; 1969; Kenmotsu 1992; Creel 1990; Tunnell and Newcomb 1969). Moreover, Fresno arrow points are also typologically quite similar to Guerrero points, which Turpin and Bement (1988:77) place in the ensuing Infierno phase. Given these factors, the presence at the Sotol site of a sherd and arrow points that are often associated with occupations dated to the Infierno phase, suggests that either the stratum was mixed or that the dates for the Flecha and Infierno phases need some adjustment. We suspect both.

Table 3. Arrow Points from Two Rockshelters in Northern Mexico.

Features typically associated with the Flecha phase present yet another problem. Turpin (1991:35) opines that cairns, ring middens, and crescentic piles of burned rock "consistently date to the Flecha or later periods" based on their presence in sites that have yielded Scallorn points and because some sites with cairns have radiocarbon dates within the Flecha phase. However, the evidence suggests that these features may date to other time periods as well. Only one cairn in the micro-region (41VV364) has been excavated and it yielded one Perdiz, one Scallorn, and two dart point fragments (Turpin 1982:154), suggesting that it may date to any of several chronological periods. Assuming that the cairn at 41VV364 was erected no earlier than the most recent diagnostic (Perdiz), then this cairn tenuously does date to the Flecha Phase. That said, most other sites with cairns have not yielded any projectile points. Like cairns, crescent and ring middens may date to several periods. Some clearly date to the Flecha phase, and outside of the micro-region, Kenmotsu (1993) has identified several in the Guadalupe Mountains that are attributed to the Mescalero Apache. Hester (1989:61) and Shafer (1986) believe they began to be constructed prior to the Flecha phase. Moreover, in surrounding regions, crescent and ring middens have been dated to both the Late Archaic and Late Prehistoric periods. For example, at O. H. Ivie Reservoir, most annular (ring) middens dated to the Late Prehistoric period, but a Late Archaic point (Pandale) was recovered from the ring midden at 41RN169, and an Early Archaic (La Jita) point was recovered from a ring midden at 41RN3 (Lintz et al. 1993:214, 328). Crescent middens in Sanderson Canyon (Terrell County) typically yielded Late Archaic points such as Ensor, Frio, and Paisano (Shafer 1971), as did the ring middens reported by Moore (1983) from Musk Hog Canyon, situated in Crockett County. While these data do not prove that ring middens date to Archaic times, they are sufficient to question a universal assignment of these features to the Late Prehistoric. In sum, it seems premature to assume that all cairns, ring middens, or crescent middens relate to Late Prehistoric and Historic period occupations of the micro-region.

Turpin (1991:36-37) calls the subsequent time period the Infierno phase, beginning around A.D. 1500. She believes that the phase ended around 1700 and separates it from the Historic period, which she initiates at 1600, implying that groups exhibiting characteristics of each of the two phases resided side by side for the better part of a century. More recently, the end date of the Infierno phase has been extended to 1780 (Mehalchick and Boyd 1999:153; Turpin and Robinson 1998:86), although the only absolute dates obtained for sites assigned to this temporal period are those of the upper component at the San Felipe Springs site (Mehalchick and Boyd 1999). Turpin (1991:37-38) associates this phase with an archeological assemblage consisting of "small stemmed arrow points, steeply beveled end scrapers, prismatic blades, and plain brown ceramics" that are typically found on high promontories and often contain stone circles or cairns. Arrow points assigned to the phase include Perdiz, Toyah, Livermore, Sabinal, Bonham (rare), Infierno, Fresno (Guerrero?), and Starr. Dorso end scrapers also date to this phase (Bement and Turpin 1987). Turpin does not assign any rock art to this phase.

The phase is named for the type-site, the Infierno Camp (41VV446), which Dibble recorded in 1974 and mapped in 1976 but did not report. It was recently investigated as part of the 1999 field school of the Texas Archeological Society (Collins et al. 2000). During the latter effort, three clusters of features were identified and mapped at the site. The north complex contained 44 wickiup rings (i.e., small stone enclosures of less than 3 m in diameter), six tipi rings (discontinuous rings of stones that measure from 3-10 m in diameter), and seven burned rock features. The middle complex contained 56 wickiup rings, one tipi ring, one burned rock feature, and one unburned circular stone pavement. Twenty-four wickiup rings were recorded in the south complex, but more are believed to be present. The artifactual material recovered by Dibble and additional material observed during the 1999 investigations at Infierno Camp is sparse, but includes a few small arrow points, small blade flakes, abrading tools, and a few plain potsherds. Turpin (1982:167) considers the Infierno phase intrusive into the Lower Pecos, largely due to its "radically different features and artifact assemblage." Collins et al. (2000:12) do not speculate on the arrival of new groups but do note certain distinctions when Infierno assemblages are compared to earlier archeological assemblages: "the striking . . . low-frequency of burned rocks on [the Infierno Site] compared to most sites in the Lower Pecos region—evidently [represents] an indication of a contrast in subsistence technology" from earlier periods.

A few other Infierno phase sites have been investigated, including several investigated during the 1999 Texas Archeological Society field school (Collins et al. 2000), the upper component at the San Felipe Springs site (Mehalchick and Boyd 1999), and another site near Live Oak Creek (Turpin and Bement 1988). At the latter site, Turpin and Bement (1988) excavated the most intact of several stone rings recorded at the site (Figure 24). The site is situated on a bank of Live Oak Creek with little soil cover and consists of several burned rock middens and fice stone features. Feature A, the excavated ring, consisted of a "double ring of 71 large blocks . . . 2.5 meters in diameter" with a southeasterly opening or break in the ring. Like most stone rings (Collins et al. 2000; Oetelaar 2000:45-49), few artifacts (n=27) came from the excavation of the ring, and they consisted of burned and unburned secondary and tertiary chert flakes. Artifacts diagnostic of the Infierno phase were recovered elsewhere on the surface of the site and include a brownware sherd, a Guerrero point, Sabinal point, a stemmed arrow point, and a dorso scraper along with several dart points. The lack of diagnostics or other material recovered from the ring clouds the chronological placement of this or other stone rings.

Figure 24. Stone ring (Feature A) excavated at the Tonto or Live Oak Hole consists of several burned rock site (41VV828) by Turpin and Bement (1988): a, plan view; b, photograph of middens and five stone features. plan view (Courtesy of the Texas Archeological Society).

Although stone rings of this size have yielded diagnostics of the Infierno phase (Collins et al. 2000; Turpin and Robinson 1998), at least one ring structure (41VV1099) that has been investigated can be dated to the Late Archaic (Wayne Bartholomew, personal communication 1991). The site is located on a high saddle overlooking a tributary of Deadman's Canyon, and sits on bedrock with no soil cover. It consists of a single stone ring (measuring approximately 2 m in diameter) with a sparse scatter of lithic reduction debris and a Paisano dart point. The dart point and the lithic debris are all from the same lithic source material. Some flakes even re-fit onto a core, also recovered at the site. Fully mapped and collected, but not yet reported, the site appears to represent a small Archaic period camp where one and possibly more dart points were manufactured. There is no evidence that the site (or the ring) was used for a signal fire or that it was used during the Late Prehistoric. Stone rings elsewhere are also not restricted to the Late Prehistoric or Historic periods (see Oetelaar 2000:44; Davis 1983). Nonetheless, Turpin and Bement (1988) assign the excavated ring at 41VV828 to the Infierno phase on the basis of surface-collected artifacts found elsewhere at the site, the integrity of the stone ring, and the historic pictographs at the nearby Live Oak Hole site (41VV169).

Dibble (1978) first outlined the Infierno phase because he believed that the pottery in the Lower Pecos was distinct from the pottery found to the north and northeast. That pottery is known as Leon Plain and it is one of the diagnostic artifacts of the Toyah phase. While Turpin and Robinson (1998:91) assert that "the Infierno phase is roughly contemporaneous with, but not an extension of, the Toyah Phase of central and south Texas," a summary of the Toyah phase by Johnson (1994:187-277) suggests to us that this archeological construct fits well with the archeological data from the Infierno phase. As Johnson points out, the Toyah phase was a Late Prehistoric to early Historic phase that swept across most of Texas south of the Southern Plains, extending into the Lower Pecos, Trans-Pecos, and Central Texas regions between A.D. 1300 and 1780. Using data from a series of well reported, excavated sites, Johnson (1993:271) argues that the material culture of this phase is distinct from earlier horizons in South and Central Texas and portions of the eastern Trans-Pecos, appearing "abruptly in the archeological record." The Toyah phase appeared at a time (ca. A.D. 1300) when climatic conditions prompted buffalo to migrate south (Johnson 1994:258). This was also a time when there is some evidence of a growing trade between Plains buffalo hunters and the pueblos of eastern New Mexico (Kenmotsu 1994, 2001; Spielmann 1982). The Toyah material culture—well made arrow points manufactured from blades and flakes, end scrapers, perforators, knives, and plain pottery—was "well adapted for the hunting and processing of buffalo" (Johnson 1994:271). Perdiz arrow points, informal knives and scrapers, dominate the lithic toolkit at Toyah sites as well as a variety of tools (end scrapers, perforators/drills, and points) fashioned from flakes, along with Harahey and Covington beveled knives, and a blade technology. Other arrow points are sometimes present on Toyah sites, including Livermore, Harrell, Garza, Soto, Fresno (or Guerrero?), Bonham, and Sabinal. Ceramics from Toyah phase sites are few, but often exhibit vessel smoothing using a wide stick, beveled rims, application of a thin wash to vessel interiors, and frequent use of bone temper (Johnson 1994:269). Through careful examination of excavated Toyah complex sites and an analysis of the Buckhollow site, Johnson hypothesized other Toyah traits. These include the evidence that the Toyah folk did not restrict their diet to bison or deer, but rather "gathered, killed, grew, and ate . . . what comestibles were locally available, and in what season of the year its people found themselves" (Johnson 1994:262). Groups generally consisted of small or extended family households, and group mobility appears to have been limited. Distinguishing between a "classic" Toyah phase in Central Texas and other areas to which the Toyah culture spread, Johnson (1994:279) places the Lower Pecos on the margins of the spread. He believes the spread was likely the result of both migration of non-resident groups south from the Plains and "conversion" of local residents to a Plains tool kit that allowed their participation in a regional interaction network that focused on bison hunting.[2]

These traits fit well with the evidence from the Lower Pecos (the micro-region). The material from the final centuries of the Flecha period and from the Infierno Phase—Perdiz arrow points, beveled end scrapers, prismatic blades, bone-tempered ceramics—is unlike the material of the earlier part of the Late Prehistoric. However, they mesh with the material from Toyah sites (cf. Johnson 1994; Creel 1990). Recently, Mehalchick and Boyd (1999:152-157) conclude that the Infierno Phase is a regional variant of the Toyah phase. Their conclusion is based on their work at the San Felipe Springs site (41VV444) in modern Del Rio and on comparisons of their data with Turpin's (1991; 1989, 1982; Turpin and Robinson 1998) work and data from nearby sites. They identified a Late Prehistoric component post-dating A.D. 1300 at the springs. This component "is distinguished primarily by Cliffton and Perdiz points, ceramics, and formal scrapers" (Mehalchick and Boyd 1999:152). Their case is strengthened by their ceramic analysis. Seven sherds from San Felipe were subjected to petrographic analysis (Robinson 1999) and compared to petrographic results of other Infierno sherds from the Lower Pecos sites and from sites 41RG26, 41FL1, 41CX9, and Mission San Lorenzo de Santa Cruz (Turpin and Robinson 1998). The results are striking and "indicate that the Infierno Phase plainwares are essentially identical (with only minor variations due to local production) to the bone-tempered pottery manufactured by Toyah people across south and central Texas" (Mehalchick and Boyd 1999:153). They also point out that the beveled end and side scrapers from Infierno sites are defining characteristics of the Classic Toyah phase as well and commonly attributed to bison hunters. Similarly, Perdiz arrow points (and their preforms, e.g., Cliffton), present in Infierno sites, predominate in Toyah phase sites, and two and four-beveled knives are present in sites of both phases. Bison remains, while not ubiquitous on Infierno sites, are nonetheless, sometimes present in Late Prehistoric deposits (Collins 1969:1-11), and Spanish documents of the late seventeenth century indicate bison were present in the region at that time.

Differences between the sites associated with the Infierno versus the Toyah phases include the lack of structural remains on Toyah phase sites and the presumed chronological differences between the two phases. Mehalchick and Boyd (1999:154-155) propose that the former difference is likely a sampling factor, and believe that the dating distinction (A.D. 1300-1650 for Toyah versus A.D. 1500-1780 for Infierno) relates to the deflated surfaces on which many Infierno sites have been found and the lack of any absolute dates for Infierno. Thus, as Mehalchick and Boyd (1999:155) remark, Johnson's dates for Toyah are based on a relatively precise chronology, but the age assessment for Infierno "is largely speculative." Still, two calibrated radiocarbon dates (Mehalchick 1999) from the Toyah zone at San Felipe Springs (neither was from feature context) date between A.D. 1295 and 1450 and suggest that the date of A.D. 1500 for the initiation of the Infierno phase is too late. Mehalchick and Boyd (1999:157) conclude by stating:

The evidence is far from conclusive, but there are two equally plausible explanations. The Infierno Phase may represent a distinctive but intrusive Protohistoric Plains culture sharing many lifestyle and material culture characteristics with the Toyah Phase, or it may be little more than a continuum or late (after A.D. 1500) variant of Toyah in the Lower Pecos.

Given the contrast between the features and artifactual assemblage of the Infierno phase and earlier occupations, Turpin (1995:553) has stated that the "Infierno people clearly came into the Lower Pecos region late in prehistory," and goes on to associate these remains with "native northern Mexican people traveling en masse to the mouth of the Pecos River for annual bison hunts." Presumably she was referring to the large groups of Native Americans encountered by Fray Manuel de la Cruz, Fray Juan Larios, and Lieutenant Fernando del Bosque who traveled to and from the Rio Grande during the late 1600s. These Spaniards encountered relatively sizable groups both north and south of the river (Wade 1999a). However, as we noted in the preceding chapter, many of those groups resided north of the Rio Grande, some from areas close to the river (e.g., Catujano, Tereodam, Gueiquesale), while others (e.g., Geniocane, Sana, Bagname, Xoman, Yorica) were from regions well to the north of the river (Wade 1998, 1999a; Kenmotsu 1994). Hence, during the closing years of the Infierno phase, there is documentary evidence that while some native groups traveled north, other Native Americans were moving south through the Lower Pecos. The documentary evidence, however, does support Turpin's belief that their journeys to and from the Lower Pecos were to hunt bison.

In summary, we, like Mehalchick and Boyd (1999:157), find the similarities between the Classic Toyah phase artifactual assemblages and the Infierno phase artifactual assemblages too close to discount. Moreover, we find that the documentary evidence shows: (1) the presence of bison herds in and close to the Lower Pecos by at least A.D. 1650 (Wade 1999a, 1999b); (2) the hunting of bison in the area by resident and non-resident groups; and (3) the territorial conflicts between those resident and non-resident groups as each sought greater access to the bison herds (Wade 1999a). Together, these data suggest to us that the Infierno phase reflects either, as Turpin (1982:167) noted, the artifactual assemblage of new groups moving into the Lower Pecos with their technology, or Toyah phase groups outside of the Lower Pecos influencing the technology of local groups. The timing of this change has yet to be fully worked out, but radiocarbon dates associated with Perdiz, beveled knives, and dorso end scrapers at sites like 41VV260, 41VV67, and 41VV444, suggest that it may have been as early as A.D. 1400. While bison hunting was not their exclusive economic endeavor, the Infierno tool kits show that it was an important one. Like Mehalchick and Boyd, we believe that these sites, and the Infierno phase, represent a regional variant of the Toyah phase.

In contrast to the Infierno phase, Turpin (1991:38) believes that archeological sites during the Historic period "consist largely of rock art panels with little or no accompanying occupational debris." Her exceptions are a few scattered metal arrow points and an occasional gunflint, or sites related to the military activity in the region, such as 41VV1428, the Bullis Trail site. She finds evidence of two periods in the historic rock art, with the first one beginning ca. A.D. 1600 and ending in the early 1700s. She dates the second period of historic rock art from the mid-1700s to the latter half of the nineteenth century. This legacy is depicted in the rock art at a few sites that contain drawings of structures that appear to be missions with crosses on their roofs, pictorial elements that are dramatically different from the rock art of earlier periods. Crosses are often present at these sites as well, and may be linked to the mission-like structures, although crosses are also present in the complex art of earlier periods. Other elements assigned to the early historic rock art are anthropomorphs that appear to be from the historic era. One such figure is present at Vaquero shelter (41VV77; Figure 25); this figure is described as a grandee by Turpin (1988:52). The individual is drawn with European-style clothing in a naturalistic style (see costumes in Boucher [1987:278-284]). Other anthropomorphs assigned to the early historic rock art have rectangular bodies, some with v-necks, crosses, and ears.

Figure 25. Figure with European-style dress at Vaquero Shelter (41VV77) (photograph courtesy of Texas Historical Commission).

We fully endorse Turpin's distinction between early and late historic pictographs in the micro-region. However, we tentatively suggest that the early pictographs may relate to the Infierno phase. That is, Turpin's dating of these early historic scenes has parallels in the documentary record of the seventeenth century. In 1658, in Saltillo, some Babane and Jumano individuals requested that the Spanish establish mission settlements for them and their allies (Wade 1999a:30). Kenmotsu (2001) describes the similar attempts of the Jumano elsewhere to entice the Spanish into establishing other missions in their lands along the Pecos and Concho rivers. Those efforts date to the seventeenth century and were an attempt to prevent Apache encroachment of their lands. In addition, Wade (1999a) points out that there were a variety of Native American coalitions on both sides of the Rio Grande during that century and each coalition sought to maintain their access to the bison herds in the region. Finally, in the 1670s, settlements and missions were established south of the Rio Grande. While their success was equivocal, Spanish priests and missions were central to the effort as were administrative and military leaders. Hence, we tentatively suggest that some of the early historic rock art in the Lower Pecos is related to these events, perhaps recording the requests for missions, missionaries, and settlements.

In contrast to the early historic pictographs, rock art panels from the late historic period exhibit realistic drawings of bison (which are rarely, if at all, present in the Archaic period art of the region), suggesting either that bison were present in greater numbers than previously, or that they were accorded importance in the lives of the artists. These rock art panels contain figures and elements that are drawn in the Plains Ceremonial and Biographic styles (Keyser 1987:45-48). They often contain pictographs with horses, occasional bison, and human-like figures (often with muskets/rifles, etc.) (Figure 26). Some of the human-like figures wear leggings and long flowing headdresses. They also carry shields or have shields located close by; others have Spanish or United States army issued hats, robes, or other European garb. Turpin (1995, 1988), following Jackson (1938) and Kirkland and Newcomb (1967), recognizes these action scenes as ones that relate to the presence and activities of Plains (perhaps Comanche and Kiowa) tribes in this region.

Figure 26. Rock art panel at site 41VV485, Dolan Springs showing a bison over a human figure with a musket. (Photograph courtesy of The University of Texas at Austin, Texas Archeological Research Laboratory, site files).

As Turpin (1988) and we (see Ethnohistoric Review) have noted, by the early eighteenth century Spanish interest in the micro-region waned, allowing intrusion of Plains tribes who were being displaced from these lands. That change appears to be reflected in the rock art of the Lower Pecos. Although Turpin (1988:54) suggests that the change began as early as 1650, documentary evidence (see Wade 1999a; AGI 1682-1683) indicates that it is more likely to have begun around 1730 (see Appendix 5). Documents prior to that date mention the presence of a variety of native groups in the northern Coahuila and west central Texas regions, but Plains tribes were not among them. By 1729, however, the Barriero map calls the Pecos the "Rio Salado o del Natagee" (see Figure 7) and several Apaches were found south of the Rio Grande that same year (Ayer 1729). Apache presence in the Lower Pecos continued throughout much of the nineteenth century. By 1756, documents note that the Comanche were beginning to travel south to the San Saba River, then southwest through the Lower Pecos into Coahuila (QA 1756:227), and subsequent documents (e.g., Wallace n.d.:322; Swanson n.d.:267) indicate that the Comanche continued to access these lands. Unlike the Apache, however, the Comanche used the Lower Pecos as a route of travel during raiding excursions, but generally resided to the north (Winfrey and Day 1995, Vol. 1:24-25, and Vol. 2:110). Other Native Americans in the Lower Pecos during the nineteenth century include the Kiowa, Seminole, Kickapoo, Cherokee, and Shawnee.

THE SITES

For this study, we reviewed the site forms for Val Verde, Edwards, Crockett, Kinney, Sutton, and Terrell counties to identify the sites dating from the final part of the Flecha phase, along with sites dating to the Infierno Phase and the Historic period, i.e., those years dating after A.D. 1200 that span the final decades of the Late Prehistoric as well as the Historic periods. We reiterate that it was necessary to include this broad time frame because it is not possible to distinguish archeologically between sites of the late portion of the Late Prehistoric and the early part of the Historic period. It should also be noted that all sites from Edwards, Crockett, Kinney, Sutton, and Terrell counties were reviewed to identify and include sites of the period of concern even though not all of the land base of these counties can be included in the micro-region. This was because, outside of Crockett County, these counties have so few recorded sites that the time needed to selectively search for only those sites present in the micro-region did not seem warranted.

The data on the sites were obtained from research in site files of the Texas Archeological Research Laboratory (TARL) at The University of Texas at Austin with the assistance of Don Wade, and from the Historic Sites Atlas, a joint effort of the Texas Historical Commission, the Texas Department of Transportation, and the Federal Highway Administration. Appendix 7 provides a listing of these sites, and Table 4 summarizes that data. Since few of these sites have been subjected to subsurface investigations, any site with materials considered diagnostic of these two periods—Perdiz, Toyah, Livermore, Garza, Fresno (Guerrero?), Bonham, Sabinal, Infierno, or metal arrow points, historic rock art, or pottery sherds—was included in the list of sites. Sites with dorso end scrapers also were included since Bement and Turpin (1987), Turpin (1991), and Mehalchick and Boyd (1999) consider these artifacts to be part of the Infierno tool kit. Sites with no artifacts diagnostic of the period after A.D. 1200, but had radiometric dates that fell into this time period were also included. Because Turpin believes cairns and stone rings to be diagnostic of Late Prehistoric or Historic periods occupations, sites with those features were included whether or not they had artifacts diagnostic of these periods. Burned rock middens and ring middens were included only if artifacts or radiocarbon dates from those periods were present. Finally, sites with European features or artifacts were included when those remains were believed to be associated with historic Native Americans.

Table 4 shows a wide disparity in the number of sites recorded. In Val Verde County, over 1,890 archeological sites have been recorded whereas, combined, only 1,842 have been recorded in the other five counties. Most of the sites recorded in Val Verde County are located within or in close proximity to the Amistad NRA, and the majority of these were recorded as a direct result of the archeological studies completed prior to the construction of Amistad dam or during subsequent cultural resource studies undertaken by the Amistad NRA to better manage the rich array of sites under their jurisdiction (Labadie 1994). The recent field school of the Texas Archeological Society (Collins et al. 2000) has augmented those studies in the Amistad NRA. Two other studies that added a sizable number of sites to the county's total are: 1) the dissertation research completed by Saunders (1986); and 2) the survey of the Dolan Springs Wildlife Management Area, a property of Texas Parks and Wildlife Department (Turpin and Davis 1993). Sites in the other counties were largely recorded during small CRM efforts (cf. Shafer 1971) or by the efforts of avocational archeologists. The higher quantity in Crockett County reflects the large number of sites recorded by the 1976 field school of the Texas Archeological Society in and around Musk Hog Canyon (Moore 1983). The low number of sites in each of the other counties is due to the limited archeological investigations that have taken place in those counties.

Table 4. Sites in the Lower Pecos with Features, Artifacts, Radiocarbon Dates, or Rock Art Post-Dating A.D. 1200.

One of the striking aspects of the Late Prehistoric and Historic period sites is their remarkably low number (see Table 4 and Appendix 7). In each county, they represent less than 8 percent of the total sites recorded. On the other hand, because most sites have only been subjected to surface survey, it should not be surprising that most have not yielded chronologically diagnostic artifacts, rock art, or military paraphernalia, and only a rare few have been radiocarbon dated. Thus, we caution that the total number of recorded sites in the micro-region that date between A.D. 1200 and 1880 may or may not be accurate. Only 184 sites contain materials that date from the later part of the Late Prehistoric (ca. A.D. 1200-1500) or Historic Indian (A.D. 1590-1880) periods, representing 4.9 percent of all sites recorded in these counties. The highest percentage of sites with materials dating between A.D. 1200 and 1880 is in Kinney County, where they constitute 7.52 percent of the recorded sites, most of which are sites that were occupied by Seminole Maroon in the 1872-1881 period (Mock 1994). The remaining four counties have even lower proportions with 7.2 percent in Val Verde County, 2.9 percent in Terrell County, 1.6 in Edwards County, and 1.8 percent in Crockett County. Sutton County, with only 73 total sites recorded in the entire county, has no sites that contain materials from the relevant periods. Together, these 183 sites underscore the need for both additional archeological survey and controlled subsurface investigation, and support the assertion of Mehalchick and Boyd (1999:157) that there are sampling biases that affect the database.

Keeping in mind these biases, some tentative trends are suggested by the data. First, arrow points (n=60 sites) or tipi/wickiup rings and/or cairns (n=61 sites) are the most commonly recorded site materials and these are diagnostic of the A.D. 1200 to 1880 time span. Sites with arrow points diagnostic of these periods (Late Prehistoric and Historic Indian) were mostly recorded in Val Verde County (n=52 sites). Perdiz is the diagnostic arrow point most often recorded at Val Verde County sites, being present on 49 of the 52 sites. Of the 49 Val Verde County sites with Perdiz only 11 contained other arrow points (including Toyah, Harrell, Garza, Fresno, or Infierno) diagnostic of these periods, while only three sites with other arrow points failed to include Perdiz. Another trend is that while 61 sites contain one or more types of arrow points, only 16 sites contain pottery sherds. Many of these same sites contain dart points, rock art, or other materials that are characteristic of earlier periods.

Sixty-one sites contained tipi/wickiup rings and/or cairns. Most (n=43) did not contain any artifacts or rock art dating to the relevant time periods. Since these features may or may not all date to this time period (as noted above), and since only a few of the sites with tipi/wickiup rings and/or cairns (n=17) had diagnostics dating from A.D. 1200 to 1880, the total number of sites that were included in this category may be inflated. Fifteen sites with rings (41VV398, 404, 409, 446, 635, 649, 869, 1723, 1724, 1860, 1875, 1876, 1881, 1884, and 1889) are located in or close to the Amistad NRA, most within or adjacent to Seminole Canyon State Park. Turpin (1994) has interpreted these features as loci where signal fires were built, but further examination of this interpretation is sorely needed and other hypotheses are possible. Seventeen of the 61 sites have cairns. Of these, only five did not also contain tipi/wickiup rings.

The number of rings present is not recorded on eight sites. On the remaining 53 sites, the number of stone rings varies somewhat but most have relatively few rings. Forty-five sites (one in Crockett County, three in Terrell County, and 41 in Val Verde County) had one to five rings. One of the sites with a single ring is 41VV880 on Dolan Creek, a site believed to represent the remains of a U.S. Cavalry site, possibly associated with the 1857 engagement between the U.S. Army and the Comanche (Turpin and Davis 1993:7). Three sites with tipi/wickiup rings and/or cairns had six to 10 rings (one in Terrell County and two in Val Verde County). The remaining three sites with rings all have in excess of 100 rings; no cairns have been reported at any of the three. All three sites are in Val Verde County, and two (41VV1723 and 41VV1724) overlook the Rio Grande while 41VV446 overlooks Seminole Canyon just a short distance north of the Rio Grande. Both 41VV1723 and 41VV1724 are described as containing several hundred tipi/wickiup rings, along with a number of hearths, and mortar holes. Recent recording of Infierno Camp (41VV446) shows more than 140 tipi/wickiup rings at that site. Perdiz and Infierno arrow points were recorded at the Infierno Camp along with several sherds. No arrow points were recovered at 41VV1723 or 41VV1724, although at least one sherd was found at 41VV1723.

Interestingly, the presence of these three sites (41VV446, 1723, and 1724) with over 100 tipi/wickiup rings on or close to the Rio Grande, conforms to some of the archival information that place large groups of Apache and Comanche camping on or close to the Rio Grande during the late eighteenth and nineteenth centuries. Table 5 lists those reports and their sources. Together, the reports demonstrate the presence and affiliation of these Native Americans with the Amistad NRA. As we noted in the opening chapter, however, the reports indicate that these groups moved frequently within the Amistad NRA and the micro-region and were often present in the macro-region as well.

Table 5. Reports of Apache or Comanche Camping on or Crossing Lands of the Amistad NRA.

Given these reports, it is tempting to conclude that the three sites with over 100 tipi/wickiup rings represent Apache or Comanche encampments since these Native Americans are known to have been present in the area, in relatively large numbers, intermittently, for a long period of time. Moreover, they are known to have erected small, conical structures at their camps (Kavanagh 1996; Merino in John and Wheat 1991:150), and tipi/wickiup rings have a long history in the northern Plains regions from which they migrated (Oetelaar 2000). Nonetheless, the data are insufficient to prove that sites 41VV446, 41VV1723, or 41VV1724 are the remains of Apache or Comanche camps, and several factors cloud any such conclusion. First, in 1875 and 1876, large numbers of Comanches were reported present at the Eagle's Nest crossing of the Rio Grande near modern Langtry. The 1875 report even documents an engagement of the Comanche with Bullis and his Seminole Maroon scouts at the mouth of the Pecos. Nonetheless, no sites with tipi/wickiup rings or with U.S. military artifacts have yet been recorded in or close to Langtry (see Appendix 7). Second, several reports refer to a popular crossing at the mouth of the Pecos. Again, no sites with tipi/wickiup rings or military artifacts have been recorded at the mouth of the Pecos. The closest site with tipi/wickiup rings is 41VV446 (Infierno Camp), located ca. two miles east of the confluence of the Pecos with the Rio Grande on the eastern side of Seminole Canyon and ca. one mile north of the Rio Grande. To date, however, no reported artifacts from 41VV446 are of military origin. A brass trigger guard and beveled end scrapers were recovered from the Old Saloon ruins (41VV544), a site situated in Seminole Canyon. The closest site reported to date that compares to the reports that Comanche and Apache crossed at the mouth of the Pecos is 41VV1428, Bullis Crossing. Supporting evidence that this may indeed correspond to the crossing at the mouth of the Pecos is the presence of another site close to Bullis Crossing that is believed to be a U.S. military camp from 1875. This is 41VV1651, which contains historic U.S. military artifacts and features from the late nineteenth century and is believed to have been one of the encampments used by Bullis and his Seminole Maroon scouts (see Crimmins Collection; Swanson n.d.:209-210).

In sum, despite some tantalizing evidence, it is not yet possible to know if these three sites, with their large numbers of stone rings, represent the remains of Apache or Comanche encampments in or near the Amistad NRA. Since 100 percent archeological survey of the Amistad NRA was not undertaken prior to inundation of the reservoir, since sites from the period A.D. 1200 through 1880 tend to yield few artifacts, and since more recent settlements at Langtry and elsewhere may have destroyed or masked earlier components, it is not clear if we will ever be able to associate specific reports of Native American camps with specific sites.

Painted Cave (41VV7, also called Castle Canyon) is one of the few sites that can be associated with specific historic reports. The site is mentioned in an 1875 report (Crimmins Collection; Swanson n.d.:205). It was also mentioned in an 1849 report of Captain French who was conducting a reconnaissance of routes from San Antonio to El Paso (Jackson 1938:231). Located on a tributary of the Devils River, the site consists of several rock shelters fronted by a burned rock midden. The rock art is present in several of the shelters and dates from the prehistoric era, but also from historic times (Greer 1966:17-18; Jackson 1938:238). The latter consists of a mission-style structure and a possible Spanish figure, but horses and cows are also present (Greer 1966:17-18). Excavations at the site recovered Perdiz, Toyah, and Harrell arrow points as well as material from Archaic occupations.

Several sites in the micro-region also can be related to historic military activities. In addition to 41VV1428 (Bullis Crossing) and its associated camp site (41VV1651) mentioned above, 41VV795 is an historic crossing of the Rio Grande known as Dweese's Crossing or Dweese's Vega (vega means trail in Spanish). Seven of the 11 sites in Kinney County included in Table 4 and Appendix 7 represent house sites of the Seminole Maroon settlements on Las Moras Creek at Fort Clark, and another is a ditch they excavated to improve the irrigation of their fields. Yet another site (41KI19) is the Seminole Indian Cemetery with graves of both Seminole and Seminole Maroon families from 1872 to the present. The other eight sites with historic military remains are: two sites (41VV841 and 41VV405) with possible gun flints; Snake Springs (41VV880) where the 1857 battle between the Comanche and the U.S. military occurred; Baker's Crossing (41VV424) of the Devils River near Camp Hudson (dating to ca. 1857); Howard's Well (41CX273) where the 1872 Kiowa battle with the U.S. Army was fought; and Howard's Stage Depot (41CX774), dating from the 1870s.

Several rock art sites are also considered relevant to the present study. In all, 22 sites contain Historic Period rock art (Table 6). Although some of the 22 also contain rock art or artifacts diagnostic of Archaic periods, few contain arrow points or sherds, and, hence, support Turpin's (1991:38) assertion that rock art is the only evidence of the use of the micro-region by Native American groups during the Historic period. Historic era rock art sites in Val Verde County (n=19) can be generally divided into early and late periods, and many of these are in or close to the Amistad NRA. The three Terrell County sites all contain historic rock art from the late period, but one site (41TE9) also has some panels with the early style. The early Historic period rock art sites depict mission-like square or rectangular structures with steep, pitched roofs and crosses extending skyward. Others exhibit human figures that are thought to represent soldiers or priests. Some of the latter are associated with individual crosses and/or hands.

Table 6. Historic Era Native American Rock Art Sites in the Micro-Region.

Eight sites in Val Verde County contain rock art from the early part of the Historic period. Vaquero shelter (41VV77) is located in Seminole Canyon State Historical Park in a shallow overhang. The early artwork here depicts a mission and three crosses (Figure 27). Other mission structures are depicted in rock art panels at 41VV7 (described above), 41VV343, and 41VV570. The mission-like structure depicted at 41VV570 is the only historic rock art at the site; the remainder of the rock art at the site appears to be prehistoric. At 41VV343, a panel with the mission-like structure also contains a horse carrying a man whose hat appears to be that of a priest (Figure 28). Site 41VV343 is located on the Dolan Creek Ranch, north of the Amistad NRA. Located just outside the boundaries of the Amistad NRA on lands owned by Texas Tech University and known as Rattlesnake Canyon, 41VV180 is one of a series of rock art sites in a narrow canyon. The site was recorded some years later as 41VV205 and named Missionary Shelter; it is, however, the same site as 41VV180. Today, the site is almost completely faded, but when first recorded by A.T. Jackson, it held a missionary-like figure pierced by a cross (Figure 29). The artwork also contained several crosses and a horse drawn in a boat-like style. One of the figures drawn in a rock art panel at 41VV327, a shelter on the Hussie Miers Ranch outside of the Amistad NRA, appears to be a Spanish soldier holding a musket. In addition to this Spanish-like individual, a Native American on a horse is shown. However, the native and horse appear to represent a latter style (discussed below), perhaps suggesting that Turpin's (1989) identification of the European as a Spaniard is erroneous. Site 41VV339, on this same ranch, depicts several horseshoes, horses with Spanish ring bits, and riders who have been impaled with spears (Figure 29).

Figure 27. Mission and three crosses drawn at Vaquero Shelter (41VV77) (Photograph courtesy Texas Historical Commission).

Figure 28. Mission structure at site 41VV343 (Photograph courtesy The University of Texas at Austin, Texas Archeological Research Laboratory, 41VV343-C10).

Figure 29. Missionary and church-like structure at site 41VV180 (Photograph courtesy The University of Texas at Austin, Texas Archeological Research Laboratory, 41VV180-23).

One site in Coahuila should be mentioned with these four early historic Indian rock art sites: the Cruz Electrica site near Musquiz, Coahuila (Sayther 1998:90-91, and Figure 3). The dominant artwork at the site consists of a cross approximately one meter tall surrounded by a zigzag line (giving it the appellation "electrica"). Seen alone, the cross might not be considered historic. However, beneath one arm of the cross is a human figure wearing a tank-like garment that extends to its knees and has wide sleeves. Its arms are raised above its head and the figure is wearing a hat that appears similar to those worn by seventeenth century Catholic priests. This is the only rock art panel identified to date in northern Coahuila that contains this early style of historic rock art. Whether the large cross also dates to the same time period is unclear.

As noted above, we believe these early style historic rock art panels relate to the period prior to 1750. Spanish priests began to invest time and interest in the region by the 1670s, and several missions were established in or close to the micro-region during this period. While most failed, the sites represent, in our opinion, expressions of the Native American interest in these institutions.

The bulk of the rock art in the micro-region is of the Plains Biographic style (Keyser 1987), found at 18 Val Verde County sites. Elsewhere in the micro-region these sites are restricted to three sites in Terrell County. No other sites in the micro-region have been recorded that contain these examples of later period rock art, although three have been recorded and reported from south of the river (Labadie et al. 1997; Turpin 1987). Whether the paucity of sites reflects a bias of archeological sampling or reflects the fact that, as Berroteran found in 1729, the area around modern Del Rio had few occupants but was consistently used as a favored crossing of the river, is unclear.

This late group of historic Indian rock art sites frequently contain one or more horses and/or cattle with horns. Some of the horned animals depict bison. According to Turpin (1986) these pictographs represent scenes showing hostility and aggression. Sometimes the horses in these scenes are riderless; sometimes they carry Anglo riders. In other cases, the riders are Native American, often with Plains style clothing and/or headdresses (e.g., 41VV327 as described in Turpin 1989). Lances, bows and arrows, muskets, and shields are among the other elements drawn in these scenes. Only three sites have illustrations of bison, and at least one of these (41VV485) is especially life-like in its depiction. Within the Plains Biographic Style, Keyser (1987) has chronologically subdivided the art work. The earliest generally date from 1775 to 1830, and a later, more refined, style of this art dates from 1830 to 1860. In the former, humans are typically square bodied with a v-neck, and horses are drawn like boats with stubby legs and often exhibit little or no neck (Keyser 1987:52). Change to a more realistic representation occurred as a result of Native American exposure to early Anglo-American artists (Bodmer, Catlin, etc.). As a result, horses exhibited rounder bodies, longer legs, necks, and hooves. Often horse trappings (reins, bridles, and saddles) became more carefully detailed, as well, and humans were depicted as rounder and less rigid figures (Figure 30).

Figure 30. Riders impaled with spears at site 41VV339 (Photograph courtesy Texas Historical Commission).

Five Val Verde County rock art sites contain scenes that can be associated with the Plains Indian Biographic Style (Turpin 1989; Keyser 1987; Parsons 1987); several additional sites of this era have been recorded in Coahuila (Turpin 1987). Two of the sites (41VV18 and 41VV910) are inadequately described. Site 41VV18 is A.T. Jackson's site 99, located on the Devils River, just outside the lands of the Amistad NRA. The site is described in the TARL site files as "a possible historic Indian pictograph," and has a human figure drawn in a realistic style. The site form for 41VV910 indicates that the site is located under the waters of the reservoir and that it too contained an historic pictograph but no additional details were provided. A single bison is present at 41VV202 in the Devils River drainage, close to Amistad NRA. The drawing could be from any period since bison were present during several mesic intervals (including the historic era, see Ayer [17291). However, it is presumed here to represent the final period of Indian rock art since its lifelike proportions best fit the Plains Indian Biographic Style (Keyser 1987:54), and likely the latter portion of that period. Finally, 41VV485 contains another bison that is quite lifelike, as well as a man with a musket (see Figure 26).

The Hussie Miers site (41VV327), is one of the more well known of the late Historic Indian rock art sites. Five panels are present at the site and each has drawings that match the latter style of the Plains Biographic art with more realistic humans, often with weapons and horses depicted with longer necks, hooves, tails, and manes and led via reins. The humans either wear native garb with long headdresses or are United States soldiers in military attire. "Turpin believes these panels are Plains combat autobiography style . . . commemorating the exploits of a single individual who defeated . . . all foes" (Labadie et al. 1997:18).

Finally, several additional sites from Coahuila should be noted. The first is the Caido Site, located just south of the Rio Grande in Coahuila, and described and recorded by Labadie et al. (1997). Uniquely, the site appears to contain both early and late styles of Plains Biographic style art. In the early depictions (Panel B), the artwork includes small tipis, individuals attired in what appear to be loincloths, and several individuals with headdresses and lances. In the later effort (Panel A), individuals are shown in larger format, mounted on horses, and with "conspicuous symbols of individual achievements in warfare" (Labadie et al. 1997:29). Given the style, the authors concluded that the latter panel was likely drawn in the late nineteenth century. Moreover, they speculated that one of the mounted horsemen in Panel A (with long, red, decorated hair) can be related to the Hussie Miers site "hero" described above. Other examples of the Plains Biographic style are present in pictographs in central Coahuila at Los Alamos, La Tinaja de Acebuches (Sayther 1998:93-5) and at Arroyo de los Indios (Turpin 1987).

SUMMARY

As we have attempted to show in the preceding pages, the archeological data corresponds with some of the documentary data. The earliest documents strongly suggest that the region was populated by groups of hunters and gatherers who relied on local resources. In our opinion, some of these may represent the Toyah-folk described by Johnson (1994:187-277). Although some documents suggest that the population was low, others indicate the presence of groups with as many as 100 to 500 individuals or more. The documentary evidence also suggests that bison were present in the region by A.D. 1650. Later documents continue to note the presence of bison, and substantiate the presence of Apache and Comanche in the Lower Pecos. While the documents indicate they moved frequently, the Apache were present from at least the mid-eighteenth century, according to the documents, and maintained their presence until the mid-nineteenth century, residing throughout the lands of the Lower Pecos as well as in northern Mexico. In contrast, the Comanche, while present from time to time, appear to have used the Lower Pecos as a travel corridor, traveling to and from Mexico where they raided for horses and other material objects, then consistently returning to their northern homelands.

Contrary to the documentary data, few sites dating after A.D. 1500 have been identified and even when the data from the period A.D. 1200 to 1880 are included, the total number of sites remains low: 184 sites. Given the fact that over 3,700 sites are recorded in the five counties considered, the total number of recorded sites dating to the late periods is markedly small. The rock art for the same periods is similarly sparse (n=22). The artwork can be divided into an early missionary (or Spanish) period and a later (or Plains Indian) period. Infierno Camp (41VV446) with its multiple tipi/wickiup rings, certainly demonstrates that large groups of historic Native Americans did occupy the region. Unfortunately, this is the only one of three such large sites that has been recorded to date. While the three sites support the notion that large groups occupied the region, either their occupation was infrequent, or their stay was sufficiently brief that their camps have gone undetected, or both.

NOTES

1. It should be noted, however, that several cairn burials have been investigated in Reeves County and northern Chihuahua (Mallouf 1987:5-9). The latter was accompanied by Perdiz arrow points; the former was accompanied by Livermore arrow points. However, Mallouf's (1987:5-6 and Table 1) synthesis of scientifically investigated burials throughout the Big Bend region reveals that only the Las Haciendas burial in northeastern Chihuahua was of a cairn type.

2. While this hypothesis is the one Johnson favors, the reader should note that he also recognizes that others are feasible (Johnson 1994:187, 277-281).