Browse this book with Google Preview »
The discoveries at the Folsom site in New Mexico from 1926-1928 clearly established human antiquity in the New World to late Pleistocene times (Wormington 1957; Meltzer 1983). The discoveries also ushered in a new era in North American archaeology: Paleoindian studies on the Great Plains. Some of the best-known and most intensively studied early sites in the Western Hemisphere are on the open grasslands of the midcontinent, and they have provided archaeologists with many presently held conceptsand biases—regarding Paleoindian typologies, chronologies, and subsistence (Wormington 1957; Wilmsen 1965; Meltzer and Smith 1986; Meltzer 1993).
Within the Great Plains, one of the highest concentrations of Paleoindian sites and research is on the Southern High Plains of northwestern Texas and eastern New Mexico (figs. 1.1, 1.2). The first excavations at the Clovis site, New Mexico, just a few years after the Folsom finds, began one of the longest traditions of Paleoindian archaeological studies in any region in North America. Moreover, the work at Clovis established a tradition of integrating geoscientific studies with Paleoindian research that still thrives on the Southern High Plains over 60 years later. Interdisciplinary Paleoindian research has been almost continuous in the region since 1933 and has included work at several type localities (Clovis, Plainview, San Jon, Midland, Milnesand), other well-known, key sites (Lubbock Lake, Miami), and a number of lesser-known but nevertheless significant sites (fig. 1.3).
This volume is a synthesis of the available data on the stratigraphic, geomorphic, chronologic, and paleoenvironmental contexts of Paleoindian occupations on the Southern High Plains. The data base includes information from many previous studies as well as new data that I gathered from most of the reported, stratified Paleoindian sites. The results of this study are intended to provide a physical and temporal basis for more expressly archaeological investigations of Paleoindian occupation. Some archaeological interpretations dealing with issues of artifact chronologies, site settings, and environmental changes are offered, but this study is not an exhaustive treatment of Paleoindian archaeology. Such a task is left to archaeologists.
A synthetic treatment of Paleoindian studies on the Southern High Plains is warranted because of the long and varied history of Paleoindian research in the region and because of the importance of this research in North American archaeology. The archaeological and geoscientific investigations of Paleoindian sites on the Southern High Plains are significant for several reasons. These studies were the first to: establish the stratigraphic relationships among some of the various Plains Paleoindian occupations, including the now-classic sequence of Clovis-Folsom-Lanceolate projectile points; provide radiocarbon age control for Paleoindian occupations; and yield clues to the very different and changing environments in which Paleoindians lived. Paleoindian discoveries at several sites also led to broader investigations of late Quaternary geohistory and paleoenvironments, such as the decades of geoarchaeological work at Clovis by C. V Haynes (Haynes and Agogino 1966; Haynes 1975, 1995; Stanford et al. 1990), and the first systematic treatment of playa basin origins and development by S. Judson, which began with his research at San Jon (Judson 1950, 1953). Paleoindian finds also led to two of the few systematic, regional studies of Paleoindian archaeology and paleoenvironments in North America by E. H. Sellards and colleagues (late 1930s through 1950s) (Sellards 1952; Sellards and Evans 1960), and by F. Wendorf and his associates (late 1950s and 1960s) (Wendorf 1961a; Wendorf and Hester 1975).
Since 1933 there have been scores of investigators working on dozens of Paleoindian sites on the Southern High Plains under a variety of conditions with varying aims. As a result, there are more data on Paleoindian archaeology, geology, and paleoenvironments than for most other regions of comparable size in the Americas. However, there have been few attempts to synthesize the results. The first summaries of the cultural sequence and integration of the geologic framework were by Sellards (1952) in his well-known volume Early Man in America and by Sellards and Evans (1960), who proposed the classic Paleoindian cultural and environmental succession for the region. Subsequent investigators (Wendorf 1961b; Kelley 1964; Collins 1971; Hester 1975a; Wendorf and Hester 1975; Hester and Grady 1977) discuss and synthesize aspects of the Paleoindian cultural succession, geoarchaeology, social patterns, and environments on the Llano Estacado, but these studies now are dated, summarizing work completed before or in the 1960s. Haynes (1968a, 1975) and Stafford (1981) deal with the geologic environments of Paleoindians, but in geographically limited areas of the Southern High Plains. Hofman (1989a) presents the most comprehensive chronology of Paleoindian occupations, based on the considerable research by many investigators in the 1970s and 1980s, but this study is not within a geologic framework and the paleoenvironmental context is limited.
Another goal of this volume is to synthesize the paleoenvironmental data for reconstruction of regional environments during the Paleoindian occupation. Most of the data on Paleoindian environments of the Southern High Plains are site-specific, and regional paleoenvironmental reconstructions are relatively rare and based on comparatively sparse data. This point is well illustrated by the results of COHMAP (1988; Wright et al. 1993) As part of the COHMAP research, the vegetation reconstructions based on proxy data for western and eastern North America (Thompson et al. 1993 Webb et al. 1993) display very little information for the Great Plains and essentially none for the southern Plains.
Regionally, a synthesis of the data for the Southern High Plains is timely, given the record of Paleoindian occupations emerging from the northern Plains (Frison 1991a, 1993). Moreover, many of the "well-known" sites and "classic" studies on the Southern High Plains are not all that well known or well reported (similar to the situation on the northern Plains; Tiffany 1993:2-3). The stratigraphic and geochronology data from many of these sites are decades old, and much of the data received only cursory discussion or none at all. A reexamination of the geologic record of each site for its implications regarding paleoenvironments and cultural chronology is warranted in view of the considerable interdisciplinary research in the area since the early 1970s.
The long record of interdisciplinary Paleoindian research on the Southern High Plains is unique in North America and is due to several circumstances. Following the discoveries at the Folsom site, the impetus for Paleoindian studies on the Great Plains resulted from repeated finds of Folsom or Folsom-like artifacts or additional discoveries of human artifacts associated with Pleistocene fauna. On the Southern High Plains these finds included the Clovis site (Howard 1935a,b), the Lipscomb site (Barbour and Schultz 1941), and the Miami site (Sellards 1938).
Perhaps more important to the development of interdisciplinary studies on the Southern High Plains, many of the archaeological sites investigated have thick, well-stratified deposits that provide evidence of markedly different depositional environments in the past (e.g., meandering streams where arroyos now prevail or perennial, freshwater lakes in presently dry basins) owing to the dramatically different environmental conditions at the end of the Pleistocene (Haynes 1990; Ferring 1994). These striking contrasts between past and present depositional environments drew the attention of archaeologists and earth scientists alike who recognized the paleoenvironmental implications. Indeed, during the first excavations at Clovis, E. B. Howard (1935a:62) was explicit in his belief that interdisciplinary research was a prerequisite to understanding Paleoindian archaeology, and especially focused on geology:
Geology, particularly its allied branches of palaeontology, physiography, and glacial geology, must be called upon to explain many phases of the subject [of the peopling of the New World] that involve a wide variety of converging lines of research, presenting many peculiar difficulties. The archaeologist, starting from the point where the historian usually leaves off, soon finds it necessary to lengthen his perspective, and eventually he is faced, so far as America is concerned, with a geological problem. The recognition on his part of the importance of special studies relating to such factors as climatic changes, studies of invertebrates, analysis of diatoms, or pollen that may be found in a given deposit marks a step in the right direction. Therefore the archaeologist must familiarize himself with these and other phases of geology which bear upon the problem, such as the study of terraces, buried soil levels, loess deposits, varved clays, ancient lakes and shore lines, and any other factors which may give a clue to the environment in which early man lived in America ... [T]he importance of a field of investigation which lies somewhere between geology and archaeology ... is becoming increasingly apparent as a number of scientists recognize.
Rephrased in a more contemporary context, "the prominent role of geology in Paleoindian archaeology ... is explained [in part] by ... the distinctive archaeological, paleoenvironmental, and evolutionary problems that are addressed by students of the Paleoindian period" (Ferring 1994:57).
The geologic records at the Paleoindian sites stimulated interdisciplinary geoarchaeological research for several other reasons. One of the most direct applications of earth science in the early decades of Paleoindian studies was for dating. Geologists and paleontologists were called upon to provide age estimates of sites in the absence of other forms of numerical age control such as radiocarbon dating (Hofman 1989a:44; Ferring 1990a; Haynes 1990). More significantly, however, the Paleoindian sites were inviting because many investigators had interests in Pleistocene stratigraphy and paleontology (e.g., Haynes 1990). The presence of a Pleistocene archaeological record in North America was "an enormous stimulus to research" (Bryan 1941a: 508) by geologists on Paleoindian sites (Wilmsen 1965). Many of the Paleoindian investigators on the Southern High Plains had academic backgrounds in geology in addition to an abiding interest in human prehistory, especially the issue of "Early Man" (Claude C. Albritton, Glen L. Evans, F. Earl Green, C. Vance Haynes, Jr., E. B. Howard, Grayson E. Meade, and E. H. Sellards). Geologic research was inseparable from their approach to archaeology. Such backgrounds also added immeasurably to their interpretations of paleoenvironments and geochronology. The Paleoindian archaeology of the Southern High Plains also happened to attract the attention of archaeologists trained in an interdisciplinary approach (E. B. Howard, Eileen Johnson, and Fred Wendorf ). Geoarchaeology was an integral part of their research programs.
Interdisciplinary Paleoindian investigations on and near the Southern High Plains in recent decades provide considerable new data on archaeology, chronology, and paleoenvironments, and this volume is a summary of much of this work along with the older data in light of the new information. The emphasis is on summarizing and comparing the geoarchaeology of Paleoindian occupations. From 1980 to 1995 I investigated most Paleoindian sites in the region for which data are published, focusing on the relationships of the archaeology to the soils, stratigraphy, and geochronology. These and previously published data are used for stratigraphic correlation, for paleoenvironmental summaries, and for refining or redefining the cultural chronology.
There are four main components to this study. Following introductory material on the physiography of the Southern High Plains, Chapter 2 provides a history of Paleoindian research in the region. Many Paleoindian researchers worked in the area between 1933 and 1995, usually independent of one another, resulting in a vast and often confusing array of data and ideas. The chapter, focusing on geoarchaeological and geochronological aspects of Paleoindian studies, provides an historical perspective on the individuals and institutions involved, on the discoveries, and on the evolving interpretations. Chapter 3 is a discussion of the geoarchaeology of the individual Paleoindian sites that integrates the new and old data on site geomorphology, stratigraphy, geochronology, and paleoenvironments. Chapter 4 is a survey of Paleoindian geoarchaeology and cultural chronology in neighboring areas of the Great Plains. The discussion is intended to place the Paleoindian record of the Southern High Plains into a broader chronological and paleoenvironmental perspective, in light of long-term and well-known Paleoindian studies from the continental interior of North America. Chapter 5 then presents my views on Paleoindian cultural chronology and paleoenvironments on the Southern High Plains based on the stratigraphic and geochronologic data along with other archaeological and paleoenvironmental data. Two appendixes also are included, presenting descriptions of site settings and stratigraphy (Appendix 1) and a discussion of stone-tool resources (Appendix 2).
The Southern High Plains or Llano Estacado probably is the most environmentally homogeneous region of its size in North America, with flat topography, relatively stable, uniform regional geology, and low, smooth environmental gradients. The Llano Estacado is a vast, level plateau covering approximately 130,000 km2 and comprising the southernmost portion of the High Plains physiographic section (Fenneman 1931; Hunt 1974) (figs. 1.1, 1.2). The plateau is defined by escarpments 50 to 200 m high on three sides. The western escarpment separates the plateau from the Pecos River valley, and the northern escarpment separates the plateau from the Canadian River valley. The eastern escarpment, formed by headward erosion of tributaries of the Red, Brazos, and Colorado rivers, separates the Southern High Plains from the Rolling (or Osage) Plains. The eastern escarpment provided the name Llano Estacado or "stockaded plains"—the prominent topographic break took on the appearance of an immense stockade to Spanish explorers traveling west across the Rolling Plains (Bolton 1990:243). To the south, the surface of the Southern High Plains merges with the surface of the Edwards Plateau province of Central Texas with no topographic demarcation. The southern boundary is defined by the northernmost outcrops of Edwards Limestone (Fenneman 1931; Hunt 1974).
The climate of the Southern High Plains is continental and semiarid, classified as BScDw: steppe with dry winters, mainly mesothermal years (mean temperature of the coldest month is 32-64º F) with occasional microthermal years (mean temperature of the coldest month is below 32º F) (Russell 1945). There are relatively uniform gradients in precipitation and temperature across the region: precipitation generally increases from west to east and temperature usually increases from northwest to southeast (Lotspeich and Everhart 1962; Carr 1967).
The natural vegetation of the Llano Estacado is a mixed-prairie grassland (Blair 1950; Lotspeich and Everhart 1962). The dominant native plant community is short-grass, which includes types of grama (Bouteloua sp.) and buffalo grass (Buchloe dactyloides). Trees are absent except along the escarpments and reentrant canyons. The floristic composition varies somewhat from north to south due to changes in climate and soil texture. Native Plant communities of the region occur in very few areas today, however, because most of the Southern High Plains is under cultivation. On a geologic time scale, the Llano Estacado probably was a grassland throughout the Holocene and probably varied from a subhumid to semiarid savanna to semiarid grassland in the Pleistocene as effective precipitation varied (Johnson 1986, 1987a; Holliday 1987a, 1989a).
The Southern High Plains is an almost featureless plain, "...the largest level plain of its kind in the United States" (NOAA 1982:3). There is a regional slope to the southeast with altitudes ranging from 1700 m in the northwest to 750 m in the southeast. Slight topographic relief is provided by lake basins, dunes, and dry valleys (Reeves 1965, 1966, 1972, 1991; Wendorf 1975a; Hawley et al. 1976; Walker 1978; Holliday 1985a) (fig. 1.2). There are about 25,000 small (<5 km2) depressions, or "playas;" dotting the landscape and containing seasonal lakes, and about 40 larger (tens of km2) basins, also called "playas" or "salinas" (fig. 1.2). The playa and salina basins contain the only available surface water on the Llano Estacado, although the water is seasonal and often brackish or saline, especially in the salinas. Lunettes (fringing dunes) often are found adjacent to the playas, and several large sand dune fields are present along the western Llano Estacado (fig. 1.2). The dry valleys or "draws" are northwestsoutheast trending tributaries of rivers on the Rolling Plains to the east (fig. 1.2).
Most of the rocks and sediments of the Llano Estacado are Cenozoic deposits. These layers overlie Mesozoic sedimentary rocks and are separated from them by an early Tertiary erosion surface (Harbour 1975; Gustavson and Finley 1985; Reeves 1972; Hawley et al. 1976; Winkler 1987; Gustavson and Winkler 1988). The bulk of the Cenozoic deposits are MiocenePliocene eolian and alluvial sediments of the Ogallala Formation (fig. 1.4), ultimately derived from mountains to the west in New Mexico. The upper Ogallala has a thick, highly resistant, pedogenic calcrete. This Ogallala "Caprock Caliche" is a prominent ledge-forming unit near the top of the escarpments bordering the plateau.
Another extensive Pliocene deposit is the Blanco Formation, a layer of lacustrine dolomite and clastic sediment deposited in large basins cut into the Ogallala (Evans and Meade 1945; Harbour 1975; Hawley et al. 1976; Pierce 1974; Holliday 1988a) (fig. 1.4). A calcrete also formed at the top of the Blanco Formation. Other, more localized lacustrine deposits of the region include the Tule Formation (early to middle Pleistocene) and the Double Lakes and Tahoka formations (both late Pleistocene), and other, unnamed deposits (Harbour 1975; Reeves 1976; Gustavson et al. 1990). Differentiating these Pliocene and Pleistocene lake sediments on lithological bases is very difficult, however (Harbour 1975; Holliday 1995a).
The Blackwater Draw Formation (Reeves 1976) (fig. 1.4) is the major surficial deposit of the Southern High Plains, blanketing all older units. The formation is a widespread eolian deposit heavily modified by pedogenesis (Holliday 1989a). It was derived from the Pecos River valley and varies in thickness and particle size from a thin veneer of sandy loam in the southwest to a thick deposit of clay loam in the northeast (Reeves 1976; Holliday 1989a). The Pleistocene lake deposits underlie, are interbedded with, or are inset against these eolian sediments (fig. 1.4). The Blackwater Draw Formation probably is the upland equivalent of the Judkins Formation in the Pecos Valley to the southwest of the Llano Estacado (Huffington and Albritton 1941) and the Lingos Formation on the Rolling Plains to the east (Caran and Baumgardner 1990).
The late Quaternary (post-Blackwater Draw Formation) stratigraphic record of the Southern High Plains, which contains the in situ archaeological record, is found in draws, playas, salinas, and dunes (Harbour 1975; Hawley et al. 1976; Holliday 1985a, 1995a) (fig. 1.4). The draws of the region are dry tributaries of the Red, Brazos, and Colorado rivers (fig. 1.2), and are "elongate stream valleys having narrow drainage basins" (Gustavson and Finley 1985:33). Between the draws are expansive areas of the High Plains surface with no integrated drainage. The draws are inset into the Blackwater Draw Formation and locally cut into lake beds or the Ogallala Formation. Within these dry tributaries are a variety of late Pleistocene and Holocene lacustrine, paludal, alluvial, and eolian deposits, many containing buried soils (Holliday 1995b). The late Quaternary stratigraphy of the region is best known from the draws, based largely on archaeological research.
Playas and salinas are lake basins inset into the Blackwater Draw Formation and locally into older units (Reeves 1990, 1991; Gustavson et al. 1995 Holliday 1995a). These basins contain late Pleistocene and Holocene lacustrine and paludal sediments. Some of the late Pleistocene sediments are considered members of the Tahoka Formation, but other late Pleistocene fills are not Tahoka and all of the Holocene sediment is considered post-Tahoka (Reeves 1991), further illustrating the difficulties in using formal lithostratigraphic terminology for lake sediments of the region.
Dunes on the Southern High Plains occur as sand dune fields or as lunettes in or adjacent to some playa basins and most Salina basins (Holliday 1995a). These two types of dunes have different origins and different stratigraphic records. The lunettes are localized accumulations of eolian sediment deflated from playas in the late Pleistocene and Holocene. They are on the northeast, east, or southeast sides of the lake basins. The dune fields primarily consist of Holocene sands that probably originated in the Pecos Valley, immediately west of the Llano Estacado (figs. 1.1, 1.2). There are three sets of extensive dunes on the western side of the Southern High Plains (Hawley et al. 1976; Holliday 1985a, 1995a) (fig. 1.2).
Geoarchaeologists, especially in Paleoindian studies, traditionally have not dealt with issues of lithic resources in a systematic, regional manner (e.g., Waters 1992: xix-xx). Although geoscientists have been involved in studies of stone-tool resources, this work usually either is locality-specific (focused on a single resource or type of resource or a particular type of method) (e.g., Bryan 1950; Godfrey-Smith et al. 1993) or is done by archaeologists (e.g., Luedtke 1992; Church 1994). Indeed, the focus of this monograph is on aspects of geoarchaeology and Paleoindian studies far removed from the study of stone-tool resources. Nevertheless, the source areas and types of rocks used to fashion artifacts clearly are geoarchaeological issues, and with regard to the Southern High Plains, matters of resource availability frequently are raised in studies of artifact manufacture, reuse, and typology (Hofman 1992; Johnson and Holliday 1980, 1981; see Chapter 5), as well as matters of settlement, mobility, and adaptations (Meltzer 1986; Hofman 1992). To provide some background and perspective on the location and types of knappable stone resources available to prehistoric inhabitants of the Southern High Plains, Appendix 2 was prepared.
Field work aimed at gathering new data on lithic resources was not a part of the research that led to this monograph, and little new data are presented in Appendix 2. Rather, this discussion is a summary of the litera ture published on stone-tool resources on, near, or otherwise available to Paleoindians on the Llano Estacado. There have been no systematic studies specifically focusing on this topic, but there are several regional overviews (Holliday and Welty 1981; Banks 1990) and numerous site-specific or lithology-specific studies that form the basis for the presentation.
A variety of materials suitable for manufacturing lithic artifacts are available along or beyond the margins of the Southern High Plains. The material includes flint, chert, agate, jasper, various quartzites, and some opal (Appendix 2). Paleoindians in particular favored the high-quality Alibates agate, Tecovas jasper, and so-called Edwards chert (Hester 1975a; Hofman 1989b, 1991, 1992). The Ogallala Formation has the greatest variety and abundance of material in the area, however. The availability of such a variety of material places some limitations on the use of rock types found in archaeological sites for estimating possible interregional contact. A notable example is the gravel of high-quality chert and flint in the Ogallala and in the gravel derived from the Ogallala and redeposited in terraces on the Rolling Plains. The material is macroscopically identical to chert found on the Edwards Plateau in Central Texas and probably was derived from outcrops of Cretaceous limestone in the area during deposition of the Ogallala. Hofman (1989b:7) suggests that the Ogallala and younger gravels were not exposed until post-Paleoindian times by drying conditions and erosion. No stratigraphic or geomorphic evidence is available to support this hypothesis, however. Such a cover over Ogallala outcrops along the High Plains escarpment should be preserved locally, but is not, and massive stripping of a conjectured cover along the escarpments seems unlikely. Moreover, the extensive gravel deposits on the Rolling Plains almost certainly have been exposed since deposition.
The importance of raw material availability and types has long been recognized as a factor in determining the quality and quantity of tools and debitage in archaeological sites on the Southern High Plains (e.g., Green and Kelley 1960). Of particular significance is the paucity of suitable material for stone-tool manufacture except along or beyond the margins of the region (Collins 1971; Hester 1975a; Holliday and Welty 1981; Johnson and Holliday 1987b). This geological characteristic of the region is a direct result of the late Cenozoic landscape evolution. The region was blanketed by relatively young eolian sediments (the Blackwater Draw Formation) and has undergone very limited erosion except along the margins in the past few million years (Holliday 1989a, 1990b; Holliday and Gustavson 1991). Rock outcrops, therefore, are common only along the escarpments that form the west, north, and east boundaries of the Llano Estacado. Other rock outcrops that were sources or potential sources are found on the Rolling Plains and possibly in the mountains of western Oklahoma to the east, the Edwards Plateau to the southeast and south, the mountains of far western Texas and central New Mexico to the southwest and west, and the Pecos River valley, immediately west (Appendix 2). Besides raw material available as in situ outcrops, gravel derived from suitable stone-tool materials also is gaining recognition as an important source (e.g., Holliday and Welty 1981; Wyckoff 1993). But the gravel sources also are available only at or beyond the boundaries of the region. In any event, to procure material, inhabitants of the interior of the Southern High Plains had to travel some distance to the outcrops or engage in trade for the resources.
Paleoindian studies, like other areas of archaeology and most other disciplines, have a confusing array of terms and jargon. This monograph focuses on Paleoindian archaeology, but is not intended to resolve definitional matters. These issues must be dealt with by archaeologists who focus on artifact technology and subsistence (e.g., Hofman 1989a). However, by focusing on the cultural stratigraphy and chronology in a geographically restricted area containing key Paleoindian sites, some order may appear that will help resolve some issues of typology and dating. In order to proceed with the discussion some terminological matters are addressed. Additional typological discussion is presented in Chapter 5.
Roberts (1940) was the first to use the term "Paleo-Indian." It was introduced in the title of his paper ("Development on the Problem of the North American Paleo-Indian") and used a few times in the lengthy article (65 pages), but nowhere was it explicitly defined. The paper reviews all sites in North America of "some antiquity" (Roberts 1940:54), however, summarized as consisting
of artifacts and skeletal materials in deposits dateable by geologic means, in association with bones of extinct species of animals and invertebrates... and also of indications of cultures that were adapted to conditions totally unlike those prevailing in modern times. (Roberts 1940: 54)
In later usage he provided a more explicit though broader definition, using "Paleo-Indian" simply to refer to "older" or "early" Indians (Roberts 1953:256).
The term "Paleoindian" did not come into wide use for over a decade after its introduction. Indeed, Roberts (1951) did not use the term in his review of "Early Man" radiocarbon ages. In the 1950s several definitions appeared, as did alternate terms (Krieger 1964:51-52). For example, Sellards (1952) continued to use the term "Early Man," though considering it the equivalent of Paleoindian (Sellards 1952:10). Sellards, like Roberts (1940), considered Early Man sites to be "of some appreciable antiquity in a geologic sense" (Sellards 1952:5). "Paleoindian" probably gained its biggest foothold in archaeological terminology when used by H. M. Wormington in the fourth edition of her influential book Ancient Man in North America (Wormington 1957), although she did not use the term in her third edition (1949). By 1965 the term was "firmly entrenched" (Wilmsen 1965:182).
Though well established as a term, the concept of "Paleoindian" still is poorly defined, probably because Paleoindian archaeology is so poorly known relative to later occupations. For the purposes of this study the Paleoindian stage, occurring in the very late Pleistocene and very early Holocene, is a time of ameliorated climate when humans coexisted with now-extinct large mammals and prepared a variety of distinctive, lanceolate, unnotched projectile points (modified from Hofman 1989a: 25). "Big-game hunting" often is cited as a characteristic of the Paleoindian stage (e.g., Stephenson 1965; Willey 1966; Wedel 1983), but the native inhabitants of the Great Plains hunted "big game" until the nineteenth century, and there is ample evidence of a more diverse Paleoindian subsistence (Johnson 1986a, 1991; Meltzer 1993; Meltzer and Smith 1986).
The Paleoindian stage is subdivided on the basis of projectile point styles (e.g., Wormington 1957; Hofman 1989a), due to the lack of other obvious archaeological traits. The principal subdivisions used in this monograph are Clovis, Folsom, and Late Paleoindian. The Clovis and Folsom periods are defined by the presence of distinctive fluted point styles that occupy relatively discrete time intervals. Archaeological sites or features that date to one of these two discrete time intervals but lack diagnostic artifacts are referred to as "Clovis age" or "Folsom age." The Late Paleoindian period is defined by the occurrence of a variety of generally lanceolate, unfluted, well-made projectile points. "Late Paleoindian" is equivalent to the "Piano Complex" of Jennings (1955, cited in Krieger 1964). There seems to be considerable temporal overlap among some of the unfluted, lanceolate styles and some temporal overlap with fluted styles (Chapter 5). Moreover, Late Paleoindian projectile point types, typology, and nomenclature present some vexing problems in dealing with this period (Chapter 5). In the absence of more or other kinds of data, however, these gross subdivisions based on artifact styles provide the best first approximation of Paleoindian culture history on the Llano Estacado.
The Paleoindian period was one of dramatic environmental change. These changes were of such magnitude that they are the basis for some definitions of major chronostratigraphic units used in the earth sciences and in archaeology. In general, the Paleoindian period includes the last millennia of the Pleistocene and first few thousand years of the Holocene, although the age of the Pleistocene-Holocene boundary is the subject of some debate (Fairbridge 1983; Farrand 1990). Chronostratigraphic conventions in this monograph are: the Pleistocene-Holocene boundary is 10,000 yrs BP (after Hageman 1972); early Holocene is 10,000-7500 yrs BP.
Some of the data synthesized in this monograph are from the work of other investigators, but the rest is the result of my field research. Most of the archaeological research and some of the geoscientific information are from others, but no other individual investigator studied all of these sites or integrated the archaeological data base with interpretations from stratigraphy, sedimentology, pedology, geomorphology, and geochronology. Geoarchaeological investigations were conducted at all sites on the Llano Estacado discussed in Chapter 3. At Clovis, however, most field work focused on strata older than the Clovis occupation. Geoarchaeological research at the site has been in the hands of C. V Haynes for the past 30 years. Off the Llano Estacado, I conducted no work at Rex Rodgers or Horace Rivers (Chapter 3), but did investigate the Lake Theo and Lipscomb sites (Chapter 4).
Most of the field work focused on the three principal sites of late Quaternary deposition: draws, playas, and dunes. Paleoindian sites are found elsewhere on the Llano Estacado (e.g., Hester 19752; Hester and Grady 1977; Meltzer 1986; Polyak and Williams 1986; Largent et al. 1991), but not in settings that allow stratigraphic or chronometric evaluation. These settings are on uplands, resting on or mixed (via bioturbation) into older Pleistocene deposits such as the Blackwater Draw Formation, or buried by thin layers of sediment lacking stratigraphic distinctiveness. Archaeologists have long recognized the problem of interpreting sites found on the vast, long-stable upland surface of the Llano Estacado (Kelley 1964; Collins 1971; Kibler 1991, 1992). Given the focus of the present study on dating and environmental reconstruction, therefore, most work reported in this monograph emphasizes settings which yielded in situ Paleoindian sites. The sand dunes of the region present special problems of interpretation because they have the potential of yielding archaeology in place, but more often produce early sites in deflated and disturbed contexts (e.g., Pearce 1936; Fritz and Fritz 1940; Polyak and Williams 1986). As a result, the only dune sites discussed are those where Paleoindian materials were found in place or where reasonable stratigraphic inferences could be made.
A variety of field and laboratory techniques were used in this study. The primary stratigraphic correlations and environmental interpretations are based on lithologic and pedologic characteristics of the sediment pre served at each site. The lithologies allow a first approximation of both stratigraphic relationships and depositional environments. The degree and nature of soil development also are useful indicators for dating and correlating some strata, for reconstructing landscape evolution, and for making some paleoenvironmental reconstructions (Holliday 1985b,c,d,e, 1995a,b).
The field data that I gathered came from investigation of natural and artificial exposures, from cores, and from hand-auger samples. All cores and sections were measured and minimum basic descriptive data (i.e., pedologic horizonation and structure, boundary, bedding, and color characteristics) were recorded in the field. Samples from most sections and cores also were brought back to the laboratory for further description and characterization using standard pedologic and geologic nomenclature (Soil Survey Division Staff 1993; AGI 1982; Birkeland 1984; Birkeland et al. 991) with several modifications noted in Holliday (1995b).
Most of the samples were subjected to a variety of analyses for further assessment of sedimentologic and pedologic characteristics, although the field characteristics and descriptions were the most informative kinds of data (e.g., Holliday 1985b,c,d,e). Laboratory analyses included particlesize distribution (sand-silt-clay content), carbonate content, organiccarbon content, bulk density, and clay mineralogy, following methods described in Singer and Janitzky (1986) and Jacobs (1995). Thin-sections were prepared for some samples and were analyzed under a petrographic microscope. The results of the laboratory analyses are available on request from the author.
Significant clues to the Paleoindian paleoenvironment of the Llano Estacado also come from paleontological and paleobotanical studies, which were an important component of many Paleoindian research programs in the region (e.g., Stock and Bode 1936; Patrick 1938; Wendorf and Hester 1975). The study of large vertebrates has been an especially important component of Paleoindian studies owing to the recovery of such remains in most of the Paleoindian sites in the region (e.g., Stock and Bode 1936; Lundelius 1972; Slaughter 1975; Johnson 1986a, 1987c, 1991). Microvertebrate studies are a relatively recent phenomenon in the region, however, pioneered largely by Johnson (1986a, 1987c) at Lubbock Lake. Beetle studies in the region also were pioneered at Lubbock Lake by Elias and Johnson (1988) and attempted at other sites (Elias 1995a), but the recovery was very low. Recovery and analysis of gastropod, bivalve, and diatom remains have been a part of some research programs on the Llano Estacado, including the early work at Clovis (e.g., Patrick 1938), the High Plains Paleoecology Project (Hohn and Hellerman 1961; Wendorf 1961c; Drake 1975; Hohn 1975), the more recent research at Lubbock Lake (Pierce 1987), and the recent regional study of the draws (Neck 1995; Winsborough 1995).
Paleobotanical research also has been an important component of Quaternary research on the Southern High Plains, focusing on pollen. Most of the palynological research was conducted during the High Plains Paleoecology Project (e.g., Hafsten 1961; Oldfield and Schoenwetter 1975) with subsequent work at Lubbock Lake (Stafford 1981; Bryant and Schoenwetter 1987). The results varied significantly and were largely unreproducible (Bryant and Schoenwetter 1987). Additional attempts at pollen recovery from draw samples were made by Hall (1995), and the results largely were negative. The problems with pollen recovery appear to be related to differential pollen production, differential pollen preservation, and extraction techniques (Hall in Holliday, Johnson, Hall, and Bryant 1985; Bryant and Schoenwetter 1987; Holliday 1987a; Hall 1995). In particular, the late Quaternary deposits on the Southern High Plains are conducive to pollen degradation because they are typically alkaline and well drained (and oxidized), but occasionally saturated. These characteristics are among the worst for pollen preservation (Hall 1981; Bryant and Holloway 1983; Bryant and Schoenwetter 1987; Bryant and Hall 1993; Bryant et al. 1994).
In light of the problems associated with pollen preservation and recovery, other approaches to paleobotany were attempted, including examination of seeds from Lubbock Lake (Thompson 1987) and recovery of phytoliths from several other draw localities (Bozarth 1995). The results of these attempts at non-pollen-based paleofloral recovery are encouraging. An especially promising approach to paleovegetation and paleoenvironmental reconstructions is the analysis of stable-carbon isotopes from soil organic matter. There is a strong and positive correlation between δ13C in soils and the proportions of C3 and C4 biomass (Kelly et al. 1993; Nordt et al. 1994). The two groups of plant species are broadly indicative of two distinct environments and therefore can provide clues to past environments. The C4 plants are mainly warm-season grasses and indicative of warm, semiarid environments, whereas the C3 plants include cool-season grasses, most aquatic plants, and all trees, and are linked to cooler, more temperate settings (Kelly et al. 1993; Nordt et al. 1994). Isotopic data are available from paludal and lacustrine mud in a few draw localities (Holliday 1995b), from buried A-horizons in some lunettes, and from lacustrine mud in several playas (Chapter 5). These settings vary in their drainage characteristics and lithologies, so direct comparisons of specific isotopic values from them cannot be made, because each could have inherently different vegetation assemblages. Comparison of isotopic trends, however, should reveal evidence of significant vegetation changes.
An important component of this study is better documentation of the geochronology and especially the cultural chronology of Paleoindian occupations. To achieve these goals, and in addition to previously reported ages, a number of new radiocarbon ages were determined for many of the sites discussed in Chapter 3. Most of the ages were determined by the Southern Methodist University (SMU) radiocarbon laboratory. Others were determined by the University of Arizona radiocarbon laboratories (A, the conventional lab, and AA, the AMS lab). A few ages were determined by the Smithsonian Institution (SI) and by Beta Analytic, Inc. (Beta), the latter for Brown (1993)
All of the recently determined or earlier unpublished ages are from samples of organic-rich sediments and soil horizons. Both humate (NaOH-soluble) and residue (NaOH-insoluble) fractions were extracted and dated. Preferred materials for radiocarbon dating such as charcoal or wood are very rare on the Llano Estacado. Shell and bone are locally more abundant but were avoided due to the uncertainties in interpreting ages from these materials (Taylor 1987). There are problems in dating organic-rich sediments and soil horizons (e.g., Campbell et al. 1967; Scharpenseel 1971, 1979; Matthews 1985), but with proper care in sampling, laboratory processing, and interpretation they can provide reliable age control, particularly in nonleaching environments. Organic-rich sediment occurs as silty and clayey, homogeneous deposits that accumulated in slowly aggrading, probably marshy settings and as clayey lenses interbedded with sand or diatomite. Dating these types of samples at Lubbock Lake, where archaeological data and ages on wood and charcoal are also available, shows that they usually yield reasonable, approximate ages of deposition, but sometimes provide only minimum ages (Holliday, Johnson, Haas, and Stuckenrath 1983, 1985; Haas et al. 1986; see also Holliday et al. 1994). In situations where ages appear reversed or where two ages are available from a single sample (e.g., for residue and humate fractions), the older age is assumed to be a better indication of the age of the sediment (Matthews 198o; Hammond et al. 1991; Martin and Johnson 1995). Contamination with dead carbon from groundwater, precipitated in calcium carbonate, is the only known, common means of yielding falsely old ages in the region. Calcium carbonate is removed during processing of samples, however.
Organic matter in the A-horizons of buried soils was incorporated into the surface of the soil parent material during a period of landscape stability and pedogenesis. A radiocarbon age from such "homogenized" horizons is the "mean residence time" of organic material from this zone, plus the time since burial by overlying sediments (Scharpenseel 1971). Data from Lubbock Lake show that the maximum age of burial (and therefore the approximate age of the overlying sediments) can be reasonably estimated by collecting samples from the top of these zones (Haas et al. 1986). More younger carbon appears to be preserved at the top of the buried Ahorizons. Older carbon probably was oxidized or mixed via bioturbation deeper in the profile. Radiocarbon ages from the lower portions of a buried A-horizon provide an intermediate age between the beginning and end of pedogenesis (Haas et al. 1986), indicating that some of the older carbon (and probably most of the oldest) were oxidized.
All of the ages determined on samples that I collected are based on a radiocarbon half-life of 5568 years and are corrected for δ13C fractionation. Tree-ring calibration of radiocarbon ages can have a significant effect on the dating of Paleoindian occupations on the Great Plains (Eighmy and LaBelle 1996), but the radiocarbon ages presented in this volume are not calibrated for several reasons. (1) Calibrations would confuse comparisons with other dated sequences from the region and surrounding areas, none of which is calibrated. (2) Calibrations often require correction (especially the more recently published calibrations), rendering published calibrated ages inaccurate (e.g., Stuiver and Pearson 1992; Stuiver 1993) (3) Many of the radiocarbon ages from Paleoindian sites are in the range of only tentative calibrations (>10,000 yrs BP) (Becker 1993; Stuiver 1993). Calibration of only part of the sequence is not a useful exercise. Eighmy and LaBelle (1996) show that calibrated ages for Plains Paleoindian complexes are ca. 2000 years older than uncalibrated ages, but that the chronological sequence and duration of individual artifact styles are not significantly changed.