The Harrell Site

From the present vantage point we can see that the archeologists of the 1940s misunderstood several key aspects of the Harrell site. The critical problem is that they did not have any way to accurately gauge the age of the deposits. Radiocarbon dating was not developed as a research tool until the 1950s. And only in the last few decades have modern geological methods been applied routinely in archeological investigations. Without these two critical research tools, it is hardly surprising that the Harrell site was seen by archeologist Alex Krieger as representing a single culture which he called the Henrietta Focus.

Alex Krieger was trying to make sense of extremely broad cultural patterns and link the chronology or time line of the poorly dated moundbuilding cultures of the southeastern United States with that of the tree-ring-dated Puebloan cultures of the southwest. He used the Harrell site and the Henrietta focus to do just that by arguing that the Plains Villagers of north Texas were contemporaneous with late Puebloan peoples and late Caddoan groups. In a general sense, this view has been borne out by a great deal of research over the last 65 years. But modern researchers have also learned many things that enable us to reevaluate the dating of the Harrell site.

Krieger argued that the Henrietta Focus dated to A.D. 1450-1650 based on the relative dates of several Southwestern pottery sherds and at least one Caddo sherd that were found at the Harrell site. From the fact that the site's distinctive shell-tempered pottery (which he named Nocona Plain) was found in both the midden and in the underlying red clay, Krieger concluded that the entire 8- to 10-foot thick artifact-bearing deposit in Excavation 3 could have formed in as little as two hundred years, mainly as the result of the periodic addition of flood deposits. He, like Jack Hughes, recognized that certain of the dart points were more characteristic of central Texas, where they were regarded as comparatively early artifact styles that dated to the Archaic period before the introduction of the bow and arrow. But Krieger assumed that the presence of these styles at Harrell along with arrow points and pottery in the same deposit meant that these dart point styles continued to be used by latter peoples. In contrast, Hughes correctly reasoned that earlier central-Texas-related peoples had lived at the Harrell site before the later Plains-related peoples who used the bow and arrow and made pottery. Hughes' views were stated cautiously in his unpublished Master's thesis and were soon overshadowed by Krieger's grand synthesis.

Today we have a much better understanding of chronology. The earliest distinctive artifacts found at the Harrell site are the beveled-stem Nolan dart points now known to date to the Middle Archaic period about 4500-5000 years ago (2500-3000 B.C.). Also found among the Harrell site's artifacts are examples of Pedernales, Bulverde, Castroville, Marcos, Ensor, Fairland and Darl dart points (among others), indicating that the site was occupied intermittently during much of the Late Archaic, roughly 1200 to 4500 years ago (2500 B.C. to A.D. 800).

Nonetheless, most of the artifacts at the Harrell site do date to the Late Prehistoric era, between about A.D. 800 and A.D. 1500. According to Hughes, some 555 arrow points were recovered from the site compared to only 81 dart points. All of the almost 600 pottery sherds also date to the Late Prehistoric, probably after A.D. 1200. In other words, the site's major occupation period was after A.D. 800.

From the present perspective, we guess that the third or high terrace where most of the Harrell site excavations were carried out probably existed as a more or less stable landform by about 5000 years ago (roughly 3000 B.C.). We further surmise that during the subsequent 4500 years (from about 3000 B.C. to A.D. 1500) a maximum of five feet of sediment accumulated on the terrace top. As Hughes realized, by far the bulk of all the artifacts occurred in the upper five feet in the dark midden soil as well as all 135 hearths and all 32 graves. The only things found deeper were relatively small quantities of scattered artifacts, some of which were found within the apparent pits that intruded into the underlying red clay. Whereas Krieger assumed that river floods continued to add sediment to the upper terrace while it was occupied, we doubt this was the case to any significant degree during the Late Prehistoric era. The midden layer is clearly what archeologists today call an anthrosol (literally "man-soil"). That is to say that the refuse, artifacts, hearth rocks and so on that people hauled to and left on the Harrell site were mixed in with the existing soil resulting in a distinct cultural layer.

We suspect that the Middle and Late Archaic materials were buried fairly quickly and may have been neatly stratified in a "layer-cake" fashion. However, after A.D. 800 or so, the site was used more intensively by a series of Late Prehistoric peoples, including the Plains Villagers who were responsible for the bulk of the debris and midden accumulation. There is obvious evidence that the Late Prehistoric occupants were habitually digging holes—pits for hearths, graves, and probably for houses and storage. These actions intruded into and greatly disturbed the earlier (Archaic) layers.

While no obvious evidence of houses or structures was recognized, there were several tantalizing clues. Excavators noticed two odd features in the southeast section of the excavation block, in the area of Burial 31. One was an area of fire-burned red clay with impressions of grass and twig marks. Fox noted that it "appears a fireplace but looks like wattle work." In the unit just north, he recorded an oval of cemented white ashes with two cut depressions in the interior. Whether these features pertained to a house (with wattle and twigs) with an interior hearth is unknown; no further mention of the features was made in subsequent analyses. At the north end of the cemetery, Fox also recorded a number of possible post-hole stains in the area of several graves holding very incomplete skeletal remains. The stains were in groups or clusters, and several appeared to align in rough arcs. These may represent the supports for some sort of arbor-like mortuary structures similar to those recognized at Zimms Complex sites in northwestern Oklahoma.

Nonetheless, houses and other types of structures were almost certainly built at the Harrell site and probably placed in shallow pits like those that have been documented at many other Plains Villager sites. All this digging and earth moving coupled with the continuous burrowing of rodents (of which much evidence was seen) resulted in the site deposits becoming badly mixed through time. While the overall trends are clear—most of the dart points were more deeply buried than most of the potsherds—the mixing makes it impossible to neatly sort out the site into separate time periods. The only things at the Harrell site that we can confidently date are those that have been dated elsewhere.

Age of the Cemetery

Before attempting a trial reconstruction of the human history of the Harrell site, we must raise a thorny dating problem that cannot be satisfactorily resolved without further research. That is the age of the cemetery. Krieger and Hughes both assumed that the graves dated to the latest and main period of site occupation. In broad terms this is undoubtedly true: the graves are Late Prehistoric in age. But today we recognize that the Late Prehistoric era spans at least 700-800 years and probably 1,000 or more. The bow and arrow may have been introduced into the southern and central Plains before the time of Christ followed by the introduction of cord-marked pottery during the first centuries A.D. These technologies are the hallmarks of the Late Prehistoric era.

The early part of the Late Prehistoric (sometimes called Late Prehistoric I), from perhaps A.D. 500 or so to A.D. 1100-1300 is seen by most experts as a time of change during which the hunting and gathering traditions that characterize the long archaic era were gradually altered by the introduction of a new weapon system (the bow and arrow), new storage and cooking technology (pottery), and, even more importantly, agriculture and semi-sedentary life. Some areas, such as most of the Edwards Plateau in central Texas, continued to be occupied by purely hunting and gathering cultures at the same time that more-settled village life was becoming prevalent elsewhere.

By the later part of the Late Prehistoric (sometimes called Late Prehistoric II), after A.D. 1200-1250, settled villagers lived along virtually all of the river valleys of the southern Plains. Most sites show evidence of corn farming along with heavy reliance on buffalo hunting, a mixed economy that sustained village life. This broad cultural tradition is known as the Plains Villager horizon and it has been subdivided into dozens of named phases, complexes, and foci reflecting different local variations and changes through time over the A.D. 1200-1500 span of the era. Most, but certainly not all, of what Krieger included within his Henrietta focus dates to this period.

Stepping back in time, we see the transition between Late Prehistoric I and Late Prehistoric II (the Plains Villager era) was a time of upheaval and extreme violence. Skeletons riddled with arrow points, mass graves, trophy skulls, and mutilated skeletons have now been documented at dozens of sites ranging from the Texas Panhandle across north Texas and western Oklahoma to central and southern Oklahoma and southward well into central Texas. Broadly, most of these sites seem to date to between A.D. 1000-1300 within the latter part of what is generally considered to be Late Prehistoric I. One of the most common arrow point types implicated in this period of violence is the Scallorn type, which is also the only arrow point type found with the graves at the Harrell site.

The graves at the Harrell site also share several other traits in common with Late Prehistoric I sites. These include mass graves, cyst graves, and the removal of the lower jaws and other forms of skeletal mutilation perhaps including the heretofore unrecognized evidence of scalping. Because of these similarities, we suspect that the cemetery at the Harrell site dates prior to A.D. 1300 and prior to most of the Plains Villager era including most of what Krieger included within the Henrietta focus. It is possible, however, that the cemetery dates to the beginning of the Plains Villager occupation. If so, it may not be a coincidence that the Scallorn point used by the aggressors is associated with older traditions that lingered to the south in central Texas, while the villagers represent a new tradition derived from the north.

The Prehistory of the Harrell Site

With all of the above in mind, here is our reinterpretation of the prehistoric record at the Harrell site, concentrating on the Late Prehistory. First and foremost we must emphasize that without further research our educated guesses are just that. Some of our inferences could be tested by further analysis of the Harrell site collections. But what we really need are more well-controlled excavations of prehistoric sites in north Texas. Today we would use modern geological methods to help choose sites that have better depositional contexts than that of the Harrell site—less mixing and fewer repeated occupations. Today we would collect charcoal samples and soil samples from individual hearths. Today we would carefully study the animal bones and the charred plant remains. And so on.

Based on today's knowledge, we infer that by 5,000 years ago (3,000 B.C.) hunters and gatherers were regularly visiting the confluence of the Salt and Clear Forks of the Brazos. In all likelihood, the area was a favored stopping place much earlier, probably beginning in Paleoindian times, but traces of such earlier occupations that may have existed at the Harrell site were probably destroyed by the meandering Brazos river. By 5,000 years ago the high terrace upon which the bulk of the prehistoric record at the Harrell site accumulated had stabilized and was above the ordinary flood level. The near edge of the river itself was probably further north than it is today based on the fact that the river is continuing to swing southward.

During the Middle and Late Archaic or about 5,000 to 1,200 years ago (3000 B.C. to A.D. 800) hunting and gathering groups who shared much in common with people living further south in central Texas repeatedly stopped at the river junction and stayed for short stays. They probably did many different things at the site and in the area, but their habit of baking roots and bulbs in earth ovens left the most obvious archeological signatures—countless hearths and, in one area near the river, a burned rock midden. The so-called "great midden" is remarkable because it was covered by flood deposits before it was used for very long, making it easy to see the roasting pits that generated the spent fire-cracked rocks that make up the midden. The age of the midden is unknown, but it is probably associated with the Archaic period at the site. The general absence of artifacts suggests that the "great midden" was a special purpose cooking area that may have been created on the lower river terrace to keep the area and the activity separate from the main camp.

The main area of human habitation at the Harrell site was atop the high terrace, even during Archaic times. We would guess that most of the deeper hearths found there were Archaic in age, based on the distribution of the dart points. Among those hearths are a few that are comparable in size and shape to the roasting pits found in the burned rock midden as well as many smaller ones of various forms that must have been used to cook different kinds of foods. There is little doubt that Archaic life at the Harrell site was much more interesting and involved than the simple inferences we draw here, but further speculation serves little purpose.

By the early part of the Late Prehistoric, perhaps around A.D. 800, the Harrell site had become a major habitation site, a place where people regularly visited and stayed for weeks or even longer. Over the next 700 years, a tremendous quantity of cultural debris accumulated on the high terrace overlooking the river confluence. Based on the large numbers of basally and small side-notched triangular arrow points and on the preponderance of shell-tempered pottery, we can infer that most of the debris accumulated after A.D 1200. That said, the cemetery could date somewhat earlier than this; if so, the presence of a well-defined cemetery suggests that people were already living there for extended periods of time. In other words, village life at the Harrell site may have begun prior to the Plains Villager period.

Some aspects of daily life were little different from the Archaic era. Numerous hearths were found in the upper few feet of the refuse midden including likely baking pits and smaller hearths as well, several of which resemble hearths found in the interior of houses at other Plains Villager sites. But other things changed. The Plains Villagers at Harrell were apparently farmers, but it is hard to gauge the relative importance of agriculture. Although there were scapula hoes and other possible farming implements, only a few corn cobs were found. Judging from the relatively large numbers of bison bones as well as the many beveled knives and hide scrapers, buffalo hunting was very important during this period, perhaps because more bison were present in the area.

Despite our misgivings about certain of Krieger's assumptions, he was right to emphasize the Plains Villager period at the Harrell site. The distinctive assemblage of shell-tempered pottery, basally and side-notched triangular arrow points (Harrell and Washita points), bison-scapula hoes, handheld mussel tools, diamond beveled knives, hide scrapers, and small drills has been found at other sites in north Texas. While the Henrietta focus is still poorly understood, there is little doubt that these materials are closely related to other Plains Villager assemblages. In particular, there are many shared similarities among Washita River phase sites along the Washita and Canadian River valleys in south-central Oklahoma. These are comparatively well dated to broadly A.D. 1200-1500 and perhaps mainly A.D. 1300-1450, according to recent syntheses by Richard Drass.

The presence of extensive Plains Villager materials at the Harrell site can probably be best explained as the result of the movement of new people into the area, probably from western or south-central Oklahoma. We cannot rule out the possibility that local peoples simply adopted the trappings of more-sophisticated neighboring groups, but we think it more likely that the Plains Villagers who settled at the Harrell site shared many customs with other Plains peoples because they were themselves Plains Indians, people who moved into the area from the north.

The ethnic identity of the Plains Villagers who lived at the Harrell site is unknown. In the archeological literature, the possibility that the people who lived at the site might be ancestors of the historic Wichita is often mentioned. While this idea was discounted by Krieger, Jack Hughes argued in his 1968 dissertation, Prehistory of Caddoan-Speaking Tribes, that the Harrell site represents an early Kichai (Kitsai) village. The Kichai spoke a minor Caddo language, but were one of at least six named groups who became consolidated during the historic era into the Wichita tribe (the others were the Taovaya, Tawakoni, Iscani, Wichita proper, and Waco). Given the shared similarities among these groups, the great difficulty in tracing the movement of peoples even in historic times, and the radical changes the Wichita groups experienced as the result of the introduction of European guns, horses, and diseases, we will probably never know for sure.

While we may never know the precise ethnic identity of the Plains Villagers at the Harrell site, we can guess that they spoke a Caddoan language (as did many groups living on and near the southern Plains including the various Caddo and Wichita groups). Based on linguistic evidence we know that Caddoan languages were spoken by many groups over a very large area stretching from east Texas and southeastern Arkansas across northern Texas and much of Oklahoma and north into Kansas and the central Plains. Broadly these people all shared a common cultural heritage, in much the same way that most Europeans do. Just as in Europe, the peoples who lived in the southern and central Plains moved repeatedly over the centuries, split apart, fought with one another, intermarried, and otherwise transformed their cultures and their identities in complex ways that defy simple classification. In other words, the PlainsVillagers who lived at the Harrell site were part of a prehistoric era that dates hundreds of years before the earliest historic accounts of named ethnic groups. We cannot give them a familiar name, which is one of the reasons why Alex Krieger coined the term Henrietta focus.

There is little or no indication that the Harrell site continued to be occupied after A.D. 1500. A single blue glass trade bead was found, but it is of a late style dating no earlier than the Kickapoo village (1850s). Other late historic Indian remains may well have been mixed in with the burned remains of the Harrell family's farmhouse. The WPA archeologists did not recognize any obvious Indian artifacts and discarded almost all of the historic debris they found.

We close by repeating Jack Hughes' shrewd observation: "Reconstructing the history of the part of the Brazos on which the Harrell is located is like trying to visualize a motion picture on the basis of a few still photographs." Archeologists may never know enough to replay the movie version of the long and varied human history of the upper Brazos. But we can and will learn more as research moves forward into the 21st century. The artifacts and records from Harrell site collections from the 1937-1938 WPA excavations remain a critical and largely untapped source of information awaiting further study.