The Gilbert Site

An Eighteenth-century French Connection in East Texas

In 1962, a handful of metal and pottery artifacts from a Rains County site in East Texas was sent to the University of Texas for evaluation. At the time no one could have anticipated these few objects would be the first faint clues toward the identification of a significant commercial enterprise with international connections, based upon this site in the East Texas woods. Yet that is exactly what subsequent archeological investigations have shown. Some 250 years ago, the Gilbert site was the focus of a successful market-oriented endeavor linking East Texas with Europe.

What was the market-driven and apparently successful product supplied by the people who occupied the Gilbert site? When did this take place, who were they, and how were the answers to these and other such questions sought and derived from the archeological evidence?

Before attempting to answer these questions, let's take a look at the site. On the surface, the Gilbert site wasn't much to look at. Part of the site was a pasture and part was covered in woods, mainly post oaks and blackjack oaks. In places, one could make out low rises, about 30 feet across, which we euphemistically called "mounds," although "low heaps" would have been more appropriate. There were 20 of these scattered along the northern end of a sandy ridge overlooking Lake Fork Creek and one of its tributaries. The tallest "mound" towered about 18 inches above the surrounding terrain—most were even less conspicuous. We soon learned that these low heaps were actually middens—trash deposits that sometimes had firepits within them and almost always had lots of animal bone, most of it deer. Oddly, a few of the middens were capped with a layer of sterile (artifact-free) red clay—obviously an intentional addition, but for what purpose?

In the 1960s, it was assumed that each of the middens probably represented the location of a house and that the 20 house-middens represented a village that had been occupied during the mid-eighteenth century. The preponderance of French trade goods mixed with items of native manufacture including stone arrowheads, pottery, and bone tools was similar to that found at several other sites in north Texas including another Rains County locality, the Pearson site. These sites, including Gilbert, were considered part of the Norteño focus, believed to be linked to eighteenth- or early nineteenth-century occupations by Southern Wichita, Caddo peoples who apparently moved south into what is today Texas and Oklahoma from Kansas. Spanish authorities used the collective term Norteños, nations, or peoples of the north, in the eighteenth century. It referred to the Wichita and various other groups known to live within provincial Texas to the north of San Antonio beyond the effective control of the Spanish. The Norteño focus concept is no longer used much today; we have come to recognize that the archeological and ethnohistoric record is much more complex than was envisioned in the early 1960s.

Trade Goods and Native Artifacts

The Gilbert site has yielded what has proved thus far to be the most remarkably large and varied inventory of "trade" goods documented from an eighteenth-century Native American site in Texas. The extensive list of such European-supplied trade goods identified from the Gilbert site includes firearms, gunflints, lead bullets and shot (in ball form), axes and hatchets. There are also knives, brass kettles, brass and iron projectile points, hawk bells, glass beads, buttons, a finger ring, hoes (notably appearing unused), horse gear, and different sizes of cuprous (containing copper) wire bracelet stock. Many of these things had been discarded in apparently unused or still useable condition.

One of the major artifact classes at Gilbert was firearms. All were flintlock smoothbores which proved to be of a lighter, hunting-type of gun as compared to the larger and heavier military "muskets" of the era. Detailed comparative study showed a major French influence in the design and decoration of these particular firearms. None were intact and it could be estimated from parts that at least 20 different long guns were discarded and mostly scattered about the site. Since so many gun parts had been cut or broken into pieces, as were other metal artifacts, it seems apparent that the Native Americans here were experimenting with the use of metal for a variety of purposes. The large quantities of diverse metal items appear to have been treated as a novel raw material and used in a variety of ways.

Despite the obvious presence of firearms in numbers, no clear bullet damage was found on deer skeletal remains. One small triangular flint arrow point of the Fresno type was found in direct association with deer bones. A large number of the same type arrow point was found throughout the site. These, and the native-made gunflints, clearly show that a high level of skillful flintworking was maintained during this period despite the acculturation process indicated by so many European goods. While guns may have been the new weapons of choice, the bow and arrow was still in use.

Remnants of French-type brass kettles showed extensive wear. Some of the discarded kettles had been broken and cut up to create tinkler cones (ornaments attached to clothing). Examples of more than one variety of native-made pottery, including broken and sometimes repaired pieces, also were found. The presence of metal and ceramic containers again suggests that, during the mid-eighteenth century, the Native Americans at the Gilbert site were adopting foreign materials when convenient and desirable, but still maintaining traditional crafts.

Native pottery was extremely common at the site. Over 2000 pottery sherds representing at least 47 individual vessels were recovered from the 1962 work alone. Among the many pottery types present in this collection are shell- and bone-tempered wares characteristic of the eighteenth century and a small number of grog-tempered (grog is ground-up pottery) vessels. The latter is probably Caddo pottery dating to several centuries earlier that represents an earlier use of the area. The grog-tempered sherds were found in only one small area of the site, the shell- and bone-tempered pottery was found in every excavation. In fact, one of the reasons that only a single group is thought to be responsible for the Gilbert site is because of the uniform distribution of ceramics and other types of artifacts across the site.

Most of the shell- and bone-tempered pottery from Gilbert most closely resembled pottery from several sites along the Red River that were originally attributed to the Norteño focus as well. Several other vessels were recognized as historic Caddo pottery and were assumed to be trade pieces. Since the 1960s, many more late Caddo sites have been studied, and it is now recognized that bone- and shell-tempered pottery is relatively common at historic Caddo sites. In other words, the ceramics do not help pin down the identity of the group that lived at Gilbert.

What was the source of the European trade goods?

The European artifacts at the Gilbert site strongly point to French sources. While some degree of affiliation with French trade activity was expected because of the site's geographical location and anticipated general age, the 1962 excavation turned up evidence for a major commercial connection between the Gilbert site and France. In fact, the initial Gilbert site report published in 1967 was the first detailed archeological documentation of eighteenth-century French trade in the south-central United States.

The apparent surplus of European goods hints that French traders themselves may have been present at the site, together with their merchandise. Alternatively, the trade goods could have been brought to the site by those transporting the hides to French traders living along the Red River. The hides would have been moved by horseback—horse bridle bits and the bones of at least one horse were found at the site. The use of horses also helps explain how such a quantity of heavy metal items—guns, hatchets, and kettles—arrived at the site. While the details of the trading are not known, it is clear that the deer-hide trade created such a wealth of foreign items that those who occupied the Gilbert site could afford to be wasteful.

This circumstance bears a striking resemblance to the wealth of French trade goods present at the Trudeau site near the confluence of the Red River and the Mississippi in south-central Louisiana. Trudeau was the principal village of the Tunica between 1731-1764 and main source of the so-called "Tunica Treasure," a vast amount of high quality European trade goods that were surreptitiously removed from Tunica graves. The Tunica were middlemen in the French-Indian trade network and were especially prominent because they controlled the trade in horses. It is possible that the Tunica at Trudeau were involved, at least indirectly, in the deer-hide trade that figured so prominently at the Gilbert site. Anyone seriously interested in the topic should read The Tunica Treasure, a marvelous 1979 study by archeologist Jeffery Brain that tells the intriguing story of the Tunica and the treasure.

The deer bones represent every part of the skeleton, indicating the animals were butchered at the site. It is extremely telling that the least-common bones were antlers and caudal vertebrae (tail bones). These bones were probably removed along with the hides. This, together with the finding of an unusually large number of end scrapers, the particular kind of stone tool that is most suitable for hide dressing, suggests that the site served as a major deer-hide production locale. To give you some idea of the magnitude, there were 418 scrapers from the 1962 work alone, suggesting that the entire site must have contained thousands of scrapers.

Most of the deer bones were those of relatively young adults ranging in age from1.5 to 3.5 years old—the prime age for hides and meat. Among the bones at the site were those of a variety of other species including rabbits, raccoons, squirrels, rodents, fish, turkey, and numerous turtles, especially box turtles. Interestingly, box turtles overwinter in burrows, a fact which suggests that they must have been collected at other times of the year (see discussion below on time of year the camp was occupied). Other animals of note include horse (one individual), puma, bobcat, and dog/coyote. Bones of most species (besides deer) belonged to only a single individual. The two most numerous mammals second to deer were bison (or cow), represented by four individuals, and bear, of which there were at least five individuals. Both bison and bear have large, luxurious hides that also could have been traded.

Numerous arrow points and gunflints evidence a ready arsenal of hunting weapons. The metal knives found at the site indicate the full range of deer hunting, butchering, and hide-preparation that took place there. While it is assumed that the large quantity of deer meat was also an important by-product, historical records suggest that hides were the major product exported from Texas to European markets during the middle-eighteenth century. The scope of this enterprise is indicated by the incredible number and variety of trade goods found at the Gilbert site, items which are assumed to have been received in payment for the hides.

Village or Warm-Season Deer Hunting Camp?

Several lines of evidence suggest that the Gilbert site does not represent a village occupied year-round. For instance, only a single mano was found and no metates or grinding slabs. Such artifacts are normally common at the villages of peoples who depended on corn as a staple. There was also no clear evidence of substantial houses, although temporary shelters must have been present. Not a single posthole was found despite purposeful searching, including machine scraping and shovel-skimming; such techniques are often used in East Texas to detect houses. Low midden/mounds at other somewhat similar sites in the region have been interpreted as abandoned pithouse depressions that had been subsequently filled with trash. But with one possible and tenuous exception, no subterranean house pits were found at Gilbert.

If substantial mud-plastered wattle and daub houses with thatched roofs (known to be characteristic of Caddo and Wichita villages) had been present, it is expected that numerous daub fragments would have been found. While some daub was found, none of it bears the tell-tale impressions of the woven-limb framework (wattle) upon which the mud (daub) was applied, such as that found at other sites in the region. Additional evidence arguing against a winter-season occupation and a village-sized population level is the presence of only a modest number of hearth/fire pits and a single bell-shaped storage pit. Such features are common at village sites. Also absent are human burials, another common occurrence in village sites.

And what of the peculiar red-clay caps that topped several of the middens? These were clearly intentional and labor-intensive additions that happened after the middens beneath had accumulated. In other words, these were added late in the use of the site. We considered various explanations. While the idea that this cap was added as a floor for a structure was discounted because there were no signs of postholes, it could be that the structures were temporary grass huts that didn't need to be securely anchored. Another idea is that the clay may have been added to seal in the stench of rotting deer remains. This is, perhaps, a stretch considering the amount of effort that must have been involved. Digging through the red clay at the site was hard-enough work, even with a sharp metal shovel and a wheel barrow at hand to move the dirt. And it wouldn't have taken more than a thin layer to seal what lay below. Perhaps you have a better idea?

The sum of evidence strongly suggests that the Gilbert site resulted from a series of warm-season encampments by a single group of Native Americans for the main purpose of harvesting deer hides for export. Although not a permanent village, the Gilbert site may have been a temporary home for weeks, perhaps even months at a stretch. The quantity and variety of materials, presence of 20 substantial middens, and presence of artifacts linked with both male hunters and female hideworkers, strongly suggests that the group was sizeable and included men, women, and their children. It was an eighteenth-century deer hunting camp with a French connection. It is possible, perhaps even likely, that French traders did visit the Gilbert site and maybe even lived there at times as they did at many Caddo villages of the period.

Who Occupied the Gilbert Site?

Certainly one of the most common questions asked of us by the general public is, "What tribe did these people belong to?" This is, of course, particularly to be expected concerning sites of the historic period. At least equally common is our inability to supply a specific and certain answer. The Gilbert site illustrates why answering this question is often so troublesome.

As explained, the artifact evidence from the initial 1962 work was considered typical for a component of the Norteño focus, a focus based upon similarity of the archeological remains from a group of mostly eighteenth-century Native American sites in the Southern Plains. During the eighteenth century, much of north Texas and adjacent areas of Oklahoma to the west of the Caddo area proper were occupied by affiliated southern Wichita groups including the Tawakoni and Yscani. Wichita is one of the four shared tongues of Caddo peoples (the others are Caddo-proper, Pawnee and Arikara). A closely affiliated group were the Kichai who were a somewhat more independent and mobile folk who spoke a minor, separate Caddo tongue. In the 1967 report, Ed Jelks proposed that the tribal identity of the people at this site was probably "Tawakoni, Kichai, or Yscani."

I continue to believe the Kichai are the most satisfactory candidate for the tribal identity at this site. As described by the ethnologist John Swanton: "The Kichai seem to have been in the habit of attaching themselves now to one tribe and now to another, sometimes with a Caddo group but more often with the Wichita and their confederates." It is just such a group that we might expect to find serving as middlemen in the hide-gun-horse trade and possessing a mix of artifacts found in both Caddo and Wichita sites. It has been suggested, however, that an as-yet-unidentified Caddo group may have occupied the Gilbert site. During historic times there were 25 different named groups who spoke Caddo. Given the preponderance of Caddo and Caddo-like ceramics at Gilbert and the known presence of French traders at every Caddo village in the eighteenth century, this suggestion cannot be dismissed.

The fact is that Caddo- and Wichita-speakers shared many elements of culture beyond language. Most of the distinctive things that set them apart in the eyes of the French and Spanish—language, tattoos, hairstyles, clothing, and customs—are archeologically invisible. Nonetheless, as more eighteenth-century Native American sites and early historic accounts are studied and compared with one another, we may one day be able to offer a more definitive answer.