NATCHEZ TRACE PARKWAY Excavations at the Pharr Mounds and the Bear Creek Site

The Pharr Site consists of eight large dome-shaped mounds and an area of domestic habitation. Four mounds were investigated: Mounds A and H were partially excavated, and Mounds D and E were dug almost completely. The occupation area was briefly tested.

The mounds were erected to mantle interments and/or supposed crematory basins, both of which had been placed on a (usually) prepared facility. This preparation took the form of a low platform at Mound E, and perhaps at Mound A, and the replacement of the old humus with a thin layer of red and yellow clay at Mound D. The exception to this practice occurred at Mound H, where inhumations and a crematory basin were placed directly on the old ground surface.

The covering of crematory basins was more than accidental. It may have been as important an objective as the covering of burials. Mound D, where a probable crematory facility but no cremations were uncovered, suggests this.

Secondary additions to the mounds were made to cover additional interments placed at the side of the primary mound, or in shafts which penetrated the primary mound. Mound D, the addition was made very soon after completion of the primary mound, whereas at Mounds A and E a pause ensued of sufficient length to permit weathering of the earlier structure.

Both inhumation and cremation were practiced at Pharr. Judging from the number of empty graves, which we assume held inhumations, and the remnants of bone at Mounds A, D, and E, the former would appear to have been more common.

The lack of skeletal remains made it impossible to determine the number of burials present and the function of the mounds, whether they were status repositories or the tribal cemetery. Nor was it possible to tell whether the burials were deposited simultaneously, as might have been the case had they been removed from a charnel house, or separately. It might even be argued that we cannot be really certain that inhumations were not included in the mound fill while they were being constructed, as were two cremations at Mound A. The total absence of artifacts and the failure to locate a single scrap of bone in the fill argues against this possibility. However the mounds were used, their construction, or that of their components, seems to have been a signal event.

Since the number of burials was undetermined, the occurrence of grave goods can only be vaguely characterized as being fairly common. At Mound H, the greenstone platform pipe was the only find. This sparseness was repeated at Mounds A and D where, respectively, a copper-covered wooden artifact and the spools were the only finds. Mound E was more fruitful; several graves contained vessels, one grave produced a cache of flint blades, and the burial platform yielded a variety of artifacts.

The mortuary tradition manifested in the mounds and burials clearly required considerable energy to fulfill. The sheer size of the earthworks (or their components) and the fact that they were built as a result of a continuous effort are perhaps the most obvious testimony of this. Further evidence to the importance attached to burial is presented the inclusion of what must have been highly treasured possessions, some of which--pottery copper spools, and the conjoined tubes--were traded over great distances.

The excavations and surface collecting in the area of domestic habitation showed that the site had been intermittently occupied from Late Archaic through Middle Mississippian times. The most intensive occupation was apparently that of the mound-building people, who, we have hypothesized, lived here intermittently at times of mortuary rites and seasonal exploitation of nearby food resources.

The pottery recovered in the excavations was studied to determine the site's relationship to the Miller ceramic sequence. The initial step was to determine whether or not the sherd assemblages recovered from the mounds reflected the true age of the structures. Examination of the sherd counts showed that the percentage distribution of types was relatively consistent in all excavation units in the mounds and habitation area, and this strongly suggested that the occupation of the area from which fill was obtained did not significantly predate the construction of the mounds.

The ceramic assemblage, characterized by high percentages of Saltillo Fabric Impressed and Baldwin Plain and low percentages of Furrs Cordmarked did not fit into the Miller sequence as it had been described. A review of this archeological unit and a key site, Bynum, was necessary to resolve the discrepancies.

The Miller sequence as depicted by previous writers consisted of time periods characterized by diagnostic pottery types. This scheme was much oversimplified, proposing as it did that pottery types died out and were abruptly replaced by other wares.

To obtain a clearer understanding of the Miller sequence, we seriated surface collections from appropriate sites in northeast Mississippi. This seriation made it possible to define the time periods on the basis of the percentage distribution of types rather than on their mere presence or absence.

Miller I was poorly represented in the seriation but could be loosely characterized by large amounts of Saltillo Fabric Impressed. It was also shown that small percentages of Furrs Cordmarked, while not actually present in the collections seriated, were not incompatbile. On this admittedly rather shaky basis, Pharr was assigned to a late Miller I age.

The Miller I ceramic assemblage was shown to have a different composition than as previously described. Not only was it pointed out that Furrs Cordmarked could have appeared in small quantities late in the period, but also that fiber tempered pottery is absent from the assemblage. Pharr demonstrates plainly that, whenever the period began, late Miller I is on a Middle Woodland level, much too late for fiber tempering.

A review of the Bynum data in light of the revised definition of Miller I suggests that this site, too, fits more comfortably here than in the succeeding period, as previously suggested. The presence of Furrs Cordmarked and the absence of fiber tempered pottery at Bynum, which appears to have been instrumental in the Miller II assignment, have been removed as objectives. Further, a study of the percentage distribution of types among Bynum's excavation units suggested that the large amounts of Saltillo Fabric Impressed recovered from the mounds is truly characteristic of their age and not merely chance inclusions from an earlier habitation area.

Inspection of the "tradeware" pottery from Pharr indicated contacts with the Lower Mississippi Alluvial Valley and northern Alabama. Contacts with the Marksville culture included both the importation of vessels and the borrowing of decorative techniques and vessel shape. The significance of the limestone tempered pottery, whether it was traded or was a minor native ware at the site, could not be determined. It was, however, demonstrably contemporaneous with the Miller I sand tempered assemblage.

Comparison of mound construction and mortuary traits affirmed the affinity between the Pharr and Bynum sites, which had first been suggested by the ceramic analysis. Despite differences in detail, Pharr and Bynum manifest an unmistakable "flavor" of similarity. At both sites, the mounds were built over prepared burial repositories and crematory facilities, and their construction was a signal event. Artifacts were commonly placed with burials at both sites, and cremation as well as inhumation was practiced. It seems clear that the Pharr and Bynum people shared a common mortuary tradition.

The Miller Site, on the other hand, stands by itself, the hallmarks of the Miller I tradition having all but disappeared.

Since we now know more about the units than what kind of pottery was made, it has been proposed to designate Miller I and Miller II as Phases. Previously, these units had been nothing more than subdivisions of a ceramic sequence. This proposition is being put forward cautiously and tentatively. Later work may well show that the mortuary tradition of Pharr and Bynum does not coincide with the Miller I ceramic assemblage. However, use of the present terms seems preferable to cluttering up the literature with still another name.

The appearance of certain artifacts and cultural traits indicates that the Miller I Phase was, in contact with Hopewellian and/or Hopewell-inspired units. It is equally obvious that the traits received were unequally distributed, for, generally speaking, Pharr had the Hopewellian artifacts and Bynum had the Hopewellian-like mounds. More important than these differences is the sharing of the mortuary tradition which, itself, has a generalized Hopewellian flavor.

A cursory examination of Copena and Marksville showed that these units shared little with the Miller I Phase beyond the fact that all three were Hopewellian influenced. Apart from the possible link of the burial platform at Pharr with similar structures at Marksville and Crooks, few traits were shared.

The genesis of the Miller I Phase mortuary tradition must remain uncertain until more is known of other nearby archeological units. The knowledge presently available indicates its closest ties are to Ohio Hopewellian. The northern trade goods and the crematory facilities beneath the mounds appear unmistakable in this regard.

The radiocarbon dates obtained for Pharr and Bynum were considered to be unacceptable. The Marksville vessel at Pharr and the Hopewellian traits at both Pharr and Bynum suggest a date of A.D. 1-200 for the Miller I Phase.