ARCHAEOLOGY SERIES, NO. 2

THE HARROUN SITE

A Fulton Aspect Component of the Caddoan Area, Upshur County, Texas

By Edward B. Jelks and Curtis D. Tunnell

This report was prepared in accordance with a Memorandum of Agreement between The University of Texas and the National Park Service providing for salvage excavations in advance of construction at Ferrell’s Bridge Reservoir, Texas.

Department of Anthropology
The University of Texas • Austin, Texas • 1959

Foreword

Excavation and analysis of the Harroun Site were carried out in 1957, 1958, and 1959 as a part of the nationwide Inter-Agency Archeological Salvage Project. Mound A was excavated by the National Park Service in 1957; Mounds B, C, and D were excavated by The University of Texas in 1958 and 1959 under the terms of a Memorandum of Agreement between the National Park Service and The University of Texas providing for archeological salvage at Ferrell’s Bridge Reservoir.

This report was prepared in accordance with the terms of the Memorandum of Agreement (Contract No. 14-10-333-422) and was submitted to the National Park Service in April, 1959, under the title, “The Harroun Site: A Fulton Aspect Component, Ferrell’s Bridge Reservoir.” As provided in Article I (f) of the contract, the letter of transmittal from The University of Texas and the letter of acceptance from the National Park Service are here reproduced.

Letter of Transmittal

The University of Texas

Austin, Texas

April 16, 1959

Mr. Hugh M. Miller

Regional Director

National Park Service

P. O. Box 1728

Santa Fe, New Mexico

Three copies of the report, The Harroun Site: A Fulton Aspect Component, Ferrell’s Bridge Reservoir, by Edward B. Jelks and Curtis D. Tunnell, are enclosed herewith. This report is submitted in partial fulfillment of the provisions of Contract No. 14-10-333-422, Article 1d, between the National Park Service and The University of Texas.

Sincerely yours,

(Signed)

T. N. Campbell, Director

Texas Archeological Salvage Project

Letter of Acceptance

Department of the Interior

National Park Service

Region Three

Santa Fe, New Mexico

May 1, 1959

Dr. T. N. Campbell, Director

Texas Archeological Salvage Project

University of Texas

Austin 12, Texas

Thank you for the three copies of the report, The Harroun Site: A Fulton Aspect Component, Ferrell’s Bridge Reservoir, by Edward B. Jelks and Curtis D. Tunnell.

Please convey our thanks to Messrs. Jelks and Tunnell for this excellent report prepared under Contract No. 14-10-333-422. It is in keeping with the fine work you are doing in the Ferrell’s Bridge Reservoir.

Sincerely yours,

(Signed)

Hugh M. Miller

Regional Director

List of Tables and Figures

Introduction

The Harroun Site (University of Texas Site No. 41UR10), in the extreme northeastern corner of Upshur County, Texas, is one of several sites excavated in the Ferrell’s Bridge Reservoir area as a part of the Inter-Agency Archeological Salvage Project. The site, which consisted of four small mounds on the south floodplain of Cypress Creek, was located and recorded by E. Mott Davis and Bernard Golden in October, 1957. It promised to produce valuable archeological data, and since it was scheduled to be completely submerged by Ferrell’s Bridge Reservoir in the summer of 1959, immediate steps were taken to provide for salvage excavations prior to its inundation.

In December, 1957, a National Park Service field party excavated the smallest of the four mounds, Mound A. A single extended burial containing two pottery vessels and an arrow point, was found in a shallow grave beneath the mound. It appeared that Mound A had been erected for the purpose of covering the burial. While Mound A was being excavated, the entire site was mapped, several trenches and test pits were dug in the floodplain of Cypress Creek near Mound A, and some of the trees and bushes were cleared from the other three mounds.

In September, 1958, under terms of a co-operative agreement between The University of Texas and the National Park Service, a crew of the Texas Archeological Salvage Project returned to Harroun to complete the investigation of the site. Because of time limitations it was apparent that all three of the remaining mounds (B, C, and D) could not be entirely excavated. Therefore, it was decided to concentrate on Mound C since it appeared superficially to be the least disturbed of the three. After excavation of Mound C was well under way, a portion of the crew was moved to Mound B and it was also opened. Both mounds were found to contain burned remains of house structures.

Fig. 1

HARROUN SITE
41 UR 10
PLAN OF SITE
solid lines mark measured locations
dashed lines mark approximate locations

Investigation of Mound D was begun late in September a few days before the termination of the dig. Time did not permit complete excavation of Mound D that season, but it was tested sufficiently to reveal that it—like Mounds B and C—contained the burned remains of at least one house.

During the 1958 season, in addition to the work at Mounds B, C, and D, several trenches and test pits were dug in the floodplain near the mounds in a fruitless search for additional occupational features.

A final trip was made to the site in February, 1959, by L. F. Duffield, W. A. Davis, and E. Mott Davis of the Texas Archeological Salvage Project. They spent three days exposing and recording the portion of the house at Mound D which had been left unexcavated the previous fall.

As a result of the investigations at the Harroun Site in 1957, 1958, and 1959, all four of the mounds were completely excavated except for certain marginal portions and several check blocks. In addition, the area surrounding the mounds was tested sufficiently to show that there was no general area of occupation near the mounds. The excavation of the Harroun Site was supervised by the senior author. The junior author served as an assistant archeologist during both the 1957 and 1958 seasons.

The artifacts recovered indicate affiliation with the Fulton Aspect of the Caddoan Area (Suhm et al., 1954: 151-161). The Fulton Aspect is the later of two aspects that have been recognized in the Caddoan Area as belonging to the agricultural, ceramic Mississippi culture pattern of the Southeastern United States.

To the Corps of Engineers, whose personnel at Ferrell’s Bridge extended many courtesies and co-operated with the archeological field parties in every possible way, we express our sincere gratitude. Special acknowledgment is due Dr. E. Mott Davis and Dr. J. F. Epstein, staff archeologists of The University of Texas, who visited the site while the dig was in progress, and who not only aided the progress of the excavations by flexing their muscles over shovels and wheelbarrows, but also offered much valuable advice toward solving the technical problems encountered in the field. Their prowess with their guitars contributed greatly to the conviviality of the field camp in the evenings.

A word of thanks and appreciation is due the shovel hands who worked at the Harroun Site in 1957, 1958, and 1959. All extended themselves beyond normal expectation in order to accomplish a maximum 4 amount of work in the limited time available. They are John B. Johnson, Robert L. Brockman, Thomas V. Loveday, W. Brent Hempkins, Floyd W. Sharrock, Andrue Moore, A. C. Harvey, and W. C. Jones. We wish also to thank Mr. Sam Whiteside of Tyler, Texas, a frequent visitor to the site who spent many hours on the dig as a volunteer shovel hand.

Especially to be commended are the assistant archeologists, John Allen Graham, L. F. Duffield, W. A. Davis, and LeRoy Johnson, Jr., all of whom carried out their duties in exemplary fashion despite the continuous pressure under which they were working.

The Site

ENVIRONMENT

Ferrell’s Bridge Reservoir is located in the northwestern part of the Gulf Coastal Plain (Fenneman, 1938: 109-110), which is characterized topographically by rounded hills sculptured from the superficial clays and sands of the region. The subsoil—a sandy clay in various shades of yellow, orange, and red—is capped by a thin mantle of gray sand which evidently derived by differential erosion from the sandy clay. The exposed geological formations recognized in the general region are clays and sands of the Eocene Claiborne group (Sellards et al., 1958: 606-666).

The reservoir is situated in the Austroriparian Biotic Province (Blair, 1950: 93-117). The uplands are thickly timbered, principally with pines, while the stream valleys sustain heavy stands of mixed hardwoods (oaks, cypress, gum, walnut, hickory, holly, and others) in addition to some pines. All the virgin forests were completely timbered out years ago. Bear and panther, which were formerly common, have long since disappeared from the area, but a large population of deer, raccoon, opossum, fox, rabbit, beaver, and other small mammals survives to the present day. The streams abound with several varieties of fish.

The climate is relatively humid, the annual rainfall at the Gilmer station averaging 43.5 inches (U.S. Dept. of Commerce, Climatological Data, Texas, V. 63, No. 13: 361). The mean annual temperature for Upshur County is 65 degrees (Ibid.: 357).

SITE DESCRIPTION

The Harroun Site was situated on the south floodplain of Cypress Creek on the outside of a large bend (Fig. 1). The only occupational features visible on the surface were four small mounds, round to slightly 6 oval in shape, with rounded tops. They were designated Mounds A, B, C, and D. A long, narrow lake, evidently surviving from an old cut-off channel of Cypress Creek, lay beside the creek channel in the northwestern part of the site.

Mound A, located 75 feet south of the creek and 350 feet east of the lake, was by far the smallest of the four mounds. It measured about 30 feet in diameter and rose to a maximum height of approximately two feet above the surface of the floodplain. Mound B stood at the south end of the lake, Mound C was on the west bank of the lake, and Mound D was situated in a precarious position on the brink of the floodplain at the creek channel, 150 feet downstream from Mound A. A shallow depression beside Mound C and two small depressions by Mound B marked possible borrow areas. Mounds B, C, and D were all approximately the same size, about 50 feet in diameter and 2.5 to 3.5 feet high.

The floodplain of Cypress Creek in the vicinity of the Harroun Site was overgrown with an almost impenetrable tangle of underbrush and second growth timber. Old-timers, however, reported that many years ago, before the virgin timber was cut, the stream valleys in this region supported a tall growth of timber with virtually no underbrush.

GEOLOGICAL CONTEXT

Excavation and Recording Methods

The same general procedure was followed in excavating each of the four mounds at the Harroun Site. A stake was placed near the center of the mound and a grid of 5-foot squares was established which tied in with the centrally located stake. Then each quadrant of the mound was excavated separately. Beginning at the top of the mound, an entire quadrant was taken down by regular vertical intervals, usually of 0.5 feet each. The floor of the excavation was cleaned and examined after each level was removed, and measured drawings were prepared to record any zoning or occupational features that were observed in the excavation floor.

The four profiles radiating in the cardinal directions from the central stake were always left intact until measured drawings had been prepared, and other profiles were recorded when deemed necessary. Strategically located check blocks were left at all the mounds, at least until the structure of the mound was determined. In some cases the check blocks were ultimately removed in order to completely expose a house floor or other feature.

For vertical reference, the base stake at Mound A was assigned an arbitrary elevation of 100.0 feet, and all vertical measurements for the entire site were keyed to that stake. For horizontal control a separate grid of 5-foot squares was established for each mound and the area adjacent to it. While the use of a separate grid for each mound had some disadvantages, this method was adopted for two main reasons: (1) so that a key stake with co-ordinates in whole numbers could be located at the center of each mound, and (2) to avoid the use of unwieldy 4-digit numbers for co-ordinates. For each grid a base stake, set well away from the mound, was assigned an arbitrary designation of 0-0, and all other stakes of that grid were labeled with the co-ordinates measured from the base stake in the cardinal directions (as N100-W50, N85-W80, 9 etc.). All the grids were oriented on magnetic north. The designation for each 5-foot square was taken from the co-ordinates at its southeast corner.

Because of the press of time it was impossible to screen all of the excavated soil. Each structural component was spot screened, however, in order to obtain a representative sample of artifacts and other material. Both ⅓-inch and ¼-inch hardware cloth were used for screens.

All artifacts and other specimens were placed in paper bags, on which were recorded the square, vertical interval, geological or structural zone (wherever possible), associated features (if any), appropriate grid, date, and any other pertinent data. For specimens found in situ the exact vertical and horizontal position was also recorded to the nearest ¹/₁₀th foot.

In addition to work in the mounds themselves, several exploratory trenches were dug in the immediate vicinity of Mounds A, C, and D. In each case these tests were tied in to the grid for the appropriate mound. Small tests were made with a post hole digger or a shovel over the entire area between the mounds, and also for some distance beyond the general mound area. These small tests were irregularly spaced from 10 to 100 feet apart. Recording the location of each of them would have required the clearing of a vast amount of underbrush; consequently, since they were all unproductive, only the general areas tested were noted.

Measured drawings, descriptive notes, and photographs were made of the mounds, the burial at Mound A, the house plans at Mounds B, C, and D, and the other occupational features. General site notes were also taken, and a daily log of activities was maintained.

Mound A

Mound A, the smallest of the four mounds, was situated 75 feet south of Cypress Creek and 100 feet west of Mound D (Fig. 1). It was roughly oval in shape with the long axis running due east and west (Figs. 2 and 11). The length at the base was 35 feet; the basal width was 28 feet. The maximum height above the modern floodplain was approximately two feet. No potholes or other disturbances were evident from the surface.

After several trees and bushes had been cleared away, the Mound A area was staked on a grid of 5-foot squares. A stake near the base of the mound on the south side was arbitrarily selected as the base or 0-0 stake, and the designations for the other stakes were determined by measuring their distance from the base stake in terms of the cardinal directions. The elevation of the ground surface at the 0-0 stake was assigned an arbitrary value of 100.0 feet, and all vertical measurements for the entire site were keyed to that point.

As an initial test, most of the southwestern quadrant of the mound was excavated to undisturbed sub-mound soil so that two radial profiles were left standing, one running south, the other west, from the approximate center of the mound. Since time did not permit complete excavation of Mound A, the only squares excavated in the remaining three quadrants were the three 5-foot squares cornering on the approximate center point of the mound and square N10-E5 which was excavated in order to expose Burial No. 1 (Fig. 2). Two additional 5-foot squares (N15-W40 and S25-E0) beyond the limits of the mound structure were carried down several feet below the surface of the floodplain. Each of the latter squares was aligned with one of the two major profiles radiating from the center of the mound. A 1-foot wide check strip, running east-west along the N20 line, was left standing.

All squares were dug in 0.5-foot levels. Those squares in the mound structure were leveled successively on each ½-foot interval of the vertical reference system, while the two squares dug into the floodplain near Mound A were dug by 0.5-foot levels measured from the surface of the ground. All the soil from squares N15-W40 and S25-E0, and an estimated 50% of the excavated mound fill from other squares, was screened through hardware cloth of ¼-inch mesh. All digging was done with shovels except for close work around Feature No. 1 and Burial No. 1 where trowels were used.

In the mound, horizontal plans were drawn to scale at 0.5-foot intervals as the squares were leveled, all features, soil changes, and major disturbances being recorded. Representative profiles were also drawn to scale so as to provide a record of the mound structure and of the relationship of the mound to the floodplain on which it rested.

The major profiles, the burned area near the center of the mound (Feature No. 1), and the sub-mound burial (Burial No. 1) were photographed in both color and black and white.

Fig. 2

HARROUN SITE
41 UR 10
MOUND A
contour interval = 0.5 feet
shading indicates excavated area

STRUCTURE OF MOUND A

The profiles revealed that the bulk of the mound fill comprised a single structural member composed of dark gray, humus stained, sandy, midden soil (Fig. 3). It reached a maximum height of 1.4 feet above the surface of the surrounding floodplain and extended down to an average depth of 1.4 feet below the floodplain surface. The fill of Mound A contained many tiny fragments of mussel shell, bone, charcoal, and stone chips, as well as a few potsherds, projectile points, and other artifacts. Some of the shell and bone fragments showed evidence of burning. The mound fill was unquestionably derived from an occupational area containing an appreciable quantity of cultural detritus.

The top of soil Zone IIb in the area surrounding the mound was, on an average, about 0.7 feet below the surface of the floodplain. The bottom of the mound structure, however, extended to an average depth of 1.4 feet below the floodplain surface where it terminated within Zone IIb (Fig. 3). Thus the surface of Zone IIb immediately beneath the mound formed a shallow, saucer-shaped depression which must have resulted from digging away of the topsoil before the mound was erected. This shallow pit (perhaps originally 1.0 to 1.5 feet deep) was approximately the same size and shape as the base of the mound.

OCCUPATIONAL FEATURES

Besides the mound itself, two occupational features were found in the Mound A area: (1) an area of burned soil within the mound fill (Feature No. 1), and (2) a sub-mound burial (Burial No. 1). Each is described separately below.

Feature No. 1

This was an elongated area of heavily burned, sandy clay lying within the matrix of the mound fill (Fig. 3). Since the northern end of Feature No. 1 was not completely excavated, the exact dimensions were not determined; but the maximum length was evidently between 9 and 10 feet, while the maximum width was 4.3 feet. The long axis ran approximately north-south. Profiles revealed a lenticular cross section with a pronounced thickened area near the mid-point of the east edge. Near the center the burned zone was 0.5 feet thick; the thickened area near the east edge reached a maximum thickness of 1.5 feet.

Fig. 3

HARROUN SITE—41UR10
PROFILE OF MOUND A
(ALONG N20 LINE)
dark gray, sandy mound fill
Feature No. 1
Zone IIb sand, sub-mound
Zone IIa sand

HARROUN SITE
41UR10
PROFILE OF MOUND B
(IDEALIZED SECTION THROUGH CENTER OF MOUND)
floor of House No. 3
central hearth
post molds
stump disturbance
humus
gray, sandy mound fill
whitish, sandy mound fill
undisturbed sub-mound soil

The surface of Feature No. 1 was burned to a conspicuous degree of hardness and was sharply demarcated from the soft, unfired mound fill which overlay it. Beneath the central portion, heat had produced a thin zone of reddish sand which merged gradually with the underlying grayish sand of the mound fill.

Feature No. 1 was situated in the lower portion of the body of the mound. It did not have the appearance of a carefully prepared hearth, but the presence of clay in the burned soil suggests that an irregular-shaped clay base had been laid down where the fire was to be built. It appears that after a layer of sand about a foot thick had been piled up to form the base of the mound, further work on the mound was temporarily interrupted, a crude hearth of sandy clay was prepared near the center of the basal layer of sand, and a fire of considerable intensity was kindled on it. The hardness of the burned, sandy clay of the hearth indicates that the fire was quite a hot one and that it must have burned—continuously or intermittently—for a period of many hours at least. After an unknown interval of time the fire was extinguished and the construction of the mound was resumed and carried to completion. The sharp definition of the hearth surface and the homogeneity of the mound fill above and below the hearth indicate that no appreciable time elapsed between the extinguishing of the fire and the addition of the upper part of the mound fill: otherwise the surface of the hearth should have shown evidence of weathering and the two different stages of mound construction should have been visible in the profiles as separate zones.

Burial No. 1

Beneath the southeast quadrant of Mound A a single burial was found (Fig. 12, A). The skeleton lay in extended, supine position, with the head to the northeast and the feet to the southwest. Preservation of the bones was poor, and several of them (including the left femur, most of the arm and hand bones, the lumbar vertebrae, and the foot and ankle bones) had been destroyed or displaced by gophers whose runs interlaced the entire burial area. As a result of this disturbance the original position of the arms could not be determined.

Two pottery vessels—a carinated bowl and a bottle, both of the type Ripley Engraved (Fig. 12, B-C) had been placed beside the left hip as burial offerings, and an arrow point of the Perdiz type (Fig. 15, O), 15 lying near the outer side of the left knee, appeared also to have been included intentionally with the burial.

The first evidence that a burial was present was the discovery of several foot and ankle bones in a rodent run in square N10-E0. The burrow was traced toward the north for several feet where the distal end of a human tibia was exposed in the northeast corner of square N10-E0. Since it was apparent that the major portion of the burial lay in square N15-E5, that square was taken down. At 0.5-foot intervals the floor of the square was scraped clean with trowels and carefully examined for evidence of a grave outline. However, none was detected in the mound fill.

The burial was finally exposed at a depth of 3.6 feet below the surface of the mound, the floor of the grave lying at an average depth of 1.0 feet below the base of the mound. A vague area of discolored soil (which contrasted faintly with the surrounding undisturbed IIb sand) marked the location of the lower portion of the grave. The edges of the burial pit were quite indefinite, having been considerably disturbed by roots and rodents, but its appearance—both in flat plan and in profile—suggested that a shallow grave about a foot deep and just large enough to accommodate the body had been dug from the floor of the shallow sub-mound pit, the body had been placed in the grave, and then earth had been heaped over both the body and the shallow pit to form the mound.

The skeletal remains from Burial No. 1 were examined by Dr. T. W. McKern, Assistant Professor of Anthropology, The University of Texas. He has kindly provided the following statement:

“The skeletal material from Burial No. 1, Site 41UR10, Upshur County, Texas, consists of one skeleton in a state of poor preservation. Not only are the bones highly fragmented but not one, including the skull, has escaped the destructive teeth of rodents. The brain case is complete but the entire face is missing. Only parts of the mandible are present including both left and right molars (3rd not erupted) and a lower left 2nd premolar. The lower left dentition is in situ. No single bone in the postcranial skeleton is anatomically complete. Also, due to the young age of the remains, most of the epiphyses are missing.

“So far as possible, metric and morphological observations were taken and recorded. But because of the incomplete nature of these observations, they will not be reproduced here.

“The skeleton is that of a 14 year old male with a cranial index of 82 (brachycranic). Although the cranium is slightly distorted there is no evidence of artificial deformation.

“For pathology, the teeth show very little wear which is consistent with the observed age. One pronounced cavity was found on the mesio-disto-occlusal surface of the lower left 2nd molar.

“Because of the almost complete lack of knowledge concerning the range of physical types for the prehistoric populations of Texas it is impossible to associate this skeleton with any known Indian group on a strictly morphological basis.”

DISCUSSION

Mound A was erected on an alluvial terrace of Big Cypress Creek for the purpose of covering a grave. Prior to construction of the mound approximately a foot of humic topsoil was dug away from the surface of the terrace at the selected spot. A shallow grave was then dug in the excavated area, the body was placed in the grave along with two pottery vessels and an arrow point, and the mound was formed over the grave. When the basal portion of the mound was in place, the work was halted temporarily while a fire of considerable intensity was kindled on the incompleted mound—perhaps for ritual purposes. After the fire had been extinguished work on the mound was resumed and brought to completion.

The presence of artifacts in the sub-mound Zone IIb formation indicates that the surface of the terrace in the Mound A area had been lightly occupied prior to construction of the mound. The fill of the mound contained artifacts similar to those in the sub-mound formation, but also contained burned bone scraps, mussel shells, and fragments of charcoal in some quantity. This circumstance shows that the soil of which the mound was built came from a more intensively occupied area than any discovered in the terrace around or under the mound. The source of the mound soil was not determined.

Mound B

This low, approximately circular mound was located on the floodplain of Cypress Creek, about 125 feet south of the stream channel, at the south end of the lake (Fig. 1). A shallow depression about eight feet across lay just southwest of the mound, and a similar but smaller depression was recorded at the northeast edge of the mound. These features probably are borrow sources. An intrusive pothole, located near the center of the mound, was three to four feet in diameter at the surface, but fortunately it proved to be quite shallow and damage to the mound was slight. The maximum diameter of Mound B, measured north-south, was 55 feet (Fig. 4), and the mound reached a maximum height of 2.8 feet above the floodplain.

Mound B was staked for excavation with the usual grid of 5-foot squares oriented on magnetic north. The base stake was 100 feet south and 100 feet east of the approximate center of the mound. Excavation was carried out by the quadrant method as previously described, the southeast and northwest quadrants being excavated first, the southwest and northeast quadrants last. Each quadrant was taken down by ½-foot levels which were keyed to the vertical reference datum. In addition to work in the mound itself, the area immediately south of the mound was tested by means of trenches (Fig. 4). The trenches were dug in ½-foot levels measured from the surface of the floodplain at each square.

The excavating and recording methods used at Mound B were generally the same as described for Mound A.

STRUCTURE OF MOUND B

The excavations in Mound B revealed clearly its internal structure (Fig. 3). An old soil surface, unmistakably defined by a dark humic zone, underlay the entire mound at an average elevation of 99 feet, or approximately the same elevation as the modern surface of the floodplain. This evidently represents the surface humic zone (Zone IIc) of the floodplain at the time the mound was built. Yellow-brown sand (Zone IIb) extended below the buried humic zone to an undetermined depth. Zones IIb and IIc beneath the mound contained a few scattered stone chips and an occasional artifact, but there were no concentrations of cultural material.

Fig. 4

HARROUN SITE
41 UR 10
MOUND B AREA
contour interval = 0.5 feet
shading indicates excavated area

Resting directly on the old floodplain surface was the basal structural component of the mound, a rather compact, circular lens of dark brown sand up to a foot or more thick and averaging about 17 feet in diameter. This lens, which contained abundant charcoal, burned clay daub, bone, shell, and a few artifacts, represented the floor level of a house, designated House No. 3. In and above the floor level were the remains of several charred poles, presumably derived from the burned framework of the house. A burned area approximately four feet in diameter in the center of the lens proved to be the remains of a central fire hearth. It was filled with complex lenses of various shades and textures. A large post mold was found beneath the hearth in the approximate center of the house.

Completely encircling the house outline was a poorly defined zone of yellow-brown sand which lay directly on the buried surface of the floodplain and extended upward a foot or two where it gradually blended into the upper component of the mound fill. This light-colored sand may have been banked against the outside of the house while it was still standing; or it may have resulted from uneven, subsurface staining by charcoal and other organic material of that portion of the mound lying directly above the house. In any event, it was virtually devoid of cultural material, only a very few stone chips, widely scattered, being found in it.

A well defined humic zone, resulting from organic staining after the mound was built, appeared at the surface of the mound. It averaged about 0.5 feet in thickness.

Except for the clay in the hearth and in the house floor, the entire mound was constructed of sandy soil like that of the surrounding floodplain, whence it undoubtedly was derived. The depressions on the northwest and southeast sides of the mound are probably the borrow sources for the sandy soil. The clay could have easily been obtained from exposures in the cut banks at the edge of the creek channel.

OCCUPATIONAL FEATURES

Besides the two possible borrow pits mentioned above, the only occupational feature found at Mound B was House No. 3.

House No. 3

This house was erected on the surface of the floodplain before the mound was built. The purpose of the mound apparently was to bury the remains of the house after it had burned.

Beneath the house floor zone, which was described in the preceding section, were found 59 post molds measuring from 0.25 to 1.3 feet in diameter and extending from 0.3 to 2.5 feet below the floor (Fig. 5). The faint gray stain of the post molds was quite dim and difficult to distinguish. They were located by cutting a vertical face completely around the house area, then carefully cutting the face inward from all sides. As the post molds were located, they were plotted on a horizontal plan and a measured profile drawing of each was prepared.

Twenty-three of the post molds formed a circular outline representing the perimeter of a house approximately 17 feet in diameter (Fig. 5). The peripheral molds averaged 0.5 feet in diameter and were spaced, as a rule, about two feet apart. At the southeast edge of the house were two parallel lines of three molds each which defined an extended entranceway. Because of disturbance in the entranceway area by tree roots, only the bottom portions of the entrance molds were preserved. Their arrangement suggests that some of the post molds related to the original entranceway were not discovered.

Within the external ring of post molds were 30 irregularly spaced molds, including four very large ones which probably held the bases of relatively heavy roof supports. Two concentrations of smaller post molds (one on the northeast side of the house, the other on the southwest side) possibly mark the location of interior structures such as sleeping or storage platforms. In the center of the house was a relatively large post mold, over which the fire hearth had been built. This probably represented a center post used in construction of the house and then removed when the house was completed.

The hearth was located in a shallow depression at the center of the house. It was in the form of a basin about four feet in diameter and one foot deep. The sandy soil underlying the hearth had been burned to a deep reddish color.

Fig. 5

HARROUN SITE
41 UR 10
PLAN OF HOUSE NO. 3
MOUND B
post mold, exterior wall
post mold, interior
post mold, entrance
central hearth
stump disturbance

From all indications this house was constructed in a manner similar to that described by early Spanish and French explorers in the Caddoan Area (Swanton, 1942: 148-154). A ring of poles, each with its base end set in a deep hole, was placed in upright position around a tall center 22 post used as a work platform. The tops of the poles were drawn together at the center and bound. Small tree branches were then woven, horizontally, between the upright poles, grass thatching was applied, and, in some cases, the exterior was plastered with a coat of clay mud. (Many pieces of burned clay daub, some bearing impressions of sticks and grass, were found on and above the floors of all the houses at the Harroun Site.) After the house was completed, the center post, used only to facilitate construction, was removed. Interior support posts may have been added, and platforms for sleeping or storage were built inside the house.

DISCUSSION

House No. 3, a circular, wattle-and-daub structure with a southeastern entranceway, was built on the surface of the Cypress Creek floodplain. There were probably four interior roof support posts, two or more interior platforms for sleeping or storage, and a centrally located, prepared hearth with a clay base. Possibly, a low embankment of sand was thrown against the wall around the exterior of the house.

The period of occupation at the house is unknown, but the scarcity of artifacts suggests that it was of short duration, or else that it was used for specialized—perhaps ceremonial—purposes. A domiciliary structure ordinarily would have much more cultural refuse about it than did House No. 3, unless it was occupied for only a very brief period of time. Stone chips and a few artifacts in the floodplain beneath the house floor indicate that the spot had been lightly occupied prior to the construction of the house.

That House No. 3 burned is evident from the charred poles and bits of heavily burned, wattle-impressed, clay daub lying on and above the house floor. Shortly after the burning, a mound of sandy soil, undoubtedly derived from the adjacent surface of the floodplain, was heaped over the house ruins.

Burial of the house remains beneath a mound implies that the house had a special significance, possibly of a ceremonial nature. Consequently it may be conjectured that perhaps House No. 3 was a small temple or chapel which was ceremonially burned and buried.

Mound C

Mound C was situated on the west bank of the lake, 350 feet northwest of Mound B (Fig. 1). This mound was in the shape of a broad oval with its long axis oriented in an east-west direction. It measured 62 by 52 feet at the base and reached a maximum elevation of 102.6 feet, or slightly more than three feet above the modern surface of the floodplain (Figs. 6 and 11, B).

There was a circular depression approximately nine feet in diameter in the top of the mound where pothunters had been at work. Excavation revealed that the pothole had been dug to a depth of 4.8 feet and had later been partially filled by natural agencies. Unfortunately, the pothole had destroyed most of the central hearths associated with the two house floors found at the base of the mound.

After the trees and bushes had been cleared from Mound C the standard grid of 5-foot squares was established with a base stake set 125 feet south and 100 feet west of the approximate center point of the mound. The initial step in excavating the mound was to dig the southwest quadrant down to elevation 100.0 feet. Next, the southeast quadrant was excavated to the same level so that an east-west profile remained standing completely across the mound. After the profile had been studied and recorded, the other two quadrants were removed and the entire mound was levelled at elevation 100.0 feet, where a circular zone of dark soil containing a large amount of charcoal marked the outline of what later proved to be the remains of two houses, one superimposed on the other.

A narrow east-west trench was next dug across the house area, the north edge of the trench being on the N125 line so that it matched the bottom of the major east-west profile which had already been removed. This trench revealed two thin layers of dark midden soil, each of which represented the floor level of a house (Fig. 7). The two floor levels were separated by a layer of clean, yellow sand. The lower floor rested on undisturbed soil at the base of the mound. Numerous charred segments of poles lay in a jumble on and just above the upper floor as though the house walls had burned and collapsed.

Fig. 6

HARROUN SITE
41 UR 10
MOUND C AREA
contour interval = 0.5 feet
shading indicates excavated area

The house floors were completely excavated, the artifacts and other material associated with each floor being collected and sacked separately wherever possible. Two concentric rings of post molds at the periphery of the house area were exposed and recorded, as were several interior post molds (Fig. 8). An entrance passageway was delineated at the west side of the house area.

In order to determine the relationship of the mound to the floodplain several short trenches were carried from the edge of the mound out into the floodplain formation (Fig. 6). A depressed area in the surface of the floodplain between the mound and the lake was also trenched in an effort to determine whether it may have been a borrow pit. Several other trenches were dug south and west of the mound in an unfruitful search for any middens, houses, burials, or other occupational features that might have been located near the mound.

Throughout the excavation of Mound C, major profiles, horizontal plans at ½-foot intervals, and occupational features were described in the field notes and drawn to scale. Major profiles and features were photographed. Most of the digging was done with shovels, but trowels were used in part for excavating the two thin floor zones and for several other situations where close attention to detail was desirable. Because of time limitations only representative, spot screening was attempted.

STRUCTURE OF MOUND C

Profiles of Mound C revealed remnants of an old stabilized surface with a well developed soil profile (including a superficial humic zone) lying immediately beneath the mound fill (Fig. 7). The elevation of the old surface averaged approximately 99.4 feet which is also the average elevation of the modern floodplain surface around the mound: therefore it appears certain that the first of the two houses was built directly on the floodplain surface and that there has been no appreciable change in the surface elevation of the floodplain since the mound was built. The first house burned, after which the second house was built over its remains; then the second house burned and the mound was erected over the ruins of the houses.

Fig. 7

HARROUN SITE
41 UR 10
PROFILE OF MOUND C
(IDEALIZED SECTION THROUGH CENTER OF MOUND)
pothole
floor of House No. 2
floor of House No. 1
sterile zone between house floors
post mold
humus (note buried humus zone between N130 and N140)
gray, sandy mound fill
whitish, sandy mound fill
undisturbed sub-mound soil

HARROUN SITE
41 UR 10
PROFILE OF MOUND D
(IDEALIZED SECTION THROUGH CENTER OF MOUND)
pothole
floor of House No. 4
dark brown sand
gray-brown sand
post mold
humus
gray, sandy mound fill
whitish, sandy mound fill
undisturbed sub-mound soil

A low embankment of sand similar to that at Mound B encircled the house area just outside the peripheral ring of house post molds (Fig. 7). Apparently this embankment was built while one of the houses was still standing since its inner edge is almost vertical in places as though it had been banked against the outside wall of the house. After the later house burned, the mound was heaped over both this embankment and the house ruins.

The geologic structure of the floodplain at Mound C was apparently the same as previously described although none of the excavations were carried deep enough to expose Zone I, the basal member of reddish clay. The mound structure rested on the surface of Zone IIc (Fig. 7) and was composed of four distinct structural units as follows (in order from bottom to top):

OCCUPATIONAL FEATURES

The only occupational features discovered at Mound C were the two house patterns.

House No. 1

The lower house floor at Mound C, designated House No. 1, rested directly on the old surface of the floodplain (Fig. 7). The floor zone was a circular lens of dark gray—almost black—sand with a greasy texture. It averaged 0.4 feet in thickness and measured some 18 feet across. This floor zone contained numerous bits of charcoal and burned clay daub, a few stone chips, mussel shells and garbage bones, and a small number of artifacts.

Around the perimeter of the floor was a ring of post molds representing the exterior house wall (Fig. 8). Average diameter of the ring was 18 feet. Each post mold extended downward below the floor level into the sub-mound floodplain. The individual molds ranged from 0.35 to 0.75 feet in diameter, the bottoms being from 1.3 to 2.0 feet below the floor level. There was a total of 29 definite molds plus one probable mold in the peripheral ring, and disturbances on the west and south sides of the house appeared to have obliterated at least five others. The posts had been set about 1.5 to 2.0 feet apart on an average. Time did not permit vertical sectioning of all the molds, but several were carefully sectioned and studied to determine the level from which they had been dug. All began at the floor of House No. 1, none extending above that level.

The large pothole observed in the top of the mound continued downward entirely through the floor of House No. 1, although it had narrowed to a diameter of less than four feet where it intercepted the floor (Fig. 8). Unfortunately the pothole had destroyed the major portion of a centrally located hearth that must have been associated either with House No. 1 or the overlying House No. 2. Actually, there was probably a hearth for each house, the later one constructed directly above the earlier one. But since only a narrow segment of burned soil remained to mark the eastern margin of the hearth (or hearths), the structural details could not be ascertained. As nearly as could be estimated by the surviving portion of the hearth, it must have been approximately three feet in diameter.

Fig. 8

HARROUN SITE
41 UR 10
PLAN OF HOUSES NO. 1 & 2
MOUND C
post mold, House No. 1
post mold, House No. 2
post mold, House No. 1 or House No. 2
probable post mold
remnant of central fire pit
ash lens
disturbance
disturbance
pothole

Beneath the pothole—which luckily terminated a foot or so below the floor of House No. 1—were the bottom portions of two post molds (Fig. 7). These were undoubtedly from the center posts used during construction of Houses No. 1 and 2. Although the exact circumstances could not be reconstructed because of disturbance, the center posts presumably were removed when the houses were completed and the hearths placed over the molds.

In addition to the two center molds, there were two other post molds within the interior of the houses at Mound C. One was just east of the hearth area, the other was northwest of the hearth (Fig. 8). Both were exposed in the excavation floor at the level of House No. 1, and since they were not encountered above that level both probably relate to the earlier house.

An extended entranceway on the west side of the houses was delineated by an elongated area of organically stained soil and by two parallel rows of post molds (Fig. 8). The stained area was clearly discernible in the mound fill above both house floors. Despite extremely careful excavation of this stained area, however, only the bottom portions of the post molds—well below the floor level of House No. 1—could be seen. Consequently the level from which the entranceway post holes were dug could not be determined and it is uncertain to which of the two houses they belonged. House No. 2 must have had its entranceway on the west side because the organically stained outline showed clearly in the mound fill well above the House 2 floor level. Possibly both houses had their entranceways in this same area.

House No. 2

House No. 2 was represented by a distinct floor zone and by a circle of post molds. The floor zone (Fig. 8) consisted of a lens of brownish sand averaging about 15.5 feet in diameter, with a maximum thickness near the center of almost a foot. It lay directly above the floor of House No. 1, but was separated from it by a thin layer of clean, sterile sand 0.1 to 0.3 feet thick. The sterile sand layer was possibly placed over the burned ruins of House No. 1 in order to provide a clean floor for House No. 2.

The peripheral ring of post molds (Fig. 8) averaged a little less than 14 feet in diameter (or almost four feet less than that of the underlying House No. 1) and lay entirely inside the exterior wall of House No. 1. The two rings were not quite concentric, however, the center point of House No. 2 being slightly to the west of the center point of House No. 1. The post molds of House No. 2 were from 0.45 to 0.85 feet in diameter, and they extended from 1.6 to 2.0 feet below the level of the related house floor. Several of the molds were sectioned vertically to determine the level from which they had been dug. They could be clearly traced from the floor of House No. 2 down through the floor of House No. 1 into the sub-mound floodplain.

As was pointed out above in the description of House No. 1, there was probably a circular, centrally located hearth associated with House No. 2, and one of the two center posts whose molds were found beneath the hearth area must have been used in the construction of the later house. There appeared to be no other interior post molds associated with House No. 2. The entranceway was probably on the west side.

DISCUSSION

Excavation of Mound C revealed that a circular house (House No. 1) was built on the south bank of the Harroun Site lake, was occupied for an unknown period of time, then was burned—perhaps intentionally. After a thin layer of sand had been strewn over the burned ruins, a second, smaller house (House No. 2) was erected on the remains of the earlier house. House No. 2 was likewise destroyed by fire, after which the remains of both houses were buried under a mound of sand.

Both houses probably had centrally located hearths, and one or both of them had an entranceway opening to the west. As at Mound B, a low pile of sandy soil may have been banked around the outside of one or both houses before they were destroyed. Architecturally the houses at Mound C were quite similar to the one at Mound B.

The sparse occurrence of artifacts and other cultural refuse suggests that neither House No. 1 nor House No. 2 was an ordinary domicile. It appears likely, rather, that both were ceremonial structures of some sort. This hypothesis is supported by the fact that the houses were considered important enough to be afforded burial beneath a mound, probably after having been ceremonially “cremated.”

Mound D

This low circular mound was located on the south bank of Cypress Creek about 150 feet east of Mound A (Fig. 1). It was perched at the very edge of the floodplain overlooking the creek channel. In recent years the channel had been migrating laterally and had begun to encroach on the north edge of the mound. The average diameter of Mound D at the base was about 60 feet, and its highest point was at a relative elevation of 100.6 feet, or about 2.5 feet above the surface of the surrounding floodplain (Fig. 9). A shallow depression about 12 feet across in the top of the mound marked the location of the usual pothole. This pothole had originally been only 5 to 6 feet in diameter, but had been considerably enlarged at the surface of the mound by recent erosion.

Excavation of Mound D was begun shortly before the end of the 1958 field season. It was dug, like the other mounds, by the quadrant method; but, because there was not enough time for thorough excavation, only the southwest quadrant was carried down to the sub-mound level in 1958. The other three quadrants were taken down to the 98-foot level, however, where a circular zone of dark, organically stained soil, 19.8 feet in diameter, clearly outlined the location of a house structure (House No. 4) similar to those at Mounds B and C. During the final work at Harroun in February, 1959, the entire northwest quadrant was exposed, excavated, and recorded. Only the peripheral ring of post molds was exposed in the other two quadrants.

The southwest quadrant of the mound was excavated in 0.5-foot levels; all other portions were taken down in 1.0-foot levels. Horizontal plans were recorded at all levels and photographs were taken. Vertical walls 1.5 feet thick were preserved across the mound along the W100 and N100 lines (Fig. 9), and trenches three feet wide were extended north, south, and west of the mound in order to obtain complete vertical profiles. Excavation and recording methods were generally the same as previously described for the other mounds.

Fig. 9

HARROUN SITE:
41 UR 10
MOUND D AREA
CYPRESS CREEK

STRUCTURE OF MOUND D

The structure of Mound D was clearly indicated by the vertical cross sections (Fig. 7). An old humus-stained surface underlying the marginal portions of the mound was sharply defined at an average elevation of 98.0 feet. No artifacts or cultural refuse were found in the floodplain below this surface. Prior to construction of the mound a shallow, circular pit had been excavated in the surface of the floodplain to an average depth of 1.5 feet. The sides of the pit sloped sharply downward, and the floor was approximately level. An embankment of yellow-brown sand, possibly composed of back-dirt from the pit, was mounded up about 1.5 feet high and four to six feet wide around the perimeter of the pit. This light sand zone contained a few artifacts but little or no charcoal.

A hard-packed house floor about 0.2 feet thick lay on the bottom of the pit. This floor zone was composed of compact sandy clay which contrasted sharply with the overlying mound fill. Charcoal, ash, and burned clay daub were found in quantity in the floor zone, but only a few artifacts were recovered. Just above the house floor was a 1-foot thick layer of dark gray-brown sand containing several charred poles and a large amount of charcoal, ash, and burned clay daub. Above that was the sandy fill making up the bulk of the mound. A mantle of surface humus from 0.2 to 0.8 feet thick covered the mound.

A ring of post molds was discovered around the edge of the floor, and other molds on the west side of the house marked the position of an extended entranceway (Fig. 10). No interior post molds were discovered.

The pothole, which extended downward through the center of the floor, had apparently removed a centrally located hearth, only slight evidence of burning at the edge of the pothole remaining to show that the hearth had been there.

The entire mound fill, including the embankment around the house, was composed of various shades and textures of sand. All of this material was probably derived from the sandy floodplain surrounding the mound. Small quantities of clay around the hearth and on the house floor could have been acquired at nearby outcrops in the stream channel.

As at Mounds B and C, the circular shape of the house at Mound D was outlined by an area of organically stained soil which extended upward 35 from the house floor almost to the surface of the mound. The flanks of all three mounds were of light colored sand which contrasted sharply with the dark, circular house outlines. The only reasonable conjecture thus far advanced to explain this circumstance is that a low embankment of relatively clean sand had been piled against the exterior wall of each house. Thus when a house burned the embankment would remain standing, well above the house floor, as a sort of mold of the lower portion of the house. Then when a mound was erected over the burned house remains and the standing embankment, the outline of the house might appear in the mound fill as a cast of the house, delineated by the circular embankment.

OCCUPATIONAL FEATURES

House No. 4

This house was circular in shape, with an exterior wall formed of upright poles, wattle, and clay daub. Post molds indicated that there were at least 27 of the upright poles in the exterior wall (Fig. 10). They were 0.3 to 0.6 feet in diameter at the base and were set about two feet apart on an average. An interior hearth near the center of the house was probably circular in shape and an estimated three to four feet in diameter. Its exact dimensions could not be determined because of disturbance by the pothole. No interior post molds were found.

Remains of an extended entranceway on the west side of the house consisted in five post molds which outlined the two parallel sides of the entranceway. The entranceway was slightly less than three feet wide, and it sloped downward from the surface of the floodplain into the house pit.

Burned poles and burned clay daub with wattle impressions showed that House No. 4 had been destroyed by fire.

DISCUSSION

Investigation of Mound D revealed that the following sequence of events had taken place. A round, Caddoan type of house (House No. 4) with an extended entranceway on the west side was built in a shallow, excavated pit on the south bank of Cypress Creek. Architecturally the house was quite similar to those at Mounds B and C. Sand was probably banked against the outside wall of the house to a height of somewhat more than a foot. After an unknown period of time the house was destroyed by fire, and then the house remains and the surrounding embankment of sand were buried beneath a mound of sandy soil. This duplicates essentially the events reconstructed for Mounds B and C, the only unique feature at Mound D being the pit in which the house was built.

Fig. 10

HARROUN SITE
41 UR 10
PLAN OF HOUSE NO. 4
MOUND D
post mold
unexcavated
burned area
pothole

Probable ceremonial burning, burial beneath a mound, and a scarcity of domestic artifacts and refuse suggest that House No. 4 was not an ordinary residence, but a small temple, chapel, or similar structure used for ceremonial purposes.

Description of the Artifacts

A total of 610 artifacts was recovered from the Harroun Site, consisting of ceramics, chipped stone implements, and a few milling stones and pitted stones. The first step in ordering the artifacts was to lump them all together in one heap. Then they were separated into general groups such as pottery, dart points, arrow points, scrapers, pitted stones, etc. Next, each group was further divided and subdivided into as many categories as seemed warranted until a number of small groups resulted, each containing a series of individual specimens with similar basic characteristics.[1] Finally, each of the small groups was compared to similar material from other sites and identified with specific types wherever possible.

The artifacts are described below by groups. Provenience of the groups and types within the site is discussed in the succeeding section.

CERAMICS

In addition to the two vessels from the burial at Mound A, ceramic specimens comprise a total of 562 sherds. The paste of these sherds is characteristically sherd tempered, occasionally with the addition of small quantities of sand and/or bone particles. There is no shell tempering. Study of the sherds indicates that bottles, jars with outcurved rims, carinated bowls, and possibly other forms are represented. Exterior surface treatment includes brushing, smoothing, polishing, and red filming; smoothing and red filming also occur as interior surface treatments. Techniques used in applying decorations are incising, engraving, appliquéing, and punctating.

The small quantity of sherds did not permit reconstruction of any vessels nor of any complete design elements: consequently correlations between techniques of decorating, design elements, vessel shapes and surface treatment were impossible as a rule, and a comprehensive typological analysis of the ceramics could not be made.

The ceramics were separated on the basis of decorative technique into six groups: brushed, incised, appliquéd, punctated, engraved, and plain. Each group is described separately below.

Brushed Pottery

Of the 141 brushed sherds, 13 are rimsherds and 128 are from body areas. The brushing is always on the exterior surface, the interior surfaces being poorly to fairly well smoothed. Wall thickness ranges from 5 to 9 mm. Lips are rounded and slightly everted.

Clay lumps of varying sizes—evidently ground up sherds are visible in the paste of most sherds, and 39 of the 141 brushed sherds also contain bone tempering. There are particles of sand in all the sherds, a few having so much that their surfaces have a distinctly sandy feel when rubbed between the fingers. Paste colors range from creams and buffs to fairly dark grays, with most sherds falling into the lighter shades of buff, brown, and gray—indicative of oxidation during firing. Some sherds have light exterior surfaces and dark interior surfaces, suggesting that the vessels stood upside down during firing.

Most of the brushed sherds could not be definitely identified with any specific pottery types; however, several sherds were assigned to the types Bullard Brushed of the Frankston Focus (Suhm et al., 1954: 252 and Pl. 9) and Pease Brushed-Incised of the Bossier Focus (Webb, 1948: 110-113 and Pls. 11 & 12; Suhm et al., 1954: 338 and Pl. 53).

There are 17 Bullard Brushed sherds, 13 of them from the body of a single vessel, the other four from the rim of another vessel (Fig. 13, A-B). All were found at Mound C. Both vessels were barrel-shaped with a slight, evenly curved constriction in the neck area. There were one or more horizontal rows of punctations made with a blunt stick separating the body area from the rim area on both vessels, but there was no angle at the juncture of the body and the rim. On the vessel represented only by body sherds, the brushing consisted of short, overlapping strokes in random directions, creating a roughened exterior of uneven appearance. The rim treatment of this vessel could not be determined. The other 40 Bullard Brushed vessel was represented by four rimsherds which fitted together. The rim of this vessel curved outwardly and was evenly brushed in a diagonal direction. A horizontal row of punctations appeared at the bottom of the rim. Both Bullard vessels were relatively large with wide mouths.

Six of the brushed body sherds (Fig. 13, C, D, G) were identified as type Pease Brushed-Incised because they have vertically brushed sections separated by vertical appliqué strips. Five are from Mound D, the other from Mound A. Five of the six have closely spaced punctations or indentations pressed into the strips. One of the Pease body sherds (Fig. 13, D) is attached to a portion of the rim which is brushed horizontally. On this sherd there is a marked angle at the juncture of the body and the rim, and a horizontal row of small punctations made with the blunt end of a stick is impressed along the line of the angle. Other Pease sherds with incising instead of brushing are described later.

The other 118 brushed sherds were not assigned to definite types, but will be described here as a group. In all or most of the vessels represented by the miscellaneous brushed sherds the coiling method was employed. Fractures along coil lines, and vessel curvature on some of the larger sherds, made it possible to orient 30 of the brushed body sherds with respect to the vessels from which they came. The brushing on all 30 is in an approximately vertical direction (Fig. 14, C-D). The nine rimsherds, in contrast, are all brushed horizontally (Fig. 14, A-B) except for one which is brushed diagonally. On one sherd containing portions of both body and rim, the body is brushed vertically and the rim horizontally. The body and rim areas are separated on this sherd by a horizontal row of small, closely spaced punctations made with a pointed instrument. On three of the nine rimsherds there are similar single rows of punctations just below the lip.

The miscellaneous brushed sherds appear to have come, by and large, from jars with outcurved rims, the bodies brushed vertically and the rims brushed horizontally. The body and rim areas were probably separated in most cases by a horizontal row of closely spaced punctations made with the end of a stick, and similar rows of punctations were placed on some rims just below the lip at the top of the brushed zone. The juncture of the body and the rim usually formed a distinct angle. There is the possibility that some vessels with brushed bodies had plain 41 or incised rims, or, conversely, that some with brushed rims had plain or incised bodies. The horizontally brushed rims, some with punctations, are quite similar to the rims of type Pease Brushed-Incised, and it is quite likely that some of the brushed sherds came from Pease vessels. It is also possible that some of the brushed body sherds are from vessels with incised rims of the Maydelle Incised type (Suhm et al., 1954: 324 and Pl. 46) described later.

Incised Pottery

Thirty-nine sherds with incised lines were found at the Harroun Site, 31 of them body sherds and the other eight from rims. The incised sherds are all sherd tempered with varying amounts of sand included in the paste. Bone tempering is also present in eight. Surface colors are predominantly light browns and grays, indicating an oxidizing atmosphere during firing. The characteristic surface treatment of the exteriors is smoothing (done before incising), and all the interiors are smoothed. Wall thickness varies from 4 to 8 mm. Two sherds have red slips.

Eleven of the incised sherds have vertical or diagonal appliqué strips marking off the vessel body into panels, each panel being decorated with parallel incised lines (Fig. 13, E-F). These have all the characteristics of Pease Brushed-Incised body sherds, and they have all been assigned to that type.

One sherd (Fig. 14, E) with punctation-filled incised panels is unmistakably from a bowl of type Crockett Curvilinear Incised of the Alto Focus, Gibson Aspect (Newell and Krieger, 1949: 98-101 and Fig. 36). This sherd has a straight rim with a squarish lip; the exterior was smoothed before decorating and the interior is poorly smoothed. Decoration consists of a portion of one curvilinear panel outlined with incised lines and filled with small, crescentic punctations. Part of a crack-lacing hole is retained on one edge of the sherd. This specimen was found in a disturbed area at Mound B.

A sharply incurving rimsherd (Fig. 13, H) with four parallel incised lines in the broad, flat lip is from a vessel which was not of traditional Caddoan shape or decoration. It was found over four feet deep in Zone IIb of the floodplain near Mound A. The incurving rim, the flat lip, and the position of the incised lines are all quite similar to styles of the Lower Mississippi Area—especially as exemplified by the types Coles Creek Incised and Chase Incised (Ford, 1951: 74-77). Another interesting 42 feature of this sherd is a bright red slip which covers both the interior and the exterior surfaces.

A second sherd (Fig. 13, I) with characteristically Lower Mississippi design is also from Zone IIb of the floodplain. This sherd came from the neck area of a jar and has portions of a decorated rim and a plain body. The decoration consists of two sets of parallel lines crossing each other at an angle so as to form a series of diamond-shaped elements. Inside each diamond is a triangular punctation made with the corner of an angular instrument. There is an abrupt decrease in wall thickness at the bottom of the rim so that a typically Lower Mississippian “overhanging line” effect is produced. In design and general execution this sherd is similar to the type Beldeau Incised (Ford, 1951: 81-83) of the Coles Creek period in the Lower Mississippi Area, but its paste appears to be more in the Caddoan than in the Baytown tradition.

The 25 incised sherds not assignable to any specific type comprise five rimsherds and 20 body sherds. Fifteen of the body sherds bear thin lines sliced into the plastic clay with a sharp instrument; the other 10 were incised with a blunt-tipped implement which gouged out, rather than sliced, the lines. Two sherds (Fig. 14, G) have a horizontal row of closely spaced punctations in the neck area. Of the five rimsherds, one has three widely spaced, horizontal, incised lines; three (Fig. 14, F) have a design of widely spaced, cross hatched incised lines; the fifth bears traces of two horizontal incised lines on the lower part of the rim above a plain body. Some of the smaller body sherds could have come from Pease Brushed-Incised vessels and the three rimsherds with cross hatched design could well be from Maydelle Incised vessels.

Thus the 39 incised sherds include at least 11 from vessels of type Pease Brushed-Incised, one is type Crockett Curvilinear Incised, and two appear to be intrusions from the late Coles Creek period of the Lower Mississippi Area. The unidentified sherds are all typically Caddoan in their general characteristics, and three of them may represent type Maydelle Incised of the Frankston Focus.

Appliquéd Pottery

The decorative technique of appliquéing occurs commonly at the Harroun Site, principally in combination with brushing and incising on the type Pease Brushed-Incised described above. However, there are five sherds with appliqué strips but with no traces of brushing or incising 43 (Fig. 14, J). Paste characteristics of these sherds are the same as for the previously described appliquéd sherds of the Pease type, and it is believed that they are from vessels similar to Pease Brushed-Incised except that the panels on the body were left plain instead of being filled with brushing or incised lines.

Punctated Pottery

As stated in previous sections, horizontal rows of punctations occur commonly in combination with brushing on the rims of jars, and punctations also appear in vertical rows on appliqué strips applied to the bodies of Pease Brushed-Incised jars. Thus punctations seem to occur most commonly in association with incising, brushing, and appliquéing. However, eight sherds have punctations as the only decorative technique. Four of them have sections of single rows of closely spaced punctations, all made with the ends of sticks or similar implements (Fig. 14, H). The other four sherds (Fig. 14, I) are covered with small, free punctations. On one of the latter the punctations were made with a blunt stick; the other three have paired fingernail impressions.

The punctated sherds are all similar in paste characteristics. All are sherd tempered and one also has a small amount of bone temper. Exterior colors are light to medium brown and gray, while the interiors tend toward darker shades of the same colors. The exterior surfaces were smoothed before the punctations were applied; the interiors are also smoothed.

Engraved Pottery

Only two complete pottery vessels were found at the Harroun Site, a carinated bowl and a bottle, both engraved and both associated with the burial beneath Mound A.

The carinated bowl (Fig. 12, B) has a flat, round base and a compound rim which turns sharply inward at the shoulder to form a narrow, almost vertical panel approximately 1.5 cm. high. Above this panel the rim turns sharply outward to form a second panel extending to the lip. Four equally spaced peaks rise from the upper panel of the rim. The bowl stands 9.5 cm. high and measures 21.0 cm. wide between opposing rim peaks. Both the exterior and the interior surfaces have been well 44 smoothed, and marks of the smoothing tool are clearly visible both inside and outside the vessel.

The lower rim panel of the carinated bowl bears a stylized version of the interlocking scroll design, featuring broad, deep, engraved lines with small excised zones. The upper rim panel has elongated triangular designs on the rim peak areas with broad, parallel, vertical, engraved lines within the triangles. An almost identical bowl is pictured by Suhm et al., (1954: Pl. 57, I) as an example of the type Ripley Engraved.

The engraved bottle (Fig. 12, C) has a broad, squat body and a tall neck with expanded rim. Total height is 23.1 cm. The body is 12.8 cm. high by 18.3 cm. wide; the height of the neck is 10.3 cm., its minimal diameter is 4.5 cm., and the oral diameter is 5.5 cm. An interlocking scroll design is repeated twice (slightly asymmetrically) on the body, and some of the engraved lines have small, pendant triangles which are hachured or excised. There are also several cross hatched, triangular elements. The exterior surface is dark gray in color and has been well smoothed. The bottle has been identified as an example of type Ripley Engraved (Suhm et al., 1954: 346 and Pl. 59).

In addition to the two vessels from Burial No. 1, examples of the engraving technique appear on 107 sherds from the Harroun Site. The paste of these sherds is fairly consistent in being fine grained and relatively hard, and all appear to have sherd temper. The paste of the engraved sherds also contains moderate amounts of sand, and 23 of them have bone particles added as a supplementary tempering agent. Surfaces are smoothed, both on the interior and exterior, and the exterior surfaces of several sherds are highly polished. Fractures along coil lines indicate that manufacture was by the coiling method. Wall thickness ranges from 3 to 7 mm.

A big majority of the engraved sherds are from the rims of carinated bowls with rounded, out-turned lips, but several are from the bodies of bottles and one is from the rim of a jar. Most of the sherds are small, having sections of from one to four engraved lines which are too incomplete to reveal any distinctive design elements: consequently no typological affiliations can be determined for them. There are some, however, which can definitely be assigned to previously recognized typological categories.

On four sherds (Fig. 14, L) are small, excised, diamond-shaped elements enclosed by concentric diamond-shaped lines, and two sherds 45 (Fig. 14, K, M) are decorated with swastikas enclosed by circles. Both of these designs are known only on the type Ripley Engraved; therefore there is no hesitation in identifying these six sherds as Ripley. Two other sherds with portions of Ripley-like designs were assigned to the same type.

One sherd (Fig. 14, O) from a small carinated bowl is decorated with a curvilinear interlocking scroll design characteristic of the type Taylor Engraved (Suhm et al., 1954: 360-362 and Pl. 65). Another sherd (Fig. 14, N) from an engraved bottle appears also to be of the Taylor type, as does an engraved rimsherd (Fig. 14, P) from a jar.

Four sherds came from the lower neck region of a bottle. A single, fairly heavy, engraved line filled with red pigment encircled the base of the neck, and the neck contracted sharply toward the top in typically Gibson Aspect style. The paste is fine grained in texture and almost black in color. The exterior is well smoothed and polished, but the interior is very poorly smoothed, as is usual for Caddoan Area bottles. The wall of the neck is 6 mm. thick. This bottle is almost certainly a Gibson Aspect form, possibly type Hickory Fine Engraved of the Alto Focus (Newell and Krieger, 1949: 90-91 and Fig. 33; Suhm et al., 1954: 294 and Pl. 31). It was associated with the floor of House No. 3 at Mound B.

The other 92 engraved sherds could not be identified with any specific types. However, they all are from carinated bowls and bottles characteristic of the Fulton Aspect, the forcefully engraved lines of many suggesting Titus Focus in particular. An interesting note is the occasional widening of an engraved line by a series of closely spaced, gouged out lines, creating small zones which are not quite completely excised (Fig. 14, K). The identical technique was noted by E. Mott Davis (1958: 61) at the Whelan Site, located on Cypress Creek about 15 miles below the Harroun Site. This treatment is similar in a general way to that of the type Poynor Engraved of the Frankston Focus, but the design elements on which it occurs, both at Harroun and Whelan, are characteristic of Titus Focus (types Ripley, Taylor, and Wilder Engraved) and not of Frankston Focus.

In general, the engraved pottery at the Harroun Site indicates Titus Focus affiliation, the only exception being the one Gibson Aspect bottle fragment. Ripley Engraved is the most common type, but type Taylor Engraved and probably type Wilder Engraved are also present.

Plain Pottery

A total of 260 plain potsherds was recovered from the four mounds and the trenches in the floodplain. Paste of the plain pottery contains varying amounts of sand, and all or most of the sherds are tempered with ground potsherds. Bone tempering is present in 31 plain sherds. Wall thickness varies from 3 mm. for the thinnest body sherds to 13 mm. for some basal sherds.

The surfaces are smoothed and some are highly polished. Sixteen plain sherds are red filmed, seven of them on the exterior surface only and the others on both the inner and outer surfaces. Paste colors are mostly browns and grays, with shades ranging from very light to quite dark.

Carinated bowls, bottles, and probably other vessel shapes are represented. Many of the plain sherds undoubtedly came from vessels which were partially decorated; others probably are from entirely plain vessels. Of the 14 rimsherds, seven are large enough to show that the rims of some vessels were not decorated. No definite types were recognized.

Miscellaneous Ceramic Objects

A perforated pottery disc (Fig. 14, Q) made from a sherd was found at Mound C. It is 32 mm. in diameter, 8 mm. thick, and has a biconically drilled hole 10 mm. in diameter in the center. The outer edge has been partially ground smooth and the two flat sides are fairly well polished. The sherd from which this artifact was made is buff in color, clay tempered, and the paste is fine textured and compact.

A small, conical, ceramic object (Fig. 14, R) was unearthed at Mound B. It appears to be the tip of an appendage that has broken off an effigy vessel or a pipe bowl. It is oval in cross section, and the distal end contracts to a blunt point. The buff-colored paste is fine grained and compact; the surface is poorly smoothed. This object measures 18 mm. long and its maximum diameter at the proximal end is 8 mm.

STONE ARTIFACTS

The 46 lithic artifacts include dart points, arrow points, bifacial blades, worked nodules, pitted stones, and other objects. All the chipped stone implements are made of local quartzites and cherts which occur as very small nodules in the older stream terraces near the Harroun 47 Site. The sandstone and hematite employed for the other stone artifacts were most likely collected from local sources also.

Dart Points

Of the 19 dart points recovered, 15 have contracting stems, 3 have expanding stems, and one has a rectangular stem. Eight of the contracting stem series (Fig. 15, A-D) fall within the shape range of the Gary type (Newell and Krieger, 1949: 164-166 and Fig. 57; Suhm et al., 1954: 430 and Pl. 94), but are smaller (3 to 4 cm. long) than most Gary points reported from other sites. The Gary type has been used as an inclusive group embracing most of the contracting stem dart points of the eastern United States. Several investigators (Ford and Webb, 1956: 52-54 and Fig. 17; Baerreis et al., 1958: 65-69 and Pls. 14-18; Bell, 1958: 28 and Fig. 14) have recognized variants within the broad Gary group, but only a bare beginning toward the definition of the different varieties of Gary has been made.

Three of the Gary points from the Harroun Site (Fig. 15, B-D) are quite similar to a small variety of Gary which seems to be restricted to northeastern Texas. The shoulders are slight and project laterally; the stem and blade are of approximately equal length. Similar points from the Hogge Bridge Site, Wylie Focus, have been illustrated by Stephenson (1952, Fig. 95, A). Many specimens of this variety were also recovered from the Yarbrough Site on the upper Sabine River by The University of Texas in 1940, and others have been reported from sites in the Iron Bridge Reservoir area on the upper Sabine (Johnson, 1957: 7 and Pl. 3, H-L).

Two of the contracting stem points from the Harroun Site (Fig. 15, F) have been assigned to the Wells type (Newell and Krieger, 1949: 167 and Fig. 58; Suhm et al., 1954: 488 and Pl. 123). They feature long, narrow stems which are rounded off at the base and the stem edges are ground smooth. One specimen is virtually complete except for a small portion of the tip. This point has narrow shoulders and a blade with slightly convex edges. The second Wells point is represented only by the stem, but it was probably attached to a blade similar to that of the more complete specimen.

Four of the contracting stem dart points (Fig. 15, J-M) are not assignable with certainty to any recognized type. All are relatively small for dart points. One (Fig. 15, J) is slender and shoulderless; the stem area 48 is somewhat reminiscent of the Wells type. The other three are vaguely suggestive of the Gary type, but are too aberrant to be identified affirmatively with that or any other type.

The other contracting stem point (Fig. 15, N) has a concave base, basal thinning, and ground stem edges. At first glance it reminds one of the Plainview type (Krieger, 1947; Suhm et al., 1954: 472 and Pl. 116). However, a drastic expansion just above the base is characteristic of the San Patrice type (Webb, 1946: 13-15 and Pl. 1) and we are confident that this specimen is a San Patrice point.

One of the expanding stem dart points (Fig. 15, H) has a triangular blade, slight shoulders, and a fairly large stem with smoothed edges. This point is similar to the Trinity type (Suhm et al., 1954: 484-486 and Pl. 82) but is also somewhat reminiscent of type Yarbrough (Ibid.: 492 and Pl. 125).

Another point (Fig. 15, E) of the expanding stem series has been assigned to the Ellis type (Newell and Krieger, 1949: 166-167 and Fig. 58; Suhm et al., 1954: 420-422 and Pl. 89).

The third expanding stem dart point (Fig. 15, I) is the crudest of the series. The stem is relatively small and the basal portion is missing. It falls in the general range of the Palmillas type (Suhm et al., 1954: 462 and Pl. 110).

The dart point with a rectangular stem (Fig. 15, G) is easily the largest projectile point found at the site. The triangular blade has mildly convex edges, and the moderate sized shoulders are slightly barbed. We are reluctant to identify this specimen with any specific type, but in general style it is suggestive of the Bulverde type (Suhm et al., 1954: 404 and Pl. 81). Extreme varieties of the Yarbrough and Morrill types also approach the form of this specimen.

Arrow Points

Only six arrow points were found, including the one associated with Burial No. 1. The burial point (Fig. 15, O) is of the Perdiz type (Suhm et al., 1954: 504 and Pl. 131). It has a relatively short pointed stem and sharp barbs.

Of the remaining five arrow points, three (Fig. 15, P-R) have contracting stems and are of the Perdiz type; the other two (Fig. 15, S-T) have expanding stems and could not be identified with any known type. The three Perdiz points are almost identical in form and are remarkably 49 uniform in size, all falling between 18 and 19 mm. long by 11 to 12 mm. wide at the shoulder. One of the expanding stem arrow points (Fig. 15, S) is in the same size range as these three Perdiz points, the other is somewhat larger. All of the arrow points except the one from the burial have serrated blade edges.

Bifacial Blades

The two bifacial blades could have been used as small knives, scrapers, or even projectile points. One (Fig. 16, E), represented by the basal portion, is a triangular blade with a straight base. It is 3.6 cm. wide at the base and is estimated to have been approximately 7 cm. long when complete. It is fairly thin and of reasonably good workmanship. The second bifacial blade (Fig. 16, F) is smaller than the other, measuring 4.2 cm. long by 2.8 cm. wide at the base. It is crudely pointed at the distal end and has a convex base. The blade edges are sinuous and show little evidence of wear.

Worked Nodules

Six small nodules of chert have been worked and show signs of wear along the worked edges (Fig. 16, A-D). All were fashioned from small elongated nodules by chipping a sharp edge at one end of the nodule, leaving the basal end smooth and unworked. They are from 4 to 6.5 cm. long. Two of them (Fig. 16, A-B) are chipped only across one end of the nodule; the others are chipped across one end and down both sides, only the basal end of the nodule being unaltered. Similar artifacts are quite common in sites over most or all of East Texas, but their purpose is unknown.

Drills

An elongated, pointed implement (Fig. 16, G) with the basal portion missing appears to be the shank of a drill. It has been chipped from gray chert. This fragment is 4.3 cm. long and is from 5 to 13 mm. wide. It is triangular in cross section and the distal end is slightly worn along the edges as though from use.

Fragmentary Chipped Stone Artifacts

Four fragments of chipped stone implements are too incomplete for accurate description. Some or all of them are probably blade fragments from projectile points or bifacial blades.

Milling Stones

One incomplete milling stone is made of light gray quartzite (Fig. 16, J). It has been pecked around the edges into a broad oval shape and it is smooth from use on both faces. It is 9.8 cm. long, 8.2 cm. wide, and 3.6 cm. thick.

Three small stone fragments smoothed on one face are probably pieces of milling stones, but all are too fragmentary for their original shapes to be determined.

Grooved Stones

An irregular shaped piece of hematite (Fig. 16, I) has several narrow, intersecting grooves running across one face. The grooves are set at apparently random angles. On the opposite face of this fragment is part of a deep, gouged out pit where the red pigment was evidently scraped away for use as paint.

A piece of fine grained sandstone (Fig. 16, H) has a broad U-shaped groove across one face. The groove is 20 mm. wide and 6 mm. deep.

Several small pieces of hematite bearing faint scratches were probably used as sources of pigment.

Pitted Stones

There are four pieces of sandstone and hematite with more or less flat sides that have small, circular pits pecked into them (Fig. 16, K). Three have one pit each, the other has two pits on opposite sides of the stone. The pits are all between 2.5 and 3.0 cm. wide and they vary from 4 to 8 mm. deep.

Miscellaneous Ground Stone Artifacts

Three small pieces of stone are smoothed on one face. One is a cobble measuring 17.7 cm. long, 5.8 cm. wide, and 3.3 cm. thick. The others are too fragmentary for reconstruction, but seem to be pieces of small grinding slabs.

PROVENIENCE OF THE ARTIFACTS

The provenience of the artifacts at the Harroun Site is summarized in Table 1. It is clear that the artifacts associated with each house, with the fill of each mound, and with the upper part of the floodplain deposits are quite similar, in the main, throughout the site. Or put another 51 way, each major type or category of artifacts is more or less evenly distributed over the site. This supports the conclusion that the burial, the four houses, the four mounds, and most of the artifacts in the upper part of the floodplain are associated with a single occupation of the site by one cultural group. Architectural and structural data from the mounds point toward the same conclusion.

The only apparent variation from the general provenience pattern is the occurrence of all 17 of the Bullard Brushed sherds at Mound C. However, only two vessels are represented by the Bullard sherds, and because of the small sample it is probably of no particular significance that they all were found at one mound.

Some of the projectile point types may have derived exclusively from a light pre-mound occupation of Archaic affiliation. But the Gary and Perdiz types are unquestionably associated with the mounds and the houses. The Coles Creek Incised (?) and Beldeau Incised (?) sherds may pre-date the mounds.

Table 1
Provenience of the Artifacts

Summary and Discussion

Excavations at the Harroun Site in Upshur County, Texas, revealed abundant evidence of a Fulton Aspect occupation related to four small mounds on the floodplain of Cypress Creek. An earlier pre-mound occupation was indicated by the presence of a few scattered artifacts and stone chips buried as deeply as four feet below the surface of the floodplain. Remains of the pre-mound occupation are very sparse, however, and it is not possible to make an accurate statement of its character. The predominance of stone chips and crude stone artifacts suggests Archaic affiliation, but Fulton Aspect sherds also occurred well down in the floodplain and no pure Archaic zones were found.

The internal structure of each of the four mounds was determined in some detail. Beneath Mound A, the smallest one, was an extended burial of an adolescent male. Offerings associated with the burial were a Perdiz arrow point, a small carinated bowl, and a bottle with an expanding neck. Both vessels are of the Ripley Engraved type. The grave had been dug from the bottom of a broad, shallow pit excavated in the surface of the floodplain; the mound had then been erected over the grave.

A prepared clay hearth in the middle of the mound fill indicated that Mound A had been built in two stages. However, the uneroded condition of the hearth and the absence of a discernible break between the upper and lower portions of the mound disallow the possibility of an appreciable lapse of time between the two construction stages. Since the mound fill contained a quantity of cultural refuse, it must have been taken from a nearby area of fairly heavy occupation. The floodplain near Mound A was tested by means of trenches and small pits, but the assumed occupation area was not discovered.

Mounds B, C, and D each contained evidence of at least one circular house structure which had been burned and then mounded over with 55 sand. Because of the consistent pattern of burning, paucity of domestic artifacts, and burial of the house ruins beneath mounds, it is believed that the structures were ceremonial in function and that the burning was intentional. The few artifacts associated with the house structures indicate that they all were built by a single group of people related to the Titus Focus of the Fulton Aspect.

The house at Mound B was 17 feet in diameter. It had an extended entranceway on the southeast side, a centrally located hearth prepared of clay, and several interior roof supports. This house had been built directly on the surface of the floodplain.

There were two houses at Mound C, the smaller one (14 feet in diameter) superimposed over the larger one (18 feet in diameter). Each apparently had a centrally located hearth and one, or possibly both, had an extended entranceway on the west side. Two interior roof support molds were related to the earlier house, but none were found for the later one.

Beneath Mound D was a single house with traces of an interior hearth situated near the center and an extended entranceway on the west side. Instead of being built directly on the surface of the floodplain as were the houses at Mounds B and C, the house at Mound D had been built in a shallow excavated pit.

Underneath each of the interior hearths associated with the houses at Mounds B and C was a relatively large post mold. It is uncertain whether there was a similar post mold at Mound D because the large pothole there had removed the central portion of the house floor, including the hearth. These molds at Mounds B and C apparently mark the locations of center posts which were used as work platforms during construction of the houses and then removed after the houses were completed. Ridges of sand around the perimeters of all the houses seem to have been banked against the exterior walls while the houses were standing. The floodplain between and around the mounds was tested by pitting and trenching, but no occupational features or concentrations of cultural material were found away from the mounds.

Circular houses of the same general architecture as those at the Harroun Site are typical of the Caddoan Area, especially during the Fulton Aspect period (Webb, 1940; Harrington, 1920; Newell and Krieger, 1949; Goldschmidt, 1935; Davis, 1958). Harrington (1920) reported 56 several circular houses with extended entranceways found beneath sand mounds one to three feet high in southwestern Arkansas. Some of these houses had been built on the surface of the ground, some had been built in shallow pits, and others had been placed on low mounds. Most of them had been burned, and all were associated with typically Caddoan artifacts and with burned clay daub. Harrington thought the houses were earth lodges which had burned and collapsed, the earth from the walls and roofs falling over the house floors so as to form mounds. Webb (1940) reported architecturally similar houses at the Belcher Site in northwestern Louisiana, but presented a strong argument that they were wattle-and-daub houses and not earth lodges.

It appears certain that the Harroun houses were also wattle-and-daub structures without any covering of earth. This conclusion is based on the following points:

1. The bodies of the mounds were composed of soft sand entirely unsuited for covering the sides and roofs of houses. It is doubtful if sand of this consistency would stick to a vertical or steeply sloping wall at all; but even if it did, it would surely be washed away with the first heavy rain.

2. The central portions of the Harroun mounds stood from two to three feet above the floors of the houses. If all this sand had fallen in from the tops of earth lodges, then the lodges must originally have had sand piled at least two or three feet thick on the middle of their roofs. This does not seem probable.

3. Fragments of burned, wattle-impressed, clay daub at all the Harroun houses indicate that the houses were plastered with clay, presumably on the outside. Burned clay daub apparently does not occur archeologically in association with true earth lodges in the plains.

4. Remains of true earth lodges in the Plains area show superficially as depressions, often with a low ring-shaped mound around the perimeter (Wedel, 1936: 24; Lehmer, 1954). Sometimes the depressions result in part from the shallow pits in which the lodges were built. But even when an earth lodge was built directly on a flat surface rather than in a pit, the mound left behind when the lodge collapsed has a concavity in the center instead of being convex as were the mounds at the Harroun Site. It is significant that Mound D was prominent and convex in shape even though the house it covered had been built in a pit. Certainly it is 57 difficult to visualize an earth lodge—whether built in a pit or not—collapsing in such a manner as to produce a smoothly convex mound like those at the Harroun Site.

In view of the foregoing factors, it is concluded that Mounds B, C, and D at the Harroun Site were purposely erected over the ruins of the burned houses.

It appears certain that the four houses at the Harroun Site were typical Caddoan houses. Perhaps they were of the traditional “beehive” shape, or possibly they had wattle-and-daub walls and thatched roofs like those photographed by Soule about 1870 and pictured in Webb (1940: Pl. 8, 1).

Historical descriptions and sketches of Caddoan houses indicate that they did not usually have extended, covered entranceways as do a majority of the prehistoric houses that have been excavated in the Caddoan Area. This suggests that the extended entranceway was used at a relatively early period but was abandoned prior to the 17th century. However, Caddoan houses of the early historic period will have to be excavated before a definite statement can be made in this regard.

Ceramics at the Harroun Site consisted mainly of brushed, incised, engraved, and appliquéd styles, including types Ripley Engraved, Taylor Engraved, Bullard Brushed, Pease Brushed-Incised, and Maydelle Incised. One sherd of Crockett Curvilinear Incised was also found, and two other sherds are similar to the types Coles Creek Incised (or Chase Incised) and Beldeau Incised. Ripley, Taylor, Bullard, and Maydelle are all indigenous types of the Titus Focus (Suhm et al., 1954: 192), Ripley in particular being considered diagnostic of the focus. Beldeau Incised and Coles Creek Incised are Lower Mississippi types and they are surely intrusive in the site. Both were buried between 2.5 and 4.0 feet deep in the floodplain (but at different locations) and they may pre-date the mounds. The sherd of Crockett Curvilinear Incised came from a disturbed area at Mound B, but four sherds from a Gibson Aspect engraved bottle are anomalies that are apparently associated with the mound period at Harroun.

The most common type in the small sample of arrow points is Perdiz, generally considered to equate in time (in East Texas) with the Fulton Aspect, but usually thought of as a trait of the Frankston Focus—not the Titus Focus. Dart point types Cary, Ellis, and Wells—all found at 58 the Harroun Site—are widely distributed in East Texas, and any or all of these types could be affiliates of the Titus Focus or related complexes, although such associations have not been previously demonstrated. The few miscellaneous stone artifacts are relatively non-distinctive in form.

The Titus Focus has been defined on the basis of data derived almost entirely from burials (Suhm et al., 1954: 191). As pointed out by Davis (1958: 67) there is a possibility that mortuary offerings of pottery, arrow points, and other objects may represent selected items and do not necessarily provide a complete catalog of traits actually used by the Titus Focus people. Trait lists compiled from burial data include the arrow point type Talco and the pottery types Ripley Engraved and Harleton Appliquéd as focus diagnostics. Other types listed as Titus Focus traits are shared with other foci.

Davis (1958: 67-68) has noted that Talco points are reported by local collectors to occur only in burials. If this is so, the absence of Talco points in the occupation zones at Harroun does not necessarily negate Titus Focus affiliation for the site. The Perdiz arrow point associated with the burial, however, does seem out of character for Titus Focus as it has been defined.

By and large, the ceramics at the Harroun Site are typical forms and styles of the Titus Focus. However, the absence of diagnostic pottery type Harleton Appliquéd and the presence of Pease Brushed-Incised are incongruous with previous concepts of the Titus Focus.[2] At the Whelan Site, on Cypress Creek 15 miles below the Harroun Site, Davis (1958) has recently reported a series of superimposed houses within a small mound, associated with an assemblage of artifacts remarkably similar to those at Harroun. Ripley Engraved and Pease Brushed-Incised were both present in significant quantities at Whelan, while Harleton Appliquéd was totally absent. No Talco arrow points were found, but six arrow points with expanding stems and one Perdiz point were recovered. Since more than 15,000 artifacts were collected from the Whelan Site, it adds considerable substance to the inventory of artifacts from Harroun, and virtually eliminates any possibility that the Harroun 59 inventory, because of the smallness of the sample, is not truly representative.

On a low ridge near the edge of the Cypress Creek valley, about a half mile west of the Harroun Site, R. R. Nicholas and E. M. German (personal communication) recently excavated several burials. They reported finding vessels of Ripley Engraved and Pease Brushed-Incised associated in the same graves. This spot may be the location of the main village occupation related to the Harroun mounds; in any event the burials there confirm the association of Titus Focus and Bossier Focus ceramic types found at the Harroun and Whelan Sites.

The traits observed at the Harroun Site indicate affiliation with the Titus Focus, but with the following notable deviations from previous conceptions of the focus:

The following alternative hypotheses were advanced by Davis (1958: 67-68) as possible explanations of the circumstances found at the Whelan Site. They are equally applicable to the Harroun Site.

We believe that the first and second hypotheses are most likely to be the correct ones, with a distinct possibility that a combination of the two may best explain the association of traits found at the Harroun Site.

Conclusions

1. Principal occupation was by Fulton Aspect people closely related to—or identical to—people of the Titus Focus. There is an excellent possibility that this is a relatively early Titus Focus site.

2. The four houses probably were used for ceremonial purposes; ultimately each was “cremated” and buried beneath a mound of sand.

4. If the above conclusions are correct, the following archeological traits may be added to those previously recognized for the Titus Focus:

References Cited

Baerreis, David A., Joan E. Freeman, and James V. Wright, 1958. The Contracting Stem Projectile Point in Eastern Oklahoma. Bull. Okla. Ant. Soc., 6: 61-82.

Bell, Robert E., 1958. Guide to the Identification of American Indian Projectile Points. Special Bull., Okla. Ant. Soc., No. 1.

Blair, Frank W., 1950. The Biotic Provinces of Texas. Texas Journal of Science, 2, No. 1: 93-117.

Davis, E. Mott, 1958. The Whelan Site, a Late Caddoan Component in the Ferrell’s Bridge Reservoir, Northeastern Texas. Unpublished report to the National Park Service, on file at the Regional Office of the National Park Service, Santa Fe, New Mexico, and at the Department of Anthropology, University of Texas.

Ford, James A., 1951. Greenhouse: a Troyville-Coles Creek Period Site in Avoyelles Parish, Louisiana. Ant. Papers, Amer. Mus. Nat. Hist., 44, pt. 1.

Ford, James A. and Clarence H. Webb, 1956. Poverty Point, a Late Archaic Site in Louisiana. Ant. Papers, Amer. Mus. Nat. Hist., 46, pt. 1.

Goldschmidt, Walter R., 1935. A Report on the Archeology of Titus County. Bull. Texas Arch. and Paleo. Soc., 7: 89-99.

Harrington, M. R., 1920. Certain Caddo Sites in Arkansas. Mus. Amer. Ind., Heye Foundation, Misc. Series, No. 10.

Johnson, Leroy, Jr., 1957. Appraisal of the Archeological Resources of Iron Bridge Reservoir, Hunt, Rains, and Van Zandt Counties, Texas. Mimeographed report of the National Park Service.

Krieger, Alex D., 1947. Artifacts from the Plainview Bison Bed. In “Fossil Bison and Associated Artifacts from Plainview, Texas,” Bull. Geol. Soc. Amer., 58: 927-954.

Lehmer, Donald J., 1954. Archeological Investigations in the Oahe Dam Area, South Dakota, 1950-51. Smithson. Inst., Bur. Amer. Ethn. Bull. 158.

Newell, H. Perry and Alex D. Krieger, 1949. The George C. Davis Site, Cherokee County, Texas. Memoirs Soc. for Amer. Arch. No. 5.

Sellards, E. H., W. S. Adkins, and F. B. Plummer, 1958. The Geology of Texas, Vol. 1: Stratigraphy. Univ. of Texas Bull. 3232.

Stephenson, Robert L., 1952. The Hogge Bridge Site and the Wylie Focus. Amer. Ant., 17, No. 4: 299-312.

Suhm, Dee Ann, Alex D. Krieger, and Edward B. Jelks, 1954. An Introductory Handbook of Texas Archeology. Bull. Texas Arch. Soc., 25.

Swanton, John R., 1942. Source Material on the History and Ethnology of the Caddo Indians. Smithson. Inst., Bur. Amer. Ethn. Bull. 132.

U. S. Department of Commerce, Weather Bureau, 1958. Climatological Data, Texas, 63, No. 13, Annual Summary for 1958.

Webb, Clarence H., 1940. House Types Among the Caddoan Indians. Bull. Texas Arch. and Paleo. Soc., 12: 49-75.

——, 1946. Two Unusual Types of Chipped Stone Artifact from Northwest Louisiana. Bull. Texas Arch. and Paleo. Soc., 17: 9-17.

——, 1948. Caddoan Prehistory: the Bossier Focus. Bull. Texas Arch. and Paleo. Soc., 19: 100-147.

Wedel, Waldo H., 1936. An Introduction to Pawnee Archeology. Smithson. Inst., Bur. Amer. Ethn. Bull. 112.

Fig. 11. A, Mound A prior to excavation, view looking southwest; B, Mound C prior to excavation, view looking northeast. Mounds B and D were of approximately the same size and shape as Mound C.

Fig. 12. A, Burial No. 1, Mound A, looking northwest; B and C, pottery vessels of type Ripley Engraved, associated with Burial No. 1.

Fig. 13. A and B, sherds of type Bollard Brushed; C-G, sherds of type Pease Brushed-Incised; H, sherd of Coles Creek Incised or Chase Incised (?); I, sherd of Beldeau Incised (?). Profile exteriors are to the left.

Fig. 14. A and B, brushed rimsherds; C and D, brushed body sherds; E, sherd of type Crockett Curvilinear Incised (exterior of profile to the left); F, sherd of type Maydelle Incised (?); G, incised body sherd; H and I, punctated sherds; J, appliquéd sherd; K-M, sherds of type Ripley Engraved; N-P, sherds of type Taylor Engraved; Q, perforated disc made from sherd; R, fragment of appendage from pipe or effigy vessel.

Fig. 15. Projectile points. A-D, Gary dart points; E, Ellis dart point; F, Wells dart point; G, rectangular stem dart point; H, Trinity (?) dart point; I, Palmillas dart point; J-M, Unidentified contracting stem dart points; N, San Patrice dart point; O-R, Perdiz arrow points (specimen O was associated with Burial No. 1); S and T, expanding stem arrow points.

Fig. 16. Stone artifacts. A-D, worked nodules; E and F, bifacial blades; G, drill; H and I, grooved stones; J, mano; K, pitted stone.

FOOTNOTES

[1]In the preliminary sorting an effort was made to disregard all previously described types insofar as possible. It is believed by the writers that more realistic results can be obtained if the artifacts from a specific site are compared and grouped on a basis of their own characteristics, and not on a basis of preconceived forms, styles, and types recognized at other sites.

[2]The anomalous presence of two Lower Mississippi sherds and five Gibson Aspect sherds at the Harroun Site must be considered intrusive, although a reasonable hypothesis to explain their occurrence in a site of Fulton Aspect date does not, it must be confessed, come readily to mind.

		A1	A2	A3	A4	A5	A6	A7	B1	B2	B3	B4	C1	C2	C3	C4	C5	D1	D2	D3	D4	F1	F2	F3	T

Ripley Engraved Vessels					2																				2
Potsherds:
	Miscellaneous Brushed	11	10	24		4	1	4	5	4	2	8	—	3	14	10	12		2				1	3	118
	Miscellaneous Incised	4	2	5		2	1			4	—	1	1	—	1	1									22
	Miscellaneous Appliquéd	—		—						1	1	1		1	1	—									5
	Miscellaneous Punctated	2	—	3		—	—	—				1		1	—	1	—	—	—					—	8
	Miscellaneous Engraved	9	4	16		5	1	2	2	4	3	5		5	12	8	6	3	5		—		—	2	92
	Plain	27	31	64		6	1	10	3	16	6	32		19	14	2	8	4	3		3		6	5	260
	Bullard Brushed	—		—							—			1	1	1	14		—					—	17
	Pease Brushed—Incised	2		3							5	—		2	2	1			1					1	17
	Crockett Curv.—Incised											1												—	1
	Coles Cr. or Chase Incised	—												—										1	1
	Maydelle Incised (?)	1		1										1											3

	Ripley Engraved	2	1	2				—		—		—			1	1	1	—	—	—		—	—		8
	Taylor Engraved							—	—	1	—	1				1		—	—		—	—			3
	Gibson Aspect, Engraved							—	—		2	2		—		—	—	—	—	—	—		—		4
	Beldeau Incised (?)	—	—					—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	1	—	1
Perforated Ceramic Disc		—	—			—	—	—	—	—		—	—	—	—	—	1	—	—	—	—	—	—	—	1
Conical Ceramic Object								—		1								—	—		—				1
Dart Points:
	Gary		2				—	—	1	—					1	1			1	1			1	—	8
	Wells							—		1									—	—				1	2
	San Patrice	—						—		1									—		—				1
	Trinity (?)	1						—	—						—					—					1
	Ellis						—	—	1										—	—					1
	Palmillas (?)	—								—		—		—	—		—	—	—	1	—	—	—		1
	Unident. Contract. stem	1							—	1		1						1							4
	Rectangular Stem						—		1							—				—		—			1
Arrow Points:
	Perdiz	1		—	1			—			—	1										1			4
	Unident. expanding stem			1			—				1		—						—						2
Bifacial Blades		—					—	—	—		—		1			—		—	1				—	—	2
Worked Nodules							—	—	1		1							2			—		1	1	6
Chipped Stone Drills								—											—					1	1
Fragmentary chipped
	St. Arts.	1				—		—						1	—	—		—				—		1	3
Milling Stones							—	—				—							—	—	—			1	1
Pitted Stones							—	—				1		—		—	—	—		—				2	3
Grooved Stones		—	—	—	—	—	—	—					—	1		—		1						—	2
Ground Stone Fragments		—	—	—		—	—	—	—	—	—	—	1	1	—	—	—	—	—	—	—	—	—	1	3
TOTALS		62	50	119	3	17	4	16	14	34	21	55	3	36	47	27	42	11	13	2	3	1	10	20	610

THE HARROUN SITE

Foreword

Letter of Transmittal

Letter of Acceptance

Table of Contents

List of Tables and Figures

Introduction

The Site

ENVIRONMENT

SITE DESCRIPTION

GEOLOGICAL CONTEXT

Excavation and Recording Methods

Mound A

STRUCTURE OF MOUND A

OCCUPATIONAL FEATURES

Feature No. 1

Burial No. 1

DISCUSSION

Mound B

STRUCTURE OF MOUND B

OCCUPATIONAL FEATURES

House No. 3

DISCUSSION

Mound C

STRUCTURE OF MOUND C

OCCUPATIONAL FEATURES

House No. 1

House No. 2

DISCUSSION

Mound D

STRUCTURE OF MOUND D

OCCUPATIONAL FEATURES

House No. 4

DISCUSSION

Description of the Artifacts

CERAMICS

Brushed Pottery

Incised Pottery

Appliquéd Pottery

Punctated Pottery

Engraved Pottery

Plain Pottery

Miscellaneous Ceramic Objects

STONE ARTIFACTS

Dart Points

Arrow Points

Bifacial Blades

Worked Nodules

Drills

Fragmentary Chipped Stone Artifacts

Milling Stones

Grooved Stones

Pitted Stones

Miscellaneous Ground Stone Artifacts

PROVENIENCE OF THE ARTIFACTS

Table 1 Provenience of the Artifacts

Summary and Discussion

Conclusions

References Cited

FOOTNOTES

Transcriber’s Notes

Table 1
Provenience of the Artifacts