Contributed by Matthew Maley
"Three Slipper Shelter"
"Here is the first report on the shelter with the preserved plant remains (where the slipper came from).
Actually, the name should be "Three Slipper Shelter", however, I provided the "official" name before I knew there were three
and before I knew they were slippers and not sandals.
The site has been destroyed by artifact hunters who used shovels to dig into the midden.
A massive amount of prehistoric history/understanding was thus destroyed."
TWO SANDAL SHELTER 15CR173 PRELIMINARY REVIEW OF MATERIALS OBTAINED DURING A SALVAGE EXCAVATION OF A ROCK SHELTER SITE A portion of the salvage conducted between August 1995 and February 1996 was sponsored by a Kentucky Heritage Council grant #21-94-90012-B Salvage Conducted by: Matthew P. Maley, MS, CEE This report is dedicated to Charlie Stamper who’s permission and assistance made the Salvage study possible. Harold Plummer, also, was instrumental in gaining permission to visit this site and many others over the 23 years of the study (1984 -2007) P.1 The Two Sandal Shelter Site – 15CR173 is located in a deep canyon that forms an intermittent stream side-branch to the Tygarts Creek in central Carter County. The site may be located at the western edge of the Grahn Quadrangle (US Geological Survey). There are three additional archaeological rockshelter sites on the property and at least one field site. All but the field site have been severely disturbed; however, the northern portion of 15CR173 had only been disturbed to a depth of 21-33 cm +/- (7-12 inches). The shelter is composed of a Mississippian age sandstone overhang with a deep sandy floor. Figure 1. The nearest limestone outcropping is located at least 6 meters (19.5 feet) below the floor of the site. There is a small cave under the north end of 15CR173 where water that percolated through the sandstone drained. Thus, the northern end of the site was very dry and accounted for the exceptional preservation of plant and other cultural remains. The site was first visited on an area reconnaissance designed to locate and map archaeological sites within the Tygarts Creek drainage. In April of 1995, two local residents and the owner were pointing out various known sites and specific geological features of the area. The Two-Sandal Shelter was the last in the series of sites to be visited and refuge from a cold spring rain was sought in the shelter site. During a discussion about the site and items found by two of those in the party, it was observed that preserved plant remains were scattered in back dirt associated with disturbance of the site. The owner indicated that his grandchildren were excavating parts of the site but that evidently others likewise were digging into the site. Permission was obtained from the owner to obtain several liters of the back dirt for analysis and samples were gathered from four locations within the shelter. Two were associated with a fairly deep hole that suggested that undisturbed stratified remains of plant materials might be present. Subsequent examination yielded a large number of burned and unburned nut shells, pieces of wood, egg shell, animal hair, feathers, chert flakes, grasses, leaves, and insects. The owner was contacted and written permission was obtained to visit and undertake a more in-depth and organized review of the deep hole to obtain additional samples. This was accomplished and it was very apparent that the site had the potential to provide a large amount of preserved plant remains (unburned). Likewise, it was apparent that a stratified midden existed adjacent to at least two of the holes dug by relatives of the owner. At that time, the survey of Carter County was functioning under a Heritage Council Grant and a proposal was submitted to alter the grant to undertake a salvage excavation of potentially undisturbed portions of the shelter. This was approved and a formal review began in late summer 1995. Much of the surface of the site had been disturbed and establishing a “surface” point (datum) to begin measurement presented several problems. Thus, a point on the ceiling of the site was established as a datum and accurate measurements were made to the existing floor surface under the datum. A laser was used to mark off specific areas of the site to establish reference points to begin the salvage activity. The northern end of the site was mapped to indicate the position of rocks that had been in place before and during any historic disturbance of the site. These rocks served as additional reference points and areas immediately adjacent to some of the rocks served as the first areas for review. Figure 2. Assessments were made to determine the depth of historic disturbance of the site. It was determined that in most areas the depth of disturbance ranged between about 21 to 42 cm. (8-17inches). The disturbed soil was removed and 100% was passed through a 6-7 mm. (0.25 inch) sieve. After small rock fragments were removed the remaining material on the sieve was bagged and labeled. Initially, each segment (layer) of the excavation was in 5 cm. (1.9 inches) units. However, when undisturbed midden was encountered the units/segments were reduced to 2 cm. (0.8-0.9 inches). It was difficult to establish standard 1-meter (39 inch) grids between the rocks because the space between them was too narrow. However, a pit was excavated which had a length of 1-meter (39 inches) with adjoining areas maintained to provide a cross sectional view of the stratified cultural remains. The dry sandy characteristic of the site produced considerable dust and it was necessary to use respiratory protection. Also, slumping of site walls was P.2 in some cases a continuous problem. Care was necessary when working adjacent to large rocks to control slumping and subsequent movement of the rocks into the excavated area. The initial or preliminary excavations were conducted around rocks A, B, and C. (Figure 2). However, these areas were for the most part disturbed the deepest. Excavations adjacent to and between rocks C and D provided the deepest undisturbed soils. Also, a good example of an undisturbed stratified midden was apparent under rocks C and D. Figure 3. No excavations were initially conduced under rocks C or D due to the risk of the rocks moving. This was both a safety and midden preservation concern. However, once the excavation around these rocks was completed, some of the midden under the rocks was sampled. Eventually, rock D did move but in a direction that did not hinder the salvage. Rock E was part of the northern wall of the site and it was possible to excavate under this rock to determine how far cultural remains were distributed. It was interesting to note that the sands under E were generally yellow and suggested “sterile” soil, however, a few chert flakes and large pieces of bone were observed back to the actual shelter rock face. During the review adjacent to and between D and E, the relatives of the owner began digging along E in a western-NW direction and dumped the back dirt onto the study area. Fortunately, a barrier had been placed over the un-excavated area and it was possible to continue the review. However, their digging by rock D caused it to move and drop into part of the previously studied area. They did not dig under D so it was possible to salvage some of the undisturbed midden under D. It was obvious that all of the rocks, at the northern end of the site, had fallen at various times after the site had begun to be occupied. This preserved much of the midden and permitted a review of segments of the site that would have eventually been lost to the digging of the relatives of the owner. Very few “diagnostic” projectile points were obtained in the salvage areas; but scattered pieces of pottery suggested a possible Early Woodland association. Also, two ovate base points were recovered. One had evidently been exposed to high heat and was damaged, however, the other was in pristine condition and had a blade of grass (bluestem?) wrapped around the base. Figure 4. The point and the intact grass were sent to OSU for review and preservation. The disturbed soils that had been reviewed prior to excavation of the undisturbed midden yielded several broken Late Woodland projectile points. Also, shell tempered pottery was observed in the context of these blades. The specific origin of these items within the shelter could not be determined but it provided a suggestion of a temporal association of occupation. ITEMS FOUND BY THE OWNER OF THE SHELTER Charlie Stamper had farmed the property for much of his life and had collected artifacts from the field site and three of the shelters on the property. He and local friends excavated 15CR173 between 1968 and 1971. The artifacts collected from the northeastern and southern areas of the site suggested an occupation from the Early Archaic to the Late Woodland Periods. Plant fiber items were collected from at least two parts of the shelter. In one case, a bag containing hickory nuts was excavated from a narrow space between a large rock that had fallen and split. In the north-western part of the shelter Charlie described finding a thick bed of chestnut leaves that was about 2.0 meters (6.5 feet) long and 1.5 meters (4.8 feet) wide and at least 15-16 cm. (6 inches) thick. At the one end of the leaf bed there was a mummified dog and on one side of the bed were two twined slippers (Figure 5)[the identification of the site as Two Sandal is related to these but they are slippers and not sandals]. P.3 Also, he found a number of bone awls and punches, animal jaws and teeth (including beaver) and in 1971 he and three local friends excavated a human skeleton located in a naturally occurring “crypt” in the central area of the site. He reported that the individual was about 6 feet tall (1.8 meters) and was a male (determined by the coroner). He had called the sheriff and it was reported that the skeleton had been taken to the University of Kentucky; however, there is no record of the skeleton from the site. Those who found the skeleton reported that it had been covered with bark pieces and a slab of wood. Also, the details related to the few artifacts found under the skeleton suggest a possible Late Woodland Period association but this could not be confirmed. It is interesting that three other skeletons were reported to have been excavated within the proximity of 15CR173, with one additional coming from the Stamper farm (15CR171) and two from a shelter on another areafarm (15CR241). Several individuals in Carter County have heard of and believed the circulating legends about their being silver and gold buried in rock shelters or caves in the county. Thus, Charlie and friends spent considerable time digging for the treasure and, at the same time, digging for and finding prehistoric Native American artifacts. The slippers were purchased from Charlie and one was donated to a Kentucky museum. However, an arrangement was made with the Kentucky Heritage Council to have the remainder of the salvage grant funds transferred to OSU and the slipper was sent to Dr. Kristen Gremillion at OSU. Subsequently she used the grant funds to have the slipper dated, which resulted in a suggested date of 2580 +/- 60 B.P.: 630 B.C.) (Beta-92918). (Gremillion 1997). The calibrated age is 790 B.C., with a two-sigma age range 825-525 B.C. (Vogel et al. 1993; Talma and Vogel 1993; Stuiver et al. 1993; calibration provided by Beta Analytic, Inc.) This date coincides with the observation of Early Woodland pottery fragments and points. RESULTS OF MATERIAL ANALYSIS Plant The focus of the salvage was the recovery of preserved plant remains related to prehistoric Native American occupation of the site. To this end, over 200 liters of material was recovered. Much of this was associated with the KHC grant but some was recovered before the grant was issued and after the grant had been transferred to OSU. A limited analysis has been conducted of samples from different undisturbed areas as well as disturbed areas and it is obvious that the site was a focus point for the use of and processing of native plant remains. The largest amount of material observed in all areas reviewed was nutshells. Most of the nutshells were not burned but were sufficiently preserved to permit identity of the genus and species of the plant from which they originated. Also, plant fiber items were observed during the excavation, which included a braided cord and rather decayed segments of a woven slipper .Figure 6. A few modified sticks had been found by the owner and additionally some were found during the salvage excavation. The specific purpose of these items, however, is left to speculation. Table 1. Animal A considerable amount of burned and unburned bone was observed throughout the site and within the salvaged midden. Some of the bones had been modified to produce awls and punches. Deer antler, also, was recovered and one tine may have been used as a flaking tool. A specialist has not reviewed the bone but it is possible to identify some of the animal species represented. Additionally, egg shell fragments, mussel shells, hair, turtle shell, and fish parts were recovered. P.4 Lithics Chert flakes and chunks were scattered over the entire site due to the extensive disturbance of the site. Literally, thousands of flakes were found and a limited analysis of the flakes has suggested that chert had been obtained from outcrops at 15CR172 which is located in the central area of a very large shelter 0.1 km (300 feet) to the south of 15CR173. Also, types commonly found along the James Branch Creek 0.4-km (1200 feet) to the east represented much of the chert that was examined. Some of the chert in the shelter had been exposed to high heat and it was difficult to determine the source due to color and texture changes. Table 2 provides a summary of an analysis of chert flakes that were found in the excavations between rocks C & D and F & E. Table 3 provides an analysis of the source of chert flakes relative to areas presented in Table 2. Three projectile points were found within undisturbed midden strata. One, as mentioned above, one was an ovate base point wrapped with a single blade of grass. The other ovate base point was found adjacent to a hearth and was damaged by exposure to high heat. The third point may be a Bakers Creek and it was found at a 2.0 cm. (0.7 inches) under the central area of rock F (see Figure 2). Additional broken points were observed in previously disturbed areas where the leaf bed had reportedly been located. All of these were associated with the Late Woodland Period with several triangular and one Jack’s Reef point being represented. The ovate base point with the grass wrapped around the base (see Figure 4) presented an interesting opportunity for speculation as to the reason the grass was present. The first seemingly obvious reason was that it was used as a pendant, however, attempts to duplicate the way the grass was wrapped using a similar ovate base point indicated that this was not very practical. The second speculation, however, seemed more plausible. What if the grass was wrapped around the base prior to hafting to more efficiently secure the point to the stick? To test this speculation, two identical ovate base points were obtained from a local collector. A blade of blue stem grass was wrapped around one point and both were hafted on identical sticks. Each was bound to the stick with deer sinew and allowed to sit for several days. Then the end of each point was grasped and pushed from side to side to determine the tightness of the fit. The point with the grass seemed to be much more secure. A presentation related to 15CR173 was given at an ASK meeting at NKY (ASK Meeting, October 10-11, 1997, NKU) and the two hafted points were passed among the 20 individuals in the room. They were asked to indicate which point felt more secure. All those present selected the point that was hafted with the grass wrapped around the base. Obviously, this does not prove the specific speculated use but it does provide a strong suggestion of possible application. A number of whole and broken chert hammer stones were observed. No glacial deposit material was observed in any of the lithic assemblage. Sandstone, which is present in the shelter, had been used for grinding/milling stones and “nutting” stones. Limestone, also, was utilized. One rounded, oblong stone was battered on one end suggesting that it was utilized for some form of pounding. However, the most interesting limestone utilization was for a possible cooking surface. In at least three segments of the salvaged area, slabs of limestone were found that had the same characteristics. They were generally 35.5-40.5 cm. (14-16 inches) wide, 45-64 cm. (18-25 inches) long, and 6-8 cm (2.5-3.0 inches) thick. One surface (top) was polished and discolored by exposure to heat. The outer edges of each stone were rounded suggesting that they may have been obtained in a stream or weathered outcrop. As the result of exposure to high temperature some were cracked into various shaped small blocks which could easily be pieced together. In one case, there was evidence of a hearth being present directly on top of the slab (Figure 7, 7a), however, in the others wood charcoal and ash was generally found around the edges of the stone. It is speculated that a shallow pit was dug into the sand floor of the shelter, the limestone slab was laid at the bottom of this hole and a fire was built on top of the slab. After the fire had burned to coals, the coals and ashes were moved to the edges of the slab and food was placed onto the slab similar to using a griddle. The “polished” top surface of the limestone slabs suggested that the stone could easily have served as a griddle. P.5 To test this speculation, an equivalent size slab of Ordovician limestone was obtained, a hole dug, and 65 pounds of maple/walnut logs were burned until they were ashes with a few coals. The maple/walnut was used to match the wood charcoal found in one of the excavated hearths. The ashes were pushed to the side and two eggs, three slices of bacon, and two pieces of bread/toast were easily cooked on the surface. The eggs were gritty but could be eaten (the dog enjoyed them). A hole had been drilled into the slab to permit the insertion of a very high temperature mercury thermometer. At the end of the heating cycle the stone had reached 214 C. (417F.). The temperature remained above 93 C. (200 F.) for over 120 minutes. An egg will coagulate/cook at 60.5 C. (141 F.), thus, the temperature when the ashes were removed was quite sufficient for cooking. (Figure.8) Other Cultural Remains A few shards of thick, unmarked, limestone tempered pottery were found in both the stratified midden and in the back dirt of prior excavations. Cord marked, shell tempered potter was observed in the same area where the Late Woodland Period points were found in back dirt. Conversations with the owner of the shelter and two others who had dug in the site suggested that pottery was sparse and no vessels were observed. (Figure 9.) A “blob” of clay that somehow had been fired was found in the back dirt associated with the triangular points and it had what looked to be the impression of the bottom of a slipper. It looked like someone had stepped on the clay and it was then exposed to a fire. A black tarry residue was found in several areas of the northern section of the site. The material dissolved in hot water and testing suggested it was probably a residue of tannin /gallotannin. Because there was a considerable amount of the material found it was speculated that Native People might have collected the material to use in a tanning process. However, this was dismissed since such material may be obtained when some green wood, which contain high levels of tannins, are burned. A viscous/tarry fluid exudes from the cambium/bark area and will oxidize to a dark brown color after exposure to air. There were a large number of hearths present in the site suggesting that much wood had been burned and the residue may have been associated with some of the wood such as chestnut or oak. TABLE 1 – Plant Remains CULTURAL REMAINS OBTAINED FROM 15CR173 PLANT/COMMON PLANT/SCIENTIFIC PLANT/COMMENT-DETAILS Basswood Tilia americana Modified sticks, braided cord, loose bark fiber Beech Fagus grandifolia "Nuts", leaves, wood Black Walnut Juglans nigra Nuts, nutshells, wood Blackberry/Raspberry Rubus species Seeds Bluestem Grass Andropopgon species Stems, leaves (leaf wrapped around point base) Buckeye Asculus species Capsule/husk, "nut shell" Butternut Walnut Juglans cinerea Nuts, nutshells Cane Arundinaria gigantea Stems, "split" burned Chestnut Castanea dentata Husk, "nut shells", leaves reported as bedding Greenbrier Smilax rotundifolia Stems, seeds? Hazelnut Corylus americana Nutshells Hickory (pignut) Carya cordiformis Husk, nutshells - burned and unburned Hickory (shagbark) Carya ovata Husk, nutshells - burned and unburned Hickory Carya species Nutshell fragments Magnolia Magnolia species Seeds Oak, White Quercus alba Acorn Oak, Scarlet Quercus coccinea Acorn Oak Quercus species Acorn fragments Paw-Paw Asimina triloba Seeds and possible modified sticks Red cedar (Juniper) Juniperus virginiana Wood Squash/gourd Cucurbita species Stem, seed Wild Grape Vitis species Stem, seed Yellow Poplar Liriodendron tulipifera Seeds, possible wood TABLE 2 – Lithic Analysis SITE AREA DIMENSIONS (cm) DEPTH (cm) SAMPLE VOL. (L) # OF FLAKES C/D 70 x 100 260-263* 1.5 54 F/E 50 x 100 251-255 0.5 40 251-255 0.5 16 251-255 1.5 92 251-255 2.0 79 255-259 1.2 23 255-259 1.2 36 255-259 1.2 27 255-259 1.2 35 259-266 1.5 23 259-266 1.5 21 259-266 0.75 05 Fire/hearth site 257-269 1.5 83 Totals 16.95 Liters 529 Flakes * Depth is based on measurement from the datum that is on the roof of the shelter. Also, where a similar depth is indicated the sample was divided to better determine the distribution of material across the sample area. A sub-grid had been established within the indicated dimensions of the sample area. TABLE 3: Chert resource analysis AREA DEPTH TOTAL LOCAL OTHER 15CR172 “EXOTIC”** C/D 260-263 83 47 28 0 08 F/E 251-255 227 157 47 18 05 255-259 116 78 30 0 08 259-266 49 22 19 0 08 Fire/Hearth 54 47 02 03 02 Totals 5 areas 529 351 126 21 31 %/Total 66.4 23.8 4.0 5.9 ** “EXOTIC” refers to chert not from the area or form a quarry site that has yet to be determined, or had been altered in color and texture due to exposure to excess heat. . LOCAL and OTHER refers to known quarry sites in Carter County. The total percentage of Carter County associated chert was 94.2%. TABLE 4: Testing of limestone slab TEMPERATURE OF LIMESTONE USED AS COOKING SURFACE Degrees F. Time/min. Degrees C. Time/min Decrease in temp. 417 0 214 0 per unit time (deg.F) Min. 370 15 192 15 47 15 342 30 176 30 28 30 317 45 162 45 25 45 292 60 147 60 25 60 266 75 133 75 26 75 242 90 119 90 24 90 225 105 110 105 17 105 214 120 103 120 14 120 This data represents an experiment to evaluate the potential for a piece of limestone to retain sufficient heat to permit the cooking of food. 60-65 pounds of maple/walnut (hardwood) were burned for a little over two hours. The coals and ashes were then brushed to the side. Two chicken eggs, three slices of bacon, and two pieces of bread were subsequently placed onto the heated limestone surface. Each cooked rapidly and well. The eggs stuck somewhat to the stone but if the bacon had been placed on the stone first this could have been prevented. Also, the eggs contained grit from ash and the stone. The stone reached a maximum temperature of 417 F. (214 C.) with an initial drop of 47 degrees in the first 15 minutes, followed by a mean decrease of 25.6 degrees F. over the next five intervals. Followed by a less rapid decline over the subsequent 30 minutes. The temperature was maintained above the boiling point of water for about 125 minutes. (see Fig. 7 & 7a) References US Geological Survey, Geological Map of the Grahn Quadrangle, Carter County, Kentucky, K.J. England, 1976 Grimillion, K.J., The Research Potential of Textile Artifacts: An Example from Carter County, Kentucky, Kentucky Heritage Council Annual Meeting, Natural Bridge State Park, Kentucky, March 14-16, 1997. Maley, M.P., Plant remains and other Cultural Material tell of the Life and Times of Prehistoric People in Carter County Kentucky, ASK Meeting, October 10-11, 1997, Northern Kentucky University. Figures: Not included due to issues relative to size of document and sending via E-mail.