Tuesday, April 1, 2008

The Kenton Caves of Western Oklahoma

The following text is reproduced under the "fair use" provision of the copyright act and is presented for educational, informational and scholarly use only.

The Kenton Caves of Western Oklahoma
by Christopher Lintz and Leon George Zabawa
from Prehistory of Oklahoma (1984)
edited by Robert E. Bell
Chapter 7, pages 161-174


        The Kenton Caves are seven of many rockshelters located in the northwestern corner of the Oklahoma Panhandle (Figure 7.1; Table 7.1). Expeditions during the late 1920s through early 1940s recovered a large amount of perishable materials because of the dry conditions within the caves. Thus, the region represents one of the few areas in western Oklahoma from which archaeologists have information about both perishable and nonperishable aspects of material culture.

        The caves are located in the Tesesquite Creek Valley, a tributary of the Cimarron River, in the Black Mesa area of Cimarron County. This area is characterized by flat-topped mesas bordering gradually tapering stream valleys 8–10 km in length. The mesas are erosional remnants of sandstone and orthoquartzite de-posits formed during the Mesozoic era (Fenneman 1931:40; Stovall 1943:22). The valleys were created prior to the Pleistocene, but soils in the valley bottoms range from Pleistocene to Recent in age (Wilson 1972). The vegetation is an extension of Rocky Mountain vegetation composed primarily of the pinon–juniper type with scattered trees predominantly confined to the upper slopes and rock outcrops. The floodplains and uplands are mixed grassland savannas (Duck and Fletcher 1943; Bruner 1931; Kapustka 1974; Risser, Risser, and Goodnight 1980). The fauna is that typical of the short grass and pinon–juniper ecotonal area.
        Past environmental conditions in the Tesesquite Creek Valley and adjacent areas may have been somewhat different than the present. Geomorphic studies in Tesesquite Creek Valley have shown that major changes have occurred in vegetation (Wilson 1972). Buried willow, cottonwood, hackberry, juniper, and possibly pine stumps measuring up to 1 m in diameter are indicative of a valley forest. The bottomland forest extinction has been radiocarbon dated at 327 to 621 years ago (circa A.D. 1329 to 1723). These dates only reflect the termination of the bottomland forest, and not the time for development. If several centuries were necessary for the development of the forest, then woodland conditions may have been present when the caves were occupied (Wilson 1972:206-207).

History of Investigations

        Although local ranchers and sheepherders have known about the caves for years, their importance was only recognized through the persistent efforts of William "Uncle Billy" Baker, the panhandle farm agent and respected avocational archaeologist. Baker's son, Ele, is credited with "discovering" Basket-maker Cave No. 1 (34Ci-50) in June 1928. Immediately, Baker contacted Dr. Leslie Spier at the Department of Anthropology, University of Oklahoma (Baker 1929a). Previous commitments, however, prevented the University from initiating fieldwork in the panhandle, so Baker agreed to monitor and preserve the site.
        The 1926–1927 excavations at the Folsom early man site in New Mexico, and the development of a Southwest cultural classification at the 1927 Pecos Conference renewed interest in cultural surveys throughout the Southwest. The Colorado Museum of Natural History launched an extensive archaeological survey program starting in northeastern New Mexico and adjacent areas. By November, 1928, Nelson J. Vaughan, the scout for the Colorado Museum party, reached the Kenton area and made inquiries about Bat Cave (34Ci-69). William Baker met Mr. Vaughan and showed him the newly discovered cave, 34Ci-50, after obtaining assurances that the Colorado Museum party would not disturb it (Baker 1929a).
        However, on July 14, 1929, the Colorado Museum party under the direction of Dr. Etienne B. Renaud quietly conducted a brief survey and limited testing at three rockshelters in Oklahoma. The survey apparently located four rockshelters and five open sites (Renaud 1929d). Bat Cave (34Ci-69) was not tested because its northern orientation was believed to make it unsuitable for habitation. Renaud's excavations at Basketmaker Cave I (34Ci-50) and the Twin Caves (34Ci-39, 34Ci-68) lasted less than a month (Renaud 1929d; Baker 1929a). A series of excavation summaries were quickly released (Renaud 1929a, 1929b, 1929c, 1930b, 1930c), with more detailed descriptions appearing in both English and French (Renaud 1930a, 1930d, 1932).
        Outraged by this breach of agreement, Baker organized a small crew of Boy Scouts in July 1929 and spent 1 day salvaging additional materials from 34Ci-50. Newspaper accounts indicate that Baker was planning to excavate other nearby shelters with deep bedrock grinding basins (Rucker 1929). Although no notes exist, his collection contains materials from the Twin Caves (34Ci-39). At a later date, he was also involved in digging other caves in the Tesesquite Valley as well as several shelters north of Boise City (Baker n.d.; Ele Baker and Bill White, personal communications, 1980).
        Baker was instrumental in publishing newspaper articles concerning the removal of important Indian artifacts from state school land by the Colorado Museum party (Anonymous 1929a, 1929b; Baker 1929a; Rucker 1929). To investigate these charges, Governor Holloway directed Dr. Joseph Thoburn, historian from the Oklahoma Historical Society, to visit and report on the status of the Kenton Caves. After a brief visit in July, 1929, Thoburn sought private funding to continue excavations. Following the 1929 Pecos Conference, Thoburn re-turned to the Kenton area and appointed Baker to direct the official excavations for the State of Oklahoma. The length of the Baker–Thoburn excavations at 34Ci-50 is uncertain; the results, however, were published in detail (Baker 1929b).
        During the winter, Thoburn sought funding for additional excavations under the Cooperative Ethnological and Archaeological Investigations Act (Public Law 248, 70th Congress 1928). The Smithsonian Institution matched a $1000 fund established by the Oklahoma Historical Society. With a $2000 budget, Thoburn spent the 3 summer months of 1930 directing excavations at Basketmaker Cave I (34Ci-50), Basketmaker Cave II (34Ci-49), and perhaps Pigeon Cave (34Ci-48). Unfortunately, little has been published about this lengthiest of all organized archaeological expeditions, and nothing is known about the work or the materials recovered.
        Professional interest in the Kenton Caves waned after the 1930 excavations. However, the publicity about the cave's contents attracted many local people. A favorite school, church, or community pastime of the 1930s was to picnic at the caves and to search for souvenirs. During the course of such activities, three burials were found in 1933. In May, Mrs. I. W. Mathews discovered a desiccated or naturally mummified child at 34Ci-50 while on a school outing from Keyes; and a short time later, two adults—a mummified man and a partially mummified woman—were unearthed from the same shelter by Merrill McGee, John Tharp, and Pard Collins (Anonymous 1968:10D; Hollis 1979).
        The recovery of mummified remains prompted Forrest Clements from the University of Oklahoma to sponsor a Federal Emergency Relief Administration (FERA) supported field party in June, 1934. Ele Baker directed four other workers in the clearing of 34Ci-50. Because much of the cave had been previously excavated by Renaud, William Baker, Thoburn, and various groups on weekend outings, only a small amount of materials were recovered (Bell 1953c).
        Excavations near Kenton shifted from archaeology to paleontology during the latter half of the 1930s. Under the direction of Dr. J. Willis Stovall, Works Progress Administration (WPA) crews worked on sites in Ellis, Cimarron, and Texas counties, Oklahoma. The Cimarron County crew was under the supervision of Mr. Crompton Tate. After 35 months, the quantity of dinosaur remains began to diminish. In an effort to continue the WPA project, Tate shifted to the excavation of Indian remains from a cave in April 1938. Only a small quantity of prehistoric materials and almost no perishable items were recovered (Tate n.d.). This departure from paleontology was short lived. With much of the cave deposits left intact, Stovall terminated the Cimarron County field crew after less than a month at the cave (Tate 1938; Stovall 1938). Correspondence indicates that Dr. Forrest Clements from the University of Oklahoma was planning additional work at the cave in June, 1938, but apparently the project was never funded (Stovall 1938).
        By late May, 1940, Stovall personally supervised five laborers on a 2-week excavation at 34Ci-70 for the WPA (Stovall 1940). It is not certain if this is the same site excavated by Crompton Tate because there are no records describing the site location. Although the shelter was totally cleared, the results have never been published and all but four items have been misplaced.
        With the outbreak of World War II, major excavations in the Kenton area ceased. Several articles were published describing unique artifacts from the shelters (Baker and Kidder 1937; Renaud 1939), and a number of cultural summaries have been written (Clements 1943; Renaud 1947; Bell and Baerreis 1951; M. A. Ray 1961a; Campbell 1969, 1976; Saunders 1982). Within the last 20 years, several surveys have been conducted in adjacent areas (M. A. Holmes 1972; Campbell 1973; Lintz 1976a; Ludrick 1966; Muto and Saunders 1978; Saunders 1978; Saunders and Saunders 1982; Haury 1982). Limited excavations at two rockshelters and two open sites in North Canyon were conducted in the Fall of 1973 (Nowak and Kranzush n.d.). Recently, the University of Tulsa excavated a trench through the talus midden at Thoburn's Basketmaker Cave II (34Ci-49), and tested the terrace in front of 34Ci-115 near Lake Carl Etling (Don Henry, personal communications 1981).
        The recovered artifactual remains from the Kenton Caves have been scattered. Materials from Renaud's excavations are in the University of Denver Department of Anthropology. William Baker's materials are in the No Man's Land Museum at Panhandle State University, Goodwell, Oklahoma. Some of Thoburn's materials are on display at the Oklahoma Historical Society in Oklahoma City; but the small quantity of remains suggests that much has been loaned or lost. The three mummies were divided among the principle finders. The mummified child was kept for a time under a bed at J. R. Butler's home, in John Tharp's garage, in Pard Collins' house, and even hung on the back of an upstairs door at the Kenton Hotel before being donated to the No Man's Land Museum (Hollis 1979). For over 15 years the mummified woman was kept in the attic storage area at the Kenton Hotel before being moved to a "museum" in Brownie Collins' house at Kenton. During the 1950s, Willis Stovall allegedly took the adult mummies; however, there is no trace of them at the University of Oklahoma. Materials from Ele Baker's 1934 excavations are at the Woolaroc Museum near Bartlesville and the Stovall Museum at the University of Oklahoma. The location of Crompton Tate's materials is unknown, but they may be with the WPA dinosaur materials at the University of Oklahoma. Only four items from Stovall's 1940 excavations are available at the Stovall Museum, University of Oklahoma. The location of the atlatl reported by Baker and Kidder (1937) is unknown. Finally, the recently excavated materials from 34Ci-49 are at the University of Tulsa.


        Little is known about the internal structuring of the cave's deposits. This is due to the general lack of records, the excavation techniques (which were primitive by today's standards), and the changes in excavation methods (which varied from one project to another). To test each of the three caves, Renaud dug four vertical trenches; one along the central cave axis, another across the back wall, and diagonal trenches from the center mouth to each back corner. His artifact recovery techniques are not specified. William Baker used shovels and sifted the soil, but the methods of provenience control are not discussed. Thoburn's recovery techniques are largely unknown, except that he "blasted" the rocks blocking the cave. Unpublished maps by Stovall are the only records of horizontal and vertical location of materials from any of the early excavations.
        The size of the caverns and depths of cultural deposits vary considerably. The best stratigraphic information is from 34Ci-50 (Baker 1929b:18; Bell 1953c). This shelter had 1.7 m of deposits, but artifacts were only found in the upper 1.2 m. Although the relationships of the complex strata are unclear, multiple occupations are indicated in the stratigraphic descriptions. Several of the caves have internal structures reflecting different activity areas. In describing the east cave at 34Ci-68, Renaud (1929d, 1930a:130) mentions the presence of two roughly circular, vertical rock enclosures measuring 2.20 m in each corner of the cave. They may represent storage cists or small rooms (Campbell 1976). The west cave at 34Ci-39 and 34Ci-49 have clusters of bedrock grinding basins, indicating mealing areas. Both Baker (1929b:18) and Renaud (1930a) mention a series of charcoal hearths on the north side of 34Ci-50, separate from layers of grass, sage brush, small twigs, and leaves on the cave's south side, which may have been used as flooring or bedding. Although discrete activity areas are suggested, spatial patterning within the shelters are not well understood from these few passing comments.

Material Remains

        A wide variety of material has been secured from the caves. Lacking clear artifact provenience, it is impossible to delineate material change through time within any shelter, or even to demonstrate that the six shelters were occupied by the same groups. The following description of materials must reflect a mixture of several cultural assemblages.
        Plants provided both food and raw-material resources for a variety of utilitarian items. Foodstuffs include both domesticated products and wild floral resources. Cultivated plants include ears and shelled kernels from several varieties of corn, and stems, seeds, and rinds from either squash or pumpkin gourds (Renaud 1930a; Baker 1929b; Thoburn 1930). Analysis of corn indicates that both small and large ears of 8-, 10-, 12-, and 14-row varieties are present (Cutler 1959; J. C. Winters 1975) (Figure 7.2 a–g). A comparison of the corn varieties indicates that the "temporal provenience could range anywhere from Basketmaker III (A.D. 300) through late Prehistoric Plains, 'Puebloid,' Apache or related cultures" (Winters 1975:3). Several specimens of "early" pod corn have also been recovered (Anonymous n.d.). The pod corn resembles the earliest corn from Bat Cave in New Mexico; but this variety could also be contemporaneous with the other corn varieties (Paul Minnis, personal communications 1982). Thoburn (1930) reports the recovery of beans; if the identification is accurate, however, beans are very rare. These cultigens were supplimented by acorns, hackberry, cactus, sand plum, crocus bulbs, and other wild fruits and berries (Renaud 1930a). Often the cultigens were recovered in paired woven-grass or prairie-dogskin bags. These bags and their contents may either represent one means of storing garden seed by suspending it away from rodents, or reflect the means of transport. Some idea of food preparation is also indicated. Corn cobs were found encased in baked clay, and three circular cakes made of acorn and wild plum or cherry were found at 34Ci-50 (Baker 1929a; Renaud 1930). The cakes measured 10–15 cm in diameter and less than 5 cm thick. A central hole through each cake may represent a suspension hole for cooking or storage. Corncobs impaled on short sticks may have served as bait for snaring small game.
        Other plant remains were used to make a variety of items. Clothing and ornaments made from plants include several varieties of square- and round-toe sandals made of yucca leaves and padded with bark; a piece of twined fabric representing either clothing or a bag; and chinaberry-seed beads (Renaud 1930a). Although sandals are most common, mats and baskets are also known. The few examples were made of yucca or bark by four methods: coiling, simple plaiting, twilled plaiting, and single-strand twining. Small bundles of grass within loose netting formed "bags" for storing seeds. Bark and cattail leaves were occasionally used to make string, but yucca-leaf cordage served as the common binding material. The majority of cordage is two ply, but some three-ply and two-strand, two-ply cords are known. The recovery of knotted yucca and grass shows that the half hitch, double half hitch, square wreath, and a number of other knots and loops were known. True textile fabric is rare and is only represented by the previously mentioned clothing article. No nets have been found.
        Game was obtained by the curved nonreturnable "boomerang" throwing stick, the atlatl (spear thrower), and the long bow. No snares or traps have been conclusively identified. Slotted and unslotted hardwood foreshafts reflect the use of compound darts, and the nock portion of at least one arrow shaft has been found, in addition to a wooden knobbed bunt.
        Wood was also used in friction fire-drill basins and in a wide range of "crochet needles," hairpins, pegs, and other pointed sticks. Sticks with flattened blades may have been used as weaving battons, or in making baskets. Some sticks were used as pine pitch applicators. Several wooden tubes and wooden beads were also found. Many worked and unworked sticks were found with knotted yucca leaves wrapped around them, as well as small wooden shafts wrapped with shredded vegetable fibers.
        The uses of tied bundles of grass, twigs, and sagebrush are uncertain. Some may reflect brushes or brooms, whereas others may merely reflect the means of transporting and storing raw materials for subsequent use. Some thick wads of knotted corn leaves, tied with yucca, show extensive wear as if they were used as general cleaning or scouring pads. The use of pine cones is uncertain. Finally, the layers of grass within the caves may have served as beds, or padded activity areas.
        As with floral resources, faunal materials were well preserved. Pieces of fresh-water mussel shell and a crayfish claw have been recovered, along with the bones of bison, deer, antelope, elk, jackrabbit, cottontail rabbit, coyote, wildcat, badger, rodents, turtles, eagles, and turkeys (Renaud 1930a). Thus, a wide exploitation pattern involving a number of different settings is indicated.
Besides food, fresh-water mussel shells were perforated and shaped into pendants, or notched along the edge and used as scrapers and fleshers. Other ovate and circular pieces of ground mussel-shell may be bead or pendant blanks, or shell inlays. Large marine-shell disc beads and Olivella shell beads indicate that the people were part of an exchange system extending to the ocean.
        Land animals were used in a wide variety of ways. Animal skins were fashioned into fur-lined moccasins and bags. The bags were made from prairie-dog skins complete with fur, legs, tail, and head, with cordage for a carrying strap. Leather pouches were made with and without the hair from deer, antelope, buffalo, or prairie-dog hides and used to store corn and gourd seeds. Skins were also cut into strips and braided together into furry thongs, or incorporated with yucca cordage. A small piece of a rabbit-fur blanket was recovered. A deer-hoof pendant tinkler was found on a thong with several bird-bone beads. The function of a feather bundle tied with yucca fiber is unclear. The bone technology is limited. Mammal leg-bones were split and sharpened into several varieties of awls that may have been used to manufacture basketry and leather goods. A perforated bone "shaft wrench" was also recovered. Bird bones were commonly used to make tubes and beads.
        Pottery was not found by most of the expeditions. Newspaper accounts, however, mention prehistoric ceramics (Rucker 1929) and nearly 40 sherds were listed by Crompton Tate (n.d.) on the shipping roster. The Tate sherd sample has been lost and no descriptions are available. The William Baker collection from 34Ci-50 contains one cordmarked sherd with angular quartz temper, six corn cob or cordmarked sherds with fiber tempering, and three black plainware sherds with fine black-sand temper. Two fragments of an unusual black-sandtempered clay pipe are also present. The pipe has a groove with punctations around the outside rim of the bowl, a broad smooth bowl, and at least three encircling grooves with punctates below the bowl. It is uncertain whether the specimen is an elbow or straight pipe form.
        Local orthoquartzite was used to make a variety of chipped- and ground-stone tools. Primarily Tesesquite and Dakota quartzites were employed to make chipped-stone tools; small quantities, however, of opalite and jasper were also used (Lopez and Saunders 1973:3; Lintz 1976b). The lithic assemblage is rather crude with flake and bifacial knives, drills, borers, gravers, scrapers, choppers, cores, modified flakes, and manufacturing debris often being represented. A few stone arrowpoints and an occasional dart point were recovered by Tate (n.d.), and William and Ele Baker. The dart points include an unfluted "fish-tail" form attributed to Paleoindian occupations, and the Williams type plus other unidentified kinds of large corner-notched forms. In comparison, the small corner-notched Scallorn, and to a lesser extent the small side-notched Washita and Harrell points and unnotched Fresno points are more abundant than the dart points.
        Ground-stone items include grooved honing stones for sharpening awls and sticks, grooved shaft-smoothers, and oval grinding basins with one-handed manos. Hammerstones and "polishers" were also found, as were anvil stones with vegetal material still adhering to the working surface. One selenite pendant with a yucca string attached through the perforation and a small fragment of a possible black slate gorget were found at 34Ci-50.
        Historic trade-items are very rare. The William Baker collection, however, has a small box of white, green, light blue, dark blue and black glass seed and pony beads from 34Ci-39.
        Although occasional isolated human bones and teeth were found during the sponsored excavations, the best articulated burials were unearthed by amateurs. Nothing is known about the adult male burial. Both the woman and the child were buried in semiflexed positions (Anonymous 1968:10D; Hollis 1979). The woman was buried with a mat or coiled basket over her head (Bell 1953c; Hollis 1979). The child was underneath a heavy rock and was buried with a necklace of 16 Olivella shell beads and a grass bag of corn ears under his legs. Based on stature, the child may be about 2 1/2 years old (Archie Hood, personal communication 1982). He does not appear to be clothed, although close examination shows what may be part of a G-string. The lack or scarcity of clothing suggests that death may have occurred during the spring or summer. The skeletons do not have deformed heads (Bell 1953c).

Rock Art

       Several of the caves have painted or scratched-rock art motifs. A reported polychrome painting at 34Ci-49 was apparently removed by the WPA crews. Most of the identifiable painted figures are anthropomorphic, although the styles vary considerably (Figure 7.3). Some are solid-painted, square-shouldered adults dressed in kilts or mantas in association with naked children. Others are outlines of long eared or horned figures. Other painted motifs include the sun or flower, Najahe crescent, and phi designs. Most of the grooved petroglyphs include a series of tally marks, zigzags, feathers, and X motifs. One exception is the lightly scratched horned anthropomorphic figure holding a shield at the west shelter of the Twin Caves (34Ci-39). In general, rock-art sites are abundant in the Black Mesa region (Burchardt 1958; Campbell 1973; T. Clark 1969; Lawton 1962; Moorehead 1931:117; Renaud 1937). The variations in the stylistic designs suggests that several different traditions are represented.

Interpretation and Conclusions

        Cultural assignment of the Kenton Cave material has been extensively debated. Heavily influenced by the newly constructed Pecos classification and noting some similarities in the perishable assemblage, Renaud (1930a:34) thought the materials represented a stage above the seminomadic hunter and below the sedentary Pueblos. He suggests that the materials were "an early, very primitive phase of the Basketmaker culture, an incipient stage preceding its more complete characterization elsewhere." He suggests a date of circa 1500 B.C. for the Kenton caves based on the presence of basketry, painted figures, and maize, and on the absence of pottery from his 1929 excavations. Similarly, the absence of pottery and small points from the published literature has led M. A. Ray (1961a) to attribute the materials to an Archaic lifestyle. Another summary acknowledges considerable antiquity, but stops short of dating the materials (Bell and Baerreis 1951:17). These earlier summaries were based on materials from one or two expeditions and secondary sources, and have not considered the full range of materials available from the shelters. In general, cultural affiliations are tenuous because there is little stratigraphic information available to segregate the assemblages, and there is a paucity of comparable perishable materials from the Plains region to the east.
        Undoubtedly the caves have considerable antiquity. Stovall (1943:119) found late Pliocene-age horse and camel bones beneath the cultural remains at 34Ci-70. The recovery of an unfluted "fishtail" point from 34Ci-50 may indicate some late Paleoindian or early Archaic use of the shelters. The large corner-notched points, wooden atlatl, and pod-corn may be indicative of later Archaic occupations; these items, however, are rare and could well coincide with subsequent use of the caves. Although projectiles and pottery were not abundant, most of the styles are post-Archaic. The many Scallorn and, to a lesser extent, Washita, Harrell, and Fresno points, along with the long bow, suggest occupations by both Woodland and Plains Village stage groups. In addition, numerous Woodland and Plains Village sites have been recorded in the adjacent Black Mesa State Park on South Carrizo Creek (Saunders 1978), and in the environmentally similar Chaquaqua Plateau in Southeastern Colorado (Campbell 1969; 1976). The recovery of thin, black-sand-tempered plainware pottery and glass beads from the shelters, as well as numerous adjacent pictograph sites depicting mounted Indians, suggests both protohistoric and historic use of the caves. Undoubtedly even sheepherders and cowboys used the caves, although nothing in the material assemblage reflects their use.
        No radiocarbon dates have yet been reported from the six Kenton Caves. One date of A.D. 560 ± 150 (no lab number reported) was obtained from a rockshelter in North Canyon, some 12.8 km northeast of the Tesesquite Valley (Nowak and Kranzush n.d.). The radiocarbon dates obtained by the University of Tulsa from Thoburn's cave 34Ci-49 have yet to be published.
        Research to further our understanding on the Kenton Caves based on firm stratigraphic information may soon be available from the University of Tulsa excavation. The caves, however, are only a small part of a larger settlement system within the Tesesquite Valley. The regional prehistory should also concentrate on studying other site situations within a broader area base. A number of site types in various topographic situations are already known—including stone structures, bison kills, terrace sites, lithic quarries, and tipi rings (Muto and Saunders 1978). These other sites must be studied in conjunction with the caves to correct the distorted exploitative-settlement situation in the mesa and canyon land area. Intensive survey of the Black Mesa State Park has tentatively defined shifts in the settlement pattern between the Woodland and Late Pre-historic periods (Saunders 1978). Between A.D. 200 and 1000, Woodland groups located major habitation camps on stream terraces and caves, whereas specialized activity sites were located on bluff tops, talus slopes, and stream ter-races. The Late Prehistoric groups, however, (A.D. 1000–1500) located camps and villages on bluff tops, in addition to the stream terraces and shelters. Perhaps defense became a concern in the selection of campsite locations (Campbell 1969; 1976).
        Additional study of the existing Kenton Cave materials would not be without merit. A detailed botanical analysis of the number and variety of floral remains could determine what resources were grown, which areas were exploited, and perhaps the seasons of occupations. Similarly, a study of faunal remains might be helpful in interpreting the exploitation patterns at the caves. A detailed description and analysis of the remaining artifacts would help establish a comparative basis for materials from other sites in the region.
        Little definitive information can be provided from a literature review. Scattered artifact collections, inadequate material descriptions, poor provenience information, and a lack of specialized studies severely limits our present knowledge of the cultural resources located in the mesa and canyon lands of the northwestern Oklahoma panhandle.


       The organization of this chapter is based on a draft report prepared by Roger Saunders, and it is with extreme regret that he declined coauthorship of this version. Many people opened their files and institutions during our studies. We thank Ele Baker, Bill White and Mrs. I. W. Mathews for their personal recollections about the early work at the caves, and Don Henry for information about the recent University of Tulsa fieldwork. Hellen Pushmueller of the Department of Anthropology, Denver University; Lynn Barnes and Martha Blaine of the Oklahoma Historical Society; Harold and Joan Kachel of the No Man's Land Museum; K. D. Meeks of the Woolaroc Museum; and Candace Greene and Robert Bell of the Stovall Museum gave us access to records and allowed us to examine collections on display and in storage. In addition, Waldo Wedel of the Smithsonian Institution; Jack Haley of the Western History Archives, University of Oklahoma; Don Wyckoff of the Oklahoma Archaeological Survey; Jiri Zedik and L. R. Wilson of the Oklahoma Geological Survey; and Rose King, secretary of the Oklahoma Anthropological Society, checked additional records and helped run down leads. Archie Hood and Paul Mennis provided their specialized expertise. Roger Saunders provided the artifact plate. Finally, Charles Wallis, Jr., listened and challenged many ideas during the formative stages of this chapter.