A juniper bark fiber net discovered in a cave on Sheep Mountain in the Absaroka Mountains of Wyoming was reported in 1986 (Figure 1; Frison et al. Reference Frison, Andrews, Adovasio, Carlisle and Edgar1986). The original report noted “discovery of the net was accidental, and the circumstances of preservation were unique. Packrat (Neotoma cinerea) midden covered and protected the object, and there was no associated evidence of human occupation of the cave” (Frison et al. Reference Frison, Andrews, Adovasio, Carlisle and Edgar1986:352). The report supplied no additional details as to the discovery, removal, and subsequent curation of the net. The net is currently curated at the University of Wyoming Archaeological Repository in Laramie, on behalf of Old Trail Town Museum, Cody, Wyoming. The artifact and its presumed site of recovery were assigned site number 48PA1022 in 1988.
Generalized map of western North America, showing location of 48PA1022 and distribution of large game and rabbit hunting by use of nets from ethnographic reports (adapted from Anell Reference Anell1969).
The Sheep Mountain Net
The Sheep Mountain find consists of cordage loosely woven into a net with three small stakes, which had been folded into a compact rectangular bundle. The 1986 article states that the net had been folded in on itself on each long side and then folded accordion-style with the stakes still attached. The deteriorated condition of the net precluded unfolding it (Frison et al. Reference Frison, Andrews, Adovasio, Carlisle and Edgar1986:352). The original report estimated the height at 1.5–2.0 m and the length at 50–65 m (Frison et al. Reference Frison, Andrews, Adovasio, Carlisle and Edgar1986:354).
Lengths of the three stakes are not reported in the original report. We measured one stake (Stake 1) at 31.5 cm long and a second (Stake 2) at 45.2 cm (Figure 2). Stake 1 is broken and the proximal end is missing. The third stake is not visible in published photographs and cannot be accessed for measurement (Frison et al. Reference Frison, Andrews, Adovasio, Carlisle and Edgar1986:355; Kornfeld et al. Reference Kornfeld, Frison and Larson2010:313). Using photographs of the net, we have measured the mesh size (center of knot to center of next knot) at 5.5–7.0 cm (Figure 3).
Photographs of net taken in (left) 1986, (center) 2010, and (right) 2016, showing deterioration of netted material. (Left photo by George Frison; center and right photos by Danny N. Walker.)
Detail photograph of net showing cordage with knots used to estimate mesh size. (Photo by Rick Weathermon.)
The net was further described in 1984 by Andrews and collaborators, quoted here by permission of Dr. James Adovasio:
The net specimen from Sheep Mountain, Wyoming, was produced via the free end process and is a looped rather than a linked fabric. Following the format of Andrews and Adovasio (Reference Andrews and Adovasio1980:31), the specimen is allocated to the “knotted loop” category, which is also called “knotted netting.” It is ascribed to a single numbered structural type within the “knotted loop” category based on the type of knot used in the mesh.
… This net specimen consists of a quadrilateral mesh constructed of a series of sequential, regularly spaced and fixed fishnet knots (Emery Reference Emery1966:38–39). The knots are asymmetrical, and the knot faces are dissimilar. On each net surface, knots in consecutive rows exhibit alternate faces.
The specimen is made with two ply, Z spun, S twist (S ZZ) cordage of variable diameter. Cordage splices are made by securing the new raw material beneath the exhausted cordage ply. No body splices are evident in the exposed portion of the net.
Cordage of various diameters is used throughout the net, and in fact, it appears to grade continuously from relatively thick cordage to that which is relatively thin. To facilitate discussion of this cordage, the diameters are grouped into four analytical categories labeled simply “large” (4.10 mm to 5.20 mm), “medium/large” (2.65 mm to 3.00 mm), “medium/small” (1.10 mm to 1.65 mm) and “small” (0.70 mm to 1.00 mm). In the portions of this folded net that are visible, “large,” “medium/large,” and “medium/small” to “small” cords constitute, respectively, 5%, 15%, and 80% of the exposed surface area. As found, the net was folded into a rough rectangular configuration measuring 21 cm × 38. The “large” diameter cordage predominates along both margins of the long axis of the rectangle, while the “medium/large” cordage is found near one of the shorter edges. This grades into smaller and smaller diameter cordage toward the center and “bottom” of the specimen. Mesh gauge varies from ca. 0.71 cm to 3.01 cm and is a function of cordage diameter.
The Sheep Mountain net includes three nearly straight sticks positioned along the long axis of the folded net. In a three-dimensional perspective, the sticks actually form a triad. . . . All three sticks are decorticated and detwigged. Each of the paired sticks has been sharpened to a single point, but the points are oriented in opposite directions. One of these pointed sticks was initially intact. A length from the end opposite the point was [removed by making a cut] perpendicular to the stick's long axis and was used for radiometric dating. The butt end of the other one of the paired sticks is fragmented. Details on the modification or working of the third twig are precluded by the friability of the net; however, the stick appears to be cut at a ca. 35° angle on one end. Some 15 cm and 35 cm (5.9 in and 13.8 in) from the sharpened ends of the fragmented and whole stick, respectively, the net has been wrapped two or more times about the sticks. The wrapped net portion consists predominantly of large cords, but the presence of some small cords suggests that manufacture of the net preceded its attachment.
… The net shows no visible signs of mending, but it is charred at one end. Cords in the “large” diameter category seem to be stained red or reddish-brown; this may represent either intentional staining or unintentional, post-depositional staining (Andrews et al. Reference Andrews, Adovasio, Ronald C. Carlisle and Edgar1984:3–8) [James Adovasio, personal communication 2020].
The stakes were identified only as “hardwood” in the original report but are identified here as greasewood (Sarcobatus vermiculatus), based on macroscopic examination. The visible, exterior portion of the net was later examined under magnification. No blood, tissue, or hairs were observed on it. No attempt has been made to extract animal DNA from the fibers because the net's postexcavation environment was not conducive to DNA preservation.
Age of the Net
The original article reported a radiocarbon date of 8860 BP (originally reported as 6910 ± 170 BC; RL-396) for the net, placing it in the Late Paleoindian period. Calibration using OxCal gives a range of 8350–7583 cal BC for the date, within 95% probability. The dated sample came from a small piece of wood detached from the charred end of Stake 2 (Frison Reference Frison1991:258; Frison et al. Reference Frison, Andrews, Adovasio, Carlisle and Edgar1986:354). The sample was taken parallel to Stake 2's long axis and did not shorten the stake. No attempt was made until recently to verify the Late Paleoindian date, although—as the original study noted—“the probability that such an extremely perishable item of such great age would be preserved at all is extremely low” (Frison et al. Reference Frison, Andrews, Adovasio, Carlisle and Edgar1986:352).
A recent radiocarbon sample on the juniper fiber netting itself called into question the Paleoindian age for this net. The sample was a cordage fragment that had fallen away from the net. This yielded a date of 1320 ± 30 BP (Beta-443460; juniper fiber), with two possible calibrated date ranges: AD 655–720 and AD 740–765, both within the 95% probability level. Because of the large discrepancy between this and the date reported in 1986, a third sample was submitted for dating. This sample consisted of a small piece of cordage removed from well inside the bundled net, and it yielded a date of 1340 ± 30 BP (Beta-511402; juniper fiber), with the same two calibrated ranges as the previous date.
When compared with the original 1986 date, these two Late Prehistoric dates suggested the possibility of old wood having been used for the stakes, resulting in the discrepancy. Wood samples removed from both accessible stakes again yielded Late Prehistoric dates, although slightly younger than the two dates from net fiber: 1201 ± 21 BP (D-AMS 036134; wood) and 1191 ± 25 BP (D-AMS 036135; wood; Figure 4). All four dates were obtained using accelerator mass spectrometry (AMS; Table 1). The slightly younger dates of the two stakes are likely the product of breakage and replacement. Whereas the net would have been mended before reuse, any broken or split stakes would simply have been replaced, resulting in a slight age discrepancy between netting and stakes. The dates for the stakes and the net converge at 1220 cal BP but cover a range of up to 240 years. Whatever the explanation for the variation in the four dates, the artifact is clearly of Late Prehistoric age. The reason for the older date returned in 1986 remains unknown.
Calibrated dates of the Absaroka Mountain net from 2017 to 2018. (Calibration provided by Madeline Mackie of the University of Wyoming.)
Radiocarbon Dates on Net from 48PA1022.
Net Function
The original study proposed that the net had been designed for taking large animals, such as deer or mountain sheep (Frison et al. Reference Frison, Andrews, Adovasio, Carlisle and Edgar1986:354–355, 359). Although the net's function cannot be conclusively demonstrated, current data do not support its hypothesized use for deer or mountain sheep. Because no animal hair, tissue, or blood adheres to the accessible portion of the Sheep Mountain net, its form is the only clue to its use. There are no documented examples or descriptions of nets used for mountain sheep. A few rock art panels from the Great Basin show long nets in conjunction with mountain sheep; however, these cannot be relied on for specific net metrics, and the presence of pictures of other animals on the same panels muddies the connection between the nets and mountain sheep. Archaic rock art in the Black Hills shows detailed scenes of deer hunting using stand nets, providing reliable, but much earlier, evidence for this hunting method in the region (Sundstrom Reference Sundstrom2004:53–62).
Net Metrics
The exact height of the net cannot be determined. As noted, the length of the longer stake is 45.2 cm. A witness to a Great Basin rabbit drive noted that the supporting stakes were “a few inches” longer than the height of the extended net (Egan Reference Egan1917:236). This suggests a net height of between 50 and 60 cm. The original article presents contradictory data about the net height. The net is described as forming a bundle about 21 × 31 cm and having been folded in thirds lengthwise before being accordion-folded and bundled. The outside edges must have been bound with a supporting string that had been strung through the outermost meshes; however, no such support cord is visible on the specimen or in photos of it. If the net had instead been folded with the two outside edges meeting in the middle, these outside edges and support strings would not be visible, but the net would have been in half—not in thirds. Assuming it had been folded in thirds, the unextended height would be 114 cm (38 × 3). If it had been folded in half, the unextended height would be 76 cm (38 × 2). Use height—with the net extended so that the meshes were roughly square—would be less: 80.6 cm if folded in thirds, and 54 cm if folded in half (unextended height × 0.707). The net could be made lower if extended farther or allowed to droop between the supports. In any case, the original estimate of “1.5 m to 2 m high” is not congruent with the available metrics for either netting or stakes.
James Teit described deer-hunting nets from two Columbia Plateau tribes, both made of Apocynum fiber: the first as about 2 m high and from 14 m to 180 m long with “large” meshes; the second as 2 m to 3 m high and from 15 m to 60 m long, mesh size not specified (Teit Reference Teit1900:245–258, Reference Teit1930:246). Museum specimens identified as rabbit nets vary from 0.45 m to 1.8 m high and from 6 m to 120 m long with a 5–7.6 cm mesh size, “approximately the size of a rabbit's head” (Table 2; Connelly et al. Reference Connolly, Kallenbach, Parker and McCabe2017:148). Wildlife biologists recommend 2.4 m high fencing to inhibit mountain sheep movement and note that these animals can “readily hop over” a 1 m high fence (Clevenger and Huijser Reference Clevenger and Huijser2011:178; Paige Reference Paige2015:16). Deer require a 2.1–2.4 m high fence (Clevenger and Huijser Reference Clevenger and Huijser2011:178; Paige Reference Paige2015:47). The Sheep Mountain net lies within the metrics for known rabbit nets, but it is less than half as tall as known deer nets or modern fences for mountain sheep.
Game Nets from Archaeological and Ethnographic Collections with a Known or Proposed Function.
a Calculated from folded net dimensions as reported in Frison and colleagues (Reference Frison, Andrews, Adovasio, Carlisle and Edgar1986).
b Measurement based on photos, not as reported in Frison and colleagues (Reference Frison, Andrews, Adovasio, Carlisle and Edgar1986).
1. Sheep Mountain, Park County, WY (Frison et al. Reference Frison, Andrews, Adovasio, Carlisle and Edgar1986)
2. Hinds Cave, Val Verde County, TX (Andrews and Adovasio Reference Andrews and Adovasio1980)
3. Chewaucan Cave, OR (Connolly et al. Reference Connolly, Kallenbach, Parker and McCabe2017)
4. Hogup Cave, UT (Adovasio et al. Reference Adovasio, Andrews, Illingworth and Hockett2009; Aikens Reference Aikens1970)
5. Etna Cave, NV (Adovasio et al. Reference Adovasio, Andrews, Illingworth and Hockett2009; Wheeler Reference Wheeler1973:22)
6. White Dog Cave, AZ (Adovasio et al. Reference Adovasio, Andrews, Illingworth and Hockett2009; Guernsey and Kidder Reference Guernsey and Kidder1921:77)
7. Kayenta Cave, AZ (Anell Reference Anell1969:47)
8. Cave Creek, AZ (Kaemlein Reference Kaemlein1971:39–42)
9. Cave 10, AZ (Guernsey and Kidder Reference Guernsey and Kidder1921:106)
10. High Cave, AZ (Guernsey Reference Guernsey1931:79)
11. Brooks Cave, TX (Jackson Reference Jackson1937:88)
12. Deep Creek Gosiute (Adovasio et al. Reference Adovasio, Andrews, Illingworth and Hockett2009; Egan Reference Egan1917:235–237)
13. Washo (Adovasio et al. Reference Adovasio, Andrews, Illingworth and Hockett2009; Kroeber Reference Kroeber1925:572)
14. Surprise Valley Paiute (Adovasio et al. Reference Adovasio, Andrews, Illingworth and Hockett2009; Kelly Reference Kelly1932:88) and Little Lake and Koso Mountain Shoshone (Adovasio et al. Reference Adovasio, Andrews, Illingworth and Hockett2009; Steward Reference Steward1938:82)
15. Washo (Adovasio et al. Reference Adovasio, Andrews, Illingworth and Hockett2009; Lowie 1939:327); Lemhi Shoshone, Fort Hall Shoshone, and Bannock Northern Paiute (Adovasio et al. Reference Adovasio, Andrews, Illingworth and Hockett2009; Steward Reference Steward1943:267)
16. Small Creek Shoshone, Battle Mountain Shoshone, and Mill City Northern Paiute (Adovasio et al. Reference Adovasio, Andrews, Illingworth and Hockett2009; Steward Reference Steward1941:329)
17. Thompson River Indians (Teit Reference Teit1900:245–248)
18. Salishan Indians (Teit Reference Teit1930:246)
19. Paiute net at Peabody Museum (Guernsey and Kidder Reference Guernsey and Kidder1921:78)
20. Wasson net (Connolly et al. Reference Connolly, Kallenbach, Parker and McCabe2017)
21. Walapai (Kniffen et al. Reference Kniffen, MacGregor, McKennan, Mekeel and Mook1935:63)
22. Washo (Kroeber Reference Kroeber1925:572)
The stakes included with the Sheep Mountain net also suggest use for small animals. The stakes are between 1 cm and 1.8 cm in maximum diameter, which is comparable to known supports for rabbit nets. The longer of the two accessible stakes is notched at the butt end, as would be expected for a rabbit or sage grouse net. Ethnographic accounts state that for rabbit drives, these long nets were strung along notched sticks such that the net would collapse when the drove of rabbits ran into it, entangling the animals (Anell Reference Anell1969:45; Riddell Reference Riddell1960:38; Steward Reference Steward1938:38; Wheat Reference Wheat1967:41).
Faunal Assemblages
Faunal assemblages from the area indicate consumption of both small and large animals. A study of faunal assemblages from 58 archaeological sites in the Green River Basin of western Wyoming listed bison, rabbits, rodents, and pronghorn in order of frequency, with cottontails often outnumbering jackrabbits (Lubinski Reference Lubinski, Madsen and Metcalf2000; McKibbin Reference McKibbin, Madsen and Metcalf2000:158). Studies in the Wind River Basin of central Wyoming gave similar results, with faunal material dominated by rabbits—both cottontail and jackrabbit—but again with cottontails dominant (Eakin et al. Reference Eakin, Francis, Larson, Larson and Francis1977; Rood Reference Rood2018; Rood et al. Reference Rood, McNees and Peterson2012; Walker Reference Walker1997). Although bone carried in by carnivores may skew leporid counts (Hockett Reference Hockett1994), the use of rabbit flesh for food, skins for woven robes, and bone for beads is well established archaeologically and ethnographically (Hockett Reference Hockett1991:668–669; Wheeler Reference Wheeler1997:18–34). Faunal material from at least two sites in the Big Horn Basin also suggests a stronger dependence on small mammal exploitation than on large game, at least seasonally (Stuckenrath and Mielke Reference Stuckenrath and Mielke1972; Walker Reference Walker, Davis and Stanford1988, Reference Walker, Frison and Walker2007).
Mummy Cave, in the Absaroka Mountains, had rabbit remains in all levels (Harris Reference Harris, Husted and Edgar2002; Hughes Reference Hughes2003; Husted and Edgar Reference Husted and Edgar2002). Three excavated sites in the Green River and Big Horn Basins had bone counts suggesting communal rabbit hunting: the Dick Myal housepit site, 48FR6256, with a minimum of 26 jackrabbits (Rood Reference Rood2018); the Raptor site, with a minimum of 60–69 cottontails (Lubinski Reference Lubinski2003); and Rabbit Bone Cave, with 44 cottontails and four jackrabbits (Stuckenrath and Mielke Reference Stuckenrath and Mielke1972; Walker Reference Walker, Davis and Stanford1988). A fourth probable communal rabbit hunting site contained 51 jackrabbit tibiae (Frison Reference Frison1991:264; Kornfeld et al. Reference Kornfeld, Frison and Larson2010:335–336).
Except for Mummy Cave, these studies refer to lower-elevation sites and consequently cannot be directly applied to the higher-elevation Sheep Mountain area; however, they show that rabbits made up a significant proportion of animals hunted throughout the precontact history of the region (Frison Reference Frison1991:263–265; Kornfeld et al. Reference Kornfeld, Frison and Larson2010:335–337; Lubinski Reference Lubinski2003). Deer and mountain sheep are relatively rare in these basin sites, with deer occurring about twice as frequently as mountain sheep. Several Late Prehistoric sites at higher elevations in the Absaroka Mountains indicate greater dependence on mountain sheep, including 48PA853 and 48PA919 (Eakin Reference Eakin1989; Eakin and Eckerle Reference Eakin and Eckerle2012). Mummy Cave contained high proportions of mountain sheep in levels 30 and 36, the latter corresponding in age to the Sheep Mountain net (Harris Reference Harris, Husted and Edgar2002:166; Husted and Edgar Reference Husted and Edgar2002:168–170). Remains recovered from these and other sites suggest that mountain sheep were relatively abundant at the time the Sheep Mountain net was in use. Although the Sheep Mountain net was found outside of jackrabbit habitat, it could either have been used at a lower elevation, or it could have been used for cottontails—which are common in faunal assemblages from the area—or both. Nets can be, and were, used in basins and higher elevations for rabbits, regardless of species.
None of these observations would preclude the use of the Sheep Mountain net for other small animals, such as sage grouse (Centrocercus urophasianus), which also occurs in archaeological assemblages from the area (Fenner Reference Fenner, Frison and Walker2007; Husted and Edgar Reference Husted and Edgar2002:167–170; Kornfeld et al. Reference Kornfeld, Frison and Larson2010:338–340; Lubinsky Reference Lubinski, Madsen and Metcalf2000:182–183; McKibbin Reference McKibbin, Madsen and Metcalf2000:159–160). Shoshone groups in Nevada employed a rabbit net to take sage grouse (Steward Reference Steward1941:222).
Discussion and Conclusions
Overall, reanalysis of the net's age is a cautionary tale about relying on a single, anomalous radiocarbon date. Because of the proposed Paleoindian age and the uniqueness of an animal trapping net from that time, the occurrence has been referenced many times (for example, Connolly et al. Reference Connolly, Kallenbach, Parker and McCabe2017:143; Driskell and Walker Reference Driskell, Walker, Walker and Driskell2007:230; Fiedel Reference Fiedel, Walker and Driskell2007:30; Frison Reference Frison and Carlisle1988:88, Reference Frison, Stanford and Day1992:35; Frison et al. Reference Frison, Reher, Walker, Davis and Reeves1990:208; Hughes Reference Hughes2003:73–74; Kauffman et al. Reference Kauffman, Meacham, Sawyer, Steingisser, Rudd and Ostlind2018:37; Kornfeld Reference Kornfeld, Walker and Driskell2007:53; Kornfeld and Larson Reference Kornfeld and Larson2008:25; Kornfeld et al. Reference Kornfeld, Frison and Larson2010:312–313, 349; Lupo and Schmitt Reference Lupo and Schmitt2002:160–161; Sundstrom Reference Sundstrom2004:51; Walker Reference Walker, Davis and Stanford1988:6.08.3; Whitley Reference Whitley2000). The four additional AMS radiocarbon dates yielded an age between 1300 and 1200 years for the netting and around 1150 cal BP for the stakes, placing the Sheep Mountain net in the Late Prehistoric period.
Based primarily on the thickness of some of the cordage, Frison and colleagues (Reference Frison, Andrews, Adovasio, Carlisle and Edgar1986) proposed that the Sheep Mountain net had been used on animals up to the size of mountain sheep or deer rather than rabbits or other small game. Although its cordage is thick in comparison with known rabbit nets from other locations, the latter are made of materials other than juniper fiber. Juniper may require (or result in) a thicker cordage. This possibility has not been explored, but it provides a reasonable alternative explanation for the thick cordage. Overall, the size, height, and form of the net argue for use for small game—probably rabbits or sage grouse.
Acknowledgments
First, we would like to acknowledge Mr. Larry Edgar—secretary-treasurer of Old Trail Town in Cody, Wyoming—for permission to collect samples from the net for the four dates. We appreciate the financial assistance of both Dr. Larry Loendorf of Sacred Sites Research and the Wyoming Association of Professional Archaeologists in obtaining the four AMS dates. Sylvia Huber, researcher and office manager of Old Trail Town, located the 1984 unpublished manuscript in the Old Trail Town files. We also appreciate Dr. James Adovasio for taking the time to discuss the history and details of the original research on the net. Dr. Jody Clauter, collections manager of the University of Wyoming Archaeological Repository, provided access to the net in 2018 to collect the first of the radiometric samples, and later, Collections Manager Dr. Marieka Arksey assisted in accessing the other three samples. Senior Academic Professional Research Scientist Dr. Rick Weathermon, Department of Anthropology, University of Wyoming, assisted in collecting the second fiber sample from the net and provided Figure 4. Michael T. Bies assisted with various aspects of the project. He and Dan Bach assisted with identification of the wood. Dr. Caroline Solazzo of the Smithsonian Institution advised on the item's potential to contain usable animal proteins and DNA. Last, but not least, we would like to acknowledge the late Dr. George Frison and his longtime commitment to understanding the prehistory of Wyoming.
Data Availability Statement
All original data used in this study are presented in the text.
