The extraordinary antiquity of the fiber artifact assemblage from the Paisley Caves (Figure 1) offers a unique opportunity to document what plants were used for technological applications (principally basketry and cordage) by First Americans and later cave occupants. In this study, microscopic fiber identification, together with radiocarbon dating, serves as an effective means to understanding landscape use and plant selection for fiber-based technologies. Preservation of fiber artifacts in cave sites throughout this region has driven decades of research in Great Basin basketry, focusing on population movements, subsistence and mobility, and cultural attributes (style and structure), with an ongoing radiocarbon dating program that continues to advance regional chronologies (Adovasio Reference Adovasio and d'Azevedo1986; Adovasio and Pedler Reference Adovasio, Pedler, Madsen and Rhode1994; Camp Reference Camp2018; Connolly Reference Connolly1994, Reference Connolly2013, Reference Connolly2022; Connolly and Barker Reference Connolly, Barker, Jenkins, Connolly and Aikens2004, Reference Connolly, Barker, Fowler and Fowler2008; Connolly et al. Reference Connolly, Barker, Fowler, Hattori, Jenkins and Cannon2016; Fowler and Hattori Reference Fowler, Hattori and Hughes2011; Ollivier et al. Reference Ollivier, Smith and Barker2017; Smith et al. Reference Smith, Ollivier, Barker, Camp, Harvey and Jones2016).
Map of sites discussed in the text: 1. Paisley Caves; 2. Fort Rock Cave; 3. Connley Caves; 4. Cougar Mountain Cave; 5. Roaring Springs; 6. Dirty Shame Rockshelter; 7. Paulina Lake Site; 8. Chewaucan Cave; 9. Buffalo Flats; 10. Boulder Village; 11. Bergen Site; 12. Catlow Cave; 13. LSP-1 Rockshelter; 14. Bergen Site; 15. DJ Ranch; 16. Big M Village.
Despite the substantial body of work that has contributed to basketry studies, less attention has been given to which plants were used in textile manufacture. Textiles are often documented only by artifact type; specific plant materials are generally not identified, or assumptions are made based on macroscopic observations, and fine cordage is most often described as “bast” or “hemp” fiber. Exceptions to this include analysis of Basketmaker II Yucca and Apocynum fiber cordage from Boomerang Rockshelter in Southeast Utah (Haas 2006); identification of bast, bark, and monocot fiber cordage and raw materials from Four Siblings Rockshelter in Nevada (Coe Reference Coe2012); and fiber analysis of cordage from Bonneville Estates Rockshelter, Hogup Cave, and Danger Cave in Utah (Coe Reference Coe2021; Lawlor Reference Lawlor2020).
Extensive fiber assemblages from Roaring Springs Cave, Catlow Cave, and Dirty Shame Rockshelter have served as iconic collections on which Northern Great Basin (NGB) typologies are based (Andrews et al. Reference Andrews, Adovasio and Carlisle1986; Connolly et al. Reference Connolly, Fowler and Cannon1998; Cressman Reference Cressman1942), but systematic identification of plant taxa has yet to be conducted beyond cursory macroscopic analysis. By contrast, archaeobotanical research using seed, phytolith, and pollen datasets has flourished in recent years in the NGB, offering new insights on human health, diet, landscape use, and the paleoenvironment (Beck et al. Reference Beck, Bryant and Jenkins2018; Blong et al. Reference Blong, Adams, Sanchez, Jenkins, Bull and Shillito2020; Kennedy Reference Kennedy2018; Kennedy and Smith Reference Kennedy and Smith2015; McDonough Reference McDonough2019; McDonough et al. Reference McDonough, Kennedy, Rosencrance, Holcomb, Jenkins and Puseman2022; Taylor et al. Reference Taylor, Hutson, Bryant and Jenkins2020), particularly at sites such as LSP-1 Rockshelter, Connley Caves, Cougar Mountain Cave, and Paisley Caves. This focus on paleobotanic microscopy and lack of taxa identification for textile technologies has prompted a systematic approach to fiber identification of the Paisley Caves archaeological textiles using direct and polarized light microscopy and scanning electron microscopy (SEM) with energy-dispersive X-ray spectroscopy (EDS).
This study addresses the following research questions:
(1) What plant taxa are represented in Paisley Caves archaeological textiles;
(2) Is there change in plant selection for textiles over time; and
(3) What can this tell us about landscape use in relation to changing climatic conditions?
To address these research questions, I systematically examined and identified more than 300 fiber artifacts using microscopy, considered the ecology of the represented taxa (e.g., wetland, dryland), and refined the chronology through radiocarbon dating of selected artifacts. Results offer a more nuanced look at seasonality and settlement-subsistence models within the Summer Lake Basin throughout the Holocene, and they suggest continuity of plant selection for technology—in particular, the use of stinging nettle and dogbane in the construction of fine cord by 11,000 years ago.
Textile Plants of the Northern Great Basin
For millennia, people have collected fiber from a wide range of plant communities within the NGB. Sagebrush (Artemisia spp.), bitterbrush (Purshia tridentata), cliffrose (Purshia mexicana), sumac (Rhus spp.), and milkweed (Asclepias spp.) were collected from dry steppe shrub zones and juniper (Juniperus occidentalis) and blue flax (Linum lewisii) from mid- to high-elevation open forests. Monocots from wetlands include Juncus rush (Juncus effusus), tule (Schoenoplectus acutus), cattail (Typha latifolia and T. angustifolia), sedge (Carex spp.), and reed or cane (Phragmites communis). Riparian corridors and lowlands included black and white cottonwood (Populus spp.), willow (Salix spp.), stinging nettle (Urtica dioica), and dogbane (Apocynum spp.). Hard-stem bulrush is commonly used in anthropological literature (Fowler Reference Fowler1992) to refer to tule (Schoenoplectus acutus, formerly Scirpus acutus)—a large sedge with rounded, hard stems.
Sagebrush and tule-twined basketry (e.g., containers, mats, bags, footwear, infant carriers) characterize much of NGB textiles (Figure 2), a region that is well-known for the Fort Rock–style sagebrush sandals (rarely tule) dating to approximately 9400 cal BP (based on an average of 58 dates). The introduction of Multiple Warp and Spiral Weft tule sandals around 9400 cal BP, continuing to the Late Holocene, were accompanied by the regionally diagnostic Catlow Twined tule basketry (Connolly et al. Reference Connolly, Barker, Fowler, Hattori, Jenkins and Cannon2016), described as plain twined basketry with Z-twist tule wefts and Z-twist, s-spun tule cordage warps (Figures 2 and 3).
A representative sample of Paisley Caves textiles. (Color online)
Woven structures: (a) tule matting; (b) whole tule matting selvage; (c) sagebrush weft element; (d) tule Catlow twined basketry; (e) tule weft element; (f) tule Catlow twined basketry; (g) unidentified fiber, weft element. (Color online)
Klamath and Modoc tule caps, bowls, and trays were often made with nettle or dogbane cordage starts; decorative elements include peeled and split reed or cane, white bear grass, feather quill, porcupine quill dyed yellow with wolf lichen, and mud-dyed tule stem or root (Margaret Mathewson, personal communication 2018). Tule mats were also made with nettle or flax cordage warps (Spier Reference Spier1930). In early summer, sedge was peeled and split for making mats (Coville Reference Coville1897). Round Juncus rush (common name) stem is stiffer but is also used in mats.
Whereas twined tule basketry is associated with Klamath and Modoc peoples, coiled willow basketry is associated with Northern Paiute and Shoshone. Late Holocene Lovelock Cave wickerware in western Nevada is also peeled willow rod and peeled and split willow (Tuohy and Hattori Reference Tuohy and Hattori1996). Willow sticks were also used in open twined seed beaters, burden baskets, and fish traps, and in making snowshoe frames (Barrett Reference Barrett1910; Coville Reference Coville1897).
Coarse cordage (twine and rope) was often made from shredded sagebrush bark, juniper, antelope bitterbrush, or cliffrose, as well as willow, juncus rush, cottonwood, sumac, and grasses (Poaceae). Like bitterbrush and juniper, sagebrush bark was peeled and shredded before being twisted into cord. The Nez Perce in the Columbia Basin and Northern Paiute and Klamath peoples used bitterbush and cliffrose in cordage and basketry. Woody stems were split and could be used as warp or weft elements in basketry (Chamberlain Reference Chamberlain1911; Cummings Reference Cummings and Minnis2004; Kelly Reference Kelly1932; Rhode Reference Rhode2002).
Strong and pliable fine cordage made from the inner fibers of herbaceous dicots was needed for sewing thread, nets, fishing line, snares, hairnets, bowstrings, and warp elements in twined basketry and matting, and it was needed in lashing and tying (Barrett Reference Barrett1910; Coville Reference Coville1897; Downs Reference Downs1966; Fowler Reference Fowler1992; Fowler and Dawson Reference Fowler, Dawson and d'Azevedo1986; Rhode Reference Rhode2002; Spier Reference Spier1930; Stewart Reference Stewart1939; Wheat Reference Wheat1967). Historically, Klamath, Modoc, and Northern Paiute peoples used long, linear nets to catch rabbits, fish, and waterfowl (Coville Reference Coville1897; Oetting Reference Oetting, Melvin Aikens and Jenkins1994; Spier Reference Spier1930). In the NGB, the most common bast fibers used historically were dogbane and stinging nettle, although blue flax and milkweed were also used (Barrett Reference Barrett1910; Coville Reference Coville1897; Downs Reference Downs1966; Fowler and Fowler Reference Fowler and Fowler1970; Rhode Reference Rhode2002; Spier Reference Spier1930; Stewart Reference Stewart1939; Wheat Reference Wheat1967).
The Paisley Caves
The Paisley Caves are located within the ancestral homelands of the Klamath, Modoc, and Northern Paiute peoples, and they geologically consist of eight wave-cut rockshelters that were at one time located along the shoreline of pluvial Chewaucan Lake in eastern Oregon, now overlooking the mostly dry Summer Lake. Fieldwork by the University of Oregon (UO) Field School between 2002 and 2011 provides evidence of pre-Clovis human occupations, based on DNA and biomarkers from a coprolite around an ash feature in Cave 5 dating to 14,200 cal BP and a tule basketry weft element (Gilbert et al. Reference Gilbert, Jenkins, Götherstrom, Naveran, Sanchez, Hofreiter and Thomsen2008; Jenkins et al. Reference Jenkins, Davis, Stafford, Campos, Hockett, Jones and Cummings2012, Reference Jenkins, Davis, Stafford, Campos, Connolly, Cummings, Hofreiter, Graf, Ketron and Waters2013, Reference Jenkins, Davis, Stafford, Connolly, Jones, Rondeau, Cummings, Kornfeld and Huckell2016; Shillito et al. Reference Shillito, Whelton, Blong, Jenkins, Connolly and Bull2020). Early occupation periods of the caves are interpreted as brief, temporary visits during fall and spring for upland seed collecting (Kennedy Reference Kennedy2018). Caves 2 and 5 were the most prolific for the Paisley period occupation (15,700–12,800 years ago), including cordage, bone tools, botanical remains, obsidian and chert flakes, Pleistocene animal bone, and stemmed points (Beck and Jones Reference Beck and Jones2010; Jenkins et al. Reference Jenkins, Davis, Stafford, Campos, Hockett, Jones and Cummings2012; Wriston and Smith Reference Wriston and Smith2017). The Paisley period is therefore defined as 15,700 to 12,800 years ago, the Connley period as 12,800–11,500 years ago, and the Fort Rock period as 11,500–9000 years ago, based on dates acquired from Paisley, Connley, and Fort Rock sites (Aikens et al. Reference Aikens, Connolly and Jenkins2011; Jenkins et al. Reference Jenkins, Connolly, Aikens, Jenkins, Connolly and Aikens2004).
A Younger Dryas botanical lens rich in organic material in Cave 2 offers a unique glimpse into the lives of early occupants. The lens is a matrix of sagebrush matting with hairs, coprolites, and rat and bat feces located between two mud lenses with dates of 11,960 and 12,930 cal BP (Hockett et al Reference Hockett, Adams, Lubinski, Butler and Jenkins2017; Jenkins et al. Reference Jenkins, Davis, Stafford, Connolly, Jones, Rondeau, Cummings, Kornfeld and Huckell2016). The lens also contains butchered pronghorn and cut hair, jackrabbit cut bone, fish bone, insects, obsidian flakes, and two unlined hearth features (Hockett et al. Reference Hockett, Adams, Lubinski, Butler and Jenkins2017). Evidence of fiber industries includes sagebrush braided rope and strips of sagebrush bark—likely material for textiles or fuel. A bone needle or awl may relate to the production of sagebrush artifacts such as sandals, clothing, and matting, or sewing activities with sinew or bast fiber (Gahr Reference Gahr2006; Puseman and Cummings Reference Puseman and Cummings2003).
Artifact assemblages spanning the Fort Rock period (11,500–9000 years ago) and Lunette Lake period (9,000–6,000 years ago) include an abundance of fine cordage for netting and sewing, coarser rope, and tule basketry. However, brief occupations characterize the Lunette Lake period, evident by hearths with limited faunal bone and lithic scatters. Textiles become more diverse and prolific during the later Bergen period (6,000–3,000 years ago) and Boulder Village period (3,000 years ago to the postcontact era).
Paisley Caves Textile Assemblage
The fiber artifact assemblage generated from the 2002–2011 UO field school excavations total 453 specimens (Table 2). The Museum of Natural and Cultural History's (MNCH) legacy collections and private donations (an additional 125 fiber artifacts) are excluded from the current study. Most twined structures were recovered during the 1930s and 1940s by MNCH curator Luther Cressman and other amateur collectors, possibly due to the higher number of Late Holocene basketry in higher strata or to periodic looting of the caves. This includes all of the sandals, which are Multiple Warp style dating to the Late Holocene. The field school collections include just 13 additional Catlow twined basketry fragments, one decorated with overlay and feather quill false embroidery (7560–7420 cal BP; Connolly et al. Reference Connolly, Fowler and Cannon1998) and three fragments of Z-twist open-twined matting.
Fine cordage makes up 33% of the collection, although this number may be inflated due to fragmented segments from the same structure (Figures 2, 5, and 6). Fine cordage (string-like) is defined as 8–14 twists per cm, ranging in diameter from 0.6 to 3.3 mm; is most often z-spun, S-twist, with some exceptions; and is almost exclusively made from bast plant fiber (Connolly et al. Reference Connolly, Barker, Fowler, Hattori, Jenkins and Cannon2016; Emery Reference Emery1966). Of the 135 fine cord fragments, 49 have at least one weaver's knot; only 12 have two or more knots (interpreted as a net fragment) in which a mesh gauge measurement was recorded. One possible looped cordage may be from a knotless net or other structure (Figure 6a).
Coarse cordage and braids include both the wider rope-like cords (greater than 7 mm in diameter) primarily made from sagebrush bark and the medium-sized cords (mean diameter of 2.6 mm) that are often made from tule or other monocots. Medium-sized cordage fragments made from monocot stems are likely basketry warp trimmings (Connolly et al. Reference Connolly, Barker, Fowler, Hattori, Jenkins and Cannon2016; Figures 2 and 7). Three-strand braids from Paisley Caves are all temporally unique to the Late Pleistocene / Early Holocene (EH) and are also contemporaneously present at Cougar Mountain Cave (Rosencrance et al. Reference Rosencrance, Smith, Jenkins, Connolly and Layton2019). The average date of braids from both sites (n = 23) is 11,985 cal BP, though braids from other sites, such as Roaring Springs Cave, also date much younger.
Other modified fibers (n = 92) include twisted and bent fiber, coils, and overhand knots. Most notable are two coils, a fiber-wrapped quill, four bent elements that are likely basketry elements, a hide strip wrapped with fiber at one end (Figures 2 and 8), and five fragments of plaited cotton fabric (Figure 9), two of which were radiocarbon dated to the Late Holocene (Jenkins et al. Reference Jenkins, Davis, Stafford, Campos, Connolly, Cummings, Hofreiter, Graf, Ketron and Waters2013; Table 1).
Paisley Caves Radiocarbon Dates.
a Previously unpublished.
Artifact Type by Temporal Component and Taxa.
Materials and Methods
Fiber artifacts were divided into (1) phloem fibers from herbaceous dicot plant stems (used in making nearly all fine cordage); (2) bark and leaves used in making coarse rope, braids, and some knotted fibers; and (3) monocots used in making most basketry and coarse cordage. Reference collections were established based on herbarium samples and loose fibers collected from nineteenth-century basketry housed at MNCH. Specimens were first examined under low magnification to select those requiring further analysis. Approximately 328 were sampled for examination with a Leica DM polarizing light microscope; 125 were not sampled (mostly tule basketry, cordage, and sagebrush coarse rope and knots).
Bast Fiber Identification
Bast fiber identification was conducted by the author at MNCH and the Center for Advanced Materials Characterization in Oregon (CAMCOR) following previously established methods (Bergfjord and Holst Reference Bergfjord and Holst2010; Florian et al. Reference Florian, Kronkright and Norton1990; Haugan and Holst Reference Haugan and Holst2013; Jakes and Mitchell Reference Jakes and Mitchell1996; Jakes et al. Reference Jakes, Sibley and Yerkes1994; and Suomela et al. Reference Suomela, Vajanto and Räisänen2018). Control samples for dogbane, nettle, flax, milkweed, and fireweed from herbaria were first examined using polarized light microscopy (PLM) and scanning electron microscopy (SEM) with consecutive energy dispersive X-ray spectroscopy (EDS) to establish fiber characteristics. Results from the control samples, together with previously reported characteristics, were then compared to fibers from artifacts (Kallenbach Reference Kallenbach2023). Fibers too difficult to identify using only PLM were also examined under SEM with EDS. Elemental composition provided an additional tool to aid in distinguishing between nettle, milkweed, and flax in some cases. Assistance in flax and milkweed characterization was also provided by Sandra Koch (McCrone Research Institute). All samples obtained from artifacts for plasma ashing and for SEM with EDS were less than 5 mg.
Bark Identification
Bark identification was conducted in consultation with wood scientist Suzana Radivojevic (UO, Historic Preservation Program), in which salient microscopic characteristics were established for three taxa of bark: sagebrush, bitterbrush, and juniper based on herbarium reference samples, literature, and wood anatomy databases. Longitudinal views of both unprepped bark and macerated bark fiber were examined. It was not feasible to examine bark cross sections due to the small fiber samples and because of the risk of unnecessary alteration of the artifacts. Subsets of samples were also analyzed by archaeobotanists Kathryn Puseman (Paleoscape Archaeobotanical Services Team) and Linda Scott Cummings (PaleoResearch Institute) as independent verifications of the author's identifications.
Monocot Identification
Microscopic identifying criteria were established for three monocots: tule stem, cattail leaf, and juncus, based on herbarium reference samples, literature, and online sources. Additionally, a subset of samples was analyzed by archaeobotanists Kathryn Puseman (Paleoscape Archaeobotanical Services Team) and Linda Scott Cummings (PaleoResearch Institute).
Results: Fiber Identification
The textile assemblage includes 453 artifacts. Sagebrush, tule, dogbane, and stinging nettle account for 74% of all taxa; the remainder includes milkweed, flax, cattail, grass, bitterbrush, sumac, juniper, cottonwood or willow, cotton (commercial/modern and ancient), commercial jute, animal hair, and unidentified fibers (Table 2). Unidentified taxa (about 13%) were documented as monocot, bark, bast fiber, or unidentified plant fiber. Modern jute (n = 1) and cotton cord (n = 25) were identified; these are not included in the analysis.
Woven Structures
Woven structures include twined matting and basketry constructed of tule stem (13%), sagebrush bark (2%), cattail leaf (1%), and cotton plaiting (4%; Figures 3 and 4). Some twisted and bent fibers grouped under “Other” may be basketry weft or warp elements. Catlow twined basketry is constructed of tule stem; two matting fragments are made from cattail leaf and sagebrush bark.
Cattail matting, possible edge, open twined, 6660–6485 cal BP. (Color online)
Fine Cordage and Netting
Fine cord and netting (n = 182) are constructed almost exclusively from bast fibers (Figures 5 and 6)—primarily dogbane (47%) and stinging nettle (28%), followed by flax (2.6%) and milkweed (2%). Other fibers include unidentified bast fibers (7.8%), cotton (2.6%; Figure 9), unidentified plant fiber (2.6%), tule stem (1.6%), and unidentified bark (1%). All 49 cords with at least one weaver's knot are constructed of dogbane or nettle, or a mix of these two plants. Fifteen cords include blended dogbane and nettle fibers (Table 2); based on the current study, it is unclear if mixing of these two taxa was intentional.
Fine cordage: (a) polarized light microscopy; (b) dogbane cordage, 11,105–10,700 cal BP; (c) dogbane cordage, 7785–7615 cal BP; (d) stinging nettle cordage; (e) stinging nettle cordage, 11,070–10,715 cal BP; (f) stinging nettle cordage, 670–560 cal BP; (g) stinging nettle cordage, 670–560 cal BP; (h–k) milkweed cordage, 2670–2350 cal BP; (l) dogbane; (m) flax, 1825–1630 cal BP; (n) flax, 685–570 cal BP; (o) flax, 560 ± 20 BP; (p) dogbane, 3230–3070 cal BP; (q) dogbane, 8520–8360 cal BP; (r) dogbane, 11,270–11,205 cal BP; (s) quill-wrapped with dogbane cordage; (t) flax, 6295–6115 cal BP. (Color online)
Netting and knotted cordage: (a) dogbane cordage with loops, suggesting unknotted net/bag, 10,279–10,212 cal BP; (b) dogbane, 10,250–10,205 cal BP; (c) dogbane net fragment in knotted bundle; (d) stinging nettle; (e–h) dogbane; (i) dogbane, 2345–2180 cal BP; (j) dogbane net with three knots; (k) dogbane net with multiple knots; (l–p) dogbane; and (q) dogbane, 625–515 cal BP. (Color online)
Very few fine cords are made from nonbast fibers but still fall within the metrics defined for fine cordage—more than eight twists per cm and less than 3.3 mm in diameter. These include tule stem (n = 3), unidentified bark (n = 1), unidentified plant fiber (n = 4), and cotton (n = 5). Of the 18 cotton cords identified, 15 are considered modern intrusions based on evenness of ply, coloration, and disturbed contexts, and they are likely remnants from articles of clothing or string left behind by the numerous visitors to the caves since the early twentieth century. Five cotton cords were included in the analysis based on precontact radiocarbon dates (Table 1).
Coarse Cordage and Braids
Coarse cordage and braids (n = 128) are constructed almost exclusively from tule stem (39.8%) and sagebrush bark (30.4%), followed by sumac stem (2.3%), bitterbrush bark (1.5%), juniper bark (1.5%), cattail leaf (1.5%), grass (0.7%), unidentified monocot (7%), unidentified bark (3.1%), and unidentified fiber (11.7%; Figure 7). Rope-like and coarser cord is most often constructed from bark, whereas medium-sized coarse cordage is constructed from tule stem, and in some cases, these cords could be basketry elements. Sumac cord (Figure 7i, j) is constructed from whole stems. All braids are constructed from sagebrush bark.
Coarse cordage and braids: (a) tule; (b) bitterbrush knot; (c) sagebrush three-strand braid, 12,585–12,045 cal BP; (d) sagebrush three-strand braid, 12,100–11,988 cal BP; (e) sagebrush; (f) juniper, 12,440–11,815 and 11,995–11,820 cal BP; (g) unidentified bark, 11,970–12,310 and 11,860–12,110 cal BP; (h) juniper bark; (i) sumac stem; (j) sumac stem, 1180–1060 cal BP; (k) tule possible weft with warp element. (Color online)
Other
Other fiber artifacts include twisted, crimped, and bent fiber, coils, wraps, and knots, and they represent a more diverse group of fiber material. This is perhaps because of the diversity in structure types, which likely represents a wide range of artifact functions, although these are unknown (Figure 8). Sagebrush (28%) includes primarily overhand knots, a half hitch knot, folded bark, twisted fibers, and one coil. Tule (34%) includes twisted, crimped, and bent fibers (likely weft or warp elements); and one knot and coil. Bitterbrush (2.2%) includes two knots; goosefoot (2.2%) includes twisted fibers likely from a structure; cattail (1.1%) and unidentified grass (1.1%) include twisted fiber; cottonwood or willow (2.2%) includes two coils; unidentified monocots (10.1%) and unidentified plant fiber (5.6%) include bent and crimped fibers, and wrapped sticks and stems. Five fiber twists or wads were identified as animal hair; only two of these five (5.6%) are interpreted as intentionally twisted.
Other fiber artifacts: (a) twisted animal hair; (b) twisted or bent animal hair; (c) multiple s-spun tule roots in loose knot; (d) sagebrush Zss cord with loop start of Zssss 4 ply; (e) s-spun split tule strand in loose knot; (f) tule wrapped stem or coil; (g.) monocot wrapped stick or coil; (h) two willow coils. (Color online)
Cotton cordage: (a) 460–310 cal BP; (b) 645–540 cal BP; (c) 740–675 cal BP; Cotton plaiting: (f) 305–418 cal BP; (g) 980–940 cal BP; (d–e and h–i) were not dated due to size limitations. (Color online)
Radiocarbon Dates and Temporal Distributions of Plant Taxa
Table 1 lists 54 radiocarbon dates from fiber artifacts from field school assemblages, both previously reported dates and 34 new dates. The 28 dates obtained for this study were sampled in consultation with Richard Rosencrance (University of Nevada, Reno), who oversaw sample pretreatment and combustion to CO2 at the University of Nevada, Reno, Human Paleoecology and Archaeometry Lab (HPAL), following standard protocol (McDonough et al. Reference McDonough, Kennedy, Rosencrance, Holcomb, Jenkins and Puseman2022). The HPAL then sent the CO2 to the Penn State Radiocarbon Laboratory for graphitization and measurement. All results listed in Table 1 are rounded following the conventions of Stuiver and Polach (Reference Stuiver and Polach1977) and calibrated with OxCal v4.4 (Bronk Ramsey Reference Bronk Ramsey2009) using the IntCal20 curve.
Previously reported early dates from the Paisley Caves include a basketry weft fragment, likely tule, dating to 14,660–14,080 cal BP (Shillito et al. Reference Shillito, Whelton, Blong, Jenkins, Connolly and Bull2020); unidentified Szz cordage (12,584–12,231 cal BP); twisted grass (12,390–12,655 cal BP); three-strand sagebrush braids (12,398–12,016 cal BP and 12,100–11,988 cal BP); and Middle and Late Holocene tule basketry and cotton plaiting (Jenkins et al. Reference Jenkins, Davis, Stafford, Campos, Connolly, Cummings, Hofreiter, Graf, Ketron and Waters2013). Dates obtained for the current study provide new evidence for the use of dogbane, stinging nettle, juniper, and sumac cordage by 11,000 years ago; and milkweed, flax, and cotton cordage by the Late Holocene.
The oldest bast fibers include two stinging nettle cords (11,930–11,410 cal BP and 11,070–10,715 cal BP) and four dogbane cords (11,690–11,255 cal BP, 11,270–11,205 cal BP, 11,080–10,715 cal BP, and 11,105–10,700 cal BP). The combined 11 total dates for dogbane cordage and netting confirm its consistent use throughout the Holocene, with the youngest dogbane net dating to 625–515 cal BP. Similarly, the five dates for stinging nettle cordage confirm its use in the Early and Late Holocene, with the youngest stinging nettle net dating to 670–560 cal BP. All five flax fine cords were dated, and they include Late Holocene dates of 1825–1630 cal BP, 960–835 cal BP, 685–570 cal BP, and 560 ± 20 BP, and one Middle Holocene date of 6295–6115 cal BP. Two of the four milkweed cords were dated, with Late Holocene dates of 2670–2350 cal BP and 1410–1345 cal BP.
Late Pleistocene fiber other than sagebrush includes juniper bark (12,440–11,815 and 11,995–11,820 cal BP) and sumac-stem coarse cordage (11,970–12,310 and 11,860–12,110 cal BP) and one late-period sumac cord (1180–1060 cal BP). The only identified cattail fiber includes a new date for twined matting dating to 6660–6485 cal BP. Of the seven dated cotton textiles, two returned recent dates and are considered modern intrusions. Three cotton cords returned dates of 645–540 cal BP, 460–310 cal BP, and 255–30 cal BP—the latter likely historic. Two of the plaited fabrics returned dates of 305–418 cal BP and 980–940 cal BP. This is the first identification of precontact-era cotton from the Paisley Caves and the NGB.
Temporal Distributions of Plant Taxa
Textiles were grouped by artifact type, taxa, and temporal component to examine plant selection over time (Table 2). Age ranges of fibers are based on direct radiocarbon dates or associated temporal component largely based on obsidian hydration and stratigraphy (Jenkins et al. Reference Jenkins, Davis, Stafford, Campos, Connolly, Cummings, Hofreiter, Graf, Ketron and Waters2013). Nearly 40% of textiles are from disturbed components, which refer to krotovina disturbance and pits made by looters during the early 1900s. Overall, there is consistent use of a limited number of taxa since the Late Pleistocene. Fine cord and netting are represented by dogbane and nettle during the Late Pleistocene and throughout the Holocene, with milkweed and flax occurring much later during the Late Holocene. Coarse cordage and braids made from bark and monocot stem indicate consistent use of sagebrush since the Late Pleistocene, and although juniper, sumac, and bitterbrush represent less than 2%, they also occur during the Early Holocene. Other textiles (coils, knots, bent and twisted fiber) are dominated by sagebrush knots and tule twisted and bent fiber, which are present continuously since the Late Pleistocene; bitterbrush and grass are also present during the Late Pleistocene. Although beyond the scope of this article, it should be noted that bast cordage technology remains constant since the Early Holocene; twist direction is consistently Szz, and fine cord metrics (cord diameter and number of twists per cm) for knotted and unknotted cordage does not vary based on dogbane and nettle selection.
Plant taxa were also grouped by ecology and temporal component, excluding artifacts from disturbed components. Data are presented by number of artifacts (Figure 10); these broad ecological groupings indicate continuity overall, but with a clear pattern of increased wetland plants after 12,000 years ago. The Lunette Lake period (9,000–6,000 years ago) has the highest frequency of wetland plants, decreasing significantly during the Bergen and Boulder Village periods (after 6,000 years ago; see Discussion).
Taxa distribution by ecology and chronology. (Color online)
Discussion
Late Pleistocene–Early Holocene Transition
Late Pleistocene and Early Holocene fiber artifacts at Paisley Caves are limited in number but already suggest knowledge of a variety of textile fiber resources represented by sagebrush rope and three-stand braids, tule basketry weft elements, bitterbrush knotted bark, and twisted grass threads. Dogbane and stinging nettle fine cordage, along with sumac stem and juniper bark coarse cordage, are present between 12,000 and 10,000 cal BP. All three-strand braids are made from sagebrush bark (Figures 4, 5, 6). These plants continue to dominate the NGB textile record throughout the Holocene and into the nineteenth century.
During the Bolling/Allerod (>14.7–12.9 ka), deep-bodied lakes filled lowland basins, and Lake Chewaucan was at its highest levels as the Earth warmed at the end of Last Glacial Maximum. Lake-level high stands during the Paisley period were followed by lake contraction during the Younger Dryas cold snap between approximately 12,900 and 11,600 years ago when the Earth's climate returned temporarily to a glacial state (Goebel et al. Reference Goebel, Hockett, Adams, Rhode and Graf2011; Hudson et al. Reference Hudson, Emery-Wetherell, Lubinski, Butler, Grimstead and Jenkins2021). During the Late Pleistocene / Early Holocene transition, expanded wetlands, spring fed lakes, and seasonal ponds overall may have improved conditions considerably for animal and plant resources (Hockett et al. Reference Hockett, Adams, Lubinski, Butler and Jenkins2017), including fiber plants. Riparian and wetland textile fiber well suited for basketry and cordage—such as tule, cattail, rushes, dogbane, and stinging nettle—likely thrived in these new landscapes. This dramatic period in ecological change coincides with the earliest evidence of humans populating the NGB by 14,500 years ago (Goebel et al. Reference Goebel, Waters and O'Rourke2008; Hockett et al. Reference Hockett, Adams, Lubinski, Butler and Jenkins2017; Jenkins et al. Reference Jenkins, Davis, Stafford, Campos, Hockett, Jones and Cummings2012; Shillito et al. Reference Shillito, Whelton, Blong, Jenkins, Connolly and Bull2020) and the Columbia Plateau by approximately 16,000 cal BP (Davis et al. Reference Davis, Madsen, Sisson, Becerra-Valdivia, Higham, Stueber and Bean2022).
Earliest dates for stinging nettle and dogbane fine cordage are 11,900–11,200 years ago, with an increase in the presence of tule and other monocot artifacts after 12,000 years ago. This coincides with expanded wetlands and lowland riparian areas—ideal habitats for these plants. The decline in Lake Chewaucan lake levels (to the immediate southeast of Paisley Caves) during the Younger Dryas likely resulted in expanded wetlands (in place of deep-water lakes) to the north and south as well as around Lake Abert, with steppe shrub habitat moving upland. Although an increase in the presence of wetland and riparian plant fiber artifacts could be due to preservation issues, it may also reflect an increase in or introduction of tule basketry and fine cordage manufacture. Whether this represents cultural innovations or changes in fiber-based technology due to environmental adaptations, or both, is unknown.
Research at Paisley Caves and Connley Caves suggests that a broad-spectrum diet was already established by the Younger Dryas (Hockett and Jenkins Reference Hockett and Jenkins2013; Kennedy Reference Kennedy2018; McDonough et al. Reference McDonough, Kennedy, Rosencrance, Holcomb, Jenkins and Puseman2022), including bison, pronghorn, grouse, rabbit, fish, insects, and a range of plants. The botanical lens in Paisley Cave 2 includes pronghorn and jackrabbit cut bone, and possible fishbone and insects also deposited by humans (Hockett et al. Reference Hockett, Adams, Lubinski, Butler and Jenkins2017). At Paisley Cave 2, human coprolites with dates of 9,620 ± 30 and 10,833 ± 59 BP include edible taxa of the sunflower family (Asteraceae), the mustard family (Brassicaceae), carrot (Apiaceae), Great Basin wild rye (Leymus cinereus), wild buckwheat (Eriogonum sp.), and the evening primrose family (Onagraceae; Taylor et al. Reference Taylor, Hutson, Bryant and Jenkins2020). Uncharred stinging nettle seeds were present in the hearth features and Cave 5 earth oven (Kennedy Reference Kennedy2018); this plant resource was used for nettle cordage during the Younger Dryas. This diverse diet was likely established in tandem with textile technology, including netting suited for rabbit hunting and fishing.
Younger Dryas macrofloral remains from a Paisley Cave 1 hearth suggest a diet of Chenopodium, bunch grass (Poaceae), needlegrass (Achnatherum), Sesuvium, borage (Boraginaceae), and saltbush (Kennedy Reference Kennedy2018). These upland charred seed taxa reflect the seasonal use of the caves during early spring and fall when these plants were most abundant (Jenkins et al. Reference Jenkins, Davis, Stafford, Connolly, Jones, Rondeau, Cummings, Kornfeld and Huckell2016; Kennedy Reference Kennedy2018). It is noteworthy that raw bast material is not present in the unmodified macrobotanical remains (sticks, bark, grass), as is seen at Bonneville Estates Rockshelter (Coe Reference Coe2021), which includes unmodified milkweed, dogbane, and flax phloem fibers. The absence of raw bast fiber material at Paisley Caves may suggest a lack of cord-making activity at the site, which supports the interpretation that the caves were used during the spring and summer months for short periods, with requisite tools being brought to the site rather than being made on site.
Foothill shrub zones and upland areas also provided resources for coarse cordage technology, including sagebrush, bitterbrush, sumac, and juniper. The presence of five cords of juniper, sumac, and bitterbrush during the Younger Dryas—although few in number—illustrate that sagebrush was not the only dryland shrub or upland resource for fiber (Figure 7). Younger Dryas pollen studies of Paisley Cave 2 coprolites include juniper, the sagebrush family (Asteraceae), bitterbrush, alder (Alnus), pine (Pinus), fir (Abies), poplar (Populus), hemlock (Tsuga), and willow (Beck et al. Reference Beck, Bryant and Jenkins2018; Blong et al. Reference Blong, Adams, Sanchez, Jenkins, Bull and Shillito2020; Taylor et al. Reference Taylor, Hutson, Bryant and Jenkins2020)—evidence that these taxa were available for technological use.
Early Holocene and Seasonal Fiber Collection
During the Fort Rock and early pre-Mazama period (11,000–7,600 years ago), populations increased, people were highly mobile, and they traveled long distances based on obsidian sourcing (Connolly Reference Connolly1999). People also accessed a diverse range of habitats, including marshlands, dryland, step shrub zones, and high-elevation forests to maintain a broad-spectrum diet. Textile fiber collection and processing was likely an important component of travel to different ecological zones for seasonal foraging and hunting activities. Winters were spent primarily at lakesides and marshes where fish and waterfowl were available year round. Tule stem, one of the most common monocot basketry and cordage fibers in the Paisley assemblage, is gathered from marshes in the fall when the stems turn brown; stands of tule in standing water could be accessed from dugout canoes. Other wetland fiber plants such as cattail, sedge, willow, and juncus would have been collected in late summer through early fall (Downs Reference Downs1966; Fowler and Fowler Reference Fowler and Fowler1970; Stewart Reference Stewart1939; Wheat Reference Wheat1967). Travel to high elevations in summers for large game hunting of ungulates (antelope, mountain sheep, deer, and bison) included foraging for seeds, roots, and berries, and this may have been an opportunity to collect juniper bark and sumac. The abundance of sagebrush and bitterbrush lowland and upland shrub zones were ideal for collecting wood fuel and bark fiber for textiles.
Dogbane and stinging nettle were well established as the primary bast fiber resources for fine cordage and netting by 11,000 years ago (Figures 5 and 6), and they continued to be used throughout the Holocene and into the contact era. Fine cordage netting was likely used in fowling and fishing year round, and possibly in fall rabbit drives as early as the Late Pleistocene. In addition to Late Pleistocene rabbit bone and carcasses in the Paisley botanical lens (Hockett et al. Reference Hockett, Adams, Lubinski, Butler and Jenkins2017), the Early Holocene site Buffalo Flats in the Fort Rock Basin is known for pit features with a sizeable number of jackrabbit and cottontail bone. These features are interpreted as processing areas following repeated hunts in which rabbits were driven into pits (Oetting Reference Oetting, Melvin Aikens and Jenkins1994). Long linear nets made from dogbane and nettle, ethnographically known to be anywhere from 6 to 121 m (20 to 400 ft.) long, were used in communal hunts to catch rabbits as they were driven from brush (Connolly et al. Reference Connolly, Elizabeth, Barker and J. McCabe2017; Kelly Reference Kelly1932; Wheat Reference Wheat1967).
Based on the continuity and quantity of dogbane and nettle cordage and netting from Paisley Caves, summer and fall activities likely included the essential task of gathering these bast fibers as part of larger foraging activities. Both of these plants grow adjacent to wetlands and in riparian habitats; Apocynum cannabinum, the wetland species of dogbane (Franklin and Dyrness Reference Franklin and Dyrness1988; Gilkey and Dennis Reference Gilkey and Dennis2001; Rhode Reference Rhode2002), was likely targeted over A. androsaemifolium (spreading dogbane) and A. pumilum (low dogbane). The tall stems of A. cannabinum are ideal for extracting long strands of phloem fiber, especially if stands are cut down each fall. Dogbane cordage manufacture would have required repeated visits to wetland stands in the late fall when stems die and turn brown; they could then be cut back to the ground. Their rhizome root structure allows for seasonal return to the same patches. Historically, Northern Paiute peoples collected long, dried dogbane stalks and scraped bark away or crushed and then peeled or scraped away bark. After stalks are hammered or retted, they are split open, and the inner pith is removed (Rhode Reference Rhode2002; Wheat Reference Wheat1967). Stinging nettle can be collected throughout the summer around the periphery of wetlands, near shady spring-fed creeks in the foothills above the basin floor, and as an understory within the Fremont National Forest and Owyhee Mountains (Gilkey and Dennis Reference Gilkey and Dennis2001; OregonFlora 2021).
Middle Holocene Climatic Change
During the Lunette Lake period, vegetation records indicate drought-like conditions, with a shift toward a drier, more arid environment, with lower lake levels and increased marshlands and spring fed basins around 8,000–6,000 years ago. Mazama ash fall likely reduced plant diversity temporarily and contributed to overall population decrease. There was a transition from cave and rockshelter habitation sites to open-air locales closer to permanent lakes (such as Paulina Lake), seasonal springs, and vernal pools, perhaps in response to this mid-Holocene climatic change (Aikens et al. Reference Aikens, Connolly and Jenkins2011; Jenkins et al. Reference Jenkins, Connolly, Aikens, Jenkins, Connolly and Aikens2004; Ollivier et al. Reference Ollivier, Smith and Barker2017). Groups were highly mobile, with ephemeral use of caves sites, and temporary camps around lakeshores.
The textile assemblage from Paisley Caves offers support for this model, with an increase in wetland plants during the Lunette Lake period, including a twisted cattail leaf, tule basketry fragment, and the only identified cattail woven structure dated to 6660–6485 cal BP (Figures 4 and 10). Although this may be due to inflated numbers from fragmentary monocot (primarily tule) and bast artifacts, the increase in wetland taxa is significant. There is also continued use of upland areas, represented by one (undated) juniper cord from a Lunette Lake component, and one flax cord dated to 6295–6115 cal BP. This broader range of upland, dryland, and wetland plants may represent the more transient or temporary use of Paisley Caves during this time (Jenkins et al. Reference Jenkins, Davis, Stafford, Connolly, Jones, Rondeau, Cummings, Kornfeld and Huckell2016), in which people may have brought resources from a more diverse range of habitats.
Previous research has also used textiles as a proxy for illustrating changing land-use patterns during the Middle Holocene. At Last Supper Cave in Nevada, the sequence of dates from Fort Rock, Spiral Weft, and Multiple Warp sandals represents a hiatus during the Middle Holocene, which suggests that the gap in the sequence for these sandals may represent more short-term use of caves and rockshelters (Ollivier et al. Reference Ollivier, Smith and Barker2017). There is also a hiatus in Catlow Twined basketry, likely due to the more ephemeral use of dry caves rather than a reduction in production of these fiber artifacts (Camp Reference Camp2018).
During the Bergen period, Lake Chewaucan shorelines rose slightly, close to historical levels at 1,290 m, as did Lake Abert after 3,900 years ago (Hudson et. al. Reference Hudson, Emery-Wetherell, Lubinski, Butler, Grimstead and Jenkins2021). This general trend toward a wetter and cooler climate has continued to the present (Aikens et al. Reference Aikens, Connolly and Jenkins2011) with oscillating dry and wet climatic periods. At Connley Caves, plant taxa present in human coprolites dating to the Bergen period indicate seasonal use of uplands and continued reliance on wetlands, including juniper and sumac upland plants, and cattail and tule seeds (McDonough Reference McDonough2019). Year-round village sites—such as the Bergen Site, Big M Village, and DJ Ranch—were adjacent to wetlands, ponds, springs, and lakeshores, allowing for continued access to wetland and riparian fiber and dietary resources. Catlow Cave and Roaring Springs Cave, known for their rich and diverse textile assemblage, are also situated close to lakes and springs.
There is a significant increase in fine cordage and tule cordage and basketry in the Paisley Caves assemblage during this period. However, the slight decrease in wetland fiber plants following the Lunette Lake period may indicate that fewer marsh resources were brought out of lowlands to temporary cave sites, and more upland plants may have been brought to seasonal camps such as Paisley Caves, away from lowland settlements. No upland textile fibers are represented at Paisley during the Bergen phase; however, two sumac artifacts remain undated and from disturbed components, and upland plants overall (sumac, juniper, flax) represent less than 3% of the entire assemblage.
Late Holocene Upland Resources
Unexpected results of this study include the presence of milkweed, flax, and cotton fine cordage dating to the Late Holocene. Though limited, these fibers make up 8% of all Late Holocene fine cordage (milkweed and flax just 5%). During this period, people increasingly use upland areas for root gathering due to cyclical wet and dry periods, reduction in marshlands, and increases in population (Brashear Reference Brashear, Melvin Aikens and Jenkins1994; Byram Reference Byram, Melvin Aikens and Jenkins1994; Jenkins Reference Jenkins, Melvin Aikens and Jenkins1994; Prouty Reference Prouty, Melvin Aikens and Jenkins1994) centered around Warner Valley, Chewaucan Basin, and Klamath Basin. An expanded resource area during this period may relate to increased populations and dispersal in the NGB generally. The acquisition of cotton cordage and plaited fabric radiocarbon dated within the last 1,000 years may reflect increased interactions with exterior groups. Dietary plants are also most diverse during this time; at LSP-1 Rockshelter (Kennedy and Smith Reference Kennedy and Smith2015), people may have been motivated by nutritional benefits, not only caloric return based on the diversity of dietary plants including Chenopodium sp., saltbush, and mustards, as well as marsh foods such as cattail, and upland resources including chokecherry seeds.
Flax, a high-elevation plant, was likely collected during visits to upland root-gathering areas, such as Boulder Village, the largest spring root collecting site (Byram Reference Byram, Melvin Aikens and Jenkins1994). Flax prefers drought-like conditions and mid to high elevations in association with sagebrush, bitterbrush, and pinyon-juniper habitats between 1,370 and 2,290 m (4,500 and 7,500 ft.) in elevation (Rhode Reference Rhode2002; OregonFlora 2021; United States Department of Agriculture 2021). Juniper bark was also used for cordage during this time, and upland sites show significant use of juniper for structural material (Stenholm Reference Stenholm, Melvin Aikens and Jenkins1994). Milkweed at Paisley Caves only makes up a total of four identified cords, but like dogbane, milkweed is also gathered in late summer or fall from dry open meadows and rocky slopes or in areas with moist, sandy, loamy soils (Rhode Reference Rhode2002).
Conclusions
This research provides long-term data on culturally significant native plants, with continuity in plant selection spanning the last 14,000 years. During the Late Pleistocene / Early Holocene transition, the archaeological record suggests a cultural “explosion” with Western-stemmed lithic technology; basketry and fine cordage; bone needles; and a broad-spectrum diet including waterfowl, rabbits, fish, artiodactyls, seeds, and roots (Hockett et al. Reference Hockett, Adams, Lubinski, Butler and Jenkins2017; Rosencrance et al. Reference Rosencrance, Smith, Jenkins, Connolly and Layton2019). This cultural expansion included knowledge of several key textile plants, including tule, sagebrush, juniper, bitterbrush, stinging nettle, and dogbane. Expanded marshlands during the Late Pleistocene and Early Holocene created suitable plant communities ideal for fiber technology—specifically wetland monocots (tule, sedges) and herbaceous dicots including dogbane and stinging nettle. This technology is key to subsistence activities and craft production throughout the Holocene.
By the Early Holocene, a limited group of plants ideal for strong string fiber, rope, and basketry are well established; there is continued reliance on this select group of taxa throughout the Holocene, whereas the near absence of cattail, juncus, reed, and willow for fiber at Paisley is notable. Despite climatic events during the Lunette Lake period, in which people transitioned from caves to sites centered around lakeshores and wetlands (a trend exhibited in the increased number of wetland plants in the Paisley assemblage), the suite of fiber plants and their technological application remains constant. This overall continuity parallels what we know about dietary plants; much of the taxa present in the Late Pleistocene diet from Paisley Caves is present in the Middle Holocene diet at Connley Caves. At the Paisley Caves, dogbane, nettle, and tule fiber were used for millennia and are primary fibers in nineteenth-century Klamath baskets. During the Late Holocene, bast fiber material diversified with the addition of flax and milkweed. The presence of flax in particular—a high-elevation plant—may reflect the increased use of upland root-collection areas as populations increased.
The Great Basin is a region rich in textiles made from plant material; this study reveals new and important aspects of people's lives that have largely been unexplored archaeologically. Textile plant identification allows a more nuanced understanding of how people engaged with diverse and changing landscapes, illustrates the importance of fiber collection as part of a larger system of settlement and subsistence practice, and complements ongoing research in dietary studies. In the Northern Great Basin, ecological knowledge required for fiber technology was established by 14,000 years ago and continues today, exemplified through traditional arts and technology.
Acknowledgments
Thank you to Suzana Radivojevic, Kathryn Puseman, and Linda Scott Cummings for assistance with macrobotanical analysis. Special thanks to Anne Lawlor, Marion Coe, and Hana Lukešová for sharing their expertise in bast fiber analysis. Richard Rosencrance oversaw radiocarbon sample preparation. Kurt Langworthy oversaw SEM and EDS instrumentation and data interpretation; Sandra Koch provided additional EDS data on milkweed and flax reference samples. Thomas Connolly and Christopher Ruiz assisted with textile images. Thank you to my committee members Thomas Connolly, Katelyn McDonough, and Scott Fitzpatrick for their insight and editorial comments, and to Dennis Jenkins for his support. Rosemary Ruiz generously provided a Spanish language abstract translation.
Funding Statement
This research was supported in part by the National Science Foundation Grant No. 2232056, the University of Oregon Museum of Natural and Cultural History, the Oregon Archaeological Society, and the Nevada Association of Archaeology and Am-Arcs of Nevada.
Data Availability Statement
All data used in the analyses are available online at tdar.org upon request from the author.
Competing Interests
The author declares none.
