Worldwide, some of the earliest coastal archaeological sites are found in Mediterranean climatic zones (e.g., Dillehay et al. Reference Dillehay, Ramírez, Pino, Collins, Rossen and Pino-Navarro2008; Erlandson et al. Reference Erlandson, Rick, Braje, Casperson, Culleton, Fulfrost, Garcia, Guthrie, Jew, Kennett, Moss, Reeder, Skinner, Watts and Willis2011; Marean et al. Reference Marean, Bar-Matthews, Bernatchez, Fisher, Goldberg, Herries, Jacobs, Jerardino, Karkanas, Minichillo, Nilssen, Thompson, Watts and Williams2007; Stiner Reference Stiner1994). Geophytes are well adapted to the seasonal nature of precipitation in such areas, and they were often important sources of calories and carbohydrates (Anderson Reference Anderson2016; De Vynck et al. Reference De Vynck, Cowling, Potts and Marean2016; Larbey et al. Reference Larbey, Mentzer, Ligouis, Wurz and Jones2019; Rundel Reference Rundel1996). Evidence for the early human use of edible geophytes has grown exponentially in recent years. In South Africa, Wadley and colleagues (Reference Wadley, Backwell, d'Errico and Sievers2020) identified carbonized remains of starchy rhizomes in Border Cave strata dated to approximately 170,000 years ago, and Larbey and colleagues (Reference Larbey, Mentzer, Ligouis, Wurz and Jones2019) documented geophytes in hearth features at Klasies River Mouth Caves dated between approximately 120,000 and 65,000 years ago. In Southeast Asia, Barker and colleagues (Reference Barker, Hunt, Barton, Gosden, Jones, Lloyd-Smith, Farr, Nyirí and O'Donnell2017) reported the remains of geophytes (yam, taro) in levels of Niah Cave dated to as much as 50,000 years ago. In Australia, Florin and colleagues (Reference Florin, Fairbairn, Nango, Djandjomerr, Marwick, Fullagar, Smith, Wallis and Clarkson2020) described geophyte remains in the basal levels of Madjedbebe Rockshelter dated to about 53,000 years ago. In Chile, Ugent and colleagues (Reference Ugent, Dillehay and Ramirez1987) identified the remains of wild potato (Solanum maglia) from hearths at the roughly 14,000-year-old Monte Verde II site, and Louderback and Pavlik (Reference Louderback and Pavlik2017) identified starch grains from another wild potato (S. jamesii) on ground stone tools from an approximately 10,500-year-old stratum at Utah's North Creek Shelter.
These studies suggest that as early modern humans spread around the globe, including some of the earliest Native Americans, they were fully aware of the food potential of geophytes. Despite growing evidence for their widespread and ancient use, some archaeologists have argued that geophytes were relatively low-ranked resources. In this article, we report on carbonized geophyte and other plant remains recovered from an approximately 11,500-year-old (cal BP) cultural feature at CA-SRI-997/H on Santa Rosa Island. First, however, we describe the modern and ancient settings of the site, along with the context of the feature from which botanical remains were recovered. After presenting archaeobotanical data from CA-SRI-997/H and documenting their antiquity, we discuss archaeological and ethnographic evidence for geophyte use on California's islands and then present experimental return rates from blue dicks (Dipterostemon capitatus [Benth.] Rydb., [Dichelostemma capitatum (Benth.) Alph. Wood]) corms on Santa Cruz Island, arguing that they were a high-ranked plant food on the islands for millennia.
California's Northern Channel Islands
Today, the Northern Channel Islands (NCI) are located 19–44 km off the California coast (Figure 1). Terrestrial mammals endemic to the islands are limited to the island fox (Urocyon littoralis), spotted skunk (Spilogale gracilis amphialus), and a deer mouse (Peromyscus maniculatus), but marine ecosystems around the NCI are highly productive, and they support a diverse array of shellfish, fish, marine mammals, and birds. Island plant communities have long been characterized as relatively impoverished, but Gill and colleagues (Reference Gill, Erlandson, Niessen, Hoppa, Merrick, Gill, Fauvelle and Erlandson2019) demonstrated that their diversity is higher than mainland areas of comparable size. Modern Santa Rosa Island has a mosaic of oak and pine forests, chaparral and coastal scrub, and extensive grassland/prairie communities. Geophytes are found in most of these communities, but they are particularly abundant in grasslands/prairies, which the Island Chumash maintained through regular landscape burning for millennia (Anderson et al. Reference Anderson, Starratt, Brunner-Jass and Pinter2010; Timbrook et al. Reference Timbrook, Johnson and Earle1982).
Figure 1. Map of the Northern Channel Islands, including approximate paleoshorelines of Santarosae around the time of occupation at CA-SRI-997/H and locations of experimental return-rate harvests (figure by Kristina Gill).
The NCI have produced some of the earliest evidence for island colonization, seafaring, and marine resource use in the Americas (Erlandson et al. Reference Erlandson, Rick, Braje, Casperson, Culleton, Fulfrost, Garcia, Guthrie, Jew, Kennett, Moss, Reeder, Skinner, Watts and Willis2011). Several terminal Pleistocene sites provide evidence that Paleocoastal (maritime Paleoindian) peoples harvested shellfish, fish, birds, and marine mammals between 12,200 and 8000 cal BP (Gusick and Erlandson Reference Gusick, Erlandson, Gill, Fauvelle and Erlandson2019). Archaeobotanical research on plant food use by the Island Chumash and their predecessors has been limited, but recent evidence for intensive geophyte use has come from a growing number of sites. Reddy and Erlandson (Reference Reddy and Erlandson2012) reported carbonized geophyte remains (Brodiaea-type corms) from several strata at Daisy Cave (CA-SMI-261) dated between approximately 10,000 and 3,000 years ago. Gill (Reference Gill2014, Reference Gill2015, Reference Gill2016) identified carbonized Brodiaea-type corms from another roughly 10,000-year-old shell midden (CA-SMI-522) on San Miguel Island, as well as an approximately 8,000-year-old site (CA-SRI-666) on Santa Rosa Island (Erlandson, Rick, et al. Reference Erlandson, Rick, Ainis, Gill, Jew and Reeder-Myers2019), and documented heavy reliance on these geophytes at CA-SCRI-619/620 on Santa Cruz Island spanning the past 6,000 years. Other researchers have also identified Brodiaea-type corms at several coastal village sites on Santa Cruz Island occupied within the last approximately 3,000 years (Martin and Popper Reference Martin, Popper and Arnold2001; Thakar Reference Thakar2014).
CA-SRI-997/H: Setting, Methods, and Results
From the Last Glacial Maximum to about 9,500 years ago, the NCI were connected as an island known as Santarosae (Reeder-Myers et al. Reference Reeder-Myers, Erlandson, Muhs and Rick2015). Along the south shore of this roughly 125 km long island was a semiprotected embayment known as Crescent Bay. Several of the largest drainages on Santarosae flowed into this bay and a deep submarine canyon offshore. With ample freshwater and riparian, marsh, estuarine, and marine habitats, Crescent Bay was attractive to early islanders—more than 25 Paleocoastal sites have been recorded in upland areas nearby (Gusick and Erlandson Reference Gusick, Erlandson, Gill, Fauvelle and Erlandson2019).
The largest of these is CA-SRI-997/H, extending for nearly 300 m along the lower reaches of Cherry Canyon. Currently located about 150 m from the coast, CA-SRI-997/H was 4–6 km from the shore of Crescent Bay near the end of the Pleistocene, when sea levels were around 75 m lower than today (Figure 1). The site is located on a raised marine terrace, on which fluvial sediments accumulated during the terminal Pleistocene. The landform appears to have stabilized approximately 12,000 years ago, when Paleocoastal peoples occupied four sites on the rim of Cherry Canyon, including CA-SRI-997/H (Erlandson, Braje, and Gill Reference Erlandson, Braje and Gill2019). Starting around 8,000 years ago, the Paleocoastal surface was gradually buried under a light-gray silty A1 soil (Figure 2C) that probably accumulated during the mid-to-late Holocene from the settling of airborne dust carried from mainland sources (Muhs et al. Reference Muhs, Budahn, Reheis, Beann, Skipp and Fisher2007).
CA-SRI-997/H was identified when National Park Service archaeologists monitoring construction found flaked stone crescents and Channel Island Barbed (CIB) points diagnostic of Island Paleocoastal assemblages (Kulaga Reference Kulaga2020). Field surveys and excavations revealed a large Paleocoastal site with two loci. Much of Locus A was disturbed by historical ranching activities, and it produced artifacts ranging from Paleocoastal to historical times (Braje and Erlandson Reference Braje, Erlandson, Gill and Brown2019). In Locus B, Paleocoastal artifacts were found associated with a buried A2 paleosol (Figure 2C), patches of which were exposed where erosion had removed the overlying sediments (Figure 2A).
Figure 2. Archaeological investigations at CA-SRI-997/H, Locus B: (A) Location of Unit DR-1, facing north, where A2 soils were exposed on an eroded road surface (Jon Erlandson, left; David Paul Dominguez, Chumash, right); (B) Unit DR-1 excavation showing CS1 and CS2 in relation to hearth-like feature with burned rocks, heat-treated lithics, and intact A2 soils; (C) Composite profile of Unit DR-1's west wall, showing the depth of Holocene A1 soil (in the road cut adjacent to the unit) that overlay the Paleocoastal A2 anthropogenic stratum. Dashed line indicates a gradual transition between the A2 and B2 soil horizons (photos by Todd Braje; profile and figure by Kristina Gill).
Excavations in Locus B included two contiguous 1 × 1 m units (DR-1 and DR-7) placed in a dark anthropogenic soil exposed in a now abandoned ranch road. DR-1 was placed over Feature B1, a darker soil with a cluster of flaked stone artifacts that included a crescentic biface preform and 15 flakes of an igneous porphyry that appeared to be from a single core. Most of these artifacts were embedded in the dark paleosol, suggesting that Feature B1 was largely intact. Units DR-1 and DR-7 produced more than 2,000 chipped stone artifacts—including crescents and a CIB point most common in NCI sites dating between approximately 12,000 and 11,000 cal BP—associated with burned rock and charcoal. Two soil columns were collected from the southwest (CS1, 25 × 25 cm) and southeast (CS2, 20 × 25 cm) corners of Unit DR-1 (within and just outside of Feature B1, respectively; Figure 2B) in 10 cm levels, with all sediment retained for flotation.
Archaeobotanical Methods
Gill processed and analyzed six flotation samples (total volume = 41 L) from the soil columns. Samples larger than 5 L were split into two separate flotation events. Each sample was floated in freshwater using manual bucket flotation (see Pearsall Reference Pearsall2000), which decanted and poured light fraction materials (that float or are suspended in water) onto 0.4 mm mesh. Deflocculation with borax was required for the lower levels (20–30 and 30–40 cm) due to higher clay content. Each sample was decanted more than three times and agitated between pours until no suspended charcoal or other plant remains were observed. Light-fraction and heavy-fraction (which do not float) materials were air dried, the latter on 1.0 mm mesh.
Dry light-fraction materials were size sorted over nested sieves into 2.0 mm, 1.0 mm, 0.5 mm, and <0.5 mm subsamples. Wood charcoal was pulled from the 2.0 mm and 1.0 mm fractions; geophyte fragments, nutshell, and glassy carbon materials were pulled from 0.5 mm fraction and larger. Other charred botanical material, including small seeds and seed fragments, as well as termite coprolites, were pulled from all size fractions. Uncarbonized plant material was not analyzed, and unidentified wood charcoal was retained for future analysis. Most botanical remains came from the light fraction, but a few carbonized fragments were recovered from the heavy fraction. Macrobotanical identifications were made using Gill's comparative collection and Martin and Barkley's (Reference Martin and Barkley2000) Seed Identification Manual, classifying plant remains to the most specific taxonomic category possible. Raw counts and density data for each taxon are presented in Table 1; with the exception of wood charcoal, weights were all <0.01 g. Given the small number of samples, robust quantitative analyses were not possible.
Table 1. Carbonized Archaeobotanical Remains Recovered from CA-SRI-997/H, DR-1.
a Total soil volume.
b Fragments.
c Weight (g).
d Weight (g/L).
The identified archaeobotanical remains appear to be cultural in origin, with both food and fuel taxa represented. Excellent stratigraphic integrity is common in many NCI archaeological sites due to a dearth of burrowing animals (e.g., gophers, ground squirrels). Analysis of off-site soil samples could ensure that no “background noise” from naturally occurring carbonized plant material was present. The surrounding area contains virtually continuous archaeological soils for hundreds of meters, however, and the A2 paleosol is almost entirely obscured by the A1 soil.
Results
Archaeobotanical remains from CA-SRI-997/H indicate that Paleocoastal peoples utilized plants from a range of habitats, including coniferous forest, chaparral/scrub, and grassland/prairie. Eight plant taxa were identified to genus and/or species level. Food plants recovered include geophytes, nutshell, and seeds (Table 1; Figure 3). The two geophyte taxa identified, Brodiaea-type corm (n = 4) and Calochortus bulb (n = 9) fragments, are most abundant in grasslands/prairies today but are also found on rocky slopes in chaparral, coastal scrub, and other habitats. All the geophyte fragments were recovered from CS1 within Feature B1, suggesting that the feature was a hearth or earth oven in which corms and bulbs were roasted. Morphological similarity within the Brodiaea complex precludes identification to species, but the corms are most likely blue dicks (Dipterostemon capitatus [Benth.] Rydb. [Dichelostemma capitatum (Benth.) Alph. Wood]) based on modern distribution, abundance, and fecundity (Gill Reference Gill2014, Reference Gill2016).
Figure 3. Selected archaeobotanical remains and modern geophytes: (A1) Carbonized Brodiaea-type corm (~4,500 years old) and (A2) carbonized Calochortus bulb (~1,500 years old) from the Diablo Valdez site (CA-SCRI-619/620) on Santa Cruz Island; (B1) Manzanita berry pit (Arctostaphylos spp.) and (B2) termite coprolites (~11,500 years old) from CA-SRI-997/H; (C1) Modern Brodiaea-type blue dicks corms (Dipterostemon capitatus) and (C2) Catalina mariposa lily (Calochortus catalinae) bulbs from western Santa Cruz Island, with outer fibrous coatings removed (photos by Kristina Gill).
The genus Calochortus had not previously been reported in Channel Island archaeobotanical assemblages, but it was recently identified by Gill in Late Holocene deposits at CA-SCRI-619/620 on Santa Cruz Island. Two Calochortus species occur on Santa Rosa Island today: Globe lily (C. albus [Benth.] Douglas ex Benth) and Catalina mariposa lily (C. catalinae [S.] Watson). Globe lily occurs in rocky north-facing slopes, chaparral, coastal scrub, oak woodland, and pine forest, whereas the Catalina mariposa lily occurs in grasslands, rocky slopes, and coastal scrub (Junak et al. Reference Junak, Ayers, Scott, Wilken and Young1995). Recent observations on the Channel Islands and mainland coast have noted blue dicks and Catalina mariposa lily growing together in coastal grasslands/prairies.
Other food plants identified in the flotation samples include a manzanita (Arctostaphylos spp.) berry pit (n = 1), nutshell from a thin-shelled pine (n = 2; most likely Bishop pine, Pinus muricata D. Don), and a Camissonia seed (n = 1), a genus with edible greens. Manzanita grows primarily in chaparral/scrub habitat and produces berries with hard, edible pits that are rich in carbohydrates and were eaten for millennia by the Island Chumash (Gill and Hoppa Reference Gill and Hoppa2016). The carbonized remains of fuel plants, including wood charcoal, pine-cone fragments (n = 2), wild cucumber (Marah macrocarpa [Greene] Greene) seed fragments (n = 13), and a seed from the bean family (Fabaceae; n = 1) were also recovered. The bean family includes edible taxa, but a number of taxa are inedible and could have served as fuel. Wild cucumber seeds, nearly ubiquitous in NCI archaeological sites, are toxic if eaten, but the Chumash used them to make beads, bind pigments, treat baldness, and contain the spiritually potent mixture ‘ayip (Martin Reference Martin2009; Timbrook Reference Timbrook2007). The large seeds also have a high oil content and may have been used as tinder/kindling (Gill Reference Gill2015). Lady's mantle (Aphanes occidentalis [Nutt.] Rydb; n = 1), identified in some Channel Island archaeological sites, may represent natural seed rain, given that it has no known ethnographic use (Gill and Hoppa Reference Gill and Hoppa2016).
Bishop and Torrey pines (P. torreyana [Parry] ex Carrière ssp. insularis J. R. Haller) are found in the vicinity today, but pines and other conifers were more diverse and abundant on the islands until the terminal Pleistocene. The pine identified from CA-SRI-997/H includes nutshell and pine-cone scale fragments that compare favorably with those of Bishop pine, which produces small, thin-shelled nuts. Bishop pine cones are relatively small (~5–10 cm long) and serotinous (opening in the presence of fire), so the identified nutshell could result from the use of whole pine cones as fuel rather than a food source. Erlandson (Reference Erlandson1994:261) proposed that the larger thick-shelled Torrey pine nuts may have been an important food for early NCI peoples, but no Torrey pine remains were identified in the CA-SRI-997/H samples, despite the presence of a relict Torrey pine forest just 3 km to the south. Torrey pine nut remains have been identified archaeologically at CA-SCRI-333 on Santa Cruz Island, dated between 6,000 and 2,500 years ago (Gamble Reference Gamble2017). Evidence for the use of pine nuts on the NCI is very limited overall, especially compared to geophytes and small seeds (see Gill and Hoppa Reference Gill and Hoppa2016).
Stratigraphy, Soils, and Chronology
Sediment samples from the four 10 cm levels of the A2 paleosol deposits in Unit DR-1 were analyzed for grain size, organic content, and pH. The soil contains high levels of sand, especially in the upper levels, which decrease with depth as the percentages of silt and clay increase. Aside from level 3 (20–30 cm; pH = 6.3), the pH of the soil samples was neutral (7.0) or slightly alkaline (8.3). The soil pH suggests that the lack of marine shell and animal bone in Locus B is not due to soil acidity, which is supported by the presence of a small 8,000-year-old feature containing well-preserved marine shells in the same A2 soil as in Locus A, ~90 m to the south (Braje and Erlandson Reference Braje, Erlandson, Gill and Brown2019).
The western part of Unit DR-1 contained a concentration of flaked stone artifacts, burned rock, and charcoal-stained soil in Feature B1, which appears to have been a hearth or earth oven. Unit DR-1 produced 2,170 flaked stone artifacts, including 25 bifaces (4 crescentics, 1 CIB point, and 20 undifferentiated fragments), 6 flake tools, 1 core, 14 cobbles, and 2,124 pieces of debitage. Most of the 25 bifaces showed evidence of heat spalling that probably resulted from intentional heat treatment (see Jew and Erlandson Reference Jew and Erlandson2013).
We obtained six AMS 14C dates from carbonized organic materials identified in flotation samples from the lower three levels of Feature B1 (soil column CS1): two for individual pine-cone scale fragments, three for single geophyte bulb and corm fragments, and one for a clump of three termite coprolites probably derived from burned fuel wood (Table 2). Single fragments of pine cone and geophytes are short-lived samples that should not be affected by “old
Table 2. AMS 14C Chronology for Feature B1 (Unit DR-1, Soil Column 1).
Notes: All samples were analyzed at UC Irvine's Keck Carbon Cycle AMS facility. All dates were calibrated with the Calib 8.1.0 program (Stuiver et al. Reference Stuiver, Reimer and Reimer2021) and IntCal20 dataset for terrestrial samples (Reimer et al. Reference Reimer, Austin, Bard, Bayliss, Blackwell, Ramsey, Butzin, Cheng, Lawrence Edwards, Friedrich, Grootes, Guilderson, Hajdas, Heaton, Hogg, Hughen, Kromer, Manning, Muscheler, Palmer, Pearson, van der Plicht, Reimer, Richards, Marian Scott, Southon, Turney, Wacker, Adolphi, Büntgen, Capano, Fahrni, Fogtmann-Schulz, Friedrich, Köhler, Kudsk, Miyake, Olsen, Reinig, Sakamoto, Sookdeo and Talamo2020).
wood” effects (Schiffer Reference Schiffer1986). Five of the dates have calibrated age ranges between approximately 11,700 and 11,300 cal BP, with a mean age of approximately 11,450 cal BP, including the termite coprolites. A sixth date, for a carbonized Calochortus bulb fragment from the lowest excavation level, produced a somewhat younger date (9405 ± 30) and calendar age (~10,600 cal BP), which might identify a later Paleocoastal occupation of Locus B. Given its context at the base of an intact cultural feature, however, the younger date seems more likely to be a statistical outlier. Together, the 14C dates from Feature B1 suggest that the Paleocoastal assemblage in this part of Locus B results from occupation roughly 11,500 years ago, an age consistent with the artifacts recovered in and around the feature. Feature B1 at CA-SRI-997/H appears to have been a large hearth or earth oven, although it appears to contain fewer rocks than later earth ovens on the NCI.
Discussion
Ethnographic and Archaeological Evidence for Intensive NCI Geophyte Use
In California, except for the Channel Islands, archaeobotanical remains of Brodiaea-type geophytes are generally found in relatively low densities ranging from the San Francisco Bay and Sacramento Delta areas, the Central Valley, Owens Valley, and the south coast (e.g., Reddy Reference Reddy2016; Wohlgemuth Reference Wohlgemuth2016). The higher densities of geophyte archaeobotanical remains on the islands may be largely due to a lack of burrowing animals and excellent stratigraphic integrity on the islands, given that blue dicks corms and mariposa lily bulbs were noted as important geophyte food plants throughout early California ethnographies (Anderson Reference Anderson1997, Reference Anderson2005; Timbrook Reference Timbrook2007). In Chumash areas, blue dicks and mariposa lily geophytes were eaten raw, either cooked directly in the hot ashes of a cooking fire or roasted in earth ovens directly between two layers of hot coals covered with earth (Anderson Reference Anderson2005; Gill Reference Gill2015; Timbrook Reference Timbrook2007:75).
Fernando Kitsepawit Librado, a Chumash consultant who worked with ethnographer John P. Harrington in the early 1900s, indicated that blue dicks (shiq'o’n) were particularly important on the islands, with several families involved in harvesting and cooking large quantities in earth ovens measuring more than a meter across (Timbrook Reference Timbrook2007). Gill (Reference Gill2014, Reference Gill2015) identified numerous carbonized Brodiaea-type corms associated with two earth ovens dating to approximately 4,500 years ago, as well as from Late Holocene hearth-clearing pit features at CA-SCRI-619/620 on Santa Cruz Island. The earth ovens identified at CA-SCRI-619/620 had a pavement of rock at the base, which Feature B1 at CA-SRI-997H did not have, possibly indicating that the feature was a large hearth or that early earth ovens were built without a rock base. Geophyte remains have also been recovered from general midden contexts, where archaeobotanical data from 22 NCI sites spanning the Holocene indicate that Brodiaea-type corms were the single most ubiquitous plant food taxon recovered archaeologically (see Gill and Hoppa Reference Gill and Hoppa2016 for a summary of eight studies).
In early Spanish accounts, blue dicks corms are referred to as cacomites, and mariposa lily bulbs were described as “another kind of cacomite” (Timbrook Reference Timbrook2007:49). For Santa Catalina Island, an AD 1602 Spanish account noted two different geophytes used by Native peoples: “On the island there is a great quantity of something like potatoes, and small xicamas which the Indians carry to the mainland to sell” (Ascension and Wagner Reference Ascension and Wagner1928:351). Both “cacomites” and “xicamas” are names derived from Nahuatl words (Molina Reference Molina1571): cacomiti, likely referring to similar geophytes (Muilla complex) that are abundant in Mexico (Gill Reference Gill2014), and xicamatl, referring to the Mexican jicama. Brodiaea-type corms are very starchy and were commonly called “Indian potatoes” throughout the West (Anderson Reference Anderson1997, Reference Anderson2005; Todt Reference Todt1997). The Mexican jicama (Pachyrhizus erosus [L.] Urb.) is a large root typically eaten raw, with a distinctive mildly sweet, nutty taste. The geophytes called “small xicamas” may have been Catalina mariposa lily bulbs, which taste strikingly similar when eaten raw.Footnote 1
Evidence from CA-SRI-997/H suggests that both geophyte types were used, yet Brodiaea-type geophytes are better represented archaeologically later in time. Blue dicks are especially fecund, with each parent corm producing upward of 15 cormlets annually, whereas the Catalina mariposa lily reproduces less readily and is rare in California today. Fecundity differences may explain why island peoples focused primarily on Brodiaea-type corms, although both types respond well to traditional disturbance regimes, which include frequent fires, tilling, and replanting (Anderson and Lake Reference Anderson and Lake2016; Dolman Reference Dolman2016). Alternatively, Catalina mariposa lily bulbs may be underrepresented archaeologically, especially if they were more typically eaten raw, as documented historically among several southern California tribes (Anderson Reference Anderson1997, Reference Anderson2005; Anderson and Lake Reference Anderson and Lake2016).
For edible geophytes, Anderson (Reference Anderson1997:158) listed five practices Indigenous Californians used to “ensure future bulb and corm production”: replanting of bulblets or cormlets, sparing whole plants, harvesting after plants went to seed, landscape burning, and irrigation. Widespread cessation of these Indigenous practices after European colonization—along with the introduction of exotic livestock and plants, fire suppression, and urban development—has sharply reduced the productivity of many California geophytes (Imper Reference Imper2016). Despite such impacts, blue dicks and other geophytes are often characterized as relatively low-ranked plant resources compared to acorns, pine nuts, and other plant foods (e.g., Rosenthal and Hildebrandt Reference Rosenthal and Hildebrandt2019; Ugan and Rosenthal Reference Ugan and Rosenthal2016).
Experimental Return Rates: Blue Dicks as High-Ranked Resources on the NCI
On the NCI today, blue dicks appear to be much more abundant than on the mainland, and their corms are considerably larger (Gill Reference Gill2014; Gill and Hoppa Reference Gill and Hoppa2016). Their abundance, size, nearly ubiquitous distribution, low processing costs, and availability during multiple seasons made them a top-ranked plant food on the islands. To calculate their return rates, Gill and Erlandson harvested blue dicks corms on western Santa Cruz Island (Table 3) using a weighted digging stick (except for the southwest coast, where trowels were used). Guided by ethnographic accounts for late spring–early summer harvest (Anderson Reference Anderson1997:154; Anderson and Rowney Reference Anderson and Rowney1999; Timbrook Reference Timbrook2007:75) and archaeobotanical evidence for fall harvest (Gill Reference Gill2014), experimental harvests were conducted in spring, late summer, and fall. Following traditional harvest protocols to replant the cormlets, each harvest event included removal of the fibrous outer coating of each corm collected, with cormlets removed and replanted.
Table 3. Return Rates for Experimental Blue Dicks Harvests on Western Santa Cruz Island.
a Locations for harvest events shown in Figure 1; soils ranged from hard, clayey, and difficult to dig (West End) to silty and moderately stony (Centinela) to easily dug sandy silt (SW Coast).
b Values based on mean proximate nutritional analysis for various Brodiaea-type corms, including blue dicks (Gilliland Reference Gilliland1985). Return rates are per person hour.
Our experimental return rates for blue dicks (mean 890.7 kcal/hr) are 10 times higher than rates that Ugan and Rosenthal (Reference Ugan and Rosenthal2016) reported for the Mojave Desert (mean 88.3 kcal/hr). This disparity should caution archaeologists from extrapolating experimental return rates between regions. Experiments repeated with experienced harvesters, moreover, in areas regularly managed in traditional ways, would likely result in even higher return rates. Our caloric return rates for blue dicks corms are also comparable to acorn return rates on California's mainland (Basgall Reference Basgall1987; Bettinger and Wohlgemuth Reference Bettinger, Wohlgemuth and Ubelaker2006).
In many island and coastal environments where calories, fats, and complete proteins are often obtained from abundant shellfish, fish, birds, and marine mammals, the caloric value of starchy plant foods may not be the primary factor in foraging decisions (Erlandson Reference Erlandson1988; Gill et al. Reference Gill, Erlandson, Niessen, Hoppa, Merrick, Gill, Fauvelle and Erlandson2019). For coastal people who rely heavily on shellfish and other lean meats, in particular, carbohydrates are essential for effectively metabolizing such meats. The mean carbohydrate (CHO) return rate for NCI blue dicks is 202.3 CHO/hour, double the recommended daily requirement (100 g) for modern children and nonlactating adults (Institute of Medicine 2005:289). Because NCI archaeobotanical assemblages show minimal use of acorn or pine nut and ubiquitous use of geophytes, it seems likely that the carbohydrate value of geophytes was critical to Islander diets.
Historical Degradation and Recent Florescence of NCI Geophytes
Historically, the diversity and productivity of Channel Island plant communities were severely affected by heavy grazing that started before systematic botanical surveys (Junak et al. Reference Junak, Ayers, Scott, Wilken and Young1995). Modern fire suppression has allowed chaparral and coastal scrub communities to crowd out many island prairies where geophytes were abundant for millennia. Nonetheless, recent removal of introduced herbivores—along with the lack of gophers, rabbits, and other small herbivores on the islands—has enabled a florescence of blue dicks and other geophytes on the NCI, with annual superblooms even during a severe four-year drought from 2014 to 2018. Each spring, NCI grassland/prairie habitats literally turn blue or purple as millions of blue dicks flower, marking root grounds of phenomenal abundance. Ethnographic, archaeological, and ecological data now all point to geophytes as a rich source of carbohydrates and calories that complemented the wealth of marine resources found in island waters (Gill et al. Reference Gill, Erlandson, Niessen, Hoppa, Merrick, Gill, Fauvelle and Erlandson2019).
Conclusions
Archaeobotanical evidence from CA-SRI-997/H documents the use of edible geophytes by Paleocoastal people on Santarosae 11,500 years ago—currently the earliest evidence for geophyte use in North America. When combined with aquatic resources in island, coastal, or other settings, such carbohydrate-rich plant foods can be an optimal solution to meeting energy, protein, and essential micronutrient requirements (Erlandson Reference Erlandson, Cunnane and Stewart2010; Gill et al. Reference Gill, Erlandson, Niessen, Hoppa, Merrick, Gill, Fauvelle and Erlandson2019). In our view, archaeobotanical data from CA-SRI-997/H and other early NCI sites are consistent with Florin and colleagues’ (Reference Florin, Fairbairn, Nango, Djandjomerr, Marwick, Fullagar, Smith, Wallis and Clarkson2020) conclusion that “botanical knowledge . . . contributed to the adaptability and flexibility required by EMH populations to traverse continents and colonise new environments around the world.” The deep history and economic importance of geophyte use by Island Paleocoastal peoples adds to a growing body of archaeobotanical data from around the world that geophytes were important food resources for many early human groups, from South Africa and Southeast Asia to Australia and the Americas.
Acknowledgments
Our research was funded by a National Park Service contract with San Diego State University and a grant from Paleoindian Research Fund at the University of Oregon's Museum of Natural and Cultural History. We thank Laura Kirn, Gary Brown, Kristin Hoppa, Nicole Kulaga, and Reuven Sinensky for logistical and intellectual support; Chumash tribal member David Paul Dominguez for field assistance; and the Syuxtun plant mentorship collective for discussions about Chumash plant foods. Permission to collect blue dicks on Santa Cruz was granted by The Nature Conservancy, and botanists John Knapp and Jessica Peak provided information on geophyte conservation status. Soil samples were analyzed at Oregon State University, and 14C samples were analyzed by John Southon at the University of California, Irvine. Finally, we thank Lynn Gamble and six reviewers for comments that significantly improved our article, as well as Patricia Chirinos Ogata for translating our abstract into Spanish.
Data Availability Statement
Additional data pertaining to this analysis are available in a technical report (Braje and Erlandson Reference Braje, Erlandson, Gill and Brown2019) submitted to Channel Islands National Park. Archaeobotanical data and notes are available from Gill.
