The Colorado Delta, 1771–1776: Rereading Francisco Garcés

AbstractThe ethnohistory of the Colorado River delta has been substantively misunderstood, owing to the widespread neglect and/or misinterpretations of the writings of Francisco Garcés. In 1771, 1774, and 1775–1776, Garcés undertook three entradas into the delta, and wrote a series of valuable ethnographic accounts. Not only have Garcés’s locations and routes frequently been misidentified by earlier scholars, his observations on agricultural production and population size have been ignored or marginalized, enabling misconceptions about delta historical demography and adaptation to flourish. The present paper seeks to restore Garcés’s accounts, making his locations and ethnographic observation intelligible and interpretable, and to show how these can help resolve extant misconceptions. Part I focuses on some key texts, tying his locations to a master map. Part II focuses on ethnolinguistic groups and settlement sites, discusses the implications for a better understanding of historical demography and agricultural adaptation in the delta.La etnohistoria del delta del río Colorado ha sido sustancialmente malinterpretada, debido al descuido generalizado y/o malas interpretaciones de los escritos de Francisco Garcés. En 1771, 1774 y 1775-76, Garcés realizó tres entradas al delta y escribió una serie de valiosos relatos etnográficos. No solo las ubicaciones y rutas de Garcés han sido identificadas erróneamente con frecuencia por académicos anteriores, sino que sus observaciones sobre la producción agrícola y el tamaño de la población han sido ignoradas o marginadas, lo que permite que florezcan conceptos erróneos sobre la demografía histórica del delta y la adaptación. El presente artículo busca restaurar los relatos de Garcés, haciendo inteligibles e interpretables sus ubicaciones y observaciones etnográficas, y mostrar cómo estas pueden ayudar a resolver conceptos erróneos existentes. La Parte I se enfoca en algunos textos clave, vinculando sus ubicaciones a un mapa maestro. La Parte II se centra en los grupos etnolingüísticos y los sitios de asentamiento, analiza las implicaciones para una mejor comprensión de la demografía histórica y la adaptación agrícola en el delta.KEYWORDS: Colorado deltaYumanethnohistorydemographyadaptationSpanish explorationindigenous interrelationsAnza expedition AcknowledgmentsArchival and field research into Garcés’s writings since 2010 has been supported by the Ogden Mills Fund, Division of Anthropology, American Museum of Natural History. I am most grateful to archivists at the following institutions: Bancroft Library (Berkeley), University of Arizona Special Collections Library (Tucson), Office of Ethnohistorical Research, Arizona State Museum (Tucson), Newberry Library (Chicago), Huntington Library (San Marino, CA), University of New Mexico Library Center for Southwest Research (Albuquerque), Dolph Briscoe Center for American History, University of Texas (Austin), Library of Congress Manuscript Division (Washington D.C.), National Anthropological Archives (Suitland, MD), Archivo General de Indias (Seville), Real Biblioteca (Madrid), Historical Archives, OFM (Rome), British Library (London), and Archivo General de la Nación (Mexico City). I am especially grateful to delta geoscientist Steven M. Nelson for indispensable, patient guidance in June and July 2022, and February–May 2023 on the historical hydrology and geomorphology of the Colorado delta, and particularly for his reading of the 1939 aerial photographs (see Part I, Figure 3): in any instance where my interpretations depart from his, I alone am responsible. I am also most grateful to four anonynmous readers for KIVA, who provided very valuable suggestions.Disclosure StatementNo potential conflict of interest was reported by the author.Notes1 One Version B transcription (Garcés Citation1771b: October 4) here shows “tierras y arboleda muy escasas,” i.e., “very sparse lands and groves,” while the other has “tierras y arboleda muy hermosas” (Garcés Citation1771c: October 4). Version A (see Part I) here has “tierras vonisimas y mui pobladas” (Garcés Citation1771a: October 3; cf. Baroni Citation2016:75), strongly suggesting “hermosas” is the correct transcription for Version B.2 Bolton (Citation1917:326) first suggested La Merced was southwest of New River. Later, however, discussing the 1774 expedition, he corrected this to east of San Jacome (Bolton Citation1930, II:328, n.3), i.e., southeast from Anza and Garcés’s then position to the north of Cerro Prieto (see Bolton Citation1930, I:120).3 Dependent on Coues (Citation1900), Castetter and Bell (Citation1951:72), in discussing Garcés’s description of the abundant crops at La Merced, mistakenly identify Cajuenche lands as on the east side of the Colorado mainstem, “a short distance below the Gila junction”—see also below.4 Derby (Citation1850) shows Ogden’s Landing at about 32° 18′ 20″ N, while Egloffstein (Ives Citation1861:map 1) places it 4′ farther north at about 32° 22′ 30″ N, 114° 50′ 48″ W, and Sykes (Citation1937:38, figure 15; coordinates inferred from Plate I) shows the latitude as about 32° 23′ N. Egloffstein’s river course—borrowed for many depictions of the mainstem throughout the late nineteenth and twentieth century (Sykes Citation1926:240)—is also included on Sykes’s (Citation1937) Plate I, a composite map of the Colorado Delta as of 1933, on which the Ogden’s Landing label is replaced by (the later settlement of) La Islita. The approximate center of La Islita seems a reasonable estimate for Santa Rosa.5 Kroeber’s (Citation1920:475) statement that Garcés located the Halyikwamai west of the river, in contrast to Oñate, neglects Garcés’s accounts of Santa Rosa.6 As most other maps of this period, Silsbee took the mainstem depiction by Egloffstein (Ives Citation1861:map1) and overlaid it onto his map, so the river course should not be read as a representation of conditions in 1900 (Steven M. Nelson personal communication 3-15-2023). Note that my figure 1 was prepared in ignorance of Silsbee’s map, which Nelson sent me in March 2023 (Steven M. Nelson personal communication 3-15-2023): the coincidence of Silsbee’s site with my inference of Garcés’s Halyikwamai site in 1775 is thus serendipitous.7 Version A (Garcés Citation1771a: September 16) of his 1771 diary identified the fishermen as “Yumas” (Version B has “Yndios”), but this was corrected in 1775 to Cucapá.8 Garcés observed Las Llagas at 32° 18′ N, but his quadrant measures fluctuate a good deal in relation to inferred modern coordinates (see part I). Derby (1850) marked the head of tidewater at approximately 32° 04′ 22″ N, and Egloffstein (Ives Citation1861:map 1) was in close agreement (about 32° 04′ N). However, their longitudes differ: ca. 114° 44′ 45″ W (Derby), 115° W (Egloffstein). This reflects the historical problematics of longitude calculations (e.g., Sobel Citation1995), rather than any 15′ river-course shift between 1850 and 1857, as is evident from other variable longitudes for the same features on their maps.9 Forbes’s (Citation1965:6, 144) location of Las Llagas by the lower Río Hardy just above its confluence with the Colorado mainstem is clearly too far southwest. Forbes’s sketch of river courses is odd: the mainstem is evidently designed to reconstruct a pre-1905 course, and appears to borrow—but simplify, distort and re-position—Egloffstein’s depiction (Ives Citation1861:map 1). Forbes shows the Pescadero with small deltas forming along its course in the central area dividing the river’s continuity (compare, e.g., Bonillas and Urbina Citation1913 [see figures 3a, 3b herein], Sykes Citation1926:after p. 254)—a phenomenon that as such post-dates the mainstem changes beginning in 1905. Forbes does not cite any sources for his map but depends in part on Sykes (Citation1937).10 “Camp 2, Cocopa Village” was observed at 32° 04′ 17.2″ N, 115° 00′ 15″ W (Ives Citation1861:38; Appendix B, p. 6).11 For named Cocopah settlements in this area, see Gifford (Citation1933:260). Gill’s photographs of Cocopahs resident here appear in several sources: Dellenbaugh (Citation1902); Kelly (Citation1977); Williams (Citation1974, Citation1983).12 Forbes (Citation1965:144) depicted San Mateo as a single rancheria along the Pescadero River southeast of Volcano Lake: however, Garcés was clear San Mateo comprised several rancherias on both sides of the “ten-league” long lake.13 Forbes (Citation1965:163) and Kelly (Citation1977:7) also infer Garcés’s “serranos” on December 19, 1775 refers to Paipai. Kelly (Citation1942:679, n. 9) reported Cocopah-Paipai relations going back at least one hundred years (i.e., to ca. 1840): Garcés’s implicit information here pushes this back in effect about another century.14 Hedges (Citation1975:71) has suggested Garcés’s San Jacome equates with the historic Mountain Kumeyaay settlement at Jacumé, in the municipality of Tecate, Baja California, ca. 75 km west of Mexicali. However, this is surely incorrect. No other analyst infers Garcés reached that far west in 1771, and Jacumba is an autochthonous Mountain Kumeyaay placename associated with a subgroup, Jacum (Spier Citation1923:298, n. 7a); approximate homonymy with Spanish Jacome is coincidental.15 “en todas partes desde la primera rancheria de Yumas vi sandias melones, mais y frijol. y solo este Pueblo donde estaba junta toda la gente carecia de un todo manteniendose solam.te con varias semillas y raises que sacan de la tierra pechita y tornillo cuyos arboles abundan en las orillas de dho Pueblo.”16 Garcés’s explanation for San Jacome’s desertion was that the well had dried up: “Llegamos al pozo de San Jacome, a donde nos llevaron unos Yndios, que vivian cerca, y vimos que ya estava ciego, y que se havia mudado la gente a la sierra, y a las rancherias inmediatas/We arrived at the well of San Jacome, guided there by some Indians who lived nearby, and we saw that it was now blind, and that the people had moved to the mountains and to settlements near them” (Garcés Citation5-Citation21-Citation1775; compare Bolton Citation1930, II:334–335). The abandonment was likely seasonal: his visit in 1771 occurred in September, while in 1774 it was in February-March. According to Díaz, accompanying Garcés and Anza in 1774, Natives from nearby La Merced advised on March 3 that San Jacome was “at present deserted for lack of water” (Bolton Citation1930, II:276, emphasis added). Garcés did not revisit San Jacome in 1775–1776, but included it in his summary of Quemeya locations (Garcés Citation1777:12-6-1775, quoted in translation above in the text). San Jacome at this period was thus evidently a large (“pueblo”-size) seasonal foraging nexus for Kamia, rather than the deserted conditions observed in winter 1774 representing complete abandonment. By the time of Gifford’s Kamia research in the late 1920’s, however, although Cerro Prieto was a named site (Wiespa, “eagle mountain” in cognate Cocopah Wii Shpa [Gifford Citation1931:9; Wright and Hopkins Citation2016:12]), he heard no mention of any former Kamia seasonal settlement below it.17 Font played a central role in the composition of Garcés’s 1775–76 diary, with the two friars in direct discussions about such matters in January 1777 (Whiteley Citation2015:368).18 Situationally, Garcés (and other Spanish sources) tended to distinguish Opas and Cocomaricopas geographically: Opas lived about the Great Bend of the Gila and somewhat farther upstream; Cocomaricopas were downstream of the Great Bend (see Ezell Citation1963:10–26). For that reason, Ezell (Citation1963:26), who provides the most comprehensive analysis of these terms, concluded historical “Cocomaricopa” was equivalent to ethnographic “Kaveltcadom,” a term introduced by Spier (Citation1933) from Maricopa usage. However, more generically, Garcés (and other Spanish sources) also used “Cocomaricopa” to include all Yumans on the Gila above the Quechans (e.g., Garcés Citation1777:November 8, 1775), with “Opa” merely a Cocomaricopa subgroup. I examine these differences elsewhere (Whiteley Citationn.d.), but for the present have elected to keep Cocomaricopa and Opa after Garcés’s specific usages, rather than substituting Kavelchadom for Cocomaricopa. “Maricopa,” an American-period abbreviation of Cocomaricopa, refers to descendants of Opas, Kavelchadom, Halchidhomas, Kahwan, Halyikwamai and other River and Delta Yumans who fled to the middle Gila in the nineteenth century (Wilson Citation2014:4–5).19 In 1775 at southernmost La Merced, a fight broke out between Kahwan hosts and their Halyikwamai guests, with one Kahwan man fatally speared (Garcés Citation1777:12-12-1775). Garcés remonstrated with the rancheria leader, asking how this could have happened, “estando yo alli que venia â ponerlos â todos en paz/while I whose purpose in coming there was to make peace among all of them.”20 “ … los Cucapa han sido siempre amigos de los Cunyeil de la sierra, que llegan hasta la mar; y enemigos de los Papagos que viven cerca del mar de Californias, de los Jaliquamays, y Cajuenches. Los Jaliquamays, y Cajuenches siempre han sido amigos, y han conservado amistad con los Yndios Quemeyá que viven en la Sierra, y que se estienden hasta las rancherias de San Diego, y con los Jalchedunes; sus enemigos han sido siempre los Yumas, y los Papagos de la marisma. Los Yumas han tenido siempre por amigos â los Jamajabas, â los Yabipays Tejua, y â los Papagos de Sonoytac y de la marisma; y han sido enemigos mortales de los Jalchedunes, de los Cocomaricopas, de los Pimas Gileños, y de todas las Naciones del rio abajo, y tambien de los Jecuiches de la Sierra” (Garcés Citation1777: Reflexiones, Punto Segundo).21 Evidently dependent on Sykes, a recent U.S.G.S. report discussing Spanish transits through the area entirely omits Garcés, and identifies Pattie’s party of 1827 as “the first to explore the delta from the north” (Mueller and Marsh Citation2002:2). Yet all three of Garcés's explorations were from the north, and covered a much broader area than Pattie, who (pace Sykes Citation1937:16) stuck close to the Colorado mainstem along its eastern course (Pattie Citation1973 [Citation1831]). Sykes’s reading of Pattie as possibly affirming the mainstem’s westward shift into the Hardy (see below) up to at least 1827—accepted by Castetter and Bell (Citation1951:4)—is not supported by a close reading of Pattie (Citation1973 [Citation1831]). In light of the clear indications in Garcés’s texts that the mainstem continued south to tidewater below the Mesa de Andrade, Sykes’s inferences are unsustainable (cf. Part I).22 These include Fages in 1785, Arrillaga in 1796, Hardy in 1826, Pattie in 1827, Derby in 1850, Heintzelman in 1850–1853, and Ives in 1857–58. For Fages and Arrillaga—who were both only along the Río Hardy—see Forbes (Citation1965:222–225, 229–231) and Robinson Citation1969 (87–89); for the remainder, see, e.g., Sykes (Citation1937:13–23). For accounts of the 1860s-1870s, see Williams (Citation1975) and Ortega Esquinca (Citation2004:218–223); and for the 1890s-1900s, see Bendímez Patterson (Citation1995:253) and Chittenden (Citation1901).23 Cocopah settlements in Baja California in 1918 included three about the Río Hardy and four about the irrigation canals (Ortega Esquinca Citation2004:224). A local count of 1,200 Cocopahs in 1900 included those who had migrated to Mexicali, Yuma, and elsewhere (Lumholtz Citation1912:251). Major changes to the Cocopah economy since the 1850s included employment by U.S. steamships, lumber yards, railroad companies, and agricultural enterprises (e.g., Williams Citation1987; Bendímez Patterson Citation1995:250–252; Porcayo et al Citation2015). Pace Kelly (in Castetter and Bell Citation1951:55), the 300 count for 1900–1930 may not be inaccurate for those remaining on their old sites in the delta.24 Kroeber thought this an overestimate, accepting 13,000 for this period (Castetter and Bell Citation1951:74).25 Garcés’s “figures on the population of this region are high, especially for the smaller groups. It seems impossible that three or four separate tribes should each have shrunk from 2000 or 3000 to a mere handful in less than a century, during which they lived free and without close contact with the whites” (Kroeber Citation1920:476).26 Bolton (Citation1919, I:315, n. 430) inferred “Coanopa” was a reference to Cocopah, but now that Kahwan has been identified north and northwest of the Halyikwamai in Garcés’s time, Kahwan seems the far more likely referent (see also Ezell [Citation1963:14–15] on the O’odham usage of ‘opa’ as a suffix referring to Yuman-speakers, noting that Kino was accompanied by a large number of O’odham).27 Recalling that, as noted in Part I, Cocopah have frequently stood in for all lower delta inhabitants in light of the disappearance of Kahwan and Halyikwamai from the area in the nineteenth century.