Mitochondrial DNA and Prehistoric Settlements: Native Migrations on the Western Edge of North America

Human Biology, Feb 2004 by Eshleman, Jason A, Malhi, Ripan S, Johnson, John R, Kaestle, Frederika A, Et al

Abstract We analyzed previously reported mtDNA haplogroup frequencies of 577 individuals and hypervariable segment 1 (HVS1) sequences of 265 individuals from Native American tribes in western North America to test hypotheses regarding the settlement of this region. These data were analyzed to determine whether Hokan and Penutian, two hypothesized ancient linguistic stocks, represent biological units as a result of shared ancestry within these respective groups. Although the pattern of mtDNA variation suggests regional continuity and although gene flow between populations has contributed much to the genetic landscape of western North America, some evidence supports the existence of both the Hokan and Penutian phyla. In addition, a comparison between coastal and inland populations along the west coast of North America suggests an ancient coastal migration to the New World. Similarly high levels of haplogroup A among coastal populations in the Northwest and along the California coast as well as shared HVS1 sequences indicate that early migrants to the New World settled along the coast with little gene flow into the interior valleys.

KEY WORDS: COASTAL MIGRATION, MITOCHONDRlAL DNA, HYPERVARIABLE SEGMENT 1 (HVS1), HOKAN, PENUTIAN, UTO-AZTECAN, LINGUISTIC DIVERSITY, CENTRAL VALLEY OF CALIFORNIA, GREAT BASIN, COLUMBIA PLATEAU, NATIVE AMERICAN GENETICS

Before European contact, the western edge of North America exhibited an exceptionally high level of linguistic diversity (Campbell 1997; Mithun 1999), with 90 separate languages being spoken in California alone (Moratto 1984). This great diversity has been cited as evidence for a greater antiquity of human occupation in these regions than in the rest of the Americas [J. Nichols 1990; but see also Nettle (1998)]. Particular patterns of this diversity have also been used to formulate hypotheses of human migration and expansions in western North America as well as a higher-level subgrouping that is presumed to reflect closer relationships among certain groups of Native American populations.

Dixon and Kroeber (1913) first proposed the existence of the Hokan and Penutian language superstocks in California. California Penutian includes four language families-Wintuan, Maiduan, Yokutsan, and Utian (which itself includes the Miwok and Costanoan subfamilies)-which are further subdivided into approximately 30 distinct languages. Hokan languages include a number of language families in California, the Yuman languages in Baja and the southwestern United States, and Washo on the western edge of the Great Basin (Campbell 1997; Campbell and Mithun 1979; Mithun 1999). The notable nonrandom distribution of various Penutian and Hokan languages has long suggested to prehistorians a wave or waves of Penutian migrations into California. Penutian speakers eventually occupied the interior valley extending to the coast around the San Francisco Bay Area, whereas a number of Hokan-speaking populations occupied the periphery of the Central Valley (Moratto 1984). This ring of Hokan-speaking groups surrounding a contiguous body of Penutian-speaking tribes suggests to some that at one time the whole of California was occupied by Hokan speakers who were later displaced to the valley periphery by an expansion of Penutian speakers from a central California homeland (Kroeber 1935).

The "Hokan-Penutian" hypothesis has had considerable influence on interpretations of California prehistory (Breschini 1983; Moratto 1984). Glottochronological estimates for the age of the Utian branch of the Penutian stock (the Miwok and Costanoan language families) in California fall between 4500 and 5200 years B.P. (Callaghan 1997; Moratto 1984). This date coincides closely with the appearance of the Windmiller pattern in the Central Valley, suggesting that the Windmiller pattern represents the earliest presence of Penutian peoples in the area (Fredrickson 1973; Ragir 1972). Archeological characteristics of Windmiller mortuary practices include extended burials, the presence of Haliotis and Olivella beads, characteristic charm stones, red ochre, and large projectile points (Ragir 1972). Similar archeological elements at the Kramer Cave site in western Nevada further suggest cultural contact between Great Basin peoples and the coeval Windmiller cultures, whereas red ochre, charm stones, and large, contracting stem-and-leaf-shaped projectile points suggest links between Windmiller and the Dalles sites in Oregon (9800-6000 years B.P.) on the Columbia Plateau (Hattori 1982), where other languages that are sometimes assigned to the Penutian superstock of languages are spoken (Sapir 1929). Foster (1996) associated Windmiller with Proto-Utian and placed their homeland in the northwestern Great Basin or on the Columbia Plateau, as suggested by similarities in the Altithermal cultures in these regions. Mitochondrial DNA of ancient burials at Pyramid Lake and Stillwater Marsh in western Nevada closely resembles that of modern California Penutians but not that of the neighboring Washo population (a Hokanspeaking group) in the Sierra Nevada (Kaestle and Smith 2001). The argument then suggests that all these areas-the western Great Basin, the Columbia Plateau, and the California Central Valley-were once occupied by closely related people who spoke languages of the Penutian stock.