Examining the Pro to-Algonquian Migration: Analysis of mtDNA BETH A. SCHULTZ, RIPAN S . MALHI AND DAVID G. SMITH University of California, Davis The study of Algonquian history is of interest to both linguists and archaeologists who have formulated hypotheses regarding the time and place of their homeland. With recent advances in genetic technology and the examination of mitochondrial DNA of Native Americans in the north­ eastern portion of North America, we can test the likelihood of these pre­ viously suggested hypotheses. 1 Sapir (1913) first suggested the term Ritwan to denote the California languages Wiyot and Yurok, and later (1916) suggested a distant genetic relationship with the Algonquian languages of the East, and named this language family Algic. This relationship is significant primarily because of its wide geographic spread. In 1929, Sapir subdivided Native American languages into six major groups, one of which was Algonkin-Wakashan (Almosan). The similarities he noted between the West Coast Wakashan language (and other Mosan languages) and the Algonquians in the East strengthened his hypothesis for a western homeland for Proto-Algonqui­ ans. Further linguistic evidence has since provided support for the Algon­ quian-Ritwan connection and is regarded by most historical linguists as proven (Haas 1960, 1965; Goddard 1975). While the Almosan grouping has been scrutinized, no strong linguistic evidence has emerged for this grouping (Haas 1965). Most recently, Campbell (1997) regarded Sapir's classification as controversial, with a "- 75 % probability" of sharing a common ancestor; in other words, Campbell believes there is a strong likelihood that there is not a common ancestor for Almosan. Although there is little evidence for a strong genetic relationship between the Algonquian and Mosan languages, independent evidence suggests that Proto-Algonquians originated in the Northwest. The linguis­ tic diversity within Algonquian groups is lower in the East than in the West, and thus provides evidence of a West to East migration (Goddard 1994). Haas (1965) has shown cognate and sound correspondences for I. Beth A. Schultz and Ripan S. Malhi contributed equall y to the research and analysis included in this paper. PROTO-ALGONQUIAN MIGRATION: ANALYSIS OF MTDNA 471 Kutenai and Algonquian, as well as further relationships between Kutenai and Salishan. Denny (1989) highlights additional linguistic similarities between Kutenai, Salishan, and Algonquian. Rhodes (personal communi­ cation, 1999) suggests that the Algonquian languages may have been located on the Plateau before they began to diversify, which is why there are typological similarities among Algonquian, Salishan, Kutenai, and Wakashan. It is possible that this proposed spread of the Algonquian lan­ guage from West to East was a linguistic spread that did not involve the movement of people. In addition to linguistic evidence for a western homeland for Proto­ Algonquians, archaeological connections link prehistoric people of the Plateau to those of the East. The Western Idaho Archaic burial complex, dated between 4000 and 6000 ybp, shares many similarities with the Red Ocher and Glacial Kame traditions in the East (Pavesic 1985). These buri­ als are primarily found in sandy knolls with bodies in a flexed position, often covered with generous amounts of red ocher, and occasionally are alongside canine burials. Burial artifacts include marine shell beads (the Olivella type in western Idaho), cache blades (large, unnotched, bifacially chipped forms found primarily in graves), and the distinctive "turkey-tail" projectile point (Pavesic 1985). The "turkey-tail" projectile point is a large ceremonial type common to the Eastern Woodlands, and is morpho­ logically similar to those points found in the Western Idaho complex. When diagnostic features of the Red Ocher complex are compared to those of the Western Idaho complex, the features appear remarkably con­ gruent (Denny 1991). Based on the distribution in time and space of these similar burial complexes, Denny (1991) proposed that the people of the Western Idaho complex were Proto-Algonquian speakers, and they migrated to the Great Lakes approximately 4000 ybp. MITOCHONDRIAL DNA Using mitochondrial DNA (mtDNA), we can examine genetic rela­ tionships between modern and ancient peoples in North America. This allow u to determine whether or not modem Algonquian-speaking peo­ ple are genetically similar to ancient peoples of the Glacial Kame and/or We tern Idaho Archaic burial complexes, or to other modem people of the Pacific Northwes t. We can also investigate, assuming a pre-historic Proto­ Algonquian migration from the West, whether the Algonquian expansion wa pri marily one of people or of language and culture. 472 BETH A. SCHULTZ, RIPAN S. MALHI AND 0AVID G. SMITH The mitochondrial genome is small, circular and located outside the nucleus in the cell, and is characterized by features that make it ideal for addressing these questions. First, because it is extra-nuclear, mtDNA does not undergo meiosis and thus is passed on only from mother to offspring, without recombination (Giles et al. 1980); the only source of variation in mtDNA is from mutation, allowing us to directly trace relationships between maternally related individuals. Second, this genome contains non-coding regions as well as loci coding for proteins used in cellular res­ piration that occurs in the mitochondria. Little selection acts upon the non-coding regions so mutations that appear here will not affect the sur­ vival of the individual (Avi se 1994, Stoneking 1990). With minimal selec­ tion, the environment will not play a directional role in the frequency of genotypes in a population. Third, mtDNA rapidly evolves at a rate from one to ten times as fast as nuclear DNA, with the non-coding regions accumulating mutations at the fastest rate (Stoneking 1990, Brown et al. 1979, Stoneking et al. 1986). This rapid rate of mutation accumulation is useful for evolutionary studies of groups with a recent common ancestor, and particularly for studies comparing human populations (Stoneking 1994, Johnson et al. 1983). Finally, the high copy number of mtDNA cre­ ates a higher probability of extracting DNA from ancient samples that have degraded over time (results are seen in Stone and Stoneking 1993, Kaestle 1997, 1998). Studies of multiple regions of the mitochondrial genome for varia­ tions, or polymorphisms, have revealed there to be five distinct lineage clusters, or haplogroups, of Native Americans (Schurr et al. 1990, Forster et al. 1996). Each maternally descended Native American belongs to one of these five haplogroups, referred to as haplogroup A, B, C, D, and X. While all of these haplogroups are widely represented throughout the Americas, haplogroup analysis has revealed a clear pattern in North America. Haplogroup A has a high frequency in the North, whereas hap­ logroup B is frequent in the Southwest. Haplogroup D is common in the Western interior, and haplogroups C and X are highest in frequency in the East (Lorenz and Smith, 1996, Mahli et al. 2001). The gain or loss of specific restriction sites (in the case of haplo­ groups A, C, D, and X) or the presence of a nine base-pair deletion (in the case of haplogroup B) in the mtDNA defines each of the five haplo­ groups. Restriction fragment length polymorphism (RFLP) analysis may be performed by adding a restriction enzyme to a particular sample of PROTO-ALGONQUIAN MIGRATION: ANALYSIS OF MTDNA 473 DNA and allowing the enzyme to cut, or digest, the fragment as illus­ trated in Figure 1. The enzyme will only digest the fragment if it recog­ nizes a particular nucleotide sequence in the fragment; if there is a variation in that sequence, the enzyme will not digest it. Samples can then be run out on a polyacrylamide gel. DNA is negatively charged so it migrates to the anode of an electrical field, sorting the fragments by size. The digested sample of DNA appears as two small bands, and the undi­ gested sample as one large band, as Figure 2A illustrates. The restriction sites diagnostic of haplogroups A, C, D, and X are given in Table 1. The nine base-pair deletion can be detected by running out a specific fragment from region V of the mitochondrial genome on a polyacrylamide gel; individuals of haplogroup B will have a fragment that is nine base-pairs shorter than non-haplogroup B individuals, as shown in Figure 2B . A higher resolution analysis can be conducted by characterizing point mutations in the control region (largest of the non-coding regions) of the mtDNA. All control region sequences in our analysis are expressed as divergences from the Cambridge Reference Sequence (Anderson et al. 1981 ). Specific mutations in this sequence correspond with each of the five haplogroups and are shown in Table 2; together with additional muta­ tions unique to the individual, these mutations form an individual's haplo­ type. For example, individuals of haplogroup A can be identified by ancestral mutational changes at nucleotide positions (nps) 16223 (cytosine to thymine transition), 16290 (cytosine to thymine transition), and 16319 (guanine to adenine transition), but an individual may also have markers specific to their maternal lineage (i.e. the Cheyenne/Arap­ aho individual in Table 2 has cytosine to thymine transition markers at nps 16111 and 16192). Current data suggest that the Americas were colonized by at least one member of each haplogroups A, B, C, D, and X (Brown et al . 1998, Schurr et al. 1990, Torroni et al. 1992). These five haplogroups may be the result of a ingle group migration, or they may have come in various migrations each bringing in different haplogroups, potentially with multiple migra­ tion bringing in different types of these five haplogroups. Haplogroups A, B, C, and D have been found in East Asia, but haplogroup X has not (Torroni et al. 1993). Haplogroup X is found in central and western Asia and in Europe; however, in most cases the European and Asian versions of haplogroup X differ from the Native American haplogroup X by a CR marker (the ab ence of a tran ition at np 16213, Brown et al.