Mitochondrial DNA of Protohistoric Remains of an Arikara Population from South Dakota: Implications for the Macro- Siouan Language Hypothesis

Diana M. Lawrence,1 Brian M. Kemp,2 Jason Eshleman,3 Richard L. Jantz,4 Meradeth Snow,3 Debra George,3 and David Glenn Smith3

Abstract Mitochondrial DNA (mtDNA) was extracted from skeletal remains excavated from three Arikara sites in South Dakota occupied between AD 1600 and 1832. The diagnostic markers of four mtDNA haplogroups to which most Native Americans belong (A, B, C, and D) were successfully identified in the extracts of 55 (87%) of the 63 samples studied. The frequencies of the four haplogroups were 42%, 29%, 22%, and 7%, respectively, and principal coordinates analysis and Fisher’s exact tests were conducted to compare these haplogroup frequencies with those from other populations. Both analyses showed closer similarity among the Mohawk, Arikara, and Sioux populations than between any of these three and any other of the comparison populations. Portions of the first hypervariable segment (HVSI) of the mitochondrial genome were successfully amplified and sequenced for 42 of these 55 samples, and haplotype networks were constructed for two of the four haplogroups. The sharing of highly derived lineages suggests that some recent admixture of the Arikara with Algonquian-speaking and Siouan- speaking groups has occurred. The Arikara shared more ancient lineages with both Siouan and Cherokee populations than with any other population, consistent with the Macro-Siouan language hypothesis that Iroquoian, Siouan, and Caddoan languages share a relatively recent common ancestry.

The popular concept of Plains Indians is dominated by the equestrian nomads, such as the Dakota, most recently popularized in the film Dances with Wolves. However, before the 19th century, semisedentary tribes living in earth lodge villages were the primary inhabitants of the Great Plains. The Arikara, or Sahnish,

1Forensic Analytical Sciences, Hayward, CA 94545. 2Department of Anthropology and the School of Biological Sciences, Washington State University, Pullman, WA 99164. 3Molecular Anthropology Laboratory, Department of Anthropology, University of California, Davis, CA 95616. 4Department of Anthropology, University of Tennessee, Knoxville, TN 37996.

Human Biology, April 2010, v. 82, no. 2, pp. 157–178. Copyright © 2010 Wayne State University Press, Detroit, Michigan 48201-1309 key words: mtDNA haplogroup, HVSI haplotype, ancient DNA, migration, admixture, Arikara, Mohawk, Sioux, Chippewa, Cherokee, Ojibwa, Pawnee, Algonquian groups, South Dakota. 158 / lawrence et al. living along the banks of the Missouri River in South Dakota, were one of the most populous and powerful tribes to emerge into the historic period. They grew corn, beans, and squash in the fertile soils along the Missouri floodplain while continuing to follow a seasonal hunter-gatherer lifestyle collecting wild plants and hunting bison (Blakeslee 1994). This subsistence pattern was unique in North America in its heavy reliance on hunting. Lehmer and Wood (1977) have likened it to the mixed horticultural and pastoral economies of the Old World. Arikara is a Caddoan language, closely related to Pawnee (Parks 1979), from which it split between 1450 and 1650 (Parks 1979). Linguistic relatives far- ther south include Wichita, Kitsai, and Caddo proper (Parks 1979). The sites in South Dakota attributed to the Arikara represent the northernmost extension of the Caddoan languages until the group moved into North Dakota after leaving the Leavenworth site in 1832. Because the focus of our study is the relationship of the language spoken by the Sahnish to other languages, we have chosen to use Arikara to refer to the Sahnish throughout this paper. Archaeological complexes usually regarded as directly ancestral to the historic Arikara include the Extended Coalescent Variant (ca. AD 1400/1450–1650), the Postcontact Coalescent (ca. 1650–1785) (Johnson 1998), and the Disorganized Coalescent (1785–1862) (Lehmer 1971). Arikara connections to early archaeological complexes, such as the Initial Coalescent (1300–1600) or the earlier Central Plains Tradition (900–1450) in Kansas and Nebraska, have not been established, although some sort of relationship has been widely assumed. If there is a connection, it likely predates the emergence of the Arikara as an ethnic entity. Exposure of the Arikara to Europeans resulted in dramatic changes, which affected their health and well-being as well as their population structure. Initial European contact may have happened as early as 1700 but is not documented before about 1706 (Jantz and Owsley 1994). Initially, European contact was bene ficial to the Arikara. Acquisition of the horse increased bison hunting efficiency, and the Arikara eventually established themselves as middlemen in trade rela- tions, which initially brought a wave of prosperity but ultimately resulted in almost complete cultural collapse as epidemics of European diseases decimated their numbers. During the early postcontact period, the Arikara population was estimated to be about 30,000 individuals (Holder 1970: 30). By the early 19th century this number had been reduced to the approximately 2,000 individuals living in the villages along the Grand River in South Dakota, now known as the Leavenworth site. Shortly after leaving the Grand River villages, the Arikara moved their village to Ft. Clark, North Dakota, where their numbers were further reduced by a particularly virulent epidemic sweeping Plains tribes in 1837–1838 (Trimble 1994). In 1862 the Arikara, Mandan, and Hidatsa were moved to the Fort Berthold In- dian Reservation, where they now live as the Three Affiliated Tribes. The Arikara population nadir was 380 individuals, reported by the 1904 census, but by 2005 they had increased to 3,400, as reported by the Indian Health Service. Today the Three Affiliated Tribes lead lives much like those of many other Native American mtDNA of Protohistoric Arikara / 159 tribes. They operate a casino, encourage tourism, and are attempting to develop the natural resources on their lands, particularly oil and gas. Chafe (1976) proposed that Caddoan together with the Iroquoian and Siouan languages belongs to a larger language family that he called Macro-Siouan, whose origin was in the Southeast. Previous studies have indicated that populations living in the same geographic area or closely related to each other often exhibit similar mitochondrial DNA (mtDNA) haplogroup frequencies as a result of either common ancestry and/or admixture (Bolnick and Smith 2003). Although the speakers of related languages themselves need not be closely genetically related to each other, they often are when vertical transmission of language predominates, and thus relationships between languages provide testable genetic hypotheses. To date, comparative studies of the mtDNA of the Caddoan, Iroquoian, and Siouan groups have not been conducted. The Iroquoian languages are divided into two major branches: Northern Iroquoian, which includes Mohawk, and Southern Iroquoian, which includes Cherokee (Chafe 1976). The Northern Iroquoian peoples split from their southern counterparts and moved north sometime between 3,500 and 4,000 years ago (Snow 1994). Subsequent admixture with Algonquian tribes in neighboring areas might make their genetic structure resemble a hypothetical admixture between ex- tant Cherokee and Algonquian tribes. If historic Plains tribes, such as the Arikara and the Sioux, also admixed with neighboring Algonquian tribes after migrating from their hypothetical southeastern homeland, then the genetic structures of the Arikara, Mohawk, and Sioux might resemble each other but differ from those of Cherokee groups. The ancestors of all three tribes might have more closely resembled the Cherokee before migrating northward, where they encountered and may have mixed with Algonquian groups. The identification of defining mtDNA markers, using restriction analysis or sequencing, makes it possible to assign individuals in populations to their haplogroups or subhaplogroups. Additional markers or DNA sequences can further subdivide members of each haplogroup, or subhaplogroup, into different haplotypes, each defined by a unique set of genetic markers or nucleotide sequences. Comparisons of numbers of nucleotide differences between all paired sequences in a lineage or population provide an estimate of the extent of diversity among individuals within that lineage or population. Because most nonbasal (i.e., derived) mtDNA haplotypes are tribe specific or nearly so (Malhi et al. 2002), the sharing of the same or similar haplotypes reflects either shared ancestry or admixture among different populations that are geographically proximate (Schurr 2004). The frequency distributions of haplogroups can be used to genetically characterize populations. These distributions usually exhibit geographic structure, because the probability that mating occurs between members of any two different human populations is inversely correlated with the geographic distance separating them. Thus the pattern of genetic diversity within and between human populations is sometimes consistent with a model of isolation by distance. The mtDNA of modern Native Americans has been shown to fall into one of at least 160 / lawrence et al. five haplogroups: A, B, C, D, and X (Kaestle and Smith 2001; Malhi et al. 2001, 2002). These haplogroups are defined by diagnostic mutations (O’Rourke et al. 2000) in the coding region (which can be identified by restriction analysis or other techniques) or in the case of haplogroup B, by a 9-bp deletion; these mutations are shared by every direct female descendant of the same ancient ancestress. Haplogroups A, B, and X are descendants of the Old World macrohaplogroup N, whereas haplogroups C and D are descendants of the Old World macrohaplogroup M. Several haplotypes of at least two of these Native American haplogroups (haplotypes C and D) appeared among the founding population of the Americas (e.g., see Tamm et al. 2007). Four of the five Native American haplogroups also share one or more diagnostic control region mutations identifiable by sequencing the first hypervariable segment (HVSI) of the control region; the major subhaplogroup of haplogroup D, D1, does not exhibit any unique control region mutations (Lorenz and Smith 1997; Smith et al. 1999). These mutations cause members of the same haplogroup (comprising closely related haplotypes) to cluster in a phylogenetic tree. Addi- tional (derived) mutations that differentiate members of the same haplogroup are often shared by only closely related or geographically contiguous tribal groups (Malhi et al. 2002) and therefore are useful in identifying common ancestry or admixture among tribal groups. These highly derived haplotypes might postdate long-distance migration and/or, because of the well-documented rapid population growth of Native Americans, outnumber their antecedent haplotypes and be more likely to be shared through admixture than antecedent ones. Thus the sharing of highly derived (recently evolved) haplotypes, but not the haplotypes ancestral to them, among groups not thought to be closely related suggests recent admixture, and sharing of ancestral haplotypes suggests common ancestry. Haplogroups A, B, C, and D all occur together in most Native American populations, albeit in varying frequencies, with haplogroup X being the most frequently absent of the five haplogroups; however, some tribes lack haplotypes from two or more of the five haplogroups. The presence of only one or two of the five haplogroups probably reflects the role of genetic drift (Schurr 2004). Haplogroup frequency distributions vary among tribal groups because of their variable common ancestry and population histories. Haplogroup A, the most common mtDNA haplogroup in the Americas, is most common in Eskimo, Aleut, Athapaskan, and Algonquian populations of Alaska and Canada, along the Pacific Coast of North America, and in Mesoamerica. Haplogroup B is most common in the Southwest, Great Basin, and Southeast. Haplogroup C occurs in moderate frequencies in the Southwest, Southeast, and Algonquian-speaking populations. Haplogroups D and X are the least common mtDNA haplogroups in North America, but they occur in highest frequencies in western North American and Algonquian-speaking populations, respectively. Haplogroups C and D reach high frequencies in many South American populations, but haplogroup X, the least common of the five haplogroups, is absent in Mesoamerica and Central and South America. mtDNA of Protohistoric Arikara / 161

If the Arikara, Sioux, and Mohawk populations share a common ancestry, as suggested by the Macro-Siouan hypothesis, then they should share similar haplogroup frequency distributions, albeit perhaps exhibiting some evidence of admixture with other Plains populations, reflecting isolation by distance that ob- scures their ancestry with the Cherokee. Based on regional and tribal distributions of haplogroup frequencies, approximately equal levels of admixture between Algonquian and hypothetical Macro-Siouan tribes that originated in the South- east would be expected to yield descendants with moderate frequencies of haplogroups A, B, and C and low frequencies of haplogroups D and X. Lower rates of admixture with Algonquian-speaking groups should yield lower frequencies of haplogroup A and higher frequencies of haplogroup B. However, common ancestry among the Arikara, Sioux, and Mohawk should be detectable by their mutual sharing of similar haplotypes with the Cherokee that are internal to haplotype networks and thus more ancient than haplotypes on the periphery of the network. In contrast, the mutual sharing of only highly derived haplotypes, those at the periphery of the haplotype networks, with Algonquian-speaking groups would reflect evidence of recent admixture that occurred after the ancestors of the Mo- hawk, Sioux, and Arikara left their hypothetical homeland in the Southeast and migrated northward.

Materials and Methods Sites. Sites used in this analysis are Mobridge (39WW1), Larson (39WW2), and Leavenworth (39CO9), located in South Dakota along the Missouri River, as illustrated in Figure 1 (McWilliams 1982). Excavations at these sites were conducted in 1968–1971, 1966–1968, and 1965–1969, respectively, to salvage materials that would have otherwise been destroyed by construction of a dam in this region. The occupation times for the Larson and Mobridge sites are imprecisely known. Taxonomically, Larson belongs to the LeBeau Phase of the Postcontact Coalescent, which dates from 1650 to 1785 (Johnson 1998). Larson has a lengthy occupation, or series of occupations, from about 1680 to 1730, based on ceramic seriation (Johnson 1994). It was apparently abandoned because the village was attacked and its inhabitants massacred (Owsley et al. 1977). Mobridge has an exceedingly complex occupation history. Its early features have been considered late prehistoric on the basis of an absence of trade goods (Jantz 1973); presumably Mobridge belongs to the LaRoche Phase of the Extended Coalescent, although its later features belong to the Postcontact Coalescent LeBeau Phase. Total occupation may span a century, from about 1600 to 1700. Owsley (1981) documented significant craniometric heterogeneity among different temporal and spatial components of the Mobridge cemetery, presumably because of a succession of somewhat different populations. Leavenworth is the only site to have been reported in full (Bass et al. 1971). Its occupation date is firmly established from historical sources as 1802–1832. It was the final residence of the Arikara in South Dakota. Taxonomically itis 162 / lawrence et al.

Mobridge Larson Leavenworth

South Dakota

Figure 1. Map showing location of the approximate locations of the Larson, Mobridge, and Leaven- worth sites. Geographic coordinates of sites: Leavenworth, 45°67 N, 100°36 W; Larson, 45°52 N, 100°41 W; Mobridge, 45°56 N, 100°45 W.

included in the Postcontact Coalescent, although Lehmer (1971) suggested the term Disorganized Coalescent to describe the cultural collapse that accompanied more intense European American contact. The three sites cover not more than 230 years of Arikara history. Although this may seem like a relatively short period of time, it covers major transitions, from an essentially precontact subsistence pattern based on horticulture and pe- destrian hunting, to initial contact, introduction of the horse around 1740, increased warfare, and, most important, the introduction of European diseases. It is not possible to assume genetic stasis during this period. On the basis of cranial morphometrics, Jantz (1973) documented cranial change attributed to gene flow from Siouan-speaking neighbors encountered as the Arikara moved north. As contact with European Americans became more intense, the Arikara experienced major changes in population structure with depopulation as a result of epidemic disease being a major factor. The earliest historically documented epidemic was 1780–1781, which resulted in extensive depopulation of Plains groups. Village groups such as the Arikara were more affected by epidemic disease than the equestrian nomads (Trimble 1994). It is likely that epidemic disease spread in the Plains before 1780–1781; there is such evidence from the Larson site (Owsley 1979). Depopulation made village groups more vulnerable to attack, especially by the equestrian nomads. By the time of the Leavenworth site, the Arikara had lost 90% of their precontact population. The precontact population may have consisted of as many as 25 villages. It is also clear that linguistic differentiation was already considerable among different Arikara villages, based on historical reports of mutual unintelligi- bility at the Leavenworth site. Parks (1979) argued that major dialects were present at the Leavenworth site, which represent preexisting linguistic differentiation. The Leavenworth site provides evidence of that preexisting genetic variation. Hoffman (1977) considered Leavenworth the final manifestation of the Bad mtDNA of Protohistoric Arikara / 163

River Phase, consisting of sites on the right bank of the Missouri River between the Bad and Cheyenne Rivers. However, Byrd and Jantz (1994) argued that the two dialects identified by Parks (1979) correspond to LeBeau and Bad River Phase sites and that both phases are present at the Leavenworth site, although LeBeau Phase individuals are a substantial majority. Each site has burial remains from multiple features housed at the University of Tennessee. Eight features (101, 102, 103, 201, 301, Lodge 1, Lodge 21, and Lodge 23) at the Larson site, five features at the Mobridge site (101, 201, 301, 302, and 303), and nine features at the Leavenworth site (101, 102, 120, 201, 202, 213, 220, 301, and 402) were excavated. These feature designations encode burial area and year of excavation. For example, feature 101 is area 1, year 1, and feature 102 is area 1, year 2. The archaeological samples used for this study were collected and morpho- logically described by archaeologists from the University of Tennessee, Knoxville. Selected bone samples were assembled into bags and labeled at the University of Tennessee, Knoxville. Each bag contained one to eight bone specimens, predomi- nantly ribs, metacarpals, metatarsals, and phalanges, all corresponding to a single individual. The specific burial context from which the remains were recovered was noted on the bag, citing the feature with which the remains were associated and their identification number. Three features were randomly selected from each of the three sites, and from each feature 7 samples were randomly selected for study for a total of 63 samples. A list of each sample selected for extraction is provided in Table 1.

Laboratory Methods. Methods used in this study to prevent contamination include the following: use of supplies and equipment in a positively pressurized laboratory solely used for ancient DNA (aDNA) analysis; use of sterile and dis- posable labware and clothing, including face masks, hair nets, gloves, and lab coats; use of separate pre- and post-PCR areas; use of ultraviolet light to eliminate surface contamination; and running negative controls at all stages of the process to detect contaminants (Eshleman and Smith 2001; Handt et al. 1994). We have not followed all the recommendations of Cooper and Poinar (2000) to ensure the authenticity of the aDNA in our study but rather have followed only those that we regarded as reasonably necessary to provide confidence in the authenticity of our particular data. Specifically, we did not conduct amino acid racemization (AAR) on the remains from which DNA was extracted and did not clone aDNA extracted from the samples. Because the preservation of both amino acids and proteins and aDNA is highly plausible in the prehistoric remains recovered from hot, dry en- vironments, AAR cannot inform the plausibility that the aDNA in our samples is authentic. Moreover, neither of two recent critical evaluations of AAR have found it useful as a means of assessing aDNA preservation in bones and teeth (Collins et al. 2009; Fernandez et al. 2009). Because cloning identifies only the majority consensus sequence, which itself can be a contaminant, it provides no confidence in the authenticity of aDNA samples unless a target sequence is unknown (e.g., as 164 / lawrence et al.

Table 1. Description of Samples (n = 63) from Each Site from Which DNA Was Extracted and Haplogroup Assignments Based on Restriction Analysis

Sample Number Site Feature Burial Element Haplogroup 1 Larson 101 B18A femur C 2 Larson 101 B4 rib B 3 Larson 101 B16 rib B 4 Larson 101 B21B rib A 5 Larson 101 B1 rib – 6 Larson 101 B40A rib C 7 Larson 101 B22B rib C 8 Larson 201 B92B rib D 9 Larson 201 B30G rib – 10 Larson 201 B92C metacarpal B 11 Larson 201 B97E rib D 12 Larson 201 B101B phalanx – 13 Larson 201 B138C rib – 14 Larson 201 B147B rib – 15 Larson 301 B3 phalanx B 16 Larson 301 B21 phalanx D 17 Larson 301 B25 phalanx A 18 Larson 301 B29B phalanx C 19 Larson 301 B33A phalanx A 20 Larson 301 B57A phalanx B 21 Larson 301 B65 phalanx B 22 Mobridge 101 B2A phalanx C 23 Mobridge 101 B9E phalanx A 24 Mobridge 101 B10E phalanx B 25 Mobridge 101 B12B phalanx B 26 Mobridge 101 B16B phalanx C 27 Mobridge 101 B25C phalanx A 28 Mobridge 101 B25D phalanx A 29 Mobridge 201 B2B phalanx A 30 Mobridge 201 B4C phalanx A 31 Mobridge 201 B8B phalanx C 32 Mobridge 201 B11F phalanx A 33 Mobridge 201 B16H phalanx A 34 Mobridge 201 B29A phalanx A 35 Mobridge 201 B36E phalanx A 36 Mobridge 302 B5B phalanx – 37 Mobridge 302 B21B phalanx A 38 Mobridge 302 B33B phalanx C 39 Mobridge 302 B34 phalanx – 40 Mobridge 302 B37B phalanx C 41 Mobridge 302 B42 phalanx B 42 Mobridge 302 B45C phalanx A 43 Leavenworth 101 B3B phalanx A 44 Leavenworth 101 B9 phalanx A 45 Leavenworth 101 B20 phalanx B 46 Leavenworth 101 B30 phalanx A 47 Leavenworth 101 B35 phalanx B mtDNA of Protohistoric Arikara / 165

Table 1. (continued)

Sample Number Site Feature Burial Element Haplogroup 48 Leavenworth 101 B49 phalanx A 49 Leavenworth 101 B73 phalanx D 50 Leavenworth 102 B4 phalanx A 51 Leavenworth 102 B8 phalanx B 52 Leavenworth 102 B11A phalanx A 53 Leavenworth 102 B14 phalanx C 54 Leavenworth 102 B15 phalanx B 55 Leavenworth 102 B18D phalanx A 56 Leavenworth 102 B41A phalanx C 57 Leavenworth 202 B3 phalanx B 58 Leavenworth 202 B6 phalanx A 59 Leavenworth 202 B8A phalanx B 60 Leavenworth 202 B10B phalanx C 61 Leavenworth 202 B13 phalanx B 62 Leavenworth 202 B15 phalanx – 63 Leavenworth 202 B17C phalanx A

in the case of Neanderthal mtDNA sequences). For the Arikara samples, authenticity is more readily provided by finding a phylogenetically sensible sequence (e.g., a known Native American sequence or haplogroup or one closely related to it) than by a consensus sequence. The laboratory procedures described here were conducted by D. M. Lawrence, and haplogroups of all samples that were initially assigned a haplogroup were independently confirmed, using the same methods, by one of two co-authors (M. Snow or D. George). Approximately 0.5–1.0 g of bone was carefully removed from the whole sample. Surface decontamination and DNA extraction were conducted according to the methods of Kemp et al. (2007a). Half-milliliter tubes were placed under ultraviolet light for 15 min. Polymerase chain reactions and screening of the samples for the diagnostic markers of haplogroups A, B, C, D, and X were conducted according to the methods described by Kemp et al. (2007b). Sequencing of at least 237 bp (16131–16368) of the mtDNA control region was conducted at Trace Genetics Inc., Richmond, California. DNA was amplified in 25-µL PCR reactions containing 0.2 mM dNTPs, 1× PCR buffer, 2 mM MgCl2, 0.4 µM of each primer, 0.4 µL of Platinum Taq (Invitrogen, Carlsbad, California), and 2 µL of template. PCR conditions were as follows: 95°C for 2 min, followed by 40 cycles of 94°C for 10 s, 55°C for 10 s, and 72°C for 20 s, and a final extension at 72°C for 3 min. Primer coordinates are provided in Table 2 and are num- bered according to the modified Cambridge Reference Sequence (Anderson et al. 1981; Andrews et al. 1999). All samples for which a haplogroup assignment could be made were first amplified for the 411-bp and 275-bp fragments (including primers) using the L15995-H16368 and L16131-H16368 primers, respectively. Amplification of the 166 / lawrence et al.

Table 2. Primer Coordinates for Sequencing HVSI

Primer Primer Sequence (5 to 3) L15995 TCCACCATTAGCACCCAAAG L16131 CACCATGAATATTGTACGGT L16209 CCATGCTTACAAGCAAGT H16218 TGTGTGATAGTTGAGGGTTG H16368 TCTGAGGGGGGTCATCCA H16420 GCACTCTTGTGCGGGATATT

275-bp sequence in two overlapping (195 bp and 125 bp) segments using primers L16131-H16218 and L16209-H16368 was attempted for all samples for which neither the 411-bp nor the 275-bp fragment could be amplified. Any samples for which both the 195-bp and 125-bp fragments could not be amplified were elimi- nated from this part of the study. As a test of authenticity of the aDNA amplified, we independently attempted to amplify the 685-bp (15939–00018) and 191-bp (00034–00185) fragments of the hypervariable segment HVSI and HVSII re- gions, respectively, for a subsample of 18 of the samples for which either the 411- bp or the 275-bp fragment was amplified successfully. Because authentic aDNA is expected to be more easily amplified in short rather than longer sequences, we expected most samples to amplify for the 191-bp fragment but not for the 685-bp fragment. PCR product was filtered using a PerfectPrep PCR Cleanup kit (Eppen- dorf, Hamburg, Germany) by running products through filter wells for 10 min and adding 20 µL of molecular grade water to wells. Clean product was transferred to a new plate for the sequencing chemistry reaction. Each sequencing reaction contained 2.7 µL nuclease-free water, 2.7 µL of 2.5× BigDye v1.1 buffer, 0.5 µL BigDye v1.1, 0.8 µL sequencing primer (3 µM), and 1.3 µL clean PCR product. Sequencing conditions were as follows: 96°C for 1 min, followed by 30 cycles of 96°C for 10 s, 50°C for 5 s, and 60°C for 3 min. Sequenced product was cleaned using a Performa V3 96-well short plate (Edge Biosystems, Gaithersburg, Mary- land) and Sephadex (0.05 g/ml) and sequenced on an ABI 3100 Genetic Analyzer (Applied Biosciences, Foster City, California). Sequences were analyzed with Se- quencing Analysis Software, version 5.1 (Applied Biosciences). The haplogroup frequency distribution of the Arikara (combined data across all three sites) was compared with previously reported frequencies for the Mo- hawk (Northern Iroquoian) (Merriwether and Ferrell 1996), Sisseton-Wahpetan Sioux (Lorenz and Smith 1996; Malhi et al. 2001; Smith et al. 1999), and Chero- kee (Southern Iroquoian) (Lorenz and Smith 1996; Malhi et al. 2001) using a modification of Fisher’s exact test (Weir 1990) that employs a Markov chain method to estimate the exact p value over 1,000 iterations (Table 3). Calculations were performed using the GenePop on the Web software package (Raymond and Rousset 2002), and mean probability and standard error were reported. The 0.05 level of probability was regarded as the threshold of statistical significance. mtDNA of Protohistoric Arikara / 167

Table 3. Haplogroup Frequency Distributions for the Arikara and Other Comparison Populations

Haplogroup Population N A B C D X Arikara 55 0.42 0.29 0.22 0.07 0.00 Larson 16 0.19 0.37 0.25 0.19 0.00 Mobridge 19 0.58 0.16 0.26 0.00 0.00 Leavenworth 20 0.45 0.35 0.15 0.05 0.00 Mohawk 123 0.57 0.17 0.24 0.02 0.00 Sisseton-Wahpetan Sioux 45 0.56 0.20 0.18 0.04 0.02 Cherokee 56 0.14 0.38 0.46 0.02 0.00 Oklahoma Red Cross 19 0.21 0.21 0.53 0.05 0.00 Stillwell 37 0.11 0.46 0.43 0.00 0.00 Cheyenne/Arapaho 35 0.35 0.11 0.34 0.14 0.06 Chippewa/Ojibwa 149 0.41 0.08 0.25 0.02 0.24 Wisconsin 62 0.28 0.05 0.35 0.03 0.29 Turtle Mountain 28 0.57 0.18 0.18 0.00 0.07 Manutoulin Island 33 0.33 0.10 0.27 0.04 0.27 Northern Ontario 26 0.64 0.04 0.07 0.00 0.25

FST values were calculated in Arlequin, version 2.000 (Schneider et al. 2000) for all pairs of the populations (Arikara, Mohawk, Sioux, and Cherokee), two Algonquian-speaking groups that lived in close proximity to the Arikara [the Cheyenne/Arapaho and Wisconsin Chippewa (Malhi et al. 2001)], and three Al- gonquian-speaking groups that were more geographically distant from the Ari- kara [the Wisconsin Chippewa, Manutoulin Island Ojibwa, and Northern Ontario

Ojibwa (Malhi et al. 2001)]. These FST values were used to construct a distance matrix input for a principal coordinates analysis, performed in the program Dist- PCoA using the Caillez method to correct for negative eigenvalues (Legendre and Anderson 1999). Median-joining haplotype networks were constructed for each haplogroup (Bandelt et al. 1999) and included HVSI sequences that have been reported for others of the same groups included in this analysis.

Results Haplogroups could be assigned to 55 (87%) of the 63 samples studied from the three Arikara sites (see Table 1). All 55 samples exhibited the mutations diagnostic of one of the four most common haplogroups in Native Americans, and therefore none of the samples were tested for membership in haplogroup X. In- dependent extraction, amplification, and restriction analysis of each of these 55 samples by one of us (M. Snow or D. George) confirmed the original haplogroup assignments in all cases. The haplogroup frequencies varied among the three Ari- kara sites with Larson, Mobridge, and Leavenworth exhibiting lower than average frequencies of haplogroups A, B, and C, respectively, as shown in Table 3 together 168 / lawrence et al.

Table 4. Fisher’s Exact Test Results

Population Probability Standard Error Significancea Arikara/Sioux 0.47665 0.00768 – Arikara/Mohawk 0.03814 0.00240 + Arikara/Cherokee 0.00094 0.00031 + Sioux/Mohawk 0.31028 0.00815 – Sioux/Cherokee 0.00004 0.00004 + Mohawk/Cherokee 0.00000 0.00000 + Oklahoma Red Cross Cherokee/Stillwell Cherokee 0.14327 0.00475 – a. Statistically significant (+) or nonsignificant– ( ) at the 0.05 level of probability.

with the haplogroup frequency distributions of the comparison populations. When frequencies of haplogroups C and D (the haplogroups with the lowest frequencies) were combined for a chi-square test, reduced to 4 degrees of freedom to meet assumptions under which this test is valid, the differences among haplogroup frequency distributions of the three Arikara sites were not statistically significant (χ2 = 6.7, p > 0.15). Thus, for subsequent genetic analyses whose results are described in what follows, the data from all three sites were pooled. Twenty-three of the combined total of 55 samples belonged to haplogroup A, 16 to haplogroup B, 12 to haplogroup C, and 4 to haplogroup D (Table 3). Table 4 reports p values based on Fisher’s exact test for each paired comparison generated from GenePop. The haplogroup frequency distributions of the two Cherokee groups cited in Table 3 were not statistically significantly different from each other, but the distribution for both Cherokee groups combined was statistically significantly different from that of each of the other groups. The data show significant relationships (i.e., nonstatistically significant Fisher’s exact test) between the Arikara and the Sioux and between the Sioux and the Mohawk, but the haplogroup frequency distributions of the Arikara and the Mohawk are statistically significantly different at the 0.05 level of probability (i.e., p < 0.038). The principal coordinates plot, illustrated in Figure 2, shows that the Ari- kara are most closely related to the Sisseton-Wahpetan Sioux and the Mohawk, with these three being the only populations in the lower left-hand quadrant of the principal coordinates plot. The next closest population to the Arikara is the Turtle Mountain Chippewa, the most geographically proximate Algonquian-speaking tribe to the Arikara. However, the Arikara are more distant from the Cherokee and, especially, other more geographically distant Ojibwa and Chippewa tribes. These results closely concur with Fisher’s exact tests of differences between these groups’ haplogroup frequency distributions and suggest a close relationship among the Arikara, Sioux, and Mohawk as well as admixture with Algonquian groups in close geographic proximity to these groups but not those more distant. mtDNA of Protohistoric Arikara / 169

0.2 Chippewa- Wisconsin 0.15 Ojibwa- Manutoulin Is 0.1

Ojibwa 0.05 Cherokee- -N.Ontario Cheyenne/Arapaho OK Red Cross

0 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 Ojibwa- -0.05 Turtle Mountain Cherokee- Stillwell

Sisseton/Wahpeton -0.1 Arikara First Principal Coordinate (19.8%) Mohawk

-0.15 First Principal Coordinate (50.7%)

Figure 2. Principal coordinates plot based on mitochondrial DNA haplogroup distances (FST).

Of the 63 samples studied, no DNA remained for sequencing of one sample and sample contamination precluded amplification for sequencing another. Of the remaining 61 samples, either the 411-bp or the 275-bp fragment could be amplified in 51 (84%) of the samples, and 237 bp of sequence for analysis could be amplified in the two smaller, overlapping fragments for one of the failed samples. Of these 52 samples, at least 275 bp of sequence (including primers) could be generated for 42 samples. Fourteen of the subsample of 18 of these 42 samples for which attempts were made to amplify the 191-bp HVSII and 685- bp HVSI fragments yielded successful amplification of the 191-bp fragment but failure of amplification of the 685-bp fragment, suggesting authenticity ofthe aDNA extracted from the samples. Both fragments were successfully amplified for three of the samples, and neither sample could be amplified for one sample. To maximize sample size, all sequences were trimmed to 237 bp of common sequence. The HVSI sequences of the 42 samples for which at least the 275-bp HVSI fragment was successfully amplified are provided in Table 5. Each haplotype exhibited most, if not all, of the control region mutations characteristic of its haplogroup. For example, all haplogroup A haplotypes exhibited the 16223T, 16290T, 16319A, and 16362C mutations, and all haplogroup C haplotypes exhibited the 16223T, 16298C, 16325C, and 16327T mutations (except one that lacked the 16325C mutation), consistent with their haplogroup assignments based on restriction analysis. The median-joining haplotype networks are depicted in Figure 3 for haplogroups A and C. Sequences could not be obtained for any of the four samples that were members of haplogroup D, and the sequence data collected for members of haplogroup B were too short to be informative. 170 / lawrence et al. Read Sequence 16003–16367 16150–16367 16057–16367 16003–16367 16156–16367 16021–16367 16051–16367 16164–16367 16003–16367 16003–16367 16003–16367 16023–16367 16041–16367 16056–16367 16061–16367 16003–16367 16003–16367 16023–16367 16052–16367 16023–16367 T C C C C C C C C C C C C C C C C C C C C 16362 G A 16335 A A A A A A A A A A A A A A A A A A A A G 16319 T C 16311 A G 16309 T T T T T T T T T T T T T T T T T T T T C 16290 T C C 16263 T T T T T T T T T T T T T T T T T T T T C 16223 C T 16209 T T T T T C 16192 C 16187 ? ? ? T A G 16129 ? ? ? T C C C A C 16126

? ? ? T T T T T T T T T T T T T C C 16111

Sample Leavenworth B18D Leavenworth 202 B17C Leavenworth Leavenworth 102 B4 Leavenworth 102 B11A Leavenworth Leavenworth 101 B49 Leavenworth Leavenworth 101 B9 Leavenworth 101 B30 Leavenworth Leavenworth 101 B3B Leavenworth Mobridge 302 B45C Mobridge 201 B16H Mobridge 201 B29A Mobridge 302 B21B Mobridge 201 B11F Mobridge 201 B4C Mobridge 201 B2B Mobridge 101 B25D Mobridge 101 B25C Larson 301 B25 Larson 301 B33A Anderson Larson 101 B21B HVSI Sequence Variation Among Arikara Samples by Haplogroup Among Variation HVSI Sequence 55 63 50 52 48 44 46 43 42 33 34 37 32 30 29 28 27 17 19

4 Haplogroup A Haplogroup Sample ID Table 5. Table mtDNA of Protohistoric Arikara / 171 16162–16193 16152–16193 16059–16193 16003–16193 16003–16193 16158–16193 16156–16193 16003–16193 16003–16193 16155–16193 16003–16193 16021–16193 16021–16193 16003–16193 16150–16367 16150–16367 16150–16367 16150–16367 16150–16367 16150–16367 16003–16367 16150–16367 T C C C C C C C C C C C C C C 16189 C C C C C C C C C C C C C C A 16183 T T T T T T T T C C C C C C C A 16182 16327 T C C C C C C C A A A A A A A A A A A A A 16181 16325 ? ? T C G A 16156 16311 ? ? ? ? ? T T T T C C C C C C C C C 16111 16298 ? ? ? ? ? T T C C 16104 16294

? ? ? ? ? T T T T T T T T T C C 16093 16223 Larson 101 B4 Larson 101 B16 Larson 301 B3 Larsen 301 B65 Mobridge 101 B10E Mobridge 101 B12B Mobridge 302 B42 Leavenworth 101 B20 Leavenworth Leavenworth 101 B35 Leavenworth Leavenworth 102 B8 Leavenworth Leavenworth 102 B15 Leavenworth Leavenworth 202 B3 Leavenworth Leavenworth 202 B8A Leavenworth Leavenworth 202 B13 Leavenworth Anderson Larson 101 B16 Larson 301 B3 Larsen 301 B65 Mobridge 101 B10E Mobridge 101 B12B Mobridge 302 B42 101 B20 Leavenworth Larson 101 B4 Anderson 2 3 15 21 24 25 41 45 47 51 54 57 59 61

22 26 31 40 53 56 60 18

Haplogroup B Haplogroup C 172 / lawrence et al.

Haplogroup A S/W S/W

16067 A 16324

16261 A, P, S/W 16192 16187 C/O A 16129 A, C/O 16184 16126 S/W 16209 16111

16111 16111 A, C/O, Sh, A Cher, S/W 16209 Cher A, Chey 16093 16250 Sh

16325 16335 16309 16192 A C/O, S/W, M 16311 Chey A 16263 A A

Haplogroup C C/O, S/W 16189 Cher 16274 16183 16263 A C/O 16241 16311 S/W 16298

A, S/W 16325

16294 16362

16203 A, S/W, C/O, Chey

C/O Figure 3. Networks for haplogroups A and C. The central haplotype of the haplogroup A network exhibits mutations at 16111, 16223, 16290, 16319, and 16362 relative to the Cambridge Reference Sequence (Anderson et al. 1981; Andrews et al. 1999). The largest haplotype node in the haplogroup C network exhibits mutations at 16223, 16298, 16325, and 16327 relative to the Cambridge Reference Sequence (Anderson et al. 1981; Andrews et al. 1999). A, Arikara; C/O, Chippewa/Ojibwa; Sh, Shawnee; P, Pawnee; Cher, Cherokee; S/W, Sisseton-Wahpeton; Chey, Cheyenne; M, Micmac. mtDNA of Protohistoric Arikara / 173

The Arikara shared derived haplotypes of both haplogroup A and haplogroup C in common with only the two Algonquian-speaking groups most geographically proximate to them (Cheyenne-Arapaho and Chippewa) and the Siouan-speaking group (Sisseton-Wahpetan Sioux), suggesting recent admixture between the Arikara and both Siouan and Algonquian groups. Haplotypic comparison to the Mohawk was not possible because no HVSI sequences are available for them. The Arikara share the deepest lineage of haplogroup A, including its most highly derived haplotype, in common with the Pawnee (whose language is closely related to the Arikara language) and the Sisseton-Wahpetan Sioux. Almost all Cherokee haplotypes of haplogroup A were basal and therefore uninformative. The Sisseton-Wahpetan Sioux shared a highly derived haplotype of haplogroup C with the Chippewa, suggestive of recent admixture, and other Chippewa haplotypes appeared only at the periphery of the network, suggesting recent relationships to lineages represented by Arikara haplotypes. However, the most common nonbasal haplotype in the haplogroup C network, represented exclusively by Cherokee samples, derived from a one-step haplotype shared by both the Arikara and the Sisseton-Wahpetan Sioux.

Discussion The careful decontamination of samples, stringent precautions taken to minimize contamination of samples during their analysis, replicability of results by different co-authors, consistency of control region mutations with independent haplogroup assignments, and failure of most (15 of 18) attempts to amplify a 685-bp fragment of mtDNA all attest to the authenticity of the aDNA analyzed in this study. The haplogroup frequency distributions of the three Arikara sites varied, with Larson, Mobridge, and Leavenworth exhibiting lower than average frequencies of haplogroups A, B, and C, respectively, undoubtedly a result of both sampling and genetic drift. The frequencies from Leavenworth may have been influenced by population bottlenecks resulting from European American contact, and the massacre at Mobridge suggests that warfare may have precipitated precontact genetic bottlenecks. Although the three populations represent independent samples of a heterogeneous population in time and space and thus reflect different population histories, combining the three for analysis probably reduced the influ- ences of independent stochastic effects on the frequencies of the three populations and provided a more reliable estimate of their common ancestral Arikara forebears. Because all 55 samples could be assigned to one of the four most common Native American haplogroups, we assumed that haplogroup X was either absent from the Arikara or occurred with a sufficiently low frequency to not have been sampled. Moreover, no compound haplogroups (i.e., those with the defining markers for more than one of the haplogroups) were identified using restriction analyses designed to identify the four most common haplogroups. Given the reported low frequency of compound haplogroups in the Americas (Lorenz and 174 / lawrence et al.

Smith 1996), it is unlikely that any compound haplogroups involving haplogroup X were undetected in the present study. The high frequency of haplogroup A exhibited by the Arikara is not surpris- ing because haplogroup A is the most common of the five haplogroups in native North Americans (Eshleman et al. 2003; Schurr and Sherry 2004). Both the Mo- hawk (Merriwether and Ferrell 1996) and the Sisseton-Wahpetan Sioux (Lorenz and Smith 1997; Malhi et al. 2001) share similar frequencies of haplogroups A, B, C, and D with the Arikara, in support of the Macro-Siouan hypothesis. All three groups have the highest occurrence of haplogroup A (43–57%), relatively equal occurrences of haplogroups B and C (17–29%), and a low occurrence of haplogroup D (2–7%), leading to similar positions in the principal coordinates plot and nonsignificant or nearly nonsignificant Fisher’s exact tests. The Oklahoma Red Cross Cherokee and Stillwell Cherokee (Bolnick and Smith 2003; Lorenz and Smith 1996; Malhi et al. 2001) differ from the Arikara, Sioux, and Mohawk in having higher frequencies of haplogroup C (0.525 and 0.433, respectively) and lower frequencies of haplogroup A (0.211 and 0.108, respectively) than would be expected of ancestors who experienced perfectly sym- metric admixture with Algonquian-speaking groups. Under the Macro-Siouan hypothesis, the relatively high frequency of haplogroup A and relatively low frequency of haplogroup B in all three groups suggest a moderate level of admixture with Algonquian groups. Alternatively, the Cherokee might have also been influenced by stochastic processes that altered haplotype frequencies and therefore may not be representative of the hypothetical ancestral population from which the Cherokee, Mohawk, and Arikara descended. Although these results are consistent with the Macro-Siouan hypothesis, Algonquian tribes are also dissimilar from the Arikara, Mohawk, and Sioux in having the highest frequencies of haplogroup X in the Americas (Bolnick and Smith 2003; Lorenz and Smith 1996; Malhi et al. 2001, 2002; Smith et al. 1999). The Sisseton-Wahpetan Sioux also exhibit haplogroup X, albeit in low frequency, suggesting recent admixture with Algonquian-speaking groups. Given the small sample size in the present study and relatively low frequency of haplogroup X even in some Algonquian-speaking populations, it is not sur- prising that we did not observe haplogroup X in a population that experienced moderate levels of admixture with Algonquian-speaking (or Siouan-speaking) populations. It would be interesting to determine whether any of the remaining 700 Arikara samples from which DNA has not yet been extracted belong to haplogroup X and, if so, whether or not their control region sequences are the same as or similar to those of the Sioux or Algonquian representatives of that haplogroup and fall on the periphery of the haplotype network, suggesting admixture with the Arikara as well. The Cherokee groups are strikingly dissimilar to the Mohawk and the other groups, supporting the hypothesis that the ancestors of the Mohawk di- verged from the Southern Iroquoian Cherokee thousands of years ago, migrated northward, and admixed with nearby Algonquian groups. The Arikara, Mohawk, mtDNA of Protohistoric Arikara / 175 and Siouan tribes might have admixed with the Algonquian as they migrated northward at approximately the same time as the Algonquians were expand- ing southward from their homeland in southern Ontario, and the subsequent northward migration of the Arikara might have led to admixture with Siouan groups, as Jantz (1973) hypothesized. The similarities in haplogroup frequency distributions among these three tribes could result both from their recent mutual admixture with Algonquian-speaking tribes and their remote common ancestry with each other. Sharing control region sequences is far more indicative of intertribal admixture than sharing similar haplogroup frequencies, especially when the tribes that are admixing do not share similar haplogroup frequencies. The sharing of only highly derived haplotypes (i.e., those on the periphery of a network) suggests admixture, whereas sharing interior nodes of a haplotype network suggests common ancestry. The networks for haplotypes of both haplogroups A and C exhibit highly derived haplotypes but few interior nodes that are shared between either the Ari- kara or Sisseton-Wahpetan Sioux and the Cheyenne-Arapaho or Turtle Mountain Chippewa, the most geographically proximate tribes to the Arikara. The sharing of the highly derived haplotype of the deepest lineage of haplogroup A among the Arikara, Pawnee, and Sisseton-Wahpetan Sioux might reflect the common recent ancestry of the Arikara with the Pawnee and recent admixture of the Arikara with the Sioux. Although no haplotypes are shared between the Cherokee and either the Arikara or Sisseton-Wahpetan Sioux, a common nonbasal Cherokee haplogroup C haplotype can be derived from a one-step haplotype shared between the Arikara and Sioux, suggesting an ancient common ancestry among the three groups. This suggests that the Arikara, like the Sioux, have experienced appreciable recent admixture with neighboring tribes after migrating to the Northern Plains, but that Caddoan, Siouan, and Iroquoian groups share a common ancestry, as implied by the Macro-Siouan hypothesis. This study is the first to successfully amplify and type mtDNA from the Arikara. The data collected in this study provide some support for the Macro- Siouan hypothesis proposed by Chafe (1976). However, changes in language and genetic structure are two separate processes that need not correlate with each other. Thus, although the results of this study are consistent with the language hypothesis, they provide no proof of it. The Mohawk more closely resemble the Siouan populations than they resemble the Cherokee, to whose language their own language is more closely related, suggesting significant admixture after migrating northward from the Southeast. Although the Mohawk frequency data resemble those of the Arikara, the two were statistically significantly different according to the Fisher exact test. With more sampling of the Arikara specimens and sequencing of Mohawk samples, the relationship between Northern Iroquoian groups and the Arikara may become more clear. Sequence data and haplotype analysis will further our understanding of the relationships among all the tribes included in this study. 176 / lawrence et al.

Acknowledgments Thanks to the late John McDonough for his help in the laboratory. This study was supported in part by the National Institutes of Health through grant RR005090. The remains are considered to be affiliated with the Arikara, and inventories have been filed with the National Park Service and provided to the Arikara, as required by NAGPRA. The Arikara have not requested repatriation of the remains, nor have they expressed any other wishes as to their disposition. The curation and scientific use of the remains are therefore the responsibility of the University of Tennessee, where they are curated. For this reason, we were not required to seek the Arikara’s permission and did not do so.

Received 3 September 2009; revision accepted for publication 21 January 2010.

Literature Cited Anderson, S., A. T. Bankier, B. G. Barrell et al. 1981. Sequence and organization of the human mitochondrial genome. Nature 290:457–465. Andrews, R. M., I. Kubacka, P. F. Chinnery et al. 1999. Reanalysis and revision of the Cambridge Reference Sequence for human mitochondrial DNA. Nat. Genet. 23:147. Bandelt, H.-J., P. Forster, and A. Roehl. 1999. Median-joining networks for inferring intraspecific phylogenies. Mol. Biol. Evol. 16:37–48. Bass, W. M., D. R. Evans, and R. L. Jantz. 1971. The Leavenworth Site Cemetery: Archaeology and Physical Anthropology. Publications in Anthropology 2. Lawrence, KS: University of Kansas. Blakeslee, D. J. 1994. The archaeological context of human skeletons in the Northern and Central Plains. In Skeletal Biology in the Great Plains: Migration, Warfare, Health, and Subsistence, D. W. Owsley and R. L. Jantz, eds. Washington, DC: Smithsonian Institution Press, 9–32. Bolnick, D. A., and D. G. Smith. 2003. Unexpected patterns of mitochondrial DNA variation among Native Americans from the southeastern United States. Am. J. Phys. Anthropol. 122:336–354. Byrd, J. E., and R. L. Jantz. 1994. Osteological evidence for distinct social groups at the Leavenworth site. In Skeletal Biology in the Great Plains: Migration, Warfare, Health, and Subsistence, D. W. Owsley and R. L. Jantz, eds. Washington, DC: Smithsonian Institution Press, 203–208. Chafe, W. L. 1976. The Caddoan, Iroquoian, and Siouan Languages. Paris: Mouton. Collins, M. J., K. E. Penkman, N. Rohland et al. 2009. Is amino acid racemization a useful tool for screening for ancient DNA in bone? Proc. Biol. Sci. 276:2971–2977. Cooper, A., and H. N. Poinar. 2000. Ancient DNA: Do it right or not at all. Science 289:1139. Eshleman, J. A., R. S. Malhi, and D. G. Smith. 2003. Mitochondrial DNA studies of Native Ameri- cans: Conceptions and misconceptions of the population prehistory of the Americas. Evol. Anthropol. 12:7–18. Eshleman, J. A., and D. G. Smith. 2001. Use of DNase to eliminate contamination in ancient DNA analysis. Electrophoresis 22:4316–4319. Fernandez, E., J. E. Ortiz, A. Perez-Perez et al. 2009. Aspartic acid racemization variability in ancient human remains: Implications in the prediction of ancient DNA recovery. J. Archaeol. Sci. 36:965–972. Handt, O., M. Hoss, M. Krings et al. 1994. Ancient DNA: Methodological challenges. Experientia 50:524–529. Hoffman, J. J. 1977. Archaeological inference in societal modeling: Social organization of the Bad River phase. Plains Anthropol. Mem. 13:21–27. Holder, P. H. 1970. The Hoe and the Horse on the Plains. Lincoln: University of Nebraska Press. Jantz, R. L. 1973. Microevolutionary change in Arikara crania: A multivariate analysis. Am. J. Phys. Anthropol. 38:15–26. Jantz, R. L., and D. W. Owsley. 1994. White traders in the Upper Missouri: Evidence from the Swan Creek site. In Skeletal Biology of the Great Plains: Migration, Warfare, Health, and Subsis- mtDNA of Protohistoric Arikara / 177

tence, D. W. Owsley and R. L. Jantz, eds. Washington, DC: Smithsonian Institution Press, 189–201. Johnson, C. M. 1994. A Chronology of the Middle Missouri Plains Village Sites. Draft report prepared for the Smithsonian Institution. Washington, DC: Department of Anthropology, National Mu- seum of Natural History. Johnson, C. M. 1998. The coalescent tradition. In Archaeology on the Great Plains, W. R. Wood, ed. Lawrence: University of Kansas Press, 308–344. Kaestle, F. A., and D. G. Smith. 2001. Ancient mitochondrial DNA evidence for prehistoric population movement: The Numic expansion. Am. J. Phys. Anthropol. 115:1–12. Kemp, B. M., R. S. Malhi, J. Mcdonough et al. 2007a. Genetic analysis of early Holocene skeletal remains from Alaska and its implication for the timing of the peopling of the Americas. Am. J. Phys. Anthropol. 132:605–621. Kemp, B. M., C. Monroe, and D. G. Smith. 2007b. Extraction and analysis of DNA from archaeological specimens. In Archaeological Chemistry: Analytical Techniques and Archaeological Interpretation, M. D. Glascock, R. J. Speakman, and R. S. Popelka-Filcoff, eds. Oxford, U.K.: Oxford University Press, 78–98. Legendre, P., and M. J. Anderson. 1999. Distance-based redundancy analysis: Testing multispecies responses in multifactorial ecological experiments. Ecol. Monogr. 69:1–24. Lehmer, D. J. 1971. Introduction to Middle Missouri Archaeology. Anthropological Papers No. 1. Washington, DC: National Park Service. Lehmer, D. J., and W. R. Wood. 1977. Buffalo and beans. In Selected Writings of Donald J. Lehmer, W. R. Wood, ed. Reprints in Anthropology 8. Lincoln, NE: J. and L. Reprint Co. Lorenz, J., and D. G. Smith. 1996. Distribution of four founding mtDNA haplogroups among native North Americans. Am. J. Phys. Anthropol. 101:307–323. Lorenz, J. G., and D. G. Smith. 1997. Distribution of sequence variation in the mtDNA control region of native North Americans. Hum. Biol. 69:749–776. Malhi, R. S., J. A. Eshleman, J. A. Greenberg et al. 2002. The structure and diversity within New World mitochondrial DNA haplogroups: Implications for the prehistory of North America. Am. J. Hum. Genet. 70:905–919. Malhi, R. S., B. A. Schultz, and D. G. Smith. 2001. Distribution of mitochondrial DNA lineages among Native American tribes of northeastern North America. Hum. Biol. 73:17–55. McWilliams, K. R. 1982. Investigation of the Population Prehistory of the Arikara. Memoir 3. Okla- homa City: Oklahoma Anthropological Society, 163–172. Merriwether, A., and R. E. Ferrell. 1996. The four founding lineage hypothesis for the New World: A critical reevaluation. Mol. Phylogenet. Evol. 5:241–246. O’Rourke, D. H., M. G. Hayes, and S. W. Carlyle. 2000. Spatial and temporal stability of mtDNA haplogroup frequencies in native North America. Hum. Biol. 72:15–34. Owsley, D. W. 1979. A demographic analysis of skeletons from the Larson site (39WW2), Walworth County, South Dakota: Vital statistics. Am. J. Phys. Anthropol. 51:145–154. Owsley, D. W. 1981. Mobridge Site cemeteries: Controversy concerning the location of the Over and Stirling burials. Plains Anthropol. 26:43–48. Owsley, D. W., H. E. Berryman, and W. M. Bass. 1977. Demographic and osteological evidence for warfare at the Larson site, South Dakota. Plains Anthropol. 22:119–131. Parks, D. R. 1979. The Northern Caddoan languages: Their subgroupings and time depths. Neb. Hist. 60:197–213. Raymond, M., and F. Rousset. 2002. GenePop on the Web: Population Differentiation Option. http:// wbiomed.curtin.edu.au/genepop/. Schneider, S., D. Roessli, and L. Excoffier. 2000. Arlequin ver. 2.000: A Software for Population Genetics Data Analysis. Geneva: Genetics and Biometry Laboratory, University of Geneva. Schurr, T. G. 2004. The peopling of the New World: Perspectives from molecular anthropology. Annu. Rev. Anthropol. 33:551–583. Schurr, T. G., and S. T. Sherry. 2004. Mitochondrial DNA and Y chromosome diversity and the peopling of the Americas: Evolutionary and demographic evidence. Am. J. Hum. Biol. 16:420–439. 178 / lawrence et al.

Smith, D. G., R. S. Malhi, J. A. Eshleman et al. 1999. Distribution of mtDNA haplogroup X among native North Americans. Am. J. Phys. Anthropol. 110:271–284. Snow, D. 1994. The Cherokee. Malden, MA: Blackwell, 9. Tamm, E., T. Kivisild, E. Reidla et al. 2007. Beringian standstill and spread of native American found- ers. PLoS One 2(9):e829. Trimble, M. K. 1994. The 1837–1838 smallpox epidemic on the Upper Missouri. In Skeletal Biology in the Great Plains, D. E. Owsley and R. L. Jantz, eds. Washington, DC: Smithsonian Institu- tion Press, 81–89. Weir, B. S. 1990. Intraspecific differentiation. InMolecular Systematics, D. M. Hillis and D. C. Moritz, eds. Sunderland, MA: Sinauer Associates, 373–410. Copyright of Human Biology is the property of Wayne State University Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.