The human bone materials designated as the Kennewick skeleton [CENWW.97.Kennewick] were recovered in shallow water adjacent to an embankment of the Columbia River near the community of Kennewick, Washington, over a period from late July 1996 to December 1996. The skeleton was found in a disarticulated condition distributed over an area of at least 30 square meters, about 3 meters offshore and in about 45 cms of water.
Attached as Appendix 1 to this report is a summary of the various lines of evidence dealing with contextual and chronological data associated with Kennewick skeleton. As summarized in that report, this skeleton represents one of the most complete early Holocene human skeletons so far recovered in the Western Hemisphere. On morphological grounds, several physical anthropologists have argued that Kennewick exhibits characteristics that associate it with modern South Asians and Europeans rather than with modern Native Americans or contemporary indigenous populations of Northeast Asia.
Because of the interest and legal issues surrounding the disposition of the Kennewick skeleton, extensive scientific studies were initiated. One aspect of these studies sought to obtain mitochondrial DNA (mtDNA) restriction and sequencing data on the skeleton in an effort to confirm or refute the inferences developed on the basis of skeletal morphological criteria.
In conjunction with radiocarbon (14C) analysis--first in 1996 and then in 1999--it was determined that Kennewick bones exhibited a significant range in the degree of preservation of their intact collagen component. While the 1996 analysis obtained by the UCR Radiocarbon Laboratory (UCR-3476/CAMS-29578, 8410±60 BP) was obtained on a bone (5th left metacarpal) which contained a little less than 70% of the amino acid carbon content (AACC) of a modern bone standard and exhibited a clear collagen-like amino acid profile, the four 1999 14C analyses were obtained on Kennewick bones with low AACC values and at least two exhibited non-collagen-like amino acid profiles. As a result, three of the four 14C values obtained in the 1999 testing by three 14C labs appeared to be anomalous (see Table 1 in Appendix 1).
In an effort to maximize the chance that uncontaminated mtDNA could be isolated, it was determined that a series of samples from various Kennewick bones would be obtained and the amino acid content of each sample determined to provide quantitative data that would index the degree to which these bones had retained their original collagen content. It was hoped that a bone that had retained significant amounts of intact collagen would also be a good candidate for the extraction of non-contaminated mtDNA.
2. Bone Samples
Table 1 lists the three Kennewick bones on which previous analyses in 1996 and 1999 had been obtained and the eight bones from the Kennewick skeleton that were collected by the author on April 26-27, 2000, at the place of current curation of the skeleton, the Burke Museum of the University of Washington, Seattle. Because of the high sensitivity of the HPLC technology used to obtain the amino acid composition, it was possible to minimize the amount of sample bone that had to be taken. In seven of the eight samples, the amount of bone sampled was 1 gram or less.
In two cases (UCR-3878/3879 and UCR-3876/3877), two ends of the same bone were sampled to determine the degree of variability that would be exhibited in the same bone.
3. Amino Acid Composition Analyses
Table 1 lists the results of the measurement of the total amino acid composition of each of the bones selected and a determination if the sample exhibits a collagen-like (C) or non-collagen-like (NC) amino acid profile. An HP 1100 HPLC with pre-column derivatization using OPA (o-phthaldialdehyde) and fluorescence detection was employed to obtain each analysis.
Tables 2 through 9 list the data obtained from the HPLC analysis. In each case, the basis of the calculation and chromatogram are presented. Included in the computations are the ratios of aspartic acid, glycine, and alanine to glutamic acid. These ratios are employed as indices of the degree to which the collagen amino acid profile is retained in each bone.
Six of the eight Kennewick bone samples retained their collagen-like amino acid profile with an amino acid carbon content (AACC) ranging from about 46% to about 18% of a modern bone standard. Different parts of the same bone also exhibited significant variability. The distal end of the 3rd right metacarpal contained about 30% AACC of modern bone while the proximal end only retained about 10% AACC while the right 8th rib varied from about 30% AACC at the vertebral end to about 20% at the sternal end.
4. Stable Carbon Isotope Analysis
Sufficient samples were available to obtain d13C values on each of the eight Kennewick bone samples. The range in values is attributed to the differential retention of various amino acids as a function of the degree of preservation of the collagen matrix. Excluding the samples exhibiting non-collagen amino acid profiles, the mean d13C value (N = 7) is -13.5 per mil.
The d13C values exhibited by the Kennewick skeleton is interpreted to reflect the presence of a marine reservoir signal presumably reflecting a high percentage intake of marine derived biomass, i.e., salmon, in the diet of Kennewick Man. Salmon was assumed as the dominant fish in the diet based on contact period ethnographic accounts and archaeological data although steelhead (Oncorhynchus spp.) has also been suggested. Such a reservoir signal needs to be considered in evaluating the 14C values obtained on the Kennewick bones.
The data obtained on Kennewick bones collected in 2000 confirm results obtained in 1996 and 1999 which indicated that there is a significant variation in the degree of intact collagen preservation in different portions of the Kennewick skeleton. The range in d13C values is also interpreted to reflect differential diagenetic effects. Variable post-depositional effects on bones from the same skeleton have been noted previously, but the data obtained on the Kennewick skeleton represent the most detailed and extensive study of these effects reported to date.
Despite the documentation that portions of the Kennewick skeleton, e.g., 5th left metacarpal (UCR-3476) and the 3rd left metacarpal (UCR-3875) retained significant amounts of intact collagen, regretfully attempts to recover autochthonous, non-contaminated mtDNA from one or more bones from the skeleton have not as yet been successful despite multiple attempts by several laboratories (D. G. Smith, personal communication).
Table 1. Kennewick microsamples: amino acid composition and stable carbon isotope data
Kennewick Skeleton: Chronological and Biomolecular Contexts"
This paper was presented at the 17th International Radiocarbon Conference, Israel, June 19-23, 2000. It will appear in the Proceedings of that conference published in the journal Radiocarbon.
ABSTRACT. A human skeleton recovered near Kennewick, Washington, USA in 1996 has been dated to the early Holocene on the basis of multiple 14C determinations, an analysis of a style of a temporally-diagnostic projectile point found embedded in the ilium of the skeleton, and geological investigations of the locality where the skeleton was recovered. Based on morphological criteria, the Kennewick skeleton, which currently represents one of the most complete early Holocene human skeletons so far recovered in the Western Hemisphere, appears to be more similar to those of modern South Asians and Europeans than to modern Native Americans or to contemporary indigenous populations of Northeast Asia. Unfortunately, to date, mitochondrial DNA (mtDNA) restriction and sequencing analysis designed to determine whether or not the Kennewick skeleton can be assigned to an Asian-specific mtDNA haplogroup has not been successful because DNA is either absent or could not be extracted from the Kennewick bones so far examined.
In late July 1996, a human skull was accidently discovered in shallow water adjacent to an embankment on the southern shore of a lake (Lake Wallula) created by a flood control and hydroelectric dam on the Columbia River at Columbia Park within the community of Kennewick, State of Washington, USA (46°13'N, 19°10'W). Over several months following the initial discovery of the skull, post-cranial skeletal parts were periodically collected at the site primarily by Dr. James C. Chatters, a local private archaeological consultant (Chatters 2000; Nickens 1998; McManamon 1999a).
Based primarily on various conventional morphological criteria along with the presence of historic artifacts, the skeleton, which has come to be referred to as "Kennewick Man" from the "Columbia Park site" was initially thought to be that of a historic contact period Euro-American settler. However, during the cleaning of the bone, Dr. Chatters noted the presence of an object embedded in the right ilium of the pelvis. CAT-scans were interpreted as suggesting that the object appeared to be similar to a "Cascade Point," a diagnostic projectile point associated with the early to middle Holocene Cascade Phase--one of the prehistoric archaeological complexes defined for the southwestern portion of Washington. A 14C determination obtained in August 1996 on a total amino acid fraction of a metacarpal bone yielded a 14C value (UCR-3476/CAMS-29578) of 8410±60 BP (Taylor et al. 1998).
The presence of an early Holocene human skeleton in North America exhibiting skeletal morphological features determined by several physical anthropologists with long experience evaluating North American aboriginal skeletal materials as uncharacteristic of recent Native American populations engendered both wide spread popular media and scientific interest. The term "Caucasoid" was associated with the remains and so reported in the popular press. It had been initially used to characterize "gross morphology, not presumed origin" (Chatters 2000: 316). Unfortunately, such usage did not explain to the general reader the problematical historical contexts, semantic difficulties, and inappropriate connotations sometimes associated with the use of this term.
This paper will review issues primarily involving the overall contextual interpretation of the 14C data obtained on different bones from the Kennewick skeleton including problems of estimating the magnitude of reservoir corrections in light of the skeleton's d13C values and the reported wide range of protein preservation exhibited in different bones of what is assumed to be a single skeleton. We will also present the results of initial attempts to obtain mitochondrial DNA (mtDNA) restriction and sequencing analyses on the Kennewick skeleton.
The Kennewick Skeleton
The human bone materials now generally designated as the Kennewick skeleton [CENWW.97.Kennewick] from the Columbia Park Site were recovered in a disarticulated condition distributed over an area of at least 30 square meters, about 3 meters offshore and in about 45 cms of water. It has been assumed the sediments containing the remains had collapsed into the near shore water at a single point and then scattered by subsequent water action (Nickens 1998).
Based on a overall evaluation of skeletal morphometric data collected on the Kennewick skeleton, it was determined first by Chatters and subsequently confirmed by Powell and Rose (1998) that the Kennewick remains represent a single male individual, approximately 175 cm (5'9 in.) in stature, who had experienced a number of injuries throughout his life, and died between 45 and 50 years of age. A number of years before his death, he had broken two right ribs and had suffered a fracture of the right humerus. Perhaps at the same time, a projectile point was embedded within the right iliac blade of the pelvis. Most of the teeth show extreme wear with only trace amounts of enamel remaining on the incisors, canines, and premolars. Given the considerable attrition of the dentition, dental traits were difficult to characterize. From a taphonomic perspective, the Kennewick remains represent an individual who was most probably intentionally buried rather than left to decompose on the surface. In terms of the number of bone elements recovered, the Kennewick skeleton represents one of the most complete early Holocene human skeletons currently known from the Western Hemisphere.
Based on measurements of the skull, it was reported that the most similar samples appeared to be those from the south Pacific and Polynesia as well as the Ainu of Japan, a pattern from other studies of early Holocene American crania from North and South America (Steele and Powell 1992, 1994; Jantz and Owsley 1997). On the basis of an overall initial morphological evaluation, it was concluded that the Kennewick skeleton can be excluded, on the basis of its cranial morphology, from late Holocene American Native American groups (Powell and Rose 1999).
A visual examination determined that the raw material used in the manufacture of the artifact was a dark gray, medium-grained basalt or andesite. CT scans revealed that the artifact was bifacially worked and lanceolate (leaf-shaped) in overall shape. A typological assessment suggested that the artifact resembled a Cascade Point, a dart point type first characterized by Butler (1961:28-29) who considered it a diagnostic trait of the Cascade Phase, the second earliest of five phases originally defined for the Lower Snake River culture sequence of southwestern Washington (Leonhardy and Rice 1970).
Cascade phase assemblages are present in sites throughout the Pacific Northwest (Newman 1966, Nelson 1969, Rice 1972). They have often been associated with volcanic ash deposits of the Mt. Mazama eruption which critical reviews of the large corpus of 14C values place within a century of 6800 BP (Hallet et al. 1997: Table 1; Bacon 1983, Zdanowicz et al. 1999:623). However, there is some uncertainty about the temporal placement of different lithic variants associated with the Cascade phase in terms of whether they immediately pre- or post-date the Mt. Mazama eruption. Lancelolate (leaf-shaped) points with serrated edges were common in the pre-Mazaman assemblages from the Columbia Plateau. However, Fagan (1998:5) concluded that the Columbia Park specimen embedded in the Kennewick skeleton more closely resembles points most often immediately post-dating the Mt. Mazama ash. On this basis, in his view, the specimen is thought to represent "a tool made and used during Early Archaic times between 5,000 and 7,000 years ago." Chatters (2000: 298) identified serrated edges on the sample and, on this basis, suggested that it should predate the Mazama ashfall and thus date to between 5000 and 8000 BP.
and Geomorphological Analyses
The sediments from which the Kennewick Man remains were derived were analyzed as being composed of relatively fine-textured alluvium capped by an eolian/alluvial deposit, both modified by soil formation. A series of 12 embankment soil profiles (CPP = Columbia Park Profile) exposed by erosion and 6 cores (CPC = Columbia Park Core) collected from lower sediments at the water edge over a 300 meter section of river front have been examined to provide a more detailed reconstruction of the geologic context. Figure 1 presents the results of the geomorphological studies of these sections and cores along with a listing of 6 14C determinations obtained on fresh water shell and sediment humates (Huckleberry and Stein 1999: Figure 1). Figure 2 provides details on (i) 3 cores collected near to the location where the Kennewick skeletal materials were recovered, (ii) 4 14C values obtained on materials from one of these cores (CPC059.5) and (iii) two interpretations of the results of the geomorpholgical analysis (Huckleberry and Stein 1999: Figure 2). Both Figure 1 and 2 present the location of Mt. Mazama tephra in the soil sections.
Initial studies of the soil profiles (Wakeley et al. 1998) divided the sedimentary structure into five units (Units I-V) while a follow-on analysis (Huckeberry et al. 1998) characterized only two major lithostratigraphic units by grouping sedimentary Units I-III in a Lithostratigraphic Unit I and Units IV-V into a Lithostratigraphic Unit II. The later interpretation will be used in this discussion.
Lithostratigraphic Unit I is characterized as containing predominantly very fine to fine sand with no internal bedding or textural grading. Its lower portions contain discontinuous volcanic tephra which was identified as deriving from Mt. Mazama and thus dating to about 6800 BP. The underlying Lithostratigraphic Unit II was characterized as buried stratified well sorted fine sand formed by a series of overbank flood deposits mixed and modified by subsequent bioturbation and pedogenesis. A series of soil humates extracted from coring in this unit exhibited 14C values ranging in age from about 9,000 to 15,000 BP. The youngest of the soil humate 14C values (9010±50 [WW-1626/CAMS-44572]) is derived from the base of a concretion-bearing sediment within the upper zone of Lithostratigraphic Unit II which, on the basis of the comparison of organic and carbonate concentrations in the sediment and adhering to the Kennewick bones (see next paragraph), was inferred to be that which originally contained the Kennewick skeleton. Unfortunately, the exact nature of the organics comprising the soil humates used to obtain the 14C value was not specified. The investigators concluded that "if we assume that the Mazama tephra is in situ and that the 14C age [of the soil humate sample] is correct, then the geologically correlated age for the skeleton is 6700-9000 BP" (Huckleberry and Stein 1999: 22).
Sediments removed from the skeleton by Huckleberry and Stein (1999) were examined by a series of conventional optical and instrumental techniques including granulometry, thin-section (micromorphology), thermogravimetric, x-ray diffraction and trace-element analysis. In large part, the purpose of these studies was to match sediments from the skeleton and from the assumed discovery location at the Columbia Park site through a combination of physical and chemical tests. Special attention was focused on calcitic concretions which were discontinuously distributed over the surface of the bones giving them a"lumpy oatmeal" appearance. This was due to the determination that the soil profile at the discovery site contained concretion-bearing sediment which under visual inspection appeared to be very similar to that adhering to portions of the Kennewick skeleton. It was determined that the organic and carbonate contents of the concretions extracted from the skeleton and from sediment samples recovered from the upper part of the sequence designated as Lithostratigraphic Unit II were almost identical. The conclusion was that "the concretion on the skeleton was formed when the human remains were within the upper part of Lithostratigrapic Unit II" (Huckleberry and Stein 1999: 17).
There is an extensive and detailed literature on problems in the 14C dating of subfossil bone extending back for several decades (e.g. Taylor 1987: 53-61 with earlier literature cited; Stafford et al. 1988; Hedges and Law 1989; Taylor 1994; Hedges and Van Klinken 1992). All of these studies highlight the significant variability in the degree to which endogenous carbon-containing fractions in bone are retained and are, or are not, protected from contamination by a wide variety of physical and chemical diagenetic mechanisms. It is well known that obtaining accurate 14C age estimates on bone requires attention to detail in sample preparation and an appreciation that each bone may present an unique chemical challenge if the isolation of a fraction that contains only autochthonous carbon atoms is to be consistently achieved.
An important factor in obtaining accurate individual bone 14C values is the degree to which a bone sample has retained significant amounts of its principal protein component, collagen. To measure the degree to which collagen is retained in the bone sample, the UCR laboratory obtains a profile of the constituent amino acids of a total hydrolysate of the bone by ion exchange chromatography following extensive physical cleaning under magnification of the bone surface followed by sonication in dilute HCl. In the UCR laboratory, routine bone 14C analyses are undertaken on samples which exhibit a collagen-like amino acid composition and retain in excess of 5% of the amino acid carbon content (AACC) of a modern bone standard. This criterion was applied to the initial human bone analyzed by the UCR laboratory in 1996.
This sample (UCR-3476/CAMS-29578) exhibited a collagen-like amino acid profile and contained significant amounts of amino acid carbon (Table 1). However, all laboratories, including UCR, that analyzed the two Kennewick bones submitted in 1999 reported that the residual organic carbon content was very much reduced from the 1996 bone sample. For example, although UCR-3807/CAMS-60684 contained about 14.3% AACC, the amino acid profile was non-collagen-like while UCR-3806/CAMS-60683 exhibited a non-collagen-like amino acid profile and contained only 2.3% AACC. We interpret the anomalous d13C values for both of the UCR 1999 bone samples as reflecting a combined diagenetic and dietary signal. As a consequence, both of the UCR 1999 14C analyses have been expressed as fraction of modern with the equivalent 14C-concentration inferred age value reported as an "apparent 14C age." Although the Beta Analytical Laboratory reported a collagen yield of 0.3%, the 14C value reported (BETA-133993) was essentially identical to the 1996 value reported by the UCR laboratory. By contrast, both the University of Arizona 14C (AA-34818) and 13C values obtained on a sample with a carbon yield of 0.05% were significantly anomalous.
of Reservoir Effect
Assuming that early Holocene Columbia River salmon accumulated most of their biomass within the Gulf of Alaska in a manner similar to that observed today (Groot and Margolis 1991), we can calculated a marine offset correction for the Kennewick sample. Radiocarbon ages for early 20th century marine shell from the Gulf of Alaska average 860 BP (Robinson and Thompson 1981) while terrestrial ages were close to 110 years (Stuiver et al. 1998). On this basis, the marine reservoir correction for this region was 750 years.
The spread in marine shell ages from around the Gulf (Robinson and Thompson 1981; unpublished data) and results from archived pre-1950 salmon scales (Brown et al. 1988; T. Brown, pers. comm.) suggest that a reasonable estimate for the geographic variability in this correction is the equivalent of ±60 years. The scatter in paired wood-shell 14C ages from the British Columbia coast (Southon et al. 1990; unpublished data) indicate that early Holocene variations of this correction of up to ±150 years cannot be excluded. Adding these in quadrature, we calculate ±160 years as a conservative estimate of the overall uncertainly for the reservoir correction at 7000 to 8000 BP in the Gulf of Alaska.
Based on these results and the estimated marine dietary contribution, we calculate the corresponding marine reservoir offset for the Kennewick 14C values as 530±140. On the basis of these considerations, we have calculated a reservoir-corrected 14C age of 7880±150 BP for UCR-3476/CAMS-29578.
Dated Early Holocene Western North American Human Skeletons
Studies of mutational variations in mitochondrial DNA (mtDNA) in human cells have been employed to determine genetic relationships between and among contemporary and ancient populations (Schurr 2000). In contrast with mtDNA, nuclear DNA encodes in excess of 50,000 genes and is contributed from both an egg and sperm; it is thus subject to recombination during meiosis. The DNA in the mitochondria--the specialized energy producing organelle within cells--codes for only 37 genes or about 17,000 base pairs arranged in a circular structure. Most importantly, mtDNA is exclusively maternally inherited. The lack of recombination permits mutations to accumulate in maternal lineages and thus mtDNA can provide a record of human ancestry and migration. The presence of identical mutations in mtDNA from geographically separated human groups can provide compelling evidence of either common ancestry or contacts between these groups (Alberts et al. 1994: 704-16).
Studies carried out over the last decade on modern Native American populations have established that all members belong to at least 135 mtDNA haplotypes which in turn can be grouped within one of five matrilines or mtDNA haplogroups--designated as A, B, C, D, and X (Table 3, part A). Each mtDNA haplogroup can be characterized by distinctive mutations in the control (D-loop) and coding regions of the genome (Schurr et al. 1990; Torroni et al. 1993a, 1993b; Horai et al. 1993; Forster et al. 1996). Modern native Asian populations also contain individuals belonging to these same mtDNA haplogroups and the most straightforward explanation is that the mutations defining these haplogroups occurred first in Asia and were subsequently carried to the New World by ancestral Native Americans. Unfortunately, the distribution of mtDNA haplogroup frequencies among contemporary Native American populations does not correspond in any simple manner to New World indigenous language-group affiliations. Also, because of migrations, genetic drift and/or other processes, the genetic composition of ancient populations may not reflect that of the current population in the same geographic region (Schurr 2000).
Since each of these five New World mtDNA haplogroups is found in low frequencies in Asia (Ballinger et al. 1992; Merriwether et al. 1996), their presence in the New World in high frequency has been said to constitute evidence of a strong founder effect (Wallace et al. 1985), although this has been contested by others (e.g., Baillier et al. 1994; Easton et al. 1996). These views have been expressed against the backdrop of the fact that while one of the five mtDNA matrilines, haplogroup X, is found only in the southwestern and central parts of Asia, it is widespread in Europe (Torroni et al. 1996). Another, haplogroup B, while present in East Asia, is not found in modern Siberian populations, and only 3 of the 5 known modern New World mtDNA haplogroups are found in eastern Siberia (Torroni et al. 1993b).
It is exceedingly unlikely that two separate, independent migrations to the New World, even from the same point of origin, would be characterized by similar haplogroup signatures. Thus, in the event that some of the earliest Paleoindian populations are not ancestral to any contemporary Native American population, their haplogroup frequencies are unlikely to be similar. They would be most likely to include populations exhibiting one or more of the Asian-specific mtDNA haplogroups F, G, Y or Z that have not been reported in modern Native American populations (Table 3, Part B).
Obtaining ancient mtDNA from subfossil bone has been shown to be possible, due in part, to its presence in thousands of copies per cell. However, despite this relative abundance, the successful extraction and amplification of ancient mtDNA must be undertaken employing strict protocols to exclude and/or monitor contamination. As of this date, we are aware of studies that have characterized mtDNA in 12 prehistoric populations in North America including that previously obtained from 30 human skeletons ranging in age from 500 to 3500 years BP from the Congdon/Memaloose site located approximately 90 kilometers west of the Columbia Park site and from sites along the Snake River in Western Idaho dating about 2,000 BP (Malhi 2000). Other skeletal remains of comparable age and morphology as the Kennewick remains exhibit the same mtDNA haplogroups found in modern Native Americans (Smith et al. 2000).
The laboratory of one of us (DGS) has undertaken mtDNA studies on samples of Kennewick bone. Initial mitochondrial DNA (mtDNA) restriction and sequencing analyses of DNA extracted from the Kennewick skeleton have not been successful in identifying its mtDNA haplogroup since modern DNA from a known researcher who studied these skeltal remains coextracted during their analysis and was successfully sequenced. Also, primer-dimer consistently formed during amplification signifying tag polymerase activity in the PCR reaction. This outcome resulted from an absence of DNA remaining in the sample of bone studied rather than from inhibitor coextracted from the sample.
Geological, archaeological, and 14C data are consistent in assigning an early Holocene age to the Kennewick skeleton, one of the most complete human skeletons of that age so far reported from a New World site. From a morphological perspective, the Kennewick specimen appears to be more similar to those of modern South Asians and Europeans than to modern Native Americans or to contemporary indigenous populations of Northeast Asia. Regretfully, mitochondrial DNA (mtDNA) analyses have been unable to assign the Kennewick skeleton to an Asian-specific or any other haplogroup because to date no ancient DNA has been isolated from the samples of Kennewick bone so far examined.
The UCR Radiocarbon Laboratory is supported in part by the Western Center Community Foundation, the Gabrielle O. Vierra Memorial Fund, the Academic Senate Intramural Research Fund and the College of Humanities, Arts and Social Sciences, University of California, Riverside. The laboratory of DGS is supported by the National Science Foundation. Portions of this research were performed under the auspices of the US DOE, under contract W-7405-Eng-48. The dedicated laboratory work of Donna Kirner, Karen Selsor, Debra George (UCR) and Ripan Malhi (UCD) is acknowledged and very much appreciated. This is contribution 00/07 of the Institute of Geophysics and Planetary Physics, University of California, Riverside.