MATING PATTERNS AND GENE DYNAMICS OF A POPULATION ISOLATE OF NATIVE AMERICANS JEFFREY

C.

LoNG, FRANCINE

C.

ROMERO, MARGRIT URBANEK, AND DAVID GOLDMAN

Laboratory of Neurogenetics, National Insititute on Alcoholism and Alcohol Abuse, National Institutes of Health, Park V Building, Room 451, MSC 8/l0, 12420 Parklawn Drive, Bethesda, MD 20892-8/l0 (JCL, FCR, MU, DG) Department of Anthropology, University of New Mexico, Albuquerque, NM 87131 (FCR) Department of Human Genetics, University of Pennsylvania School of Medicine, Philadelphia, PA 19104-6145 (MU)

Mating structure can have important effects on population genetic phenomena, including inbreeding and genetic drift. However, data necessary to test predictions based on mathematical models or identify sensitivity to simplifying assumptions are difficult to collect. We used two sources of such data, pedigrees and genotypes, collected in a human-population isolate. The population studied was Native American and located in New Mexico. It was founded in the mid-19th century by ca. 30 individuals, primarily of Navajo origin, and its size increased steadily thereafter. A complete tribal pedigree spanning ca. 100 years (up to 1948) was collected by anthropologists starting in the 1920s. Probabilities of allelic identity by descent (IBD) within and among individuals were calculated for all generations directly from the pedigree. Wright's F-statistics were calculated from the IBD probabilities, and Ne was obtained from the statistic F ST' Genetic typings were performed on blood samples collected from the population between 1991-1993. A second set of F-statistics were calculated from genetic typings. Genetic kinship between individuals (FST) and average inbreeding within individuals (FIT) stabilized after the first two generations. However, FST was always greater than FIT of the next generation, suggesting that the net effect of social practices was inbreeding avoidance. In contrast to general expectations for growing populations, Ne increased over generations due to immigration. F-statistics estimated from the genetic typings were remarkably close to pedigree estimates, suggesting a drift-migration steady state. Key words:

human-population isolate, F-statistics, pedigrees, marker typings

Human-population isolates form natural experiments on effects of genetic drift and other processes that influence levels and patterns of genetic variation. They provide an opportunity to evaluate genetic consequences of phenomena such as mate-exchange systems, founder's effects, and population bottlenecks. These phenomena set the trajectory of evolutionary divergence in small populations but become less important and less analytically tractable as populations grow large. We studied a semi-isolated group of Native Americans from the Navajo tribe. Our analysis of its unique genealogical history and current gene freJournal of Mammalogy, 79(3):681--691, 1998

quencies provide a useful reference for understanding other populations, and we point to new directions for development and interpretation of population-genetic models. Isolation and finite population size make consanguineous matings unavoidable, regardless of mating structure. These constraints result in accumulation of allelic identity by descent (IBD), which occurs when two copies of an allele in the population are both copies of an allele that existed as a single copy in an earlier generation (Crow and Kimura, 1970). When two alleles are IBD within an individual, the person is inbred. When two IBD alleles ex681

682

JOURNAL OF MAMMALOGY

ist in different people, they share genetic kinship. Distribution of IBD alleles within and among individuals is a function of mating structure, while the rate at which IBD accumulates depends on both mating structure and population size (Cockerham, 1969, 1973). The extent of IBD in a population and its mating structure can be inferred from diverse data. We used pedigrees and genetic markers to estimate probabilities that randomly drawn genes within and among individuals from the population were IBD. Those probabilities were then analyzed with a population-genetic model. The model's parameters were F-statistics, which are functions of IBD probabilities, and effective population size (Wright, 1951, 1965, 1969). Our basic interest was in how the mating structure affected evolution and distribution of genotypes in the population. Although inbreeding for human isolates has been determined from pedigrees (Castilla and Adams, 1990; O'Brien et al., 1988), this is the first study to use such human data to compute a full array of F-statistics and effective population size. It is also the first, for a human population, to compare pedigree estimated F-statistics with F-statistics estimated from allele frequencies and incorporate effects of migration, which played an important role in maintenance of genetic diversity in this isolate. POPULATION BACKGROUND The focal population was founded between 1820-1840 when a few families moved into westcentral New Mexico where the town of Ramah is now located. They are geographically isolated by 40, 80, and 120 km from the three other Navajo communities in the region. This population was severely disrupted between 1863-1868 while the United States Army held all Navajo people captive at Fort Sumner, New Mexico (Bosque Redondo). However, the population re-established at its former location immediately following release in ca. 1870. The founders after incarceration were primarily Eastern Navajos, but there also

Vol. 79, No.3

were three Chiricahua Apaches, three Mescalero Apaches, and one Walapai Indian. Population growth by intrinsic increase and immigration of families and marriage partners followed. By ca. 1890, the population included 23 men, 30 women and 46 children. After 1890, a Laguna Indian, a Yaqui, a Zuni, and additional Navajos married into the population. Immigrants were accepted as marriage partners, but no new biological families settled after this time. By 1948, there were 614 members dispersed in small groups of nuclear and extended families over 1,334 km2 • By 1965, the population numbered ca. 1,100 individuals, and it was ca. 2,500 by 1990. Traditional Navajo culture predominated until the mid-1960s. It exists today but there is now more EuroAmerican influence. The culture, demography, and genealogy of the tribe have been studied extensively (Kluckhohn, 1966; Morgan, 1973; Spuhler, 1989; Spuhler and Kluckhohn, 1953). Social organization and marriage practices of Navajo affect the genetic structure. For example, mate choice is influenced heavily by the clan system. Clan membership is matrilineal, and post-marriage residence is matrilocal. Kluckhohn (1966) noted four general clan related prohibitions concerning appropriate marriage partners: marriage to members of one's own clan, marriage to members of their father's clan, marriage to members of clans linked with his or her own clan, or marriage to members of clans linked with his or her father's clan. Nevertheless, violations of these prohibitions occur and clan linkage is ambiguous, even among Navajo (Spuhler and Kluckhohn, 1953). Other cultural practices affecting mate choices among Navajo have potential genetic consequences. Until recent decades, polygyny (levirate and sororate) was practiced, and nuclear and extended families exhibited preferential patterns of reciprocal marriages. Cross-generational marriages between older women and younger men also were relatively common. Clan exogamy, as opposed to random mating,

August 1998

SPECIAL FEATURE-SOCIAL GENE DYNAMICS

will reduce inbreeding, but polygyny and preferential mate exchanges among nuclear families and lineages will increase inbreeding. We analyzed the balance between these opposing forces and their impact on genetic differentiation and effective population size. MATERIALS AND METHODS

Pedigree and demography.-Detailed demographic and genealogical records, unparalleled in completeness and spanning the period between 1870-1948, were collected by C. Kluckhohn and his associates. A copy of the genealogical chart of Spuhler and Kluckhohn (1953) was provided to the first author by the late Professor J. N. Spuhler. This chart recorded the complete genealogy of the population, beginning with 31 individuals (founders) in 1870 and ending with 614 individuals on 1 September 1948. In total, there were 1,105 individuals representing 317 fertile unions across seven generations. Familial relationships among some founding individuals were known and incorporated into subsequent calculations. The genealogical chart cannot be presented here because of its size and complexity, but Spuhler and Kluckhohn (1953) described it in detail and diagrammed some of the more interesting segments. Each subject was assigned to a pedigree generation as follows. Pre-founders were assigned to generation 0 and founders were assigned to generation 1. Each per~on born to a pedigree member was assigned to the first generation after the highest generation occupied by his or her parents. Migrants were identified as individuals who contributed offspring to later generations but could not be linked to the previous generations. All migrants were assumed unrelated to each other, or to pedigree members, and to be non-inbred. Migrants were assigned to the generation of their partners. Genotyping.-Blood samples and summary demographic infonnation were collected from adult Native Americans visiting the Albuquerque Indian Hospital, Albuquerque, New Mexico and other Indian Health Service facilities in New Mexico between 1991-1993. Thirty-four of these samples were donated by Navajos from the focus group. Genomic DNA was extracted directly from blood samples, or from Ebstein-Barr Virus transfonned lymphoblasts. Genotypings at

683

13 dinucleotide repeat loci located on chromosomes 9, 10, 11 and 20 were perfonned with an ABD 373a automated DNA sequencer (Applied Biosystems Division of Perkin Elmer, Foster City, CA) following PCR amplification using fluorescent dye labelled primers. Because of PCR failures and the use of multiplex procedures, the actual number of persons typed per locus ranged from 15 to 26. Allele frequencies were determined by direct counting. For comparative purposes, allele frequencies from six other populations of Native Americans in New Mexico were used in computing F ST• These included Navajos on the main reservation, a second isolate of Navajos, two populations of Apaches, and two populations of Pueblos. Allele frequencies for Pima and Cheyenne tribes and Finns and Swedes were included for FST calculations. These four latter populations were represented by control samples obtained for studies on alcoholism and other psychiatric disorders ·(Urbanek et al., 1996). F-statistics.-The basic quantities considered were Wright's F-statistics, which were computed from inbreeding and kinship coefficients. FIT equaled correlation of alleles within individuals relative to the founding population. FST equaled correlation of random alleles within the existing population, relative to the founding population. F Is equaled correlation of alleles within individuals, relative to the existing population. FIT was the average of inbreeding coefficients, F i • Thus, (1)

where N was the number of individuals in the population. FST was the average of genetic kinship coefficients, 9ij' (2)

where P = lhN(N - 1) was the number of pairs of individuals in the population. Proof that average probabilities of IBD given above are indeed correlations is given in Table 1. Following Wright (1951), F IS was obtained by manipulating the equation (1 - Ffr) = (1 - F sT)(1 - F ls ) where the prime denotes the offspring generation. Thus, (3)

Vol. 79, No.3

JOURNAL OF MAMMALOGY

684 TABLE

I.-Proof that average probabilities of identity by descent are correlations. Genic value"

GI Al Al A2 A2

G2

Probability of pair

Al A2 Al A2

p2(1 pq(I pq(1 q2(1

-

F'IT' This condition arises when there is

systematic avoidance of consanguineous mating (Wright, 1965). Population-genetic model.-A population-genetic model was developed to evaluate effects of the mating system. The model assumed a diploid population with discrete generations and without self-fertilization. Mating was random, and all individuals had the same expected number of progeny. The population size was finite but could change from one generation to the next, and there was immigration into the population. The fundamental parameters were Wright's Fstatistics, which were functions of population size and allelic identity by descent. From basic principles of probability (c.f., Crow and Kimura, 1970:102; Wright, 1969:291), the transition in FST across one generation was:

+ FIT)

Ne = ,

- FST 2

(FsTI(1 - m) - F ST )

(5)

•

The effective number defined by equation 5 was a variance effective size because FST was the quantity of change (Chepko-Sade et al., 1987; Crow and Kimura, 1970). A general formula for FST> at any generation, (G), since the isolate's formation, was obtained by applying equation 4 recursively and simplifying, G-]

2:

Fk~ =

(g, G)

+ C(O, G)

(6a)

g~O

where (g, G) = (1

+ FW)

(1 - m(g)2

2N(g) e

x

[TI (1 -~)(1l~g+]

2Ne

m(i)2] (6b)

and where the prime denotes FST in the next generation, N was the population size, and m was the proportion of immigrants. The F-statistics and their relationships with IBD probabilities were defined in equations 1-3. Under the model's assumptions, F'rr = FsT' That was not be the case with non-random mate exchange systems and violations of other model assumptions. For equation 4 to hold, N was replaced by an effective number, N e. Upon replacement with Ne and solving equation 4, the formula for Ne is

C(O G) = F(O) ,

ST

TI i~O

G-] (

1 - - 1 ) (1 - m(i)2 N~i)

(6c)

Components of this formula had important interpretations; (g, G) was the IBD contribution of the earlier generation (g) to the focal generation (G), and C(O, G) was the genetic kinship of the founders remaining in the focal generation. As such, C(O, G) + (0, G) was a measure of founder's effect and the (g, G) coefficients (g > 0) revealed bottlenecking. Statistical analysis of pedigree.-Using stan-

August 1998

SPECIAL FEATURE-SOCIAL GENE DYNAMICS

TABLE

Generation

2.-Demographic and pedigree derived statistics for the Navajo isolate. a

NB

NM

29 99 264 402 208 14

16 23 14 5

0 2 3 4 5 6

685

* *

N 31 45 122 278 407 208 14

m

FIT

0.36 0.19 0.05 0.01

0.0000 0.0000 0.0001 0.0057 0.0145 0.0153

* *

FST

F ls

0.0046 0.0152 0.0293 0.0266 0.0268 0.0292

-0.0154 -0.0301 -0.0214 -0.0127 -0.0143

Ne

NjN

47 9 42 153 159

1.52 0.20 0.42 0.58 0.40

'NB = number born, NM = number of migrants, N = NB + NM = total number, m = NMIN = migration fraction, all other symbols are introduced in the text. * = missing data, generation not completed.

dard pedigree methods, inbreeding coefficients were computed for all individuals in the population. Kinship coefficients were computed for all individuals in a generation if there were (g,G).

Generation receiving (G)a

C(O,G)

0

2 3 4 5

0.0046 0.0045 0.0016 0.0010 0.0009

0.0106 0.0039 0.0025 0.0022 0.0022

Generation contributing (g) 2 0.0228 0.0146 0.0131 0.0127

0.0078 0.0070 0.0068

3

0.0029 0.0028

4

FST

0.0031

0.0152 0.0312 0.0265 0.0262 0.0285

, Each entry is obtained from equation 6b using the data from Table 2.

the Ramah population. Ne exceeded N for the founding generation at Ramah, despite the fact that effective sizes are much smaller than actual sizes in most other populations. While the cause for this is uncertain, founders of a new deme were not a random sample of people born into the population. For example, they have escaped childhood mortality and many have already found mates. Analysis of the kinship contributions of earlier to later generations revealed several interesting phenomena (Table 3, Fig. 1C). The founder effects of generation 0 were greatly diminshed by the high migration rates in the next two generations. Generation I had the largest overall contribution

4.-Dinuckotide repeat loci analyzed in the Navajo isolate.

TABLE

and

PIS

Locus D9S273 D1OS192 D1OS547 DllS935 DllS937 D20S111 D20S114 D20S172 D20S177 D20S186 D20S193 D20S97 D20S171 Average

Number of different Sample alleles Gene size observed diversity 24 24 15 26 24 20 18 20 21 20 20 20 26

8 6 4 4 10 2 3 4 3 8 4 5 6

F]s

0.615 0.788 0.596 0.589 0.856 0.455 0.573 0.471 0.291 0.696 0.649 0.348 0.396

0.187 0.048 0.216 -0.044 0.027 -0.099 -0.454 0.151 -0.144 -0.221 -0.002 0.282 -0.167

0.563

-0.016

to inbreeding and genetic kinship, due to the small effective population size at this generation and low migration rates in later generations. Finally, a population bottleneck in the transitions from generation 3 to generation 5 was clear from the flatness of the (g,G) functions. Genotypic F-statistics.-The analysis of F-statistics from recently collected genotypes closely mirrors the pedigree predictions from ca. 40 years earlier. The point estimate of F[s averaged over the 13 loci was negative (-0.016) but it was not significantly different from zero (Table 4). FST for the Navajo isolate relative to other populations of Apaches and Navajos was 0.027, a little less than the final pedigree value (Table 5). A graph of the nested popUlation structure (Fig. 2) showed that after the reduction in heterozygosity leading from Europeans to Native Americans is accounted for, there was only a minor tendency for tribes of Native Americans to cluster into genetic supgroups. When the terminal branch lengths were converted to estimates of FST (Table 5), other populations showed a greater reduction in heterozygosity than was measured for the focal isolate. This suggested that all tribes are semi-isolates in their own right. DISCUSSION

Although the focal population became more inbred over time, on the balance, the mating system minimized inbreeding and may have provided the impetus for much of

TABLE 5.-FsT relative to nearest internal node of hierarchy (Fig. 2) for individual Native American and European populations. Navajo isolate 1 is the focal population of this study.

Population Navajo isolate 1 Navajo isolate 2 Navajo Apache 1 Apache 2 Pueblo 1 Pueblo 2 Pima Cheyenne Finn Swede

Vol. 79, No.3

JOURNAL OF MAMMALOGY

688

Gene diversity Sample size (average (average for for 13 13 loci) loci) 21 26 37 23 26 25 25 123 46 48 38

0.563 0.524 0.603 0.528 0.625 0.638 0.569 0.634 0.573 0.715 0.695

Hierarchical Population Structure

Isolate I

~-- Isolate 2 Navajo FST

Apache I

0.027 0.090 0.027 0.059 0.024 0.027 0.104 0.059 0.027 0.025 0.038

Apache 2 Pueblo I Pueblo 2 Pima Cheyenne Finn Swede

0.25

the immigration seen in the early generations when population size was small. The extent of inbreeding reduction would have been difficult to deduce by simultaneously modeling complicated factors such as polygyny, clan avoidance, and reciprocal marriage exchange between families. Also, mate exchange rules in society are often broken. Therefore, it is difficult to say how well an idealized system will predict an actual genetic architecture. Our sample of genotypes provides an F[s value that is close to the pedigree's. However, a jackknife standard error for the genotypic estimate is large, so that the result would have been equivocal in the absence of the pedigree analysis. The consequence of this mate exchange pattern is negative F[s values. This implies an excess of heterozygotes in a contemporaneous sample from the population. This is observed in other North and South American tribes (Neel and Ward, 1972; Workman et aI., 1973). Consanguinity avoidance has been suggested for these tribes, but differences in gene frequencies between sexes and other explanations for the excess of heterozygotes have been favored. A recent study on the Havasupai, a neighboring tribe of southwestern Native

I

0.30

I

0.35

I

0.40

0.45

I

0.50

I

I

Homozygosity

FrG. 2.-Hierarchy used for calculating FST of individual popUlations relative to their base (closest internal node). Isolate 1 is the focal population of this study. Isolate 2 is another Navajo isolate. Navajo is represented by a sample from the main reservation. According to the wishes of the study participants and tribal authorities, the exact identities of popUlations of Apaches and Pueblos are not given. Branch lengths measure homozygosity (loss of heterozygosity). FST is calculated from F HL = (J~ - JD/(l - JD where J~ is the homozygosity (gene identity) associated with the observed popUlation and J~ is homozgosity (gene identity) associated with its nearest internal node.

Americans, attributed an excess of heterozygotes at the HLA-A locus to balancing selection (Markow et al., 1993). In light of the present findings based on pedigrees and dinucleotide repeat loci whose dynamics are regulated by random drift, the balancing-selection hypothesis should be regarded with caution until a mating structure explanation can be ruled out. FST measures the divergence of the population from allele frequencies of the foun-

August 1998

SPECIAL FEATURE-SOCIAL GENE DYNAMICS

ders and pre-founders. The pedigree shows large increases in the first two generations, followed by an apparent plateau. Estimates of (g, G) show that after the second generation, the kinship contributed by a generation to its successors ~ently decays, but the amounts contributed by these generations is relatively small. This stability is unsurprising because equation 4 provides an eqUilibrium at FST (1 + Frs)/(4N em + 1) assuming that the product Nem remains constant across generations. With random mating in the population (i.e., Frs = 0), this reduces to the eqUilibrium for Wright's (1951, 1965, 1969) island model of population structure. However, a stable FST can arise for reasons other than a constant Ne and m. For example, Nem can remain stable if changes in population size are related inversely to changes in migration rates, as in the early generations of the study population. A base population from which genes in the isolate were drawn is implicit in FST computed from allele frequencies. We are fortunate to have sampled an array of populations of Navajos (including another isolate and the main reservation) from which to construct this base. The analysis of allele frequencies indicates that FST similar to that of the pedigree was maintained over the next 40 years. Interestingly, FST is not extreme in the focal Navajo Isolate as compared with other tribes of Native Americans. In fact, FST in the other Navajo isolate sampled is nearly three times as great, and it is considerably higher in large populations of Native Americans such as the Pima. This assuredly relates to the population disruption and unique histories of exposure to disease and population rebound that nearly all tribes experienced in the centuries following European contact (Ramenov sky, 1987). The effective population size, N e, is an important parameter that bridges complexities of natural populations with predictions of simplified models. Because several subtly different effective sizes have been de-

=

689

fined (Chesser et al., 1993; Kimura and Crow, 1963; Sugg and Chesser, 1994), it is important to point out that we have analyzed the variance effective size. Theory demonstrates that variance and other effective sizes are usually less, and often much less, than the actual size of a closed population (Kimura and Crow, 1963). In principle, the long-term effective size of a closed population that changes in size is approximated by the harmonic mean of population sizes over time. Accordingly, Ne is dominated by the smallest population size experienced, and population growth alone cannot substantially increase it. Nevertheless, the effective size of this Navajo isolate grew precipitously during the first four generations. Complexities imposed on Ne by population growth with migration are seen by examining generation 3. At this time, the effective size was 153 individuals which is only 58% of the total population number. However, the harmonic average of all effective sizes prior to this generation is only ca. 20 individuals. Growth of Ne must be attributed to the addition of independent genes by migration, but it is apparent that population growth compounds the effect of migration. To illustrate the compounding effect, consider that 53 migrants had joined this Navajo isolate by generation 3. In addition to the 31 founders, this yields 84 independent genomes, which is only about one-half of the measured effective size. These results show that, unless a population is isolated completely, the harmonic mean formula can be grossly inadequate for measuring effective size. Thus, Ne should not be estimated from population size changes alone (e.g., Chepko-Sade et aI., 1987), unless it can be demonstrated that the population is a complete isolate. This result also shows the power that can be obtained when pedigrees are available for analysis. Interestingly, recent theory and application to prairie dogs (Cynomys) shows that more complicated hierarchical social structures also can increase effective population size

JOURNAL OF MAMMALOGY

690

and retard loss of genetic variation (Sugg et aI., 1996). Genetic drift in the isolate was only slightly affected by the mating system. Differentiation of the isolate from the founders, or a base population, is quantified by FST' If we take F IS as the degree of non-random mating, the mating system influences the rate at which IBD accumulates (equation 4) through F'IT = 1 - (1 - F Is )(1 - F ST)' This effect is minimal with the slightly negative F IS observed. Similarly, equilibrium for FST == (1 + FIS)/(4Nem + 1) is not appreciably affected by the observed F IS' However, this is not to say that the relationship between the mating system and F IS is unimportant. Consanguinity avoidance encourages immigration and prevents inbred individuals, and perhaps genetic deaths, in the early generations of a newly formed isolate when population size is smallest and probability of extinction is highest. ACKNOWLEDGMENTS

We are indebted to Professor J. N. Spuhler (deceased) who enabled this study by generously providing his genealogical information. He also provided help and encouragement in the early phases of the project. The authors are responsible for all analyses and interpretations. Data collection for this project was funded in part by National Science Foundation grant BNS 91 08422. LITERATURE CITED CASTILLA, E. E., AND J. ADAMS. 1990. Migration and genetic structure in an isolated population in Argentina: Aicufia. Pp 45-62, in Convergent issues in genetics and demography (J. Adams, A. Hermalin, D. Lam, and P. Smouse, eds.). Oxford University Press, New York, 361 pp. CAVALLI-SFORZA, L. L., AND A. PIAZZA. 1975. Analysis of evolution: evolutionary rates, independence, and treeness. Theoretical Population Biology, 8: 127-165. CHEPKO-SADE, D. B., ET AL. 1987. The effect of dispersal and social structure on effective population size. Pp. 287-321, in Mammalian dispersal patterns: the effects of social structure on por" lation genetics (D. B. Chepko-Sade and Z. T. Halpin, eds.). The University of Chicago Press, Chicago, Illinois, 342 pp. CHESSER, R. K., O. E. RHODES, JR., D. W. SUGG, AND

Vol. 79, No.3

A. SCHNABEL. 1993. Effective sizes for subdivided populations. Journal of Genetics 135:1221-1232. COCKERHAM, C. C. 1969. Variance of gene frequencies. Evolution 23:72-84. - - - . 1973. Analysis of gene frequencies. Genetics 74:679-700. CROW, J. E, AND M. KIMURA. 1970. An introduction to population genetics theory. Harper & Row, New York, 591 pp. CURIE-COHEN, M. 1982. Estimates of inbreeding in a natural population. Genetics, 100:339-358. GREENBERG, J. H., C. G. TuRNER, AND S. L. ZEGURA. 1986. The settlement of the Americas: a comparison of linguistic, dental, and genetic evidence. Current Anthropology, 7: 195-200. KIMURA, M., AND J. E CROW. 1963. The measurement of effective population number. Evolution, 17:279288. KLUCKHOHN, C. 1966. The Ramah Navajo. Smithsonian Institution, Bureau of American Ethnology, Washington, D.C., Anthropological Papers, Bulletin 196, Number 79, 327-377 pp. MARKOW, T., ET AL. 1993. HLA polymorphism in the Havasupai: evidence for balancing selection. American Journal of Human Genetics, 53:943-952. MORGAN, K. 1973. Historical demography of a Navajo community. Pp. 2663-314, in Methods and theories of anthropological genetics (M. Crawford and P. Workman, eds.). University of New Mexico Press, Albuquerque, 509 pp. NEEL, J. v., AND R. K. WARD. 1972. Genetic structure of a tribal population, the Yanomama Indians. VI. Analysis by F-statistics, including a comparison with the Mkiritare and Xavante. Genetics, 72:639666. NEr, M. 1987. Molecular evolutionary genetics. Columbia University Press, New York, 512 pp. O'BRIEN, E. L., L. B. JORDE, RONNLOF, J. O. FELLMAN, A. W. ERIKSSON. 1988. Inbreeding and genetic disease in Sottunga, Finland. American Journal of Physical Anthropology, 75:477-486. RAMENOVSKY, A. E 1987. Vectors of death: the archaeology of European contact. University of New Mexico Press, Albuquerque, 300 pp. SPUHLER, J. N. 1972. Genetic, Linguistic, and geographical distances in Native North America. Pp. 72-95, in The assessment of population affinities in man (J. S. Weiner, and J. Huizinga, eds.). Clarendon Press, Oxford, United Kingdom. - - - . 1989. Update to Spuhler and Kluckhohn's "inbreeding coefficients of the Ramah Navaho population." Human Biology, 61 :726--730. SPUHLER, J. N., AND C. KLUCKHOHN. 1953. Inbreeding coefficients of the Ramah Navaho population. Human Biology, 25:295-317. SUGG, D. W., AND R. K. CHESSER. 1994. Effective population sizes with multiple paternity. Genetics, 137: 1147-1155. SUGG, D. W., R. K. CHESSER, E S. DOBSON, AND J. L. HOOGLAND. 1996. Population genetics meets behavioral ecology. Trends in Ecology and Evolution, 11: 338-342. URBANEK, M., D. GOLDMAN, AND J. C. LONG. 1996. The apportionment of dinucleotide repeat diversity in Native Americans and Europeans: a new ap-

August 1998

SPECIAL FEATURE-SOCIAL GENE DYNAMICS

proach to measuring gene identity reveals asymmetric patterns of divergence. Molecular Biology and Evolution, 13:943-953. WEIR, B. S. 1990. Genetic Data Analysis. Sinaur and Associates, Inc., Publishers, Sunderland, Massachusetts, 377 pp. WORKMAN, P. L., H. HARPENDING, J. M. LALOUEL, C. LYNCH, J. D. NISWANDER, AND R. SINGLETON. 1973. Population studies on southwestern Indian tribes. VI. Papago population structure: a comparison of genetic and migration analyses. Pp. 166-194, in Ge-

691

netic structure of populations (N. E. Morton, ed.). University of Hawaii Press, Honolulu. WRIGHT, S. 1951. The genetical structure of populations. Ann Eugenics, 15:323-354. - - - . 1965. The interpretation of population structure by F-statistics with special regard to systems of mating. Evolution, 19:395-420. - - - . 1969. Evolution and the genetics of populations. Volume 2. The theory of gene frequencies. The University of Chicago Press, Chicago, Illinois, 509 pp.