Population Studies on Southwestern Indian Tribes IV. THE ZUNI P. L. WORKMAN,' J. D . NISWANDER, K. S . BROWN AND W . C. LEYSHON Human Genetics Branch, National Institute of Dental Research, National Institutes of Health, Bethesda, Maryland 20014 KEY WORDS Amerindians. Genetic differentiation . Kinship . Distance . Zuni ABSTRACT The Zuni Indians of west-central New Mexico have been relatively isolated since their foundation by an amalgamation of individuals from different southwestern cultural areas during the Regressive Pueblo period (c.1200-1350 A.D.). Genetic analysis revealed a high frequency of blood type B in both young (0.06) and old (0.05) Zuni, but at 14 other blood group and serum protein loci, allelic frequencies including A (0.011) and Rh negative (0.001) were generally similar to those of other relatively unmixed southwestern Indian tribes. Consideration of Zuni history and demography since Spanish contact in 1540, together with genetic analyses, suggest that the high B frequency probably derives from intermixture with a small number of B, Rh positive non-Indians in the early post contact period. Genetic differentiation among four southwestern tribes, Zuni, Pima, Papago and Maricopa, was summarized by kinship analysis. Approximately 70% of the inter-tribal genetic variation could be explained by the geographic distances among these groups showing that isolation by distance has been the most important factor in determining the pattern of regional genetic differentiation. The Zuni Indians living in west-central New Mexico appear to be an amalgamation of at least two indigenous southwestern Indian groups. Cushing (1896) first described Zuni mythology and language as a fusion of distinct elements, and recent linguistic studies appear to confirm this (Newman, '64; Swadesh, '59, '67). The Zuni language is quite distinct from other New Mexico Pueblo languages and not related to the Uto-Aztecan language of the Papago, Pima and Hopi in Arizona. In addition to language affiliations, reconstruction of the history of the southwestern tribes must also account for tribal variation in cultural and biological factors. As part of an ongoing study of biological differentiation in the southwest we will restrict our focus in this paper to a genetic characterization of the Zuni and to the pattern of genetic differentiation among the indigenous southwestern Indian tribes. The genetical analyses contribute to the larger, yet unwritten, history through an understanding of the most probable patterns of AM. J. PHYS. ANTHROP., 41: 119-132. mate exchange (which may or may not have been accompanied by cultural or linguistic influences) and, by inference, on the likelihood that different groups may have descended from a common founder population. In order to interpret the pattern of biological variation it is first necessary to examine observations from archaeology, ethnohistory and historical demography. HISTORICAL OBSERVATIONS Prehistory to Spanish contact (I 540 A.D.) Man is believed to have first entered the southwest at least as early as 10,000 B . C . In the northern region (Nevada, Utah, Colorado) the first men were primarily big game hunters; their earliest known sites contain flint tools and projectile points in association with the bones of mammoth, long-horned bison, and traces of fire. In southern Arizona and New Mexico, however, the frequent occurrence of milling ' Present address: Department of Anthropology, University of Massachusetts, Amherst, Massachusetts 01002. 119 120 P. L. WORKMAN, J. D. NISWANDER, K. S. BROWN AND W. C. LEYSHON stones as well as bones, in the earliest sites, suggests a culture more dependent upon hunting and gathering. By 6000 B . C . men were widely dispersed in small bands and the heterogeneity among artifacts from different sites believed to be contemporaneous makes it likely that cultural diversification accompanied the dispersion. For several thousand years the southwestern peoples developed in relative isolation. Sites dating from 2500 B . C . show remains of baskets, sandals, milling stones and flint and bone tools. The entire southwest at that time appears to have been inhabited by small, seminomadic bands of hunters and gatherers. Under such conditions, cultural differentiation could proceed without significant biological divergence. The genetic structure might be well approximated by a model of partially isolated subgroups in which the amount of intermixture varies with the travel time between groups. During this same period there arose in the southern Mexican plateau a culture whose contributions were to dominate all further change in the southwest. The most important development in Mexico was the invention of horticulture and the domestication of squash, beans, maize and many other plants. By about 1000 B.C. these agricultural developments had been introduced into the southwest and the seminomadic groups were generally replaced by larger, sedentary agricultural settlements which may have had both greater cultural isolation and a higher degree of endogamy. Another, somewhat later, contribution from Mexico was pottery which was introduced into the region by about 200 B . C . After this period we can trace the development of three great culture areas: The Mogollon in southern New Mexico and southeastern Arizona; the Hohokam in southcentral Arizona; and, the Anasazi, centered on the four corners area, i.e., the common border of Utah, Colorado, Arizona and New Mexico. Each of these, of course, should not be taken to represent a single Mendelian population but rather a collection of groups of different sizes and degrees of isolation sharing numerous culture traits (burial practices, axe types, architectural style, pottery type, etc.). In addition, especially by analysis of pottery variants, local differentiation within these culture areas can also be recognized. There appears to have been sufficient cultural, and possibly biological, interchange among these areas so that between 500 A . D . and 1100 A . D . a generally uniform style of life obtained throughout the southwest. From about 700 A . D . there was an expansion of each of the cultures into new sites. The main thrust of the Anasazi after 900 A . D . was an expansion into the Rio Grande Valley, which, before then, had been sparsely populated. During this period the Anasazi began the construction of the first large communal houses, or Pueblos. One of the most famous, dating from about 1000 A.D. was that at Chaco Canyon in northwestern New Mexico. During the twelfth century, called the Great Pueblo period, the Anasazi rose to its highest point. The Indians moved from their small scattered settlements and congregated in the large pueblos, some of which had between 500 and 1,000 rooms. The thirteenth century, in marked contrast to the preceding period, is described as the Regressive Pueblo period. The great houses were gradually abandoned as was most of the territory which had been occupied by the Anasazi. In general, the population withdrew to the Rio Grande Valley but three somewhat isolated areas remained occupied in the west in sites now occupied by the Hopi, Zuni and Acoma tribes. A variety of reasons have been offered to explain these events: overpopulation; disease; attacks by hostile bands of Utes and Athapascans moving down from the northern plains; drought which made agriculture in the area almost impossible; and, a shift from dominant winter rain to heavier summer rain which inhibited spring planting and led to increased erosion. Despite evidence for an almost total lack of rainfall between 1276 and 1299 it is likely that all of these factors may have been of some importance. At the same time many Mogollon settlements in New Mexico were also abandoned. Throughout the southwest, the Indians congregated in large villages, or groups of related villages to a large extent both geographically and culturally isolated. Many of the contemporary southwestern Indian tribes trace their origin to this period. In the Zuni area, between 700 A.D. and 1100 A . D . , the inhabitants were dispersed in many small settlements often containing only a single family. During the tweIfth POPULATION STUDIES ON THE ZUNI INDIANS century several large pueblos were built and many small settlements were abandoned. A large pueblo at Atsinna, dating from 1100, had more than 500 rooms. All of the sites from this period demonstrate a culture similar to that at Chaco Canyon, called San Juan Anasazi. However, sites dating after 1300 show a rather marked change to a culture very much like that of the Mogollon to the west and south. Sites whose period of use spanned this period show both Pueblo and Mogollon features. For example, at Atsinna, the last structure in the Anasazi style was probably constructed about 1300; rooms dating from a later period are in the Mogollon style. It is likely that the original inhabitants were absorbed by northern groups of the Mogollon, and not merely displaced, because Zuni language and mythology appears to be a fusion of two distinct cultural traditions. Other groups may also have contributed to the population in the area. For example, there is good evidence for extensive Zuni-Hohokam trade (Bandelier, 1892). During the general Pueblo withdrawal, migrants from the north probably settled at Zuni where water was available throughout the period, and, successful farming may have been possible. Thus, by 1400, the Zuni appear to have been a rather heterogeneous mixture of southwestern peoples. However, the majority were probably of Mogollon descent and even in the late nineteenth century, the Zuni made ceremonial visits to shrines in the west and southwest, territory formerly occupied by the Mogollon. Between 1300 and 1400, many additional Zuni settlements were abandoned and the Zuni withdrew into six large villages well isolated from their nearest neighbors at Acoma. The Acoma Indians congregated on a single mesa some 75 miles from the nearest Zuni village: prior to the fourteenth century, Zuni and Acoma settlements had been only 25 miles apart. Although there was probably some contact and mate exchange with Acoma and with the Hopi to the west, the Zuni appear to have remained relatively isolated until Spanish contact in 1540. The distribution of Indian groups in the southwest just prior to that time is shown in figure 1. The archaeological observations summarized here are based primarily on the work of Reed (‘49, ’50, ’55), Wendorf and 121 Reed (’5.3, Woodbury (‘56) and Rinaldo (‘64). Spanish contact to the present: In 1540, the Spanish explorer, Coronado, accompanied by some 30 soldiers and a number of Nahuatl speaking Indians from Jalesco and Sonora in Mexico, sought the legendary “Seven Cities of Cibola.” “Shibola” was, in fact, Zuni, and Coronado, guided by a Zuni residing with the Pima, found six villages of adobe instead of seven cites of gold. The Zuni were partially subdued and when Coronado left, there were at least three Indians from Mexico known to have married into the tribe (Bandelier, 1892). This contact provided the first opportunity for gene flow with Spaniards; other contact with Mexican Indians, however, might have come earlier. Throughout the following 200 years, the Zuni resisted Spanish secular and religious influence. The first Catholic church was established at Zuni in 1628 but priests were killed in 1630, 1670, and at the time of the general Pueblo revolt against Spanish domination, in 1680. Again, in 1703, the Spanish resident policeman was killed. Each time major conflict with the Spanish appeared likely, the Zuni moved to the top of their sacred mesa which rises 1,000 feet above the surrounding plain. The mesa has water and grass and could be easily defended by very few men since only one narrow, steep trail permitted access to the top. From 1680-1692 and in 1703 the mesa became the home of the Zuni, and overall, the tribe spent several years in protected isolation. During this period, the smaller villages were effectively abandoned and the majority of the Zuni lived in a central pueblo, Zuni. These experiences enabled the Zuni to retain cultural, and, probably, biologcal integrity to a far greater degree than most other southwestern Indians. With the exception of the Hopi to the west, the Zuni is the only Pueblo group in which Spanish is not spoken and in which Spanish influences are noticeably absent. The retention of so much of their own culture encouraged numerous ethnological studies including those by Cushing, 1896; Stevenson, ’04; Kroeber, ’19; Bunzel, ’32; Benedict, ’35; and Eggan, ’50. Today, the Zuni pueblo has been relatively modernized. On reservation employment includes farming and jewelry manu- 122 P. L. WORKMAN, J. D . NISWANDER, K . S. BROWN AND W . C. LEYSHON Fig. 1 Culture areas of the southwest 1500 A.D. facture and the first phase of an industrial complex providing more technical jobs has been completed. However, the desire to retain their cultural identity persists most strongly. Adair and Vogt (‘49) in a study of returning veterans noted that “As the village has become more modern in its technology, antagonism to white culture has become increasingly apparent, especially to those values which threaten Zuni religion.” On the whole, ethnohistorical observations suggest that the Zuni are likely to be biologically quite similar to the heterogeneous ancestral population which appears to have been established during the fourteenth century. There is no evidence for significant intermixture with non-Indians and intermarriage with other Pueblo Indians has never been frequent. DEMOGRAPHY In order to use contemporary gene frequencies for inferences about relations among southwestern tribes, it is also necessary to consider the possibility of fluctuations in population size, which might have resulted in substantial genetic drift. Estimates of the population size at Zuni prior to Spanish contact vary between 3,000 and 6,000 (Cushing, ’20). For example, according to Bandelier (1892), at the time of contact there were six villages with a total of about 3,000 inhabitants. The largest village, Hawikah, had between 200 and 300 families; the smallest had 30 to 60 families. As shown in table 1 and figure 2, the population appears to have declined slowly until it reached a low point of 1,200 in 1864. Since that time there has been a steady rise, which has accelerated rapidly in the past two decades. The decline and slow return to precontact numbers must have involved considerable loss due to epidemics. As recently as the winter of 1898-1899 about 250 died from smallpox (HrdliEka, ’08). However, major improvements in medical facilities, 123 POPULATION STUDIES ON THE ZUNI INDIANS TABLE 1 TABLE 2 Estimates of Zuni population size f r o m 1540 to 1968 Age and sex distribution of Zuni born prior to Year 1540 1680 1790 1809 1850 1864 1879 1890 Size Year Size 1900 1910 1920 1930 1940 1950 1963 1968 3000 2500 1935 1598 1500 1200 1500 1547 J a n . 1,1969 Age 1523 1640 1813 1952 2205 2922 5011 6007 Source of data: 1540 (Bandelier, 1892); 1680 (Cushing, ’20);1790-1840 (AnnualReportof thecommissioner of Indian Affairs, 1867); 1850-1940 (Annual Reports of the Commissioner of Indian Affairs); 1950 (Bureau of Indian Affairs); 1963 (Public Health Service); 1968 (present study). Male Female Total 70 of Total 1 4 5 9 10-19 20-29 30-39 40-49 50-59 60439 70 364 503 746 438 300 239 187 140 137 321 488 778 497 345 184 159 88 93 685 991 1524 935 645 423 346 228 230 11.4 16.5 25.4 15.6 10.7 7.0 5.8 3.8 3.8 Total 3054 2953 6007 100.0 + ical of the population for many years. The 1921 and 1941 estimates included males and high fertility without high child mor- and females in ratios 1,040/823 and 1,245/ tality, have resulted in a current growth 1,007, respectively. In an early medical rate of over 5% per year. study, Fleming (‘24) was “surprised to find The age and sex distribution of the pop- many men of approximately 70 years of age ulation, based on the Zuni family register leading lives of extreme physical activity.” as of spring 1968, is given in table 2. A Studies on the distribution of hyperglycesmall proportion of the 6,007 Zuni may mia, to be published elsewhere, suggest actually be living off the reservation. Most that the life style of the Zuni, which inindividuals live in the main village at Zuni, cludes considerable sex differences in activbut there are also small farming settle- ity and diet, may be responsible for the ments at Ojo Caliente, Tekapo, Pescado, observed two to three times greater inciand Nutria. As expected in a rapidly in- dence of hyperglycemia in Zuni women. creasing population, the population is, on Women over 60 have a much lower average the average, quite young. Over 53% are blood glucose level than women aged 35less than 20 years old. The data of table 2 60, and it is possible that a greater mortalshow an atypical sex ratio pattern. In the ity of females related to hyperglycemia, early years, first males (1-9 years) and obesity, and hypertension may account, in then females (10-35) predominate. How- part, for the predominance of males among ever, more adults over 35 are male and the the older Zuni. Analyses of the medical obdifference is quite large for those over 60. servations on the Zuni will be presented in This distribution appears to have been typ- detail elsewhere. 6000 - ferential fertility in populations of this size may result in rather marked changes in gene frequencies. HrdliEka (‘08) noted that few women were sterile and that most families had several children. Thus any effect of differential fertility would be more likely due to variation in the reproductive rates of males. 5000 W N V, 4000- z 5 2 a x 30002000- 1000 I I I I I GENETIC ANALYSES 124 P. L. WORKMAN, J. D. NISWANDER, K. S. BROWN AND W. C. LEYSHON tensive physical examinations including detailed dental and ophthalmological examinations were done on over 450 Zuni over 20 years of age. Urine samples were obTABLE 3 Distribution of Zuni “full bloods” i n the sample population b y nge and sex Age % sample of Zuni in agegroup Male Female Total Total Zuni 13-19 20-29 30-39 4049 50-59 60-69 a70 72 50 26 13 18 18 22 137 130 49 39 39 22 27 209 180 75 52 57 40 49 997 935 645 423 346 228 230 21.0 19.3 11.6 12.3 16.5 17.5 21.3 Total 219 443 662 3804 17.4 tained for glucose and albumin screening. Briefer examinations of about 200 schoolchildren were conducted. This provided observations on 662 Zunis for genetic analysis and for comparative studies of adaptation and health status in southwestern Indian tribes. Blood was drawn for serological analyses, two hours after ingestion of a 75 gm glucose equivalent carbohydrate load. Blood samples were divided into three aliquots; one with ACD as an anticoagulant, a second containing fluoride as an enzyme inhibitor and a third clotted sample was obtained without additives for enzyme studies. Blood glucose values were determined by the Southwestern Field Studies Section of the National Institute of Arthritis and Metabolic ~ i ~ ~ TABLE 4 Distribution of Zuni phenotypes Phenotype No. quency ABO 0 A1 B 576 15 71 0.870 0.023 0.107 MNSs MMSS MMSs MMss MNSS MNSs MNss NNSS 30 156 199 4 77 154 2 4 36 0.045 0.236 0.301 0.006 0.116 0.233 0.003 0.006 0.054 CcDEe CCDee ccDEE CcDee ccDEe CCDEE CCDEe CcDEE ccdEe 198 329 39 30 11 1 48 5 0.299 0.497 0.059 0.045 0.017 0.001 0.073 0.008 1 0.001 PI ( + ) Pi(-) 533 129 0.805 0.195 Duffy Fya+ Fya - 618 44 0.934 0.066 Kell kk kkP 656 6 0.991 0.009 Lewis Lea- 662 1.000 Diego DibDib DiaDib DiaDia 631 31 0 0.953 0.047 NNSs NK S b Rh P Fre Fre System 0.000 System Phenotype + No. quency Kidd Jka Jka - 336 326 0.508 0.492 Cerumen W-(wet) ww (dry) 160 298 0.349 0.651 Haptoglobin 1-1 1-2 2-2 279 310 68 0.425 0.472 0.103 Group specific 1-1 1-2 2-2 485 157 13 0.740 0.240 0.020 Red cell Acid phosphatase PAPA PBPB 32 210 418 0.048 0.318 0.633 Transferrin TfCTf‘ TfCTf 6 O - l TfBO-1 530 121 8 0.809 0.185 0.006 Albumin AL~AL A 639 A L ~ A L ~ ~ 16 ~ 0.976 0.024 Phosphoglucomutase: PGMl 1-1 1-2 2-2 434 203 025 0.655 0.307 0.038 PAPB ~ 125 POPULATION STUDIES ON THE ZUNI INDIANS TABLE 5 Zuni gene frequencies System ABO Allele Frequency 0 At B 0.933 0.011 0.055 MS Ms NS Ns 0.213 0.546 0.020 0.221 CDe cDE cDe CDE 0.704 0.219 0.033 0.043 System Rh 0.001 cde Frequency Jk " 0.296 0.704 W W 0.193 0.807 Haptoglobin HPI HP~ 0.661 0.339 Group specific Gcl Gc~ 0.860 0.140 Red cell Acid phosphatase PA PB 0.208 0.792 Tr an sferr i n TfC TfBO-1 0.902 0.099 Albumin A L ~ ALM@X 0.987 0.012 PGMI' PGMiZ 0.809 0.191 Kidd jkb Cerumen MNSs Allele , P p1 PZ 0.559 0.441 Duffy FYa Fyb 0.742 0.258 Kell k kP 0.995 0.005 Lewis Le Lea 1.000 0.000 Diego Dia Dib 0.023 0.977 Red cells were typed for A, Al, B, M, N , S, s, U, P, C, C " , E, c, e, K, K p n , K P b , k, Lea, Fy", Jk", Di" and Dib . Haptoglobin albumin, transferrin and group specific serum protein types were determined by polyacrilamide gel electrophoresis according to methods previously described (Brown and Johnson, '70). Red cell acid phosphatase and phosphoglucomutase genotypes were determined by the methods of Karp and Sutton ('67) and Spencer, Hopkinson and Harris ('64) respectively. The recent ancestry of the Zunis in the sample was determined and 95% could be putatively described as full-blooded. The analysis in the present paper is restricted to the sample of full-blooded Zuni as described in table 3. Data on the Gc, Hp, AL and Tf systems has been reported by Brown and Johnson ('70) but, for completeness, we present all of the phenotypes and gene frequencies in tables 4 and 5, respectively. Ten codominant systems permit estimation of Fi = 1 - Hi/2piqi where Hi is the observed proportion of heterozygotes and pi, qi = 1 - pi are the frequencies for the ith allele. A s pointed out by Harpending, Workman and Grove ('73), if a population is not highly endogamous, migration effects alone will cause the expected value of this estimate to be negative. As shown in table 6 , the mean of the Fi is 0.0019 which provides support for the asTABLE 6 Estimates of F and corresponding chi-square tests for codominant loci System or locus MN ss cc Ee Ac.P. Di PGM HP Gc Tf mean (Fi) Fi x2 = F2N 0.0295 - 0.0005 0.0689 - 0.0057 0.0328 - 0.0240 0.576 0.0002 3.141 0.022 0.711 0.380 0.043 1.791 0.005 1.071 0.0081 - 0.0522 0.0028 - 0.0404 0.0019 126 P. L. WORKMAN, J. D. NISWANDER, K. S. BROWN AND W. C. LEYSHON sumption that the Zunis are relatively isolated. Chi-square tests for each locus, comparing observed frequencies with these predicted by the Hardy-Weinberg Law, are given by x ( ~ )=~ F2N and, as also shown in table 6, none of the chi-squares is significant. Genetic stratification of the Zuni population might on a priori grounds, be either spatial or temporal. The Zuni pueblo can be divided into five areas corresponding to locations separated by the Zuni River and by the major roads through the pueblo. These subdivisions are actually recognized by the Zuni for administrative aspects of their health care service. Since the Zuni are matrilineal, live in extended family units, and are subdivided into clans, there might be opportunities for local genetic differentiation. However, Kroeber ('19) found no association between clan and residence location and it appears that new households are established without regard to the location of other extended family members. Genetic analyses by contingency chi-square tests showed no significant differences among the five locations and it seems that spatial differentiation can be assumed to be absent. Because of the recent rapid increase in population size, the possibility of genetic changes due to drift or differential fertility should be considered. The Zuni sample was sufficiently representative of the different age groups that one can compare gene frequencies in the youngest group (age 13-19; N = 208) and the oldest group (age 3 60; N = 89. Table 7 shows such a comparison for several of the genes studied. No large differences can be seen, and none were significantly different, so the possibility of age stratification can be rejected. Thus, the pooled Zuni frequencies can be assumed to be representative of the contemporary genome, and, indeed, to be a good approximation to the frequencies at least in the past century. The most striking observation in tables 4 and 5 is the high frequency of type B in the ABO system. Evidence for non-Indian intermixture in southwestern Indians usually is taken from the A1 and r (cde) genes. Here, however, the frequencies of A1 and r are 0.011 and 0.001 which suggests very little non-Indian gene flow. On the other hand, B, assumed absent in American Indi- TABLE 7 Genefrequencies i n the y o u n g e s t (1 3-1 9 ; N =208) and oldest ( 260; N = 289) individuals i n t h e Zuni sample Gene 13-19 60+ Gene 13-19 60+ P1 Fya 0.58 0.58 0.74 Hpl Gc1 R1 (CDe) R2 (cDE) RO(cDe) Rz (CDE) 0.65 0.87 0.67 0.81 0.70 0.24 0.02 0.04 Jka Dib B PA 0.75 0.26 0.02 0.06 0.20 0.35 0.04 0.05 0.21 0.66 0.24 0.05 0.05 ans, has a frequency of 0.055 overall and a frequency of 0.05 in individuals over 60 years old. If intermixture were the source of the B gene, it is not likely to be due primarily to recent intermixture. In other southwestern tribes, excluding Athapascans, only one report of a similar high B frequency has been reported. Allen and Schaeffer ('35) found B = 0.054 in Tewa Indians at the Taos, New Mexico Pueblo; however, in that group, A, had a frequency of 0.192 and the evidence for intermixture is clear. In order to test for stratification with respect to the B gene the frequencies of genes at other loci were determined for individuals with type 0 and type B. The more recent the intermixture, the more likely type B individuals would have gene frequencies closer to Caucasian values, assuming B to be derived from Caucasian gene flow. No differences were found among B and 0 individuals and we must conclude that if B came in by gene flow, it is of quite distant origin. The high frequency of B alone among the non-Indian alleles suggests the possibility that very few non-Indians (B, Rh positive), and perhaps only a single individual, may have introduced the B allele into the population. The demographic history of the Zuni would easily permit the subsequent differential fertility and/or drift required to elevate the B frequency to its current level. Alternatively, postulating the presence of B in the original founder population, while not impossible, has no support in any other study of southwestern or Mexican Indian tribes. GENETIC DIFFERENTIATION AMONG SO U T H W E ST E R N I N D I A N TRIBES The southwest contains by far the greatest number of American Indians within the continental United States, including both 127 POPULATION STUDIES ON THE ZUNI INDIANS Athapascans (Navajo and Apache) who migrated into the area as early perhaps as the twelfth century and indigenous tribes whose ancestors may have been in the region for 5,000 to 10,000 years. The current reservation population sizes are given in table 8. Most of the genetic observations on these tribes were summarized by Matson, Burch, Polesky, Swanson, Sutton and Robinson ('68). There are very few observations on the Athapascans and the Yumans, and even fewer on the Pueblos. Since we are here concerned with biological differentiation among the indigenous groups, there appear to be adequate genetic data only for the Maricopa (Matson et al., '68), Pima (Matson et al., '68). Papago (Niswander, Brown, Iba, Leyshon and Workman, '70) and the Zuni. Consequently, the analysis to be presented is to be viewed only as a preliminary study of genetic differentiation in the southwest. The gene frequencies reported for these tribes are given in table 9. Gene frequencies in all four tribes are available only for the ABO, MNSs, Rh, P, Fy and Jk systems. Using these loci, an analysis of genetic variation among the tribes can be carried out as follows. For k alleles over s populations we form the ( s x k) matrix V with elements (PSk - Pk) &k (1 - ck), where Pk is the weighted mean gene frequency and Psk is the frequency of the kth allele in the Sth population. Then R = (l/k) VV' is an ( s X s) matrix whose diagonal elements, rii , describe what has been termed local kinship (Morton, Yee, Harris and Lew, 'i'l), that is, the deviation of the ith population from the mean gene frequencies; larger values of rii indicate greater relative isolation. The weighted sum, S w ir ii, an estimate of mean kinship within the total population, is an index of the total genetic heterogeneity among populations, where w denotes the subpopulation sample size relative to the total sample. A correction for sampling error, - (1 - h)/2N , where h is the fraction of the census population sampled, and N T is the total over all samples, must be added to Hw ir ii. Here we use the hypergeometric correction, assuming h = 0.10, to permit inference on the contemporary gene pool (Harpending and Jenkins, '73; Workman, Harpending, Lalouel, Lynch, Niswander and Singleton, '73) rather than binomial sampling to draw inferences on the parental gene pool as done by Morton et al. ('71) for kinship bioassay. The corrected mean kinship, called here R St (rather than the usual notation, F st, in order to avoid confusion with F taken as a probability), can also be obtained from the average value of the scaled gene frequency variances so that , where p i and up' are the appropriately weighted mean and variance of the ith allele. The matrix R with corrected diagonal elements is given in table 10. The Zuni and Maricopa have much higher values of r i i indicating greater genetic isolation. The off diagonal elements of R, r indicate the TABLE 8 Estimated Indian population on or adjacent to Reservations i n Arizona and New Mexico. (Bureau of Indian Affairs estimates.for Sept., 1968) Tribe or group Estimated population Pima Papago 10,240 6,210 Yuman 5,180 Apache san Carlos Fort Apache Jicarilla Mescalero Navaho 4,650 6,000 1,500 1,690 13,840 1 19,500 Tribe or group Estimated population Zuni 5,040 Hopi 6,080 Acoma Laguna E. Keresan Tanoan Towa Tewa N. Tiwa S. Tiwa Total Pueblo 1,920 2,880 2,400 1,380 2,400 1,270 2,210 25,580 128 P. L. WORKMAN, J. D. NISWANDER, K. S. BROWN AND W. C. LEYSHON TABLE 9 G e n e f r e q u e n c i e s in t h e Z u n i , P i m a , Papago a n d Maricopa I n d i a n s Gene Zuni + Az Pima 1 Maricopa Pap ago 3r4 0.011 0.055 0.933 0.041 0.000 0.959 0.063 0.003 0.934 0.000 0.000 1.000 MS MS NS NS 0.213 0.546 0.020 0.221 0.274 0.427 0.091 0.208 0.346 0.444 0.080 0.130 0.294 0.347 0.049 0.310 RO (cDe) Rl (CDe) RZ (cDE) R z (CDE) r (cde) 0.033 0.704 0.219 0.043 0.001 0.039 0.581 0.361 0.019 0.000 0.022 0.626 0.330 0.022 0.000 0.030 PI 0.559 0.742 0.296 0.023 0.661 0.860 0.538 0.709 0.359 0.029 0.543 0.533 0.762 0.476 - - 0.494 0.820 0.360 0.028 0.452 0.832 0.099 0.031 0.037 - 0.020 0.005 0.227 - AI B 0 ;;: Dia Hpl Gc1 ~fB0-l A W L ~ ~ ~0.012 0.193 0.208 0.809 PA PGM1' 5 0.607 0.317 0.046 0.000 - Present study; N = 662. Matson et al. ( ' 6 8 ) ; N = 909. Niswander et aL ('70); N = 709. 4 Brown and Johnson ('70). 5 Matson et al. ( ' 6 8 ) ; N = 124. 1 2 3 TABLE 10 R m a t r i x f o r f o u r s o u t h w e s t e r n I n d i a n p o p u l a t i o n s a b o v e the diagonal; Sanghvi's G2 b e l o w t h e diagonal (1 1 I (1) Zuni 0.0205 (2) (4 ) (3) - 0.0080 - 0.0087 (2) P i m a 0.0083 (3) Papago 0.0109 0.0044 0.0080 (4) Maricopa 0.0119 0.0055 0.0092 Mean kinship = - 0.0048 0.001 0.0034 - 0.0042 I 0.0208 0.0109. relationships between pairs of populations. A genetic distance between populations i and j could be taken as the squared Euclidean distance d $ = r i i rjj - 2 rij . However, interdependencies among genes suggest a more appropriate genetic distance would be given by G2 (Sanghvi, '53) or B2 (Balakrishnan and Sanghvi, '68). The corrected values of R st per gene are given in table 11. Only the B allele gives values which appear to be much larger than the mean. The high value of B at Zuni + TABLE 11 E s t i m a t e s of R,t p e r g e n e , corrected for s a m p l i n g error Gene Rst Gene Rst 0.0122 0.0359 0.0051 0.0119 0.0121 0.0137 0.0132 R1 (cDe) R1 (CDe) R2 (cDE) R z (CDE) 0.0008 0.0097 0.0151 0.0037 0.0018 0.0105 0.0066 ~ A1 B 0 MS MS NS Ns Mean kinship ( E s t ) = P Fy" Jk a 0.0109. ~~ POPULATION STUDIES ON THE ZUNI INDIANS alone is also indicative of nonrandom differentiation for that allele. The low value for RO (cDe) indicating very little differentiation may be due to chance or, perhaps, to stabilizing pressures on this allele. No interpretation of this observation can be made with the present data. The Indian populations in the southwest are located in discrete sites at which year round water is available. Intermixture among these tribes might be related to the distance separating them as already demonstrated for subdivisions within the Papago by Workman and Niswander ('70) and Workman et al. ('73). The effect of isolation by distance can be studied given the geographical distance between tribes and the R matrix. Using the longitude and latitude of the main population centers of the reservations (Sells and Sacaton for the Papago and Pima) as coordinates, the geographical locations can be plotted directly in two-dimensions. A two-dimensional reduction of the R matrix can also be ob- 129 tained using the first two eigenvectors of the centroid adjusted R matrix following Lalouel('73). Although some information is lost by the procedure, the first two eigenvectors maximize the dispersion of points in two-dimensions. The proportion of the total sum of squares retrieved by this reduction can be measured by f (A) = (A, Az) I ZA i, where h are the eigenvalues of the adjusted matrix. For the matrix of table 8, f (A) = 0.914; that is, almost all information is in the two-dimensional reduction. The two-dimensional reduction can then be rotated to best fit with the geographical coordinates as suggested by Lalouel ('73). A measure of the goodness of fit of the rotated projection of the R matrix to geographic coordinate is given by Carrol's measure of disagreement, C , = 0.30 or, equivalently by the correlation between the coordinates of the configurations, 0.85 (see Lalouel, '73, for details). The resulting projection is shown in figure 3. + goographic location gonotic prodiction Fig. 3 Two-dimensional reduction of the R matrix rotated to best fit with the geographical locations of four southwestern Indian tribes (shaded areas labelled by tribe, delimit present reservation boundaries). 130 P. L. WORKMAN, J. D. NISWANDER, K. S. BROWN AND W. C. LEYSHON MI 0 4 Fig. 4 Two-dimensional reduction, plotting alleles in four southwestern Indian tribes using first two eigenvectors of the S matrix. The positions of the Zuni, Pima and Papago based on the information on the R matrix appear very close to their present geographical locations which their ancestors inhabited for many centuries. The Maricopa, today located just south of the Pima reservation, were moved to that place about 1890 from the upper Colorado River, much to the northwest. The genetic prediction places them much closer toward their ancestral location as should be expected if geographical location is an important aspect of genetic differentiation. As also seen in table 9, the largest genetic distance is between the Zuni and Maricopa as would be expected given their geographical distance in earlier years. An alternative representation using all the genetic markers given in table 9 was made using Sanghvi’s G2, weighting the contribution from each k-allelic locus by (k - 1). That is, G2.. = L [ Z G Z / ( k - l ) ] U L where L is the number of loci.2 The G values so obtained are also given in table 10. The fit by G2 did not differ substantially from that by R and need not be presented here. The close agreement of geographical location and the location predicted by the genetic model indicates that distance has been a major factor in determining the amount of gene flow among tribes. Another reduction of the genetic data which is also useful, plots the gene frequencies as points on axes directly comparable to the plot of the populations. The genetic variation can be partitioned either This differs slightly from formulation used for these data by Sofaer et al. (Am. J. Phys. Anthrop., 37: 357366, 1962), in which each allele is weighted equally. Although the results appear trivially different we prefer the weighted formula because it gives slightly higher correlation of thedistances in two dimensions with those in the original n-space. POPULATION STUDIES ON THE ZUNI INDIANS by populations or by alleles, as was shown, for example, in formula (1) for the derivation of R S t . That is, instead of obtaining a dimensional reduction of R = l / k VV', we consider S = (l/s)V'V which is a (k X k) matrix; describing variation and covariation among allelic frequencies. Since both R and S have the same set of nonzero eigenvalues (see discussion by Harpending and Jenkins, '73) the dimensions of any reduced space are the same whether we plot populations or alleles. The two-dimensional plot of the genes used in this analysis is given in figure 4. By comparison of the location of genes and populations in figures 3 and 4 we see that Zuni is typified by its frequencies of B, R* and Ms; Maricopa by the frequencies of J k a , 0 and Ns, and the Papago and Pima are characterized by their R 2 , A, NS, and MS frequencies (see also the data in table 9). Note that RO (cDe) which differentiates none of the populations ( R s , = 0.0008) plots in the middle of the array. These two-dimensional representations appear to provide the simplest and most easily interpreted reduction of the data. They do not use all of the information but if, as in the present case, geographical considerations are of major importance, then a two-dimensional representation will contain most of the information. The analysis presented here indicates that the Zuni have been relatively isolated within the region, despite a general agreement of gene frequencies at most loci including the presence of the Albumin Mexico variant in all the tribes. Differences such as presented by the B allele, indicate for the Zuni, a generally low degree of gene flow with the tribes included in this comparison at least since the time of contact with the Spanish. The Zuni also have a high frequency of albinism (1:240) which has changed very little over the past 90 years (Witkop Niswander, Bergsma, Workman and White, '72), and only the neighboring Hopi have a similar frequency of albinism indicating, perhaps, a degree of Zuni-Hopi intermarriage. Since Brown, Hanna, Dahlberg and Strandskov ('58), found no evidence for B in a small sample of Hopi, Tewa, the exchange between these tribes may also pre-date the introduction of B into the Zuni. When genetic studies of the other Pueblos are available it is hoped 131 that some of the questions raised, but not answered, in this present study can be resolved. ACKNOWLEDGMENTS The authors are indebted to the Zuni people, their tribal leaders and the staff of the Public Health Service Indian Hospital at Zuni, particularly Dr. Errett E. Hummel, Jr., Service Unit Director. We also wish to acknowledge the technical support of Mr. L. L. Taylor. This work was supported in part by grant lROlHG06003 from the United States National Institutes of Health while the senior author was at the Population Genetics Laboratory, University of Hawaii, Honolulu. LITERATURE CITED Adair, J., and E. Vogt 1949 Navaho and Zuni veterans: A study of contrasting modes of culture change. Am. Anthrop., 51 : 547-561. Allen, F. W., and W. Schaeffer 1935 The distribution of the human blood groups among the Navajo and Pueblo Indians of the Southwest. University New Mexico Bull., Biology Series, 4 : 2. Balakrishnan, V., and L. D. Sanghvi 1968 Distance between populations on the basis of attribute data. Biometrics, 24: 859-865. Bandelier, A. F. 1892 An outline of the documentary history of the Zuni tribe. J. Am. Ethnology and Archaeology, 3: 1-1 15. Benedict, R. F. 1935 Zuni Mythology. Vol. 2. Columbia University Press, New York. Brown, K. S., B. L. Hanna, A. A. Dahlberg and H. H. Strandskov 1958 The distribution of blood group alleles among Indians of Southwest North America. Am. J. Hum. Genet., 10: 175195. Brown, K. S., and R. S. Johnson 1970 Population studies on Southwestern Indian tribes. 111. Serum protein variations of Zuni and Papago Indians. Human Heredity, 20: 281-286. Bunzel, R. L. 1932 Introduction to ZuAi ceremonialism. In: 47th Annual Report of the Bureau of American Ethnology, Government Printing Office, Washington, D.C., pp. 467-544. Cushing, F. H. 1896 Outlines of ZuAi creation myths. In: 13th Annual Report of the Bureau of American Ethnology, Government Printing Office, Washington, D.C., pp. 321-447. 1920 ZuAi Breadstuff. Indian Notes and Monographs. Vol. VIII. Museum of the American Indian. Heye Foundation, New York. Eggan, F. R. 1950 Social Organization of the Western Pueblos. University of Chicago Press, Chicago. Fleming, H. C. 1924 Medical Observations on the ZuAi Indians. Contribution from the Museum of the American Indian. Vol. VII. Heye Founation, New York. Harpending, H., and T. Jenkins 1973 Genetic 132 P. L. WORKMAN, J. D. NISWANDER, K. S. BROWN AND W. C. LEYSHON distances among Southern African populations. In: Methods and Theories of Anthropological Genetics. M. H. Crawford and P. L. Workman, eds. University of New Mexico Press, Albuquerque. Harpending, H., P. L. Workman and J. Grove 1973 Local genotypic disequilibrium in a generalized island model. Hum. Biol., 45: 359-362. HrdliEka, A. 1908 Physiological and medical observations among the Indians of Southwestern United States and Northern Mexico. Bureau of American Ethnology, Bull. 34, Smithsonian Institution, Washington, D.C. Karp, G. W., and H. E. Sutton 1967 Some new phenotypes of human red cell acid phosphatase. Am. J. Hum. Genet., 19: 54-62. Kroeber, A. L. 1919 Zuiii kin and clan. Anthropological Papers of the American Museum of Natural History, New York, 18: 39-204. Lalouel, J. M. 1973 Topology of population structure. In: Genetic Structure of Populations. N. E. Morton, ed. University of Hawaii Press. Matson, G. A,, T. A. Burch, H. F. Polesky, J. Swanson, H. E. Sutton and A. Robinson 1968 Distribution of hereditary factors in the blood of Indians of the Gila River, Arizona. Am. J. Phys. Anthrop., 29: 311-338. Morton, N. D., S. Yee, D. E. Harris and R. Lew 1971 Bioassay of kinship. Theoretical Population Biol., 2: 507-524. Newman, S. 1964 Comparison of Zuiii and Californian Penutian. Int. 3 . Am. Linguistics, 30: 1-13. Niswander, J. D., K. S. Brown, B. Y. Iba, W. C. Leyshon and P. L. Workman 1970 Population studies on Southwestern Indian tribes. I. History, culture, and genetics of the Papago. Am. J. Hum. Genet., 22: 7-23. Reed, E. K. 1949 Sources of upper Rio Grande Pueblo culture and population. El Palacio, 56: 163-1 84. 1950 Eastern Central Arizona archaeology in relation to the western Pueblos. Southwestern J. of Anthrop., 6: 120-138. 1955 Painted pottery and Zuiii history. Southwestern J. Anthrop., I I : 178-193. Rinaldo, J. B. 1964 Notes on the origins of historic Zuiii culture. Kiva, 29: 86-98. Sanghvi, L. D. 1953 Comparison of genetical and morphological methods for a study of biological differences. Am. J. Phys. Anthrop., 11: 385404. Spender, N., D. A. Hopkinson and H. Harris 1964 Phosphoglucomutase polymorphism in man. Nature, 204: 742-745. Stevenson, M. C. 1904 The Zuiii Indians. Bureau of American Ethnology Annual Report No. 23, Government Printing Office, Washington, D.C. Swadesh, M. 1959 Linguistics as an instrument of prehistory. Southwestern J. Anthrop., 15: 20-35. 1967 Lexicostatistic classification. In: Handbook of Middle American Indians. Vol. 5, Chapter 4. R. Wauchope, ed. University of Texas Press, Austin, pp. 79-115. Wendorf, F., and E. K. Reed 1955 An alternative reconstruction of Northern Rio Grande prehistory. El Palacio, 62: 131-173. Woodbury, R. B. 1956 The antecedents of Zuni culture. Trans. N.Y. Acad. Sci., Series 2, 1 8 : 557-563. Witkop, C. J., J. D. Niswander, D. R. Bergsma, P. L. Workman and J. G. White 1972 Tyrosinase positive oculocutaneous albinism among the Zuni and the Brandywine triracial isolate: Biochemical and clinical characteristics and fertility. Am. J. Phys. Anthrop., 36: 397-406. Workman, P. L., and J. D. Niswander 1970 Population studies on Southwestern Indian tribes. 11. Local genetic differentiation in the Papago. Am. J. Hum. Genet., 22: 2 4 4 9 . 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. In: Genetic Structure of Populations. N. E. Morton, ed. University of Hawaii Press, Honolulu.