Comparison of genetic and anthropological interpretations of population isolates in Aguacatenango Chiapas Mexico.код для вставкиСкачать
Comparison of Genetic and Anthropological Interpretations of Population Isolates in Aguacatenango, Chiapas, Mexico ROBERT P. ERICKSON,I SARA NERLOVE,2 WILLIAM AND A. KIMBALL ROMNEY3 Departments of Anthropology and Medicine, Stanford University, Stanford, California 94305 P. CREGER ABSTRACT Intensive demographic and genetic studies were made of Aguacatenango, a partial isolate population in Chiapas, Mexico. The census showed higher rates of birth than mortality, with a stable population maintained by emigration. Complete enumeration showed many discrepancies between the ideal and actual mating patterns. Though ideally a n endogamous town, many marriages were to outsiders. Also many occurred between the two allegedly endogamous barrios of the town. Blood group determinations on a third of the population provided a second description of the isolate. This description suggested that although the population is nearly a homogeneous genetic unit, the elder residents of one bani0 show genetic differences from those of the other. That the barrios might have indeed been genetic isolates was further suggested by a better fit to Hardy-Weinberg equilibria for each barrio separately than for the isolate as a whole. These blood group data were compared to blood group data on one other isolate in the region revealing marked gene frequency differences. It is suggested that similar studies are needed o n many populations if the effects of particular population structures on gene drift and selection are to be elucidated. Interpretation of paputation isolates There has been increasing interest in the simultaneous study of polymorphic physiological characters and demographic parameters in isolate populations (Neel et al., '64; McKusick et al., '64; Arends et al., '67). Assessment of the contribution of gene drift and natural selection to genotype frequency changes has been sought in studies of small, primitive communities which have high childhood death rates. Such communities with historical records of some depth have provided data on migration distances, consanguinity and the "founder" effect for tests of population genetic model (Cavalli-Sforza et al., '64; Roberts, '68). Cultures in different parts of the world have widely divergent customs in regard to the choice and number of mates, the stability of a particular mating, the sex and number of offspring desired, etc. The effects of these aspects of a population's mating structure (which are not fully expressed in the co-efficient of inbreeding and differential fertility of population) on changes in allele frequency are AM. J. PHYS.ANTHROP.,32: 105-120. being analyzed (Neel et al., '64; Gajdusek, '64; Yasuda and Morton, '67). The present study was made in Aguacatenango, Chiapas, Mexico, a small, primitive Tzeltal-speaking Mayan community with a high childhood death rate, for which there are some historical records available. One aim of the study was to examine this semi-isolate Mayan Amerindian population in order to provide further information on the variables involved in allele frequency changes. Another aim was to compare the allele frequencies of this semi-isolate to those of another nearby in order to contribute to the studies of divergence among related groups (Neel and Salzano, '64). In the course of the examination of allele frequencies, the focus of the study became the reconciliation of some discrepancies between the anthropological and genetic descriptions, both of which emphasize the importance of migration and thus of "open" population models. 1 Present address: National Institute of Arthritis and Metabolic Diseases, Bethesda, Maryland. 2 Present address: University of California, Irvine, School of Social Sciences. 3 See footnote 2. 105 106 ERICKSON, NERLOVE, CREGER AND ROMNEY Historical background of Tzotzil and Tzeltal-speaking groups At present Tzotzil and Tzeltal, two closely related Mayan languages, number about 160,000 speakers who are situated in the highlands of Chiapas. Archeological and linguistic evidence suggests that about 1200 years ago a Mayan group migrated from the mountainous heights that now form the northern rim of Guatemala to the Chiapas Highlands, crossing some 50 miles of low plateau that intervenes between the two highland areas. The archeological evidence indicates that as of about a millenium ago, the highland region became more densely inhabited and characterized by uniformity both in settlement pattern and ceramic inventory ( Adams and McQuown, ’59). These changes can be regarded as the consequence of an influx of population with cultural similarities (particularly in ceramics j from the zone of Classic Mayan civilization in the lowlands to the east. The immigrating Mayans spread over the highlands and differentiation of their language into “dialects” began. According to lexicostatistical indices (Swadesh, ’59), Tzotzil and Tzeltal separated about 1,200 years ago. The Tzotzil and Tzeltal Indians did not remain completely free of foreign influences after settling in the highlands. During the ninth and tenth centuries, a Nahuatl-speaking group from central Mexico penetrated into the area as far as the western edge of the Chiapas Highlands; and during the early sixteenth century, two of the northern Tzotzil groups were conquered by Montezuma I1 (Calnek. ’59). In 1524, Luis Marin began the conquest of the Tzotzil and Tzeltal Indians. His victories were facilitated by mutual animosity among subgroups (clan groups with ceremonial centers replete with ball courts and small pyramids) of Indians. The Zinacantecos helped him defeat the Chamulans and Huistecos (these 3 Tzotzil groups are still present in Highland Chiapas). Marin did not attempt to stay in the Chiapas Highlands, but he returned in 1528 to found Ciudad Real, now Ciudad de las Casas. Thus began the colonial period, the dominant characteristics of which seem to have been established dur- ing the fourth and fifth decades of the sixteenth century. Ethnohistorical evidence shows that Aguacatenango has had a continuous population almost since the time of the Spanish conquest. The first census report found for Aguacatenango is for the year 1611 AD. when the population was 720 individuals (Calnek, ’61). Later, a second town, Quesaltepec, was added to Aguacatenango during one of the directed population movements in the seventeenth century. Quesaltepec probably originated independently but first appears on the census roles as a parcialidade (land possession) of Aguacatenango. Aguacatenango: A Tzeltal community Aguacatenango, a community of Tzeltal speakers situated in the highlands of Chiapas, Mexico, constitutes the partial isolate of Mayan Amerindians of this study. In the summer of 1963, its Indian population was 1405 (fig. 1). It is a nuclear town whose residents live in an area of several acres in which there is space for a limited cultivation of fruits, vegetables and some maize, which is the staple crop. Most of the farming, however, is done in the surrounding country to which the men commute in order to cultivate their milpa or maize fields. A few families of Zadinos (people who consider themselves to be non-Indian and a part of Mexican national culture) live in Aguacatenango. They run small shops selling a variety of goods. One ladino is the “engineer” for the electric and water supplies and a power corn-grinder. Four of the Indian homes serve also as cantinas (bars) which have electric record players with public address systems over which the purchase of trago (native distilled fermented sugar cane juice) is encouraged. Most households have cold running water which comes through a faucet in the yard, a few have one electric light, and some have houses with tile roofs which are considered the height of achievement. There is a three-year primary school which teaches a moderate percentage of the children, mostly males, enough Spanish so that they can get along outside of the village, while a rare individual will learn the rudiments of reading and writing. The Isthmus '. -._. -'.. __ .'.. '... .. -._ 1 . 1 . . -.. \..' I Fig. 1 Location of Aguacatenango. Mexican National Indian Institute (I.N.I. ) has provided an Indian high school i n nearby Ciudad de las Casas, the highland commercial center. A few Indians from various villages are thus accommodated for further education and some of the graduates return to the villages as teachers or nurses. I.N.I. has also built a clinic in Aguacatenango which has such a nurse and is visited weekly by a doctor. Ethnographic data provide ideal descriptions of mating behavior in Aguacatenango 108 ERICKSON, NERLOVE, CREGER A N D ROMNEY which include lineage exogamy, and town and barrio endogamy. The kinship structure consists of a localized group of patrilineal kinsmen which constitute at least the nucleus of a lineage. This nucleus functions as a unit within which the individual seeks advice in time of indecision and protection in time of trouble (Metzger, ’56). These “loose-lineages’’ involve some three generations of patrilocal families. Four generations might be involved when a linking relative is recently deceased and still remembered. Members of a patrilineal lineage share the same family name of Spanish origin, e.g., Mendez, but one such name may be shared by several lineages in Aguacatenango. Exogamy within the patrilocal looselineage is required, but genealogies are seldom remembered past three generations so many third-cousin marriages can occur. The inhabitants of Aguacatenango are aware of a few exceptions to the ideal of patrilineal exogamy : there are known occurrences of marriages between patrilineal first cousins. Limiting the duration of a given loose-lineage are human memory and strong male sibling rivalry which tends to speed the separation of the lineage into several lineages. Termination of a lineage occurs for a lack of male offspring to continue it. Aguacatenango is divided into two barrios (territorial divisions) one on each side of the main central plaza in which the Catholic church is situated. The plaza forms part of the barrio boundary and the remainder of it runs between bordering household plots in an unmarked, but well recognized, manner (fig. 2). The origin of the division of Aguacatenango into two barrios is unknown although it may be a remnant of old clan organizations, or it may be the results of the incorporation of the parcialidade, Quesaltepec, into the town. It is marital, not political or religious interaction, which is limited between the two barrios. The normative descriptions of behavior indicate barrio endogamy although, again as with the ideal of lineage exogamy, exceptions are known. Following the Mexican Revohtion, the town acquired lands about four miles away on a lower plateau. Some families moved to be near their milpa on this plateau, forming two “colonies” there. These colonies are gradually developing their own rules of endogamy although they often immigrate back to Aguacatenango. The colonists still account for the major portion of immigration to Aguacatenango as well as emigration from it. One colony is named El Puerto; the other remains unnamed. SAiMPLE AND METHODS Anthropological field work in Aguacatenango began in 1956 and has been continued by several different anthropologists to the time of the present study. Demographic data for this study were collected with the aid of interpreters in the summer of 1963 during a period of seven weeks of intensive field work. The genealogical data were gathered during the course of a complete census of the 253 households then comprising the town of Aguacatenango. The simultaneous gatherings of genealogies and census and the availability of genealogical data from previous studies provided maximal opportunity for crosschecking these data. For each household, questions asked included women’s fertility histories (including number of still births), birth place of each member of the families of orientation and procreation, estimated date, and age at death of all deceased relatives, number and order of spouses, duration and outcome of marriages, and the present location of siblings or children who have left the village. Estimated ages of elderly and deceased persons, as well as numbers of still births and infant deaths are unavoidably crude. Birth records were available for over a third of the population and were used whenever possible for establishing ages. Death certificates were seldom issued. Initial blood group studies were performed on 250 samples of blood drawn on several weekends during the summer of 1962 (Erickson et al., ’63). These samples were tested with suitable antisera for A, B, M, N, S, C, D, E, c, e, Kell, Cellano, Duffy and Diego antigens. Known blood cells (Pano-cells, Knickerbocker Co., New York, N. Y.) and duplicate samples (single blind) were used to control the groupings studied. These controls showed many inaccuracies in the Rh system. I N T E R P R E T A T I O N S O F P O P U L A T I O N ISOLATES TO Outlier SE 4 ~o Outlier P I\ 4 Fig. 2 Map of Aguacatenango, Chiapas, Mexico. 109 110 ERICKSON, NERLOVE, CREGER AND ROMNEY Demography The results of the preliminary blood group study led to a n investigation of a The complete census revealed that the larger sample of the population for the an- barrio division is not the only spatial divitigens which were found to vary signifi- sion with social implications of the popucantly in the isolate. Many biochemical lation of Aguacatenango (fig. 2). There traits, such as red blood cell proteins are six outliers, four of which are named. (Lisker et al., '66) or serum proteins (11, 111, IV, and V numbered for purposes (Lisker et al., '67; Melartin et al., '67) are of reference). Each of the outliers has its not useful markers for this population be- own cross, which is a n indication of some cause of insufficient variability in geno- autonomy. type frequencies. Blood samples were colOutlier I, which is adjacent to Barrio 2, lected from 460 individuals on two consec- is comprised of eight households of Tzeltalutive days during the summer of 1964. speakers. All members of these households They were taken from individuals who were were born outside or had parents who were willing to give blood in exchange for five born outside Aguacatenango. Outliers 11, pesos (4OC U. S . ) . The Mexican national 111, and IV form a semicircle around the election day was selected for the first day periphery of Barrio 1. Two additional outof blood collection as everyone gathered in liers are across the road from Aguacatethe town for the balloting and ensuing fes- nango (Outliers V and VI). Outlier V is tival. The town elders (officials), with comprised of seven households of Aguacawhom the anthropologists had maintained tenangueros who moved from a group of good rapport through the years, were the households on the periphery of Barrio 1. first to donate. About 20 more older men, Outlier VI is comprised of 12 households closely associated with the elders, followed made up of people from Huistan, a Tzotzilthem, but all other mature males were ex- speaking town several miles from Aguacatenango. tremely reluctant to donate. Instead, they In summary, the named Outliers 11, 111. sent their wives and children, who thus IV, and V have more intimate connections made up three-quarters of the sample. both spatially and socially, with the two Young males were particularly eager to do- barrios of Aguacatenango. Outliers I and nate blood as they could keep the money VI have no blood or affinal ties with either for themselves to spend at the festival. of the two barrios. The clotted samples were packed in ice, Three households cannot properly be shipped, and arrived at Stanford Univer- considered part of either outliers or barrios sity five days later. These cells were tested by virtue of both birth place and interacfor A, B, M, N, S, C, D, E, c, and e anti- tions of the residents. There is one family gens. Of the sample of 460 donors, 400 of Tzotzil Indians from the town of were definitely identified on the geneal- Huistan near Outlier 111, and there are two ogies and their blood samples were used families cf Tzeltal Indians of unknown for computations of phenotype frequen- origin near the northeastern periphery of cies. Of the 60 donors dropped from the Barrio 2. The household and population distribusample, blood group data from 50 of these tion across barrios and outliers in Aguacawere not used because the individuals were not from Aguacatenango proper, but from tenango is presented in table 1. Population pyramids of age and sex composition for El Puerto, one of the two colonies nearby. Barrio 1, Barrio 2, and Outliers 11-V are The remaining ten blood samples were dis- presented in figure 3 . Fifty-seven per cent carded because they were duplicates or of all the males are 20 years of age or there had been a mix-up in sample num- younger. The birth rate is 67 per 1,000. bers. Blood group data from each of the A five-year period was selected for investi400 subjects studied were punched on key- gation of mortality rates among nonsort cards along with the donor's age, emigrant Aguacatenangueros ( 1300 indibarrio number and the name of his patri- viduals). The total number of deaths in each age interval of five years was divided lineal lineage. 111 INTERPRETATlONS O F POPULATION ISOLATES TABLE 1 Distribution of the households and population of Aguacutenango ucross burrio and outliers No. of households Barrio I Barrio I1 Outliers 11-V Outlier I Outlier V I Misc. No. of males No. of females 212 255 164 16 218 75 95 60 8 12 Totals 3 8 276 175 24 29 7 253 676 729 v) 51-60 a a 41-50 0 w ” 31-40 4 21-301 b l r t n - 10 L zoo 160 120 80 40 o 40 NUMBER DO 120 160 200 240 280 Fig. 3 Age and sex distribution of barrios I and I1 and outliers 11-V (for 1300 individuals, 631 males and 669 females - read 233 males under 10 rather than 205). by five to give a n average estimate for deaths occurring in one year. A separate interval was used for one year and under. The deaths occurring in this interval include still births. The mortality in Barrio 1, Barrio 2, and Outliers 11-V is presented in table 2. Fertility was tabulated on the basis of completed families for the population of Barrio 1, Barrio 2, and Outliers 11-V (table 3 ) . The sample upon which the tabulation is based includes all women who were 50 years and over during the summer of 1963. None of these women had given birth within the last five pears. There were 53 such women, one of whom was dropped from the sample because of inadequate information. Twenty-six of the 52 women in the sample were currently living with their husbands. The number of children included on a woman’s fertility history includes still-borns and deceased as well as all living children. The particular father of the child and the child’s present residence were not considered. 21 No. of individuals 430 53 1 339 40 50 15 1,405 Population structure of Barrios 1 and 2 The genetic contributions from the outliers, from the other barrio, and from outside villages to Barrios 1 and 2, respectively, were not considered. These genetic contributions were, at best, difficult to calculate. The time depth which i t is feasible to include in such calculation is quite limited. Data for the calculation were available for 961 individuals. The contributions from the outliers were computed on the basis of the location in which subjects’ parents reside. The contribution from one barrio to another was computed from all cross-barrio marriages. Those offspring from cross-barrio marriages who live in Barrio 1 are considered contributions of Barrio 2 to Barrio 1. Similarly, those offspring from cross-barrio marriages who reside in Barrio 2 are considered contributions of Barrio 1 to Barrio 2. The genetic contributions from outside the town have been tabulated on the basis of outsiders married to Aguacatenangueros with offspring living in Aguacatenango. These genetic contributions to each barrio from these various sources are shown in table 4. The chi-squares show significant differences between the barrios in the various rates of genetic contributions. The general picture of genetic contributions indicates that Barrio 2 is more isolated than Barrio 1. The proportion of the population of the outliers originating in either of the two barrios could not be ascertained. As noted, the connections of Outlier V are with Barrio 1. It seems that the populations of Outliers 11, 111, and IV also derive from Barrio 1. Note, however, that a genetic contribution from groups which have derived from the group which 112 ERICKSON, NERLOVE, CREGER AND ROMNEY TABLE 2 Mortality in Aguacatenango froin interview data Deaths Deaths FeMales males 5 year 1 and under 23 + 15 4 2 3 1 0 1 1 3 2 2 2 0 27 15 7 0 2 2 0 1 1 1 0 2 1 0 Totals 59 59 2-5 6-10 11-15 16-29 21-25 26-30 31-35 36-40 4 1-50 51-60 61-70 71-80 80 Totals 1 year 4.60 5.40 3.00 3.00 0.80 1.40 0.00 0.40 0.40 0.60 0.20 0.40 0.00 0.00 0.20 0.20 0.20 0.20 0.60 0.20 0.40 0.00 0.40 0.40 0.40 0.20 0.00 0.00 Population at risk as of Summer 1963 10.00 6.00 2.20 0.40 1.00 0.60 0.00 0.40 0.40 0.80 0.40 0.80 0.60 0.00 Estimated death rate per 1,000 169.9 29.0 10.4 2.9 7.3 6.3 0.0 5.6 6.3 6.1 12.9 5.3 5.0 0.0 59 207 211 138 137 96 92 72 64 131 31 15 12 7 1,272 TABLE 3 Fertility in Aguacatenango f r o m interview data Frequency distribution of no. of husbands (average no. of husbands: 1.64) Number of husbands a woman has had 0 1 2 3 4 5 Number of women over 50 years having had that number of husbands 0 30 13 7 1 1 Frequency distribution of no. of children (average no. of children: 6.94) Number of children 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 52 is the recipient of their contribution is characterized by an element of circularity. In contrast, Barrio 2 receives a greater genetic contribution from non-Aguacatenangueros. It is clear that people are mov- Number of women over 50 years having had that number of children 1 3 1 5 4 2 3 14 6 3 3 4 1 0 0 0 0 0 1 0 1 52 ing out of the barrios as the size of the population has remained relatively stable for the last seven years (Metzger, '56 for 1956 census; Mexican government for 1960 census; present study for 1963 cen- 113 INTERPRETATIONS O F POPULATION ISOLATES TABLE 4 Genetic contributions to barrio I and 11 from outliers, other barrio, and non-Aguacatenangueros as tabulated from interview data Number of individuals now living in Birthplace of Barrio I (pop. 430) Barrio11 (pop.531) 14 7 54 (41 ’) 9 4 0 6 Outlier I1 Outlier I11 Outlier IV Outlier V 5 x2 =72.2 < 0.01 (0.001at least) Total from Outliers 84 15 p Barrio I Barrio I1 - 21 xa=6.86 - p 13 28 p 64 12.1% xe=75.2 131 30.5% Total Total % 1 Indicates 34 < 0.01 x2~2.83 Non-Aguacatenangueros and offspring p < 0.05 < 0.01 (0.001 at least) the contribution of the Giron family which is discussed in the text below. sus) though the birth rate is 67 per 1,000 and the adjusted rate is 21 per 1,000. A comparison of the population pyramid to the age-related death rate suggest that the narrowing of the population pyramid from ages 20-50 is largely due to emigration. A particularly well-documented case of the immigration, multiplication and spread of a kindred of outsiders in Aguacatenango is that of the Girons who came from San Rafael and account for a large part of the outside contributions to Barrio 1 (table 4). The distribution by generation and location of the 72 members of the Giron family who are in Aguacatenango is presented in TABLE 5 Distribution of girons i n Aguacatenango by generation and location (total 7 2 ) Generation G-1 (oldest) 6-2 Barrio 1 IV v 2 0 1 0 2 0 0 2 2 4 4 2 6 0 0 0 7 12 0 8 1 Total 15 22 20 1 Males Females 11 4 Total 15 0 0 0 0 1 1 Males Females 0 1 Total Males Females Total 6-3 6-4(youngest) -::? -:;? Males Females table 5. Not all Girons in Aguacatenango are descendants of the three siblings of the oldest generation (G-1) who live in Aguacatenango. Some Girons living in Aguacatenango are descended from the daughter of another sibling of that generation who has never lived in Aguacatenango. Blood group typing It was not feasible to double-test the blood samples, and controls detected certain laboratory errors. Control typings for anti4 (direct agglutination) revealed false negatives. Four out of the 460 CDEce typings showed “impossible” phenotypes although there had been no trouble in typing controls. The inaccuracy in the N and S antigen typings affects MNS comparisons, between the data of this study and that of other studies, but does not affect the intragroup comparisons that have been made because the members of families and of the various subgroups that have been defined came to donate blood in random order. Therefore, errors in typing should be random with each group. Considering the degree of mating across barrios, it seems likely that the population of Aguacatenango would be genetically homogeneous, However, the chi-square analysis of blood group allele frequencies between the two barrios in the preliminary data suggested that they should be considered as partial genetic isolates (Erickson et al., ’ 6 3 ) . 114 ERICKSON, NERLOVE, CREGER AND ROMNEY The present blood group study does not substantiate at statistically significant levels that the two barrios are partial genetic isolates. Nonetheless, there is an interesting trend of genetic difference when one compares the barrios and the various sub-samples, i.e., corrected (described below), over 35 years, and under 15 years, within them for Rh and MNSs frequencies. Table 6 shows the composite phenotype frequencies of blood groups for Barrios 1 and 2, while table 7 presents their derived genotype frequencies. Chi-squares for differences between the barrios and between the various sub-samples examined within the barrios are presented in table 8. The biggest differences between barrios are seen in the sub-sample of individuals 35 years and older from each barrio. When only individuals of 15 years or younger are considered, the barrios seem almost identical in both blood group systems. To remove the bias in the data caused by disproportionate representation of various families, a corrected sub-sample for each barrio was constructed. The two sub-samples could not contain more than two individuals from each patrilineal lineage, nor, where possible, could they contain more than two members of the same nuclear family. Using these two sub-samples to compare barrios for Rh and MNSs frequencies, it is shown in tables 6 and 8 that the phenotypes are not affected by correcting for family representation. Calculations were made to see if phenotype frequencies fit Hardy-Weinberg equilibria. The results of these calculations are presented in table 9. It can be seen that each barrio approaches Hardy-Weinberg equilibrium for Rh but the total population does not. Through Barrio 2 approaches equilibrium, neither Barrio 1 nor the total population of Aguacatenango is in equilibrium as determined by the calculations for the MNSs data. The MNSs data are suspect because of trouble with the anti-N typings. Nonetheless, the errors in MNSs typing should equally affect the calculations for each barrio. The MNSs data, as well as the Rh data, however, fit the Hardy-Weinberg equilibria better for each barrio separately. Kell Cellano, Duf€y, Diego and Gm and Inv allotype frequencies are presented in table 10. The Duffy and Diego phenotype frequencies can be utilized along with the total Aguacatenango Rh frequencies to compare this partial isolate to the one other partial isolate that has been sampled in the Tzotzil-Tzeltal area: Chalchihuitan, a TABLE 6 Composite o f blood group phenotype frequencies f m barrios 1 and 2 i n Aguacatenango Phenotype 0 A ccddee ccDee ccDEE ccDEe CCDee CcDee CCDEe CcDEE CcDEe NNs NNS NMS MMS MNs MNS Total Barrio 1 Barrio 2 243 4 0 1 34 21 69 20 150 1 1 1 14 14 49 17 3 3 12 84 23 10 100 42 39 32 247 6 452 20 6 44 24 33 24 151 Barrio Barrio Corrected 1 Corrected 1 35 years Barrio 1 35 years Barrio 2 15 years Barrio 3 15 years Barrio 2 57 2 0 0 7 117 1 16 3 37 0 0 0 2 4 16 6 78 0 0 1 11 7 22 0 0 1 8 5 2 6 37 a4 2 0 0 11 10 22 10 1 4 27 11 2 23 20 15 15 56 0 0 0 6 5 21 4 3 4 13 7 1 19 9 12 a 3 22 3 2 23 14 7 10 86 56 59 a 9 7 a 0 0 16 7 35 8 3 7 39 2 15 3 50 13 21 16 118 a 3 3 222 12 2 27 6 22 9 78 Corrected for family sampling by taking no more than two individuals from each identified patrilineal lineage and, when possible, not from the same nuclear family within that lineage. 2 Three Rh mistypes in Barrio 1 and 1 in Barrio 2 deleted. 3 MNSs mistype i n Barrio 1 deleted. 1 INTERPRETATIONS O F POPULATION ISOLATES 115 Tzotzil town studied by Matson and Swanson ('63). Table 11 presents these data. Wide differences are observed. DISCUSSION The demographic data show a broadbased population pyramid which narrows quickly because of high infancy and child mortality. Most of the remaining population survives through middle age. The mortality figures do not account for all of the further narrowing of the pyramid. Part of this narrowing is the result of emigration. The population data show an excess of females, which probably relates both to the "social" deaths of males (e.g., murder for witchcraft) and to the usual rate of differential viability. Fertility is noted to be variable with a mode of seven children, with a range of none to 20. The age of 15 years is used as the initial age of reproduction. The data in figure 3 and tables 2 and 3 can then be used to calculate indices of selection intensity following Neel and Chagnon ('68). I,, the portion of the selection intensity due to mortality (of females) prior to the age of reproduction equals 0.767; while If, the portion due to fertility differences among women reaching the age of reproduction, equals 2.682. These indices are considerably higher than those of the "primitive" Xavante (Neel et al., '64). We have been unable to calculate the breeding size of the population accurately, although only a few mature individuals have had no offspring. Although the genealogies show that there are a fair number of third-cousin marriages, we have also been unable to calculate a coefficient of inbreeding. If one could state the coefficient of inbreeding as well as the differential fertility accurately, the desirable next step would be to analyze their respective contributions to the evolution of the present gene frequencies in the population. There are numerous barriers to such an analysis. First, it is necessary to know what the selection coefficients are in this population for certain blood group phenotypes or genes closely linked to them. In order to know these selection coefficients, knowledge of the phenotypes of both parents and children in a number of families is needed. Such knowledge would enable the detec- 116 ERICKSON, "ERLOVE, CREGER AND ROMNEY TABLE 8 Chi-square values and probabilities Between barrios Comparison Rh' Chi-square Barrio I to Barrio I1 Corrected Barrio I to Probability 0.5 X s Z =5.23 0.20 ~ 5 % 0.4 xs3= 1.96 0.75 x f =7.98 0.08 x 2 =6.94 0.15 xs2=1.75 0.89 x4z =3.94 0.40 =5.23 corrected Barrio I1 1 Chi-square xsZ~ 4 . 9 6 Older than 35, Barrio I to Barrio 11 Younger than 15, Barrio I to Barrio I1 MNSs Probability Rare Rh phenotype pooled as in table 9. TABLE 9 Closeness o f fit t o Hardy-Weinberg equilibria i n Aguacatenango barrios 1 and 2 considered both separately and together Aguacatenango Observed Expected Barrio 1 Observed Expected 0 3 6 45 49 17 14 14 0.168 5.438 3.088 xS=7.053 51.151 Dr=5 44.031 p 0.05 18.485 14.200 10.497 1.940 24 84 69 20 34 21 1 0.273 8.158 6.149 x2=12.946 93.188 Dr=5 60.831 p == 0.025 22.138 34.561 16.687 2.014 64.118 59.918 124.990 x2=31.278 109.865 D r = 2 13.966 p 0.001 24.143 42 32 100 39 10 23 40.302 25.922 87.772 x2=24.105 62.929 D,=2 7.795 p 0.001 11.279 66 56 MNSs Observed 0 0.441 6 13.593 18 9.237 x2= 19.718 129 144.898Dr=5 118 104.684 p 0.01 40.430 37 48.340 49 27.474 35 3.904 2 0 3 12 < Rh Barrio 2 Expected 'g 16 43 < 1 24 44 33 6 20 -< > 23.926 24.132 37.745 x2= 7.350 45.527 Dr=2 5.942 p 0.025 13.728 < TABLE 10 Additional blood group and serum allotype data f o r Aguacatenango System Phenotype KellCellano K(+ )k( - ) + + K( )k( ) K(-)k(-) no. % 0 0 0 100 0 132 Genotype R 12 Allele Frequency + 0 1.0 * Duffy Fy(a+ 1 Fy(a- 1 81 99 44.96 55.04 FY" not F p 0.258 0.051 0.7420.051 Diego Di(a+) Di(a-) 9 124 6.8 93.2 Dia not Dia 0.0352 0.023 0.965 f 0.023 Gm Allotypes Gm 1 , 2 , 1 3 Gm 1,13 Gm 1 , 2 Gm 1 2 3 74 77 1.28 1.92 47.44 49.36 Gml Gm1.2 Gm'Ja 0.7000 0.2839 0.0161 Inv Allo- Inv(l+a+) types Inv( 1+ a - ) 85 3 49 62.04 2.19 35.77 Inv(1-a-) 117 INTERPRETATIONS O F POPULATION ISOLATES TABLE 11 Comparison of blood group phenotypes between two semi-isolates: Chachihuitan and Aguacatenango Expected Observed Phenotype CCDee CcDEE CcDEe CcDEe ccDEE ccDEe ccDee, ccdee, CCDEe 2:- Chalchinango hitan Chalchi- tz:z hitan nango 15 0 36 3 22 2 2 118 18 129 37 48 35 9 22.45 3.04 27.35 6.75 11.81 6.24 1.86 110.55 14.96 137.15 33.25 58.19 30.76 90.14 80 394 79.50 475.00 76 4 81 99 52.35 27.65 80 180 80.00 2.472 3.040 2.335 2.083 8.782 2.881 0.011 0.502 0.618 0.434 0.423 1.784 0.584 0.001 ~e’=26.060 p 117.70 62.30 10.62 24.73 xie=67.02 15 65 9 124 9.02 10.98 14.96 118.04 80 133 20.00 133.m 3.95 0.50 < 0.01 10.15 21.52 p < 0.001 2.37 0.30 tion of statistically significant shifts from behavior and differential fertility on allele the theoretically expected (based on the frequency changes, our primary involvehypothesis of no selection) segregation ment became that of defining the variables ratios (Morton, ’59). Selection coefficients under study. Verbal statements about might be taken from Morton’s data on a breeding behavior recorded in the genealNortheastern Brazilian population which ogies incorporate significant error and culshowed selection only in the ABO system turally defined social entities may be (Morton et al., ’66). It is not a safe as- widely discrepant with the actual populasumption, however, that the same selec- tion entities. Such a discrepancy has been tion forces are at work in the two dissim- noted previously with respect to marriage ilar environments. It is only after one patterns (Kunstadter et al., ’ 6 3 ) . knows what portion of allele frequency difThe problem of “illegitimacies” is a genferences might be due to selection that the eral one, which serves particularly to comeffects of total population size, breeding size and coefficient of inbreeding on ge- plicate studies of differential fertility. For netic drift can be considered. The contribu- example, the Girons, who have made such tion of the population’s mating system to a disproportionately large apparent contrigenetic drift is similarly unamenable to bution to the population, may have made a still larger contribution by way of “illegiticalculation. Secondly, the mathematics for analysis macies.” For 17 families there are blood of population genetics are predicted on samples for both parents and a total of 37 sampling a fairly large population, not on of their offspring. Children with blood typ complete enumeration of a small popula- ings which are multiply incompatible with tion. The difficulty of trying to analyze a those of their parents included one whose small population has been noted (Sutter, blood is incompatible with that of the ’63; Arends et al., ’67) but the solution mother. It is usually assumed that the mother indicated is a biological parent, but does not appear to be forthcoming. Although the goal of this study was to this rule can be in error if step-parentage provide data for comparative studies of or adoption are not indicated by informthe effects of parameters such as mating ants (Buettner-Janusch et al., ’64). 118 ERICKSON, NERLOVE, CREGER AND ROMNEY If an Aguacatenaguero is asked if there are any “outsiders” who live in town, he would indicate ladinos living in the town and the inhabitants of Outliers I and VI. He considers the town to be endogamous. He also considers each barrio to be endogamous but close questioning would elicit the fact that he knows of exceptions to the endogamy rule. It is only when a complete census is taken that the disparity between real and ideal social entities is discovered. It is important to consider, however, how much of that disparity has to do with the exogamous marriages necessitated by imbalances in the spouse pool. Even geographically, the apparent isolate can be shown to consist of micro-units. One-third of Aguacatenango’s population lives outside of the barrios in outliers. The barrios show a high degree of intermarriage and outsiders are not only present, but their offspring are a considerable proportion of the population. The suggestion, then, of genetic isolation between the barrios found in the preliminary study is not only interesting but also somewhat puzzling. In the present study, the differences do not reach signscance at usually accepted levels of probability for chi-square tests, but there is a clear trend of increasing differences with increasing age of the individuals. Such a trend suggests that perhaps there had once been greater isolation between the barrios and that it is now breaking down. Such age-related genotype differences are not found in populations which are not isolated. For instance, studies of several serum protein factors in West Greenland Eskimos show no regional variations, even for individuals over 55 years of age (Persson, ‘68). The evidence of better fits to the Hardy-Weinberg criterion for genetic equilibria in each barrio separately is also suggestive of barrio isolation. If the possibility is considered that each barrio might have a separate isolate, then it is surprising to find for each barrio a fit to Hardy-Weinberg equilibria which is premised on a large population for panmixis (Sutter, ’63). Yet good fits to Hardy-Weinberg equilibria in much smaller and more highly inbred populations are being found (Nee1 et al., ’64; Arends et al., ’67). It will not be possible to understand the genetic structure of small populations until further genetic studies of small populations are made. There are three hypotheses that would explain gene frequency differences between the two barrios: 1. differential immigration, 2. separate origin or 3. differential selection. In distinguishing between these hypotheses, it is important to decide whether the gene exchange between barrios observed in 1963 is a recent innovation. If the inter-barrio exchange is of long standing, large initial allele frequency differences would be expected or the mechanisms maintaining allele frequency differences need to be very strong. It seems highly unlikely that the population living in each barrio is exposed to a selective environment distinct enough to maintain allele frequency differences. The forces of modern communication and transportation, however, which can serve to contribute to a breakdown of old customs of endogamy conceivably could have greatly increased the degree of genetic contribution to Aguacatenango from the outside. Along with a breakdown in town endogamy could be a breakdown in barrio endogamy and thus account for a recent onset of higher gene exchanges between barrios. If one barrio has always received more outsiders or each barrio has received outsiders from a different area, differential genetic contribution to the two barrios could explain blood group allele frequency differences. A t present, Barrio 1 has a significantly higher rate of barrio exogamy (mostly involving outliers) while Barrio 2 has a greater number (not significant) of outsiders. Further blood group studies of “outsiders” and individuals who have changed barrios would allow us to determine whether their movements are important for the possible genotype differences that we find between the two barrios. More data on phenotypes of complete families such as the Girons are needed for better assessment of the effects of differential fertility on modal genotypes in this village. In Japan, Yanase (’65) has found decreased fertility for immigrants rather than the increased fertility we find in this one case. Perhaps a combination of differential immigration and differential fertility (with or without selection) would determine barrio allele frequency differences. INTERPRETATIONS OF POPULATION ISOLATES It has been mentioned that the origin of the barrios is unknown. If it represents an old clan distinction, it is probably that the clans had similar allele frequencies (by reason of common ancestry, exposure to nearby identical selective environments and a continued small degree of inter-clan gene exchange. Note that this argument represents the sort of hypothesis that our methods could help substantiate.) It is also possible that the barrios might have originated when Quesaltepec was added to Aguacatenango during the directed population movements in the seventeenth century. If so, two barrios, with rules of barrio endogamy, might have been set up to maintain isolation. It is interesting to note that paricialidades had some political autonomy (Calnek, ’61) which provides more reason to believe that the combined populations would not immediately merge. The comparison of modal genotype frequencies between Aguacatenango and Chalchihuitan, a northern Tzotzil semiisolate, show significant differences of varying degree for different blood group systems. The differences found suggest that outsiders, but from within the Tzotzil-Tzeltal area, would contribute significant blood group heterogeneity to the population. Blood group data from many other semi-isolates in this region are needed before a meaningful discussion of gene clines and their relaticnship to local historical events and geographical location is possble. The other blood group genotype and allotype frequencies presented in table 9 may be helpful in this regard. Blood group data (Matson and Swanson, ’59, ’61, ’63) and allotype data (Steinberg et al., ’61) on samples of Tzotziles and Tzeltales are available as a preliminary indication of the genotype variation expected. ACKNOWLEDGMENTS The work was supported by a grant from the Stanford University School of Medicine, Dean’s Office Fund, by the Pamela Weigl Memorial Fund for Research in Diseases of the Blood, Stanford University, by a U.S.P.H.S. grant, No. CH 00050-04, to Dr. A. Kimball Romney, and by two Russell Sage Foundation grants to Robert P. Erickson. Antisera for Kell, Cellano, and Duffy antigens were kindly donated by 119 Ortho-Pharmaceuticals, Raritan, New Jersey. The antisera for determination of Diego antigen was due to the courtesy of Dr. Miguel Layrisse, Caracas, Venezuela. The allotype determinations were kindly performed by C. Ropartz. We are indebted to Robert Armstrong, John Hotchkiss, and Vicente Gomez for help with the field work. Jo Anne Dalziel, Joan Harris, and David Clark are to be thanked for excellent laboratory work. Paul Comer and Peter Workman have aided us with calculations. Duane Metzger and Howard Cann have provided valuable discussion. LITERATURE CITED Adams, R. M., and N. A. McQuown 1959 Prehistory, and post-conquest developments. Report of the University of Chicago, “Man-inNature” Project in Chiapas, Mexico, mimeographed. Arends, T., G. Brewer, N. Chagnon, M. L. Gallango, €I. Gershowitz, M. Layrisse, J. Neel, D. Shreffler, R. Tashian and L. Weitkamp 1967 Intratribal genetic differentiation among the Yanomama Indians of Southern Venezuela. Proc. Natl. Acad. Sci. U. S.A., 57: 1252. Buettner-Hanusch, J., J. R. Bove and N. Young 1964 Genetic traits and problems of biological parenthood in two Peruvian Indian tribes. Am. J. Phys. Anthrop., 22: 149-154. Calnek, E. E. 1959 Ethnohistorical notes. Report of the University of Chicago “Man-inNature” Project in Chiapas, Mexico, mimeographed. 1961 Distribution and location of the Tzotzil and Tzeltal pueblos of the highlands of Chiapas from earliest times to the present. University of Chicago, Anthropology Department, mimeographed. Cavalli-Zforza, L. L., I. Barrai and A. W. F. Edwards 1964 Analysis of Human Evolution Under Random Genetic Drift. Cold Spring Harbor Symp. Quant. Biol., XXM: 9-20. Erickson, R. P., D. Clark, W. P. Creger and A. K. Romney 1963 The comparison of genetic and anthropological interpretations of breeding patterns and breeding isolates. Abstracts, 62nd Annual Meeting of the American Anthropological Association: 16. Gajdusek, D. C. 1964 Factors governing the genetics of primitive human populations. Cold Spring Harbor Symp. Quant. Biol., XXIX: 121136. Kunstadter, P., R. Buhler, F. F. Stephan and C. F. Westoff 1963 Demographic variability and preferential marriage patterns. Am. J. Phys. Anthrop., 21: 511-520. Lisker, R., G.Zarate and A. Loria 1966 Studies on several genetic hematological traits of Mexicans IX. Abnormal hemoglobins and erythrocytic Glucose-6-phosphate dehydrogenase deficiency in several Indian tribes. Blood, 27: 824. 120 ERICKSON, NERLOVE, CREGER AND ROMNEY Lisker, R., G. Zarate and E. Rodriguez 1967 Studies on several genetic hematological traits of the Mexican population XIV. Serum polymorphisms in several Indian tribes. Am. J. Phys. Anthrop., 27: 27-32. McKusick, V. A., J. A. Hostetter, J. 0. Egeland and R. Eldridge 1964 The distribution of certain genes in the old order Amish. Cold Spring Harbor Symp. Quant. Biol., XXIX: 99114. Matson, A., and J. Swanson 1959 Distribution of hereditary blood antigens among the Maya and non-Maya Indians in Mexico and Guatemala. Am. J. Phys. Anthrop., 17: 49-74. 1961 Distribution of hereditary blood antigens among American Indians in middle America: Lacondon and other Maya. Amer. Anthrop., 63: 1292. 1963 Distribution of hereditary blood antigens among Indians in Middle America 11: Tzotzil and other Maya. Am. J. Phys. Anthrop., 21: 1-14. Melartin, L., B. S. Blemberg and R. Lisker 1967 Albumin Mexico, a new variant of serum albumin. Nature, 215: 1288. Metzger, D. 1956 The social organization of Aguacatenango, Chiapas: a preliminary statement. Unpublished. Morton. N. E. 1962 Segregation and linkage. In: Methodology and Human Genetics. W. J. Burdett, ed. Holden-Day, San Francisco, 17. Morton, N. E., H. Krieger and M. P. Mi 1966 Natural selection on polymorphisms in Northeastern Brazil. Amer. J. Hum. Genet., IS: 153. Neel, J. V., and N. A. Chagnon 1968 The Demography of Two Tribes of Primitive Relatively Unacculurated American Indians. Proc. Natl. Acad. Sci., 59: 680. Neel, J V., and F. M. Salzano 1964 A prospectus for genetic studies of the American Indian. Cold Spring Harbor Symp. Quant. Biol., XXIX: 85-98. Neel, J. V., E. M. Salzano, P. C. Junqueira- F. Keiter and D. Maybury-Lewis 1964 Studies on the Xavante Indians of the Brazilian Mato Grosso. Am. J. Hum. Genet., 16: 52-140. Persson, I. 1968 The distribution of serum types i n West Greenland Eskimos. Acta Genet., Basel, 18: 261. Roberts, D. F. 1968 Genetic effects of population size reduction. Nature, 220: 1084. Steinberg, A. G., M. S. Cordova and L. Ruben 1967 Studies o n several genetic hematologic traits of Mexicans XV. The Gm allotypes of some Indian tribes. Amer. J. Hum. Genet., 19: 747-756. Sutter, J. 1963 The relationship between hum a n population genetics and demography. In: The Genetics of Migrant and Isolate Populations, E. Goldschmidt, ed. Williams and Wilkins, New York, 160. Swadesh, M. 1959 Linguistics as an instrument of prehistory. Southwestern J. of Anthrop., 15: 20-35. Yanase, T. 1965 Degree of endogamy and effective migration rate in isolated populations. Proc. Japan Acad., 41: 169. Yasuda, N., and N. E. Morton 1967 Studies on human population structure. Proc. I11 Intern. Congr. Human Genet., Johns Hopkins Press, Baltimore, 249.