Human Lymphocyte Antigens: Apparent Afro-Asiatic, Southern Asian, & European
HLAs in Indigenous American Populations
by
James L. Guthrie
[1]
New England Antiquities Research Association
Human Lymphocyte Antigens: Apparent Afro-Asiatic, Southern Asian, & European HLAs in Indigenous American Populations
by James L. Guthrie [1]
Section 1
In our pursuit of cultural and physical human diffusion around the globe, NEARA encourages research exploring "hard" scientific evidence. Over the years, Jim Guthrie has published numerous articles in the NEARA Journal on many subjects. Now, his long interest in micro-biology has culminated in a comprehensive article on human lymphocyte antigens and their dispersal into indigenous American populations, published in Pre-Columbiana, Volume 2, Number 2 & 3, December 2000 & June 2001. Pre-Columbiana, like the NEARA Journal has a limited circulation and the NEARA editorial team felt that this work is so important, that it must reach as wide an audience as possible, scientist and layman alike. In collaboration with Pre-Columbiana editor Stephen Jett and with permission, we are pleased to make this ground-breaking research available on through the internet.
The NEARA editors
Abstract
Studies have shown that the number of human lymphocyte antigen (HLA) alleles characteristic of indigenous American populations is relatively small, and that some isolated South American tribes possess only a few types that are common throughout the Americas. But other groups, especially those near sites of former Mesoamerican and Andean urban societies, exhibit HLA alleles that are rare in America but common in certain Afro-Asiatic, South Asian, and European populations. These unexpected genes account, on the average, for 6-7% of the American HLA total, but range as high as 24%.
The atypical genes are postulated to have been acquired by assimilation of foreign populations at various times after initial colonization of the hemisphere but prior to the sixteenth-century influx of Europeans and Africans, because they suggest gene-flow from places some scholars claim to have been in ancient contact with the Americas, such as North Africa and Southeast Asia. The occurrence of parallel anomalies in blood groups such as Rhesus, Kell, and Duffy, as well as in serum proteins such as transferrin and immunoglobin, supports this interpretation to some degree, but the small number and poor distribution of samples in all systems, including HLA, preclude conclusive results. Other explanations are considered possible but less likely. Key Words: Genetics, Americas, Migrations, Pre-Columbian.
Introduction
Human lymphocyte antigens (HLAs) are part of the histocompatibility system, whose main function is to produce antibodies. They are proteins on white blood cells that play a role in tissue and organ transplantation similar to that of the more familiar blood groups in transfusion. HLA distributions differ among world populations to such a degree that careful typing and matching must be done at transplant centers in order to minimize adverse reactions. This diversity gives population geneticists a powerful tool for tracing ancient migrations, and, at present, HLA distributions are more informative in this regard than are any other genetic system except DNA.
The purpose of this paper is to point out that certain indigenous American populations have HLA alleles that are rare in America but common in parts of the world not usually associated with American Indian origins, and many of the unexpected HLAs are characteristic of populations sometimes claimed, on the basis of other kinds of evidence, to have had ancient contacts with Americans. In other words, there seems to be genetic support for the idea of ancient interhemispheric mobility. I propose that the “non-Indian” HLAs were introduced from the outside at various times between the initial colonizations of the hemisphere and the late fifteenth century A.D. I also consider other possible explanations, but they seem less likely to me. The percentage of apparently foreign HLAs averages only 7% in the populations tested so far, and this observation does not contradict the supposition that the founding American populations were overwhelmingly Asian.
Many competent scholars have published evidence for intermittent contact between Native Americans and others over several millennia. Among the proposals are: contacts between Indonesia as well as Japan with the Pacific coast of America, circumpolar migrations, trans-Atlantic incursions from Africa and the Near East, and numerous interactions with China, India, and Pacific Oceania.[2] These publications are little known among academics, being largely ignored for reasons recently analyzed by Alice Kehoe (1998:190-207).
The prevailing view of anthropologists is that Native Americans were almost entirely cut off from developments elsewhere after the submersion of Beringia. Persistence of this belief is due in part to the fact that contrary evidence from diverse areas of scholarship has not been integrated and argued in a sober and convincing manner, so that few professional academicians are aware of more than a fraction of the pertinent literature. They find it easy to dismiss as isolated curiosities the parts that they know about. Full synthesis of the evidence for early transoceanic contacts would be an enormous job, one that seems unlikely to be done in the near future.
Another problem has been the adversarial nature of discussions, tempting participants to take extreme positions. Advocates of early contact sometimes make claims that go far beyond the evidence, while critics tend to base their rebuttals on preconceptions and outdated information. It is rare to find informed evaluation of new, unexpected data about possible early voyaging.
The HLA material presented here is based on the compilations of L. Luca Cavalli-Sforza, Paolo Menozzi, and Alberto Piazza in The History and Geography of Human Genes (1994). I think that the distributions of HLA types, combined with supporting data from other genetic systems, provide strong evidence that some American populations have assimilated significant numbers of foreigners in places where early contacts have been claimed on the basis of other considerations. I am not aware that anyone has discussed the potential of HLA data in evaluating these proposals except Cavalli-Sforza et al., in passing.
Cavalli-Sforza and his colleagues published The History and Geography of Human Genes to summarize decades of work on distributions of genetic polymorphisms, including those of HLAs. Their Appendix 2 lists worldwide frequencies of the 29 HLA families for which there are enough data to construct useful distribution maps, but there are many more types that are less well mapped at the present time.[3] This discussion is limited to those at the A and B loci tabulated by Cavalli-Sforza et al. (hereafter, “CS”). For my purposes, I will use the term “HLA type” to mean the family designation—e.g., A*1.
The objective of the CS study was to collect and compare, using mathematical models, gene-frequency distributions of world populations as a way of illuminating their population structures. Although data were gleaned from nearly a thousand publications, the samples are not evenly distributed and some regions are poorly represented, causing conclusions to be less reliable than would have been possible with a more complete database. The set of genes widely enough documented for their aims comprised 128 alleles from 491 individual populations and 116 aggregates (pp. 393-468); however, there were sufficient HLA data from only 132 individual populations and 78 aggregates, including 30 indigenous American groups. Only twelve type-A and 17 type-B HLAs were sufficiently well sampled for useful worldwide comparisons. The problems and caveats associated with the use of whatever data are available from the literature are thoroughly discussed by CS and should be kept in mind by anyone who uses their tables and maps for further analysis.
The set of HLA data has large gaps. For example, there are no data from Egypt and few from Northeast Asia. In North America, interpretation is severely limited by the absence of HLA samples from several important linguistic groups. Cherokee is the only eastern Amerind population in the set, and Navajo is the only representative of the set of populations commonly known as Na-Dene. Na-Dene is a linguistic term, and some find the category to be without much linguistic basis (a phantom), objecting to its use in classifying populations. I use it in this report as it is used in much of the cited literature to designate a genetically distinct set of populations of northwestern North America (northern Na-Dene) plus Navajo and Apache (southern Na-Dene).
In South America, there are no samples from Arawakan groups and there is only one sample each from Panoans and Tupians. In addition to the HLA data presented by CS, I have used results published by Belich et al. (1992) for the Guaraní (Tupian) and Caingáng (Ge) people. Specific linguistic classifications used in this report are those of Greenberg (1987) but as far as I am aware do not involve any disputed terminology. Designations are intended mainly for identification. As Greenberg, Cavalli-Sforza, and others have established, genetic and linguistic distributions are strongly associated, and the reasons for this have been explained in detail by Cavalli-Sforza, Piazza, Menozzi, and Mountain (1989) and by Greenberg (1995), who cited the statistical analysis by Penny, Watson, and Steel (1993).
In general, an individual with several HL antigens will be better able to resist disease than one with fewer. Many populations have all, or nearly all, of the 29 HLA types listed by CS; however, this is not true of indigenous Americans, who typically lived at population densities too low to maintain HLA diversity. Genes are apt to be lost whenever a small group splits from a parent population (the founder effect), so two long-separated tribes in the same geographic region may have different sets of a few HLA types at exaggerated frequencies. CS point out (p. 335) that in the total absence of cross-migration, drift eventually might lead to the survival of only one HLA allele in each population. Gene loss is minimized in large, stable populations, especially if missing or low-level genes are reintroduced from the outside. CS think there was gene loss during migrations to America but that later gene flow reintroduced some missing genes (p. 130).
American populations listed in CS have from seven to 26 HLA alleles. Those near the ancient urban societies of Mesoamerica and the Andes have the most, and some marginal tribes of South America have the fewest. The degree of HLA diversity in a population may be a measure of its former size and cosmopolitan nature. According to Parham and Ohta (1996:71), the large number of alleles in modern urban populations is “predominantly the result of admixture bringing together alleles that evolved under natural selection in previously separated populations.”
Some American HLA types remain undetected for technical reasons, and some with restricted world distributions were omitted from the CS tabulations (A*24, A*68, B*39, B*48, B*51, B*52). For these reasons, HLA totals given average 91% rather than 100%. Of those listed, four types account for 94% of the American A HLA-A total and six types account for 93% of the HLA-B total. I will refer to these ten as the “American” alleles. Some South American tribes apparently have only these alleles, whereas those near former urban centers tend to have significant percentages of HLAs that now are most common in the Near East, India, Africa, Northwest Europe, or Southeast Asia (including Pacific Oceania). I will refer to these as the “non-Indian” or atypical alleles.
Table 1 ranks the American samples according to the number of alleles reported. Except for Greenland Eskimos, who seem to have acquired some European genes, those with 16 to 26 alleles (first column) are from present or former urban populations. Those with ten or fewer types (third column) are exclusively South American samples from comparatively isolated places. Locations of the South American populations are shown in Map 1. Tribal/group designations used throughout this report are normally those of the original authors, even though a few names may no longer be in use.
Table 1. Number of HLA families tabulated by CS for American indigenes.a
Nahua
26
Inupik
13
Caingáng
10
Quechua 24 Ticuna 13 Trio 10 Cherokee 22 Navajo 13 Guaraní 10 Greenland Eskimo
21 C Amerind 12 Parakana 10 Eastern Maya 21 Warao 12 Mataco 9 Mapuche 19 NW Eskimo 12 Cayapó 9 Araucano 18 Aymara 12 Oyampí 9 Papago 17 Atacama 11 Yanomama 9 Pima 16 Yupik 11 Yupa 9 Zuni 16 E Eskimo 11 Bari 8 Makiritaré 8 Emerillon 7
aCaingáng and Guaraní from Belich et al. (1992).
Cherokee origins are controversial. Cherokees are genetically diverse and may, at least in part, have migrated to the U.S. Southeast from Mesoamerica. According to CS's genetic-distance table (p. 327), they are closest to, and not statistically different from, the Nahua, and Spuhler (1979) linked them genetically with Indians of the American Southwest.[4] Overall, they are more like Amerindians of the southwest than like neighboring Algonquians (CS 327). The HLA data suggest to me that they came from Mesoamerica or perhaps even Venezuela. But without HLA data from other eastern Amerindians, it is impossible to tell whether Cherokees are more diverse than the others.
American and Other HLAs
Of the 29 HLA types that CS listed, nine occur widely in the Americas and all but one of these has its highest frequencies there. They are likely to have been present in the earliest populations. A tenth (B*27) has its highest frequencies in the Navajo and Eskimo samples. These ten “American” alleles have interesting distributions outside of America that are clues to American origins and migration paths, but that topic will not be treated here except for brief comments in the appendix. The focus of this report is on the possible significance of the 18 “non-Indian” HLAs. A 29th allele, B*41, so far seems to be absent in America.
Table 2 lists the American samples in decreasing order of their “non-Indian” content. The numbers show percentages of these alleles in the total tabulated for that sample. For example, 24% of the Nahua total comes from 16 HLA alleles that seem “out of place” in America. The Nahua are at one extreme of the spectrum, while six South American tribes are at the other, showing none of the atypical alleles. Except for the northwest Canadian Eskimos and the Cherokees, populations with 10% or more are clustered either in the Andean region or in southwestern North America.
Table 2. “Non-Indian” HLA content of American samples, by percent.a
Nahua 24.0 Cherokee 17.4 Mapuche 17 Atacama 15.8 NW Canadian Eskimo 14.9 Araucano 13.4 Papago 12.4 Eastern Maya 11.5 Quechua 11.5 C Amerind composite 10.9 Pima 9.8 Caingáng 9.4 E Canadian Eskimo 7.6 Navajo 7.3 Greenland Eskimo 4.4 Oyampí 4.0 Aymara 3.7 Zuni 2.4 Warao 2.2 Guaraní 2.1 Inupik 2.1 Ticuna 1.7 Yupik 1.1 Trio 0.7 Parakana 0.5 Yupa 0.1
aThose with exclusively “American” HLAs are: Bari, Cayapó, Emerillon, Mataco, Makiritaré, and Yanomama.
The 18 atypical alleles account for only 6% of the American HLA-A total and 7% of the HLA-B total. Seven of these are termed “Eurasian” by CS (p. 130) and four are said to be characteristic of Southeast Asia or Pacific Oceania (p. 369). CS make virtually no comment on these in their chapter on the Americas, because they are concerned with main effects rather than with anomalies. However, both the Eurasian types (A*1, A*3, A*32, B*7, B*8, B*12, B*14) and the Oceanic types (A*10, A*11, B*13, B*22) have distributions in America that seem to reflect pre-Columbian contacts of the kinds that have been advocated on the basis of other kinds of evidence. There are other possible explanations that must not be ignored, however. One is that the first Native Americans possessed all of the HLA alleles and that some were lost, leaving a random pattern with no useful information about prehistory. Another is that reintroduction of missing genes took place recently through unrecorded post-Columbian contacts.[5] This is a serious problem and requires careful case-by-case study before conclusions are possible.
Until about 30 years ago some largely “unmixed” American populations could be studied, but CS used a few samples thought to have as much as 10% of recent admixture. They say (p. 341) that “In general, we have tried to avoid using populations in which admixture of some magnitude was suspected, but it was impossible to avoid mixed populations entirely without introducing an unwarranted bias.” Some anomalies may be explainable as recent admixtures, but I think that most are not. The apparently foreign HLA alleles are usually less characteristic of Spain, Portugal, or West Africa than of places alleged to have had earlier contact, such as Pacific Oceania, North Africa, or Southwest Asia, and in many instances other “marker” genes of modern European and West African populations are absent. Also, veteran linguistic scholars such as Key (1999), Foster (1999), and Stubbs (1998) are advancing reasons to think that elements common to Afro-Asiatic and Austronesian languages were present in certain Mexican and South American Indian languages long before post-1492 contact. Nevertheless, the possibility of modern admixture and other explanations must be kept in mind.
The CS volume is masterful. Except for Mourant, Kopec, and Domawewska-Sobczac's pioneering work of 1976 (hereafter, “Mourant”), there has been nothing like it. However, its very scope required selectivity and minimal discussion of several interesting distributions and relationships, and the authors expressed their hope that others will use the database for further investigations. In the discussion of HLA data, for example, there is no interpretation of the world distributions of B*21 or A*33, which (combined) contribute a fifth of the “non-Indian” data in America.
The apparently Arabian or North African B*21 reaches frequencies of about 10% in three samples of Uto-Aztecan speakers, yet CS say only that it peaks in the Caucasian-occupied portions of Africa (p. 187), that it averages 1.5% in America with a “maximum above 10% in the extreme southwest of the United States” (p. 334), and that it is absent from South America (p. 369)—the last despite their data showing traces in the Mapuche, Araucano, and Yupa samples.
A*33 seems to trace movement of a Near-Eastern population to Southeast Asia and South America,[6] but CS do not mention this. They say only that A*33 peaks in the Middle East and in Southeast Asia (pp. 288, 247) and that it is one of 13 alleles occurring in America at frequencies “significantly different from zero” (p. 334). Yet, the key to the synthetic maps summarizing the statistical treatment (p. 338) shows A*33 contributing 70-80% to the second principal component, with its strongest effect in eastern North America and Panama. Interpretation is left to others. Like B*21, A*33 is said to be absent from South America, despite a 5% level in the Quechua sample and a 3% level in the Aymara sample. These erroneous statements are probably based on the computer-smoothed distribution maps from which the actual data have disappeared. As recognized by CS, it will be profitable for others to examine parts of the data in more detail than they were able to do.
The authors seem to accept conventional ideas about the peopling of the Americas, avoiding much mention of possible post-initial-settlement gene flow from the outside. Negroid and Caucasoid genes are assumed to be recent acquisitions. This is understandable in discussions of eastern Venezuela, the Guianas, and eastern North America, where centuries of admixture are well documented, but readers who know the evidence for more ancient contacts may find the commentary inadequate. For example, one synthetic map (p. 339) shows apparent African admixture near the northeast coast of South America and in the southwestern United States. The South American admixture is considered “likely to have taken place,” but there is no comment on the Southwestern phenomenon, for which the conventional model provides no explanation. Acknowledgement that several early explorers reported seeing “Negroes” near South American coasts and in Central America would have been helpful.[7] The only departure from orthodox views appears on page 340, where Polynesian influence is invoked to explain an anomalous area on a color map of the Americas “in the middle of the Andes near Bolivia and Peru” caused by atypical data in several genetic systems for the Atacama, Mapuche, Araucano, and Aymara samples south of Lake Titicaca.[8] The use of genetic data to reconstruct population movements is at an early stage. Events that shaped present distributions are complex and controversial, but it is my view that reconstructions that ignore possible pre-Columbian gene flow between Native Americans and others from overseas will prove defective and ultimately will be revised.
Classification of Atypical HLAs
To organize this presentation, I have grouped the “foreign” HLAs into three somewhat arbitrary categories: those now concentrated in North Africa or Southwest Asia (Afro-Asiatic), those now prevalent in Northwest Europe (“European”), and those with their highest frequencies in South and Southeast Asia and the Pacific islands (“southern Asian”). These assignments are not completely clear-cut, because of overlapping distributions. For example, CS term A*1 “European” but it appears to me to have been dispersed from North Africa or Southwest Asia, so I have assigned it to the “Afro-Asiatic” set. The same may be true of A*33, about which CS say little. I attribute its present distribution to Asian intrusion into both Africa and America and have grouped it with the southern Asian HLAs. These problems are purely organizational and do not affect the central demonstration that unexpected genes are present in America.
Table 3 lists all 18 “non-Indian” alleles in decreasing order of their contribution to the American total. The most important is Afro-Asiatic B*21, which contributes 10.4% of the atypical HLAs found in American samples. Six alleles account for more than half of the total. The nine Afro-Asiatic types together contribute 47%; the five southern Asian types, 28%; and the four European types, 25%. These percentages are only approximate as they stand and would doubtless change with more complete sampling or with changes in classification.
Table 3. "Non-Indian" HLA alleles, in order of importance.
Allele % of Contribution Designation B*21 10.4 Afro-Asiatic A*33
9.6 Southern Asian B*7 9.1 European A*30 8.0 Afro-Asiatic A*32 7.9 Afro-Asiatic B*14 7.0 Afro-Asiatic B*12 6.9 European A*1 6.5 Afro-Asiatic B*22 6.4 Southern Asian A*11 5.4 Southern Asian A*3 4.7 European B*8 4.5 European A*10 4.0 Southern Asian B*17 2.6 Afro-Asiatic B*13 2.4 Southern Asian A*29 2.0 Afro-Asiatic B*18 1.9 Afro-Asiatic B*37 0.6 Afro-Asiatic
The atypical HLAs of the eastern Canadian and Greenland Eskimos are predominantly European, and those of the Nahua and Eastern Mayans are heavily southern Asian. Those with mainly Afro-Asiatic types are the Mapuche, Atacama, Papago (Tohono O’odham), Pima, Oyampí, Warao, and a composite sample of Central Americans (“Central Amerinds”).[9] Atypical HLAs of the Cherokee and Eastern Mayans are mainly of the Afro-Asiatic and southern Asian types, while those of the Araucano and northwest Canadian Eskimos are largely European and southern Asian. The Quechua display the greatest diversity, with significant levels from all three sets (see Table 11, in the appendix).
In the following discussion, I review each category and provide comments on individual alleles as well as two tables. One table shows the major American occurrences, with minor ones indicated in footnotes. The other table shows the highest worldwide frequencies CS listed, including American data. The number of world entries has been arbitrarily limited to about ten Old World populations for each allele, which is sufficient in most cases to give a good idea of its distribution. The reader should consult the CS volume for more detail. Supplementary Tables 11 and 12 are in the appendix. Table 11 is an expansion of Table 3, subdivided by regional designation, and Table 12 ranks American samples in each category.
CS represent some important populations by multiple samples—e.g., Bantus and Lapps. For these I have used the average values that CS tabulated unless a single high frequency seemed especially noteworthy. In a few cases, extreme genetic drift seems to have caused exaggerated frequencies of some alleles in small populations, such of those of Sardinia and of Easter Island. Also, space has been saved in Tables 5 and 9 by using average frequencies of A*1, B*7, B*8, and B*12 for three European composite samples in which values are tightly bunched. The composites are designated as “British” (Ireland, Scotland, Wales, England), “Germanic” (Germany, Netherlands, Austria, Switzerland), and “Scandinavian” (Sweden, Norway, Denmark).
Many references cited in the endnotes discuss evidence for early interhemispheric travel, but they represent only a small sample. Some topics have been discussed in far too many publications to be listed here, but the better ones may be found easily through the annotations of Sorenson and Raish (1996), which contain much more information than is included here. Readers familiar with this literature can form an opinion about the degree of fit between the distributions of atypical HLAs and the main proposals made by students of early voyaging. Older references are included, to show how early some perceptive postulates were made. It is important to realize that some scholars who compiled striking trait comparisons strongly opposed prehistoric contact as an explanation.
It is unfortunate that ideology has been allowed to color discussions of early voyaging and that participants are often assumed to have a political agenda. It has become fashionable to find “subtle racism” or “inherent racism” in any presentation of evidence for early contacts between Native Americans and others. This school of thought seemingly attaches less importance to evaluation of the evidence than to the supposed ideological consequences of interpretations. One view may be considered comfortable and therefore correct, while another is “disturbing” and therefore false. Members of this school take offense at claims of early contact, saying that such claims deny Native Americans the capacity to have developed their own civilizations without “help.” This position is so devoid of logic that investigators trained in scientific disciplines are incredulous when they first encounter it. Nevertheless, it is a stance that has become entrenched through repeated assertion.
David Kelley has commented on this phenomenon a number of times. His remarks in a Smithsonian publication (1995) may be the most insightful analysis to date. He pointed out how few anthropologists are aware of contributions by competent scholars from fields other than anthropology, and advises: “To gain some idea of the problems of appraising intercontinental relationships, one should study the comprehensive annotated bibliography of Sorenson and Raish (1990), the only bibliography I have ever sat down and read right through.” A second edition of Sorenson and Raish appeared in 1996, and both editions have been invaluable resources for me.
Support From Other Genetic Systems
Following each discussion of HLA alleles are comments about supporting data from other systems. For example, it is noted that five American populations with seemingly Afro-Asiatic HLAs also have genes in the immunoglobin, Duffy, Rhesus, and Kell systems that are rare outside of Africa. Other examples are taken from the transferrin, acid phosphatase, and adenylate-kinase compilations of Mourant and of CS. To my knowledge, these things have not been pointed out before in the context of ancient contacts between Native Americans and others. They may be only curiosities, but I would like to see a rigorous examination by specialists who are open to the idea of outside influences. A few more detailed comments on certain genes are appended, but a comprehensive review of the atypical genes is beyond the scope of this report.
Recent results from studies of mitochondrial DNA (mtDNA) are not considered here except for a brief summary in the appendix. Because only females transmit mtDNA, only the most prevalent types survive for long. Work so far has identified three basic female lineages with close relatives in northeast Asia (now designated A, C, D), one associated with southern Asia (B), and a possible fifth (X, proposed to be called E) that seems European. Results from both nuclear and mtDNA studies are stimulating new and controversial hypotheses too rapidly for confident inclusion in this report.
An exhaustive study of all American data would go far toward establishing a more reliable classification of indigenous groups, their order of entry, and migration paths. CS’s brief sketch (pp. 340-342) is an excellent start. My own studies have convinced me that remnants of the earliest colonists are concentrated in southernmost South America and that later arrivals, coming by sea, occupied the coasts first and then moved inland, especially up the South American rivers and the Mississippi.
Using genetic information by itself to prove assimilation of transoceanic voyagers is difficult, even when data from several different systems are in agreement. Because human distributions have changed with time, arguments based on the present situation are not convincing unless combined with other kinds of evidence, as I have tried to do in some of the endnotes. I am not competent to make a critical assessment of all of the cited material, although much of it has been scrutinized by specialists and found to be acceptable. Often there is not enough information to judge the validity of a given finding except in the general sense that it is consistent with other evidence considered important by the more experienced scholars.
Apparent Afro-Asiatic HLAs in America
Table 4 shows the occurrence, in American populations, of nine HLA types that currently display most of their highest frequencies in Southwest Asia, North Africa, or northern India. They account for 47% of the “non-Indian” HLA data in Native America, but these occurrences are concentrated in five populations of southwestern North America and three of the Southern Andes. Table 5 shows the highest world frequencies listed in CS’s Appendix 2. The following observations regarding HLA distributions are in decreasing order of importance in America of the relevant alleles. Numbers in parentheses are percentages contributed to the total “non-Indian” HLA content of 32 American samples.
Table 4. Frequencies (%) of “Afro-Asiatic” HLAs in American samples.
Sample B*21a A*32 A*30b B*14 A*1c B*17d A*29e B*37 B*18f C Amerind 7.5 6.5 6.5 Papago 12.5 0.7 0.7 1.6 Pima 9.4 1.3 0.3 Nahua 3.5 0.5 0.5 2.0 3.5 n.d. 1.5 2.5 Navajo 2.5 1.5 8.5 Cherokee 3.0 5.0 1.0 2.0 1.0 1.0 East. Maya 0.3 0.7 2.8 1.6 0.9 0.3 1.6 Mapuche 1.7 9.0 7.8 4.7 Atacama n.d. n.d. 10.7 7.1 Quechua 1.0 2.0 1.0 1.0 1.0 2.0 Araucano 0.5 n.d. n.d. 0.5 n.d. 0.5 Grd Eskimo 0.2 0.2 1.3 0.6 0.2 Oyampí 6.6 Parakana 0.5 0.5 Warao 3.4 0.5 Zuni 0.6 0.6 Caingáng
1.3
5.2
0.4
a Yupa, 1.0%.
b Guaraní ,1.5; Trio, 0.7%.
c Northwest Eskimo, 2.8; Aymara, 1.0; Inupik 1.0; Ticuna, 0.2%.
d Eastern Eskimo, 0.3%.
e Yupik, 1.0; Guaraní, 0.5%.
f Northwest Eskimo, 0.2%.
Table 5. Highest world frequencies (%) of Afro-Asiatic HLAs
(New World groups/composites underlined).
B*21
A*32
A*30
Saudi Arabia 22.2 Samoab 36.4 Samoab 26.0 Tigre 21.3 Tuareg 15.8 San 23.5 Jordan-Palest. 16.0 Honshuc 12.0 Central Bantu 21.3 Papago 12.5 Mapuche 9.0 Sardinia 16.3 Tuareg 12.1 Sardinia 8.3 Biakad 12.6 Berber 12.0 Punjab 6.8 Ibo 10.7 Mbutia 10.7 Oyampí 6.6 Basque 8.9 Iraq 9.5 C Amerind 6.5 Navajo 8.5 Pima 9.4 Khoi 6.5 Saudi Arabia 6.6 Turks 8.9 Greece 6.1 C Amerind 6.5 C Amerind 7.5 Brahman 5.9 Greece 5.4 Lebanon 6.8 Yugoslavia 5.8 Cherokee 5.0 Sardinia
6.7
Italy
4.9
North China
4.9
aPygmies of Zaire.
bSamoan outliers; absent elsewhere in Polynesia.
cChubu sample. Japanese samples are highly variable; the Kyushu sample displays 8%; overall, Japanese samples average only 2.7%, the Ainu 0.3%.
dPygmies of Cameroon.
Table 5, continued.
B*14 A*1 B*17 Atacama 10.7 Britishe 22±3 Koya (India) 25.5 Portugal 9.5 Jordan-Pales.f 19.0 Khoi 24.0 Sardinia 8.2 Berberg 18.0 Bantu 23.8 Mapuche 7.8 Scandinaviah 16±4 Ibo 17.8 Berber 7.0 Germanich 15±1 Dravidian 14.4 Iraq 6.2 Tunisia 15.0 Sardinia 13.5 Ireland 5.8 Punjab 15.0 Altaic comp. 11.4 Spain 5.5 Central India 14.5 Sherpa 10.9 Tigre 5.3 Czekoslovak. 14.5 Pygmies 10.8 Jordan-Palest. 5.0 Tibet 14.4 Tuva 10.0 Central Bantu 4.4 Belgium 13.9 Basque 9.0 France 4.3 Hungary 13.8 Berber 8.0 Tunisia 4.0 Spain 13.1 Malaysia 7.9
eAverage of four; see text. Irish highest (26.4%).
fAlso, Lebanon, 12.7; Iraq, 11.6: Iran, 8.8%.
gAlso, Tigre, 11.7; Tuareg, 7.2%.
hAverage of four; see text.
Table 5, continued.
A*29 B*37 B*18 Basque 14.2 Pygmies 5.4 Sardinia 28.2 Pygmies 11.2 Marathi 4.4 Basque 13.7 Spain 7.7 Iran 3.3 Sumatra 11.9 Lebanon 7.4 Sweden 2.1 Greece 11.6 Bantu 6.8 Mande 1.7 Italy 9.3 France 6.7 Nahua 1.5 Bali 8.9 Vietnam 6.0 Portugal 1.4 Berber 8.0 Uzbekistan 5.7 Belgium 1.4 Yugoslavia 7.8 Scotland 5.2 Russia 1.3 Lebanon 7.3 Caingáng 5.2 North China 1.2 Malaysia 7.3 Iraq 5.1 Greece 1.2 Melanesia 6.9 Portugal 5.0 Italy 1.2 Austria 6.8 Wales 4.9 England 1.2 Hungary 6.7
B*21 (10.4%). Old World occurrences of B*21 are concentrated in regions of strong Arab presence or influence. Frequencies of more than 15% are confined to populations in Saudi Arabia, Ethiopia (Tigre), and Jordan-Palestine, but influence extends across North Africa and into Spain, Portugal, and Italy (5-6%). In America, 84% of occurrences are clustered in four Uto-Aztecan populations (Papago, Pima, Nahua, and a Central Amerind composite).[9] The Papago have the fourth-highest frequency in the world, comparable to that of Tuaregs and Berbers.[10] CS’s Central Amerind composite sample is unique in that all of its “non-Indian” HLAs are of the Afro-Asiatic set (B*21, A*30, A*32). Thus, significant Afro-Asiatic contact with western Mexico and/or the Caribbean almost certainly occurred, probably from Arabia or North Africa.[11]
Traces of B*21 also appear proximate to the Pacific coast of South America (Mapuche, Araucano) and near Lake Maracaibo of Venezuela’s Caribbean littoral (Yupa) but not in CS’s fifteen other South American samples. The Mapuche sample has the highest total content of the Afro-Asiatic HLA alleles reported in America (13%; see Table 12). Much evidence of other kinds has been presented for an early North African presence in parts of South and Central America.[12]
A*30 (8.0%). The relatively rare A*30 allele is sparsely sampled, and CS have little to say about its peculiar distribution. I interpret it as a signature of a Caucasoid population that reentered sub-Saharan Africa at an early date, then reached America by voyages to the Caribbean and parts of South America, and also entered from Asia with the ancestral Na-Dene. Highest Eurasian values are found in Sardinia, Spain (Basques), Saudi Arabia, Greece, and northern China. But four of the globally highest levels appear in Africa, especially among the San (Bushman) of Botswana. This fits CS’s finding (pp. 175, 176) that the San have about 50% (Asiatic) Caucasoid genes. There are no data from the Caucasus or Southwest Asia that might help define A*30’s original range. CS found no A*30 data from North Africa, but computer interpolation indicated 6%. A*30 is absent from the Ainu, Eskimo, and Lapps, as well as from populations of Australia, New Guinea, and Pacific Oceania except for one exaggerated frequency found in Samoan outliers, possibly the result of one or a small number of early voyages.
The Navajo have the highest American frequency of A*30, presumably reflecting the “Dene-Caucasian” expansion postulated by Ruhlen and others[13] linking Athapaskans with Sino-Tibetan and other Dene-Caucasian languages. If this postulate is correct, A*30 should be present in many indigenous populations of western Canada, but at present no data are available.
Other significant North American occurrences are in the Central Amerind composite, the Cherokee, and the eastern Maya,[14] supporting the proposition that the Cherokee entered North America from the south. CS stated (p. 334, on the basis of their map on p. 249) that American A*30 peaks in the southeastern United States (where Cherokee is the only sample). I postulate that A*30 was carried across the Atlantic to the Caribbean and also to the Guaraní and Trio people near the mouths of the Plata and Amazon rivers, then up the rivers to the Quechua. It appears in only four of 14 South American samples, being strongest (3.4%) in the sample from the Warao of the Venezuelan coast. The Warao were skilled canoe voyagers (Wilbert 1977) and are Paez-speakers like the Timuca of Florida (Greenberg 1987). Granberry (1991) also found Timuca to be close to Warao but containing (Arawakan) Maipuran elements as well. He postulated that the Timuca migrated to the southeast directly by sea from the region of Puerto Hormiga, about 2000-1500 B.C. It would be valuable to know whether Indians of the southeastern United States other than the Cherokee possess the A*30 allele.
A*32 (7.0%). The A*32 allele seems to indicate a Mediterranean or specifically Aegean impact in the Caribbean region (including on the Cherokee) as well as on Tupians of the lower Amazon (Oyampí and Parakana). It seems to connect this set and the Central Amerind composite with northern India, Sardinia, the Tuareg of Algeria, and with populations around the Adriatic Sea in Greece, Yugoslavia, and Italy.[15] A*32 is absent from other South American samples except the Mapuche. Carriers of A*32 seem also to have reached parts of Japan, where isolated spikes appear in contrast to the low Japanese average.[16] An unexpected link between East Asia and India has been discovered recently through studies of HLA*2 subtypes (Narinder Mehra, reported in Anonymous 1997) and also through virology. Miura, et al. (1994) have identified two varieties of human T-lymphotropic virus type HTLV-I that connect India and Japan, and one of these has also been transmitted to Colombia and the Caribbean.[17] Migrations that could account for these data could also explain why dolmens and other “megalithic” phenomena, present in India by 1000 B.C., then appeared in Korea, Japan and Colombia, about 600 B.C. (described by Joussaume 1985:267-94; see also map in Scarre 1988:35).
A*32 levels in the Mapuche, Oyampí, and the Central Amerind composite samples are among the top nine in the CS tabulation (7-9%), and the Tupian Oyampí near the mouth of the Amazon River have the second highest American frequency. If this is not an artifact of sampling, it implies connection of Tupians with the Mapuche by river commerce.[18] Unfortunately there are no data from the Atacama or Araucano, whose histories are intertwined with that of the Mapuche (Araucano subsumes Mapuche).
Both A*32 and A*30 are found at significant levels in Greece, Sardinia, and in the Central Amerind composite. They also appear at anomalously high frequencies in Samoan outliers but are not documented elsewhere in Pacific islands. This may reflect limited exploration of the Pacific by Mediterraneans who otherwise left few traces except (controversial) petroglyphs. The strong presence of A*32 but not A*30 among the Mapuche and Oyampí indicates a different source, perhaps India, for A*32 in the Amazon region.
B*14 (7.0%). The B*14 allele appears to link the western Mediterranean with the Atacama, Mapuche, and Quechua. CS describe it as peaking in the Middle East, Sardinia, and southern Spain, but it actually reaches its highest world frequency (11%) in the Atacama population of Chile. The Mapuche sample, with 8%, is fourth, after those from Portugal and Sardinia. Other high frequencies occur in North Africa, Iraq, and Ireland.
The presence of B*14 in the Andes might be attributed to recent Iberian influence, except that if this were the case it should be widely common through South America. In fact, outside of the Andean cluster it has been reported only in a Caingáng sample near the mouth of the Río de la Plata. It seems likely to me that B*14 was carried to its present locations by a more ancient population, with roots in the Near East.[19] It appears also in eastern Maya, Nahua, and Cherokee populations; but unlike A*32, B*14 was not found in the Papago, Zuni, Navajo, Central Amerind composite, Oyampí, or Parakana samples.
Most of the “non-Indian” HLA content of the Atacama and Mapuche samples comes from B*14 and A*1, both of which are important in North Africa.[20] These two alleles have similar distributions in America and are likely to have been brought by the same people. The Atacama also have an 8% frequency of the African FY*0 allele of the Duffy system, which is found as well among the Ge-speaking Cayapó of Brazil.[21]
A*1 (6.5%). CS refer to the A*1 antigen as “typically European” (p. 288); but, like B*14, it is equally important in North Africa and in the Near East. A*1 reaches high frequencies in the Caucasoid regions of the world, but is rare in East Asia, America, and Oceania. Both A*1 and B*14 display high frequencies among Berbers, Tunisians, the Tigre, and populations of Jordan, Palestine, Spain, and Ireland, and both reach significant levels in the Andean cluster. A*1 has apparently not been reported elsewhere in South America.
Inexplicably, CS say that A*1 is absent in America (p. 369), despite its presence in six of their Andean listings as well as in six others in the hemisphere. As in the case of B*14, the highest American levels of A*1 were found among the Atacama and Mapuche (7%, 5%), with lower frequencies from the nearby Quechua, Aymara, and Araucano samples, and also among the Ticuna. If this were a consequence of recent contact with the Spanish or Portuguese—who carry 12-13% of A*1—it would be difficult to explain the absence of A*1 in Brazil. I suspect that its Andean distribution is due to an older, unrecognized contact with the Near East.[22]
In North America, the highest level of A*1 (2.8%) came from a sample of northwestern Canadian Eskimos. The Inupik of Alaska had 1% as well, but the allele was not found among eastern Canadian Eskimos. One possibility is that A*1 (but not B*14) was brought to northwestern America by western Asians related to the later Scythians, Tocharians, or Celts.[23] A*1 is probably scattered throughout indigenous populations of the North American northwest who have yet to be tested. The Greenland Eskimos have traces of several unexpected HLAs, including 1.3% of A*1, that probably came from later contacts with Europeans.
B*17 (2.6%). The highest frequencies of B*17 occur in Africa and India, in an apparently ancient pattern. Low levels were found in ten American samples, but more than half was in two Uto-Aztecan populations (Nahua and Papago). Traces in the Orinoco delta (Warao) and in the lower Amazon (Tupian Parakana) again suggest trans-Atlantic contact.
A*29 (2.0%). A*29 is spread thinly over Southwest Asia, Africa, and Western Europe, with highest current frequencies among Basques and certain African pygmies. In America, A*29 has been reported only in southern Brazil (Caingáng 5%, Guaraní 0.5%) and for the Cherokee and Yupa samples (1% each).
B*18 (1.9%). B*18 appears to be an ancient Caucasoid antigen, linking Basques, Berbers, Sardinians, Greeks, and Southern Europeans. It also went along the Asian coast, especially to Indonesia, then apparently to Ecuador and Mexico. The overall distribution suggests involvement of Mediterranean seafarers. In America, it appears above the 1% level only among the Nahua, Quechua, and eastern Maya, with traces among the Araucano, northwest Canadian Eskimos, and the Greenland Eskimo. I suggest that it came to the Pacific coast by way of Indonesia.[24]
B*37 (0.6%). The B*37 allele is scattered worldwide at low frequencies, with highest present levels being found in Africa and western Asia. In the Americas, B*37 is found only in the Uto-Aztecan Nahua and Pima samples, and for the Caingáng of Brazil, supporting the postulate of contacts from Africa, India, or Indonesia. It is somewhat surprising to find that the Nahua sample had the sixth-highest world frequency of this relatively rare antigen.