BLOOD GROUPS O F BR,AZILIAN INDIANS A. M. P A N T I N AND P. C. JUNQUEIRA Department of Pathology, Cambridge University, and Znstituto Oswaldo Cruz, Rio de Janeiro This paper reports a study of the blood-groups of 73 Brazilian Indians. The region in which the blood samples were collected is that bounded by the rivers Batovi, Ronuro, Kurisevu and Kuluene, which are tributaries of the Xing6 in the Matto Grosso. The4ribes of this region are the following: Bakairi, Kalapalo, Nahukwi, Kuikuru and Matipuhy of the Cariba family; W a u r i and Mehinaku of the Aruaque family; Auety and Kamaiur6 of the Tupi family; and the isolated group Trumai. Of these tribes, only two were accessible enough to provide blood samples which could reach the laboratory in suitable condition for testing. These were the Kamaiuras living near Camp Jacar6 of the Fundacjio Brasil-Central, and the Kalapalos living in a fishing village near Camp Kuluene on the Kuluene river. From Camp Jacar6 we collected 35 blood samples, comprising 32 KamaiurA (Tupi), 1 Trumai, 1 W a u r i (Aruaque) and * This work was made possible by the generosity of the Brazilian Government, through Dr. Olympio d a Fonseca, Director of the Instituto Oswaldo Cruz; Dr. Walter Oswaldo CNZ, Director of the Department of Haematology of that Institute; Dr. Modesto Donatini Dias d a Cruz, Director of the ServiGo de ProteGLo dos Indios; General Borges Fortes, Presidente da FundaQLo Brasil-Central; Orlando VilasbBas, Chefe d a ExpediqHo Xingd-Roncador, and Dr. Pedro E. Lima. To all these, and t o the Brazilian Air Force which supplied us with air-transport, we offer sincere thanks. We are especially indebted to Dr. A. E. Mourant, Director of the Blood Group Reference Laboratory, the Lister Institute, who not only provided the very valuable sera for our tests but calculated the gene frequencies from our results, and whom we would gratefully thank for his help. 395 396 A . M . PANTIN A N D P. C . JUNQUEIRA 1 Sui& (GB). The first tests on these samples were set up in Rio d e Janeiro 28 hours after bleeding, and completed two days later. From Camp Kuluene, 104 samples were taken. Storms delayed our air-transport, and only 38 samples reached Rio in first-class condition for tests. These included 24 Kalapalos, 3 Nahukw&, 2 Matipuhy, and interbreeds as follows: 4 Kalapalo/Kuikuru, 2 Kalapalo/Nahukwb, 1 Kalapalo/Matipuhy, 1 Nahukwb/Matipuhy, and 1 Nahukwb/Kuikuru. All belong to the Cariba family. One sample only was from a cross between different families, viz. Kalapalo (Cariba) /Mehinaku (Aruaque). This second batch of tests was set up three days after bleeding and finished two days later. MATERIAL AND METHODS 1. CeZEs. These Indians would not permit venepuncture, so finger-puncture was used. Blood from deep finger-stabs was allowed to drop into dry sterile tubes which were closed with sterile rubber bungs and placed upright in sealed tins immersed in cool water (about 10.C.) in thermos containers. These containers were transported by air from the Xingfi t o the laboratory in Rio de Janeiro. F o r tests, only bright firm blood-clots in clear colorless serum were used. All' clots not absolutely first-class were discarded. Each clot was shaken up in a little sterile 0.9% saline, the loose cells removed and washed, and an approximately 2% cell suspension, as judged by eye, reconstituted in sterile saline. 2. Sera. The sera used in these experiments were supplied by the kindness of Dr. A. E. Mourant, of the Blood Group Reference Laboratory, the Lister Institute. They were freezedried, and only made up as needed in sterile distilled water. They were specific for the antigens A, A,, B, C, c, D, E, e, M, N, S, P, Lutheran, Kell, and D~iffy(FY). All except anti-hll and anti-N, which were animal sera, were human sera of high BLOOD GROUPS O F BR.AZILIAN I N D I A N S 397 titer, all except anti-Kell were absorbed to remove the natural agglutinins, and all except anti-Kell and anti-Fya were saline-aggl-utinating. Tests for Kell and Fy" were done by the Coombs test (adding rabbit anti-human-globulin serum to the sensitized cells) and by the direct action of the testing serum on cells suspended in 20% human albumin. All the sera were used undiluted, except anti-A and anti-B, which were diluted to half-strength, lest these very powerful sera should lyse the cells. 3. Methods. The small-volume technique devised by Race was employed throughout, whereby equal volumes (approuimately 0.008 ml) of serum and of 2% cell-suspension in saline are delivered from a marked Pasteur pipette, under direct vision, into clean d r y precipitin tubes and kept for the necessary time at the appropriate temperature. Tests for A, A,, and B were kept for two hours at room temperature. Tests for Rh antigens were incubated for 90 minutes at 37"C., and never centrifuged. Larger volumes were used in the tests f o r M and N. These, and tests for S, P and Lutheran were kept f o r two hours in a refrigerator at approximately 18°C. F o r Kell. and Fy", 1 volume of 10% cell-suspension was incubated with 2 volumes of serum for 30 minutes at 37"C., and the cells, after 3 washings in saline, exposed to rabbit anti-human globulin Anti-Kell was also serum (titer $3, no zoning; used a t -io-). used in equal volumes with a 2% suspension of cellms in 20% human albumin. When the test was ready, the contents of each tube were placed on a slide with minimal disturbance and read under the microscope (Q" objective). Controls of known cells were set u p with each test. All the results were viewed by one person. Further comments on the tests will be made under the separate headings. The gene-frequencies were calcul-ated from our data by Dr. A. E. Mourant, using his method which is described in his papers on Basques and Indians (Chalmers et al., '49; Prasad et al., '49). 398 A . M . PANTIN AND P. C. JUNQUEIRA RESULTS The ABO groups The cells only, aiid not the serum, of the blood samples were examined. All these Brazilian Indians belonged to Group 0. Of 27 White controls, the ABO groups were : GROUP N O . OBSERVED FREQUESCY 4 4 0.2962 1 1 0.0741 3 14 0.1111 0.5186 37 1.0000 The “ Rhesus” Blood groups The sera used in these tests were of high titer and absolute specificity, aiid were used undiluted; the resu1,ts were completely definite in all cases. An acute shortage of anti-e made it impossible to test all, or many, of the bloods for this antigen, and such tests as were made were disregarded in calculating chromosome frequencies. Apparently little information is thus lost. The phenotype ccddee ( r r ) was not observed in these Indians. RE.4CTIONY anti C D E c e ++++ ++-++-+ -++++++-+-+ XO. OBSERVED I,HENOTYPE COMPIOXEST GEPI’OTYPE CeDE CCDee CeDee ccDEE CCDE ccDce Ri Rz 17 R, Ri Ri R o R3 R, 11 Ri R‘ R, R, Tupi TOTAL 32 23 6 0 4 15 11 3 6 0 0 - - 0 - 38 35 73 0 9 6 4 399 BLOOD GROUPS O F BRAZILIAN INDIANS ~ c~~~~~~ NO. OBSEBVED CcDE CCDee CcDee ecDEE CCDE ecDee NO. FREQUENCY BXPECTED OBSERVED 28.14 25.06 5.63 9.30 4.56 0.32 73.00 ~~~ FREQUENCY EXPECTED 0.4384 0.3014 0.1233 0.0822 0.0548 0.0000 0.3855 0.3433 0.0770 0.1275 0.0624 0.0043 1.0001 1.0000 Chromosome frequencies: CDe cDE eDe CDE % R, R. R. 0.5859 0.2973 0.0657 0.0510 0.9999 Of 27 White control bloods, the Rh groups were as follows: PHENOTYPE COXNONEST GENOTYPE NO. OBSERVED ~~ 6 4 1 6 4 4 2 CCDee CcDE ccddEe ceDE ccddee CeDee eeDee 27 The M N S groups The interpretation of the MNS results presents a problem. The anti-hi and anti-N sera used in these experiments were animal sera prepared at the Blood Group Reference Laboratory a t the Lister Institute. Two different sera of the same specificity were used for every test, and these same sera had been in routine use at the Lister with satisfactory results against controls. In addition, these same sera gave the following results with 27 White European bloods used as controls in Brazil: 11MN, 9 M, and 7 N, a distribution very close to the expected; and these results agreed with tests made on the 400 A . M . PANTIN A N D P. C. JUNQUEIRA same control bloods with commercial sera. The physical conditions of the tests were rigidly controlled. Nonetheless, the Brazilian Indian bloods gave a number of N-positives which were neither in accord with previous findings of other workers, nor statistically likely, and this feature recurred in the first 19 of a series of North American Indian bloods (Dieguefios) tested with the same sera. When a different pair of anti-N sera was employed to test the remaining Diegueiios, the numbers of N-positives dropped. On the other hand, a re-test of N-negative Brazilian Indians with the original sera still gave a negative resu1.t. There a r e therefore three possibilities : 1. That the first pair of anti-N sera, which a r e notoriously difficult to make, contained non-specific agglutinins and gave a number of false positives. 2. That these sera contained in addition to anti-N an agglutinin against some unknown blood antigen which was picked up in different proportions of different populmations. 3. That the high incidence of N in the peoples tested is real. I n view of the reaction with the controls, and the fact that MM Indian bloods still gave a negative reaction on re-test with the doubtful anti-N sera, this is at least a possibility. Without further experiment, it is impossible to determine which is the right answer, and the matter must be left there for the present. The anti-M serum, and the human anti-S serum gave excellently clear-cut results. Indians: MNS group PHENOTYPE NO. OBSERVED MS Mas MNS MNss NS NSS 8 6 35 34 0 1 74 ' One additional sample was tested f o r MNS, making 74. 401 BLOOD GROUPS O F BRAZILIAN I N D I A N S Probably only one N, the rest being MN NO. OBSERVED M+ M- FREQUENCY 73 0.9865 0.0135 1 - - 74 1.0000 + M. GENE FREQUENCIES M N 0.8837 0.1163 1.0000 The evidence is insufficient t o show how S genes are distributed between M and N. Results of the White control tests were : PHENOTYPE NO. OBSERVED FREQUENCY MS 4 MI38 MNS MNss 5 6 5 NS 1 Nss -6 0.1482 0.1852 0.2222 0.1852 0.0370 0.2222 Total: 27 1.0000 Th.e Lu t hmwn antigen This antigen was discovered in 1946 by Callender and Race ('as), who found its antibody in the serum of a man who had been many times transfused. This antibody has since been found, always a s an immune response, on 5 further occasions. The Lu" antigen is inherited a s a Mendelian dominant character (Callender and Race, '46) and the phenotypes Lua+ and Lua- can be recognized by the anti-Lutheran serum. Race and Sanger ('50) give the incidence of the Lutheran group in Engl'ish blood as follows : Total no. LU.+ 105 (7.65%) 1373 Lu'1268 (92.35%) And the gene frequencies : Lub = 0.9610 and Lila = 0.0390 1.0000 Mainwaring ancl Pickles ( '48) describe strongly and weakly reacting types of cells, and deduce the existence of three allelomorphic genes determining the nature of the Lu" substance. 402 A . M . P A N T I N A N D P. C. JUNQUEIRA I n the Brazilian Indians in the present study, 3 of the 12 positive results were weak, 9 strong. Indians : PHENOTYPE NO. OBSERVED FREQUENCY LU.+ Lu'- 12 61 0.1644 - 0.8356 73 1.0000 Gene frequencies : Lub = 0.9141 Lu' = 0.0859 1.0000 Controls : NO. OBSERVED LU*+ Lu'- FREQUENCY 2 0.0740 0.9260 37 25 1.0000 The P untigeta A single human anti-P serum was used, and the test carried out at temperatures between 10°C. and 18°C. Positive results graded from large coarse agglutinates comprising all cells to weak agglutinates of a few cells among many unagglutinated ones. Four categories were established (#, [ and weak) and a positive was recorded if any signs of general agglutination were present. But it was not always easy to interpret the results with this serum, as the border-line between a very weak positive and a negative was difficult to determine. +, +] ~ PHENOTYPES NO. ORSERVED FREQUENCY P+ P- 30 43 0.4110 0.5890 73 1.0000 ~ Indians : Gene frequencies: ~- p = 0.7675 P = 0.2335 Controls: PHENOTYPES N O . OBBERVED EREQUENCY P+ 14 13 0.5185 0.4915 P- I 27 1.0000 403 BLOOD GROUPS OF B R A Z I L I A N I N D I A N S T h e Kell antige9z I n 1946, the Kell antigen was identified by Coombs, Mourant and Race. Anti-Kell, in the “incomplete” form, had developed in the serum of a Kell-negative mother whose child was thought to have haemolytic disease. The Kell antigen was found to be inherited as a Mendelian dominant character. Race and Sanger ( ’50) give the frequency of the Kell antigen in Caucasian bloods (White American and English) as f OllO\VS : Kell+ Kell- NO. OBRBRVED FREQUENCY 43 380 0.1017 0.8983 423 1.0000 - and the gene frequencies k = 0.9478 K = 0.0522 Dunsford (’49), testing 566 Group-0 bloods with a different anti-Kell serum, found the percentage of K positives to be 7.24, while Bertinshaw e t al. ( ’50), using the same serum on a further random sample of 475 bloods, found 6.95% were E positive. The antibody in the anti-Kell serum used to test these Brazilian Indian bloods was “incomplete,” and the serum, Group 0, was not absorbed; as the Indians were all Group 0, this was immaterial. PHEIOTYPES h0. OBBERVED FEEQUENCY K’ K- 17 56 0.2329 0.7671 73 1.0000 Indians : Gene frequencies: k = 0.8759 = 0.1241 K 1.0000 Controls : PHEPFOTYPES NO. OBSERVED FREQUENCY K+ 2 25 0.0740 0.9260 27 1.0000 K- - 404 A . M . PANTIN A N D P. C. JUNQUEIRA The "D26fy" antigen, Fy" The most interesting thing about these Indian bloods is that they all appear to lack the antigen Fy". The example of anti-Fy" used t o test these Indian bloods did not agglutinate cells in saline, and the reaction on Fy-positive cells in 20% albumin was either extremely weak o r negative. Each test was therefore done by the Coombs method. All 73 Brazilian Indian bloods gave negative results. The same serum used by the same method gave strong positives with 8 out of 16 White control, bloods (Portuguese, English, French and German). The blood-group antigen Duffy, o r Fy", was discovered by Cutbush, Mollison, and Parkin ( '50), who found the corresponding antibody in the blood of a man immunized by transfusion. The most complete published figures, those of Race and Sanger ( '50),show that Fy" is present in the blood of 65% of English people. Cutbush and Mollison ('50) have found it in 91% of persons from Northern Pakistan; Miller, Rosenfield and Vogel ( '51) in 99% of Chinese and 26% of American Negroes ; and Hartmann and Brendemoen in 97% of Norwegian Lapps. It is a n interesting fact, which may prove to be of some anthropological importance, that the antigen Fy" has not been found in this group of Brazihian Indians. SUMMARY Blood from 73 pure-bred Brazilian Indians from the hlatto Grosso (Kalapalos and Kamaiuras) has been examined for the antigens A, A,, B, C, c, D, E, M, N, S, P, Lutheran, Kell, and DUffY (FYI. All these Indians belonged to group 0. All carried the antigen D. The phenotype ccddee (rr) was not observed. An unexpectedly high number of N-positives was found. Whether this finding was real or due to non-specific agglutination is discussed. Of these Indians, 58.9% were P-negative. The most interesting fact was that all the Indians tested were Duffynegative. This fact may prove to be of anthropological8importance. a Personal cominunication. BLOOD GROUPS O F BRAZILIAN INDIANS 405 L I T E R A T U R E CITED BERTINSHAW, DOREEN,S. D. LAWLER,H. A. HOLT, B. H. KIRMANAND R. R. RACE 1950 The combination of blood-groups in a sample of 475 people i n a London hospital. Ann. Eugen., 15: 234-242. CALLENDER,SHEILA T., AND R. R. RACE 1946 A serological and genetical study of multiple antibodies formed in response t o blood transfusion by a patient with lupus erythematosus diffusus. Ann. Eugen., 13' : 102-117. CHALMERS, J. N. M., E. W. IKINAND A. E. MOURANT 1949 The ABO, MN, and Rh blood groups of the Basque people. Am. J. Phys. Anthrop., n.8. 7: 529-544. COOMBS,R. R. A., A. E. MOURANTAND R. R. RACE 1946 In-vivo isosensitization of red cells in babies with haemolytic disease. Lancet., i: 264-266. CUTBUSH,MARIE, AND P. L. MOLLISON 1950 The Duffy blood group system. Heredity, 4: 383-389. CUTBUSH,MARIE,P . L. MOLLISONAND D. M. PARKIN1950 A new human blood group. Nature, 165: 188. DUNSFORD,I. 1949 A saline-agglutinating Kell antibody. Nature, 163 : 962. MAINWARINQ, U. R., AND M. M. PICKLES1948 A further case of anti-Lutheran immunisation, with some studies on its capacity f o r human sensitization. J. Clin. Path., 1 : 292-294. MILLER,E. B., R. E. ROSENFIELD AND P. VOQEL 1951 On the incidence of some of the new blood agglutinogens in Chinese and Negroes. Am. J. Phys. Anthrop., n.s. 9 : 115-126. PRASAD, C. H., E. W. IKINA N D A. E. MOURANT 1949 The Rh and M N S blood groups of some students from India. Am. J. Phys. Anthrop., n.s. 7: 553-558. RACE,R. R., AND R. SANQEE1950 Blood Groups in Man. Blackwell Scientific Publications, Oxford. POLYNESIAN-AMERIND RELATIONSHIP.The blood grouping figures favor neither Melanesia nor Micronesia [as the probable path of the ancestors of the Polynesian]. If no other facts were known they would almost compel one to look for another alternative. And the only obvious alternative, the very long landless ocean path from America, gains strong support from blood gene frequencies. The serological similarity of these two groups which is not unsupported by likenesses in other criteria suggests that the Polynesian 406 AMERICAN ASSOCIATION O F P H Y S I C A L ANTHROPOLOGISTS and the Amerind have a coinmon dominant. component. Further, this component, perhaps somewhat differentiated during centuries in the widely diverse environments of the Americas, could have reached Oceania a t any time with the aid of prevailing ocean currents. That ocean currents have had a major influence on the migratory journeyings of the Polynesian has gained much support in recent years following the successful enterprise of Heyerdahl and the involuntary voyages of two other groups quoted by Avias, 1949. These three remarkable ocean voyages took place within a space of 12 months. How many similar sailings have taken place over the centuries we have no idea, but we know that long ocean voyages taking advantage of ocean currents and prevailing winds were well within the capabilities of the seafaring early Polynesian. Avias ( ’49) has suggested that the Polynesian has resulted from admixture of an Amerind component with a proto-Melanesian group which had arived in Polynesia before the Amerinds. H e described the latter group as a negrito-ainoid complex with possibly a dash of Papuan. Our results suggest that the proto-Melanesian component must be weak in Polynesia, as it has left little serological evidence of its presence in the Maoris. Nevertheless there appears to be much to indicate that another race reached some of the islands of Polynesia. This group, though perhaps outnumbered by the “Amerinds,” may well have been responsible for some of the similarities of culture and language that have been quoted in support of an eastward course for the Polynesian. It is also likely that since his original settlement in the central Pacific the Polynesian has, on occasion, driven westward into both Melanesian and Micronesian territories and possibly even into Indonesia. In conclusion it is submitted that the serological evidence presented in this paper strongly supports a Polynesian-Amerind relationship, making it probable that the islands of Polynesia have been settled largely by migrations f rom continental America. Prevailing ocean currents and other factors suggest the coast of Peru as the starting point of such migrations.- J. J. Graydon. Blood groups and the Polynesians. lIankind, vol. 4, no. 8, March, 1952, pp. 329-339.