BLOOD GROUPINGS AND EACIAL CLASSIFICATION R. RUGGLES GATES King's College, Universiiy of London, England The blood groups are simply an additional means of comparing races. They are definite and fixed units, whose method of inheritance is known and whose descent can be traced. As indices of racial relationships they have a great advantage over physical measurements ; for the latter are not only quantitative and subject to fluctuations, but very little is known of the manner of their inheritance. The relation of the cephalic index or the nasal index to genic differepces is largely a matter of conjecture, and much more needs to be known of the inheritance of head shapes before any very definite meaning can be attached to differences in cranial measurements. The blood groups appear to be typical genic mutations, and as such must follow the general laws of mutations. The use of the blood groups in connection with racial differentiation and relationship must then be based upon the theory of mutations, and I have shown elsewhere (Gates, '36 a) that the theory of mutation frequencies can be developed further in connection with the study of the blood groups and their racial distribution. Like other mutations, the blood groups must arise repeatedly, and the evidence appears to be best interpreted on the hypothesis that A and B have arisen as dominant mutations from 0, which would then represent the primitive condition. If this is true, we must assume that A and B have arisen in the same way in anthropoids, as parallel mutations. There is no difficulty about this, because it is now recognized that l P a p e r read before the Section of Anthropology at the Oalcutta meeting of the Indian Science Congress, January, 1938, a8 a delegate from the British Asaociation and the American Association f o r the Advancement of Science. 385 A X E R I C A N J O U R N A L O F PHYSICAL ANTHROPOLOOY, VOL. XXIV. S O . JANUARY-MARCH, 1939 3 386 R. RUGGLES GATES parallel mutations in related species or groups are of frequent occurrence, and it seems evident (Gates, '36 b) that they have played a more important role in evolution than has yet been recognized. As regards anthropoids, the fact that of seventysix chimpanzees tested seventy-one were A (= 93.4%), only five 0 and none B, indicates that the distribution differs in important respects from what has been observed in man, although pure Blackfeet Indians showed 76.5% A. Not only do particular mutations recur with a particular frequency which is characteristic, the frequency being much higher with some mutations than with others; but we must also assume that at times mutation rates change, and that new mutations begin to appear, for otherwise there would be no evolution, or at any rate no differentiation. We have a t present little d e h i t e evidence regarding changes in mutation rate, although it is necessary to assume that such changes occur. The Amerindians of Darien, who produce a high percentage of albinoes in every generation, appear to be most reasonably interpreted as a case of a greatly increased mutation rate; and in the Blackfeet and related tribes of Indians in Western Canada, which have a very high percentage of the A blood group (hlatson and Schrader, '33)' this condition may have been produced by a great increase in the A mutation rate. The evidence suggests that the frequency of the A and B mutations may be of the order of perhaps 1in 100,000. I n comparing the blood groups of different peoples and explaining the differences which arise between types occupying different areas, it is necessary to take into account not only migration and interracial crossing but also the occurrence of fresh mutations with a rate which is probably characteristic of each racial type over considerable periods. The blood groups appear to be quite non-selective in themselves, although they might conceivably be genetically linked to some racial character or condition which is selective. The distribution of A, B and 0, however, in various races leads one to conclude that they are probably entirely non-adaptive in nature. If this is the case, the rate of spread of the A or B will depend BLOOD OROUPINQS AND RACIAL CLASSIFICATION 387 entirely on their mutation frequency. As pointed out earlier (Gates, '36 a), the B mutation seems to have begun to appear at a much later period than the A, and A is therefore spread much more widely through the primitive races surviving to the present time. The statistics of blood grouping make it probable that the B mutation has never appeared in such peoples as the Australian aborigines, the Bushmen of South Africa and the Basques of Europe, which have a high percentage of A while their very low percentage of B has probably come in through crossing. The American Indians have long been regarded as having aboriginally only the 0 blood group. But more recent results indicate that certain tribes in Western Canada are much higher in A than are Europeans. They can therefore not have derived their A from such a source. These tribes form a group in Western Canada east of the Rocky Mountains, including the Blackfeet and Bloods, the Beavers and other tribes on the Peace River, the Slaves and Dogribs in the southern part of the Mackenzie basin. F o r this purpose I am permitted to quote some unpublished results of Dr. J. C. B. Grant, as well as published results of his and Matson's and mine. Surrounding this area high in A (40 to 76%) are the Stony (Assiniboine) Indians with 30.8% A, the Chipewyan and Yellowknives with about 24% ; and the Crees on the East and Loucheux on the North with apparently no original A. On the agreed hypothesis that America was originally populated from across Behring Strait in post-glacial times, the fact that the Indians of pure blood showed neither the A nor the B blood group was somewhat anomalous. From the recent results we may assume either that a center of the A mutations arose in this part of Western Canada, or that the tribes with 40% or more of A are descendants from an original people migrating from Asia and already having, like many other primitive peoples, a high percentage of A. The latter hypothesis is perhaps less probable, because the more northern tribes such as the Loucheux, and the British Columbia coastal Indians (Gates and Darby, '34) which are in the path of a migration from Asia, both appear to be originally without A. 388 R. RUGGLES GATES Another difficulty in accounting for the high percentage of A is that the Blackfeet are Algonkian Indians while the Beavers, Slaves and Dogribs are Athapascans, and the Stonies are classed with Siouan tribes. The latter are "to some extent amalgamated with Cree" (J.C. B. Grant in litt.), which would reduce their percentage of A. Jenness ('37) has recently shown good reason for concluding that the Algonkian tribes were the earliest arrivals in America. Among these people he includes the Micmacs, who (Gates, unpubl.) were probably 0 before crossing wit.h Europeans began. The Athapaskans, if they came later, might have brought the A blood group with them. But there are difficulties with this view, as already pointed out, and the fact that the Blackfeet and Bloods, which are high in A, belong to the Algonkians, favors the conclusion that a fresh center of the A mutation arose on the American continent. Other recent evidence bearing on this question is supplied by the examination of mummies and skeletons excavated in America. Matson ('36) examined sixteen such mummies by absorption methods and concluded that they were all 0, while Boyd and Boyd ( '37) find among 196 mummies 2% A, 3.6% B and 1% AB. The latter results are so improbable on general grounds that these data can only be accepted when the accuracy of the method has been confirmed. Candela ('37) has recently examined six excavated early Indian skeletons and concluded that five of them were A and one 0. If these tests are reliable they would of course show that a high percentage of A existed in the early aboriginal population of America, but we must wait for further evidence. The great number of racial types in India and adjacent lands makes this a fertile area for the study of blood-grouping, and already results of much interest have been obtained. The Paniyans of the Wynaad Plateau have 62.4% of A (Aiyappan, '36), which confirms the relationship of these pre-Dravidian people to the Australian aborigines and contrasts strongly with the 26.7% of A in the adjacent Maharattas of Goa. The Tibetans of Gyantse (Gates, '36c) on the eastern border of BLOOD GROUPINGS AND RACIAL CLASSIFICATION 389 Tibet are extraordinarily low in 0 and high in AB, a result which may be accounted for by crossing of an A people with the Chinese high in B. Work which is still in progress in Assam indicates that the various racial types of the Naga Hills can be distinguished by their blood group ratios. The results so f a r obtained at Kohima by Doctor Vieyra and at Mokokchung by Doctor Kundu under the general direction of Mr. J. P. Mills, appears to support the latter’s conceptions of the relations between these native peoples. Thus he concludes on general grounds (in litt.) that the Konyaks are ‘older,’ i.e., more primitive, than the Angami, and this is confirmed by the higher percentage of 0 and A among the Konyaks and their lower percentage of B. Again, he concludes that the Konyaks and Aos have much in common, and the fifty-seven Aos hitherto bloodgrouped indicate considerable similarity although they appear to be somewhat lower in A and higher in B. That the Thado Kuki are an entirely different race is confirmed by their extremely low frequency of 0 and the fact that they are highest of all in B and AB. Finally, I should like to point out that no country in the world is likely to surpass India in the interest of its blood group results. This being the case, it is desirable that the work of blood grouping be organized in different parts of the country and that a concerted effort be made to obtain from each racial type a sufficient number of blood tests to constitute a statistically significant sample of each population. This is a work in which anthropologists can cooperate with medical workers and others in making the necessary tests, but it is all-important that the results be of the utmost accuracy and that every possible source of error be eliminated. 390 R. RUGQLES GATES LITERATURE CITED AXYAPPAN, A. 1936 Blood-groups of the pre-Dravidians of the Wynaad Plateau, India. Man, XXXVI, 191-194. BOYD,WILLIAM C., AND L. G. BOYD 1937 Blood grouping tests on 300 mummies with notes on the precipitin test. J. Immunology, XXXII, 307-319. CAXDELA, P. B. 1937 Blood group determination upon Minnesota and New Pork skeletal material. Am. J. Phys. Anthrop., XXIII, 71-78. GAmS, R. Rnaoms 1936a Recent progress in blood group investigations. Genetica, XVIII, 4745. 1936 b Mutations and natural selection. Am. Naturalist, LXX, 505-516. 1936c Tibetan blood groups. Man, XXXVI, 110-111. 1939 The blood groups and features of the Micmac Indians. J. Roy. Anthrop. Inst. ( I n press.) GATES,R. R., AND G. E. DARBY 1934 Blood groups and physiognomy of British Columbia Coastal Indians. J. Roy. Anthrop. Inst., LXIV, 2344. JENNESS, D. 1937 Tha Indian background of Canadian history. Canada Dept. of Mines and Resources (Nat. Museum of Canada), Bull. 86 (Anthrop. series no. 21), p. 46. MATSON,G. A. 1936 Blood-grouping of mummies. J. Imrnunol., XXX, 459-470. 1933 Blood grouping among the ‘BlackMATSON, 0. A., AND H. F. SCHRADER feet’ and ‘Blood’ tribes of American Indiana. J. Jmmunology, XXV, 155-163.