The aim of the study was to test hypotheses of the origin of the modern population of the Ryukyu Islands. For this purpose, data from two systems of anthropological features - craniometry and cranioscopy-were integrated using multidimensional statistics. The results obtained indicate the Mestizo origin of the population of the islands. It seems that it was formed on the basis of at least three different groups of the population. The first two of them - carriers of the Jomon and Yayoi cultures - are also ancestral to modern Japanese. The pronounced anthropological peculiarity of the Ryukyush people may be due to the presence of a southern Mongoloid admixture.

Keywords: Japanese Archipelago, Ryukyu Islands, integration of craniometric and cranioscopic features.

Introduction

The Ryukyu Archipelago is a chain of small islands stretching from northeast to southwest of Kyushu Island. Until the beginning of the 17th century, an independent state existed here, which was finally annexed to Japan only in 1879. The long period of autonomous development has led to a significant peculiarity of the material and spiritual culture of the indigenous population of the Ryukyu Archipelago; many features, despite the strong cultural and political influence of the Japanese state, have survived to this day.

Until recently, all questions of the population history of the indigenous population of the Ryukyu Islands seemed to have been resolved. As early as 1911, E. Baelz, based on the analysis of somatological features, put forward a hypothesis about the relationship of the modern population of the Ryukyu Islands with the Ainu (Baelz, 1911). Among the features that indicate a greater similarity between the Ryukyushis and the Ainu than with the Japanese, he attributed a significant increase in tertiary hair, darker pigmentation, weakening of the Mongoloid features of the eye area, small face height and short stature. This view dominated Japanese physical anthropology until the end of the twentieth century. It fits well into the currently recognized so-called dual structural model of the formation of the population of the Japanese Islands (Hanihara, 1991). The model suggests that the population diversity of the Japanese archipelago arose in the process of interaction between two main groups of the ancient population.

The first group dates back to the Upper Paleolithic population, which entered the Japanese Archipelago around 35-30 thousand years AGO via a land bridge connecting Southeast Asia and the Japanese Islands at that time. The Ryukyu Archipelago is a remnant of this bridge. It was here, at the Minatogawa site (Okinawa Island), that the most representative paleoanthropological materials of the Upper Paleolithic period were collected for Japan. The remains, which are approximately 18-16 thousand years old, belong to three individuals. The best preserved male skeleton (Minatogawa 1). This discovery is a key one for recon-

Table 1. Sources of data used

* The data is supplemented with the author's measurements (face height and nose protrusion angle).

structures of the early population history of Japan. According to its morphological features, the Minatogawa 1 skull is similar to the presumably Upper Paleolithic skull from Liujiang (Southern China), which indicates the direction of migration of the Upper Paleolithic population to Japan. At the same time, the Minatogawa 1 skull is similar to the skulls of the Neolithic Jomon culture, which shows a clear continuity with the Upper Paleolithic of Japan (Baba and Narasaki, 1991; Suzuki and Hanihara, 1982).

A second group of populations other than the Dzemons began to enter Japan, namely Kyushu, around 300 BCE, from Northern China and the Korean Peninsula. The migrants brought with them a new culture - yayoi. The Yayoi's economic strategy was based on rice farming; the Dzemons were primarily hunter-gatherers. Carriers of the Yayoi culture differed sharply from the Dzemons in craniological features: they had more pronounced Mongoloid features. It is assumed that as a result of the assimilation of the native Dzemonese population by the Yayoi culture carriers, modern Japanese were formed; the Ainu are descendants of the Dzemonese displaced to the north, who were only slightly affected by Japanese influence. Within the framework of the "dual structural model", the Ryukyushis are considered as descendants of the ancient Dzemonsky population, and their difference from the Ainu consists in a larger share of the Mongoloid component.

This concept was confirmed by the analysis of various systems of anthropological traits and genetic markers (Levin, 1961; Kozintsev, 1992a; Higo et al., 2003), but at the end of the XX century. It has been questioned by a number of researchers who have denied the presence of a Dzem component in the composition of Ryukyushis [Dodo, Doi, Kondo, 1998, 2000; Manabe et al., 1999; Pietrasewsky, 1999; Hatta et al., 1999].

The aim of this work is to test hypotheses of the origin of the modern population of the Ryukyu Islands by integrating data on two systems of anthropological features-craniometry and cranioscopy.

Material and methodology

The analysis is based on data on craniological series representing the Jomon, Yayoi, and Kofun cultures, Edo-era Tokyo turtles (XVIII century), and near-modern Japanese (including Ryukyu Islanders), Hokkaido and Sakhalin Ainu, Chinese (mostly northern), and Indonesians. All series were cranioscopically examined by A. G. Kozintsev according to the program proposed by him [Kozintsev, 1992a]. I analyzed the values of five signs - the frequency of the sphenoid-maxillary suture (CWS), posteroscullar suture (HSS), suborbital pattern type II (PGUII) and supraorbital foramina (NO), as well as the index of the transverse palatine suture (IPNS). The occipital index (ZI) was not taken into account: preliminary analysis showed that this feature does not affect any of the first three vectors, which contain the most important information about the intergroup variability of the analyzed groups. This confirms the conclusion that there are no noticeable gradients in the distribution of SI in Northern Asia (Kozintsev, 1988), and therefore its application in this case is not justified.

During the study, 14 craniometric parameters were also taken into account: length, width and height of the skull, width and height of the face, minimum width of the forehead, width and height of the orbits and nose, nasomalar and zygomaxillary angles, simotic index and angle of protrusion of the nose. Published data from various researchers were used. In the absence of some features in the publications, my own measurements were used (Table 1).

Methods for integrating data from various systems of anthropological features have been described by me in a number of works [Moiseev, 2001, 2004].

Results

At the first stage of the analysis, in addition to data on ancient and modern groups from the territory of Japan, as well as the Ainu of Sakhalin, values were used that characterize the series of skulls of the North Chinese, since it is North China, as well as Korea, that are currently considered as territories from which the mass migration of the ancestors of the Yayoi culture carriers to the Japanese Islands took place.

The results of studying cranioscopic and craniometric signs were quite similar. The first canonical vector (I KB), as well as the first main component (I GC), contrast the Dzemonian and North Chinese series (Fig. 1). The first of these vectors was obtained by analyzing craniometric data, and the second - cranioscopic data. Dzemons are characterized by a unique combination of both cranioscopic and craniometric features. Cranioscopically, they are distinguished by a combination of extremely high values of the HSS with low values of the PSU, but also the HSS (Table 2).

The craniometric peculiarity of the Dzemons is determined by the large width of the forehead and face on the Asian scale, by the rather strong protrusion of the nose at a low height (Table 3).The correlation between I GC and I KB is high and statistically significant (r = -0.70; p < 0.05). This indicates a high level of anthropological originality of the Dzemons, which probably took a very long period of isolation to form. There are no statistically significant correlation coefficients between other GC and KB (Table 4).

At the opposite pole of I KB and I GC is the North China series, while the ancient and modern series from the Japanese Archipelago occupy intermediate positions. This arrangement of groups within the vectors under consideration is well consistent with the hypothesis of the formation of the population of Japan on the basis of two main components-carriers of the Jomon culture and migrants from Northern China and the Korean Peninsula. Unfortunately, I don't have any data on the ancient and modern Korean series at my disposal.

Figure 1. Location of groups in the space of I GC (according to cranioscopic signs) and I KB (according to craniometric signs). The North Chinese model.

Table 2. Correlation coefficients between cranioscopic features and the first two main components

* Significant at p < 0.05.

Table 3. Correlation coefficients between the initial craniometric features and the first three canonical vectors

* Significant at p < 0.05.

See Table 4. Spearman's nonparametric correlation coefficients between the first three GC and HF

* Significant at p < 0.05.

The Ainu show the greatest similarity to the Dzemons in both systems of characteristics, which once again confirms the widespread opinion that the Ainu are descendants of the Dzemons. However, the gap between the Dzemonic and the Ainu series closest to it is quite large in both systems, and the Ainu show a rather pronounced North Chinese tendency, which is explained by the Mongoloid admixture. The source of the latter may be not only the Japanese, but also other Mongoloid groups, especially carriers of the early medieval Okhotsk culture of Southern Sakhalin and Hokkaido (Moiseev, 2008). Epochal microevolutionary changes may have played a role in shaping the differences between the Ainu and the Dzemons.

The remaining groups from the territory of Japan form a fairly dense group in the space of the First KV and I GC, which occupies an intermediate position between the Dzemons and Ainu, on the one hand, and the Northern Chinese, on the other. The differences between them are small and unstable at the intersystem level. Thus, the Yayoi series was closest to the Northern Chinese in terms of craniometric characteristics (IQ) and the farthest from them in terms of cranioscopic characteristics (IQ). The situation of the Ryukyushu people is also unstable. In terms of cranioscopic features, they are superior to the Japanese and Kofun people in terms of the degree of expression of the Dzem features, but in terms of craniometric features, the opposite is observed. However, whatever the actual position of the Ryukyushis within the limits of the Jōmon-North Chinese scale, the presence of a Jōmon trend in them, as in all Japanese series close to modern times, as well as in the carriers of the Yayoi and Kofun cultures, is beyond doubt. In my opinion, the only question is the amount of this admixture among the Ryukyushis and other Japanese. When integrating the results of the analysis of cranioscopic and craniometric features, the Ryukyushans, in particular, in terms of the first integral GC (IGC), which is mainly determined by the first GC and First KV described above, are slightly superior to modern Japanese and Edo-era series in terms of the severity of Dzemonic features (Fig. 2, Table 5). However, the difference is small and it can be explained by random factors.

2. Arrangement of groups in the space of I and II IGCS. The North Chinese model.

See Table 5. Correlation coefficients of the initial vectors and the first two integral principal components (North China model)

* Significant at p < 0.05.

3. Location of groups in the space of I GC (according to cranioscopic signs) and I KB (according to craniometric signs). An Indonesian model.

Table 6. Correlation coefficients of the initial vectors and the first three integral principal components (Indonesian model)

* Significant at p < 0.05.

The meaning of the second IGC is to contrast the Ainu of Sakhalin with the Ryukyusans. Other series on this IGC occupy an intermediate position. Their separation from both groups is quite large and indicates that they are neutral in this vector of variability. In other words, the II IGC indicates the presence of some anthropological components in the composition of the Ainu of Sakhalin and the Ryukyushans, which are absent in the other groups. As for the Ainu of Sakhalin, their deviation from the Ainu of Hokkaido can be confidently attributed to the presence of a significant Siberian (most likely Tungus-Manchurian) admixture [Kozintsev, 1992b]. The anthropological peculiarity of the Ryukyushans can theoretically be explained either by microevolutionary processes in the conditions of island isolation, or by the inclusion in their composition of some anthropological component that differs from all the analyzed groups. Significant correlation coefficients with this IGC are II GC and II KB (Table 5).

Unfortunately, I do not have craniological data on the aborigines of Taiwan and the Philippines, and it is with them that the Ryukyu Islanders most likely maintained contact. A series of Indonesian skulls was used to assess the possibility of southern admixture in the Ryukyushans. A second analysis was carried out, in which this group replaced the North Chinese one. The Indonesian model turned out to be slightly worse than the North Chinese one: the correlation coefficient between the new QIQ and the new QIQ decreased from 0.70 to 0.68, while remaining reliable at the 5% level. The Indonesian series is slightly more distant from the inter-group regression line than the North Chinese one (Figure 3). As in the case of the North Chinese model, the new I IGC is determined mainly by the first vectors obtained in the intra-system analysis (I KB and I GC), but the correlation coefficients between these vectors and I IGC are significantly lower (Table 6). Based on the above, it seems inappropriate for me to dwell on a detailed description of the location of groups within the I IGC.

Apparently, the North Chinese model is more suitable for explaining the source of Mongoloid admixture in most Japanese groups. This does not exclude the possibility that the Indonesian model may be more appropriate for some groups from Japan. In addition, it is impossible to exclude the presence of several components in the anthropological composition of the Ryukyushans. To find out this possibility, I used the method proposed by A. G. Kozintsev for determining the pair of most probable ancestral populations of a mestizo group (Kozintsev, 1992b).

The proposed RD (Relative Difference) criterion is based on the assumption that the ratio of the sum of Euclidean distances between the mestizo group (B) and each of the assumed ancestral groups (A and C) to the distance between the latter can serve as an indicator of the degree of suitability of a particular Mestizo model. The criterion is calculated using the following formula:

RD = (AB+SUN)/AC - 1, where AB and BC are the distances between the mestizo group and the two supposed parent groups, and AC is the distance between the latter. The smaller the RD, the closer the mestizo group is to the straight line connecting the ancestral groups and, therefore, the more suitable this model is.

In this case, one of the ancestral groups was considered to be the Dzemons, and the other-the Chinese and Indonesians. RD value when using Indonesian

Fig. 4. Location of IGCS in space II and III. An Indonesian model.

The model was almost half as large as when using the North Chinese model (0.23 and 0.40, respectively). Apparently, the Mongoloid component in the composition of the Ryukyushans is closer to that of the Indonesians. Thus, the results obtained suggest the presence of a South Mongoloid admixture in the Ryukyushans. It should only be repeated that, most likely, we are talking about kinship not with Indonesians, but with the population of Taiwan and the Philippines.

Of undoubted interest for the topic under discussion are the II and III IGCS obtained during the verification of the Indonesian model (Fig. 4). Thus, the II IGC separates the ancient Yayoi and Kofun groups from the rest of the groups. These groups clearly share a common origin; they are descendants of migrants from Korea and Northern China. The extreme position is occupied by the Yayoi people, while the Kofun group is somewhat shifted towards the other groups, probably due to the presence of a greater Dzemonic admixture among the Kofun people compared to the carriers of the Yayoi culture. Of the other groups, it is the Ryukyushans who are characterized by the most pronounced Yayoi tendency. In terms of the values of this IGC, the Dzemons are also close to the Ryukyushans, which makes it possible to interpret the II IGC as a vector separating the most ancient groups. Obviously, the Ryukyushans retained the anthropological characteristics characteristic of early migrants from Korea and Northern China. The second IGC is mainly determined by II GC and III KB obtained in the intrasystem analysis (Table 6).

The meaning of the third IGC is to fix the southern trend in the Ryukyushans, and the main significance for its determination was played by craniometric data (Table 6). At the negative pole of this IGC are the Ryukyushans and Indonesians, and at the positive pole are the Japanese groups, except for the Yayoi and Kofun series.

Thus, the Indonesian model allows us to specify the sources of Non-Japanese admixture in the population of the Ryukyu Islands, while the North Chinese model only indicates the anthropological differences between the Ryukyuans and other Japanese populations.

Conclusions

1. The modern population of the Ryukyu Archipelago, as well as the Ainu and Japanese, has a single anthropological component, which is related by origin to the carriers of the Neolithic Jomon culture.

2. The value of the Dzem component in the Ryukyushis is close to that of the modern Japanese.

3. The Mongoloid component in the composition of the Ryukyushans is probably related in origin to the carriers of the Yayoi culture.

4. The Ryukyusans have a pronounced anthropological peculiarity, which may be due to the presence of a South Mongoloid admixture in their composition.

Acknowledgements

I express my sincere gratitude to Professor T. Amano (Hokkaido University), Dr. H. Ono (Hokkaido University) and Dr. H. Matsumure (Sapporo Medical College) for the opportunity to study craniological collections. Special thanks to Professor H. Ishide (University of Okinawa) for providing unpublished data. I am also deeply grateful to A. G. Kozintsev for his critical comments and assistance in the work.

List of literature

Alekseev, V. P., Materials on Craniology of New Guinea, the Sunda and Moluccas Islands, and the Malay Peninsula, Part II, Sb. MAE, 1984, vol. 39, pp. 26-60.

Debets G. F. Antropologicheskie issledovaniya v Kamchatskoi oblasti [Anthropological studies in the Kamchatka Region], Moscow; Leningrad: Izd-vo AN SSSR, 1951.

Kozintsev A. G. Ethnic cranioscopy (racial variability of skull sutures). - L.: Nauka, 1988. - 168 p.

Levin M. G. Nekotorye problemy etnicheskoi antropologii Jap'anii [Some problems of ethnic anthropology of Japan]. SE. - 1961. - N 2. - pp. 63-75.

Masafumi N. Nihon minzoku (The Japanese and their Culture). Tokyo: Rokko Publ., 1988, vol. 2, 320 p. (in English).

Moiseev V. G. Severnaya Evraziya: yazykovaya differentsiatsiya i dannye fizicheskoi antropologii [Northern Eurasia: linguistic differentiation and data from Physical Anthropology]. - 2001. - N4. - p. 154-159.

Moiseev V. G. Opyt integratsii razlichnykh sistem antropologicheskikh chastov [Experience of integration of various systems of anthropological features]. St. Petersburg: MAE RAS Publishing House, 2004, pp. 186-200.

Moiseev V. G. O proiskhozhdenii naseleniya okhotskoi kul'tury Severnogo i Vostochnogo Hokkaido (po kraniologicheskim dannym) [On the origin of the population of the Okhotsk culture of Northern and Eastern Hokkaido (based on craniological data)]. - 2008. - N 1. - p. 133-140.

Cheboksary, N. N. Ethnic Anthropology of China. Moscow: Nauka Publ., 1982, 301 p. (in Russian)

Baba H., Narasaki S. Minatogawa man, the oldest type of modem Homo sapience in East Asia // Quarter Results. - 1991. -Vol. 30. -P. 221 - 230.

Baelz E. Die Riu-Kiu-Islunaner, die Aino und andere kaukasierahnliche Reste in Ostasien // Korrespondenz-Blatt der Deutschen Gesellschaft fur Anthroplogie, Ethnologie und Urgeschichte. - 1911. - Bd. 42. - S. 187 - 191.

Dodo Y., Doi N., Kondo O. Ainu and Ryukyuan cranial nonmetric variation: evidence which disputes the Ainu-Ryukyu common origin theory // Anthropological Science. - 1998. -Vol. 106. - P. 99 - 120.

Dodo Y., Doi N., Kondo O. Facial flatness of Ryukyuans // Anthropological Science. - 2000. - Vol. 108 (2). - P. 183 - 198.

Fukumine T., Doi N., Ishida H., Zukeran C., Sensui S., Saso A., Higa T. Human skeletal remains from the Yacchino-Gama and Kanjinbaru grave sites // Okinawa Prefectural Archaeological Center Research Report. - 2001. - N 6. -P. 345 - 385.

Hanihara K. Dual structure model for population history of the Japanese // Japan Review. - 1991. - N2. - P. 1 - 33.

Hatta Y., Ohashi J., Imanishi T., Kamiyama H., Iha M., Simabukure Т., Ogawa A., Tanaka H., Akaza T., Gojobori T., Juji T., Tokunaga K. HLA genes and haplotypes in Ryukyuans suggest recent gene flow to the Okinawa Islands // Human Biology. - 1999. -Vol. 71. -P. 353 - 365.

Higo Т., Hanihara Т., Sukawa H., Ishida H. Dental variation of Ryukyu Islanders: a comparative study among Ryukyu, Ainu, and other Asian populations // Amer. J. of Physical Anthropology. - 2003. - Vol. 15. - P. 127 - 143.

Kozintsev A.G. Ethnic epigenetics: A new approach // Homo. - 1992a. - Vol. 43 (3). - P. 213 - 244.

Kozintsev A.G. Prehistoric and recent populations of Japan: multivariate analysis of cranioscopic data // Arctic Anthropology. - 1992б. - Vol. 29(1). - P. 104 - 111.

Manabe Y., Kitagawa Y., Oyamada J., Kato K., Ito R., Kobayashi S., Rokutanda A. Genealogical positioning of the Okinawa main islanders in East Asian populations based on nonmetric tooth crown traits // Anthropological Science. -1999. - Vol. 107. - P. 85.

Pietrusewsky M. A multivariate craniometric study of the inhabitants of the Ryukyu Island and comparisons with cranial series from Japan, Asia, and the Pacific // Anthropological Science. - 1999. - Vol. 107. - P. 255 - 281.

Suzuki H., Hanihara K. The Minatogawa Man: The Upper Pleistocene Man From the Island of Okinawa // University of Tokyo Bulletin. - Tokyo: University Museum, 1982. - N 19. -P. 7 - 49.

The article was submitted to the Editorial Board on 29.05.08.

Permanent link to this publication: