Population History of the Northern and Central Nansei Islands

Home , Sinodonty and Sundadonty, Tanegashima

ANTHROPOLOGICAL SCIENCE Vol. advpub, 000–000, 2007

Population history of the northern and central Nansei Islands (Ryukyu island arc) based on dental morphological variations: gene flow from North Kyushu to Nansei Islands Yoshitaka MANABE1*, Yoshikazu KITAGAWA1, Joichi OYAMADA1, Kazunari IGAWA1, Katsumoto KATO2, Naohiro KIKUCHI3, Hiromi MARUO3, Shigeru KOBAYASHI3, Atsushi ROKUTANDA1 1Department of Oral Anatomy and Dental Anthropology, Unit of Basic Medical Sciences, Course of Medical and Dental Sciences, Graduate School of Biomedical Sciences, Nagasaki University, Nagasaki, 852-8588 Japan 2Department of Physical Therapy, Unit of Physical and Occupational Therapy, Course of Health Science, Graduate School of Biomedical Sciences, Nagasaki University, Nagasaki, 852-8588 Japan 3Division of Anatomy, Kyushu Dental College, Kitakyushu, 803-8580 Japan

Received 9 March 2007; accepted 20 July 2007

Abstract This study investigates dental morphological variations among the modern inhabitants of Tanegashima Island and Okinawa Main Island of the Nansei Islands (Ryukyu island arc), which is the southern gateway of the Japanese archipelago. Temporal variations within Tanegashima Island and regional variations of the northern half of the Nansei Islands were used to study population history. The late Aeneolithic Yayoi to protohistoric Kofun populations (c. 0–700 AD) of Tanegashima was found to be similar to the native Japanese populations, such as the Neolithic Jomonese (c. 10000–300 BC) and Hokkaido Ainu; however, the modern population of Tanegashima was similar to the migrant Jap- anese populations such as the post-Jomon people (c. 300 BC–present time) of the Japanese main islands. Using statistically sufficient materials, this study confirmed that a substantial change had occurred from the prehistoric to the modern period in Tanegashima Island. It is suggested that temporal change occurred in Tanegashima over approximately 1000 years after substantial change (c. 300 BC) in the Japanese main islands. The delay in temporal change in Tanegashima indicates that dispersal of migrant populations from northern Kyushu to its southern end occurred at a time remarkably later than the northeastward dispersal of migrant populations to central Honshu Island. Furthermore, the geographical cline extending from North Kyushu to Okinawa Main Island via Tanegashima Island suggests southward gene flow from North Kyushu to the central Nansei Islands via the northeast end of the Nansei Islands.

Key words: Nansei Islands, Ryukyu-Okinawa, Tanegashima, nonmetric dental traits, population history

Introduction Main Island, from Pinza-Abu (c. 26000 BP) (Sakura, 1985) in Miyako Island, and from other sites in the Nansei Islands. The Nansei Islands, otherwise known as the Ryukyu is- Affinities between the Pleistocene humans of the Nansei Is- land arc, belong to Kagoshima and Okinawa prefectures and lands and the Jomonese (c. 12000–2300 BP) of the Japanese comprise a chain of Japanese islands situated in the western main islands can be perceived as one of the grounds support- Pacific Ocean at the eastern limit of the East China Sea. ing the southern origin theory regarding the Jomonese. In These islands extend southwest from Kyushu Island to Tai- addition, it is very important to investigate the southern dis- wan. It is of great importance to elucidate the physical an- persal of migrant populations from North to South Kyushu thropological characteristics on the Nansei Islands in order and the Nansei Islands from the viewpoint of dispersal of the to clarify the population history of the Japanese archipelago migrant populations that entered South Honshu and North and its peripheral regions, because these islands, which are Kyushu after the shift from the Jomon to the Yayoi period. located in the southern end of the Japanese archipelago, con- An investigation of the relationship between the Nansei stitute the southern gateway of gene flow to Japan. islanders (Ryukyu arc islanders) and other populations led to Some Pleistocene human fossil remains have been ex- the ‘Ainu–Ryukyuan common origin theory’. This theory is cavated from Minatogawa (c. 18000 BP) (Suzuki and K. one of the basic components of the now widely accepted Hanihara, 1982; Baba and Narasaki, 1991) in Okinawa ‘dual structure model’ for the population history of the Jap- anese by K. Hanihara (1991). Ever since a physical similari- * Corresponding author. e-mail: [email protected] ty between the Ryukyuan and Ainu populations was indicat- phone: +81-95-849-7626; fax: +81-95-849-7629 ed by von Baelz in 1911, many anthropological studies using Published online 14 November 2007 various methods have been conducted on Nansei islanders. in J-STAGE (www.jstage.jst.go.jp) DOI: 10.1537/ase.070309 The methods used in these studies comprised somatological

© 2007 The Anthropological Society of Nippon 1 2 Y. MANABE ET AL. ANTHROPOLOGICAL SCIENCE measurements and observations (Newman and Eng, 1947; 2003a). Suda, 1950; Ikeda and Tagaya, 1980), osteological measure- In the present study, modern inhabitants of Okinawa Main ments and nonmetric traits (Hsu, 1948; Ikeda, 1974; Island, which is centrally located in the Nansei Islands, were Kohhara, 1976; Tagaya and Ikeda, 1976; Yamaguchi, 1982, investigated as a representative population of the Nansei Is- 1992; Mouri, 1986; Dodo, 1992; Pietrusewsky, 1994, 1997, lands. Furthermore, modern inhabitants of Tanegashima— 1999, 2004; Doi et al., 1997; Dodo et al., 1998, 2000; Doi, located at the entrance to the Nansei Islands from the south- 1998, 2003, 2004; Asato and Doi, 1999; Matsushita, 2001; ernmost Kyushu Island in the Japanese main islands—were Nakahashi, 2003), dental measurements and nonmetric studied to estimate the two-way genetic influence between traits (Yamada et al., 1970; K. Hanihara et al., 1974; K. Kyushu Island and the Nansei Islands. We investigated not Hanihara, 1976; Tsuru, 1978; Matayoshi, 1982; Turner, 1987; only the Ainu–Ryukyuan common origin theory based on T. Hanihara, 1989, 1991, 1992; Matsumura, 1995; Manabe the variations in the modern populations of Okinawa Main et al., 1999a, b, 2001, 2003a; Higa, 2003; Haneji, 2007), Island and Tanegashima Island but also the geographical and genetics (e.g. Omoto et al., 1976; Omoto, 1978, 1992; cline that indicates migratory direction. Horai et al., 1996; Omoto and Saitou, 1997; Hatta et al., The materials from the Aeneolithic Yayoi to the protohis- 1999; Tokunaga, 2003). However, the results that were toric Kofun period (0–700 AD) excavated from Tanegashi- obtained using these different methods and materials led to ma constitute a rare collection of materials in that they are diverse opinions, and there was no consensus with regard to available in statistically sufficient quantities among the pre- the anthropological positioning of the Nansei islanders in the historic human remains excavated from the Nansei Islands. Japanese archipelago and East Asia. These materials show dental characteristics that are similar In addition to the differences in the methods used, the lim- to those found in the Neolithic Jomonese (12000–2300 BP) itations involved in the different material gathering tech- of the Japanese main islands (Manabe et al., 2002). The ma- niques were also responsible for the lack of consensus. It is terials from the Yayoi–Kofun skeletal remains and from the crucial to collect data pertaining to each island in order to modern inhabitants of Tanegashima were compared using elucidate the regional variations in the arc-shaped Nansei Is- nonmetric dental traits in order to clarify the temporal lands, which spread over 1300 km, but difficult to collect changes that have occurred from the prehistoric to the mod- statistically sufficient data sets for each island from skeletal ern period in the area that forms the northern entrance to the remains. This is because, except for a few sites, the Nansei Nansei Islands. The period during which the temporal Islands do not yield an adequate amount of excavated skele- change occurred in Tanegashima was compared to the peri- tal materials for research purposes. However, tooth crown od when such changes occurred on the central Japanese morphology, for which samples can be gathered from the main island, with the aim to examine the possible causes that present inhabitants, can be used for collecting statistically underlie the regional differences in the timing of temporal sufficient data for each island. Moreover, to elucidate the changes. Finally, from the results of this study, as well as by population history of the Nansei Islands, it is as important to considering some recent osteological evidence (Doi et al., explain the temporal changes that occurred on each island as 1997; Doi, 1998, 2003, 2004; Dodo et al., 1998, 2000; Asato to reveal the regional variations. However, thus far, no study and Doi, 1999; Matsushita, 2001), we examine the charac- has inspected the temporal changes from the prehistoric to teristics of the prehistoric population of the Nansei Islands, the modern period on any of the Nansei Islands using statis- and discuss the validity of the ‘Jomonese–Ryukyuan com- tically sufficient data. mon origin theory’ from a prehistoric perspective. In order to inspect the temporal changes that have occurred from the past to the modern period, it is necessary to Materials and Methods conduct a simultaneous analysis of data from skeletal remains and present-day living persons. Tooth crown morpho- We investigated the modern inhabitants of Okinawa Main logical data enable this simultaneous analysis by using the Island (106 males/98 females), which constitutes the geo- same method of data collection for skeletons and living bod- graphical and political center of the Nansei Islands, and ies, while the simultaneous analysis of osteological and so- those of Tanegashima Island (68 males/61 females), which matometrical data is impossible. Nonmetric traits of tooth is located near the northern end of the Nansei Islands, at the crowns were considered in this study because variations in entrance to the Nansei Islands from Kyushu Island these traits in East Asia have been investigated in detail, and (Figure 1, Table 1). Plaster casts prepared by means of oral the traits are known to express the genealogical characteris- impression were obtained from the present Nansei islanders. tics of a population. Young individuals with minimal dental wear or antemortem The nonmetric traits of tooth crowns in the skeletal re- tooth loss were selected as sources of highly reliable data. mains excavated from the Nansei Islands were first investi- Using the Arizona State University dental anthropology gated by Yamada et al. (1970), and those of living inhabit- system (ASUDAS) (Turner et al., 1991), 17 nonmetric traits ants were investigated by K. Hanihara et al. (1974), Tsuru of the tooth crown were observed and classified, with special (1978), Matayoshi (1982), and others. Following the gradual attention to the 11 traits that typically indicate marked dif- elucidation of the variations in the dental morphology in ferences between the native and migrant Japanese popula- East Asia by Turner’s large-scale investigations (Turner, tions in the Mann–Whitney test (Manabe et al., 2002). Inci- 1987, 1990, 1992), the anthropological positioning of dences were counted using the criterion developed by Nansei islanders in East Asian populations has been clearly Turner and Scott (1977). Sex-pooled frequencies of the 17 indicated subsequently (Manabe et al., 1999a, b, 2001, traits are shown in Table 2, since neither a chi-square test Vol. 115, 2007 DENTAL VARIATION IN NORTHERN NANSEI ISLANDS 3

Figure 1. Map of Nansei Islands in the Japanese archipelago.

Table 1. Materials from the Japanese archipelago investigated based on dental morphology Sample size Population abbreviation Region or ethnic group Period Reference (individuals) Native Japanese IE-JOMON Total Japan Neolithic initial-early Jomon 73 Manabe et al. (2002) (c. 8000–3000 BC) MLF-JOMON Total Japan Neolithic middle-late-final Jomon 451 Manabe et al. (2002) (c. 3000–300 BC) RH-AINU Hokkaido Ainu (northern island) Early Modern (1568–1867 AD) 153 Manabe et al. (2002) Migrant Japanese YA-YAYOI Yamaguchi in southwestern Honshu Aeneolithic Yayoi (c. 300 BC–100 AD) 186 Manabe et al. (2002) (main island) NK-YAYOI Northern Kyushu Aeneolithic Yayoi (c. 100 BC–300 AD) 106 Manabe et al. (2002) KOFUN Total Japan Protohistoric Kofun (c. 300–700 AD) 199 Manabe et al. (2002) KAMAKURA Total Japan Medieval Kamakura (c. 1192–1338 AD) 64 Manabe et al. (2002) MI-MODERN Honshu (main island) Modern (1867 AD–present) 192 Manabe et al. (2002) NK-MODERN Northern Kyushu Modern (1867 AD–present) 93 Manabe et al. (2002) Nansei Islanders TA-YAYOI-KOFUN Tanegashima Island in southern Aeneolithic Yayoi to Protohistoric Kofun 117 Manabe et al. (2002), Kyushu (c. 0–700 AD) present study TA-MODERN Tanegashima Island in southern Modern (1867 AD–present) 129 Present study Kyushu OK-MODERN Okinawa Main Island Modern (1867 AD–present) 204 Present study 4 Y. MANABE ET AL. ANTHROPOLOGICAL SCIENCE nor Fisher’s exact test indicated that any sex difference ex- by Manabe et al. (2002), as constituting the only statistically isted in trait expression. comparable materials of prehistoric Nansei islanders. The degree of dissimilarity among populations was evalu- The Yayoi period is an era in the history of the Japanese ated using Smith’s mean measure of divergence (MMD) sta- main islands that followed the Jomon period (c. 10000–300 tistic (Berry and Berry, 1967) with Freeman–Tukey trans- BC) and extended from approximately 300 BC to 300 AD. formation (Freeman and Tukey, 1950). Green and Suchey However, a recent study using accelerator mass spectrome- (1976) demonstrated that the Freeman–Tukey transforma- try (AMS) radiocarbon dating method discovered that the tion should be used to stabilize variances in small samples. beginning of the Yayoi period, in fact, dated back to c. Spearman’s ranked correlations among the 17 traits were 1000–800 BC, nearly 700–500 years earlier than what was examined for over 3800 individuals participating in this previously believed (Imamura, 2001; Harunari, 2003). In analysis, because MMD assumes that traits are independent. this study, for the sake of convenience, the Yayoi period was Only one pair—UM2 hypocone/LM2 Y-groove—showed a considered to stretch from 300 BC to 300 AD, since its actu- moderate correlation (r = 0.304), while the other pairs al dates are still the objects of controversy among archeolo- showed weak or no correlation. gists. When this controversy is resolved in the future, the pe- Multidimensional scaling (MDS) based on the MMDs riod specified in this article should be revised and was conducted using the 17 traits in order to express two-di- appropriately interpreted by readers. mensional interpopulational affinities. The relative contribu- Periodization of the Amami and Okinawa areas located to tion of each of the 17 traits that contributed to the first and the south of Tanegashima Island was different from that of second axes of MDS was identified using a linear multiple the Japanese main islands. The Kaizuka (Shellmound) peri- regression analysis on each trait (Kruskal and Wish, 1978). od of the hunting-gathering economy stage in the Amami This analysis was conducted using the coordinate values of and Okinawa areas corresponds to the time between the ear- the first and second axes as two independent variables and ly Jomon and middle Heian period (c. 6500 BC–1000 AD), the incidences of the traits as a dependent variable. according to the dating system of the Japanese main islands. Two modern Nansei Island populations were compared The subsequent period comprising approximately 400 years with ten other Japanese populations (Manabe et al., 2002) in is called the Gusuku period, and is characterized by multiple the analysis of variations in the Japanese archipelago crop-cultivation, the onset of the ceramic industry, and iron (Figure 1, Table 1). The ten Japanese populations were di- production. vided into two groups of native and migrant populations During the analysis of variations in 36 Asian-derived pop- (Manabe et al., 2002) based on their variations in the non- ulations, the Thai tribe (Manabe et al., 1997a), the Northeast metric traits pertaining to tooth crowns and roots. The abbre- Chinese (Manabe, personal data), the Neolithic Northern viations used for the Japanese populations in this study are Chinese (Manabe et al., 2003b), the Taiwan Yami tribe given in Table 1. (Manabe, 1989), the Taiwan Bunun tribe (Manabe et al., Skeletal remains excavated from the Hirota site and be- 1991), the Taiwan Ami tribe (Manabe et al., 1992), the pre- longing to the late Aeneolithic Yayoi and the protohistoric Inca Peruvian (Manabe and Rokutanda, 1992), and 17 other Kofun period (c. 0–700 AD) (Morizono, 2003) in Tanega- East-Asian populations (Turner, 1987) were compared with shima were investigated using tooth crown morphology the two modern Nansei islander groups in this study.

Table 2. Frequencies of 17 non-metric traits in tooth crowns of the Okinawa Main Islanders and Tanegashima Islanders Frequency % (number examined) Trait Tooth Expression dichotomy Tanegashima Okinawa Yayoi-Kofun1 Modern2 Modern2 Shoveling UI1 (3–7/0–7) 18.0 (50) 68.5 (124) 63.3 (196) Double-shovel UI1 (3–6/0–6) 1.8 (56) 16.3 (123) 16.2 (191) Interruption groove UI2 (1–4/0–4) 38.2 (34) 17.7 (113) 24.6 (179) Mesial ridge UC (1–3/0–3) 3.4 (29) 5.2 (96) 7.6 (171) Distal accessory ridge UC (2–5/0–5) 50.0 (10) 79.2 (96) 77.3 (181) Hypocone UM2 (2–5/0–5) 97.8 (46) 84.1 (82) 92.6 (149) Cusp 5 UM1 (1–5/0–5) 64.0 (25) 35.5 (107) 57.0 (165) Carabelli trait UM1 (2–7/0–7) 39.1 (46) 32.0 (125) 39.7 (194) > 1 lingual cusp LP2 (1–9/A, 0–9) 93.8 (48) 89.0 (118) 72.6 (186) Y groove pattern LM2 (0/0–2) 15.5 (58) 4.7 (86) 7.8 (167) Cusp 6 LM1 (1–5/0–5) 22.2 (54) 38.7 (93) 36.3 (168) 4-cusp LM2 (4/4, 5) 49.2 (59) 25.8 (89) 38.6 (171) Deflecting wrinkle LM1 (3/0–3) 0.0 (28) 25.3 (87) 21.3 (94) Distal trigonid crest LM1 (1/0, 1) 5.0 (40) 30.0 (90) 8.2 (85) Protostylid LM1 (1–7/0–7) 29.2 (48) 32.7 (104) 36.1 (133) Cusp 7 LM1 (1–4/0–4) 5.2 (58) 17.0 (100) 12.9 (170) Odontome U, LP1, 2 (1/0, 1) 2.0 (49) 0.8 (124) 2.2 (184) 1 Manabe et al. (2002), 2 present study. Vol. 115, 2007 DENTAL VARIATION IN NORTHERN NANSEI ISLANDS 5

Nichol and Turner (1986) discussed the interobserver er- Interpopulational distance analysis ror and suggested that when scoring is performed by differ- Smith’s MMDs among 12 Japanese populations based on ent observers, sample comparisons are possible at various their trait incidences were calculated in order to synthesize levels of confidence. We have previously examined the po- the characteristics of the 17 traits. Abbreviations for the Jap- sitions of typical sinodonty and typical sundadonty in the anese populations, such as TA-MODERN, were used for the MDS in order to observe interobserver concordance purpose of clarity (Table 1). With regard to the MMDs from (Manabe, 2003b). We concluded that it was possible to ana- the TA-MODERN population (Figure 2), the TA-MODERN lyze both Turner’s and Manabe’s data because of the ap- population was particularly similar to the NK-MODERN, proximate agreement between these two sets of observa- KAMAKURA, and MI-MODERN populations and consid- tions. erably similar to the KOFUN, OK-MODERN, and NK- YAYOI populations. However, TA-MODERN was dissimi- Results lar to the IE-JOMON, RH-AINU, MLF-JOMON, and TA- YAYOI-KOFUN populations. Comparison with other Japanese archipelago populations In all, the modern Tanegashima islanders were similar to Comparison of each trait the migrant Japanese group but not the native Japanese The frequencies of the 17 traits are shown in Table 2 and group. The modern Okinawa main islanders exhibited ten- Table 3. The incidences of the traits in the three populations dencies similar to the modern Tanegashima islanders. It was of the Nansei Islands, i.e. modern populations of Okinawa interesting that among the Japanese populations considered, Main Island and Tanegashima Island and the population of the modern Tanegashima islanders were highly dissimilar to the late Yayoi–Kofun period (c. 0–700 AD) in Tanegashima the Tanegashima islanders of the Yayoi–Kofun period. Island, were compared with three native and six migrant Figure 3 shows the 12 Japanese populations arranged into Japanese populations (Manabe et al., 2002) using either the two dimensions by MDS, based on the MMDs of these 12 chi-square test (Yate’s correction) or Fisher’s exact test populations. On the first axis, the migrant Japanese popula- (Table 3). tions such as NK-YAYOI, YA-YAYOI, KOFUN, Based on the comparative results of the 11 traits, which KAMAKURA, MI-MODERN, and NK-MODERN are exhibited marked differences between the native and mi- located in the positive direction, while the native Japanese grant Japanese populations (Manabe et al., 2002) (Table 3), populations such as IE-JOMON, MLF-JOMON, and RH- the similarities found between the three Nansei islander pop- AINU are located in the negative direction. The first axis dis- ulations and the native or migrant Japanese populations are tinctly divides the two genealogical groups of the Japanese presented in Table 4. archipelago. On the first axis, TA-YAYOI-KOFUN is includ- In the instance of UI1 shoveling, the Tanegashima Yayoi– ed in the native Japanese population group. However, OK- Kofun population (TA-YK) was classified as a population MODERN and TA-MODERN are not included in this group similar to the native Japanese populations (Table 4), because but are located in close proximity to the migrant Japanese while this population was significantly different from the as compared to the native Japanese population. Based on migrant Japanese populations, it was not significantly differ- the coordinate values of the first axis, OK-MODERN is ent from the native Japanese populations (Table 3). In the located slightly closer to the native Japanese population than case of traits that showed nearly no significant difference, TA-MODERN. such as the LM1 protostylid presented (Table 3), the similar- On the second axis, among the native Japanese popula- ity of population was judged from the degree of similarity to tions, IE-JOMON and MLF-JOMON are located in the neg- the trait’s incidence in the native Jomonese or migrant Yayoi ative direction, while RH-AINU are located in the positive population (as presented in Table 4). direction. Similarly, among the migrant Japanese popula- The modern Okinawa main islanders showed native-like tions, NK-MODERN is located towards the positive direc- features in only two traits, migrant-like features in six traits, tion, and YA-YAYOI is located towards the negative direc- and intermediate features in the remaining three traits. The tion. Based on this arrangement, it is suggested that the modern Tanegashima islanders showed migrant-like fea- second axis reflects a trend of temporal change in the Japa- tures in seven traits, intermediate features in the remaining nese populations. four traits, and no native-like feature. The Tanegashima islanders of the Yayoi-Kofun period showed native-like fea- Interpretation of the axes tures in ten traits, no migrant-like feature, and intermediate Multiple regression analysis identified the relative contri- features in only one trait. bution of each trait to each MDS axis (Table 5, Figure 4). Analysis of each trait revealed that the modern popula- The traits that indicated both significant multiple correlation tions of Okinawa Main Island and Tanegashima Island have coefficients and multiple regression coefficients at the 5% more migrant-like than native-like characteristics. The mod- and 1% levels on the first axis were predominantly the traits ern Okinawa main islanders had marginally more native-like revealing obvious differences between the native and mi- characteristics than the modern Tanegashima inhabitants. In grant Japanese populations in previous analyses, e.g. UI1 contrast, the Tanegashima islanders of the Yayoi–Kofun pe- shoveling, UI1 double shovel, UC distal accessory ridge, riod exhibited native-like characteristics that were very sim- LM1 cusp 6, LM2 4-cusp, and LM1 deflecting wrinkle ilar to those of the Jomonese. (Manabe et al., 2002). Thus, the first axis can be used to distinguish between the native and migrant Japanese populations. 6 Y. MANABE ET AL. ANTHROPOLOGICAL SCIENCE OK-M* OK-M* OK-M** OK-M** < > < < TA-M* ·· TA-M* TA-M** > > < ulations, using chi-square TA-YK* ·· TA-YK** < < OK-M* 4.2 48 ·· ·· ·· OK-M* 10.0 40 ·· ·· ·· OK-M* 19.1 47 ·· OK-M** 32.0 125 ·· — ·· OK-M** 38.8 67 ·· ·· ·· OK-M** 37.4 107 ·· ·· ·· OK-M** 34.8OK-M** 33.3 69 48 ·· ·· ·· ·· ·· ·· OK-M** 27.3 55 ·· ·· ·· OK-M** 19.7 76 OK-M** 10.2 59 ·· ·· ·· OK-M** 0.0 9 ·· ·· ·· > > > < < < < < < < > > hree native and six migrant pop ve (UI2) Mesial ridge (UC) TA-M* ·· 39.1 46 — ·· ·· TA-M* ·· 3.4 29 — ·· ·· M1) Carabelli trait (UM1) TA-M** — 39.7 194 ·· ·· — TA-M** ·· 17.9 56 ·· TA-M** TA-M** TA-M** TA-M** TA-M** > > > > > > > > > TA-YK* — TA-YK* ·· TA-YK* ·· TA-YK* — ·· 5.2 96 ·· — ·· TA-YK* ·· ·· 6.7 45 ·· ·· ·· TA-YK* TA-YK** ·· TA-YK** ·· TA-YK** ·· TA-YK** ·· < < < < < > < < < < nese populations, including the t OK-M** 22.9 48 OK-M** 38.2 34 — OK-M** 50.0 56 ·· OK-M** 60.0 125 < < > < TA-M* ·· 64.0 25 — TA-M* ·· 48.2 85 ·· ·· ·· 17.8 152 TA-M* TA-M** TA-M** > > < vel (UI1) groo Interruption > < test or Fisher’s exact TA-YK* — ·· 35.5 107 TA-YK* ·· ·· 28.2 78 TA-YK* ·· TA-YK* — ·· 17.7 113 TA-YK* ·· ·· 43.1 51 ·· TA-YK** ·· — 24.6 179 ·· ·· — 7.6 171 ·· ·· — TA-YK**····20.035······10.040······ TA-YK**····29.987······12.985······ TA-YK** ·· ·· 39.2 51 ·· TA-YK** TA-YK** ·· ·· 44.2 52 ·· < < < > > > > > > > > 0.05; ··, not significant. nders and the nine Mainland Japa < P TA-YAYOI-KOFUN; TA-M, TA-MODERN; OK-M, OK-MODERN. TA-MODERN; TA-M, TA-YAYOI-KOFUN; OK-M* 84.8 66 ·· ·· ·· 21.2 52 OK-M** 80.0 80 OK-M** 1.8 56 — OK-M** 25.0 36 OK-M** 35.7 56 OK-M** 3.2 31 ·· ·· ·· 10.3 29 OK-M** 1.0 105 ·· OK-M** 5.0 20 ·· ·· ·· 51.9 27 ·· > < < > > < < < C) Hypocone (UM2) Cusp 5 (U *, significant at > TA-M* TA-M** TA-M** TA-M** TA-M** TA-M** TA-M** > * and < < > < < < < 0.01; Shoveling (UI1) Double-sho TA-YK* ·· ·· 84.9 106 TA-YK* ·· ·· 94.5 55 ·· ·· ·· 27.6 29 < TA-YK** ·· TA-YK** ·· — 16.2 191 TA-YK** — ·· 16.3 123 TA-YK** TA-YK** ·· ·· 16.3 49 TA-YK** ·· ·· 27.6 76 TA-YK** ·· ·· 25.5 51 TA-YK** TA-YK** ·· ·· 26.4 53 > > > > > > > > > > > P Distal accessory ridge (U % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M **, significant at > ** and Population abbreviations as shown in Table and TA-YK, 1, < Table 3. Comparisons between the incidences of three Nansei Isla TA-MODERNOK-MODERN 79.2 77.3 96 181 ·· ·· — ·· ·· — 84.1 92.6 82 149 ·· ·· — 57.0 165 ·· TA-YAYOI-KOFUN 50.0 10 — ·· ·· 97.8 46 — NK-MODERN 96.6 29 KAMAKURAMI-MODERN 79.3 88.3 29 60 ·· ·· ·· 89.1 55 ·· ·· ·· 23.5 34 YA-YAYOI 60.0 20 ·· ·· ·· 93.0 71 ·· ·· ·· 46.5 43 ·· ·· ·· 25.3 79 ·· ·· OU 249··· ·· 72.429·· ··KOFUN 88.972·· ·· 30.253 NK-YAYOI 92.6 27 Population RH-AINU 40.0 25 ·· IE-JOMONMLF-JOMON 42.9 72.4 29 7 ·· ·· ·· ·· ·· ·· 93.5 95.8 169 24 ·· ·· ·· ·· 37.5 8 ·· ·· ·· 15.0 20 ·· ·· ·· OK-MODERN 63.3 196 TA-MODERN 68.5 124 TA-YAYOI-KOFUN 18.0 50 — NK-MODERN 88.2 34 KAMAKURA 74.5 47 MI-MODERN 75.0 76 KOFUN 72.5 51 YA-YAYOI 93.3 60 NK-YAYOI 70.0 50 RH-AINU 6.9 29 ·· MLF-JOMON 14.4 97 ·· IE-JOMON 12.5 16 ·· Population Vol. 115, 2007 DENTAL VARIATION IN NORTHERN NANSEI ISLANDS 7 OK-M* OK-M* OK-M* OK-M** OK-M** OK-M** < < > < < > TA-M* ·· TA-M* ·· TA-M** TA-M** ·· TA-M** TA-M** < < < > > > TA-YK** — ·· TA-YK** ·· TA-YK** ·· TA-YK** ·· ·· TA-YK** ·· TA-YK** ·· ·· < < < < < < OK-M* 4.4 90 ·· OK-M* 11.9 177 ·· ·· ·· OK-M** 57.3 75 ·· < < < TA-M** < TA-YK* ·· ·· 21.3 89 TA-YK* ·· ·· 25.0 76 TA-YK** ·· ·· 24.0 50 TA-YK** ·· ·· 49.8 201 ·· > > > > OK-M* 39.6 48 ·· ·· ·· 4.5 44 ·· ·· ·· OK-M* 16.7 60 ·· OK-M** 32.7 104 ·· — ·· 17.0 100 ·· — ·· OK-M** 39.3 28 ·· ·· ·· 3.1 32 ·· ·· ·· OK-M** 20.7 58 ·· ·· ·· 9.4 64 ·· ··· ·· > > > > > n (LM2) Cusp 6 (LM1) (LM2) 4-cusp est (LM1) Protostylid (LM1) Cusp 7 (LM1) TA-M*··43.966······14.888······ TA-M* TA-M** — 36.1 133 ·· ·· — 12.9 170 ·· ·· — TA-M** ·· 29.2 48 — ·· ·· 5.2 58 — ·· ·· TA-M** ·· 34.9 63 ·· ·· ·· 1.6 62 ·· TA-M** ·· 21.2 132 ·· ·· < > < < < < Table 3. (continued) TA-YK* ·· TA-YK** — TA-YK** ·· TA-YK** ·· > > > > OK-M* 5.0 40 — OK-M*OK-M* 25.9 27 6.8 44 ·· OK-M* 5.2 77 ·· OK-M* 12.3 73 ·· ·· ·· 42.9 70 OK-M** 14.3 63 ·· OK-M** 4.7 86 ·· — ·· 38.7 93 ·· — ·· 25.8 89 OK-M** 15.5 58 — ·· ·· 22.2 54 — ·· ·· 49.2 59 — OK-M** 8.4 83 ·· ·· ·· 43.8 73 OK-M** 6.3 48 ·· ·· ·· 41.3 46 ·· ·· ·· 20.8 48 OK-M** 8.7 46 ·· ·· ·· 53.1 32 OK-M** 8.8 204 ·· ·· ·· 44.9 158 < > > < > > > > > > > > TA-M* TA-M** TA-M** TA-M** — 7.8 167 ·· ·· — 36.3 168 ·· ·· — 38.6 171 ·· ·· — > < < < 1 lingual cusp (LP2) groove patter Y TA-YK* ·· — 8.2 85 ·· TA-YK* ·· ·· 33.3 24 TA-YK* ·· ·· 42.9 35 TA-YK** — ·· 30.0 90 TA-YK** ·· ·· 20.0 45 ·· ·· ·· 19.4 67 ·· ·· TA-YK** ·· ·· 14.3 35 ·· ·· ·· 37.5 40 ·· ·· ·· 3.8 52 ·· TA-YK**TA-YK** ·· ·· TA-YK** TA-YK** > > > > > > > > > > < Deflecting wrinkle (LM1) trigonid cr Distal Odontome (U and LP1, 2) % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M % N TA-YK TA-M OK-M IE-JOMONMLF-JOMONRH-AINUNK-YAYOIYA-YAYOI1.661······ 0.0 0.0KOFUN 137 22KAMAKURAMI-MODERN 0.0 2.3 ··NK-MODERN ·· 68 44TA-YAYOI-KOFUN 5.7TA-MODERN 2.0 ·· 2.0 53 1.4 ··OK-MODERN 49 0.0 ·· 99 74 ·· 48 ·· — 0.8 ·· ·· 124 2.2 ·· ·· ·· 184 ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· ·· — ·· ·· ·· ·· ·· ·· ·· — Population OK-MODERN 21.3 94 TA-MODERN 25.3 87 TA-YAYOI-KOFUN 0.0 28 — NK-MODERN 25.0 20 KOFUN 33.3 36 KAMAKURA 35.7 14 NK-YAYOIYA-YAYOI 46.2 42.9 26 35 MI-MODERN 44.2 52 RH-AINU 25.0 28 IE-JOMONMLF-JOMON 14.3 6.9 7 58 ·· ·· ·· ·· 20.0 5 ·· ·· ·· 30.0 20 ·· ·· ·· 5.0 20 ·· ·· ·· OK-MODERNPopulation 72.6 186 TA-MODERN 89.0 118 ·· — NK-MODERN81.527······3.231······33.330······33.333······ TA-YAYOI-KOFUN 93.8 48 — ·· MI-MODERN 87.4 103 ·· ·· KAMAKURA 94.4 54 ·· ·· KOFUN 87.7 81 ·· ·· YA-YAYOI 85.5 55 ·· ·· ·· 15.4 52 ·· ·· ·· 41.5 53 ·· ·· ·· 13.1 61 NK-YAYOI 94.9 59 ·· ·· RH-AINU 84.0 75 ·· ·· ·· 17.8 73 ·· IE-JOMON76.913······15.632······25.020······45.231······ MLF-JOMON 94.3 140 ·· ·· Population 8 Y. MANABE ET AL. ANTHROPOLOGICAL SCIENCE

Table 4. Similarity to the migrant or native Japanese on each trait of the Tanegashima Islanders and Okinawa Main Islanders Similarity to migrant or native Japanese Trait Tooth Tanegashima Okinawa Yayoi–Kofun1 Modern2 Modern2 Shoveling UI1 Native Migrant Migrant Double-shovel UI1 Native Migrant Migrant Interruption groove UI2 Intermediate Migrant Migrant Mesial ridge UC Native Intermediate Intermediate Distal accessory ridge UC Native Migrant Migrant Hypocone UM2 — — — Cusp 5 UM1 Native Intermediate Native Carabelli trait UM1 — — — > 1 lingual cusp LP2 — — — Y groove pattern LM2 Native Migrant Migrant Cusp 6 LM1 Native Intermediate Intermediate 4-cusp LM2 Native Migrant Intermediate Deflecting wrinkle LM1 Native Intermediate Native Distal trigonid crest LM1 — — — Protostylid LM1 Native Migrant Migrant Cusp 7 LM1 — — — Odontome U, LP1, 2 — — — 1 Manabe et al. (2002), 2 present study. Native, similar to the native Japanese populations; Migrant, similar to the migrant Japanese populations; Intermediate, intermediate between the native and migrant Japanese populations.

in populations from the Southeast Asian archipelago and the continent (Turner, 1985, 1987, 1989, 1990, 1992), is located along the negative direction. The South Chinese and native Taiwanese populations are located in the intermediate area between sinodonty and sundadonty. In the analysis pertaining to East Asia, the migrant Japa- nese group is located in the sinodonty area, overlapping the North Chinese and Korean populations, while the native Jap- anese group is located in the sundadonty area. The modern populations of the two Nansei Islands are located in the intermediate area between sinodonty and sundadonty, near the native Taiwanese group. Compared to the modern population of Tanegashima Island, that of Okinawa Main Island is situated closer to the sundadonty area. The Tanegashima is- Figure 2. Biological distances between the modern Tanegashima landers of the Yayoi–Kofun period are included in the native Islanders and other Japanese populations using Smith’s MMD. Japa- nese populations are abbreviated as shown in Table 1. Japanese group in the sundadonty area at a distance from the modern populations of the two Nansei Islands. The second axis appears to express the separation between the New World–Siberia division and the East Asia di- The LM1 distal trigonid crest, UM2 hypocone, UI2 inter- vision in sinodonty because the New World and Siberia ruption groove, and UM1 cusp 5 contributed to the forma- group shows a negative value, while the East Asia group ex- tion of the second axis. The contribution of the UM2 hypo- hibits a positive value. cone, which is associated with tooth reduction, suggests that the second axis reflects temporal change in the Japanese Interpretation of the axes populations, as indicated by the MDS. Multiple regression analysis was applied to the MDS of 36 Asian-derived populations (Table 6, Figure 6). The eight Positioning in East Asia significant traits on the first axis, namely UI1 shoveling, UI1 Interpopulational distance analysis double shovel, UM1 cusp 5, LM2 Y-groove, LM1 cusp 6, MDS was applied to the MMDs of the 36 Asian-derived LM2 4-cusp, LM1 deflecting wrinkle, and LM1 protostylid, populations in order to clarify the positioning of the Nansei are predominantly the traits that have shown obvious differ- islanders in East Asia (Figure 5). On the first axis, typical si- ences between sinodonty and sundadonty in previous analy- nodonty, such as that found in North China, Korea, North ses (Turner, 1987, 1990; Manabe et al., 1991, 1992). The Asia, the New World, and Siberia, is located along the posi- first axis can be used to distinguish the populations exhibit- tive direction, while typical sundadonty, such as that found ing sinodonty from those exhibiting sundadonty. Vol. 115, 2007 DENTAL VARIATION IN NORTHERN NANSEI ISLANDS 9

Figure 3. Multidimensional scaling analysis based on Smith’s MMDs among the 12 Japanese populations. Japanese populations are abbreviated as shown in Table 1.

Table 5. Multiple regression analysis between the frequencies of traits and the coordinates of the multidimensional scaling analysis among the 12 Japanese populations Regression coefficient Multiple correlation Trait Tooth Axis 1 Axis 2 rP β P β P Shoveling UI1 0.981 0.000 ** 0.980 0.000 ** −0.057 0.398 Double-shovel UI1 0.955 0.000 ** 0.954 0.000 ** −0.066 0.517 Interruption groove UI2 0.762 0.020 * −0.069 0.755 −0.758 0.007 ** Mesial ridge (lingual) UC 0.513 0.253 0.512 0.107 −0.042 0.888 Distal accessory ridge UC 0.736 0.030 * 0.736 0.010 ** −0.003 0.988 Hypocone UM2 0.846 0.003 ** −0.196 0.300 −0.821 0.001 ** Cusp 5 UM1 0.838 0.044 * −0.405 0.053 −0.730 0.003 ** Carabelli’s trait UM1 0.465 0.333 0.458 0.155 −0.085 0.779 > 1 lingual cusp LP2 0.236 0.773 −0.034 0.919 −0.233 0.490 Y groove pattern LM2 0.556 0.189 −0.545 0.081 −0.106 0.710 Cusp 6 LM1 0.760 0.021 * 0.665 0.013 * −0.374 0.118 4-cusp LM2 0.940 0.000 ** −0.922 0.000 ** 0.195 0.119 Deflecting wrinkle LM1 0.837 0.004 ** 0.820 0.001 ** 0.157 0.412 Distal trigonid cest LM1 0.970 0.000 ** 0.036 0.670 0.969 0.000 ** Protostylid LM1 0.681 0.060 0.675 0.022 * −0.097 0.700 Cusp 7 LM1 0.251 0.745 −0.185 0.581 0.173 0.606 Odontome U-LP1-2 0.494 0.285 0.372 0.231 −0.328 0.287 ** Significant at P < 0.01; * significant at P < 0.05.

In contrast, the significance of the second axis is more beria and the East Asia divisions with regard to sinodonty, as complex. In total, seven traits, namely UC mesial ridge, revealed by the MDS, because four of the seven traits consti- UM1 cusp 5, LP2 > 1 lingual cusp, LM2 Y-groove, LM2 4- tute traits that express a considerable difference between cusp, LM1 distal trigonid crest, and LM1 protostylid, con- these divisions (Turner, 1985). tribute to form the second axis. It is suggested that the second axis reflects the separation between the New World–Si- 10 Y. MANABE ET AL. ANTHROPOLOGICAL SCIENCE

Figure 4. Regression coefficient (β) in multiple regression analysis between the frequencies of traits and the coordinates of the multidimensional scaling analysis of among the 12 Japanese populations.

Discussion theory was later supported by Newman and Eng (1947). A large-scale anthropological investigation involving the Inspection of the ‘Ainu–Ryukyuan common origin the- Ryukyuans was carried out by Kyugakkairengo (association ory’ in the modern period of nine academic societies) in 1971. Using the dental mor- The dual-structure model for the population history of the phology method, K. Hanihara and his colleagues found that Japanese (K. Hanihara, 1991) is now widely accepted. The Ryukyuans were relatively more similar to the Ainu than to native Japanese population was broadly distributed through- the Japanese main islanders (K. Hanihara et al., 1974; K. out the Japanese archipelago during the Jomon period (ap- Hanihara, 1976). Using blood genetic markers, Omoto et al. proximately 12000–2300 BP). Following this period, the mi- (1976) showed that Ainu and Ryukyuans are included in a grant populations entered North Kyushu and Yamaguchi and common group within worldwide variation, even though dispersed northward and southward in the Japanese archipel- there is no close relationship between them. Kouhara (1976) ago, in conjunction with some absorption of native genes in supported this theory based on variations in the functions of the Yayoi period (approximately 2300–1700 BP). Eventual- facial expression muscles. ly, a population with more migrant than native genetic back- Subsequently, the Ainu–Ryukyuan common origin theory ground evolved as characteristic of the general Japanese has been further supported by many studies in the fields of population. osteomorphology (Yamaguchi, 1982, 1992), human genetics According to this model, due to the northward and south- (Omoto, 1978, 1992; Omoto and Saitou, 1997), and dental ward dispersal of the migrant population in the Japanese ar- morphology (T. Hanihara, 1989, 1991, 1992a, b; chipelago, populations with more native than migrant genet- Matsumura, 1995). Furthermore, Hinuma (1985, 1986) sup- ic background was established in Hokkaido and the Nansei ported this theory by showing that Ryukyuans show more Islands, which are located at the northern and southern ends similarities to the Ainu than to the Japanese main islanders in of the Japanese archipelago, respectively. Based on this the distribution of the human T-cell leukemia virus (HTLV). viewpoint, the Ainu–Ryukyuan common origin theory be- The Ainu–Ryukyuan common origin theory was regarded comes a basic component of the dual-structure model for the as a likely theory by many researchers until recently; there population history of the Japanese. The Ainu–Ryukyuan currently exist several studies that challenge or are not sup- common origin theory was initially proposed on the basis of portive of the theory from the viewpoint of somatology the similarity identified by von Baelz (1911) in the morpho- (Suda, 1950; Terakado Y., 1976; Ikeda and Tagaya, 1980), logical features seen among the living populations, and this craniometry (Hsu, 1948; Ikeda, 1974; Tagaya and Ikeda, Vol. 115, 2007 DENTAL VARIATION IN NORTHERN NANSEI ISLANDS 11

Figure 5. Multidimensional scaling analysis based on Smith’s MMDs among the 36 Asian-derived Populations. Japanese populations are abbreviated as shown in Table 1.

Table 6. Multiple regression analysis between the frequencies of traits and the coordinates of the multidimensional scaling analysis among the 36 Asian-derived populations Regression coefficient Multiple correlation Trait Tooth Axis 1 Axis 2 rP β P β P Shoveling UI1 0.961 0.000 ** 0.949 0.000 ** 0.089 0.072 Double-shovel UI1 0.768 0.000 ** 0.770 0.000 ** −0.080 0.479 Interruption groove UI2 0.348 0.118 0.279 0.098 −0.235 0.161 Mesial ridge (lingual) UC 0.536 0.004 ** 0.218 0.149 0.471 0.003 ** Distal accessory ridge UC 0.500 0.009 ** 0.283 0.070 0.388 0.015 * Hypocone UM2 0.230 0.407 −0.186 0.282 0.154 0.373 Cusp 5 UM1 0.579 0.001 ** −0.426 0.005 ** 0.432 0.005 ** Carabelli’s trait UM1 0.366 0.093 0.047 0.775 0.359 0.034 * > 1 lingual cusp LP2 0.848 0.000 ** −0.014 0.833 0.849 0.000 ** Y groove pattern LM2 0.683 0.000 ** −0.378 0.006 ** −0.536 0.000 ** Cusp 6 LM1 0.554 0.002 ** 0.515 0.001 ** −0.252 0.091 4-cusp LM2 0.891 0.000 ** −0.768 0.000 ** 0.525 0.000 ** Deflecting wrinkle LM1 0.733 0.000 ** 0.692 0.000 ** 0.189 0.121 Distal trigonid cest LM1 0.566 0.002 ** 0.194 0.186 0.515 0.001 ** Protostylid LM1 0.699 0.000 ** 0.388 0.004 ** 0.548 0.000 ** Cusp 7 LM1 0.213 0.464 0.118 0.494 0.167 0.335 Odontome U-LP1-2 0.377 0.079 0.278 0.095 −0.281 0.092 ** Significant at P < 0.01; * significant at P < 0.05. 12 Y. MANABE ET AL. ANTHROPOLOGICAL SCIENCE

Figure 6. Regression coefficient (β) in multiple regression analysis between the frequencies of traits and the coordinates of the multidimensional scaling analysis of among the 36 Asian-derived populations.

1976; Pierusewsky, 1994, 1999, 2004; Doi et al., 1997; Doi, populations in Okinawa Main Island and Tanegashima Is- 1998, 2003, 2004; Asato and Doi, 1999), cranial nonmetric land, and that these populations were much more similar to traits (Mouri, 1986; Dodo, 1992; Dodo et al., 1998, 2000), the migrant Japanese as compared to the native Japanese mitochondrial DNA (Horai et al., 1996), and the human leu- population, even though they exhibited characteristics of kocyte antigen (HLA) gene (Hatta et al., 1999; Tokunaga, both Northeast and Southeast Asia. Therefore, this study 2003). Some arguments against the theory have been high- does not support the Ainu–Ryukyuan common origin theory lighted, particularly in the recent studies by Dodo (1992), with regard to the modern populations. Doi et al. (1997), Doi (1998, 2003, 2004), Dodo et al. (1998, Regarding the dental morphology of the modern Nansei 2000), and Asato and Doi (1999). islanders, the data derived from young, living inhabitants From the viewpoint of dental morphology, Manabe et al. possess the advantage of accuracy and statistically sufficient (1999a) performed a study using dental plaster casts of mod- quantity when compared with the data obtained from skele- ern, young, living Okinawa main islanders; their study sug- tal remains. Therefore, it is necessary to consider results de- gested that this population was much more similar to the mi- rived from the living inhabitants, such as those obtained in grant Yayoi and modern Japanese main island populations this study. Moreover, in skeletal morphology, the results of than to the Jomonese and Hokkaido Ainu populations. Sub- recent studies by Dodo (1992), Doi et al. (1997), Doi (1998, sequently, some evidence that the modern Nansei islanders 2003, 2004), Dodo et al. (1998, 2000), and Asato and Doi show a greater affinity to the migrant rather than to the na- (1999), who reexamined skeletal remains from the Nansei tive Japanese was additionally provided by Manabe et al. Islands, should be taken into account. These studies that are (1999b, 2001, 2003a). Furthermore, very similar conclu- based on sufficient material do not support the Ainu– sions were drawn by Higa et al. (2003) and Haneji et al. Ryukyuan common origin theory in the modern period. (2007), although their study differed from the aforemen- The two contrasting opinions are most likely to have re- tioned studies in some details. sulted from the misclassification of the modern Nansei Is- In the present study, it was noted that there was a resem- land population as exhibiting characteristics of the native blance between the tooth crown morphologies of modern Japanese group, based on either the condition of materials or Vol. 115, 2007 DENTAL VARIATION IN NORTHERN NANSEI ISLANDS 13 the method of data collection. This is because the modern (2003) have suggested that the decorative patterns on shell Nansei islanders exhibit not only many migrant Japanese products from Tanegashima during the Yayoi period were characteristics but also a few native characteristics, albeit to probably influenced by the Chinese continental cultures. a slight degree. This suggests a complicated population his- The archeological evidence summarized above appears tory in the Nansei Islands. inconsistent with the dental morphological similarity between the Tanegashima Yayoi–Kofun population and the Inspection of the ‘Jomonese–Ryukyuan common origin Jomonese. However, this evidence leads to the following as- theory’ in the prehistoric period sumptions. The native populations such as the Jomonese Characteristics of the Yayoi–Kofun period population in lived in Tanegashima from the Jomon, through the Yayoi, to Tanegashima the Kofun period. They absorbed the cultures of the Chinese Both Manabe et al. (2002) and this study indicate that the continent, North Kyushu, South Kyushu, and Kinki districts Tanegashima Yayoi–Kofun population considerably resem- during these periods; however, they were genetically barely bles the Jomonese and Ainu populations. We can assume influenced by the above-mentioned populations. Therefore, that this population was one of the local types of Jomon de- gene frequencies were maintained without any substantial scendants. change. The dental morphology of the Tanegashima Yayoi–Kofun population was first investigated by Yamada et al. (1970). Characteristics of the prehistoric Nansei islanders Their study clarified that this population exhibited many Near-complete skeletal human remains (c. 18000 BP) characteristics that were dissimilar to those of modern Japa- were excavated from Minatogawa in Okinawa Main Island nese main islanders, based on the evidence of a smaller tooth (Suzuki and K. Hanihara, 1982). The Minatogawa man was size, lower incidences of incisor shoveling, lower molar pro- similar to the Jomonese specimens in terms of skeletal and tostylid, and lower molar cusp 6. Thereafter, the resem- dental morphology (Suzuki and K. Hanihara, 1982; Turner, blance of this Yayoi–Kofun population to the Jomonese was 1987; Baba and Narasaki, 1991). The morphological simi- shown by T. Hanihara (1991a, b), Matsumura (1995), and larity between the two groups suggests that a population Manabe et al. (2002). The consensus from these dental mor- similar to the native Japanese dwelt on the Pleistocene phological studies supports the view that the Tanegashima Okinawa Main Island. Yayoi–Kofun population belongs to the native Japanese Recent studies have indicated that the Okinawa Shell- group. Through osteomorphological studies, Doi (1998, mound population had many characteristics in common with 2003, 2004), Asato and Doi (1999), and Nakahashi (2003) the Japanese main island Jomonese population (Matsushita, indicated that the Tanegashima Yayoi–Kofun population 1993, 2001; Doi, 1998, 2003, 2004; Asato and Doi, 1999), shows a greater similarity to the native Japanese than to the and also that the former were similar to the Tanegashima migrant Japanese; however, this group also shows a few Yayoi–Kofun population. Thus, to a certain extent, the characteristics that are dissimilar to those of the Jomonese. Minatogawa man and the Shellmound population in the In contrast to the results of the above studies, the osteo- Nansei Islands and the Japanese main island Jomonese logical study by Matsushita (2001) proffered a different showed many common characteristics. opinion, namely that the Tanegashima Yayoi–Kofun popu- However, it is suggested that the Minatogawa man either lation was a migrant population from the East Asian conti- became extinct or migrated to another region because of the nent, which differed from the type of Yayoi population rapid reduction of island areas in the latest Pleistocene found in North Kyushu and Yamaguchi. This opinion was (Takamiya, 1996, 1997). In order to elucidate the continuity based on the morphological discontinuity of the Tanega- between periods in the Okinawa Main Island and with re- shima Yayoi–Kofun population from the Jomonese of the spect to regional evolution, it is necessary to find hitherto Nansei Islands. However, since this morphological disconti- undiscovered skeletal remains within the time-span of ap- nuity was not excessive as compared to the difference be- proximately 10,000 years spanning the relevant periods. In tween the native and migrant populations, it cannot be de- any case, it is estimated that during the Okinawa Shell- nied that the Tanegashima Yayoi–Kofun population belongs mound and Japanese main islands Jomon periods, a common to a local native type. Japanese native population was distributed across both the According to the archeological study of pottery at the Nansei Islands and the Japanese main islands. Hirota site in Tanegashima (Morizono, 2003), the local The Ainu–Ryukyuan common origin theory suggested a Tanegashima culture was influenced by the North and South similarity between the Ainu and Ryukuan groups based on Kyushu cultures during the Yayoi period, and by the Kinki strong inheritance from the Jomonese, and it was further di- region (southern-central region of Honshu, the Japanese vided into the Jomonese–Ainu and Jomonese–Ryukyuan main island) cultures during the Kofun period. Many of the common origin theories. The validity of the latter in the ‘Gohoura’ shells collected in Okinawa as materials of shell modern period has been examined by using statistically suf- bracelets were found in Tanegashima, and many finished ficient data; however, its validity in the prehistoric period shell bracelet products were found in some Yayoi sites in have not been examined due to the comparative lack of pre- North Kyushu and Yamaguchi. Based on these findings, it historic materials excavated from the Nansei Islands. The is thought that Tanegashima was a relay point for the trade characteristics found in the extremely meager prehistoric of shell products between Okinawa and North Kyushu– materials excavated from the Nansei Islands (Matsushita, Yamaguchi during the Yayoi period (Kinoshita, 1996). In 1993, 2001; Doi, 1998, 2003, 2004; Asato and Doi, 1999) do addition, Kanaseki (1966), Kokubu (1978), and Kinoshita not contradict the Jomonese–Ryukyuan common origin the- 14 Y. MANABE ET AL. ANTHROPOLOGICAL SCIENCE ory in the prehistoric period. Timing of the substantial change The present study, which verified that a population simi- The results of the present study indicate that the substan- lar to the Jomonese inhabited Tanegashima in the Yayoi– tial change from the late Yayoi–Kofun period to the modern Kofun period, supports the Jomonese–Ryukyuan common period occurred in Tanegashima approximately between the origin theory. The Jomonese–Ryukyuan common origin the- 7th century AD and the present-day period. Assuming that ory is assumed to be not supported by the modern period ev- the shift from the Jomon to the Yayoi period in North idence, but it is accepted from the prehistoric period evi- Kyushu and Yamaguchi occurred approximately during the dence of the present study. These results do not contradict 3rd century BC, the populational changes must have begun the foundation of the dual-structure model for the population in Tanegashima ten centuries or more later than it did in history of the Japanese in Japan (K. Hanihara, 1991), al- North Kyushu and Yamaguchi. though a minor revision may be required. Assuming a southward gene flow from the Kyushu main island to the Ryukyu Islands (Asato and Doi, 1999; Doi, Factors involved in the substantial temporal change in 1998, 2003, 2004), it is geographically estimated that the Tanegashima change involving sinodontification in Tanegashima started Populations that influenced the substantial change before the Gusuku period in Okinawa Main Island. The re- The temporal change from the Yayoi–Kofun to the mod- sults of the present study show that sinodontification in ern period in Tanegashima reflected a considerable change Tanegashima appears to have occurred between the 7th and from typical sundadonty to an intermediate position between 10th centuries. sundadonty and sinodonty in East Asia, and also reflected a Matsumura (1998) suggested that the migrant populations considerable shift from the native to the migrant Japanese had influenced the central Honshu (west Kanto) district by populational conditions. Among the Japanese populations, the middle Yayoi period. Therefore, the genetic influence of the two (Yayoi-Kofun and modern) populations of Tane- the migrant population probably reached the central Honshu gashima were extremely different from each other. district within approximately 400 years since the first migra- None of the known regions showed a rapid and large tem- tion to North Kyushu and Yamaguchi (c. 300 BC). In con- poral change from sundadonty to sinodonty, as seen in the trast, the genetic influence of the migrant population on post-Jomon change in Japan (Turner, 1987, 1990; Manabe et Tanegashima was delayed and occurred considerably later al., 1989, 1991, 1992, 1997a, 2002). The change that oc- than it did in the central Honshu district. curred between the Jomon and the Yayoi periods in Japan The substantial delay in the genetic influence of the mi- cannot be explained without considering the influence of the grant population into Tanegashima cannot be attributed to migrant sinodont genes. Hokkaido Ainu inherited almost all geographical distance. This delay could also have resulted of the Jomonese characteristics, although they exhibited a from the Tanegashima islanders’ opposition to being geneti- slight change from the Jomon to the early modern period cally assimilated by the North Kyushu administration during (Manabe et al., 2002). The large temporal change in Tane- the Yayoi and Kofun periods. Another cause for this delay gashima took place most likely under influence of the mi- could be the genetic barrier maintained between North grant sinodont genes. Kyushu and Tanegashima by the South Kyushu prehis- The population that contributed to sinodontification after toric population. It is interesting that the South Kyushu prehis- the Yayoi–Kofun period in Tanegashima probably migrated toric population, which was opposed to the genetic assimila- from a geographically adjacent area of Tanegashima. Three tion mandated by the Japanese central administration until directions of migrations may be suggested geographically, a the 7th or 8th centuries, was described as Hayato in the sec- southward migration from the Japanese main islands, an ond- and third-oldest books on classical Japanese history— eastward migration from the Chinese continent, and a north- the Nihon-Shoki and Shoku-Nihongi (Ujitani, 1988, 2004). ward migration from Taiwan. Based on assessments of dental morphology and the MDS analysis of East Asian popula- Gene flow into the northern and central region of the Nansei tions shown in Figure 5, populations of these three can be Islands from the viewpoint of the geographical cline of mod- regarded as those that potentially caused sinodontification in ern populations Tanegashima. In the Ryukyu Islands, from Amami to Oki- The modern populations of both Okinawa Main Island nawa, Doi (1998, 2003, 2004) identified a discontinuity in and Tanegashima Island are similar to the migrant Japanese the skeletal morphology from the Shellmound to the Gusuku but not to the native Japanese, as shown in Figure 5. With re- period and suggested that the characteristics of the Shell- gard to details of the variation found in the Nansei Islands, mound population were altered by the population that mi- the modern Okinawa main islanders, as compared to the grated from the Japanese main islands after the Gusuku peri- Tanegashima islanders, have more similar traits to the native od. The results of the archeological study by Asato and Doi Japanese populations (Table 4), and the modern Okinawa (1999), which indicates that the Gusuku culture was estab- main islanders are synthetically closer to the native Japanese lished under the influence of the Kyushu island during the populations than are the Tanegashima islanders (Figure 5). period from the 10th to the 12th centuries, are consistent A geographical cline exists from the north to the central re- with the viewpoint of Doi (2004). gions of the Nansei Islands. Based on the above-mentioned factors, the substantial The large temporal change in Tanegashima is the key fac- change in Tanegashima is probably due to the southward mi- tor for the geographical cline in the Nansei Islands. It is sup- gration of populations from North Kyushu or Yamaguchi to posed that gene flow from the Kyushu main island to Tane- Tanegashima by land or sea routes. gashima in the north of the Nansei Islands spread southward Vol. 115, 2007 DENTAL VARIATION IN NORTHERN NANSEI ISLANDS 15 along the Nansei Islands after the Kofun period and eventu- the Ryukyus (ed.), The History and Medica1 History of Oki- ally reached Okinawa Main Island. nawa. [Okinawa no Rekishi to Iryou-shi]. University of Assuming that the genetic influence of the Japanese mi- Kyushu Press, Fukuoka, pp. 89–l03 (in Japanese). Doi N. (2003) Population history of the Ryukyuans as viewed grant population faded with the southward spread of gene from human skeletal remains. In: Foundation for the Promo- flow, the geographical cline can be interpreted to reflect the tion of Okinawa Prefectural Culture (ed.), Okinawa Prefec- magnitude of southward migration of the Japanese migrant ture History, Archaeology. Educational Board of Okinawa population from the northern to the central Nansei Islands. Prefecture, Naha, pp. 575–610 (in Japanese). Furthermore, the cline in the northern and central Nansei Doi N. (2004) Prehistoric and Gusuku people in Okinawa as Islands continues from North Kyushu, as shown in Figure 5. viewed from skeletal morphology. Anthropological Science, 112: 291. The continuous cline of dental variation observed from Doi N., Dodo Y., and Kondo O. (1997) Amami-Okinawans as North Kyushu to Okinawa Main Island via Tanegashima Is- viewed from cranial measurements. Anthropological Science, land suggests a common directional gene flow involving the 105: 79. Kyushu main island and Nansei Islands. Freeman M.F. and Tukey J.W. (1950) Transformations related to In this study, a part of the population history of the Nansei the angular and square root. Annals of Mathematical Statis- Islands was estimated using dental morphology, based on tics, 21: 607–611. Green R.F. and Suchey J.M. (1976) The use of inverse sine trans- the geographical cline in the northern and central Nansei Is- formations in the analysis of non-metric cranial data. Ameri- lands and the substantial temporal change on the island lo- can Journal of Physical Anthropology, 145: 61–68. cated in the northern gateway region of the Nansei Islands. Haneji K., Hanihara T., Sunakawa H., Toma T., and Ishida H. In the future, we anticipate that sufficient prehistoric skeletal (2007) Non-metric dental variation of Sakishima Islanders, materials will be excavated from the Nansei Islands to verify Okinawa, Japan: a comparative study among Sakishima and our interpretations. neighboring populations. Anthropological Science, 115: 35– 45. Hanihara K. (1976) The Okinawa Islanders as viewed from tooth Acknowledgments crown characters. In: Kyugakkai Rengo Okinawa Chosa Iin- kai (ed.). Okinawa: Nature, Culture, and Society. Kobundo, We would like to express our grateful thanks to the inhab- Tokyo, pp. 112–117 (in Japanese). itants of Tanegashima and Okinawa Main Island for their Hanihara K. (1991) Dual structure model for the population his- kind cooperation in the collection of dental plaster casts. We tory of the Japanese. Japan Review, 2: 1–33. Hanihara K., Masuda T., and Tanaka T. (1974) Affinities of dental wish to particularly thank Prof. Y. Shibata and Dr. H. Koga characteristics in the Okinawa Islanders. Journal of Anthro- (Graduate School of Medical Sciences, Kyushu University); pological Society of Nippon, 82: 75–82. Prof. Y. Tanaka, Prof. T. Nakahashi, and Prof. K. Mizoguchi Hanihara T. (1989) Comparative studies of geographically isolated (Graduate School of Social and Cultural Studies, Kyushu populations in Japan based on dental measurements. Journal University); and Prof. N. Doi (Graduate School of Medicine, of Anthropological Society of Nippon, 97: 95–107. Ryukyu University) for allowing us to investigate the prehis- Hanihara T. (1991) Dentition of Nansei Islanders and peopling of the Japanese Archipelago: the basic populations in East Asia, toric materials from Tanegashima and providing indispens- IX. Journal of Anthropological Society of Nippon, 99: 399– able information regarding these materials. We also wish to 409. thank Prof. C.G. Turner II for not only permitting us to use Hanihara T. (1992a) Negritos, Australian Aborigines, and the the data collected by him but also providing valuable infor- ‘Proto-Sundadont’ dental pattern: the basic populations in mation regarding the data. This study was supported in part East Asia, V. American Journal of Physical Anthropology, by Grant-in-Aid for Scientific Research (No. 13640715; No. 88: 183–196. Hanihara T. (1992b) Biological relationships among Southeast 16570194) from the Japan Society for Promotion of Science. Asians, Jomonese, and Pacific populations as viewed from dental characters: the basic populations in East Asia, X. Jour- nal of Anthropological Society of Nippon, 100: 53–67. References Harunari H. (2003) The beginning of the Yayoi period. Rekihaku, 120: 6–10 (in Japanese). Asato S. and Doi N. (1999) Where did the Okinawan people come Hatta Y., Ohashi J., Imanishi T., Kamiyama H., Iha M., Shima- from? [Okinawa-jin wa doko kara kitaka]. Border Ink, Naha bukuro T., Ogawa A., Tanaka H., Akaza T., Gojobori T., Juji (in Japanese). T., and Tokunaga K. (1999) HLA genes and haplotypes in Baba H. and Narasaki S. (1991) Minatogawa Man, the oldest type Ryukyuans suggest recent gene flow to the Okinawa Islands. of modern Homo sapiens in East Asia. The Quaternary Human Biology, 71: 353–365. Research, 30: 221–230. Higa T., Hanihara T., Sunakawa H., and Ishida H. (2003) Dental Berry A.C. and Berry R.J. (1967) Epigenetic variation in the variation of Ryukyu Islanders: a comparative study among human cranium. Journal of Anatomy, 101: 361–379. Ryukyu, Ainu and other Asian populations. American Journal Dodo Y. (1992) Population history of the Japanese as viewed from of Human Biology, 15: 127–143. the cranium: minor morphological variants, an alternative Hinuma Y. (1985) Natural history of the retrovirus associated with scale. Kagaku, 62: 245–249 (in Japanese). a human retrovirus. BioEssays, 3: 205–209. Dodo Y., Doi N., and Kondo O. (1998) Ainu and Ryukyu cranial Hinuma Y. (1986) The Tales of Viruses: Searching for the Origin nonmetric variation: evidence which disputes the Ainu– of the Japanese [Shin Uirusu Monogatari: Nihonjin no Kigen Ryukyus common origin theory. Anthropological Science, o Saguru]. Chuou kouron sha, Tokyo (in Japanese). l06: 99–120. Horai S., Maruyama K., Hayasaka K., Matsubayashi S., Hattori Y., Dodo Y., Doi N., and Kondo O. (2000) Flatness of facial skeletons Fucharoen G., Harihara S., Park K.S., Omoto K., and Pan I-H. of Ryukyuans. Anthropological Science, 108: 183–198. (1996) mtDNA polymorphism in East Asian populations, Doi N. (1998) Skeletal morphology of the Nansei Islanders. In: with special reference to the peopling of Japan. American Research Center for Comprehensive Medicine, University of Journal of Human Genetics, 59: 579–590. 16 Y. MANABE ET AL. ANTHROPOLOGICAL SCIENCE

Hsü H.L. (1948) Anthropologische Untersushungen uber die Ryukyu Islanders in Japan as viewed from dental morphol- Schadel der Riukiu-Insulaner. Bulletins of the Anatomical ogy. 3. The modern inhabitants of the Ishigaki Island. Anthro- Department of the National Taiwan University, 22: 227–330 pological Science, 109: 110. (in Japanese with German title). Manabe Y., Kitagawa Y., Oyamada J., and Rokutanda A. (2002) Ikeda J. (1974) Craniometry of Miyako Islanders, the Ryukyus. Population history in the Japanese Archipelago based on Journal of Anthropological Society of Nippon, 82: 150–160 regional and temporal variations of dental morphology. Den- (in Japanese with English summary). tistry in Japan, 38: 9–16. Ikeda J. and Tagaya A. (1980) Geographic variation of the anthro- Manabe Y., Ito R., Kikuchi N., Kitagawa Y., Oyamada J., Tsutsu- pometric measurements in the Japanese Islands. Journal of mida A., Kato K., Rokutanda A., and Kobayashi S. (2003a) Anthropological Society of Nippon, 88: 397–407 (in Japanese Population history of the Ryukyu Islanders in Japan as with English summary). viewed from dental morphology. 4. The modern inhabitants Imamura M. (2001) From the Jomon to the Yayoi period: prob- of the Miyako Island. Anthropological Science, 111: 423. lems and perspectives in determining absolute ages. The Qua- Manabe Y., Oyamada J., Kitagawa Y., Rokutanda A., Kato K., and ternary Tesearch, 40: 509–519 (in Japanese with English Matsushita T. (2003b) Dental morphology of the Dawenkou summary). Neolithic population in North China: implications for the ori- Kanaseki T. (1966) Culture of the Hirota site in Tanegashima gin and distribution of Sinodonty. Journal of Human Evolu- [Tanegashima Hirota-iseki no Bunka]. Fukuoka UNESCO, 3: tion, 45: 369–380. 39–52 (in Japanese). Matayoshi H. (1982) A physical anthropological study of the teeth Kinoshita N. (1996) Shell-work Culture in the Nansei Islands: the in the inhabitants of Ishigaki Island in Okinawa prefecture. Archaeology of the Shell Route [Nantou Kai-bunka no Ken- The Journal of The Kyushu Dental Society, 36: 1–25 (in Japa- kyu: Kai no Michi no Koukogaku]. Hoseidaigaku Shuppan- nese). kyoku, Tokyo (in Japanese). Matsumura H. (1995) A microevolutional history of the Japanese Kinoshita N. (2003) The Hirota site from the viewpoint of shell- peoples as viewed from dental morphology. National Science work accessories. In: Scientific Research Group of Hirota Museum Monographs, 9: 1–130 Site (ed.), The Hirota site in Tanegashima Island [Tanega- Matsumura H. (1998) Native or migrant lineage? The Aeneolithic shima Hirota Iseki]. Kagoshima Prefectural Museum of Cul- Yayoi people in western and eastern Japan. Journal of ture, Reimeikan, Kagoshima, pp. 329–366 (in Japanese). Anthropological Society of Nippon, 106: 17–25. Kokubu N. (1978) Archaeology of the shell route. Ethnos in Asia, Matsushita T. (1993) Human skeletal remains excavated from the 10: 5–12 (in Japanese). Gushikawajima sites. In: Educational Board of Izena-mura Kouhara Y. (1976) Facial expression in the Okinawa Islanders. In: (ed.), The Gushikawajima Sites [Gushikawajima Isekigun]. Kyugakkai Rengo Okinawa Chosa Iinkal (ed.), Okinawa: Izena-mura, Okinawa, pp. 215–244 (in Japanese). Nature, Culture, and Society. Kobundo, Tokyo, pp. 127–129 Matsushita T. (2001) Talking by Crania. Roots and Future of the (in Japanese). Japanese [Sharekoube ga kataru. Nihonnjin no ruutsu to Kruskal J.B. and Wish M. (1978) Multidimensional Scaling. Sage, mirai]. Nagasaki Shinbunsha, Nagasaki (in Japanese). London, (translated into Japanese by Takane Y., Asakura Morizono N. (2003) Chronology of pottery excavated from the shoten, Tokyo). Hirota site. In: Scientific Research Group of Hirota Site (ed.), Manabe Y. (1989) Dental characteristics in the Yami tribe, Taiwan The Hirota Site in Tanegashima Island [Tanegashima Hirota aborigine. The Journal of the Kyushu Dental Society, 43: Iseki]. Kagoshima Prefectural Museum of Culture, Reimei- 489–522. kan, Kagoshima, pp. 297–299 (in Japanese). Manabe Y. and Rokutanda A. (1992) The teeth from the pre-Inca Mouri T. (1986) Geographical and temporal variation in Japanese Chancay site. In: Kato K. (ed.), Morphological Studies of The populations as viewed from nonmetric traits of the skull. Human Skeletal Remains from The Chancay Site (Pre-Inca Ph.D. thesis, Kyoto University, pp. 1–79. Period) in Peru. Report of Research Project, Grant-in-Aid for Nakahashi T. (2003) Physical anthropological study of the skeletal Scientific Research (C) from the Ministry of Education, Sci- remains excavated from the Hirota site in Tanegashima, ence and Culture, Japan, pp. 56–81 (in Japanese). Kagoshima prefecture. In: Scientific Research Group of Manabe Y., Rokutanda A., Kitagawa Y., and Oyamada J. (1991) Hirota Site (ed.), The Hirota site in Tanegashima Island Genealogical position of native Taiwanese (Bunun tribe) in [Tanegashima Hirota Iseki]. Kagoshima Prefectural Museum east Asian populations based on tooth crown morphology. of Culture, Reimeikan, Kagoshima, pp. 281–294 (in Japa- Journal of Anthropological Society of Nippon, 99: 33–47. nese). Manabe Y., Rokutanda A., and Kitagawa Y. (1992) Nonmetric Newman M.T. and Eng R.L. (1947) The Ryukyu people: a biolog- tooth crown traits in the Ami tribe, Taiwan aborigines: com- ical appraisal. American Journal of Physical Anthropology, 5: parative with other East Asian populations. Human Biology, 113–158. 64: 717–726. Nichol C.R. and Turner C.G., II (1986) Intra- and Interobserver Manabe Y., Ito R., Kitagawa Y., Oyamada J., Rokutanda A., concordance in classifying dental morphology. American Nagamoto S., Kobayashi S., and Kato K. (1997) Non-metric Journal of Physical Anthropology, 69: 299–315. tooth crown traits of the Thai, Aka and Yao tribes of northern Omoto K. (1978) Genetic polymorphism of the Japanese. In: Ikeda Thailand. Archives of Oral Biology, 42: 283–291. J (ed.), The Japanese II. Anthropology 6 [Zinruigaku Kouza]. Manabe Y., Kitagawa Y., Oyamada J., Kato K., Ito R., Kobayashi Yuzankaku Press, Tokyo, pp. 217–263 (in Japanese). S., and Rokutanda A. (1999a) Genealogical positioning of the Omoto K. (1992) Some aspects of the genetic comparison of the Okinawa main islanders in East Asian populations based on Japanese. In: Hanihara K. (ed.), Japanese as a Member of the nonmetric tooth crown traits. Anthropological Science, 107: Ainu and Pacific Populations. International Research Center 85. for Japanese Studies, pp. 139–145. Manabe Y., Kitagawa Y., Oyamada J., Kato K., Ito R., Kobayashi Omoto K. and Saitou N. (1997) Genetic origins of the Japanese: a S., and Rokutanda A. (1999b) Regional variations in the partial support for the dual structure hypothesis. American Nansei Islanders of Japan based on nonmetric tooth crown Journal of Physical Anthropology, 102: 437–446. traits. 2. The modern inhabitants of theTanegashima Island Omoto K., Misawa S., and Ishimoto G. (1976) Hereditary blood and the Okinawa Main Island. Acta Anatomica Nipponica, markers of Okinawa inhabitants. In: Kyugakkai Rengo Oki- 74: 64. nawa Chosa Iinkai (ed.), Okinawa: Nature, Culture, and Soci- Manabe Y., Kitagawa Y., Oyamada J., Kato K., Ito R., Kobayashi ety. Kobundo, Tokyo, pp. 93–111 (in Japanese). S., and Rokutanda A. (2001) Population history of the Pietrusewsky M. (1994) Pacific–Asian relationships: a physical Vol. 115, 2007 DENTAL VARIATION IN NORTHERN NANSEI ISLANDS 17

anthropological perspective. Oceanic Linguistics, 33: 407– history of East Asia based on dental variation. American 429. Journal of Physical Anthropology, 73: 305–321. Pietrusewsky M. (1999) A multivariate craniometric study of the Turner C.G., II (1989) Teeth and prehistory in Asia. Scientific inhabitants of the Ryukyu Islands and comparisons with cra- American, 263: 70–77. nial series from Japan, Asia, and the Pacific. Anthropological Turner C.G., II (1990) Major features of sundadonty and sin- Science, 107: 255–281. odonty, including suggestions about East Asian microevolu- Pietrusewsky M. (2004) Multivariate comparisons of female cra- tion, population history, and late Pleistocene relationships nial series from the Ryukyu Islands and Japan. Anthropologi- with Australian aboriginals. American Journal of Physical cal Science, 112: 199–211. Anthropology, 82: 295–317. Sakura H. (1985) Pleistocene human fossil remains from Pinza- Turner C.G., II (1992) Microevolution of East Asian and European Abu (Goat Cave), Miyako Island, Okinawa, Japan. In: populations: a dental perspective. In: Akazawa T., Aoki K., Department of Education, Okinawa Prefectural Government and Kimura T. (ed.), The Evolution and Dispersal of Modern (ed.), Report on Excavation of the Pinza-Abu Cave. Depart- Humans in Asia. Hokusen-sha, Tokyo, pp. 415–438. ment of Education, Okinawa Prefectural Government, Naha, Turner C.G., II and Scott G.R. (1977) The dentition of living Eas- pp. 161–176 (in Japanese). ter Islanders. In Dahlberg A.A. and Graber T.M. (eds.), Oro- Suda A. (1950) The Ryukyu people as viewed from anthropology. facial Growth and Development. Mouton Publishers, The Minzokugaku Kenkyu, 15: 109–116 (in Japanese). Hague, pp. 229–249. Suzuki H. and Hanihara K. (1982) The Minatogawa man: the Turner CG, II, Nichol C.R., and Scott G.R. (1991) Scoring proce- Upper Pleistocene Man from the Island of Okinawa. Univer- dures for key morphological traits of the permanent dentition: sity of Tokyo Press, Tokyo (in Japanese). the Arizona State University dental anthropology system. In: Takamiya H. (1996) Initial colonization, and subsistence adapta- Kelley M.A. and Larsen C.S. (eds.), Advances in Dental tion processes in the late prehistory of the island of Okinawa. Anthropology. Wiley-Liss, New York, pp. 13–31. Bulletin of the Indo-Pacific Prehistory Association. Chang Ujitani T. (1988) The Chronicles of Japan: Translation into Con- Mai Paper 2, 15: 143–150. temporary Style. [Nihon-shoki. Zen-gendaigo-yaku]. Kodan- Takamiya H. (1997) From hunter-gatherer to farmer. a case in Oki- sha, Tokyo, Vol. 1-2. nawa. Nantou Kouko, 16: 27–46 (in Japanese). Ujitani T. (2004) Continuation of the Chronicles of Japan. Transla- Tagaya A. and Ikeda J. (1976) A multivariate analysis of the cra- tion into Contemporary Style. [Shoku-Nihongi. Zen-gen- nial measurements of the Ryukyu islanders (males). Journal daigo-yaku]. Kodansha, Tokyo, Vol. 1-3. of Anthropological Society of Nippon, 84: 204–220. von Baelz (1911) Die Riu-Kiu-Islander, die Aino und andere Terakado Y. (1976) The Okinawa Islanders as viewed from the kaukasierahnliche Reste in Ostasien. Korrespondenz-Blatt fingerprints. In: Kyugakkai Rengo Okinawa Chosa Iinkal der Deutschen Gesellschaft fiur Anthropologie, Ethnologie (eds.), Okinawa: Nature, Culture, and Society. Kobundo, und Urgeschichte, 42: 187–191 (in German). Tokyo, pp. 87–92 (in Japanese). Yamada H., Nakagawa K., Waniishi H., Sumiyoshi Y., and Tokunaga K. (2003) HLA and migration of Human. In Hinuma Y. Nakanishi H. (1970) Anthropological studies on the teeth of and Sakiya M (eds.), Anthropological Diversity in the Japa- the prehistoric Yayoi—ancients excavated from the Hirota nese Archipelago. Bensei Shuppan, Tokyo, Science of site, Tanegashima, Kagoshima prefecture. The Journal of The Humanity Bensei, 42: 4–9 (in Japanese). Kyushu Dental Society, 24: 34–62 (in Japanese). Tsuru H. (1970) An anthropological study on the physical traits of Yamaguchi B. (1982) A review of the osteological characteristics teeth in the inhabitants of Tanegashima Island. The Journal of of the Jomon population in prehistoric Japan. Journal of The Kyushu Dental Society, 32: 227–260. Anthropological Society of Nippon, 90 (Suppl.): 77–90. Turner C.G., II (1985) The dental search for Native American ori- Yamaguchi B. (1992) Skeletal morphology of the Jomon people. gins. In: Kirk R. and Szathmáry E (eds.), Out of Asia. Peo- In: Hanihara K. (ed.), Japanese as a Member of the Asian and pling the Americas and the Pacific. The Journal of Pacific Pacific Populations. International Research Center for Japa- History: Canberra, pp. 31–78. nese Studies. Nan’un-do, Kyoto, pp. 53–63. Turner C.G., II (1987) Late Pleistocene and Holocene population