RAW MATERIAL UTILIZATION FOR STONE IMPLEMENTS OF THE JOMON CULTURE IN 889 .

in Kanto, Japan- Part 2, Journal of the Faculty of raw materials of stone implements], in Kamaki Science sec. V, vol. iv, part 4: 395469. Tokyo: Yoshimasa Sensei Koki Kinen Ronshu: University of Tokyo. Kokogaku to Kanren Kagaku: 447-91. Okayama; TAKANEZAWA-MACHIKYOIKU IINKAI. 1982. Ishigami Kamaki Yoshimasa Sensei Koki Kinen Ron- lseki [The Ishigami site]. Tochigi: Takanezawa- bunshu Kankoukai. machi Kyoiku Iinkai. WARASHINA,TETSUO et al. 1978a. Keikou x-sen bun- TSUBOI,KIYOTARI. 1984. Introduction, in K. Tsuboi seki niyoru sanukite sekki no gensanchi suitei (ed.), Recent Archaeological Discoveries in (Ill) [Sourcing of sanukite stone implements by Japan: 1-4. Tokyo: The Centre for East Asian x-ray fluorescence analysis (III)], Koukogaku to Cultural Studies and UNESCO. Shizenkagaku 10: 53-81. WARASHINA,TETSUO. 1972. Sanukite no keikou x-sen 1978b. Keikou x-sen bunseki niyoru sanukite bunseki [X-ray fluorescence analysis of sanu- sekki no gensanchi suitei (IV) [Sourcing of sanu- kite], Koukogaku to Shizenkagaku 5: 69-75. kite stone implements by x-ray fluorescence WARASHINA,TETSUO & TAKENOBUHIGASHIMURA. 1973. analysis (IV]]. Koukogaku to Shizenkagaku 11: Keikou x-sen bunseki niyoru sanukite sekki no 33-47. gensanchi suitei [Sourcing of sanukite stone YAMAGATA-KENKYOIKU IINKAI. 1981. Higashikouya-B implements by x-ray fluorescence analysis], Iseki [The Higashikouya-B site]. Yamagata: Koukogaku to Shizenkagaku 6: 3342. Yamagata-ken Kyoiku Iinkai. 1975. Keikou x-sen bunseki niyoru sanukite sekki YAMAMOTO,KAORU. 1989a. Jomon jidai no sekki ni no gensanchi suitei (111 [Sourcing of sanukite tsukawareta ganseki oyobi koubutsu nitsuite stone implements by x-ray fluorescence analysis [Raw material utilization for stone implements (1111, Koukogaku to Shizenkagaku 8: 61-9. in the Middle Jomon period of Japan], Chigaku 1983. Sekki genzai no sanchi bunseki [Sourcing of Zasshi 98.7: 79-101. raw materials of stone implements], Koukogaku 1989b. Jomon jidai no sekki seisaku niokeru to Shizenkagaku 16: 51-89. sekizai no rivou nitsuite fLithic resource utili- 1986. Isoyamajou iseki shutsudo no sanukite zation for stone implements in the Jomon oyobi kokuyouseki ibutsuno sekizai sanchi bun- period], Tsukuba Daigaku Senshigaku Kou- seki [Sourcing of sanukite or obsidian used for kogaku Kenkyu 1: 45-96. stone implements in the Isoyamajyo site], in YAMANAKA-MACHIKYOIKU IINKAI. 1985. Yamanaka- Isoyamajo Iseki: 205-13. Shiga: Maibara-machi machi Uehara-A Iseki [The Uehara-A site]. Ishi- Kyoiku Iinkai. kawa: Yamanaka-machi Kyoiku Iinkai. 1988. Sekki genzai no sanchi bunseki [Sourcing of

Introduction Japanese prehistorians alike. In the latter case, Processes of acculturation and assimilation in research has emphasized the transition, begin- contact situations have been the subject of ning about 1000 BC, to the wet-rice-focussed considerable interest to North American and Yayoi (Akazawa 1981, 1986) (see TABLE1 for

Department of Anthropology, University of Toronto, Toronto, Ontario. Canada. GARY W. CRAWFORD & HIROTO TAKAMIYA scientific name English name Japanese name cultigens Cannabis sativum hemp asa Carthamus tinctorius safflower benibana, safurawa Cucurnis me10 melon meron, makuwa-uri Fagopyrum esculentum buckwheat soba Hordeum vulgare barley ou-mugi Linum usitatissimum flax ama Oryza sativa var japonicum rice kome, ine Panicum miliaceum common or broomcorn millet inakibi Perilla frutescens var crispa beefsteak plant shiso Setaria italica ssp. italica foxtail millet awa Sorghum bicoior sorghum morokoshi Triticum aestivum wheat, bread wheat ko-mugi Vigna radiatus var radiatus mung bean ketsuru-azuki V. angularis var angularis adzuki bean azuki Zea maize maize tou-morokoshi,tou-kibi weedy grainslgreens Chenopodium sp. chenopod akaza zoku Hordeum murinum wall barley mugi-kusa Polygonurn sp. knotweed tade ka P. densiflorum inu-tade P. sachalinense ou-itadori P. cuspidatum itadori Rumex sp. dock (sheep sorrel) gishi-gishi zoku Setaria italica ssp. viridis foxtail grass enokoro-gusa Setaria talica ssp. glauca foxtail grass kin-enokoro fleshy fruits Actinidia sp. silvervine matatabi zoku Aralia sp. udoltara-no-ki Cornus sp. dogwood mizu-ki zoku Empetrum nigrum crowberry gankou-ran Phellodendron amurense Amur corktree kihada Physalis sp. Chinese lantern plant houzuki Rubus sp. bramble ki-ichigo zoku Sambucus sp. cf. S. sieboldiana elder, elderberry niwatoko Solanum nigrum black nightshade inu-houzuki Vitis sp. grape budou fyama-budouj others Alliurn monanthum wild onion, leek hime-nira Arctium lappa great burdock gobou Castanea crenata chestnut kuri Corydalis ambigua corydalis Ezo-engosaku Juglans ailanthifolia walnut onigurumi Quercus sp. oak (acorn) donguri Rhus sp. sumac urushi zoku

TABLE1. Plant nomenclature used in text. LATE PREHISTORIC PLANT HUSBANDRY IN NORTHERN JAPAN

process of culture change took place, involving among other processes, the introduction of food production. By the 9th century AD,food produc- tion was more extensive in southwestern Hok- kaido than previously thought (Crawford & Yoshizaki 1987). At one Ezo site, Sakushu- Kotoni River, large quantities of carbonized plant remains are evidence of a high degree of dependence on millets, wheat, barley, beans, and several other upland crops, rather than wet rice. The genesis of this plant husbandry complex is not documented. In this paper we explore how the introduction of food production proceeded, relying mainly on plant remains data collected in the last few years in Hokkaido. We speculate that the devel- opment of the non-rice-based system was in part a response to selection pressures in Tohoku. The processes we examine took place well northeast of the northern Tokai boundary identified by Akazawa (1981; 1986) in his model of agricultural expansion in Japan. We have a number of objections to Akazawa's model, in particular to its applicability to northern Tohoku and Hokkaido. He relies on secondary evidence for plant procurement in FIGURE1. Overview of japan showing regions and the Chubu District and southern Tohoku, the northern limit of rice expansion delineated by including the Pacific coast, in order to propose Akazawa (1982:1621 two types of procurement systems: a narrow spectrum and a broad spectrum system plant nomenclature used in this paper). The (Akazawa 1986: 200). Without systematically- spread of agriculture to northeastern Japan is collected plant remains, adequate assessment of usually viewed as a northeastward progression the proposed contrast cannot be made, no of a frontier that reached northern Tohoku by matter what clustering of fishing- the Middle Yayoi (FIGURES1 & 2). However, the specializations or resource potential is situation is more complex than this, in our observed. Akazawa (1981; 1986) makes the view, and involves a spatial and cultural assumption, as do many others, that the spread dichotomy between Hokkaido and northern of agriculture is the spread of rice agriculture, Tohoku on the one hand and southern Tohoku- and that the appearance of Yayoi traits at sites southwestern Japan on the other. Furthermore, signals the change to a food-producing ' we interpret Ainu culture (as distinct from the economy. It is clear that in northeastern Japan, a Ainu biological population) of Hokkaido and non-rice-based food-production system arose. , Sakhalin to be an outcome of a long period of Any model of food-production origins in social interaction along this boundary. northeastern Japan must take this system into A broad frontier extended through south- account and explain its relation to the Yayoi western Hokkaido and northern Tohoku begin- culture. ning in the early half of the 1st millennium AD A model involving agricultural spread to (FIGURESI, 2 & 3). To the northeast of the western Europe has been proposed by Zvelebil frontier lived so-called native populations. To (1986). This model, while not explanatory, is the southwest were fully agricultural societies heuristic in that it orovides a wav to describe who, by AD 500, were articulated with the the processes that ensued during protohistoric Yamato state. At the end of the Final Jomon in contact periods in northeastern Japan. Zvelebil northern Tohoku and Hokkaido, a significant (1986) outlines a three-phase partition of agri- 892 GARY W. CRAWFORD & HIROTO TAKAMIYA

SOUTHWESTERN HOKKAIDO JAPANfTOHOKU Phase/Culture Subphase

MUROMACHI AINU 1333 -m-mmmm6m-mmmm-m-m (Transitional) KAMAKURA Satsumon

794 mmmm6mmm.mmmmE-mmm NARA EZO 710 mssmmm-mmmmm~mmmm-m Ezo-Haji

KOFUN

Southwestern FINAL JOMON

FIGURE2. Japanese 1000 periodization from 1000 BC t0 ADl568. cultural frontiers: first there is an availability precious few plant remains have yet been phase; this is followed by a substitution phase, retrieved from Tohoku. Most such remains in and finally by a consolidation phase. The period northeastern Japan are from Hokkaido but often of Yayoi-Jomon interaction (about 400 BC to AD from cultural contexts with similar counter- 300) is the phase of availability in northern parts in Tohoku. Tohoku and Hokkaido regions. The substitution and consolidation phases were initiated with Culture history the demise of the Hokkaido Jomon beginning The temporal emphasis of this paper is from about AD 300 and the subsequent appearance of about 1000 BC to the proto-historic period of a northern agricultural complex in the latter northern Tohoku and Hokkaido which began in - . half of the 1st millennium AD. the 7th century AD, and lasted until the early In order to examine the origins of food pro- 20th century in parts of Hokkaido. We also duction in Tohoku and Hokkaido in the late examine the relationship of these phases to .- prehistoric and early proto-historic periods, archaeologically- and historically-documented this paper is set out in two sections. First we periods in southwestern Japan. Of particular review the cultural history of Tohoku and concern are the Final Jomon, Tohoku Yayoi, Hokkaido in the context of subsistence change. Zoku (Epi-) Jomon, the Ezo-Haji and Satsumon The second section summarizes our recent subdivisions of Ezo, and Ainu. research on subsistence in Hokkaido. Previous researchers such as Akazawa (1981; 1986) were Final Jomon (Tohoku and Hokkaido) not able to use plant-remains data for models of The end of the Jomon in northeastern Japan agricultural development in northeastern involved the demise of two cultures: the Final Honshu. Considerable.data now exist, although Jomon in Tohoku and Hokkaido and the later LATE PREHISTORIC PLANT HUSBANDRY IN NORTHERN JAPAN 893

NORTHERN TOHOKU & HOKKAIDO F^?

Nishitsukigaoka

Chibutashinaizaw

Botanical Gardens Engineering Department

FIGURE3. Map of Hokkaido Island and YAMAGATA Tohoku (i.e. the + Kofun Tomb northern Honshu isiand) showing locations of sites mentioned in text.

Zoku Jomon, mainly in Hokkaido. The Final are somewhat distinct from each other at the Jomon ceased everywhere in northeastern Japan end of the Final Jomon. with the beginning of the Tohoku Yayoi, an The Final Jomon of Tohoku and Hokkaido is intensive agriculture phase intruding into often typified by the rich Kamegaoka site in Tohoku from the southwest. After this intrusion western Aomori Prefecture and from which the began, most of Hokkaido was occupied by the so-called 'Kamegaoka Culture' takes its name. Zoku Jomon while the Zoku Jomon had only a Local archaeologists today rarely refer to a peripheral presence in Tohoku. Kamegaoka Culture', however. Sites of this The Final Jomon period in Aomori Prefecture period are usually referred to as 'Obora' phase. may be divided into seven phases (Suzuki Nevertheless, the Final Jomon in this region is 1986). They are called from the earliest to the quite distinctive. For example, Okamoto (1966: latest: Obora B, Obora BC, Obora Ci, Obora Cz, 420) points out that 'the Kamegaoka culture had Obora A, Obora A' and Sunazawa. In southern reached the highest level of culture which can Tohoku, Obora is succeeded by the be achieved by hunting and gathering Fukurashima phase whereas in northern populations'. Setting aside any discussion of Tohoku, Fukurashima; although present -is the veracity of this statement, there is little difficult to place chronologically. The Sima- doubt that the Obora-phase Jomon was quite zawa phase usually closes out the Final Jomon successful. Obora material culture is char- there (Ito 1984: 23; Suto & Sahara 1987: 201-4). acterized by thin-walled, well-made in a In other words, northern and southern Tohoku variety of forms with zoned, cord-marked 894 GARY W. CRAWFORD & HIROTO TAKAMIYA motifs; stone technology which included sites in this region (Suzuki 1986: 83-4; Suto & polished stone batons; a variety of ; Sahara 1987: 206; Katoh & Suto 1986: 168; Ito pottery figurines; and a substantial technology 1984: 1; Takahashi 1986: 35). It0 (1984: 23) based on perishable materials evidenced at notes that Yayoi beads are found in Obora-A several wet sites. Obora-phase people were contexts at the Kamegaoka site. In southern . involved in exchange with outlying areas. For Tohoku, on the other hand, it is not clear that example, artefacts made from shell which is rice was known to the latest phase of the Final , available only in southwestern Japan have been Jomon there, but likely it was. We may postulate - recovered from Obora sites. Obora pottery has that gardening, adoption of rice and exchange been reported from as far south as the Kinki with Yayoi peoples were components of a region (Hayashi 1976: 191-2). Katoh & Suto mechanism that was maintaining a distinctive (1986: 168) feel that this exchange network northern socio-economic system in the face of provided an avenue whereby Yayoi cultural Yayoi expansion in the southwest. Although a elements were introduced to Tohoku. Tohoku Yayoi is recognized after the Final Final Jomon subsistence is assumed to have Jomon, it is distinguished only by rice agri- relied upon hunting, gathering and fishing. culture and a material culture that is Yayoi-like However, evidence indicates that plant- hus- but is still similar to the preceding Jomon. A bandry was part of subsistence activities in the local identity is maintained. Early and Middle Jomon of Hokkaido (Crawford 1983) as well as the Early Jomon of southwes- Tohoku Yayoi tern Japan (Tsukada et al. 1986). We have no The presence of Yayoi culture in northern reason to believe that plant-husbandry would Tohoku was not generally accepted until rice not have continued into the Late and Final production was confirmed in the archaeological Jomon, assuming it was present earlier. A record there. Certain traits at sites were similar proper understanding of Final Jomon subsis- to Yayoi traits, but the assemblages still resem- tence must await systematic collection of both bled the Final Jomon, particularly the Final plant and animal remains, in particular by Jomon pottery. Ito (1984: I), as early as the flotation. Pollen analysis, however, suggests 1950s, believed that rice agriculture began in that buckwheat husbandry was carried out Tohoku immediately after the Jomon-period. (Yasuda 1988: 53-4; 16-4). The late Final His argument was based first on pottery sherds Jomon peoples, who were contemporaries of the from the Inakadate site. The pottery was similar Early Yayoi, were undoubtedly aware of rice. to western Japanese Yayoi pottery rather than Large quantities of carbonized rice remains Final Jomon pottery types (1984: 2-3). Many have been reported from sites as far north as the researchers believed that this region was too Kamegaoka locality itself in association with cold for rice agriculture and that the Yayoi Obora A pottery (Sato 1984). Obora-A-phase elements were simply a consequence of trade pottery with rice impressions is also reported (1984: 1).In 1955, Ito recovered two carbonized (Tasaki 1986: 9; It0 1986: 351-2; Sasaki 1983: rice grains and several pieces of pottery with 53; Ito 1984: 21-2; Katoh& Suto 1986: 167).Rice rice impressions from the Tareyanagi site (1984: paddies have recently been reported from the 3). However, rice-impressed pottery sherds and -- Sunazawa-phase type site. The full impli- rice grains were not enough evidence to confirm cations of the discovery have not yet been the presence of rice agriculture in northern explored but at least one archaeologist thinks Tohoku. that this may indicate an experimental stage of It was not until 1983, when actual rice paddy rice agriculture in Aomori Prefecture fields were found at Tareyanagi, that archaeo- (Murakoshi 1988: 211-13). Yayoi glass beads. logists were convinced that rice was cultivated and spindle whorls are associated with the in northern Tohoku as early as the Inakadate Sunazawa phase as well (Hayashi 1986: 118). phase (Komoto & Yamazaki 1984: 41; Ito 1984: Moreover, Ongagawa-type pottery, the earliest 7; 1985: 350; Katoh & Suto 1986: 170; Tasaki Yayoi pottery in western Japan - the spread of 1986: 19).As of 1986,654 paddy fields compris- which has been used as evidence for the diffu- ing approximately 3967 sq. m had been sion of rice agriculture throughout Honshu - unearthed. Moreover, the paddy fields are well has been reported from several Sunazawa phase organized. Most of them exhibit either regular LATE PREHISTORIC PLANT HUSBANDRY IN NORTHERN JAPAN 895 square or rectangular shapes, ranging in size Katoh & Suto 1986: 167). This sequence seems from 2.98 sq. m to 18 sq. m, averaging about to have been strongly influenced by the Kanto 10 sq. m, In addition, the site occupants also region Yayoi (Katoh & Suto 1986: 166). On the installed waterways along the paddy fields other hand, the northern Tohoku Yayoi retained (Komoto & Yamazaki 1984: 41; Ito 1984: 7; 1985: Jomon elements despite influence from the 346-50). Sasaki (1983: 46) reports that south (Ito 1966: 211-14). The of Fujiwara, who conducted plant opal analysis of northern Tohoku may be divided into five the site, concludes that rice agriculture must phases: Nimaibashi, Utetsu 11, Inakadate, Nen- have been an important aspect of the Tareyanagi butsuma and Chokaiyama (Itoh & Suto 1982: subsistence system for a considerable period of 361). According to Katoh & Suto (1986: 167),the time. Nimaibashi pottery type is the progenitor of the The processes and results of the expansion of Esan pottery type and is closely related to the rice production throughout Japan are signifi- Esan culture, the earliest Zoku-Jomon phase on cantly different between the southwest and Hokkaido (Ito 1966: 211-14). It seems that the northeast (Akazawa 1986: 163). As Kanaseki & main reason to separate Esan from Nimaibashi Sahara (1978: 20-21) point out, the spread of the is not by ceramic attributes, but by the apparent Early Yayoi from Kyushu to a line near modern lack of rice agriculture in association with Esan Nagoya (Akazawa's Tokai boundary) was rapid. pottery. Regional complexes continued to One type of pottery, Ongagawa, is shared maintain local identities which, in the case of throughout Japan west of this line at this time. northernmost Tohoku and southernmost Hok- Not until the Middle Yayoi does Yayoi become kaido, are sometimes difficult to distinguish in established in Tohoku (after 100 BC). This the archaeological record. occurs during what Akazawa (1986: 163) calls a In southern Tohoku, rice agriculture was second phase of rice agriculture expansion. likely established by the Matsukata Gakoi Yayoi Contrasting the first, rapid spread of rice agri- phase, which is contemporaneous with Ina- culture with the second phase of northeastward kadate of northern Tohoku (Hayashi 1986: 114; expansion of the Yayoi, Akazawa sees an initial Katoh & Suto 1986: 168-70; It0 1966: 208-11; resistance to food production by narrow spec- Tasaki 1986: 19). Rice paddies have been trum food procurers, particularly along the reported from the Tomizawa site, Miyagi pre- Pacific coast (1986: 200). Along the boundary of fecture, for example (Katoh & Suto 1986: 170). the Tokai district, rice appears to move first During this period, the sites tend to be distri- through inland areas where people were more buted on flat alluvial planes less than 10 m receptive to agricultural innovation. Akazawa above sea level (Ito 1966: 208-11). Furthermore, hypothesizes that this is due to their experience out of 75 sites with rice grain impressed pottery with 'intensive plant collecting and/or plant sherds, 59% of the sites are located in southern cultivation' (1986: 200). In the areas he is TO~O~U(It0 1985: 337-42). concerned with, there are no substantial collec- If we are to assess the origins and develop- tions of plant remains to support or disconfirm ment of the non-rice-based food production the suggestion. Finally, Akazawa does not indi- system documented in Hokkaido for the 1st cate whether his model is applicable to Tohoku millennium AD, it is important to know what in general and northern Tohoku specifically. crops, other than rice, may have been grown We argue that it is not. In fact, as has already during the Tohoku Yayoi. To date, little work been pointed out, rice had made its way into has been done to examine the range of crops Tohoku as early as the Obora-A phase of the used by the Tohoku Yayoi - we know more Final Jomon, contemporaneous with the Early about Yayoi crops in southwestern Japan. Even Yayoi. so, the lack of detailed quantification of plant The dichotomy between southern and remains in most site reports makes it difficult to northern Tohoku, first recognized during the assess the relative abundance of plant taxa from late Final Jomon (post-Obora), is apparent Yayoi sites. Grains reported from southwestern during the Yayoi period as well. In southern Japan include rice, wheat, barley, millet and Tohoku, the Fukurashima phase is succeeded beans (Kasahara et a). 1986: 101--6; Crawford by the Tanakura, Matsukata Gakoi, Enda and n.d.). However, one of us (Crawford) has exam- Sakurai phases (Ito 1966: 208-11; '215-19; ined a number of collections of Yayoi plant 896 GARY W. CRAWFORD & HIROTO TAKAMIYA remains assemblages, and it is clear that wheat barnyard millet in Iwate Prefecture (Yabuno is quite rare. Barley and millet are usually not 1987: 488), indicating that barnyard millet had recovered in much quantity either. There become an important part of the economic appears to be no Yayoi crop assemblage in system by this time. The suggestion of substan- southwestern Japan during the Yayoi analogous tial non-rice agriculture from the Yayoi period to those assemblages being recovered in Hok- and later in Tohoku is, in our view, quite viable kaido. Phytolith analyses provide some sug- in light of later assemblages of plant remains in gestion that rice was not the only crop grown by Hokkaido (Crawford & Yoshizaki 1987). the Tohoku Yayoi. Panicum sp. phytoliths, for example, are reported from the Babano I1 site in Zoku Jomon fffokkaido) Iwate (Ito 1984: 23). Rice is notably absent. A later Jomon variant, the Zoku Jomon, Barnyard and foxtail millet played an impor- appears in southwestern Hokkaido by 150 BC. tant role in Tohoku subsistence until recently Most radiocarbon dates for the Zoku Jomon in (Yabuno 1987). Barnyard millet not only pro- Hokkaido are older than AD 300 (FIGURE6). vides higher protein levels (twice that of The earliest Zoku Jomon stage is contem- polished rice) and can be stored for long poraneous with the Yayoi and late Final periods, it is also better adapted to areas in the Jomon. The earliest phase is characterized by Tohoku region where rice agriculture is not Esan pottery, first recognized at the Esan site suitable (Yabuno 1987: 486). Yabuno (1987: on the Oshima Peninsula (see Yoshizaki 1984). 486) points out that barnyard millet provided Non- acculturated Zoku Jomon peoples on insurance against rice crop failure in some areas Hokkaido continued to survive, but not at the and that in Iwate Prefecture, barnyard millet level of complexity of the preceding Final was a substitute and an extender for rice until Jomon. A North American analogue to this the 1940s. Until the Meiji period in Shimokita, Final Jomon- Zoku Jomon transition might be Sanbongihara, and in the mountainous areas of the Hopewell- Late Woodland transition in Kitakami, barnyard millet was the major crop the midwestern United States. Hopewell was a grown in paddy fields (Ichikawa 1985: 111).In ranked society with plant husbandry based on Ichikawa's opinion (1985: ill), the interpreta- crops other than maize. The complex and per- tion that paddy fields always mean rice produc- haps competitive cultural system involving a tion does not necessarily apply until the Late broadly based interaction sphere declined (the Meiji period in Tohoku. In northern Iwate 'Hopewell decline'), to be followed by a less Prefecture where cold weather prevents rice complex set of cultures, called the Late production, barnyard millet, foxtail millet, Woodland. The Late Woodland is a period wheat, and soy bean were grown instead of rice with varying degrees of dependence on maize, (Yabuno 1987: 429). depending on location. Although agricultural society in northeastern Oddly, few living sites are known for the Japan rapidly evolved after the mid Middle Zoku Jomon in Hokkaido. Most sites are Yayoi period (Suto & Sahara 1987: 212-13), the cemeteries or cemetery- (e.g. Esan, crops may not have been the same as, or, at least, Poplar Namiki). The best known non-cemetery may not have been grown in the same propor- site is K135, discussed later in this paper. tions as in southwestern Japan. During the Early Interaction between Zoku Jomon and south- Yayoi, rice, barley, wheat, millet (foxtail, western Japanese groups is seen in exchange - common, and barnyard), and bean appeared in that brought metal tools, glass beads, and pot- the southwest, but rice dominates the plant tery as far north as Hokkaido, while Zoku Jomon remains assemblages from Middle and Late pottery is the only recognizable Hokkaido and Yayoi sites in the southwest. There is evidence northern Tohoku product on Tohoku Yayoi to suggest that barley, wheat, millet, and bean sites. Two reports of cultigens on Zoku Jomon were more common than rice in northeastern sites, as well as cultigen pollen indicate that Japan during the Yayoi (and perhaps during the perishable products were moving through the following ) (Terasawa 1986: 24-7). system as well (Crawford 1987; Yamada 1986; Three members of the Fujiwaras, a powerful 1987). Thus, cultigens, in addition to metal, family during the 12th century in northeastern glass, and Yayoi pottery were available to Zoku Japan, were buried not with rice but with Jomon people from their contemporaries - the LATE PREHISTORIC PLANT HUSBANDRY IN NORTHERN JAPAN

Tohoku Yayoi, thus making this a stage of populations. As a result, northern Japan was a availability in Zvelebil's (1986) terms. relatively hostile environment for the state- sanctioned expansion of Japanese interests. Historic period in Tohoku Aboriginal peoples south of present-day Akita The Japanese began intruding into Tohoku soon and Iwate often lived amicably along side their after the Kofun period began in southwestern Japanese counterparts; however, in Akita and Japan. Each successive regime of the Yamato Iwate and further north lived 'a very large state made its mark in northern Honshu. As number of turbulent people, including Japanese early as the Heian period, the regions beyond who had thrown in their lot with the Ainu' the frontier were often considered havens for (Sansom 1958: 104). Palisaded NaraIHeian- people to escape taxation and control of the period communities such as the c. AD 800 state (Sansom 1958).There was no single conso- Hottanosaku site were built by Japanese in lidation of the Japanese state's presence in Tohoku. The people they were encountering are Hokkaido until the 1600s. Many southern-style called Emishi (Yemishi) or Ezo in early records kofun (burial mounds) are known in Tohoku such as the Nihongi (Aston 1896) and they are (FIGURE3), but little more is known about this usually assumed to be Ainu. In AD 642, the first early state's presence in the region. The appear- mention of indigenous peoples - in Akita Pre- ance of kofun in Tohoku corresponds with the fecture - was made (Peng & Geiser 1977). An widespread appearance of Haji pottery (Hajiki) attack by the Japanese is recorded in AD 658, and in the Middle Kofun that is characteristic of the many Ezo died (Chamberlain 1887: 45). Other southwestern Japanese Kofun period (Okada clashes include an attack on Tagajo, a Japanese 1984). Early Kofun Hajiki is not common in fortification near Sendai, Miyagi Prefecture, Tohoku. Yayoi pottery in Tohoku ceased to be which was overrun by Ezo in the late 8th ' made during the Tohoku early Kofun but just century AD. At the end of the following century, how and why this happened is unknown. In in AD 875, 80 boat-loads of Ezo raided Akita northern Aomori Prefecture, the Zoku Jomon (Peng & Geiser 1977). persisted for a short time, but Hajiki is quite Many of the earliest accounts depict more common elsewhere in Aomori (Okada 1984). peaceful relationships. The Nihongi describes Okada feels that by the late Kofun (late 6th efforts by the Japanese to peacefully bring the century AD), Hajiki developed a local Tohoku Ezo into submission. For example, in the appearance with, for example, dark interiors summer of AD 658, Abe no Omi is said to have (Okada 1984). commanded 180 ships against the people of In a period of less than 200 years, northern what is now Akita Prefecture (Aston 1896: Tohoku peoples had come under influence from XXVI, 252). The Ezo apparently submitted to the Yamato state. The relationship between the Abe no Omi's forces without a struggle. During state and Tohoku peoples is unknown. Okada the same campaign,"there is mention of the (1984) feels that the change from Yayoi to a 'Yemishi of Watari no Shima' who were pro- culture under some form of Yamato influence vided a great feast by Abe at an unknown was relatively peaceful. The northern culture location known as 'Arima' (Aston 1896: XXVI, was, in all likelihood, the local Yayoi society 252). Aston (1896: XXVI, 252) believes that acculturated to some extent by contact with, Watari no Shima is the island of Yezo, or and influence from, Kofun-period Japanese. Hokkaido as it is now known. In AD 658,200 Ezo ,A There is little evidence for extensive migration bearing presents reportedly appeared before the of Japanese into northern Tohoku, although empress (Aston 1896: XXVI, 254). These may some migration likely took place. After the have been Ezo from Akita who had submitted to Kofun period ended in southwestern Japan, the Abe no Omi. Nara period Japanese made their own pio- At least three groups of Ezo were known to the neering efforts in Tohoku. The seat of Japanese, and one, the Nigi-Yemishi, brought government at that time was in the Nara Basin. annual tribute to the Yamato court (Aston 1896: Kanto, the current seat of Japanese government, XXVI, 262). This Ezo group is thought to have was considered a frontier territory. been the most 'civilized', while the Ara- Post-Kofun-period Japanese incursions to the Yernishi, and Tsugaru-Yemishi were less north met with great resistance from local 'civilized'. The Hokkaido Ezo are not men- 898 GARY W. CRAWFORD & HIROTO TAKAMIYA tioned in this reference. The identification of Tohoku before AD 600, rice production appears three native groups may not only be a measure to have been risky with occasional crop failures of the degree of hostility each group showed forcing the Japanese to retreat occasionally from toward the Japanese, but it may also be a Tohoku (Okada 1984). Sansom speculates measure of geographic distance from the Japan- (1958: 106) that the 'Ainu', as he calls the ese. Furthermore, the Final Jomon and Yayoi of natives of Tohoku, were able to hold out for so northern Tohoku show regional variation as long in Tohoku due to the assistance of local well and the possibility that their ethnohistoric pioneers who resisted the intrusion of the counterparts are recorded in the Nihongi central government. Furthermore, the Ezo's through the Japanese recognition of distinct relative success in the area is probably in no native groups in the 7th century AD should small way due to a food-production system remain open. based on barley, wheat, millets, and bean In AD 659, the Ezo suggested that a seat of among other temperate crops, rather than rice. government be established in Shiribeshi, Plant remains from the 9th century AD Ichinohe thought to be on the west coast of Hokkaido, Baipasu site in Iwate Prefecture are predomin- near the mouth of the Ishikari River (Aston antly wheat (Sato 1986). At Hottanosaku in 1896: XXVI, 260). Abe no Omi apparently did Akita Prefecture, one of us (Crawford) was able so. Given that Shiribeshi is, indeed, on Hok- to observe barley, millet, beans and hemp in a kaido, the expedition provides us with the first sample from a house floor. written account of a Japanese visit to Hokkaido. From the beginning of the Kofun period The encounter with the Ishikari Ezo seems to nearly 800 years passed before Tohoku and have been a cooperative one. The Ezo were said parts of Hokkaido were under Japanese control. to have asked for assistance to thwart an attack By the mid 7th century AD, only three centuries by a group Aston calls the Su-shen (presumably after the end of the Yayoi in southwestern the Ju-chen of Manchuria). Abe no Omi, with as Japan, the Japanese were encountering power- many as 3000 troops according to one estimate ful and, at times, hostile Ezo. Culturally, we can (Yoshizaki 1984), defeated the Ju-chen and postulate that 7th-century populations in presented 49 captives to the empress (Aston Tohoku were Kofun-acculturated descendants 1896: XXVI, 260 & 264). The legacy left by the of Yayoi who, to a large extent, were accultu- government post at Shiribeshi is unknown. rated Obora Final Jomon. The Ezo appear to Little about Ezo subsistence was recorded in have maintained their identity and territorial the Nihongi. During an AD 659 Japanese visit to hold on northern Tohoku and southern Hok- the Tang emperor of China, two Ezo, a man and kaido in part because of an economic system a woman, accompanied the envoys (Aston based not on wet rice or only foraging, but on a 1896: XXVI, 261). The Tang emperor asked if dry-food-production system that was relatively there were five kinds of grain in their (Ezo) secure in the northern regions. country. The envoys answered that there were not. In the same discussion the envoys said that Archaeological Ezo: the Ezo-Haji the Ezo dwellings were 'under the trees in the By AD 800, the indigenous Ezo appear to have recesses of the mountains' (Aston 1896: XXVI, been extirpated in Tohoku (Takakura 1960) - -- 262). The archaeological record of 7th-century assimilated into Japanese society in Tohoku Hokkaido shows this statement to be mislead- and/or killed, but there is evidence of emigra- ing, considering that Ezo hamlets consisted of tion of this population to Hokkaido at the end of "* substantial houses in locations other than the Zoku Jomon. This corresponds to the recesses of mountains. The Ezo were not speak- decline of the Zoku Jomon in Hokkaido. The ing for themselves, so we have a report of the Ezo-Haji period has only recently been recog- envoys' conversation about what they thought nized as a distinctive phase in Hokkaido and as of Ezo lifeways or of what they thought the Tang a result, its extent and meaning are not well emperor should hear. In view of these observa- understood (Yoshizaki 1984). This phase likely tions, we suspect that other details of Ezo life begins earlier, with some overlap between the coming from this conversation in China cannot Zoku-Jomon and Ezo-Haji likely occurring be considered of much value. (FIGURES2, 6 & 7). Ezo-Haji appears to have Although rice agriculture is known in been distributed primarily in the Ishikari Plain LATE PREHISTORIC PLANT HUSBANDRY IN NORTHERN JAPAN 899 south into the Oshima Peninsula (Yoshizaki the Nishitsukigaoka site in the northeast 1984). The northern coastal areas of Hokkaido (Kohno 1959; Iwasaki 1966). The cultigens from were occupied by the Okhotsk Culture. The these sites were fortuitous discoveries so the better known Satsumon terminology is now extent to which the site occupants were usually reserved for the latter half of this period involved in plant husbandry is not known. We (Yoshizaki 1984; Crawford & Yoshizaki 1987). have obtained flotation samples from one other Satsumon sites are distributed as far north in Satsumon site in northern Hokkaido, Omusaro- Hokkaido as the coast of the Sea of Okhotsk. C, but the samples contained no cultigens. The name 'Ezo-Haji' is a metaphor for the Interestingly, barley or wheat, broomcorn strongly-acculturated form that this early phase millet, and foxtail millet were recovered from exhibits. The Ezo-Haji terminology implies that flotation samples from the Futatsuiwa site, an this pottery is not the true Haji ware of the Okhotsk culture occupation near Abashiri. The Japanese. It was locally produced , site also contained some Satsumon pottery. For often turned on a potter's and bearing the moment we cannot eliminate exchange as strong morphological resemblance to the origi- an explanation for the presence of cultigens in nal Haji pottery. House forms, however, are the northern extremes of Hokkaido. The extent similar to Kanto and Tohoku Kofun-period to which food production spread into northern dwellings. They are semi-subterranean, rec- Hokkaido remains an important issue in Hok- tangular, and have one or occasionally two kaido . ovens in one wall. By AD 1200 this house form is found all over Hokkaido. Grey Sue stoneware Ainu (Sueki) pottery made in Tohoku, fired in kilns, With the establishment of districts, or basho, in is a rare component of Ezo-Haji ceramic assem- Hokkaido in 1615,the Matsumae government of blages. Few stone tools are found at these sites; Hokkaido attempted to assert control over Hok- tool technology was based on metal and wood. kaido and its people. Trading posts were estab- Our research documents that their subsistence lished on the coast of each basho [Takakura base included agriculture (Crawford & Yoshi- 1960; Watanabe 1972: 86). In bashos with good zaki 1987). Three burial mounds are known fishing, Ainu were placed into compulsory from the Ishikari Plain during this period but service in the fishing [Watanabe 1972: their implications for the Ezo-Haji phenom- 86). The basho system was eventually a fun- enon are unknown. Future research should damental influence on Ainu culture change; by consider the socio-political organization of the the 18th century the Ainu and their territory Ezo-Haji. There is some indication of a hier- were being exploited to such an extent that archical settlement organization but no data survival became difficult for them [Peng & have been compiled to test this suggestion. Geiser 1977: 10-11). Presently, life for the estimated 17,000 Ainu Satsumon (Watanabe 1972: 1) in Hokkaido is much like The manufacture of Ezo-Haji ware ceases about that of the Japanese. Traditional Ainu society in AD 1000 and Sueki is no longer found in Hokkaido is usually a reconstruction of life Hokkaido after this time. Instead, pottery, some before the basho system. Until the turn of the with Haji-like forms, is no longer made on a 20th century, the Sakhalin Ainu lived in a potter's wheel and is decorated with a variety of manner thought to have been relatively trailing and incised patterns and patterns pro- unchanged by political events to the south on duced by scraping the clay surface with a piece Hokkaido (Ohnuky-Tierney 1974). It is gener- of wood. The latter technique is called 'Satsu- ally understood that Ainu society was egali- mon,' after which this phase is named. Bowl tarian and not organized under a centralized forms are more common as well. Otherwise this power. Their religion was animistic, with the phase is much like the preceding Ezo-Haji. It is bear and owl deities being among the most widely distributed in Hokkaido. Satsumon also powerful spiritual forces [Ohnuki-Tierney continues the Ezo-Haji pattern of plant hus- 1974: 90, 97). Birds, symbolized by the sacred bandry, possibly taking it into northern Hok- shaved sticks [inau and ikupasai ), were inter- kaido. Cultigens are reported from the Toyo- mediaries between the Ainu and the deities. tomi site in northwestern Hokkaido and from Houses were one-room, rectangular struc- 900 GARY W. CRAWFORD & HIROTO TAKAMIYA tures usually oriented upstream (towards the North Asian plant-husbandry project mountain-dwelling deities) or to the east. Stor- One outcome of the recent attention being paid age facilities were separate from the dwelling, to the late prehistoric period in Hokkaido has constructed with floors raised well above the been the examination of the organization and ground and oriented in the same direction'as development of the dry, temperate plant- hus- the dwellings (Watanabe 1972: 9-10). Ainu bandry system of the Ainu ancestors (Crawford settlements were composed of 10 or fewer & Yoshizaki 1987). Over the last five years, households and each often consisted of a systematic flotation has been conducted as an single family. In the Tokapchi Valley, site integral component of excavations at many late location was usually close to dog-salmon prehistoric sites in Hokkaido. An underlying spawning grounds. Within a settlement, assumption to this research is that plants, parti- houses were often separated by 100-500 m cularly domesticated plants and their associ- (Watanabe 1972: 9-10). ated weed complexes, can provide good About the time the basho system was indications of relationships between geogra- abolished (1869), the Japanese government phically separate areas. Along with this anal- encouraged the Ainu to take up farming (Wata- ysis, flotation produces a quantity of items that nabe 1972: 87). This government policy seems are excellent for radiocarbon dating - cultigens to imply that the Ainu had no form of agri- that are artefacts with a single-year lifespan that culture until the late 1800s, but Watanabe when dated provide fine chronological con- (1972) and Hayashi (1969) both note that sub- trols. sistence farming in Hokkaido appears to have Zoku Jomon sites sampled so far in the North a longer history than this. Certainly, the arch- Asian Plant Husbandry Project include the Esan aeological record documents plant-husbandry Shellmound, Mochiyazawa, Poplar Namiki, by Ainu ancestors as early as the 1st millen- and the K135-4 and K135-5 sites (two localities nium AD. We suggest that the 19th-century at the Sapporo Station, North Entrance). Ezo- Japanese government wanted the Ainu to par- period sites sampled to date are: Chibutashi- ticipate in a farming economy, a system in naizawa, Kagawa 3-Sen, Kagawa 6-Sen, Kashi- which men could be considered by the Japan- wagigawa 11, K67, K441, Okawa, Omusaro-C, ese government to be employed as full-time Sakushu-Kotoni River, Sapporo Botanical Gar- farmers. The Ainu resisted this transition to dens, and Taiikukan. Two sites of mixed or what amounted to full-time farming, but not indeterminable contexts are Idenshikougaku because they were unfamiliar with farming and Jouhoushori on the Hokkaido University practices. Peng & Geiser (1977: 14-16) note campus. Most of these sites have been exca- that a change from an egalitarian, subsistence vated as part of Hokkaido salvage archaeology level economy to participation in a redistri- programmes. butive, state-level production system would Sporadic reports of carbonized cultigen require fundamental changes in sex roles, remains from several Ezo-period sites exca- social status, regulatory mechanisms, and reli- vated before our project began, made it clear gion. Under these conditions, resistance to that cultigens were widespread in Hokkaido by change is not surprising. AD 1200. Such remains fortuitously (no flotation Osteological research has established a close or screening) came to light from the Nishitsuki- relationship between the Ainu and the Jomon gaoka, Wakatsuki and Toyotomi sites (Iwasaki (Brace & Nagai 1982; Brace et al. 1989; Hanihara 1966; Kohno 1959; Ishibashi 1974; Matsutani -- 1986). Modern Japanese are closely related to 1980). Barley, broomcorn millet and beefsteak Koreans and Chinese rather than Jomon or Ainu plant seeds were recovered from Wakatsuki. (Brace & Nagai 1982; Brace et al. 1989).Further- Seeds of Vigna sp., safflower, and broomcorn more, the Tohoku Japanese population exhibits millet have been found at Toyotomi. Broomcorn similarities to both the Jomon and Ainu millet has been identified from the Nishitsuki- (Hanihara 1986: 83). The Sakhalin Ainu do not gaoka site. The problem with such fortuitous fall neatly into these groupings. One reason may discoveries is that without corroborative data, be their long-time association with the Gilyak the significance of these discoveries is difficult with whom they often intermarried (Ohnuky- to assess. For example, a case for widespread Tierney 1974: 10). plant-husbandry (as opposed to widespread LATE PREHISTORIC PLANT HUSBANDRY IN NORTHERN JAPAN 901

cultigens) in Ezo-period Hokkaido depends on rare in the later occupation levels. This suggests supporting evidence such as the presence1 that fishing was less important toward the absence of weed seeds and cultigen by-products end of the Zoku Jomon at K135. The other plants such as chaff, the distribution of cultigen and represented at the site are not as cyclically weed remains, and site locational analyses. The productive as nuts and would have been avail- salvage excavation in the early 1980s at able at forest edges, on banks, and at Sakushu-Kotoni River provided such an oppor- other disturbed habitats in the fall of most tunity. As analysis began on the flotation years. samples from Sakushu-Kotoni River in 1984, The archaeological horizons at both sites are we quickly realized that cultigen remains and separated by thick layers of alluvial deposits weed seeds were distributed throughout the site which allow clear occupational distinctions to and in high densities (Crawford 1986; Crawford be made. Flotation samples from two Zoku & Yoshizaki 1987). Jomon localities provide a detailed view of As the results from Sakushu-Kotoni River subsistence variation during the Zoku Jomon became known, colleagues made quick checks (Crawford 1987).An older and a younger stage of three sites. Soil from Obiratakasago left over of the Kohoku C2-D phase are present at from pollen sampling was floated. A single 4-Chome. Two horizons are present at 5-Chome small sample from Kamuinai was processed as well; the upper level is equivalent to the and a number of soil samples from the Satsumae younger Kohoku C2-D phase at 4-Chome. The site were washed. Five or six cultigens occur in older horizon at 5-Chome is the Esan phase flotation samples from Satsumae: broomcorn which precedes both Kohoku phases. The millet, barley, foxtail millet, rice, wheat, and samples total just over 3700 kg of soil; about possibly sorghum (Matsumae-cho Kyoiku three-quarters of this is from K135-4. To date, Iinkai 1986; Yamada & Tsubakisaka 1989a). One 2244 carbonized seeds have been sorted, about taxon each is reported from Ebetsu Buto, 70% of these from K135-4. Nearly 572 grams of Kamuinai, and Obiratakasago. These are hemp, carbonized nut remains have been recovered so foxtail millet, and Vigna sp., respectively (Ishi- far. zuki 1984; Yamada 1986). To date, cultigens have been recovered from 23 Zoku Jomon and K135-4 Chome. The samples from K135-4 Ezo-period sites in Hokkaido. Only Omusaru-C Chome are from the flotation of 2910 kg of soil and Taiikukan have no cultigens and the anal- and represent several sequential occupations ysis of samples from Mochiyazawa is in pro- and two types of contexts: burned soil lenses gress. In the following section we outline details and unburned fill. Most of the samples (1880 kg of completed, or nearly completed, research at of soil) are from the early Kohoku C2-D-period these sites. occupation (level VIIc). In general, the samples are composed of variable quantities of walnut shell, chestnut meat, seeds of fleshy fruits Sapporo Station: K.135-4 Chome and K135-5 (mainly grape, silvervine, and elder), and a few Chome seeds of weedy annuals (mainly knotweed). Flotation samples from two Zoku Jomon locali- Four whole, carbonized grapes occur in unit ties, K135-4 Chome and K135-5 Chome, at the D-6, level VIIc. No acorn remains have been Sapporo railway station provide a detailed view identified in samples other than in one sample

I ' of subsistence variation during the Zoku Jomon from level 11. Seed concentrations in the burned and a contrast with the succeeding Ezo-Haji soil lenses and other samples are similar, on (Crawford 1987). The localities are stratified average. The seed concentrations are quite vari- and represent multiple occupations over only able, ranging from 0 to 56 seeds per 10 kg of soil two or three centuries. The Zoku Jomon peoples in level VII. were scheduling their visits to this location for Within level VII, distinctions in spatial dis- reasons yet unknown, but nut harvesting or a tribution of plant remains pertain mainly to nut combination of nut harvesting, hunting, and concentrations and types of nuts. Grid unit D-6 fishing may have been involved in the decision (unburned fill) has nut concentrations nearly 80 to move to this locale. Salmon bones are times that of the adjacent grid unit D-5. In common in the earlier occupation levels while addition, walnut is the only type of nut remains 902 GARY W. CRAWFORD & HIROTO TAKAMIYA

Hokkaido. The specimen measures 6.1 (L) by 2.7 (B) by 2.0 (T) mm. One other grass seed that is from a plant introduced to Hokkaido is a specimen of wall barley in level VIIc, burned soil lens no. 120. It measures 5.3 (L)by 1.8 (B) by 1.7 (T) rnrn. This weed form of barley is not native to Japan and likely was introduced, in association with crops such as barley and bread wheat. By Zoku Jomon times, rice agriculture D-6 (Burned soil lenses) D-5 (unburned fill) was present in Tohoku (e.g. Tareyanagi site), FIGURE4. Comparison of nut composition between and weed communities associated with food Units D-5 and D-6 at K135-4. production would have been present as well. The presence of an introduced grass in the early in the unburned fill of unit D-5, and this nut Kohoku assemblage hints at some form of shows up in small quantities in about half the communication with food producing areas in samples from this unit. The unit D-6 samples early 1st-millennium AD Hokkaido. One other are about 93% chestnut (FIGURE4). Walnut taxon from this component is not considered to appears in only seven of the D-6 samples and be native to Hokkaido. Carbonized bulbs one sample (Block 9) accounts for most of the (numbering 12) of a wild onion or leek (AJJium D-6 walnut. Chestnut dominates the samples monanthum), occur in screened samples from from this unit and is present in rather high units E7 and D7, level VIIc. The only other concentrations (up to 21.64 g per kg of soil). archaeological occurrences of this plant in Only one of the 10 level-VII, D-5 burned soil Japan are at the Sakushu-Kotoni River site lens samples (no. 90) contains any nut remains, (Crawford and Yoshizaki 1987), and at the and these are mostly chestnut. The highest Ezo-period Chibutashinaizawa site in Otaru. concentrations of nuts in the samples are found These 4-Chome leek specimens represent their in unit D-6, although the burned soil lens nut earliest known occurrence in Hokkaido. concentration, on average, is lower than that in non-burned soil lenses in D-6. Again, the D-6 K135-5 Chome. Twenty-two soil samples weig- samples are predominantly chestnut, although hing a total of 813 kg were floated from K135-5 more walnut occurs in the burned soil lens. Chome. One set of nine samples is from Esan Unit-6 samples, than in the unburned fill from contexts while the remaining samples are from unit D-6. About 70% of the total nut remains the later Kohoku phase. The Kohoku samples from the burned soil lenses, level VII, is contain fewer taxa and much lower concentra- chestnut meat. tions of plant remains than the preceding Esan Level-by-level comparison shows few samples. Seed densities are 85 times higher in obvious differences that might represent subsis- the Esan samples than in the Kohoku samples. tence or ecological change through time. The The Kohoku samples, in terms of plant taxa and upper levels are low in both nut and seed quantities, are generally within the range of concentrations. The seeds are mainly grape, variation of the samples from K135-4. Acorn is elderberry and some knotweed. The nut somewhat more common at 5-Chome. As in the remains are mainly walnut shell. Seedstg nuts, a early Kohoku 4-Chome samples, the Esan statistic which usually increases with ecologi- samples contain no acorn. Chestnut is conspi- cal disruption and agricultural development, cuous in its absence from this site, considering does not appear to have any systematic vari- its abundance at 4-Chome. One other char- ation at this site. acteristic makes the Esan samples at 5-Chome The only unusual carbonized remains from unusual compared to the other samples: the the upper levels is a single grain of barley in D-5, high concentrations of knotweed. stratum IIIa. Late Yayoi pottery has been recov- ered from stratum VII (Sapporo-shi Kyoiku Poplar Namiki Iinkai 1987: 311) so the barley likely dates Poplar Namiki is a Zoku Jomon cemetery on the sometime after AD 200 and before AD 400. This Hokkaido University campus. Flotation samples specimen is the earliest record of barley in were taken from grave pits, other pits, and areas LATE PREHISTORIC PLANT HUSBANDRY IN NORTHERN JAPAN 903 where carbonized remains were noted during number about 200,000; cultigen seeds alone the excavation. A variety of wild and dom- total 186,000. Nearly 93% of the cultigens are esticated plant seeds are present in the samples. barley, bread wheat, broomcorn (common) So far identified are chenopod, knotweed, dock millet, foxtail millet, indistinguishable wheat/ (sheep sorrel), Chinese lantern plant, bramble, barley, indistinguishable broomcornlfoxtail elderberry, crowberry, Aralia, grape, dogwood, millet and hemp, with the remainder being silvervine, legume, sumac, broomcorn millet, adzukilmung bean (

FIGURE5. Sakushu- Kotoni River grain type proportions as percentage of total number 1186.000) and Barley Wheat ? Cereal Broorncorn Foxtail Millet ? Millet percentage of total Millet weight of carbonized GRAIN TYPES remains (339 gJ. 904 GARY W. CRAWFORD & HIROTO TAKAMIYA

sis on spatial analysis of the Sakushu-Kotoni by 1.6 (W) by 1.3 (T) mm. The single barley River remains. measures 4.9 (L) by 3.1 (W) by 2.5 (T) mm. Weed Some variation in plant remains suggesting seeds are few in number, represented by che- crop processing has been found. Barley breadth nopod and knotweed only. distributions indicating sieving by-products can be recognized here. At least one sample, Botanical Gardens from Pit 4, has fewer than 100 barley grains and In October of 1986, the salvage archaeology centre they all have widths less than 2.3mm. The at Hokkaido University conducted an excavation mean barley width at the site is 2.5 mm. ofa house pit in the Botanical Gardens in Sapporo. Otherwise, the main distributional pattern is a The main purpose was to retrieve subsistence contrast between house contexts and pit1 related data from anundisturbed context dating to samples. House samples are relatively the same period as Sakushu-Kotoni River. An iron clean while external house samples appear not for a wooden hoe was recovered from the to be primary contexts for plant remains but are house floor and is the first well-documented find secondary refuse deposits with remarkable of its kind from a house in Hokkaido. Half the similarities from sample to sample. house was excavated and we managed to collect 44 (174 1) flotation samples. The samples are not Kagawa 3-Sen and Kagawa 6 sites particularly rich, averaging 6.5 seeds per 10 litres. Examination of a few flotation samples from Seven of the 113 seeds are cultigens: broomcorn two late Ezo sites in Tomamae near the Japan and foxtail millet. Most of the remaining identi- sea coast of Hokkaido have revealed the pres- fied seeds are from weedy knotweeds and fleshy ence of four cultigens (Crawford 1986; Yoshi- fruits. zaki 1987). One flotation sample was collected from an L-shaped storage pit in House 5 at K441 Kagawa 3-Sen. Most of the plant remains is Situated in northern Sapporo, K441 is a 10th- wood charcoal (59.38 g), amounting to a density century AD Ezo occupation (Sapporo-shi of 11.88 gllitre of soil. Carbonized seeds number Kyoiku Iinkai 1989: 70). Twenty-three flotation 24, amounting to a density of 48 seeds110 litres samples from level VIb contain 668 carbonized of soil. Nine of these specimens are un- cultigen seeds (Yoshizaki 1989). These are identifiable. That is, they are too badly pre- mainly foxtail (187) and broomcorn millet served to be identified. The remaining 15 seeds (366). The remaining seeds are beefsteak plant are all from cultigens: barley (3), broomcorn (9),probably wheat (Z), and a single specimen of millet (5), foxtail millet (1)and 6 specimens of buckwheat. The only other archaeological unidentifiable millet. One specimen appears to buckwheat dating before AD 1000 in Hokkaido is be a carbonized leek bulb. from the Early Jomon Hamanasuno site (Craw- Only one of the barley grains is preserved ford et al. 1978; Crawford 1983). One other well enough to measure in three dimensions. It potential cultigen from K441 is Chinese lantern measures 5.1 (L) by 3.8(B)by 2.9 (T)mm. Two of plant (6 seeds). Herbaceous weeds are repre- the broomcorn millet measure 2.0 (L) by 1.7 (B] sented by barnyard grass (611), chenopod (55) by 1.4 (T) mm and 1.6 (L) by 1.7 (B) by 1.4 (T) and knotweed (31). mm. The foxtail millet measures 1.3 (L) by 1.2 (B) by 0.9 (TI mm. All measurements fall within Satsumae the size range of the respective grains from the Satsumae is an Ezo site approximately contem- Sakushu-Kotoni River site (Crawford 1986). poraneous with Sakushu-Kotoni River. Situ- Sixty-eight litres of soil from Kagawa 6 pro- ated on the southern tip of the Oshima duced 255 carbonized cultigen seeds. The Peninsula, it is the southernmost site in Hok- samples are from ovens, , burned soil kaido so far sampled for flotation. Fourteen lenses, house floors, and pits. The remains are small samples were analyzed by Yano (1985) mainly foxtail millet (75), broomcorn millet and 11 samples were analyzed by Yamada & (89), unidentifiable millet (95), and one speci- Tsubakisaka (1989a). Yano reports an unspe- men each of barley and beefsteak plant. The cified number of foxtail millet grains, three foxtail millet averages 1.3 (L) by 1.2 (W] by 1.0 barley grains and 1 sorghum caryopsis from (T) mm. The broomcorn millet averages 1.9 (L] houses 6,10,11 and 14 (1985: 303-4). Yamada& LATE PREHISTORIC PLANT HUSBANDRY IN NORTHERN JAPAN 905

Tsubakisaka (1989a) report a total of 96 speci- ]omon and Ezo periods are summarized in mens from hearths, ovens and floors of houses FIGURES6 & 7 (see also Ishizuki 1986). Radiocar- 34 and 36 and from a burned soil lens (K-14). bon dates from individual sites andlor keishiki The seeds are identified as barley (42), rice (5), (pottery assemblages) are arranged together on sorghum (I), and wheat (48). This is the only the charts in order to indicate intra-sitelkeishiki substantial sample of Ezo bread wheat besides variation of dates. In an attempt to examine the that from ~akushu-~otoriRiver. The wheat is local chronology in greater detail, seeds from identical to the compact type of bread wheat several flotation samples were submitted to the from Sakushu-Kotoni River. University of Toronto (ISOTRACE) accelerator dating facility. Results are summarized in Idenshikougaku TABLE2 and are indicated in FIGURE6 by solid An oblong structure ( 1)and pit (Feature ellipses. Seeds were chosen for dating when 2) were excavated at Idenshikougaku on the possible because of their short life span. The Hokkaido University Campus. Artefacts are all cellulose carbon component was extracted for Ezo-period (Hokkaido Daigaku Maizo Bunkazai dating. One sample from K135 did not have Chosashitsu 1988). Flotation samples were sufficient cellulose carbon for a radiocarbon taken from levels Xi and Xi of Feature 1. The date. shallower level Xi contained 223 cultigen One date from level VIIc at K135 (TO-627) is seeds. These are barley (37), wheat (204), in the early range for the Kohoku phase. The broomcorn millet (I),possible beefsteak plant other date from level VIIC (TO-624) appears to (I),unknown bean (2)and maize (1).Wild plant be too young, indicating contamination of some remains are silvervine, Amur cork tree, sort. The level I11 date (TO-626) is about a knotweed, elderberry and sumac. The maize is century younger than the TO-627 date, con- from an 8-row variety common in North Amer- firming our suspicions that the deeply stratified ica after AD 800. Its undoubted New World K135 site represents a short time-span within origin points to a relatively recent date for level the Zoku Jomon. Until we are able to secure Xi or mixing of Ezo and recent soil in level Xi. many more dates from this site, the radiocarbon All but 38 of the wheat grains are large bread chronology remains rather uncertain. The Esan wheat. There are only two reports of such wheat Shellmound date corresponds roughly with one in an archaeological context in Hokkaido. The date on Esan Shellmound material (N-3388) other is from Poplar Namiki on the Hokkaido that falls within the 4th and 5th centuries BC. University campus which has been radiocarbon The single 9th-century BC date (N-3786) for Esan dated to 210k50 BP. is unacceptably old. One date from Poplar Namiki (TO-630) from deep pit fill falls within Kashiwagigawa 11 the Zoku Jomon range but the wheat from a Kashiwagigawa 11, excavated in 1989, is an shallower level (111) at the same site is historic Ezo-period site in Shimamatsu-cho, Eniwa city. (TO-687). This was expected because the wheat The analysis is in progress samples from House is non-compact bread wheat that is not found in 1 indicate that cultigen remains are present in definite Ezo-period contexts. Flax was found in high densities. House 1was destroyed by a fire, similar deposits at Poplar Namiki. Similar so it was an excellent opportunity to examine deposits at Idenshikougaku on the Hokkaido the distribution of remains on or near the house University campus produced 8-row maize, a , ' floor. The floor was sampled in 50-cm squares new-world crop that was introduced to Japan and 90 of these samples have been examined to sometime after the 16th century AD. The date. Cultigen seeds number 10,000 and are Okhotsk sample (TO-628), from a burial, is mainly foxtail (54%) and broomcorn millet within its expected time frame. (44%). The remainder are hemp, unidentified The chronological details of the Zoku Jomon millet and one unidentified bean. A large and Ezo-periods in Hokkaido are still being number of weed seeds include chenopod, worked out but it is becoming clear that we are grasses, and knotweed. not dealing with a simple linear periodization. Esan has some overlap with Kohoku indicating Radiocarbon dates that at least two populations are responsible for Radiocarbon dates for the Hokkaido Zoku these pottery assemblages. Furthermore, the 906, GARY W. CRAWFORD & HIROTO TAKAMIYA

HOKKAIDO ZOKU JOMON RADIOCARBON DATES (Uncorrected) 14001

LABORATORY-NO.

FIGURE6. Hokkaido Zoku Jomon radiocarbon dates [uncorrectedj. Dates within the same shade of grey are from the same site; unshaded dates represent sites with single dates. White squares and bars indicate the mean and standard deviation for each radiocarbon date; black squares are University of Toronto dates. 1 earliest Ezo dates range from AD 200 to 400 and in Tohoku followed by a subsequent northward the latest Zoku Jomon dates cluster around AD expansion into a Hokkaido occupied by Zoku 200 to 300, assuming that the two late Ebetsu I11 Jomon populations. dates and the GaK-9201 date with a wide error margin are incorrect (FIGURES 6 & 7). There Discussion appears to be some temporal overlap of no more The data from the many intensively researched than a century at the end of the Zoku Jomon and late prehistoric sites excavated in the last few beginning of the Ezo period. This would not be years in Hokkaido are providing a new perspec- outof line with a model postulating Ezo origins tive on the Japanese northern frontier in the 1st LATE PREHISTORIC PLANT HUSBANDRY IN NORTHERN JAPAN 907

HOKKAIDO EZO PERIOD RADIOCARBON DATES (Uncorrected)

LABORATORY-NO.

FIGURE7. Ezo-period radiocarbon dates [uncorrectedJ. Dates within the same shade of grey are from the same site; unshaded dates represent sites with single dates. White squares and bars indicate the mean and standard deviation for each radiocarbon date.

millennium AD. Detailed subsistence and other agricultural Tohoku Yayoi and later Kofun- archaeological data reported here represent col- period Japanese but plant husbandry seems to

I lections from the beginning of the Zoku Jomon have been peripheral to their way of life. From through the two phases of the Ezo period. Zoku the limited living-site evidence available, parti- Jomon subsistence clearly contrasts with that of cularly that from K135, greater seasonal mobil- the Ezo-Haii which immediately succeeds it in ity than in the preceding Final Jomon is central and southern Hokkaido. In many ways, apparent. K135 evidence indicates repeated use including the degree of sedentism, ceremonia- and abandonment of the site. Nut, small grain lism, subsistence, and technology, the Hok- and berry harvesting, in addition to salmon- kaido Zoku Jomon differs from the preceding fishing and deer-hunting were common pur- Final Jomon as well. suits. Further data are required in order to The Zoku Tomon is a local Hokkaido assess whether short encampments such as the population who had some interaction with the ones at K135 were linked to long-term villages 908 GARY W. CRAWFORD & HIROTO TAKAMIYA

sample laboratory b.p. calibrated number (uncalibrated)

K135 C-5 Level 111 acorn TO-626 1830k50 AD 127-260 D-5 Level VIIc grape seed TO-624 1240k50 AD 759-781 C-4 Level VIIc grape seed TO-627 1940k50 AD 17-81 Esan SheJJ Mound Amur cork TO-628 2260k50 395 BC-353 BC tree seed 294~~-232BC PopJar Namiki 01-04(31) walnut TO-630 1920k50 AD 48-129 Level 3 wheat TO-687 210k50 ~~1603-1706 (large) Oumugawajiri [Okhotsk) corydalis TO-688 1620k50 AD 386-434 tuber

These results are the average of two machine-ready targets measured on different occasions. They have been corrected for natural, preparation, and splutteringfiactionation to be a baseof &I3 = -25 0100. Theuncalibratedages are quoted inradiocarbon years using the Libby "C meanlife of 8033 years. The errors represent 68.3% confidence limits. The calibrations are courtesy of R. Beukens, ISOTRACE laboratory, University of Toronto.

TABLE2. Accelerator radiocarbon dates from Hokkaido. that may have had some plant husbandry. The ties found at Sakushu-Kotoni River. A conso- presence of a few cultigen seeds is not adequate lidation phase, when agricultural peoples evidence to allow us to conclude that the Zoku occupied all arable land in southern Hokkaido, Jomon people were growing the crops. They could may never have happened in proto-historic have been imported. Nevertheless, Zoku Jomon times. It is possible that there were varying peoples were aware of cultigens. Besides the degrees of dependence on food production from barley from K135, only one other Zoku Jomon site northern Hokkaido, where there was none at all has produced cultigen remains - hemp and great (Omusaro-C produced no cultigen remains) to burdock from Ebetsu Buto (Yamada 1986). the Ishikari Plain, where plant-husbandry was Cultigens from Poplar Narniki appear all to be better established. Pottery manufacture even- recent. This phase of availability, however, did tually declined, not as a devolution of local not lead progressively to acceptance and conso- technology, but because Japanese ceramic and lidation of agriculture on Hokkaido. metal technology provided replacements for The succeeding Ezo-Haji phase brings to vessels and tools in much the same way as in the Hokkaido an entirely new technology and sub- New World after European contact. sistence system. The change appears to be Ethnographers generally assume that Ainu sudden and represents a substitution phase of plant-husbandry is a recent introduction to % - agricultural origins in Hokkaido. Sakushu- Hokkaido, perhaps not being established until Kotoni River is our best example of an agri- the 19th century (e.g. Watanabe 1972). This is culturally-oriented community in Hokkaido at clearly an oversimplification. Hayashi (1969), this time. The agriculture is based on a variety of without the benefit of archaeological data, sug- crops. Rice, although present, seems to have gests that there is reason to believe that Ainu been of lesser importance than barley, wheat, food production is, in fact, Yayoi-like in many millet, beans and other crops. To what degree ways, and therefore has a relatively long his- agriculture became consolidated in Hokkaido is tory. Based on linguistic, ethnographic (folk- difficult to assess. Later Satsumon sites have lore), and ethnohistoric data, Hayashi evidence of cultigens but not in the large quanti- concludes that here is little doubt that Ainu LATE PREHISTORIC PLANT HUSBANDRY IN NORTHERN IAPAN 909 agriculture was derived from a Japanese agri- known in the archaeology of northeastern Japan. culture system (1969: 191). The Ainu use of Shortly thereafter, written documents refer to shell-sickles for harvesting and storage houses the natives of the region as distinct from early with raised floors, for example, are 'characteris- Japanese. The natives are the Ezo-period occu- tics of the Yayoi agriculture system' (Hayashi pants who soon spread to Hokkaido, bringing 1969: 187). In addition, he sees morphological about the termination of the Zoku Jomon there. similarities between the of the The emergence of the Ainu is therefore a set of Yayoi and Ainu (1969: 187 & 215). He suggests processes involvingthe Jomon, the Yayoi, and the that, while the Yayoi agriculture system evol- Yarnato state- processes not clearly understood at ved through time in western Japan, the Ainu the moment. Until the confirmation of food agriculture system remained 'primitive' in Hok- production as a significant aspect of Ezo culture, kaido until recently (1969: 187). the Ezo and their Ainu descendants were easy to The complexities of southwestern Hokkaido set apart from the Japanese. The cultural system- prehistory indicate that it is an oversimpli- including agriculture - that emerged by the 8th fication to assume that the Ainu are a remnant century AD in Tohoku and Hokkaido, because of Jomon population who continued hunting and its close historical relationship to the Yayoi gathering practices until recently. Biologically, chiefdoms and Yamato state, was quite similar to there is continuity from the Jomon to the Ainu that of the 8th-century Japanese. Archaeological (Brace et al. 1989), but culturally, the Ainu are research on late prehistoric and proto-historic not the product of a linear descent from the period in Hokkaido is still young, so we expect to Jomon of northeastern Japan. The Final Jomon demonstrate many new aspects of culture change began to feel the effects of contact with Yayoi in general and agricultural history specifically as peoples, a contact first evidenced with the research continues. appearance in Aomori of Ongagawa pottery, Acknowledgements. The research for the Northern Asian Yayoi beads, and rice as early as the Obara A Plant Husbandry Project was sponsored by the Social phase. Subsequently, the establishment of rice Sciences and Humanities Research Council of Canada (Grant paddies by the first century AD in Aomori No. 410-86-0769)and a bilateral exchange grant from signalled the end of the Final Jomon in Tohoku SSHRCC and the Nihon Gakujutsu Shinkoukai (Grant No. 473-86-0007).The research continues with support from and Hokkaido. With the collapse of the Final SSHRCC (Grant No. 410-89-0786)and Earthwatch. The first Jomon came the appearance of local groups draft of the paper was presented at the 1987 annual meeting whose pottery bears varying degrees of resem- of the Society for American Archaeology in Phoenix, Ariz- blance to Jomon and Yayoi pottery and who are ona. A conference travel grant from Erindale College, Uni- versity of Toronto made it possible to attend. now usually referred to as Tohoku Yayoi. The cooperation and support of many colleagues and Assimilation and acculturation likely took friends have made this research possible. They include place, and the Yayoi's local quality, particularly Yonemura Tetsuhide, Ishibashi Takao, Kato Kunio, Uwaya in northern Tohoku, seems to result from the Shuichi, Ueno Shuichi, yamamoto Yasuko and the staff of development of local, independent chiefdoms the numerous excavations in Hokkaido where we sampled. Supporting institutions include the Hokkaido, Sapporo, from a substantial Jomon base. Chitose, Eniwar and Yoichi Boards of Education as well as The contemporaries of the Tohoku Yayoi on Hokkaido University. Sapporo Subaru supplied a vehicle at Hokkaido were the Zoku Jomon whose pottery minimal cost each field season. We would also like to in southern Hokkaido is Yayoi-like [Esan). In express our gratitude to Tsubakisaka Yasuyo who has helped in many ways, particularly by managing the Hokkaido central Hokkaido Kohoku pottery bears little University palaeoethnobotany laboratory. Yarnada Goro has resemblance to Yayoi pottery. At this time, an helped with many botanical/archaeological problems. A. C. intensive dry plant-husbandry system may well D'Andrea helped analyse much of the data reported here and have been beginning to evolve in northern continues to do research in Hokkaido and Tohoku. Yoshi- Tohoku. A new phase of change came about zaki Masakazu and the staff of the Hokkaido University Salvage Archaeology Center have unselfishly given of their when the Yamato State extended its influence time and resources. We also wish to thank Gina Barnes for into parts of Tohoku. This period is the least her valuable comments on our paper.

References toric hunter-gatherers and their transition to AIKENS,C.M. & T. HIGUCHI.1982. The prehistory of agriculture in Japan, in S. Koyama & D.H. Thomas Japan. New York (NY): Academic Press. (ed.), Affluent foragers: 213-58. Osaka: National AKAZAWA,T. 1981. Maritime adaptation of prehis- Museum of Ethnology. .# GARY W. CRAWFORD & HIROTO TAKAMIYA

1986. Hunter-gatherer adaptations and the transi- HAYASHI,Y. 1969. Ainu no Noukou Bunka [Ainu tion to food production in Japan, in M. Zvelebil Plant Cultivation]. Tokyo: Kokominsosho 4. (ed.), Hunters in transition: 151-65. Cambridge: HOKKAIDODAIGAKU MAIZO BUNKAZAI CHOSASHITSU. Cambridge University Press. 1988. Hokudai Konai no Isekii [Hokkaido Uni- ASTON,W.G. 1896. Nihongi, chronicles ofJapan from versity Campus Sites]. Hokkaido Daigaku. the earliest times to AD 697. Tokyo: Charles E. ICHIKAWA,T. 1985. Nihon no bunatai ni okeru noukou Tuttle [1972]. to seikatsu-bunka [Agriculture and life style in BRACE,C.L., M.L. BRACE& W.R. LEONARD.1989. Japan's Beech Forest Zone], in Bunatai-Bunka: Reflections on the face of Japan: a multivariate 107-123. Tokyo: Shisakusha. craniofacial and odontornetric perspective, ISHIBASHI,T. 1974. Tokachi-bu Wakatsuki: dai ni ji American Journal of Physical Anthropology 78: hakkutsu chosa [Wakatsuki, Tokachi Region: the 93-113. second excavation]. Hokkaido: Obihiro Kyoiku BRACE,C.L. & M. NAGAI.1982. Japanese tooth size, Iinkai. past and present, American Journal of Physical ISHIZUKI,K. 1984. Shizen ibutsu [Natural remains], in Anthropology 59(4): 399-411. Kamuinai Kannon Dokutsu. Hokkaido: CRAWFORD,G.W. 1983. Paleoethnobotany of the Kamuinai-cho Kyoiku Iinkai. Karneda Peninsula Jomon. Ann Arbor (MI): 1986. Ainu Bunka no Genryu [The Origin of the Museum of Anthropology, University of Ainu]. Sapporo: Miyama Shobo. Michigan. Anthropological Papers 73. ITO,GENZO. 1966. Tohoku, in S. Wajima (ed.): 203-20. 1986. Sakushu-Kotoni River site: the Ezo-Haji 1~0,NOBUO. 1984. Aomori-ken ni okeru inasaku noko component plant remains. Sapporo: Hokkaido bunka no keisei [Development of rice agriculture Daigaku Bungakubu. in Aomori Prefecture], in T. Katoh (ed.), Hoppo 1987. K135 Iseki kara shutsudo sareta shokubutsu Nihon Bunka no Kenkyu: 1-26. Miyagi: Tohoku shushi ni tsuite [Plant seeds excavated from the Gakuin Daigaku Tohoku Bunka Kenkyusho. K-135 site], in Sapporo-shi Kyoiku Iinkai (ed.), 1986. Tohoku chiho ni okeru inasaku noukou no K135 Iseki: 565-81. Hokkaido: Sapporo-shi seiritsu [The emergence of rice agriculture in Kyoiku Iinkai. [In Japanese and English]. Tohoku], in N. Egami (ed.], Nihonshi no Reimei: In press. Prehistoric plant in East 335-365. Tokyo: Rokko Shuppan. Asia: the Japanese perspective, in P.J. Watson & IWASAKI,T. 1966. Nishitsukigaoka Iseki [The Nishit- C.W. Cowan (ed.), The origins of plant dom- sukigaoka Site]. Hokkaido: Nemuro-cho Kyoiku estication in world perspecttive. Washington Iinkai. (DC): Smithsonian Institution Press. KAMAKI,YOSHIMASA. 1965. Jomon, Jidai [The Jomon CRAWFORD,G.W., W. HURLEY& M. YOSHlZAK1. 1978. Period]. Tokyo: Kawaide shobo. Implications of plant remains from the Early KANASEKI,H. & M. SAHARA.1978. The Yayoi Period, Jomon, Hamanasuno site, Asian Perspectives Asian Perspectives 1976 19.1: 15-26. 1976 19(1): 145-55. KASAHARA,Y., M. BUDA& A. FUJIZAWA.1986. Yonago- CRAWFORD,G.W. & M. YOSHIZAKI1987. Ainu ancestors shi Megumi iseki no shushi no bunrui dotei and early Asian agriculture, Journal of Archaeo- . [Analysis and identification of seeds from the logical Science 14: 201-13. Megumi site, Yonage city], in Megumi Iseki. DENNELL,R. 1985. The hunter-gatherer/agricultural Tottori-ken: Yonage-shi Kyoiki Iinkai. frontier in prehistoric temperate Europe, in S. KATOH,MINORU & T. SUTO.1986. Tohoku, in Iwanami Green & S.M. Perlman fed.), The archaeology of Koza Nohon Kokogaku (extra vol. I): 155-97. frontiers and boundaries: 113-40. New York Tokyo: Iwanami Shoten. (NY): Academic Press. KOHNO,H. 1959. Hokkaido Shutsudo no Ogata U jikei HANIHARA,KAZURO. 1986. The origins of the Japanese Tekki ni tsuite [A U-shaped iron tool excavated in in relation to other ethnic groups in East Asia, in Hokkaido]. Sapporo: Hokkaido Gakkai Daigaku R.J. Pearson, G.L. Barnes & K.L. Hutterer (ed.), Kokogaku Kenkyukai Renrakushi. Windows on the Japanee past: studies in KOMOTO,M. & S. YAMAZAKI.1984. Yayoi-jidai no kiso archaeology and prehistory: 75-83. Ann Arbor chishiki [Basic knowledge of the Yayoi period]. (MI): Center for Japanese Studies, University of Tokyo: Tokyo Bijutsu. Kokogaku series 5. Michigan. MATSUMAE-CHOKYOIKU IINKAI. 1985. Satsumae. Hok- HAYASHI,KENSAKU. 1976. Kamegaoka Bunkaron [A kaido: Matsumae-cho Kyoiku Iinkai. comment on the Kamegaoka Culture], in Tohoku MATSUTANI,AKIKO. 1980. Tokachi-bu Wakatsuki Iseki Kokogakkai (ed.), Tohoku Kokogaku no Sho- Shutsudo Tankabutsu no Shikibetsu ni tsuite rnondai: 168-203. Tokyo: Azuma Shuppan. [The identification of carbonized remains from 1986. Kamegaoka to Ongagawa [Kamegaoka and the Wakatsuki site, Tokachi region], Urahoro-cho Ongagawa], in Y. Kondo et ul. (ed.), Nihon Kyodo Hakubutsukan Hokoku vol. 16. Kokogaku 1: 93-124. Tokyo: Iwanami Shoten. MURAKOSHI,KIYOSHI. 1988. Sunazawa iseki [The Sun- LATE PREHISTORIC PLANT HUSBANDRY IN NORTHERN JAPAN 911

azawa site], in Yayoi-Bunka no Kenkyu 10: etic studies on the Japanese. Tokyo: University of 211-13. Tokyo: Yuzankaku. Tokyo Press. OHNUKI-TIERNEY,EMIKO. 1974. The Ainu of the North- WATANABE,HITOSHI. 1972. The Ainu ecosystem. west coast of southern Sakhalin. New York (NY): Seattle: University of Washington Press. Holt, Rinehart & Winston. YABUNO,TOMOSABURO. 1987. Japanese barnyard OKADA,ATSUKO. 1984. Tohoku chiho to Ezo chii [The millet (Echinochloa utiiis, Poaceae) in Japan, Tohoku district and the Ezo region], Sozo no Economic Botany 41.4: 484-93. Sekai 49: 97-105. YAMADA,GORO. 1986. Hokkaido ni okeru senshi jidai OKAMOTO,ISAMU. 1966. Yayoi Bunka no seiritsu [The no shokubutsu shokuryo ni tsuite [Prehistoric establishment of the Yayoi Culture], in S. Wajima plant foods in Hokkaido], Hokkaido Kokogaku 2: (ed.): 424-41. 87-106. PENG,F.C.C. & P. GEISER.1977. The Ainu: the past in 1987. K135 iseki kara sanshutsu shita kafun, the present. Hiroshima: Bunka Hyoron houshi [Pollen and spores from the K-135 site], in Publishing Company. Sapporo-shi Kyoiku Iinkai (ed.), K135 Iseki: SANSOM,G. 1958. A History of Japan to 1334. Stan- 583-95. Hokkaido: Sapporo-shi Kyoiku Iinkai. ford: Stanford University Press. YAMADA,GORO & Y. TUSUBAKISAKA.1989a. Satsumae SASAKI,TAKAO. 1983. Inakadate Tareyanagi iseki [The iseki kara sanshutsushita saibai shokubutsu ni Inakadate Phase Tareyanagi site], in Yayoi Jidai tsuite [Cultivated plants from the Satsumae site], 1: 35-59. Tokyo: Shakai Shisosha. in Matsumae-cho Kyoiku Iinkai (ed.), Satsumae SATO, TOSHIYA.1984. Kamegaoka iseki Sawane 11: 33-8. Hokkaido: Matsumae-cho Kyoiku- Chiku-B-ku shutsudo no ine eika narabi ni tanka Iinkai. kome tsubu [Rice caryopses and carbonized rice 1989b. Ohotsuku Bunka no iseki ni tomonatta kernels from the Kamegaoka site, Sawane Local- zakkoku [Grains that accompaned an Okhotsk ity B.], in Kamegaoka Sekki Jidai Iseki. Aomori Culture site], Project Seeds News 1: 3. Kenritsu Kyodokan, Archaeology Report 16: YANO,B. 1985. Shutsudoshita shokubutsu itai ni 216-24. tsuite [Recovered plant remains], in Matsumae- 1986. A-6 tateana jukyo shutsudo komugi-you cho Kyoiku Iinkai (ed.), Satsumae: 303-6. Hok- ibutsu [Wheat-like remains excavated from kaido: Matsumae-cho Kyoiku Iinkai. House A-61, in K. Takada (ed.), Ichinohe Baipasu YASUDA,YOSHINORI. 1988. Jomon Jidai no kankyo to Kankei Maizo Bunkazai Chosa Hokokusho 111: seigyo [Environment and subsistence during the 341-6. Iwate: Ichinohe-cho Kyoiku Iinkai. Jomon period], in K. Sasaki &T.Matsuyama (ed.), SUTO,TAKASHI & M. SAHARA.1987. Higashi Nihon ni Hatasaku Bunka no Tanjo: 25-63. Tokyo: Nihon okeru Yayoi bunka no seiritsu to tenkai [The Hoso Shuppan. establishment and development of the Yayoi YOSHIZAKI,MASAKAZU. 1984. Kokogaku ni okeru Ezo Culture in Eastern Japan], in Yayoi Bunka no to Ezo-chi 1: Kodai Ezo-chi no bunka [The Ezo Kenkyu 4: 201-22. Tokyo: Yuzankaku. and Ezo region from an archaeological perspec- SUZUKI,KATSUHIKO. 1986. Nihon no Kodai Iseki 29: tive I: the ancient Ezo culture], Sozo no Sekai 49: Aomori [Ancient Sites in Japan 29: Aomori]. 80-97. Tokyo: Hoikusha. 1986. Sakushu-Kotoni-Gawa Iseki [The Sakushu- TAKAHASHI,MAMORU. 1986. Yayoi bunka no hirogari Kotoni River Site]. Sapporo: Hokkaido Uni- [The spread of the Yayoi culture], in Yayoi Bunka versity. no Kenkyu 9: 35-44. Tokyo: Yuzankaku. 1987. Tomamae-cho Kagawa 6 iseki shutsudo no TAKAKURA,SHINICHIRO. 1960. The Ainu of Northern shokubutsu shushi [Plant seeds from the Toma- Japan, Transactions of the American Philosophi- mae town Kagawa 6 site], in Tomamae-cho cal Society n.s. 50(4). Kyoiku Iinkai (ed.), Kagawa 6 Iseki: 171-85. TASAKI,HIROYUKI. 1986. Yayoi jidai no shokuryo Hokkaido: Tomamae-cho Kyoiku Iinkai. [Foods during the Yayoi-period], Kikan 1989. K441 iseki kita 34 chome chiten shutsudo no Kokogaku 14: 18-22. shokubutsu shushi [Plant seeds recovered from TERASAWA,KAORU. 1986. Hatake sakumotsu [Dry field the K441 site, North 34 Street locality], in crops], Kikan Kokogaku 14: 23-31. Sapporo-shi Kyoiku Iinkai (ed.), K441 Iseki Kita TSUKADA,M., S. SUGITA& Y. TSUKADA.1986. Oldest 34 Chome Chiten: 70-79. Hokkaido: Sapporo shi primitive agriculture and vegetational environ- Kyoiku Iinkai. ments in Japan, Nature 322: 6324. ZVELEBIL,M. 1986. Mesolithic prelude and WAJIMA,SEIICHI. 1966. Yayoi Jidai [The Yayoi Period]. revolution, in M. Zvelebil (ed.),Hunters in Tran- Tokyo: Kawaide Shobo. sition: 5-15. Cambridge: Cambridge University WATANABE,S., S. KONDO& E. MATSUNAGA.1975. Press. Human adaptability 2: anthropological and gen-