Home | Index Japanese Archaeology last revised:
May 27, 1997

Prehistoric and Historic Settlement Patterns in the
Takase River Drainage, Northeastern Japan

by Charles T. Keally
Sophia University, Tokyo

This is a modified version of a paper originally presented to the meeting "Cultural Adaptation at the Pleistocene/Holocene Boundary," at the World Archaeological Congress, The 11th Congress of the International Union of Prehistoric and Protohistoric Sciences, Southampton, England, 1-6 September 1986. This report is based on the 1986 database, but it includes discussion of more recent periods than the original paper did.


Human settlement in the Takase River drainage was sparse and temporary during the Pleistocene and early Holocene when only the small rivers and streams, and the boreal or birch forest provided food resources. A rather sedentary settlement pattern is apparent in the many shellmiddens along the sheltered shores of the shallow Palaeo-Ogawara Bay that formed as the ocean flooded the lower Takase Valley after about 7700 B.P. The molluscan shells, and the fish, bird, and mammal bones found in the early middens are predominantly of species abundantly available within a radius of 1-3 km of the sites. Although plant remains are not reported from these sites, walnuts, among others, certainly were collected, stored and eaten regularly. The relative permanence of the settlements is indicated by the seasonally balanced representation of species and by the small but noticeable proportion of molluscs, fish, and occasionally sea mammals derived from the Pacific Ocean 5-10 km away. Shellfishing was most important from late Earliest Jomon through Early Jomon, roughly 7500-5000 B.P. The population rose and fell during the Jomon period, reaching its greatest density in Late Jomon during the 2nd millennium B.C. After that it fell to almost zero as the forager way of life disappeared from Japan and wet rice farming became the economic base. From early Historic times the population rose again, but there are few indications these people depended much on the resourses of the lakes or the ocean.


The Takase River is a small drainage system at 41Enorth latitude, on the northeast Pacific coast of Honshu, the largest of the four main Japanese islands. The river and its tributaries drain the eastern flanks of the Hakkoda volcanoes and a small plain between the highlands of the Shimokita Peninsula and the mountains of Iwate prefecture. Rising sea levels at the end of the Pleistocene transformed the lower Takase River valley into the Palaeo-Ogawara Bay. When the sea receded again a few thousand years ago, the bay became a lake. The bay and subsequent lake provided a rich and locally varied biotic environment in the lower reaches of the Takase River. This was complemented by the diversity of niches in the surrounding dissected plains and hills, and in the nearby mountains. The rising and falling of the level of the bay expanded and contracted the marshlands along the shores, giving further variability to the ecology over time.

Changes in human settlement patterns reflect this variability. Evidence of human activity in the drainage prior to 8500 B.P. is almost nonexistent. Several small collections of potsherds and stone tools dated 7700 to 8500 B.P. tell of people with a subsistence economy based on bow and arrow hunting and net fishing, and perhaps plant collecting. The numerous shellmiddens that accumulated after about 7700 B.P. show exploitation of many kinds of molluscs, fish, land mammals, and birds, and of advanced fishing tools and skills. The appearance of these shellmiddens seems to coincide with the beginning of the formation of the bay. Their expansion inland between 4000 and 6000 B.P. also seems to parallel the maximum expansion of the bay, while their disappearance follows the contraction of the bay into a brackish lake sometime after about 4,000 years ago. The foragers virtually disappeared from the drainage during a period of climatic deterioration in the last centuries B.C., and only a handful of potsherds has been found to account for the early rice farmers in the first six or seven centuries A.D. But from at least the 8th century A.D., numerous farming villages were established near the marshes around the lake, and by the 11th or 12th century rich landowners and powerful overlords were building manors and castles on the river bluffs on the plain near the mountains.

The research reported here was conducted mostly in 1984 and 1985. Published prefectural site registers (Aomori-ken 1978; Bunka-cho 1981) provided most of the information used for the research. But I obtained additional details from local specialists (Segawa Shigeru, Tajima Kazuo, and Peter Kunkel) and town offices (Misawa city), and from publications on excavation work conducted at some of the sites.1 I visited many of the sites to confirm their locations. A small amount of surface reconnaissance turned up 12 new sites not recorded in the registers. Altogether I have been able to get at least some information on 218 sites so far, ranging from Palaeolithic to Medieval in age. Some of these sites were contiguous locations with the same kinds of materials. I merged these into single sites for the analysis. The final list of 208 sites combined has 352 broadly defined occupations.2 Thirty-nine of the occupations are "unknown" and 14 more were given only as the overly broad "Jomon." Subtracting these from the final data base left 299 occupations, of which 218 belong to the Palaeolithic and the five Jomon periods. There are only 3 sites dated to the succeeding Yayoi and Kofun periods, but there are 78 Historic period sites, about half from the Nara-Heian period and half from the Medieval period.

Geology and Ecology of the Takase Drainage:

The drainage basin of the Takase River is only 25 by 43 km, with an area of about 900 km2, quite small even for Japan. Geologically the Takase River drainage has two major parts, the Ou Mountain Range and the Sanbongi Plain (Kitamura et al. 1972; Geological Survey 1977: 6-7; Kankyo-cho 1982: 13-21). The mountains are early Miocene effusive rocks and middle Miocene sedimentary rocks that have been uplifted and folded. Over these are layers of basalt, dacite, porphyry, Tertiary and Quaternary andesites, and large Middle Pleistocene tuff flows, all probably from eruptions in the Mt. Hakkoda region. The plain is largely uplifted late Pliocene to early Pleistocene marine sediments, flat layers of almost pure sand, extensively eroded into a mosaic of older and younger terraces. Middle Miocene dacite and early Pliocene sediments are exposed in ravines in the northern part of the plain. Middle Pliocene sediments are exposed in ravines in the southern part. Terraces and related deposits from the Middle Pleistocene (Higher Terrace), early Upper Pleistocene (Middle Terrace), late Upper Pleistocene (Lower Terrace), and Holocene have been eroded deeply into the older marine deposits. A major fault line separates the mountains and plain. East of this fault, marking the eastern fringe of the mountains, are late Miocene sediments with Early or early Middle Pleistocene terrace deposits (Highest Terrace) crowning them. Black ando soils cover the plain; brown forest soils are well developed in the mountains; and podzols are found only on the peaks at the western edge of the drainage area. Volcanic activity at the Towada and Hakkoda volcanoes has blanketed the whole area many times with layers of yellow, clayey ash called loam.

Temperatures average 22-24°C in August and -2°C in January (Wadachi 1958: 426; Kizo-cho 1972: 27-29). The summers are hot and humid, with daytime temperatures frequently exceeding 30°C. There are 165-185 days with average temperatures over 5°C (Momose 1972: 168). Snowfall is relatively light in the Takase drainage, and snow is on the ground only about 80 days a year. But the winters are quite windy and feel cold, with nighttime temperatures usually dropping to -6°C. Temperatures generally are milder in summer and winter nearest the coast. Rainfall is especially heavy from July to September. Precipitation is lightest in winter and spring. The total annual precipitation is 1,000-1,200 cm, making this the driest region in northern Tohoku (Kankyo-cho 1982: 23).

In the drainage basins to the west and northwest of the Takase, the summers are hotter, the winters colder, there is a considerably greater accumulation of snow, and snow is on the ground a month longer (Wadachi 1958; Kizo-cho 1972). Precipitation there is highest in December and January, but it is also quite high from July to September. The climate in the drainage immediately to the south of the Takase is about the same as in the Takase drainage. But it is cooler and wetter in the highlands further to the south and to the north.

The natural vegetation of the area is northern deciduous broadleaf forest, dominated by oaks and beech (Kankyo-cho 1982; Yasuda 1978: 204, 218). Mongolian oak (Quercus mongolica) and beech (Fagus crenata) would be especially abundant under natural conditions. The coastal localities would have typical herbaceous dune vegetation (Glehnia littoralis, etc.). Alder, willow, reeds and aguatic vegetation (Alnus japonica, Salix, Phragmites, and Potamogeton) would dominate the floral communities in the marshlands around the lake. Beech and Aomori cedar (Hiba arbor vitea: Thujopsis dolabrata v. hondae) would characterize the vegetation of the mountains.

Japanese Serow (Capricornis crispus), shika deer (Cervus nippon), black bear (Selenarctos thibetanus), Japanese monkeys (Macaca fuscata), fox (Vulpes vulpes), racoon dog (Nyctereutes procyonoides), and badger (Meles meles) are native to the region, but the present northern limit of wild boar (Sus scrofa) is a bit to the south (Kankyo-cho 1982: 44-48). Many other smaller mammals can also be seen in the fields and woods, and water fowl are common on the lake (Tataka 1978). Fishing is good both in the lake and just off the Pacific coast.

Lake Ogawara is the most outstanding feature of the drainage (Misawa-shi 1971; Tataka 1978). This lake is Japan's 11th largest. It is roughly 7 by 14 km, has an area of about 65 km2 and is 26 m at its deepest point. There is a rather wide, sandy shelf, or submerged terrace, along most of the shore, where the water is less than 5 m deep. The surface of the lake is just 54 cm above the level of the sea, and only a very short segment of the Takase River separates the lake from the Pacific Ocean. Hence the salinity of the water is high, especially in the northern end where tides sometimes back up into it. The southwestern end of the lake is freshwater, as is the adjoining small Anenuma lake, because of the inflow of water from the streams. This causes the fauna and flora in and near the lake to vary considerably over short distances. In prehistoric times this lake was an arm of the sea.

Topographic Zones in the Takase Drainage:

The topography of the Takase River drainage has five major sectors: coastal plain and beach, marshlands, dissected Pleistocene terraces, hills, and mountains. However, it is useful to divide these further into 22 topographic zones in order to get a clearer picture of prehistoric land use, or settlement patterns. The narrow Pacific beach zone has (1) rolling dunes in the north, (2) rather level, straight beach in the center, and (3) dissected beach and terrace in the south. Just behind the beach is a coastal plain, (4) flat to the north but (5) gently rolling to the south. The western edge of the coastal plain is (6) rolling dunes on a high bluff that drops 25 to 35 m to the shore of Lake Ogawara. There are large marshlands on (7) the northeastern and (8) southwestern ends of the lake. These marshlands are only a few meters above sea level in most places; 10 m elevation is thought to mark the maximum transgression of the Holocene sea. (9) The extant lakes, ponds and swamps and the narrow marshlands around them can be counted together as a micro-zone. (10) Deep ravines with marshy bottomlands are another type of marshland micro-zone. Much of the drainage area is dissected tablelands: (11) north bank, (12) west bank and (13) south bank of the lake; (14) the large west plain and (15) the small tableland between it and the lake; and (16) several small headlands that extend into the southwestern marshland from the west plain. In the northwestern sector of the drainage the land is variously (17) deeply rolling, (18) gently rolling, or (19) extensively dissected. The mountain zone consists of (20) mountains that reach elevations of 800 to 1,000 m, (21) high hills, and (22) a fringe of low hills (the Highest Terrace).

Prehistory of the Takase Drainage:

There are 218 known archaeological sites in the Takase River drainage. About one-third of them are reported to have multiple occupations, with a total of 299 separate occupations that are identified clearly to period of habitation. There are 2 Palaeolithic occupations, 0 Incipient Jomon, 20 Earliest Jomon, 48 Early Jomon, 40 Middle Jomon, 77 Late Jomon, 31 Latest Jomon, 3 Yayoi, 0 Kofun, and 78 early Historic and Medieval occupations. The relative proportions of sites in each period do not parallel general trends in either Aomori Prefecture or in most other regions of Japan (Koyama 1978: 8). There are many possible explanations for the incongruities, ranging from errors in the data to very local differences in modes of adaptation. The highly unusual ecology and ecological history of the Takase River drainage might prove to be the best explanation for the patterns of fluctuations in site quantities in that locality. Explanations, however, will mostly have to wait until the data base can be greatly improved and reanalyzed considering more factors than were used here.

Palaeolithic: Only two sites are known from the Pleistocene. They constitute a meager 0.7% of the total occupations in the drainage. Dating them is difficult: one probably dates to 13,000-16,000 B.P., but the other might be as young as 10,000 years.

The oldest is Sabishiro, represented by two knife-shaped stone tools rumored to have been found on the present Pacific beach (Misawa City Museum display, 1985). These tools are of types characteristic of Phase IIb of the Japanese Palaeolithic, the basis for estimating their age at 13,000 to 16,000 years (Oda and Keally 1979).

The other site is Chojakubo and it is known mostly through surface collection; only one small excavation has been conducted there (Hirao, Tsunojika and Sato 1960; Tsunojika and Watanabe 1980: 7-13). Partially ground ax/adzes, bifacial spear points, scrapers, gravers, and flakes are prominant in this collection. Most of these artifacts appear to belong to the final phase of the Japanese Palaeolithic. This phase dates about 10,000-12,000 B.P. (Oda and Keally 1979). Chojakubo is in the northwest hill zone, but the reports on it leave doubt about the precise location with regard to the bluff (top or bottom) where the artifacts have been found.

Palaeolithic sites are scarce all over Aomori prefecture at the moment, and none of them seems to predate 16,000 B.P. (Oda and Keally 1979). Consequently, their rarity in the Takase drainage is not surprising. Perhaps the local Pleistocene environment was not conducive to much human settlement. At 20,000 B.P. this area was sub-alpine or boreal forest consisting mostly of spruce (Picea), fir (Abies), and hemlock (Tsugu) (Yasuda 1978: 158, 161), or grassland, and the temperate forest did not come in until after the end of the Pleistocene (Yasuda 1978: 158, 161; Kamei et al. 1981: 191, 194, 200; Machida 1982: 28-29). Spruce (Picea), fir (Abies), and pine (Pinus) still dominated the forest at 11,000 B.P. (Yasuda 1979: 162-163). This forest was replaced by one dominated by birch (Betula) by 9000 B.P. (Yasuda 1978:177-179). The mixed temperate forest with lots of oak and beech came in sometime after that. Another consideration is the inundation and alluviation of the lower river valleys in the early Holocene. The lowlands in Aomori possibly were the most hospitable localities in the late Pleistocene, and they are all under Holocene alluvium now; many Palaeolithic sites might be buried too deeply for us to find. Moreover, Palaeolithic sites in the Takase drainage are in topographic zones where occupations of other periods are relatively few, meaning that they are absent in those zones where they are most expected, judging from the patterns of later periods. This difference most likely reflects a considerably different adaptation during the final phases of the Palaeolithic.

Incipient Jomon: There are no sites in the Takase drainage that can be assigned to the time between 10,000 and 8,500 years ago. In northern Tohoku (Aomori, Akita and Iwate prefectures) only two sites report pottery that might predate 8500 B.P. One of these sites is Kamotaira in Hachinohe City, in the drainage area just south of the Takase. It yielded quite a few sherds of nail-impressed pottery (Suzuki 1982: 54-55; Aomori Museum of Local History display, 1982), which can be estimated at about 10,000 B.P. (Keally and Muto 1982). The other site is Odai Yamamoto 1 on the western side of Aomori prefecture (Miyake and Iwamoto 1979). This site is classified as terminal Palaeolithic, but 32 tiny potsherds were found there. One of them has what looks like linear-relief design on it. The rest of the sherds are plain and two are bases, both flat. Odai Yamamoto 1 is typologically roughly contemporary with Chojakubo, at the end of the Palaeolithic. The linear-relief sherd at Odai Yamamoto 1 suggests an age of 12,000 years for these two sites, although a single radiocarbon-based obsidian hydration age measurement gave 8,600 years for this pottery. The reason for the hiatus during Incipient Jomon is not clear, but this period corresponds to the early post-glacial birch forest in this area (Yasuda 1978: 177-179).

Earliest Jomon: Twenty Earliest Jomon occupations have been recorded in the Takase drainage, 6.7% of the total occupations and 9.3% of the Jomon. This is half the number of the next lowest Jomon period, Middle Jomon. The age of these occupations can be estimated at 6,000 or 6,300 years to 8,500 years (Keally and Muto 1982). The standard pottery sequence for this period of roughly 2,500 years has seven phases in northern Tohoku, all of them represented in the sites along the Takase River. Potsherds of the four oldest styles (7700-8500 B.P.) turn up in sites without shellmiddens, or in strata below the shell layers in excavated midden sites. At the Noguchi shellmidden, Mushiri I potsherds (estimated at 7200-7700 B.P.) seem to come from the oldest shell layer (Okamoto and Kato 1963). But at the Waseda shellmidden there is no question about the association of the Mushiri I pottery and the shells (Sato, Tsunojika and Nihonyanagi 1956).

The Earliest Jomon settlers in the Takase drainage were responding to radically changing ecological conditions there. When the first of these people camped at sites along the lower course of the river, the climate had warmed to modern temperatures, a cool temperate broadleaf forest was replacing the conifer and birch forests (Yasuda 1978), and perhaps the sea was beginning to invade the river valley. The accumulation of shellmiddens at some sites after 7700 B.P. certainly means that the Palaeo-Ogawara Bay was already developed enough to have tidal flats where the shallow water of the lake is today. These middens also show that the people quickly learned how to exploit this rich, abundant and readily available source of protein.

Seventy percent of the Earliest Jomon sites are located along the north, east, and south shores of the modern lake. They are relatively (and absolutely) most common on the eastern shore, and relatively common also on the northern shore. A considerable absolute number of the Earliest Jomon sites are located on the western plain. Another relatively well used habitat at this time was the northwest hilly zones, quite away from the bay, and the only point of similarity any of the Jomon periods has with the Palaeolithic pattern.

A majority of the sites are on bluffs, usually next to shallow erosion notches. These bluffs are 5 to 40 m high, but the sites tend to concentrate on bluffs about 10 m high and 90% are on bluffs 10-30 m high. As a rule, wide marshlands exist today below the Earliest Jomon sites. Relatively very few of the sites are found along narrow ravines, which are common on the dissected terraces and tablelands.

Nearly half of the Earliest Jomon sites are shellmiddens. This is by far the highest proportion of shellmiddens in any of the five Jomon periods, and these constitute one-quarter of all Jomon shellmiddens. If our data even roughly represent prehistoric reality, the obvious conclusion is that the products of the bay were unusually important in the lives of the late Earliest Jomon inhabitants of the Takase drainage.

A preponderance of the molluscan shells found in the middens is hamaguri venus clams (Meretrix lusoria). This species inhabits the shallow intertidal zone in bays with sandy bottoms, where the salinity is not too high (Okada 1965a: 267; Kira 1978: 140), exactly the conditions of the bay margin near the Earliest Jomon sites. The next most common molluscan species found in the sites are the shiofuki surf clam (Mactra veneriformis) and oonogai soft-shelled clam (Mya (Arenomya) arenaria oonogai), which prefer intertidal-zone habitats in bays. Many other species are found in small numbers: asari venus clams, also called Japanese littlenecks (Tapes (Amygdala) philippinarum), and oysters (Crassostrea gigas) from shallow, intertidal, sandy bottom habitats in bays; ark shells (Anadara (Scapharca) broughtonii) and rock or dye shells (Rapana thomasiana), from deeper water in bays; and horn shells (Batillaria cumingii and B. multiformis), kagamigai venus clams (Dosinia (Phacosoma) japoinica), pearly top shells (Umbonium (Suchium) costatum), giant Ezo scallops (Patinopecten (Muzuhopecten) yessoensis), ubagai surf clams (Spisula (Pseudocardium) sachalinenis), and tellins (Peronidia venulosa), most of which come from deeper water near the mouth of the bay or from the Pacific coast about 5 km from the sites they are found in. These rare species are conspicuously absent in the Tengumori shellmidden, which is over 5 km from the coast (Misawa-shi 1979). But the okishijimi venus clams (Cyclina sinensis), reported only at Tengumori, derive from brackish water habitats at the heads of bays or estuaries, where freshwater and salt water mix.3 Tengumori is on a high bluff above a small cove on Lake Anenuma, near the head of the Palaeo-Ogawara Bay. It is the only one of the shellmiddens studied that is near such a habitat.

The fish species found in the various middens show use of a wider range of environments. The most common species is the Japanese sea bass, or sea perch (Lateolabrax japonicus). This is a coastal species; it winters in deeper water in the ocean but is in inner bays in summer (Okada 1965b: 278; Kamohara 1961: 30; Akazawa 1980: 334-335). The young sometimes enter the freshwater zone near the mouths of rivers and streams. Many of the other species of fish come from coastal and open ocean habitats, as well as from the bay: rockfish (Sebastes sp.), Japanese flounder (Paralichthys olivaceus), grunts (Parapristipoma trilineatum), black mullet (Mugil cephalus), red sea bream or porgy (Pagrus major), puffers (Takifugu pardalis), Japanese tilefish (Branchiostegus japonicus), and possibly sardines (Sardinops melanostictus) and bluefin tuna (Thunnus thynnus).

The tuna4 almost certainly indicates these people had the ability to go out into the deep ocean to fish, probably in dugout canoes. The oldest such canoe so far reported in Japan is Early Jomon (Watanabe 1973: 218), only a few hundred years after the Takase Earliest Jomon shellmiddens. Fishhooks, including composite types, are reported in several sites on Lake Ogawara and nearby in eastern Aomori prefecture (Watanabe 1973: 114-115 and attached charts). Simple toggle harpoons appear in sites in the vicinity in Early Jomon, and in Hokkaido just to the north in Earliest Jomon (Watanabe 1973: 144). Moreover, bipolarly notched pebbles thought to be net weights are reported in the oldest Earliest Jomon sites in the Takase drainage, about 8,000 to 8,500 years ago. Although this evidence is sparse and mostly indirect, it hints at the likelihood that the Earliest Jomon peoples in the Takase drainage were participating in the regionŐs development of fairly advanced fishing skills and toolkits.

On land the people hunted deer (Cervus nippon) regularly, and also wild boar (Sus scrofa), the latter no longer native to the area (Geological Survey 1977: 20; Kankyo-cho 1982: 45). They took rabbits (Lepus brachyurus), racoon dogs (Nyctereutes procyonoides), serow (Capricornis crispus), and black bear (Selenarctos thibetanus japonica), too. The dog (Canis familiaris) was probably their companion. The Noguchi shellmidden had some bones of whale (Cetacea) and sea lions (Zalophus californianus) in the oldest shell layer, about 7200-7700 B.P. (Okamoto and Kato 1963). These were probably scavenged or killed on the Pacific beach about 5 km from the site. Large numbers of bird bones are reported at several of the sites. Some of these are identified as those of pheasants (Phasianus soemmerringii and P. colchicus), which are common in the area even today. Many of the others might be from migratory water birds, some of which still are seen on the lake throughout the year.

There is no evidence of plant use, at least none that has been reported. However, Watanabe (1975: 16-19) provides us some indirect evidence: 15 of the 16 sites he lists in Aomori prefecture report walnuts, including a late Earliest Jomon site near the Takase drainage. We can surmise that the people in the Takase drainage at this time also were eating walnuts. The stone grinders reported at some of the sites also could indicate use of plants.

The frequencies of animal species by habitat indicate most food was obtained very close to the settlements. The hamaguri, shiofuki and oonogai clams probably thrived in the tidal flats just below the sites and would have been a ready resource from spring to autumn. The sea bass could be caught with hooks and lines or nets from simple canoes just off the shore in the bay in summer or near the mouth of the bay in winter. Sometimes the people traveled the 5-7 km to the Pacific beach for molluscs, fish and sea mammals not available in the bay. Most of these probably provided food, but some, like the ark shells (Anadara), might have served other purposes. In fact, ark shells were being collected and used for decorating pots many centuries before the first shellmiddens formed. At times the people also undertook dangerous expeditions to the open sea after tuna. Most of the hunting could have been done within a 5-km radius of the camps, although occasional forays to the mountains 25 km away might be indicated by the bear and serow. But especially in winter when hunting would have been the main food acquision activity, these animals were probably concentrated in the sheltered ravines near the bay, not far away. Walnuts, an eminently storable plant available in mid autumn, and many other types of edible plants, certainly could have been collected at different seasons in the immediate vicinity of the camps.

There was hardly a dense human population in the Takase drainage in Earliest Jomon, with only 20 sites in a period of 2,200 to 2,500 years. And the small size of the sites suggests they were used for only short periods of time. Nevertheless, given the low population desity and the predominant food resources, the Earliest Jomon peoples here could have been quite sedentary, seasonally exploiting different plants and animals as they became available, rarely moving beyond 5 km from a fixed settlement.

This pattern of food resource exploitation is seen in Jomon sites elsewhere in Japan (Akazawa 1980: 338-342; Akazawa and Komiya 1981). Akazawa speculates that the gathering of molluscs was the work of the women, children and old people, hence the species are almost exclusively from very near the sites. Fishing is men's work and consequently shows use of a larger territory, up to 5-10 km in radius. Still most of the fish could have been caught less than 5 km from the site. The land mammals are hard to assign distances from the sites but invariably the main ones are generally thought to inhabit niches not too far from the settlement yielding them. Hunting is most likely the responsibility of the men.

One final point of indirect evidence concerning the subsistence adaptation of the Earliest Jomon people in the Takase drainage needs to be considered. The rivers of Tohoku and southern Hokkaido are salmon spawning grounds (Kamohara 1961: 8; Okada 1965b: 180). Salmon bones are nearly nonexistent in Jomon sites, but 4 of the 9 sites listed by Watanabe (1973: 210) are in Aomori prefecture, and one of them is a late Earliest Jomon site in Hachinohe, literally next door to the Takase drainage. It is very likely that the tools and other evidence of salmon fishing were left at the water's edge where they were used (MacDonald, personal communication, 1983), accounting for the paucity of evidence and making even a rare find evidence of considerable exploitation of this food resource. The discovery of thousands of dog salmon (Oncorhynchus keta) teeth at the terminal Palaeolithic or early Incipient Jomon site at Maeda Kochi in Tokyo (Kato 1985: 84-85; Akigawa-shi 1983: 252-255) greatly strengthens the inference that the Earliest Jomon people in the Takase drainage took advantage of this food resource. Salmon runs occur from September to the following January (Kamohara 1961: 8; Okada 1965b: 180), and salmon fishing would have filled the seasonal gap from mid autumn to early winter.

Early Jomon: The number of sites in the Takase drainage more than doubles to 48 in Early Jomon, 16.1% of the total sites and 22.2% of the Jomon, the second highest number of Jomon sites after Late Jomon. Sites of this period also are common all over Aomori prefecture. An enhanced archaeological visibility is part of the cause of this increase, but absolute numbers definitely do get larger. Their approximate age is 4,600 to 6,000 or 6,300 years, according to radiocarbon measurements (Keally and Muto 1982).

Early Jomon sites in the Takase drainage follow much the same distribution pattern as Earliest Jomon sites, with a tendency to be a bit more inland. The largest absolute number is on the south bank, but Early Jomon sites are also relatively and absolutely numerous on the east shore of the lake. Compared to other Jomon periods, they are relatively common on the west bank, and in the coastal dunes near the mouth of the bay, too, and relatively low in number in the mountains.

Like Earliest Jomon sites, the Early Jomon ones are on bluffs, usually near small erosion notches. But site landforms are more varied than before. Again, like Earliest Jomon, the sites tend to be on bluffs about 10 m high; 64.6% are on 10-25 m high bluffs, and the range is 0 to 50 m. Lowlands below the sites are usually lake margin or marshland, typical of shellmidden sites and of Earliest Jomon sites, and they are relatively rarely narrow valleys, also the same as Earliest Jomon.

Shellmiddens are found at 22.9% of the Early Jomon sites, a much smaller proportion than the roughly 50% in Earliest Jomon. But Early Jomon shellmiddens are absolutely more numerous than those of Earliest Jomon. They are also located in a greater variety of topographic environments, including the west bank and the coastal dunes near the mouth of the bay. In fact, these Early Jomon shellmiddens in the coastal dunes are the only middens found there, probably indicating that the tidal flat habitats of the preferred molluscan species were only briefly present there, that is, that the topography of the entrance to the bay changed rapidly in the early Holocene (Misawa-shi 1971: 4).

At some of the Early Jomon middens, the accumulaton of shells was more or less continuous from the end of Earliest Jomon. Others belong to younger phases of the period. Fauna remains seem to be about the same as those from Earliest Jomon layers, as far as the limited evidence allows us to guess. The only apparent difference is the sudden increase in asari venus clams (Tapes (Amygdala) philippinarum) and the corresponding decrease in hamaguri venus clams (Meretrix lusoria). Since these two species are found in much the same intertidal sandy bottom habitats, no important change in diet is implied. But over exploitation of the hamaguri venus clam in Earliest Jomon might be. The fact that these venus clams are prevalent in sites near the mouth of the bay as well as along the inner shores suggests that they were the preferred species; both are delicious.

Early Jomon probably coincides with the maximum expansion of the Palaeo-Ogawara Bay and with the advance of the warm temperate forest northward almost to the southern edge of the drainage (Machida 1982: 29). These factors could account for the increase in the number of sites and for the minor changes in settlement pattern from Earliest to Early Jomon.

Summarizing Remarks Focused on Earliest-Early Jomon:

Accumulation of shells at many of the Earliest Jomon middens continued without interruption into the early part of Early Jomon. But the number of Early Jomon sites more than doubles, and their locations shift inland toward the western plain. However, shellmidden sites drop from nearly half of the sites to only 22.9% of the total. At the same time the most commonly collected clam was no longer the larger hamaguri venus clam but instead became the smaller asari venus clam. Since these two species are found in much the same intertidal sandy bottom habitat, the implication of the change is over use of the hamaguri during Earliest Jomon rather than a significant change in environment or subsistence strategy. The appearance of shellmiddens further inland during Early Jomon reflects the further expansion of the bay as the Climatic Optimum approached. On the other hand, the singularity of the Early Jomon middens near the mouth of the bay almost certainly indicates the encrouchment of coastal sand dunes that first produced the tidal flats habitat for the preferred molluscan species and then destoyed that habitat by closing off the mouth of the bay, turning it into the brackish lake that exists today. On the whole, the trends in settlement pattern following Earliest Jomon support the interpretations that have been made from the data directly related to that period.

The Earliest Jomon people in the Takase River drainage began to exploit the rich and stable protein resources of the emerging Palaeo-Ogawara Bay and tidal flats as rapidly as these resources became available. The major additions to their diet were clams and sea bass, and probably tuna and salmon. The use of these resources does not seem to be a local innovation; all of them have precedents elsewhere. For example, the oldest shell layer at the Natsushima shellmidden on Tokyo Bay yielded Pacific oysters, various species of clams and other shellfish, bluefin tuna, black mullets, red sea breams, and other fish (Sugihara & Serizawa 1957: 34-35). This layer is 9,000-9,500 years old (M-769, 770, 771). Salmon are reported at about 13,000 BP at the Maeda Kochi site in Tokyo (Kato 1985, Akigawa-shi 1983: 252-5, Keally & Miyazaki 1986) and at about 10,500 BP from Layer VI at the Ushki site on Kamchatka Peninsula in northeastern Siberia (Dikov 1979: 192, Vasilyevskiy, personal communication, 1987). Contrarily, the toggle harpoon and composite fishhook do seem to be major technological innovations of the Earliest Jomon peoples along the Pacific coast of northeastern Japan (Watanabe 1973: 112, 144). These implements were probably parts of a complex of technology and skills being developed for exploiting large marine animals such as tuna and sea mammals. This new strategy appears to have spread southward along the coast of northeastern Japan and northward around much of the North Pacific rim, where it formed part of the basis of the marine adaptations of many modern peoples (Vasilyevskiy, personal communication, 1987).

The Earliest Jomon people in the Takase River drainage also quickly adopted the easiest subsistence strategy possible, given their particular culture, population density, and environment. Initially, their lithic-based culture, their very low population density, and the existence of perennially nutritionally complete microenvironments along the eastern shore of the Palaeo-Ogawara Bay allowed them to settle there in small, relatively permanent camps, exploiting mostly resources within a few kilometers of their living sites, only occasionally going as far as 5-7 km for species from the open ocean. As their technology improved, their population grew, and the environment changed, they adjusted that strategy to maintain the easiest way of life.

Middle Jomon: Sites of Middle Jomon age are less common than Early Jomon sites; there are 40 of them, 13.4% of the total sites and 18.5% of the Jomon sites. Middle Jomon in Tohoku is estimated at 4000-4600 B.P. (Keally and Muto 1982).

Despite the drop in numbers, Middle Jomon sites are absolutely and relatively more concentrated on the south bank than are sites of any other period. They are also relatively common on the north bank, and they are the oldest sites reported in the mountains. Middle Jomon sites are relatively few on the west bank, and they have the smallest relative number of any period on the west plain. Veiwed as a whole, however, Middle Jomon people occupied the greatest variety of environments.

The landforms of the site locations are also quite varied. The heights of the sites above the nearest lowlands show a more diffuse concentration between 10 and 25 m, ranging from 0 to 50 m. Further, a greater variety of lowlands is found below the sites. Notably, the lake margin is relatively quite rare near Middle Jomon sites. The overall impression is that these people preferred environments further from the coast and large bodies of water than their predecessors.

The great drop in shellmiddens in this period coincides well with the changes in settlement pattern. There are only 5 known middens with Middle Jomon layers; a small 12.5% of all Middle Jomon sites. Two of the middens are on the south bank and two are on the west bank, a distinct shift to the brackish water marshes at the head of the bay. Only one of these middens has been excavated and reported. This is the Furuyashiki site which sits on a hill above a ravine at the extreme inner reaches of the Palaeo-Ogawara Bay. Nevertheless, asari venus clams (a clam found in shallow water in bays) are the predominant shells in the midden. The brackish water yamato shijimi marsh clams are common and more expectable; they probably come from just below the site. Hamaguri venus clams, shofuki surf clams, and oysters are also fairly common in the midden. A large number of other molluscan species are present in miniscule quantities.

The fish were not quantified in the report but are said to be scarce compared to land mammals. The species come from a wide range of habitats, from inner bay to the deep sea: eels, Japanese flounder, Japanese sea bass or perch, red sea bream or porgy, puffers, herring, scorpian fish, and sharks. The nearest marine coast, it should be noted, is 11 km overland from Furuyashiki. The land mammals are mostly deer and wild boar. But a few teeth or bones of wolves, dogs, black bears, northern fur seals and Japanese otter have been identified. Pheasants and several kinds of large and medium-sized migratory water birds are also reported. The nearest shores of present lakes Anenuma and Ogawara are 2.5 and 3 km away.

The changes in settlement pattern, the decrease in number of sites, the relatively even greater decrease in number of shellmiddens, and the make-up of the faunal assemblage all indicate shrinkage of the tidal flats around the margin of the bay, reducing a primary source of animal protein. There presently are no geological data to corroborate this. Further speculation about Middle Jomon adaptation is difficult.

Late Jomon: There is an extreme increase in sites in Late Jomon, about 3250-4000 B.P. (Keally and Muto 1982). They account for one-quarter of all sites in the Takase drainage, and 35.6% of the Jomon sites. Their relative numbers show a definite shift to the west, especially to the west plain and the mountains. Late Jomon sites are relatively more common than others on straight bluffs, a consequence of location on the west plain. Bluff heights tend to be lower, centering around 10-20 m, and the lowlands below the sites are frequently narrow ravines, one probable cause of the lowered heights of the bluffs.

Shell middens are relatively very rare in Late Jomon. Two are located on the north bank and two on the south bank of the lake. This undoubtedly reflects the post-maximum contraction of the bay and the need to readapt to a much changed set of ecological conditions. But if the molluscs from the bay ceased to be of great importance to the people, the source of sustencance for this largest Jomon population in the Takase drainage is a mystery.

Latest Jomon: Population drops sharply from Late to Latest Jomon, evidenced by a 60% decrease in the number of sites. There are only 31 Latest Jomon sites, 10.4% of the total and 14.4% of the Jomon sites. Radiocarbon gives 2250-3250 B.P. as the approximate dates for this period (Keally and Muto 1982).

Site distribution is more varied in this period than in earlier periods, with no one zone seeming to be the favorite. Latest Jomon sites are relatively more common on the coast near the mouth of the bay and in the mountains and foothills than are other Jomon sites. They also show a distincly lower use of the south bank, which until Late Jomon had been a zone of estuaries at the head of the Palaeo-Ogawara Bay. This changed settlement pattern argues for a shift in subsistence strategy. The scarcity of shell middens (there are only 3, and one or two of these might be misidentified) further attests a shift.

Very likely land mammals gained importance as a source of animal protein, over fish and molluscs. This probably coincides with the continued shrinkage of the bay into a small freshwater lake, greatly reducing the food resources in that part of the drainage. This reduction would have affected the settlement pattern causing a relative shift toward the mountains. However, the percipitous drop in the numbers of sites in Latest Jomon is a much a mystery as their sudden increase in the preceding period.

Yayoi & Kofun: The only thing certain about the Yayoi (about 1st to 3rd centuries A.D.) presence in the Takase drainage is that it left almost no trace. Only a few sherds have turned up at two, or perhaps three, sites so far. No Kofun sites (4th to 7th centuries A.D.) are reported as such in the published register (Aomori-ken 1978). I suspect, however, that some of the sites indentified as "haji" might prove to be at least Late Kofun in age, even though the registers list them as Historic. Haji is a pottery type distinctive of the Kofun and early Historic periods.

Historic Periods: Seventy-eight sites are reported as Historic, a very large number of the total sites in the Takase drainage. They are located in a wide range of environments: on the Pacific coast, all around the lake, and in the northwest hills and the mountains. But Historic sites are absolutely and relatively most abundant on the dissected west plain. Sites identified as villages concentrate (60%) on the east and south sides of the lake, and are relatively uncommon on the west plain. Contrarily, manor, castle and tomb sites are overwhelmingly abundant (70-90%) on the west plain. Probably a large proportion of the village sites are early Historic (8th to 11th century), whereas the manors, castles and tombs for the most part are surely Medieval (12th to 16th century). Some were listed as such in the register (Aomori-ken 1978), but I have not yet been able to check the details individually for those sites not specifically recorded in the general sources.

The Historic settlement pattern shows a sharp constrast to that of the Jomon period, concentrating so heavily on the west plain. Even until after World War II the largest settlements in the drainage were there, a continuation of the settlement patterns established in the Medieval period.

Concluding Remarks:

On a map of the world, Japan is little more than a remote speck far from the dynamic cultural centers of Europe and Asia. Consequently, foreign and Japanese archaeologists alike often tend to treat Japan's prehistoric and historic cultures as if they lacked any significant originality or regional variability. "The Jomon culture of Japan" and "waves of migration (influence) from the continent" are frequently heard and reflect widely held misconceptions about the homogeneity and lack of creativity of Japan's prehistoric (and present) cultures. As the preceding discussion makes clear, these statements hardly apply to the prehistory of the Takase River drainage, itself a very small and remote part of Japan; they seriously misrepresent the extreme diversity and creativity of local Palaeolithic, Jomon and Historic cultures and their highly effective adaptations to a wide range of natural environments and econiches.


  1. These are far too many to list in the text and only a partial listing is given here:
    Aomori-ken 1973;
    Hirao, Tsunojika and Sato 1960;
    Kamikita-machi 1983;
    Kunkel 1985;
    MacCord 1955a, 1955b;
    Misawa-shi 1979;
    Misawa-shi 1985a, 1985b, 1985c;
    Nagao 1983;
    Narita 1984;
    Okamoto and Kato 1963;
    Sato, Tsunojika and Nihonyanagi 1956;
    Tanaka 1983.
  2. An occupation, as used here, includes all materials assignable to the same subperiod, for example, all materials at one site that belong to Early Jomon. In future analyses, as the data base improves, the definition will be refined to include all materials from the same pottery style phase, a much narrower band of time.
  3. There are three shell layers at Tengumori, but the published report does not specify which species came from which layer (Misawa-shi 1979). However, since Lake Anenuma was probably freshwater when the two youngest layers formed, the reported molluscs, fish, and mammals very likely came from the Earliest Jomon layer.
  4. See footnote 3, which holds for most species. The significance of the tuna as an indicator of advanced fishing tools and skills suggests we should have reservations about this species until its provenience can be confirmed.

Table of Site Quantities in the Takase River Drainage

Period Number
of Sites
of Total
of Jomon
Percent That Are

Palaeolithic 2 0.7
Incipient Jomon 0
Earliest Jomon 20 6.7 9.3 < 50%
Early Jomon 48 16.1 22.2 22.9
Middle Jomon 40 13.4 18.5 12.5
Late Jomon 77 25.8 35.6 rare
Latest Jomon 31 10.4 14.4 9.7
Yayoi 3 1.0
Kofun 0
Historic 78 26.1

299 100.2 100.0